Proceedings of the Galactic Center Workshop 2002: The Central 300 Parsecs of the Milky Way

A. Cotera, S. Markoff, T. R. Geballe, and H. Falcke (Eds.) Proceedings of the Galactic Center Workshop 2002 The Centr...

Author: Angela Cotera | Sera Markoff | Thomas R. Geballe | Heino Falcke

10 downloads 876 Views 16MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

A. Cotera, S. Markoff, T. R. Geballe, and H. Falcke (Eds.)

Proceedings of the

Galactic Center Workshop 2002 The Central 300 parsecs of the Milky Way

Astronomische Nachrichten, Supplementary Issue 112003

WILEYVCH WILEY-VCH Verlag CmbH & Co. KCaA

This Page Intentionally Left Blank

A. Cotera, S. Markoff, T. R. Geballe, and H. Falcke (Eds.)

Galactic Center Workshop 2002 The Central 300 parsecs of the Milky Way Astronomische Nachrichten, Supplementary Issue 112003

This Page Intentionally Left Blank

A. Cotera, S. Markoff, T. R. Geballe, and H. Falcke (Eds.)

Proceedings of the

Galactic Center Workshop 2002 The Central 300 parsecs of the Milky Way

Astronomische Nachrichten, Supplementary Issue 112003

WILEYVCH WILEY-VCH Verlag CmbH & Co. KCaA

Editors DI:Angela Cotera SETI Institute, Arizona State University [email protected]

This book was carefully produced. Nevertheless, editors, authors and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

DI:Sera Markoff Max Planck Institute, Center for Space Research smarkoff @ mpifr-bonn.mpg.de Pro$ DI:Heino Falcke Max Planck Institute, Center for Space Research [email protected] Thoinas R. Geballe Gemini Observatory [email protected]

Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library. Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at

0 2004 WILEY-VCH Verlag GmbH & Co.

KGaA, Weinheim All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form - nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Printed in the Federal Republic of Germany. Printed on acid-free paper

Composition Druckhaus >>ThomasMiintzercc GmbH, Bad Langensalza Printing Strauss Offsetdruck GmbH, Morlenbach Bookbinding Litges & Dopf Buchbinderei GmbH, Heppenheim

ISBN 3-527-40466-X

Preface

The Galactic Center (GC) is one of the most scientifically intriguing regions available for astrophysical research. First and foremost, it is by far the closest galactic nucleus available for study, observable at spatial resolutions unapproachable in other galaxies. At the distance of the GC (-8 kpc), spatial resolutions of %0.1”, now achievable with the latest generation of radio and near-infrared telescopes, corresponds to s 800 AU. The highest achievable resolution for the closest external galactic nucleus, M31, is a factor of 100 times lower. Thus the relative proximity of our Galactic Center, combined with the high angular resolutions now available, provides the opportunity to differentiate and study a wealth of objects; many of which appear to be highly unusual, possibly unique, but could in fact be merely the standard occupants of a normal galactic nucleus. At the very center of the Galaxy lurks the (currently) weakly luminous, supermassive black hole candidate Sgr A*. Surrounding this object is the densest stellar cluster in the Galaxy, much of it immersed in ionized interstellar gas. At much greater scales of up to 100 pc, clusters of bright stars and synchrotron-brightfilaments are prominent. All of these phenomena are surrounded by the galactic stellar bulge and dense giant molecular clouds extending out to kiloparsec scales. Proximity to the GaIactic Center does not, however, make observations of this region trivial. Due to large amounts of dust and gas along the line of sight to the GC - through the Galactic disk - the central region is inaccessible with optical astronomy, even using the most the sensitive and sophisticated techniques. Observational exploration of the GC had to await, and must rely on, the development of radio, sub-millimeter, infrared, and most recently, X-ray and y-ray astronomy. The first subarcsecond observations of the Galactic Center were made in the 1970’s with radio interferometers. In the 1980’s the VLA led the way in high angular resolution studies, while infrared observations began to reach the sub-arcsecond regime. Technological developments in infrared observing with both ground- and space-based telescopes, as well as an improved understanding of atmospheric seeing, have allowed significantly higher resolution observations of the GC in the near-infrared (NIR, 1-3 pm) and mid-infrared (MIR, 3-20 pm). More recently, remarkable improvements in resolution at far-infared (FIR, 20-200 pm), sub-millimeter, millimeter, and within the last few years, X-ray, wavelengths, have resulted in remarkable progress in addressing many long standing questions about the Galactic Center. For example, studies of stellar orbital motions have confirmed that the GC harbors what is by far the most secure candidate for a supermassive (-3 x lo6 M,) black hole: Sgr A*. In very recent years, the detection of weak X-ray emission undergoing daily flares has prompted numerous revisions to theoretical models of the accretion processes onto this object. The Galactic Center also contains a collection of some of the most luminous stars in the Galaxy, which are responsible for both compact and diffuse X-ray emission and the ioniza-

VI

Preface

Preface

VII

tion of large quantities of interstellar gas. The obvious youth of these stars raises questions about how close to the center they formed, and how they formed in an environment that currently appears to be quite hostile to starbirth. These luminous stars ionize the surfaces of regions of molecular gas, some of which appear to be interacting with both the hot gas and strong ambient magnetic fields. Thus, the Galactic Center is complex laboratory containing a multitude astrophysical phenomena. Currently the links between some of these are obvious, and between others are less so. The detailed investigation of the GC is fraught with exciting questions, whose answers have ramifications for the structure and evolution of our galaxy, the phenomena seen in more distant galactic nuclei, and even fundamental physics. The study of the GC brings together many disciplines of astronomy which are tycpically distinct, and many astronomers whose research does not ordinarily overlap. The Galactic Center Workshop 2002: The Central 300 Parsecs, follows a series of conferences which started in the 1970’s, dedicated specifically to the center of our Galaxy. From one and two day symposia at Caltech and the University of California at Berkeley (1982, 1984, and 1986), these meetings have continued to expand, with the fiist week-long conference held in 1988 at the University of California at Los Angeles. The meetings have also become more international, with the 1992 meeting in Ringsburg, Germany, the 1995 meeting in La Serena, Chile, as well as a symposium during the 1997 IAU General Assembly meeting in Kyoto, Japan. In 1998, The Galactic Center Workshop 1998: The Central Parsecs, held in Tucson, Arizona, was the first full meeting dedicated solely to the very heart of our Galaxy. The Galactic Center Workshop 2002: The Central 300 parsecs, held at the Ohana Keauhou Beach Resort in Kona, Hawai’i on 3-8 November, consisted of eight topical sessions. The format of each of these was 6-8 talks followed by 30-45 minutes of moderated discussion, which was recorded. These Proceedings are arranged in the same order as the talks, with the posters from each session following. The discussions were informative and often lively; a CD of the recorded discussions is also being made available as part of these Proceedings. This current workshop was extremely fortunate to have been sponsored by so many prestigious institutions, most of them Hawai’i observatories. The Gemini Observatory was the primary host of the meeting. Significant financial support was provided by Gemini, the Institute for Astronomy, University of Hawaii; the Subaru Telescope; the Joint Astronomy Centre; the W. M. Keck Observatory; the Sub-Millimeter Array; and the National Science Foundation. The Galactic Center Newsletter played a key role in disseminating information about the meeting to astronomers worldwide. As conference co-chairs, we are grateful to the more than 100 attendees for their enthusiastic response to this conference. We would like to especially thank the tireless work of the local organizing committee, without whom “GC02” would not have been such a success. Angela Cotera Sera Markoff Thomas Geballe Heino Falcke

VIII

Scientitic Organizing Committee Angela Cotera (co-chair), Anzona State University Andreas Eckart, University of Cologne Heino Falcke (co-chair), Max Planck Institute, Center for Space Research Tom Geballe (co-chair), Gemini Observatory Rolf Kudritzki, Institute for Astronomy, Honolulu Sera Markoff (co-chair), Max Planck Institute, Center for Space Research Mark Moms, University of California at Los Angeles Ramesh Narayan, Harvard University Torno Oka, University of Tokyo Mark Reid, Harvard University Francois Rigaut, Gemini Observatory Kris Sellgren, Ohio State University Farhad Yusef-Zadeh, Northwestern University Jun-Hui Zhao, Harvard-Smithsonian Center for Astrophysics

Local Organizing Committee Andy Adamson, Joint Astronomy Centre Tracy Beck, Gemini Observatory Tom Geballe (chair), Gemini Observatory John Hamilton, Gemini Observatory Janice Harvey, Gemini Observatory Kalena Quinones, Gemini Observatory Peter Michaud, Gemini Observatory Francois Rigaut, Gemini Observatory Antony Schinckel, Sub-Millimeter Array

Preface

Contents

Chapter 1: Large Scale Structures: Surveys and Interactions High-resolution Hi Absorption Observations towards the Central 200 pc of the Galaxy Cornelia C. Lang, Claudia Cyganowski, W. M. Goss, Jun-Hui Zhao

.......................................................

High Resolution, High Sensitivity Imaging of the Galactic Center at 330 MHz Michael E. Nord, Crystal L. Brogan, Scott D. Hyman, T. Joseph W. Lazio, Namir E. Kassim, T.N. LaRosa, K. Anantharamaiah, Neboja Duric

..

Spatially Resolved Very Large Array 74 MHz Observations Toward the Galactic 17 Center C. L. Brogan, M. Nord, N. Kassim, J. Lazio, K. Anantharamaiah

.......................................................

Chandra view of the central 300 pc of our Galaxy

.......................

25

Q. Daniel Wang

..............

Two Thousand X-ray Stars in the Central 20 pc of the Galaxy 33 M. P. Muno, F. K. Baganoff, M. W. Bautz, W. N. Brandt, P. S. Broos, E. D. Feigelson, G. P. Garmire, M. R. Morris, G. R. Ricker, L. K. Townsley Magnetic field in the Galactic Centre: Rotation Measure observations of extra41 galactic sources Subhashis Roy, A. Pramesh Rao, Ravi Subrahmanyan

.................................................

........................

47

......................

53

Study of the Nuclear Bulge region of the Galaxy K. S. Baliyan, S. Ganesh, U. C. Joshi, I. S. Glass, T. Nagata

A morphological Study of the Galactic Inner Bulge Kiran S. Baliyan, Shashikiran Ganesh, Umesh C. Joshi, Ian S. Glass, Mark R. Morris, Alain Omont, Mathias Schultheis, Guy Simon Warm molecular gas, dust and ionized gas in the 500 central pc of the Galaxy N. J. Rodriguez-FernBndez, J. Martin-Pintado, A. Fuente, T. L. Wilson

..

.................

Prospects for LOFAR Observations of the Galactic Center N. E. Kassim, T. J. W. Lazio, M. Nord, S. D. Hyman, C. L. Brogan, T. N. LaRosa, N. Duric

59

65

X

Contents

43 GHz SiO masers in late-type stars with 86 GHz SiO masers and astrometry 73 with VERA in the Galactic center LorAnt 0. Sjouwerman, Maria Messineo, Harm J. Habing

................................... ....................

A Search for Radio Transients at 0.33 GHz in the GC Scott D. Hyman, T. Joseph W. Lazio, Namir E. Kassim, Michael E. Nord, Jennifer L. Neureuther

79

Chapter 2: Molecular Clouds and Magnetic Fields

..................

A Molecular Face-on View of the Galactic Center Region Tsuyoshi Sawada, Tetsuo Hasegawa, Toshihiro Handa, R. J. Cohen

....................................

The Inner 200pc: Hot Dense Gas Christopher L. Martin, Wilfred M. Walsh, Kecheng Xiao, Adair P. Lane, Christopher K. Walker, Antony A. Stark Gravitational Stability of Molecular Clouds in the Galactic Center Tomoharu Oka, Tetsuo Hasegawa

85

93

........... 101

Spectroscopy of Hydrocarbon Grains toward the Galactic Center and Quintuplet Cluster 109 J. E. Chiar, A. J. Adamson, D. C. B. Whittet, Y .J. Pendleton

.......................................................

X-rays from the HII Regions and Molecular Clouds near the Galactic Center Katsuji Koyama, Hiroshi Murakami, Shinichiro Takagi

. . 117

Reflected X-ray Emissions on Giant Molecular Clouds - Evidence of the Past Activities of Sgr A* 125 Hiroshi Murakami, Atsushi Senda, Yoshitomo Maeda, Katsuji Koyama

..............................................

Observation of Toroidal Magnetic Fields on 100 pc Scales in the Galactic Center 133 G. Novak, D. T. Chuss, J. L. Dotson, G. S. Griffin, R. F. Loewenstein, M. G . Newcomb, D. Pernic, J. B. Peterson, T. Renbarger Extended photoionization and photodissociation in Sgr B2 J. R. Goicoechea, N. J. Rodriguez-Fern6ndez, J. Cernicharo Propagation of charged particles from the Galactic Center W. Bednarek, M. Giller, M. Zielifiska

................. 139

................. 145

Discovery of New SNR Candidates in the Galactic Center Region with ASCA and Chandra 151 Atsushi Senda, Hiroshi Murakami, Katsuji Koyama

......................................................

XI

Contens

............. 157

Molecular Lineobservations of the Tornado Nebula and its Eye J. Lazendic, M. Burton, F. Yusef-Zadeh, M. Wardle, A. Green, J. Whiteoak

The Search for Water and Other Molecules in the Galactic Centre with the Odin Satellite 161 Aa. Sandqvist, P. Bergman, A. Hjalmarson, E. Falgarone, T. Liljestrom, M. Lindqvist, A. Winnberg, the Odin Team

......................................................

Chapter 3: Sagittarius A and its Environs

... 167

Deep X-Ray Imaging of the Central 20 Parsecs of the Galaxy with Chandra Mark Morris, Fred Baganoff, Michael Muno, Christian Howard, Yoshitomo Maeda, Eric Feigelson, Marshall Bautz, Niel Brandt, George Chartas, Gordon Garmire, Lisa Townsley

........... 173

Mapping Magnetic Fields in the Cold Dust at the Galactic Center David T. Chuss, Giles Novak, Jacqueline A. Davidson, Jessie L. Dotson, C. Darren Dowell, Roger H. Hildebrand, John E. Vaillancourt

The Galactic Center Nonthermal Filaments: Recent Observations and Theory T. N. LaRosa, Michael E. Nord, T. Joseph W. Lazio, Steven N. Shore, Namir E. Kassim

. . 181

Interaction between the Northeastern Boundary of Sgr A East and Giant Molecu189 lar Clouds: Excitation Mechanisms of the H2 Emission Sungho Lee, Soojong Pak, Christopher J. Davis, Robeson M. Hermstein, T. R. Geballe, Paul T. P. Ho, J. Craig Wheeler

...................

Sgr A East and its surroundings - a view with XMM-Newton Masaaki Sakano, Robert S. Warwick, Anne Decourchelle

.............. 197

......................

205

.....................

211

Chandra ACIS Imaging Spectroscopy of Sgr A East Y. Maeda, K. Itoh, F. K. Baganoff, M. W. Bautz, W. N. Brandt, D. N. Burrows, J. P. Doty, E. D. Feigelson, G. P. Garmire, M. Morris, M. P. Muno, S. Park, S. H. Ravdo, G. R. Ricker, L. K. Townsley A Census of Dust Absorption at the Galactic Centre Andy Adamson, Rachel Mason, Emily Macdonald, Gillian Wright, Jean Chiar, Yvonne Pendleton, Tom Kerr, Janet Bowey, Doug Whittet, Mark Rawlings

Thermal SiO observations of a shell attached to the nonthermal filaments in 217 SgrA Toshihiro Handa, Masaaki Sakano, Masato Tsuboi

........................................................

Absorption and Emission in the Four Ground-State OH Lines Observed at 18 cm 223 with the VLA Towards the Galactic Centre R. Karlsson, Aa. Sandqvist, L. 0. Sjouwerman, J. B. Whiteoak

............................

XU

Contents

Constraints on distances to Galactic Centre non-thermal filaments from HI 229 absorption Subhashis Roy

....................................................

........... 235

Discovery of a non-thermal X-ray filament in the Galactic Centre Masaaki Sakano, Robert S. Warwick, Anne Decourchelle High-negative velocities in the inner 25 pc of the Galactic center Lorint 0. Sjouwerman

............ 241

Chapter 4: Stars and Star Formation Really Cool Stars and the Star Formation History at the Galactic Center Robert D. Blum, Solange V. Rm’rez, Kristen Sellgren, Knut Olsen Massive Stars and The Creation of our Galactic Center Donald F. Figer

...... 247

...................255

.....................

The Galactic Center Source IRS 13E: a Star Cluster Jean-Pierre Maillard, Thibaut Paumard, Susan Stolovy, Franqois Rigaut X-ray Emission from Stellar Clusters Near the Galactic Center Casey Law, Farhad Yusef-Zadeh

............. 271

.......................

Simulated X-ray emission from the Arches cluster Pablo F. Velrizquez, Alejandro C. Raga, Jorge Cant6, Elena Masciadri, Luis F. Rodriguez

SiO Maser Sources within 30 pc of the Galactic Center Shuji Deguchi, Hiroshi Imai

263

279

................... 285 ...................293

86 GHz SiO masing late-type stars in the Inner Galaxy M. Messineo, H. J. Habing, L. 0. Sjouwennan, K. M. Menten, A. Omont CNO Abundances in the Quintuplet Cluster M Supergiant 5-7 S. V. Ramirez, K. Sellgren, R. Blum, D. M. Terndrup

............. 299

........................

303

........................

309

..........................

315

New results on the Galactic Center Helium stars Thibaut Paumard, Jean-Pierre Maillard, Susan Stolovy

Ten Thousand Stars Toward the Galactic Center Franqois Rigaut, Robert Blum, Tim Davidge, Angela Cotera

Stellar Orbits at the Center of the Milky Way N. Mouawad, A. Eckart, S. Pfalzner, J. Moultaka, C. Straubmeier, R. Spurzem, R. Schodel, T. Ott

XI11

Contens

Dynamical Friciton near the Galactic Center Sungsoo S. Kim, Donald F. Figer, Mark Moms

...........................

321

Near-infrared adaptive optics observations of the Galactic Center with NAOS/CONICA (ESO) and GriF (CFHT) 327 Y. Clknet, D. Rouan, F. Lacombe, E. Gendron, D. Gratadour

.............................

.................................

Radio Pulsars in the Galactic Center T. Joseph W. Lazio, James M. Cordes, Cornelia C. Lang, Eric V. Gotthelf, Q. Daniel Wang Review of low-mass X-ray binaries near the Galactic center A. Lutovinov, S. Grebenev, S. Molkov, R. Sunyaev

333

................ 337

Neutrons, neutrinos, and gamma-rays from the Galactic Center W. Bednarek

............. 343

Chapter 5: Sgr A* I: New Observational Results

...............................

349

.................................

355

Linear and Circular Polarization from G. C. Bower Intrinsic Radio Variability of Sgr A* Jun-Hui Zhao

Flares of Sagittarius A* at Short Millimeter Wavelengths Atsushi Miyazaki, Takahiro Tsutsumi, Masato Tsuboi

.................. 363

Limits on the Short Term Variability of Sagittarius A* in the Near-Infrared S. D. Hornstein, A.M. Ghez, A. Tanner, M. Morris, E. E. Becklin

.... 371

A New X-Ray Flare from the Galactic Nucleus Detected with XMM-Newton A. Goldwum, E. Brion, P. Goldoni, P. Fernando, F. Daigne, A. Decourchelle, R. S. Warwick, P. Predehl

..........................

Searching for Structural Variability in Sgr A* Zhi-Qiang Shen, M. C. Liang, K. Y. Lo, M. Miyoshi

Observations of the Galactic Centre at 610 MHz with the GMRT Subhashis Roy, A. F’ramesh Rao

... 377 383

........... 391

Closure Amplitude Analysis of 15, 22 and 43GHz VLBA Observations of Sagit397 tarius A*: Size is Consistent with the Scattering Law G. C. Bower

.....................

XIV

Contents

....................

VLBA observation of a radio intraday flare of Sgr A* Makoto Miyoshi, Hiroshi Imai, Junichi Nakashima, Shuji Deguchi, Zhi-Qiang Shen

403

A Chandra View of Diffuse X-Ray Emission in the Central 20 Parsecs of the Galaxy 407 Sangwook Park, Frederick K. Baganoff, Mark W. Bautz, Gordon P. Garmire, Yoshitomo Maeda, Mark Morris, Michael P. Muno

.......................................................

Investigating the submillimetre variability of Sagittarius A* with SCUBA Douglas Pierce-Price, Tim Jenness, John Richer, Jane Greaves

...... 413

............. 419

Near-Infrared Flux Limits for Sgr A* Based on NICMOS Data Susan Stolovy, Fulvio Melia, Donald McCarthy, Farhad Yusef-Zadeh

The wavelength dependence of Sgr A* size and the unified model of compact radio 425 sources Fedor Prigara

.......................................................

Search for Circular Polarization toward Sagittarius A* at 100 GHz M. Tsuboi, H. Miyahara, R. Nomura, T. Kasuga, A. Miyazalu

.......... 431

Chapter 6: Sgr A* 11: Theoretical Models Radiatively Inefficient Accretion Flow Models of Sgr A* Eliot Quataert

....................................

445

.....................................

453

.......................

459

Jet Models for Flaring in Sgr A* Sera Markoff, Heino Falcke A Jet-ADAF Model for Sgr A* F. Yuan, S. Markoff, H. Falcke

.................. 435

A model for polarised radio emission from Sgr A* T. Beckert

On the Chandra Detection of Diffuse X-Ray Emission from Sgr A* M. E. Pessah, F. Melia A Relativistic Disk in Sagittarius A* Siming Liu, Fulvio Melia

.......... 467

.................................

The frozen (inactive) disk in Sgr A*: freezing the accretion of the hot gas too? Sergei Nayakshin Gamma-ray emission from an ADAF around a Kerr black hole Kazutaka Oka, Tadahiro Manmoto

475

. . 483

............. 491

xv

Contens

Chapter 7: The Central Parsec

..........................................

497

.............................

505

The Discovery of Sgr A* W. M. Goss, Robert L. Brown, K. Y. Lo

The Position, Motion, and Mass of Sgr A* Mark J. Reid, Karl M. Menten, Reinhard Genzel, Thomas Ott, Rainer Schodel, Andreas Brunthaler Tidal processes very near the black hole in the Galactic Center Tal Alexander

............. 513

...................

New MIR Excess Sources north of the IRS 13 Complex 521 A. Eckart, J. Moultaka, T. Viehmann, C. Straubmeier, N. Mouawad, R. Genzel, T. Ott, R. Schodel Full Three Dimensional Orbits For Multiple Stars on Close Approaches to the 527 Central Supermassive Black Hole A. M. Ghez, E. Becklin, G. DuchCne, S. Hornstein, M. Moms, S. Salim, A. Tanner

...................................

The Galactic Center stellar cluster: The central arcsecond R. Schodel, R. Genzel, T. Ott, A. Eckart

................. 535

Stellar Dynamics in the Galactic Center: 1000 Stars in 100 Nights Thomas Ott, Reinhard Genzel, Andreas Eckart, Rainer Schodel A Bow Shock of Heated Dust Surrounding IRS 8 F. Rigaut, T. R. Geballe, J.-R. Roy, B. T. Draine

........... 543

........................

551

.....................

Monitoring Sagittarius A* in the MIR with the VLT 557 A. Eckart, J. Moultaka, T. Viehmann, C. Straubmeier, N. Mouawad, R. Genzel, T. Ott, R. Schodel, F. K. Baganoff, M. R. Morris The magnetic field in the central parsec A. C. H. Glasse, D. K. Aitken, P. F, Roche

..............................

563

Mid-Infrared Imaging and Spectroscopic Observations of the Galactic Center with 567 SubardCOMICS Y. Okada, T. Onaka, T. Miyata, H. Kataza, Y. K. Okamoto, S. Sako, M. Honda, T. Yamashita, T. Fujiyoshi

...............................................

Physical Conditions in the Central Parsec Modeled from Mid-Infrared Imaging 573 Photometry Dan Gezari, Eli Dwek, Frank Varosi

....................................................

LINCINIRVANA - The LBT Near-Infrared Interferometric Camera C. Straubmeier, A. Eckart, T. Bertram, T. Herbst

......... 577

XVI

Contents

Chapter 8: Morphology and Dynamics of the Central 10 Parsecs Hot Molecular Gas in the Central 10 Parsecs of the Galaxy R. M. Herrnstein, P. T. P. Ho

................ 583

.........................

The ISM and Stellar Distributions Near Sgr A* Nick Scoville, Susan R. Stolovy, Micol Christopher

Resolving The Northern Arm Sources at the Galactic Center Angelle M. Tanner, A. M. Ghez, M. Morris, E. E. Becklin

............... 597

Structural analysis of the Minispiral from high-resolution Br/ data Thibaut Paumard, Jean-Pierre Maillard, Mark Moms Gas physics and dynamics in the central 50 pc of the Galaxy B. Vollmer, W. J. Duschl, R. Zylka

591

.......... 605

...............613

The First Measurement of Radial Acceleration of Ionized Gas Near Sagittarius A* 621 Doug Roberts, Farhad Yusef-Zadeh

............. 629

Simple hydrodynamical Simulations of the Circumnuclear Disk Robert F. Coker, Michol H. Christopher, Susan R. Stolovy, Nick Z. Scoville

Astron. Nachr./AN 324. No. S I . 1 - 7 (2003) / DO1 I0.1002/asna.200385065

High-resolution HI Absorption Observations towards the Central 200 pc of the Galaxy Cornelia C. Lang*’, Claudia Cyganowski**2,W. M. G O S Sand ~ , Jun-Hui Zhao’ I Department of Physics & Astronomy. University of Iowa, Iowa City, IA 52242 * National Radio Astronomy Observatory, Socorro, NM 87801 Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Camhndge, MA 01238

’

Key words interstellar medium, Galactic Centet

-

Abstract. We report the first results of an HI absorption survey of the central 200 pc of the Galaxy. These Very Large Array (VLA) data have a resolution of 15” (0.6 pc at the Galactic Center (GC) distance) and a velocity resolution of -2.5 !an s-’ . This study provides HI data with high spatial resolution, comparable with the many high resolution observations which have been made of GC sources over the past ten years. Information on the velocities, relative di\tunces of Sources and H I column densities will be studied. Such data serves to clarify the nature of physical associations between the unique radio continuum features in this region and the atomic and molecular components. In particular, we are interested in the arrangement of sources in the active Radio Arc region: the atomic, ionized and molecular gas and their relation to the massive stellar clusters and magnetic filaments.

1 Introduction The bright and unusual radio continuum sources in the central few hundred parsecs of the Galaxy provide an opportunity to measure the 21 cm line of atomic hydrogen in absorption. Previous HI absorption studies have been crucial to our understanding of Galactic structure and rotation and the nature of atomic gas in the inner parts of the Galaxy. There are several well-known HI components in the direction of the GC (Cohen & Davies 1979): ( 1 ) the expanding 3 kpc arm appears at v=-53 km s-l and is thought to be -5.5 kpc from the Sun, (2) HI components near v=+l35 km s-l are thought to be beyond the G C by distances of a few hundred parsecs to 2 kpc, and (3) the “nuclear disk” and “molecular ring” components appearing at velocities of 160 to -200 km s-’ and 135 km spl are located within a few hundred parsecs of the GC. In the inner Galaxy, atomic gas is most often associated with regions of molecular gas where it serves to shield the molecular gas against photodissociation (Dickey & Lockman 1990). Therefore, HI absorption features not described above may possibly be identified with known G C molecular emission features using correlations in velocity structure. The recent Oka et al. (1998) CO survey made with the Nobeyama 45-m telescope provides the best spatial resolution, velocity, and spatial coverage of any survey of molecular gas within the central Galaxy. In addition, the multitude of “forbidden” (e.g. sign opposite to galactic rotation) velocity components in the GC region are thought to represent the response of the molecular gas in the GC to the Galaxy’s strong stellar bar (Binney et a]. 1991 ; Bally et a]. 1988). The CO survey data of Oka et al. (1998) illustrated that the molecular gas traced by C O emission in the central 200 pc is organized into filamentary and shell-like features. This morphology and kinetic structure indicates that violent kinetic activity (such as supernova explosions and stellar winds from Wolf-Rayet type ~

~

* Corresponding author: e-rnai1:[email protected] * * C.Cyganowski was a REU summer research student at NRAO @ 2003 WILEY~VCHVerlag GinbH & Co. KGaA. Weinhem

C. Lang et al.: HI towards the Galactic Center

2

stars) plays an important role in shaping the ISM. In addition to the Radio Arc region (where the Quintuplet and Arches clusters are located), the GC region is filled with sites where compact thermal radio and midinfrared sources have been observed (e.g. Sgr B, Sgr C; as well at many positions along the Galactic plane; LaRosa et al. 2000; Egan et al. 1998) and it is likely that massive stars are either forming or have formed in these regions. In addition, the spectrum of diffuse X-ray emission in this region suggests that the ISM is being strongly influenced by massive star-forming activities (Wang, Gotthelf & Lang 2002). In order (1) to understand the physical urrungement and interactions between the stellar and interstellar components, (2) to put constraints on distances to radio continuum features in the field of view, and (3) to make detailed images and estimates of the HI opacity toward well-studied Galactic center sources, we have carried out a complete HI absorption study of the central 200 pc of the Galaxy. This HI absorption survey represents the highest resolution and most complete study of H I absorption toward the GC, and will form the basis for the comprehensive study of the physical line-of-sight locations and interactions between interstellar features in the GC. The 15” spatial resolution and a 2.5 km s-’ velocity resolution are a vast improvement over the previous HI absorption study of Lasenby et al. ( 1 989) which included only a single pointing toward the Sgr A complex and had spatial and velocity resolutions of 50” and 10 krn S C ’ , respectively.

2 Observations & Imaging We have observed five overlapping pointings along the Galactic plane at 2 1 cm using the Very Large Array (VLA) of the National Radio Astronomy Observatory’. The VLA primary beam at 21 cm is 30’, and thus the observed region covers 100’ x 50’ (which corresponds to the central 25Ox 125 pc of the Galaxy at a distance of 8.0 kpc). These observations were made using both the DnC and CnB array configurations, giving a final spatial resolution of 15”. 127 channels were used with a total bandwidth of 1.5 MHz, corresponding to -300 km s-lof velocity coverage, centered at 0 km s-*, and a velocity resolution of 2.5 km s-’. The data were calibrated using the standard AlPS routines. The line-free channels (channels 1026) were used to fit a constant continuum level, and continuum subtraction was performed using the AlPS task UVLSF. The data were deconvolved jointly using the rniriad software package to take full advantage of the additional sensitivity from overlapping fields. Each plane of the HI absorption cube was cleaned to a uniform level using mossdi in miriad. N

N

3 Assembling an HI Absorption Catalog 3.1 Continuum Emission Figure 1 shows the continuum image constructed from -10 line-free channels. In addition, the wide field of view provided by the five mosaicked pointings allows comparison with large-scale surveys at other frequencies, such as the 90 cm image of LaRosa et al. (2000). The high sensitivity and resolution of the continuum image also allowed us to identify nine new compact sources. 3.2 HI Absorption

The continuum image has been used to guide the assembly of a catalog of continuum-weighted, line-tocontinuum HI absorption spectra towards -40 distinct continuum sources in the field. Typical rms noise for the spectra are in the range of 0.01-0.05, where the units are line-to-continuum ratio. Figure 2 shows a sample of two integrated and continuum-weighted HI profiles toward Sgr A East and Sgr A West. As Figure 2 shows, a distinct absorption feature at -40-50 km s-’ is present in the Sgr A East spectrum, but not in that of Sgr A West. This is consistent with the known interaction of Sgr A East with the so-called 40

‘

The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.

Astron. Nachr./AN 324, No. S I (2003)

3

-20 10

20

30

40

s 8 N

2

50

z g I-

9 -

-2900

4

Y 0

10

20

30

40

50

-

49

I 40

I

I

47 46 45 44 RIGHT ASCENSION (52000)

43

I

I

I

-1

Fig. 1 21 cm continuum image showing the cenfral 250 x 125 pc (100’ x 50‘) of the Galaxy, with a spatial resolution of 15”. The greyscale represents total intensities ranging from 0 to 125 niJy beam-’. The inset image is also a 20 cm VLA image of the central SO pc of the Galaxy (from Lang, Moms & Echevdma 1999) showing the key components in the active Radio Arc region of the GC.

km s-’cloud, and with previous conclusions that Sgr A West lies at the near cdge of Sgr A East (though still embedded in the Sgr A East shell). Figure 3 shows a sample of the continuum image (with the region used for the H I profile marked by greyscale) and corresponding HI absorption profile. Many of the profiles have very complex velocity structure. Similar profiles for -40 sources provide an overall sample o f the dataset and can guide more detailed analysis. Wc arc currently generating optical depth profiles for these regions and fitting these components to solve for the central velocities of components in each profile for identifying GC and line-of-sight features toward each of thc 40 regions.

4

C. Lang et al.: HI towards the Galactic Center

Sgr A East HI spectrum

-150

-100

-50

0 h (km/4

Sgr A West HI Spectrum

50

100

150

-150

-100

-50

0

50

1W

150

b [km/sl

Fig. 2 (left panel): Integrated, continuum-weighted HI absorption spectrum for Sgr A East region: (right panel): Integrated, continuum-weighted HI absorption spectrum for Sgr A West region

Sgr El1 HI Spectrum

7

Fig. 3 (left panel): Continuum image (greyscale >5u) used for determining the continuum-weighted profile of HI ahsorption toward Sgr B 1 ; (right panel): Integrated, continuum-weighted HI absorption spectrum for Sgr B1 region.

Astron. Ndchr./AN 324, No. S1 (2003)

5

4 Preliminary Science Results 4.1

The Atomic Component of the Arched Filament Complex

Preliminary results from the HI absorption cube already illustrate that this dataset is rich in detailed velocity and opacity information. Figure 4 shows sample data from our HI absorption survey toward a portion of the Arched Filament HII complex. This H I I complex represents one of the most active sites in the GC, with curved filamentary arcs of ionized gas tracing out the edge of a dense molecular cloud. The extraordinary Arches cluster is responsible for the heating of this cloud (Lang, Goss & Morris 2001) but is embedded in the complex structure of filamentary molecular and ionized gas (Lang, Goss & Morris 2002). The presence of two components of atomic gas (at V N -25 kin spl and -40 km s-') and higher HI opacities toward the W1 and W2 filaments indicate that part or the atomic gas lies on the near side of this complex and that the cluster is indeed embedded within this complex. The molecular gas is presumably separated from its ionized edge by a layer of atomic material which has been photodissociated by the ionizing flux of the cluster and must have a distribution such that some of the cloud surfaces along our line of sight have not been exposed to the ionizing radiation. Figure 5 shows a sketch (from above the cloud) of a possible arrangement of various components (molecular, atomic and ionized gas in addition to the stellar cluster) that can explain the velocity profiles toward different parts of this complex (from Lang, Goss & Morris

2002).

Fig. 4 A sample of Hr absorption profiles lrom the new HI dataset toward the Arched Filament complex, which consists of molecular, ionized, atomic gas components and the dense Arches stellar cluster. The central panel shows the identification of features in 8.3 GHz radio continuum from Lang, Goss & Morris 2001) and the left and right panels show HI absorption profiles integrated over two of the filamentary features, E2 and the W1 and W2 regions. The idenfitication of the HI components are given in the figure.

C. Lang et al.: HI towards the Galactic Center

6

Ionized Gas

4 Observer

3

4

Fig. 5 A schematic diagram showing a view of the ionized, molecular, atomic, and stellar components in the Arched Filament complex from a position “above” the molecular cloud, looking down its long axis. The dotted lines represent certain sightlines through the complex where complicated, double profiles are observed in the molecular and ionized gas. The numbering scheme (1-4) corresponds to sightlines of the CS(2-1) and H92cu profiles in Figure 10j, h, g, and d from Lang, Goss & Morris (2002).

4.2 Constraining Distances to Galactic Center Sources The unusual nature and morphology of the radio continuum features detected across the central 200 pc of the GC still provides the strongest case for these features being related to the center of the Galaxy. Such peculiar phenomena (e.g., the Radio Arc region comprised of the unusually-shaped “Sickle” and “Pistol” HII regions as well as the many-stranded magnetic NTFs) are not detected in radio emission toward any other region of the Galactic plane (Gray 1994). So far, we have relied on this morphological uniqueness to indicate that these sources actually reside within the central parts of the Galaxy. However, the radio continuum alone does not provide a way to adequately constrain the location of these features along our line of sight or their arrangement with respect to each other. Because of velocity crowding near .f=O degrees, the method of using galactic rotation to determine distances to GC sources can also not be used. However, the detection and identification of known HI absorption components (such as the “3 kpc-arm” and the “expanding molecular ring”; see Introduction) present in the spectra of known radio continuum sources can be used to place constraints on the distance to the radio continuum sources. Roy (2003) has recently demonstrated that distances to several of the GC magnetic filaments (e.g., Sgr C, G354.54+0.18 and G359.79+0.17) can be well-constrained to be within a few hundred pc of the GC region from measurements of HI absorption using data from the Giant Metrewave Telescope in India. We plan to carry out such anaylsis for our catalog of sources.

Astron. Nachr./AN 324, No. SI (2003)

4.3

I

A Wide HI Component Toward Sgr A

The HI absorption profiles toward parts of the SgrA complex show the presence of a large velocity dispersion (LVD) (AV -50 km s-') HI component. Several followup observations were made with the VLA in A,B,C and D array configurations using a much broader velocity coverage of -600 km s-lin order to try to characterize this unique HI component. The new data show that a low-level HI feature (~-0.3210.12), centered at -4&15 km s-lis distributed over the central I0 x 5 pc (4' by 2'). This dispersion is ten times larger than that observed in most of the cold, diffuse HI concentrations. The dispersion also appears to increase away from SgrA' indicating that the LVD HI is likely to be rotating around a distributed mass near SgrA* (Dwarakanath et al. 2003).

5

Ongoing Work on the Galactic Center HI Absorption Survey

The analysis of this rich dataset has just bcgun and the following steps outline what is planned for the thorough analysis and discussion of this H I Absorption survey: Assemble a complete catalog of HI absorption profiles toward all continuum sources in the central 200 pc. Identify the well-known sources of H I absorption toward all sources in the G C region. Such identification will also rely heavily on the molecular data of Oka et al. (1998) and other GC datasets, such as recombination line data. Constrain the distances to as many continuum sources as possible using this multi-wavelength HI absorption component identification. Construction of images showing the distrihution of HI opacity ovcr appropriate velocities. Determination of physical associations by careful comparisons of the velocity signatures of the HI , molecular and ionized gas structures. Synthesis of the HI line absorption data with the available multi-wavelength datasets is the key to determining physical interactions. 0

Quantitative estimates of the energy balance in stellar and gas components can also be derived using models of massive star energy input or Norman & Ikeuchi (1989).

Acknowledgements The authors would like to thank K. S. Dwarakanath for his assistance with putting together the catalog and analyzing the HI profiles, and S. Kim for assistance with the initial imaging. References Bally, J., Stark, A.A., Wilson, R.W., & Henkel, C. 1988, ApJ, 324, 223 Binney, J., Gerhard, O., Stark, A., Bally, J., & Uchida, K. 1991, MNRAS, 252, 210 Cohen, M. & Davies, R. 1979, MNRAS, 186,453 Dickey, J. & Lockman, F. J. 1990, Annual Reviews of Astronomy & Astrophysics, 28,215 Dwarakanath, K., Goss, W.M., Zhao, J.H. & Lang, C.C., 2003, ApJ, submitted Egan, M. P.. Shipman. R. F., Price, S. D., Carey, S. J., Clark, F. 0. and Cohen, M. 1998, ApJ, 4Y4, L199 Gray, A. 1994, MNRAS, 270, 822 Lang, C.C., Morris, M.. & Echevarria, L. 1999, ApJ, 525,727 Lang, C.C., Goss, W.M. &Moms, M. 2001. ApJ, 121, 2681 Lang, C.C., Goss. W.M. & Moms, M. 2002, AJ, 124, 2677 LaRosa, T. N., Kassim, N. E.. Lazio, T. J. W. and Hyman, S. D. 2000, AJ, 119,207 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyazaki, A. 1998, ApJS, 118,455 Norman, C. A. & Ikeuchi, S. 1989, ApJ, 345, 372 Roy, S. 2003, A&A, in press. Wang, Q. D., Gotthelf, E.V & Lang. C.C. 2002, Nature, 415, 148

This Page Intentionally Left Blank

Astron. NachrJAN 324, No. S1,9- 16(2003) / DO1 10.1002/asna.200385077

High Resolution, High Sensitivity Imaging of the Galactic Center at 330 MHz. Michael E. Nard*'.*, Crystal L. B r ~ g a n * *Scott ~ , D. Hymad, T. Joseph W. Lazio', Namir E. Kassim' ,T.N. LaRosa5, K. Anantharamaiah***6,and Neboja Duric2

' Naval Research Laboratory, Code 7213, Washington, DC 203755351 USA ' University of New Mexico, Department of Physics and Astronomy, 800 Yale Blvd.

'

NE, Albuquerque, NM, 87131, USA National Radio Astronomy Observatory, 1003 Lopezville Rd. Socorro, NM, 87801, USA Department of Physics, Sweet Briar College, Sweetbriar, VA, 24595, USA Department of Biological & Physical Sciences, Kennesaw State Univ, 1000 Chastain Rd., Kennesaw, GA 30144 USA Raman Research Institute, C.V. Raman Avenue, SaddShiVanagdr Post Office, Bangalore 560-080, India

Key words Galactic Center, Low Radio Frequencies, Sgr A*, Non-Thermal Sources, Wide-field Imaging.

-

Abstract. We present results derived from a wide field, sub-arcminute resolution VLA image of the Galactic Center region at 330 MHz (A = 90 cm). With a resolution of 7" x 12" and an rms noise of 1.6 mJy bean-', this image represents a significant increase in resolution and sensitivity over the previously published VLA image at this frequency (eg. LaRosa et al. 2000). The improvement in sensitivity has significantly increased the census of sinall diameter sources in the region, resulted in the detection of two new Non-Thermal Filaments (NTFs) and 18 new NTF candidates, and resulted in the lowest frequency (tentative) detection of Sgr A*.

1 Introduction The Galactic Center (GC) was first imaged with the VLA at 330 MHz in 1989 (Pedlar et al. 1989; Anantharamaiah et al. 1991) and represented a major improvement in sensitivity and resolution over past meter wavelength images. However, wide field imaging algorithms at the time were unable to compensate for the non-coplanar nature of the VLA. Hence the full primary beam of the VLA at 330 MHz (-- 2.2' radius) was not imaged and only the very center of the GC region was studied. The G C image published by LaRosa et al (2000) represented a major improvement. It led to the discovery of many new sources, and provided an unparalleled census of both extended and small diameter, thermal and non-thermal sources within 100 pc (projection) of the GC. This major step forward was afforded by significant advances in wide-field imaging algorithms, coupled with greatly increased computational power. However that effort fell short of utilizing the full resolving power of the VLA and the commensurate greater sensitivity that it would have afforded. Since those data were obtained and published, significant improvements in software, hardware, and computational power have continued 10 be realized. This inspired us to revisit the G C at 330 MHz.

* Corresponding author: e-mail: Michael.NordOnrl.navy.mil,Phone: +1 202767 7310, Fax: +01202404 8894 * * NRAO Jansky Postdoctoral Fellow. * * * Deceased.

@ 2003 WILEY-VCH Verlag GmhH & Co K O I A Weinhem

10

M. E. Nord et al.: High Resolution, High Sensitivitv Imagine of the Galactic Center at 330 MHz.

Fig. 1 The Galactic Center Region at 330 MHz as imaged with the VLA. This image has aresolution of 7" x 12" and an rms noise of 1.6 mJy bean-', and is a sub-image of a much larger image. This sub-image is roughly 0.8" x 1.0". The entire image covers an area of nearly 15.5 square degrees. Two new Non-Thermal Filaments and 18 new NonThermal Filament candidates were discovered in this image. A full explaination of how the image was constructed as well as the full details on results will be published in Nord et al., in prep.

Astron. Nachr./AN 324. No. S1 (2003)

I1

We present here the image generated from ncw A and B configuration data sets, which are appropriate for generating an image with a minimum of confusion noise and maximum sensitivity to relatively smaller scale (< 1’) structure. For the first time, the entire GC region contained by the primary beam of the VLA has been imaged at the maximum possible resolution and sensitivity. The image has a resolution of 7” x 12’’ and an rms noise of 1.6 mJy beam-’, an improvement by roughly a factor of 5 in both sensitivity and resolution over the LaRosa et al. (2000) image. Figure 1 is a subimage of the central GC region. Here we describe new results pertaining 10 Sgr A* and to newly discovered NTFs and NTF candidates. A full description of data reduction techniques and results pertaining to compact source distribution, spectral indeces, greater detail in extcnded sources, and Galactic Center scattering will be presented in Nord et al. (2003, in preparation). Efforts to image the GC from the combination of data obtained in all four VLA configurations, which is more appropriate for the study of more extended (> 10’) emission, are under way.

2 Observations Three sets of observations were obtained. Thc tirst was observed at 330 MHz in the A configuration of the VLA (maximum baselines 35 km) in Octoher of 1996. Data were obtained in two IF’Swith dual circular polarization, a bandwidth of 3 MHz, and the total bandwidth was split into 32 channels in order to enable later radio frequency interference (RFI) excision as well as to mitigate the effects of bandwidth smearing. The 1996 data were obtained from a series of observations designed to find candidate GC pulsars (Lazio & Cordes, in prep.). Between March 1998 and May 1999, the Galactic Center was observed in all four VLA configurations at a single IF centered at 330 MHz with 3 MHz bandwidth and 32 channels, while the other IF was dedicated to 74 MHz. All observations wcre full synthesis which tracked the Galactic Center as long as it was visible to the VLA (- 6 hours). and unlike the archival data re-processed by LaRosa et al. (2000). all these new data were obtained using all 27 antennas of the VLA. These details are summarized i n Table 1. Images made from the 74 MHz data are presented in Brogan et al., these proceedings. ~4

Table 1 Observational Summary

Epoch 1996 October 1998 September 1998 March

VLA Configuration A

B A

I/

(MHz) 332.5 327.5 327.5

3 SgrA”

#of IFs 2 1 1

Integration (Hours)

Beam

5.55 5.47 5.38

9” x 5’’

36” x 20” 9’l x 5”

-

Below 10 GHz, Sgr A* has a nearly flat spcctrum with a flux density of 0.6 Jy. In 1989, Pedlar et al. suggested that at 330 MHz, the flux density of Sgr A* could be no more than 100 mJy. It is not fully understood why the flux density at 330 MHz should be so low. Pedlar et al. suggest that this is due to foreground free-free absorption of optical depth T 2.5, while Beckert et al. (1996) suggest that the turn over may be modeled by synchrotron self-absorption intrinsic to the source itself. In order to more accurately measure the flux density of Sgr A*, confusion with the large scale components of flux i n the region must be considcrcd. The data presented here are already chosen to minimize the contribution of large scale flux. The super-uniform weighting scheme (Briggs et al. 1999) can further reduce the contribution of short spatial frequencies in the image, thereby decreasing the size of the synthesized beam at the cost of increasing the overall noise level in the image. The region around Sgr A* was imaged with super-uniform weighting and through this technique the size of the synthesized beam was 10 mJy beam-’. decreased to 3.4“ x 8.2” (RA, DEC) while increasing the rms noise of the image to The super-uniform image is shown in Figure 2.

-

-

12

M. E. Nord et al.: High Resolution, High Sensitivity Imaging of the Galactic Center at 330 MHz.

-mm

-

Jf fY

01

Fig. 2 Super-uniformimage of the Sgr A* region at 330 MHz. Image display is inverted: white areas represent areas of low flux due to absorption. The grey scale is linear between -2 and 100 mJy bean-'. The resolution is 3.4" x 8.2'' and the rms noise is -10 mJy beam-'. Note the area of diffuse eniission at

Fig. 3 Model created from 6 cm data of the Sg A* region as it should appear at 330 MHz. Image display is inverted; white areas represent areas of low flux due to absorption. See Section 3 for details.

the center. The oval represents the location, predicted scattering diameter and orientation of Sgr A* at 330 MHz.

The scatter-broadened size of Sgr A* at 330 MHz was estimated using the measurements at 20.7 cm by Yusef-Zadeh et al. (1994). Taking the 20.7 cm size of 624 x 350 milliarcseconds and scaling by A', the size of Sgr A* is calculated to be 11.8" x 6.6". The position angle was assumed to be constant at -80". The scatter-broadened size, location, and position angle of Sgr A" are represented by the oval in Figure 2 Of note in the super-uniform image is a diffuse region of flux having a peak intensity within 5" of the location of Sgr A*. Slices along the major and minor axes of the scattering disk are shown in Figures 4 and 5. These slices strongly suggest that there is a narrow region of 330 MHz flux near the position of Sgr A*. A gaussian fit with baseline subtraction to this region gives a source size of 19" f 4 x 11" f 1, a position angle of -30", a peak intensity of 50110 mJy beam-' and a flux density of 330f100 mJy. The expected size of Sgr A' convolved with the super-uniform beam gives an expected source size of 10.5'' x 12". With the correct position, nearly the correct size, and a flux density near that of higher frequency values, we decided to test the hypothesis that this diffuse source is Sgr A*. In order to check that the source being observed is in fact Sgr A*, a 330 MHz model of the Sgr A* region was created from a 6 cm (v = 5 GHz) image. Starting with a 6 cm image (Yusef-Zadeh et al. 1987), a 0.75 Jy model of Sgr A* was subtracted. The remaining flux should be ionized gas from the 'thermal spiral', Sgr A West (Lo & Claussen 1983). As this ionized gas will be in absorption against the non-thermal flux from the Sgr A East supernova remnant at 330 MHz, the 6 cm image is inverted, so that positive flux becomes negative. A model of Sgr A* with a flux density of 0.75 Jy, and the appropriate 330 MHz scatter-broadened size was added back into the image, and then the image was convolved to have the resolution of the super uniform image at 330 MHz. The resulting image is shown in Figure 3. This model shows significant similarities to the observed super uniform image. The region of absorption caused by the thermal gas matches nearly exactly. More interestingly, at thc position of Sgr A*, there is a region of extended flux similar in size, shape and location as in the 330 MHz super uniform image. Morphologically, N

Astron. Nachr./AN 324, No. S1 (2003)

Fig. 4 Slice through the assumed major axis of Sgr A* in the super-uniform image (Figure 2). The slice is centered at 17h 45"' 40" -29"00'28"(52000) with a position angle of -90".

13

Fig. 5 Slice through the assumed minor axis of Sgr A* in the super-uniform image (Figure 2). The slice is centered at 17h 45'" 40" -29"00'28"(52000).

the diffuse region in the 330 MHz super uniform image at the position of Sgr A* appears similar to what this model predicts. Before identifying this diffuse source as Sgr A*, the possibility that the source is background flux from the Sgr A supernova remnant was investigated. Figure 6 is a greyscale image of the Sgr A* region with the 6 cm, Sgr A* subtracted image of the previous paragraph superimposed in contours, and the position, scatter broadened size, and position angle of Sgr A* represented by an oval. At the position of Sgr A", the 6 cm intensity is near maximum, suggesting that this should be the area of maximum absorption at 330 MHz. However, what is seen at 330 MHz is a region of extended emission. This observation makes it unlikely that the flux originates from behind the absorbing ionized region. We therefore conclude that we are detecting Sgr A* at 330 MHz for the first time. The position of maximum intensity, size, and shape of the dirfuse region all agree with our model of how the source should appear at 330 MHz. We fit a maximum intensity of 504110 mJy beam-', and a flux density of 3301100 mJy. However, we leave open the possibility that the flux is from, or is contaminated by, background non-thermal flux from the Sgr A East supernova remnant. See Figure 7 for a radio (0.33 < v < 2 3 GHz) spectrum of Sgr A*.

4 Non-Thermal Filaments Among the most fascinating of the unique structures in the Galactic Center are the non-thermal filaments. NTFs are remarkably coherent magnetic structures that extend tens of parsecs and maintain widths of only a few tenths of parsecs (e.g., Lang et al. 1999). It has been hypothesized that the NTFs are part of a globally ordered, space filling magnetic field (e.g., Morris & Serabyn 1996). If so, they would be the primary diagnostic of the GC magnetic field. An alternative idea is that the NTFs are magnetic wakes formed from the amplification or a weak global field through a molecular cloud galactic center wind interaction (Shore & LaRosa 1999; see also LaRosa et al., these proceedings).

M. E. Nord et al.: High Resolution, High Sensitivity Imaging of the Galactic Center at 330 MHz.

14

17 45 44

42

40 38 RIGHT ASCENSION (J1000)

36

Fig. 6 Greyscale image of the Sgr A* region at 330 MHz. Image display is inverted white areas represent areas of low flux due to absorption. The grey scale is linear between -5 and 400 mJy beam-’. The resolution is 6.81” x 12.58” and the rms noise is ~ 1 . mJy 6 beam-’. Contours are 6 cm with a resolution 3.39” x 2.93” and levels of 1, 2, 3, and 3 times SO mJy beam-’. A model of Sgr A’containing 750 mJy has been subtracted from the 6 cm contours. Note the area of diffuse emission at the center. The oval represents the location, predicted scattering diameter and orientation of Sgr A*at 330 MHz.

Fig. 7 The low frequency (0.33 < u < 23 GHz) spectrum of Sgr A*. The 0.61 GHz value is from Subhashis Roy (private communication). The 1.4 < u < 23 GHz values are from Zhao et al. (2001). Also included is the projected LOFAR detection limit at 240 MHz (see Kassim et al., these proceedings).

Nine NTFs were known before this work was completed. Of those nine, we detect eight as we do not have sufficient surface brightness sensitivity to detect G359.85+0.39 (LaRosa et al. 2000, 2001). In addition to higher resolution and higher sensitivity measurements of the known NTFs, we report the discovery of two new NTFs and 18 new NTF candidates. For the two new NTFs, confirmation of NTF status was made by observations of 6 cm polarization which is reported on in greater detail in LaRosa et al. (2003a, these proceedings). If all are confirmed, these represent a tripling of the number of known NTFs. Table 2 summarizes the new NTFs and NTF candidates. Of particular interest is the orientation with respect to the Galactic Plane of these new NTF candiates, and what this can tell us about the GC magnetic field. This topic as well as images of several of the new NTFs and NTF candidates is covered In LaRosa et al., (2003a, thcse proceedings). Figure 8 is a histogram of the NTF’s maximum intensity. We use intensity instead of flux density for two reasons. Firstly, detection of these sources is based on maximum intensity, not overall flux and secondly baseline subtraction can be very difficult for very extended sources that pass through regions of diffuse flux, so the total integrated flux density of the NTFs is uncertain. Though there are not enough sources to definitively fit a surface brightness curve, the increase towards low surface brightness rises faster than SP2. By increasing the sensitivity by a factor of -5, we have tripled the linear and appears to be number of known NTFs, suggesting that the total number of NTFs rises at minimum as N ( S ) S-3/5.

-

N

15

Astron. Nachr./AN 324, No. S I (2003) Table 2 New Non-Thermal Filaments and Candidates

Name

Intensity (mJy beam

NTF 359.12+0.66" NTF 359.22-0.16' NTF 359.33-0.42 NTF 359.36+0.09 NTF 359.40-0.03 NTF 359.40-0.07 NTF 359.43+0.13 NTF 359.59-0.34 NTF 359.66-0.1 1 NTF 359.85-0.02 NTF 359.86-0.24 NTF 359.88-0.07 NTF 359.90+0.19 NTF 359.99-0.54 NTF 0.02+0.04 NTF 0.06-0.07 NTF 0.37-0.07 NTF 0.39+0.05 NTF 0.39-0.12' NTF 0.43+0.01

11.9 23.4 13.9 10.7 11.9 40.6 18.8 20.8 9.9 8.5 11.2 33.8 11.9 9.4 22.3 10.5 14.1 17.8 16.1 11.6

')

Flux Density (dy) 647.8 268.5 80.6 65.2 93.8 229. I 264.5 188.2 226.0 172.5 205.1 930.0 129.2 88.8 227.8 162.6 128.1 231.5 731.2 43.9

Size

Plane Angle"

(9

(")

15.6 x 0.2 1.8 x 0.5 2.0 x 0.2 2.1 x 0.2 1.6 x 0.2 1.7 x 0.3 2.4 x 0.3 2.3 x 0.2 3.5 x 0.5 1.8 x 0.2 8.1 x 0.2 1.6 x 0.2 2.4 x 0.2 8.6 x 0.2 2.0 x 0.3 2.1 x 0.2 1.1 x 0.3 4.1 x 0.3 10.1 x 0.3 1.6 x 0.3

35 55 55 60 5 40 0,90d 25 20 90 35 5 35 30 0 15 5 5 5 5

Plane Angle is the angle of the NTF with respect to the normal to the Galactic Plane. This filament was first detected at higher frequencies ( M . Morris 2002, private communication). Source observed to have significant 6 cm polarization, and therefore is confirmed an NTF (LaRosa et al 200%. these proceedings). This source may be two interacting NTFs with orientations of Oo and 90' to the Galactic plane.

We conclude that just the tip of the NTF luminosity distribution is being detected and we hypothesize that there may be hundreds more NTFs in the G C region (see Kassim et al. 2003, these proceedings).

5

Conclusions

Rcsults from a new high resolution, high sensitivity imaging of the Galactic Center region at 330 MHz were presented. In addition to more than tripling the number of NTF candidates, we report the tentative lowest frequency detection of Sgr A*. Further results will be presented in Nord et al., in prep. In addition to shorter baseline VLA data that is LO be added to this dataset, time has been allocated on thc Green Bank Telescope for adding single dish information to these data. This will allow us to make the connection from the structures seen in our interferometer images to those larger scale features seen on single-dish images which have been interpreted as being related to, for example, episodic periods of star-burst activity that can also be traced in X-ray images (Sofue 2000).

Acknowledgements This project was originally conceived under the guidance of K. Anantharamaiah. Anantha passed away during the data reduction phase of this project and will be missed greatly. The authors would like to thank Manana S . Lazarova and Jennifer L. Neureuther, students at Sweet Briar College for their assistance in small diameter source location and quantification.

M. E. Nord et al.: High Resolution, High Sensitivity Imaging of the Galactic Center at 330 MHz.

16

Fig. 8 NTF maximum intensity histogram, including the 9 previously known NTFs, 2 newly discovered NTFs and 18 NTF candidates.

Basic research in radio astronomy at the NRL is supported by the Office of Naval Research. S.D.H was supported by a grant from the Jeffres Memorial Trust and Research Corporation. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.

References Anantharamaiah, K. R., Pedlar, A,, Ekers, R. D., & Goss, W. M. 1991, MNRAS,249,262 Beckert, T., Duschl, W. J., Mezger, P. G., & Zylka, R. 1996,A&A,307,450B Briggs, D. S., Schwab, F. R., & Sramek, R. A. 1999, ASP Conf. Ser. 180: Synthesis Imaging in Radio Astronomy 11, 127 Brogan et al. 2003, these proceedings Kassim et al. 2003, these proceedings LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, AJ, 119, 207 LaRosa, T. N., Lazio, T. J. W., & Kassim, N. E. 2001, ApJ, 563, 163 LaRosa et al. 2003a, these proceedings LaRosa et al. 2003b (in preparation) Lazio & Cordes 2003 (in preparation) Lo, K. Y. & Claussen, M. J. 1983, Nature, 306,647 Liszt, H. S. & Spiker, R. W. 1995, ApJS, 98,259 Moms, M. & Serabyn, E. 1996, ARA&A, 34,645 Nord et al. 2003, (in preparation) Pedlar, A,, Anantharamaiah, K. R., Ekers, R. D., Goss, W. M., van Gorkom, J. H., Schwarz, U. J., & Zhao, J. 1989, ApJ, 342,769 Shore, S. N. & Larosa, T. N. 1999, ApJ, 521,587 Sofue, Y. 2000, ApJ, 540, 224 Yusef-Zadeb, F. & Morris, M. 1987, ApJ, 320,545 Yusef-Zadeh, E, Cotton, W., Wardle, M., Melia, F., & Roberts, D. A. 1994, ApJ, 434, L63 Zhao, J., Bower, G. C., & Goss, W. M. 2001, ApJ, 547, L29

Astron. Nachr./AN 324, No. S 1, 17 - 24 (2003)/ DO1 10. I002/asna.2003F5023

Spatially Resolved Very Large Array 74 MHz Observations Toward the Galactic Center C. L. Brogan*’,M. Nord’, N. Kassim’, J. Lazio’, and K. Anantharamaiah**3

’ National Radio Astronomy Observatory, 1003 Lopezville Rd., Socorro, NM 87801 Navel Research Laboratory,4555 Overlook Ave., Code 7213, Washington, DC, 20375 ’ Raman Research Institute, C. V. Raman Avenue, SadashivanagarPost Office, Bangalore 560-080, India

Key words Cosmic Rays, Galactic Center, H ‘ 1 regions, Supernova Remnants PACS 04A25

We present the highest resolution and sensitivity low frequency image (< 300 MHz) of the Galactic center to date using the Very Large Array at 74 MHL in its A, B, C, & D configurations. The resulting images have a resolution of 2.1’ x 1.2 and a dynamic range of 400. From this data we have been able to identify a region of enhanced 74 MHI emission about 5” in extent that is coincident with the high density molecular gas surrounding the Galactic ccntcr known as the Central Molecular Zone. In addition to giving an unprecedented view of the extended nontherinal emission surrounding the Galactic center, the 74 MHL image shows deep free-free absorption across the Galactic center itself, as well as, part of the Galactic center radio lobe, and a number of H II regions in the field. This absorption is due to ionized thermal gas in front of, or in some cases embedded in, the nonthermal Galactic center (CC) emission. Such absorption allows us to unambiguously place some of the H I I regions in the direction of the GC along the line of sight for the first time. The morphology, nature, and relationship to the Galactic center of the 74 MHz absorption and emission is discussed. N

1 Introduction Low frequency (< 100 MHz) observations of the Galactic center (GC) provide a unique window on distinct synchrotron sources like supernova remnants, as well as, the diffuse G C synchrotron emission. In addition, because ionized thermal gas can absorb low frequency synchrotron emission, these data can also be used to unambiguously place H I I regions along the line of sight toward the GC. In the past, radio observations bclow I00 MHz were of limited usefulness due to their low resolution and sensitivity. The recent addition to the Very Large Array (VLA) of a 74 MHz receiver system has allowed for high resolution and sensitivity low frequency observing for the first time. The 25 mcter size of the VLA antennas also provides a very large 14” field of view at 7 4 MHz, thus allowing for a complete picture of the whole GC region. Since the detection of absorption in radio continuum data is relatively rare, we briefly describe the physics behind such observations. For objects observed in 74 MHz absorption with an interferometer, the where , T, is the electron tempcraturc of the observed brightness temperaturc will be T o h s = T, - T G , ~ ionized thermal gas (H I I region) and T c : ,is~ the temperature of the Galactic synchrotron crnission behind the H 11 region. Therefore, if T G ,<~ T, the observed brightness temperature will be negative and an H 11 region will appear as an absorption “hole”(c.f. Kassim 1990). N

* =*

Corresponding author: e-mail: Deceased

[email protected],Phone: cO15OS 835 7224, Fax: +(I1505 835 7027

@ ZOO3 WlLEY-VCH Verlag Gmhtl B Co KGaA, Weiritieini

C. L. Broedn et al.: 74 MHz Observations of GC

18

2 Observations In order to sample the Galactic center region on a wide range of spatial scales, we have observed this region in all four VLA configurations (A, B, C, & D). The images resulting from this combination of data have a resolution of 2.1' x 1.2,an rms noise of 0.12 Jy beam-', and a dynamic range of 400. This image has a synthesized beam area 17 times smaller and an rms noise 40 times smaller than any previous synthesis image of the GC with v < 100 MHz (LaRosa & Kassim 1985). Figure 1 shows the full 14" 74 MHz GC field of view with several of the prominent sources labeled for reference. Figure 2 shows a close up of the inner 4" with many of the sources mentioned in the results section below labeled. For comparison, Figures 3a and 3b show a VLA C and D configuration image at 330 MHz, and a 10 GHz Nobeyama image from Handa et al. (1987), respectively. The VLA images at 74 and 330 MHz have been corrected for primary beam attenuation. Note that these VLA 74 MHz images are not sensitive to emission on size scales larger than 6" due to spatial filtering by the interferometer. Low frequency observations offer a number of unique challenges including the very large fields of view, ionospheric refraction, strong interference, and 3-D effects from non-coplanar arrays (c.f. LaRosa el al. 2000). The details of the data reduction techniques used to make these images will be described in a future paper (Brogan et a]. 2003, in prep.).

-

-

-

-

-

3 Results 3.1 GC 74 MHz Emission Concentrated Inside the CMZ

-

One of the most prominent features of the GC 74 MHz image shown in Figure 1 is a 5" sizescale region of diffuse emission. Interestingly, this diffuse synchrotron emission is coincident with the region of high density, high temperature, high velocity dispersion molecular gas in the GC called the central molecular zone (CMZ; c.f. Morris & Serabyn 1996). Figure 4 shows a 74 MHz image of the GC with integrated CO emission contours superposed showing the extent of the CMZ from Dame, Hartmann, & Thaddeus (2001). The CMZ, which has a radius of 200 pc, contains about 10% of the total neutral gas in the Galaxy. The outskirts of the CMZ has also been identified as the expanding molecular ring (EMR) which according to some models is expanding at a velocity of 160 km s-' and rotating at 70 km s-l (Sofue 1995). The EMR is thought to be the source of some of the very high velocity features seen toward the Galactic center region. Previous low frequency images toward the GC have been of insufficient resolution and sensitivity to recognize this diffuse, yet distinct synchrolron emission component due to confusion with the ubiquitous Galactic plane synchrotron emission, distinct synchrotron sources like SNRs, extragalactic sources, and at the lowest frequencies free-free absorption. However, comparison of the 74 MHz image shown in Fig. 1 with previous low frequency surveys using total power data at 35 MHz (Dwarakanath & Udayashankar 1990) and 408 MHz (Haslam et al. 1982) show conclusively that this diffuse emission is distinct from the Galactic synchrotron background (to which we are not sensitive). By comparing our 74 MHz D configuration image to that of Haslam et al. (1 982) at 408 MHz (resolution 0.9") our preliminary analysis suggests that the diffuse emission has a spectral index a -0.7 to -0.5 in regions that are free of absorption and known SNRs. The concentration of diffuse synchrotron emission inside the CMZ is likely due to the enhanced density, magnetic field, or cosmic ray acceleration due to the increased star formation rate (and most probably a combination of all three) within this region.

-

-

-

-

3.2 Galactic Center Omega Lobe

-

-

Sofue & Handa (1984) were the first to identify a large, faint loop like structure (diameter 1.5 degrees) jutting out from the Galactic center region toward positive latitudes from the 10 GHz Nobeyama GC survey images (see Fig. 3b). The eastern half of this lobe (w.r.t. the Galactic plane) appears to be connected to the radio arc, while the western half extends down to the plane near the Sgr C region (see Figs. 2 & 3b). Sofue

Astron. NachdAN 324, No. S 1 (2003)

19

& Handa named this structure the GC Omega Lobe, due to its morphology. The GC Omega Lobe is not apparent in the 330 MHz VLA C & D configuration image shown in Fig. 3a. This could be due to either a high 330 MHz optical depth, or more likely spatial filtering since the 330 MHz image is not sensitive to size scales 2 1”.We have been approved for 330 MHz GBT observations in order to add zero spacing data to this image, in part to address this issue. Tsuboi et al. (1986) observed the polarized emission from this region at 10 GHz and curiously found that the eastern half of the lobe (w.r.t. the Galactic plane) is highly polarized while the western half of the lobe is not (also see Seiradakis et al. 1989). In excellent agreement with this finding, our 74 MHz image shows that the eastern half of this lohe is only weakly absorbed, while the western half is very heavily absorbed, indicative of the presence of ionized thermal gas. The 74 MHz absorption data alone could also be explained by a difference in the pathlength to each side of the lobes along the line of site, e.g. the western lobe could be closer to us and therefore absorb more of the G C synchrotron emission. However, coupled with the polarization data, these results suggest that this mysterious source is composed of primarily synchrotron gas on one half and i s mostly thermal on the other. The GC Omega Lobe was also recently detected at 8.3 p m from the Midcourse Space Experinlent (MSX) satellite (Bland-Hawthorn & Cohen 2003). These mid-IR data show both the eastern and western halves of the positive latitude lobe, but the wcstern lobe is much more distinct. Bland-Hawthorn & Cohen (2003) find that the dust temperature is higher than can be accounted for from the disk stellar radiation field and favor a starburst driven wind model for the lobe with a shell velocity of 150 km s-’ and a total energy of 2 los4 erg. There is also some indication in both the polarization data and the mid-IR data of a negative latitude component to the Omega Lobe (Seiradakis et al. 1989; Bland-Hawthorn & Cohen 2003), unfortunately, this region is confused by local H I I regions in the radio continuum (see $3.3). N

3.3

74 MH7, Absorption by H 11 Regions

-

The location of the Sgr D H 11 region along the line of site has been a matter of contention for some time. The Sgr D H II region has a radio recombination line (RRL) velocity of -20 km s-’, in agreemcnt with a narrow velocity component observed in a number of high density molecular lines in this direction (Downes 1980; Lis 1992). However, since the velocity of Sgr D H I I does not agree with the velocity of the bulk of high density molecular gas in this direction (ranging from +20 to +I60 km s-I), it also seems clear that the Sgr D H 11 region does not lie within the nuclear disk (Liszt 1992; Lis 1992). The arguments for its location beyond the GC rely on high velocity (2 +50 km s-’) HzCO absorption measurements that could only originate from the nuclear disk gas (Downes 1980; Mehringer et al. 1998). Howcver, since the excitation temperature for HzCO i s quite low ( w 1.7 K), it is possible that the absorption of this high velocity gas is due to the blackbody background and not the H I I region. A recent narrowband 2 pm study of this region by Blum & Damineli (1999) suggests that there is a deficil of stars toward this region which can only be accounted for by placing the H I I region on the near side of the GC. Additionally, the kinematic model of the Galactic spiral arm clouds in the direction of the Galactic center by Greaves & Williams (1994) suggests that the Sgr D H 1 1 region could lie within the “4 kpc” arm of the Galaxy. With our new 74 MHz data, we can say unambiguously that at least the extended emission associated with this source must be located on the near side of the GC, and likely on the near side of the CMZ since we see it in strong ahsorption against the diffuse G C synchrotron emission. The Sgr B 1 and Sgr B2 H I I regions, which are known to lie within the G C nuclear disk, show up only weakly in 74 MHz absorption, with Sgr B2 more obvious than Sgr B 1. Higher frequency continuum observations of these two H I I regions suggest that Sgr B2 has a higher electron temperature than Sgr B I (c.f. Downes 1980), inconsistent with the fact that Sgr B2 appears more deeply absorbed at 74 MHz assuming that they arc at the same distance and that the 74 MHz synchrotron emission behind them is the same. This result suggests that Sgr B2 is somewhat closer to us than Sgr B1. The average RRL velocities of these two regions are +45 and +G5 kin s-l for Sgr B1 and Sgr B2, respectively. Given the 20 N

-

C. L. Brogan et al.: 74 MHz Observations of GC

20

km spl difference in their RRL velocities it is not surprising that they may be at different distances within the nuclear disk. The Sgr A East and West regions are strongly absorbed at 74 MHz. This is not surprising since these regions are optically thick even at 330 MHz with T 1- 2 for Sgr East and T 4 for Sgr A West (Pedlar et al. 1989; LaRosa et al. 2000; also see Roy et al. and Nord et al. these proceedings). The region of the Radio Arc denoted GO. 16-0.15 shows up very clearly in 74 MHz emission, consistent with its detection at even lower frequencies (LaRosa & Kassim 1985), while the rest of the Radio Arc is absorbed. (30.16-0.15 is also coincident with the position of the Radio Arc peak at 330 MHz, but not at higher frequencies where the peak emission shifts to higher latitudes (see Figs. 2 & 3a,b). This shift in the location of the peak is presumably due to free-free absorption at 330 MHz (Anantharamaiah et al. 1991). This finding is in good agreement with the fact that GO. 16-0.15 is the only portion of the radio arc that shows strong linear polarization (Tsuboi et al. 1986; Seiradakis et al. 1989), a fact that has been attributed to strong Faraday rotation due to ionized gas toward the positive latitudes of the Radio arc where it connects to the thermal Arched filaments (c.f. Anantharamaiah et al. 199 1). Based on previous higher frequency continuum observations it is known that the Sgr C complex is composed of both thermal and nonthermal gas (c.f. Liszt & Spiker 1995). The Sgr C H 11 region (G359.430.09) and the middle and western parts of the Sgr C threads show weak absorption at 74 MHz, consistent with it being located within the nuclear disk. Evidently there is enough ionized thermal gas near, or in front of this part of the steep Sgr C synchrotron threads to cause absorption at 74 MHz (at our current resolution). The presence of thermal gas in front of the middle part of the threads is confirmed by the detection of emission in this region at 22 GHz (Tsuboi et al. 1991). In contrast, the eastern half of the Sgr C complex which appears at 330 MHz as a faint, diffuse extension of the threads, shows up very clearly at 74 MHz. Comparison between the 74 MHz image shown in Fig. 2 and the 330 MHz image shown in Fig. 3a suggests that the spectral index of this eastern extension of Sgr C is fairly steep with a -0.9. Interestingly, the average spectral index of the Sgr C threads between 330 and 1420 MHz is only LY -0.55 (Anantharamaiah et al. 1991). The Sgr E H 11 region complex does not appear in 74 MHz absorption, suggesting that this region is on the far side of the GC. However, it is likely that it lies within the nuclear disk given its very high RRL velocity of -210 km spl (Cram et al. 1996). We mention that the assessment that Sgr E lies on the far side of the GC is not completely conclusive, because the H 11 regions in this complex have very small size scales (Liszt 1992). This suggests that beam dilution could play a role. A number of the deepest 74 MHz absorption features apparent in Figures 1 & 2 are due to free-free absorption by local H I I regions along the line of site to the GC. For example, the region of strong 74 MHz absorption just east of the W28 SNR is due to the H 11 region M8, while the region of absorption near 17h25n118S,-34’10‘57” (52000) is due to the H I I region complex NGC 6357 (see Fig. 1). The absorption hole 1’west of Sgr D H II region is coincident with thc “diffuse H 11” region G0.64+0.623 detected by Lockman, Pisano, & Howard (1996) via radio recombination lines at a velocity of 3.7 km spl (see Fig. 2). Additionally, the whole region of deep absorption near G0.5-0.5 (17h4811’20s,-28’45’00‘‘ 52000) and (3359.7-0.3 (17h46m20.5S,-29’18’08” 52000) is composed of a collection of at least twelve local HI1 regions that have RRL and HzCO absorption velocities of +10 to +20 km s p l , indicating that they are within about 2 kpc of the sun (Downes 1980; Figs 2,3a,b).

-

N

-

N

N

N

-

4

Summary

We present the highest resolution and sensitivity low frequency image (< 300 MHz) of the Galactic center to date using the Very Large Array at 74 MHz in its A, B, C, & D configurations. The resulting images have a resolution of 2.1’ x 1.2‘ and a dynamic range of 400. From this data we have been able to identify a region of enhanced 74 MHz emission about 5” in extent that is coincident with the high density molecular gas surrounding the Galactic center known as the Central Molecular Zone. In addition to giving N

Astron. Nachr./AN 324, No. S1 (2003)

21

an unprecedentcd view of the extended nonthcrmal emission surrounding the Galactic center, the 74 M H z image shows deep free-free absorption across the Galactic center itself, as well as, part of the Galactic ccnter radio lobe, and a number of H 11 regions in the field. This absorption is d u e to ionized thermal gas in front of, or in some cases embedded in thc nonthermal Galactic center (GC) emission. This absorption allows us t o unambiguously place some of these H 11 regions along the line of sight for t h e first time. For example, due its strong 74 MHz absorption, we can definitively place the Sgr D H 11 region in front of the GC nuclear disk. Additionally the relative absorption difference between the Sgr B2 and Sgr B 1 H II regions suggests that the former is closer to us along the line of sight (but still within the GC nuclear disk). Acknowledgements The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. Basic research in radio astronomy at the Naval Research Laboratory is supported by the Office of Naval Research.

References Anantharamaiah, K. R., Pedlar, A,, Ekers, R. D., & Goss, W. M. 1991, MNRAS, 249,262 Bland-Hawthorn, J. & Cohen, M. 2003, ApJ, 582. 246 Blum, R. D. & Damineli, A. 1999, ApJ, 512,237 Cram, L. E., Claussen, M. J., Beasley, A. J., Gray, A. D., & Goss, W. M. 1996, MNRAS, 280, 1 1 10 Dame, T. M.. Hartmann, D., & Thaddeus, P. 2001, ApJ, 547,792 Downes, D., Wilson, T. L., Bieging, J., & Wink, J. 1980, A&AS, 40, 379 Dwarakanath, K. S., Udayashankar, N., 1990, JAPA 1 I.323 Greaves, J. S. & Williams, P.G. 1994, A&A, 290, 259 Handa, T., Sofue, Y., Nakai, N., Hirabayashi, H., Inoue, M. 1987, PASJ 39,709 Haslam, C. G . T., Salter, C. J., Stoffel, H., Wilson, W. E., 1982, A&AS 47, 1 LaRosa, T. N. & Kassim, N. E. 1985, ApJ, 299, L13 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, AJ, 119, 207 Lis, D. C. 1991, ApJ, 379, L53 Liszt, H. S. 1992, ApJS, 82,495 Liszt, H. S. & Spiker, R. W. 1995, ApJS, 98,259 L o c h a n , F. J., Pisano, D. J., & Howard, G. J. 1996, ApJ, 472, 173 Kassim, N. E. 1990, in Low Frequency Astrophysics from Space, N.E. Kassjm and K. W. Weiler (eds.), Lecture Notes in Physics Vo1.362 (Springer, Berlin), p. 144. Mehringer, D. M., Goss, W. M., Lis, D. C., Palnier, P., & Menten, K. M. 1998, ApJ, 493,274 Moms, M. & Serabyn, E. 1996, ARA&A, 34,645 Pedlar, A,. Anantharamaiah, K. R., Ekers, R. D., Goss, W. M., van Gorkom, J. H., Schwarz, U. J., & Zhao, J. 1989, ApJ, 342, 769 Seiradakis, J. H., Reich, W., Wielebinski, R., Lasenby, A. N., & Yusef-Zadeh, E 1989, A&AS, 81, 291 Sofue, Y. 1995, PASJ, 47, 551 Sofue, Y. & Handa, T. 1984, Nature, 310, 568 Tsuboi, M., Inoue, M., Handa, T., Tabard, H., Kato, T., Sofue, Y., & Kaifu, N. 1986, AJ, 92, 818 Tsuboi. M., Kobayashi, H., Ishiguro, M., & Muratd, Y. 1991, PASJ, 43, L27

22

C. L. Brogan et a].: 74 MHz Observations of GC

-24'

-26'

n

0 0 0

-28'

-

(u J W

C

.-0

U

0

-30'

C

-32'

- 34'

-36' 1 8h06m

1 7h54m

42m

SOm

Right Ascension (JZOOO) Fig. 1 VLA A, B, C, & D configuration image of the GC region at 74 MHz with 2.1' x 1.2'resolution. The rms noise is 0.1 Jy beam-' and the peak flux density is 18.5 Jy beam-'. This image has been corrected for primary beam attenuation and has been masked for I b 12 1.5" to emphasize the emission along the Galactic plane.

-

23

Amon. NachrJAN 324, No. S I (2003)

-27O30

00'

-g

-28O30

0 N 7 v

00' 0 ._

I

0

.-

-29O30

0 aJ

00'

-30'30

-31'00 17h52m

4Em 44m Right Ascension (J2000)

40m

Fig. 2 VLA A, B, C. & D configuration close up view of the GC region at 74 MHz with 2.1' x 1.2' resolution (see Fig. I ) . A number of the features mentioned in the text are labeled for reference.

- 2 7'30

- 2 7O30'

00

0

00'

- 28O30'

- 2 8'30

0 0

0 0 N 7

7 N v

v

00

C

0

00'

0

c 0

0

5aJ - 2 9'30

-29'30' 0 W

n

00'

00

-3 0'30

-7

-30'30'

il00O

-31°00' 44m

Right Ascension (J2000)

40m

a)

Right Ascension (J2000)

Fig. 3 a) VLA C, & D configuration image of the GC region at 330 MHz convolved to 2.1' x 1.2' resolution. b) Nobeyama image of the GC region at 10 GHr. with 3' resolution (Handa et al. 1989).

b)

C. L. Brogan et al.: 74 MHz Observations of GC

24

4

3

2

1

0

- 1

-2

GLON ( c i o g ~ e s s )

Fig. 4 A subimage of the 74 MHz image in Figure 1 in Galactic coordinates. The contours show the integrated CO emission (resolution g ) from Dame, Hartmann, & Thaddeus (2001) at 0.03, 0.1, 0.3, 0.5, 0.7, and 0.9 times the peak value of I757 K bean-’.

-

Astron. Nachr./AN 324, No. S 1,25 - 3 1 (2003)/ DO1 10. I002/asna.200385 1 16

Chandra view of the central 300 pc of our Galaxy

'

Q. Daniel Wang*

'

Astronomy Department, University of Massachusetts Ainherst

Key words ISM, Galactic Center, pulsars, X-ray, X-ray binaries, hot gas

-

Abstract. We have conducted a Chandra survey of a 0."7 x 2" swath along the Galactic central ridge. This survey has resulted in the first arcsecond-resolution X-ray panorama of high-energy objects and their interaction with the unique environmcnt of the center region. We review our ongoing sludies and follow-up observations as well as the results from the survey.

1 Introduction The Galactic center (GC) of the Milky Way is a mecca of high energy activities. The GC has experienced a mildly active starburst over the past several million years (e.g., Figer et al. 1999). The three young massive stellar clusters (Arches, Galactic center, and Quintuplet) alone, for example, contribute a stellar wind luminosity of order los9 ergs s-', comparable to that from the central cluster of the LMC 30 Doradus nebula - the most luminous HI1 region in the Local Group of galaxies (e.g., Wang 1999). Massive stars (-W2 8M0) are also short-lived and explode as supernovae at ages from a few to a few tens of million years, creating high-energy products: pulsars, accreting neutron stardblack holes, and supernova remnants (SNRs). Although the massive black hole at the dynamic center of the Galaxy is weakly active at present (Baganoff et al. 2001), it might have been many orders of magnitude brighter in the recent past (e.g., lo2 years ago). The AGN activity may be imprinted in scattered emission, which has been proposed to explain the unusually strong 6.4 keV Fe Ktr fluorescent line observed from the G C region (Koyama et al. 1996; Wang et al. 2002; Cramphorn & Sunyaev 2002). Therefore, the GC, only 8 kpc away, provides an excellent laboratory for understanding such high energy activities, which are suspected to be important both in galaxy evolution and for the thermal and chemical feedback from galaxies to the intergalactic medium. With its unprecedented spatial and spectral resolutions as well as the broad energy coverage, Chandra provides us with an unique opportunity to comprehend the complex G C region. As a Chandra cycle-2 large project, we have mapped out the Galactic central ridge with 30 overlapping ACIS observations (12 ksec exposure each; Fig. 1 ; Wang et al. 2002a). The resultant data, complemented by deeper pointing on several specific targets with the same instrument (e.g., Yusef-Zadeh el al. 2002; Takagi, Murakami, & Koyama 2002; Baganoff et al. 2001), represent the first high resolution X-ray view of the G C with sensitivity more than two orders of magnitude higher than any pre-Cbandra observations (e.g., ASCA). While the Chandra data analysis and interpretation are still ongoing, below we highlight progress made so far (see also Wang 2002a,b):

-

-

* Corresponding author: e-mail: wqd Qastro.umass.edu

D. W a g : Chandra View of the central 300 uc

26

~

~

~

~

~p

Fig. 1 Chandra X-ray intensity map of the Galactic central ridge in the 1-8 keV range. The plot is in the Galactic coordinates and is centered near Sgr A * .

2 Discrete source detection and identification

-

We detect N 103 discrete sources down to a 2-8 keV flux limit of 9 x c’rg spl ciri-’. Only a handful of X-ray sources are previously known. Our spectral analysis shows that about 30% of our detected sources are foreground objects (e.g., CVs and normal stars). An estimate based on millimeter and far-infrared intensity in the GC field further suggests that less than 10% of the remaining sources are extragalactic or located far beyond the GC. Therefore, the bulk of the sources belong to the GC. We have detected X-ray emission from all known young stellar clusters at the GC as well as relatively nearby ones that are embedded within emission nehulae Sh2-20 and Sh2-17 (Wang et al. 2002a; Wang 2002h and referenccs therein). In particular, much of the X-ray emission from the compact Arches cluster seems to be produced by colliding winds from individual massive close binaries and from cluster stars collectively - the so-called cluster wind (Raga et al. 2001; Yusef-Zadeh et al. 2002). Strong 6.4 keV line-bearing emission is detected from the vicinity of the cluster. An ongoing in-depth study of the spatial correlation between the 6.4 keV emission with a high-resolution CS molecular map we have obtained with the OVRO will allow for a detailed modeling of the important fluorescencelscattering processes. Furthermore, we noticed an extended feature to the northwest of the cluster, in a direction away from the Galactic plane (Fig. 2); most of the individual peaks in the image are not statistically significant. This apparently rim-brightened balloon-shaped feature is also apparent i n the XMM/Newton image (R. Warwick; private communications). While the exact nature of the feature is unknown, its morphology may suggest that it arises in the interaction of the cluster wind with the ambient medium. We have identified several of our Chandra sources as quiescent counterparts of previously luminous X-ray and gamma-ray sources, which include GRO 51744-28, GRS 174 1.9-2853, and SAX 51747.0-2853. The analysis of these sources have allowed us to place important constraints on the X-ray emission mechanisms of neutron stars in the poorly studied states of low to very low accretion rates (Wijnands & Wang 2002; Wijnands et al. 2002). The sub-arcsecond position accuracy of the Chandra sources now allow for identifications with both radio and infrared sources. A substantial number of the sources are probably neutron stars accreting from the winds of main-sequence stellar companions (mid-0 to late-B spectral types), and many more may be discovered with deeper X-ray observations (Pfahl et al. 2002). In collaboration with R. Bandyopadhyay (PI) et al., we have obtained VLT observing time for multi-color near-IR imaging of 30 selected candidates for such objects to test this scenario.

Astron. Nachr./AN 324, No. S 1 (2003)

27

47 A

0 0 0

N

3

v

1

I

55s

45s

I

nh45m

35s

RIGHT ASCENSION (52000) Fig. 2 Cliandla ACIS-I intensity map of the Arches cltisler and the vicinity in the 4-8 keV hand. Thc overlaid contours represenr the 6.4 keV line intensity distribution (Wang 2002a).

An estimate of the population of X-ray sources will potentially provide important constraints on thc star formation history oP the GC. The bulk of the sources in the GC is believed to be accreting compact binaries, which arc thc end-products of massive stars. This independent star formation tracer, though relatively untcstcd, is badly needed for the C C region. All observed young massive clusters have ages 5 5 x 10" yrs; older ones have presumably been evaporated by the strong gravitational tidal forcc in t h e GC (c.g., Kim et a]. 2002). There is little consensus o n the star formation rate of the G C on time scales of , . - lo7years. which are about the same as the dynamic ages of various large-scalc structures observed in the inner region of the Galaxy (Wang 2002b; Bland-Hawthorn & Cohen 2003). Therefore, it is particularly important to see whether or not there is a large population of high-mass (e.g., B-type) X-ray binaries, which may be used to constrain the recent star formation history in the GC region.

D. Wanr: Chandra View of the central 300 pc

28

03' 30"

-29O 05' 30"

i -1

,

I

40'

445

, 17h 45m 3€

RIGHT ASCENSION (JZOOO)

Fig. 3 a) Chandra ACIS-I image of GO. 13-0.1 I in the 2 - 6 keV band. The contours represent the smoothed distribution of 6.4 keV line emission (Wang 2002a). The rectangle and ellipse mark the regions from which the source and background spectra are extracted, which are presented in Fig. 4a. b) Chandra ACIS-I image of (3359.89-0.08in the 4.0-9.0 keV band, with 20 cm radio intensity contours overlaid. The gray-scale bar changes logarithmically from 0.29 to 2.9 cts arcsec-' (Lu et al. 2003).

3 Massive star end-products The GC is expected to host a large population of pulsars, both recently formed and recycled. However, no pulsar is yet known within 1 O of Sgr A*. The traditional method of finding pulsars at radio wavelengths in (or beyond) the GC becomes problematic, because of interstellar beam scattering and pulse broadening

Astron. Nachr./AN 324, No. SI (2003)

29

(Cordes & Lazio 1997). It is also not effective to conduct a blind search at high frequencies (v 2 10 GHz), where the scattering is not as important, due to the typically steep radio spectra of pulsars and smaller telescope beam sizes.

D

u x,

1

2 Chonnei energy (keV)

5

2

5 chonnel energy (keV)

Fig. 4 The ACIS-I spectra of a) G0.13-0.11 and b) G359.89-0.08, both of which are well described by power laws In a), the (bottom) spectrum of the embedded pulsar candidate is also included (Fig. 3a).

Fig. 5 A smoothed 2.0-9.0 kcV Chandra ACIS-I intensity map of the X-ray emission associated with G359.54+0.18, overlaid on the 6 cm radio continuum image fi-om YuseT-Zadeh et al. ( I 997).

We are pursuing a program first to find X-ray-emitting pulsar or PWN candidates, which are often associated with corresponding radio sources or features, and then to carry out a sensitive, high-frequency periodicity search of such targets on 100 m Efflesberg and Green Bank Telescopes, together with highfrequency VLA imaging observations (Lazio et al. 2003). Several X-ray candidates have been identified so far based on their possible associations with radio sources or features. The detection of pulsars would serve as useful probes of the magnetoionic medium and possibly the spacetime structure as well as the supcrnova rate and star formation history of the GC. In particular, we have detected and analy7ed various linear nonthermal X-ray threads (Yusef-Zadeh et al. 2002; Wang et al 2002b; Lu et al. 2003). The three brightest threads (Figs. 3-5) all seem t o bc associated with nonthermal radio filaments (NTFs) or radio “wisps”. Some of these associations may represent the hcst candidates for pulsar wind nebulae, possibly shaped by the strong magnetic field of

D. Wang: Chandra View of the central 300 pc

30

the GC. This PWN scenario in a strongly organized magnetic field environment, as indicated by radio polarization measurements, naturally explains the apparently nonthermal X-ray emission and the linear morphology of these features. However, we cannot yet rule out possibilities that some of the X-raykadio filament associations may be caused by locally enhanced inverse-Compton scattering of far-IR photons off relatively low-energy electrons (y lo2) or by interstellar magnetic field reconnection. Our study is still ongoing. Our objective is to firmly determine the nature of both X-ray threads and NTFs, which may then be used to trace the origin, propagation, and fate of ultra-relativistic particles as well as the strength and configuration of magnetic field in the GC. N

ooo

10'

00

-ooo

10'

15

00' 05' Galactic Logitude

-OOo 05'

Fig. 6 6.4 keV line intensity contours (Wang 2002a) overlaid on the 20 cm continuum image (Yusef-Zadeh et al. 1984) of the central region near Sgr A*.

4 Nature of diffuse X-ray emission Probably the greatest mystery of the Galactic X-ray astronomy is the hard Galactic ridge X-ray emission (GRXE; e.g., Worrall et al. 1982; Valinia & Marshall 1998; Tanaka 2002; Dogie1 et al. 2002). which

Astron. Nachr./AN 324, No. S1 (2003)

31

-

has a scale height of 100 pc and is greatly enhanced toward the GC. The overall spectrum of the ridge emission suggests a characteristic thermal temperature of 7-10 keV. Such diffuse hot gas, if not confined (say by magnetic field), would escape fast into the Galactic halo, demanding a very efficient heating sourcc that is not yet identified at the GC. The high tcinperature may not be required, however, if a large fraction of the hard GRXE i s nonthermal in origin (e.g., Valinia & Marshall 1998). Even with Chandra observations, we cannot yet draw any firm conclusion about the nature of the GRXE. The intensity fraction contributed by discrete sources varies strongly among individual observations. av10% (Wang et al. 2002a; Ebisawa et al. 2001). This variation may be partly due to the yet eraging uncertain non-cosmic X-ray background subtraction of these observations and to the differential line-ofsight absorption at energies smaller than a few keV. We find that unresolved “diffuse” X-ray emission clearly consists of multiple components: thermal emission characterized by distinct K(I lines from highly-ionized species (e.g.. SXV, ArXVII, and FeXXV), reflection marked by the 6.4 keV line (Fig. 6). and possible nonthermal bremsstrahlung radiation of lowenergy cosmic rays (Wang et al. 2002a). In addition, inverse-Compton scattering may also play a rule, although there is no detailed correlation of the hard X-ray emission with either radio or far-IR intensity in the field (e.g., Fig. 6). The thermal diffuse X-ray emission, in particular, clearly extends beyond the region covered in the existing mapping (Wang 2002b; not apparent i n the broadband image of Fig. I , though). A detailed modcling effort is under way to quantify the importance of these components in individual regions.

-

Acknowledgements I thank my research collahorators, especially E. Gotthelf, C. C. Lang, F.-J. Lu, and R. Wijnands, for their contributions to the project presented above.

References Baganoff, F., el al. 2001, Nature, 413, 45 Bland-Hawthorn,J., & Cohen, M. 2003, ApJ, 582, 246 Cordes, J. M., & Lazio, T. J. 1997, ApJ, 475, 557 Cramphorn, C. K.. & Sunyaev, R. A. 2002, A&A. 389,252 Dogiel, V., et al. 2002, ApJ, 581, 1061 Ebisawa, K., et al. 2001, Science. 293, 1633 Ghosh, P., & White, N. E. 2001, ApJL, 559.97 Figer, D., et al. 1999, ApJ. 525, 750 Koyama, K., et al. 1996, PASJ, 48, 249 Lazio, T. J. et al. 2003, AN, 324, 3 Lu, F., Wang, Q. D.. & Lang, C. C. 2003, AJ. submitted Pfahl, E., et al. 2002, ApJL, 571, 37 Tanaka, Y. 2002, A&A, 382, 1052 Valinia, A . & Marshall, F. E. 1998, ApJ, 505, 134 Wang, Q. D. 1999, ApJL, 5 10, 1 39 Wang. Q. D. 2002a. in New Vision of the X-ray Universe in the XMM-Newton and Chandra Era (astro-ph/0202317) Wang, Q. D. 2002h, in High Energy Processes and Phenomena in Astrophysics (astro-pMO211630) Wang, Q. D., Gotthelf, E., & Lang, C. 2002a, Nature, 415, 148 Wang Q. D., Lu, F., & Lang, C. C. 2002b. ApJ, 581, 1 148 Wijnands, R., & Wdng, Q. D., 2002, ApJL, 568,93 Wijnands, R., Miller, J. M., & Wang, Q. D. 2002, ApJ, 579, 422 Worrnll, D. M., et a]. 1982, ApJ, 255, 1 I I Yusef-Zadeh, F., et al. 2002, ApJ, 570, 665 Yusef-Zadeh, F., et al. 1987, ApJL, 475, 119 Yusef-Zadeh, F., et al. 1984, Nature, 3 10, 557

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1, 33-39 (2003) / DO1 10.1002/asna.200385074

Two Thousand X-ray Stars in the Central 20 pc of the Galaxy

',

M. P. Muno" F. K. Baganoff', M. W. Bautz', W. N. Brandt2, P. S. Broos2, E. D. Feigelson*, G. P. Garmire*, M. R. Morris3, G. R. Ricker', and L. K. Townsley2

' Center for Space Research, Massachusetts Institute of Technology, Cambridge, MA 02139

' Department of Astronomy and Astrophysics, The Pennsylvania State University, University Park, PA 16802 ' Department of Physics and Astronomy, University of California, Los Angela, CA 90095 Key words Galactic Center, X-ray catalog, stellar X-rays, X-ray binaries

Abstract. We present a catalog of 2357 point sources detected during 620 ks of Chandra observations of the 17 x 17 arcminutes field around Sgr A * . This field encompasses a physical area of 40 x 40 parsecs at 8 kpc. The completeness limit of the sample at the Galactic Center is N 10" ergs-' (2-8 keV), while the detection limit is an order of magnitude lower. We estimate that 20-100 of the observed sources are background AGN. The spectra of the majority of the Galactic Center sources are very hard and are best described by a power law (Fr) with photon index r < 1. Such hard spectra are only seen from magnetically accreting white dwarfs (polars and intermediate polars) and neutron stars (pulsars). The spatial density of the Galactic Center X-ray sources falls off as R p 2 ,similar to that of the infrared stars in the field. This highlights the possibility that these X-ray sources can be used to study the history of star formation at the Galactic Center. The log(N) - log(S) distribution of the Galactic Center sources is extremely steep, such that point sources could account for all of the previously rcported diffuse emission if the number counts extend down to erg s-l with the same slope. However, there are numerous filamentary structures in the field that also contribute to the total flux, which implies that the luminosity distribution between 2-8 keV must turn over below our completeness limit.

1 Introduction The X-ray emission from galaxies is produccd by a mixture of stellar sources at various phases of their life-cycles, diffuse interstellar plasma heated by supernovae and galactic collisions, and accretion onto super-massive black holes in galactic nuclei (see Fabbiano 1989). With its 0."5 angular resolution, the Chundru X-ruy Obsewutory is particularly well-suited to separating the diffuse and point-like components of this emission. CIzandru observations can therefore be used to estimate more accurately the amount of hot, X-ray emitting interstellar matter in galaxies, and to trace the histories of star formation in galaxies using their stellar X-ray populations. Over the past three ycars, we have observed the 17' x 17' field around Sgr A* for over 620 ks using thc Advanced CCD Imaging Spectrometer imaging array (ACIS-I) aboard Chundm. The resulting image (Figure 1) makes it possible for the first time lo distinguish the X-ray emission from the accrcting supermassive black hole Sgr A* from the surrounding early-type stars, the remnant of a lo5* erg explosion (Sgr A East; Maeda et al. 2002), and numerous filamentary features (Baganoff et al. 2001 ; Baganoff et al. 2003; SCC also Morris et al., Park et al., and Maeda et al. in these proceedings). In this paper, we describe the nature of the point-like sources of X-rays (Section 2), and discuss their implications for the history of star formation at the Galactic Center (Section 3) and the origin of the "hot" (10 keV) component of the diffuse Galactic ridge X-ray emission (Section 4). * Corresponding author: e-mail: munoQmit.edu. Phone: + I 6172533429, Fax: + I 6172530861

@ 2003 WILEY-VCH Vcrlilg GmhH Ri Co KG.2h. Wsmhrm

34

M. P. Muno et al. Galactic Center Chandru Sources

Fig. 1 Three-color image of the inner 8.5' by 8.5' field around Sgr A*, created using photons between 2.0-3.3 (red), 3.3-4.7 (green), and 4.7-8.0 keV (blue). North is up, and east to the left. The image has been corrected for variations in exposure due to bad columns and chip gaps and has been adaptively smoothed to allow point sources and diffuse emission to be distinguished more easily. The color scale is logarithmic. The lack of sources to the east of the Galactic Center is due to the presence of the molecular clouds M-0.02 - 0.07 and M-0.13 - 0.08 (Gusten, Walmsley, & Pauls 1981) and confusion with the diffuse emission from Sgr A East. See Park et al. in these proceedings for an annotated version of this image.

2 The Sample of Point Sources Using standard techniques (e.g., Brand1 et al. 2001; Feigelson et al. 2002) we have identified 2357 individual point sources in the full 17' x 17' ACIS-I field around Sgr A* (see Muno et al. 2003 for the full catalog). Thanks to the high spatial resolution and good detector response above 2 keV, this sample is more than two orders of magnitude larger than the 17 sources that were observed in this field by previous X-ray missions (Watson et al. 1981; Pavlinsky, Gebenev, & Sunyaev 1994; Predehl & Truemper 1994; Sidoli et al. 1999; Sidoli, Belloni, & Mereghctti 2001; Sakano et al. 2002). The sources in this image are detected with fluxes between 3 x and 2 x erg cmp2 s-l (2.0-8.0 keV). At the Galactic Center distance of 8 kpc, the luminosities of these sources range from 3 x 10"" to 2 x erg s-l, if one accounts for an omp2 of absorption toward the Galactic Center. Of average decrease in flux of 30% due to the > 6 x

Astron. Nachr./AN 324, No. S 1 (2003)

35

Table 1 Galactic X-ray Point Sources

log( L y )" Spectru m0 lOg(C'Kg . - ' ) MS Stars' 25 - 30 :I AT < 1 keV Plasma 29 31.1 YSOS kT = 1 - 10 keV Plasma RS CVdAlgol 2!) :31 7 kT = 0.1 2 keV Pla5ma 31 kT = 0.1 - 6 keV Plasma WR/O Stars 20.5 - 32 (1 k T = 1 25 keV Plasma cvs Pulsars 29 3 - :K) r = 1 2.5 PL ; kT = 0 . 3 keV BB NS LMXBs 31.6 - 38 kT 0 3 keV BB ; r = 1 2 PL'' 30 - :
Object

~

~

~

:rl

~

~

~

N

~

~

r

r

-

these sources, 28 I can be identified as foreground sources, because they are detected below I .5 keV. We cstimate that only 20-1 00 ofthe 2076 sources only detected above 2 keV arc background AGN, based upon the log(N) - log(S) distribution of sources in deep Chandra observations at high Galactic latitudes (e.g., Brandt et al. 2001) and accounting conscrvatively Tor the >30% reduction in the flux from background sources due to Galactic absorption. A wide range of Galactic sources emit X-rays with luminosities greater than 103' erg 5 - l (see Table 1). Therefore, in order to understand which sources are present in the field, we examine the properties of their X-ray spectra. I n Figure 2, we plot two X-ray colors Tor all of the sources that are detected with > 90% confidence in three energy bands above 2 keV. along with the colors expected for an absorbed power-law spectrum. Simulations indicate that the thermal models expected for the coronal sources and for most CVs all produce hard colors less than 0.1. This includes ionized plasmas with k T < 25 keV, blackbodies with kT < 2 keV, or Bremsstrahlung emission with kT < 50 keV (see Table 1). Although these softer sourccs should be the most numerous in our field, only a small fraction are expected to be bright enough above 2 keV to observe through the Galactic absorbing column of 6 x 10" cm-' of H. On the other hand, over half of the Galactic Center sources cluster i n a region consistent with absorption columns l o g ( N ~> ) 22.5, and very flat spectra with photon indices r < 1 (where negative values indicate rising numbers of photons with energy). Such hard spectra are unusual for X-ray point sources. but have been seen previously from cataclysmic variables (CVs) containing magnetized white dwarfs (polars and intermediate polars; e.g., Ezuka & Ishida 1999) and from magnetized neutron stars accreting from the winds of massive companions (High-Mass X-ray Binary [HMXB] pulsars; e.g., Campana et al. 2001). Supporting this hypothesis, 8 of the brightest 285 sources exhibit coherent X-ray pulsations with periods ranging from 300 s to 4.5 hours (Muno et al. 2003b, in preparation).

3 Spatial Distribution and the Star Formation History Within about 8' from the aim point, we estimate that we can detect all sources with photon fluxes greater than 5 x l o p 7 photons cmp2 s-* with a signal-to-noise of at least 71, 3 i n the 2.0-8.0 keV band. About 40%' of the Galactic Center sources detected within 8' of Sgr A* have photon fluxes greater than this value. In Figure 3 we plot the number of Galactic Center sources above this flux limit per unit solid angle as a 1

M. P. Muno et al. Galactic Center Chandru Sources

36

-0.5L 10 -1.0

L

L

I

m

-0.5

,

I

I

+ I

,

,

1

,

0.0

I

0.5

.

I

,

,

1 .0

Medium Color

Fig. 2 Comparison of the observed hard and medium colors to those expected from an absorbed power-law spectrum for sources detected with greater thdn 90% confidence in all three energy bands above 2 keV. The colors are defined as the fractional difference between the count rates in two energy bands, ( h s ) / ( h + s), where h and s are the numbers of counts in the higher and lower energy bands, respectively. The medium color is defined using counts with energies between 3.34.7 keV and 2.0-3.3 keV, and the hard color using counts between 4.7-8.0 keV and 3.34.7 keV. Open circles denote data from 39 foreground sources and filled circles from 785 sources at the Galactic Center. The crosses connected with solid lines indicate the expected colors for absorbed power laws, with the photon indices and absorption columns indicated on the plot. The median uncertainty for these sources is displayed at the bottom of the plot. We note that the sources in the upper-left comer of this Figure all have uncertainties a factor of 2-3 larger. ~

function of angular separation from Sgr A*. We have fit this surface density distribution with a power law of the form

where H is the angular separation from Sgr A* in arcminutes. Both the normalization and the power-law of 26 for 30 degrees of freedom. slope were allowed to vary. The resulting fit is acceptable, with a If we assume that these sources are distributed with spherical symmetry about the Galactic Center, the implied spatial density falls off with radius as R-2. The spatial density of stellar sources observed in the infrared also decreases approximately as R-a.o*o.3within 30 pc of Sgr A* (Serabyn & Morris 1996). This suggests that the X-ray sources lie primarily in the Nuclear Stellar Clustcr (Mezger, Duschl, & Zylka 1996; Launhardt et a1.2002) and that their spatial distribution traces that of infrared stars. This suggests that the stellar X-ray sources can be used to understand the star formation history in the inner tens of parsecs of the Galaxy, where it is uncertain how the large tidal forces and the milliGauss magnetic fields affect star formation, and where many traditional observational tracers of star formation have been difficult to find (Mezger et al. 1996; Morris 1993; Serabyn & Morris 1996). For instance, if there are magnetic CVs among the hard X-ray sourccs, they would trace low-mass stars in the nuclear bulge (Warner 1995). On the other hand, if some of them are wind-accrcting neutron stars, they would provide an important constraint on the amount of massive stars formed near the Galactic Center in the last lo7 - loy years (Pfahl, Rappaport, & Podsiadlowski 2002).

xz

Astron. Nachr./AN 324. No. S 1 (2003)

60

"

'

I

"

"

'

'

'

676 sources with F

in

OF

I

0

,

I

1

,

2

,

,

>

50x10.'

,

.

~~

I

4

37

1

"

'

ph cm-' s-'

, +

6

a

Offset (arcmin)

Fig. 3 Surface density of Galactic Center poinl sources as a function of offset angle (8) from Sgr A*. The number of sources in each annulus was divided by the solid angle over which a source could he detected above 5 x lor7 photons cmr2 srl with a signal-to-noise of 3 in that annulus. The dotted linc indicates the best-fit power law, 2: cx Or' "*".').

3.1

Fig. 4 Cumulative log(N) - Iog(S) distribution of sources at the Galactic Center (filled circles,

2-8 keV) and in the foreground (open circles, 0.5-8 keV). The best-fit models determined by a maximum-likelihood method are over-plotted with solid lines. The expected extra-galactic contribution from Brandt et al. (2001) is indicated with the dashed line (see also Rosati et al. 2002).

Point Source Contribution to the Diffuse X-ray Emission

A hot (lo8 K) component of the ISM is thought to be responsible for the He-like Fe 6.7 keV emission that is observed all along the Galactic ridge. However, the temperature of this putative diffuse plasma is much higher than that typically produced in supernova shocks, and it is too high for the plasma to he gravitationally hound to the Galactic disk (Worrall et al. 1Y82; Koyama et al. 1986). If the plasma is unbound, the power required to sustain this hard Galactic ridge emission is approximately 10" erg spl, equivalent to the kinetic energy of one supernova occurring every 30 ycars (e.g., Valinia & Marshall 1998). This input would have to be provided by exotic processes, such as interactions between lowenergy (10 keV) cosmic-rays and the ISM (Valinia & Marshall 1998; Tanaka, Miyaji, & Hasinger 1999) or magnetic reconnection driven by turbulence in the ISM (Tanuma et al. 1999). However, there are currently no independent means of observing either low-cncrgy (10 keV) cosmic rays or magnetic reconnection (but sce Serabyn & Morris 1994 for the latter), so it is important to estimate the energy in the hot plasma as accurately as possible. The point sources in Figure 1 could contribute significantly to the Galactic Ridge X-ray emission, and t h u s lessen the energetic requirements on thc hot plasma. In Figure 4, we plot the cumulative number counts as a function o f flux for sources at the Galactic Center (filled circles, fluxes between 2-8 keV) and in the roreground (open circles, 0.5-8 keV), normalized to the solid angle of the survey 12 in units of arcmin-2. We have been conservative in our source selection to avoid incompleteness in our sample caused by the varying sensitivity over our image (see Muno et al. 2003 for complete details). Note that only about one-third of hoth the foreground and Galactic Center sources detected in our image satisfy our selection criteria, since thc high background in the image adds significant uncertainty to our flux measurements We focus on the distribution of Galactic Center sources, since these are most relevant to understanding the origin of the hard component of the diffuse X-ray emission. Using the un-binned flux valucs, we modeled the log AT log S distributions using the maximum likelihood technique described in Murdoch, Crawford. & Jauncey (1975). The distribution of Galactic Center sources were consistent with two power ~

M. P. Muno et al. Galactic Center Chandrcl Sources

38

laws of the form

where S is in units of photons cm-' s-', and the normalization is in units of sources arcinin-'. We estimate the total flux produced by these point sources by integrating the flux convolvcd with Equation 2, converting the photon fluxes into energy fluxes by assuming l photons cmp2 s-l = 8 x lo-" erg cm-2 s-l (2.0-8 keV). We find that point sources with fluxes greater than 3 x 10W" erg cm-' s-' contribute a mean surface brightness of 4 x erg cm-' 5-l arcmin-' over the inner 9' around Sgr A*. This is about 10%)of the diffuse flux from the inner regions of the Galaxy derived by Koyama et al. (1996; erg cmP2 s-* arcmin-') and by Sidoli & Mereghetti (2001; erg cm-2 s-' in a I " field, or 3 x 1x erg cm-' s-' in a 190 arcmin-' field, or 5 x lo-'" erg cm-2 s-' arcmin'). A similar result was obtained by Ebisawa et al. (2001) in a rcgion at I = 28" and h = 0.2". However, the steep slope of the flux distribution ( S - ' 7 , below 6 x lo-'' erg cm-2 s-', implies that lhe integrated flux from point sources in the field will diverge if the distribution extends to arbitrarily low fluxes (see Figure 4). Point sources would account for all of the diffuse emission reported by Koyama et al. (1996) and Sidoli & Mereghetti (1999) if the distribution in Equation 2 extends a factor of 40-100 lower in flux. However, from the image in Figure 1, it is clear that filamentary features contribute a significant liaction of the diffuse emission, which implies that the flux distribution in the 2-8 keV band (where most erg cm-' of the diffuse emission is observed) must turn over between fluxes of 3 x 10-'7 and 3 x s-l, or luminosities o f 2 x loz9 to 2 x 10"' erg s-' at the Galactic Center. We have also compared the spectra of the point sources and the diffuse emission, in order to constrain the relative contribution of each to the 6.7 keV iron emission from the Galactic Center. The equivalent widths of the apparently diffuse 6.7 keV emission range from 350-500 eV, even in regions that are otherwise relatively free from filamentary features (see Park et al. 2003, in these proceedings). The iron emission in the combined spectrum of the Galactic Center point sources is at the low end of this range, about 330 eV. Thus, it appears that unresolved point sources are unlikely to account for all of the 10 keV component of the diffuse Galactic X-ray emission (Muno et al. 2003c, in preparation).

4

Conclusions

We have detected 2357 X-ray point sources during 620 ks of Chundru observations of the 17' x 17' field around Sgr A* (Figure I ) . The completeness limit of our survey at the Galactic Center is about 3 x erg cm12 ~ ~ ' ( 2 -keV), 8 while sources are detected with fluxes nearly an order of magnitude lower. Only 20-100 of these sources are expected to be background AGN. The large number of sources in this field probably results from the high stellar density at the Galactic Center. Indeed, we have demonstrated that the surface density of Galactic Center X-ray sources decreases as N l/O away from Sgr A* (Figure 3). just as the surface density of infrared stars does. We have also shown that the l o g ( N )- log(S) distribution of the Galactic Center sources is very steep, rising as S-1.7near our completeness limit (Figure 4). This indicates that unresolved point sources can contribute significantly to the diffuse component of the Galactic X-ray emission. More than half of the sources for which we have spectral information are very hard, with spectra that are consistent with I' < 1 power laws (Figure 2). Such hard spectra have only been observed previously from magnetically accreting white dwarfs and wind-accreting neutron stars. If they are magnetic CVs among these X-ray sources, they would he the first low-mass stars identified in the nuclear bulge. If they are wind-accreting neutron stars, these systems would provide an important constraint on the amount of star formation that has taken place near the Galactic Center in the last lo7 - lo8 years. This highlights the importance of identifying the nature of the Galactic Center sources with more certainty. The X-ray spectral

Astron. Nachr./AN 324, No. S1 (2003)

39

and timing properties of these sources will be reported in detail in the near future, and w e are in the process of identifying these sources at radio and infrared wavelengths.

Acknowledgements We gratefully thank D. Schwartz, P. S h e , and the Churidru Mission Planning group for their efforts in scheduling thc observations in late May so that they would have nearly identical roll angles and aim points. We also thank K. Getman and F. Batter for developing software that aided us in collating and visualizing these results, V. te Velde for helping to construct a map of the diffuse background, P. Schechter for valuable advice on how to treat the log(N) - log(S) distribution, and E. Pfahl and J. Sokoloski for helpful discussions about the possible natures of these sources. This work has been supported by NASA grants NAS 8-39073 and NAS 8-00128. W.N.B. alho acknowledges the LTSA grant NAG 5-81 07 and the Alfrcd P. Sloan foundation.

References Baganoff, F. K.. Bautz, M. W., Brandt, W. N.. Chartas, G., Feigclson, E. F., Garmire, G. P., Maeda, Y., Morris, M., Ricker. G. R., Townsley, L. K., &Walter, F. 2001. Nature, 413, 45 Baganoff, F. K. et al. 2003, ApJ, 591. 891 Brandt, W. N. et al. 2001, AJ, 122, 2810 Campana, S., Gastaldello, F., Stella, L., Israel, G. L., Colpi, M., Pizzolato. F., Orlandini, M., & Dal Fiume. D. 2001, ApJ, 561.924 Ebkawa, K., Maedd, Y., Kaneda, H.,& Yamauchi. S. 2001 b, Science, 293, I633 Ezuka, H. & Ishida, M. 1999, ApJS, 120, 277 Fabbiano, G. 1989, ARA&A, 27, 87 Feigelson, E. D., Broos, P., Gaffney, J. A. 111, Garinire, G . , Hillenbrand. L. A,, Pravdo, S. H., Townsley, L., & Tsuboi, Y. 2002, ApJ, 574, 258 Giistcn, R., Walmsley, C. M., & Pauls, T. 19x1, A&A, 103, 197 Koyama, K., Maeda, Y., Sonobe, T., Takeshima, T., Taneka, Y., & Yamauchi, S. 1996, PASJ, 48. 249 Koyaina, K., Makishinia, K., Tanaka, Y., & Tsunemi, H. 1986, PASJ, 38, 121 Launhardt, R., Zylka, R., & MeLger. P. G. 2002, A&A, 384, I12 rf Maeda, Y. et al. 2002, ApJ, 570, 67 I Mewe, R., Lemen, J. R., & van den Oord, G. H. J . 1986, A&AS. 65.5 I I Zylka, R., 1996, AARev, 7. 289 Muno, M. P., Baganoff, F. K.. Bautz. M. W.. Ricker, G. R., Monis, M., Garinire, G. P., Feigeison, E. D., Brandt, W. N., Townsely, L. K., & Broos, P. S. 2003a, ApJ, 589, 225 Murdoch, H. S., Crawford, D. F., & Jauncey, D. 1973. ApJ, 183, 1 Pavlinsky, M. N., Grebenev, S. A.. & Sunyaev, R. A. 1994, ApJ, 425, 110 Pfahl, E., Rappaport, S., & Podsiadlowski, P. 2002, ApJ, 571, L37 Predehl, P. & Truemper, J. 1994, A&A, 290, L29 Raymond, J. C. & Smith, B. W. 1977, ApJS, 35,419 Rosati, P. et al. 2002, ApJ, 566, 667 Sakano, M., Koyama, K.. Murakami, H., Maeda. Y.. & Yamauchi, S. 2002, ApJS, 138, 19 Serabyn, E. & Morris, M. 1994, ApJ, 424, L91 Serahyn, E. & Morris, M. 1996, Nature, 382,602 Sidoli, L., Belloni, T., & Mereghetti, S . 2001. A&A, 368, 835 Sidoli, L. & Mereghetti, S. 1999, A&A, 34’). L49 Sidoli, L., Mereghetti, S., Israel, G. L., Chiappctti, L., Treves, A., & Orlandini, M. 1999, ApJ, 525, 215 Tanuma, S., Yokoyama, T., Kudoh, T., Matsumoto. R., Shihata, K., & Makishima, K. 1999, PASJ, 51, I61 Tanaka, Y., Miyaji, T., & Hasinger, G. 1999. Astron. Nachr., 320, 181 Valinia, A. & Marshall, F. E. 1998, ApJ, 505, 134 Wdmer, B. 1995, Curaclv,rmic Variable .Star.\, Cambridge University Press Watson. M. G., Willingale, R., Hertz, P., & Grindlay, J . E. 1981, ApJ, 250, 142 Worrall, D. M., Marshall, F. E., Boldt, E. A , , & Swank, J. H. 1982, ApJ, 255, 1 I I

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S I , 41-46 (2003)/ DO1 10.1002/asna.2003851IS

Magnetic field in the Galactic Centre: Rotation Measure observations of extragalactic sources Subhashis Roy * I , A. Pramesh Rao

I,

and Ravi Subrahmanyan2

' National Centre for Radio Astrophysics (TIFR), Pune University Campus, Post Bag No.3, Ganeshkhind, Pune 41 1 007, India.

' Australia Telescope National Facility, CSIRO, Locked bag 194, Narrabri, NSW 2390, Austrd~ia Abstraet. We have estimated the Faraday Rotation Measure (RM) towards 45 extragalactic sources seen through the central - 6 O < 1 < 6", - 2 O < h < 2" region of the Galaxy using the ATCA and the VLA. This is the first systematic study of RM i n the central kpc of the Galaxy and has lead to the detection of a large scale magnetic field in the region. The magnetic field does not undergo any reversal of sign across the rotation axis of the Galaxy, which is consistent with the bisymmetric spiral model for the magnetic field in our Galaxy. If this component prevails over the central 2 kpc region of the Galaxy, and the average electron density in the region is taken to he 0.4 cm-:', then the line of sight component of this regular magnetic field is estimated to he 0.7 p G . In addition to this large scale field, there are also small scale fluctuations, whose coherence scale length is estimated to he -20 pc. If the scattering material is unilbrmly distributed in the GC. then we estimate the random component of the field to have a strength of 10 pG. On the other hand. if the scattering medium is clumpy, the random component will have similar strength as the regular component. Constraining the filling factor of [he ionised medium is required for estimating the random component of the magnetic field from the R M data.

1 Introduction Magnetic fields can be strong enough to have a significant influence on the dynamics and evolution in the central region of our Galaxy. Magnetic pressure can contribute significantly to the overall pressure balance of the interstellar medium (ISM) and can even influence the distribution of gas (Beck et al. 1996). Strong systematic magnetic fields in the Galactic Centre (GC) region are believed to be responsible for the creation and maintenance of the unique non-thermal filaments (NTFs) (Morris & Serabyn 1996, and references thercin). Measurement of the field geometry and its strength in the central part of the Galaxy is important for estimating its effect on the G C ISM and discrete objects in the GC. There are no estimates of the magnetic ficlds in the inner 5 kpc region of the Galaxy (Davidson 1996) except in the central 200 pc region which are based mainly on observations of non-thermal filaments (Yusef-Zadeh & Morris 1987; Anantharatnaiah et al. 1991; Gray et al. 1995; Yusef-Zadeh et al. 1997). The R M estimated along these NTFs are -1 000 rad m-2 (Yusef-Zadeh & Morris 1987; Gray et al. 1995; Yusef-Zadeh et al. 1997). From the pressure balance argument, (Yusef-Zadeh & Morris 1987) derived a magnetic field strengths of about 1 mG in these NTFs. Previous studies of NTFs have shown that the field lines are oriented along the length of the NTFs. Since, all the NTFs found within a degree from the GC arc oriented almost perpendicular to the Galactic plane, it suggests that the field lines in the surrounding ISM are also perpendicular to the Galactic plane (Morris & Serabyn 1996). However, in the NTF Pelican (G358.85+0.47) located beyond a degree from the GC, the field lines are almost parallel to the Galactic plane. This may indicate that the field lines change their orientation beyond one degree from the G C and become parallel to the disk, as usually seen in the rest of the Galaxy. However, it should be noted that if the * Suhhashis Roy: e-mail: [email protected] 02001 WILEY-VCH

Verlap GmhH C To K(iaA. W a n h a m

S. Rov et al.: Magnetic field in the Galactic Centre

42

NTFs are manifestations of a favourable local environment (Shore & Larosa 1999) (much higher magnetic field), inferences drawn from these observations can be misleading. While Zeeman splitting of spectral lines directly measure the magnetic field, this method is sensitive to high magnetic fields. Therefore, estimates of mG magnetic field based on Zeeman splitting (Killeen et al. 1992; Yusef-Zadeh ct al. 1999) of the HI or OH lines towards the GC could be atypical, being from local enhancement of field (e.g., near the core of high density molecular clouds) To estimate any systematic magnetic field in the region, it is necessary to use an observational method, which is sensitive to the large scale field. The Faraday Rotation Measure, which is the integrated line of sight (los) magnetic field weighted by the electron density (RM=0.81 x Jn,B,ldl, where n, is the electron density, Bl1 is the 10s component of the magnetic field, and the integration is carried out along the los), is one such method. If a model for the electron density is available, such observations towards the background extragalactic sources are well suited to estimate the 10s magnetic fields prevailing in the ISM near the GC. However, no systematic studies of the Rotation Measures (RM) towards the extragalactic sources seen through the GC region have been carried out in the past. We have carried out a survey of Rh4 towards 65 suspected extragalactic sources seen through the central -6"< 1 < 6", -2"< b <2" region of the Galaxy. The survey has provided 45 polarised sources (63 components) for which RM has been estimated. These observations were designed to estimate a RM as high as 15,000rad m p 2 without any nrr ambiguity. To be able to observe with large enough bandwidth without significant bandwidth depolarisation, these observations have been carried out at higher radio frequencies using the C and X band of the ATCA and the VLA. The coherence scale length of the magnetoionic medium near the NTFs has been estimated to be =lo" (Gray et al. 1995; Yusef-Zadeh et al. 1997). Therefore, to avoid any beam depolarisation introduced by the ISM of the Galaxy, these observations have been made using higher resolution array configurations of these telescope ensuring that the synthesised beam sizes are considerably smaller than the coherence scale length ofthe Faraday screen near the GC. The RM towards the source have been estimated by using the formula: RM=(+$2)/(X? - A;), where, $1 and &I are the polarisation angles at wavelengths XI and X2 respectively.

2 Sample selection When this project was undertaken, not many extragalactic sources were known in the GC region. Therefore, we selected candidate extragalactic sources from VLA Galactic plane survey (Zoonematkermani et al. 1990; Helfand et al. 1992; Becker et al. 1994) on the basis of their small scale structure (angular size 5 10") and non-thermal spectra ( N 5 -0.4, where, it denotes the spectral index and is defined as F(v) c( P).The spectral indices have been determined from the estimated flux densities of the sources in at least two of the three frequency bands of 325 MHz, 1.4 GHz and 5 GHz. The VLA GC map at 330 MHz (Larosa et al. 2000), Galactic plane survey at frequencies of 1.4 GHz (Zoonematkermani et al. 1990; Helfand et al. I992), 5 GHz (Becker et al. 1994), sources observed by (Lazio & Cordes 1998a), and the Texas survey at 365 MHz (Douglas et al. 1996) have been used for this purpose.

3 Results From the ATCA observations, twenty four sources were found to have at least one polarised component. At 5 GHz, the resolution of the ATCA maps are ~ 6 x" 2", and the RMS noise (Stokes I) in the maps are typically 0.2 mJylbeam. From VLA observations, twenty one sources were found to have at least one polarised component. At 5 GHz, the resolution of the VLA maps are -2" x 1.5", and the RMS noise in the maps are typically 75 pJy/beam. Maps of two of the sources are presented in Fig. 1 and 2.

Astron. Nachr./AN 324, No. SI (2003)

3. I

43

Contribution of the intrinsic Faraday Rotation to our observations:

Compared to the ISM of our Galaxy, the clectron density of the intergalactic medium is very small, and consequently the Faraday Rotation introduced by i t is negligible. However, if the synchrotron clectrons in the source are mixed with thermal plasma. or if the emission is seen through (i) the ISM of the parent galaxy, or (ii) some other galaxy along our l o s , it can introduce additional RM to the data. Since, the source is located far away from us, the cmission from different parts of the source will be observed within one synthesised beam. Here, differential Faraday Rotation through different parts of the intervening thermal plasma will give rise to beam depolarisation (Kronberg et al. 1972). Since. differential Faraday Rotation increases at lower frequency, any intrinsic Faraday Rotation is likely to cause a decrease of the polarisation fraction at lower frequency (Kronberg et al. 1972). From the ratio of the polarisation fraction at 8.5 and 4.8 GHz in images made with the same resolution, we computed the depolarisation ratio of the source components. If any of the component is found to have a depolarisation ratio of more than I .5, the RM of that component has not been used any further in this paper. 3 sources have been rejected following this criterion.

3.2 Features in the RM sky near thc GC: The estimated RM towards 45 sources are plotted in Fig. 3. In the plot, the positive RMs have been indicated by 'cross (X)' and the negative RMs by 'circle (O)', where the size of the symbols increases linearly with the InAII. Clearly, the region is dominated by positive RM, as estimated towards most of the sources in both positive and negative galactic longitude. The observed RMs towards the sources located within 1.5" from the Galactic plane are quite high -1000 rad ni-2, and do not appear to decrease along the Galactic plane, even several degrees away from the GC. Therefore, to estimate this large scale RM, we calculate the median value of the RMs towards all these sources, which is estimated to be 467537 rad m-? We note that the median RM estimated towards thc sources with positive galactic longitude is 5791131 rad m-2, whereas the median RM towards the sources in negative galactic longitude is 45 1 i l l 0 rad m - - 2 . To detect small scale fluctuations in the RM, we have plotted rhe RM structure function in Fig. 4 & 5. As can be seen from the plots, the RM structure function tend to zero as the separation of the polarised components tend to zero. This signifies an interstellar origin (plasma turbulence) of the RM, and indicates that contribution of the intrinsic RM of the sources must be small. The structure function tends to saturate for components separated by more than 0.3". From the plot, a lower limit of 0.05" (8 pc) and an upper limit -0.3" (40 pc) are estimated for the coherence scale of the Faraday screen.

4 4.1

Discussion Geometry of the magnetic field along the Galactic plane

Our observations show a large scale positive bias i n the RM, ( its median value being -5 times highcr than its uncertainty) which suggests that the 10,s field is mostly pointed towards us in [his region. The absence of any reversal of sign of the RM (4 0 result) across the rotation axis of the Galaxy (I=O"), indicates that thcre is no reversal of magnetic field across the rotation axis of our Galaxy. This rules out configurations ot magnetic field, which are symmetric around thc rotation axis of the Galaxy. Two models, the axisymmetric and the bisymmetric spiral model ofthe magnetic fields, are generally invoked to explain the field geometry in galaxies (Beck et al. 1996). It is thought that current loop generated by rotation of the mattcr around the centre (dynamo model) gives rise to the axisymmetric spiral configuration of the magnetic field. However, if the magnetic fields at the present time is the result oT concentration of a primordial magnetic field due to collapse of matter while the Galaxy was formed, then it gives rise to the bisymmetric spiral configuration of the magnetic field in galaxies. In our Galaxy, the magnetic field configuration has been suspected to be bisymmetric (Simard-Normandin & Kronherg 1980), in which case, the field does not undergo a reversal

S. Roy et al.: Magnetic field in the Galactic Centre

44

G358.1-2 IPOL 4760 100 MHZ 0358.1-2.OHGEO 1 -31 23 00 05

g

10

0

3

15

$

20

5

25

P

30 35 40 17 47 50

49

40 47 RIGHT ASCENSION (JZOOO)

46

45

POI line 1 arcsec = 2.5OWE-03 JYBEAM Peak flux = 1.35568-02 JYIBEAM

Levr~3.000E-04'(-2,-1.1,2,4,6,8,12,16, 20, 24, 32, 40, 40. 64.96,128, 160,192,224, 256, 320)

Fig. 2 4.8 GHz continuum map of the source G358.1-2 made from the VLA data

Fig. 1 4.8 GHz continuum map of the source G356.7-1 made from the ATCA data. I

1

x

i

X

X

C~

0

,

I

;

-

.

W

.:

1'.V

X

I

O

0

2

.'

0

I

;

,

I

,

I

XX X0

" x

6-

N

I :

x x $ xX A :

l - 0

I

I

" 2 I

of sign across the centre. Our observations are consistent with the bisymmetric spiral configuration in the inner parts of our Galaxy. 4.2 Estimation of the magnetic field from the RM:

In order to estimate the 10s component of the magnetic field from the RM, a model of the electron density distribution is required. Therefore, we first discuss the available models of the electron density distribution near the CC, and then estimate the magnetic field from our data. Taylor & Cordes (1993) have made a model of electron density in our Galaxy. However, their model does not include the electron density in the GC region. Therefore, we consider alternate models of the

45

Astron. Nachr./AN 324, No. S1 (2003)

Fig. 4 Plot of the RM structure function with angw lar separation less than a degree

Fig. 5 Same as the tigure at left, but with angular separation less than 0.3".

-

electron density distribution in the GC region.Using a likelihood analysis, Lazio & Cordes (199%) showed that there is a 'hyperstrong' scattering screen ( r i , 10 cm-") with an angular extent of about half a degree along our l o s towards the GC. They estimate a distance of 13:3?$? pc for the screen from the GC. Bower et al. (2001) have carried out VLBA obscrvations of 3 extragalactic sources, and report an intermediate scattering region covering 2 S o i n longitude and 5 5" in latitude. The estimated scattering size of the 3 extragalactic sources are almost 2 orders of magnitude less than what is predicted by the 'hyperstrong' scattering and about 1.5-6 times more than the prediction by the Taylor & Cordes ( 1993) model. Following Bower et al. (2001), we believe that other than the GC scattering screen of size O S " , there is a region of enhanced scattering covering 25" in Galactic longitude and 5 5" in Galactic latitude. However, since the scattering size of a source is a function of the electron density and the clumpincss of the medium, in the absence of further constraints, estimating the electron density of this component is difficult. Since the scattering angle measured by Bower et al. (2001) is 1.5-6 times higher than the prediction from Taylor & Cordes (1993), and since except one, all other sources in our sample are seen outside the 'hyperstrong' scattering region, we have taken an electron density 0.4 cm-:', which is twice of the Taylor & Cordes (1993) model for the inner Galaxy. The corresponding dispersion measure due to the inner 2 kpc of the Galaxy is 800 pc cm-'3. The estimated los averaged magnetic field strength is 0.71rG from our RM data. To compute the random component of the magnetic field, we note that the typical scale size of the Faraday screen as estimated in $3, is -20 pc. Hence, if the scattering medium is uniformly distributed, then there will be about lo2 cells of irregularities within the central 2 kpc of the Galaxy. Assuming that the RM contributions from the individual cells along the los are random, the cell contributions would have an RMS lUvl value of 67 rad m-'. If the RMS electron density within such a cell is 0.4 cmP3, the estimated magnetic field is about 10 pG, which is more than an order of magnitude higher than its 10s average value. On the other hand, if the scattering material is distributed in lumps (like the outer envelopes of thc HI1 regions), then there can be - 1 cells along line of sight, and the corresponding random component of the magnetic field will have strengths similar to the regular component. Therefore, unless the filling factor of the ionised medium is known, the random component of the magnetic field cannot be estimated from the RM data.

References Anantharamaiah. K. R., Pedlar, A,, Ekers, R. D., & Goss, W. M. 1991, MNRAS, 249, 262 Beck, R., Brandenburg, A., Moss, D., Shukurov, A,, & Sokoloff, D. 1996, ARA&A, 34, 155 Becker, R. H., White, R. L.. Hclfand, D. J., & Zooneniatkermani, S. 1994, ApJS, 91, 347 Bower, G. C.. Backer, D. C., & Sramek, R. A. 2001, ApJ, 558, 127

46

S. Roy et al.: Magnetic field i n the Galactic Centre Davidson, J. A. 1996, in ASP Conf. Ser. 97: Polarimetry of the Interstellar Medium, 504 Douglas, J. N., Bash, F. N., Bozyan, F. A., Torrence. G. W.. & Wolfe, C. 1996, VizieR Online Data Catalog, 8042. 0 Gray, A . D., Nicholls, J., Ekers, R. D., & Cram, L. E. 1995, ApJ, 448, 164 Helfand, D. J., Zoonematkermani, S., Becker, R. H., &White, R. L. 1992, ApJS, 80,21 I Killeen, N. E. B . , Lo, K. Y., & Crutcher, R. 1992, ApJ, 385,585 Kronberg, P. P., Conway, R. G., & Gilbert, J. A. 1972, MNRAS, 156,275 Lang, C. C., Moms, M., & Echevarria, L. 1999, ApJ, 526,727 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hvman, S. D. 2000. AJ, 119,207 Lazio, T. J. W. & Cordcs, J. M. 1998a, ApJS, 118,>0l -. 199% ApJ, 505,715 Moms,M. &-Serahyn, E. 1996, ARA&A, 34,645 Saikia, D. J. & Salter, C. J. 1988, ARAkA, 26, 93 Shore, S. N. & Larosa, T. N. 1999, ApJ, 521,587 Simard-Normandin, M. & Kronherg, P. P. 1980, ApJ, 242, 74 Taylor, J. H. & Cordes, J. M. 1993, ApJ, 41 1, 674 Yusef-Zadeh, F. & Morris, M. 1987, ApJ, 322,721 Ynsef-Zadeh, E, Roberts, D. A., Goss, W. M., Frail, D. A,, & Green, A. J. 1999, ApJ, 512, 230 Yusef-Zadeh, F., Wardle, M., & Parastaran, P. 1997, ApJL, 475, L119 Zoonematkermani, S., Helfand, D. J., Becker, R. H., White, R. L., & Perley, R. A. 1990, ApJS, 74, 18 I

Astron. Nachr./AN 324. No. S 1.47- 5 1 (2003)/ DO1 IO.1002/asna.20038S024

Study of the Nuclear Bulge region of the Galaxy K. S. Baliyan’ I, S. Ganesh’, U.C. Joshi’, 1.S. Glass’, and T. Nagata’

’ Physical Research Laboratory, Ahmedabad-380 009, India ’ South African Astronomical Observatory, South Africa ’ Nagoya University, Nagoya, Japan

Key words Near Infrared, Milky Way Galaxy, Nuclear Bulge, Photometry, Extinction. PACS 04A25 A near infrared survey of the inner 3OOpc of Nuclear Bulge region of the Milky Way is being carried out as a core program of the SIRIUS Camera mounted on the IRSF telescope at the South African Astronomical Observatory, Sutherland. The SIRIUS camera has three I K x 1K detectors for simultaneous imaging in the J, H and Ks bands. With pixel scale of 0.45” and good seeing most of the time, these observations present the deepest views of a large area of the Nuclear Bulge. The aim of this survey is to overcome the incompleteness and confusion limited nature of the undersampled near infrared surveys, such as DENIS and 2MASS in order to get better estimate of the extinction in these lines of sight and distinguish between

various stellar populations. Preliminary results from the survey demonstrate the capability of the camera and photometric procedures for the crowded fields.

1 Introduction The Galactic Center region of our Galaxy provides the best opportunity to study the astrophysical processes i n galactic nuclei with high spatial resolution. The study of the inner regions of the Galaxy is also very important for understanding the structure, dynamics, kinematics & energetics of the Milky Way, as well as galactic evolution and star formation processes (Serabyn & Morris 1996, Mezgcr, Duschl & Zylka 1996, Ellis 2001, Launhardt et al. 2002). Due to the large extinction, studies of the Galactic structure based on optical observations have been restricted to high Galactic latitudes. However, the longer wavelenglhs observations, including near infrared, can penetrate the relatively high extinction at low Galactic latitudes towards the nuclear bulge. In the recent past there have been several NIR surveys, i.e. DENIS (Epchtein et al. 1997), 2MASS (Skrutskie et al. 1997), etc, which either lack depth or suffer from confusion due to high source densities in the Galactic Center Region. Recently Schultheis et al. (1999) prepared a map of the interstellar extinction towards the inner Galactic Bulge using DENIS data and reported Av> 25 with a clumpy, inhomogeneous nature. However, the J band data i n DENIS is undersampled in this region of high extinction. A large number of Ks sources do not have counterparts in 1 & J in DENIS. The situation has not improved much with the availability oE 2MASS data. To overcome these problems, and to gain a better understanding of the distribution of stellar populations in the inner bulge region, we are carrying out a deep imaging survey of this region i n J, H bz Ks bands with particular emphasis on the fields covered by the ISOGAL survey at 7 p n and 15 p m discussed here by Baliyan et al. (2003) and the X-ray survey by thc Chandra observatory. The deep imaging survey in J, H & Ks bands was carried out using IRSF telescope of the South African Astronomical Observatory at Sutherland during June-July 2002. These observations have been planned so as to reach the tip of the RGB at the distance of the Galactic center. Another key ingredient for studying the spatial structure of * Corresponding author: e-mail:

baliyanQprl.res.in. Phone: +9l 796302 129, Fax: +91 796301 502

02003 WILEY-VCH VrrLtg Grnblj L Co

KGaA. Wrinhwm

K . S. Baliyan et al.: Near Infrared Survey of the Nuclear Bulge

48

the Galactic bulge is deriving accurate distances to the stars in these regions. Such information will help identify foreground and background sources. In this paper here we present fust results from this near infrared survey underlining the capabilities of the SIRIUS camera and the photometric procedures adopted in this work.

2 Thesurvey Our survey of the inner 300 pc of the bulge region within [ I ( = 1.5 deg, (b( = 0.5 deg is carried out using the IRSF (InfraRed Survey Facility) 1.4 m telescope at Sutherland, South Africa. The IRSF is jointly operated by the SAAO, South Africa and Nagoya University, Japan. The telescope is equipped with a three channel camera (Nagashima et al. 1999) known as SIRIUS (Simultaneous InfraRed Imager for Unbiased Survey), capable of imaging in J ( I .25 pm), H ( I .63 p m ) & Ks (2.14 pm) bands, simultaneously with 1024x1024 pixels array detectors. The FOV is 7.8’x7.8’ with a scale of 0.45 arcseclpixel. The region covered in the present survey is shown in Fig. 1 where each square rcpresents the center of a mosaic of 3x3 images. Righl ascension 1:

3

46:OO

44:OO I

I

I

30:OO

-29:oo:oo 2 L

-C .0

0

30:OO

-3O:OO:OO

,

I

Fig, 1 Image of the Inner Milky Way showing the surveyed region.

3 Observations We performed imaging observations for the inner Galactic Bulge at near infrared wavelengths during June 25-July 1 & July 9-15, 2002 using the InfraRed Survey Facility (IRSF) at SAAO. The observations were conducted under good seeing (better than 1.3 arcsec) conditions most of the time with airmass values between 1.0 and 1.5. For each field, 10 dithered images were taken (one at the center and 9 on a circle around it) with 0.1 and 5 secs exposures. Towards some directions we have also taken 10 second exposures

Astron. Nachr./AN 324, No. S 1 (2003)

49

to reach fainter levels of completeness in the highly extincted J band. Observations were also made for

sky, dark, twilight flat fields and on a set of standard stars (Persson et al. 1998) for data reduction and calibration. Observations for sky were taken towards a dark nebula not far from the object of interest. Some fields were also observed towards low extinction regions symmetrically placed perpendicular to the galactic plane to serve as reference.

4 Data Reduction and Analysis The images were reduced using the IRAF package along with scripts written in-house. Standard procedures of near-infrared array image reduction are adopted, including dark subtraction, sky subtraction and Rat fielding. The SIRIUS Pipeline (Nakajima, private communication) was used for combining the 10 dithered images into a single image to obtain a higher signal to noise ratio for each position. Since the Nuclear Bulge is highly crowded, special care has to be taken while doing photometry. We noticed that using IRAF’s DAOPHOT package for source detection and photometry resulted in a large number of residual sources. We, therefore, performed source detections and photometry using Christophe Alard’s (Alard 2000) software which applies PSF fitting to the image after dividing it into several regions. Thus it uses space varying kernels. This procedure results in far less number of residuals than what is obtained wirh IRAF/Daophot. The limiting magnitudes in J, H and Ks are 17.2, 17 & 16, respectively. For the purpose of this paper, the measured fluxes were calibrated, for preliminary results, using the 2MASS photometry for corresponding obiects.

5 Sample Results & Discussions Here we show some sample results obtained from the 8’ x 8’ colour image (Fig. 2) of the Galactic Center region. The image is coniposcd of J (blue), H (green) and Ks (red) band images. The emission at near infrared wavelengths is much less extincted as compared to optical and is mainly from the photosphere of late spectral type, evolved stars. The central bright star cluster is clearly visible in the image. The image itself indicates the non-uniform, highly varying nature of the extinction. It is easy to see the filamentary and clumpy, inhomogenous distribution of molecular material. The J band shows maximum extinction in the near infrared, decreasing towards longer wavelengths. The sources seen in this band are either blue foreground sources or very bright stars o f the central stellar cluster. The H and Ks band images show overcrowding of the sources in this region. Some morphological details of these molecular clouds is also apparent. The source extraction results i n more than 7500 sources in Ks. The magnitude values used in these preliminary results are calibrated with 2MASS data as mentioned. The Ks versus (J-Ks) colormagnitude (CMD) and (J-H) versus (H-Ks) color-color diagrams are plotted in Figs. 3 and 4, respectively, to distinguish between stellar populations. Only the stars detected in all three bands are used in these plots. The sharp cut off in the Ks versus J-Ks CMD is due to J band limitations. The CMD also shows foreground bright stars (to the Icft) and a small number of highly reddened sources towards left. I t is also possible that there are some massive high luminosity bluer, young stars hidden by the large population of red giants (which are easily detectable in NIR). The color-color diagram shows that stars arc distributed in a much wider range than one finds in less extincted regions. A large number of thc sources right to the main concentration indicates to thc large amount of circumstellar material. The region, therefore, harbours sources belonging to various populations. 11 is to be noted that our survey is deeper by more than 2-magnitudes when compared to 2MASS & DENIS i n the corresponding bands. We are therefore reporting measurements of a large number of sources (fainter than Ks o f 14 mag) for the first time. We also imaged a field lying at higher latitude (I=0.0, b=I.O) with much lower extinction (Av 6, see for example Omont et al. 1999) Tor refercncc purposes. Towards this field, with 50 secs (10 images of 5 secs exposure average combined) integration, we detect more than 18000, 13000 & 12000 sources in J,

-

K. S. Baliyan et al.: Near Infrared Survey of the Nuclear Bulge

50

Fig. 2 J, H, Ks colour composite image ~ ( 8 ’ x 8 ’of ) the Galactic Center Region centered at GC

H & Ks bands with more than 9000 detections common to ail three bands. The CMD of this region (Fig 5 ) shows well defined giant branch, red clump and foreground sources in contrast to the central region where a large scatter is observed. Also notice that the CMD of GC rcgion shows stars distributed in a much wider range in colour due to varying and non-uniform extinction across the field compared to that in this reference field. The information on the distribution of various classes of the objects is very useful to get a handle for estimating the extinction towards the Galactic Center and distances to the bulk of the sources in these lines of sight.

6 Conclusions The present work reports on a deep near infrared survey of the inner 300 pc of the nuclear bulge region of the Milky way using SIRIUS camera mounted at the I .4 M lRSF telescope of the SAAO. The preliminary results show unprecedented view of the Galactic Bulge region in the J, H and Ks hands, deeper by more than two magnitudes as compared to DENIS and 2MASS. Therefore, we report the measurement of a large

Astron. Nachr./AN 324, No. SI (2003)

O

Z

I

8

.

3

J-Ka

Fig. 3 Ks versus J-Ks CMD showing sharp cut off on the lower right due to high extinction at J.

51

0

1 H-K,

2

3

Fig. 4 J-H versus H-Ks colorcolor diagram for the Galactic Center Region

Fig. 5 Ks versus J-Ks diagram for the reference field (P = 0, b = 1).

number of sources for the first time. Color-magnitude and color-color diagrams are plotted for sources i n

the sampled region. Acknowledgements This work was funded by the Department of Space, Govt. of India. The observation time at IRSF facility awarded by SAAO is gratefully acknowledged. KSB & SG are thankful to SAAO for hospitality at Cape Town and to Observatory staff for help. We also express our thanks to C. Alard for his codes, M. Schultheis and A. Omont for useful discussions and the 2MASS & DENIS teams for the respective data. ISOGAL provided the background image of the inner galaxy used in the figure 1. Partial funding by DST enabled KSB to attend the GC-2002 conference in Hawaii.

References Alard, C. 2000, A&AS, 144,363 Baliyan, K.S., et al. 2003, these proceedings Ellis, R.S. 2001, PASP 113, 515 Epchtein, N., et al. 1997, Messenger 87, 27 Launhdrdt, R., Zylka, R., Mezger, P.G. 2002, AXrA, 384, 112 Mezger, P.G., Duschl, W.J., Zylka, R. 1996, A&AR, 7, 289 Nagashima, C., et al. 1999, in, Srur Formarion 1999. Ed. T. Nakamoto, Nagama: Nobeyama radio Observatory), 397 Omont, A., et al. 1999, A&A, 348,755 Persson, S.E.,Murphy, P.C., Krzeminski, W., Roth, M. & Rieke, M.J. 1998, AJ, 116, 2475 Serabyn, E. & Morris, M. 1996, Nature, 382, 602 Skrutskie M.F. et a1 1997, in The impact oflurge .scale Near IR sky survry, Eds. Garzon F, Epchtein N B Omont A, K1uwer:Netherlands. 117

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 53 -57 (2003) / DO1 10.1002/asna.200385025

A morphological Study of the Galactic Inner Bulge Kiran S. Baliyan*', Shashikiran Ganesh', Umesh C. Joshi', Ian S. Glass', Mark R. Morris3, Alain Omont4, Mathias Schultheid, and Guy Simon5 I

' '

Physical Research Laboratory, Navarangpura, Ahmedabad, India South African Astronomical Observatory, Cape Town, South Africa Department of Astronomy, UCLA, USA Institut d' Astrophysique de Paris, Paris, France Observatoire de Paris, Paris, France

Key words Galactic Inner Bulge, mid infrared, stars, molecular clouds PACS 04A2.5

A sizable region of the Inner Galaxy was observed during the ISOGAL survey using ISOCAM narrow band imaging at 7 and 15 pm with 3 arcsec resolution. Due to very low, though non-negligible, extinction at mid-infrared wavelengths, this window is very suitable for exploring the morphological features in the inner Galaxy. The images from this survey display spectacular mid-IR emission and absorption features and highly crowded star fields. The starforming regions of Sgr B1 and C stand-out at mid-IR wavelengths. Also very obvious are quiescent molecular clouds which are seen as dark regions in the mid-infrared images. The diffuse mid-IR emission is from ionixd gas, warm dust (mainly at 15 pm) and PAH (7 pm). The emission from point sources is mainly due to circumstellar dust, but at 7 pm stellar photospheres also contribute to it. As a first step to understand this complex ensemble of sources, we compare these images with previously known information at other wavelengths particularly the VLA 90 cm map. A good correlation between the mid-IR starforming regions and their thermal counterparts in the 90 cm VLA map is seen.

1 Introduction Attempts at understanding the inner Milky Way remain as a forefront research area with several good reviews on the subject of the Galactic Inner Bulge and the nuclear regions are such as by Mezger et al. (1996),Morris & Serabyn (1996) and Launhardt et al. (2002). The MSX survey (Price et al. 1997, Egan et al. 1998) of the Galactic Plane covered the entire plane of the Galaxy. The ISOGAL Survey has revealed a new view of the Inner Milky Way at mid-infrared wavelengths (Omont et al. 2003). This survey at 7 and 15 p m used the ISOCAM camera onboard the Infrared Space Observatory. With a pixel-field of view of 6" (3" in some of the more crowded regions) and sensitivity better than IOmJy, the survey resulted i n a view that is better by two orders of magnitude in spatial resolution and sensitivity as compared to IRAS at similar wavelengths. With its smaller pixel sizes, the ISOGAL images are of a superior resolution compared to MSX. A first look at the results in the disk of the Milky Way was presented by PCrault et al. ( 1996). Subsequently Glass et al. (1999) and Omont et al. (1999) provided a good insight of the relatively outer regions of the Bulge of the Galaxy at mid-infrared wavelengths. They standardized the procedures for the analysis of mid-IR data from the survey (in combination with near-infrared ground based observations) in thc presence of low interstellar extinction (see also Ojha et al. 2003). Many of the stars detected by ISOGAL are late M-type giants on the asymptotic giant branch(AGB) in the Galactic Bulge and central disk but the most numerous class of sources detected at 7 p m are red giants with luminosities just above or close to the RGB tip with weak mass-loss rates. * Corresponding author: c-mail:

[email protected]

K. S. Baliyan et al.: A morphological Study of the Galactic Inner Bulge

54

Interstellar extinction poses a major difficulty at optical wavelengths in studies of the inner regions of the Galaxy. The very high and non-homogeneous extinction in these regions (see for example Catchpole et al. 1990, Schultheis et al. 1999) is significant at near infrared wavelengths and non-negligible even at mid-infrared wavelengths along some lines of sight. ISOGAL observations have been used by Hennebelle et al. (2001) to provide constraints on the interstellar extinction curve at mid infrared wavelengths towards the Inner Bulge. Indeed the ISOGAL images present the deepest and clearest views (over a large area) of the Nuclear regions of the Galaxy till date. Version 1.0 of the five wavelength ISOGAL-DENIS point source catalog is now published (Schuller et al. 2003) with discussion of the scientific results (Omont el al. 2003). In this work we present some of the ISOGAL survey images and compare them with already known information at other wavelengths. The images exhibit a good correlation between locations of the dark clouds in mid-IR and continuum emission at millimeter wavelengths. These mid-IR images also confirm the asymmetric distribution of the stellar sources, dust and gas in this region.

2 Observations Observations of the Galactic Nuclear Bulge regions (0.80' > f! > - 1.2", -0.4" < b < 0.25" were made with the 60cm telescope and the camera ISOCAM (Cesarsky et al. 1996), on hoard the Infrared Space Observatory under the ISOGAL program. While most of the ISOGAL observations employed 6" pixels, in the dense regions of the Nuclear Bulge 3" pixels were used. Narrow band filters, LW5 at 7 p m and LW9 at 15 pm, were used for the observations discussed here. Further details of the observations and the data reduction techniques are discussed elsewhere (Schuller et al. 2003, Ganesh et al. in prep.)

3 Discussion of the images

'grB2

I

Sgr B l

G0.25+0.02

GC

Fig. 1 A mosaic of the 7 pm ISOGAL images superimposed VLA 90 cm contours. This figure appears in colour in the electronic version

Figure 1 shows the mid infrared view of the Inner Milky Way constructed from the ISOGAL observations at 7 pm. Very evident are spectacular emission features including Sgr B and Sgr C. Superimposed on the emission features are also regions marked by strong absorption at 7 pm. One well studied (see for example Lis & Carlstrom, 1994) giant molecular cloud G0.25+0.02 is marked in Figure 1 where it

Astron. Nachr./AN 324. No. S1 (2003)

55

appears as a prominent dark absorption feature. Apart from diffuse emission, we also see largc number of stars resolved for the first time by the superior dctcctors+telescope combination of the [SO. Most of these were unresolved by IRAS and some also by MSX. We note that with relatively large pixel size, some of the bright ISOGAL sources could also be unresolved compact clusters. For example, the Arches Cluster would be covered by a single 6” pixel of ISOCAM. However, many of such young objects appear as slightly extended sources (Schuller 2002). It is clear that the number density of stars on both sides of the Galactic Centre is non-uniform. Towards the negative longitudes we have larger number of stars as compared to that at positive longitudes. This asymmetry is perhaps due to the higher extinction towards the positive longitudes where there are relatively larger number of dark globules and patches indicating giant molecular clouds. Superimposed on 7 p i image in Figure 1 are contours from the VLA obscrvations at 90 cm (LaRosa et al. 2000). The correspondence of the strong emission at 90 cm with extended 7 pm emission is seen very clearly lor the Sgr B 1 starforming region and also the Sgr C region. Towards Sgr B2 however, the 7 pm mosaic shows a lot of absorption by intervening molecular cloud layers as discussed above.

Fig. 2 An enlarged view of the Sgr B region at 7 pi from ISOGAL superimposed with VLA 90 cm contours. Orientation is in Galactic Coordinate system for all ligures.

Since the VLA is sensitive to both the thermal and non-thermal emission, interesting correspondences are seen with mid-IR features. Several ridge-like extended emission features appear correlated with VLA

56

K. S. Baliyan et al.: A morphological Study of the Galactic Inner Bulge

flux contours. Above the Sgr Bl & B2 complex, towards positive latitudes, a dark cloud, with low 90 cm flux, appears to he relatively quiescent with little, if any, star formation activity.

Fig. 3 Magnified view of the Sgr C region from figure 1

Among the extended emission features, the star forming regions, Sgr B1 (see Figure 2) and Sgr C (Figure 3), stand-out at mid-IR wavelengths and exhibit interesting correspondences with structures at 90 cm seen in the images obtained by the VLA. In the interest of keeping the figures clear and to display the large number of mid-infrared point sources in proper contrast, we do not overlay astrometric grids on these figures. In the case of Sgr B 1, a computational artifact in the VLA contours runs across the molecular cloud but the correspondence between the contours and the 7 pm extended emission features is clearly

Astron. Nachr./AN 324, No. S 1 (2003)

57

seen. High extinction, even in the mid-infrared, apparently hides from view the strong VLA source Sgr B2. There are a fcw mid-infrared poinl sourccs scattered around the actual peak of radio emission. In Figure 3, some 7 p m sources appear aligned along the non-thermal filament(NTF) associated with the Sgr C. Whether there are any correspondences between mid-IR sources and radio NTFs is however, still an open question, as the apparent alignment could be purely coincidental. Further discussion of these images including comparison with molecular data will be published elsewhere (Ganesh et al. 2003). An electronic-version-only figure' shows a color image constructed by mapping mosaic of 7 pin to green and 15 pm to red. The result shows colors of the point sources and the extended emission. At 15 pm, there is a ridge-like feature extending Crom (I,b =0.05, -0.1) to (l,b=0.22,-1.0), touching b=-0.18 midway, which is almost absent in the 7 pni image. It appears like a compressedlshocked shell perhaps driven by nearby supernova remnant. This feature seems to be in the same line of sight as the base of the radio arched filament. MSX images show this feature as a super bubble, perhaps powered by the Quintuplet cluster. Acknowledgements This work is supported by the Indo-French Centre for the Promotion of Advanced Research under project 1910-1. Financial support from Department of Space & Department of Science & Tech, Govt. of India made it possible to present this work at the GC-2002 conference at Hawaii. The VLA 90 cm data were obtained from the web pages of Lazio (http://rsd-www.nrl.navy.mil/72 I3/lazio/GCatlaa/)

References Catchpote, R. M., Whitelock, P.A., Glass, I.S. 1990, MNRAS, 247,479 Cesarsky, C.J., et al. 1996, A&A, 315, L32 Egan, M.P., Shipman, R.F., Price, S.D., Carey, S.J., Clark, F.O., Cohen, M. 1998, ApJ, 494, 199 Ganesh, S., et al. 2003, A&A, in preparation Glass, I.S., et al. 1999, MNRAS, 308, I27 LaRosa, T.N., Kassim, N.E., Lazio, J.W., Hyman, S.D. 2000, AJ, 119, 207 Launhardt, R., Zylka, R., Mezger, P.G. 2002. A&A, 384, 1 12 Lis, D. C. & Carlstrom, J.E. 1994, ApJ, 424, 189 Mezger, P.C., Duschl, W.J., Zylka, R. 1996, A&AR, 7,289 Moms, M.R., & Serabyn, E. 1996, ARA&A, 34,645 Ojha, D.K., et al. 2003, A&A, 403, 141 Omont, A,, et al. 1999, A&A, 348,755 Omont, A,, et al. 2003, A&A, 403, 975 PCrault M., et al. 1996, A&A 315, L165 Price, S.D., et al. 1997, IAU Symp. 179, 1 IS Schuller, F. 2002, PhD Thesis (University of Paris). Schuller, F., et al. 2003, A&A, 403,955 Schulthcis, M., et al. 1999, A&A, 349, L69

' BaliyanK.2.eonly.jpg

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 59-63 (2003) / DO1 10.1002/asna.200385026

Warm molecular gas, dust and ionized gas in the 500 central pc of the Galaxy N.J. Rodriguez-Fernandez * I , J. Martin-Pintado2, A. Fuente3, and T.L. Wilson4 ' LERMA, Ohservatoire de Paris,61, Av de I'Observatoire, 75014 Paris, France

' Instituto de Estmcturd de la Materia, CSIC, Serrano 121, 28006, Madrid, Spain

' Observatorio Astron6mico Nacional. LGN, Apdo. 1143, 28800 Alcali de Henares, Spain Max-Planck-Institutfur Radioastronomie,Auf dem Huge1 69,53121 Bonn, Germany

Key words ISM: lines - ISM: Infrarcd - Galaxy: center

PACS 04A25

We present infrared and millimeter ohservations of molecular gas, dust and ionized gas towards a samplc of clouds distributed along the 500 central pc of the Galaxy. The clouds were selected to investigate the physical state, in particular the high gas temperatures, of the Galactic center region clouds located far from far-infrared of thermal radio continuum sources. We have found that there is ionized gas associated with the molecular gas. The ionizing radiation is hard 35; 000 K) but diluted due to the inhomogeneity of the medium. We estimate that- 30% of the warm molecular gas observed in the Galactic center region clouds is heated by ultra-violet radiation in photo-dissociationregions.

+

1 Introduction The interstellar medium in the central SO0 pc of the Galaxy (hereafter Galactic Center, GC) is mainly molecular gas. The molecular clouds in the GC exhibit an extended gas component with high temperature ( IS0 K). On the contrary, the dust temperature is lower than 30 K. The large line widths of the molecular lines, the high gas phase abundance of molecules linked to the dust chemistry, and the difference between gas and dust temperature suggest that some kind of shocks could he responsible for thc high gas temperalures of the molecular gas (Wilson et al. 1982, Martin-Pintado et al. 2001). The possible influence of radiation in the heating of the molecular gas is usually ruled out due to the lack of far infrared and thcrma1 radio continuum sources in the G C others than the well known H I I regions associated with the Sgr complexes (A-E) or ionized nebulae like the Sickle To investigate the heating of the molecular clouds in the GC we have studied a sample of 18 clouds located all along this region. The clouds were selected as molecular peaks located far from far infrared or radio continuum sources. Thosc sourccs were observed with the spectrornetcrs on board the lnfrural Spncc~ Observatory (ISO) and with the IRAM 30-in telescope. The data obtained with IS0 are ohservations of the lowest H2 pure-rotational lines, dust continuum spectra from 40 to 190 pm and a number of fine-structure lines from neutral atoms or ions (e.g. 0 I 6.1 pm, C II IS8 p m , Ne II 12 p m , 0 III 52 pm). With the IRAM 30-m antenna we have observed C"0, " C 0 , H 3 S a and H41 a. I n this paper we will review the results already published by Rodriguez-Fernandea et al. (2001a, 2001b. hereafter RFOla, RFOIb, respectively) and present some new results that will b e extended elsewhcrc (Rodriguez-Fcrnandczet al. 2003, RF03).

-

* e-mail: nemesio.rodriguezQobspm.fr, Phone: +33 140512061, Fax: +33 140512002 @ 2003 WILEY-VCH Vsiiag tinihH B Co. KGdA, Wcinliciin

60

N. J. Rodriguez-Fernhdezet al.: Warm molecular gas. dust and ionized gas in the 500 central uc of the Galaxv

2 Warm molecular gas Before ISO, the warm gas component of the GC molecular clouds was mainly studied by means of NH3 observations. The multilevel study of Huttemeister et al. (1993) showed that the temperature structure of the GC clouds can be characterized with two gas components at different temperatures: a cold gas component with a temperature close to that of the dust ( N 20 K) and a warm gas component whose temperature ranges from 100 to 250 K. However, since the abundance of ammonia is known to vary significantly is difficult to estimate the column density of warm gas in the GC clouds. We have, for the first time, obtained the column density of warm gas in the GC clouds by observing Hz pure rotational lines (RFOla). Columns 3 and 4 of Table 1 show the total column density and the temperature of the warm gas, respectively. We have also estimated the HZ density and the total column density of molecular gas in these clouds by observing 13C0 and C"0 and doing aradiative transfer analysis for kinetic temperatures between 15 and 200 K. The results are shown in columns 1 and 2 of Table 1. The fraction of warm H? to the total H2 column density as traced by CO varies from source to source but is 30% on average. As discussed in RF0 1a, it is difficult to explain the large column density of warm gas in the GC clouds. Several low velocity (- 10 Km s-l) C-shocks, photo-dissociation regions (PDRs) or both should be present in the line of sight. Comparing the energy of the turbulent motions in the GC clouds with the cooling by H2 (which at the moderate density of the GC clouds is comparable to that by CO), one finds that dissipation of supersonic turbulence could account for the heating of the warm Hz. However, with the available data is not possible to rule out heating in PDRs, indeed the observed temperature gradient in the GC clouds can be appropriately reproduced in a context of a PDR (RFOI a).

3 Dust temperatures The dust continuum emission peaks at wavelengths of 100 to 80 pm for all but two sources, whose spectra peak at SO-60 pm. The dust luminosity in the observed range is listed in column 5 of Table 1. It is not possible to fit the spectra with just one grey body. Thus we have used a model with two grey bodies like that described in section 2 of Goicoechea et al. (2003). Figure 1 shows the data and the grey bodies fits for two sources. To explain the emission at long wavelengths it is needed a dust component with a temperature of 15 to 18 K. The temperature of the warmer component varies from source to source from 26 to 39 K. Due to the uncertainties in the dust emissivity it is not easy to determine a total column density of dust (on the contrary, temperatures are almost independent on the dust emissivity). Nevertheless, the column density of dust with temperatures higher than SO K is less than 500 times lower than the column of dust at 15-35 K.

50

1 00 150 lung onda (micros)

200

Fig. 1 Left panel: Dust continuum emission towards M+0.21-0.12 (black solid triangles) and grey-body fit (green) with two temperature components ar 39 (red) and 16 K (blue).Right panel: same for M+0.76-0.0.5 with a 24 K (red) and 1.5 K (blue) components.

61

Astron. Nachr./AN 324, No. S I (2003)

Table 1 Physical parameters derived from the observations. See texi [or explanation. Number in parentheses are errors of the last significant digit. Typical errors of TI and Tz are 5 and 1 K, respectively.

M-0.96+0.13

3.5-4

0.6-1.1

I.lO(9)

157(6)

2.9

30

15

547.8

52.5

...

M-0.55-0.05

3.8-4.4

4.3-6.0

2.7(3)

135(5)

12.3

31

16

547.5

57.1

...

M-0.50-0.03

3.4-3.7

2.4-3.0

2.3(2l

135(4)

10.5

34

17

548.0

52.9

...

M-0.42+0.01

4-4.5

2.1-3.4

1.03(8)

167(6)

9.8

30

16

548.0

530

...

M-0.32-0.19

>3

1.1-2.2

1.03(51

188(5)

7.7

35

16

547.6

4.4

34

M-0.15-0.07

3.7-4.1

6.6-8.4

2.6(4)

136(6)

...

._.

._.

547.6

...

...

M4.16-0.10

3.8-4.2

3.7-4.9

1 17(13l

157(7)

9.9

32

17

547.3

9.4

35 35

Mc0.21-0.12

>3

0.8-1.5

0.64(7)

186(13)

13.5

39

16

547.6

15

M+0.24+0.02

3.5-3.9

4.8-7.1

1.736)

163(2)

16.1

30

16

547.4

2.5

M+0.35-0.06

3.8-4.8

1.7-2.7

O.66(5)

1931I )

14.5

36

18

548.1

520

...

M+0.48+0.03

3.8-4.2

3.2-3.6

1.03(9)

174(7)

9.8

26

14

547.5

516

...

M+0.58-0.13

3.6-3.9

3.1-3.9

1.32)

149(5)

7.2

27

15

541.5

52.5

..

Mc0.76-0.05

2.8-4.2

6.6-8.6

1.77(8)

lXl(4)

8.1

24

15

547.4

II

...

M+0.83-0.10

3-4.5

4.8-6.5

l.59(6)

178(5)

6.3

26

15

548.0

586

...

M4.94-0.36

>3

1.3-2.9

0.95(10)

146(7)

...

...

...

548.0

...

...

M+1.56-0.30

3-4.5

2-0.5

omin)

26000)

1.5

27

17

548.0

5140

...

M+2.99-0.06

3-4

1.0-2.1

I .40(9)

152(3)

1.7

29

18

548.0

527

...

Mc3.06+0.34

3-3.8

1.6-1.2

0.21(8)

250(20)

1.1

34

I8

548.0

5140

...

4 Radio recombination lines To study the possible presence of ionized gas associated with the GC molecular clouds we have observed the H41a and H35a recombination lines using the IRAM 30m antenna (RF03). However, we have no1 detected any of the lines in any of the sources. From our 3 a limits to the line fluxes we have determined upper limits to the flux of Lyman continuum photons in the 30m beam (column 8 of Table I ) . The comparison of thosc limits with stellar atmospheres models rules out the presence of stars of an earlier type than B0.5 and effective temperatures higher than 32,000 K. N

5

Fine structure lines

We have detected lines of atoms and low excitational potential (< 13.6 eV) ions like 0 I 63 pm, C II I58 p m or Si I I 34 p m in all of the sources. In most of them we have also detected N 11 122 pm or Ne 11 12 p m (excitational potential of 14 and 21 eV respectively). In 1 1 of the 18 sources we have even detected the 0 I I I 88 p m line (the excitational potential of 0 I I I is 35 eV). Column 9 of Table 1 shows the electron densities of the ionized gas component as derived from the 0 III 52 to 0 III 88 p m lines ratios. The densities, around 10-100 crnp3, are lower than those found in the Radio Arc region (RFOlb) and in Sgr B2 (Goicoechea et al. 2003). Column 10 of Table 1 shows the effective temperatures derived from the N I I I 57 p m to N 1 1 122 p m lines ratios for the threc sources were the N I I I line has been detected. These temperatures have been calculated following the H 11 regions models by Rubin et al (1994). However, as pointed out by Shields and Ferland ( 1 993), the observed lines ratios can even be reproduced with higher

62

N. J. Rodriguez-Femandez et al.: Warm molecular gas, dust and ionized gas in the 500 central pc of the Galaxy

effective temperatures of the ionizing radiation if the ionization parameter is low, i.e., if the medium is clumpy and inhomogeneous and the ionizing sources are located far from the ionized nebulae. These seems to be the case in the Radio Arc region, were we have shown the presence of an extended component of gas ionized by the combined effect of the Quintuplet and the Arches clusters (RFOlb). The radiation can reach large distances due to the inhomogeneity of the medium (in part due to the presence of a large bubble clearly seen in infrared images). Photoionization model calculations showed that the lines ratios observed in this region can be explained with a constant effective temperature of the ionizing radiation but a different ionization parameter for each cloud, consistent with the different distances of the clouds to the ionizing sources. The analysis is more difficult for the clouds located far from the Radio Arc region since the geometry of the medium and the possible ionizing sources are unknown. However, photoionization simulations for 35,000 K are possible if the ionization many lines ratios demonstrate that effective temperatures of RF03). parameter is low (Some of the fine structure lines have been observed in the Fabry-Perot mode. The spectral resolution of this mode (- 30 km s-l) give us the possibility of studying the line profiles of the broad lines from the GC clouds. Figure 2 shows a sample of the lines observed in this mode towards two sources. Taking into account the moderate spectral resolution, the line profiles and the line centers of highly excited ions (like 0 III ) are in good agreement with those of the neutral or low excitational potential ions (like 0 1 and C II ). Furthermore, Fig. 3 shows that the agreement of the line profiles of the weakly ionized gas and the molecular gas is rather good. This fact suggest that the three components are associated, and that radiation could play a role not only in the ionization of the ionized gas but also in the heating of the neutral gas. We have compared the observed far infrared continuum and the C 11 , 0 I and Si 11 lines fluxes with the predictions from PDR models. The power radiated by lines is 0.5% of that radiated by the continuum. While the C II /O I ratio is 5. Plotting both quantities in a plot like that of Fig. 2d of Goicoechea et al. (2003), one finds that the GC clouds exhibit similar properties to those of Sgr B2, with a far ultra-violet field lo3 times larger than the local interstellar radiation field and a hydrogen density of lo3 cmP3. W cmP2)are also consistent The absolute fluxes of the C 11 (- lo-’’ W cmP2)and the 0 1 line (- 2 with the Hollenbach et al. (199 I ) PDRs models predictions.

-

-

-

-

N

6 Discussion: heating and ionization The IRAM-30m “radio view” of the GC interstellar medium seems to confirm the “classical” idea of a mainly neutral and dense gas. However, the global picture arising from the IS0 observations is a complex interstellar medium with associated molecular, atomic, and ionized gas components with decreasing den~ n f ~ lo3, n, 10’-’ cmP3). The ionizing radiation is hard (effective temperatures sities ( n ~ lo4, close to 35,000 K) but diluted due to the large distances from the ionizing sources to the nebulae (- 50 pc in the Radio Arc region). The large distance effect of the ionizing radiation can only be explained if the interstellar medium is very inhomogeneous. This scenario is also necessary to explain the low number of Lyman continuum photons derived from the radio recombination line observations. The radiation must also influence the atomic and molecular phases. As we have seen, the 0 I and C I1 lines are well explained by a PDR with a H density of lo3 cmP3 and a far ultra-violet incident field lo3 times higher than that in the local interstellar medium. The absolute 0 I and C I I lines fluxes predicted by the PDR models are also similar to the observed fluxes arising from the GC clouds. Those models also predict a warm H2 layer with temperatures of 150 K, as we have derived from the pure-rotational lines. However, the total column density of warm Hz predicted by the models is 3 x lo2’ cm-’ while the total lo2’ cm-’. Thus, column density of warm H2 that we have derived from the pure-rotational lines is we conclude that approximately 30% of the warm molecular gas in the GC clouds arises in PDRs in the external layers of the clouds. It is important to note that the discrepancy of dust and gas temperatures only rules out gas heating by gas collisions with hot dust, but it does not rule out ull radiative heating mechanisms. In the external layers 150 K by photo-electric effect on the dust of the proposed PDRs the gas is heated to temperatures of

--

-

-

-

-

-

-

-

63

Astron. Nachr./AN 324. No. SI (2003) M+O 58-0 13

M+O 24+0 02

:::mi [C 111 157741 pm

-, 3 E

[C 111 1 5 7 7 4 1

w

300 200

-200

[N I11

0

200

-200

0

200

121 898 I”

9

150

“:ip$“ -200

[o

0

200

-200

1111 aa 358 pm

0

200

[0 1111 88 356 ~ r n

100

100

80 -200

0

200

-200

0

200

L . . . . . . . . . . . . . . . -xa

Fig. 2 Fdbry-Perot spectra

-!m

c &.)

. . . . . . 7m

* . . I

m

Fig. 3 A comparison of the line profile3 of ‘3CO(l-O) (green) and C II 158 ,urn (blue) towards M+0.24+0.02 (up) and M+0.580.13 (down). Intensity in arbitrary units.

-

grains without heating the dust to temperatures higher than 35 K (Hollenbach et al. 1991). In fact, the 30 K dust component in the GC can be associated with the 150 K gas. At least a fraction of the other 7 0 % of warm gas should be heated by shocks. The main evidence for shocks in the G C are the high degree of turbulence revealed by the large line-widths and the high abundance in g a s phase of molecules linked to the dusr chemistry as S O , NH3 or C ~ H S O H(Martin-Pintado et al. 2001), which are easily photo-dissociated i n the presence of ultra-violet radiation.

-

Acknowledgements NJK-F has has been supported by a Marie Curie Fellowship of the European Community program “Improving Human Research Potential and the Socio-economic Knowledge base” under contract number HPMFCT-2002-01677. NJK-F acknowledges useful discussions with J.K. Goicoechea.

References Goicoechea, Rodriguez-FernBndez and Cernicharo 2003, these proceedings Hollenhach, D. J., Takahashi, T., & Tielens, A. G. G. M. 1991, ApJ, 377, 192 Huettemeister, S., Wilson, T. L., Bania, T. M . , & Mai-tin-Pintado, J. 1993,A&A, 280, 255 Martin-Pintado, J., Rizzo, J. K., de Vicente, P., Rodriguez-FemBndez, N. J., & Fuente, A. 2001, ApJ, 548, L65 Rodrfguez-Fem2ndez. N. J., Martin-Pintado, J.. Fuente, A., de Vicente, P., Wilson, T. L.. & Hiittemeister, S. 2001n. A&A, 365, 174 Rodriguez-Fernindez, N. J., Martin-Pintado, J., & de Vicente, P. 2001, A&A, 377,63 I Rodriguez-FernBndez 2003 (in prep). Kubin, K. H., Simpson, J. P., Lord, S. D., Colgan, S. W. J., Erickson, E. F., & Haas, M. K. 1994, ApJ, 420, 772 Shields, J. C. & Ferland, G. J. 1994, ApJ, 430, 236 Wilson, T. L., Ruf, K., Walmsley, C. M., Martin, R. N., Batrla, W., & Pauls, T. A. 1982, A&A, 1 15, 185

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S I , 65 - 71 (2003 ) / DO1 IO.l002/asna.200385063

Prospects for LOFAR Observations of the Galactic Center N. E. Kassim*',T. J. W. Lazio', M. Nord'.2,S. D. Hyman', C. L. Brogan**4,T. N. LaRosa', and N. Duric'

' Code 7213, Naval Research Laboratory, Washington, DC, 20375-5320, USA Department of Physics and Astronomy, University of New Mexico, Albuquerque, NM, 87131, USA

' Department of Physics, Sweet Briar College, Sweet Briar, VA 24595, USA

' National Radio Astronomy Observatory, Socorro, New Mexico, 87801, USA ' Department of Biological and Physical Sciences,Kennesaw State University, Kennesaw, GA, 30144, USA Key words LOFAR, low radio frequencies, Galactic center, nonthermal radio emission, transients, cosmic rays Abstract. Continued improvements in existing low frequency radio interferometers are expected, but limits of sensitivity, angular resolution, and frcqucncy range impose fundamental restrictions which cannot easily be overcome. This has inspired the development of the Low Frequency ARray (LOFAR) which will provide significantly increased imaging power over present low frequency systems. In light of advantages offered by recent low frequency observations of the Galactic center, we consider how LOFAR might impact the field by the end of this decade.

1 Introduction The breakthrough to long wavelength, high resolution imaging (Kassim et al. 1993) has ushered in a quiet renaissance in low frequency radio astronomy making contributions to several areas of Galactic center (GC) research. The 330 MHz VLA image presented in Tucson (Kassim et al. 1999; LaRosa et al. 2000) provided a striking and practical atlas of the inner few hundred parsecs. It also revealed new nonthermal sources including the first parallel nonthermal filament (NTF) (Lang et al. 1999), and the first NTF to exhibit a uniformly decreasing spectral index with length (LaRosa, Lazio, & Kassim 2001). This meeting has seen results from continued improvementsin low frequency imaging. Nord et al. (2003, hereafter N03) presented a higher resolution 330 MHz image nearly tripling the number of NTF candidates and revealing hundreds of small diameter sources. Additional randomly-oriented NTF candidates challenge the model of a simple, ordered GC magnetic field (LaRosa et al. 2003), and radio transients (Hyman et al. 2002) have motivated a focused GC monitoring program (Hyman et al. 2003). The first 74 MHz VLA image (Brogan et al. 2003) provides insights into previously poorly understood sources and affords opportunity to determine the radial locations of GC sources. Continued low frequency improvement at the VLA and GMRT are forthcoming, hut fundamental limitations are imposed by spot frequency coveragc, limited collecting area, and array size. This has inspired the development of the Low Frequency ARray (LOFAR), which will provide significantly increased imaging power. We consider how LOFAR will impact GC research as it becomes operational over the course of this decade. * Corresponding author: e-mail: Narnil:Kassim~~irl.riavy.~i~il, Phone: +01 202 767 0668, Fax: +01 202404 8894 * * NRAO Jansky Postdoctoral Fellow.

(?J 2003 WILEY-VCH V e r l q GinbH & Cu. KGilA, Wwnhnm

N. E. Kassim et al.: LOFAR and the Galactic Center

66

2 LOFAR Capabilities LOFAR will revolutionize low frequency radio astronomy by advances in three key areas: 1 ) broad-band frequency coverage from 10-240 MHz; 2) improved angular resolution, by extending maximum baselines to 400 km; and 3) significantly enhance sensitivity, afforded by lOOx increased collecting area over previous or current systems. Towards the GC, nominal LOFAR capabilities will be modified relative to extragalactic observations. Sensitivity will be affected by increased sky noise, but will still significantly exceed current capabilities. Meaningful comparisons for the GC are between the VLA’s sensitivity (- 6 hr) at [330, 741 MHz of approximately [3, 1001 mJy, and LOFAR’s projected sensitivity at [240,74] MHz of [0.1, 31 mJy, respectively. Arcsecond angular resolution will be tempered by frequency and position dependent scattering, and absorption will render certain regions opaque. However, these latter effects can be employed to map the distribution of ionized gas and to determine the radial locations of sources. Furthermore, LOFAR’s broad-band frequency coverage will provide measurements at lower and intervening frequencies currently unavailable. Frequency flexibility is important towards the GC, where it can disentangle competing effects, both intrinsic and extrinsic, which will modify the observed nature of emission and absorption processes. Finally, its -5000 baselines will exceed the baseline coverage at the VLA and GMRT by 1 OX, resulting in much better image fidelity.

-

3 Sgr A Complex While scattering and absorption will be significant towards the Sgr A complex at frequencies below 100 MHz, LOFAR can avoid much of these efrects by observing at higher frequencies. Even below 100 MHz, the absorption is patchy, and many regions around Sgr A will be open to investigation. For example, while Sgr A East is visible at 123 MHz but absorbed at 1 10 MHz, portions of the Radio Arc are visible down to 57.5 MHz (LaRosa & Kassim 1985; Kassim et al. 1986; Brogan et al. 2003) and other GC nonthermal sources can be traced to lower frequencies. N

3.1

Sgr A*

The low frequency flux density of Sgr A* anchors continuum spectra predicted by models of the black hole accretion disk and its surroundings (e.g., Yuan et al. 2003). Alternatively, if the spectrum turns over from either extrinsic (e.g.. thermal absorption, Pedlar et al. 1989) or intrinsic (e.g., synchrotron-self absorption, Beckert et al. 1996) propagation effects, this can be used as a further diagnostic of physical processes in the Sgr A environment. NO3 summarize low frequency measurements of Sgr A*, including their possible new 330 MHz detection. If that result is robust, it represents the lowest frequency at which Sgr A* has been detected. However, NO3 acknowledge that the complex effects of confusion require independent verification of their measurement. Scaling the scattering diameter of Sgr A* to 240 MHz, we expect a mean angular diameter of 17”, corresponding to baselines of approximately 15 km. Approximately 50% of LOFAR’s collecting area will be on baselines shorter than 15 km, with an 8-hr (at 4 MHz bandwidth) GC thermal noise limit of 0.2 mJy. Allowing for a further 1OX reduced sensitivity due to possible foreground confusion from intervening regions of the Sgr A complex, we estimate LOFAR could detect or set a 10 d y upper limit on the flux density of Sgr A* at 240 MHz. This is a conservative estimate since LOFAR’s frequency and resolution flexibility, and excellent baseline coverage will significantly mitigate confusion. LOFAR could thus confirm the new VLA result, albeit at a somewhat lower frequency where intrinsic or extrinsic absorption could play a more important role. However, even the existing spectrum of Sgr A* (N03, Figure 7) indicates that any such absorption turnover would have to be very sharp in frequency to hide a detection by LOFAR.

-

-

N

Astron. NachrJAN 324, No. Sl (2003) 3.2

67

Sgr A East and West

Pedlar et al. ( 1 989) placed Sgr A West in front of Sgr A East because the thermal gas in the former absorbed the nonthermal emission from the latter. Similar measurements also can be used to constrain H 11 region electron temperatures independent of LTE assumptions (Subrahmanyan & Goss 1996). Measurements at additional frequencies ( w 150 MHz) would constrain the Sgr A West electron temperature as a function of depth in the source, and, by extending measurements to lower frequencies, track the transition from optically thick thermal emission to thermal absorption. Such measurements, also applicable to other thermal GC sources, can provide improved estimates of H 1 1 region filling factors, emission measures, and distances (Kassim et al. 1989). Radio spectral index is a valuable tracer of the relativistic energy spectrum. A high fidelity image of Sgr A East at 240 MHz could be combined with existing higher frequency images for a robust spectral index study. In classic shell-type supernova remnants (SNRs) the relativistic electrons are generated via diffusive shock acceleration and the spectral signatures are well known. While recent X-ray results have reduced the initial controversy over whether Sgr A East is a SNR, there remains some debate on its nature. A spatially resolved spectral index map, made accurate by a broad frequency baseline anchored at low frequencies, will provide a good confirmation of the presumed SNR identity.

4 Nonthermal Sources 4.1

Nonthermal Filaments (NTFs)

NTFs remain mysterious structures, unique to the GC. An improved understanding of them should reveal important insights into the GC magnetic field. The first half dozen NTFs were discovered at centimeter wavelengths and, in retrospect, were perhaps only the brightest and most centrally located NTFs. NO3 have discovered a population of NTF candidates, which if confirmed, would roughly triple the population, suggesting we are probing the tip of a luminosity function. NO3 (their Figure 8) find a luminosity function for the number of NTFs of N x I - " / 5 , where I is the NTF radio brightness at 330 MHz. Moreover, the new candidates were revealed in an image optimized to detect sub-arcminute scale structure. We consider NTF detection at an intermediate LOFAR frequency of 120 MHz, which avoids most free-free absorption, while taking advantage of nonthermal NTF spectra. Assuming 50% of LOFAR's collecting area available to detect structure at this scale, attenuating the sensitivity by -1OX due to sky noise, we estimate I 1.5 mJy beam-' at 120 MHz. Assuming an NTF spectral index of -0.7 implies LOFAR should more than triple again the number of known NTFs, possibly detecting 100 objects at frequencies 2 120 MHz. Given the baseline and frequency coverage which will probe a large range ot NTF parameter spacc, we cxpcct the known source population to be increased significantly. If NTF orientation traces GC magnetic lields, randomly oriented, though generally less bright and shorter NTFs might suggest superposition of a large scale, ordered field upon a smaller scale, tangled component, or it may be related to NTF age. Alternatively, NTFs have been explained as manifestations of local field conditions (Shore & LaRosa 1999). Regardless of the correct explanation, increasing the known population by lifting selection effects influencing current interpretations will clearly benefit understanding of the phenomena.

-

4.2

-

-

SNRs

SNR catalogs are affected by severe selection effects and are incomplete at surface brightness levels 8x W m-' Hz-' sr-' (Green 199 I ), which LOFAR will cxcecd hy two orders of magnitude. Gray (1994a) has shown that the number of SNRs in the GC is enhanced, tracing an increased star formation rate. Nonetheless, the census of SNRs ai the GC remains incomplete, both for older SNRs and bright. compact, young SNRs; a more complete census will improve constraints on GC SNlSNR birthrates, statistics, energy input to the interstellar medium (ISM), and enable comparison with progenitor populations.

N. E. Kassim et al.: LOFAR and the Galactic Center

68

The usefulness of low frequency observations for revealing new GC SNRs is well documented (Kassim & Frail 1996; Gray 1994b; Roy & Rao 2002; Bhatnagar 2002), and LOFAR's much greater sensitivity and baseline coverage will easily uncover many new SNRs. 4.3

Small Diameter Sources

NO3 also have increased the census of low frequency, small diameter sources by -3X (to 240 sources) compared to LaRosa et al. (2000). While most are extragalactic, a Galactic population has also been identified, including a steep spectrum population of pulsar (PSR) candidates. If confirmed, this would be consistent with the increased star formation rate reflected in the SNR density (Gray 1994a) and the suggestion that the GC may harbor a large PSR population (Cordes & Lazio 1997). Estimates of the luminosity function from NO3 suggest that LOFAR will increase the census to over 1000 sources. With its frequency flexibility, it will be possible to use these and other background sources as probes of ionized gas, through separation of frequency dependent scattering and absorption effects.

5

Transients

Enhanced stellar densities make the G-C a promising region to search for transient radio emission. The strongest evidence is from X-rays, where successively more sensitive observations have pointed towards a high population of white dwarf, neutron star, and black hole binaries with a density peaked towards the GC (Skinner 1993; Fender & Kuulkers 200 I). This is evident in new results from Chandra and XMM-Newton (Wang et al. 2003; Predehl et al. 2003), which reveal a large population of small diameter sources, including >2000 hard X-ray sources likely associated with accreting white dwarfs and neutron stars (Muno 2003). These results imply a large number of X-ray binary systems, known sources of transient radio emission. It is therefore natural to expect a corresponding concentration of radio transients. Unfortunately, given the poor efficiency of conventional radio observations, only a handful of radio transients have been identified. The limitations of previous radio searches have motivated the GC monitoring program reported by Hyman et al. (2002, 2003). Their observations exploit the large 330 MHz VLA field of view (- 2.5") and have revealed two bright transient sources and a few variable candidates. Their work implies radio transients above 50 mJy are either very infrequent ( w one every few years) or have time scales much shorter than a month. However these results are tightly constrained by the sensitivity of the VLA and the impracticality of much more frequent monitoring. Both of these severe selection effects are lifted by LOFAR. LOFAR will be a powerful instrument for studying the bursting and transient universe. The multibeaming design incorporates a dedicated All Sky Monitor (ASM) which could continuously observe the GC for at least several hours a day, efficiently monitoring for transients over a wide spectrum of time-scales. Its hardware buffer will provide a "look back" capability that could be triggered by transient detections at other wavelengths or by LOFAR itself. Therefore LOFAR will enable the first sensitive and unbiased survey for GC transients. While scattering and absorption will limit searches at lower frequencies, searches above 120 MHz. where the sensitivity is enhanced by -1OX over the 330 MHz VLA, should be profitable. In regions more than N 1" from Sgr A*, there will be significant transparency at lower frequencies, enhancing the possibility of discovering a bursting population of coherent, intrinsically steeper spectrum sources. It i s difficult to estimate the number of GC transients LOFAR may detect, but a lower limit comes from considering the number of low mass X-ray binaries (LMXBs). Of the small diameter X-ray sources identified by Sidoli et al. (2001). -25% are LMXBs. If a similar percentage of the mainly unidentified -2000 new Chandra sources are LMXBs, this suggests a population of at least several hundred radio transients. Therefore LOFAR's ability to conduct dedicated, efficient GC monitoring may reveal a potentially large GC radio transient population.

-

Astron. Nachr./AN 324.No. S 1 (2003)

69

6 Lower Frequencies Low frequencies (< 100 MHz) offer a dramatically different view of the inner Galaxy, as shown at this meeting by comparison of 74 and 4800 MHz images (Brogan 2003). The difference is due to H 11 regions becoming opaque at low frequencies, which can be used as an important tool for Galactic astrophysics (Kassim 1990). 6.1

3D View of the Galactic Center

The appearance of sources in absorption on low frequency GC maps constrains their properties and radial position. Examples include Sgr A East and West (53), the H II regions Sgr D and Sgr E, placed in front, and behind the GC, respectively, and the appearance of the GC Omega Lobe in absorption, delineating a thermal component (Brogan 2003). These latter results emerge from a VLA 74 MHz image with -20X higher angular resolution and higher sensitivity, respectively, then previously available. LOFAR will push resolution and sensitivity at least 1-2 orders of magnitude further, providing measurements towards many inore sources. Moreover its broad-band antennas will permit observations at successively lower frequencies to smoothly trace the onset of frequency dependent propagation effects. This can be used to investigate H 1 1 region physical properties and to disentangle the relative superposition of thermal and nonthermal sources (Kassim & Weiler 1990). We note the discrete region of extended nonthermal emission, coincident with the central molecular zone (CMZ), identified on Brogan’s 74 MHz image. Delineating its properties, including its spectrum, is challenging from observations at only one frequency. LOFAR’s frequency and resolution flexibility will allow its properties to be far better constrained. Its understanding could provide useful insights on the filling factor of hot gas and the magnetic energy density within the CMZ. 6.2 Cosmic Rays Kassim (1990) has described how absorption measurements of H I I regions at known distances probes the radial (3D) distribution and spectrum of Galactic cosmic ray clectrons. This constraint is akin to H I velocity measurements in determining the distribution of H 1 in the Galaxy. Combined with measurements of soft gamma-rays generated by collisions of thc synchrotron emitting electrons with matter, the strength and distribution of the magnetic field can also be extracted (Longair 1990; Webber 1990; Duric 2000). The aim is to link the high energy particles to the presumed acceleration sites in SNRs, addressing a key problem of Galactic cosmic ray origin. More than a dozen absorption regions have been detected at 74 MHz with the VLA, but LOFAR will increase the number of such measurements to hundreds (though not all of these will be in the GC). Frequency flexibility and resolution are crucial. The VLA is limited to a single frequency where the background emission is weak, and its limited surface brightness sensitivity renders only the largest H i I regions visible in absorption. For example, for LOFAR an H I I region of size 10’ at 30 MHz will produce an absorption depth of approximately a few Janskeys. However, for a more distant or smaller H 11 region only 1’ in size, the absorption depth is roughly 20 mJy. The 74 MHz VLA can detect only the largest such absorption regions while LOFAR should have the sensitivity to probe a range of sizes, even smaller than l’,thereby encompassing a much greater number of H 1 1 regions. 6.3

Large Scale Structure

Existing low frequency images (LaRosa et al. 2001) suggest that extensions of the Radio Arc and Sgr C appear to connect to larger scale structure seen on single dish maps, but the image fidelity falls short of quantifying the connection. Sofue (2000) has linked thesc extensions to large scale X-ray structure interpreted as energetic outflows driven by episodic bursts of star formation. LOFAR will provide surface brightness sensitivity capable of connecting structure seen on interferometer maps to those on single dish

N. E. Kassim et al.: LOFAR and the Galactic Center

70

radio and large scale X-ray maps, addressing the possible link. This could be important in tracing manifestations of star formation possibly linked to previous episodes of black hole activity. The latter would be impossible to study directly because of Sgr A*’s current, presumably quiescent state.

7 Recombination lines

-

Meter wavelength (- 120 MHz), stimulated recombination lines probe envelopes of extended H 11regions, delineating the distribution of hot (T lo4 K), low density (n 1cm-3) ionized gas. In the 20-100 MHz range, both hydrogen and carbon recombination lines occur from high quantum number transitions, in 100 cmp3) gas. At such high quantum numbers, atoms have radii cooler (T N 20-100 K), denser (7z of order 0.1 mm and are extremely sensitive to their surroundings, making them excellent probes of the ambient physical conditions of a cool, poorly delineated ISM component. Erickson et al. ( I 995) have mapped GC carbon line emission, which unexpectedly appears to trace 5 11 keV positron annihilation-line emission (Purcell et al. 1997; Erickson 2002). However, it is difficult to make a definitive association based on the current, single dish radio measurements. LOFAR could map this cool gas far more extensively, and investigate any correspondencc with the gamma ray emission. Physically, an association between positron annihilation emission and carbon line absorption is not predicted but is not unreasonable. The carbon lines trace extensive regions of fairly high density, cool gas which could be positron source regions via radioactive decay or regions where fast positrons thermalize and annihilate. If an association between the radio carbon lines and the gamma-ray emission could be established, it could provide kinematic distances and detailed physical information lacking from the 5 1 1 keV observations alone (Erickson 2002). N

8 Summary We anticipate that LOFAR will make important contributions to many areas of GC research. It may provide the lowest frequency detection of Sgr A*, improve our understanding of Sgr A West and East, and likely detect many new SNRs, PSRs, and NTFs. It will be able lo map the relative radial distribution of discrete sources as well as the spectrum and distribution of cosmic ray electrons towards the GC. Combined with gamma-ray measurements the latter may offer unique understanding of the origin of cosmic rays and the strength and distribution of the magnetic field towards the GC. LOFAR will detect many new small diameter sources and utilize them to map the distribution of ionized gas through separation of frequency dependent scattering and absorption effects. These will be complimented by recombination lines from low density, hot gas, and from lower frequency lines from cooler regions of hydrogen and carbon. Lastly, one of LOFAR’s key applications is as a transient detector, and it will provide the first sensitive and unbiased survey for GC radio transients. Its all-sky field of view, broad-band sensitivity, and ability to probe a continuous range of time scales for at least several hours each day, make it ideal for uncovering a hidden transient source population which current radio observations would never have seen. If recent X-ray observations serve as a proxy, there are indications of a potentially massive transient population awaiting discovery by LOFAR. Acknowledgements Basic research in radio astronomy at the NRL is supported by the Office of Naval Research. This research is supported at Sweet Briar College by Research Corporation and the Jeffress Memorial Trust. The National Radio Astronomy Observatory is a fdcility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.

References Beckert, T., et al. 1996, A&A, 307, 450 Bhatnagar, S. 2002, MNRAS, 332. 1 Brogan, C., el al. 2003, these proceedings

Astron. Nachr./AN 324, No. S 1 (2003)

71

Cordes, J. M. & Lazio, T. J. W. 1997, ApJ, 47.5. 557 Duric, N. 2000, in: Radio Astronomy at Long Wavelengths, R.G. Stone, K.W. Weiler, M.L. Goldstein, and J.-L. Bougeret (eds.), Gcophysical Monograph 1 19 (American Gcophysical Union, Washington, DC), p. 277 Erickson, W.C., et al. 1995, ApJ, 4.54, 125 Erickson, W.C. 2002, private communication Fender, R. P. & Kuulkers, E. 2001, MNRAS, ApJ, 324, 923 Gray. A. D. 1994a, MNRAS, 270, 861 Gray, A. D. 1994h, MNRAS, 270, 847 Green. D. A. 1991, PASP, 103,209 Haynes, R. F., et al. 1978, Aus. J. Phys. Astr., 45, 1 Hyrnan, S. D., ct al. 2002, AJ, 123, 1497 Hyman, S. D., et al. 2003, these proceedings Muno, M. P. 2003, these roceedin s Kassim, N. E., et al. l 9 8 g Nature,!322, 522 Kassim, N. E., et al. 1989, ApJ, 338, 152 Kassim, N. E. & Weiler, K. W. 1990, ApJ, 360, 184 KassIm, N. E., et a]. 1993, AJ, 106, 2218 Kassim, N. E. & Frail, D.A. 1996, MNRAS, 283, L51 Kassim, N. E. et al. 1999, in "The Central Parsecs of the Gaaxy", ASP Conference Series, Vol. 186, edited by H. Falcke, A. Cotera, W. J. Duschl, F. Melia, and M. J. Rieke, p. 403 Lang et al. 1999, AJ, 521, L41 LaRosa, T. N. & Kassim, N. E. 1985, ApJ, 299, L13 LaRosa, T. N., et al. 2000, AJ, 119, 207 LaRosa, T. N., Lazio, T. J. W., & Kassim, N. E. 2001, ApJ, 563, L163 Nord, M., et al. 2003, these proceedings Predehl, P., et al. 2003, these proceedings Purcell, W. R., el al. 1997, ApJ. 491, 725 Roy. S. & Pramesh, A. R. 2002, MNRAS, 329, 775 Shore, S. & LaRosa, T. N. 1999, ApJ, 521, 587 Sidoli, L., et al. 2001, A&A, 368, 835 Skinner, G. K., 1993, A&ASuppl., 97. 149 Sofue, Y. 2000, ApJ, 540, 224 Subrahmanyan, R. & Goss, W. M. 1996, MNRAS, 281,239 Yuan et al. 2003, these proceedings

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1, 73 -77 (2003) / DO1 10.1002/asna.200385027

43 GHz SiO masers in late-type stars with 86 GHz SiO masers and astrometry with VERA in the Galactic center

'

Lorant 0. Sjouwerman* ,Maria Messineo', and Harm J. Habing2

' National Radio Astronomy Observatory, P. 0. Box 0, Socorro, NM 87801;

* Leiden Observatory P. 0. Box 9513. 2300 RA Leiden, The Netherlands;

Key words Masers, Galactic center, AGB stars, Circumstellar matter, Astrometry

We present Very Large Array (VLA) observations of 43 GHz SiO ( J = 1 --t 0 , v = 0 and v = 1) maser emission in a sub-sample of late-type stars in the Inner Galaxy. This sample of stars has relatively strong (> 0.5 Jy) 86 GHz SiO masers, and is located well within 2 degrees of the Galactic center (see contribution of Messineo et al. in these proceedings). It is our main intent to study the different maser properties of the sample, but as almost all 86 GHz SiO inasing stars are detected in at least one of the 43 GHz SiO maser lines, we here suggest that our selection criteria are a very efficient way to discover circumstellar 43 GHz SiO masers that can be used as gravitational mass probes. The 43 GHz SiO masers are extremely useful for VLBI astrometric and proper motion measurement\, for example with the Japanese "VLBI Exploration of Radio Astrometry" (VERA) VLBT network, allowing determination of Galactic structure, geometric distances (to individual objects, but also e.g. to the Galactic center), and types o f orbits of individual stars, in particular in the central part of the Galaxy.

1 Introduction As shown by Lindqvist et al. (IY92a), OH/IR stars are very useful objects to probe the gravitational potential in the central part of the Galaxy. OH/IR stars are the more massive Asymptotic Giant Branch (AGB) stars, that at the tip of the Giant branch are variable and loose mass at a very high rate due to Hydrogen shell burning and Helium flashes (thermal pulses). Their circumstellar envelope totally obscures the central star in the visible, re-radiates the stellar radiation in the mid-infrared. and - if proper conditions are met frequently sustains maser emission at 1612 MHz from the OH-molecule (e.g. Habing 1996). Circumventing the high visible extinction i n the Galactic center region, the OHllR stars are very prominent in the mid-infrared and radio wavelengths. With some effort the bolometric magnitudes can be measured (Blommaert et al. 1998; Wood et al. 1998; Ortiz et al. 2002) and their masers reveal their line-of-sight velocities accurate to a fraction of one km s-' instantaneously (e.g. Baud et al. 1981). About 200 OH/IR stars are known in the central degree of the Galactic center (Lindqvist et al. 19Y2b; Sjouwerman et al. IYYXb and references therein), and a few hundred more in the Inner Galaxy. In this paper we refer to the Inner Galaxy as Galactic longitudes 1, with 30" < I < -30" (e.g. Sevenster et al. 1997). Because of severe interstellar scattering, which scales as A', in the Galactic center the 1612 MHz OH masers arc not useful for astrometry (e.g. van Langevelde et al. 1992). OWIR stars frequently also harbor 22 GHz H a 0 andlor 4 3 GHz SiO masers in their circumstellar shell (e.g. Habing 1996 and refercnccs therein). While the HzO masers are highly variable, the 4 3 GHz ( J = 1 + 0) SiO maser is more stable. Thus in particular the 43 GHz S i O maser can be used for astrometry, most notably in the Galactic center region (Menten et al. 1997; Sjouwerman et al. 1998a, 2002; Reid et al. 2003). However, the number of43 * Corresponding author: e-mail:

lsjouwermanQnrao.edu @ 2003 WILEY-VCH Vcrlng GmhH B Co KGaA. Wr~ahem

14

L. 0. Siouwerman et al.: 43 GHz SiO masers and astrometrv with VERA in the GC

GHz masers in OH/IR stars in the Galactic center is only a very small fraction of the known OH masers (Sjouwerman 1997; Sjouwerman et al. 2002), which has triggered searches for 43 GHz masers in other types of mid-infrared sources and in blind surveys in the Galactic center (Menten et al. 1997; Sjouwerman 1997, and the Japanese groups as published in Shiki et al. 1998; Izumiura et al. 1998; Miyazaki et al. 2001; Deguchi et al. 2002; Imai et al. 2002). In particular, it is believed that the OWIR stars arc just the tip of the iceberg, and that the 43 GHz maser, or even the 86 GHz ( J = 2 + I , 'u = 1) SiO maser, is readily observable in the much more numerous Mira stars, and in some Semi-regular AGB variables and red supergiant stars (Habing 1996). The number of 43 GHz SiO masers available for astrometric studies, kinematics, dynamics and proper motions in the Galactic center has been increasing - slowly, but steadily, and recently very rapidly (Imai et al. 2002; Sjouwerman et al. 2002 and references therein; this work).

2 86 GHz SiO masers Unlike the 43 GHz SiO maser surveys of the Japanese groups mentioned above, our group has chosen to focus on the 86 GHz SiO maser to obtain more stellar mass probes to investigate the structure of the Galaxy (Messineo et al. 2002, and Messineo et al. i n these proceedings). The 86 GHz masing stars in our sample have a lower mass-loss rate than the OH/IR stars. Our survey is therefore complementary to previous OH maser surveys, although we trace-a slightly different population (e.g. Habing 1996). The combination of the OWIR stars with the population of stars defined through their 86 GHz SiO maser, should allow a proper study of the kinematics and structure of the Inner Galaxy. Because a targeted survey is more efficient than a blind survey (Messineo et al. 2002), we selected mid-infrared sources with colors and magnitudes characteristic for Mira-like stars from the ISOGAL (Omont ct al. 1999) and MSX (Egan et al. 1999) catalogs. That is, characteristic for dense circumstellar envelopes that could harbor masers, but avoiding the highest mass-loss OH/IR stars. The results of the 86 GHz SiO maser survey are given in this volume by Messineo et a]. (see also Messineo et al. 2002).

3 43 GHz observations of 86 GHz SiO masers Apart from studying Galactic structure, we are also interested in what determines whether an AGB star has a particular maser. It is possible that, among other factors, the mass-loss rate has a major influence on the type of maser apparent in the circumstellar environment. The high mass-loss rate (- lop4 Mayr-l) OWTR stars may be too turbulent close to the star to sustain 86 GHz masers, whereas semi-regular (SR) low mass-loss rate (Moyrpl) stars may never build up a dense enough shell at large distance to form an OH maser. For this reason, we arc conducting a more general search for the different masers in the different types of late-type AGB stars. Here we report on using the Very Large Array (VLA) to target the brighter 86 GHz masers found by Messineo et al. (2002) for 43 GHz (.I = 1 + 0, u = 0 and v = 1) SiO masers in January 2002; we are currently continuing our search in the less bright 86 GHz masers. Because of a limited relative bandwidth at 43 GHz, the VLA is not very effective in searching for 43 GHz maser emission ranging from -400 to +400 km spl in the ISOGAL and MSX sources. However, because we have a-priori information on the stellar velocity from the 86 GHz maser detections, the VLA, with its large collecting area, is a very powerful instrument to detect the corresponding 43 GHz masers within a few minutes of observation. The only limitation is imposed by the rapid phase changes over the array, requiring a phase calibrator within about two degrees of the targeted source. Fortunately, Sgr A* itself, the black hole in the dynamical center of the Galaxy located at the highest apparent stellar density, happens to be a very good calibrator. We searched 39 of our 86 GHz SiO masers and found 43 GHz SiO masers in 38 of them. Typical flux densities range from 0.1 to 2 Jy, which makes them observable with the VLBI Exploration of Radio Astrometry (VERA) network. Although we targeted the stronger 86 GHz masers in our first 43 GHz maser search, because of the high detection rate (38/39) it is likely that a large fraction of the less strong 86 GHz

75

Astron. Nachr./AN 324, No. S I (2003)

1

0

1

-1

Fig. 1 The 86 GHz SiO masers from Messineo ct al. (2002; big dots) compared to the location of the OH sources of both Lindqvist et al. (1992b; crosses), Sjouwerman et al. (199%; also crosses) and Sevenster et al. (1997; triangles). This figure shows that the 86 GHz SiO masers fill the gap in kinematic data between the deep, but small sky-coverage OH surveys of Lindqvist et al. (199%) and Sjouwernian e~ al. (1998b), and the large but less sensitive OH sinvey of Sevenster et al. (1997) covering the full Inner Galaxy. Most 43 GHz SiO masers within about one degree of the Galactic center found to date form a subset of the 86 GHz SiO masers (this work) and known Long-Period variables -only a few have been found in blind surveys (references in the text).

0

-1

Galactic Longitude (degrees)

Galactic Longitude (degrees)

Fig. 2 Current status of the 43 GHz SiO masers in the Galactic center. The top (square) panel shows a blow-up of the central region of the lower panel, and symbols used are indicated by the name of the PI reporting the 43 GHz SiO maser detection (although Messineo et al. 2002 refers to this 43 GHz work). Note that no ISOGAL sources are located close to the very center, because the observations avoided the most luminous infrared sources to protect the detector and save on cooling fluid. However, this niche has nicely been filled by Imai et al. (2002) who have targeted the Long-Period variables identified after a few year's monitoring program by Glass et al. (2001). A few detections are not shown because only a crude position is known (e.g. Izumiura et al. 1998), or because they blend in in the central region (e.g. Menten et al. 1997). The 43 GHz masers found here extend thc possible use of 43 GHz masers out from the inner 30 pc to more than 60 pc from the Galactic center without the need of long near-infrared monitoring programs (Glass et al. 2001).

masers in o u r follow-up V L A observations will also show 43 GHz S i O maser emission . This makes a targeted survey for 43 GHz masers in stars with 86 G H z masers, selected using their mid-infrared colors, a very efficient way to find 43 G H z masers for follow-up VLBI observations. Here w e suggest that an efficient way to discover 43 G H z S i O masers is t o target the 86 G H z SiO masers found in the sources selected from the ISOGAL/MSX-DENIS color-magnitude diagram (Messineo et al. 2002).

4

Astrometry with 43 GHz SiO masers

Once the position of an AGB star is known accurately, say to an arcsecond or better, the star becomes very attractive for astrometry purposes. Menten et al. (1997) have demonstrated nicely how the infrared and radio reference frames in the Galactic center can b e aligned by using AGB and supergiant stars; the only objects that radiate both in the infrared and in the radio through their obscuring circumstellar shell and

L. 0. Sjouwennan et al.: 43 GHz SiO masers and astrometry with VERA in the GC

16

masers therein. Because the position and the line-of-sight velocity of the AGB stars can be measured very accurately using the 43 GHz SiO maser (e.g. Sjouwerman et al. 1998a), it should be possible to obtain information on the individual orbits of the stars (Reid et al. 2003). This orbital information is important in deriving the three dimensional characteristics of the gravitational potential in the Galactic center. As mentioned above, in the Galactic center the 1612 MHz OH maser is too scattered to be used for accurate astrometry (e.g. van Langevelde et al. 1992). Not only is the scattered OH maser too extended (500 mas) to be detected with VLBI, but also the intrinsic position of the OH maser cannot be determined better than a few 100 mas. For the Galactic center, a typical line-of-sight velocity of 100 km s-l translates to 2.6 mas yr-', i.e. too small relative to the intrinsic position uncertainty to be measured accurately in a few year's time. At higher frequencies, the interstellar scattering becomes less dominant and much more accurate positions can be determined. For example, the intrinsic and scattered size of the circumstellar 22 GHz HzO maser is about 2 mas. But the H2O maser is probably too variable to be a good candidate for position monitoring. With an one mas intrinsic (and scatter) size, the 43 GHz SiO maser is the most promising candidate for determining accurate stellar positions in the Galactic center. Although a position of an 86 GHz SiO maser would be completely intrinsic, with a negligible scatter size (to a fraction of a mas), the current instruments and weather in the upper atmosphere seem to hinder mas-accuratc position measurements in the Galactic center.

5

VLBA observations of 43 GHz SiO masers

Driven by the prospect of measuring stellar proper motions at 43 GHz, Sjouwerman et al. (1998a) have used the Very Large Baseline Array (VLBA) and phased VLA to obtain first epoch mas-accurate stellar positions using the circumstellar 43 GHz SiO maser. See also Reid et al. (2003) for a more recent result. To accommodate for the rapid tropospheric phase fluctuations caused by the troposphere before fast-switching at the VLA was implemented, Sjouwerman et al. (1998a) divided the VLA up into 2 subarrays - one observing the targeted 43 GHz maser, the other observing the phase calibrator - while the VLBA was phase-referencing with a 20/20 second cycle. Although only 2 mas-accurate positions out of 10 sources were measured, the feasibility was demonstrated (Fig. 3). Since then the available instruments

!.I

I

u

10

5

"

I

I

I

$; j

B

-10

10

5

0 Mllli-arcseconds

5

-lo

a)

0 MIIIi-arcsesonds

5

-lo

b)

Fig. 3 Two detections in the 43 GHz SiO maser VLBA astrometry campaign (taken from Sjouwerman et al. 1998a).

Astron. Nachr./AN 324. No. S 1 (2003)

77

have improved; in particular the dual beam VERA will eliminate the need for fragile phase-referencing techniques.

6 Prospects with the VERA network With projected baseline lengths of up to approximately 1700 km, the VERA network will have a resolution ofjust slightly less than one mas at a frequency of43 GHz. This is exactly the proper resolution to observe circumstellar 43 GHz SiO masers in the Galactic center! Most of the 43 GHz SiO masers in the Galactic center found to date are located within 15 arcminutes, or 30 parsecs in projection, of the dynamical center, the black hole Sgr A*, and are concentrated toward the stellar density maximum. However, this (still continuing) work has also located many additional circunistellar 43 GHz SiO masers out to about 2 degrees (300 pc) or the Galactic center. Typical Hux densities of the circumstellar 43 GHz SiO masers are several 100 mly, and about 4 krn s p l wide. It is fortunate that Sgr A*, point-like and with a continuum Hux density of about 1-2 Jy at 43 GHz, is a good calibrator and located in the middle of all circumstellar 43 GHz SiO masers in the Galactic center. As Sgr A* is the dynamical center, relative astrometry with respect to Sgr A* is absolute astrometry with respect to the dynamical center of the Galaxy! The sensitivity of the VERA network, with four 25 meter antennas, will at first be a bit low to detect the majority of the circumstellar 43 GHz SiO masers i n the Galactic center. However, utilizing the dual beam receivers, spanning an angle of 2 degrees between the calibrator (Sgr A*) and the target source, all the currently known circumstellar 43 GHz SiO masers in the Galactic center can be observed coherently with long integration times. Milli-arcsecond accurate astrornetric positions can be measured with VERA and will provide proper motions of individual stars i n the Galactic center, and possibly even individual stellar orbits in the inner few tens of parsecs of the Galactic center in the next decade.

References Baud, B., Habing, H. J., Matthews, H. E., Winnherg, A . 1981, A&A 95, 156 Blommaert, J. A. D. L., van der k e n , W. E. C. J., van Langevelde, H. J . , Habing, H. J.. Sjouwerman. L. 0.. 1998, A&A 339,991 Deguchi, S., Fujii, T., Miyoshi, M., Nakashiina. J. 2002, PASJ 54, 61 Egdn 1999, A S P Conf. Ser. 177,404 Glass, I S . , Matsumoto, S., Carter, B. S., Sekiguchi, K . 2001, MNRAS 321, 77 Habing, H. J. 1996, A&A Rev 7,97 lmai, H., Deguchi, S., Fujii, T., Glass, 1. S., Ita, Y., Izumiura, H., Kameya, O., Miyazaki, A,, Nakada, Y., Nakashima, J. 2002, PASJ Let 54, L19 Izumiura, H., Deguchi, S., Fujii, T. 1998, ApJ Let 494, L89 Lindqvist. M., Habing, H. J., Winnberg. A. 19928, AXrA 259, 118 Lindqvist. M., Winnberg. A., Habing, H. J., Matthews, H. E. l992b, A & A Sup 92,43 Menten, K. M., Reid, M. J., Eckart, A,, Genzel, R. 1997, ApJ Let 475, LI 1 1 Messineo, M., Habing, H. J., Sjouwerman, L. O., Omont, A,, Menten, K. M. 2002, A&A 393, 115 Miyazaki, A,, Deguchi, S., Tsuboi, M., Kasuga, T., Takano, S. 2001, PASJ 53, 501 Omonl, A., GdneSb, S., Alard, C., Blommaert, J. A. D. L., Caillaud, B., Copet, E., Fouque, P., Gilmore, G., Ojha, D., Schultheis, M., et al. 1999, A&A 348, 755 Ortiz, R., Blommaert, J. A. D. L., Copet, E.. Ganesh, S., Habing, H. J., Messineo, M., Omont, A., Schultheis, M., Schuller, F. 2002, A&A 388, 279 Reid, M. J., Menten, K. M., Genzel, R., Ott, T., Schiidel, R., & Eckart. A. 2003, ApJ in press, (aslro-ph 0212273) Sevcnster, M. N., Chapman, J. M., Habing, H. J., Killeen, N. E. B.. Lindqvist, M. 1997, A&AS 122, 79 Shiki, S., Ohishi, M., Deguchi, S. 1998, ApJ 487, 206 Sjouwerman, L. O., 1997, PhD thesis Onsala, Sweden Sjouwerman, L. 0.. Lindqvist, M., van Langevelde, H. J., Diamond, P. J. 2002, A&A 391, 967 S.jouwerman, L. O., van Langevelde, H. J., Diamond. P. J. 1998a. A&A 339, 897 Sjouwerman, L. O., van Langevelde, H. J., Winnberg, A,, Habing, H. J. 1998b, A&A Sup 128, 35 van Langevelde, H. J., Frail, D. A,, Cordes, J. M., Diamond, P. 5. 1992, ApJ 396, 686 Wood, P. R., Hahing, H. J., McGreggor, P. J. 1998, A&A 336, 925

This Page Intentionally Left Blank

A\tron. Nachr./AN 324, No. S I , 79 - 83 (2003) / DO1 10.1002/asna.200385028

A Search for Radio Transients at 0.33 GHz in the GC Scott D. Hyman*',T. Joseph W. Lazio', Namir E. Kassim2, Michael E. Nard'.', and Jennifer L. Neureuther'

' Department of Physics, Sweet Briar College. Sweet Briar, VA 24595 USA

' Naval Research Laboratory. Code 72 13, Washington.DC 20375-5351 USA University of New Mexico. Department of Physics and Astronomy, 800 Yale Blvd. NE, Alhuquerque, NM 87131 USA

Key words radio transients. low radio frequencies, Galactic center PACS 04A25

We report on a search for transient and variable radio sources in the Galactic center using a number of 327 MHz VLA observations made during the 1990's, and a series of monthly VLA observations made during Spring and Summer 2002. A typical yield of compact sources in a given epoch is roughly 200. We have detected one new bright radio transient, GCRT 51746-2757, located only 1.1 degrees north of the Galactic center. We discuss our on-going transient monitoring program and the implications of'this work for constraining the Galactic center population of transients.

1 Introduction Known classes of highly variable and transient radio sources include radio counterparts of X-ray sources and microquasars. Although there are many examples of variable radio sources discovered as a result of high-cncrgy observations, there are surprisingly few radio surveys for highly variable or transient sources. A radio survey of the Galactic plane (Gregory & Taylor 1981, 1986) discovered 4 variable sources including GT 0236+610, a Galactic X-ray binary, and 1 candidate transient. The MIT-Green Bank surveys (Langston et al. 1990; Griffith et al. 1990; Griffith et al. 1991) discovered a number oT variable sources (< 40% variable). An on-going program at NRAO Green Bank monitors the Galactic planc at 8.4 and 14.4 GHz (Langston et al. 2000). The Galactic center (GC) is a promising region in which to search for highly variable and transient sources. The stellar densities are high, and neutron star and black hole binaries appear as (transient or variable) X-ray sources concentrated toward the GC (Skinner 1993). Previous surveys have been ill-suited for detecting radio transients toward the GC, however. Typically they have utilized either single dish instruments. which suffer from confusion in the inner Galaxy, or they have utilized the VLA for only a single epoch (e.g., Zoonematkermani et al. 1990; Becker et al. 1994). The first two radio transients detected toward the GC were A1742-28 (Davies et al. 1976) and the Galactic Center Transient (GCT, Zhao et al. 1992). These two transients had similar radio properties, but only the former was associated with an X-ray source. More recently, radio counterparts to the Xray transients, XTE 51748288 (Hellming et al. 1998a, 199%; Rupen et al. 1998) and GRS 1739278 (Hellming et al. 1996) have been detected in the GC. The G C also contains many exotic phenomcna not seen elsewhere in the Galaxy and may contain additional previously undetected novel classes of radio sources. This paper reports on the current status of our transient monitoring program utilizing low-frequency VLA radio observations. We summarize the detection and properties of a new, bright f-200 mJy) radio * Corresponding author: c-mail:

shyrnanQsbc.edu. Phone: + I 434381 6158, Fax: + I 434381 6488 @ 2003 W I L R Y ~ V C HVerlag GmhH B Cn KCiaA. Wcinhcim

S. D. Hvman et al.: Radio Transients

80

transient, GCRT 51746-2757 (Hyman et al. 2002), and the low-frequency detection of the radio counterpart to the X-ray transient, XTE 51748-288. We present one notable variable radio source candidate, and describe efficient, automated techniques we are developing to search for additional transient and variable sources. We also discuss the impact of our results on our long-term goal to constrain the nature of the transient and variable source population(s) based on their individual and group properties, and outline our future monitoring plans.

2 Radio Observations In Spring and Summer 2002, we initiated a high resolution, high sensitivity monitoring program of the GC with the VLA at 0.33 GHz in order to detect highly variable and transient radio sources. Recent developments in low frequency 3-dimensional imaging techniques (LaRosa et al. 2000; Nord et al. 2003a) allow us to produce wide-field images (= 2") with uniform and high resolution across the field. The large field-of-view covers the entire G C region with a single observation. Low frequency observations also increase the likelihood of detecting transient sources since they typically have steep spectra. We have used the 2002 observations, along with archival VLA observations made during the 1990's, to search for variable and transient candidates. We have established temporal baseline measurements for -200 sources and detected two bright (-200 mJy) radio transients (Hyman et al. 2002) and a few variable candidates. Table 1 describes our database of A- and B-configuration VLA observations consisting of various integration times, resolutions, and sensitivities. Table 1 VLA 0.33 GHz Observations Epoch 1989 March 1995 August 1996 October 1997 February 1998 March 1998 September 2002 March 2002 April 2002 May 2002 June 2002 July 2002 combined

VLA Resolution

20" x 40" 5" x 10" 51' x 10" 10" x 20" 51' x 10" 20/' x 40" 5" x 10" 5" x 10" 10" x 20" 15" X 60" 15" x 40" 8" x 14"

Duration (hr) 5.5 1.0 5.x 1.3 5.5 6.7 1.1 I.4 I .4

0.6 I .0 5.5

rmy

(mJy bm- ') 5 II 3 14 3 3

7 7 5 I4 9 3

While the inhomogeneity of the observing parameters listed in Table 1 has hindered reliable detection of variability for fainter sources, the observations have allowed us to provide constraints on the timescales of brighter transient and variable sources. For example, Figure 1 depicts the variable flux density detection threshold (70) for sources varying during our 1998 March observation. Simulated transient sources with timescales, T, shorter than the total integration time, T = 5.5 hr, were added to the visibility data. After reimaging, the detection threshold was verified to increase by a factor of T/r, as indicated by the negatively sloped line in the figure. Note that this short timescale threshold would be lower still by (T/T)'/' for images synthesized from a subset of the observation time equal to T. We are presently developing an iterative, automated routine to accomplish this, in order to improve our sensitivity to extremely short timescale transients. Figure 1 also shows the constant 7 u threshold curve for sources varying with timescales, T > T, although presently our database includes only a few observations with sensitivities similar to that of 1998 March with which to sample this temporal regime. Constant brightness temperature curves are also indicated based on source size upper limits (= CT). The temperature upper limit for an incoherent synchrotron emitter is 10l2K.

Astron. Nachr./AN 324, No. S1 (2003)

81

I / ow1 0001

'

4 ""I

' ' .

i

. . . . , L u , , & , . 0,

10

1

Timescale,

t

IW

1WO

(days)

Fig. 1 Threshold for transient detection as a function of timescale, T .We are sensitive to detecting sources with flux

densities in the region above the solid lines. The four parallel dashed lines are constant brightness temperature curves (left to right: 10"K, 10l°K, 10'K. and 106K) based on source size upper limits (= CT).

3 Search Methodologies Various methods were employed to search for transient and variable sources on each or the epoch images. First, the CLEAN components of the 5.5 hr 1998 A- and B-configuration images were subtracted from the visibility data obtained in similar configurations for each of the other epochs using the task UVSUB in AIPS. Wide-field imaging requires that many small, overlapping facets be imaged across the field-ofview in order to prevent significant distortion of sources (LaRosa et al. 2000). Accordingly, the residual data were imaged using 512 ( 5 5 ) facets for the higher (lower) resolution epochs. Each facet was then searched for transients identifiable either as a bright source, indicative of newly detected source, or as a "hole" in the image, indicative of a source no longer detectable. This "model subtraction" technique mitigates confusion, removes the time-consuming and computation-intensive CLEANing step, and results in superior sensitivity than would be obtained otherwise for epochs with shorter integration times. For epochs of similar integration time (e.g., 1996 October), the removal of the deconvolution (CLEAN) step is still a clear advantage. Figure 2 shows the bright transient, GCRT 51746.2757, visible only in the 1998 September image (see Sec. 4), appearing as a "hole" source in the 1989 March residual image. The other sources in the field are largely removed by the model subtraction technique, although not completely, due to incomplete CLEANing of the 1998 model image. Because of this limitation and other systematic uncertainties leading to imaging artifacts, the model subtraction technique is adequate only for detecting transients or highly variable sources. Therefore, only sources exceeding a threshold peak flux density of 7 ~ 7were designated as transient or variable candidates by this mcthod. In the second detection method, used primarily to search for variable sources, model subtraction was not employed. Instead, an image was synthesized for each epoch and the automated source detection program SAD (AIPS task Search and Destroy) was run. We compared source measurements with those listed in our reference database. The latter consists of -250 sources detected on an image synthesized from the combination of our 1996 and 1998 observations (Nord et al. 2003b). This comparison enabled us to quickly remove false detections resulting from the SAD procedure. Conversely, sources reliably detected on the superior sensitivity combined image, lend credence to marginal detections made on individual epoch images. All sources varying by more than 50 between any two epochs were remeasured, and if confirmcd, these sources were designated as variable candidates. Sources that differed due to unmatched resolutions and/or confusion with extended emission were removed from the list. Only a few variable candidates were

82

S. D. Hvman et al.: Radio Transients

Fig. 2 Images of GCRT 51746-2757 at 0.33 GHz. (Lefr) 1989 March residual image after subtraction of 1998 September visibility data. The radio transient appears as a “hole” in the image. (Right) 1998 September detection.

Fig. 3 Images of the variable candidate G0.490-1.043 at 0.33 GHL. (Lefr) No detection on 1989 March 18 image with rms senstivity of 7 mJy beam-’. (Righf)80 mJy detection on 1998 September 25 image.

detected. One is G0.490- 1.043, which has consistent -80 mJy detections except for a non-detection in 1989 March, as shown in Figure 3.

4 Confirmed Transients Our monitoring program has detected two transients, GCRT 51746-2757 and XTE 51748-288, both in the 1998 September image, and the latter also in a lower resolution 0.33 GHz image made from 1998 Novembcr observations. Details of these detections are presented in Hyman ct al. (2002). In summary, GCRT 51746-2757 is located only 1.1” (150 pc in projection) north of Sgr A* with a flux density at detection of 21 6 f 20 mJy. Non-detections in follow-up VLA observations at higher frequencies in 2000 July and December indicate that the source faded significantly in the interim. While GCRT 51746-2757 was detected at only one epoch and at only one frequency, a number of diagnostic imaging tests were done which provide evidence that the detection is robust. Contemporaneous searches revealed no X-ray counterpart. We conclude that GCRT 51746-2757 was either highly Doppler-boosted in the radio or was a “fast” X-ray transient, or that it is a member of a class of radio transients with no associated X-ray emission. The radio counterpart to the X-ray transient, XTE J1748-288, has been monitored extensively at 1.5 GHz and higher frequencies (Hjellming et al. 1998~1,1998b; Rupen et al. 1998). Our 0.33 GHz detections

Aatron. Nachr./AN 324, No. S I (2003)

83

(-250 mJy) in 1998 September and November together with those at higher frequencies (kindly provided by M. Rupen) indicate that the spectrum of X T E 51748-288 steepened significantly over a period of months after its radio peak in 1998 June. Based o n the ratio of its radio and X-ray fluxes, Fender &L Kuulkers (2001) have classified X T E 51748-288 as a black-hole binary candidate. Of the -30 Low-Mass X-ray Binary sources reported within our 0.33 GHz field-of-view, w e have only detected a radio transient counterpart t o X T E 51748-288. We have detected the radio counterpart to the X-ray transient G R S 1734-292, however, which underwent a burst in 1992 September. Based on its redshift, Marti et al. ( 1 998) classified this source as a Seyfert 1 galaxy. They found n o radio variability at 1 .5 GHz and higher frequencies, consistent with our nearly constant measured flux density of -150 mJy at 0.33 GHz. Evidently G R S 1734-292 is presently in a quiescent state.

5

Conclusions

V L A observations of the Galactic center at 0.33 GHz provide high-resolution, high dynamic range, large lield-of-view radio images which are well-suitcd for transient monitoring. From recent and archival 0.33 G H z observations, w e conclude that radio transients above -50 mJy are either very infrequent (approximately one every few years) o r have timescales much shorter than a month. T h u s far, w e have found only a few radio variable candidates out of -250 sources detected, but continued monitoring may yet reveal additional variables. The task is more difficult than detecting transients; e.g., a 50 mJy source visible in only one epoch is a more reliable detection than a 50 mJy variation in a 300 mJy source. T h e latter could b e due to a number of systematic uncertainties. The scarcity of radio transients/variables detected in this survey so far, underscores the need to initiate a far more extensive monitoring program. Increased sensitivity, and inore frequent observations, are required in order to detect, monitor, and identify a large number of radio sources transientlvariable over very short (< 1 hr) to very long (> 1 yr) time scales. Acknowledgements J.L.N. and S.D.H. thank Grant Denn and Manana Lazarova for further Large Array of the National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement with Associated Universities, Inc. This research was supported at NRL by the Office of Naval Research, and at SBC by Research Corporation and the JetTress Memorial Trust.

References Becker, R. H., White, R. L., Helfand, D. J., & Zoonematkermani, S. 1994, ApJ Suppl., 91, 347 Davies, R. D., Walsh, D.. Browne, I. W. A., Edwards, M. R., & Noble, R. G. 1976, Nature, 261,476 Fender, R. P., & Kuulkers, E. 2001, MNRAS, 324,923 Gregory, P. C . , &Taylor, A. R. 1986, Astron. .I.92, , 371 Gregory. P. C . , &Taylor, A. R. 1981, ApJ, 241. 596 Griflith, M., Hetlin, M., Conner, S., Burke, B., & Langston, G. 1991, ApJ Suppl., 75, 801 Griflith, M., Langston, G., Heflin, M., Conner, S., Lehhr, J., & Burke, B. 1990, ApJ Suppl., 74, 129 Hjellming, R. M., Rupen, M. P., & Mioduszewski, A. J. 1998a, IAUC 6934 Hjellming, R. M., Rupen. M. P., Ghigo, F., Waltnian, E. B., & Mioduszewski, A. J. 1998b, IAUC 6937 Hjellming, R. M., Rupen, M. P., Marti, J., Mirabel, F.. & Rodriguez, L. F. 1996, IAUC 6383 Hyman, S. D., Lazio, T. J. W., & Kassim, N. E. 2002, Astron. J., 123, 1497 Langston, G., Minter, A., D’Addario, L., Eberhardl, K., Koski, K., & Zuber, J. 2000, Astron. J., 1 19, 2801 Langston, G . I., Heflin. M. B., Conner, S. R.. LehBr. J., Carilli, C. L., & Burke, B. F. 1990, ApJ Suppl., 72, 621 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, Astron. J., 119, 207 Lazio, T. J . W., & Cordes, J. M. 1998, ApJ, 505, 715 Marti, J., Mirabel, 1. F., Chaty, S., & Rodriguez, L. F. 1998, A&A, 330, 72 Nord, M. E., et al. 2003a. these proceedings. Nord, M. E., et al. 2003b, in preparation. Rupen, M . P., Hjellming, R. M., & Mioduszewski, A. J . 1998, IAUC 6938. Skinner, G. K. 1993, A&A Suppl., 97, 149 Zhao, J-H., Roberts, D. A., Goss, W. M., Frail, D. A., Lo, K. Y., Subrahmanyan, R., Kesteven, M. J., Ekera, D. A,, Allen, D. A., Burton, M. G., & Spyromilio. J. 1992. Science, 255, 1538 Zoonernatkermani, S., Helfand, D. J., Becker, R. H., White, R. L., & Perley, R. A. 1990, ApJ Suppl., 74, 181

This Page Intentionally Left Blank

Astron. NachrJAN 324. No. SI. 85-91 (2003)/ DO1 10.1002/aana.200385088

A Molecular Face-on View of the Galactic Center Region Tsuyoshi Sawada* I , Tetsuo Hasegawa2, Toshihiro Hands?, and R. J. Cohen'

' Nobeyama Radio Observatory

' National Astronomical Observatory of Japan ' Institute of Astronomy, University of Tokyo University of Manchestcr

Key words Galaxy: center, ISM: molecules

Abstract. We present a method to derive positions of molecular clouds along the lines of sight from a comparison between 2.6 mm CO emission lines and 18 crn OH absorption lines, and apply it to the central region of the Milky Way. With some simple but justifiable assumptions, we derive a face-on distribution of the CO brightness and corresponding radial velocity in the Galactic center without the help of kinematical models. The derived face-on distribution of the gas is elongated and inclined so that the Galactic-eastern (positive longitude) side is closer to us. The gas distribution is dominated by a barlike central condensation, whose size is about 500 x 200 pc. The major axis of the condensation is tilted with respect to the line of sight by an angle of ru 70" (tilled by = 10-50" from the large-scale stellar bar). This geometry resembles the central regions of barred galaxies. The velocity field shows highly noncircular motion in the central condensation. These characteristics agree with a picture in which the kinematics of the gas in the central hundreds of parsecs of the Galaxy is undcr the strong influence of a barred potential.

1 Introduction The behavior of molecular gas, in particular its physical conditions and kinematics, in central regions of galaxies is key information to understand the star forming activity which occurs there. The Galactic center can be observed in much greater detail compared with central regions of other galaxies. It has long heen argued that the Galaxy has a bar, and recent studies strongly suggest the existence of a bar, whose Galactic-eastern (positive longitude) end is closer to us (see Gerhard 1999, for reviews). Therefore the Galactic center is also important for thc study of phenomena in central regions of barred galaxies. However, its inevitable edge-on perspective sometimes complicates the interpretation of the data. In particular, a face-on image o f the Galactic center is very hard to construct, though such an image would be very helpful to understand its kinematics and to make a comparison with central regions o f other galaxies. Attempts have been made to construct models of gas kinematics (see, e.g., Liszt & Burton 1980; Binney et al. 1991). Kinematical models can be used to project position-velocity diagrams of molecular line data into a face-on view (see, e.g., Cohen & Dent 1983; Sofue 1995). This is an indirect method to investigate the spalial distribution of the gas: it would be invaluable if we could derive positions and motions of molecular clouds independent of kinematical assumptions. Wc present a method to derive a molccular face-on view o f the Galactic center without any help of kinematical models. In $ 2 we describe the hasic methodology. Using that, we draw a face-on distribution of the molecular gas from existing data and discuss the resultant face-on view in 3 3. * e-mail: sawadaQnro.nao.ac.jp

T. Sawada et al.: Molecular Face-on View

86

2 TheMethod 2.1 Qualitative Inspection of Molecular Gas Distribution

The essence of our method lies in comparing molecular emission and absorption spectra. We compare the CO 2.6 iiiiii emission with the OH 18 crn absorption. Because the Galactic center region itself i s an intense diffuse 18 crii continuum source, strong OH absorption arises preferentially from the gas that lies in front of the continuum regions, rather than gas that lies behind them. On the other hand, the CO emission samples the gas both in front and back of the continuum sources equally. Thus the OH/CO ratio carries information on the position of the gas along the line of sight rclative to the continuum sources. We used the "CO J = 1 - 0 data by Bitran et al. (1997) and an absorption survey of the OH main lines (1665 MHz and 1667 MHz) made by Boyce & Cohen ( I 994). The CO data were smoothed and resampled with a 0?2 grid to match the OH data. Figure 1 shows the spectra of the OH absorption and the CO emission at (a. b) = (-002.000). For example, we may draw attention to the velocity components near U L S R E -130 kin s-' and 160 knis-' (shadowed in Fig. l), both of which belong to the so-called expanding molecular ring (EMR; Kaifu, Kato, & Iguchi 1972, Scoville 1972). The CO intensities of these components are almost the same: i.e., the amounts of molecular gas are similar. On the other hand, the OH absorption depths are strikingly different. We immediately deduce from this fact that the negative-velocity component, at which deep absorption i s seen, is located in front of strong continuum source surrounding the Galactic center, while the positive-velocity component is behind it. This logic led Kaifu, Kato, & Iguchi (1972) to conclude that this feature is expanding away from the center. Figure 2 shows the longitudc-velocity (l-w) diagram of the ratio between the OH apparent opacity ( T =~Tat,s/Tc,,,,t,; ~ ~ ~where Tabs and Tc,,,t are line absorption and continuum antenna temperatures, respectively) and the CO J = 1 - 0 line intensity. The ratio has a clear trend; high ratio tends to he seen at positive !, negative 21. The hatched velocity ranges in Fig. 2 are excluded from the following analysis because the ratio is affected by the foregroundgas in the Galactic disk. 300

I

0

-100 6

'

5

'

1 4

'

/ , I , I , I , l , I 1 1 , 1 ,I , 3 2 I 0 1 - 2 - 3 - 4 - 5 - 6 GaldCtlC

0 00

LSR Velocity [km s-'1

Fig. 1 Sample spectra of CO emission (thin line) and OH absorption (thick line) at ( B , b) = (-0?2.0?0). Shadowed velocity components demonstrate the difference of cloud position along the line of sight (see text).

0 05

Longitude [degrees]

0 10

n 15

(OH Apparent Opacity)/(CO J = 1 4 Tim) [K '1

Fig. 2 The ratio between the OH apparent opacity and the CO J = 1 - 0 intensity at b = 0". Contours are TL,u(CO)= 1 , 2 , 4 , 7 , 1 0 ,and 14 [K]. Hatchedvelocity

ranges were excluded from the analysis because of contamination by clouds well outside the Galactic nucleus.

Astron. Nachr./AN 324, No. S1 (2003) 2.2

87

Deriving Distances to Clouds

We extend the logic shown above to quantitative estimation of molecular gas distribution. In order to determine the distances to molecular clouds quantitatively, we adopt the following four assumptions. (1) At a given P,emission observed at each velocity bin comes from a single position along the line of sight. (2) The CO line intensity Tc.0 at a given velocity is proportional to the amount of molecular gas in unit velocity width, N(Hl )/ d1,,and the OH opacity T()H (not T ~ , , , , ,but the real opacity) at a given velocity is also proportional to N(H2)/&:: thus T ~ ) H= ZT(.o where 2 is a constant. (3) The excitation temperature of OH, T,,,(OH), is uniform. (4) The 18 ciii continuum emission is optically thin and arises from a distributed, axisymmetric volume emissivity, j ( r ) [ r is the Galactocentric radius]. Now 2 and T,,(OH) are unknown parameters: how to determine them is described in $ 2.3. When a cloud whose OH opacity is T ( ) H is located at s = sg ( s is the position along the line of sight), the cloud absorbs the continuum intensity behind it, J"; j ( r ) d s . The absorption depth is written as ,f[1 ~ ~ ) ( - T ~ H ,j('r)d,s ) ] [ J- T,,(OH)] " : L where f is the beam filling factor of OH absorbing gas. The : j ( ~ ) d sand , the apparent opacity of the cloud is expressed as observed continuum intensity is:J -

and Tc.0(and thus 7()H = ZT&) are known through the observations, we can obtain the value Since 7EL,,p j ( r ) d s if we know f , 2,and Tpx(OH).Then s o is derived from the value of::J , j ( r ) d s using a -w distribution model of j ( r ) . We assume that the 18 cm continuum cmissivity , j ( r ) around the Galactic center is described as a sum of several axisymmetric Gaussians:

"fJ'""

Figure 3 shows the schematic relation of geometrical parameters. The Galactic longitude !is in degrees. Here T , s, and (T! are in units of projected distance corresponding to I" at the distance to the Galactic center: i.e., 150 pc at 8.5 kpc. The observed longitudinal distribution of the continuum brightness Tcont(P) can be written as Tcoll+(P) = ;:/, j ( ~ds. ) Since the continuum emissivity beyond the solar circle should j ( r )ds can be replaced with JTxj ( T ) d s . Thus be negligible compared with that in the center. ;:"J

Now n , and ~7~can be drawn so that the Eq. (3) reproduces the observed longitudinal distribution o f t h e continuum brightness. It is found that ohserved continuum distribution is well fitted by up to three components (iV = 3 ) . For b = 000,we obtain nl = 98.3, a 2 = 34.0, o y = 11.4; (TI = 0.120, g 2 = 0.677, (73 = 7.18 ( a zare in Kelvins, 0,are in degrees). Figure 4 shows the result of our fit. The model reproduces the observed data quite well. Since C T ~ 0, 2 are small enough (6 100 pc) and al,a2 are large, the continuum distribution due to the first ( i = 1) and the second ( i = 2) components is well defined. On the other hand, because of large ( ~ and ~ 1 small 0 3 , the third (i = 3 ) component should suffer from possible non-axisymmetric distribution on the largest scale andor individual sources; furthermore, small dT,.,,+/ds causes large errors in the derived positions of clouds. Therefore, i n longitudes where the contribution from the first and the second components is negligiblc (i.e., (el 2 105, corresponding to f 2 2 0 pc), the positions of clouds obtained using this model are rather uncertain.

T. Sawada et al.: Molecular Face-on View

88

Line of Sight A

I’ lSO

RADIO CONTINUUM SOUHCEAHOUND

5

:loot h“ so

0

6

4

2

0

-2

-4

-6

Galactic Longitude [degrees] Fig. 3 The schematic relation of gcoinetrical parameters in the face-on view of the Galactic center region.

Fig. 4 Longitudinal distribution of the 18 cm continuum in units of antenna temperature at b = O?O by Boyce & Cohen (1994) (open circles) and a fit by 3 Gaussians (solid line).

2.3 Choice of the 2 and T,,(OH) Values Following the procedure, we can draw a face-on (2-y) distribution of the molecular gas by putting each data point of the !-v diagram onto the x:-y plane with a projection of (P. s) + (:r. y). However, f, Z,and T,,(OH) are still unknown. Here we assumc ,f = 1. There is no bottom-up scheme to determine 2 and T,,(OH). We employ a trial-and-error scheme, making face-on maps at b = 0?0. 1t012. and *0?4 with various values of 2 and T,,(OH). Trials have been done for 2 = 0.04 to 0.70 [K-’1 and T,.,(OH) = 0 to 10 [K]. We have chosen an appropriate set OF (Z. T,-,(OH)) so that the following three conditions are satisfied: ( I ) The resultant face-on distribution of the CO brightness is not too asymmetric between the near and far sides with respect to the centcr; (2) The features extending above and below the Galactic plane are placed in similar face-on positions a1 different latitudes; and (3) Most of the CO emission has a solution for the position so. By combining these conditions, we have chosen Z = 0.15 & 0.03 [K-’] and T,,(OH) = 4 f l [K]. Sets of(larger 2,smaller T,,(OH)) and (smaller Z,largcrT,,(OH)) are rejected from conditions (1) and (3), respectively. If we adopt smaller or larger Trx(OH) values, condition (2) is not satisfied. The validity of the parameters, f,2 , and T,,(OH), is discussed in 5 3.3.

3 Results and Discussion 3.1 Overall Structure Figure S shows the resultant face-on view of the central 1kpc x 1kpc: of the Milky Way at b = 0?0, seen from the north Galactic pole. The CO brightness of each pixel in Fig. 2 is projected onto the y y plane and smoothed by m = 20 pc Gaussian. Figures Sn and Sh show the distributions of CO brightness and corresponding radial velocity, respectively. As seen in Fig. 2, high O H / C O ratios are more widespread at positive longitudes. This comes about because the face-on CO distribution is tilted so that the gas in the Galactic-eastern (positive 2 ) side lies closer to us, as qualitatively suggested by Cohcn & Few (1 976). An elongated molecular cloud condensation (“central condensation”), whose size is approximately 500 x 200 pc, dominates the CO emission in the Galactic center region. It should be compared with “twin peaks” in central regions of barred galaxies (Kenney et al. 1992). The minor axis length of the condensation might be smaller since the face-on map involves positional errors along the lines of sight. The major axis of the condensation is tilted with respect to thc line of sight (n. = 0) by E 70” so that the Galactic-eastern side

Astron. Nachr./AN 324, No. S1 (2003)

89

3uo

1

J

-400

-400

1 IPLI

I IPCI

no

05

in

200

100

‘ T O J=I-0

ILW

n Lkm \

100

‘I

?on

Fig. 5 The resultant molecular face-on view 01 the Galactic center at b = 0’0 ( a ) Distnbution of CO J = 1 - 0 emission (normalized by the peak value). Contours are 0 01,0.02, 0 04, 0 08, 0.16, 0.32, and 0 64 of the peak value (h) Correvponding radial velocity The solar system is located dt (z, y) = (0. -8500)

is closer to us: this angle does not change significantly in the acceptable parameter space of (2. Tc:,(OH)). The condensation includes the Sgr A ( B E 0?0),Sgr B (B N 0?6), and Sgr C (! ru -0”) molecular cloud complexes and a huge molecular cloud complex at P r-. 1:s (“1:s region”). The face-on distribution of radial velocity (Fig. 5b) shows that the gas motion in the condensation is strongly noncircular. The gas on the far-side has larger receding velocity, while the gas on the near-side is approaching. This velocity field can be explained if the gas orbits are elongated along the major axis of the condensation. A similar trend is often seen in both numerical simulations of gas kinematics in a barred potential and observations of barred galaxies (see, e.g., Athanassoula 1992; Lindblad, Lindblad, & Athanassoula 1996). This fact agrees with the arguments that the Milky Way is barred. The EMR feature corresponds to the peripheral region surrounding the central condensation.

3.2

Nature of the Gas in the Central Hundreds of Parsecs

Oka et al. (1996) argued that formation of massive stars may have been taking place in the central 100 pc of radius (“star-forming ring”), based upon comparisons between their CO data, hydrogen recombination line emission (Pauls & Mezger 1975), and OWIR stars (Lindqvist et al. 1991). The “star-forming ring” forms a part of the central condensation. Sofue (1995) identified a pair of arm-like features in the central condensation from the I 3 C 0 data taken by Bally et al. (1987), though these “arms” are not clearly separated in our face-on map. This may he due to insufficient spatial resolution of the present analysis and to deviation from the assumed smooth, axisymmetric continuum emissivity because of embedded discrete sources. Considering Sofue’s two-arm model and our face-on map, active star-forming regions Sgr B and C are both located on the leads of the arms. Conccntration of clouds onto the arms would he occurring due to some kind of gas orbit crowding. The locations of the sites of active star formation on the leading edges of the inner arms may suggest a time lag between the arrival of the gas into the central orbit and the beginning of star formation as discussed by Kohno et al. (1999) for NGC 695 1. Our results are schematically summarized in Figure 6. We have found that the major axis of the central condensation is tilted by an angle of N 70” with respect to the line of sight. Therefore, the inclination between the major axes of the large-scale stellar bar and the central condensation is ru 40”-50”, if our viewing angle of the bar is ru 20”-30” (see Gerhard 1999, for reviews). In some barred galaxies, there are central molecular gas concentrations (“twin peaks”, Kenney et al. 1992). In a face-on projection of high resolution images, some of these concentrations tend to consist of N

T. Sawada et al.: Molecular Face-on View

90

a pair of arms and their major axes are tilted with respect to the bar major axes by moderate angles: such as IC342 by lshizuki et al. (1 990), M 10 1 etc. by Kenney et al. ( 1 992) NGC 695 1 by Kohno et al. (l999), and NGC 4303 by Koda et al. (in preparation). They are similar to our face-on view of the Galactic center. Hydrodynamical simulations of the barred galaxies also produce similar gas distributions. Central Condensation /, .. A pair of arms7

c

/

SgrB

Observer

300 pc

Fig. 6 A picture of molecular gas in the Galactic center region proposed on the basis ofthe results from this work. Linear scales are approximate.

3.3

I [pc]

Fig. 7 Contours show the integrated continuum emissivity j ( i - ) d s at b = 0eO.Levels are 2.5, 5.0, 7.5, 10.0, and 20.0 K in units of brightness temperature. The center (I-,y) = (0,O) is shown by a cross. Dashed lincs indicate I = +1?5.

Validity of Parameters

Beam filling factor. We have assumed that f , thc beam filling factor of the OH absorbing gas, is unity. From Eq. ( l ) , T ~ ~gives , ~ , the lower limit o f f . For some lines of sight, T:~,,,, is rather large: 0.64 toward Sgr B; 0.35 toward the 1:s region and Bania's Clump 2 (Bania 1977) at B r , 3". Sawada et al. (2001) estimated the beam tilling factor of CO J = 1 - 0 emission to be 0.4-0.7 based upon a large velocity gradient analysis and high resolution CO data taken by Oka et al. (1998b). Since it is expected that the OH absorption arises also from less densc cloud envelopes compared with thc CO emission, thc beam filling factor of OH absorbing gas is at least similar and can be even largcr. These facts indicate that f is large and can be reasonably replaced with unity. The "Z" factor. There are three previously-known relations. ( I ) The column density of molecular hydrogen N(H2) is proportional to the CO .J = I - 0 integrated intensity. The conversion factor (X-factor) is measured to be about 2 x 102"[ ~ i i i (K - ~kiris-')-'] for molecular clouds in the Galactic disk (see, e.g., Dame, Hartmann, h Thaddeus 2001). For the Galactic center clouds, however, it is reported that the X-factor is smaller than that for the Galactic disk: i.e., X = scveral x 10'' (see, e.g., Oka et al. 1998a). (2) The relative abundance of OH to molecular hydrogen, [OH]/[H2], is typically sr,vpra,l x lop7 (see, e.g., Herbst & Leung 1989). (3) The OH column density is derived from the OH 1667MHz opacity 7 as N(OH) = 2 x 1014T,, / T ~ U Using . these relations, the 2 factor is written as Z [K-l] = 5 x lo-:' ([OH]/[H2]/10-7) ( X [cnir2( K k ~ i i s - ' ) - ~ ] / l O (lO/T,, ~~) [K]). This value becomes several x lo-', several times smaller than our value, 0.15. This discrepancy can be understood if we consider the higher OH abundance (sevcml x lo-"), which may be due to interstellar shocks (see, e.g., Wardle 1999) and/or a high ionization rate caused by X-rays (Lepp h Dalgarno 1996). Excitation temperature. Figure 7 shows the integrated continuum emissivity, j ( r ) d s in Eq. ( I ), at b = 0?0. It is expected that at each location in the iigure, any gas whose T,,(OH) is higher than the

L:J

Astron. Nachr./AN 324, No. SI (2003)

91

integrated continuum emissivity at that location would be seen in emission. However, no significant OH emission is observed. Thus, if T,,(OH) 2 (i[K],the face-on gas distribution is almost restricted to near side of the Galactic center, g < 0, which is unrealistic. On the other hand, Boyce & Cohen (1994) noted that the OH absorption was only detected at positions where the continuum temperature exceeded 5 K of antenna temperature (7.5 K of brightness temperature). Lower limit of T,, (OH) is provided from this fact: T,,(OH) is higher than the integrated continuum emissivity, which is typically half of the observed continuum brightness for the Galactic center clouds. Thus T,,,(OH) should not be far below (7.5/2) K. These two constraints are consistent with our value, 4 K.

4

Conclusions

We have developed a method to derive positions of molecular clouds along the lines of sight. The method is completely independent of any kinematic model and based on observable data alone; the C O emission line, the OH absorption line, and 18 (:m continuum distribution. It is applied to the central region of the Milky Way to obtain a molecular face-on map. Most of the CO emission comes from the “central condensation”. It is elongated, and its major axis is tilted wilh respect to the line of sight by = 70” so that the Galacticeastern end is closer to us. The gas within it shows highly noncircular motion: the gas in the far side is receding whereas the gas in the near side is approaching. This noncircularity of the gas motion is most likely induced by a barred potential. The results give a new evidence for the existence of a bar in the Milky Way Galaxy based on direct distance derivation independent of kinematic models. Acknowledgements We acknowledge Leonardo Bronfman for providing us the CO d = 1 - 0 data in a computer readable form This work was supported by a Grant-in-Aid for Scientific Research of the Ministry of‘ Education, Culture, Sports, Science, and Technology 08404009 and 10147202.

References Athanassoula, E. 1992, MNRAS, 259, 345 Bally, J., Stark, A. A., Wilson, R. W., & Henkcl, C. 1987, ApJS, 65, 13 Bania, T. M. 1977, ApJ, 216, 381 Binney, J., Gerhard, 0. E., Stark, A. A., Bally, J., & Uchida, K. I. 1991, MNRAS, 252, 210 Bitran, M., Alvarez, H., Bronfman, L., May, J., & Thaddeus, P. 1997, A&AS, 125, 99 Boyce, P.,& Cohen, R. J. 1994, A&AS, 107, 563 Cohen, R. J., & Dent, W. R. F. 1983, in Surveys of the Southern Galaxy, ed. W. B. Burton & F. P. Israel (Dordrecht: Reidel), 159 Cohen, R. J., & Few, R. W. 1976, MNRAS, 176,495 Dame, T. M., Hartmann, D., & Thaddeus, P. 2001, ApJ. 547,792 Gerhard, 0. E. 1999, in ASP Conf. Ser. 182, Galaxy Dynamics, ed. D. Memtt, M. Valluri, & J. Sellwood (San Francisco: ASP), 307 Herbst, E., & Leung, C. M. 1989, ApJS, 69, 27 I Ishizuki. S., Kawahe, R., Ishiguro, M., Okumura, S . K., Morita, K.-I., Chikada, Y., & Kasuga, T. 1990, Nature, 344,224 Kaifu, N., Kato, T., & Iguchi, T. 1972, Nature Phys. Sci., 238, 105 Kenney, J. D. P., Wilson, C. D., Scoville, N. 2.. Devereux, N. A,, & Young, J. S. 1992, ApJ, 395, L79 Kohno, K., Kawabe, R., & Vila-Vilar6, B. 1999, ApJ, 5 I I , 157 Lepp, S., & Dalgamo, A. 1996, A&A, 306, L21 Lindblad, P. A. B., Lindhlad, P. O., & Athanassoula, E. 1996, A&A, 313, 65 Lindqvist, M., Winnherg, A., Hahing, H. J.. & Matthews. H. E. 1991, A&AS, 92, 43 Liszt, H., & Burton, W. B. 1980, ApJ, 236, 779 Oka, T., Hasegawa, T., Handa, T., Hayashi, M., & Sakamoto, S. 1996, ApJ, 460, 334 Oka, T., Hasegawa, T., Hayashi, M., Handa, T., & Sakamoto, S. 1998a, ApJ, 493, 730 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyazaki, A. 1998t1,ApJS, I 18, 455 Pauls, T., & Mezger, P. G. 1975, A&A, 44. 255, Sawada, T., et al. 2001, ApJS, 136, 189 Scoville. N. Z. 1972, ApJ, 175, L127 Sofue, Y. 1995, PASJ, 47, 527 Wardle, M. 1999, ApJ, 525, LlOl

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 93-99 (2003) I DO1 10.1002/asna.200385109

The Inner 200pc: Hot Dense Gas Christopher L. Martin* I, Wilfred M. Walsh', Kecheng Xiao', Adair P. Lane', Christopher K. Walker', and Antony A. Stark'

' Harvard-Smithsonian Center for Astrophysics, 60 Garden St., MS-12, Cambridge, MA 02138 ' Steward Observatory,University of Arizona, 933 N. Cherry Ave., Tucson, AZ 85721 Key words Ga1axy:center 1SM:molecules

-

Ga1axy:kinematics and dynamics - 1SM:atoms - 1SM:general

-

Abstract. We present fully-sampled maps of 461 GHz CO J = 4 --7' 3 , 807 GHz CO J = 7 + 6, and 492 GHz [C I] 'PI --t "POemission from the inner 3 degrees of the Galactic Center region taken with the Antarctic Submillimeter Telescope and Remote Observatory (AST/RO) in 2001-2002. The data cover -1P3 < e < 2', -0P3 < b < 0y2 with 0.5' spacing, resulting in spectra in 3 transitions at over 24,000 positions on the sky. The CO .I = I 3 emission is found to be essentially coextensive with lower-J transitions of CO. The CO J = 7 i 6 emission is spatially confined to a far smaller region than the lower-J CO lines. The [C 11 'PI + 'PO emission has a spatial extent similar to the low-J CO emission, but is more diffuse. Bright CO d = 7 + 6 emission is detected in the well-known Galactic Center clouds SgrA and SgrB. Analyzing our CO J = 7 + 6 and CO J = 4 + 3 data in conjunction with J = 1 + 0 " C O and I3 C/O data previously observed with the Bell Laboratories 7-m antenna, we apply a Large Velocity Gradient (LVG) model to estimate the kinetic temperature and density of molecular gas in the inner 200 yc of the Galactic Center region. Typical pressurcs in the Galactic Center gas are n(EI2) . Tklr, 10" K cin-". We present an % !( b ) map of molecular hydrogen column density derived from our LVG results.

-

7

-

1 Introduction

-

Much has been learned about dense gas in the Galactic Center region through radio spectroscopy. Early 2) OH absorption (Robinson et al.(l964),Goldstein et aL(1964)) suggested the observations of F ( 2 existence of copious molecular material within 500 pc of the Galactic Center. This was confirmed by detection of extensive J = 1 4 0 "CO emission (Bania(lY77),Liszt & Burton(1978)). Subsequent CO surveys ( Bitran(1987), Stark et a1.(1988), Oka et a1.(1998), Bitran et a1.(1997)) have measured this emission with improving coverage and resolution-these surveys show a complex distribution of emission, which is chaotic, asymmetric, and non-planar; there are hundreds of clouds, shells, arcs, rings, and filaments. On scales of 100 pc to 4 kpc, however, the gas is loosely organized around closed orbits in the rotating potential of the underlying stellar bar (Binney et al.( 1991)). Some CO-emitting gas is bound into clouds and cloud complexes, and some is sheared by tidal forces into a molecular inter-cloud medium of a kind not seen elsewhere in the Galaxy (Stark et al.( 1989)). This diffuse inter-cloud medium appears in absorption in F ( 2 + 2) OH (McGee(1970),Robinson & McGee(l970)), in (110 + 111) H2CO (Scoville et al.(l972)), and in J = 0 + 1 HCO+ and HCN (Linke et aL(l981)). In contrast, the clouds and cloud complexes are dense, as they must be to survive in the galactic tide, and they appear in spectral lines which are tracers of high density (n(H2) > lo4 (m-')), such as NH3 (1, 1 ) (Giisten el aL(1981)) and CS cJ = 2 + 1 (Bally et al.( 1988)). The large cloud complexes, Sgr A, Sgr B, and Sgr C, are the among the largest molecular cloud complexes in the Galaxy ( A 1 2 10"' M a ) . Such massive clouds must be sinking * Corresponding author: e-mail:

cmartin @cfa.harvard.edu @ 2003 WILtY-VCH Vrrlag GwhH B Co K G . A Weinhem

C. Martin et al.: AST/RO: Inner 200pc

94

toward the center of the galactic gravitational well as a result of dynamical friction and hydrodynamic effects (Stark et al.( 1991)). The deposition of these massive lumps of gas upon the center could fuel a starburst or an eruption of the central black hole (Genzel & Townes( 1987)). As prelude to further study of the Galactic Center molecular gas, we would like to determine its physical state-its temperature and density. This involves understanding radiative transfer in CO, the primary tracer of molecular gas. Also useful is an understanding of the atomic carbon lines, [C I], since those lines trace the more diffuse molecular regions, where CO is destroyed by UV radiation but Hz is still present. The .J = 1 + 0 I 2 C 0 line is often optically thick. Its optical depth can be estimated by studying its isotopomers, '"CO and C " 0 . In the Galactic Center region, l'C0 is 24x less abundant than l 2 C 0 (Penzias( 1980),Wilson & Matteucci(l992)), and C l 8 0 is 250x less abundant than 12C0 (Penzias(1981)). Since the radiative and collisional constants of all the isotopomers are similar, the ratio of optical depths in their various spectral lines should simply reflect their relative abundances. Where the lines are optically thin, the line brightnesses should be in the same ratio as the isotopic abundances; where the lines are thick, deviations from the abundance ratios are a measure of optical depth. Bally et a].(19871, Bally et al.( 1988) and Stark et al.( 1988) produced fully-sampled surveys of "CCO and '"C0 in the Galactic Center region. They find that the ratio of the " C O .I = 1 + 0 to "CO J = 1 + 0 line brightness temperatures (TtZo/T&) is typically 10 3:2 in Galactic Center gas that is far from dense cloud cores. This a),especially in indicates much of the Galactic Center l 2 C 0 emission is only moderately thick 6 (Polk et al.( 1988)) is smaller, comparison to the galactic disk outside 3 kpc radius, where T;:"/T;:(, even though the isotope ratio 12C/13C N 40 (Penzias( 1980)) is larger. Heiligman( 1982) and Dahmen et aL(1998) made surveys in ClSO .J = 1 40. These show "C0 J = 1 -+ 0 to C"0 J = 1 0 line brightness temperature ratios (T:?+o/T;50)which vary from 40 to over 200, with typical values near 70, indicating values of T::(, which vary from 3 to less than I , while the core region of SgrB2 shows 7;:" 10. Determining the excitation temperature of CO works best if emission lines from several J levels have been measured. Lacking such observations, what is often done is to use the brightness temperature of the "CO J = 1 40 line as a lower limit to the excitation temperature of the J = 1 state, T e x , ~ =This l. estimate can be misleading, because the emission may not fill the telescope beam, diluting the brightncss temperature and causing it to be many times smaller than Tex,.~=l; as will be apparent from the data to be presented here, this is thc usual case for gas in the Galactic Center region. Moving up the energy ladder, Sawada et a1.(2001) surveyed the Galactic Center region in l 2 C 0 J = 2 4 1. They compare their data to the .J = 1 0 data of Bitran et aL(1997) and find Ti51/T:?+o = 0.96 0.01, with little spatial variation. What this means is that almost all the CO in the Galactic Center region has l o w 4 states which are close to local thermodynamic equilibrium (LTE), so that the excitation temperatures Tex, J of those states are all close to the kinetic temperature, Tki,, and , the ratio of line brightnesses for transitions between those states are near unity and therefore independent of Tkill (cf. Goldreich & Kwan(l974)). LTE in the low-.J states of CO does not occur under all circurnstanccs in the interstellar medium, but it is very common and appears to be the rule for Galactic Center gas. For each value of Tkin and ~ ( H Z )there , will, however, be some value of J above which all higher-J states fail to be populated, because their Einstein A coefficients (which increase as J3)are so large that the collision rate at that value of Tki,, and n(H2) cannot maintain those states in LTE, and they must therefore be subthermally excited, i.e., T e x ,<< ~ Tkin. The brightness temperature of the J + J - 1 line from those states will be significantly less than that of the lower-J states, and the line ratios T:EJ-l/T::o will be much smaller than unity. Unlike the low-J states, the value of the line ratios from those higher4 states will vary from place to place, depending on T k i n , 71.(H2),and radiative transfer effects. Higher still on the energy ladder, Kim et aL(2002) used the ASTRO telescope to survey a strip at b = 0" in l 2 C 0 J = 4 + 3 and l 2 C 0 J = 7 i 6. They found that even the Ti?+3/T:?+o ratio is not far from unity and shows little spatial variation. In contrast, the distribution of .J = 7 + 6 line emission was found to be markedly different from the lower-J transitions. Temperatures and densities could therefore

-

-

N

-

-

*

--f

Astron. Nachr./AN 324, No. S1 (2003)

9.5

be calculated as a function of position and velocity using the varying value 0fT:!+,;/T25~ and an estimate of 7;:” from the Bally et a1.(1987), Bally ct a1.(1988) and Stark et aL(1988) data. In the current paper, we extend this work to a fully-sampled (P, 1 7 ) map of the Galactic Center region, and estimate kinetic temperature, T k i n r and density, 11 (H2). throughout our mapped area.

2 Observations The observations were performed during the austral winter seasons of 2001 and 2002 at the Antarctic Submillimeter Telescope and Remote Observatory (AST/RO; Stark et a1.(2001)), located at 2847 111 altitude at the Amundsen-Scott South Pole Station. This site has very low water vapor, high atmospheric stability and a thin troposphere making it exceptionally good for submillimeter observations (Chamberliu et a).( 1997), Lane( 1998)). AST/RO is a I .7 in diameter, offset Gregorian telescope capable of observing at wavelengths between 200 pm and 1.3 IIIIU (Stark et aL(1997)). A dual-channel SIS waveguide receiver (Walker et a].(1992),Honingh et al.( 1997)) was used for simultaneous 4 6 1 4 9 2 GHz and 807 GHz observations, with double-sideband noise tempcratures of 320-390 K and 1050-1 190 K , respectively. Telescope efficiency, r p , estimated using moon scans, skydips, and measurements o f the beam edge taper, was 81% at 4 6 1 4 9 2 GHz and 71% at 807 GHz. Atmosphere-corrected system temperatures ranged from 700 to 4000 K at 461-492 GHz and 9000 to 75,000 I< at 807 GHz. A multiple position-switching mode was used, with emission-free reference positions chosen at least 60’ from regions of interest. These refercnce positions were then shared by a strip of points at constant galactic latitude which were observed five at a timc. This mapping mode caused each point in the map to be observed for 60 s per pass through the map. In an attempt to obtain uniform noise over the entire region of interest, subregions were reimaged as often as required. Emission from the CO .J = 4 i 3 and CO ,J = 7 + G lines at461.041 GHz and 806.652 GHz, together + ,”Poand [C I] 3P2 + ‘3P,lines at 492.262 GHx and 809.342 GHz, was imaged over with the [C I] the Galactic Center region -SP3 < B < 2 O , -0?3 < b < OP2 with 0.5’ spacing in P and h; i.e., a spacing of a half-beamwidth or less. Smaller selected areas were also observed with longer integration times in the [C I] 3P1 ‘Ppo line. Maximum pointing errors were no larger than S’, and the beam sizes (FWHM) were 103-109” at 461-492 GHz and 58” at 807 GHz (Stark et a1.(2001)). To facilitate comparison of the various transitions, the data were regridded onto a 0.25’ grid and smoothed to a FWHM spatial resolution of 2’ with a Gaussian filter function. Two acousto-optical spectrometers (AOSs; Schicder et al.( 1989))were used as backends. The AOSs had I .07 MHz resolution and 0.75 GHz effective bandwidth, resulting in velocity resolution of 0.65 kin s-’ at 46 I GHz and 0.37 kin s-l at 807 GHz. To facilitate comparison, the data were then smoothed to a uniform velocity resolution of 1 kriis-’. The high frequency observations were made with the CO .J = 7 + G line in the lower sideband (LSB). Since the intermediate frequency of the AST/RO system is 1.5 GHz, the [C I] “P2 4“PI line appears in the upper sideband (USB) and is superposed on the observed LSB spectrum. The local oscillator frequency was chosen so that the nominal line centers appear separated by 100 krii s-l in the double-sideband spectra. The standard chopper wheel calibration technique was employed, implemented at AST/RO by way of regular (every few minutes) observations of the sky and two blackbody loads of known temperature (Stark el a1.(2001)). Atmospheric transmission was monitored by regular skydips. and known, bright sources were observed every few hours to further check calibration and pointing. At periodic intervals and after tuning, the receivers were manually calibrated against a liquidnitrogen-temperature load and the two blackbody loads at ambient temperature and about 100 K. The latter process also corrects for the dark current of the AOS optical CCDs. The intensity calibration errors became as large as +S5% during poor weather periods. Once taken, the data in this survey were reduced using the COMB data reduction package. After elimination of scans deemed faulty for various instrumental or weather-related reasons (52% of the total dataset), linear baselines were removed from the spectra in all species by excluding regions where thc --f

96

C. Martin et al.: AST/RO: Inner 200pc

-200

-100

0 vLSK

100

-200

-100

0

100

200

( km s-I)

Fig. 1 Spectra toward the CO J = 7 + 6 peak in the SgrA (! = O", b = -0:07) and SgrB clouds 2!( = 0:6G, b = -0P05) in 3 different transitions, as indicated by the color idcntilications at upper left, and from top to bottom: the 461 GHz CO .I = 4 43, 807 GHz CO = 7 + 6, and the 492 GHz [C 11 3P1 + "Po transition.

CO J = 1 -+ 0 spectra showed emission greater than TL 3 1K . This allowed known emission in the Galactic Center region to be readily excluded from the baseline fitting procedure and was generally for a given reduced spectrum, this sufficient. In a few cases, usually due to higher than average T,.,,,, method fails and artifacts (e.g., vertical lines in the longitude-velocity maps) appear. While the original intent was to make T,,,, as uniform as possible across the entire map, this was not always possible. For the CO J = 4 i 3 transition, T,,,,, in 1 k m - ' wide channels with 2' spatial smoothing is on average 5 0.3K except in the region 1:s > t! > 1:s where T,.,,,, 5 0.8K.The [C I] 'PI -+ 3Po transition has T,,,,, 5 0.5 K in 1 kriis-l channels for the central region of lP0 > e > -0P5, T,,, 5 1.0K f o r e > lY0, and T.,,, 5 2 K f o r e < -0;s. Finally, for CO J = 7 + li (ignoring the occasional baseline feature), T,,,,, 5 0.8K in I kn1s-l channels for P > 0.7 and -0.5 > ?! > -0.8, and T,,,, 5 2 K elsewhere.

3 Data Presentation Sample spectra at the position of peak C O .J = 7 + 6 emission toward the Sgr A (t = 0P0,6 = -0707) and SgrB (C = 0166, h = -0P05) molecular complexes are shown in Fig. 1. All of the spectra are from datacubes smoothed to 2' resolution. The CO *J = 4 4 3 profile in SgrB has a prominent self-absorption and CO J = 7 + G lines show peak feature at VLSR = GO kms-', in contrast, thc [CI] "PI + "?I' emission at the self-absorption vclocity, suggesting these lines are less optically thick. It is important to note that the strong feature at negative velocities in the CO d = 7 + 6 spectrum is due to superposed emission in the 809 GHz ('Ppz +"PI) line of [C I] in the image sideband. Fig. 2 presents spatial-spatial (B. 6) maps integrated over velocity for the three transitions observed with ASTRO. All three maps have been smoothed to the same 2' spatial resolution. The most striking result is that CO ,J = 4 3 emission in the Galactic Center region is essentially coextensive with the emission

-

Astron. Nachr./AN 324, Yo. S1 (2003)

Fig. 2 Spatial-spatial

97

P (dcgrccs) (e, b ) integrated intensity maps for the 3 transitions observed with AST/RO. Transitions are

identified at left on each panel. The emission is intcgrated over all velocities where data are available. These values of ([I,,,,,, 7 ) m 7 r L ) are: [Cr], (-90,150); C0(7-6), (-30,120); CO(4-3), (-150, 150). All 3 maps have been smoothed to the same 2' resolution. from the lower J transitions of CO. This contrasts sharply with the outer Galaxy where CO ,J = 4 + 3 emission is rather less extensive than CO J = 1 i 0. In all three maps, the brightest emission occurs primarily at negative latitudes. Four major cloud complexes are seen in the maps, from left to right: the complex at !cx 1; 3, the Sgr B complex near F ru 0; 7, the Sgr A cloud near !ru OP0, and the Sgr C cloud near ( B cx -0P45, b ru -0P2). As noted by Kim et a1.(2002), the CO .I = 7 + 6 emission is much more spatially confined than the lower-J CO transitions. In contrast, the [CI] emission is comparable in spatial extent to the low-J C O emission, but its distribution appears somewhat more diffuse (less peaked). The Sgr C cloud is much less prominent in the [C I] map than in the other two transitions. The noise in the [C I] map at P < -0; 7 is greater than that in thc rest of the map due to shorter integration times.

4 LVGModel We use the LVG methodology of Goldreich & Kwan( 1974) to estimate the kinetic temperature, T kir l, and the number density of molecular hydrogen, ,,(Ha), throughout the Galactic Center region. Due to the high velocity dispersions characteristic of the Galactic Center, the LVG approximation is most likely valid over much of the mapped region. The LVG approximation does not apply to some of our data: the foreground absorption by spiral arms is clearly not local to the emitting gas i n the Sgr A cloud. It should be kept in mind that the LVG approximation is an Ansatz which allows us to estimate the physical properties of the Galactic Center gas in what otherwise would be an untenably complex modeling problem. Our cloud model was developed by M. Yan and S. Kim. It has plane-parallel cloud geometry. It uses C O collisional rates determined by Turner( 1995) and uses newly-derived values for the Hz ortho-to-para ratio (= 2) (Rodriguez-Fernfindez et aL(2000)) and for the collisional quenching rate of C O by Hz impact (Balakrishnan et aL(2002)). The model has two input parameters: the ratio of "CO to '"CO abundancc, and the ratio X(CO)/VV, where X ( C 0 )is the fractional C O abundance parameter and VV is thc velocity gradient. As discussed in 51, the abundance ratio 12CO/13C0is 24 in the Galactic Center region (Penzias( 198O),Langer & Penzias(l990),Wilson & Matteucci(l992),Langer & Penzias( 1993)). We will use a value X ( C O ) / O V= lop"." pc kiiip's, assuming that the "COIHz ratio is and the velocity gradient within the Galactic Center gas is a uniform 3 kms-' pep'. Dahmen et al.( 1998) estimated that VV is 3 killsp1 pcpl to 6 kriis-l p ~ 'and . indeed these are typical slopes of position-velocity features in Galactic Center maps. There is, however, no reason to suppose that a single value of VV applies throughout the Galactic Center region, or even that VV is constant within a single cloud. This is a weak

C. Martin et a].: AST/RO: Inner 20Opc

98

.--.

N

-

e (degrees)

M 3

Fig. 3 False color velocity-channel map of log[,f n , ( H z ) d v / V Vas ] determined by the LVG model described in the 150 < ULSR < -60 kn3s-l and 20 < VLSR < 150 krns-' in order to avoid contamination hy the foreground material for which the LVG analysis is invalid. This value is then divided hy VV in order to make a map comparahlc to the expected column density in units of c ~ I - ~ .

text. n(Hz) is integrated over the ranges

~

point in the analysis, because in the LVG analysis 7 ~ ( H z )x (0V)"-", when all other parameters are held fixed. For each observed point, we take the brightness temperature ratios T&/Ti53 and T;tn/T;?+,,,using the same methodology as Kim et a1.(2002), to determine Tkinand n ( H 2 ) . Note that our model, using the T&j/T& ratio, is particularly sensitivc to variations in density and teinperature near T,
-

N

5

Conclusions

1. We have mapped the inner 3" of the Galaxy in 461 GHz CO J = 4 + 3, 807 GHz CO J = 7 4 6 , P ' po emission. SgrA, SgrB, SgrC, and the 300pc molecular ring are and 492GHz [CI] 'PI easily identified.

2. For each observed point, T;56/Ti::3 line ratios, together with Ti:o/T:5n line ratios, were used to estimate kinetic temperatures and molecular hydrogen volume densities. Kinetic temperature was found to decrease from relatively high values (70 K) a1 cloud edges to low values (20 K) in the interiors. Molecular hydrogen densities, n,(H2),ranged up to the limit of our ability to determine via our LVG analysis, 10' c i ~ i - ~ . N

,'

-

3 . Typical pressures in the Galactic Center gas are n.(H2) . T k i n 10"' Kern-:'. This is two orders of magnitude larger than the pressure ofthe interstellar medium near the Sun. Acknowledgements We thank the receiver group at the U. of Arizona for their assistance; R. Schieder, J. Stutzki. and colleagues at U. Koln for their AOSs; J . Kooi and R. Charnberlin of Caltech, G. Wright of PacketStorm Cornrnunications, and K. Jacobs of U. Koln for their work on the instrumentation; M. Yan and S. Kim for the use of their LVG code; and J . Carlstrom for comments on the manuscript. This research was supported in part by the National Science Foundation under a cooperative agreement with the Center for Astrophysical Research in Antarctica (CARA), grant

Astron. Nachr./AN 324, No. SI (2003)

99

number NSF OPP 89-20223. CARA is a National Science Foundation Science and Technology Center. Support was also provided by NSF grant number OPP-0126090.

References Balakrishnan, N., Van, M., & Dalgarno, A. 2002. ApJ, 568,443 Bally, J., Stark, A. A.. Wilson, R. W.. & Henkel, C. 1987. ApJS, 65, 13 -. 1988, ApJ, 324,223 Bania, T. M. 1977, ApJ, 216, 381 Binney, J., Gerhard, 0. E., Stark, A. A., Bally, J.. & Uchida, K. I. 1991, MNRAS, 252, 210 Bitran, M., Alvarez, H., Bronfman, L., May, J., & Thaddeus, P. 1997, A&AS, 125,99 Bitran, M. E. 1987, PhD thesis, University of Florida Chamberlin, R. A,, Lane, A. P., & Stark, A. A. 1997, ApJ, 476, 428 Dahmen, G., Hiittemeister, S., Wilson, T. L., & Mauersberger, R. 1998, A&A, 33 I , 959 Genzel, R. & Townes, C . H. 1987, ARA&A, 25, 377 Goldreich, P. & Kwan, J. 1974, ApJ, 189, 441 Goldstein, S. J., Gundermann, E. J., Penzias, A. A,, & Lilley, A. E. 1964, Nature, 203, 65 Giisten, R., Walmsley, C. M., & Pauls, T. 1981, A&A, 103, 197 Heiligman, G . M. 1982, PhD thesis, Princeton University Honingh, C. E., Haas, S., Hottgenroth, D., Jacobs, K., & Stutzki, J. 1997, in Eighth International Symposium on Space Terahertz Technology, 92 Kim, S., Martin, C. L., Stark, A. A., & Lane, A. P. 2002, ApJ, 580, 896 Lane, A . P. 1998, in Astrophysics From Antarctica, ed. G. Novack & R. H. Landsberg, Vol. 141 (ASP Conference Series), 289 Langer, W. D. & Penzias, A. A. 1990, ApJ, 357, 477 -. 1993, ApJ, 408,539 Linke, R. A., Stark, A. A., & Frerking, M. A. 1981, ApJ, 243, 147 Liszt, H. S. & Burton, W. B. 1978, ApJ, 226,790 McGee, R. X. 1970, Aust. J. Phys., 23,541 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyaraki, A. 1998, ApJS, 1 18, 455 Penzias, A. A. 1980, Science, 208,663 -. 1981, ApJ, 249,518 Polk, K. S.. Knapp. G. R., Stark, A. A,, & Wilson, R. W. 1988, ApJ, 332, 432 Robinson, B. J., Gardner, F. E, vanDamme, K. J., & Bolton, J. G. 1964, Nature, 202,989 Robinson, B. J. & McGee, R. X. 1970, Aust. J. Phys., 23, 405 Rodriguez-Fernindez, N. J., Martin-Pintado, J.. de Vicente, P., Fuente, A,, Hiittemeister, S., Wilson, T. L., & Kunze, D. 2000, A&A, 356,695 Sawada, T., Hasegawa, T., Handa, T., Morino, J., Oka, T., Booth, R., Bronfman, L., Hayashi, M., Luna Castellanos, A., Nyman, L. .. Sakamoto, S., Seta, M., Shaver, P., Sorai, K., & Usuda, K. S. 2001, ApJS, 136, 189 Schieder, R., Tolls, V., & Winnewisser, G. 1989, Experimental Astronomy, 1, I01 Scoville, N. Z., Solomon, P. M., 62 Thaddeus, P. 1972, ApJ, 172, 335 Stark, A. A., Bally, J., Balm, S. P., Bania, T. M., Bolatto, A. D., Chamherlin, R. A,, Engargiola, G., Huaiig, M., Ingalls, J. G., Jacobs, K., Jackson. J. M., Kooi, J. W., Lane, A. P., Lo, K.-Y., Marks, R. D., Martin, C. L., Mumma, D., Ojha, R., Schieder, R., Staguhn, J., Stutzki, J., Walker, C. K., Wilson, R. W., Wright, G. A., Zhang, X., Zimmermann, P., & Zimmermann. R. 2001, PASP, 113, 567 Stark, A. A., Bally, J., Gerhard, 0. E., & Binncy, J. 1991, MNRAS, 248, 14P Stark, A. A., Bally, J., Knapp, G . R., & Wilson, R. W. 1988, in Molecular Clouds in the Milky Way and External Galaxies, ed. R. L. Dickman, R. L. Snell, & J. S. Young (Springer-Verlag), 303 Stark, A. A., Bally, J., Wilson, R. W.. & Pound, M. W. 1989, in IAU Symp. 136: The Center of the Galaxy, ed. M. Morris (Kluwer), 129 Stark, A. A., Chamberlin, R. A,, Cheng, J., Ingalls, J., &Wright, G. 1997. Rev. Sci. Instr., 68, 2200 Turner, B. E. 1995, ApJ, 455, 556 Walker, C. K., Kooi, I. W., Chan, M., Leduc, H. G., Schaffer, P. L., Carlstrom, J. E., & Phillips, T. G . 1992, Int. J. Infrared Millimeter Waves, 13, 785 Wilson, T. L. & Matteucci, F. 1992, A&A Rev., 4, 1

This Page Intentionally Left Blank

Astron. Nachr.iAN324. No. SI. 101 - 107 (2003)/ DO1 10.1002/asna.2003851I 1

Gravitational Stability of Molecular Clouds in the Galactic Center Tomoharu Oka*' and Tetsuo Hasegawa**2

' Research Center for the Early Universc and Department of Physics, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033,Japan. National Astronomical Observatory of Japan 2-21-1 Osawa, Mitaka, Tokyo 181-8588, Japan.

Key words ISM, molecular clouds Abstract. The data from the Nobeyama Radio Observatory 45 m telescope Galactic Center CO survey have been analyzed to generate a compilation of molecular clouds with intense CO emission in this region. The measured parameters of identified clouds are analyzed and cross-correlated to compare with those of clouds in the Galactic disk. We diagnosed gravitational stability of identified clouds using their viral theorem mass and CO luminosity. If we assume that the disk clouds are nearly at the onset of gravitational instability, all the clouds and cloud complexes in the Galactic center must be gravitationally stable. This indicates that the GC clouds could be in equilibrium with high pressure in the Galactic center environment. The velocity dispersion of a cloud correlates inversely with the degree of gravitational instability. Gravitationally less Stdbk clouds follow the main ridge of intense CO emission, part of which define two rigidly-rotating molecular arms. It is concluded that mechanisms such as orbit crowding at a resonant radius could promote dynamical evolution of clouds and thereby trigger subsequent star formation.

1 Introduction The spatial structure and dynamical properties of molecular clouds could be important clues to understand the mechanisms of formation and evolution of molecular clouds, and formation of stars within them. The identification of molecular clouds in the Galactic disk has shown their size and mass spectra, as well as scaling laws of the type ITV 'x and hfv, rx LCO (Scoville et al. 1987; Solomon et al. 1987). The tight correlation between virial theorem mass MVTand CO luminosity LCOconsidered as evidence for virialization, and has been used to derive the CO-to-Hz conversion factor (Young & Scoville 1991). Molecular gas is concentrated in the central 500 pc of the Galaxy, exhibiting highly complex distribution and kinematics as well as remarkable variety of peculiar features (Oka et al. 2001a, 2001b). Molecular clouds in the Galactic center (hereafter, GC clouds) are characterized by large velocity widths, high temperature, and high density, which differ considerably from those in the Galactic disk. It must be meaningful to compare the gross properties of the GC clouds with those of the disk clouds. However, few studies of cloud identification have been done for the Galactic center region, since the region is highly crowded and thus the identification by a uniform criterion i s difficult. Oka et al. (1998a) manually identified 15 large GC clouds using the CO J=2-1 survey data with 9' resolution, and found that the r r ~ / S "and . ~ MVT/LCO ratios far larger than those of the disk clouds. From the unusually high MVT/LCOratio, they concluded that identified large molecular clouds are in pressure equilibrium with ambient hot gas andor magnetic field. Miyazalu & Tsuboi (2000) also concluded that * Corresponding author: e-mail: tomoQphys.s.u-tokyo.ac.jp. Phone: +81 35841 4217, Fax: +81 35841 4178 * * Corresponding author: e-mail: [email protected],Phone: +XI 422 34 3780, Fax: +81422 34 3764

@ 2003 WREY-VCH Vrrlag GmhH & C o KC'A. Weinhemi

T. Oka and T. Hasegawa: Gravitational Stability of Molecular Clouds

102

strong magnetic field in the Galactic center region bind the dense molecular clouds identified in their CS

J=1-0 survey data. Although CO J=1-0 spectra of the Galactic center suffers severe contamination of foreground gas in the Galactic disk, cloud identification in those data is still crucial to discuss the effects of environment to cloud conditions and evolution, since a number of such studies have been done for clouds in the Galactic disk and extragalaxies based on CO J=1-0 data. The Nobeyama Radio Observatory 45 m telescope key program, 'A Large-Scale CO Imaging of the Galactic Center' (Oka et al. 1998b),provided extensive highresolution data set which is ideal for studies of morphology and statistical properties of clouds. In this data set, we identified 165 molecular clouds (hereafter '45 m GC clouds'). The measured parameters of identified clouds are analyzed and cross-correlated to compare with those of disk clouds. We diagnosed gravitational stability of identified clouds using their viral theorem mass and CO luminosity.

0

00

0 Ooo 0

-2001. 2.0

8

a

3

1

1.5

,

fi

0

1

1.o

,

0

8

I

0.5

I

'

8

8

1

'

"

0.0

'

1

-0.5

'

"

1

'4

-1.0

Galactic Longitude [degrees] Fig. 1 The 1-h and I-V distributions of the clouds identified in the 45 m data set. Sizes of circles denote their CO luminosities, LCO = lo"', lo5-", 106-7 K km s-' pc2, with increasing size, respectively. Broken lines denote the I-V loci of the Galactic arms.

Astron. Nachr./AN 324, No. S I (2003)

103

2 Cloud Identification Clouds are defined as topologically closed surfaces of antenna temperature in the (1. h. V )space. We took three boundary threshold intensities Tn,in= 5.7.5.10 K, and required minimum peak temperatures twice those of the thresholds, T p e & 2 10,15,20 K, respectively. Details of the identification procedure as well as the effect of blending of unrelated clouds were described in Oka et al. (2001~).98 clouds were identified with the Tnlin= 5 K threshold, 45 clouds with the T,,,,,, = 7.5 K threshold, and 22 clouds with the Z,,ill= 10 K threshold. Clouds with peak velocities between -60 and +20 km s-l are excluded from the analyses to avoid the contamination. The dispersions were computed from the intensity-weighted variances over all elements T(2.b. V ) within the clouds. Explicitly, 0,.= (&T2)*", S = T.c/ T , s = 2 . h. V. The total CO luminosity,

LCO= D2

]I/

Tidl2dV,

(1)

is obtained by integrating over all elements T(I.b. V )within the cloud boundary. D is the distance to the cloud (in our case D = 8.5 kpc). The measured dispersions and the total CO luminosities have been corrected for the effect of truncation using simulated clouds with triaxial Gaussian shapes. The size parameter S is defined by

a).

S = Dt,an(

(2)

A virial theorem mass of a purely gravitationally-bound cloud is calculated by

where f , is a projection factor (we employed f j j = 2.9; Solomon et al. 1987).

3 Cloud Properties 3.1

Cloud Sizes and Velocity Dispersions

Figure 2 shows a plot of size versus velocity dispersion for the GC clouds with the same plot for the disk clouds. The GC clouds reside above the S-cr~,line of the disk clouds, ov = l.OSO-"(Solomon et al. 1987)). A linear least-squares regression to the logs using the 45 m GC clouds with T,,,i,, = 5 K leads to the result 0v = 4.211.1 ~ 0 . 4 0 1 0 . 1 0 (4) (km spl). -0.9 where S is in parsecs, and the uncertainties are lrr. The exponent of the S-0"relation i s the similar to that of the disk clouds (Y 0.5). If we fix the exponent lo be 0.5 in the regression, we get the size-line width coefficient (rr"/S".') 3.8 times larger than that of the disk clouds.

3.2

CO Luminosities and Virial Theorem Masses

Figure 3 shows a plot of CO total luminosity versus virial theorem mass for the GC clouds with the same plots for the disk clouds. Again, all the identified GC clouds reside above the LCO-MVTline of the disk clouds, MVT= 3 ' 3 ( L c 0 ) ~The , ~ ~L. c o - M ~ lloci ~ l . of less luminous GC clouds are closer to the disk clouds, while they shift upward with increasing LCO. A linear least-squares regression to the logs using the 45 m GC clouds T,,,,, =5 K leads the result hfVT= 1141:p

( L ~ ~ ) O . R X * O . ~ ~(Aft>)

( 51

where LCOi s in K km spl pc2, and the uncertainties are 1c. The exponent of the LCO-MVTrelation is the similar to that of the disk clouds (= 0.81).

T. Oka and T. Hasegawa: Gravitational Stability of Molecular Clouds

1 04

I o9

100

I ox 10' c

0

lo6

2

3

10'

I o4

0. I 0.1

10

1

s

100

LCO [K km s-1 pcz]

[PCI

Fig. 2 A S-uv plot is shown for the clouds identified both in the 45 m (filled polygons) and 1.2 m (open circles) data sets, with the same plots for the disk clouds (dots) and the large GC clouds identified manually in the CO J=2-1 survey data (crosses). Shapes of filled polygons indicate boundary intensities, T,,,,, = 5 K (filled circles), 7.5 K (filled rhombus), and 10 K (filled triangles).

4

Fig. 3 A LCO-MVTplot is shown for the clouds identified both in the 45 m (filled polygons) and 1.2 rn (open circles)data sets, with the same plots for the disk clouds (dots) and the manually identified large GC clouds (crosses). Shapes of filled polygons indicate boundary intensities, T,,,, = 5 K (filled circles), 7.5 K (filled rhomhus), and 10 K (filled triangles).

Discussion

4.1 Cloud in Equilibrium with External Pressure The larger MVTvalues of the GC clouds than those of the disk clouds with the same LCOcan not be solved by introducing a large CO-to-Hz conversion factor, since y-ray (Blitz et al. 1985), far-infrared (Sodroski et al. 1995), and X-ray observations (Sakano et al. 1997) suggest CO-to-HZ conversion factors in the Galactic center smaller than the standard value. ~ can be understood in terms of model clouds in pressure The large variations in the L ~ o - b f "plot equilibrium with an intercloud medium (Oka et al. 1998a). The virial equation for a spherically symmetric non-magnetic cloud of mass M and radius R, embedded within an intercloud medium of pressure p can be written as 1 dzI -= 3a;M 2 dt2

~

G M ~ - 4nR"p R

a-

where I is the generalized moment of inertia of the cloud, ov is the velocity dispersion of turbulent motion (thermal motion has been neglected), and a is a dimensionless coefficient of order of unity. The virial equation can easily be extended to the case of a magnetized cloud by substituting a to a,if = a[l (@/@cr)2(1 - R/Ro)](Nakano 1998). In the equilibrium state, d 2 1 / d t 2=O, we have two equilibrium masses ,

where j 3 = 16.rraR2p/9a$ is a dimensionless parameter, and Mo = YRrr$/G is the commonly used virial theorem mass which is the same as Afv, in eq.(3). /? is also expressed by three dynamical parameters as

105

Astron. Nachr./AN 324,No. S 1 (2003)

( P / T ) ,where U is the gravitational potential energy, T is the internal kinetic energy, and P = 47rR3p . It is easily confirmed that when the gas behaves isothermally, an equilibrium state with a+

[ j = ( / U I / T )x

(> 1/2) is gravitationally unstable and that with

(1- (< l/Z) is gravitationally stable. This N denotes the degree of gravitational instability. For an opaque cloud with uniform CO brightness of TCO, the total CO luminosity can be expressed as Lpo = 27rR2Tcoo1,,. From this expression, and equating the virial mass &fvT = trl& to (4/3)7rR3pp, where p is the mean mass density, we get the linear Lco-A'lv~ relation,

Introducing the CO-to-Hz conversion factor X = N(Hz)/Ico, on the other hand, the total molecular mass including the helium correction 1.36 is written as

M

= 2.2 x 1 0 - 2 " x L c o

(&J)

(10)

where Lc0 is in K km pc2. Using the virial mass Afv, = a h f , as the total molecular mass, and comparing eqs.( I 1) and (12), we get

x

=

a+xo,

X o stays constant when rewritten as Mo

= 2.2 x

r

)

p x / T c ~ doe5 not vary in the mean from cloud to cloud. Thus eq.( 1 1 ) is

lO-*"tr-fXoLco.

The large variation in the LCO-MVT relation for each region or scale. 4.2

(1 3 ) in

eq.( 15) is explained by choosing different

(Y

in the eq.( 15)

Gravitational Stability

A theoretical study has shown that interstellar clouds which are close to gravitational instability exhibit precisely the same scaling laws, provided that they are in equilibrium with a constant pressure environment (Chikze 1987). The disk clouds actually follow the same scaling laws, defining a base of highly scattered distributions (Figs.2, 3). The paucity of clouds below the Lco-MvT trend of the disk clouds suggests that such clouds are gravitationally unstable ((Y > 1/2) and may have moved to stable zone by contraction or have collapsed to form new stellar clusters. Assuming that the disk clouds are nearly at the onset of (a linear relation between R a n d gravitational instability, ( Y = 1/2, and using AJxJ, defined by eq.(3) as M,, cm-2 (K km s-')~'. This gives a the size parameter, S, is implicitly assumed), we get X O= 1.6 x CO-to-Hz conversion factor X = 1.2 x 10'" cmP2 (K km s-l)-' for the disk clouds, which is smaller than the standard value by a factor of 3, but is close to that derived from the y-ray observations of the Orion molecularcloud, X=1.06 x em-' (K km s - ' ) ~ ' (Digel, Hunter, & Mukherjee 1995). Although we do not have compelling evidence that the disk clouds are nearly at the onset of gravitational instability, diagnosis of gravitational stability by the parameter (1 must be meaningful in a reiative sense. We diagnose the degrees of gravitational stability of the CC clouds by (Y using X" = 1.6 x 102" cm-2 ( K km s - l ) - ' . All but one G C clouds have (1 smaller than 1/2, which suggests they are gravitationally stable. They might be in equilibrium with the external pressure of intercloud medium. This fact may explain the paucity of star formation activity in this region. We show the 1-V distributions of the GC clouds with three a ranges separately (Fig.4). Clouds with (Y > 0.01 follow the main ridge of intense CO emission, part of which defines two rigidly-rotating molecular

106

T. Oka and T. Hdsegawa: Grrlvitdtional Stability of Molecular Clouds

arms (Sofue 1995). These molecular arms are associated with a ring of HI1 regions, which is called the star forming ring (Oka et al. 1996). On the other hand, clouds with cy < 0.01 spread over the 1-V plane irrespective of the location of the Sofue’s molecular arms. quit

-v I)

E

200 t

-I

100

Y

0 -100 -200

-*.*

2.0

1.5

1.0

0.5

0.0

4.5

Galactic Longitude [degrees]

j -1.0

.

0

1.5

1.0

0.5

0.0

j

2‘ ”

-0.5

’

-1.0

Galactic Longitude [degrees]

Fig. 4 The l-V distribution of the 45 m GC clouds with ( a ) a 2 0.01 and (b)a < 0.01. Shaded areas denote Sofue’s molecular arm I-IV (Sofue 1995).Broken lines show the 1-Vloci of the Galactic a r m .

Figure 5 shows the correlation plots between basic cloud parameters and their degrees of gravitational = 5 K sample of the 45 m GC clouds. The largest five clouds with S 2 20 pc, instability for the Trnin which might be categorized as cloud complexes, were indicated by open circles. No correlation is seen in the S - n or Lc0-a plots, while gv correlates inversely with except for the largest five clouds. A anti-correlation in h f ” ~ - plot n is a result of the r v - u anti-correlation. These suggest that the degree of gravitational instability of the GC clouds is intimately related to the velocity dispersion. Neither size nor mass is the determining factor for gravitational instability. These results imply that gravitational instability of the GC clouds grows with dissipation of velocity dispersion. We conclud that dissipation processes associated with the star forming ring promote gravitational instability of clouds and subsequent star formation. Mechanisms such as orbit crowding at the inner Lindblad resonance may play a role. lkuta & Sofue (1997) studied kinematic properties of giant molecular clouds (GMCs) associated with star forming regions in the Galaxy, and found that the star formation efficiency ( S F E ) is inversely correlated with the velocity dispersion and the virial mass of a GMC. Kohno et aL(1999) found, in the central region of the barred spiral galaxy NGC 6951, an anti-correlation between the dense molecular gas fraction and the gas velocity dispersion. They claimed the formation of dense molecular gas is caused by gravitational instability, which is intimately related to the dissipation of velocity dispersion. These results also imply that the dissipation of random motion within a GMC promotes the gravitational instability and subsequent star formation.

5

Conclusions

Our analyses of the large-scale CO data by the NRO 45 m telescope have yielded samples of molecular clouds in the Galactic center region. The major results of our analyses are the following: 1 . All clouds and cloud complexes in the Galactic center, exept for one cloud, are gravitationally stable, being in equilibrium with the external pressure of intercloud medium. 2. Less stable clouds in the Galactic center concentrate on the Galactic plane, following the main ridge of intense CO emission, part of which defines the two rigidly-rotating molecular arms.

Astron. Nachr./AN 324, No. S 1 (2003)

107

Fig. 5 Relations between size, velocity dispersion, CO luminosity, virial mass, and the parameter u1 for the T,,,,, = 5 K sample of the 45 m GC clouds. Open circles indicate the largest 5 clouds with S > 20 pc, which might be categorized as cloud complexes.

3. The velocity dispersion of a cloud correlates inversely with the degree of gravitational instability. Dissipation processes associated with the star forming ring may promote the gravitational instability of molecular clouds and subsequent star formation. telescope key program,

References Blitz, L., Bloemen, J. B. G. M., Hermsen, W., & Bania, T. M. 1985, A&A, 143, 267 Chi&, J. P. 1987, A&A, 171,225 Digel, S. W., Hunter, S. D., & Mukherjee, R. 1995, ApJ, 441,270 Ikuta, C., & Sofue, Y. 1997, PASJ, 49, 323 Kohno, K., Kawabe, R., & Vila-Vilaro, B. 1999, ApJ, 51 1, 157 Miyazaki, A,, & Tsuboi, M. 2000, ApJ, 536,357 Nakano, T. 1998, ApJ, 494,587 Oka, T., Hasegawa, T., Hayashi, M., Handa, T., & Sakamoto, S. 1996, ApJ, 460, 334 Oka, T., Hasegawa, T., Hayashi, M., Handa, T.. & Sakamoto, S. 1998a. ApJ, 493,730 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyazaki, A. 1998b, ApJS, 118, 455 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyazaki, A. 2001a, PASJ, 53, 779 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., & Miyazaki, A. 2001b, PASJ, 53, 787 Oka, T., Hasegawa, T., Sato, F., Tsuboi, M., Miyazaki, A., & M. Sugimoto 2001c, ApJ, 562, 348 Sakano, M., et al. 1997, IAU Symp. 184, The Central Regions of the Galaxy and Galaxies, 227 Scoville, N. Z., Yun, M. S., Clemens, D. P., Sanders, D. B., & Waller, W. H. 1987, ApJS, 63, 821 Sodroski, T. J., et al. 1995, ApJ, 452,262 Sofue, Y. 1995, PASJ, 47,527 Solomon, P. M., Rivolo, A. R., Barret, J., & Yahil, A . 1987, ApJ, 319, 730 Young, J. S., & Scoville, N. Z. 1991, ARA&A, 29, 581

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. SI. 109- 115 (2003)/ DO1 10.1002/asna.200385056

Spectroscopy of Hydrocarbon Grains toward the Galactic Center and Quintuplet Cluster J.E. Chiar*'.*,A.J. Adamson3,D.C.B. Whittet4,and Y.J. Pendleton' I

NASA Ames Research Center, Mail Stop 245-3. Moffett Field, CA 94035

* SET1 Institute, Mountain View, CA 94043

Joint Astronomy Centre, 660 N. A'ohoku Place, University Park, Hilo, Hawaii 96720 Rensselaer Polytechnic Institute, Department of Physics, Applied Physics, and Astronomy, Troy, NY 12180

Key words extinction, interstellar medium dust and molecules, circumstellar dust Abstract. Our view of the Galactic center (GC) is affected by extinction from both diffuse interstellar medium (ISM) dust and dense molecular clouds along the line of sight. The enormous visual extinction present toward the center of our Galaxy (- 31 magnitudes) necessitates a study of the interstellar dust properties as well as an investigation into the distribution of the different dust components. We have built upon the historic spectroscopy o f Willner et al. (19791, Butchart et al. (1986), and McFadzean et al. (1989) in order to investigate the distribution of these dust components across the GC field. Specifically, we employ spectroscopy in the 3 pm region to investigate absorption features at 3.0 pm and 3.4 pm in lines of sight toward the GC central cluster and the Quintuplet cluster to the northeast. The 3.4 pm feature is one of the primary spectral signatures of the organic component of interstellar dust and is, to date, only observed in the cold diffuse interstellar medium. The 3.0 ice feature is carried by dense molecular cloud material, and can therefore be used to loosely trace the distribution of such material across the GC field. By obtaining spectra for multiple sightlines we have been able to deconvolve the diffuse ISM and dense molecular cloud components. Our study shows that differences exist in the spectra of relatively nearby lines of sight in the Galactic center central cluster. The depth o l the 3.0 pm water-ice feature varies by a factor of almost 5 across a 2 parsec (in projection) region, perhaps reflecting the clumpy nature of the dense clouds. In addition, we found that the 3.4 pm hydrocarbon feature varics in depth across the areas studied toward the central cluster, whereas the depth is relatively constant toward the Quintuplet cluster. This is likely a reflection of the distribution of extinction from the foreground diffuse ISM. Our ground-based and space-based spectroscopy reveals differences in absorption features in the 3 and 6 pm regions between sightlines toward the GC central cluster and those toward the Quintuplet cluster. While the 3 pm spectra of both regions show a broad absorption feature blueward of the 3.4 pm absorption, only the Quintuplet spectra show a distinct absorption feature at 3.28 pm. This feature is indicative of the presence of polycyclic aromatic hydrocarbons (PAHs) along the line of sight. The Quintuplet-proper sources have 6 p m spectra that are markedly different than that of GC IRS 7 in the central cluster, and instead strongly resemble the spectra seen toward dusty late-type carbon-class (WC) Wolf-Rayet stars. This is the first hint of some spectroscopic similarity between the Quintuplet sources and dusty WC stars.

1 Introduction T h e center of our Galaxy is obscured by some 30 magnitudes of visual extinction. Investigations of objects located near the Galactic center (GC) require an understanding of the physical properties of the intervening dust through which all observations are made. Some of the dust is local to the G C , while much of it lies along the line of sight. Bright infrared sources located at the G C are used to probe this dust. T h e line of sight toward the G C is dominated by diffuse ISM dust (Lebofsky 1979), however it was suggested by * Correaponding author: e-mail: chiarQmisty.arc.nasa.gov, Phone: + I 6506040324, Fax: +1 6506046779 0xm WILEY VCH V C T I ~G~m m

co K G A wrlnhrlt~~

Chiar et al.: Spectroscopy of Dust Components toward the Galactic Center

110

McFadzean et al. (1989) that there is a molecular cloud component that contributes to the extinction toward the GC. Space-based spectroscopy with the Infrared Space Observatory (ISO) confirmed the presence of molecular cloud material by revealing absorption due to icy grain mantles (de Graauw et al. 1996; Lutz etal. 1996; Gerakines et al. 1999). It has been estimated that as much as one-third of the visual extinction arises in molecular cloud material (Whittet et al. 1997). Much of the molecular cloud extinction presumably arises in dust clouds located within 4 kpc of Earth (Sanders, Scoville & Solomon 1985). If this is so, then the clouds are not associated with the infrared sources that provide the continuum against which molecular absorption features can be observed. The diffuse interstellar medium is devoid of ices and instead contains only the refractory grain component, which includes aromatic [ring-like) and aliphatic (chain-like) hydrocarbons and silicates (Figs. 1 and 4). Hydrocarbons in the form of aromatics and aliphatics are a significant component of the diffuse ISM (e.g., Pendleton & Allamandola 2002). Aromatic hydrocarbons, observed throughout our Galaxy and other galaxies, are characterized by a family of bands normally observed in emission around 3.3, 6.2, 7.7, 8.6, 1I .3 and 12.7 pm (Allamandola, Tielens, & Barker 1989). Short-chained aliphatic hydrocarbons [Fig. 4, rightmost insert) are characterized by absorption at 3.4 pm and subfeatures due to CH2 (methylene) and CH3 (methyl) stretching modes at 3.38 and 3.48 pm (methylene) and 3.42 pm (methyl) (Sandford et al. 1991). Their spectral signature is seen not only toward the GC (Butchart etal. 1986), but in diffuse ISM dust throughout our Galaxy (Pendleton et al. 1994) and other galaxies (Wright et al. 1996). The 3.4 pm absorption feature observed in the diffuse ISM is distinct from that seen in dense molecular clouds (Brooke et al. 1996; Chiar et al. 1996). The latter is a smooth feature centered at 3.47 pm and the camer is thought to coexist with H2O-ice in the grain mantle. In contrast, the diffuse ISM hydrocarbons are most likely carried by a population of very small unaligned grains, rather than refractory mantles on silicate cores (Chiar et al. 1998; Adamson et al. 1999; Ishii et al. 2002), although the interpretation of these results remains controversial (Li & Greenberg 2002). I

I

I

I

I

I

! I Hydrucarbons

Dense Cloud Material

------

Diffwe Interstellar Medlurn

t

!

I 5

I 10

1

Wavelength ( p i )

Fig. 1 ISO-SWS spectrum centered on CC IRS 7. The beam size was 14” x 20“. Absorption features arising in the diffuse ISM and dense molecular cloud material are noted. Figure adapted from Lutz et al. 1996 ‘and Chiar et al. 2000.

2 Dense Cloud Absorption Features Ices in the cold molecular cloud material along the GC line of sight are characterized by absorption features due to HzO (3.0, 6.0 prn), COz (4.27 and 15.3 pm), and CH4 (7.67 pm) (Chiar etal. 2000; Gerakines et al. 1999; de Graauw etal. 1996; Lutz etal. 1996). The Short-Wavelength Spectrometer (SWS) on IS0

Astron. Nachr./AN 324, No. S1 (2003)

111

h

0

In

0 ) --28

59 00

-29 00 00

17 42 34 17 42 32 17 42 30 17 42 28 17 42 26

Right Ascension (1950)

Fig. 2 Positions of infrared sources overlaid on a map o f HCN J= 1-0 emission (solid contours) and ionized gas at IS GHz (dottedcontours) [Reproduced from Giisten et al. 1987.1. HCN emission traces the high density molecular gas in the circumnuclcarring. Contour interval for the velocity-inlegratedHCN emission is 0.15 K averaged over 300kms-’ i n a 2” beam. Dotted contours of 15 GHz emission at 20 K intervals in a 3.6” x 3.4” beam. The radio-source SgrA* is located at cr = 17”42’”29.2” 6 = -28’59’19’’ The GC Quintuplet sources are located 14’ northeast of the GC, 30 pc away (assuming a distance of 8.5 kpc)

-

observed the line of sight toward GC IRS 7 with a 14” x 20” beam. The spectrum from 2.6 to 12.5 pm is shown in Fig 1. Absorption features due to solid C 0 2 and CH4 are present; analysis of the profiles showed that the abundances of these molecules relative to HsO-ice are similar to those observed in local molecular clouds (Boogert et al. 1998; Gerakines et al. 1999; Chiar et al. 2000). The 6.0 y m absorption feature has been previously studied with the Kuiper Airborne Observatory (using FOGS and HIFOGS: Willner et al. 1979; Tielens et al. 1996), then later by ISO-SWS (Chiar et al. 2000; Fig. 5 ) . There is some controversy surrounding the precise identification of the 6 pm profile. although it is generally accepted that it can at least be partially attributed to HyO-ice with possible trace amounts of other ices (Chiar et al. 2000). A comparison between the 6 prn absorption profile observed toward G C IRS 7 and the Quintuplet-proper sources is shown in Fig. 5 and discussed in section 4. While airborne and space-based observations enabled us to study the average ice properties along the line of sight, spatial information could not be ascertained. The ISO-SWS observations, centered on GC IRS 7, were made with a large beam and included the sources GC IRS 1, IRS 3, much of the ionized bar and the ionized northern arm (Lutz et al. 1996). The KAO observations were carried out with a similarly large beam with FOGS and HIFOGS ccntered on GC IRS 3 (Tielens et al. 1996). In order to gain insight into the distribution of diffuse ISM dust and dense cloud material, we undertook a program of 3 Itm spectroscopy of the positions shown in Fig. 2 (Chiar et al. 2002) at the United Kingdom Infrared Telescope (UKIRT) using the the 0.6” slit of the Cooled Grating Spectrometer, CGS4. The 3 pm region includes absorption features due to H20-ice and aliphatic and aromatic hydrocarbons. Aliphatic hydrocarbons reside in the diffuse ISM and are discussed further below. Water-ice is the most abundant mantle constituent in any dense cloud, so we use it to trace the dense cloud material.

112

Chiar et al.: Spectroscopy of Dust Components toward the Galactic Center

Our spectroscopy shows that the HzO-ice profile shape is remarkably consistent across the seven lines of sight studied, including GC IRS 8 and IRS 19, which are located apart from the central cluster, closer to the GC circumnuclear ring (Fig. 2; Geballe et al. 1989). Compared to local molecular cloud ice profiles, e.g., in Taurus, the GC profile is broader and peaks at shorter wavelengths (Fig. 3). Laboratory spectra of pure HaO-ice are not able to account for the observed absorption in the GC ice profile; additional absorption is present shortward of 3 pm and from 3.2 to 3.6 p m (in addition to the 3.4 pm hydrocarbon feature). The possibility that the blue excess is due to NH3-ice has been discussed by Chiar et al. (2000), although other explanations may be plausible (Dartois & d'Hendecourt 2001). We find that the optical depth of the ice band is the greatest toward GC IRS 19 ( 7 3 . 0 = 1.5), a factor of almost 5 greater than the weakest ice band (Chiar et al. 2002). The variation of the 3.0 pm profile depth across the central GC field has been noted previously by McFadzean et al. (1989). We also obtained a 4.5-5 pm spectrum of the line of sight toward GC IRS 19, to search for evidence of solid CO absorption, since it has the deepest H2O-ice band (and therefore a high abundance of icy grain mantles). In addition to unresolved gas-phase CO lines, our spectrum shows a weak solid CO feature at 4.67 pm and an X-C=N feature at 4.62 pm (Chiar et al. 2002). The solid CO to HzO-ice ratio is similar to that observed in local molecular clouds. In the solar neighborhood, the X-C-N feature is seen only toward some deeply embedded protostars. Toward the GC, it may indicate the serendipitous presence of such an object in the line of sight to IRS 19, or it might conceivably arise from the processing of ices in the circumnuclear ring of the GC itself. rn'

1

2.6

,

I

'

1

,

2.8

'

I

1

,

'

I

1

,

'

I

1

3.0 3.2 34 Wavelength (pm)

,

' I l l

I

1

1

1

3.6

Fig. 3 Comparison of the 3 pm ice feature observed in the local Taurus molecular cloud (toward the background star, Elias 16; Smith et al. 1993; dashed line) and in the molecular clouds along the line of sight toward the Galactic center (KO-SWS spectrum from Chiar et al. 2000; solid line). The absorption feature centered at 3.4 pm in the GC spectrum is indicative of aliphatic hydrocarbons in the diffuse ISM (see text).

0.5 NU

1. Ammatic

network

2. Aliphatic bridge

Fig. 4 The two classes of hydrocarbons thought to be present in the diffuse interstellar medium. Aliphatic (chain-like) hydrocarbons exhibit C-H stretching vibration modes near 3.4 pm, whereas the C-H stretching vibration of aromatic (ringlike) hydrocarbons is centered near 3.3 pm. Adapted from Pendleton & Allamandola 2002

Astron. Nachr./AN 324, No. S1 (2003)

I13

3 Diffuse Interstellar Medium Absorption Features The structured 3.4 pm absorption feature (e.g., Fig. 5 , left panels) has been the focus of many laboratory investigations into the exact nature of the hydrocarbon material (see Pendleton & Allamandola 2002 for a review). While many laboratory analog materials have provided insight into the carrier of the interstellar band based on absorption signatures at 3.4 pm, longer wavelength spectroscopy obtained from space using ISO’s SWS revealed vital information regarding the corresponding deformation modes at 6.85 and 7.25 pm toward the GC (Chiar et al. 2000). The relative strengths of these three features (Chiar et al. 2000), along with a detailed analysis of laboratory data produced via competing processes, have revealed that hydrogenated amorphous carbon produced through plasma processing, closely matches the interstellar data (Pendleton & Allamandola 2002). Our UKIRT-CGS4 spectroscopy of the GC central cluster sources, described i n section 2, shows that the depth of the aliphatic hydrocarbon absorption at 3.4 p m varies by a factor of I .7, indicating significant changes in the foreground extinction across the small field. Our spectroscopy of multiple sightlines allowed for deconvolution of the diffuse ISM absorption from the dense cloud absorption component (Chiar etal. 2002). Many previous studies of the 3.4 pm feature relied on fitting a local continuum over the 3.3 to 3.7 pm region. Our method helped us uncover broad absorption on the blue shoulder of the 3.4 pm absorption feature in the approximate spectral region where polycyclic aromatic hydrocarbons (PAHs; Fig. 4, left insert) are expected to absorb. The depth of the absorption does not correlate with the depth of the ice band, thus it is a characteristic of the diffuse ISM dust along the line of sight (Chiar etal. 2002). However, the width of the “feature” is too great to be simply reconciled with PAHs in solid grain material. The nature of this broad absorption is still an open question. Our combined UKIRT-CGS4 and ISO-SWS spectroscopy in the 3 and 6 p i regions revealed significant differences between the spectra of the GC central sources and the Quintuplet sources (Fig. 5 ) . For instance, in addition to the broad shoulder described above, a distinct narrow 3.28 Irm absorption feature is present in the Quintuplet cluster spectra (Chiar et al., in preparation), but is (probably) absent in the spectra of the GC central cluster sources. The central wavelength and width of the absorption feature are indicative of the C-H stretching vibration in PAHs. Whether the carrier of the absorption is intrinsic to the Quintuplet cluster sources or a widespread diffuse ISM dust component is unclear. The 6 pm spectra of the Quintuplet-proper sources exhibit an absorption feature centered at 6.2 pm, markedly different than the symmetric absorption feature present in the GC IRS 7 spectrum. We discuss a possible explanation for these profile differences below.

4 The Enigmatic Quintuplet Sources The five bright Quintuplet sources were discovered in near-IR surveys by Okuda et al. (1990) and Nagata ct al. (1990). We will refer to these original sources as the Quintuplet-proper sources. The surrounding cluster was revealed by later surveys (e.g., Moneti et al. 1992), and most recently by the Hubble Space Telescope which revealed hundreds of sources (Figer et al. 1999). Some of these stars have been classified as Wolf-Rayet stars, Luminous Blue Variables and OB supergiants (Figer, McLean, & Morris 1999). However, the nature of the Quintuplet-proper sources remains uncertain due to non-detection of pholospheric features which would enable their spectral classification. Each of the proposed identifications - massive dust-enshrouded young stars, OH-IR stars, dusty late-type carbon-class (WC) Wolf-Rayet stars - has its problems (Nagata etal. 1990; Figer, McLean, & Moms 1999; Moneti etal. 2001). We favor the dusty late-type WC star hypothesis. Figer, McLean, & Moms (1999) were the first to suggest that the Quintuplet sources may be dusty late-type WC stars. Moneti etal. (2001) analyze the spectral energy distributions (SEDs) of the Quintuplet sources and find that they are best reproduced by disk models similar to those used by Williams et al. ( I 987) to model SEDs of dusty late-type WC stars. However, the lack of the nearIR emission lines in the Quintuplet spectra normally used to classify the WC stars is problematic (Figer, McLean, & Morris 1999).

114

Chiar et al.: Spectroscopy of Dust Components toward the Galactic Center

0

0. I

0

w c Star 3

3.1 3.2 3.3

0.05

3.4 3.5 3.6

3.7

5.6

5

R

6

6.2

h.4

6.6

Wavelength (pin)

Fig. 5 The 3 and 6 p m spectral regions for the Quintuplet-proper members, the GC central cluster sources, and a dusty late-type carbon-class Wolf-Rayet star (WR 118). Spectra are from Chiar et al. 2003 (Quintuplet-proper,3 pm; UKIRT-CGS4), Chiar et al. 2000 (Quintuplet (GCS3) and GC Central, 6 pm; ISO-SWS), Pendleton et al. 1994 (WR 118, 3 pm; Infrared Telescope Facility), Chiar & Tielens 2001 (WR 1 18, 6 p m ; [SO-SWS), Chiar et al. 2002 (GC Central, 3 pm; UKIRT-CGS4). The 3 pm Quintuplet spectrum is an average of the individual spectra centered on each of Quintuplet-proper members. The GC Central 3 pm spectrum is an average of 5 spectra, each has been scaled to T = 1.0 near 3.4 pm. The 6 pm spectrum of GCS3 was centered on GCS3-I; the ISO-SWS beam included GCS 3-11 and GCS3-I11(partially).

Our 6 pm spectroscopy reveals a similarity between the Quintuplet proper sources and dusty WC stars (Fig. 5). A distinct 6.2 pm absorption feature is seen toward several dusty WC stars (Schutte et al. 1998; Chiar & Tielens 2001) and the Quintuplet proper sources. The absorption feature has been attributed to the C-C stretch in amorphous carbon in the hydrogen-deficient circumstellar material associated with the WC stars, rather than PAHs in the interstellar dust along the line of sight (Chiar & Tielens 2001). Fig. 5 displays the 6 pm spectra from ISO-SWS of the lines of sight toward the Quintuplet source GCS 3-1, GC IRS 7, and the WC star WR 118. We note that the 3.4 pm hydrocarbon feature observed toward WR 1 18 is carried by the 12 magnitudes of interstellar visual extinction along the line of sight, and is not circumstellar in nature. Due to the extreme hydrogen deficiency in the WC star circumstellar environment, it is not possible to form hydrocarbon material. A broad symmetric 6.0 pm absorption feature is seen in the GC IRS 7 spectrum; the absorption is mostly accounted for by ices in the dense cloud material along the line of sight. The similarity between the WC star spectra and the GCS 3 spectrum and the dissimilarity of the GC IRS 7 spectrum is striking. These spectra are the first hint of some spectroscopic similarity between the Quintuplet sources and any of the proposed classifications, and lends support to the suggestion that they are dusty WC stars.

5

Conclusions and Future Work

Our recent spectroscopy of lines of sight toward the GC central cluster and the Quintuplet cluster has given us great insight into the chemical composition, characteristic absorption profiles, and distribution of both the diffuse ISM and dense clouds components along the line of sight. In addition, our group has carried out a program of narrow-band imaging in order to fully map the variation of the ice, hydrocarbon, and silicate dust components toward the GC central cluster, including the circumnuclear ring (Adamson et al. 2003). Future spectroscopic observations from airborne (SOFIA) and ground-based observatories will answer such questions as 1. Do icy grain mantles in the circumnuclear ring contribute to the deep ice features and

Astron. Nachr./AN 324. No. SI (2003)

115

X-CEN feature observed in those lines of sight'? 2. Is the 6 pm absorption profile of the Quintuplet-proper members unique to those sources and dusty late-type WC stars? 3. Is the distinct 3.28 pm feature observed in the Quintuplet-proper spectra due to PAH absorption in the diffuse ISM? 4. Will this feature b e present in high signal-to-noise spectra of other diffuse ISM lines of sight such as heavily extincted WC stars? 5. Is this feature really absent in lines of sight toward the GC central cluster? T h e answers to questions 2 through 5 will not only tell us more about aromatic hydrocarbons in the diffuse ISM, but will also bring us closer to uncovering the nature of the enigmatic Quintuplet sources. Acknowledgements J.E. Chiar gratefully acknowledges support from NASA's Long Term Space Astrophysics Program under grant 399-20-61-02.

References Adamson, A. J., et al., these proceedings. Adamson, A. J., Whittet, D. C. B., Chrysostomou, A., Hough, J. H., Aitken, D. K., Wright, G. S., & Roche, P. F. 1999, ApJ, 512,224 Allamandola, L. J., Tielens, A.G. G. M., & Barker, J. R. 1989, ApJS, 71,733 Boogert, A. C. A., Helmich, F. P., van Dishoeck, E. F., Schutte, W. A,, Tielens, A. G. G. M., & Whittet, D. C. B. 1998, A&A, 336,352 Brooke, T. Y., Sellgren, K., & Smith, R. G. 1996, ApJ, 459,209 Butchart, I., McFadzean, A. D., Whittet, D. C. B., Geballe, T. R., & Greenberg, J. M. 1986, A&A, 154, L5 Chiar, J. E., Adamson, A. J., & Whittet, D. C. B. 1996, ApJ, 472,665 Chiar, I. E., Pendleton, Y. J., and Geballe, T. G., and Tielens, A. G. G. M. 1998, ApJ, 507,281 Chiar, J. E., & Tielens, A. G. G. M. 2001, ApJ, 550, L207 Chiar, J. E., Tielens, A. G. G. M., Whittet, D. C. B.. Schutte, W. A,, Boogert, A. C. A., Lutz, D., van Dishoeck, E. F., & Bernstein, M. P. 2000, ApJ, 537, 749 Chiar, J. E., Adamson, A. J., Pendleton, Y. J., Whittet, D. C. B., Caldwell, D. A,, & Gibb, E. L. 2002, ApJ, 570, 198 Dartois, E., & d'Hendecourt, L. 2001, A&A, 365, 144 de Graauw, T., Whittet, D. C. B., Gerakines, P. A., et al. 1996, A&A, 3 IS, L345 Figer, D. F., McLean, I. S., & Moms, M. 1999, ApJ, 514,202 Figer, D. F., Kim, S. S., Moms, M., Serabyn, E., and Rich, R. M., & McLean, I. S. 1999, ApJ, 525,750 Geballe, T. R., Baas, F., & Wade, R. 1989, A&A208, 255 Gerakines, P. A,, Whittet, D. C. B., Ehrenfreund, P., er al. 1999, ApJ, 522, 357 Gusten, R.,Genzel, R., Wright, M. C. H., Jaffe, D. T., Stutzki, J.. & Harris, A. 1987, in AIP Conf. Proc. 155: The Galactic Center, ed. D. C. Backer (New York, AIP). 103 Ishii, M., Nag_ata,T., Chrysostomou, A., & Hough, J. H. 2002, AJ, 124, 2790 Lebofsky, M.J. 1979, AJ, 84, 324 Li, A., & Greenberg, J. M. 2002, ApJ, 577, 789 Lutz, D., Feuchtgruber, H., Genzel, R., et al. 1996, A&A, 3 15, L269 McFadzean, A. D., Whittet, D. C. B., Bode, M. F., Adamson, A. J., & Longmore, A. J. 1989, MNRAS, 241,873 Moneti, A., Glass, I., & Moorwood, A. 1992, MNRAS, 258,705 Nagata, T., Woodward, C. E., Shure, M., Pipher, J. L., & Okuda, H. 1990, ApJ, 351, 83 Okuda, H. et al. 1990, ApJ, 351, 89 Pendleton, Y. J., & Allamandola, L. J. 2002, ApJS, 138, 75 Pendleton, Y. J.. Sandford, S. A., Allamandola. L. J . , and Tielens, A. G . G. M., & Sellgren, K. 1994, ApJ, 437,683 Sanders, D. B., and Scoville, N. Z . , & Solomon, P. M. 1985, ApJ, 289,373 Sandford, S. A., Allamandola, L. J., Tielens, A.G.G.M., Sellgren, K., Tapia, M., &Pendleton, Y.J. 1991, ApJ, 371, 607 Schutte, W. A., van der Hucht, K. A., Whittei, D. C. B., et al. 1998, A&A, 337,261 Smith, R. C., and Sellgren, K., & Brooke, T. Y. 1993, MNRAS, 263,749 Tielens, A. G. G. M., Wooden, D. H., Allamandola, L. J., Bregman, J., & Wittebom, F. C. 1996, ApJ, 461,210 Whittet, D. C. B., Boogert, A. C. A., Gerakines, P. A., et al. 1997, ApJ, 490, 729 Williams, P. M., van der Hucht, K. A., & The, P. S. 1987, A&AI82, 91 Willner, S. P., Russell, R. W., Puetter, R. C., Soifer, B. T., & Harvey, P. M. 1979, ApJ, 229, L65 Wright, G., Geballe, T., Bridger, A., & Pendleton, Y.J. 1996, in New Extragalactic Perspectives in the New South Africa, eds. D.L. Block & J.M. Greenbeg (Dordrecht:Kluwer), 143

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 117- 123 (2003) / DO1 10.1002/asna.20038S064

X-rays from the HI1 Regions and Molecular Clouds near the Galactic Center Katsuji Koyama* I , Hiroshi Murakami** 2 , and Shinichiro Takagi***

'

' Department of Physics, Kyoto University, Kyoto 606-8502, Japan Institute of Space and Astronautical Science(ISAS), Kanagawa 229-85 10, Japan

Key words X-rays, Molecular clouds, HI1 regions, Young stars

Abstract. We report measurements by C,'hnn,dra of a variety of X-ray sources in the molecular clouds and HI1 regions of the Sgr B2, Arches, Quintuplet and the Galactic center clusters. Moderately bright X-ray sources are present in the Sgr B2, Quintuplet and the Galactic center clusters at the positions of ultra compact HI1 regions and bright infrared sources. Their X-ray spectra are fitted with models of a thin thermal plasma with 2-10 keV temperatures and luminosities of 10"2p""erg s-'. The X-ray properties are typical of those of high-mass young stellar objecta or clusters of such objects. The Arches Cluster has three bright X-ray sources, at the positions of bright IR and radio stars, with X-ray luminosities of a few x l0""erg spl each, which may indicate an unusual X-ray emission mechanism from high mass YSOs. A unique X-ray feature of molecular clouds and HI1 regions near the Galactic center is the presence of diffuse emission with a strong 6.4 keV line; in Sgr B2 this is attributable to the fluorescence of gas irradiated hy external sources in the Galactic center, while the diffuse emission from Arches is puzzling.

-

1 Introduction A unique X-ray feature in the Galactic center (GC) region was discovered by the Ginga satellite. The X-ray spectrum near the GC exhibits a bright iron line (Koyama et al. 1989). Figure 1 shows scan profiles taken with the classical proportional counter of beam size about I degree. The top panel is the wide hand X-ray flux (2-18 keV), and there is no large peak at the GC. However, a narrow band X-ray profile that includes the 6.7 keV iron line shows a bright peak at the GC, by far the brightest source (middle panel). The emission near 6.7 keV is extended and positive longitudes (north-east of the GC) emit lower energy (about 6.6 keV) than do negative longitudes, which emit at 6.7 keV (see lower panel of figure 1). In order to determine the origins of the iron line emission and the energy variations, we have used the ASCA satellite to obtain imaging spectroscopy of the GC region. Figure 2 (left) shows the ASCA image of the GC region in the 2-10 keV band. As expected, we found bright diffuse X-rays near the GC extending over a 1 square degree area. Figure 3 is the X-ray spectrum of this diffuse component. It shows many emission lines from highly ionized Si, S, Ar, Ca and Fe. The spectrum implies that the origin of the diffuse X-ray is a thin hot plasma of 107-108K. The continuum X-rays accurately trace this high temperature plasma. A surprising discovery is that there is a 6.4 keV line just below the 6.7 and 6.9 keV lines. The latter two lines are due to He- and H-like irons, and hence support the presence of very high temperature plasma (near IO'K), while the former line is due to neutral or low ionization states of iron. To investigate t h e origin of the 6.4 keV line, we obtained a narrow band image, which is shown in figure 2 (right panel). The

-

* e-mail: koyama 0crscphys.kyoto-u.ac.jp * * e-mail: [email protected] * * * e-mail: takagi9&cr.scphys.kyoto-u.ac.jp

@ 2003 WILEY-VCH Verlag GnihH & Co KGaA, Wcinhrim

118

K. Koyama et al.: X-rays near the Galactic Center

Fig. 1 The scan profile of the X-ray emission in the 2-18-keV energy band ( u: upper panel), the iron-line emission (b: middle), and the line energy along the Galactic plane (c: lower panel), (from Koyarna et al. 1989).

Fig. 2 The ASCA images of the GC region. Courtesy: Yoshitomo Maeda (see http://www-cr.scphys.kyoto-u.ac.jplIAUlgallely/gallery.html)

figure shows that the 6.4 keV line emission is clumpy, and is more concentrated to the northeast of the G C than to the opposite side (south west). This asymmetry is the origin of the Ginga asymmetry of the iron line energy.

I19

Astron. Nachr./AN 324, No. S I (2003) G C sum 8 SPECTRUM

0.1

I 1

U

2

5

10

channel energy (keV)

Fig. 3 The X-ray spectrum of the GC region (from Koyama et al. 1996)

By comparing the X-ray and radio data, we found that the 6.4 keV line emissions are well correlated with the giant molecular clouds (MC), which are high mass star forming regions. The region of brightest 6.4 keV line emission is the Sgr B2 cloud, about 0.65 degree (or lOOpc in projection) from the GC. The next brightest region is just inside of the Radio Arc, which corresponds to a newly found molecular cloud (MC) MO. I3+0.13. There also is faint emission near the Arches cluster, which contains young high mass stars and many compact HI1 regions. In the flowing sections, we report more details of individual clouds, in particular the X-ray emission from high mass young stellar objects (YSO) and the diffuse 6.4 keV line emission. Since MCs may harbor high mass YSOs, we briefly review the X-ray emission from high mass YSOs observed with Chnndl-a. Although the samples are still limited, the X-ray luminosities of YSOs appear to increase with increasing stellar mass, and saturate at about l0"'org s-' for massive stars or stars associated with compact HI1 regions (e.g. Kohno et al. 2002). The X-ray spectra of high mass YSOs are generally harder (a few keV) than stars on the main sequence (less than 1 keV).

2

SgrB2

Although ASCA found diffuse X-rays with strong 6.4 keV line emission from Sgr B2, X-rays from many unresolved point sources (possibly YSOs) may contaminate the diffuse emission. We thus made high spatial resolution observation of Sgr B2 with Chandra (ObsID = 944), and found 17 new point sources in the cloud. Figure 4 is the C h m d r a image near the center of Sgr B2 overlaid on the radio contour map. The two brightest X-ray sources (Nos. 10 and 13) lie near Sgr 8 2 Main (M), which contains the largest complex of ultra compact HI1 (UCHII) regions Figure 5 is a close-up view of Sgr B2 (M). The X-ray emission originates from several point sources, some of which coincide with ultra compact radio sources, while others have no radio counterpart. The

120

K. Kovama et al.: X-rays near the Galactic Center

composite X-ray spectra near from the UCHII complexes 10 and 13 are fitted with models of thin thermal plasmas with temperatures of 5-10 keV and large absorption column densities of 4 x H cm-2. The high temperatures are consistent with the high-mass YSOs in the Orion Nebula and the Mon R2 cloud (e.g., Shultz et al. 2001; Kohno et al. 2002). The X-ray luminosities from these UCHII complexes are about 10"crg spl. Since the UCHII complexes are typically comprised of about 10 sources, the X-ray luminosity of each point source is about 103'erg s-', which is roughly consistent with high mass YSOs. These luminosities, together with the high H indicate that some of these hard X-ray sources are absorption column densities of 4 x likely due to high-mass YSOs in the core of the Sgr B2 (M). Others are UCHII regions, and hence are probably zero-age main sequence (ZAMS) stars. N

The other UCHII complexes, Sgr B2 North (N) and South (S), also show hard X-rays with high abrorpH cmP2 and luminosities of a few to 10 times 10"crg s-'. tion column densities of about (5-6) x Thus these are also clusters of high mass YSOs located in the cores of the UCHII complexes. We note that the absorption column density of 5 x loz3H cm-2 for the Sgr B2 X-rays is the largest among the known stellar X-ray sources. Our observations demonstrate that hard X-rays are a very effective way to discover deeply embedded, YSOs even those that are located as far away as the GC and suffer from optical extinctions as large as of A v -200-300 mag. N

-

-2822 00-

0

0 0

?

v

C

2-2823 002 0

0" 28 24 00-

174725

17 47 20

174715

Right Ascension(J2000)

Fig. 4 The 2-10 keV band X-ray image of Sgr B2, overlaid on the contours of HI1 regions (Gaume et al.

Fig. 5 The 2-10 keV band X-ray image near Sgr 8 2 (M), overlaid on the contours of HI1 regions (De Pree et al. 1998, Gaume & Claussen 1990). The right source is No. 10, and the left source is No. 13.

1995).

In addition to the point sources, we also detected diffuse X-rays in the 6.4 keV line. The integrated flux of point sources (YSOs) is only 10%of the diffuse flux. Therefore at 6.4 keV the diffuse X-rays dominate the total emission. The morphology of the 6.4 keV line emission is a crescent with the curvature pointing toward the GC. This morphology, together with the spectrum, can be reproduced by illumination by strong X-rays from the GC direction. We call this class of X-ray source an "X-ray Reflection Nebula (XRN)."

Astron. Nachr./AN 324, No. S1 (2003)

121

3 Arches Cluster In the Chnrrdru mosaic image of the GC region (obsID= 242+945+1561), there is also a complex of Xray emission from two HI1 regions, the Arches and Quintuplet clusters. Figure 6 is a close-up view of the Arches cluster. Three of the bright X-ray sources coincide with the positions of bright IR andlor radio sources. We fitted the X-ray spectra for the 3 sources with a thin thermal plasma model with solar metal abundances. The best-fit parameters are given in Table I . The temperature of 1-3 keV and the large X-ray H ciiir2 are consistent with these sources being high mass YSOs absorption column density of about in the cluster. The X-ray luminosities (in the 2-10 keV band) are, however, a few x 10””t.i-gsrl, which are the highest among known YSOs. They are similar to those of the 30 Dor YSOs, which possibly are high-mass YSO binaries producing strong X-rays in the collisions of energetic stellar winds.

2

1

R3

Fig. 6 The 2-10 keV band X-ray iinage of the Arches cluster. The bin size is 0.”5 x 0.”5. The diamond ( 0 )and cross (+) show the position of the bright infrared sources (Blurn et al. 2001) and the radio sources (Lang et al. 2001), respectively.

Table 1 The best-fit parameters for the YSOs in the Arches Clustei

ID:IR(Radio)

26

2 1 (AR4)

23(AR1) _____

_____

N H( 10”Hcrr~-2)

6.9(4.5-11.2)

9.3(5.6-15.5)

9.9(6.5-14.8)

kT(keV)*

2.4(1.3-5.6)

1.3(0.6-2.5)

1.6(1.0-2.7)

L,(lO”erg s-l)

2.8(2.3-3.4)

3.9(3.14.7)

5.0(4.2-6.0)

* in the 2-l0keV

band

Surrounding the Arches Cluster is a diffuse structure (a circle of about 1’ radius), which shows strong 6.4 keV line emission with deep absorption at low energy (Yusef-Zadeh et al. 2002). The diffuse spectrum is fitted with a power-law and a narrow Gaussian line at 6.4 keV. This spectral feature is similar to Sgr B2, an XRN. However the X-ray morphology is different from Sgr B2, and is roughly a circle around the three brightest YSOs. Therefore, unlike Sgr B2, the irradiating source may not be external. YSOs within the

K. Koyama et al.: X-rays near the Galactic Center

122

Arches cluster may be the sources. For simplicity, we assume an uniform density spherical cloud, of size similar to the diffuse emission of 50" radius (2.1 pc at 8 kpc). From the observed NH the mean density Using the total flux from YSOs (the brightest three sources), we of the cloud is estimated to be 10'c&'. calculate the expected diffuse 6.4 keV line luminosity to be 1.3 x 10"erg s-', which is only 10% of the 103'erg s-l. Yusef-Zadeh et al. (2002) proposed an exotic scenario in which the observed value of diffuse X-rays are shock-heated gas created by the collision of individual 1000 km/s stellar winds in the dense cluster environment. However this hot gas should emit the 6.7 keV line, not the 6.4 keV line. Thus the origin of the 6.4 keV line is a puzzle and a deeper exposure Cliaridra observation is needed to help resolve it.

-

N

4 The Quintuplet Cluster The Quintuplet cluster is one of the most massive clusters in the Galaxy and contains bright IR sources and radio emission from the HI1 region (e.g., Figer et al. 1999 and references therein). Most of the cluster sources are thought to be 09 and WR stars (Figer et al. 1999); hence X-rays from high-mass YSOs are expected from this region. In the 50-ksec pointing observation (obsID= 945), the Quintuplet is located in a gap between CCD chips. Therefore, we have examined the GC survey data (obsID=2276, 12-ksec exposure). Although there are excess X-rays from this region, the limited exposure time makes it difficult to resolve individual sources. Assuming the same temperature and absorption as those of the Arches cluster, we estimate their luminosities to be 4 x 10"erg s-', which may be typical of high-mass YSO clusters. To make further progress, a deeper exposure is needed.

5 The Galactic Center Clusters In the GC itself X-rays are found from some of the IR bright stars (or star clusters), IRS 3, 13, 31 and 16SW (Baganoff et al. 2001). The GC region is known to be a massive star-forming region (e.g., Morris 1993; Paumard et al. 2001), and these IR sources are likely high mass stars. For example, IRS 13 is known to consist of many stars including the WR star IRS 13E (Paumard et al. 2001). IRS 16SW is probably an eclipsing He-star binary, hence the X-rays are probably due to the colliding stcllar winds (Ott et al. 1999). Therefore the observed X-ray sources are likely individual YSOs or clusters. Their average absorption column densities of 1023Hc1n-l and luminosities of 103't5rg s-' (Baganoff et al. 2001) are typical of high-mass YSO clusters in the GC region.

-

-

6 Summary We report C h a ~ ~ d robservations a. of the Sgr B2, Arches, Quintuplet and the Galactic center clusters. Sgr 9 2 exhibits X-ray emission coincident with three UCHII complexes (Main North and South). The X-ray spectra are fitted with thin thermal plasma models of 5-10 keV temperature, which are consistent with the sources being high-mass YSOs. Diffuse emission with strong 6.4 kcV lines is round, which is most likely due to external irradiation from the direction of the GC. The Arches Cluster contains three bright X-ray sources, at positions of bright IR and radio sources, with X-ray luminosities of a few x l0"erg 8 - l (2-10 keV band) each: these are the highest among any known YSOs. There is also diffuse emission, which includcs strong 6.4 keV line emission. Unlike Sgr 92, the 1' radius, suggesting that the irradiating sources are internal. However the morphology is a circle of YSOs can produce only 10% of the observed line flux, hence the true origin is still puzzling. There also are excess X-rays from the Quintuplet cluster. Assuming the same temperature and absorption as those of the Arches cluster, we estimate the luminosity to he 4 x 103'erg s-l, which may be typical of high-mass YSO clusters. N

Astron. Nachr./AN 324, No. SI (2003)

12.1

References Baganoff, F. K. , Maedd, Y. , Moms, M. , Bautz, M. W., Brandt, W. N.,Cui, W., Doty, J. P. , Feigelson, E. D . , Garmire, G. P. , Pravdo, S. H. , Ricker, G. R. ,& Townsley, L. K. 2001, astro-ph/0102151 Blum, R. D., Schaerer, D., Pasquali, A., Heydari-Malayeri, M., Conti, P. S., & Schmutz, W. 2001, AJ, 122, 1875 De Pree, C. G., Goss, W. M., & Gaume, R. A. 1998, ApJ, 500,847 Figer, D. F., McLean, I. S., & Morris, M. 1999, ApJ, 5 14,202 Gaume, R. A. & Claussen, M. J. 1990, ApJ, 351, 538 Gaume, R. A., Claussen, M. J., De Pree, C. G.. Goss, W. M., & Mehringer, D. M. 1995, ApJ, 449,663 Kohno, M., Koyama, K., & Hamaguchi, K. 2002, ApJ, 567,423 Koyama, K., Awaki, H., Kunieda, H., Takano, S., & Tawara, Y. 1989, Nature, 339, 603 Koyama, K., Maeda, Y., Sonobe, T., Takeshima, T., Tanaka, Y., & Yamauchi, S. 1996, PASJ, 48,249 Lang, C. C., Goss, W. M., & Rodriguez. L. E 2001, ApJL, 551, L143 Morris, M. 1993, ApJ, 408,496 Murakami, H., Koyama, K., I%Maeda, Y. 2001, ApJ, 558, 687 Murakami, H., Koyama, K., Tsujimoto. M., Maeda, Y., & Sakano, M. 2001, ApJ, 550, 297 Ott, T., Eckart, A., & Genzel, R. 1999, ApJ, 523, 248 Paumard, T., Maillard, J. P., Moms, M., & Rigaut, F. 2001, A&A, 366, 466 Senda, A., Murakami, H., & Koyama, K. 2002, ApJ, 565, 1017 Schulz, N.S., Canizares, C., Huenemoerder, D., Kastner, J. H., Taylor, S. C., & Bergtrom, E. J. 2001, ApJ, 549,441 Takagi, S., Murakami, H., & Koyama, K. 2002, ApJ, 573, 275 Yusef-Zadeh, F., Law, C., Wardle, M., Wang, Q. D., Fruscione, A,, Lang, C. C., & Cotera, A. 2002, ApJ, 570, 665

This Page Intentionally Left Blank

Astron. NachrJAN324, No. SI, 125- 131 (2003) I DO1 10.1002/asna.200385075

Reflected X-ray Emissions on Giant Molecular Clouds -Evidence of the Past Activities of Sgr A* Hiroshi Murakami*’, Atsushi Sends**?, Yoshitomo Maeda***’,and Katsuji Koyamat’

’ High Energy Astrophysics Division, Institute of Space and Astronautical Science 3-1 -1 Yoshinodai, Sagamihara, Kanagawa 229-8510, Japan ’ Department of Physics, Graduate School of Science, Kyoto University Kitashirakawa, Sakyo-ku, Kyoto, 606-8502, Japan

Key words molecular clouds, X-ray, refection, Sgr A*, Sgr B2, Sgr C, Radio Arc, MO.ll-0.08 Abstract. We have found strong 6.4-keV line emissions from the giant molecular clouds in the Galactic center region: Sgr B2, Sgr C , and M0.I 1-0.08 (at the Radio Arc region). The high angular resolution of Chcrndra reveals that the 6.4-keV line emissions are indeed coincident with the clouds, and shifted towards the Galactic center. The X-ray spectra have very strong 6.4-keV lines with equivalent widths 2 1 keV and are attenuated by larger column densities of interstellar gas. These characteristics imply that the massive molecular clouds are irradiated by an external X-ray source in the direction of the Galactic center and emit fluorescent and scattered X-rays. These clouds are new category of X-ray source: “X-ray Reflection Nebula”. The reflected X-ray flux reveals the recent luminosity history of the primary irradiating source, which may be the massive black hole Sgr A*, according to the light travel time to each cloud. Making use of the radio determinations of the cloud masses, we find that Sgr A* was as luminous as 10” erg s-’ a few hundred years ago, and has gradually decreased to present value.

1 Introduction With its wide energy X-ray band (0.5-10 keV) imaging capability, ASCA found 6.4-keV line emissions in the Galactic center (GC) region (Figure 1; Koyama et al. 1996). There are two distinct peaks, the Sgr B2 region and the Radio Arc region, which roughly agree with the locations of giant molecular clouds. For Sgr B2, by comparison with a radio observation of the molecular cloud, we have found that the distribution of 6.4-keV line is indeed correlated with the cloud, and is slightly shifted from the cloud core (Koyama et al. 1996;Murakami et al. 2000). Because the 6.4-keV line is characteristic of radiation from neutral iron, it i s natural to think that X-rays are emitted by molecular clouds. However, the clouds are very cold, and cannot emit high energy X-rays in themselves. Koyama et al. (1996) suggested that the clouds are irradiated by an external X-rays, and emit fluorescent X-rays in the 6.4-keV line. Thus the Sgr B2 cloud may b e a new category of X-ray source, an X-ray Reflection Nebula (XRN). To explain the luminosity of 6.4-keV line from the Sgr B2 cloud, a strong X-ray source with the luminosity of ,-- lo3’ erg sP1 is required (Murakami et al. 2000), but there is no bright X-ray source in the G C region. The most plausible explanation is that our Galactic nucleus Sgr A*, which is a massive black hole * e-mail: hiroQastro.isas.ac.jp,Phone: +81 427598134, * * e-mail: sendaQcr.scphys.kyoto-u.ac.jp *** e-mail: ymaedaQastro.isas.ac.jp t e-mail: koyamaQcr.scphys.kyoto-u.ac.jp

FdX:

+ X I 427598455

@ 2001 WILEY-VCII Vcrlag CmhH & Co. KGaA. Wemhrm

126

H. Murakami et al.: Reflected X-ray Emissions on Giant Molecular Clouds

(MBH) of 1O6A4a,was much brighter some hundreds of years ago, the light travel time between Sgr A* and Sgr B2, and is dim at present. N

The limited spatial resolution of ASCA, however, could not exclude possible contamination of many point sources such as young stellar objects in Sgr B2, which might deform the spectrum and morphology of the diffuse emission. Hence we have analyzed Chandra observations of giant molecular clouds in the GC region: Sgr B2, Sgr C, and M0.11-0.08 (at the Radio Arc region). With the higher angular resolution of Chandru, we have been able to constrain the contribution of point sources to a negligible level. Throughout this paper, the distances to the clouds are assumed to be 8.5 kpc, the same as to Sgr A*.

Fig. 1 The distribution of the 6.4-keV line intensity in the GC region observed with ASCA (Koyama et al. 1996).

2 Observations

-

Chandra observed Sgr B2 and MO. 11-0.08 with long exposure times of 100 ksec (ObsID=944) and 50 ksec (ObsID=945), respectively. The Sgr C region was observed as a part of GC survey, with a net exposure time of 20 ksec (ObsID=2267+2270+2272). We screened the observed data using standard criteria, and made point source subtracted spectra and 6.4-keV line images. N

N

Astron. Nachr./AN 324, No. S1 (2003)

127

3 Results 3.1

SgrB2

Sgr B2 is one of the most massive giant molecular clouds in the Galactic center region, and the most luminous in the 6.4-keV line band. With Cliandm's high angular resolution, we detected 20 point sources in the cloud. Some of the sources are coincident with star forming regions, and are considered to be young stellar objects. The integrated luminosity of all the resolved point sources is 3 x 10"" erg s - l , which is only 3% of the luminosity of the diffuse X-rays (see below). The details for the point sources are given in a separate paper (Takagi et al. 2002). Figure 2a shows the 6.4-keV line image of Sgr B2. In this energy band most of the X-rays are from diffuse emission. The diffuse X-rays come mainly from the south-west half of the cloud and the emission region has a concave shape pointing towards the Galactic center. The point source subtracted spectrum (Figure 3 ) exhibits strong Kcu and K,G emission lines of neutral iron. The center energy of K u line (Ec;) is 6.38 keV, and the feature has an equivalent width (EW) of 1.7 keV. The continuum component can be reproduced by a power-law model with a fixed photon index of 2.0 and a large absorption column density of NH 8.8 x cmp2. The luminosity is about erg spl (4-10 keV).

-

-

-

Fig. 2 (a) 6.4-keV line image of Sgr B2 observed with Chandru. The contours show the density distribution of the molecular cloud ("CO, Sato et al. 2000). The point source at the center of the cloud is coincident with the star forming region Sgr B2 Main (Takagi et al. 2002). (hj A simulated image with the X-ray retlcction nebula model. The irradiating source is assumed to be in the direction of CC.

To confirm the XRN scenario, we made a numerical simulation of the reflected X-rays for the casc of the molecular cloud being irradiated by an external X-ray source in the direction of the GC. The simulated image is shown in Figure 2b. The shifted distribution and the crescent shape are well reproduced by the simulation. The simulated spectrum with the XRN model exhibits very strong 6.4-keV line and a large absorption, and fits the observed spectrum well, with a reduced x2 of 1.14 (solid line in Figure 3).

H Murakami et al.: Reflected X-ray Emissions on Giant Molecular Clouds

128

These results support the idea that the Sgr B2 emission is fluorescent and scattered X-rays due to irradiation by an external source in the direction of the GC. The details for Sgr B2 cloud and XRN simulations are already published in Murakami et al. (2001).

>

2 10-2

3

>5I10-3 c 0

1o

-~ 2

x

o -2

I 10

L

5 Energy (keV)

Fig. 3 The spectrum of Sgr B2. The solid line shows the simulated spectrum with an XRN model (Murakami et al. 2001), and the dashed line shows the contribution of point sources.

3.2 SgrC This cloud is located at the opposite side of the GC from Sgr B2. The 6.4-keV line image (Figure 4a) shows diffuseX-ray emission on the GC side (at the upper left) of the cloud. The distribution has a crescent shape, which is oriented similarly with respect to the GC as that of Sgr B2. The spectrum (Figure 4b) is well fitted with the same model as Sgr B2, and exhibits strong iron line (Ec 6.38 keV; EW 3.2 keV) and a large absorption ( N H 1.4 x loz3 cmp2). These features are also same as Sgr B2. Thus this cloud also is a candidate XRN. More detailed analysis, such as comparison with a numerical simulation, is meaningless because of the poor photon statistics. An additional Chundru observation with longer exposure time is required.

-

-

N

3.3 M0.11-0.08 MO. 11-0.08 is a molecular cloud near the Radio Arc, and the nearest giant molecular cloud to the GC. Figure 5a, b show the 6.4-keV line image and the spectrum, respectively, of MO.ll-0.08. Diffuse X-ray emission fills the entire region of the cloud. In addition, filamentary structure is found at the edge of the cloud facing the GC. The spectrum is roughly reproduced by a thin thermal plasma model, but there are residuals at 6.4 keV and at higher energy. Therefore, we fitted the spectrum with a thermal component and the same model as Sgr B2 and Sgr C (a power-law and two Gaussian lines of K N and KB). The best-fit parameters of the K n line are, Ec 6.36 keV and EW 1.1 keV. The absorption column density for the power-law component is also large ( N H 2.4 x loz3cm-2). The thermal component is reproduced by emission from a thin thermal at a temperature of 3 keV and having a luminosity of 3 x erg s-'. Thus we find that the X-ray spectrum of MO. 1 1-0.08 exhibits strong 6.4-keV line. Although there is also a thermal component, fluorescent X-rays must be emitted from the cloud. By using the same model as Sgr B2 and Sgr C, we are able to reproduce the spectrum. MO.ll-0.08 also must be an XRN. N

-

N

-

-

-

129

Astron. Nachr./AN 324. No, S I (2003)

10

5

9

Fig. 4 (a) The 6.4-keV line image of Sgr C superposed on the radio intensity of the CS line (radial velosity of -120 km s-' to -110 km s-'; Tsuboi, Handa. Ukita 1999). Diffuse X-ray emission is seen on the GC side of the cloud. (b) The X-ray spectrum of Sgr C. The solid line shows the best-fit model spectrum using a power-law and two Gaussian lines.

lo-'

x

I 2

5

I 10

Energy (keV)

Fig. 5 (a) The 6.4-keV line image of MO. I 1-0.08. The contours are the CS line intensity over the radial velociry range of20 kin s-' to 30 hi sC1 (Tsuboi. Handa. & Ukita 1999). (b) Thc X-ray spectrum of MO.ll-0.08. The dotted line shows the thermal component, and the dashed lines show the reflected component (a power-law with two narrow lines of neutral iron KCYand K L ~ ) .

-

T h e thermal component might be related to the expansion of the cloud. T h e total energy of the plasma erg. A part of the expansion energy of 10" erg (Tsuboi, Ukita, & Handa 1997) could be transformed into thermal energy and result in the emission of X-rays.

is about 3 x

11. Murakami et al.: Reflected X-ray Emissions on Giant Molecular Clouds

130

4 Discussion In the previous section, we found that three molecular clouds are probably XRNe. If so, the luminosity of the 6.4-keV line from each cloud indicates the past luminosity of the irradiating source at a time corresponding to the light travel time from the source to the cloud. As already mentioned, the irradiating source is required to be very luminous, and considered to be the Galactic nucleus Sgr A*. We thus convert the distance from Sgr A* to elapsed time, and make a light curve for the X-ray luminosity of Sgr A* during the past 500 years (Figure 6). The luminosity was as high as lo3' erg s-' a few hundreds years ago, and seems to have decreased gradually to the present value. However, there are only four data points, and thus we cannot discuss the detailed variability. The time span of each data point corresponds to the size of each cloud. Our method is insensitive to variability on shorter time scales. The past activity of Sgr A" could have been generated by a surge of accretion onto the MBH due, for example, to the passage of dense shell of a young supernova remnant Sgr A East as discussed by Maeda et al. (2001) based on the new Chnndru results on the GC. They estimated that the age of the SNR is lo4 years, and that the dense shell reached Sgr A* lo3 years ago. After the shell swept the surrounding matter past Sgr A* the luminosity would have become anomalously low. Their time scale agrees with the derived light curve by the XRN model. Hence our drivation of the past activity of Sgr A* suggest that Sgr A* is in a low luminosity phase at present.

-

-

1o2 10

Time (year) Fig. 6 The past luminosity of Sgr A* estimated from the luminosity of 6.4-keV line of each cloud.

5

Summary

We observed three molecular clouds, Sgr B2, Sgr C, and M0.11-0.08 with Chundvu. With its high angular resolution, we found diffuse neutral iron line emissions from all of these clouds. These giant molecular clouds are "X-ray Reflection Nebulae".

Astron. Nachr./AN 324, No. S1 (2003)

131

From the intensity of reflected X-rays, we obtained the luminosity history of the Galactic nuclei Sgr A* during the last 500 years. Sgr A* was as luminous as erg s-l a few hundreds years ago, and has dimmed gradually since then

References Koyama, K., Maeda, Y., Sonohe, T., Takeshima, T., Tdnaka, Y., & Yamauchi, S. 1996, PASJ, 48, 249 Maeda, Y. et al. 2002, ApJ, 570,671 Murakami, H., Koyama, K., Sakano, M., Tsujimoto, M., & Maeda, Y. 2000, ApJ, 534, 283 Murakami, H., Koyama, K., & Maeda, Y. 2001, ApJ, 558, 687 Sato, F., Hasegawa, T., Whiteoak, J. B., & Miyawaki, R. 2000, ApJ, 535, 857 Takagi, S., Murakami, H., & Koyama, K. 2002, ApJ, 573,275 Tsuhoi, M., Handa, T., & Ukita, N. 1999, ApJS, 120, 1 Tsuhoi, M., Ukita, N., & Handa, T. 1997, ApJ, 48 I , 263

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1, 133 - 137 (2001) / DO1 10.1002/asna.200385078

Observation of Toroidal Magnetic Fields on 100 pc Scales in the Galactic Center

',

G . Novak"',D. T. Chuss', J. L. Dotson 3, G. S. Griffin4, R. F. Coewenstein M. G. Newcomb 5 , D. Pernic 5 , J. B. Peterson 4, and T. Renbarger6 I

Department of Physics and Astronomy, Northwestern University

* NASA, Goddard Space Flight Center

'

NASA, Ames Research Center Physics Department, Carnegie Mellon University Yerkes Observatory,University of Chicago School of Physics and Astronomy, University of Minnesota

Key words Magnetic Fields, Galactic Center, Submillimeter Polanmetry, Interstellar Dust Abstract. We present new submillimeter polarimetric observations of the Galactic center region, made using the SPARO polarirneter that operates at the South Pole. Compared with previous submillimeter polarimetry of this region, our measurements cover much more sky area, and they imply that the molecular gas in the central few hundred pc is threaded by a large scale toroidal magnetic field. We consider this result together with radio observations that show evidence for poloidal fields in the Galactic center, and with Famday rotation observations. We compare all of these observations with a magnetodynamic model for the Galactic center.

1 Introduction Our contribution to the Galactic Center Workshop 2002 was to present new submillimeter polarimetric observations of the Galactic center, obtained using the SPARO instrument at South Pole station. SPARO (the Submillimeter Polarimeter for Antarctic Remote Observations), is a 9-pixel submillinieter array polarimeter incorporating 3He-cooled detectors (Renbarger et al. 2003). It is operated on the Viper telescope (Peterson et al. 2000). Submillimeter thermal emission from interstellar dust grains is generally polarized, due to magnetic alignment of grains. Thus, submillimeter polarimetry provides a method for mapping interstellar magnetic fields. The SPARO map extends over a much larger sky area than has been covered in previous submillimeter polarimetric maps, so it provides new information on the large-scale configuration of the magnetic field in the Galactic center. This can be compared with the results of radio synchrotron observations and Faraday rotation observations, both of which also give information about magnetic fields. We have already published a paper that presents our SPARO results, and compares them with synchrotron and Faraday observations (Novak et al. 2003). Accordingly, here we will merely summarize the Novak et al. (2003) paper.

2

SPARO results

As is typical for submillimeter continuum observations, measurements made using SPARO are not absolute, but rather are differential. Specifically, the flux that SPARO measures is the difference between the * e-mail: [email protected],Phone: 847 491 864s

@ 2003 WILEY-VCH Vcrlilg GrnhH Rr Ca KGaA. Wetnhcioi

G . Novak et al.: Toroidal Magnetic Fields in the Galactic Center

I34

flux at the main observing position and the average flux for the two sky reference positions, which are separated from the main position by +0.5” and -0.5” in cross-elevation, respectively. We observed the Galactic center for a total of five weeks during the interval April-July 2000. The results are presented in Figure 1. The contours correspond to a 450 p m photometric map made using SPARO, and they clearly show the large concentration of molecular gas that is associated with the innermost few hundred pc of the Galaxy. The highest column density occurs at the position of Sgr B2, displaced from the center of the Galaxy toward positive Galactic longitudes. The SPARO polarization results are shown using bar symbols. The orientation of each bar gives the inferred magnetic field direction, that is orthogonal to the E-vector of the measured polarization, and the length of the bar is proportional to the degree of polarization. The SPARO polarization results imply that the magnetic field permeating the Galactic center molecular gas, when projected onto the plane of the sky, is for the most part parallel to the Galactic plane. The most natural way to account for this is to suppose that the molecular gas in the Galactic center is threaded by a large scale magnetic field having a toroidal configuration. This had already been suggested, based on earlier polarimetry results at far-infraredhbmillimeter wavelengths (Morris et al. 1992, Novak et al. 2000). However, the SPARO results cover much more sky area than the previous observations, and thus they provide the strongest evidence yet obtained for the existence of this toroidal large-scale field. Figure 2 shows the SPARO magnetic “vectors” superposed on a radio continuum image (gray scale) showing Galactic center non-thermal radio filaments. These filaments trace magnetic fields running preferentially perpendicular to the Galactic plane. They appear to delineate a large scale magnetic field with a poloidal configuration. It is clear from Figure 2 that the magnetic field in the central few hundred pc is neither purely toroidal, nor purely poloidal. Rather, there appear to be regions in which toroidal fields dominate as well as regions in which poloidal fields dominate. In particular, the field seems to be toroidal in the denser molecular material that is concentrated near the Galactic plane, and poloidal in the more diffuse, hotter, and more tenuous synchrotron-emitting regions.

3 The model of Uchida, Shibata, and Sofue A theoretical model for the Galactic center that may be able account for the separate “poloidal-dominant” and “toroidal-dominant” regions that we see in Figure 2 is the magnetodynamic model developed by Uchida, Shibata & Sofue (1985), and further refined by Shibata & Uchida (1987). This model was developed in order to explain the “Galactic Center Lobe” (GCL), that is a limb-brightened radio structure with a size of several hundred pc extending from the plane of the Galaxy up towards positive Galactic latitudes (Sofue & Handa 1984). In the model of Uchida et al. (1985), the GCL represents a gas outflow that is magnetically driven. The model consists of nonsteady axisymmetric magneto-hydrodynamic simulations in which the field is assumed to be perpendicular to the Galactic plane at high Galactic latitudes, but acquires a toroidal component near the Galactic plane due to differential rotation of the gas to which it is coupled via flux-freezing. The stress of the resultant magnetic twist is what drives the outflow. In Novak et al. (2003), we argue that this model is fundamentally consistent with both the poloidal field seen by radio observers and the toroidal field seen by SPARO, provided that we make allowances for the clumpy distribution of the molecular gas.

4 Faraday rotation It is possible to probe the line-of-sight component of the magnetic field in any region of the Galactic center that contains thermal gas, provided that this region lies along the line-of-sight to a synchrotron source. This is because polarized radio emission suffers Faraday rotation as it passes through thermal gas. The model of Uchida et al. (1985) makes a specific prediction regarding the line-of-sight component of the magnetic field. To visualize this prediction, imagine dividing the Galactic center region into four quadrants

135

Astron. Nachr./AN 324, No. S I (2003)

m

0

a

-0.5

(

-1.0 -

-

2% Inferred B

I

I

1 .o

0.5

0.0

-0.5

-1

.o

A a (degrees) Fig. 1 Results of450 /mi polarimetry (bars) and photometry (contours) of the Galactic center. obtained using SPARO. The distribution of 450 pm flux closely follows the Galactic plane. that lies at a position angle of +31”. Coordinate offsets are measured with respect to the location of Sgr A* (that lies at the inmsection of the horizontal and vertical dotted lines). Each bar is drawn parallel to the inf‘eired magnetic field direction (i.e. perpendicular to the E-vector of the measured submillimeter polarization), and the length of the bar indicates the measured degree of polarization (sce key at bottom left). Contours are drawn at 0.075, 0.15, 0.30, 0.60. and 0.95 times the peak Rux, which is located at the position of Sgr 6 2 . For clarity, negative contours are not shown. The reference beam offsets were the same for polanmetry and photometry and are given in 3 2. The S’ beam of SPARO is shown in the key. Positive Galactic latitudes lie towards the upper right of the figure. and positive Galactic longitudes lie towards upper left.

according to the signs of Galactic longitude and latitude. According to the model, the sign of the line-ofsight field (i.e., towards or away from the observer), should be the same within a quadrant, and opposite in adjacent quadrants. We carried out a survey of the literature on Faraday rotation measurements toward Galactic center synchrotron soiirces, and we discovered a pattern of observed reversals in the sign of the Faraday “rotation measure” (hereafter, RM) that matches these predictions. We show this, using plus and minus signs, in Figure 3. The plus sign represents positive RM, and the minus sign represents negative RM. References to the Faraday rotation observations are given in Novak et al. (2003). We note that for RM measurements at positive Galactic longitudes, the asymmetry with respect to the Galactic plane had been noted previously, and Uchida et al. ( 1 985) pointed out the agreement with their model. To our knowledge, Novak et al. (2003)

G . Novak et al.: Toroidal Magnetic Fields in the Galactic Center

136

0.8

0.6

0.4

0.2 CQ

U

0

-0.2

-0.4

0.6

0.4

0.2

0

-0.2

-0.4

-0.6

Act (degrees) Fig. 2 450 pm polarization measurements (bars) shown together with 90 cm radio continuum image (gray scale, LaRosa et al. 2000), and 850 pm continuum emission (contours, Pierce-Price et al. 2000). As in Fig. 1, the orientation of each bar is parallel to the inferred magnetic field direction (i.e., orthogonal to the measured direction of polarization) and its length is proportional to the degree of polariration. The radio continuum image shows about six locations where non-thermal filaments can be seen. These non-thermal filaments trace magnetic fields in hot ionized regions. The gray scale image is loganthmically scaled, and the contours of 850 pm emission are also logarithmic. Coordinate offsets are measured with respect to the position of Sgr A*. The location of the brightest bundle of non-thermal filaments (referred to as the non-thermal filaments of the Radio Arc) is indicated in the figure.

were the first to compare the signs of the RM measurements that lie towards negative Galactic longitudes with the predictions of the Uchida et al. ( I 985) model. The pattern that we see in Figure 3 is the one that results when the poloidal field points toward positive Galactic latitude. If the poloidal field is Laken to point in the negative latitude direction, then all of the RM signs should be reversed. Thus, if our interpretation of these data in the context of the Uchida et al. (1985) model is correct, we can conclude that the large scale poloidal field points towards Galactic North. Considering this pattern of RM reversals together with the SPAR0 polarization map and the evidence for poloidal fields derived from radio synchrotron maps, we conclude that the available observations do

Astron. Nachr./AN 324, No. S1 (2003)

137

Fig. 3 Contours show distribution of 3.5 cm radio continuum emission from the central 500 pc (Haynes et al. 1992). The plus and minus symbols refer to the sign of the Faraday rotation measure, as discussed in 5 4.

support the general picture given by Uchida et al. (1 985) for the large-scale magnetic field in the Galactic center. The data, however, are very sparse. More observations, especially of Faraday rotation, are needed. The SPAR0 project was funded by the Center for Astrophysical Research in Antarctica (an NSF Science and Technology Center; OPP-8920223), by an NSF CAREER Award to G.N. (OPP-9618319), and by a NASA GSRP award to D.C. (NGT5-88). References Haynes, R. F., Stewart, R. T., Grey, A. D., Reich. W., Reich, P., & Mebold, U. 1992, A. & A,, 264, 500 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, AJ, 119,207 Morris, M., Davidson, J. A., Werner, M., Dotson, J., Figer, D. F.. Hildebrand, R. H., Novak, G.. & Platt, S. 1992, ApJ, 399, L63 Novak, G., Chuss, D. T., Renbarger, T., Griffin, G. S., Newcomb, M. G., Peterson, J. B., Loewenstein. R. F., Pernic, D., & Dotson, J. L. 2003, ApJ, 583. L83 Novak, G., Dotson, J . L., Dowell, C. D., Hildebrand, R. H., Renbarger, T., & Schleuning, D. A. 2000, ApJ, 529, 24 1 Peterson, J. B., Griffin, G. S., Newcomb, M. G.. Alvarez, D. L., Cantalupo, C. M., Morgan, D., Miller, K. W., Ganga, K., Pernic, D., & Thoma, M. 2000, ApJ, 532, L83 Pierce-Price, D., Richer, J. S., Greaves, 1. S., Holland, W. S., Jenness, T., Lasenby, A. N., White, G. J., Matthews, H. E., Ward-Thompson, D., Dent, W. R. E, Zylka, R., Mezger, P., Hasegawa, T., Oka, T., Omont, A,, & Gilmore, G. 2000, ApJ, 545, L121 Renbarger, T., Chuss, D., Dotson, .I.L., Hanna, J. L., Novak, G., Malhotra, P., Marshall, J., Loewenstein, R. F., & Pernic, R. 2003, in preparation Shibata, K., & Uchida, Y. 1987, PASJ, 39, 559 & Handa, T. 1984, Nature, 310, 568 Sofue, Y., Uchida, Y., Shibata, K., & Sofue, Y. 1985, Nature, 317, 699

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S 1, 139 - 143 (2003) / DO1 IO.IO02/asna.200385087

Extended photoionization and photodissociation in Sgr B2 J.R. Goicoechea*l,N.J. Rodriguez-Fernandez* * 2 , and J. Cernicharo***I

’ Departamento de Astrofisica Molecular e Infrdrroja, IEM/CSIC, Serrano 121, E-28006 Madrid, Spain

* Observatoirede Pans - LERMA. 61, Av. de I’Observatoire,75014 Paris, France Key words Galaxy: center - infrared: ISM: lines and bands ISM: individual (Sagittarius B2)

- ISM:

dust, extinction-H

II

regions -

Abstract. We present large scale 9’ x 27’ (25 pc x 70 pc) far-IR observations of the Sgr B2 complex using the spectrometers on board the fnfrcired Spuce Obsen.atnr$ (ISO). The Par-IR spectra are dominated by the strong continuum emission of dust and by the fine structure lines of high excitation potential ions (N I I , N 111 and 0 111 ) and those of neutral or weakly ionized atoms (0I and C I I ). The line emission has revealed a very extended component of ionized gas. The study of the N I I I 57 p n / N I I 122 p,m and 0 111 52pm /88pm line intensity ratios show that the ionized gas has a density of 11,-.10~-~cm-3 while the ionizing radiation can be characterized by a diluted but hard continuum, with effective temperatures of ~ 3 5 , 0 0 0K. Photoionization models show that the total number of Lyman photons needed to explain such an extended component is approximately equal to that of the H 11 regions in Sgr B2(N) and (M) condensations. We propose that the inhomogeneous and clumpy structure of the cloud allows the radiation to reach large distances through the envelope. Therefore, photodissociation regions (PDRs) can be numerous at the interface of the ionized and the neutral gas. The analysis of the 0 I (63 and 145 pm) and C II (158 pm) lines indicates an incident far-UV field (Go, in units of the local interstellar radiation field) of 103-4 and a H density of cm-8 in such PDRs. We conclude that extended photoionization and photodissociation are also taking place in Sgr B2 in addition to more established phenomena such as widespread low-velocity shocks. ’

1 Introduction The Sgr B2 complex represents an interesting burst of massive star formation in the inner 400 pc of the Galaxy (the Galactic Center region, or GC) and may be representative of other active nuclei. Large scale continuum emission studies show that Sgr B2 is the brightest emission and the most massive cloud of the region (-.lo7 Ma;Lis & Goldsmith 1989). The main signposts of star activity are located within three dust condensations labelled as Sgr B2(N), (M) and (S). They contain all the tracers of on-going star formation: ultracompact H I I regions driven by the UV field of newly born OB stars, hot cores from embedded protostars, molecular masers, and high far-IR continuum intensity. These core regions are surrounded by a low density ( n ~ , < I 0cm-:’) ~ extended envelope (-lS’), hereafter Sgr 8 2 envelope, o f warm gas (Tk.200 K) and cool dust (Td=20-30 K; Hiittemeister et al. 1995). A summary of the different components present in Sgr B2 and their main charactcristics i s shown in Figure 1 (I@). The origin of the observed rich chemistry in the Sgr B2 envelope and its possible heating mechanisms are far from settled and several scenarios have been proposed. Low-velocity shocks have been traditionally invoked to explain the enhanced abundances of SiO or NH3 and the differences between gas and dust temperatures (cf. Floweret al. 1995). The origin of shocks in the Sgr B2 envelope have been associated * Corresponding author: e-mail: javierQdamir.iem,csic.es, Phone: +0034 91 561 68 00 ,Fax: +0034Yl 564 55 57 ** e-mail: Nemesio.Rodriguez-FernandezQobspm.fr * * * e-mail: cerniOdamir.iem.csic.es

@ LOO? WLEY-VCH Vzrlag GmhH b Co KGaA. Wriohcim

J. Goicoechea et al.: Photoionization and photodissociation in Sgr B2

140

17h45m

17h44m

17h43m

17h42m

R.A. (1950)

Fig. 1 Rzght : Large scale IRAS image at 60 p n (Gordon et al. 1993) and IS0 target positions across Sgr B2 region. L e f t . Sketch showing the different structures and components in the Sgr B2 complex Hot cores are shown black shaded and H I I regions are the structures enclosing the stars. (Adapted from Huttemeister et al. 1995)

either with large scale cloud-cloud collisions or with small scale wind-blown bubbles produced by evolve massive stars (Martin-Pintado et al. 1999). The effect of the UV radiation in the Sgr B2 envelope has been traditionally ruled out because of the gas and dust temperature differences, the unusual chemistry and the abscense of thermal radio-continuum and ionized gas outside the H 11 regions and hot cores within the central condensations. Our observations reveal the presence of an extended component of ionized gas detected by fine structure emission. All this new data suggest that UV radiative-type processes are also important in the heating of the Sgr B envelope in addition to mechanical mechanisms, as can be in other GC clouds (see Rodriguez-Fernandez et al. 2003). In this contribution we present a brief summary of the results obtained by ISO' in the Sgr 8 2 envelope (Fig. 1 [right])concerning the ionized gas and the effects of the UV radiation.

2 Extinction corrections The large HZ column density (up to loz5cmP2) found in Sgr B2 suggests that even in the far-IR, fine structure lines can suffer appreciable reddening. We have estimated the prevailing extinction in each position by converting the continuum opacity into visual extinction. The spectra cannot be fitted with a single gray body. Thus, we have modeled the observed continuum spectrum as a sum of two gray bodies. The total continuum flux in the model is: S A = (I - e-TTorm) L3A(Twarm) nu,,,

+ (1

-

e?ioLd )

BA(Tcold)flcold

(1)

where Bx(T,) is the Planck function, 7: is the continuum opacity and 0, is the solid angle subtended by the i dust component. The continuum opacity is given by r~ = 0.014 A" (30/A)' where /3 is the grain emissivity exponent of each dust component. The observed spectral energy distributions are best fitted with a dust component with a temperature of 13-22 K and a warmer component with a temperature of 24-39 K. The warmer component contributes less than 10 % of the total optical depth. The higher dust temperatures are those measured in the southern part of Sgr B2. Depending of the position, the derived extinction varies from -20 to -1000 magnitudes (see Table 1 for lower and upper limits to the visual extinction).

' Based

on observations with ISO, an ESA project with instruments funded by ESA Member States (especially the PI countries: France, Germany, the Netherlands and the United Kingdom) and with participation of ISAS and NASA.

Astron. Nachr./AN 324, No. S I (2003)

I41

3 The ionized gas The far-IR spectra exhibit several fine structure lines from the ionized material. We have clearly detected the 0 I 63 and 145 pm, and C I I 158 pm lines in all observed positions. In addition, lines coming from the N 11 122 ILm, N I I I 57 l m , 0 111 52 and 88 pm lines are also detected in most positions, revealing a prominent component of ionized gas in the southern and eastern regions of Sgr B2. Table 1 lists the extinction corrected 0 I[I 52 pm/88 pm line intensity ratios for the two derived limits to the extinction across the region. For extended emission sources and lines excited by collisions with electrons (see Rubin et al. 1994) we derive electron densities (n,!) in the range ~ 1 0 ' ~c n" p 3 for the extended envelope. At the limited spectral resolution of the LWS/grating mode, 0 111 lines are hardly detected in the central positions. Nevertheless, Fig. 2b shows their Fabry-Perot detection in Sgr B2(M). Both 0 111 lines appear centered at V ~ s ~ z 5 15 0 km f s-' and do not show emission/absorption at more negative velocities (foreground gas). Considering A" > 1000 magnitudes, we found n,> lo"." cmp3 in Sgr B2(M). Table 1 also lists the extinction corrected N 111 57/N I[ 122 line intensity ratios. The minimum averaged ratio is 0.77 while the upper limits are dependent to the maximum extinction affecting the lines. For those ratios, we derive minimum effective temperatures (T,ff) for the ionizing radiation of 32.000 36.000 K . Those T,ff should be considered as a lower limit to the actual T,ff of the ionizing source if this is located far from the nebulargas. We have carried out CLOUDY (Ferland 1996) simulations showing that the observed line ratios are consistent with an scenario where almost all ionizing photons arise from the H 11 regions within Sgr B2(M) and (N). The total flux of Lyman continuum photons is 5-l and T,ff = 35,000 K . The differences in the observed N 111 /N 11 ratios are due to the dilution of the incident radiation (lower ionization parameter). Hence, the size of the ionized region can only be explained if the medium is highly inhomogeneous. This suggests that the clumpy nature of the cloud allows the radiation to reach large distances through the envelope. In this scenario, several PDRs can be expected in the interface between the ionized and the neutral gas. N

Table 1 Selected line ratios after correcting for the estimated minimum and maximum extiction. The different beam sizes of each LWS detector are taken into account and extended emission is considered. Offsets are in arsec. map

(0,-90) (0,-180) (0,-270) (0,-450) (0,-630) (0,-810) (270,O) (180,O) (90,O) (-90,O) (-I 80,O) (-270.0)

warm dust A V (mag)

1.2-1.5 1.6-2.0 1.1-1.8 5.2-8.0 0.7- 1.7 16-18 3.7-5.2 2.8-3.8 2.7-3.2 3.5-3.9 1.1-1.3 27-28 4.9- 6.6 7.4-9.0 28-34 21-26 2 1-25

cold dust A v (mag) 15-55 23-84 25-92 41-112 131-294 367-877 148-493 59-205 28-102 23-85 16-59 156-536 78-276 168-565 228-579 62- I68 45-124

0 111 R(52/88) <2.40 1.79-2.23 I .36-1.73
I . 14-2.38 0.62-2.64 0.85-2.97 0.43-0.63 0.38-0.50

n,(O

111 )

N 111 / N U

log(cmp3) R(57/122) <3.18 2.83-3.10 2.49-2.79 <2.70 <3.66 2.41-4.67 1.99-3.53 1.03-1.67 2.28-2.61 2.85-3.13 1.39-1.58 ~2.24 2.28-3.18 1.53-3.30 1.92-3.44 1.10-1.55 0.94-1.27

<0.72 <0.72 <0.26 < 1.77 <2.73 I .60- 1 1.3 0.78-2.95 0.79- I .39 0.18-0.24 0.55-0.71 0.18-0.2 I <3.13 ~0.95 1.31-6.07

TF,"EL -2 2

(+lo+K ) <33.2 <33.2 ~31.8 <35.6 <35.2 (35.0-37.2) (34.9-35.3) (35.0-35.8) (31.6-32.2) (32.9-33.2) (32.8-33.1) <36.6 <33.6 (35.7-36.3)

J. Goicoechea et al.: Photoionization and photodissociation in S e r B2

142

' I

P

0",180"

[OI] 3P,-3P2 63 184 prn

10P

40

80

120

180

- 100

100

200

Fig. 2 ISOLWS-grating mode raster map of the 0163, 0111 88 and C I I 158 pin lines. Offset positions are given in arcsec respect to Sgr B2(M) (O", 0"). (a) Averaged continuum flux of each ISOLWS detector and gray-body best fits for some selected positions. The error bars correspond to a 30 % of flux uncertainity. (b) Fine structure lines detected with ISOLWS-FP mode in Sgr B2(M). (c) [Top]Correlation between the N I I 122 pm and C I I I58 pin lines. SgrB2(M) and (N) are not included . [Bottom]Main features of the grating raster map labelled. (d) C I I /O I intensity ratio vs. (C 11 +O I +Si II )/FIR for several PDR models of varying FUV fields and hydrogen densities (from Wolfire et al. 1990). The space parameter ocuppied by Sgr B2 envelope is represented by gray squares (see text).

Astron. Nachr./AN 324, No. S 1 (2003)

143

4 The Photodissociation regions The large OH column density of warm (Tkz-300 K) and low density gas (nH,< lo4 crnp3) recently found in the envelope of Sgr 8 2 suggests that its external shells are illuminated by a strong far-UV field that photodissociate the large amount of water vapor found in the region (Goicoechea & Cernicharo 2002). However, the main properties of the associated warm PDRs could not be derived. For that porpose, we have compared the far-IR continuum emission, and the C I I 158 pm, 0 I 63 pni and 145 pm lines at each position with PDR theoretical models. Figure 2d shows our results in comparison with predictions from PDR models of Wolfire, Tielens and Hollenbach ( I 990). The gray squares show the parameter space occupied by the Sgr B2 positions for intensity ratios corrected for the minimum (filled) and maximum (not filled) visual extinction estimated from the continuum analysis. The observational points scatter around a far-UV and n ~ - . l O “ - ~ cm-:’ depending on the use of all observed C I I flux (dark gray) or flux, Go, of Ihe remaining C I I flux after subtracting the C 11 emission arising in (non-PDR) low-density ionized gas and correlated with the N 11 emission. From the observed correlation (see Fig. 2c) we estimate that 20 to 70 L70 of the C 11 emission arises in the PDRs of Sgr B2. In addition, both the observed C 11 158/01 63 and (C 11 158+0 I 63)FIR intensity ratios shown in Fig. 2d are not predicted by shocked gas models (Hollenbach & McKee 1989) while they arc commonly reproduced in PDR models.

5 Summary and perspectives We have presented new far-IR observations of the Sgr B2 region that reveal a new perspective of the less known extended envelope of the complex. The I S 0 data show the presence of a widespread component of ionized gas reaching very large distances from the H If regions of known massive star formation. Molecular tracers and atomic fine structure tracers do not show evidences of high-velocity shocks. Hence, the ionized gas can not he explained in terms of high-velocity dissociative shocks. It seems that the well established widespread low-velocity shocks are not the only mechanism heating the gas to temperatures larger than those of the dust. We now have evidence that low-velocity shocks and large scale radiative processes, such as the photoelectric effect, coexist in the whole Sgr B2 envelope, increasing the difficulty of interpreting the astronomical data, but providing one of the richest and most peculiar clouds in the galaxy. Coexistence of mechanical and radiative-type heating mechanisms, based on the effects of a far-UV field that permeates an inhomogeneous and clumpy medium, seems to he the rule in the envelope of Sgr B2 and in the hulk of GC clouds observed by I S 0 (Rodriguez-Fernindez el al. 2003). Acknowledgements We thank J. Martin-Pintado for stimulating discussions about Sgr B2 and the Galactic Center, and M.A. Gordon for providing us the IRAS maps of Sgr B2. NJR-F has has been supported by a Marie Curie Fellowship of the European Community program “Improving Human Research Potential and the Socio-economic Knowledge base” under contract number HPMF-CT-2002-01677.

References Ferland, G.J. 1996, A Brief Introduction to Cloudy, U. Kentucky, Dept. Physics and Astronomy Internal Report Flower, D.R., Pineau des Forgts, G., & Walmsley, M.C. 1995, A&A 294,815 Goicoechea J.R., & Cernicharo, J . 2002, ApJ, 576, L77 Hollenbach, D. & McKee, C.F. 1989, ApJ, 342,306 Hiittemeister, S..Wilson, T.L., Mauersberger, R., Lemme, C., Dahmen G., & Henkel, C. 1995, A&A, 294, 667 Lis, D.C., & Goldsmith, P.F. 1989, ApJ, 337, 704 Martin-Pintado, J., Gaume, R.A., Rodriguez-Fernandez, N.J., de Vicente P., &Wilson, T.L. 1999, ApJ, 519, 667 Rodriguez-FemBndez N, Martin-Pintado, J., Fuente A., Wilson T.L., 2003, these proceedings Rubin, R.H., Simpson, J.P., Lord, S.D., et al., 1994, ApJ, 420, 772 Wolfire, M.G., Tielens A. G. G. M., & Hollenbach, D. 1990, ApJ, 358, 116

This Page Intentionally Left Blank

Astron. NachdAN324, No. S1, 145- IS0 (2003) / DO1 10.1002/asna.200385053

Propagation of charged particles from the Galactic Center W. Bednarek*, M. Giller, and M. Zielinska Department of Experimental Physics, University of L6di. Poland

Key words The Galactic Center, pulsars, cosmic ray origin, cosmic ray sources.

Abstract. Recent analysis of the anisotropy of cosmic rays at lo1' eV (the AGASA and SUGAR data) show significantexcesses from regions close to the Galactic Centre and Cygnus. Our aim is to check whether such anisotropies can be caused by a single source of charged particles. We investigate propagation of protons in two models of the Galactic regular magnetic field (with the irregular component included) assuming that the particles are injected by a short lived discrete source lying in the direction of the Galactic Centre. We show that apart from a direct (prompt) image of the source, the regular magnetic field may cause indirect (delayed) images at quite large angular distances from the actual source direction. The observed image is strongly dependent on the time elapsed after ejection of particles and it is also very sensitive to their energies. For the most favourable conditions for particle acceleration by a young pulsar the predicted fluxes are two to four order of magnitudes higher than that observed. The particular numbers depend strongly on the Galactic magnetic field model adopted but it appears that a single pulsar in the Galactic Centre could be responsible for the observed excess.

1 Introduction Recent analysis of the AGASA data shows anisotropy in the arrival directions of cosmic rays with energies 1017.9- 1018." eV, with excesses from thc direction near the Galactic Centre (Hayashida et al. 1999; Bellido et al. 2001). In this paper we concentrate on the details of propagation of protons with energies lo1* eV from a source located at the G C applying two models of the galactic magnetic field in the Galactic Plane (GP) and halo. The existence of a large magnetic halo extending several kpc out of the G P in the perpendicular direction z is suggested by observations of nearby spiral galaxies, e.g. NGC 253 z > 10 kpc, l3 7pG (Beck et al. 1994), NGC 4631 - z > 8 kpc, l3 = 2pG (Golla & Hummel 1994), NGC 891 and NGC 4561 - z 3 kpc, B = 1pG (Sukumar & Allen 1991). We will consider here the possibility that the observed excess is due to cosmic ray protons and study their propagation from some point sources located in the direction of the G C see also Bednarek, Giller & Zieliliska 2002). In particular, we have studied the point source image as a function of time. We have also studied its dependence on the model of the regular magnetic field in the Galaxy as well as on the proton energy. We have also considered whether the observed excess near the G C could be caused by particle emission from a single pulsar. N

-

N

2 Propagation of EHE protons Protons ejected from a point-like source propagate in the Galactic magnetic field which consists of regular (Breg)and irregular (Birr)components. We have adopted two different models for the regular Galactic magnetic field: model I is that proposed and described in detail by Urbanik, Elstner & Beck (1997), model I1 is the bisymmetric field model with field reversals and odd panty (BSS-A) proposed by Han * Corresponding author: e-mail: [email protected]

@ 2003 WILEY-VCH Verlag GmhH B Co KGaA. Wcinhem

W. Bednarek et al.: Propagation of charged particles from the Galactic Center

146

1

lO*R

0

50

time delay

100

150

200

in le3 years

10

P $"., 0

200 time delay

400

I 6 0 0 800 1 0 0 0 in l e 3 y e a r s

Fig. 1 Amval directions of protons with energies 3 x 10'' eV injected by a point source in the GC (marked by the large dot) for model I (left) and model I1 (right). Maps (in galactic coordinates) from a) to f ) show directions of particles arriving in consecutive time delay intervals of 5 x lo3 yr, i.e. a) is for 0 - 5 x lo3 yr, _._,f) (2.5- 3) x lo4 yr, and from i) to 0) with intervals of 2 x lo4 yr, i.e. i) is for 0 - 2 x lo4 yr, ... , 0) (1 - 1.2) x lo5 yr. g) and o) Arrival directions integrated over time. h) and p) Delay time distribution of arriving particles; time in units of lo3 yr.

& Qiao (1994, see also Stanev 1997). We calculate numerically the proton trajectories within the range of energies corresponding to those of the AGASA excess eV) and the SUGAR excess - 101s.5 eV) from the direction of the GC. For 1.2 x lo6 protons ejected isotropically from a point-like source located at the Galactic Center we record the parameters (numbers, directions, and arrival times) of particles intersecting a sphere with the radius of 250 pc centred on the Earth. These events are considered as observed by a detector on the Earth. The numbers of the arriving particles as a function of travel time (the time of flight along the straight line being subtracted) are displayed in the form of histograms in Figs. Lh, 2h, and 3h for model I, and in Figs. lp, 2p, and 3p for model 11. It becomes evident that the distribution of the arrival times of particles with energies by a factor 2-3 larger is completely different. The bulk of particles arrive to the observer within 2.5 x lo4 years (model I) and 5 x lo4 years (model 11) for 3 x 10" eV (Figs. Ih and lp), up to ,-- lo5 years (model I) and lo6 years (model 11) for lo1* eV (Figs. 3h and 3p). In Figs. 1,2, and 3 from a) to 9 (model I) and from i) to n) (model 11) we show maps (in galactic coordinates, with longitude increasing to the left) with the arrival directions of protons intercepting the sphere around the Earth within consecutive time delay intervals chosen accordingly to the particle energy and magnetic field model (see figure captions). Maps summed up over time are in Figs. g) and o) showing the direction distribution in the case of a steady source.

-

Let's first concentrate on the results for model I. The most interesting feature for protons with energies (2 - 3 ) x 10" eV is the particle clustering in multiple images of the source. These images appear at different places on the sky at different times after injection. A large number of protons reach the Earth's vicinity from directions close to the GC (shifted by about N loo towards positive longitudes) creating an

147

Astron. Nachr./AN 324, No. S1 (2003)

I

lO*H

o

50 time delay

100

ISO

ma

in l e 3 y w r s

0

LDi +--II zoo

o

time delay

400

600

800

inoo

in le3

eV. Maps for model I are for time delay intervals 2 x lo4 Fig. 2 As in Fig. 1 but for protons with energies 2 x yr, for model I1 - lo5 yr. Maps for model I1 start from l o 5 yr. extended source with the radius of about 20" -30". For protons with energies 10'' eV the amval directions become much more scattered (see Figs. 3a to 3h). In this case, protons arrive from a large part of the sky, almost independently of time after injection, apart from the peak for the first 2 x lo4 years. Increasing slightly the field strength will cause the discovered features shifting to higher energies and fitting better to the energy range where the actual excess of particles has been detected. The arrival directions of protons and their time distributions are completely different for the field model 11. There are no protons arriving directly from the actual position of the source at the GC (out of lo6 ejected) up to 3 x 10'' eV. For protons with 3 x eV, only a single image of the source is visible at a high negative latitude. It is created mainly by protons amving with relatively small time delay with respect to the rectilinear propagation, i.e. within less than 2 x lo4 years (see Fig. lp). For lower proton energies the image of the source is also centred on high galactic latitudes becoming broader and stronger. Particles arrive to the Earth much later than for model I i.e. after (2 - 4) x lo5 years for 2 x eV and ( 2 - 7) x lo5 years for 3 x 1018 eV. In spite of the instantaneous injection the anisotropy due to the source would be visible in the same directions for a long time. By comparing our calcufation results on the proton anisotropy for the two magnetic field models we conclude that the propagation of charged particles is very sensitive to their energies and to the structure and strength of the magnetic field. The two models, both based on experimental observations, give totally different predictions concerning the particle angular distribution on the sky, meaning that one should be very careful with drawing any conclusions based on one particular model of the Galactic magnetic field.

-

-

3 Single pulsar as a plausible source We assume that particles with different energies are ejected isotropically by a short lived source, most likely a very young pulsar. Such very young pulsars (with milisecond periods) are presumably formed during the

W. Bednarek et al.: Propagation of charged particles from the Galactic Center

148

zoo 0

50

0

50 t i m e delay

100

150 200 in 1e3 years

n

Fig. 3 The differential spectra of neutrons (dashed curves) and y-rays (full curves) observed from the Galactic Center on Earth at the time 10, lo2,lo3, 3 x lo3, and lo4 yrs (from the thinnest to the thickest curve) after the formation of the pulsar. The thick dotted curve shows schematically the observed cosmic ray spectrum within the 20n circle.

supernova type Ib/c explosions. The precursors of these supernova types are probably lower mass WolfRayet or oxygen-carbon stars rotating very fast and which have small mass explosion envelopes. Let's assume that at least one of them had parameters allowing acceleration of protons to energies above 10l8 eV. We follow the suggestion that pulsar winds are able to accelerate particles to energies E corresponding to the full potential drop available across the polar cap region (Gunn & Ostriker 1969, Blasi, Epstein & Olinto ZOOO),

where 0 = 27r/P, P = 10p3Pn,, s is the pulsar period, R = lo6 cm is the radius of the neutron star, B = 1013BB13 G is its surface magnetic field, e is the elementary charge, and c is the speed of light. The above equation allows us to constrain the parameters of the pulsar able to accelerate protons to energies E 2 10'' eV. The following condition has to be fulfilled P,, 5 8B:,/2.If the pulsar loses its rotational only on emission of the dipole radiation with the power L, then its period at specific time t is energy, ErOt, determined by the equation

Erot,= L=+IO(~

=

B2R6f14

-

6c3

149

Astron. Nachr./AN 324, No. S1 (2003)

where 1 = 1.4 x g cmP2 i s the neutron star moment of inertia. As a result of this energy losses the period of the pulsar changes in time according to P&(t) = Po",,,,, + 3.5 x 10p8B?,t, where f is in seconds, and f i ~ , ? is ~ the, pulsar initial period. By using the above equations, we estimate the time elapsed from the pulsar formation, t,,,, during which protons will be accelerated above 10Is eV (assuming that Po.,,, << P,,,s(tacc)) to he f,,, xz 6OB;' yr. Since this time is relatively short, when compared to the time scale of particle propagation, we can consider such injection of relativistic protons as instantaneous. Let us consider whether a single pulsar could he responsible for the observable excess from the GC. The power emitted by the pulsar accelerating protons can be expressed by the proton energy E , L = cE'//lir' = 5.5 x 104"E?' ergs-', where E = 10'*El' eV. The rate, robs, at which protons arrive to m-2 spl the Earth from the 'AGASA-SUGAR source' can be estimated from the flux Fobs = 9 x derived from the SUGAR observations. It is = T R ~ F , ,= ~ , 1.7 ~ x 1O2'pparticles <5-'. On the other side the number of particles AN injected by the pulsar with energies between El and Ez would be

where Ze is the particle charge, and q is the efficiency of particle acceleration defined as the ratio of the number of injected particles to that of the maximum number possible given by the Goldreich & Julian (1969) density at the pulsar light cylinder. Our present calculations have shown that the source image changes significantly even if proton energy goes from 3 to 4 x 10'' eV (i.e. changes by 30%). Therefore, we think that it i s quite likely that the observed excess is actually due to a narrower energy band than the quoted factor of 4 (unless the source is quite close to us) and we adopt that E2/El = 1.2 (with E 3 x 10l8 eV). Then we have AN = 5 x 10"71/B1:i protons, For a pulsar located in the GC the fraction f of these particles giving the image close to the source (for model I, Fig Ih) is 5 x l o p 4 (about 550 particles out of 1.2 x lo6) i.e. 2.5 times larger that that for a straightforward propagation. As these particles arrive to the Earth vicinity within A f 5 x lo3 years, we should expect for their average rate N

Comparing this with we obtain that T ~ / B I lop4 : ~ for this case (a value probably not unreasonable!). It may be, however, that the observed excess is due to a more delayed image of a source located somewhere else. Let's assume (somewhat arbitrarily) that we can apply our calculation results to this case as well. For example, the secondary compact images (Fig. Ib,c,d,e) are visible for At 2 x lo4 years and the fraction of the emitted particles producing them equals ,f N 7 x lop5(- 80 particles per one secondary , y p l , and, to agree with image, out of 1.2 x lo6). Then their average rate is 7'2 5 x 1OZ7'r//Bl~prot,ons rot>s,r//Bls 3 x lop3. We can not, however, exclude a possibility that the observed excess is being produced by even more delayed protons with a smaller intensity and a longer arrival time At. If the total observed CR flux was not produced by pulsars (but by some other sources) then such a hunch of particles could produce an increase on the average CR flux. For the bisymmetric field model proposed by Han & Qiao (1994) the image of the source located at the GC does not appear in the direction to the real source up to 3 x 10" eV, hut is shifted from its real position by a large angle. Therefore, in this case protons can not be responsible for the observed excess. N

4

Conclusions

1) Protons with energies between (1 3 ) x 10'' eV injected instantaneously by a point-like source at the Galactic Centre can form multiple images at directions completely different from those towards the source, as well as images shifted only slightly from the position towards the source. 2) The results of particle propagation for the two considered magnetic field models give totally different predictions for the particle angular distribution on the sky. ~

W. Bednarek et al.: Propagation of charged particles from the Galactic Center

150

3) The application of the Han & Qiao magnetic model produces strong north-south anisotropy of parti(2 - 3) x 10l8 eV. Most of the delayed particles arrive from the southern Galactic

cles with energies

N

hemisphere. 4) Our results do not depend on the particular distribution of the irregular magnetic component, providing that its magnitude is not larger than assumed here. 5) The maximum proton fluxes predicted for a pulsar model are larger than that observed. Thus, the model of isotropic and instantaneous particle injection by a pulsar, located close to the GC, could explain the observed flux of particles in the AGASA-SUGAR excess.

Acknowledgements The work is supported by the grants: KBN No. 5P03D02521 and the University of t 6 d i .

References Beck, R. et al., A&A 292,409 (1994). Bednarek, W., M. Giller, M. Zielinska, J.Phys.G, 28,2283 (2002). Bellido, J.A. et al.,Astropart.Phys.15, 167 (2001). Blasi, P., Epstein, R.I.,Olinto, A.V., ApJ 533, L123 (2000). Goldreich, P., Julian, W.H., ApJ, 157,869 (1969). Golla, G., Hummer, R., A&A 284,777 (1 994). Han, J.L., Qiao, G.J., A&A, 288,759 (1994). Hayashida, N., et al., AstropactPhys. 10,303 (1999). Stanev, T., ApJ 479,290 (1997). Sukumar, S., Allen, R.J., ApJ 382, 100 (1991). Urbanik, M., Elstner, D., Beck, R., A&A 326,465 (1997).

Astron. NachrJAN 324, No. S1. 151 - 155 (2003) / DO1 10.1002/asna.200385029

Discovery Of New SNR Candidates in the Galactic Center Region with ASCAand Chandra

',

Atsushi Senda* Hiroshi Murakami2,and Katsuji Koyama'

' Cosmic Ray Group, Dept. of Physics, Kyoto University, Sakyo-ku, Kyoto, 606-8502, Japan

' Institute of Space and Astronautical Science (ISAS),3-1 - 1 Yoshinodai, Sagamihara,Kanagawa 229-85 10, Japan

Key words X-ray, ISM, supernova remnamts, G0.570-0.018 PACS 04A25

We report the discovery of diffuse X-ray features which are possible SNR candidates near the Galactic Center (GC) observed with ASCA and Chandra. G0.570-0.018 has extremely small (20" diameter) shell-like morphology. Its X-ray spectrum exhibits strong Fe-K line emission and is well fitted by an NEI model with a temperature of about 6 keV. These characteristics suggest that G0.570-0.018 is a quite young (t 100 year) SNR. Diffuse hard X-rays were also detected from G359.92-0.09. Its X-ray spectrum also exhibits strong Fe-K line emission. The X-ray excess coincides with the shell-type feature observed in the radio continuum (e.g.. Ho et al. 1985), which is attributable to a new SNR. In addition, we have discovered several soft X-ray clumps. Their X-ray spectra are thermal (kT I keV) and clearly show atomic line features such as Si, S, Ar and Ca. The origin of the diffuse X-ray emission from the GC region has been an unresolved issue for over a decade. Hard clumps such as (30.570-0.018 are likely to be young/middle-aged SNRs, and could produce the hot component of the GC plasma, while relatively soft (- 1 keV) clumps, which also may be SNRs, could contribute to the cool component of the GC plasma.

-

-

1 Introduction The origin of the diffuse X-ray emission from the Galactic center (GC)region has been an open issue for over a decade. Giiiga and ASCAfound a large-scale (1" x 1.8") thin-thermal plasma with strong line emissions from highly ionized atoms (Koyama et al. 1989, 1996; Yamauchi et al. 1990). The total thermal energy of the plasma is as large as ergs. Koyama et al. (1996) proposed that the plasma was created either by an energetic explosion that occurred at the central massive black hole (Sgr A * ) or by multiple 105 year. supernova explosions that took place within the past With its superior spatial resolution, C h a n d ~ ahas successfully resolved thousands of point sources in the GC region. However, most (- 90%) of the X-ray flux from the GC region is attributable to a diffuse component (Ebisawa et al. 2002; Wang et al. 2002). On the other hand, Chandvu also has revealed that the diffuse X-rays from the GC region are rather clumpy (Bamba et al. 2002). The presence of the clumpy structures may favor a multiple-SNe scenario. In fact, new X-ray supernova remnants (SNRs) have been discovered with Chandra (e.g. Sgr A East; Maeda et al. 2002). In this paper, we investigate newly discovered clumpy structures near the GC, which reveal the origin of the diffuse X-ray emission.

-

' Corresponding author: e-mad: sendaQcr.scphys.kyoto-u.ac.jp,Phone. +8175 753 3869, Fax: +8175 753 3799 @ 2007 WILEY-VCH Verldg GmhH & Cu KGaA Weinhem

A. Senda et al.: New SNR Candidates in the Galactic Center Reeion

152

2 Observations We used the archive data of seven fields of views (FOVs) of Chandra ACIS-I observations; Sgr B2, Sgr A*, and five FOVs of the Chundru GC Survey. The on-axis position and total exposure time of each observation are given in Tab. 1 . Table 1 Chandra ACIS observations

Target Name Sgr A* Sgr B2 GCS 13 GCS 14 GCS 16 GCS 17 GCS 19

Position (1, b) [deg] (359.94, -0.05) (0.59, -0.02) (0.00, -0.20) (0.00,0.00) (359.80, -0.20) (359.80,O.OO) (359.61, -0.20)

Exposure [sec] 48720 98989 10762 10762 10762 11261 11261

3 Results and Discussions 3.1 G0.570-0.018 G0.570-0.018 was discovered by ASCA (Sakano et al. 2002). Chandra observations resolved this source as a small shell-like structure with a diameter of -20”. The X-ray spectrum exhibits an extremely strong Fe-K line emission with an equivalent width of about 4 keV. The X-ray spectrum is well reproduced by a high temperature (- 6 keV) thin-thermal non-equilibrium ionization (NEI) model; hence the source is likely to be a young SNR. A detailed discussion of (30.570-0.0 I8 is given by Senda et al. (2002). 3.2 G359.92-0.09 Ho et al. (1985) detected a non-thermal radio continuum filament, called the “wisp,” 4’ south of Sgr A*. In addition, the inward curve of Sgr A East and the presence of other condensations (including the “wisp”) arranged in a circular shape imply the presence of a shell-like slructure, as shown by the solid circle in Fig. I. NH3 observations of the dynamics of nearby molecular material support this morphology, hence Coil et al. (2000) identified the source, G359.92-0.09, as a new SNR candidate. As shown in Fig. 1, the Chandra observations show that an X-ray excess fills the eastern half (EH) and southwest part of the radio shell of (3359.92-0.09 (Murakami et al. 2002). Although the northwest part shows no clear excess within the shell, this is due to the contamination of the intense X-ray emission from the SNR Sgr A East (Maeda et al. 2002). On the southwest edge of the shell, an X-ray bright filament is also discovered, which clearly corresponds with the non-thermal “wisp.” We have extracted X-ray spectra from three different regions; Eastern half (EH), Southwest quadrant, and the “wisp.” A thermal NEI model yields an acceptable fit for a spectrum from each region (Fig. 2 and Tab. 2). The observed emission measure from the EH is I . 1 x cmP3 using a distance of 0=8.5 kpc for G359.92-0.09). From the radius of the shell (2’ 5 pc), the total volume of the plasma is determined to be T/toral 1.6 x cm3. Assuming that the half of the X-ray emission comes from the EH, we calculate the electron density and the thermal energy of the EH to be n, 0.4 and EEH= 37&,kTVEH 1.6x 1050 ergs, where kT is the best fit value for the EH spectrum. Even though we cannot estimate the X-ray properties of the NW quadrant, a crude extrapolation of the EH result suggests that the thermal energy from the whole shell would be reasonably expected from a typical supernova explosion. Assuming that the expansion velocity of the SNR shock front is the sound velocity (1000 km s-l at 10 keV), the age of G359.92-0.09 is

-

N

-

-

-

Astron. Nachr./AN 324, No. SI (2003)

153

x

o N

1

5

2

10

channel energy (keV)

Fig. 2 The X-ray spectrum of the Eastern-Half (EH) of the Fig. 1 The C h a n d ~ aACIS grayscale image of (359.92-0 .09 in the 3.0-8.0 keV band. The overlaid contours are from the VLA,6cm continuum map (Ho et al. 1985).

G359.92-0.09 obtained with the Chandra ACIS. The best-lil spcctrum with the NEI model is shown by solid line.

Table 2 Best-tit Results of the X-ray Spectra o f G3.59.92-0.09 with Chamfra

kT (keV) T (10’’ s cmp3) NH cm-’) Fluxn (10-12 erg sP1 cm-2 )

Eastern Half 11.4 (>3.4) 1.4 (0.8-5.4) 6.0 (3.7-8.3) 1.5

Southwestern Quadrant 2.6 (> 1.9) 2.1 (>0.01) 18 (7.0-33) 0.4

”wisp” 12.7 0 7 . 9 ) 88 (>0.001) 37 ( 3 2 4 4 ) 0.4

An NEI model was employed to produce the estimated wlues. Quantities in parentheses are 90% confidence limits. “Flux (no correction for absorption) i n the 2.0-10.0 Lev hand.

determined to be 3800 year. On thc other hand, the ionization parameter indicates that the age of the plasma of the EH is 1.2 x I O4 year, although the uncertainties of this estimate is largc. N

3.3 G359.79-0.26 and (3359.77-0.09 The soft band image from the Charrdra GC Survey shows that diffuse emission extends from Sgr A East to the southward direction. The extended emission is relatively soft and clumpy (Fig. 3). From these clumpy structures, we havc identified three prominent soft clumps named G359.79-0.26, (3359.77-0.09, and (3359.73-0.35. The X-ray spectrum of each clump exhibits K-line emissions from He-like and/or H-like ions of Si, S, Ar, and Ca. These clumps were also detected previously by ASCd and ROS.4T. We have combined the X-ray spectra and tried to fit them with one model. The best fit results arc shown in Fig. 4 and Tab. 3. The large (NH 5 x lo2’ cm-2) absorption columns of (3359.79-0.26 and G359.77-0.09 indicate that they are located near the GC, while the significantly smaller ( N H lo2’ cm-’) absorption column of (3359.73-0.35 suggests that this clump is a foreground object. The results or spectral fitting with a thermal NEI model show that physical parameters of G359.79-0.26 and (3359.77-0.09 (NHand metal abundances) are similar to each other. In addition, the 1-3 keV band image suggests that (3359.79-0.26 and G359.77-0.09 are the southeastern and northwestern parts of a largc (- 30 pc) elliptical shell. This indicates that the two clumps have the same origin, an energetic N

-

A. Senda et al.: New SNR Candidates in the Galactic Center Region

IS4

I ,

1

2

.; 5

channel energy (keV)

Fig.

The image Of the ’Oft (1’0-3’0 kev) band near the GC region. The image center corresponds to an on-axis position of the FOV of GCS16.

Fig. 4 Combined X-ray spectrum of G359.79-0.26 obtained with ACIS-I (black), A S C A GIS2 (red), GIS3 (green), and ROSAT PSpC (blue), Solid lines show the best-fit thermal NEl model,

explosion such as a supernova, which occured at the center of the large shell. However, their temperatures are slightly different, so the interpretation of their origin is still preliminary. Table 3 Best-fit Results of the Combined X-ray Spectra of the Soft Clumps

kT (keV) T (10’~ s~ m - ~ ) N H (10” cm-’) FIUX~ ergs-’ cm-2 -Abundances (solar)-

Si S Arc Ca

)

G359.79-0.26 G359.77-0.09 (3359.73-0.35 0.84(0.75-0.93) 1.31 (1.03-1.79) 1.4(1.12-1.52) 0.5 (0.1-1.6) 9.9(-) 64 (>5.6) 4.9(4.65.2) 5.8 ( 5 . 1 4 . 5 ) 1.2(0.8-1.4) 2.3 2.2 I.Ib 0.4(0.3-0.6) 0.8(0.5-1.0) 1.3 (0.4-2.2) 3.2(1.4-5.6)

0.6(0.4-0.9) 0.8(0.5-2.1) 0.3 (< 1.0) 1.4 (< 3.3)

2.1 (1.2-3.8) 5.4(359.3) I 1 (3.1-27) 9.3(< 34)

Values in parentheses are 90% confidence limit\. aFlux (no correction of absorption) in the 2.&10.0 keV band. bFlux obtained with the ASCA GIS data. C A VMEKAL model is applied to obtain the results.

4 0

0

0

Summary

Combining Chandra observations and archive data from ASCAand ROSAT,we have discovered several X-ray clumps in the GC region. Some of these clumps ((30.570-0.018 and G359.92-0.09) show thermal spectra from high temperature (-1 0 keV) plasmas; others ((3359.77-0.09 and G359.79-0.26) show thermal spectra from lower temperature (- 1 keV) plasmas. The X-ray emission from G359.92-0.09 is a counterpart to the non-thermal radio shell. Its energetics suggest that (3359.92-0.09 is a young/middle-aged (3800-1.2~lo4 year old) SNR.

Astron. Nachr./AN 324, No. S I (2003)

155

The X-ray properties of G359.77-0.09 and G359.79-0.26 appar to be typical of Galactic SNRs, but the natures of these two objects are still uncertain. The X-ray spectra of G0.570-0.018 and G359.92-0.09 are similar to that of the hard component of the GC plasma, while the spectra of (3359.77-0.09 and G359.79-0.26 are similar to that of the soft component. These characteristics suggest that the GC plasma may be largely resolve into individual clumps, which arc likely SNRs. However, the total diffuse emission from the GC region is greater than the sum of detected GC SNRs by 1-2 orders of magnitude.

References Bamba, A,, et al. 2001, PASJ 53, L21 Bamba, A,, et al. 2002, Proc. of “New Visions of the X-ray Universe in the X M M - Newton and Chandra era”, in press (astro-ph/0202010) Coil, A. L, & Ho, P. T. P. 2000, ApJ 533, 245 Ebisawa, K., et al. 2002,Proc. of “X-ray Surveys in the light of new observations”, in press (astro-ph/0210681) Koyama, K., et al. 1989, Nature 339, 603 Koyama, K., et al. 1996, PASJ 48,249 Ho, P. T. P., et al. 1985, ApJ 288, 575 Maeda, Y., et al. 2002, ApJ 570, 671 Murakami, H. 2002, Ph.D thesis, Kyoto University. Sakano, M., et al. 2002, ApJS 138, 19 Senda, A., Murakami, H., & Koyama. K. 2002. ApJ, 565, 1017 Wang, Q. D., et al. 2002, Nature 415, 148 Yamauchi, S., et al. 1990, ApJ, 1990, 365,532

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1. 157- 160 (2003) / DO1 10.1002/asna.200385030

Molecular Line Observations of the Tornado Nebula and its Eye J. Lazendic’

‘.536,

M. Burton’, F. Yusef-Zadeh3,M. WardIe4, A. Green’, and J. Whiteoak‘

’ Harvard-SmithsonianCfA, 60 Garden Street, Cambridge MA 02138, USA

’ School of Physics, University of New South Wales, Sydney NSW 2052, Australia Department of Physics and Astronomy, Northwestern University, Evanston, IL 60208, USA Department of Physics, Macquarie University, NSW 2109, Australia School of Physics, University of Sydney, Sydney NSW 2006, Australia Australia Telescope National Facility

Key words shock waves, molecular clouds, supernova remnants, star formation, Tornado Nebula, G357.70.1, Eye of Tornado, G357.63-0.06 PACS 04A25 We present millimetre and NIR molecular-line observations of the Tornado Nebula and its Eye. The observations were motivated by the presence of OH( 1720 MHz) maser emission towards the nebula, believed to be an indicator of interaction between a supemova remnant and a molecular cloud. We found that the distribution of molecular gas around the Tornado complements its radio morphology, implying that the nebula’s appearance has been influenced by the structure of the surrounding molecular gas. Our NIR HZ ObSerVdtions revealed the presence of shocked molecular gas at the location where the nebula is expanding into the surrounding molecular cloud. It has been suggested that the Eye of the Tornado is related to the nebula on the basis of their apparent proximity. Our NIR and millimetre-line observations show that the two objects are not spatially related. Bry line emission, in conjunction with IR data at longer wavelengths and high-resolution radio continuum observations, suggests that the Eye is a massive protostellar source deeply embedded within a dense molecular core.

1 Introduction The Tornado nebula (G357.7-0.1) is a peculiar radio source located towards the Galactic Centre region. It has been classified as a supernova remnant (SNR) due to its steep radio spectrum and linear polarization (e.g., Kundu et al. 1974, Caswell et al. 1980, Shaver et al. 1985a), hut its unique morphology has led to other interpretations (e.g., an accretion powered nebula, Becker & Helfand 1985). The Eye of the Tornado (G357.63-0.06) is a compact radio source located 30” from the emission peak of the nebula. It was initially thought to he responsible for the formation of the nebula (e.g., through mass ejection from a pulsar or accreting binary system), hut was instead found to have a flat radio spectrum and was suggested to he an HII region (Shaver et at. 1985b).

2

Tornado

Frail et al. (1996) found a single OH(1720 MHz) maser at the northwestern tip of the Tornado (see Figure 1). When not accompanied by maser emission from the other three OH ground-state transitions at 1612, 1665 and 1667 MHz, the detection of this maser has been recognized as a signature of SNWmolecular * Corresponding author: e-mail: jlazendicQcfa.harvard.edu

02003 WILEY-VCH Vcrlag GmbH L Co

KGdA. Weinhem

158

J. Lazendic et al.: Tornado and its Eye

cloud interactions (see Koralesky et al. 1998 and references therein). Its presence may support the classification of the nebula as an SNR. The maser has a velocity of -12.4 km s-', implying a distance of 11.8 kpc to the nebula and placing the Tornado behind the Galactic Centre. 52'

53

/ 54

55'

56'

57' 58'

\ -30'59' 1?"3 RA (B1950)

U (B1950)

Fig. 1 (lefr) A 20 cm VLA image of the Tornado Nebula. The square marks the part of the nebula covered by the U N S W observations. frighr) Contours of' H2 line emission superimposed on a greyscale 20 cm radio continuum image of the northwestern part of the Tornado. The contour levels are: 1.6, 4.7, 6.3, 7.9, 9.5, 1 1 . 1 , 12.6, 14.2 x ergs-' cm-2 sr-'. The white and black crosses mark the location of the OH(1720 MHz) maser.

Using the University of New South Wales Fabry-Perot narrow-band tunable filter (UNSWIRF), we detected 2.12 pm Hz 1-0 S(l) emission towards the OH(1720 MHz) maser in the Tornado Nebula (see Figure 1). The correlation of the emission peaks in the radio continuum and H2 images suggest that the H2 emission originates from an expansion of a shock wave and is most probably shock excited, as found in other SNRs associated with the OH( 1720 MHz) maser (e.g., Lazendic et al. 2002a,b). The OH( 1720 MHz) maser is located at the western edge of the Hz emission, which is more sharply defined than the rest of the ring, probably delineating the leading edge of the shock front. Molecular transitions at millimetre wavelengths were also detected at the maser velocity of -12 km s-l using the 15-m Swedish-ESO Submillimeter Telescope (SEST). Emission from molecular species other than l 2 C 0 and 13C0, e.g., HCO+, HCN and H2C0, was found to be very weak (see Lazendic et al. 2003 for more details). Molecular gas associated with the OH(1720 MHz) maser and HP emission is optically ) . density is in agreement with the requirements for the thick, cold (-7 K) and dense (- lo5 ~ m - ~ This OH(1720 MHz) maser production in the post-shock gas behind the SNR shock front (Lockett et al. 1999), but the temperature is much lower than that expected in the post-shock gas in which the maser is created (50 - 125 K). However, since the cloud is optically thick, our CO observations are probing only the envelope of the cloud. Observations of more optically thin transitions of "CO and 13C0 are needed to examine the whole cloud temperature. The structure of the associated molecular gas complements the radio morphology of the Tornado Nebula (see Figure 2), implying that the distribution of the surrounding medium has influenced the nebula's unusual appearance. In particular, two minima in the molecular gas distribution, located symmetrically on each side of the nebula, coincide with large arc-like filaments in the nebula and point to locations where the shock could perhaps expand more readily than in the other regions.

Astron. Nachr./AN 324, No. S 1 (2003)

159

54'

55' 0

in

m

56

57'

58'

~ 31°59' 0 17h37m18'

12s

37m05

6'

545

RA (B1950) Fig. 2 An image of I3CO 1-0 emission (shown in greyscale and black contours) integrated between -16 and -9 kms-', overlaid with contours of the 20 cni radio continuum emission. The 1 3 C 0 contour levels are: 3.8, 5.7, 7.6, 9.5, 11.4, 13.3, 15.2 and 17.1 K km s-l. The lightest colours represent the weakest emission. The 20 cm contour levels are: 2. 7, 8.1, 13.5. 18.9, 24.3 x Jy heam-'. The cross marks the location of the OH( 1720 MHz) maser.

3 Eye of Tornado 14

14

16

16

18

D

-E 0

16

-

20

D "7

20

u

22

-2

22

1

CI

24

24

26

26

30'57 28 17h36m52'6

30'57 28 52'4

82'2

52'0

RA (si95a)

51.8

61'6

51.4

i7h36m52'6

52'4

52'2

52'0

51'8

51'6

31'4

RA (B1950)

Fig. 3 NIR and radio images of the Eye. ( I t $ ) 2.16pm continuum image of the field centred on the Eye, overlaid with the Bry line image. Contours are at 3, 5. 7, 9. 12, 15 and 16 x lo-'' W m-2 arcsec-'. (righr) 6 cm VLA image Jy beam-'. overlaid with contours of 20 cm VLA image. Contours are at 8,26, 52, 104, 156,208 and 258 x

160

J. Lazendic et al.: Tornado and its Eye

Using the 3.8-m UK Infrared Telescope in conjunction with the CGS4 spectrometer we found 2.16pm Bry emission towards the Eye. The emission peaks at a velocity of N -200 kms-l, which drastically differs from the velocity of the molecular gas associated with the Tornado. The velocity of the Bry emission indicates the distance to the Eye is 8.5 kpc, which makes the Eye foreground to the Tornado Nebula. A similar velocity towards the Eye has also been measured using the H92a radio recombination line (Brogan & Goss 2003). The Eye is resolved by the NIR and radio measurements (see Figure 3) as a compact HI1 region, and therefore must be undergoing massive star formation. It consists of four knots of emission, each about 1.5” across and of similar brightness, placed symmetrically about the perimeter of a 6” diameter circle. There are faint extensions -2” to the south and to the west in the Bry image, but no emission from the centre. A fit to flux measurements from our M R and other IR data obtained from Midcourse Space Explorer (MSX)and Znfrured Astronomy Satellite (IRAS)is consistent with the Eye being a warm (- 190 K). unresolved ( w 0.05”) blackbody source at the core of an extended ( w 5.5”), cold (- 35K) greybody. The best fit value of this two-component greybody gives an angular size for the Eye which is similar to the size derived from the NIR and radio images. The Eye’s integrated infrared luminosity of 2 x 104La suggests it harbors a massive (-12Ma) protostellar source, perhaps a BO star (see Burton et al. 2003 for more details).

-

References Becker, R. H. & Helfand, D. J. 1985, Nature, 313, 1 15 Brogan, C. & Goss, W. M. 2003, AJ, 125,272 Burton, M. G., Lazendic, J. S., Yusef-Zadeh, F. & Wardle, M. 2003, in preparation Caswell, J. L.; Haynes, R. F.; Milne, D. K.; Wellington, K. J. 1980, MNRAS, 190,881 Frail, D. A,, Goss, W. M., Reynoso, E. M., Giacani, E. B., Green, A. J. & Otrupcck, R. 1996, AJ, 11 I , 165 1 Koralesky, B., Frail, D. A., Goss, W. M., Claussen, M. J. & Green, 1998, AJ, 116, 1323 Kundu, M. R., Velusamy, T. & Hardee, P. E. 1974, AJ, 79, 132 Lazendic, J. S., Wardle, M., Burton, M. G., Yusef-Zadeh, F., Whiteoak, J. B., Green, A. J. & Ashley, M. C. B. 2002a, MNRAS, 331,537 Lazendic, J. S., Wardle, M., Green, A. J., Whiteoak, J. B. & Burton, M. G. 2002h, in “Neutron Stars in Supernova Remnants”, Eds. P. 0. Slam and B. M. Gaensler, p.339 Lazendic, J. S., Wardle, M., Burton, M. G., Yusef-Zadeh, F., Whiteoak, J. B., Green, A. J. 2003, MNRAS, in preparation Lockett, P., Gauthier, E. & Elitzur, M. 1999, ApJ, 51 1,235 Shaver, P. A., Salter, C. J., Patnaik, A. R., van Gorkom, J. H. & Hunt, G. C. 198Sa, Nature, 3 13, 113 Shaver, P. A,, Pottasch, S. R., Salter, C. J., Patnaik, A. R., van Gorkom, J. H. & Hunt, G. C. 1985b, A&A, 147, L23

A a o n Nachr./AN 324, No. Sl. 161 - 165 (2003)/ DO1 10 1002/aana.200385031

The Search for Water and Other Molecules in the Galactic Centre with the Odin Satellite Aa. Sandqvist*I , P. Bergman’, A. Hjalmarson’, E. Falgarone3,T. Liljestrom4, M. Lindqvist2, A. Winnberg?, and the Odin Team’ Stockholm Observatory, SCFAB-AlbaNova,SE-106 91 Stockholm, Sweden Onsala Space Observatory, SE-439 92 Onsala, Sweden Ecole NomialcSupCrieure,FR-75005 Pans, France Metsahovi Radio Observatory, Helsinki University of Technology,FIN-02 150 Espoo, Finland

’ ’ ’ ‘ ’ http://www.snsb.se/Odin/Odin.html

Key words HzO, Sgr A Complex, CND, +20 km s-l cloud, +SO km s-’ cloud PACS 04A25

Observations with the Odin midsubmni space telescope have been made towards the Sgr A Complcx (the CircumNLiclear Disk, the +20 and +SO kin s- molecular clouds) in the Galactic Centre and we report here on the results of searches for HaO, H2”O and other molecules in these regions.

1 The Odin Satellite Odin is a millimetreisubmillimetre wave spectroscopy astronomy and aeronomy satellite, launched with a START-I rocket on 20 February 2001 from Svobodny, Russia in far-eastern Siberia. It has a I .I-m highprecision telescope with a beam efficiency ol‘ about 90% and beamwidths ol‘ 2‘.1 and 9 ’ 5 at submm and mm wavelengths, respectively. Its pointing uncertainty is < 10“. The submm radiometer consists of four cryo-cooled submm receivers tunable in the frequency range of 486 - 580 GHz with a single sideband tcmperature of M 3000 K. A cryo-cooled HEMT receiver, which is tuned to 1 19 GHz and dedicated to the search for 0 2 , has a single sideband temperature of M 600 K. The backend spectrometers consist oi an acousto-optical spectrometer (AOS) with a total bandwidth of 1040 MHz and two auto-correlators with bandwiths in the range of 100 - 800 MHz, corresponding to velocity resolutions of 0.08 - 1 .0 km s p L . The satellite is described in detail by Frisk et al. (2003) and the receiver calibration by Olberg ct al. (2003). Figure 1 is a photo montage of Odin in Earth orbit. Odin is a Swedish-led satellile project funded jointly by the Swedish National Space Board (SNSB), the Canadian Space Agency (CSA), the National Technology Agency of Finland (Tekes) and the Centre National d’Etudes Spatiales (CNES, France). The Swedish Space Corporation was the prime contractor i’or development and launch of Odjn and is responsible for the operation of the satellite.

2 Odin Observations of the Galactic Centre The molecular complex associated with Sgr A consists predominantly of a molecular belt comprising the “+50 km cloud” (M-0.02-0.07), the “+20 km spl cloud’ (M-O.13-0.08), and the Circuinnuclear Disk *

Corresponding author: e-mail: [email protected]: +46 - 8 5537 851 I . Fax: +46 - 8 5537 8510

@ 1001 WILEY-VCH M r l a g GmhH & Co. KGaA. Wrmhcua

Aa. Sandqvist et al.: HzO towards S.gr A

162

Fig. 1 The Odin millimetre/submillimetresatellite

(CND) which surrounds Sgr A West and has a rotational velocity of the order of 100 km s-' in the same direction as the rotation ofthe Galaxy (see Fig. 2). These warm and high-density Galactic Centre molecular clouds are intimately entwined and interact with the continuum complex of Sgr A (e.g. Sandqvist 1989; Zylka et al. 1990) Three positions towards Sgr A have so far been observed with Odin, namely Sgr A* with the CND, the +20 km s-' molecular cloud, and the +50 km spl molecular cloud. The observed positions are marked on Fig. 2. Observations have been made in the spectral lines of 119-GHz 02,487-GHz 0 2 , 492-GHz C I , 548-GHz Ha80, 557-GHz Hi60, 572-GHz NH3, and 576-GHz ( J = 5 - 4) CO. However, only the data for H i 6 0 and Hi8O have been fully calibrated and reduced so far and they are presented in the next section. These results have recently been published as part of a special Odin Letters issue of Astronomy & Astrophysics (Sandqvist et al. 2003).

3 Water in the Sgr A Complex and the Expanding Molecular Ring Strong emission and absorption lines have been observed in the Ha60 line at all three Sgr A positions and they are presented in Fig. 3. However, no spectral line features can be detected in the H;80 line towards Sgr A* CND down to the rms noise limit of x 0.02 K. The line of sight towards the Sgr A Complex also crosses the massive x 180-pc Expanding Molecular Ring (EMR) surrounding the Galactic Centre and various spiral arm features further out in the Galaxy. A Gaussian analysis has been performed on the Sgr A' CND H!j60 profile in Fig. 3 using four absorption components and two emission components. The continuum emission was first subtracted out by fitting a linear baseline to the outermost channels on either side of the profile. The results were: Feuture Source = (Velocity (km spl), T i (K), Halfwidth (kms-l)) - I CND = (f73.2, f0.32, 88.5); IZ CND = (-31.6,

Astron. Nachr./AN 324, No. SI (2003)

I63

-28 54

56

z 0

t;r

z

58

A

0 W P

-29 00

02

04 1 43 00

45

30

15

RIGHT ASCENSION Fig. 2 Observed positions (epoch 1950.0) in the Galactic Centre Sgr A region are marked by circles whose diameter indicates the Odin beam (2'.1) - Sgr A* CND: (17"42"'29".3, -28"59'18"); +20 kms-' Cluud: (17"42"29*.3, -29"02'18"); +50 kms-' Cloud: (17h42"41s.0, -28"8'00''). The Complex consists of the 20-cm continuum radiograph, the CND (thin lines) in HCN, and the molecular belt in 2-mm HzCO (solid lines) with isovelocity contours (broken lines). The cross marks the position of Sgr A* (adapted from Sandqvist 1989).

f0.24, 47.9); III Local Sgr Arm = (-4.8, -0.24, 13.4); IV -30 kms-' Arm = (-30.2, -0.25, 11.0); V 3 - k p c A m = (-53.5, -0.21,8.1); VIEMI?= (-132.2, -0.09,60.0). The first two components, I and 11, both seen in emission, are believed to originate in the rapidly rotating CND. The northeastern part of the CND is receding and the southwestern part approaching, which gives the asymmetric, somewhat double-peaked line profile structure. The 2. I-arcmin beam of Odin encloses fully the CND and the resulting velocity structure of the profile is reminiscent of that seen in many other molecular lines. The two H 2 0 profiles towards the +20 and +50 km S K ' clouds in Fig. 3 are marked by the characteristic emission component from these molecular clouds at velocities near +20 km s-' and +50 km s-', respectively. Furthermore, a new molecular feature in the Galactic Centre can now been identified. It is detected as broad H i 6 0 absorption in the velocity range of M +120 to +220 kms-'. We shall call this feature the High Positive Velocity Gas (HPVG). This feature is not seen in the Sgr A* CND profile, which we interpret as being due to the background continuum emission seen at this position being somewhat lower than towards the dust continuum peak emission from the Sgr A +20 and +50 km sP1 molecular clouds (seen in 800 and 350 pm continuum maps of Lis & Carlstrom 1994 and Dowel1 et al. 1999, respectively). Evidence for the existence of the HPVG in the Sgr A region, seen in other spectral lines, is scarce although some is present in IS0 mid- and far-infrared observations of H 2 0 (Moneti et al. 2001), and VLA observations of H I (Dwarakanath et al. 2003, in preparation) and OH (Karlsson et al. 2003). The HPVG should not be confused with the molecular gas in the far side of the EMR whose velocity falls inside the

164

Ad.

-- -

F -*

$-

$ 0

2

N

b "2 a

:: ti

d

elB

5z ? 0

Sandqvist et at.: HzO towards Sgr A

$I 2 , -200

,

,"

,

,

-100

L'',Y,

,

, ,

0

,

, 100

,

,

,

,

, 200

-200

-100

0

100

200

100

7-00

Velocity [Irm/.]

Velocity [hm/s]

-200

-100

0

Velocity [hm/s]

Fig. 3 The Ha60 profiles towards the Sgr A Complex.

same range but whose emission lines are narrower. Also, the HPVG is seen in absorption which places it in front of the Galactic Centre continuum sources and thus it cannot be part of the far side of the EMR. The broad H2O absoption component (VI), seen at velocities near-132 km sP1, is observed in all three positions. This feature has its origin in the near side of the EMR.

4 Water Abundance in Spiral Arms Three narrow HzO absorption components, 111, IV and V, seen at velocities near -5, -30 and -53 kms-l, respectively, are observed at all three positions (see Fig. 3) and are well-known Galactic spiral arm features, which were first identified in early 21-cm H I observations. They originate along the line of sight crossing the so-called Local Sgr, -30 kms-' and 3-kpc spiral arm structures. The absorption feature (111) at -5 km s-l appears to be the strongest and, judged by estimated continuum levels, this feature has an optical depth of at least one. With this assumption, a lower HzO column density limit of > 2 x 1013 is found for feature 111 in the Sgr A* CND profile, using the appropriate line width and an excitation temperature of 15 K. We can estimate Hz column densities using the C180-profiles in Fig. 4 which are the C l 8 0 profiles resulting for the three H 2 0 positions from a convolution of the SEST Cl80 (1 - 0) survey of the Galactic Centre spectra (Lindqvist et al. 1995) to a resolution of 2' (corresponding to the Odin beam size).The integrated intensities have been determined over the regions corresponding to the three narrow H20 absorptions. The H2 column densities have then been calculated by assuming optically thin emission, an excitation temperature of 15 K and a C l 8 0 abundance of 2 x lop7 with respect to H a (Frerking et al. 1982). This results in an H2 column density of 1.0 x loz2 for feature I11 in the Sgr A* CND profile. The corresponding H2O abundance limit for this feature is then > 2 x loP9.

I65

Astron. Nachr./AN 324. No. SI (2003) a] t50 k m

-

Q-'

2-

b) Sgr A' CND

Cloud

Y

I

2e : zi : -

?j:€P2 2 -

Our non-detection of Hg80 towards Sgr A* CND provides an upper limit on the HzO abundance in the narrow absorption features. Using a lsO/lsO ratio of 500 (Wilson & Rood 1994) for this local absorbing cloud (111) and an excitation temperature of 15 K, we obtain an upper limit for the H 2 0 column density of 5 x l O I 4 cm-'. Hence, for the Local Sgr Arm absorption, the 3a upper limit of the H 20 abundance becomes 5 x lo-* while the lower limit was found to be 2 x lo-'. The average H.20 abundance estimated for the foreground gas towards Sgr B2 by Neufeld et al. (2000) is 6 x 10W7, which is about an order of magnitude higher than our range towards Sgr A. On the other hand, our range is in better agreement with HzO abundances found in giant molecular cloud cores by Snell et al. (2000) and in a local diffuse molecular cloud by Neufeld et al. (2002).

References Dowel1 C.D., Lis D.C., Serabyn E., Gardner M., Kovacs A,, Yamashita S. 1999, in ASP Conf. Ser. 186, The Central Parsec of the Galaxy, eds. H. Falcke, W.J. Cotera, W.J. Duschl, F. Melia, M.J. Rieke, 453 Frerking M.A., Langer W.D., Wilson R.W. 1982, ApJ 262,590 Frisk U., Hagstriim M., Ala-Laurinaho J. et al. 2003, A&A, 402, L27 Karlsson R., Sjouwerman L.O., Sandqvist Aa., Whiteoak J.B. 2003, A&A 403, 101 1 Lindqvist M., Sandqvist Aa., Winnberg A., Johansson L.E.B., Nyman L-A. 1995, A&AS 113, 257 Linke R., Stark A.A., Frerking M.A. 1981, ApJ 243, 147 Lis D.C., Carlstrom J.E. 1994, ApJ 424, 189 Moneti A., Cernicharo J., Pardo J.R. 2001, ApJ 549, L203 Neufeld D.A., Ashhy M.L.N., Bergin G. et al. 2000, ApJ 539, L111 Neufeld D.A., Kaufman M.J., Goldsmith P.F., Hollenbach D.J., Plume R. 2002, ApJ 580, 278 Olherg M., Frisk U., Lecacheux A. et al. 2003, A&A, 402, L35 Sandqvist Aa. 1989, A&A 223,293 Sandqvist Aa., Bergman P., Black J. et al. 2003, A&A, 402, L63 Snell R.L., Howe J.E., Ashby M.L.N. et al. 2000, ApJ 539, L101 Wilson T.L., Rood R. 1994, ARA&A 32, 19 1 Zylka R., Mezger P.G. Wink J.E., 1990, A&A 234, 133

This Page Intentionally Left Blank

Astron. Nachr.iAN324, No. Sl, 167- 172 (2003) / DO1 IO.lO02/asna.20038.51 10

Deep X-Ray Imaging of the Central 20 Parsecs of the Galaxy with Chandra

',

Mark Morris* Fred Baganoff', Michael Muno2, Christian Howard', Yoshitomo Maeda", Eric Feigelson4, Marshall Bautz2, Niel Brandt4, George Chartas4, Gordon Garmire4, and Lisa Townsley4

' Division of Astronomy, UCLA, Los Angeles, CA 90095-1562. USA Center for Space Research, Massachusetts Institute of Technology, Cambridge, MA 02139-4307, USA Institute of Space & Astronautical Science, 3- 1-1 Yoshinodai, Sagamihara, Kanagawa, 229-8510, Japan Department of Astronomy &Astrophysics, Penn. State Univ., 525 Davey Lab., University Park, PA 16802, USA

Key words Galaxy: center, X-rays: ISM, Sgr A*

Abstract. A deep observation toward the Galactic center with the Chandra X-Ray Observatory revealed a number of extended features, in addition to Sgr A* and SgrA East. Here, we focus on two curious, extended X-ray structures: large-scale (-10 pc) bipolar lobes centered on Sgr A* and a bright cometary source located 0.3 pc from Sgr A*, CXOGC 5174539.7-290020. The bipolar lobes consist of a number of emission clumps oriented along a line perpendicular to the Galactic plane, suggesting that a series of ejections has taken place on characteristic time scales o f hundreds to thousands of years. The clumps are embedded in a low-intensity, edge-brightened lobe which is most evidcnt in a flux ratio map. At two locations along the lobe, nonthermal linear features are present, suggesting that relativistic electrons may be impinging on the compressed, magnetic wall of this structure. The cometary X-ray source has no counterpart at other wavelengths; its orientation is consistent with a high-velocity neutron star ejected from the grouping of stars at IRSl3, but there are problems with that hypothesis, and other models warrant consideration.

1 Introduction The ACIS instrument aboard the Chandra X-Ray Observatory has been pointed at the Galactic center on several occasions since 1999, culminating in a deep observation in 2002 May-June. With a total useful exposure time of 590 ksec, a rich variety of phenomena are revealed, and are reported here: Sgr A*, its extent, its time variations and its possible jet (Baganoff et al. 2003a, 2003b), stellar point sources (Muno et al. 2003), SgrA East (Maeda et al. 2003), and extended continuum and line sources emission (Park et al. 2003). In this paper, we focus on a few extended X-ray sources of particular interest, both of which have been newly revealed by Chandra: bipolar lobes centered on Sgr A* and an unusual cometary source located near Sgr A* (Baganoff et al. 2 0 0 3 ~ ) .

2 The Bipolar Lobes Figure 1 shows a broad-band image of the center of the ACIS field of view, using data from 3.3 to 4.7 keV. The region around Sgr A*, including SgrA East, IS saturated in order to emphasize the fainter structures oriented perpendicular to the Galactic plane. A number of diffuse, extended structures are present in the image, but the most prominent of them are aligned along a line passing through Sgr A* and oriented perpendicular to the Galactic plane. Overall, these features appear to form a roughly point-symmetric * Corresponding author: e-mail: [email protected],Phone: 1 310 825-3320, Fax: 1 3 10 206-2096

@ 2003 WILEY-VCH Verlag GmhH

Co KGaA. Wemhcm

168

M. Morris et a].: Extended X-Ray Structures

structure centered on Sgr A*, reminiscent of bipolar nebulae around mass-losing stars. The bipolar lobes are apparently filled structures having in places a discernible sharp edge. The well-defined edge is better revealed by figure 2, which shows the flux ratio in two bands, emphasizing features with relatively soft spectra. The ratio map is affected somewhat by absorption by molecular clouds in the near foreground to the lobe emission (Howard, in preparation), which probably contributes at least partially to the asymmetry of the ratio map. Thermal fits to the ‘‘blobs’’ along the central axis of the bipolar lobes give temperatures of 2 keV, on average, with variations of a factor of 2 or so. The uncertainties in the lobe temperatures are considerable because of the difficulty of finding an appropriate local background to subtract. At present, no radial trends in gas temperature are evident.

Fig. 1 Broad-hand (3.3 - 4.7 keV) X-ray emission from a 35 x 28 pc region centered o n Sgr A * . The image has heen adaptively smoothed, and all point sources identified by Muno et al. (2003) have been removed. The true Galactic plane and the positioii of Sgr A * are indicated in black.

Along the northern edge of the northwestern lobe lie two linear X-ray features, shown in the unsmoothed image in Figure 3 as features “e” and “f”. These features have nonthermal spectra with power law indices of 1.3. Source “ d is very similar, but it bears no obvious relationship to one of the bipolar lobes. It is not clear whether these structures are truly filamentary, or whether they are bright by virtue of being edge-on surfaces. If the former view applies, thcn they are likely to be magnetic structures, by analogy with the nonthermal radio filaments located in this region. If the latter picture is applicable, then sources “e” and

Astron. NachrJAN 324, No. S1 (2003)

169

Fig. 2 "Softness Ratio" map of the bipolar lobes, consisting of the intensity ratio of the 2 - 4.7 keV band to the 4.7 - 8 keV band. The region shown is 25 x 44 pc (10 x 17.5 arcmin). For both bands, adaptively smoothed images are used, and all point sources meeting our selection criteria have been removed (see Muno et al. 2003). The dark spots represent unsnbtracted, hard sources. The central region around Sgr A* is not well represented by the point-source extraction routine because of the extremely high surface density of point sources there. Note the well-defined edge to both lobes.

"f" may be strong shocks occuring where relativistic particles occupying this lobe strike the compressed medium at the edge of the lobe. A compelling interpretation for the bipolar lobes straddling Sgr A* is that they result from energetic mass ejections from the immediate environment of Sgr A*, presumably from an accretion disk (see Baganoff et al. 2003a for a discussion of a possible jet which may be feeding the X-ray lobes). The 1 - 2 arcminute radial separations of the blobs corresponds to time intervals of (2500 - 5000 years)x(1000 km s - ' N ) , where V is the unknown outflow velocity of the X-ray emitting material. These characteristic times are somewhat less than the expected time between successive tidal disruptions of stars by the central black hole, 1-3 x lo4 years (Alexander & Hopman 2003), unless V is unreasonably small, so it may be difficult to ascribe the mass ejections to stellar tidal disruptions. The time scale for arrival at the outermost observable extent of the bipolar lobes is (1.5 x lo4) x (1000 km s-lN),which can be compared to the expansion time of SgrA East, -lo4 years (Maeda et al. 2003; Mezger et al. 1989; Uchida et al. 1998). This possible correspondence in time scale is interesting in the context of the suggestion of Maeda et al. (2002,2003) that the expansion of the shell of SgrA East past the location of Sgr A* might have provoked an accretion event in the recent past, although that event was hypothesized to have been only a few hundred years ago. The bipolar X-ray lobes have counterparts in the radio, but the detailed correspondence is better at long radio wavelengths (90 cm; Pedlar et al. 1989; Nord et al. 2003) than at 20 cm or shorter wavelengths (Yusef-Zadeh & Morris 1987). It is only with the most recent 90-cm data of Nord et a]. (2003) that the bipolar lobes have appeared to be anything else but an extension of the radio halo of the SgrA complex. The steep-spectrum, low-frequency radio emission can be interpreted as synchrotron emission from a population of energetic electrons that are closely related to those responsible for the presumably thermal emission

170

M. Moms et al.: Extended X-Ray Structures

Fig. 3 Linear X-ray features near Sgr A*, displayed in an unsmoothed image of SgrA. The features labelled “a” corresponds to the “Plume” discussed by Baganoff et al. (2003~)and by Maeda et al. (2003), while feature “b” is the jet described by Baganoff et al. (2003a). Feature c lies at the base of the nortbwest X-ray lobe, and is coincident with the radio “streamers” discussed by Yusef-Zadeh & Moms (1987). Features d, e, and fare all short, bright segments having nonthermal spectra with power law indices of 1.3.

in the X-ray lobes. The total amount of mass involved in the X-ray lobes is quite small; the column density of electrons required to give the observed intensities, assuming a thermal emission model, implies a space density of only about 1 cmp3 if the line-of-sight depth of the blobs is equal to their widths on the sky. This, in turn, implies that each blob has only about 1 Mo.

3 The Cometary Source, CXOGC 5174539.7-290020 The central 18” x 20” of the Chandra field (Figure 4) shows a concentration of at least 4 bright sources, one of which is Sgr A*. Approximately 8” (0.3 pc) to the northwest of Sgr A* lies a remarkable ridge

Astron. Nachr./AN 324, No. S1 (2003)

171

of X-ray emission. The intensity of this structure i s continuous along this ridge, increasing monotonically from 2” W of Sgr A* to a bright tip at IS”W, 8.5’’ N. At this point, we do not rule out the possibility that it consists of an unresolved string of 3 or more point sources, but there is no hint of that in the intensity image, and there is no evidence that the spectrum is also not continuous or homogeneous. The spectrum is well fit by a foreground-absorbed power-law having no indication of spectral line emission. With an overall spectral index of 1.7, it stands out as the source with the hardest spectrum in the immediate vicinity of Sgr A*. The absorption-corrected luminosity of this source is 1 .I x 1 034ergs s p l , almost 5 times larger than that of Sgr A* . The cometary geometry of CXOGC 5174539.7-290020 suggests a rapidly moving source of energetic particles, The spectral variations, shown in figure 4, are consistent with this, in the sense that the spectrum is hardest at the bright tip of the structure, and declines into the “tail”. This suggests that the source of particles is at the tip, that those particles are deposited in the source’s wake, and that they experience a net loss of energy as they radiate and distance themselves from the source. The nature of the source remains a mystery. There is no counterpart in the radio (e.g., Zhao & Goss 1998, Yusef-Zadeh, Roberts & Biretta 1998),in the mid-infrared (Morris et al., unpublished Keck data: see http://irastro.jpl.nasa.gov/GalCen/galcen.html), and the Gemini near-infrared images (Rigaut et al. 2003) show no obvious source associated with either the tip or the tail of CXOGC 5174539.7-290020. One clue may be provided by its orientation. The tail orientation is inconsistent with an origin based on linear ejection from Sgr A*; however, the tail does point back to the X-ray/infrared/radio source 1RS13, which appears to be a cluster of young, high-mass, stellar objects, possibly the surviving core of a substantial cluster (Maillard et al. 2003). A high-velocity neutron star could have been produced in such a cluster, either as the result of a supernova explosion, or by being gravitationally scattered out of the evaporating, tidally disrupting cluster (Kim & Morris 2003). If we posit that the speed of the neutron star is 300 km s-l in the plane of the sky, then the time scale for it to displace itself from IRS13 by the observed amount is I400 years. A supernova of this age should still be recognizable, even in this complex region, so ejection in some relaxation process would therefore be the more likely origin. Indeed, X-ray nebulae associated with pulsars can have a cometary morphology. For example, the pulsar B 1957+20, which has an X-ray powerlaw photon index of 1.9 f 0.5, has a cometary tail about half as long as that of CXOGC J 174539.7-290020 (Stappers et al. 2003), although the luminosity ofB1957+20 is three orders of magnitude smaller than the galactic center object. Furthermore, the absence of a radio source coincident with 5174539.7-290020 is mysterious if it i s truly a young neutron star. Consequently, other models warrant consideration. We mention three others here, although none of them are very compelling at this stage: 1) that the ridge results from the emission from hot electrons in a unidirectional jet emanating from some source in IRS13, 2) that the ridge is the laterally moving shock site where a rapidly precessing jet originating at Sgr A * impacts some (currently unseen) ambient medium (this model is inconsistent with the straightness of the jet on the opposite side of Sgr A*; Baganoff et al. 2003a), and 3) that the ridge is an essentially edge-on shock front resulting from a high-velocity wind, perhaps from the mass-losing stars in the IRS I6 cluster, impacting an ambient medium. Clearly, additional work remains to be done before the nature of this interesting cometary source can be clarified.

Acknowledgements This work has been supponed by NASA grants NAS8-00128, NAS8-38252 and G02-3115B.

References Alexander, T. & Hopman, C . 2003, ApJL, suhrnitted Baganoff, F.K., et al. 2003a, these proceedings Baganoff, EK., et al. 2003b. these proceedings Baganoff, F.K., et al. 2003c, ApJ, 590, in press Kim, S.S. & Moms, M. 2003, these proceedings Maeda. Y., et al. 2002, ApJ, 570, 671 Maeda, Y., et al. 2003, these proceedings Maillard, J.-P. et al. 2003, these proceedings

172

M. Morris et al.: Extended X-Ray Structures

r =1.55 +/- a.15

Fig. 4 Unsmoothed X-ray image of an 18” x 20” region showing the several sources near Sgr A* (marked with an “x”). The individual pixels have widths of 0.5 arcsec. CXOGC 5174539.7-290020 is the vertical ridge in the upper middle of the figure. The spectra of the entire source, and of the upper and lower parts, as indicated, are shown at the left, along with their power-law indices, r. Mezger, P.G. et al. 1989, A&A, 209,337 Muno, M., et al. 2003, these proceedings Nord, M.E., et al. 2003, these proceedings Park, S., et al. 2003, these proceedings Pedlar, A., et al. 1989, ApJ, 342, 769 Rigaut, F., et al. 2003, these proceedings Stappers, B.W., Gaensler, B.M., Kaspi, V.M., van der Klis, M. & Lewin, W.H.G. 2003, Science, 299, 1372 Uchida, K.I., Moms, M., Serabyn, E., Fong, D. & Meseroll, T. 1998, in IAU Syinp. No. 184: The Central Regions of the Galaxy and Galaxies, ed: Y. Sofue, (Dordrecht: Kluwer), p. 317 Yusef-Zadeh, F. & Moms, M. 1987, ApJ, 320,545 Yusef-Zadeh, F., Roberts, D.A. & Biretta, J. 1998, ApJL, 499, L159 Zhao, J.-H. & Goss, W.M. 1998, ApJL, 499, L163

Astron. Nachr./AN 324, No. S1, 173- 179 (2003) / DO1 10.1002/asna.200385099

Mapping Magnetic Fields in the Cold Dust at the Galactic Center David T. Chuss*', Giles Novak', Jacqueline A. Davidson3, Jessie L. Dotson', C. Darren Dowell', Roger H. Hildebrand6, and John E. Vaillancourt'

' NASA Goddard Space Flight Center ' Northwestern University

' USRA

NASA Ames Research Center

Caltech

' University of Chicago

' University of Wisconsin Key words Magnetic Fields, Galactic Center, Submillimeter Polarimetry, Dust

Abstract. We report the measurement of 158 new 350 pm polarimetry vectors in the central 30 parsecs of the Galactic center. These data were obtained at the Caltech Submillimeter Observatory using Hertz. Morphologically, these results show a consistency with previously published far-infrared and submillimeter results. We find that the angle of the magnetic field inferred from these observations is related to the 350 pm flux as obtained by SHARCKSO in the following way. At low fluxes, the magnetic field angle is consistent with that of a poloidal field as seen in nearby features such as the Galactic Center Radio Arc and the Northern and Southern Threads. At high fluxes, the magnetic field is oriented parallel to the plane of the Galaxy. This relationship suggests a model in which an initially poloidal field is sheared out in dense regions which are dominated by gravity. If this model is correct, it implies a characteristic field strength for the region of3 mG.

1 Introduction Ever since the discovery of the Galactic center Radio Arc (Yusef-Zadeh, Morris, & Chance, 1984), there has been much interest in the structure and strength of magnetic fields in the Galactic center. Even so, relative to other physical processes and structures in the Galactic center, comparatively little is know about the magnetic fields in this region. The study of the structure of the field was initiated by the observation that the Radio Arc and a number of other NTFs in the Galactic center (see LaRosa et al. 2000 and references therein) are oriented in a direction perpendicular to the Galactic plane. Such observations led to the hypothesis that the Galactic center is threaded by a global poloidal field. More recent results have given clues that the magnetic field structure in the Galactic center is more complicated than that of a simple poloidal field. Preliminary far-infrared observations of the Circumnuclear Disk (Morris et nl. 1988) indicated that the field running through regions of dust in the Galactic center does not trace a uniformly poloidal field. More recently, Novak et al. (2003) have used submillimeter polarimetry to trace the structure of the field in the central 200 parsecs and have found that the field in the cold dust on these scales is predominantly parallel to the Galactic plane. To reconcile their findings with the observations of the NTFs, these authors invoke a model (Uchida, Shibata & Sofue 1985) in which an * e-mail: chussQstars.gsfc.nasa.gov, Phone: 301 286 1858

@ 2003 WlLEY-VCH Veriag GmhH & Co KGaA. Wcmhem

D. T. Chuss et al.: Mapping Magnetic Fields in the Cold Dust at the Galactic Center

174

initially poloidal field can be sheared into a toroidal field in regions of relative high density by differential rotation and infall. The study of the strength of the field has a similarly interesting history. An estimate of the strength of the field associated with the Radio Arc was obtained by Yusef-Zadeh and Morris (1987). They estimated a lower limit of the field strength of - 1 mG by noting that the filaments maintain their structure against the ram pressure from interacting molecular clouds. The presence of such a large field strength led to two possibilities. First, it is possible that the mG field is the characteristic strength of a global field. However, the magnetic pressure implied by such a large global field strength is substantial and would require a large force for containment. If, on the other hand, the magnetic field in the filaments is enhanced with respect to that in the surrounding regions, the filaments would tend to dissipate on short timescales. There have been many Zeeman measurements of the line-of-sight field strength that have given values of between 10 pG and a few mG for various regions of the Galactic center (Plante, Low, & Crutcher 1995; Killeen, Lo, and Crutcher 1992). In some cases, such measurements provide a lower limit for the field strength, but without knowledge of the strength of the field in the plane of the sky, they do not help differentiate between the two ideas mentioned above. Far-infrared and submillimeter polarimetry provides a method for measuring the structure of the magnetic field as projected onto the plane of the sky. Rotating asymmetric dust grains become partially aligned such as to emit radiation in the far-infrared and submillimeter that is polarized in a direction perpendicular to the aligning magnetic field. Because of the extended distribution of dust in the Galactic center (Pierce-Price et al. 2000), this technique is an excellent way to obtain information about the structure of the magnetic field in this region. Unfortunately, because of uncertainties in dust grain physics and in the magnetic field component parallel to the line of sight, it is difficult to obtain a direct estimate of the magnetic field strength at any one point. However, using a model-based approach, we are able to obtain an indirect estimate of the magnetic field strength by looking at the overall structure of the measured field. We present submillimeter polarimetry of the central 30 pc of the Galactic center. We find that the magnetic field direction is neither entirely poloidal nor toroidal, but rather, the direction of the field depends on the submillimeter flux in the following way: for low fluxes, the field has an orientation perpendicular to the Galactic plane (poloidal); for high fluxes, the field’s orientation is parallel to the plane (toroidal).

2 Observations The 158 new polarimetry vectors were obtained at the Caltech Submillimeter Observator in May 2001. They are shown along with the data from Novak et al. (2000) in region 111of Figure 1. These data were obtained with the University of Chicago polarimeter Hertz, a 350 pm, 32 element array polarimeter with a ANA = 0.1 and FWHM resolution of 20”. For details of the observing procedure and data reduction see Chuss et al. (2003) and references therein.

3 Discussion The magnetic field structure inferred from all of the Hertz data obtained in this region to date are shown plotted over 850 pm continuum emission (Pierce-Price et al. 2000) in region I11 of Figure 1. Also shown are 60 pm polarimetry data (region I) and 100 p m polarimetry data (region 11) (Dotson et al. 2000). The field in the central 30 pc shows neither a predominantly poloidal nor toroidal direction. It does, however, appear to have structure on scales significantly larger than that of the molecular clouds (typically of the order of 5-10 pc). Also, the submillimeter data appear to trace magnetic fields that are spatially consistent with those traced by far-infrared polarimetry. Together, these observations indicate that there is some decoupling between the magnetic fields and the dense clouds. This decoupling is reinforced by

175

Astron. Nachr./AN 324, No. SI (2003)

16 14

12 10

8

n

6

*z ii

4

a Q

2

d

W

0

-2 -4

-6 10

5

0

-5

A a (arcmin) Fig. 1 The inferred magnetic field directions for polarization measurements in the Galactic center are displayed on 850 jim contours from SCUBNJCMT (Pierce Price et al. 2000). Region I shows 60 pm polarimetery of the Sickle (Dotson et al. 2000). Region TI shows 100 pin polarimetry of the Arched Filaments (Dotson et al. 2000). Region 111 shows new 350 pm inferred magnetic field vectors along with the 350 pm vectors from (Novak et al. 2000). Important dust features are shaded and labeled. The axes scales are offsets in arcminutes from the position of Sgr A* (02000 = 17”45”’40504,6~000 = -29”00’28!’07).

noting that in Figure 1, the magnitudes of the 350 p m polarization vectors tend to be lower in regions of high submillimeter flux. Though the field measured by Hertz is not local to the individual clouds, the dense clouds in the central

30 pc do interact with the field. One example of this interaction is the region of the Sickle (See Fig. 1) Here, the magnetic field direction lies in a toroidal orientation parallel to the long axes of the molecular cloud. This direction is perpendicular to that of the poloidal field implied by the Radio Arc (see Fig. 2). Such an orientation is indicative of a field that has been sheared by the cloud’s motion.

176

D. T. Chuss et al.: Mapping Magnetic Fields in the Cold Dust at the Galactic Center

15

10

5

0

-5 15

10

5

0

-5

A a (arcminutes) Fig. 2 Inferred B vectors are superposed on a 20 cm map (Yusef-Zadeh,Moms, & Chance 1984) taken with the VLA. Important thermal and non-thermal structures are labeled. The 100 pm vectors appear to trace the Arched Filaments. Note that the field in the molecular cloud associated with G0.18-0.04 is perpendicular to the field associated with the Radio Arc.

Similar shearing is observed in M-0.13-0.08 (Fig. I). It has been noted by Novak et al. (2000) that the flaring of the field near the southern edge of this cloud indicates that the field in this region initially had a different configuration but was sheared by the inward motion of M-0.13-0.08 into a direction parallel to the cloud’s trajectory. Along these same lines, the effect of the motion M+0.07-0.08 on the magnetic field lines is noticed at the southwestern edge of the cloud. The magnetic field lines are observed to trace the cloud edge indicating that the cloud is moving towards the region of low flux and poloidal field to the northeast of M-0.02-0.07. In the process of moving through a low density region threaded by a poloidal field, this cloud is “sweeping up” poloidal magnetic flux and shearing it into a toroidal configuration.

Astron. Nachr./AN 324. No. S 1 (2003)

177

The toroidal magnetic field structures in dense molecular regions OF the Galactic center indicate that the gravitation energy density in these regions is strong enough to overcome the magnetic energy density such that the fields (which are frozen into the matter) are slretched out by the motion of the clouds. This relationship is quantified in Figure 3. Here. the absolute deviation from a poloidal field is plotted againsl the 350 pm SHARC flux, and we find that for high fluxes, the field is in general toroidal, while for low fluxes, the field is poloidal. 100

80

60

40

20

C

-2c

50

100

150

200

250

300

F (Jy/SHARC beam) Fig. 3 The absolute value of the deviation of each ineasurement from a poloidal field is plotted against 350 p m flux in a 15" SHARC beam. Here the angles are used from all of the GC polarization measurements in Figure I. The assumption is that the polarization angle will not change significantly from 60 pm to 350 pm.

This result can be interpreted in a way that reflects the model of Uchida et al. (1985) referenced in the introduction. We start with the assumption that the magnetic field in thc Galactic center was initially poloidal. In overdense regions, gravitational energy density can be strong enough to shear the field into a configuration that is toroidal (along the direction of the motions of the clouds). In underdense regions, the magnetic field energy density is high enough to maintain its poloidal structure againsL gravitational forces. This argument implies the existence of a critical density at which the magnetic energy density and the kinetic energy densily are equal. In this case, the following holds.

We assume this equilibrium occurs at a flux corresponding to a field angle of 45" with respect to the poloidal direction. Fitting a line to Figure 3 allows us to calculate the flux density associated with this

D. T. Chuss et al.: Mapping Magnetic Fields in the Cold Dust at the Galactic Center

178

Collapse along Field Lines

Molecular Material

I

I

I

Relativistic Electrons

Fig. 4 Molecular clouds can produce relativistic electrons necessary for the illumination of NTFs by the following process. In regions of low density, the molecular material is dominated by the magnetic field, and we observe a poloidal field (A). MHD allows for movement of material along the lines of flux. In this way, the material can form clouds and gravity can begin to compete with the magnetic field energy density (B). At this stage, velocities of the molecular material with respect to the poloidal field can distort the field. This process continues (C) as the poloidal fields become sheared into toroidal ones in the vicinity of the cloud. Finally, oppositely-oriented magnetic fields near the cloud centers will be forced into contact by gravity and will reconnect, thereby releasing energy that energizes relativistic electrons. These electrons spiral along the external field and produce synchrotron radiation that we observe as an NTF.

angle to be 125 Jy/SHARC beam. This flux can be converted into a mass density assuming a line-of-sight dimension and a typical cloud velocity. This leads to a characteristic magnetic field strength estimate.

4

Conclusions

This estimate of the magnetic field strength is in agreement with that obtained by Yusef-Zadeh and Morris (1987) for the field strength i n the Radio Arc. O u r data seem to suggest that m G fields penetrate much of

Astron. Nachr./AN 324, No. S1 (2003)

179

the inner 30 pc of the Galaxy, but that conditions there are such that in regions of high density, the fields are sheared by the gravitational motion of clouds. The idea that the field measured in the cold dust may be part of the same global field that is seen in the NTFs suggests a possible source of the relativistic electrons required to “light up” the filaments. Magnetic reconnection has been posited as the source of the energy for the relativistic electrons (Serabyn & Gusten 1991); however, it is most often invoked as a cloud-flux tube collision in which the magnetic fields of a flux tube become bent and magnetically reconnect. Figure 4 depicts an alternative scenario derived from the idea that poloidal and toroidal fields are part of a global magnetic field. Here, we have a situation in which some underdense molecular material is threaded by a poloidal field. In this case, the field is strong enough to maintain its structure against the relatively weak influence of gravity. However, the material is free to collapse along the field lines. If this occurs, the gravitational energy density can begin to overcome the magnetic pressure and the field will begin to be sheared into a toroidal configuration (parallel to the long axis of the cloud.) As this shearing continues, oppositely-oriented toroidal magnetic fields are brought into contact with one another. At this point, magnetic reconnection can occur releasing energy that can produce the relativistic electrons necessary for NTF formation. Such a scenario fits observations of the Sickle (G0.18-0.04) region well. It can be seen that the toroidal magnetic field vectors are coincident with the molecular cloud shown in Figure I . Also, the poloidal filaments of the Radio Arc extend outward from the Sickle in both directions (see Fig. 2) showing a similar relationship between a set of NTFs and a molecular cloud as exhibited in Figure 4D.

References Chuss, D. T.. Novak, G., Hildebrand, R. H., Dowell, C. D., Vaillancourt, J. E., Davidson, J. A., & Dotson, J. L. 2003, ApJ, accepted Dotson, J. D., Davidson, J., Dowell, C. D., Schleuning, D. A., & Hildebrand, R. H. 2000, ApJS, 128, 335 Dowell, C. D., Hildehrand, R. H., Schleuning, D. A., Vaillancourt, J. E., Dotson, J. L., Novak, G., Renbarger, T., & Houde, M. 1998, ApJ, 504,588 Dowell, C. D., Lis, D. C., Serahyn, E., Gardner, M.. Kovacs, A,, & Yamashita, S. 1999, in ASP Conference Series 186: The Central Parsecs of the Galaxy, ed. H. Falcke, A. Cotera, W. Duschl, F. Melia, & M. Rieke, 453465 Killeen, N. E. B., Lo, K. Y., & Crutcher, R. 1992, ApJ, 385,585 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, ApJ, 119, 207 Novak, G., Chuss, D., Renbarger, T., Griffin, G. S., Newcomb, M. G., Peterson, J. B., Loewenstein, R. F., Pernic, D., & Dotson, J. L. 2002, in press Novak, G., Dotson, J. L., Dowell, C. D., Hildebrand, R. H., Renbarger, T., & Schleuning, D. A. 2000, ApJ, 529, 24 1 Pierce-Price, D., Richer, J. S., Greaves, J. S., Holland, W. S., Jenness, T., Lasenby, A. N., White, G. J., Matthews, H. E., Ward-Thompson,D., Dent, W. R. F., Zykla, R., Mezger, P., Hasegawa, T., Oka, T., Omont, A,, & Gilmore, G. 2000, ApJ, 545, L121 Plante, R. L.,Lo, K. Y., & Crutcher, R. M. 1995, AaJ, 445. L113 Schleuning, D. A. 1998, ApJ, 493,81 I Schleuning, D. A., Dowell, C. D., Hildebrand, R. H., & Platt, S. R. 1997, PASP, 109, 307 Serabyn, E., & Gusten, R. 1991, A & A, 242, 376 Uchidd, Y., Shibata, K., & Sofue, Y. 1985, Nature, 317, 699 Yusef-Zadeh, F., Moms, M., &Chance, D. 1984, Nature, 310, 557 Yusef-Zadeh, F. & Moms, M. 1987, AJ, 94, 1178

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S 1, 18 1 - 187 (2003) / DO1 10.1002/asna.200385066

The Galactic Center Nonthermal Fi1aments:Recent Observations and Theory T. N. LaRosa”, Michael E. Nord*.j, T. Joseph W. Lazio’, Steven, N. Shore4, and Namir E. Kassim3

’ Department of Biological & Physical Sciences, Kennesaw State Univ, 1000 Chastain Rd., Kennesaw, GA 30144 USA

* Department of Astronomy, Univ. of New Mexico

’ Code 7213, Naval Research Laboratory, Washington, DC 20375-5351

‘ Dipartimento di Fisica “Enrico Fermi”, Universita di Pisa, 56100, Pisa, Italy Key words Non-Thermal Filaments

Abstract. The large-scale topology and strength of the Galactic Center magnetic field have been inferred from radio imaging of the nonthermal filaments (NTFs). These objects, which seem to be unique to the Galactic center, are defined by extreme aspect ratios and a high degree of polarization. Recent high resolution, wide-field VLA imaging of the GC at 90 cm has revealed new candidate NTFs with a wide rangc of orientations relative to the Galactic plane. We present follow up 6 cm polarization observations of 6 of these candidates and confirm 4 as new NTFs. Together the new 90 and 6 cm results complicate the previous picture of largely perpendicular filaments that trace a globally ordered magnetic field. NTF observations in general do not rule out any particular models for the origin of the NTFs. Hence we explore the idea that the NTFs are local, individual structures: magnetic wakes generated through the interaction of molecular clouds with a Galactic Center wind. Numerical simulations of the evolution of a magnetized wake will be discussed and compared with NTF observations.

1 Introduction Beginning with the discovery of the bundled nonthermal filaments in the Galactic Center Radio Arc (GCRA) by Yusef-Zadeh, Morris and Chance (1984) it was recognized that the Galactic Center (GC) is the site of some unusual magnetic phenomena. The identification of isolated nonthermal filaments (NTFs), at various locations within the central few hundred pc (Yusef-Zadeh & Morris 1985; Liszt, 1985; Balley & Yusef-Zadeh 1989; Gray et al 1991; Lang et al 1999; LaRosa, Lazio & Kassim 2001; Reich 2003) shows that magnetic phenomena are widespread within the GC. The NTFs are defined by extreme aspect ratios (10-100) and high intrinsic polarization (e.g., Morris & Serabyn 1996). Although they are unique to the GC, their origin and relationship to other phenomena there continues to excite considerable speculation (e.g., Heyvarts, Pudritz and Norman 1988; Uchida, Shibata & Sofue 1985; Chevalier 1992; Benford 1988, Rosner & Bodo 1992; Serabyn & Morris 1994, Shorc & LaRosa 1999; Chandran 2001; Bickncll & Li 200 1a). In this paper we present new 6 cm polarization observations of candidate NTFs revealed by recent advances in low frequency wide-field, VLA imaging of the G C (Nard et al2003a,b; LaRosa et al2000). We then place the new results in context by providing a general overview of the observed NTF characteristics. We do not find that the observations can rule out any particular NTF model. Finally, we describe our * Corresponding author: e-mail: [email protected], Phone: + I 770423 6038, Fax: t I 770423 6625

020113 WILEY-VCH Verlag GnihH & Co. KGaA. Weinheir

182

T. LaRosa et al.: Nonthemal Filaments

group’s efforts exploring the possibility that the NTFs are dynamic structures created by the interactions between molecular clouds and a GC wind.

2 Observations Recently improved wide-field, high resolution, VLA imaging of the GC at 90 cm (Nord et al2003a,b) has revealed a number of new NTF candidates. Their Figure 1 covering a few 100 pc, was constructed from a combination of VLA A & B array data giving a high resolution of 12”x 7”. Although it is not sensitive to large-scale low surface brightness features, this image shows all previously known NTFs except for G359.44+0.39 which was significantly resolved. The previously known NTFs are mainly perpendicular to the Galactic plane. However, the orientations of the candidate NTFs are more diverse, with several nearly parallel to the plane. To verify their status, we have made follow up VLA observations of several of these sources in 6 cm polarization in Oct of 2002 in CnB configuration in dual polarization mode with significantly better resolution (- 4”x 3”) than at 90 cm (for details see LaRosa et a1 2003). We concentrated on several candidates in the Sgr C region. Figure 1 shows the 90 cm AB subimage of Sgr C.

G359.22-0.16 is located about 20 pc in projection south of the Sgr C H I1 region. This source was detected at 18 cm (Liszt & Spiker 1995) and appears 5 pc long. Its surface brightness is not uniform: the peak occurs at the southern terminus where it’s slightly wider. The 90 cm morphology is similar. -0.3 ( S P).The Using the peak flux at 18 and 90 cm gives an estimate for the spectral index N 6 cm polarized intensity image of (3359.22-0.16 is shown in figure 2; it is 40% polarized, consistent with previous measurements of other NTFs. It is 7 pc long and roughly 0.5 pc wide. Based on this morphology, its polarization, and its nonthermal spectral index we confidently classify this source as an NTF. This is an important result since this is only the second confirmed NTF that is parallel to the Galactic plane. Furthermore, since the end of this source is less than 10 pc in projection from the Sgr C filament, if both are at the same distance they cannot be tracing a simple globally ordered field. There is considerable evidence that the magnetic field in the neutral medium along the Galactic plane in the Sgr A and Sgr B regions is parallel to the plane (e.g., Novak et al 2003a,b). (3359.22-0.16 could be related to this toroidal field, but SCUBA images (Pierce-Price et a1 2000) along the Galactic plane indicate that the thermal emission from the Sgr C environment is much less than in the Sgr B or Sgr A regions so Sgr C may not be dominated by the neutral medium. N

N

N

G359.43-0.13 lies northwest of the Sgr C filament. At 90 cm it has a distinctive X-shape. At 6 cm the X resolves into two, or perhaps, three filaments with significant curvature (see Figure 3). Unfortunately, our low signal to noise ratio prevented a detection of this system in polarization. Thus we hesitate to classify this system as an NTF although morphologically similar to other NTFs. Deeper observations of this system are warranted. G359.40-0.07, seen 5 pc south in projection from the Sgr C filament is the brightest NTF candidate at 90 cm. This source was also detected at 18 cm (Lizst & Spiker 1995) and we estimate a 18/90 cm -0.3. The faint extension of this source in the 90 cm image (G359.40-0.03) was spectral index of a not detectable at 6 cm over our integration time due to the intensity of the background emission along the Galactic plane. N

We also observed the three linear features in the Sgr B region, G0.39-0.12, G0.37-0.07, and G0.39+0.05 (figure 4) Only one, 60.39-0.12, was detected in 6 cm polarization at the 10% level. Their morphologies and similar orientation to the filaments in the GCRA strongly suggests they are NTFs. If so, they are the first N7Fs to be found north of the GCRA and extend the volume over which the NTF phenomenon occurs to over 300 pc along the Galactic plane.

1x3

Astron. Nachr./AN 324. No. S 1 (2003)

Fig. 1 90 crn subirn of the Sgr C region. Candidate NTFs are labelled

3 Review of NTF Characteristics We now consider these observations together with previous work to formulate a compendium of observed characteristics. The population: The region over which the phenomenon occurs is 300 pc x 75 pc on the sky. There are now 14 confirmed isolated NTFs and another dozen or so filaments in the GCRA. Two of the isolated NTFs are parallel to the Galactic plane. There are another 14 candidate NTFs. These are considerably shorter and have lower surface brightness. Several of these are parallel to the plane and a few are quite close in projection to other NTFs. It seems unlikely these are tracing the same globally ordered magnetic field. Magnetic fields aligned longitudinally: Rotation measure studies show that the NTF magnetic fields are aligned along the long axis of the filaments (Tsuboi et a1 1986; Yusef-Zadeh, Parastaran & Wardle 1997; Lang, Morris & Echevarria 1999). Estimates for the strength of the magnetic field from equipartition yield 100 pG (e.g, Gray et a1 1995)The rigidity of the filaments against the ram pressure of the surrounding highly turbulent molecular clouds suggest magnetic field strengths of 1 mG (Yusef-Zadeh & Morris 1987a,b).

-

-

Subfilamentation and braiding (e.g., Liszt & Spiker 1995; Yusef-Zadeh, Wardle & Parastaran 1997): High resolution images of isolated NTFs invariably exhibit subfilamentation, and flaring at the ends. Moreover, the subfilaments cross and appear to be braided around each other. If the NTFs were tracing a global field a reasonable assumption is that this field will have relaxed to a potential or force free state but braiding suggests that the field can be tangled and that magnetic reconnection is taking place. The stability of such structures within a global field remains an open question. Peak intensity located near geometric center (LaRosa et a1 2000): This property may be related to the braiding. Where two filaments overlap the surface brightness must increase.

184

T. LaRosa et al.: Nonthermal Filaments

Fig. 2 6 cm total polarkation of G359.22-0.16. The rcsolution is 3.75” x 3’’

5 . Nonthermal Spectra with curvature above 5 GHz: The isolated NTFs have 20/90 cm spectral indicies -0.45 a 2 -0.6 (e.g., LaRosa et al. 2000). However, they are very weak at 6 and 2 cm, showing a clear spectral turnover at these higher frequencies (Lang, Morris & Echevarria 1999). This is not the case for the bundled filaments: the NTFs in the GCRA have been observed up to 43 GHz and their spectra are flat, suggesting ongoing particle acceleration or extreme youth (Reich, Sofue & Matsuo 2000).

6. Association with molecular clouds: Nearly a11 the well studied NTFs appear to be associated with, and could be interacting with, molecular clouds (e.g., Serabyn & Gusten 199 1 Staguhn et a1 1998). One explanation for the origin of the high energy electrons illuminating the NTFs is particle acceleration driven by magnetic reconnection occuring at the interface between a global magnetic fieid and some local cloud magnetic field. Examples of such interactions are inferred from observations of molecular clouds at several locations along the GCRA (Serabyn & Morris 1994; Tsuboi, Ukitd & Handa 1997). While these data are interesting and compelling, there is no obvious brightness variation in the NTFs at the purported interaction sites. 7. Lengthwise variations of spectral index: If electrons are injected at one end of a filament, then given the synchrotron lifetime (determined by magnetic field strength) and the lengths of the NTFs one might expect to see some spectral variation. Lang, Morris & Echevarria (1999) found no significant variation in the spectral index as a function of length in both the northern and southern threads, and LaRosa et a1 (2000) studied the Sgr C filament and found a constant spectral index with position. For a 1 mG magnetic field the synchrotron lifetime of electron emitting at 20 cm is N 3 x lo4 years. If the electrons can stream at the local Alfven velocity, they can travel roughly 65 pc in this time. Since the NTFs are shorter than 65 pc, LY variation may not be expected but in the longest NTF, the Snake, the spectral index does vary (Gray et al 1995). At the locations of the kinks the spectrum is flatter, suggesting that acceleration is taking place there (see Bicknel & Li 20014. For G359.43+0.2

Astron. Nachr./AN 324, No. S 1 (2003)

-29

174354

52

50

48

46

44

42

40

38

36

RIGHT ASCENSION (J2000) Grey scale flux range= -0.228 1.221 MilliJYlBEAM

Fig. 3 6 cm total intensity image of G359.43f0.13. The resolution is 3.86" x 2.8"

Fig. 4 90 cni subim of the Sgr B region. Candidate NTFs are labelled

LaRosa, Lazio & Kassim (2001) found that the 20/90 cm spectral index decreased un~formlywith the distance away from the Galactic plane. The most natural explanation is that the electrons are radiating in a weakening magnetic field. If so, thc length scale for the variation is only a few pc suggesting this system may be locul field.

T. LaRosa et al.: Nonthermal Filaments

186

These characteristics do not, however, point to any dejinitive model for the NTFs and fundamental questions remain open. Are the NTFs tracing a global or local field? Are these structures static or dynamic? Why are the NTFs observed only at the Galactic center? Are the same physical mechanisms that generated the isolated NTFs also responsible for the filaments in the GCRA? What acceleration mechanism generates the high energy electrons? Is the acceleration local or is it distributed along the length of the NTF? Although a number of ideas have been proposed (for recent reviews see Morris 1998; Bicknel & Li 2001b) we confine our discussion to the cometary model advanced by Shore & LaRosa (1999).

4 Theory: The Comet Model The NTF phenomenon must be related to special conditions that arise at the Galactic center and not in the general interstellar medium. The Galactic center has the highest concentration of young massive stars in the Galaxy and there is growing evidence for bursts of star formation there (e.g., Simpson et a1 1999). It is therefore reasonable that the combined stellar winds and supernova explosions will collectively generate a large-scale Galactic wind. Continuum X-ray observations have detected hot thermal gas and spectral line measurements indicate that it is moving at velocities of a few thousand km s-l (Koyama et a1 1996). Recent evidence for a nuclear starburst and associated outflows is summarized in Bland-Hawthorn & Cohen (2003). If this wind advects a weak magnetic field and encounters an obstacle, such as a molecular cloud, the resulting cloud-wind interaction will generate a magnetic wake (Shore & LaRosa 1999). The magnetic field diffusion time through a cloud is orders of magnitude longer than the fow timescale so the flow wraps the magnetic field around the cloud, forming a long thin wake. Detailed 3-D numerical simulations of a molecular cloud-wind interaction (Gergori et al 2000) confirm this picture (see their figure 9) and show the stretching and amplification of the magnetic field. In this cometary model the NTFs are observed in projection and can therefore have any orientation with respect to the Galactic plane. The advantage of this model is that since the NTFs represent local amplification of a weak field, the total magnetic energy density in the GC is greatly reduced compared to a strong pervasive field. However, the viability of this model depends on the stability of the wakes that are produced. Dahlburg et a1 (2002) have numerically simulated the evolution of a 2-dimensional wake for Galactic center conditions. The initial conditions were an exterior wind speed of 2000 km spl, an external density of I w p 3 ,an external field of mG, and an interior field of 1 mG. The code, which follows a piece of plasma in time, was initialized with small perturbations to insure the unstable modes were naturally excited using a linear code to determine the fastest growing wavelength (essentially a nonlinear magnetic KelvinHelmholtz instability) with which to normalize the length scales for a full nonlinear calculation. The results are shown in their figure 7 where the quantity pD2 (which is proportional to the synchrotron emissivity ) is plotted as a function of time (note that since the plasma is flowing longer times translate into increasing length along the filament). It is clear the field begins to shred, and the wake expands until it ultimately joins with the surrounding flow. However, the shredding begins after the the plasma travels approximately 50 times it own width so it is possible to obtain structures with aspect ratios similar to the NTFs. Furthermore, the peak brightness of the filament should occur where instability begins to grow, downstream from the interaction site. This may explain why the peak brightness of the NTFs occur near the midpoint rather than at one end. The instability may also be important for driving magnetic reconnection and creating the conditions for significant particle acceleration, which we are now studying.

5 Conclusions Based on morphology and polarization there are now 14confirmed NTFs, in addition to those in the GCRA. Two of the isolated NTFs are parallel to the Galactic plane. There are also several NTF candidates that show curvature and are also parallel to the Galactic plane. The candidate NTFs are considerably shorter and less intense than those previously identified. Several are quite close, in projection, to larger NTFs and

Astron. Nachr./AN 324, No. SI (2003)

187

would be difficult to explain if the NTFs are tracing a globally ordered field. At present observations d o not strongty restrict the theoretical possihilites. Continuing observational work combined with more delailed theoretical modeling will be required to elucidate the nature of these unusual objects.

Acknowledgements We thank Russ Dahlburg and Giorgio Einaudi for their willing and capable collaboration. Basic research in radio astronomy at the Naval Research Laboratory is supported by the Office of Naval Research.

References Bally, J. & Yusef-Zadeh, F. 1989, ApJ, 336, I73 Benford, G. 1998, ApJ, 333, 73 Bicknell, G., & Li, J. 2001a. ApJ, 548, L69 Bicknell, G. & Li, J. 2001 h, PASA, 18, issue 4. 43 I Bland-Hawthron. J. & Cohen, M. 2003, ApJ, 582, 246 Chandran, B. D. G. 2001, ApJ, 562,737 Chevalier, R. A. 1992, ApJ, 397, L39 Dahlburg, R., Einuadi, G., LaRosa, T.N. & Shore, S.N. 2002, ApJ, 568, 220 Gregori, G.. Miniati, F., Ryu, D. &Jones, T. J. 2000, ApJ, 543, 775 Gray, A. D., Cram, L. E., Eckers,R. D. & Goss, W. M. 1991, Nature, 353, 237 Gray, A. D., Nicholls, J., Ekers, R.D. & Cram, L. E. 1995, ApJ, 448, 164 Heyvaerts, J., Norman, C. & Pudritz, R. E. 1988, ApJ, 330, 718 Koyama, K., Maeda, Y., Sonobe, T. Takeshima, T. Tanaka, Y., & Yamauchi, S. 1996, PASJ, 48, 249 Lang, C. C., Moms, M. & Echevarria L. 1990, ApJ, 526, 727 Lang, C. C., Anantharamaiah, K. R., Kassim, N. E., Lazio, T. J. W., & Goss, W. M. 1999, ApJ, 521, L41 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W. & Hyman, S. D. 2000, AJ, 119, 207 LaRosa, T.N., Nord, M. E., Lazio, T. J. W. & Kassim, N. E. 2003, in preparation LaRosa. T. N., Lazio, T. J. W. & Kassim, N. E. 2001. ApJ, 563, 163 Liszt, H. 1985, ApJ, 298,128 I Liszt, H. & Spiker, R. 1995, ApJS, 98,259 Moms, M. 1998, in The Central Regions of the Galaxy and Galaxies, IAU Symposium 184, ed. Y. Sofue (Dordrecht: Kluwer Academic Publishers), p. 33 I Morris, M. & Serabyn. E. 1996, ARA& A, 34,645 Nord, M. E., Lazio, T. J. W., Kassim, N. E., Hyman, S. D., LaRosa, T. N. & Duric, N. 2003a, AJ submitted Nord, M. E., Brogan, C . , Hyman, S. D., Lazio, T. J. W., Kassim, N. E., LaRosa, T. N., Anantharamaiah, K. & Duric, N. 2003b, these proceeedings Novak, G., et al. 2 0 0 3 , ApJ, 583, L83 Novak, G., et al. 2003b, these proceedings Pierce-Price, D., et al. 2000, ApJ, 545, L121 Reich, W. 2003, A& A in press Reich,W., Sofue, Y. & Matsuo, H. 2000, PASJ. 52, 355. Rosner, R. & Bodo, G. 1996, ApJ, 470, L49 Serabyn, E. & Moms, M. 1994, ApJ. 424, L91 Serabyn, E. & Gusten, R. 1991, A&A, 242, 376 Shore, S. N. & LaRosa, T. N. 1999, ApJ, 521.587 Simpson, J. P.,Whitteborn, F. C., Cohen, M. & Price, S. D. 1999 in ASP Conf. Ser. 186, The Central Parsecs of the Galaxy, eds. H. Falcke, A. Cotera, W.J. Duschl, F. Melia, & M.J. Rieke (San Francisco: ASP) p. 527 Staguhn, J., Stutski, J., Uchida, K.I., & Yusef-Zadeh, F. 1998, A& A, 336. 290 Tsuboi, M., Inoue. M., Handa, T., Tabara, H.. Kato, T., Sofue, Y. & Kaifu, N. 1986, AJ, 92, 818 Tsuboi, M., Ukita, N. & Handa, T. 1997. ApJ, 481, 263 Uchida, Y., Shibata, K. & Sofue, Y. 1985, Nature, 3 17, 699 Yusef-Zadeh, F., Moms, M., & Chance, D. 1984, Nature, 310, 557 YLisef-Zadeh, F. & Morris, M. 1985, AJ, 90,251 I Yusef-Zadeh, F. &Morris, M. 1987a, ApJ, 322, 721 Yusef-Zadeh, F. & Moms, M. 1987h, AJ, 94, 1 128 Yusef-Zadeh, F., Wardle. M. & Parastaran, P. 1997, ApJ, 475, L119

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1, 189- 195 (2003) / DO1 10.1002/asna.200385068

Interaction between the Northeastern Boundary of Sgr A East and Giant Molecular Clouds: Excitation Mechanisms of the H2 Emission Sungho Soojong Pak2, Christopher J. Davis3, Robeson M. Herrnstein4, T. R. Geballe’, Paul T. P. Ho4, and J. Craig Wheeler6

’ Astronomy Program in SEES, Seoul National University, Shillim-Dong, Kwanak-Gu, Seoul 151-742, South Korea ’ Korea Astronomy Observatory, Whaam-Dong, Yusong-Gu, Daejeon 305-348, South Korea Joint Astronomy Centre, University Park, 660 North A’ohoku Place, Hilo, HI 96720, USA Harvard-SmithsonianCenter for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA Gemini Observatory, 670 N. A’ohoku Place, Hilo, HI 96720, USA Astronomy Department, University of Texas, Austin, TX 78712, USA

Key words Galaxy: centre - ISM: individual(Sgr A East), molecules -infrared: ISM: lines and bands Abstract. We have detected the v = 1 i 0 S(1) (A = 2.1218 pm) and v = 2 + 1 S(1) (A = 2.2477 pm) lines of HZin a region between the northeastern boundary of Sgr A East and the giant molecular cloud (GMC) M-0.02-0.07, in the Galactic centre. The broad line widths, the low peak velocities relative to the molecular clouds, and the moderate Hz v = 2 --t 1 S(1) to v = 1 + 0 S(1) line ratios can be best explained by a combination of C-type shocks and fluorescence. The detection of shocked H2 in this region is clear evidence that Sgr A East is driving material into both the GMC M-0.02-0.07 and the northern ridge found by McGary, Coil, & Ho (2001).

1 Introduction Sgr A East has frequently been interpreted as a supernova remnant due to its shell structure and non-thermal spectrum (Jones 1974; Goss et al. 1983 and references therein; and see the more recent references in Maeda et al. 2002). Some recent research, however, has suggested that the energetics, size, and elongated morphology (3’ x 4’ o r 7 pc x 9 pc at d = 8.5 kpc) of Sgr A East cannot have been produced by a typical supernova (Yusef-Zadeh & Morris 1987; Mezger et al. 1989). The origin of Sgr A East still seems to be an open question and the required energy to produce it is a key parameter in this issue (Mezger et al. 1989). In principle, the energy of the explosive event can be directly measured by studying regions where Sgr A East is colliding with ambient interstellar material. By tracing the dynamics of molecular gas, an interaction between the eastern part of Sgr A East and the giant molecular cloud (GMC) M-0.02-0.07 (also known as the ‘50 k m sP1cloud’) has been inferred (Genzel et al. 1990; Ho et al. 1991; Serabyn, Lacy, & Achtermann 1992; Mezger, Duschl, & Zylka 1996; Novak 1999; Coil & Ho 2000). Recent observations of NH3(3,3) emission in the region show that Sgr A East impacts material to the north and west as well (see Fig. I , also McGary, Coil, & Ho 2001). As direct evidence of this interaction, several 1720 MHz OH masers, which are a good diagnostic of the continuous, or C-type, shock excitation (Frail et al. 1996; Wardle, Yusef-Zadeh, & Geballe 1999), have been detected along the southern edge of Sgr A East and to [he north of the circumnuclear disk (CND) (Yusef-Zadeh et al. 1996). * Correspondin&author: e-mail:

1eeshOastro.snu.ac.kr.Phone: +8242 865 3248, Fax: +8242 861 S610 @ 2003 WILEY~VCHVcrkig GmhH & Co KO&.. Weinhelm

190

S. Lee et al.: Interaction between Sgr A East and GMCs

i

c

,%I

~J2(i'JOI

Fig. 1 Central 10 x 15 pc region of the Galaxy. Contours representing the velocity-integrated map of NH3(3,3) emission are overlaid on a 6 crn continuum image of Sgr A complex from McGary et al. (2001). The black dot at the centre of image is Sgr A* and the mini spiral is Sgr A West. The CND is traced by the brighter part of continuum surrounding them and Sgr A East is seen by the outer part extended to the boundary of NH3 contours. The GMC M0.02-0.07 lies to the east of this region. The dashed box at the northeastern edge of Sgr A East encloses the 90" x 27" region observed in Hz. The solid line in the box is Slit 9 from which the Hz spectra shown here were extracted. Letters mark the positions of OH( 1720 MHz) masers with error ellipses scaled up by a factor of 15 (Yusef-Zadeh et al. 1999a).

Wardle et al. (1999) and Yusef-Zadeh et al. (1999b, 2001) detected H2 line emission in regions where OH-masers have been detected and they interpreted the emission as thermal. It is therefore likely that Sgr A East is indeed driving shocks into the adjacent GMCs to the south and into the CND. However, the fields observed by Wardle et al. (1999) and Yusef-Zadeh et al. (1999b, 2001) are restricted to the vicinity of the CND and cover only some of the regions where interaction of the Sgr A East shell with surrounding material is expected. Before one can hope to estimate the energy released in the event that created Sgr A East, it is necessary to observe additional interaction regions in diagnostic lines of Hz.

2 Observations We observed the Hz v = 1 --t 0 S(1) (A = 2.1218p.m) and the Hz v = 2 1S(1) (A = 2.2477pm) lines at the 3.8 m United Kingdom Infra-Red Telescope (UKIRT) in Hawaii on 2001 August 3 and 4 (UT), using ---f

191

Astron. Nachr./AN 324, No. S I (2003)

2 1.5 1

0.5 0 2 1.5 1

0.5 0 2 1.5 1

0.5 0 2 1.5 1

0.5 0 2 1.5 1

0.5 0 2 1.5 1

0.5 0 -200

0

-

200

V,

(km s-l)

Fig. 2 €12 v = 1 + 0 S ( l ) and Ha v = 2 1 S ( l ) spectra from six positions along Slit 9. Indicated positions are relative to (Y = 17”45”45?3, 6 = -28d58“’58”; 52000. Left panels show the H P v = 1 3 0 S(1) spectra. The right three panels present both the HZ v = 1 0 S(1) and Hz v = 2 + 1 S(1) spectra from the positions where H:! v = 2 4 1 S(1) emission is detected; these are averaged over 3.4” on the sky to improve the S/N ratios. The dotted lines are Gaussian fits to the observed line profiles. The spectra are not corrected for instrumental broadening. -i

the Cooled Grating Spectrometer 4 (CGS4; Mountain et al. 1990) with a 31 I/mm echelle grating, 300 mm focal length camera optics and a two-pixel-wide slit. The spatial resolution along the slit was 0.90” for HZ v = 1 --* 0 S(1) with the grating angle of 64.691 degree and 0.84” for H2 v = 2 -+ 1 S(1) with 62.127 degree, respectively; the slit widths on the sky were 0.83” and 0.89”, respectively, for these two 90”. Thc instrumental resolutions, measured from Gaussian fits to sky configurations. The slit length is lines in our raw data, were 17 krri s-’ for Hz v = 1 + 0 S( 1) and 19 kin s-’ for HZ v = 2 + 1 S( l ) , respectively.

-

N

N

192

S. Lee et al.: Interaction between Sgr A East and GMCs

-60

-40

-20

0

20

40

60

Position along Slit 9 (arcsec) Fig. 3 Derived line parameters for the spectra along Slit 9: (a) line center velocity; (b) line width; and (c) integrated line intensity. Indicated positions are as in Fig. 2; positive towards NE. Thc filled circles and the open circles represent the +50 km sC1 component and the 0 kin s-’ Component of Ha v = 1 --* 0 S(l) spectra, respectively, and the filled and open squares (the +50 km s-’ and 0 k m s-’ component) of the NH:3(3,3), from McGary et al. (2001). In the panel (c), the solid and dashed lines dcnote the total intensity of HZ v = 1 + 0 S ( l ) and NH:3(3,3), respectively, the latter scaled by The range of our Hz observation is denoted by two vertical lines. The decrease in NHs flux at positions greater than +30” is a result of reduced sensitivity at the edgc of the mosaic (McGary et al. 2001).

Ten parallel slit positions were observed, sampling a 90” x 27” area on the northeastern boundary of Sgr A East. The slit was oriented 40’ east of north for each measurement; adjacent slit positions were separated by 3” perpendicular to the slit axis. The coordinates at the centre of the observed area are (Y = 17h45m45?9, 6 = -28d59”’05” (52000) (see Fig. 1). Only the ninth slit position, hereafter called ‘Slit 9’, was observed in both Hz v = 1 + 0 S(1) and H2 v = 2 + 1 S(1), the line ratio of which we can use to constrain models for the excitation of Hz. Here we present the results of Slit 9, rather than the whole data set, as a preliminary report. We aim to concentrate on the excitation mechanism ofthe detected H:! emission.

Astron. Nachr./AN 324, No. S 1 (2003)

3 H2 v = l

+0

I93

S(l) emission

Bright (1.6 - 21 x 1U-I8 W n-’ arcsw-2) HZ emission was detected from most of the observed region along the northeastern boundary of Sgr A East. From the Hz v = 1 + 0 S ( l ) spectra in Fig. 2 we measure line centers and line widths along the interaction region. Each spectrum is well fitted by one or two Gaussian components. Fig. 3 show the distributions of the derived Hz v = 1 i 0 S(1) line parameters along Slit 9. For direct comparison, we include in these figures data from the NH3(3,3) observations of McGary et al. (2001); the NH3 emission essentially traces the cool ( 5 100 K), dense ( l o 5 cnir3)),cloud material. From these data we note the following.

-

In Fig. 3a there are two velocity components, at ~ . S R 0 krri srl and +50 kin s r l . Both components are evident in Hz and NH:%.The +50 kui srl component of the NH3 emission traccs the GMC M-0.02-0.07, while the 0 kin s-l component corresponds to the “northern ridge” of McGary et al. (200 I ). The variation in the velocity of either component is less than 20 krii s r l . N

-

-

The H2 line widths in Fig. 3b are generally much broader than the NH3 widths. The NH:r component at +20 kiii s-’ seen in Fig. 3a at negative offsets seems to trace hotter gas according to the NH3(2,2) to (1,l) line ratio map by McGary et al. (2001), which they suggest may be the result of an impact by Sgr A East. The distribution of the total intensity of Hz (the solid line in Fig. 3c) is similar to that of NH:3 (dashed line). It should be noted that NH3 is attenuated at positions greater than +30” by the edge of the primary beam in the VLA mosaic (McGary el al. 2001). There is a small discrepancy between NH3 and H2 between t 1 0 ” and 1-30” which may be explained either by exhaustion of the source of excitation (e.g. shock energy or UV photons) or by obscuration, at the inner, more dense, regions of the cloud. The brightest emission along Slit 9, between offsets -20” to -45” (the southwestern part), arises from the outer Hz clumps of Yusef-Zadeh et al. (2001).

4 H2 excitation The Hz v = 2 + 1 S(1) line was detected at thrcc locations along Slit 9, at positions NE 23.”7, SW lo.”& and SW 22.”2 relative to the centre of the slit ( o = 17”45”’45?3, 6 = -28d58’1158s; J2000). From these data we measured line ratios Hz v = 2 --* 1 S(1) / v = 1 + 0 S(1) of 0.40 0.12, 0.51 0.17, and 0.27 i 0.07, respectively (see Fig. 2). At other positions only the HZ v = 1 + 0 S ( l ) line was detected, with 3 a upper limits to the ratio of 0.5.0.6, and 0. I at offsets of NE 31.”8, NE 4.”8, and SW 44.”7 along Slit 9, respectively. Fluctrescent excitation in a low-density PDR (n(H2) < 5 x lo4 cin-”) should yield a ratio of‘ about 0.6. A lower ratio is expected in a more densc PDR environment (Black & van Dishoeck 1987), or in a shock. There are two basic types of shock: ’*jump” or J-type and “continuous” or C-type (see Draine & McKee 1993 for a review). A J-typc shock is formed in a highly ionized or weakly magnetized gas. Fluid parameters such as density and temperature undergo a discontinuous change (jump) at the shock front where the molecules may be dissociated. J-typc shocks (with vclocities greater than about 24 kin s - ‘ ) will completely dissociate the molecules (Kwan 1977); H2 emission occurs from a warm, recombination plateau in the post-shock rcgion. J-type shocks typically produce low line intensities cotnparcd to C-type shocks and HP v = 2 1 S(1) / 7; = 1 0 S(1) line ratios as large as 0.5 arc possible (Hollenbach & McKee 1989). At lower shock velocities, below the H2 dissociation speed limit, J-type shocks may yield much lower line ratios; < 0.3 (Smith 1995). In a C-type shock, where the magnetic field softens the shock front via ion-magnetosonic wave propagation such that the fluid parameters change continuously across the $hock front, the H2 dissociation speed limit is much higher (- 45 kin srl: depending on the density and magnetic field strength in the pre-shock gas). Smaller line ratios of about 0.2 are then predicted (Smith 1995; Kaufman & Neufeld 1996).

*

---f

-

*

194

S. Lee et al.: Interaction between Sgr A East and GMCs

From the observed ratios alone we are not able to unambiguously distinguish between excitation mechanisms. Our results can either be explained by fast J-type shocks or dense PDRs, or by a combination of fluorescence and either C-type shocks or slow J-type shocks, since the higher line ratios associated with fluorescence will be tempered by the low H2 v = 2 + 1 S(1) intensities associated with collisional excitation in shocks. To help distinguish between the HZ excitation mechanisms, we consider kinematic information and the spatial variation of the line ratio along Slit 9 . At most positions in Fig. 3b, the Hg line widths are high; typically 40 - 70 kni s-', but as high as 120 kin s-' in some positions. This suggests shock excitation and turbulent motions in the gas and tends to exclude the pure fluorescence models, where the H2 line emission generally arises from the stationary gas at the edges of neutral clouds illuminated by Far-UV photons from early-type stars. In shocked regions the line ratio is found to be constant over a wide range of Hg v = 1 + 0 S(1) intensities and spatial positions (Davis & Smith 1995; Richter, Graham, & Wright 1995), although this is not necessarily predicted from theory (Draine & McKee 1993). Conversely, in a PDR the ratio is sensitive to the incident F W flux and the molecular gas density; the H2 v = 1 + 0 S(1) intensity increases but the Hg v = 2 + 1 S(1) / v = 1 + 0 S(1) ratio decreases with increasing gas density or UV intensity (Usuda et al. 1996; Takami et al. 2000). Thus an unchanging H2 v = 2 i 1 S(1) / v = 1 + 0 S(1) ratio is found in shocks, while a varying ratio is expected in the pure fluorescent case. The measured line ratio and the H2 v = 1 + 0 S( 1) intensity in Fig. 3c, show evidence of an anti-correlation in our data, as expected in dense PDRs. Although the wide line profiles point to shock excitation, fluorescence appears to play a significant role at at least some locations. Considering the kinematics further, we note that J-type shocks produce narrow lines that peak at the velocity of the shock, while C-type shocks produce broader lines which peak at the velocity of the preshock gas and extend up to the shock velocity. Fig. 3a shows that there are two velocity components that are similar in Ha and NH3. The H2 emission traces hot (- 2000 K) gas and the NH3 cool (5 100 K) gas. Thus, if we assume that shocks are driven by Sgr A East into adjacent molecular clouds, whose velocities are given by thc NHS data (M-0.02-0.07 at +50 km s-l and the northern ridge at 0 k m s-'), then fast J-type shocks are inconsistent with our results, due to the low peak velocities of the HS lines relative to the molecular clouds. In summary, then, the wide line profiles and low peak velocities indicatc C-type shock excitation. However, the high values of the line ratio at some positions along Slit 9 and the spatial variation in that ratio, point to a fluorescent component to the excitation in some locations. A combination of C-type shocks and fluorescence (see e.g. Fernandes, Brand, & Burton 1997) is therefore the most reasonable explanation for the Hz excitation. Our conclusion on the C-type shocks is consistent with the detection of 1720 MHz OH masers to the north of the CND and to the south of Sgr A East (Yusef-Zadeh et al. 1996). Very recently Karlsson et al. (2003) detected the 1720 MHz OH masers at two positions near our target region. For the fluorescence, the source of the UV radiation could be either nearby early type stars or J-type shocks. However, as noted above, we see no evidence of J-type shocks in our data. Also, we cannot establish whether nearby stars are the source of the UV flux due to the lack of information on where or how many early type stars there are in the region. Acknowledgements We give special thanks to Young-Sam Yu and Tae-Hyun Kim for their help with the observation. Fig. 1 is reproduced from Fig. 10 of McGary et al. (2001) by permission of the AAS. The United Kingdom Infrared Telescope is operated by the Joint Astronomy Centre on behalf of the U.K. Particle Physics and Astronomy Council. This work was financially supported by the BK21 Project of the Korean Government. TRG's research is supported by the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., on behalf of the international Gemini partnership of Argentina, Australia, Brazil, Canada, Chile, the United Kingdom and the United States of America.

References Black, J. H., & van Dishoeck, E. F. 1987, ApJ, 322, 412

Astron. Nachr./AN 324. No. SI (2003)

195

Coil, A. L., & Ho, P. T. P. 2000, ApJ, 533, 245 Davis, C. J., &Smith, M. D. 1995, ApJ, 443, L41 Draine, B. T., & McKee, C. F. 1993, ARA&A, 3 I , 373 Femandes, A. J . L., Brand, P. W. J. L., & Burton, M. G. 1997, MNRAS, 290, 216 Frail, D. A., Goss, W. M., Reynoso, E. M., Green. A. J., & Otrupcek, R. 1996, AJ, I I I , 1651 210, 565 Genzel, R., Stacey, G. J., Harris, A. I., Geis, N., Graf, U. U., Poglitsch, A,, & Sutzki, J. 1990, ApJ, 356, 160 Goss, W. M., Schwarz, U. J., Ekers, R. D., & van Gorkom, J.H. 1983, in Proc. IAU Symp. 101, Supernova Remnants and Their X-Ray Emission, eds J. Danziger & P. Gorenstein (Reidel: Dordrecht), p. 65 Ho, P. T. P., Ho, L. C., Szczepanski, J. C., Jackson, J . M., Armstrong, J. T.,& Barrett, A. H. 1991, Nat, 350, 309 Hollenbach, D., & McKee, C. F. 1989, ApJ, 342. 306 Jones, T. W. 1974, A&A, 30,37 Karlsson, R., Sjouwerman, L. O., Sandqvist, Aa., & Whiteoak, J. B. 2003, A&A, 403, 101 1 Kaufman, M. J., & Neufeld, D. A . 1996. ApJ. 456, 61 1 Kwan, J. 1977, ApJ, 216, 713 Maeda, Y., Baganoff, F. K., Feigelson, E. D., Morris, M., Bautz, M. W., Brandt, W. N., Burrows, D. N., Doty, J. P., Garmire, G. P., Pravdo, S. H., Ricker, G. R., & Townsley, L. K. 2002, ApJ, 570,671 McGary, R. S., Coil, A. L., & Ho, P. T. P. 2001, ApJ, S59, 326 Mezger, P. G., Duschl, W. J., & Zylka, R. 1996, ARA&A, 7,289 Mezger, P. G., Zylka, R., Salter, C. J., Wink, J. E., Chini, R., Kreysa, E., & Tuffs, R. 1989, A&A. 209, 337 Mountain, C. M., Robertson, D. J., Lee, T. J., & Wade, R. 1990, in Proc. SPIE Vol. 1235, Instrumentation in Astronomy VII, ed. D. L. Crawford (Bellingham: SPIE), p. 25 Novak, G. 1999, in ASP Conf. Ser. Vol. 186. The Central Parsecs of the Galaxy, eds H. Falcke et al. (San Francisco: Astron. SOC. Pdc.), p. 488 1998, Richter, M. J., Graham, J. R., & Wright, G. S. 1995, ApJ, 454, 277 Serabyn, E., Lacy. J. H.. & Achtermann, J. M. 1992, ApJ. 395, 166 Smith, M. D. 1995, A&A, 296,789 Takami, M., Usuda, T., Sugai, H., Kawabata, H., Suto, H., & Tanaka, M. 2000, ApJ, 529, 268 Usuda, T., Sugai, H., Kawabata, H.,Inoue, M. Y., Kataza, H., & Tanaka, M. 1996, ApJ, 464, 818 Wardle, M., Yusef-Zadeh, F., & Geballe, T. R. 1999, in ASP Conf. Ser. Vol. 186, The Central Parsecs of the Galaxy, eds H. Falcke et al. (San Francisco: Astron. Sac. Pac.), p. 432 Yusef-Zadeh, F., & Moms, M. 1987, ApJ, 320, 545 Yusef-Zadeh, F., Roberts, D. A., Goss, W. M., Frail, D. A,, & Green, A. J. 1996, ApJ, 466, L25 Yusef-Zadeh, F., Roberts, D. A., Goss, W. M., Frail, D. A., & Grcen, A. J. 199Ya, ApJ, 512, 230 Yusef-Zadeh, F., Stolovy, S. R., Burton, M., Wardle, M., & Ashley, M. C. B. 2001, ApJ, 560, 749 Yusef-Zadeh, F., Stolovy, S. R., Burton, M., Wardle, M., Melia, F., Lazio, T. J. W., Kassim, N. E., & Roberts, D. A. 19Y9b, in ASP Conf. Ser. Vol. 186, The Central Parsecs of the Galaxy, eds H. Falcke et al. (San Francisco: Astron. Soc. Pac.), p. 197

This Page Intentionally Left Blank

Astron. NachrJAN 324. No. S1. 197 -203 (2003) / DO1 10.1002/asna.200385032

Sgr A East and its surroundings - a view with XMM-Newton Masaaki Sakano*’x3,Robert S. Warwick**’,and Anne Decourchelle****

’ Department of Physics and Astronomy, University of Leicester, Leicester LEI 7RH, UK ’ Japan Society for the Promotion of Science (JSPS)

* CEAIDSMIDAPNIA,Service d‘Astrophysique, C.E. Saclay, 91 191 Gif-sur-YvetteCedex, France

Key words The Galactic Centre, supernova remnants, Individual: Sgr A East, X-ray, Plasma emission

We present an X-ray study of the Sgr A East region based on recent XMM-Newton observations. The spectrum of Sgr A East can be represented by a two-component thin thermal plasma model with temperatures of 1 and 4 keV, both of which have reached ionization equilibrium state. The abundance of iron is found to be higher in the central region of the nebula, with Z = 3 4 solar, than in the outer area for which Z 0.5 solar. On the other hand, the abundances of other elements appear uniformly distributed with Z 1. We also detect a weak fluorescent Ka line from neutral iron in the outer region of source. We discuss the nature of Sgr A East on the hasis of these new X-ray results.

--

1 Introduction Sgr A East is a radio bright non-thermal structure surrounding the Galactic Centre. The oval shell-like structure in the radio continuum supports the view that Sgr A East is a supernova remnant (SNR), SNR GO.O+O.O (Jones 1974; Ekers et al. 1983). On the other hand, alternative interpretations have also been proposed, for example it could be the remnant of an explosion associated with the central massive black hole Sgr A*, which is located within Sgr A East. There is quite good evidence for a physical interaction between Sgr A West and East (e.g. Yusef-Zadeh et al. 2000), although many of the details, such as the origin of Sgr A East and its past evolution, remain open questions. Hard X-ray imaging observations of the Sgr A region were first made with ASCA (Koyama et al. 1996) and then with SAX (Sidoli & Mereghetti 1999); they traced an extended distribution of X-ray emission originating from a thin thermal plasma. More recently this region has been observed by CkandmlACIS, at a spatial resolution of 0.5 arcsec (Baganoff et al. 2003a,b; Maeda et al. 2002, 2003; Park et al. 2003). The bulk of the radiated energy of a SNR appears in the X-ray band, hence X-ray observations are essential for the study of this kind of source. In this paper, we report the XMM-Newfon results for Sgr A East and its surroundings as derived from both X-ray imaging and spectral data. XMM-Newton observations benefit from the large effective area, good imaging capability (- 5 arcsec) of the XMM mirrors, and good energy resolution of the EPIC CCD detectors. With this capability, the quality of the spectrum and statistics of the image are remarkable. The XMM-Newton observation of the Sgr A region was made on 2001 September 4 as a part of the XMMNewron Galactic Centre Survey. Further details of this observation and the data screening are reported in Sakano et al. (2003a), whereas a general description of the survey is given in Warwick (2002,2003) and Sakano et al. (2003~).In this paper, we assume a value of 8.0 kpc for the distance to the Galactic Centre (Reid 1993). masQstar.le.ac.uk, Phone: +44 116 252 35 10, Fax: +44 116252 33 1 1 * * e-mail: rswQstar.le.ac.uk.Phone: +441162523517,Fax: +44116252331I . * * * e-mail: [email protected],Phone: +33 1690843 84, Fax: +33 169086577. * Corresponding author: e-mail:

@ 2001 WILEY-VCH Verlag GmhH & Co KGdA. Weinhem

M. Sakano et al.: Sgr A East & its surroundings

198

2 X-ray images

-29 07 00

Radio Shell

k M M J174540-290.

J174540-29045

17:46:00

4550

45:40

17:45:30

17:46:00

4S:XJ

45:40

17:45:30

17:46:00

4S:SO

45:40

17:45:30

17:46:00

45:SO

45.40

/7:45:30

Fig. 1 (a) XMM/MOS1+2 image of Sgr A East in the 2-9 keV band. The dashed line indicates the position of the radio shell of S g r A East and the solid lines show the main source and background regions used in the spectral extraction. The two concentric circles represent suh-regions referenced in Fig. 4a and b. The small circular region enclosed by the thick dotted line, which encompasses Sgr A*, is excluded in the spectral accumulation. (b) The hardness ratio image (4.5-9 keV/24.5 keV) overlaid with the 2-9 keV contour. Black regions are harder in X-rays. The position of an extremely hard (black) source XMM J174540-2940.5 is also indicated (see Sakano et al. 2003d, e for detail). (c,d) The He-like iron and sulfur Ka-line images, respectively, where the underlying continuum is subtracted. The 2-9 keV contour is superimposed. All the coordinates are in 12000. The Galactic Plane (bn = 0) is also indicated. Fig. la shows the X-ray image from the MOS1+2 cameras. The brightest spot in this image corresponds to the location of Sgr A’. The X-ray emission around Sgr A* is found to be extended. Note the time interval including the Sgr A* flare reported by Goldwurm et al. (2003a, b) was not included in our analysis. Sgr A* was reported to be resolved in the X-ray band by Chandra (Baganoff et al. 2003a), whereas the recent very deep Chandra observation found that Sgr A* is in fact slightly extended in the quiescent state (Baganoff et al. 2003b). At any rate, Sgr A* carries only about 10% of the total X-ray flux in the central 10” region (Baganoff et al. 2003a) except during flaring episodes. Taking the spatial resolution of XMM-Newton of -5 arcsec (for MOS; Jansen et al. 2001) into account, our result is consistent with the Chandra result.

Astron. Nachr./AN 324, No. S 1 (2003)

199

There is a further X-ray emission to the east of Sgr A*, extended nearly parallel to the Galactic plane. I t is clearly distinct in both the lower and higher energy bands. This X-ray emitting region is mostly contained within the radio shell of Sgr A East and thc molecular dust ring (see Maeda et al. 2002,2003; Park et al. 2003), which suggests a strong physical connection between the observed X-ray and radio emission. The whole X-ray emitting region of Sgr A East extends across 200 arcsec, whereas the e-folding radius of its core is 28 arcsec in the 2-10 keV band. The hardness image (Fig. Ib) shows that the regions to the north and south of Sgr A East in Galactic coordinates are enhanced in the soft X-ray band or, alternatively, that there is a narrow (< 3’ wide) strip along the Galactic plane with a harder spectral form. These features are also seen in the Ckundru image (Morris et al. 2003; Park et al. 2003). The scale height of the molecular clouds in the Galactic Centre region is 5’-9’ (Tsuboi, Handa & Ukita 1999; Sakano 2000), which corresponds to the full extent of our X-ray hardness picture. In addition the soft enhancement is significantly brighter than its adjacent regions. The most plausible interpretation would seem to bc that the soft X-ray enhancement represents an X-ray outflow. However, we note that the soft emission does not seem to emanate directly from Sgr A*, but from a region located a couple of arcminutes to the east of Sgr A * . Further discussion is found in Warwick (2002,2003). Fig. Ic and Id show narrow-band images corresponding to the He-like iron (6.7-keV) and sulfur (2.4keV) K a lines, respectively, where the underlying continuum has been subtracted using adjacent bandpasses and assuming an averaged spectral shape of the whole field of view. The 6.7-keV line is clearly more concentrated in the core of Sgr A East than the continuum (Fig. la). This implies that the core of Sgr A East is more abundant in iron, or possibly that it is higher in temperature, or the combination of the two. In contrast, the 2.4-keV line peak is located on Sgr A*, as is the peak in the continuum. This nature is quantitatively evaluated with the spatially-resolved spectral analysis described in the next section (Section 3.2). In these XMM-Newtondata, we further found more extended structure in both the 6.7-keV or 6.4-keV line, where the latter represents the fluorescent Ktu line from neutral or low-ionized iron. Detailed results are given in Warwick (2003) and Predehl (2003) in these proceedings.

-

3 X-ray spectrum 3.1

Model fitting of the whole Sgr A East spectrum

We extracted the source spectrum from a 100”-radius region, but excluded Sgr A* and its immediate surroundings (within a 24 arcsec radius region) and the response of a bright soft point source. The background spectrum is taken from an adjacent sky region at similar galactic latitude to the source region (see Fig. I and Sakano et al. 2003a for detail). Fig. 2 displays the resulting background-subtractedspectra taken with pn and MOS1 and 2. The spectra show several strong emission lines. We fitted the spectra with a phenomenological continuum and many narrow Gaussian lines, and found that the centre energies of most of these lines, except for the 6.4-keV line (see the previous section) correspond to K u lines from highly ionized ions. Then we estimated the ionization temperature for each atom from the ratio of the K-line fluxes of the helium-like and hydrogenlike atoms. Fig. 3 summarises the derived temperatures. The temperature is found to vary significantly from element to element; for example, the ion temperatures of sulfur and iron are 1.1 keV and 4.0 keV, respectively. This implies that the spectrum consists of multi-temperature components. In fact, when we tried to apply a single-temperature thermal model to the spectrum, it was clearly rejected. We next applied a two-temperature thin-thermal plasma model with a common ahsorption column, allowing the abundances of silicon, sulfur, argon, calcium, and iron in both the plasmas to be free. The abundance of nickel was linked to that of iron. In Ihe best-fitting result (Table I ) a counts excess below 2 keV is still evident which lead us to consider the possibility of patchy absorption. We assume that a

M. Sakano et al.: Sgr A East & its surroundings

200

S g r A East (all; Pcfabs(ZVmekal+6.4))

,

t 1

i i

5

2

10

channel energy (keV)

Fig. 2 The simultaneous fitting of the spectrum of Sgr A East within a radius of 100". A two-component thermal plasma model is employed with metal abundances of Si, S, Ar, Ca, and Fe allowed to vary. The low energy spectrum is modified by a partially covering absorber.

0

22

LSi '

~

"

'

~

"

"

S

~

'

'

'

Ar

'

~

"

'

Ca

Fe

Fig. 3 Ionization temperatures for each atom. These are estimated from the line ratio of K a lines from helium-like and hydrogen-likeions, assuming a thin thermal plasma in ionization equilibrium (Mewe 1985). The temperature of silicon may be subject to some systematic error due to uncertainty of the continuum model.

certain ratio 1- t (where E is a free parameter) of the emission is absorbed by a column of 7 x 1022Hcm-2 (fixed). The fitting result is found to be significantly improved with a probability by the F-test of99.996% (x2/dof=400.4/321).Table 1 and Fig. 2 summarise the best-fitting results. The bulk of the systematic trend in the residual is removed by this approach (Fig. 2).

The best-fitting heavy absorption column and its covering fraction (E) were found to be 1 4 . 3 ~ 1 0 ~ ~ H cm-' and 93%, respectively. The Galactic Centre region is known to be full of molecular clouds of

Astron. NachrJAN 324. No. S 1 (2003)

20 I

Table 1 The spectral fitting results Params. FIX( i J A;l(i) NII(ii)

Unit

- Data Whole r < 28" 0.93 ( 0 . 8 3 4 . 9 7 ) 0.93 (fixed) 14.3 ( 1.3.3-16.4) 17.4 (15.9-19.1) 7.0 (fixed) 7.0 (fixed) 4.34 (1.884.84) 3.20 (2.67-3.79) 1.44 (1.16-1.73) 0.40(0.27-0.62) 0.97 (0.90-1.05) 0.86 (0.72-0.98) 13.0(9.8-17.0) 5.3(3.4-9.l) 2.5 (0.80-6.3) 1.7 (1.1-2.7) 1.7(1.1-2.6) 1.38(1.15-1.65) 1.14(0.83-1.47) 1.1 (0.52-1 7 ) 1.7(1.0-2.4) 2.12(1.68-2.57) 3.5(2.8-4.5) 1.25(1.12-1.40) 0.2(0.0-0.4) l.Z(O.8-1.8)

Whole -

[H cm-'1 [H cm-']

13.7 (13.5-14.5) ~

4.34 (4.194.48) 1.38 (1.34-1.41) kT,(2) [keV] 0.93 (0.86-1.00) Norm(2) 11.2(9.h-13.4) Z Si 4.8 (3.7-5.4) ZS 1.71 (1.541.88) z.4 r 1.26 (0.97-1.55) ZCi( 2.55 (2.14-2.97) &, .N2 1.34(1.27-1.41) Fe-Ke 1.5(1.0-1.9) [10-'ph s-' ~ r n - ~ ] X'ldof 422.11322 Fx (MOS) 12.2 [IO-"erS s-' ~ m - ~ ] k?;(l) Norm( 1 )

lkeV]

l . = y -8

-60"

r > 60" 0.91 (fixed) 13.2 (12.414.2)

0.4(0.0-0.5)

7.0 (fixcd) 4.80 (3.94-6.04) 0.64 (0.49-0.80) 0.90 (0.79-1.01) 5.8(4.1-8.1) 1.3 (0.7-2.2) 1.32(1.03-1.71) 1.05 (0.59-1.63) 2.4(1.7-3.2) 0.47(0.38-0.61) 0.7(0.4-1.1)

258.61238 3.9

190.34/198 4.8

1.4(1.1-1.8)

235.0/180

400.4132I 12.2

ti

0.93 (fixed) 16.0 (14.7-17.3) 7.0 (fixed) 5.5 (4.6-6.4) 0.31 (0.24-0.42) 1.01 (0.92-1.16) 5.6(4.3-7.8) 2.8 ( 1 . 6 5 . 8 ) 1.4(1.0-2.1) 1.1 (0.69-1.6) 1.9(1.4-2.6)

3.3

The unit for the normalisation is lo-'' / ' n,.+n1~dI~/(47r1l2), where n,, and Z L I I are the electron and proton number densities (crK3)and D is the distance of the source (crn). The fluxes are for the 1-10 keV energy band. The quoted uncertainties are at 90'% confidencc for one interesting parameter. various sizes from radio and far infrared observations. This patchy-absorption model with this large covering fraction (i.e. small fraction for the 'hole') must, therefore, be realistic enough. We also tried a three-temperature model, including a third additional lower temperature component as an alternative to the partial absorption model but this gave neither good improvements in the fitting nor a realistic soft X-ray luminosity. Thus we found that: ( I ) temperatures of the two components are 1 keV and -4 keV, which further confirms the estimate from the line-ratios (see Fig. 3); (2) the lower-temperature component has an orderof-magnitude higher emission measure than the higher-temperature one; (3) metal abundances are a few, to several tens, percent higher than the solar value on average, except for calcium, which is significantly more abundant (-2 solar) than the other elements. N

3.2

Model fitting of the spatially split spectra

-

a)

.t '"

'"

0

_k'= 2;

0

0

% a

:i "

b) . -

0

Y O

2 0

f "

/

"

8

1:

1 0

E e

1 0

X

x

cn*nna, *"e?$? (re")

D

m0nn.i

en=,ll

(hv)

m

nmnn., *n.qr

("4

Fig. 4 The spectral fitting results for different sub-regions in Sgr A East. We have applied a two-temperature thin thermal plasma model plus a patchy absorption. (a) the core region with a radius of 28"; (b) thc annular region within radii of 28"-60"; (c) the whole source region outside of the 60"-radius circle (see Fig. la). For simplicity only the pn data are shown although the results apply to the simultaneous fitting of both the pn and MOS1/2 datasets.

The images of Sgr A East (Fig. Ic) show that the iron-line component concentrates in the core region more than continuum. Thus, we examined the spectral variation within Sgr A East. We extracted spectra

202

M. Sakano et al.: Sgr A East &its surroundings

from concentric regions of radii 28” (the core region) and 28/’-60” (the peripheral region) centred on the peak of the 6.7-keV-line image (Fig. lc). Also an outer region given by the full source region less the area within the 60’’ circle was also examined (see Fig. 1 a). The spectra from these three sub-regions were fitted with the final model defined in Section 3.1 with the covering factor for the absorption fixed, but all the other parameters free. Fig. 4 shows the resulting best-fitting models and Table 1 lists the result. We found that the iron abundance is higher for the smaller “core” region with Z = 3.5, whereas 2 M 0.5 in the outer region. The abundances of other metals do not significantly vary within Sgr A East from the solar norm. The temperatures of the plasmas are slightly lower in the core region. Neither the ratio of the intensity of the lowerand higher-temperature components varies. The 6.4-keV iron fluorescent line is significantly detected only in the outer region with an equivalent width of -75 eV. 3.3 Plasma parameters From the apparent extent of Sgr A East in the hard X-ray band in a 28” radius, we calculate the total plasma volume V to be 1 . 6 ~ 1 cm3, 0 ~ ~assuming a spherical shape. The plasma is found to comprise two components and we assume that each component exists separately in pressure balance with the other. Using the best-fitting parameters for the core region and introducing the total filling factor qttot,which is the sum of the filling factors of the two components, we estimated the filling factor of = 0.5qtot and 7114 = 0.5qtot, the density of n , , ~ = 23qtot -112 and n c , H = 6.1r,tii12, a total energy of E = 1.5 x 1049,vtot, 112 erg, and a total X-ray emitting mass of 19.17;: M a , where the subscriptions of L and H mean the lowerand higher-temperature components, respectively.

4 DISCUSSION 4.1 Is Sgr A East a SNR? Mezger et al. (1989) claimed that Sgr A East may be multiple supernovae based on their derived total energy of 6 x lo5’ erg. On the other hand, our derived total energy of the hot plasma in Sgr A East is 1.5 x 104g17toterg, which i s clearly much smaller than the Mezger et al. value and even smaller than the nominal energy for a single SNR (- 10” erg), thus one SNR can easily account for Sgr A East. Since the plasma has already reached ionization equilibrium, this estimate does apply to the full thermal energy in the observable X-ray band. On the other hand, the estimated (observed) mass of 1 .9qtot f Ma is sufficiently large to account for a supernova. With an age of 8000 yr for Sgr A East (Mezger et al. 1989),this mass may originate as either the ejecta or swept-up interstellar material. The localisation of the iron abundance enhancement suggests that the mass is predominantly that of the ejecta. This value of the mass, as well as the total energy, suggests that a single SNR is the likely scenario for the origin of Sgr A East. The most remarkable characteristic of Sgr A East is the unusually high temperature of 4 keV for a SNR of this age. One possible scenario to explain this high temperature is that the shock has interacted with an ambient plasma already preheated to a temperature of several keV, perhaps due to past activity in Sgr A*. N

4.2 Origin of the 6.4-keV line We detected significant iron fluorescent line emission at 6.4-keV only in the outer region of Sgr A East with an equivalent width of -75 eV (Section 3.2). This outer region is nearly coincident with the observed dust ringhhell (Mezger et al. 1986, 1989). The column of this dust shell is estimated to be 3 x (Mezger et al. 1989). If we assume a simple 10% cm7712 based on an average density of lo4 isotropic morphology for the cloud, this column can account for the observed equivalent width of the 6.4keV line. Hence the detection of the 6.4-keV line supports the idea that Sgr A East is actually surrounded by the dust shell. N

Astron. Nachr./AN 324, No. S 1 (2003)

4.3

203

The Metal Abundance in the Galactic Centre Region

The metal abundance in the Galactic Centre region remains uncertain with estimates ranging from near solar based on infrared star observations (Carr el al. 2000; Ramirez et al. 2000; Sellgren et al. 2003) to a few times the solar value or more based on a variety of other methods (e.g. Mezger et al. 1979; Murakami et al. 2001). Although there will inevitably be some contamination of our spectra by point sources which may have power-law like spectra and which are not spatially resolved with XMM-Newton, the impact must be small since the plasma emission dominates this region (Maeda et al. 2002). Our results should therefore be robust and show that, except for the core of Sgr A East, the iron abundance in the high-temperature interstellar matter that pervades the Galactic Centre region, is near to the solar value or possibly sub-solar (see also Sakano et al. 2003b). Acknowledgements M. S. acknowledges the financial support from JSPS.

References Baganoff, F. K., et al. 2003a, ApJ, submitted (astro-pW0102151) Baganoff, E K. et al. 2003b, these proceedings Carr, J. S., Sellgren, K., & Balachandran, S. C. 2000, ApJ, 530, 307 Dogiel, V. A., lchimura, A., Inoue, H., & Masai, K . 1998, PASJ, SO, 567 Ekers, R. D., van Gorkom, J. H., Schwarz, U. J., & Goss, W. M. 1983, A&A, 122, 143 Goldwurm, A,, et al. 2003a, ApJ, 584, 751 Goldwurm, A., et al. 2003b, these proceedings Jansen, F., et al. 2001, A&A, 365, L1 Jones, T. W. 1974, A&A, 30,37 Koyama, K., Maeda, Y., Sonobe, T., Takeshima, T., Tanaka, Y., & Yamauchi, S., 1996, PASJ, 48, 249 Maeda, Y., et al. 2002, ApJ, 570, 67 I Maeda, Y., et al. 2003, these proceedings Mezger, P. G., Penkonin, V., Schmid-Burgk, J., Thum, C., &Wink, J. 1979, A&A, 80, L3 Mezger, P. G., Chini, R., Kreysa, E., & Geniiind, H. -P. 1986, A&A, 160, 324 Mezger, P. G., Zylka, R., Salter, C. J., Wink, J. E., Chini, R., Kreysa, E., & Tuffs, R. 1989, A&A, 209, 337 Mezger, P. G., Duschl, W. J., & Zylka, R. 1996, A&AR, 7,289 Murakami, H., Koyama, K., & Maeda, Y. 2001h, ApJ, 558, 687 Park, S., et al. 2003, these proceedings Reid, M. J. 1993, ARA&A, 31,345 Sakano, M. 2000, Ph.D. thesis, Kyoto Univ Sakano, M., Warwick, R. S., Decourchelle, A,, & Predehl, P. 2003a, in preparation Sakano, M., Warwick, R. S., & Decourchelle, A. 2003b, AdSpR, submitted Sakano, M., Warwick, R. S., & Decourchelle, A. 2003c, Proc. “Japan-Germany Workshop on Galaxies and Clusters of Galaxies”, p.9 (astro-pW0212464) Sakano, M., Warwick, R. S., & Decourchelle, A. 2003d, MNRAS, 340,747 Sakano, M., Warwick, R. S., & Decourchelle, A. 2003e, these proceedings Sellgren, K., 2003, these proceedings Sidoli, L., & Meregherti. S. 1999, A&A, 349, L49 Tsuboi, M., Handa, T., & Ukita, N. 1999, ApJS, 120, 1 Wang, Q. D., Gotthelf, E. V., & Lang, C. C. 2002, Nature, 415, 148 Wang, Q. D. 2003, these proceedings Warwick, R. S. 2002, Proc. New Visions of the X-ray Universe in the XMM-Newton and Chandra era., in press (astro-phl0203333) Warwick, R. S. 2003, these proceedings Zylka, R., Mezger, P. G., & Wink, J. E. 1990, A&A, 234, 133

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 205 -210 (2003) / DO1 10.1002/asna.200385107

Chandra ACIS Imaging Spectroscopy of Sgr A East Y. Maeda", K. Itoh', F. K. Baganoff2, M. W. Bautz2, W. N. Brandt3, D. N. Burrows3, J. P. Doty2, E. D. Feigelson", G. P. Garmire', M. Morris4, M. P. Muno', S. Park-?,S. H. Pravdos, G. R. Ricker2, and L. K. Townsley'

' Institute of Space and Astronautical Science, 3-1 - 1 Yoshinodai, Sagamihara, Kanagawa, 229-85 10, Japan ' Center for Space Research, MIT, Cambridge, MA. 02139, U.S.A. '

Astronomy and Astrophysics, Penn State University, 525 Davey Lab., University Park, PA. 16802, U.S.A. Physics and Astronomy, UCLA, Los Angeles, CA. 90095, U.S.A. Jet Propulsion Laboratory, MS 306-438.4800 Oak Grove Drive, Pasadena, CA 91 109, U.S.A.

Key words Galaxy: center - ISM: clouds - X-rays: individual (Sagittarius A East) -X-rays: ISM Abstract. We report on the X-ray emission from the shell-like, non-thermal radio source Sgr A East located in the inner few parsecs of the Galaxy based on observations made with the ACIS detector on board the Chandra X-ruy 0h.vewutory. The X-ray emission from Sgr A East is concentrated within the central Y 2 pc of the larger radio shell. The spectrum shows strong Kcu lines from highly ionized ions of S, Ar, Ca, 2 keV, absorption column and Fe. A simple isothermal plasma model gives electron temperature 1x H cm-', luminosity 8 x lo3* ergs ti-' in the 2-10 keV band, and gas mass 279 M o with a filling factor 17. The plasma appears to be rich in heavy elements, over-abundant by roughly a factor of four with respect to solar abundances. Accompanied with filamentary or blob-like structures, the plasma shows a spatial gradient of elemental abundance: the spatial distribution of iron is more compact than that of the lighter elements. These Chundrn results strongly support the long-standing hypothesis that Sgr A East is a supernova remnant (SNR). Since Sgr A East surrounds Sgr A* in projection, it is possible that the dust ridge compressed by the forward shock of Sgr A East hit Sgr A* in the past, and the passage of the ridge may have supplied material to accrete onto the black hole in the past, and may have removed material from the black hole vicinity, leading to its present quiescent state.

-

-

-

-

1 Introduction The center of our Galaxy embodies a rich variety of phenomena which create diverse complex structures that are visible to us over a broad range of wavelengths. The radio emission from the central few parsecs of the Galaxy has several components, including a compact non-thermal source (Sgr A*) thought to he associated with the central massive black hole, a spiral-shaped group of thermal g a s streams (Sgr A West) that are possibly infalling to Sgr A*, and a 3 ' 5 x 2 ' 3 shell-like non-thermal structure (Sgr A East) (Ekers et al. 1975). S g r A East surrounds Sgr A* West in projection, but its center is offset by about 50" (2 pc). A number of arguments suggest that Sgr A* West is physically located very near or possibly embedded within Sgr A East. For the latter case, interaction between S g r A East and Sgr A* West would he inevitable, so S g r A East may h e a key for understanding the activity in the nucleus of our Galaxy (for a recent review, see Yusef-Zadeh, Melia & Wardle 2000).

*

Corresponding author: e-mail: [email protected],Phone: 81 42759 8150, Fax: 81 42759 8455

@ XH!i WILEY-VCH V c i l q CimbH & Co KCPA, Weinhemi

(XKL4-61i71031SIOI-00016 17 SO+ 5010

Y. Maeda et al.: Chandra View of Snr A East

206

The Chandra X-ray Observatory (Chrmdru) with the Advanced CCD Imaging Spectrometer (ACIS) 10 keV and moderate spectral resolution with the detector combines the wide-band sensitivity up to superior high spatial resolution (0”.5-1”) of Chundru’s High-Resolution Mirror Assembly (HRMA). In this paper, we present a preliminary analysis of ACIS Galactic center observations with an exposure time of over 500 ksec, and review the first-epoch observations published by Maeda et a]. 2002. The other findings using the same data are separately reported in this volume by Baganoff, et al., Morris et al., Muno et al., and Park, et al.. XMM-Newton studies of Sgr A East are also presented by Predehl et al. and Sakano et al. in this volume. Throughout this paper we adopt a distance of 8.0 kpc to the Galactic Center (Reid 1993). N

2 Results 2.1 Images The first-epoch observation of the Galactic Center region was carried out early in the Chundru mission on 1999 September 21 over a period of 5 1 ks using the ACIS-I array of four abutted, frontside-illuminated CCDs. The follow-up observations reached 590 ksec as of 2002 June. During most of the follow-up observations, the central region of Sgr A East was pointed at the gap of the CCD array. To minimize any systematic uncertainties due to the gap, we used the first-epoch observations for an overall view of Sgr A East, while all the available data has been used to investigate the selected structures presented in this paper. The smoothed broad-band ( 1 5 - 7 keV) X-ray image of the Sgr A radio complex using the first-epoch observation is shown in Figure 1, where we have overlaid the X-ray image with radio contours from a 20 cm VLA image (F. Yusef-Zadeh. private communication). The outer oval-shaped contours are due to synchrotron emission from Sgr A East (Ekers 1983). The thermal radio source Sgr A West, an HI1 region located in the central parsecs of the Galaxy, appears on the western side of the Sgr A complex. Several bright compact X-ray sources can be seen in the images. With the follow-up observations, the number of the point sources reached over 2300 point sources within -20 pc (Muno et al. 2003). One of these sources is coincident to within 0”.35 with the radio position of the compact non-thermal radio source Sgr A* (Baganoff et al. 2001). In addition to the compact sources, the Sgr A West region shows bright diffuse X-ray emission superposed on a broader region of diffuse emission which peaks 1’ east of Sgr A*, and which appears to fill the central N 2 pc of the Sgr A East radio shell (Figure 1). Based on its spatial properties, we identify this diffuse X-ray emission centrally concentrated within the radio shell as an X-ray counterpart of Sgr A East. Figure 2 shows raw images in both the total and iron-K (6-7 keV) bands without exposure corrections. A broader feature in the center of the Sgr A East shell is especially conspicuous in the 6-7 keV band, where the flux is dominated by iron-K line emission. A curious linear feature 0”.5 long, which we refer to as ’the Sgr A plume’ in Baganoff et al. (2001), can be seen extending (in projection) from the center of Sgr A East to the northeast. In addition to the Plume, we found several filamentary structures within the Sgr A East shell (Figure 2). The filamentary structures in the ACIS field of view is comprehensively reported in our companion papers Morris et al. (2003) and Park et al. (2003). Notably, neither significant X-ray continuum or line emission is seen from the location of the radio shell.

-

-

2.2 Spectra By extracting counts from a circular region of radius 40” within the shell, we made an over-all spectrum as shown in Figure 3. The spectrum exhibits a continuum plus emission lines which give critical information on the physical state of the plasma. We first fit the spectrum with a thermal bremsstrahlung model having four Gaussian emission lines of unspecified energy, all absorbed by an interstellar medium having cosmic abundances. The emission lines can be attributed to the Ka: transitions of the helium-like ions of sulfur, argon, calcium and iron. The line width for each line is consistent with being unresolved. The existence of

Astron. Nachr./AN 324. No. S 1 (2003)

207

the highly ionized ions confirms the presence of an optically thin thermal plasma. The equivalent widths of the lines are relatively small for the first three elements (EW 0.1-0.2 keV) hut very large for iron (EW 2 3.1 keV). The continuum temperature is around 3 keV and the line-of-sight column density is around 1 x loz3cm312. The large equivalent width of the iron line suggests that the Sgr A East plasma is enriched in heavy elements. Since the high ionization state of the iron K-line can be reproduced by a plasma in collisional ionization equilibrium (Masai 1994), we fit the spectrum to models of an isothermal plasma having variable elemental abundances (MEKA; Mewe, Gronenschild & van den Oord 1985), modified by interstellar absorption. The model with solar elemental abundances (Anders et al. 1989) was rejected with xz = 309 ( 1 85 d.0.f.) because it does not reproduce the large equivalent width of the iron line. The best-fit (x2= 217 with 184 d.0.f) was obtained using heavy element abundances which are 4 times solar. These results are given in Table 2 and are shown as a solid line in Figure 3.

=

Table 1 Best-fit parameters to the Sgr A East spectrum fitted with the MEKA model

Parameter [unit]

Best fit value

N H [loz2cmP2]

11.4(10.5-12.3)

kT, [keV]

2.1( 1.9-2.4)

Z

3.9(2.9-5.9)

Normalization

1.1 (0.9-1.3)

217( 184)

X2(d.o.f.)

/ ~ L , ~ L H/ (47rD2), ~V where 71, is the electron number density (cm-’)), ~ L H Normalization: is the proton number density (cm-’)). and D is the distance to the source (cm). n, = 1.8 x I L H for the best tit. 3 What is Sgr A East? The X-ray spectrum enriched by heavy elements suggests that the X-ray plasma is dominated by supernova ejecta. The small gas mass of 271; Ma and thermal energy lo4’ ergs are consistent with the ejecta by a single SN explosion. The X-ray images show filaments which are usually detected shocks in SNRs. Therefore, these results strongly supports the long-standing hypothesis that Sgr A East is a single SNR (Ekers et al. 1983). Rho and Petre (1 998) defined a new class of composite SNRs showing centrally concentrated thermal Xrays lying within a shell-like non-thermal radio structure. They called these objects “mixed morphology” (MM) supernova remnants and identified 19 members of the class. With the centrally concentrated X-ray cmission we find here, and its well-established non-thermal radio shell, Sgr A East becomes a ncw member of the class of MM SNRs. The non-thermal radio emission from Sgr A East in the direction of Sgr A West is heavily absorbed by Sgr A West (Yusef et al. 1987). This fact convincingly indicates that, along the line-of-sight, Sgr A West lies in front of the Sgr A East shell. However, the distance between Sgr A West and Sgr A East along the line-of-sight is uncertain. It is quite possible that they lie at nearly identical distances, in which case the front edge of the expanding Sgr A East shell has probably reached and passed through Sgr A*. This configuration is simply illustrated in Figure 5, see Baganoff t al. (2001) and Maeda et al. (2002) for a more thorough discussion.

-

Y. Maeda et al.: Chandra View of Sgr A East

208

Fig. 1 Smoothed X-ray image (1.5-7.0 keV) taken during the first-epoch obscrvations, overlaid with 20 cm radio contours (white: Yusef-Zadeh private communication). Both exposure and vignetting corrections were applied. The field center is at Sgr A*, and the pancl size is 8’.4 x 8l.4.

Acknowledgements This work has been supported in parts by NASA contract NAS8-OI 128, NAS 8-38252 for the Chandra X-Ray Observatory.

References Anders, E., & Grevesse, N. 1989, Geochimica et Cosmochimica Acta, 53, 197 Baganoff, F. K. et al. 2001, ApJ submitted Baganoff, F. K. et al. 2003, in this volume Ekers, R. D., van Gorkom, J. H., Schwarz, U. J. & Goss, W. M. 1983, A&A, 122, 143 Maeda. Y. et al. 2002. ADJ. 570.671 Masai, K. 1994, ApJ, 437,770 Mewe. R., Gronenschild, E.H.B.M., & van den Oord, G.H.J. 1985, A&AS, 62, 197 Morris, M. P. et al. 2003, in this volume Muno, M. P. et al. 2003a, ApJ, Submitted Muno, M. P. et al. 2003b, in this volume Park, S. et al. 2003, in this volume Predehl, P. et al. 2003, in this volume Reid, M. J. 1993, Annual Review of Astronomy and Astrophysics, 3 I , 345 Rho, J . & Petre, R. 1998, ApJL, 503, L167 Sakano, M. et al. 2003, in this volume Yusef-Zadeh, F., Melia, F. & Wardle, M. 2000, Science, 287, 85

Astron. Nachr./AN 324, No. S1 (2003)

209

Fig. 2 Deep Raw X-ray images in the total (left) and iron-K band (right) with 20 cm radio contours (green). Neither exposure nor vignetting corrections were applied. The dark lanes which crosses at the center are due to the gap of the CCD array.

L

1

2

5

10

Energy (keV) Fig. 3 The ACIS-I spectrum of Sgr A East. Error bars are 1 0.The solid line corresponds to the best-fit value with the optically thin-thermal model summarized in Table 2. Fit residuals are shown in the bottom panel.

Y. Maeda et al.: Chandra View of Sgr A East

210

data and folded model grp5Q/dinusaSmarga_grp.pha

I

1

data and folded model grp5Q/fillamant6_meq)b4rp.pha

I.

I.

I I

channel energy (kev)

1+_1IW

chonnsl energy (kev)

2 - c -

!W

Fig. 4 Spectra taken from two regions of Sgr A East. The left spectrum is taken from a central region of the shell, which shows the condensation at the iron-K band image in Figure 2. The equivalent width of the iron-K line was found as 5 keV. The right is for the filament A shown in Figure 2. The spectrum is dominated by thermal emission.

Fig. 5 A schematic diagram showing now a supernova remnant might regulate gas falling onto the supermassive black hole at the center of our Galaxy (Baganoff et al. 2001; Maeda et al. 2002). A supernova Sgr A East exploded about 10,000 yrs ago near the supermassive hlackhole in the Galactic Center. The inward shock wave heated up the ejecta that was detected in X-rays by ACIS, while the outward shock pushed gas towards the central black hole perhaps a thousand years ago.

Astron. Nachr./AN 324, No. SI, 211-215 (2003) /DO1 10.1002/asna.200385097

A Census of Dust Absorption at the Galactic Centre

',

Andy Adamson* Rachel Mason?, Emily Macdonald3, Gillian Wright2, Jean Chiar4, Yvonne Pendleton4, Tom Kerr', Janet Bowey', Doug Whittet6, and Mark Rawlings7

' Joint Astronomy Centre, Hilo, Hawaii 96720 U.S.A. ' Royal Observatory Edinburgh, Blackford Hill, Edinburgh EH9 3HJ, UK

'

University of Oxford, Department of Astrophysics, Keble Road, Oxford OX13RH, UK NASA Ames Research Center, Mail stop 245-3, Moffett Field, CA 94035, USA University College London, Dept. of Physics and Astronomy, Gower Street, London WClE 6BT. UK Rensselaer Polytechnic Institute, Department of Physics, Applied Physics and Astronomy, Troy, N Y 12180 Observatory, P.O. Box 14, University of Helsinlu, FIN-00014 Helsinki, Finland

Key words Interstellar: dust, molecules, extinction

Abstract. The Galactic Centre offers a uniquely valuable line of sight for studies of the nature of dust in the ISM, but the long-held assumption that the line of sight samples only the diffuse ISM has been subverted by ground-based and I S 0 observations demonstrating the presence of absorption bands due to solid-phase volatile ices in the field. Spatial variability of the observed features suggests a patchy distribution of even the foreground diffuse-medium absorption. To map this clumpy distribution and to produce an inventory of the dust components, we are carrying out a narrow-band imaging survey over a field which extends from the Galactic Centre to beyond the circumnuclear ring, using both IRCAM3 and Michelle on UKIRT to cover the 3 pm water ice, 3.3 pm PAH, 3.4 pm hydrocarbon and 9.7 pm silicate absorption features. This paper presents the rationale for this programme and reports on progress with analysis of the survey data.

1 Introduction

I

Through detailed infrared spectroscopy, we know much about the nature of the materials present in hydrocarbon dust grains in the diffuse ISM. The 3.4 pm absorption band, first seen in the diffuse ISM along the line of sight to the Galactic Centre (GC) (e.g. Willner et al. 1979, Butchart et al. 1986), has been thoroughly analyzed, and substructure within the band is identified with short-chained aliphatic hydrocarbons. I S 0 detections of the corresponding weak deformation modes at 6.85 and 7.25 pm (Chiar et al. 2000) have considerably narrowed the range of possible hydrocarbon materials (Pendleton & Allamandola 2002). The relationship of the diffuse-ISM hydrocarbons to other grain components is poorly known, although there is a correlation between the 3.4 p m band and the silicate feature over a range of extinctions (Pendleton et al. 1994, Sandford et al. 1995). Spectropolarimetry suggests that these materials cannot be physically associated with any form of aligned grains which are responsible for the visual and near-IR polarization of starlight (Adamson et al. 1999), and they cannot be associated with organic refractory mantles (i.e. a result of processing of ices - Chiar et al. 1998). The observations tend to favour models invoking small grains over those requiring carbonaceous mantles on large grains, though the key spectropolarimetry results will remain controversial until observations of organic and silicate polarization spectra can be obtained in exactly the same line of sight. In dark clouds, the 3.4 p m band is replaced by absorption at 3.47 pm (Allamandola et al. 1992), which is known to be associated with ices rather than the silicates (Brooke et al. 1996, Brooke et al. 1999, Chiar * Corresponding author: e-rnail: [email protected], Phone: iOOl808969651I. Fax: +OO18089616516

02003 WILEY-VCH V e r l q GmhH & Co. KGaA. Weinhem

A. Adamson et al.: Dust Absorption

212

et al. 1996), indicating a major change in the nature of the hydrocarbon-bearing grains. Absorption due to polycyclic aromatic hydrocarbons (PAHs) has been suggested on the basis of a 3.25 pm absorption band seen in some dark cloud sources (Brooke et al. 1999, Sellgren et al. 1994). While seen in sources with deep ice absorption, correlations suggest that these materials are in fact more closely related to refractory grains such as silicates (Brooke et al. 1999). In fact, dense-cloud and diffuse-cloud PAHs may be different in nature: the dense-cloud feature at 3.25 pm peaks at a shorter wavelength, indicating a lower temperature and/or different composition, than the feature observed toward GCS3 in the Galactic Centre (which peaks at 3.28 pni - Chiar et al. 2000).

2 Rationale Our CGS4 spectroscopy of dust grains in the Galactic Centre (Chiar et al. 2002) permitted complete deconvolution of the complex of ice and dust absorption features along this line of sight. Fig. 1 shows the deconvolved hydrocarbon absorption feature in the 3.2-3.7 pm region, comprising the 3.4 /m aliphatic hydrocarbon band and a broad (FWHM 0.15 pm) absorption centered around 3.3 pm. Previous work, which has relied on fitting a local continuum across the sharp 3.4 pm band in what is a very difficult part of the telluric spectrum, has inadvertently removed this broad feature. The band is indicative of aromatic hydrocarbons, although it is too broad for absorption arising in PAHs embedded in solid grain material.

1

1 4.2

Fig. 1 Spectrum of hydrocarbon absorption in the Galactic Centre, deconvolved from the ice band (Chiar et al. 2002). Sharp negative lines are ratioing and atmospheric artifacts; m o w s indicate the main 3.4 pm band and the -3.3 pm “PAH’ absorption. Subfeatures near 3.5 p m are also associated with aliphatics.

The strength of the ice band varies dramatically across the Galactic Centre field. This may be due to additional ice absorption in the circumnuclear molecular ring (IRSS and IRS19 are both situated in or behind the ring, and these two sources have strong ice bands relative to the 3.4 pm feature; see the map of this region in Chiar et al. - this volume). Our spectroscopy shows that the 3.4 pm band also vanes across the field, raising the possibility that it arises in diffuse clouds which are detectably patchy, on very small transverse scales. The aim of the current programme is to determine the variability of hydrocarbon and ice absorption across the entire Galactic Centre, and so (i) to determine the scale size of the absorbing regions and (ii) further define the relationship between the aliphatic and aromatic hydrocarbon components.

213

Artron. Nachr./AN 324, No. S1 (2003)

3 Observations We have carried out a 3 and 10 p m survey of the central 2 arcminutes of the Galactic Centre, in narrowband filters which sample all the key NIR dust features. The field of view includes the circumnuclear ring, and JHK imaging data already exist with which to constrain the extinction and short-wavelength wing of the ice absorption (Blum et al. 1996). Observations were obtained with the UKIRT facility imagers IRCAM3 and Michelle, in the summer of 2001 and 2002 respectively.

3.1 3 pm photometry IRCAM3 mosaics were obtained using the 3. I , 3.3, 3.4, 3.6 and 4.0 pm narrow-band filter sets. A mosaic in the broad L' filter was also obtianed. Fig. 2 shows a selection of images, and the filter bandpasses used. 100

:

3

32

34

36 Wsvelengthlw

38

42

a)

Fig. 2 a) The IRCAM3 filter set after convolution with atmospheric transmission; b) IRCAMS mosaics in the 3.1 & 3.3 p m filters. North is at the right, East is at the top

Fig. 3 a) L' continuum mosaic; b) Derived catalogue. Orientation is as in Fig. 2.

The L' mosaic image provides a deep L-band catalogue of the nuclear region, containing approximately 270 sources. Most of these are also detected in all the narrow-band filters. The basic catalogue is the deepest available wide-field L-band list of which we are aware in this region; only 10 of these sources have L-band photometry in the Blum et al. catalogue (Fig. 4), which covers a similar area of sky. By comparison,

A. Adamson et al.: Dust Absorption

214

the existing near-infrared data required to provide additional model constraints is more extensive: 96 of these sources are detected in J H and K by Blum et al. Work is in progress on deeper JHK imaging data over the same field, taken with UFTI on UKIRT. This will permit constraints to be placed on extinction and thus remove an apparent degeneracy between the depth of the ice band and the degree of extinction.

4 ' 10

11

12

13

14

15

16

IRCAM L' instrumental magnitude

Fig. 4 Photometry within the overlap between our work and Blum et al. L band catalogue. Instrumental magnitudes from IRCAM3 are negative.

3.2

10 pm photometry

Michelle imaging in filter bandpasses in the 7.9,9.7, 1 1.6 and 12.5 pm filters was carried out in summer of 2002. A small chop was used (of order three arcseconds) to chop out as much as possible of the extended structure which pervades the Galactic Centre. The imaging is most sensitive in the 7.9 pm filter, where more than 20 sources are detected within the immediate environs of the Galactic nucleus. The majority of these sources are associated with stellar objects also detected in the L-band catalogue. These data will provide a measure of the silicate absorption complementing the hydrocarbon and ice measurements, for comparison with the data on organic and volatile components from the 3 pm survey.

4

Modelling

Modelling of the data consists of iterative fitting of a model spectrum of a hot stellar source, subjected to both continuum extinction and the combined effect of the silicate, ice and hydrocarbon absorptions, folded through the known bandpasses of the filters used. Since the images contain sources for which these features have already been observed spectroscopically, the scaling factors between observed counts and feature depths are self-calibrating. Initial results suggest that with the continuum extinction fixed by JHK photometry, reliable ice-band optical depths can be extracted, but to isolate the three main components of the hydrocarbon band from one another will be difficult. Clearly, silicate depths will be available only for a small subset of the sources detected in the L-band survey, but the detection of stellar sources in the silicate filters at this distance is a major challenge.

Astron. Nachr./AN 324, No. S I (2003)

215

Fig. 5 a) 7.9 pm image; b) 1 I .6 pm image. As for the IRCAM3 frames, North is to the right, East is to the top. Each detected source (the best detections are shown in boxes) shows the chophod signature, with positive beam in the centre and two negative beams on either side.

5 Summary Observations with the UKIRT facililty imager IRCAM3 have produced both an L-band catalogue, and narrow-band images over the central two arcminutes of the Galactic Centre. T h e catalogue is the deepest available over such a wide area in this region: some 270 sources are detected in the broad L' filter, an order of magnitude more than are present in the previous deepest list (Blum et al.). Narrow-band detections in filters distributed across the three-micron window are available for most of the detected sources. The fluxes are being fitted to produce a catalogue of the variation of key volatile and refractory dust components as a function of position within the Galactic Centre. Work is in progress on complementary JHK imaging, which supplements the J H K catalogue of Blum et al., and permits the continuum extinction to be constrained. Finally, w e have recently obtained Michelle observations over a similar field, detecting more than 20 faint point sources exterior to the central cluster, most of which are probably stellar in nature.

Acknowledgements We gratefully acknowledge the staff of UKIRT for their support of the telescope, and the Michelle team from the UK Astronomy Technology for their excellent instrument and its commissioning.

References Adamson,A.J., Whittet,D.C.B., Chrysostomou,A.C., Hough,J.H., Aitken,D.K.. Wright,G.S., Roche,P.F. 1999, ApJ, 5 12,224 Allamandola, L.J., Sandford,S.A., Tielens,A.G.C.M., Herbst,T.M. 1992, ApJ, 399, 134 Blum,R.D., Sellgren,K., Dep0y.D.L. 1996, ApJ. 470,864 Brooke,T.Y., Sellgren,K., Smith,R.G. 1996, ApJ, 459, 209 Brooke,T.Y., Sellgren,K., Geballe,T.R. 1999, ApJ, 517, 883 Butchart.1.. McFadzean,A.D., Whittet,D.C.B., Geballe,T.R. & Greenberg,J.M. 1986, A&A, 154, L5 Chiar,J.E., Adamson,A.J., Whittet,D.C.B. 1996, ApJ, 472, 66.5 Chiar,J.E.. Pendleton,Y.J., Geballe,T.R., Tielens,A.G.G.M. 1998, ApJ, 507, 281 Chiar,J.E., Tielens,A.G.G.M., Whittet,D.C.B., Schutte,W.A., Boogert,A.C.A., Lutz,D., van Dishoeck,E.F., Bernstein,M.P. 2002, ApJ, 537, 749 Chiar,J.E., Adamson,A.J., Pendleton,Y.J., Whittet,D.C.B., Caldwell,D.A., Gibb,E.L. 2002, ApJ. 570, 198 Pendleton,Y.J. & Allamandola ,L.J. 2002, ApJS, 138.75 Pendleton,Y.J., SandfordSA., Allamando1a.L.J.. Tielens,A.G.G.M., Sellgren,K. 1994, ApJ, 437, 683 Sandford,S.A., Pendleton,Y.J.,Allamandola,L.J. 1995, ApJ, 440, 697 Sellgren,K., Smith,R.G., Brooke,T.Y. 1994, ApJ, 433, 179 Willner, S.P., RusseIl,R.W., Puetter,R.C., Soifer,B.T. & Harvey,P.M. ApJ, 229, L65

This Page Intentionally Left Blank

Astron. NachrJAN324, No. S I , 217-221 (2003) /DO1 10.1002/asna.20038S061

Thermal SiO observations of a shell attached to the nonthermal filaments in Sgr A Toshihiro Handa * I , Masaaki Sakano**2,and Masato Tsuboi***3

' Institute of Astronomy, University of Tokyo, Osawa 2-21-1, Mitaka, Tokyo 181-0015, Japan ' Department of Physics and Astronomy. University of Leicester, Leicester LE1 7RH, UK Astrophysics and Planetary Sciences, Ibaraki University, Mito, Ibaraki 310-8512, Japan

Key words Galactic Center, nonthermal filaments, molecular cloud, interstellar shock.

Abstract. The CS observations with NRO 4Sm telescope reveal that a dense molecular shell is located between Sgr A and the nonthermal filaments of the radio arc, and that the shell shows an interacting feature with the nonthermal filaments (Tsuboi et al. 1997). The shell i ociated with X-ray emission observed with ASCA and Chdndrd (Yusef-Zadah et al. 2002). These result suggests the shell is formed by a shock. The SiO lines in the v = 0 state are thermally excited and are thought to be a shock tracer. They are good probes to investigate the physical property of the shell. We observed the shell in SiO ( J = 1 0.71 = 0) and SiO ( J = 2 - 1, u = 0) using the NRO 4Sm telescope. Features associated with the shell clearly appear in both the SiO lines. The morphology of the shell in both SiO lines after adjusting the beamsize effect is similar in 2 - b - v 3-dimensional space. The intensity ratio of the two SiO lines ( J = 2 - 1 over J = 1 0) is uniform over the shell, suggesting the shell is uniform in density. Weestimate the average value of the ratio R ~ - l / l - uzz 0.9, which means the molecular &asdensity is about lo5 using multi-level population analysis with the LVG approximation. The morphology of the shell in SiO lines is quite similar to in the CS line in 1 - b - v 3-dimensional space. The intensity ratio of SiO ( J = 1 - 0) to CS ( J = 1 - 0) is almost uniform over the shell and Rs,o,cs Y 0.24. The ratio of the SiO (.I = 1 0) line over the CS ( J = 1 0) line is about 0.24, which means the relative abundance of SiO over CS, X(SiO)lX(CS) = 0.05 0.13 using multi-level population analysis with the LVG approximation. It ia close to the value found in an SNR shell. ~

~

~

~

~

1 Introduction The radio arc in the Galactic Center is a unique feature and its origin is still beyond the veil. Using spectral properties it is divided into two parts; the nonthermal filaments and the arched filaments. T h e nonthermal filaments are dominated by synchrotron emission and extend vertically to the galactic plane. The nontherma1 filaments are thought to b e a part o f the polari7ed lobe found by Tsuboi e t al. (1985). The arched filaments show thermal origin and connect Sgr A West and the nonthermal filaments. T h e morphology and polarized emission suggest the nonthermal filements were produced by a jet, like those seen in AGN, but n o point-like energy source was found on the nonthermal filaments. T h e three dimensional structure and origin of the radio arc are still behind a veil, although many investigations have been done. Millimeter astronomy shows that the Galactic center is rich in dense molecular gas. A CS line survey of the Galactic center region was made with the Nobeyama 45 meter telescope to explore the detailed structure of dense molecular gas (Tsuboi et al. 1999). O n e of the interesting features arising from the * e-mail: handaOi0a.s.u-tokyo.ae.jp,Phone: +81 422 345062, Fax: +81422 345041 * * e-mail: masOstar.le.ac.uk * * * e-mail: tsuboi O rnx.ibaraki.ac.jp

T. Handa et al.: Thermal SiO near the nonthermal filaments

218

survey is an expanding shell contacting with the nontherinal filament. Morphological evidences show the shell is interacting with the nonthermal filament (Tsuboi et al. 1997). Recent high resolution images and spectra of the Galactic center region in X-ray require uniqueness of the shell. A large scale X-ray survey with ASCA shows an extended feature located at the shell and a composite spectrum in X-ray (Sakano et al. 2001). The X-ray feature is resolved as an association of X-ray sources and filaments with Chandra observations (Yuzef-Zadeh et al. 2002). The latest analysis shows that the X-ray spectra of the galactic eastern and western edges are different; the galactic western edge shows an ion line, but the galactic eastern edge shows a featureless spectrum. Even in the galactic western edge some clumps show highly ionized iron emission hut others show neutral iron emission (Bamba et al. 2002). These results suggest that the physical properties of the molecular gas in the shell, and the distribution of shocked molecular gas, are importanl. Thermal SiO emission is thought to be a shock indicator, because a shock produces much SiO gas from dust by evaporation and spattering.

~

~.~

~

-._l_

~

Fig. 1 The shell attached to the nonthermal filaments. Thick countours are radio continuum at 43 GHz (after Sofue et al. 1986) and a gray scale image with thin contourb is in CS at ULSR = 30 - 40 !un/s (Tsuboi et a1 1999) A CS feature marked by a thick arrow IS the shell

2 Observations The observations were made in April 2002 using the Nobeyama 45-meter telescope. Both SiO ( J = 1 - 0, v = 0) and SiO ( J = 2 - I , w = 0) lines are observed simultaneously. The beamsizes are 39.5” and 17.6”. and the mainbeam efficiencies are 0.8 1 and 0.50 at 43 GHz and 86 GHz, respectively. The observed area is 4’ 5 1 5 lo’, -10‘ 5 h 5 -4‘. The grid spacing is 20” or 0.82 pc assuming a distance to the Galactic center of 8.5 kpc. The antenna pointing was checked by five-point scans of nearby SiO maser sources, OH 2.6-0.4 and VX Sgr. Typical pointing difference between two pointing check procedures was 5”. We used acousto-optical spectrometers (AOS) with 124.7 kHz resolution. Although the original spectral resolutions correspond to 0.87 kmls and 0.44 kmls at SiO ( J = 1- 0,w = 0) and SiO ( J = 2 - 1,v = 0) frequencies, respectively, we smoothed and resampled to 5 kmls velocity resolution in each line.

Asuon. Nachr./AN 324, No. S 1 (2003)

219

The line intensity was calibrated using the standard chopper wheel method. Typical rms noise levels in the atmospheric attenuation corrected antenna temperature scale are 0.1I K and 0.18 K in J = 1 - 0 and J = 2 - 1 lines, respectively. Liner or parabolic baselines were applied to all profiles between 'ULSR = - 150 krri/s and +150 km/s.

3 Results 3.1

Thermal SiO emission

Figure 2 shows the integrated intensity maps over ULSR = 15 40 km/s in SiO lines. The instensity scales are correced after main beam efficiencies. The observed half power beam width are shown in the left bottom corner of each panel. Besides the beamsize effects features in two lines are quite similar. The morphology in both lines is similar to that in CS ( J = 1 - 0) line (Figure 3). ~

,

1 (dcgrcc)

1 (degrcc)

Fig. 2 Integrated intensity distributions over ,uLsI< = 15 40 km/s in SiO J=l-O, v=0 (left panel) and SiO J=2-I, v=0 (right panel). The observed beamsize is shown at the left bottom corner in each panel. The intensity scale is corrected by the mdinbealn efficiencies. (The PDF version is in color.) ~

To avoid the beamsize difference in both transitions and pointing ambiguity, we convolve the image to be 69" beamsize by corresponding gaussians. Using the convolved images we made line intensity ratio maps of CS ( J = 2 1, u = 0 ) over CS ( J = 1 - 0) for every 5 k d s bin. The resultant channel maps show a rather uniform ratio and no significant features over 'L>LSR= 15 - 40 k d s . To check the uniformity of the line intensity ratio we made an intensity correlation diagram (Figure 4). The sample points are all pixels with 111,s~= 15 40 km/s by 5 k d s step over whole observed area by 20" grid. The majority of the pixels are along a straight line suggesting the ratio is uniform over the I - b - o space. The averaged ratio derived from the regression line is Xz-l/l-o = 0.88. Using a multilevel population calculation with the LVG approximation the ratio means typical density of the molecular gas is n(H2) = (1.0 - 1.7) x 10'ccrrir" at 40K 5 Tk 5 80K. The result is not affected by a beam filling factor if T ~ 2 0.1 ~ and~the factors ~ ~are the ~ same ~ in both , lines. The uniformity of the ratio suggests the gas density of the shell is rather uniform, although several pixels deviate from the primary component along the regression line. ~

~

T. Handa et al.: Thermal SiO near the nonthermal filaments

220

I (degree)

I (degree)

Fig. 3 Integrated intensity distributions over VLSR = 15 - 40 km/s in SiO J=I-O, v=O (left panel) and CS J=1-0 (right panel). The effective beamsize is shown at the left bottom comer in each panel.

3.2

Comparison with CS

The critical density of SiO ( 5 = 1 - 0 , v = 0) is similar to CS ( J = 1 - O), because the Einstein A coefficients of SiO ( J = 1 - 0 , =~ 0) and CS ( J = 1 - 0) are 3.0 x lop6, 1.8 x lop6, respectively. Therefore, the intensity ratio of these two lines shows the difference in abundance and indicates the shock strength. The SiO shell is seen over WLSR = 15 -40 km/s, and the shell in CS is seen over U L S R = 15 -45 km/s (Tsuboi et al. 1997). The overall features in SiO ( J = 1 - 0, II = 0) are quite similar to those seen in CS ( J = 1 - O), although the high velocity end of the integration range is different by single pixel in velocity (Figure 3). The SiO ( J = 1 - 0, v = 0) map integrated over U L S R = 15 45 km/s shows an additional component near 1 = 0.0105", b = -0.085'. This difference between CS and SiO may not be real; the integration range difference is only a single pixel in velocity and a strong emission ridge is located at the galactic northern edge of the shell at U L S R = 50 k d s both in CS and SiO. Another possiblity is that the difference is due to sampling grid between CS and SiO. Beside of this marginal difference the shell is also very similar in the position-velocity diagrams in both CS and SiO. After adjustment of beamsize to be 69" by gaussian convolution we made line intensity ratio maps of SiO ( J = 1 - 0.2) = 0) over CS ( J = 1 - 0) for every 10 k d s bin. The resultant channel maps show a rather uniform ratio and no significant features over VLSR = 15 - 35 k d s . To check the uniformity of the line intensity ratio we made an intensity correlation diagram (Figure 5). The sample points are all pixels in OLSR = 15 - 35 k d s by 10 k d s step over whole observed area by 20" grid. The majority of the pixels fall along a straight line, suggesting the ratio is uniform over the 1 - b - v space. The averaged ratio derived from the regression line is Rs,o/cs = 0.24. Using a multilevel population calculation with the LVG approximation the ratio is mainly affected by relative abundance between SiO and CS. The derived abundance is X(SiO)/X(CS) = 0.05 - 0.13 if X(SiO)/(dv/dr) = - lo-'". Comparison with typical values at L1157, a protostellar bipolar jet, ( w 3; Zhang et al. 2000) and at IC443G, an SNR shell, is ( w 0.06; Turner et al. 1992) suggests the shock strength of the shell is similar to that of a SNR shell. ~

22 I

Astron. Nachr./AN 324, No. S I (2003)

SiO( 1-0) Tinh [K]

Fig. 4 Intensity correlation diagram over U L S R = 15 - 40 km/s between SiO J=l-O, v=O and SiO J=2I , v=O. The mainbeam efficiencies are correcled in both lines. The beamsizes are adjusted to be 69’ after gaussian convolution.

CS( 1 -0)’I’mh

v( K

Lnds)

Fig. 5 Intensity correlation diagram over ULSR = 15 - 35 km/s between SiO J=1-0, v=O and CS J=l0. The mainbeam efficiencies are corrected in both lines and integration widths are the same as 10 k d s . The beamsizes are adjusted to be 69” after gaussian convolution.

Acknowledgements We thank Mr. Seiichiro Naito for his powerful assistance at the observations

References Bamba A. et al. 2002, in: New Visions of the X-ray Universe in the XMM-Newton and Chandra Era,, edited by E Jansen et al., ESA SP-488 (astro-ph/0202010) Sakano M. et al. 2002, ApJS, 138, 19 Sofue Y., Inoue M., Handa T., Tabara H., Hirabayashi H., Monmoto M., Akabane K. 1986, PASJ, 38,475 Tsuboi M., Inoue M., Handa T., Tabara H., Kato T. 1985, PASJ, 37, 359 Tsuboi M.. Ukita N., Handa T. 1997, ApJ 481,263 Tsuboi M., Handa T., Ukita N. 1999, ApJS 120, 1 Turner B. E., Chang K., Green S., Lubowich D. A. 1992, ApJ 339, I14 Yusef-Zadah F., Law C., Wardle M. 2002, ApJ, 568, L121 Zhang Q., Ho P. T. P., Wright M. C. H. 2000. AJ 119, 1345

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 223 -227 (2003) / DO1 10.1002/asna.200385033

Absorption and Emission in the Four Ground-State OH Lines Observed at 18 cm with the VLA Towards the Galactic Centre R. Karlsson*

I,

Aa. Sandqvist', L.O. Sjouwerman2.', and J.B. Whiteoak4

' Stockholm Observatory, SCFAB-Albanova,SE-106 9 1 Stockholm, Sweden ' Joint Institute for VLBI i n Europe, Postbua 2, NL-7990 AA Dwingeloo, the Netherlands NRAO Array Operations Center, P.O. Box 0, Socorro NM 87801, USA ' CSIRO. Australian Telescope National Facility, Box 76, Epping NSW 2121, Australia Key words Galactic Centre, Sy A, OH Streamer, OH-masers, molecular clouds.

The OH distribution in the Sgr A Complex has been observed in the 1612, 1665, 1667, and 1720 MHz OH transitions with the Very Large Array (VLA), in the BnA and DnC configurations. Spectral line maps have been produced with a channel velocity resolution of about 9 km sC1, and with angular resolutions of 4" x 3". and 24" x 22", respectively. Some clear results are highlighted here: i) the existence of an OH streamer inside the Circumnuclear Disk (CND) near Sgr A*, iij absorption from the CND, iiij strong absorption towards the eastern and most of the western parts of the Sgr A East shell, iv) lack of absorption towards both Sgr A West and the compact H rr-regions to the east of Sgr A East, v) a double-lobed structure of the High Negative Velocity Gas (HNVG) oriented to the northeasl and southwest of Sgr A*, and vij compact point-like maser emission in all four transitions. In particular, a 1720 MHz maser at - 1 32 km s-' in the CND as counterpart to a 1720 MHz maser at +I32 km s-' in the CND, was observed.

1 Introduction The Hydroxyl radical, OH, is relatively abundant in the molecular clouds of the Galactic Centre (GC). The four hyperfine rotational transitions at 1612, 1665, 1667 and 1720 MHz are readily observable in absorption, although the 1612 MHz is often contaminated by interference from satellite navigation systems. In Local Thermodynamic Equilibrium, LTE, for an optically thin gas the relative intensities of the four hyperfine rotational transitions are 1 :5:9: 1. Studies of OH absorption therefore provide a useful tool for investigating the kinematics and physical environment in the GC. Maser emission is also observed at all four frequencies. The OH masers at 1612, 1665, and 1667 MHz are pumped by infrared radiation and associated with evolved stars and H 11-regions. The 1720 MHz OH masers are collisionally pumped by shocks penetrating into molecular clouds. Preliminary results, mainly concerning the O H Streamer at 1667 MHz, were reported in 1987 and 1989 (Sandqvist et al. 1987, 1989),and the full results of the BnA observations can be found in Karlsson et al. (2003).

2 Observations and data reduction The main lines at 1665 and 1667 MHz were observed with the Very Large Array (VLA) in its hybrid extended (BnA) and hybrid compact (DnC) configurations. The 1612 and 1720 MHz satellite lines were observed with the BnA configuration only. The observations were made in June 1986, and in October 1989. A total bandwidth of 3.125 MHz resulted in a total velocity coverage of about 550 km s-'. A brief * Corresponding author: e-mnil: [email protected],Phone: 4 6 -

8 5537 8538, Fax: 4 6 -8 5537 8510

@ 2003 WILEY-VCH Verlag GnihH & Co. KGaA. Wcmhem

R. Karlsson et al.: 18-cm VLA observations of OH towards the GC

224

Table 1 VLA 18-cm OH observational summary: BnA configuration, 18 antennas, R-polarization, June 1986

Frequency (MHz) 1612 1665 1667 1720

HPBW (arcsecxarcsec) 4.0 x 2.7 3.9 x 2.9 4.0 x 2.8 3.7 x 2.7

PA (")

55.3 64.7 61.1 57.1

Resolution (km sP1) 9.1 8.8 8.8 8.5

tintegration

(minutes) I44 169 173 148

Table 2 VLA 18-cm observational summary: DnC configuration,27 antennas, Dual L+R-polarization,October 1989

Frequency (MHz) 1665 1667

HPBW (arcsecxarcsec) 24.0 x 22.3 30.2 x 23.7

PA (")

Resolution (km s-')

29.8 46.3

8.9 8.8

tintegration

(minutes) 144 136

summary of the observations is shown in Tables 1 and 2. NRAO's Astronomical Image Processing System, AIPS, and VLA standard calibration procedures were used for the data reduction. The 1665 and 1667 MHz lines form one continuous data set because of overlapping velocity ranges. For the 1665 and 1667 MHz main lines, the continuum map was constructed by averaging some line-free channels on each side of the lines, before subtraction from the map cube. The 1612 and 1720 MHz continuum maps were produced by subtraction of a linear fit over frequencies from the visibilities in the u , %I plane. The maximum entropy method of AIPS was used for improving the quality of the maps shown in this paper.

3 Results The main results of this study are: i) full high spatial resolution line-of-sight velocity OH absorption maps in all four ground-state OH rotational transitions (Karlsson et al. 2003), ii) the OH Streamer (Fig. 1 and Fig. 2), iii) the double-lobed structure of the HNVG (Fig. 3), and iv) 10 new point-like OH maser sources (Table 3). The OH Streamer seems, in projection, to stretch inside the CND between the southwestern part of the CND and Sgr A*. Its length is about 2 pc and its density and velocity dispersion are highest in the "head', i.e. close to Sgr A*. The streamer is clearly detected between +76.5 km s-l and +23.6 !an sC1 in the 1612, 1665, and 1667 MHz lines. The head can be traced to a velocity of -20.4 km s-'. The OH Streamer is not detected at 1720 MHz. As the line-of-sight velocity drops, the streamer seems to shorten (Fig. 1). In the vicinity of Sgr A*, the head seems to turn in an anti-clockwise direction as the velocity decreases from +41.2 km s-l to +23.6 km s-'. At even smaller velocities, the streamer finally seems to coincide with Sgr A*. The velocity of the tail of the OH Streamer differs by more than 100 km s-' from the velocity of the CND at the position where the streamer appears to merge with the CND. This may imply that the OH Streamer and the CND are not located in the same plane. Figure 2 shows -TIIT, for the 1667 MHz line at +58.8 km sP1, where the streamer is seen at its full length. The general structure of the OH Streamer is retained in the -Tl/Tc map, although a quantitative analysis has to include a more detailed model of the background continuum (Karlsson et al., in preparation). Strong absorption is also observed towards both the eastern part, and parts of the western areas of the Sgr A East shell (Fig. 1 and Karlsson et al. 2003). There is a lack of absorption towards the compact H IIregions to the east of Sgr A East, and towards Sgr A West. Extended OH absorption is, however, detected

Astron. Nachr./AN 324, No. SI (2003)

22s

Fig. 1 1667 MHz OH absorption towards the Sgr A Complex. The contour levels are IS, 93 and 500 mJy/beam, and outline Sgr A East, the H ~iregionsfurther to the east, Sgr A West and Sgr A*, respectively. The velocities are indicated in the upper right corner. The OH absorption is shown i n grey-scale, and in colour in the electronic version of this paper. The OH Streamer is seen at full length between +76.5 km s-’ and +SO km s-’, and is very clear in colour.

in the region of Sgr A West with a maximum velocity around - 170 kni spl. Two lobes are observed, both in the line and -TIT, maps, to the northcast and southwest of Sgr A* (Fig. 3). This structure is attributed to the HNVG.

In the 1665 and 1667 MHz DnC observations, extended large-scale OH absorption is also observed towards the Radio Arc, the Bridge and the Sgr A Complex. Examples of absorption towards the “Pistol” and the “Bridge” are seen in Figures 4 and 5, at the velocities of -73.2 km spl and -99.6 km s-’ . The HNVG, the Expanding Molecular Ring (EMR), and the CND can all be identified in the full set of maps.

R. Karlsson et al.: 18-cm VLA observations of OH towards the GC

226

0

200

-28 58

-28

0

34

174236

Fig. 2 The 1667 MHz OH line-to-continuum distribution (-Tl/Tc),at +58.8 km SKThe I. OH-streamer is highlighteded by contours.

00

32 30 20 26 RIGHT ASCENSION(61950)

24

Fig. 3 The 1667 MHz OH absorption at - 170 km s-', showing the double-lobed, northwest to southeast, structure of the HNVG.The contours mark the same continuum components as in Fig. 1.

10

05

-28 40

45

-

50

Y)

:

I

55

c -I

vy

-2900

05

I1 30

00

41 30

Fig. 4 OH large-scale absorption at 1667 MHz in the Sgr A Complex at -73.2 km sK1, Observations made with the DnC configuration.

174400

4

Fig. 5 OH large-scale absorption at 1667 MHz in the Sgr A Complex at -99.6 km sK1. Observations made with the DnC configuration.

Emission is detected from OH masers in all four transitions. Table 3 lists the new detections found in the 1986 BnA data set, i.e. masers not previously published using other independent observations (1612 MHz: Lindqvist et al. 1992 and Sjouwerman et al. 1998, and 1720 MHz: Yusef-Zadeh et al. 1996).

Of the ten new masers, the two I665 MHz and four 1667 MHz main line OH masers are associated with four known OWIR stars. The two 1665 MHz masers appear along with 1667 MHz masers in the O W R stars with the brightest 1612 MHz masers. Together with the two other 1667 MHz masers, these OWIR

Astron. Nachr./AN 324, No. S 1 (2003)

227

Table 3 New point-like OH masers found in the 1986 VLA-BnA data set

Name

Position in B1950

Position in 52000

'usystem

Flux

(Dec.)

(R.A.)

(Dec.)

(kms-')

(mJy)

-29 03 52.9 1 -29 00 36.25

17 45 40.472 17 45 46.440

-29 05 02.27 -29 01 45.16

-29 -91

30 35

-28 -29 -29 -28

58 35.21 03 53.40 00 36.55 58 53.57

17 45 38.618 17 45 40.442 17 45 46.444 17 45 54.367

-28 -29 -29 -29

59 44.75 05 02.80 01 45.50 00 01.94

t63 -29 -82 -3

37 103 106 29

-29 -28 -28 -28

00 05.83 58 09.87 58 28.00 58 19.67

17 45 38.764 17 45 40.403 17 45 45.586 I7 45 50. 165

-29 -28 -28 -28

01 15.27

-132 +30 +39 +56

70 37 75 93

(R.A.)

New OH 1665 MHz masers OH359.880-0.087 OH359.938-0.077

17 42 29.628 17 42 35.678

New OH 1667 MHz masers 08359.952-0.036 OH359.880-0.087 08359.938-0.077 0H359.977-0.087

17 42 27.907 17 42 29.597 17 42 35.682 17 42 43.647

New OH 1720 MHz masers 0H359.930-0.049 0H359.960-0.037 OH359.966-0.056 OH359.977-0.069

17 42 28.017 17 42 29.705 17 42 34.878 17 42 39.462

59 19.16 59 37.03 59 28.20

stars are among the reddest OWIR sources with the longest periods in the GC and probably in transition to become planetary nebulae. Of the four new detections at 1720 MHz that are associated with shocks in the GC, the most interesting one is 0H359.930-0.049 at a line-of-sight velocity of - 132 km s-l. In contrast to the 1720 MHz OH masers with line-of-sight velocities between +30 to +70 km spl which are associated with the Sgr A East shell, 08359.930-0.049 and 0H359.955-0.042 (at +I32 km SKI, Yusef-Zadeh et al. 1996) originate in the CND. Their velocity and symmetry in the CND yield an enclosed mass of at least 7 . 5 lo6 ~ Mu within 47.4" ( I 3 4 pc). Granting up to 3.7 x 1 O6 MD for the black hole at the position of Sgr A* (Schodel et al. 2002), this would then imply that at least 50% of the total mass within the CND is contained in the enclosed stellar cluster and molecular cloud complexes.

References Karlsson R., Sjouwerman L.O., Sandqvist Aa., Whiteoak J.B. 2003, A&A, 403, 1011 Lindqvist M., Winnberg A,, Hahing H.J., Matthews H.E. 1992, A&AS 92,43 Sandqvist Aa., Karlsson R., Whiteoak J.B., Gardner F.F. 1987, in AIP Conf. Proc. 155, The Galactic Center, ed. D.C. Backer (AIP, New York), 95 Sandqvist Aa., Karlsson R., Whiteoak J.B. 1989, in IAU Symp. 136, the Center of thc Galaxy, ed. M. Moms (Kluwer, Dordrecht), 421 Schodel R., Ott T., Genzel R., et al. 2002, Nature 419, 694 Sjouwerman L.O., van Langevelde H.J., Winnberg A,, Habing H.J. 1998, A&AS 128, 35 Yusef-Zadeh F., Roberts D.A., Goss W.M., Frail D.A., Green A.J. 1996, ApJ 466, L25

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S 1.229 -234 (2003)/ DO1 10.1002/asna.200385I 14

Constraints on distances to Galactic Centre non-thermal filaments from HI absorption Subhashis Roy *

’

’ National Centre for Radio Astrophysics (TIFR), Pune University Campus, Post Bag No.3, Ganeshkhind, Pune 41 I 007, India. Abstract. We have studied HI absorption towards three non-thermal filaments (NTFs)Sgr C, G359.54+0.18 and G359.79+0.17 using the Giant Metrewave Radio Telescope (GMRT). Our study, for the first time, constrains the distance of the Sgr C NTF and the HI1 region seen associated with the NTF in the sky plane, to within a few hundred parsecs from the Galactic Centre (GC). A molecular cloud with a velocity of - 100 km sC1 appears to be associated with the central part of the Sgr C NTF. G359.54+0.18 shows weak HI absorption (4 CJ detection) at a velocity of - 140 !un sC1, which is the velocity of a known dense molecular cloud seen towards the NTF. This cloud is expected to be located within 200 pc from the GC and thereby provides a lower limit to the distance. The upper limit to the distance of this NTF from the Sun is 10.5 kpc. The distance to the NTF G359.79+0.17 is between 5.1 and 10.5 kpc from the Sun.

-

1 Introduction The long narrow non-thermal filaments (NTFs) observed in high resolution radio-continuum maps are 2” region of our Galaxy. These structures are less than unique features seen towards only the central 1 pc in width, but extend up to 30 pc in length. With the exception of the NTF named the Pelican (Lang et al. 1999), which is nearly parallel to the galactic plane, all other NTFs are oriented perpendicular to the Galactic plane to within 20” (Morris & Serabyn 1996, and references therein). Since these NTFs remain straight despite interaction with nearby molecular clouds, it is believed that the molecular clouds and the NTFs are in pressure equilibrium, which indicates a magnetic field strength of a few milliGauss inside the NTFs (Yusef-Zadeh & Morris 1987). Magnetic fields of comparable strengths are thought to be present in the central molecular zone (CMZ). Before any attempt is made to relate the magnetic field in the NTFs with the processes occurring in the GC, it is necessary to establish that these NTFs arc actually located in the GC region and are not chance superpositions of foreground or background objects (Lasenby et al. 1989). HI absorption towards the GC ‘Radio-arc’ (Lasenby et al. 1989) and the ‘Snake’ (Uchida et al. 1992) have indicated that they are located close to the GC, but the distances to the remaining NTFs are not constrained. Here, we present new HI absorption measurements towards three NTFs, Sgr C, G359.54+0.18 and G359.79+0.17 made with the Giant Metrewave Radio Telescope (GMRT). These observations not only constrain the distances of these objects, but also test the association of some of the above mentioned clouds with the corresponding NTFs. From single dish HI emission observations towards the NTFs under study (e.g. Cohen & Davies 1979), several large scale features are identified. Near the Galactic longitude of 359.5”, two high velocity HI emission features known as the ‘Nuclear disk’ (Rougoor & Oort 1960) and the ‘Molecular ring’ (Scoville 1972) have been found. The ‘Nuclear disk’ shows high negative velocities ranging from zz - 160 to -200 Both these features are believed km sC1,whereas, the ‘Molecular ring’ has a velocity of z - 135 km SKI.

-

* Subhashis Roy: e-rnail: [email protected]

@ 2003 WILEY-VCH Verlag GmbH C Co KGaA. Weinhem

R. Roy: HI absorption line study o f three non-thermal filaments

230

to be nearer than the GC and located at a distance of Pew hundred parsecs from it (Cohen & Davies 1979). The emission from the ‘3 kpc arm’ (Rougoor & Oort 1960) located at a distance of zz 5.1 kpc from the Sun is identified at a velocity near -53 km spl. At positive velocities, emission near 135 km spl is seen due to the HI features ‘XVI’ and ‘I’ (Cohen & Davies 1979), both of which are thought to be located behind the GC. While the feature ‘XVI’ is likely to be located within a few hundred parsecs from the GC (Cohen & Davies 1979), the feature ‘I’ is thought to be 2 kpc behind the GC (Cohen 1975). Absorption by these HI features will be used to constrain the distances to the NTFs.

2 Results In this section, we present the absorption spectra towards the target sources and identify the velocity of the HI absorption features. In all the spectra, unless stated otherwise, the X-axis represents the velocity in where, I is the observed flux density of the background km s-l and Y-axis represents the transmission (I&), source at the given frequency and 10 is the actual flux density of the source. All the velocities are expressed with respect to the Local Standard of Rest and the GC is assumed to be at a distance of 8.5 kpc.

Sgr C HI1 reeon -29

c

‘I

Nuclear disk

-65 kmls cloud

7

0

1-

e

0 ,

-200

-150

-100

-50

,

,

,Y,

, ,

0

, 50

,

,

,

,

, , 100

, ,

, 150

, ,

,

,j 200

Velocity (km sl) 1

Fig. 1 Continuum map of the Sgr C filament after high pass filtering. The Sgr C HI1 region is resolved out. The image has a resolution of 7 x 6 arcsec2, along PA=56”. RMS noise in the map is 1.6 mJy

Fig. 2 HI absorption spectrum towards the Sgr C HI1 region and the central bright part (marked ’A’ in Fig. 1) of the Sgr C NTF. The bandwidth is 2 MHz. RMS noise in the two spectra are 0.02 I and 0.034 respectively.

beam-’.

2.1

SgrC

The absorption spectra towards various parts of the object (Fig. 1) are shown in Fig. 2. Fig. 2 shows the absorption spectrum towards part ‘ A of the NTF. It shows several absorption features at negative velocities in addition to the absorption feature at 0 km s-l (line-width FZ 30 km s-’). A strong absorption feature near -54 km s-l is seen due to HI absorption by the ‘3 kpc arm’. A broad absorption feature is seen between -100 km 8 - l and -200 km s-’, with an almost linear decrease in optical depth from FZ 0.5, at -100 km s-’, to FZ 0.0 at -200 km spl. The absorption spectrum towards the Sgr C HI1 region with a resolution of 3.3 km spl show similar absorption spectrum as the NTF with a few differences, which we note here. Towards the HI1 region, the broad absorption feature seen between -100 and -200 krn s-l

Astron. Nachr./AN 324, No. SI (2003)

23 I

shows at least three main components centred at -118 km s-', -138 km S K ' and -175 km s-' with optical depth of =0.5,0.3 and 0.2 respectively. The absorption depth at these velocities are similar to what is seen towards part 'A' of the NTF.

2.2 NTF G359.79+0.17, G359.87+0. I 8 and G3S9.54+0.18 The continuum image at 20 cm of the field of NTF G359.79+0.17 and G359.54+0.18 is shown in Fig. 3 and Fig. 5 respectively. The absorption spectrum integrated over the NTF is plotted in Fig. 4 and Fig. 6 respectively. At negative velocities, absorption features can be seen near -26 km s-', and a weaker feature at -58 km s-', which coincides with the line of sight velocity of the ' 3 kpc arm'. Fig. 4 also shows the absorption spectrum towards the extragalactic source G359.87+0.18. Lazio et al. (1999) have observed HI absorption against G359.87+0.18, and the aforementioned features match with their spectrum. However, the present observations have a wider velocity coverage than Lazio et al. (1999) and we detect an additional absorption feature at + I40 km sStaguhn et al. (1998) have found a dense molecular cloud at - 140 km SK' near the bent portion of the NTF (position of the molecular cloud is denoted by 'E' in Fig. 5 ) . HI spectrum taken towards this region of the NTF (denoted by 'F' in Fig. 5 ) shows absorption (Fig. 6) at this velocity (4a detection). 50

0

100

28 57

C359.87tO.18

'\I

3-kpr arm

58

03 30 25 20 RIGHT ASCENSION (52000) Grey scale flux range= -10.9 141.7 MilliJYBEAM Con1 peak flux = 1.4175E-01 JYncr.' II"L""7 LeYP = &OWEi-03 ' (-2, -1,1,2,4,6,8,10,12, 16.20.24.32, 40.48, €4,W, 96,128) 35

17 44 40

15

P

10

Feature-1

'

Fig. 4 HI absorption spectrum towards the extragalactic source G359.87+0.18 and the NTF G359.79+0.17. The bandwidth is 4 MHz. RMS noise in the two spectra are 0.028 and 0.026 respectively.

Fig. 3 Continuum image of the NTF G359.79+0.17 ai 1.4 GHz with a resolution of 9.7 x 6.7 arcsec*, along PA=79". The rms noise is 3.0 mJy beam-'.

3 Discussion 3.1

Identification of HI features & constraints on the distances to the NTFs

Identifications of HI absorption feature is performed by comparison with features of known velocities. Absorption indicates that the continuum source is located on the far side of the HI cloud and thereby provides a constraint on the distance to the continuum source. The velocities and the distances of the known HI emission features towards the three NTFs studied here have been discussed in 4 1, which will be used to constrain the distances to the NTFs.

R. Roy: HI absorption line study of three non-thermal filaments

232

4w

-300

-200

-100

100

0

200

300

100

Velocify

Fig. 6 HI absorption spectrum integrated over the

with a resolution of 9.6 x 6.4 arcsec’, along P k 7 9 ” . The rms noise is 1.7 m l y beam-’

of the NTF where it bends (region ‘F‘in Fig. 5). The bandwidth is 4 MHz. RMS noise in the two spectra are 0.025 and 0.1 respectively. I

3.4 kpc

I1

-2 kpc

0 Sun

VeloLlly

Fig. 7 CO emission spectra towards the central part (top panel) (marked ‘A‘ in Fig. I ) and the eastern part of the NTF (middle), along with the spectrum taken towards the Sgr C HI1 region (bottom) (Data courtesy Oka et al. (1998)).

L

GC region -400 pc

8.5 kpc

Fig. 8 Schematic diagram of the Sgr C complex with HI absorbing clouds (not to scale) as seen from bottom of the Galaxy

3.1.1 SgrC Due to absorption by the line of sight HI gas and velocity crowding near 1=0”, strong absorption is observed near 0 km s-l in all spectra towards the Sgr C NTF and the HI1 region seen associated in the sky plane. Absorption by the ‘3 kpc arm’ is observed at -54 km sP1 towards the Sgr C N l T and the HI1 region (Fig. 2). The broad absorption feature (Fig. 2 ) between - I00 and -200 km s-’ is likely to be caused by several absorption features, whose line-widths are broader than their separation. The HI absorption near - 100 km sP1 is identified in CO Oka et al. (1998) emission as shown in Fig. 7. HI absorption at this velocity is caused by a cloud of size -5’, and Roy (2003) has shown it to be associated with the NTF. Absorption near - 138 km s-’ is likely to be caused by the HI associated with the ‘Molecular ring’, which has a line-width of N 40 km s-l in emission (Cohen & Davies 1979). Detection of absorption beyond - 160 km s-’ indicates absorption by the ‘Nuclear disk’ (Rougoor & Oort 1960). As ‘Molecular ring’ and the ‘Nuclear disk’ are located within 200 pc from the GC, this provides a lower limit of -8.3 kpc to these objects. We note that despite the emission feature seen in the CO map

-

Astron. Nachr./AN 324, No. S I (2003)

233

a1 -140 km s-’ (Fig. 7), no corresponding HI absorption could be detected towards the Sgr C region. The CO emission from the molecular cloud at -140 km spl is likely to be associated with the HI features ‘XVI’ and ‘I’ and both these features are located at the far side of the GC. Absence of any absorption by the HI associated with these structures indicate that the Sgr C NTF and the HI1 region are located within a few hundred parsecs at the far side of the GC, which provides an upper limit to their distances.

3.1.2 NTF G359.79+0.17, G359.87+0.18 & G359.54+0.18 No HI absorption at high positive velocity is detected towards the NTF G359.79+0.17 or G359.54+0.18. However, CO emission has been detected near +I40 km s-l towards G359.79+0.17, G359.54+0.18 and the extragalactic source G359.87+0.18, which indicates that there is no hole in feature ‘I’ (or perhaps in feature ‘XVI’) along these directions. Consequently, the upper limit to the distance of the NTF G359.79+0.17 and G359.54+0.18 is = 10.5 kpc. The presence of absorption in the spectra of the NTFs up to a negative velocity of -58 km s-l suggests absorption by the ‘3 kpc arm’ and consequently, the lower limit to their distance is N 5.1 kpc from the Sun. HI absorption near +140 km spl is only observed (Fig. 4) towards the extragalactic source G359.87+0.18. We note that both the HI emission features ‘I’ and ‘XVI’ have velocities close to 140 km s p l at this longitude and are located farther away from the GC. Therefore, the absorption near 140 km indicates that it is caused by either one or a combination of both these HI features, and is consistent with its location outside of the Galaxy. We could detect weak HI absorption at 40 level at - 140 km s-l (Fig. 6) toward part-F of the NTF shown in Fig. 5. At the same velocity, a dense molecular cloud was detected by (Staguhn et al. 1998). HI absorption by the dense molecular cloud suggests that this NTF is either embedded or located behind the CMZ found within 200 pc from the GC.

4 Conclusions HI absorption studies of three NTFs known as the Sgr C, G359.54+0.18 and G359.79+0.17 using the GMRT have yielded the following results: (a) For the first time, the Sgr C NTF and the HI1 region are shown to be located within a few hundred parsecs from the GC. (b) A molecular cloud with a velocity of -100 km s-’ appears to be associated with the central part of the Sgr C NTF. (c) HI ab5orption by the ‘3 kpc arm’ is detected against all the three NTFs, which indicates that the NTF G359.54+0.18 and G359.79+0.17 are located at a minimum distance of 5 . I kpc from the Sun. (e) Weak HI absorption (4 (7 level) at - 140 km s-l suggests that the NTF G359.54+0.18 is located at a minimum distance of = 8.5 kpc from us. (f) The maximum distance of’ the NTF G359.54+0.18 and G359.79+0.17 are estimated to be 10.5 kpc from the Sun. The present study extends the number of NTFs, which have been found to be located near the GC region to five. With most of the known NTFs now being shown near the GC, there remains little doubt that phenomena related to the central region of the Galaxy are responsible for the creation and maintenance of the NTFs.

5

Acknowledgements:

It is a pleasure to thank A. Pramesh Rao, with whom I have discussed several aspects of this work at various stages. I thank the staff of the GMRT that made these observations possible. GMRT is run by the National Centre for Radio Astrophysics of the Tata Institute of Fundamental Research.

References Cohen, R. J. 1975. MNRAS, 171,659 Cohen, R. J. & Davies, R. D. 1979, MNRAS, 186,453

234

R. Rov: HI ahsorotion line studv of three non-thermal filaments

Lang, C. C., Anantharamaiah, K. R., Kassim, N. E., & Lazio, T. J. W. 1999, ApJL, 521, L41 Lasenhy, J., Lasenhy. A. N., & Yusef-Zadeh, F. 1989, ApJ, 343, 177 Lazio, T.J. W., Anantharamaiah, K. R., Goss, W. M., Kassim, N. E., & Cordes, J. M. 1999, ApJ, 515, 196 Moms, M. & Serabyn, E. 1996, ARA&A, 34,645 Oka, T., Hasegawa, T., Sato, E, Tsuboi, M., & Miyazaki, A. 1998, ApJS, 118,455 Rougoor, G. W. 1964, Bull. Astron. Imt. Netherlands, 17, 381 Rougoor, G. W. & Oort, J. H. 1960, Proc.Nat.Acad.Sci, 46, 1 Roy, S. 2003, A&A, 403,917 Scoville, N. 2. 1972, ApJL, 175, L127 Staguhn, J., Stutzki, J., Uchida, K. I., & Yusef-Zadeh, F. 1998, A&A, 336,290 Tsuhoi, M., Handa, T., & Ukita, N. 1999, ApJS, 120, 1 Uchida, K., Morris, M., & Yusef-Zadeh, F. 1992, AJ, 104, 1533 Yusef-Zadeh, F. & Moms, M. 1987, ApJ, 322, 721

Astron. Nachr./AN 324. No. S1.235 -239 (20031 / DO1 10.1002/asna.200385034

Discovery of a non-thermal X-ray filament in the Galactic Centre Masaaki Sakano*

',', Robert S. Warwick**' ,and Anne Decourchelle****

' Department of Physics and Astronomy, University of Leicester, Leicester LEI 7RH, UK

' CEAIDSMIDAPNIA,Service d'Astrophysique, C.E. Saclay, 91 191 Gif-sur-Yvette Cedex, France Japan Society for the Promotion of Science (JSPS)

Key words The Galactic Centre, Individual: Sgr A East, Individual: XMM J174540-2904.5, X-ray We report the discovery of an X-ray filament, XMM J174540-2904.5, in the Galactic Centre region. Images from Chandru and XMM-Newton show the X-ray source is extended and coincides with a non-thermal radio structure of somewhat larger extent. The X-ray spectrum is clearly not thermal in nature, and is well approximated as a heavily absorbed power-law continuum with a photon index % 2. Combining the radio and X-ray spectra, we concluded that the emission in both wavebands probably originates in the synchrotron process. We discuss some possible origins for this peculiar non-thermal structure.

1 Introduction Many peculiar non-thermal filamentary structures have been reported on the basis of radio observations of the Galactic Centre region. Some of the filaments are relatively compact whereas others subtend large angular scales, e.g. the Galactic Centre Lobe (e.g. Mezger et al. 1996) and the jet (Sofue et al. 1989). The nature and origin of the latter features, which appear to be unique to the Galactic Centre, remains a mystery. Morris ( 1 994) summarised the properties of several individual radio filaments seen in the Galactic Centre region. Interestingly, most of them have a similar nature; they are long, thin and, although often exhibiting some curvature, are generally aligned perpendicular to the Galactic Plane. The radio emission froin these filaments is very likely to be due to the synchrotron process in which relativistic electrons spiral in a local magnetic field of -I mG. Polarisation observations identified the direction of the magnetic field to be along the filaments, i.e. orientated perpendicular to the Galactic Plane. However, recent deep VLA observations have revealed that there are many weak radio structures in the Galactic Centre region, which are more closely aligned with the Galactic Plane (e.g. Novak et al. 2003). Some of the radio non-thermal structures seen in the Galactic Centre region can probably be identified with supernova remnants, H 11 regions, or possibly pulsar nebulae. For example, Lazendic et al. (2002) detected OH maser sources from some parts of the shell of SNR G 359.1-0.5, and estimated the magnetic field to be 0.5 mG. The shell of G 359.1-0.5 is known to be interacting with the surrounding molecular cloud. Thus, the shock in the dense cloud might create the non-thermal radio structure with a strong magnetic field. Some other filaments also show morphological evidence for interactions with molecular clouds; the concentration of the matter might again enhance the particle acceleration. The new generation of X-ray telescopes carried by missions such as Chandra and XMM-Newton enable us to resolve discrete extended X-ray sources froin point sources and the overall diffuse emission in the * Corresponding author: e-mail: rnasQstar.le.ac.uk, Phone: +44 116252 3510, Fax: 4 4 116252 331 1

* * e-mail: rswOstar.le.ac.uk, Phone: +44 1162523517, Fax: +44 116252331 I '** e-mail: adecourchelleQcea.fr, Phone: +33 I690843 84. Fax: +33 169086577,

@ 2003 WILEY-VCH Verlag GmhH C Co KGaA. Wemhclm

M. Sakano et al.: A non-thermal X-ray filament in the GC

236

Galactic Centre region. Koyama (2001) has reported the discovery of a candidate X-ray filament in this region. Bamba et al. (2002) has further found some X-ray filamentary structures in the Radio Arc region. The X-ray spectra of these X-ray filaments seem featureless consistent with a non-thermal origin. Most of these X-ray structures, however, have no clear correlation with the radio filaments. The only exception is NTF G359.54+0.18, for which there is good similarity between the X-ray and radio morphology (Wang et al. 2002). The X-ray spectrum of NTF G359.54+0. I8 was, however, not determined with any precision due to poor counts statistics. In this paper, we report the XMM-Newton and Chandra discovery of a non-thermal X-ray filament, XMM 5174540-2904.5, in the Galactic Centre region, which has an apparent radio counterpart (the Sgr A-E 'wisp'=lLC 359.888-0.086=G359.88-0.07). This source is located 4 arcmin to the Galactic west of Sgr A* (see Fig. 1 Sakano et al. 2003a). Arguably, the present discovery provides the clearest example of an X-rayhadio filament in this region, both from a morphological and spectral point of view.

2 X-ray Observations We have carried out a survey with XMM-Newton concentrating on the region along the Galactic Plane within +lo of the Galactic Centre. Further details of this programme, including the mosaiced image, are presented in Warwick (2002, 2003) and Sakano et al. (2003~).XMM 5174540-2904.5 was discovered in one of the observations, designated GC6, see Sakano et a]. (2003a). We have also employed data from a Chandra observation of this region made on 8 July 2000. Further information relating to this latter observation is given in Maeda et al. (2002) and Sakano et al. (2003a).

3 X-ray and radio images

..................... . *

i

/ /

/;,

I-: $

/ /

I

\ - + - - - - - / ..........................................

-2gd5" .......

Fig. 1 The Chandra image of XMM 5174540-2904.5 in the 2-8 keV band, overlaid with the 2 cm radio continuum measured by the VLA (Ho et al. 1985). The coordinates are in the 52000. This figure is taken from Sakano et al. (2003a).

Fig. 1 shows the X-ray image of XMM 5174540-2904.5 overlaid with the corresponding radio 2 cm contour. The X-ray and radio images show that the source is extended in both waveband bands and that the morphologies are well correlated in the northern region of the source (the part nearer to the Galactic

Astron. Nachr./AN 324, No. S 1 (2003)

237

Plane). This good spatial correlation strongly suggests that both the X-ray and radio cmissions have the same origin. The northern part of the source is aligned nearly perpendicular to the Galactic Plane; this nature is often seen in the radio filaments i n the Galactic Centre region, as noted earlier. On the other hand, the radio source is more extended than its X-ray counterpart and shows significant curvature at its southern extent. The radio source, which coincides with the X-ray source, was first detected in 2 and 6 cm by Ho et al. ( I 985) and designated as Sgr A-E 'wisp' based on its characteristic morphology. Lazio & Cordes ( 1 998) have later catalogued this source as ILC 359.888-0.086 based on their 1281 MHz (-20 cm) observation. They noted that this source has an angular extent of 47 arcsec, which is consistent with 2 cm and 6 cm measurements of Ho et al. (1985). More recently, Lang et al. (1999) confirmed this result with a 20 cin observation; in this case the source was designated as G359.88-0.07.

4 X-ray and radio spectra We extracted XMMIMOS and pn, and Chunclr~i/ACISspectra from an elliptical source region centred on the X-ray filament. The associated background spectra were taken from a near-sky region (see Sakano et al. 2003a for more details). If we use a thin thermal model to model the observed background-subtracted spectra, we find that a temperature in excess 40 keV is required; accordingly we rejected the thermal model. In contrast, the spectral fitting the data with a power-law model gives acceptable results. The bestfitting parameter values were: photon index r = 2.0'::; and hydrogen column density N f i = 3821, x 10""H cmp2. The observed 2-10 keV flux is 4 ~ 1 0 ~ ' ~ es-'r gcmp2, which converts to a 2-10 keV unabsorbed luminosity of 1 x 1034ergs-'. Here all the uncertainties quoted are at a 90% confidence level. Confidence contours

b)

-t Fig. 2 (a) The X-ray spectra with the best-fitting power-law model; circle (black), cross (red). star (green), and tliangle (blue) represent XMMlpn, MOS1, 2 and ChundrulACIS data, respectively. This figure is from Sakano et al. (2003a). (b) The confidence contour between the photon index (r)and the column density ( N H for ) the 6876, 90% and 99% confidence limits.

Fig. 2 shows the observed spectra together with the best-fitting power-law model and the corresponding confidence contour for column density versus photon index. The column density is found to be very large implying that the X-ray source is located in or behind dense molecular clouds. The flat power-law spectrum with heavy absorption makes this a very distinct hard source in the X-ray hardness map of the Galactic Centre region (see Fig.1 in Sakano et al. 2003b, 2003d). Combining the radio measurements at I5 GHz (2 cm), 5 GHz (6 cm) and 1.3 GHz (Ho et al. 1985; Lazio & Cordes 1998), we find all three data points to be well fitted with a power-law model with spectral index cy x 0.4 ( I E K E-"; i.e. photon index r of 1.4). Fig. 3 summarises the broad-band spectrum from the radio to the X-ray bands of the filament. We found that the measured X-ray flux sits a few decades

M. Sakano et al.: A non-thermal X-ray filament in the GC

238

below the extrapolation of the radio power-law into the X-ray band, but that the X-ray spectrum may well fit smoothly to the radio spectrum via a spectral break somewhere in the IR to EUV range. Broad band spectrum (Radio - X-ray)

*

s[

'"""'I

'"'"'~1

"" 'i'''''l I '""'''1 ~

'"'3"'l

' 1'T' """1 '

-'I

'

y

aradio= 0.4

--.

s lil

&J,z-~~Y+z-~TL~

'Y

Y A Y O ~~ A'''Y~L

'''~'ll' ''''yo' Energy ( e V )

,~!5

Fig. 3 The broad band energy spectrum of XMM 5174540-2904.5 from the radio to X-ray hands. The three crosses represent the radio data of 15 GHz (2 cm), 5 GHz (6 cm) (Ho et al. 1985) and 1.3 GHz (-20 cm) (Lazio & Cordes 1998), whereas the polygon represents the X-ray spectrum obtained in this work. The dashed line shows the extrapolation o f a power-law model with spectral index 0.40 (IE 0: E - Q ) which provides a best-fit to the radio data.

5 Discussion The flat X-ray spectrum of XMM J174540-2904.5 implies that the emission is non-thermal in origin. In fact, the thermal model was rejected by the spectral fitting. The broad-band spectrum from the radio to X-ray bands shows the spectral shape which can be smoothly connected somewhere between the two bands. Non-thermal X-ray emission has been detected from a number of Galactic SNRs and pulsar wind nebulae (e.g. Koyama et al. 1995). For those sources, the radio spectrum is also non-thermal and smoothly connected to the X-ray spectrum, just as in our case. The emission is most probably due to the synchrotron radiation of relativistic electrons with energy of up to -100 TeV (e.g. Reynolds & Keohane 1999). The lifetime ( T ) of synchrotron-emitting electrons is inversely proportional to the square of the magnetic field (7 c( W ' ) .In the case of the Sgr A-E 'wisp' (XMM 5174540-2904.5), Ho et al. (1985) estimated the magnetic field to be 0.3 mG, based on an equipartition argument. The lifetime of the 20 TeV electrons emitting hard X-rays through synchrotron radiation is then -7 yr. This lifetime is comparable to the observed spatial extent of the filament which is -2 light year in the X-ray band. Consequently, either the continuous injection of high-energy electrons into the source is required or, more likely, the on-going acceleration of electrons is occurring in the source itself. In fact, the position of this source coincides with the peak of the molecular cloud M-0.13-0.08 (the "20 km s-'" cloud; e.g. Mezger et al. 1986), which is only several tens of parsecs away from Sgr A' (Zylka, Mezger & Wink 1990). This together with the heavy absorption in the X-ray band strongly suggests that XMM J174540-2904.5 is embedded in, or beyond, the cloud. It is plausible that these high-density

-

Astron. Nachr./AN 324, No. S1 (2003)

239

conditions enhance the particle acceleration hence explaining why this is a bright non-thermal source right up to the X-ray band. Coil & Ho (2000) argue from their radio observations that the radio ‘wisp’ (i.e. XMM J 174540-2904.5) may be part of a SNR (designatedas G 359.92-0.09)which lies to the side of and behind the M-0.13-0.08 cloud. If this is the case, X M M 5174540-2904.5 could be a shock front of the SNR. Interestingly, a recent Chandru deep look shows an X-ray shell-like structure, which may possibly include this XMM 5174540-2904.5 (Park et al. 2003). The X-ray colour of the opposite side of the possible X-ray shell is, however, very different from XMM J 174540-2904.5. Another bright radio structure Sgr A-D designated by Ho et al. ( 1985), which is regarded as a part of their proposed SNR, has no apparent X-ray counterpart, with only XMM J 174540-2904.5 being bright in the hard X-ray band. Consequently, the interpretation of the X-ray filament in terms of a SNR shell remains uncertain. An alternative is that we are dealing with an isolated X - r a y h d i o filament. The alignment of the X-ray source, perpendicular to the Galactic Plane, is then consistent with the phenomenology of other isolated radio non-thermal filament in the Galactic Centre region. Or this source could possibly be an extragalactic background object, namely a one-sided jet emanating from a QSO, although the lack of a central point source corresponding to the QSO core somewhat weakens this argument. In any case, this is the first clear detection of an X-ray filament which has a radio counterpart and unequivocally has a non-thermal X-ray spectrum. Further detailed investigation in the X-rayhadio bands, particularly via deep high resolution imaging observations, should help reveal the true nature of this peculiar source. Acknowledgements MS acknowledges the financial support from JSPS.

References Bamba, A., Murakami. H., Senda, A., Takagi, S., Yokogawa, J., & Koyama, K. 2002, Proc. New Visions of the X-ray Universe in the XMM-Newton and Chandra era, in press (astro-ph/0202010) Coil, A. L., & Ho, P. T. P. 2000, ApJ, 533, 245 Ho, P. T. P., Jackson, J. M., Barrett, A. H., & Armstrong, J. T. 1985, ApJ, 288, 575 Koyama, K. 2001, in H. Inoue & H. Kunieda, ed., ASP Conf. Ser. Vol. 251, New Century of X-ray Astronomy, Astron. SOC.Pac., San Francisco, p.50 Koyama, K., Petre, R., Gotthelf. E. V., Hwang, U., Matsurd, M., Ozaki, M., & Holt, S. S. 1995, Nature, 378, 255 Lang, C. C., Moms, M., & Echevarria, L. 1999, ApJ, 526,727 Lazendic, J. S., et al. 2002, MNRAS, 331, 537L Lazio, T. J. W., & Cordes, J. M. 1998, ApJS. 118, 201 Maeda, Y., et al. 2002, ApJ, 570,671 Mezger, P. G., Chini, R., Kreysa, E., & GemUnd, H. -P. 1986, A&A, 160,324 Mezger, P. G., Duschl, W. J., & Zylka, R. 1996, A&AR, 7,289 Morris, M., 1994, in R. Genzel& A.1. Harris, eds., The nuclei of Normal Galaxies. Kluwer, Dordrecht, p.185 Novak, G., et al. 2003, these proceedings Park, S., et al. 2003, these proceedings Reynolds, S. P., & Keohane, J. W. 1999, ApJ, 525, 368 Sakano, M., Warwick, R. S., Decourchelle, A., & Predehl, P. 2003a, MNRAS, 340, 747 Sakano, M., Warwick, R. S., & Decourchelle, A. 2003b, AdSpR, submitted Sakano, M., Warwick, R. S.,& Decourchelle, A. 2003c, Proc. “Japan-GermanyWorkshop on Galaxies and Clusters of Galaxies”. p.9 (astro-ph/0212464) Sakano, M., Warwick, R. S., & Decourchelle, A. 2003d, these proceedings Sofue, Y.,Reich, W., &L Reich, P. 1989, ApJL, 341, L47 Wang, Q. D., Gotthelf, E. V., & Lang, C. C. 2002, Nature, 415, 148 Warwick, R. S. 2002, Proc. New Visions of the X-ray Universe in the XMM-Newton and Chandra era, in press (astro-pW0203333) Warwick, R. S. 2003, these proceedings Zylka, R., Mezger, P. G., & Wink, J. E. 1990, A&A, 234, 133

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S 1,24 1-245 (2003) / DO1 10.1002/asna.200385035

High-negative velocities in the inner 25 pc of the Galactic center Lorht 0. Sjouwerman National Radio Astronomy Observatory, P.0. Box 0, Socorro, NM 87801; Isjouwerman@nrao. edu

Key words Galactic center, high-velocity stars, high-velocity gas, kinematics

Almost three decades ago, in a survey for OH masers in the Galactic center, Baud et al. detected an OH/IR star with a line-of-sight velocity of -343 !an sK1 with respect to the Local Standard of Rest, 50 pc in projection from Sgr A*. Since then, at least three more high-velocity (IVl > 250 krn s-') O M R stars have been found within about 25 projected pc of the Galactic center, all with a blue-shifted (negative) velocity. Over the years, several authors have also found emission and absorption by high-negative velocity gas (HI,HzCO, HCO', CO, OH) toward the Galactic center; gas with IVI > 180 km sK1.This contribution attempts to explain the observations of the high-negative velocity OWIR stars by relating their remarkable spatial and kinematic alignment in the inner 25 pc to the OhSerVdtiOnS of the high-negative velocity gas.

1 Introduction: stars on highly elongated orbits While performing a survey for OH sources in the direction of the Galactic center (GC), Baud et al. (1975) observed a double peaked 1612 MHz OH absorption feature in a frequency switched observation, 25' from Sgr A* (50 pc in projection). Further analysis showed that this absorption feature was actually a double peaked emission feature in the reference band. As the double peaked profile is characteristic for OWIR stars, this star therefore had to be at a high velocity, with a velocity of about -340 km s-' (all velocities in this paper refer to line-of-sight velocities with respect to the Local Standard of Rest). Although it was known that high velocities (up to f 2 0 0 km S C ' ) are not uncommon in the dircction of the GC, the authors already remarked that the star is far outside that range. Another striking characteristic is that the object seems to be moving counter to the Galactic rotation. In subsequent papers where they analyzed their results, it was demonstrated that high-velocity stars, i.e. stars with / V / > 250 km s-', should be uncommon (Baud et al. 1979), and that a (high) negative velocity due to a star outside thc Solar circle was unlikely to explain the observation (Baud et al. 1981). About ten years later, in 1992, van Langevelde et al. published the detection of two additional highvelocity OH/IR stars. They are located even closer to the GC, within 12' or within 25 pc in projection, and also appear at negative line-of-sight velocities of about -300 to -350 km s- l ; both moving with Galactic rotation. No OWIR stars were found at high-positive velocities. Having three high-negative velocity (HNV) OH/IR stars among about 150 (from a survey by Lindqvist et al. 1992), the authors were able to distinguish between some possible explanations. As the possibility of finding three HNV stars with the same sign is 25%, thcy did not consider this as important. However, they noted that 2 3 HNV stars is a few percent, which is too high to be explained by a high-velocity tail of an isotropic velocity distribution, and that collision mechanisms also would produce velocities higher than the apparent range up to the roughly 350 km s K 1 observed. They concluded - conditional upon no significant additional detections of negative high-velocity stars without corresponding detections of positive high-velocity stars - that high-velocity stars were consistent with low angular momentum, highly elongated orbits along the line of sight in a tri-axial bulge of bound bulge stars currently passing the GC. 0Lo03 WlLEY-VCH

Verlap GmhH 8: Co KGaA. Weinhem

242

Lor& 0. Sjouwerman: High-negative velocities in the inner 25 pc of the Galactic center

2 High-velocity gas on hyperbolic orbits Two decades ago, Gusten & Downes (1981) pointed out that, toward the radio continuum of the GC, absorption by H2C0 and H I gas was seen at a velocity around - 190 km s-’. Furthermore they remarked that this velocity exceeds the velocity range of the known gaseous components, that the gas must be in front of the GC, and that equilibrium rotation about the nucleus could be ruled out (unless the center of rotation was shifted 210’ away from the nucleus). They favored an “ejection” model after they argued against a large scale foreground “field” model. About ten years later, Man et al. (1992) suggested that some of this HNV gas interacts with the GC environment, whereas Pauls et al. (1993) proposed the HNV gas close to the GC to be falling inward from behind. In 1993, Liszt & Burton discussed the observation of Marr et al. (1 992) and Pads et a1 (1 993) that Sgr A * was not found to be occulted by the HNV gas (see however Karlsson et al. 2003a, 2003b). and argued that the HNV gas was a foreground cloud and unrelated to the phenomena in the GC. They mainly based this on their CO emission mapping of the HNV gas, which showed an elongated East-West structure West of Sgr A* with a very small velocity dispersion across the feature, and which is suspiciously parallel to the rotation axis of the Circumnuclear Disk (CND), as well as to the emission of the near side kpc-scale “expanding” molecular gas. Many papers followed with different tracers and on different angular scales in which the physical association between the HNV gas and Sgr A* with its direct environment was investigated (e.g. Yusef-Zadeh et al. 1993; Marshall & Lasenby 1994; Yusef-Zadeh et al. 1995; Liszt & Burton 1995; Zhao et al. 1995; Roberts et al. 1996; Sofue 1996; Karlsson et al. 2003a, 2003b). Arguments in “favor of this”, and arguing “against that” went back and forth, and it probably now would be accepted that the HNV gas and the GC do interact, in particular at a location a few arcminutes to the South-West of Sgr A*. The HNV gas has been modeled as gas on a tilted hyperbolic orbit about Sgr A*, where the gas is tidally disrupted within the central 100 pc and is flowing toward the nucleus from the far side (Zhao et al., 1995; Roberts et al., 1996). However, none of the authors directly related the observations of the HNV stars and the HNV gas.

3 Associating the HNV stars and HNV gas: the emerging picture In a deep OH survey, Sjouwerman et al. (1998; Sjouwerman 1997) found another high-velocity OWIR star, again at a negative velocity of -301 km s-l. Actually, the new HNV O M R star appears almost exactly in the line connecting the two HNV stars of van Langevelde et al. (1992) in the sky plane. This alignment still holds when another HNV star at -346 km spl is included, a HNV star reported by Menten & Reid (1998, in their figure and recently associated with IRS 9 by Reid et al. 2003). Finding significantly more, negative-only high-velocity stars in the velocity range -300 to -350 km spl clearly undermines the tri-axial bulge, bound elongated orbit HNV explanation by van Langevelde et al. (1992) as high-positive velocities have been equally well surveyed in these studies. Investigating the striking East-West orientation of the spatial alignment of the HNV stars, in the same direction as the CO-cloud of Liszt & Burton (1993), Sjouwerman (1997) linked the observations of the HNV stars with the observed HNV gas (Fig. I). It turned out that the large scale HNV gas distribution, as well as the HNV stars, lie in a plane with the same inclination with respect to the Galactic plane, which does not intersect with the position of Sgr A* itself (but South-West of it). Moreover, the velocity structure of the HNV gas also aligns with the velocities of the HNV stars in the longitude-velocity diagram (Fig. 2). Recalling the explanation of the HNV gas with an inclined hyperbolic orbit by Zhao et al. (1995) and Roberts et al. (1996), it seems plausible to investigate explaining the HNV stars with this model. Picturing that the gas and stars are falling in from a positive longitude somewhere (North-)East and behind the GC. on an inclined orbit with the peribothron (periastron) South-West of Sgr A* and elongated along the line of sight, the gas is gravitationally more easily influenced by the GC than the stars. The gas interacts, decelerates, and dissipates, giving rise to a hyperbolic appearing orbit (a transition of xl-to-xz orbits?), and the observedinteractions of the HNV gas in the inner few pc. The unbound heavy stars, the HNV stars with

243

Astron. Nachr./AN 324, No. S I (2003)

0 7

Lo

0

-tJ

U

0 .*

0

I

0 U

0 7

I10

5

0

-5

-10

g a l a c t i c longitude o f f s e t ( a r c m i n ) Fig. 1 Galactic sky distribution of HNV OH/IR stars and HNV gas (here the "CO gas; Liszt & Burton 1993). Offsets are with respect to Sgr A*. The arbitrary contours (gray-scales)outline the HNV gas with a velocity of - 190 (-213) !an sP1. The HNV stars at a velocity of around -320 (-150) h s-l are the open (closed) symbols. Comparison with other HNV gas, for example H I absorption (Yusef-Zadehet al. 1993) or 1667 MHz OH absorption (Karlsson et al., 2003b, and these proceedings), shows a similar picture. Taken from Sjouwennan (1997).

-360 < V < -290 km s-', are hardly affected, nor decelerated, and continue almost in a straight line, i.e. more to the South (more negative longitudes) of the HNV gas. The discussion of IRS 9 (at V= -346 kin s-l) being unbound by Reid et al. (2003) would support this vision. It would explain the high velocities of the HNV stars, their location with respect to the GC and HNV gas, the bound stars at -150 km s-' (that may have decelerated and thus follow the HNV gas closer), and the asymmetry of lacking positive high-velocity components. That the gas and stars have a similar inclination as the rotation axis of (but offset from) the CND (Liszt & Burton 1993) and the kpc-scale radio lobe (Sofue 1996) should be regarded

Lorint 0. Sjouwerman: High-negative velocities in the inner 25 pc of the Galactic center

244

a

-

... .......~..........

....

~...

......~.... .....~ ......................

.... ....

. . . . ~ ~ . ..... .

0

co

0

-320 kmls

o BAUD I

I

20

~

I

I

I

0

I

-

-

I

I

-20

Galactic longitude ~ f f f s e t(arcrnin) Fig. 2 Galactic velocity distribution of OH/IR stars (Sjouwerman et a]. 1998). The HNV OWIR stars with velocities around -320 km s-l clearly stand out (open star symbols in Fig. I), as well as the group around -150 km sC1 and with a Galactic longitude between 35' of Sgr A" (filled star symbols in Fig. 1). The circles are the OWIR stars known prior to the Sjouwennan et al. (1998) survey. The spatial and velocity structure of the HNV gas nicely fits into the diagram between the -320 km s-' and - IS0 km s-' groups. Modified from Sjouwennan (1997).

as coincidental, although the 1720 MHz OH masers at +60 km s-' (A, D, E, F, and G, Yusef-Zadeh et al., 1996, 1999) that also align might hypothetically result from the deflected gas.

4 Remains to be explained: Baud's star ! Ironically, this model does not directly solve the high-velocity problem posed by Baud's star, the star that triggered one to start thinking about this problem three decades ago. Possible solutions, that all can explain

Astron. Nachr./AN 324, No. S 1 (2003)

245

the retrograde circular motion and the deviation from the straight alignment with the H N V gas and other HNV stars. are that Baud's star: resulted from a tolally unrelated single event, e.g. such a s the collision mechanism described by van Langevelde et al. ( I 992).

falls in from the other side, indeed having an opposite orbital angular momentum. This would agree with the velocity of the other H N V stars, but one would expect a better alignment in the plane of the sky.

lags in a similar orbit, still o n the other sidc of S g r A*. This would explain the difficulty of finding a red counterpart (Blommaert e t al. 1998) and is consistent with a peribothron (periastron) South of Sgr

A*. g o t deflected by Sgr A*, but still moves almost in the line of sight. This would agree with the misalignment (to the proper side of Sgr A*) with the other H N V stars, and may b e indicated by the HNV gas East of Sgr A* as observed by Yusef-Zadeh et al. (1993, their Fig. 2). It is also consisten1 with the proper motion vector of the probably unbound H N V source IRS 9 (Reid e t al. 2003).

5

Summary

The explanation of the high-velocity stars by van Langevelde et al. (1992) does not hold because significantly more negative high-velocity stars have been found. O n the other hand, the high-velocity stars d o align with the high-negative velocity gas that i s intcracting with the Galactic center. Modeling of the gas (Zhao et al. 1995;Roberts et al. 1996),indications that the high-negativevelocity stars are not bound (Reid e t al. 2003), and combination of the kinematic observations of the gas and stars, picture a cloud falling in from the far side of, and interacting with the Galactic center. T h e unbound high-velocity stars were filtered from the cloud by the interaction, where Baud's star was possibly deflected by S g r A*.

References Baud, B., Habing, H. J . , Osullivdn, J. D., Winnberg, A. & Matthews, H. E. 1975, Nature 258,406 Baud B, Habing, H. J., Matrhews, H. E. & Winnbeg, A. 1979, A&AS 35, 179 Baud B, Habing, H. J., Matthews, H. E. & Winnberg, A . 1981, A&A 95, 171 Blommaert, J. A. D. L., van der Veen, W. E. C. J., van Langevelde, H. J . , Habing, H. J. & Sjouwerman, L. 0. 1998, A&A 339,991 Gusten, R. & Downes, D. 1981, A&A 99, 27 Karlsson, R., Sandqvist, Aa., Sjouwerman, L. 0. & Whiteoak JB 2003a, these proceedings, Sect. 111 Karlsson, R., Sjouwerman, L. 0..Sandqvist, Aa. & Whiteoak JB 2003b, accepted by A&A Lindqvist M, Winnberg, A., Habing, H. J. & Matthews, H. E. 1992, A&AS 92, 43 Liwt, H. S. & Burton, W. B. 1993, ApJ 407, L25 Liszt, H. S. & Burton, W. B. 1995, ApJS 98, 679 Marr, J. M.. Rudolph, A. L., Pauls, T. A., Wright, M. C. H., & Backer, D. C. 1992, ApJ 400, L29 Marshall, J. & Lasenby, A. 1994, MNRAS 269, 619 Menten. K. M. & Reid, M. J . 1998, ASP Conf. Scr. 144 IAU Colloq. 164, 229 Pauls, T., Johnston, K. J., Wilson, T. L., Man-, J. M. & Rudolph, A. 1993. ApJ 403, L13 Reid, M. J . , Menten, K. M., Genzel. R., Ott, T., Schodel, R. & Eckart, A. 2003, ApJ in press, (astro-ph 0212273) Roberts, D. A , , Yusef-Zddeh, F. & Goss, W. M. 1996, ApJ 459, 627 Sjouwerman, L. 0. 1997, Onsala PhD thesis, Chalmers University of Technology, Gothenburg, Swedcn Sjouwerman, L. 0..van Langevelde, H. J., Winnberg, A. & Habing, H. J. 1998, A&AS 128, 35 Sofue, Y, 1996, ApJ 459, L69 van Langevelde, H. J., Brown, A. G. A,, Lindqvist, M., Habing, H. J. & de Zeeuw, P. T. 1992, ApJ 396, 686 Yusef-Zadeh, F., Lasenby, A. &Marshall, J. 1993, ApJ 410, L27 Yusef-Zadeh, F., Zhao, J.-H. & Goss, W. M. 1995, ApJ 442, 646 Yusef-Zddeh, F., Roberts, D. A,, Goss, W. M., Frail, D. A. & Green, A. J. 1996, ApJ 466, L25 Yusef-Zadeh, F., Roberts, D. A,, Goss, W. M., Frail, D. A. & Green, A. 5. 1999, ApJ 512, 230 Zhao, J.-H., Goss, W. M. & Ho, P. T. P. 1995, ApJ 450. 122

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S I , 247 - 253 (2003)/ DO1 10.1002/asna.200385054

Really Cool Stars and the Star Formation History at the Galactic Center

',

Robert D. Blum" Solange V. Ramirez**I,Kristen Sellgren***3,and Knut Olsent I

*

'

Cerro Tololo Interamerican Observatory, Cassila 603, La Serena, Chile SIRTF Science Center, JPL/Caltech, Pasadena. CA 91 125, USA Astronomy Department, The Ohio State University, 140 West 18th Ave, Columbus, OH 43210, USA

Key words Galactic Center, AGB Stars, Star Formation Abstract. We present AlAA = 550 to 1200 near infrared I1 and K spectra for a magnitude limited sample of 79 asymptotic giant branch and cool supergiant stars in the central h 5 pc (diameter) of the Galaxy. Using a set of similar spectra ohtained for solar neighborhood stars with the same range i n T..Rand MI,<,, for the Galactic center (GC) sample as calibrators, we construct the Hertzsprung-Russell diagram for the GC sample. Using an automated maximum likelihood routine, we derive a coarse star formation history of the GC. We find that roughly 75 % of the stars that have formed in the central few pc are older than 5 Gyr. In particular, the best fitting star formation history yields star formation rates in four age bins of 0.01 to 0.10 Gyr, 0.10 to 1.0 Gyr, 1.0 to 5.0 Gyr, and 5 Gyr to 12 Gyr, of 39.5 i 11 x lop4 Ma yr-', 3.6 i 2 x lop4 M a yr-', 5.7 f 2 x Ma yr-', and 13.4 i 5 x Ma yr-', respectively. The GC has been forming stars over the last 100 Myr at a rate roughly four times higher than the average rate over 12 Gyr, though the total mass formed in this recent epoch is less than three percent of the total mass formed over all times.

1 Introduction In their review of the global phenomena oii-going in the G C region, Morris & Serabyn (1996) described the properties o f the "central molecular zone," or CMZ. The CMZ is a "disk" of enhanced molecular density about 200 pc in radius centered on the GC. The gas is confined to a region near the plane of the Galaxy, but with significant non-circular motions. The distribution and presence of molecular gas in the C M Z may in large part be due to the effects of the inner Galactic stellar bar (Liszt & Burton 1980, Mulder & Liem 1986, Binney et al. 1991, Blitz & Spergel 1991, Weiland et al. 1994, Dwek et al. 1995). The material in the CMZ is fueling current star formation on this large scale at a rate of about 0.5 M, yrP1 (Gusten 1989), but it may also be the ultimate source of material which is processed into stars within a few pc of the G C (Morris & Serabyn 1996). If so, angular momentum losses must funnel the gas down to the circurnnuclear disk (CND) at radii between 2 and 8 pc (see the extensive reviews by Genzel, Hollenbach, & Townes 1994 and Morris & Serabyn 1996). This molecular structure is probably not a long-lived one, but rather periodically forms and supplies the G C with star forming material through instabilities which cause material to fall from its inner radius into the central pc (Sanders 1999); at present the CND may be ~ M o y r r ' (Gusten et al. 1987, Jackson et al. 1993). accreting about 0 . 5 lo-* ~ Mt,,,l for a magnitude limited sample of C C stars, using the twoOur goal is to determine T pand dimensional classification provided by the measured strengths of the CO and H20 absorption features N

* e-mail: rblumQctio.noao.edu, Phone: c56 51 205272, Fax: +56 51 205212 ** [email protected] *** [email protected] + [email protected]

@ 2003 WLLEY-VCH Verlag GmhH B Co. KQA, Wcinhcim

R. D. Blum et al.: Really Cool Stars

248

present in modest resolution K-band and H-band spectra. After T,,and l\/ft,ol are determined, we place the GC stars in the HR diagram and use this to constrain the star formation history (SFH) within the central few pc of our Galaxy.

2 OBSERVATIONS Spectroscopic observations of the GC stars and comparison stars were made using the facility IRS and OSIRIS spectrometers mounted on the Cerro Tololo Interamerican Observatory (CTIO) 4m Blanco telescope over several runs beginning in 1997 and ending in 2000 (see Tables 1 and 2). In addition, comparison stars were observed at the Michigan-Dartmouth-MIT (MDM) 2.4m telescope on Kitt Peak using the Ohio State University MOSAIC infrared carnerdspectrometer. The IRS, OSIRIS, and MOSAIC are described by DePoy et al. 1990, DePoy et al. (1993), and Pogge et al. (1998), respectively.

'

2.1

SPECTRAL CLASSIFICATION

Following Blum et al. (1996; hereafter BSD96), we used K-band spectroscopic indices for CO and H20 to provide a two-dimensional classification yielding Teff and ni[bol for cool, luminous GC stars (e.g., Kleinmann & Hall 1986; see also Ramirez et al. 1997). The CO index is a measure of the strength of the "CO 2.2935 pm 2-0 rotational-vibrational bandhead (Kleinmann & Hall 1986). The HzO feature is a broad depression of the continuum between the H and K-bands due to myriad blended steam absorption lines. Both H and K band coverage are used to define the HzO absorption where as BSD96 had only K-band spectra. The CO index is defined (BSD96) as the percentage of flux in the CO 2.3 pm feature relative to a continuum band centered at 2.284 pm ([ 1 - Fbnnd/Fcont] x 100). The CO band and continuum band were 0.015 pm wide and the CO band was centered at 2.302 pm. The CO index is only marginally affected by extinction since the CO and continuum bands are closely spaced. A typical CO strength of 20% changes by about 1% for a delta A K of one magnitude. This is similar to the typical uncertainty in the derived CO strength which is taken as the one-o uncertainty in feature strength derived from the pixel-to-pixel variation in the nearby continuum. The CO and associated continuum band are graphically represented in the upper panel of Figure 1. The HzO strength is defined similarly to the CO index, but with a quadratic fit to the continuum using bands at 1.68-1.72 pm and 2.20-2.29 pm (see the lower panel of Figure 1 ) and a band 0.015 p m wide centered at 2.0675 pm (indicated in the lower panel of Figure I). The uncertainty in the measured HzO is similar to that for CO (51-2 %). not including any systematic uncertainty in the derived continuum. This should be much smaller than given by BSD96 since the continuum fit spans the H and K bands and is thus insensitive to the details of the reddening. 2.2 Comparison Stars The comparison stars were chosen to span the range of expected GC star Mb01 and T,A.(BSD96, Can et al. 2000, Ramirez et al. 2000). The literature was surveyed for cool giant, AGB, and cool supergiant stars with derived Teff and Mhnl. 2.3 Galactic Center Sample The complete GC sample was chosen as all stars in the GC K-band luminosity function (KLF, taken here to be the dereddened luminosity function) as derived by BDS96 which have KO5 7.0 (146 stars). The counts at this magnitude limit are complete; however, 31 stars in this original list were not observed

' OSIRIS is a collaborative project between the Ohio State University and CTIO. OSIRIS was developed through NSF grants AST 90161 12 and AST 9218449.

249

Astron. Nachr./AN 324, No. S1 (2003)

1""""""""""I"I

0 J I 1.4

l

l

l

l

1.6

l

l

l

l

l

1.8

l

l

l

l

l

2

l

l

l

l

I

2.2

I

I

I

I

I

I

I

I

2.4

Wavelength (pm) Fig. 1 H and K spectra of the asymptotic giant branch (AGB) stars R Ser (M7 111, Mira) and BK Vir (M7 111) used to demonstrate the CO and HzO measurements for all spectra. The CO strength is determined by ratio of flux in the band centered at 2.302 pm compared to the flux in the continuum band (in the star, not the fit) at 2.2875 p m (the bands are indicated with vertical dashed lines in [he upper panel). The dmhed curves are quadratic fits to the continuum in bands at 1.68-1.72 pm and 2.20-2.29 p n (as indicated in the lower panel). The H 2 0 strength is measured using the (lux in a band at 2.060-2.075 pm relative to the flux in the fit at the same position (as indicated in the upper panel). The strong H 2 0 absorption between 1.7 p m and 2.1 p m is used to distinguish AGB stars from luminous supergiants which have simi1arCO (2.3 Dm bandhead feature) strength.

because of severe crowding with neighboring stars, and w e were unable to observe an additional 24 stars

R. D. Blum et al.: Really Cool Stars

250

1.4

1.6

1.8

2

2.2

2.4

Wavelength (pm) Fig. 2 Example spectra of stars classified as supergiants in the Galactic center. These stars were analyzed at high spectral resolution by Ramfrez et al. (2000) and Carr et al. (2000). The dashed curves are the fits to the continua used to measure HzO at the position of the vertical dashed lines (see text and Figure 1). Y axis is scaled logarithmically.

because of cloudy weather at the telescope. Eleven stars in the list are known emission-line stars or have featureless continua (BDS96). The luminosity class of each GC star was determined by examining the strength of the measured HzO and CO. Strong or extreme HzO indicates a giant or LPV type, respectively. For supergiants, the H 2 0 must be very small or absent for CO up to about 20 %. Due to the rather large scatter in HzO strengths in the comparison stars, in practice, we set the limit at H20 < 5 % for CO up to 20%. Example spectra of GC stars are shown in Figure 2.

25 1

Astron. Nachr./AN 324, No. S1 (2003)

3 THE STAR FORMATION HISTORY The spectral indices developed in the preceding section allow us derive bolometric magnitudes and T,tf for the GC stars. Tefffollows directly from the measured CO index once the luminosity class is chosen. The bolometric correction (BCK) is derived from the literature for the different luminosity classes given in Table 5. For supergiants and LPVs, the BCK is the same as given by BSD96: BCK =2.6 for supergiants and 3.2 for LPVs. For the giants, we improve on the work of BSD96 by considering a BCK which is a function of Tes. Using the BCK as a function of .I - K given by Frogel & Whitford (1987), the mean ,J - K of giants as a function of spectral type from Frogel et al. (1978), the spectral type vs T e from ~ Ramirez et al. (2000) and Dyck et al. (1998), and the Tetf derived here, we calculate BCK as a linear function of T e ~These . BCK range from 2.8 to 3.2 for the warmest and coolest GC giants. Figure 3 shows the Hertzsprung-Russell diagram (HRD) for all the GC stars and comparison stars. Overplotted are Bertelli et al. (1994) and Girardi et al. (2000) isochrones spanning ages from 10 Myr to 12 Gyr, which show that these data span a wide range in age. As can be seen in the figure, all of the GC giants are AGB stars. They are too luminous to be first ascent giants, and this is a consequence of our selection criteria. We have used the results of Figure 3 to derive the SFH implied by these observations of the GC cool stars. The calculation was carried out using Olsen’s (1999) implementation of the method described by Dolphin (1997), with some modifications. In brief, we constructed a set of models describing the expected distribution of stars in the H-R diagram within specified age bins, assuming a particular metallicity, slope of the initial mass function (IMF), and constant star formation rate ( 1 Moyr-l) within the bin and accounting for observational errors and incompleteness. We chose the best model SFH for the GC by fitting the observed data to a linear combination of the star formation within these bins. This fit was determined through a maximum likelihood analysis. Two choices of sets of age bins and metallicities were used. Models had either four age bins (10-100 Myr, 100 Myr - 1 Gyr, 1-5 Gyr, and 5-1 2 Gyr) or three age bins (1 0-50 Myr, 50 Myr - 3 Gyr, and 3-1 2 Gyr). For both sets of age bins, models were run with all stars at solar [Fern] (Models A with four age bins and Model 1 with three age bins), and then again with solar [Fe/H) for the younger stars and [Fe/H] = -0.2 in the oldest bin (Model B with 4 age bins, Model 2 with three age bins). We assumed the IMF, since our data do not span a large enough range in mass at a given age to allow it to be a free parameter. We adopted the same modified Salpeter (1955) IMF as Miralda-Escude & Gould (2000). The fits for models 1 and 2 were siginificantly worse, and are not discussed further. The resulting SFH for Models A&B, showing the SFR in age bins of 10-100 Myr, 100 Myr - 1 Gyr, 1-5 Gyr, and 5-12 Gyr, is given in Figure 4. Figure 4 indicates significant on-going star formation in the central few pc, but that the bulk of stars (roughly 75 % by mass) formed at earlier times. This is in agreement with earlier work based on near infrared number counts (Genzel et a1 1994, Mezger et al. 1999). For the oldest stars, we are sampling just the very tip of the AGB, hence to observe any stars in such a short lived phase requires a large mass to have originally formed. The details change by roughly a factor of two depending on which model SFH is chosen, but the uniform metallicity case (Model A) is preferred. If true, it could suggest that the nucleus formed largely from enriched material produced in the early formation of the bulge. The purely solar [Fern] is also consistent with the narrow distribution of [Fern] from high resolution spectra (Ramirez et al. 2000). Though the range of ages considered by Ramirez et al. (2000) is not as large as the data sct presented here, the high resolution data sample stars with ages up to 5 Gyr (Figure 3). The total mass represented by the SFH in Figure 4 is 12 i 4 x lo6 Ma for Model A. This is at the upper end of mass estimates for the central 2 pc from dynamical models and about 3-6 times larger than the most detailed dynamical models which use star counts and three-space velocities (not counting the central black hole mass). Mass loss during the lifetime of stars from about 1 Ma to 120 M a will reduce the cumulative initial mass in the cluster. If all stars with 120 Ma > M > 1 Ma are taken to have their remnant mass at the present time, then the present inass in stars is reduced by about 45 ‘ro (i.e. we infer the present mass in stars remaining in the cluster to be 0.55 x the total mass formed over all times). The mass N

N

252

R. D. Blum et al.: Really Cool Stars

-10 I ' ' ' ' ' '

i+

-81

' ' '

''''''

'

'

' '

'''''''''

'

''

' ' ' ' ' ' ' '

''

' ' '

' ' ' ' '

' ' '

' ' ' ' '

I

I

' ..

*'** - _

5000

4000 3000 TeJK)

5000

4000 3000 T,JK)

Fig. 3 Hertzprung-Russell diagram for the Galactic Center (GC) stars (shown as open circles with typical uncertainty given by the error bars in the upper left comer of each panel) and comparison stars (Comp Stars, plotted as filled diamonds). The GC stars analyzed at high spectral resolution by Ramhez et al. (2000) and Carr et al. (2000) are plotted as filled squares. Model isochrones are shown for reference. The isochrones are from Bertelli et al. (1994) for age < 100 Myr and Girardi et al. (2000) otherwise. Isochrones are plotted for ages of 10 Myr, 25 Myr, 50 Myr, 100 Myr, 500 Myr, 1 Gyr, 4 Gyr, and 12 Gyr. The models in the lefl panel have [Fern] = 0.0, and the models in the right panel have [Fern] = -0.2. Neither set of isochrones reaches the coolest stars, but the [Fe/H] = 0.0 isochrones extend to cooler temperatures and thus fit more Galactic stars than the [Fern] = -0.2 isochrones. Comparison to the isochrones shows that all the GC stars classified as giants are AGB stars; they are too luminous to be first ascent giants. This is a consequence of the selection criteria. The horizontal line segment at MI,,,^ = -7.2 in each panel indicates the approximate observed luminosity above which only supergiants lie BSD96.

lost through stellar winds could be expelled from the region andlor recycled into new generations of stars. The total present mass in stars and remnants due to the cumulative star formation history inside the central 2 pc (radius) is thus 56.6 f 2 x lo6 M, and so within about a factor of two compared to expectations from the dynamical models. Changes in the lower mass cut-off or slope in the adopted IMF could result in less total star formation. For example, simply cutting off the mass function at one Ma would reduce the total by approximately a factor of 2. Finally, we have implicitly assumed all the tracers of the SFH lie within a true radius of 2 pc, but they are actually distributed in a projected radius of 2 pc. Given the steepness of the stellar cluster radial density distribution (20.5 pc core radius), the overestimate is likely to be small. A full discussion and updated results for this work are given by Blum et al. (2003)

References Bertelli, G.. Bressan, A,, Chiosi, C., Fagotto, F., & Nasi, E. 1994, A&AS, 106, 275 Binney, J . , Gerhard, 0. E., Stark, A. A,, Bally, J., & Uchida, K. 1. 1991, MNRAS, 252,210 Blum, R. D., Sellgren, K. &DePoy, D. L. 1996, ApJ, 470,864

Astron. Nachr./AN 324, No. S1 (2003)

253

Model A

Age (Gyr)

4

Age (Gyr)

b)

Fig. 4 Star formation history (SFH) for the Galactic center. The crosses represent the results to the SFH fits to the Hertzspmng-Russell Diagram (Figure 3) for Models A (upper panel) with Solar [Fe/H] throughout and Model B (lower panel) with [Fe/H] = -0.2 in the oldest age bin. Model A is a statistically better fit; see text. The age bins corresponding to the horizontal width of the crosses are 10 Myr - 100 Myr, 100 Myr - 1 Gyr, 1 Gyr - 5 Gyr, and 5 Gyr - 12 Gyr. The vertical height of each cross is the onc sigma error in the star formation rate for the respective bin; see text. Blum, R. D., Sellgren, K. &DePoy, D. L. 1996, AJ, 112, 1988 (BSD96) Blum R. D., Ramirez, S. V., Sellgren, K., & Olscn, K. 2003, ApJ, in press Blitz, L. & Spergel, D. N. 1991, ApJ, 379. 631 Cam, J. S., Sellgren, K., & Balachandran, S. C. 2000, ApJ, 530, 307 DePoy, D. L., Gregory, B., Elias, J., Montane, A., Perez, G., & Smith, R. 1990, PASP, 102, 1433 DePoy, D. L., Atwood, B., Byard, P., Frogel, J . A., & O'Brien, T. 1993, in SPIE 1946, "Infrared Detectors and Instrumentation," pg 667 Dolphin, A. E. 2002, astro-pW0112331 Dwek. E. et al. 1995, ApJ, 445,716 Dyck, H. M., Benson, J. A., van Belle, G. T., & Ridgway, S. T. 1996, AJ, 11 I , 1705 Dyck, H. M., van Belle, G. T., & Thompson, R. R. 1998, AJ, I 16, 981 Frogel, J. A., Persson, S. E., Mdtthews, K., & Aaronson, M. 1978, ApJ, 220. 75 Frogel, J. A. ti Whitford, A. E. 1987, ApJ, 320. 199 Genzel, R., Hollenbach, D., & Townes, C. H. 1994, Reports of Progress in Physics, 57,417 Ghez, A. M., Klein, B. L., Moms, M., & Becklin, E. E. 1998, ApJ, 509, 678 Girardi, L., Bressan, A., Bertelli, G., & Chiosi, C. 2000, A&AS, 141, 371 Giisten, R. 1989, IAU Symp. 136: The Center of the Galaxy, 136,89 Giisten, R., Genzel, R., Wright, M. C. H., Jaffe. D. T., Stutzki, J., & Harris, A. 1. 1987, ApJ, 318, 124 Jackson, J. M., Geis, N., Genzel, R., Hams, A. I., Madden, S., Poglitsch, A., Stacey, G. J., & Townes, C. H. 1993, ApJ, 402, 173 Kleinmann, S. G. & Hall, D. N. B. 1986, ApJS, 62, 501 Liszt, H. S. & Burton, W. B. 1980, ApJ, 236, 779 Mezger, P. G., Zylka, R., Philipp, S., & Launhardt, R. 1999, A&A, 348,457 Miralda-EscudC, J. & Gould, A. 2000, ApJ, 545, 847 Moms, M. & Serabyn, E. 1996, ARAA, 34,645 m186 Mulder, W. A. & Liem, B. T. 1986, A&A. 157, 148 Olsen, K. A. G. 1Y9Y, AJ, 1 17. 2244 Pogge, R. W. et al. 1998, SPIE, 3354,414 Ramirez, S. V., Sellgren, K., Carr, J. S., Balachandran, S. C., Blum. R., Terndrup, D. M., & Steed, A. 2000, ApJ, 537,205 Salpeter, E. E. 1955, ApJ, 121, 161 Sanders, R. H. 1999, ASP Conf. Ser. 186: Thc Central Parsecs of the Galaxy, 250 Weiland, J. L. el al. 1994, ApJ , 425, L81

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S 1.255 -261 (2003) / DO1 10.1002/asna.200385060

Massive Stars and The Creation of our Galactic Center Donald F. Figer'

'

' STScI, 3700 San Martin Drive, Baltimore, MD 21218 Key words Galactic Center, Stellar Clusters, Star Formation Abstract. Our Galactic Center hosts over 10% of the known massive stars in the Galaxy, the majority of which are located in three particularly massive clusters that formed within the past 5 Myr. While these clusters are extraordinary, their formation repesents about half of the total inferred star formation rate in the Galactic Center. There is mounting evidence that the clusters are just present-day examples of the hundreds of such similar clusters that must have been created in the past, and whose stars now comprise the bulk of all stars seen in the region. I discuss the massive stellar content in the Galactic Center and present new data obtained with HST/NICMOS and Geniini AO, and an analysis that suggests that effects of continuous star formation in the Galactic Center can be wen in the observed luminosity functions.

1 Introduction Over 10% of the known massive stars (Mi,,it>20 M o ) in the Galaxy reside in three clusters of young stars located within 30 pc of the Galactic Center. These clusters are the most massive young clusters in the Galaxy and contain approximately 30 Wolf-Rayet (WR) stars, at least 2 Luminous Blue Variables (LBV), approximately a half dozen red supergiants, and approximately 450 0 stars. Together, they emit enough ionizing radiation to account for roughly half of the thermal radio emission in the central few degrees of the Galaxy, suggesting that the young clusters contain approximately half of the stars recently formed in this region. An additional collection of young stars exists in the region, with members scattered about the central 50 pc; some have evolved to the WR stage, while others are still deeply embedded within their natal dust cocoons. A lower bound to the current star formation rate can be approximated by dividing the mass in the clusters by their ages, i.e. 5(104) M 0 / 5 M y r ~ O . 0 1Mcj/yr, or a star formation rate density of Mo/yr p c 3 . This rate is approximately 250 times higher than the mean rate in the Galaxy, and about the same factor lower than the rate in starhurst galaxies. Clearly, the Galactic Center has formed a plethora of stars in the past 5 Myr, hut it is less apparent when the nlillions of stars in the central 50 pc formed. If we assume that the star formation rate in the past was similar to the present rate, then the total mass of stars formed over the past 10 Gyr is -1 O8 Ma within a radius of 30 pc of the Galactic Center, and an order of magnitude greater than this amount over the whole Central Molecular Zone, as first suggested by Serabyn & Morris ( I 996).

2 The Central Cluster The Central Cluster contains over 30 evolved massive stars having Minitial >20 Mo (Becklin & Neugehauer 1968, Rieke & Lehofsky 1982, Lebofsky, Rieke, & Tokunaga 1983, Forrest et al. 1987, Allen, Hyland, & Hillier 1990, Krahbe et al. 1991; 1995, Allen 1994, Rieke & Rieke 1994, Blum et al. 1995, Eckart et al. 1995, Genzel et al. 1996, Tamhlyn et al. 1996). A current estimate of the young population * Corresponding author: e-mail: figerQstsci.edu, Phone: 410-453-9321 Fax: 410-516-2829

02001 WlLEY-VCH Verlag GmhH B Co

KGaA, Wcinhriin

D. Figer: Massive Stars

256

includes 9 WR stars, 20 stars with OfpelWN9-like K-band spectra, several red supergiants, and many luminous mid-infrared sources in a region of 1.6 pc in diameter centered on Sgr A* (Genzel et al. 1996). In addition, I estimate that it contains at least 100 0-stars (07 and later) still on the main sequence. Najarro et al. (1994) and Najarro et al. (1997) modeled the infrared spectrum of 8 blue supergiants in the center, finding characteristics consistent with an “OfpeIWN9” classification, and concluding that they formed a few million years ago. More recently, Paumard et al. (2001) reviewed the emission-line stellar population in the central parsec, using new narrow-hand infrared imaging. Eckart et al. (1999) and Figer et al. (2000) identified massives stars within a few thousand AU of the supennassive black hole. Evidently, a significant fraction of this small group of stars are young ( T <~ 20~M y~r ) and require extraordinarily dense or formation sites much further away from the central black hole than pre-collapse cores ( p > lo1’ their present location would suggest. Results from proper-motion studies suggest that at least some of the stars in this cluster are bound to the black hole and are not on highly elliptical orbits (Ghez et al. 2001; Eckart et al. 2002); therefore they are likely to he near to their formation sites.

3 The Quintuplet Cluster The Quintuplet Cluster is located approximately 30 pc, in projection, to the northeast of the Galactic Center (Glass, Catchpole, & Whitelock 1987). In addition to the five bright stars for which the Quintuplet was named (Nagata et al. 1990; Okuda et al. 1990),the Quintuplet cluster contains a variety of massive stars, including four WN, five WC (possibly ten, see below), two WN9/0fpe, two LBV, one red supergiant and several dozen less-evolved blue supergiants (Figer et al. 1999a, 1999~).The five Quintuplet-proper members are massive stars (L lo5 L o ) embedded within dusty cocoons, although their spectral types and evolutionary status are unknown (Moneti et al. 2001). Figer et al. (1996, 1999a) argue that these stars are dust-enshrouded WCL stars, similar to WR 140 (Monnier et al. 2002) and WR 98A (Monnier et al. 1999). New evidence in support of this hypothesis was presented by Chiar (2003), Law (2003), and Lang (2003). In addition to these post-main sequence stars, it is likely that 100 0-stars still on the main sequence exist in the cluster, assuming a flat to Salpeter IMF. The total cluster mass is estimated to he w 1 O 4 M a . The total ionizing flux is photons spl, enough to ionize the nearby “Sickle” HI1 region ((30.18-0.04). The total luminosity from the massive cluster stars is NN 107.5La,enough to account for the heating of the nearby molecular cloud, M0.20-0.033. The two LBVs in the cluster are added to the list of 6 LBVs in the Galaxy. They include the Pistol Star (Moneti et al. 1994; Figer et al. 1995a, 1995b, 1998, 1999b; Cotera 1995; Cotera et al. 1996), one of the most luminous stars known, and a newly identified LBV (Geballe et al. 2000) that is nearly as luminous as the Pistol Star. Most of the luminous stars in the cluster are thought to be 3-5 Myr old. N

Minitial

>300Mo,

4 The Arches Cluster First discovered about 10 years ago as a compact collection of a dozen or so emission-line stars (Cotera et al. 1992; Nagata et al. 1995; Figer 1995a; Cotera 1995; Cotera et al. 1996; Blum et al. 2001), the Arches cluster contains thousands of stars, including at least 160 0 stars, according to Figer et al. (1999~).Figer et al. (1999~) used HST/NICMOS observations to estimate a total cluster mass (2104M a ) and radius (0.2 pc) to arrive at an average mass density of 3(105) M a pcP3 in stars, suggesting that the Arches cluster is the densest, and one of the most massive, young clusters in the Galaxy. They further used these data to estimate an initial mass function (IMF) which is relatively flat (r -0.61t0.1) with respect to what has been found -1.35, Salpeter 1955) and other Galactic clusters (Scalo 1998). Stoke et for the solar neighborhood (I? al. (2002) recently confirmed this flat slope by analyzing the same data and recently obtained Gemini A 0 data. Figer et al. (2002) estimated an age of 2.510.5 Myr, based on the magnitudes, colors, mix of spectral

-

N

257

Astron. Nachr./AN 324, No. S1 (2003)

~905

210

215

220

Wavelength ( f i r m )

225

10

23G

11

a)

12

13

14

15

mf205W

b)

Fig. 1 (left) K-band spectra of four Wolf-Rayet stars (WNL) in the Arches Cluster from Figer et al. (2002). Emission lines can be seen at 2.058 p m (HeI), 2.104 pm (NIII), 2.1 12/113 p m (HeI), 2.1 15 pm (NIII), 2.166 pm (HeVHI), 2.189 pm (HeII), and 2.224/225 ,urn (NIII). The sharp feature near 2.32 pm in the spectrum for star #2 is due to a detector defect, and the absorption features longward of 2.33 pm are due to imperfect correction for telluric absorption. (right) Equivalent width of the 1.87 pm feature in massive stars of the Arches Cluster (Figer et al. 2002b; see also Blum et at. 2001) as a function of apparent magnitude in the HST/NICMOS F205W filter. The feature has contributions from He1 and HI.

10 5 0 - 5 - 1 0 R.A. o f f s e t {arcseconds} Fig. 2 Difference image, F187N-FI90N. highlighting stars with excess emission at 1.87 pm. Radio sources (v 4.9 G H z ) are shown by circles, as identified in Lang et al. (2001), and as squares, as identified by Figer et al. (2002). X-ray sources are shown by diamonds, as identified by Yusef-Zadeh et al. (2002).

types, and quantitative spectral analysis of stars in the cluster. Given the current state of knowledge about this cluster, it now seems apparent that we have observed a firm upper-mass cutoff (Figer 2003a). Note Indeed, we should even expect that we should expect at least 10 stars more massive than M,nitlal=300Ma. one star with an initial mass of 1,000 Mo ! Of course, it is questionable how long such a star would live; however, it is clear that the Arches cluster IMF cuts off at around 150 Ma.Finally, even if we steepen thc IMF slope to the Salpeter value, we still should expect at least 4 stars more massive than 300 Mo. Figer et al. (2002) conclude that the most massive stars are bona-fide Wolf-Rayet (WR) stars and are some of the most massive stars known, having Mirritial> 100Ma, and prodigious winds, A? > lop5 M a y-',

D. Fieer: Massive Stars

258

that are enriched with helium and nitrogen. These findings are largely based upon the spectra and narrowband equivalent widths shown in Figure 1, and a detailed quantitative analysis of these data (also see Najarro 2002). Figer et al. (2002) found an upper limit to the velocity dispersion of 22 km s-', implying an upper limit to the cluster mass of 7(104) M a within a radius of 0.23 pc, and a bulk velocity of vcluster xi-55 km s-' for the cluster. It appears that the cluster happens to be ionizing, and approaching, the surface of a background molecular cloud, thus producing the Thermal Arched Filaments. They estimate that the cluster produces 4(1051) ionizing photons s-', more than enough to account for the observed thermal radio flux from the nearby cloud. Commensurately, it produces lo7.' La in total luminosity, providing the heating source for the nearby molecular cloud, L,-loud = lo7 Lo. These interactions between a cluster of hot stars and a wayward molecular cloud are similar to those seen in the "Quintuplet/Sickle" region. Finally, note that significant work is being done on this cluster at radio and x-ray wavelengths, i.e. shown in Figure 2.

5 The Star Formation History of the Galactic Center The evidence for recent (<10 Myr) star formation in the Galactic Center abounds. The Lyman continuum flux emitted in the central few degrees of the Galaxy is about lo5' photonsls (Cox & Laureijs 1989), with half coming from stars in the three massive clusters. This flux is about 10% of the that for the whole Galaxy, and the number of massive stars (Minitial >20 Ma)in the GC is about 10% of the number in the whole Galaxy. However, note that the star formation rate in the GC is about one one-hundredth of that for the whole Galaxy. Such a low star formation rate as a function of Lyman continuum photon production necessarily follows from the relatively flat initial mass function (IMF) slope used in estimating the mass of stars formed in the young clusters. 8

NICMOS

10 12 14 16 18

N=8320

20 22

0

1

0

1

2

3

4

5

2

3

4

5

8 10

12 14 16 18 20

22

8 10 12 14 16 18

20 22 H-K

Or

mF160W-mF205W

Fig. 3 Observed color-magnitude diagrams for Lick data (upper panel), Gemini data (middle panel), and MCMOS data (lower panel).

The recent star formation history (-rage 6 50 Myr) in the Galactic Center is relatively clear. Embedded HI1 regions trace star formation at the present time (Figer 1995,Cotera 1995, Cotera et al. 1999), while the young clusters trace star formation that occured 2.5-5 Myr ago. The lack of red supergiants (Mnol<-6.3

259

Astron. Nachr./AN 324. No. S 1 (2003)

and Mirritial>8M a ) in the region provides evidence for a very low star formation rate from 5-20 Myr ago. Indeed, a burst the size of that seen in the three clusters at 20 Myr ago would have produced 40 red supergiants, yet we see none (other than those associated with the young clusters). The constraints on this activity are very strong, i.e. there were fewer than 5(IO3) Ma formed in stars over this time period, assuming the type of star formation that spawned the three massive clusters. Looking beyond 20 Myr, the picture becomes less clear because it is difficult to separate old low-mass red giants from much younger high-mass AGB stars with photometric data alone. Using spectroscopy, Blum et al. (1996) have shown that there is a relative dearth of stars with ages between 10 and 100 Myr in the central few parsecs. Given their conclusions, I estimate a low star formation rate during this period. These same authors, and Haller (1992), identified stars with ages on the order of a few hundred Myr in the central few arcminutes. Sjouwerman (1999) identified a population of OWIR stars with a narrow range of expansion velocities, indicating intermediate ages and a starburst event a few Gyr ago. In addition, Frogel et al. (1999) identified an excess of bright stars in the fields they observed within 0?2 of the Galactic Center. Finally, note that Blum (2003) has identified evidence for intermediate-age populations in the Galactic Center. In order to infer the past star formation rate, I modeled the observed surface number density of stars (Figure 3 ) as a function of star formation history using the Geneva stellar evolution models, assuming a range of power-law initial mass functions (IMFs), metallicity, and wind mass-loss rates. I considered: an ancient burst, episodic bursts, continuous formation, and combinations of these three. I used these models to produce synthetic luminosity functions for comparison with HSTDJICMOS data. The surface number density in the observed luminosity functions has been set by dividing the number of stars per half magnitude bin by the area of the observations. The NICMOS fields vary in location from about 15 pc from the G C to 55 pc, and the sample was culled of forground and background stars by limiting the data to stars with l.O
-

Acknowledgements I acknowledge very useful discussions with Paco Najarro, Bob Blum, Laurant Sjouwerman, Mike Rich, Mark Morris, Sungsoo Kim, and Jay Frogel.

References Allen, D. A,, Hyland, A. R. & Hillier, D. J. 1990, MNRAS , 244, 706 Allen, D.A., 1994, in The Nuclei of Normal Galaxies, eds. R. Genzel& A. I. Harris (Dordrecht: Kluwer). 293 Becklin, E. E., & Neugebauer, G., 1968, ApJ, 151, 145 Blum, R. D., Depoy, D. L., & Sellgren, K. 1995, ApJ, 441,603 Blum, R. D., Schaerer, D., Pasquali, A., Heydari-Malayeri,M., Conti, P. S., & Schmutz, W. 2001, AJ, 122, 1875 Blum, R. D. 2003, these proceedings Chiar, J. 2003. these proceedings Cotera, A. S., Erickson, E. F., Simpson, J. P., Colgan, S. W. J., Allen, D. A,, & Burton, M. G. 1992, American Astronomical Society Meeting, 18 1, 8702 Cotera. A. S. 1995, Ph.D. Thesis, Stanford University Cotera, A. S., Erickson, E. F., Colgan, S. W. J.. Simpaon, J. P., Allen, D. A,, &Bunon, M. G. 1996, ApJ, 461, 750 Cotera, A. S., Simpson, J. P., Erickson, E. F., Colgan, S. W. J., Burton, M. G., & Allen, D. A. 1999, ApJ, 510, 747

:

D. Finer: Massive Stars

260

jl;:o;li

Star Formotion History

h

" f 0.10 b

1.o

0.1

0.01 107

ffl

Zsun,mossd IOSS=C) Luminosity Function L ( (metolicity=t.O ; g

.

5

10'0

109

108

time

l

!

10'

107

ffl

5

10'0

I 09

lime

j

2

h

10 KO or

15

20

15

20

mm,o

10

KO of m m . 0

?

;

.

"

t

0.10

t

0.01 107

0.1

10'

I 09

10'0

5

I0 or mrm*o

15

20

5

ID KO or mnowp

15

20

lime

1

~lo.oo&~ 1.00

d;

:

1@

I 09

'

l

I

" !! 0.10

e

ffl

0.1

0.01 107

lime

10'0

Fig. 4 Model luminosity functions (right, bold) and observed luminosity functions (righr, light) for various star formation histories in the Galactic Center (left). The models use the Geneva isochrones with solar ahundances and canonical mass loss rates.

Cox, P. & Laureijs, R. 1989, IAU Symp. 136: The Center of the Galaxy, 136, 121 Eckart, A,, Genzel, R., Hofmann, R. Sams, B. J., & Tacconi-Garman, L. E., 1995, ApJ, 445, L26 Eckart, A., Oft, T., & Genzel, R. 1999, A&Ap, 352, L22 Eckart, A,, Genzel, R., Ott, T., & Schodel, R. 2002, MNRAS, 331,917 Figer, D. F., Najmo, F., Morris, M., McLean, I. S., Geballe, T. R., Ghez, A. M., & Langer, N. 1998, ApJ, 506,384 Figer, D. F. et al. 2000, ApJ, 533, L49 Figer, D. F., Moms, M., & McLean, I. S. 1996, The Galactic Center, San Francisco: ASP 102, 263 Figer, D. F., McLean, I. S., & Morris. M. 1995b, ApJ, 447, L29

Astron. Nachr./AN 324, No. SI (2003)

26 I

Figer, D. F. 1995a, Ph.D. Thesis, University of California, Los Angeles Figer, D. F., Kim, S. S., Morris, M., Serahyn, E., Rich, R. M., & McLean, 1. S. 1999c, ApJ, 525, 750 Figer, D. F., McLean, I. S., & Morris, M. 19994 ApJ, S 14, 202 Figer, D. F., Moms, M., Gehalle, T. R., Rich, R. M., Scrabyn, E., McLean, I. S., Puetter, R. C., & Yahil, A. 1999b. ApJ, 525, 759 Figer, D. F. 2002a, in proceedings of IAU symposium 212 on “A MASSIVE STAR ODYSSEY, FROM MAIN SEQUENCE TO SUPERNOVA,” 487 Figer,D. F. 2002b, ApJ, 581,258 Forrest, W.J., Shure, M.A., Pipher, J.L., Woodward, C.A., 1987, in AIP Conf. 155, The Galactic Center, ed. D.Backer (New York AIP), i53 Frogel, J.A., Ticde, G.P., & Kuchinski, L.E. 1999, AJ. 117, 2296 Geballe, T.R., Figer, D.F., & Najarro, F. 2000, ApJ, 530,97 Genzel, R., Thatte, N., Krahbe, A., Kroker, H. & Tacconi-Garinan 1996, ApJ, 472. 153 Ghez, A. M., Kremenek, T., Tanner, A., Morris, M., & Becklin, E. 2001, Black Holes in Binaries and Galactic Nuclei, 72 Glass, I. S., Catchpole, R. M., & Whitelock, P. A. 1987, MNRAS, 227, 373 Haller, J. W. 1992, Ph.D. Thesis, The University of Arizona Krabhe, A., Genzel, R., Drapatz, S . & Rotaciuc, V. 1991, ApJ, 382, L19 Krabhe, A. et al. 1995, ApJ, 447, L95 Lang, C. C., Goss, W. M., & Rodriguez, L. F. 2001, ApJ, 551, L143 Lang, C. C. 2003, these proceedings Law, C. 2003, these proceedings Lehofsky, M. J., Rieke, G. H., & Tokunaga, A. T. 1982, ApJ, 263, 736 Moneti, A., Glass, I. S., & Moorwood, A. F. M. 1994, MNRAS, 268, 194 Moneti, A., Stolovy, S., Blommaert, J . A. D. L., Figer, D. F., Najarro, F. 2001, A&A, 366, 106 Monnier, J. D., Tuthill, P. G., & Danchi, W. C. 2002, ApJ, 567, L137 Monnier, J. D., Tuthill, P. G., & Danchi, W. C. 1999, ApJ, 525, L97 Nagata, T., Woodward, C. E., Shure, M., Pipher, J. L. & Okuda, H. 1990, ApJ, 351, 83 Nagata, T., Woodward, C. E., Shure, M., & Kobayashi, N. 1995, ApJ, 109. 1676 Najarro, F., Hillier, D. J., Kudritzki, R. P., Krdbhe, A., Genzel, R., Lutz, D., Drapatz, S. & Gehalle, T. R. 1994, A&A, 285,573 Najarro, F., Gabbe, A., Genzel, R., Lutz, D., Kudritzki, R. P., & Hillier, D. J. 1997, A&A, 325, 700 Najarro, F. 2002, in proceedings of IAU symposium 212 on “A MASSIVE STAR ODYSSEY, FROM MAIN SEQUENCE TO SUPERNOVA,” 487 Okuda, H, Shibai, H., Nakagdwd, T., Matsuhara, H., Kobayashi, Y.. Kaifu. N., Nagata, T., Galley, 1. & Gehalle, T. R. 1990, ApJ, 351,89 Paumard, T., Maillard, J. P., Moms. M., & Rigaut, F. 2001, A&A, 366,466 Rieke, G. H. & Lebofsky, M. J. 1982, AIP Conf. Proc. 83: The Galactic Center, 194 Rieke, G. H., & Rieke, M. J., 1994, in The Nuclei of Normal Galaxies, ed. R. Genzel & A. I. Harris (Dordrecht: Kluwer), 283 Salpeter, E. E. 1955, ApJ, 121, 161 Scalo, J. 1998, in The Srellur Initial Muss Function, G. Gilmore and D. Howell (eds.), vol. 142 of 38LhHerstrnoncem Conference, San Francisco: ASP, 201 Serahyn, E., & Morris, M. 1996, Nature, 382, 602 Sjouwerman, L. 0..Hahing, H. J., Lindqvist. M., van Langeveldc, H. J., & Winnherg, A. 1999, ASP Conf. Ser. 186: The Central Parsecs of the Galaxy, 379 Stoke, A., Grehel, E. K., Brandner, W., & Figer, D. F. 2002, A&A, 394,459 Tamhlyn, P., Reike, G., Hanson, M., Close, L., McCarthy, D., Reike, M. 1996, ApJ, 456, 206 Ynsef-Zadeh, F., Law, C., Wardle, M., Wang, Q. D., Fruscione, A., Lang, C. C., & Cotera, A. 2002, ApJ, 570,665

This Page Intentionally Left Blank

Astron. NachrJAN324, No. SI,263-269 (2003)/ DO1 10.1002/asna.20038S108

The Galactic Center Source IRS 13E: a Star Cluster Jean-Pierre Maillard’, Thibaut Paumard’,Susan Stolovy’, and Franqois Rigaut3

’ Institut d’Astrophysiquede Pans (CNRS),98b Blvd Arago, 75014, Paris, FRANCE SIRTF Science Center, CalTech, MS 220-6, Pasadena, CA 91 125, USA

’ Gemini North Headquarter,Hilo, HI 96720, USA

Key words Galactic Center, star cluster, WR star, infrared. adaptive optics

Abstract.High spatial resolution, near-infrared observations of the Galactic Center source, close to Sgr A * , known historically as IRS 13, are presented. These observations include ground-based adaptive optics images in the H, K’ and L hands, HST-NICMOS observations in filters between 1.1 and 2.2 pm, and spectroimaging data in the He1 2.06pm line and the Bry line. Analysis of all these data has made possible the resolution ofthe main component, IRS 13E, into a cluster of seven individual stars within a projected diameter of 0.5” (0.02 pc), and to build their SED. The main sources, 13E1, 13E2, I3E3 (a binary), and 13E4, are hot stars of different nature. 13E2 and 13E4 are emission line stars. The spectral type of the various members goes from 051 to WR, including dusty WRs like IRS 21 (Tanner et ul. 2002). All these sources have a common westward proper motion. Two weaker sources, 13ES and 13E6, are also detected within the compact cluster, with 13E5 proposed as another dusty WR and 13E6 as a O W star. An extended halo seen around the cluster, part of the mini-spiral of dust is particularly enhanced in the L band. It is interpreted as a contribution of the scattered light from the inner cluster and the thermal emission from the dust. IRS 13E is proposed to be the remaining core of a massive, young star cluster which was disrupted in the vicinity of Sgr A*, and hence, the possible source of the young stars in the central parsec, from the helium stars to the s stars.

-

1 Introduction I n the early mapping works of the central parsecs, a spot named IRS 13, bright at all near-infrared wavelengths, was reported, located approximately 3.6” south-west of Sgr A*. It was later resolved into two sources in the K band separated by 1.2”,IRS 13E and IRS 13W (Simon et al. 1990). From spectroscopic studies in the same band, IRS 13W was identified as a cool star (Krabbe et al. 1995) and IRS 13E as an emission line source with strong He I 2.058, 2. I 12ptn, Bry line and other Brackett lines up to Br12 (Genzel et al. 1996),typical of the helium stars present in the central parsec. The first adaptive optics (AO) map of IRS 13E in the K band obtained on the CFH Telescope was published by Paumard et al. (2001 ), showing that the source resolved into two equally bright components 13E1 and 13E2, plus a third weaker component called 13E3. Since the spectra oC IRS 13E does not have the same spatial resolution as the images, the identification of the associated spectral type is subject to caution. In the centimetric domain, Zhao and Goss (1 998) found IRS 13 as the brightest radio continuum source after Sgr A* at the Galactic Center. The detection of a discrete X-ray source from CHANDRA at the position of IRS 13 (Baganoffer nl. 2001) was another element making 1RS 13 a source of special interest. The high resolution images of the central parsec currently obtained at various wavelengths in the infrared provide the possibility of studying this peculiar Galactic Center source in detail. A complete description of the present work can be found in the companion paper of Maillard ef al. (2003).

-

02003 WILEY-VCH

Verlag GnibH & Co KGaA. Weinhem

264

J. P. Maillard et al.: The IRS 13 cluster

2 New data on IRS 13 Calibrated ground-based A 0 data from several telescopes and space-based NICMOS data in the near infrared, all containing IRS 13 in their field, plus some spectroscopic data, and the proper motions of the sources of the IRS 13 field, have all been combined. 2. I

High-angular resolution data

The A 0 data come from two different systems, Gemini North (Graves et ul. 1998) for the H and the Kp (2.12pm, FWHM 0.41 pn) bands, and ESO 3.6-m telescope (CICnet et al. 2001) for the L band. Medium (M) and wide-band (W) filters, respectively centered at 1.1, 1.45, 1.60, 2.22 pm (coded F11OM, F145M, F160W and F222M) and two close narrow-band (N) filters (F187N centered on the 1.87pm Paa line and F190N) were used in observing the stars at the inner parsec of the Galactic Center with the NICMOS cameras on board HST. A small N 2.5” x 2.5”region of the image, roughly centered on IRS 13E, was analyzed for each filter. All these high-resolution, multi-band images have provided spectrophotometric information on the IRS 13 sources and its environment, from 1 to 4 pm. The Gemini A 0 data were calibrated by linear interpolation based on one bright, hot star of the field under study, from the calibrated NICMOS data, between the F160W and F190N photometry for the H band, F190N and F222M for the Kp band. The star detection and photometry was made with the SturFinder procedure (Diolaiti et al. 2000) for all the images. For the A 0 images a deconvolution code called MCS (Magain et uE. 1998) was applied. For the H and the Kp images the width of the synthetic PSF was equal to 0.040” and to 0.192” for the L-band image, i.e. a gain in resolution respectively of a factor 4.5 in H, 4.3 in Kp and 1.5 in L. 2.2 Spectroscopic data The only high-spatial resolution images giving spectroscopic information are contained in the NICMOS narrow-band images from the F187N filter. By subtraction of F190N, a narrow-band filter in the nearby continuum, from Fl87N, a map of the 1.87pm Pacv emission was obtained (Stolovy et al. 1999). This map shows the distribution of the ionized gas and stellar spots from Pacv emission in the atmosphere of the hot stars. The Bry and 2.06pm He1 line profiles at IRS 13E from BEAR spectro-imagery, an imaging ITS (Maillard 2000), were used as complementary information to help define the spectral type of the underlying stars. The IRS 13 complex is located in a region of intense interstellar emission. The data cube was particularly useful to correct the two emission line profiles from the interstellar emission, leaving fully resolved stellar profiles respectively at 21.3 and 52 km spl resolution. 2.3

Proper motions

The proper motions of the IRS 13 sources and the sources contained in the surrounding 2.5”x2.5” were obtained from Ott et al. (2003) who conducted an analysis of ten years of SHARP data (Eckart et al. 1995), providing more than 1000 proper motions in the central parsec.

3 Results From the deconvolution analysis of the 2.5”x2.5” field including IRS 13, 20 individual sources were identified. IRS 13E is decomposed into seven sources, respectively named 13E1, 13E2, 13E3A and B, 13E4, 13E5 and 13E6. The name 13E3A and B is proposed for the two components of the source 13E3 which appears double only after deconvolution, in the H and Kp bands. The positions of all the sources and their proper motions, estimated for most of them, are given in Fig. 1. Their observed photometry in the H, K and L bands is presented in Table 1. In the F187N - F190N image only 13E2 and 13E4 are remaining, indicative that these two sources are emission stars.

265

Astron. Nachr./AN 324. No. SI (2003)

t

1

0.5

t 0.0

i

0.5 1.0 1.5 2.0 dacos6 (arcsec)

2.5

Fig. 1 The IRS 13 field with the star identification. The vector associated with most of the stars represents the amplitude in velocity and the direction of proper motion measured from SHARP data by Ott et a/. (2003). For E3A and E3B only the harycenter proper motion is determined. The origin of the field corresponds to -2.19"W and -3.00"s with respect to Sgr A*.

From the flux measurements made in 8 bands, bctween 1.1 and 3.5 pm, it was possible to obtain dereddened spectrophometry of the IRS I3 sources. This requires the adoption of an A\, value and an extinction law. The mean value of A v =31.1 is known to vary strongly with location within the central parsecs (Scoville et al. 2003). The local value was derived utilizing two constraints: IRS 13W is known to be a cool oxygen star, and IRS 13E2 and 13E4 are hot stars, as seen in Pan imaging. A value of Av = 35 was adopted and was assumed to be valid over the small field around IRS 13. A" could be pretty well constrained assuming that IRS 13E2 is a WR star with T,ff > 25,000K. The IRS 13E2 spectral type is based on the Pacv imaging, but also on Fabry-PCroc spectro-imaging behind the CFHT-A0 system where IRS 13E2 is detected as the only source of the broad 2.06 p m He I emission line (CICnet et ul. 2003). With rhe spectral range under study, from about I 10 4pm, beyond a temperature of 25,000K we are in the Rayleigh-Jeans regime, and the shape of the SED becomes constant in a log F ( X ) versus X diagram. A v can be adjusted to bring the data points parallel to the SED. However, the fit of the dereddened data has to be made as the sum of two black-body curves since most of the sources have an infrared excess, a signature of thermal dust emission. The adjuslments is made with four parameters for each IRS 13 star, by C o r f l x BB(T1) Cocf2 x BB(T2).The TI temperature being the high-temperature component, is mainly determined by the data points between 1 and 2.5 pm,and Tz by the 2 to 4 pm points. If for the fitting, TI becomes 2 25,000 K the temperature is fixed at 25,000K. The four final parameters are presented in Table 2. The dereddened points and the fits are shown on Fig. 2. Several stars are very hot stars, i.e. with

+

J. P. Maillard et al.: The IRS 13 cluster

266

a T,ff 2 25,000 K. The NICMOS FI 1 OM and the L data are essential to provide the maximum constraints for the SED of each source.

4 Nature of the IRS 13E sources From the main results reported in the previous section an identification of the spectral type of the seven components of IRS 13E can be derived. Table 1 H, K and L photometry of the IRS 13E cluster and the nearby field stars __ H K L H-K K-L ID __ -

W El E2 E4

5 6 7 8 E3A 10 E3B 12 13 14 15 16

E5 18 E6 20

14.51 12.71 13.02 14.34 14.28 14.68 14.38 14.82 16.44 15.51 16.93 15.89 15.91 15.91 16.61 16.62 18.15 16.77 16.20 17.26

11.22 10.90 10.95 11.64 11.82 12.05 12.10 12.14 12.38 12.71 12.98 13.19 13.26 13.33 14.12 14.13 14.20 14.43 14.48 14.70

8.92 8.59 7.73

7.50 7.92

8.48

3.29 1.81 2.07 2.70 2.47 2.63 2.28 2.68 4.06 2.80 3.95 2.69 2.65 2.58 2.49 2.49 3.95 2.34 1.73 2.56

2.29 2.3 1 3.21

4.89 5.06

5.71

Table 2 Fitting parameters of the SED of the WS 13E sources and IRS 13W The spectral type of each source as discussed in Sect. 5 is summarized in the last column

Star 1 W I El E2 E4 E3A E3B E5 E6 __

Coefi 23.00 0.500 0.450 0.070 0.460 0.375 0.070 0.008

TIK 2600 225000 >25000 2 25000 3800 3800 6000 2 25000

ndusty Wolf-Rayet star

Coefz Tz K 1700 650 12000 550 40000 550 45 1550 33000 610 29000 580 9800 630

SP. Type MSIII 051 wc9 05IIIe d. WR" d. WR d. WR OSV

Astron. Nachr./AN 324, No. S I (2003)

267

10- 4

10-5

4

I

E

y3 10 -6

E

;

10-8

Fig. 2 Dereddened flux in Wcm-' pm-' for A v =35. The top of the arrows represents the upper limit of the detectable flux in the 1.1 pm and the L-band filters. The various lines represent the best fitting between 1 and 4 p m of the data points from a two-component model with the parameters of Table 2.

4.1

13E1.13E2and 13E4

The source 13E1 is a bright, blue star, but with no detected emission at Pati. From its luminosity and its

T,ff, IRS 13E1 i s proposed to be close to a 051 main sequence star. I3E2 and 13E4 are two emission line stars, 13E2 being brighter in Paa by a factor 2. From the BEAR data, the 2.06 pm He I line is a broad line 900 km s c l FWHM) while the Bry line is narrow (- 215 km s-' FWHM). 13E2 is reported as the only He I emitter (CICnet et aZ. 2003). As a broad-line, helium-rich star, 13E2 is proposed as a Wolf-Rayet type star, from the criterion on the linewidth developed in Paumard et al. (2001). By analogy with similar stars in the central parsec the source should be more precisely a WC9 star. 13E4 is a blue star which shows a narrow emission line in Paa but no helium emission line. Therefore, this star is much less evolved than 13E2. It can be proposed as a 05IIIe, since it is weaker than El and has hydrogen lines in emission.

(-

J. P.Maillard et al.: The IRS 13 cluster

268

4.2

13E3A, 13E3B and 13E5

On Fig. 2 these three stars (dash-dotted lines) have a similar SED. They are adjusted by a strong, cool component at E 600 K (Coef, is high) and a weak, hotter component ofa few thousands K. This adjustment can be compared to the fitting of the SED of IRS 21 (Tanner et ul. 2002) fitted by a two-component model, the near-infrared scattered light from the ccntral source peaking at N 3.8 pm (760 K), and the mid-infrared re-emitted light from a dust shell at 250 K. Tanner et al. (2003) conclude that this source is a dusty WR star, experiencing rapid mass loss, and the other red, featureless spectrum sources along the Northern Arm IRS 1W, 2, 3 , 5 , and low, as well. By analogy, we conclude that the three red sources within the IRS 13E complex, also located in the dusty part of the mini-spiral are dusty WR stars. N

4.3

13E6

The source IRS 13E6 is another blue star, much weaker than 13E1, l3E2 and 13E4, with Kmng = 14.5 (Table 1). The image in the L band is not deep enough to detect it at this wavelength to confirm that this star is also embedded in the same concentration of dust than the other IRS 13E sources. From its color and magnitude IRS 13E6 can be considered as close to a 0 5 V type star. Without further indication, only from the fact that all the other stars in IRS 13E are hot stars, we assume that IRS 13E6 belongs also to the same complex.

5 Model of IRS 13E as the remaining core of a massive star cluster IRS 13E appears as only composed of hot, massive stars, with at least 7 stars within 0.5”. The common direction and comparable amplitude of the proper motions of the main components is a decisive argument to indicate that 13E1, 13E2, 13E3A/B and 13E4 are physically bounded. The source previously called IRS 13E is likely a compact star cluster. Furthermore, its composition suggests a young star cluster of a few x lo6 yr old, since several members are identified as having already reached the WR stage. The presence of such a compact cluster with a limited number of members raises the question of its origin. First, it can be noticed that each component has many other examples of stars of the same spectral type in the central parsecs. However, the large abundance of massive stars, which are very rare elsewhere in the Galaxy, remains one of the major mysteries of this region of the Milky Way. Since star formation would be difficult due to the strong tidal forces from the Sgr A* black hole, Gerhard (2001) made the interesting hypothesis that the central parsec He1 stars, the most prominent of the massive young stars, might be the remains of a dissolved young cluster, disrupted in the vicinity of the central black-hole. Kim et al. (2003) tested this idea for different cluster masses and different initial orbit radii. They came to the conclusion that some simulations can be regarded as possible candidates for the origin of the central parsec cluster. With its exceptional concentration of massive stars, very close to Sgr A*, all bounded together, we propose that IRS 13E might be the remaining core of such a massive cluster which was disrupted by Sgr A*. The analysis of the 12 other stars identified in the IRS 13 field (Fig. I ) conducted as for the IRS 13E cluster sources, made possible to separate the sources in two categories, 9 red stars (T,ff from 2800 to 5000K) and 3 blue stars (T,~J.2 25,000K). The red stars are members of the most numerous population of the central parsecs, which is an old population of K, M and AGB stars, to which belongs also IRS 13W. The blue stars should be members of the most recent stellar population. On the other hand, the blue stars are comparable in magnitude and color to the stars of the S-cluster (Gezari et al. (2002) detected around Sgr A*. The IRS 13Ecluster itself contains also one of such lower mass blue stars (Table 2). Hence, the hot stars, including the S and the helium stars, could come from the same initial massive cluster, and complete its IMF. However, more simulations are needed to validate this hypothesis. Another aspect of IRS 13 is the detection of a discrete X-ray emission within 1” positional accuracy (Baganoff et al. 2001). IRS 13E as a star cluster might be the X-ray source, by the colliding winds of all the close, hot, mass-losing stars. An example of such a source can be provided by the detection of a

Astron. Nachr./AN 324, No. S I (2003)

269

discrete X-ray source at the position of the core of the Arches cluster (Yiisef-Zadeh et al. 2002). A better astrometry of the X-ray source at IRS 13 could help confirm this assumption, consistent with IRS 13E as the remaining core of a massive star cluster.

References Baganoff, F.K., Maeda, Y., Moms, M., Bautz, M.W. Brandt, W.N., Cui, W., Doty, J.P., Feigelson. E.D., Garmire, G.P., Pravdo, S.H., Ricker, G.R., & Townsley, L.K. 2001, ApJ, astro-ph/01002151 CICnet, Y., Rouan, D., Gendron, E., Montri, J., Rigaut, F., Lena, P. & Lacombe, F. 2001, A&A, 376, 124 CICnet, Y., Lacombe, F., Gendron, E., & Rouan, D. 2003, these proceedings Diolaiti, E., Bendinelli, 0..Bonaccini, D., Close, L., Currie, D. & Parnieggiani, G. 2000, A&AS, 147, 335 Eckart, A,, Genzel, R., Hofmann, R., Sams, B.J.. & Tacconi-Garman, L.E. 1995, ApJ, 445, L23 Genzel, R., Thatte, N., Krabbe, A., Kroker, H., & Tacconi-Garman, L.E. 1996, ApJ, 472, 153 Gezari, S., Ghez, A.M., Becklin, E.E., Larkin, J., McLean, I.S., Moms, M. 2002., ApJ. 576,790 Gerhard, 0.2001, ApJ, 546, L39 Graves, J.E., Northcott, M.J., Roddier, F.J., Roddier, C.A., &Close, L.M. 1998, SPIE, 3353, 34 Krabbe, A., Genzel, R., Eckart, A., Najarro, F., Lutz, D., Cameron, M., Kroker, H., Tacconi-Garman, L.E., Thatte, N., Weitzel, L., Drapatz, S., Geballe, T., Sternberg, A,, & Kudritzki, R.P. 1995, ApJ, 447, L95 Kim, S.S., Morris, M., & Figer, D.F. 2003, lhese proceedings Maillard, J.P., 2000, in Imaging the Universe i n 3 Dimensions, ed. E. van Breughel & J. Bland-Hawthorn, ASP Conf. Ser.. 195, 185 Maillard, J.P., Paumard, T., Stolovy, S., & Rigaut, F. 2003, A&A, submitted Magain, P., Courbin, F. & Sohy, S. 1998, ApJ, 494,472 Ott, T., Genzel, R., Eckart, A,, & Schodel, R. 2003, these proceedings Paumard, T., Maillard, J.P., Moms, M., & Rigaut, F. 2001, A&A, 366,466 Simon, M., Chen, W.P., Forrest, W.J., Garnett. J.D., Longmore, A.J., Gauer, T., & Dixon, R.I. 1990, ApJ, 360, 95 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher, M., & Yusef-Zadeh, F. 2003, rhese proceedings Stolovy, S.R., McCarthy, D.W., Melia, F., Rieke, G., Rieke, M.J., & Yusef-Zadeh, F. 1999, in The Central Parsecs of the Galaxy, ed. H. Falcke, A. Cotera, W.J. Duschl, F. Melia, M.J. Rieke, ASP Conf. Ser., 186, 39 Tanner, A., Ghez. A.M, Morris, M., Becklin, E.E., Cotera, A., Ressler, M.. Werner, M., & Wizinovitch, P. 2002, ApJ, 575, 860 Tanner, A,, Ghez, A.M, Morns, M., & Becklin, E.E. 2003. these proceedings Yusef-Zadeli, F., Law, C . ,Wardle, M., Wang, Q.D.. Fruscione, A., Lang, C.C., & Cotera, A. 2002, ApJ, 570, 665 Zhao, J.H., & Goss, W.M. 1998, ApJ, 499, L163

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S 1,27I- 277 (2003)/ DO1 10.1002/asna.200385067

X-ray Emission from Stellar Clusters Near the Galactic Center Casey Law* and Farhad Yusef-Zadeh Department of Physics and Astronomy Northwestern University 2 145 Sheridan Road Evanston, 11. 60208

Key words Galaxy: center, X-rays: galaxies: clusters, stars: Wolf-Rayet, galaxies: clusters: individual (Arches, Quintuplet) Abstract. With the recent detection of X-ray emission from the Arches cluster, we examine the existing Chandru observations of the Galactic cenler region to determine X-ray properties of known stellar clusters and near-IR cluster candidates. We discuss the first detection of X-ray emission from the Quintuplet cluster, and find its emission characteristics very different from those of the Arches cluster. The X-ray and IR characteristics of these well-known clusters are used to discern if the candidate star clusters are indeed massive, Galactic center star clusters.

1 Introduction The Galactic center (GC) region is host to some spectacular stellar clusters. The Arches, Quintuplet, and IRS 16 clusters are all young and among the most dense in the galaxy. The distribution of these clusters within the central 30 pc of Sgr A* may not be coincidental. The ambient conditions of the GC region are known to be extreme (e.g., high gas and stellar densities, strong magnetic fields, intense ionizing flux; Morris & Serabyn, 1996) and should have a significant effect on the process of star formation. The detection of X-ray emission from the Arches cluster suggests that the collision between individual stellar winds can be responsible for the observed emission, as had been modeled specifically for the Arches (Canto et al. 2000; and later confirmed by Raga et al. 2001). There is also the possibility that cluster wind emission may be a unique signature of massive star clusters. Recent theoretical and observational studies have suggested that there may be many more, undiscovered, massive star clusters in the GC region. N-body simulations of these compact young clusters (CYCs) have suggested that star clusters such as the Arches dynamically evolve rapidly, dissolving into the stellar background within 10-20 Myr (Portegeis-Zwart et al. 2001; Kim et al. 1999). Thus i t was suggested that there may be another lO-SO CYCs within the central IS0 pc at various stages of dissolving into the projected background stellar field. An observational test of this idea by Dutra & Bica (2000,2001 ; hereafter DBOO and DBOI) searched 2MASS images for Arches-like objects, in order to make a list of candidate G C clusters. Their results are consistent with rough estimates of the number of CYCs from Portegeis-Zwart et al. (2001). This paper discusses the use of X-ray and IR properties of the Arches and Quintuplet clusters to search for new clusters in the central degree. In 52, the current Chandra observations of the Arches and Quintuplct clusters is discussed (X-ray emission from IRS 16 is discussed in Baganoff et al. 2001). Section 3 discusses our analysis of the candidate star clusters of Dutra & Bica, and the X-ray and IR observations can be used * Corresponding author: e-mail: [email protected]

@ 2003 WILEY-VCH Verlag CmbM B Co KGaA. Weinhem

C. Law and F. Yusef-Zadeh: X-rays from GC Star Clusters

272

to comment on their status as new, massive, G C star clusters. We conclude by commenting on the use of X-ray surveys to find new, massive star clusters in the GC.

2 X-ray Observations of Known Stellar Clusters 2.1 The Arches Cluster The Arches cluster is one of the densest and most massive star clusters in the Milky Way and is located 30 pc in projection from the Sgr A* (Cotera et al. 1996). Figure 1 shows X-ray and IR images of the cluster. All Chandra data of the region were combined for this analysis, including ObsID 945 (50 ks of exposure), 2276 (1 1 ks), and 2284 ( 1 1 ks). The first observation of the cluster found three sources: “Al” in the cluster core, “AT’ about 10” north of the core, and the diffuse halo “A3” emission (Yusef-Zadeh et al. 2002, hereafter Y02). A later observation had better resolution, and resolved the “AI” source into a northern and southern component. We found that Al N and A l S were coincident with ionized stellar wind sources seen in radio observations (sources AR4 and ARl of Lang et al. 2001). After aligning the X-ray sources with these radio sources, we correlated the X-ray sources with several previous IR observations, as summarized in Table 1.

Fig. 1 (Lejfft): Adaptively smoothed image of all Chandm observations of the Arches star cluster. (Right): HSTINICMOS image (Figer et al. 1999b) wilh X-ray contours overlaid. The X-ray contours were taken from the highest resolution image of the region (ObslD 2276), which split source A1 into a north and south component.

The spectra of the A1 and A2 sources may he fit with a two-temperature absorbed model with kT1 ,-., 0.7 keV and rCT2 5 keV. No significant difference can be detected between the fits of the A1 sources and A2. A one-temperature model has difficulty fitting both the low energy (He-like sulfur) and high energy (Helike iron) spectral lines. However, a one-temperature/powerlaw model can fit as well as a two-temperature ergs model. All models give an absorption-correctedx-ray luminosity, L,(O.5 - 8keV) N 0.5 - 3 x s-’ for the cluster. N

Table 1 Cross-identificationof X-ray sources of the Arches cluster with previous studies. Translation of abbv.: Y02, Yusef-Zadeh et al. (2002); BOl, Blum et al. (2001); C96, Cotera et al. (1996); N95, Nagata et al. (1995).

Y02name Designation A1 N CXOGC 5174550.5-2849 18 A1 S CXOGC J 174550.4-28492 1 A2 CXOGC J174550.2-284911

LGROI AR4 ARl -

B01 21 23 26

C96

5 8 I

N95 10 8 5

The highest resolution observation of the region shows three compact sources with
Astron. Nachr./AN 324. No. S 1 12003)

273

in the radio. These correlations suggest the X-ray and IR emission have a similar origin (e.g., colliding wind binaries), as opposed to diffuse emission resulting from the cluster wind collision. However, there is evidence of nonthermal diffuse radio emission from the core of the Arches cluster (Yusef-Zadeh et al. 2003), consistent with a shock acceleration mechanism. The diffuse component, “A3,” was originally suggested as a candidate “cluster wind” (Y02), although it may also be caused by scattered, fluorescent emission. The X-ray spectrum may be fit by a power law with a Gaussian component at 6.4 keV. This model fits the well known mechanism, whereby radiation is scattered from molecular material that fluoresces strongly in the iron K a line at 6.4 keV. While a thermal model can tit the emission, the equivalent widths of the thermal lines are consistent with the ambient GC levels. A recent study by Lang, Goss, & Morris (2003) has identified a possible source of molecular material along our line of sight, near the Arches cluster. This model for the A3 emission could be confirmed with an accurate measurement of the column density of the cloud and its distance from the star cluster (e.g., Sunyaev & Churazov, 1998). If A3 is scattered emission, it may confuse any search for diffuse emission from the cluster itself. 2.2

The Quintuplet Cluster

The second massive GC star cluster i n our study i5 the Quintuplet cluster. first identified in the 1R by its five most luminous members (Nagata et al. 1990). The Quintuplet is less massive, less dense, and older than the Arches cluster, suggesting it has bcen dissolved by GC tidal forces (Kim et al. 1999). Here the correlation of the X-ray and IR point sources is discussed, and we examine the spatial dependence of the X-ray emission.

Fig. 2 (L<j?:filChudrci X-ray imagc of all observations toward the Quintuplet cluster convolved with a IYS Gaussian. Ellipscs show X-ray detected sources with potential IR correlations. (Kighr) HST/NICMOS image of the Quintuplet cluster, with the same regions identified.

Figure 2 shows a comparison of the X-ray cmission to HSTINICMOS data (provided by Don Figer). This region was observed by Ch~rndruon Lhree occasions: ObsID 945 (49 ks), 2273 (1 I ks), and 2276 ( 1 1 ks). The X-ray emission is predominantly pointlike, in contrast to the Arches, but there is also diffuse emission apparent in Figure 2. The easternmost X-ray source is seen within 1” of source 21 I of FMM99, a

C. Law and F. Yusef-Zadeh: X-rays from GC Star Clusters

214

“Quintuplet proper” member. FMM99 classify this source as a possible dusty, late-type WC star (DWCL). From east to west, the other two potential X-ray/IR correlations are aligned with FMM99’s 257 and 242, a B01 supergiant and an unclassified star, respectively. Each of these coincidences is less certain, due to the density of IR sources in the cluster core and typical errors in absolute astrometry of Chandru data (-1’’ ). There is possibly a fourth X-ray source at 17h46”’14~77,-28049’ 47” (J2000), within 1’’ of source 231, which is likely to be a DWCL. The X-ray source is faintly visible on the merged dataset, but not significantly detected on any single observation (making the calculation of a detection confidence non2x ergs s-’, consistent with trivial). These sources have absorption corrected luminosities, L, IR-defined spectral types. X-ray spectra from three regions near the Quintuplet cluster were fit to examine the spatial dependence of the emission. We extracted spectra from the core of the cluster (-30” across, which includes the possible X-ray/IR correlations), the diffuse emission near the cluster, and finally a background region, located -1’ south of the cluster. We see 6.4keV emission in the spectra of the outer two regions, but at levels typical of the GC region and much less than seen towards the Arches. As seen in Table 2, there is a gradient in the gas temperature and surface brightness from the core of the cluster outward. These gradients are consistent with a cluster wind model, in which the wind collisions produce a lo7 K gas in the core, which adiabatically cools as the gas escapes the cluster. However, this analysis does not account for complications such as confusion with ambient GC emission, dust scattering, and unresolved point sources. N

Table 2 Temperature and surface brightness for three regions near the Quintuplet cluster. All regions fit simultaneously with an absorbed thermal model, assuming 2 x solar abunddnces and a single N H for all three regions. Best fit N H = 7.1 i0.5.

Region core halo background

S,(O.5 - 8 keV) [ergs cm-’ s-’ sr-l] 7.1 x lo-’ 4.2 x 10-5 3.6: x 10-5

kT [keV] 2.0 i0.3 1.6 i 0.2 1.4

0.1

In notable contrast to what is seen in the Arches cluster, radio continuum studies by Lang et a]. (1999) find no ionized stellar wind sources within 1” of any Chandru-detected sources. In total, the X-ray flux from the Quintuplet’s point sources is roughly 1150th of that in the Arches. This is larger than the scaling of their masses and IR luminisoty (1:3) and less than the scaling of their core stellar densities (1:400) (FMM99).

3 Candidate Clusters Our goal in this analysis is to use our knowledge of the Arches and Quintuplet to search for new candidate star clusters located near the GC. Our study of the Arches and Quintuplet clusters has given us general 1034-35ergs s-’, diffuse and pointlike emission, highly characteristics with which to start (e.g., L, absorbed). We have also suggested the cluster wind model describes the observed X-ray characteristics, giving us a possible model to aid us in our search. The study of Dutra & Bica searched for massive stellar clusters-using the Arches cluster as a templateby inspecting 2MASS J , H , and IC, images. Nine of their forty candidates lie within the central 2“ x0?8 covered by Chundru’s Galactic Center Survey (Wang ef al. 2002), as shown in Figure 3 . In the following sections, we describe the analysis of candidate DB00-4, 5 , 6, as an example of our method. We conclude by commenting on whether the candidates are new, massive, GC star clusters. N

3.1

Example: Candidate DB00-4,5,6

Cluster candidates 4, 5 , and 6 from DBOO are located within 1’ of each other, near (E. b ) = (0.3, -0.19). The appearance of the X-ray emission is similar to the Quintuplet cluster: numerous point sources over a

275

Astron. Nachr./AN 324. No. SI (2003)

Fig. 3 Merged data from the Galactic Center Survey with regions showing candidate and known clusters. Circles show clusters and candidates with X-ray emission: crosses show those without X-ray emission.

faint diffuse background. The close alignment of these candidates suggests that they are associated with each other. Figure 4 shows the Chundra X-ray and 2MASS IR ( K , qimages ) of the region about DBOO4, 5 , 6. By comparing sources detected in the X-ray and IR observations, a consistent 2" offset is seen, consistent with the warping of the off-axis PSF. Accounting for this effect, we find 4-5 X-ray sources with ergs cmp2 s p l , typical of point IR counterparts. They have observed (absorbed) fluxes of a few ergs cmp2 SKI). sources in the GCS, but faint compared to the Arches (a few The easiest way to comment on whether these sources are at the GC is to study the absorption of the Xray spectra. We extracted X-ray spectra of the core of DBO0-4,5,6 and fit them with absorbed thermal and absorbed power-law models. Both models give values of the hydrogen column N H = (1-1.5) x 1 0 2 2 ~ ~ ~ i p 2 , far less than the expected 6 x 1 0 2 2 m - 2 . The 2MASS color-color diagram confirms the low extinction seen in the X-ray spectra. We estimate the reddening toward the X-raylIR sources and find values consistent with our X-ray measurements of N ~ I , assuming they are main-sequence stars. Curiously, there are a number of sources with reddening values consistent with the nominal GC value ( A v = 30 mag). These sources are located within 30" of DB00-4, the southernmost of the three candidate clusters. 3.2

Comparing X-ray and 2MASS Extinctions

The combination of X-ray NH and 1R A" can determine whether the candidate clusters are near the GC. Estimating the NHrequires -- 30 X-ray counts, and estimating the A" toward a source requires photometry in three bands (or two, with assumptions). If we apply a simple correction for dust scattering and assume N H [cmp2] = 1.8 x 102'Av [mag] (Predehl and Schmitt, 1995), we find the following values for Av toward the X-raylIR sources: for DB00-4, 5. 6, A" (IR) = 0-10 mag and 4 v (X) -- 5 mag; for DBOO58, A" (IR) = 5-20 mag and A" (X) 1 2 mag; and for DB01-42, A" (IR) = 10-20 mag and Av (X) >-- 30 mag. For DB00-4,5,6 and 58, the agreement between the X-ray and IR derived values is good. Av values for these two regions suggest that they are outside of the GC. However, DBO 1-42 shows a significant difference between the X-ray and IR derived Av . Upon closer inspection, the X-ray/IR correlation is relatively weak, since the PSF is relatively large in this case. Aside from this possible false X-ray/IR correlation, we arc seeing highly absorbed X-ray emission from the candidate cluster. The color-color diagram shows that the majority of the IR sources within 30" of this candidate are highly reddened. N

C. Law and F. Yusef-Zadeh: X-rays from GC Star Clusters

276

Fig. 4 (Lejt): Adaptively smoothed X-ray image of the DB00-4, 5 , 6 candidate cluster. Circular regions denote the nominal 30" edge of the cluster, and small ellipses show X-ray sources that are possibly correlated with 2MASS sources. (Right): 2MASS K , image of the same region with the same X-ray/IR regions overlaid.

i

X

lip;

x

i

0

0

15

05 H-Ks

25

0

05

1

i 5

2

25

H-Ks

Fig. 5 (Lej?) J - H vs. II - K , for 2MASS sources near candidate cluster DB00-4, 5 , 6. Large crosses show IR sources with X-ray counterparts. Arrows indicate upper limits in II - Ks (all others have complete photometry). Theoretical tracks from Koorneef (1983) and reddening vector of Av = 20 mag from Cardelli, Clayton, and Mathis (1989). (Righl) .I - I1 vs. H - K , for 2MASS sources near all candidate clusters. Here, the crosses show sources from candidates with at least one X-ray/IR correlation. Circles show sourccs from candidates without any X-ray/IR correlations.

Figure 5 shows an IR color-color diagram comparing sources from clusters with X-ray counterparts (DB00-4, 5 , 6, 58, DB01-42) and without X-ray counterparts (DBOO-I, 53, 54, 55). Two trends are

Astron. Nachr./AN 324, No. SI (2003)

277

apparent: ( I ) non-X-ray detected clusters tend to have high reddening, consistent with GC values, and (2) all possible X-ray/lR correlated sources are at relatively low reddening. Point ( 1 ) suggests that these IR sources are probably at or near the GC, but point (2) shows that the X-ray data aren't sensitive enough to observe them. The IR data may be a more efficient method of identifying these clusters. However, i t Is also clear that these clusters are not as luminous as the Arches, and are likely to be less luminous than the Quintuplet, which is detected in similar duration Chandm observations. Moreover, we have shown that the X-ray observations can distinguish the Arches and Quintuplet clusters much more efficiently than 2MASS IR observations, suggesting that X-ray observations may provide a sensitive measurement of young and dense massive stellar clusters. This survey of the X-ray and IR properties of these candidate clusters has put limits on their elassification as massive, GC star clusters. Only one of the candidates with X-ray emission is possibly located near the GC (DBOI -42). However, the undetected candidates are highly extincted in the IR, suggesting that they may lie at the GC, but are not as X-ray luminous as the Arches or Quintuplet. Numerical studies and IR observations of the Arches and Quintuplet have shown that the mass, stellar density, and luminosity all decrease as the cluster ages. If the X-ray emission is generated by a cluster wind (or any other stellar-density based model), we may use the limits on the X-ray luminosity to constrain the ages of the cluster candidates to he older that 3-6 Myr. Alternatively, the X-ray luminosity of clusters like the Quintuplet may he simpler, single-star mechanisms (e.g., shocks in individual stellar winds; MacFarlane & Cassinelli 1989). In this case, the undetected candidate clusters are less likely to have stars with high-velocity, ionized winds, as seen in the Quintuplet. Acknowledgements This work was made possible by grant NASAX-39073. This research has made use of the NASAIIPAC Infrared Science Archive. which is operated by JPL/CalTech, under contract with NASA, and NASA's Astrophysics Data System. We thank thc Chnndrn X-ray Center for advice on data reduction and Fred Baganoff for helpful discussions on the effects of dust scattering.

References Baganoff, F. K., ct al. 2003, ApJ, 591, 891 Blum, R. D., Schaerer, D., Pasquali, A., Heydari-Malayeri, M., Conti, P. S., Schmutz, W. 2001, AJ, 122, 1875 Canto, J., Raga, A. C.,& Rodriguez, L. F. 2000, ApJ, S36, 896 Cardelli. J., Clayton, G., & Mathis, J. 1989, ApJ, 345. 245 Cotera, A. S., Erickson, E. F., Colgan, S. W. J.. Simpson, J. P., Allen, D.A,, Burton, M. G. 1996, ApJ, 461. 750 (C96) Dutra, C. M., & Bica, E. 2000, A&A, 359, L9 (DBOO) 2001, A&A, 376,434 (DBOI) Figer, D.F., McLean, I. S., & Morns, M. 1999i1, ApJ, 514,202 (FMM99) Figcr, D. F., Kim, S. S., Moms, M., Serahyn, E., Rich, R. M., McLean, 1. S. 1999h. ApJ, 525, 750 Kim, S. S., Morris, M., & Lee, H. M. 1999, ApJ, 525,228 Koorneef, J. 1983, A&A, 128, 84 Lang, C. C., Figer, D. F., Goss, W. M., Morris, M. 1999, AJ, 118, 2327 Lang, C. C., Goss, W. M., & Moii-$, M. 2002, AJ. 124, 2677 Lang, C. C., Goss, W. M., & Rodriguez, L.F. 2001, ApJ, 551, L143 (LGROI) MacFarland, J. J. & Cassinelli, J. P. 1989, ApJ, 347, 1090 Morris, M., & Serabyn, E. 1996, ARA&A, 34,64S Nagata, T., Woodward, C. E., Shure, M., Pipher, J. L., Okuda, H . 1990, ApJ, 351, 83 (N90) Nagata. T., Woodward, C. E., Shure, M., Kohayashi. N. 1995, AJ, 109, 1676 Portcgies Zwart, S. F., Makino, J., McMillan, S. L. W., Hut, P. 2001, ApJ, 546, L l 0 l Predehl. P., & Schmitt, J. H. M. M 1995, A&A, 293, 889 Sunyaev, R., & Churazov, E. 1998, MNRAS, 297, 1279 Wang, D., Gotthelf, E., & Lang, C. C. 2002, Nature, 415, 148 Yusef-Zadeh, F., ct al. 2002, ApJ, 570, 665 (Y02) Yusef-Zadeh. F., Nord, M., Wardle, M., Law, C., Lang, C., Lazio, T. J. W. 2003, ApJ, 390, L103

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1,279-284 (2003) / DO1 10.1002/asna.200385036

Simulated X-ray emission from the Arches cluster Pablo F. Velazquez*’, Alejandro C. Raga**’,Jorge Cant$, Elena Ma~ciadri***~, and Luis F. Rodriguezt4

’ Instituto de Ciencias Nucleares, UNAM, Apdo. Postal 70-543, CP 04510, MCxico D.F., MCxico.

* Instituto de Astronomia, UNAM, Apdo. Postal 70-264, CP 04510, MCxico D.F., Mexico.

Max-Plank-Institutfur Astronomie, Kiinigstuhl 17, D-69117 Heidelberg, Germany Instituto de Astronomia, UNAM, campus Morelia, Apdo. Postal 3-72, Morelia, Michoacin, CP 58089, Mt5xico

Key words Galaxy: center, hydrodynamics, methods: numerical, open clusters and associations: individual (Arches, G0.121+0.017), shock waves, stars: winds, outflows PACS 04A25

3D numerical simulations of the Arches cluster wind are presented. In order to make these simulations as close the observations as possible, we have taken the positions (on the plane of the sky) of the 60 brightest stars and we have created three possible 3D spatial distributions. For each of these configurations,we have obtained simulated X-ray maps and spectra, for comparing directly with recent observations of this cluster obtained with the Chandra satellite by Yusef-Zadeh and coworkers.

1 Introduction The Arches cluster (G0.121+0.017) is an amazing object, close to the Galactic Center, and formed by around 100 massive stars within a radius of 0.2-0.3 pc (Nagata et al. 1995; Cotera et al. 1996; Scraby, Shupe & Figer 1998; Figer et al. 1999). Several theoretical models for the multiple wind interactions of the clusters of massive stars have been recently presented in the literature. Ozernoy, Genzel & Usov (1997) carried out analytical modeling, while Cant6, Raga & Rodriguez (2000), did both analytical and numerical studies. Cant6, Raga & Rodriguez (2000) applied their study to the Arches cluster case, predicting that this cluster could be observed at X-ray wavelengths. This theoretical prediction has been confirmed with recent observations of the Arches cluster, obtained with the Chandra satellite by Yusef-Zadeh et al. (2002). Raga et a1.(2001),carried out three-dimensional numerical simulations, modeling the Arches cluster. They obtained simulated X-ray maps and spectra, which are consistent with observations. In this work, we summarize the main results of Raga et a1.(2001), and we also analyze the early stages of the evolution of the cluster wind, which exhibit structures, which are similar to filamentary structures observed in other nebulae, such the “Honeycomb nebula” in the LMC (Meaburn et al 2001). * Corresponding author: e-mail: [email protected]: +52 55 5622 4672, Fax: +52 55 5616 2233 *’ e-mail: ragaQnuclecu.unam.mx * * * email:[email protected] email:l.rodriguezQastrosrno.unammx

@ 2003 WII.EY~VCHVcrlap Frnhll & Co. KFAA,Wcinhcim

P. F. Vellzquez et al.: Simulated X-ray emission from the Arches cluster

2x0

2 A model for the Arches cluster In order to make a somewhatrealistic model for the Arches cluster, Raga et al. (2001) consider the positions in the plane of the sky (2,y ) of the sixty brighter stars of this cluster in the HST IR image of Figer et a1 (1999). These positions are measured with respect to the barycenter of the cluster. At each of these positions, identical stellar winds have been introduced. These winds have mass-loss rates of Ma yr-’ (from an estimate of Lang, Goss & Rodriguez 2001), and terminal velocities u,=1500 km sP1 (taking into account the velocity estimated by Cotera et al. 1996 and Lang et al. 200 I ) . However, this description is not complete, because the star positions along the line of sight (z-coordinate) are unknown. In order to solve this, Raga et al. (2001) employ the following statistical method. They assume that the intrinsic distribution function f for the stellar positions, is only a function of the spherical radius R (measured respect to the cluster’s barycenter). Then, knowing star positions in the plane of the sky ( 2 0 , yo), the possible values of z have to satisfy a probability function given by:

h(l.1) = -4

+

and A is a normalization constant so that Jo” h(1zl)d)zl= 1. where = (zi y;)’/’, For the distribution function, Raga et al. (2001) chose a functional form f ( R )K R-2. For this choice of f ( R ) ,the normalized distribution o f IzI i s obtained, for a star at a projected distance TO, from the barycenter of the cluster (measured on the plane of the sky), through Eq.( 1). Comparing the distribution of IzI with an uniformly distributed variable ‘I, we obtain:

z=

tan “‘I 2

(2)

.I

after carrying out the integral .Ih( lzl)dlzl = ildq to determine the normalization constant. In this way, we are able to generated appropriately distributed values for z , by first computing = (:cg yi)l/z, then choosing 11 randomly (in the interval [O, l]),and finally using Eq. (2) to determine z (positive and negative values are also chosen randomly). Furthermore, we reject the generated z-values if IzJ > r,,,,, where r,na2is the maximum distance (projected on the plane of the sky) with respect to the barycenter of the cluster for all of the cluster stars. With this process, three possible distribution were generated (hereafter models M I , M2 and M3). In Figure 1, we show the stellar positions for these three modcls.

+

3 Numerical simulation Considering the initial conditions given in $2, we have carried out three-dimensional gasdynamic numerical simulations, using the YGUAZU-A binary adaptive grid code. This code integrates the gasdynamic equations with a second order implementation of the “flux-vector splitting” method of van Leer (I982), and is described in detail in Raga, Navarro-Gonzilez & Villagran-Muniz (2000). Due to the fact that the radiative losses are unimportant in this problem (Cant6 et al. 2000), they are not included in the calculations. The numerical simulations were carried out in a five-level binary adaptive grid, with a maximum resolution of 1 . 1 7 2 ~10l6 cm along the three coordinates, covering a cubic domain of 3 x 1018 cm sides, centered on the barycenter of the stellar cluster. This domain was initially filled with an ISM of temperature of lo7 K and density of lo P 3 r r i H cmP3, except for spherical regions of radius R , = 5 x loL6cm, which are centered in each of the cluster star positions. In these spherical regions, spherically symmetric and constant velocity winds are imposed (with temperatures of lo6 K and the mass-loss rate and terminal wind velocity given in $2).

28 I

Astron. Nachr./AN 324, No. S 1 (2003)

h

4

1 1XlO'B

0

-1x10'8

10'8

0

10'8

x (cm)

x (em)

Fig. 1 Top leti panel: Spatial distribution of 60 stellar wind sources on the plane of the sky (zo, yo). The other three panels show the (z, t)distribution for models M1, M2 and M3, with the z-value given by the procedurc described in $2.

Raga et al. (2001) computed three models ( M I , M2 and M3). Steady flow configurations are achieved for these three models after a time integration of 1000 yr. In this work, we are going to focus our attention on the evolution of model M1. As an example of the resulting flows, in Figure 2 we show the density and temperature stratification on the y = 0 plane, for model M1 at t = 1000 yr. From this Figure, we see that the central region is filled by an almost homogeneus medium, with densities of l U p 2 ' g cm-3 and temperatures of 5 x lo7 K. This medium was called "the cluster wind' by Cant6 et a1.(2000). Cooler stellar wind cavities are embedded within this cluster wind. Dcnsity and temperature stratification maps (left and right panel of Figure 2, respectively) shows that the central regions have complex morphologies, resulting from multiple stellar wind interactions. However, in some parts o f the periphery, simpler "stellar wind bow shock" structures are observed. These structurcs are the rcsult of the interaction between the cluster wind (generated by the central stars) and the stellar winds of stars located in the periphery.

-

4 4.1

Results Predicted X-ray emission

From the thrcc-dimensional numerical simulation of the Archcs cluster wind. we have carried out prcdictions of the X-ray spectrum cmitted in the 0.5-8 keV photon energy rangc. The wavelength-dependent emission coefficient was computed using the CHIANTI' atomic data set (see Dcre et a1.2001 and refercnces therein) as a function of the gas tempcrature. For these calculations, Raga et a1.(2001) assumed that the gas i s in coronal ionization equilibrium and in the low density-regime. They also assumed that the abundance of heavy elemcnts is twice of the solar abundance (following Baganoffet al. 2001). I

The CHIANTI database and associated IDL procedures, are freely available at the following addresses on the World Wide Web: http://www.solar.nrl.navy.mil/chianti.html,http://www.srcetri.astro.it/science/chaint~chiaiiti.html, and http://www.dain~p.cam.ac.ukluser/astro/chianti/cliianti .litml

P. F. VelQquez et al.: Simulated X-ray emission from the Arches cluster

282

I o - * ~ 1 0-22

-1.

10

18

0

1 o6

1 OP2'

-1.

1O l 8

1 o8

I0 '

10

18

0

1 0l8

Fig. 2 Lefrpanel: density stratification on the y = 0 plane for the steady flow stratification(after 1000 yr o f integra~ . panel: the same but tion). The grayscale is logarithmic, represented by the upper bar, and is given in g ~ m - Right for the temperature stratification on the y = 0 plane. The upper bar represents the logarithmic grayscale, given in K. 0 025

15 -

m 0

+

I

"

'

I

"

'

I

"

'

0 02

1 -

0015 pi

0

0 01 m

2

0 005

#

I

-1

10l8

0

0

Fig. 3 Predicted X-ray image from model MI. This image was obtained after integrating the emmion coefficient along the line of sight (in the simulation, this line is parallel to z-axis), and over the [0.5, 81 keV energy range. The grayscale (represented by the bar to the nght) is linear and IS given in unit\ of erg sC1 cm-' sr-'.

2

4

6

8

Energy (keV)

Fig. 4 Radiative flux (per unit frequency, normalized to the peak of the continuum) as a function of photon energy in the 0.5-8 keV range, obtained from model MI. The strong line at 6.6-6.7 keV is the Fe XXV K feature

With this emission coefficient, Raga et aL(2001) generated X-ray emission maps, integrated over photon energies from 0.5 to 8 keV, in order to directly compare with X-ray images obtained by Yusef-Zadeh et a1.(2002), with the Chandra satellite. For model M1, the resulting map is shown in Figure 3. This simulated map has pointlike intensity peaks, corresponding to the inner stellar wind regions. However, most of the flux comes from a diffuse component, which exhibits several local maxima and filamentary structures. Figure 4 shows the spectrum of the emission integrated over the whole volume of the computational domain. Raga et a1.(2001), applied a frequency-dependentextinction, corresponding to an N H = 1.24 x H cmP3 for directly comparing with observational spectra (Yusef-Zadeh et a!. 2002).

283

Astron. Nachr./AN 324, No. S I (2003)

Finally, from either predicted images o r predicted spectra, Raga et al. (2001) compute a total X-ray 0 ~s-l. ~ luminosity in the 0.5-8 keV band of 2 . 9 ~ 1 erg

1o6

1 o5

-1. 10

18

0

l o l a -1. 10

18

0

Fig. 5 Column temperature maps for the early cluster wind evolution. In the first frames, a filamentary structure it i h observed, which evolves to more or less homogeneus temperature regions, with some cold wind cavities. The grey-scale is logarithmic and i s given in K pc.

4.2

Early stages of evolution for the cluster

Raga ct a1.(2001) studied the final stationary stage of a cluster wind evolution. However, i t is interesting to analyze the early stages of the evolution. Individual wind sources generate stellar wind bubbles, whosc

284

P. F. Velizquez et al.: Simulated X-ray emission from the Arches cluster

shock waves start to sweep up the circumstellar gas. Interacting regions have high temperatures and pressures, which are good tracers for following the early evolution of these shock waves. In Figure 5, we show maps of thc temporal evolution of the column temperature CT = T dz. At t =20 yr and 60 yr frames, it is seen that the central cluster region is characterized by a filamentary structure, which is made from the superposition of multiple rings. This morphology i s the result of multiple wind interactions, and resembles structures observed in some nebulae, such as the “Honeycomb Nebula” (Meaburn et a1.2001). The filamentary structure evolves to more or less stationary and more homogenous temperature distribution, with some “rays” or “fingers” far away from the cluster harycenter.

1

5

Conclusions

Three-dimensional numerical simulations were carried out for modeling the Arches cluster wind. In order to compare with recent X-ray observations (Yusef-Zadeh et al. 2002), obtained with the Chandra satellite, simulated X-ray emission maps and spectra were generated from the numerical results, and also employing the CHIANTI database (see Dere et al. 2001). Althought the simulated X-ray maps are only qualitatively similar to the observed images (this strongly depends of the employed distribution for the z - coordinate), the X-ray spectra generated for three models (MI, M2, and M3, Raga et al. 2001) are in excellent quantitative agreement with observations. For the erg s-’ was obtained, which is very close to three computed models, a total X-ray luminosity of 3 x erg spl value measured by Yusef-Zadeh et al. (2002) for the Arches cluster. the 4 x We have also analyzed the early cluster wind evolutionary stages, finding that the shock waves from individual stellar wind sources have filamentary structures, which are similar to the structures observed in H a images of the “Honeycomb nebula”. However, the high temperature interaction regions in our simulations do not show appreciable optical line emission. For different model parameters (e.g. a denser initial intra-cluster medium and/or lower velocity winds), the filaments in the initial flow configuration could easily be radiative, and therefore would be observable in optical recombination and collisionally excited lines. Acknowledgements PV and AR are supported by CONACYT grants 34566-E and 36572-E, and DGAPA-UNAM grant INI 12602. LFR acknowledges the support of DGAPA. UNAM, and CONACYT, MBxico. We thank Israel Diaz for computer support.

References Baganoff, B. K., Bautz, M. W., Brdndt W. N. et al. 2001, Nature, 413, 45 Cantb, J., Raga, A. C. & Rodriguez, L. F. 2000, ApJ, 536,896 Cotera, A. S., Erickson, E. F., Colgan, S. W. J., Simpson, J. P., Allen, D. A., & Burton, M. G. 1996, ApJ, 469, 729 Dere, K. P., Landi, E., Young, P. R., & Zanna, G. 2001, ApJS, 134,331 Figer, D. F., Kim, S. S., Moms, M., Serabyn, E., Rich, R. M., & McLean, I. S. 1999, ApJ, 525, 750 Lang, C. C., Goss, W. M. &Rodriguez, L. F. 2001, ApJ, 551, L143 Meaburn, J., Redman, M. P., Bryce, M., Lbpez, J. A., Al-Mostafa, 2. A., & Dyson J. E. 2001, ApSS, 272, 217 Nagata, T., Woodward, C. E., Shure, M. & Kohayashi, N. 1995, AJ, 109, 1676 Osernoy, L. M., Gcnzel, R. & Usov, V. 1997, MNRAS, 288,237 Raga, A. C., Velazquez, P. F., Cantb, J., Mdsciadri, E. & Rodn’guez, L. F. 2001, ApJ, 559, L33 Serabyn, E., Shupe, D., & Figer, D. F. 1998, Nature, 394, 448 van Leer, B. 1982, ICASE Rep., 82-30 Yusef-Zadeh, F., Law, C., Wardle, M., Wang, Q. D., Fruscione, A., Lang, C. C., Cotera, A. S. 2002, ApJ, 570, 665

SiO Maser Sources within 30 pc of the Galactic Center

'

'.'

Shuji Deguchi* and Hiroshi Imai ' Nobeyama Radio Observatory,Minamimdki, Minamisaku, Nagano 384-1305. Japan VERA project office, National Astronomical Observatory, Mitaka, Tokyo 18 1-8588, Japan present address: JIVE, PO Box 2, 7990 AA, Dwingeloo, the Netherlands

'

Key words Masers, Galactic Center, AGB, Mass-loss Abstract. Using the Nobeyama 45-111radio telescope, we have observed 314 large amplitude variables within 30 pc of the Galactic center in SiO maser lines. Resulting detections give the radial velocities of 174 stars; light-variation periods have been known for all of thcse stars. The SiO detection rate increases sharply with the period and it is about twice of the OH maser dctection rate. The radial-velocity data show slow and rapid rotations of the outer and inner circumnuclear-diskstars, respectively. Five high-velocity stars were found only at the negative-longitudeside of the Galactic center. Estimation of the ages of high velocity stars suggests that these stars must be be accelerated to high velocities within 10' years.

1 Introduction Mass-losing AGB stars are good probes of the Galactic center. For the last few years, we have observed steller mascr sources near the Galactic ccnter in SiO maser lines at 43 GHz (Deguchi et al. 2000) with the Nobeyama 45-m radio telescope. These observations were made toward color-selected IRAS sources. Because of the incompleteness of the IRAS survey near the Galactic center, the area within 1 degree from the Galactic center had remained unsurveyed i n SiO masers. The situation, however, has recently changed owing to new ground-based surveys with near-infrared array cameras, and to space-based mid-infrared surveys. Large numbers of candidate stars suitable for maser surveys toward the nuclear disk have been discovered in the near-infrared K band by Glass et al. (2001) making use of their characteristic largeamplitude variability. In the present paper, we report the SiO maser survey of the large amplittide variables within 30 pc of the Galactic center. Resulting detections gave the radial velocities of 174 stars. These sources are intermediate-mass stars in the Asymptotic Giant Branch phase of the ages between lo7 - 10"' y. Based on these data, we discuss on kinematics of masers stars in the Galactic center.

2 Observations and results Observations in the SiO .J = 1 - 0 o = 1 and 2 transitions at 42.821 and 43.122 GHz were made with the 45-m radio telescope during February 2001 - May 2002 i n a long-term project of the Nobeyama radio observatory. A detail description of the 45-in telescope system can be found in Izumiura et al. (1999). A preliminary result of the survey of 134 large amplitude variables in the Galactic center was published i n Imai et al. (2002). Monitoring observations of SiO sources towards Sgr A* were reported in Deguchi et al. (2002). Examples of SiO maser detections are shown in Figure 1. For most of the SiO maser sources, the intensity ratio of the SiO J = 1 - 0 'r = 1 to '1' = 2 line is near unity. It is know to weakly correlate * Corresponding author: e-mail: deguchiQnro.nao.ac.jp, Phone: +81 267 984369, Fax: +81 267 98 2884

02007 WILEY~VCHVcrlq GmbH & Co

KGaA. Weinhein,

S. Deauchi and H. Imai: SiO Masers within 30 DC of GC

286

with the IRAS 12/25 pm color (see discussion in Nakashima and Deguchi 2003). Therefore, simultaneous observations of two SiO lines secure detections. The 45-m telescope beam size is about 40” (HPFW). Because of the high source density at the Galactic center, we occasionally had double or triple detections of sources in the same beam (as shown on the right in Figure 1). Even in such cases, we can assign the source to a particular object by observing an offset position from the source of 10-15” and measuring the relative intensity variation according to the position. We have observed 314 objects, resulting in 174 detections through June 2002.

-

0.8

0.8

0.6

0.6

2 0.4

-Y, 0.4

v

F 0.2

2 0.2 0

0 -0.2

50

150

100

v

Isr

200

(kmis)

-0.2 -200

-150

-100

v

Isr

-50

0

(km/s)

Fig. 1 Spectra of SiO masers toward g6-247 (left) and g3-46 (right). The right shows a case of triple detections at KSr = -141, -98, and -45 km s-’. The number under the source name indicates the observed date in yymmdd.d

format.

2.1 Period-Detection rate Figure 2 shows a histogram of the detection rate versus variablity period. The SiO detection rate is approximately 55 %, which is similar to the SiO detection rate for the inner bulge IRAS sources. This fact suggests that the sample of the large amplitude variables in the Galactic center has very similar characteristics to the color-selected bulge IRAS sample. The SiO detection rate sharply increases with the period of light variation. We note that the average period of the Glass et al. (2000) sample is about 430 days. Because the present sample is considered to be at a uniform distance, a relation between the detection rate and the period is firmly established. Blind OH maser surveys have been made with VLA (Sjouwerman et al. 1998) in the same region of the sky. The SiO detection rate i s about twice of the OH detection rate. 2.2 Longitude-Velocity Diagram Figure 3 shows a longitude-velocity (I-u) diagram. The best fit to the velocity data is = 0.0594(Al/‘/) - 6.85 km sC1

(1)

for all of the sources, and

V,,,

= 0.274(Al/’/)

- 2.90

km s-l

(2)

for sources within 300”. The radial velocity increases with the longitude offset, Al, from Sgr A * due to Galactic rotation; on average, the rotational rate is about 214 km lip’ per degrce, which is close to the Galactic rotation obtained from CO and HI observations (Honma and Sofue 1997). However, the rotational velocity tends to increase near the Sgr A* (within 5’) as shown in equation (2) (see also Deguchi et al. (2002) within 2’). The overall velocity structure of the SiO 1-ZJ diagram is similar to the OH b-IJ diagram (for example, see Sjouwerman et al. 1998). We can recognize a hole in the SiO I-,ii diagram at

287

Astron. Nachr./AN 324, No. S 1 (2003)

100

100

80

80

% 60

a

5

40 20

20

0

0 100

200

300

400

500

600

700

800

900

Period(d) Fig. 2 histogram of the period. Shaded area shows SiO detection and blank nondetection. The line graph shows the detection rate (the scale on the right axis), and the broken lines shows the OH detection rate for the same sources in the present sample.

= 200", = 50 km s-'), which i s also seen in the OH l-r~ diagram. It i s possible that this hole is produced by some non-axisymmetric gravitational potential in the Galactic center.

(Al

400

300 200 100

. !so h

m

v L

(I)

5 -100 -200 0

-300 -400 1000

*

500

0

a

-500

-1000

A1 (")

Fig. 3 Longitude-velocity diagram for stars with periods below and above 600 d. The leasf-square tits to each subsample were shown.

S. Deguchi and H. Irnai: SiO Masers within 30 pc of GC

288

Figure 3 also exhibits a slight asymmetry in velocity dispersion with respect to Sgr A*; on the A1 < 0 side, the points seem scattered more than those on the A1 > 0 side. To check the asymmetric structure in velocity dispersion, we computed the dispersions from the best fit lines for the subsets inside and outside of r = 300”. The results were summarized in Table 1. The high velocity stars (JV,, 1 > 200 km s-l) strongly influence to the results. Therefore, we calculated the cases including and excluding the high-velocity stars. Even excluding the high velocity stars, the velocity dispersion is significantly larger on the Al < 0 side. Table 1 Velocity Dispersions.

A1 > 0, r > 300‘‘

A1 < 0, r > 300”

r < 300’’

(km s-l)

(km s - * )

(km s-l)

All sources

49.4

90.1

80.3

Excluding high-vel. sources

49.4

61.2

57.9

Sample

To check the presence of period (or mass) segregations of stars in the stellar cluster, we also computed the velocity dispersions for subsets o f stars with periods below and above 600 days. No period segregation was found in the present sample, indicating that the mixing time scale in phase space is sufficiently larger than the ages of these stars. It is notable that periods of stars circulating around the Galactic center are smaller than 4 x 106 years within 30 pc of the Galactic center. Even for high progenitor-mass stars (9 A f @ ) , the ages in the AGB phase (for example, Vassiliadis & Wood 1993) are longer than this dynamicalmixing time scale. 2.3 High-velocity Sources Figure 4 shows SiO maser spectra and 2MASS J H K images of 5 high velocity sources. These high velocity sources appear at both positive and negative velocities, therefore the previously suggested asymmetry in velocity (van Langevelde et al. 1992), can be regarded as a statistical fluctuation. It is possible that Lhese high velocity sources have highly elongated eccentric orbits around the Galactic center. From the luminosities and periods, we can estimate ages of these stars. The mass of the brightest star, g3-2855, is estimated to be more than 5 A&. The age must be shorter than lo8 y. Therefore, this high velocity star must be accelerated to V 300 km/s within lo8 y. Also, the velocity distribution of high velocity stars seems to be discreet from the low velocity stars (for the inner bulge stars ((!I > lo), the distribution of high velocity stars seems continuous to the distribution of low velocity stars). The time scale of stars circulating around the Galactic center is about 4 x lo6 years. The ages of these stars are longer than 2 x lo7 years (for M < Shf,), so the distribution of these high-velocity stars should be smoothed out enough as a result of rapid phase mixing. Therefore, the fact that these high velocity stars were found only at A1 < 0 side is unexplained. Kim and Morris (2001) proposed that vertical diffusion of nuclear disk stars is driven by the scattering of stars o f fgiant molecular clouds in the nuclear disk. This time scale is longer than lo9 y, indicating that this mechanism cannot he used to explain observed high velocity stars. An efficient mechanism to disrupt thc star cluster falling into the central blackhole was proposed in this conference by Kim, Morris, & Figer, which may possibly create the high velocity stars near the Galactic center in a free fall time scale.

-

2.4

Around Sgr A*

Figure 5 shows a time variation of SiO intensities of the sources which are located within 20” from Sgr A*. By mapping the 100” x 200” area of Sgr A*, we detected about 15 SiO sources (Deguchi et al. 2002), and obtain the positions in 5-10’’ accuracies. We detected one high velocity SiO source (-342 km s p l component; No. 5 in the lower panel of Figurc 5). This component was more clearly detected in 1997 N

2x9

Astron. Nachr./AN 324, Yo. S1 (2003) 0.6 0.5 0.4 0.3 Y 0.2 + 0.1 0 -0.1 - 0 . z F 1" ' ' " 150 200 -200

-

-350

-300

-250

-350

-200

610 d

P= 597d

- .~ 02r'

200

"

-250

-300

' I

" "

250

I ' ' I '

300

' 1 '

350

P= 472d

"

"

400

.0.2 F , -400

'

"

I

'

250

"

' I

' ' ,'1

300

318 d

'

'

I

' ' ' '

-350

'

'

'

-300

'

'

'

' ' '

-250

5

1

-200

597 d

Fig. 4 SiO maser spectra and false-color J H A ' images of.5high velocity sources. The high velocity source found toward G4-1 13 is different from G4-I 13 (OH 359.855-0.078) with Vb,. = 5.2 km s - ' ; in order to clarify the difference, additional number (.2)was attached.

(Izumiura et al. 199X). Because of it's intensity, it was quite difficult to locate the position of this star i n the mapping observations, but is estimated to be no further than 10" away from Sgr A*. The -27 km s pl component, which was attributed to IRS IOEE located about 10" NE of Sgr A" (Menten et al. 1997), showcd a large time variation. The SiO masers of IRS lOEE flarcd in March - May 2000 and again in March 2002. We have monitored the SiO maser intensity of this object for last few years. The result of this monitoring till 2001 was reported in Deguchi et al. (2002). The intrinsic maser intensity (about 1.5 Jy) of IRS lOEE SiO masers during the flares is approximately thc same as those of

S. Deguchi and H. Imai: SiO Masers within 30 pc of GC

290

the strongest SiO maser sources in our Galaxy: Orion SiO masers and Sgr B2 MD5. This fact indicates that the maser power at the flare time is comparable to the maximum seen i n SiO masers.

0.4 0.3

-s

0.2

(d

-I

0.1

0 '

-0.1 ~ ~ " " " " " " ' ~ " " " " " " " " " ' ~ -400 -300 -200 -100 0

V

Isr

100

200

300

'

" ~ 400

(km s-')

0.4 0.3 n

0.2

2

0.1

s

0 -0.1

-400

-300

-200

-100

0

100

200

300

400

VIsr (km s-')

-

Fig. 5 Time variation of the SiO maser sources around Sgr A* (within 20"). Upper panel is the spectra taken on June 9, 1999 and the lower panel on May 25, 2000. The SiO masers from IRClOEE at -27 km s C 1 flared during March-May 2000.

Near-infrared time variation of IRS lOEE was monitored by Wood et al. (1998) and the light-variation period of this star is 720 days. It was not certain whether the SiO maser flares in this period or not. SiO masers of this star flared again in March 2003. Therefore, they seem to flare in every two years. N

Astron. NachrJAN 324. No. SI (2003)

29 1

3 Conclusions We found that the SiO detection rate increases sharply with the period and it is about twice of the OH maser detection rate. SiO masers are useful tool to investigate the relatively low-mass AGB stars in the Galactic center region. The radial-velocity data show slow and rapid rotations of the outer and inner circumnuclear-disk stars, respectively. No period/mass segregation was found in the sample, indicating that a phase mixing occurs in a dynamical time scale of about lo6 y. Five high-velocity stars were found only at the negative-longitude side of the Galactic center. SiO masers of IRS IOEE were found flared i n May 2000. This SiO maser flare seems to occur periodically in every two years. Acknowledgements The observations were made for a long-term project of Nobeyama Radio Observatory with collaboration of Drs. T. Fujii, , I . S. Glass, Y. Ita, H. Izumiura, 0. Kameya, A. Miyazaki, Y. Nakada, and J. Nakashima. This research was partly supported by Scientific Research Grant (C2) 12640243 of Japan Society for Promotion of Sciences.

References Deguchi, S., Fujii, T.,Izumiura. H., Kameya, O., Nakada, Y., Nakashiina. J., Ootsubo, T., & Ukita, N., 2000, ApJS, 128,571 Deguchi, S., Fujii, T., Miyoshi, M., & Nakashima, J. 2002, PASJ, 54, 61 Glass, I. S., Matsumoto, S., Carter, B. S .. & Sekiguchi, K. 2001, MN. 321,77 Honma, M., & Sofue, Y. 1997. PASJ, 49,539 Imai, H., Deguchi. S. Fujii, T., Glass, I. S., Ita, Y., Izumiura, H., Kaineya, O., Miyazaki. A,, Nakada, Y., & Nakashima, J.. PASJ, 54, 19 Izumiura, H., Deguchi, S., & Fujii, T. 1998. ApJ. 494, L89 Izumiura, H.. Deguchi, S., Fujii, T., Kameya, 0..Matsumoto, S., Nakada, Y., Ootsubo, T., & Ukita, N. 1999, ApJS, 125,257 Kim, S.S. & Moms, M. 2001, ApJ, 554, 1059 Menten, K. M., Reid. M. J., Eckart, A., & Genzel, R. 1997, ApJ. 475, LI I 1 Nakashima, J., & Deguchi, S. 2003 PASJ, 55, No. I , in press Sjouwerman, L. 0..van Langeveldc, H. J., Winnberg, A,, & Habing, H. J. 1998, A&AS, 128,35 van Langevelde, H. J., Brown, A . G. A.. LindqvisL M., Habing, H. J., de Zeeuw, P. T. 1992, A&A, 261, L17 Vassiliadis, E., & Wood, P. R., 1993, ApJ, 413,641 Wood, P. R., Habing, H. J.. & McGregor, P. J. 1998, A&A, 336, 925

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S 1,293-297 (2003) / DO1 10.1002/asna.200385037

86 GHz SiO masing late-type stars in the Inner Galaxy M. Messineo* I, H.J. Habing', L.O. Sjouwerman', K.M. Menten', and A. Omont4

' Leiden Observatory, P.O. Box 9513,2300 RA Leiden, the Netherlands

' National Radio Astronomy Observatory, P.O. Box 0, Socorro NM 87801, USA Max-Planck-Institut fur Radioastronomie,Auf dem Hugel 69, D-53121 Bonn, Germany Institut d'Astrophysique de Paris, 98bis Boulevard Arago, F 75014 Pans, France

Key words Masers, Circumstellar matter, Galaxy: kinematics and dynamics

We present 86 GHz ( v = 1,J = 2 t 1) SiO maser line observations with the IRAM 30-ni telescope of a sample of late-type stars in the inner Galaxy (30" < 2 < -30'). The stars were selected from the TSOGAL and MSX catalogues on the basis of their mid-infrared fluxes and colours. SiO maser emission was detected towards 268 (6 1%) of our targets, thereby doubling the number of maser line-of-sight velocities measured toward the inner Galaxy. Our sample consists mostly of Mira-like stars. They are more numerous than OWIR stars which were previously observed to measure line-of-sight velocities. The revised longitudevelocity diagram of the inner Galaxy clearly shows a stellar nuclear disk.

1 Introduction Asymptotic Giant Branch (AGB) stars are good tracers of the Galactic structure and kinematics. Being bright in the infrared, they are visible in directions with high extinction. Maser emission from their circumstellar envelopes is strong enough to be detected throughout the Galaxy and reveals the line-of-sight velocity of the stars to within a few km s-' . Frequently detected maser lines are from O H at 1.6 GHz, HzO at 22 GHz, and SiO at 43 GHz and 86 GHz (e.g. Habing 1996). Until recently only a few hundred stellar line-of-sight velocities were known towards the inner regions of the Milky Way ( 30 O < 1 < -30 O and (bl < 1).These are mainly of OWIR stars, which are AGB stars with OH maser emission in the 1612 MHz line, mostly undetected at visual wavelengths. This number is too small to allow for a good quantitative multi-component analysis of Galactic structure and dynamics. Therefore, obtaining more line-of-sight velocities remains an issue of prime importance. To enlarge the number of known stellar line-of-sight velocities, using the IRAM 30-m telescope we have started a survey for 86 GHz S i O (v = 1,J = 2 + 1) masers towards an infrared-selected sample of late-type stars.

2 Source selection A large variety of names exists to indicate oxygen-rich AGB stars characterized by different pulsation properties and/or mass-loss rates: semi-regular (SR) stars and Mira stars (visual pulsation amplitude larger than 2.5 mag), large amplitude variables (LAV), long pcriod variable (LPV) stars (when their periods are longer than 100 days), and OH/IR stars (with 1612 MHz OH maser emission). In the IRAS colour-colour diagram the oxygen-rich AGB stars are distributed on a well-defined sequence of increasing shell opacity and stellar mass-loss rate (e.g. Habing 1996) which goes from Miras with the bluest colours and the 9.7 p m silicate feature in emission, to O M R stars with the reddest colours and the 9.7 pm silicate feature in absorption. * Corresponding author: e-mail: messineoQstrw.leidenuniv.nI, Phone: +31 71 5275831, Fax: +31 71 5'275819

02003 WILLY-VCH Vrrl.ig GmbH & Co

K G , A Wcmheim

M. Messineo et al.: 86 GHz maser survev

294

rp

p

2

1

-a

2 15 1 oo=

*

;

a -o:

0 0%"-

?"w;.,.

w

4

.-

4

3

-$

-I

-1.5 -2

-2.5 -3

,

,

I

,

,

,

1

,

,

*

* ,

I

,

I

,

1

,

[

/

I

, 1

,

I

,

I

*

,

,

,

l

,

1

,

1

,

1

,

Fig. 1 Location of the observed stars, irrespective of detection and non-detections, in Galactic coordinates. The MSX sources are shown as open squares and the ISOGAL sources as filled circles.

SiO maser emission is generated in the envelopes of mass-loosing AGB stars, close to their stellar photospheres. It occurs more frequently towards oxygen-rich Mira stars than towards other AGB stars. Furthermore, the relative strengths of different SiO maser lines are observed to vary with AGB type (Nyman et al. 1993 and references therein). It indicates that the SiO maser properties depend on the stellar mass loss rate and on the stellar variability and that, differently than for OWIR stars, for Mira-like stars the 86 GHz (v = 1, J = 2 + 1) SiO maser transition is a good tool to obtain stellar line-of-sight velocities. Therefore, we selected Mira-like stars. They also are far more numerous than OHlIR stars. The stars to be searched for maser emission were selected from both a preliminary version of the combined ISOGAL-DENIS catalogue (Schuller et al. 2003; Omont et a]. 2003) and from the MSX catalogue (Egan et at. 1999). ISOGAL is a 7 and 15 pm survey of -16 deg2 towards selected fields along the Galactic plane, mostly toward the Galactic centre. The Midcourse Space Experiment (MSX) is a lower sensitivity and resolution survey covering the entire Galactic plane at five mid-IR bands ranging from 4.3 pm [Bl band], to 21.4 pm [ E band]. The SiO maser search was limited to the Galactic plane between (b( < 1 O and I = +30° and 2 = -4"; the lower limit in longitude is imposed by the northern latitude of the IRAM 30-m telescope. The brightest 15 pm sources, with a magnitude [I510 < 1.0, and those with a 7ym-15pm colour, ([7]0 - [15]0) < 0.7, were excluded since they are likely to be foreground stars. Because of the general correlation of SiO maser emission and IR luminosity (Bujarrabal et at. 1987), sources with [15]0 > 3.4 were excluded since they are likely to show SiO maser emission fainter than our detection limit of 0.2 Jy. Sources with ( [ 7 ] 0- [15IO) > 2.3 were excluded since they are likely to be compact HI1 regions or young stellar objects. Furthermore, for the ISOGAL sample a range of intrinsic (Ks0- [15]0) colour was selected to avoid thick envelope OHAR stars and young stars. No similar restriction could be applied to the MSX sample because no near-infrared counterparts were available at the time of the observations. As an alternative to the ( K s o- [15]0) criterion, for the MSX sample we imposed an upper limit to the ratio of the fluxes in band E and C (FE/Fc< 1.4). For the MSX sources additional variability information was taken into account. Moreover, sources close to a known OH maser were discarded as the kinematic data are already known.

3 Results The observations were carried out with the IRAM 30-m telescope (Pico Veleta, Spain) between August 2000 and September 2001. With a detection limit of 0.2 Jy we detected SiO (,u = 1,J = 2 + 1)maser emission towards 268 stars, of which 255 were previously unobserved. In almost all cases the targeted star

Astron. Nachr./AN 324, No. S1 (2003)

295

is the only mid-infrared object falling inside the main beam (29"). The total detection rate is 61%. The SiO line width distribution ranges between 2 and 16 km s-l with a peak at 4 km s-'. We observed 15 LPVs found by Glass et al. (2001) and detected SiO maser emission from 1 1 of them (73 %). Since the observations were taken at a random pulsation phase, and since the SiO maser intensity is known to vary during the stellar phase by up to a factor ten (Bujarrabal 1994), this detection rate is a lower limit to the actual percentage of LPV sources characterized by 86 GHz SiO maser emission. Only 16 % of those LPV stars have associated OH emission (Glass et a]. 2001). and only 23 % among those within our defined colour-magnitude region. Among large amplitude variable AGB stars, the 86 GHz SiO masers are much more frequent than OH masers.

4 Infrared measurements The combination of near- and mid-infrared photometry permits us to study the nature of the stars, to derive luminosities, mass-loss rates, and provides a good discrimination of foreground stars. We searched for possible counterparts of our SiO targets in the recent extensive infrared data catalogues: DENIS, 2MASS, ISOGAL and MSX. We found (Messineo et al. in preparation) that our stars are intrinsically very bright in the near-infrared, and not heavily obscured by circumstellar matter, unlike the OWIR stars which are often obscured even in the K s band (e.g., Ortiz et al. 2002). Using near and mid-infraredphotometry it is possible to calculate the interstellar extinction and apparent bolometric magnitudes. This will add information on the source distances and improve the understanding of the Galactic longitude-velocity diagram. Therefore, 86 GHz SiO masers from Mira-like stars are better tracers of the Galactic structure than OWIR stars.

2 -

18

16

14

12 10 J (DENIS)

8

6

-2

K, (2MASS)

0 -

2

K, (DENIS)

Fig. 2 Left Panel: The difference between the 2MASS and DENIS J magnitudes versus the DENIS J magnitude. Squares show the position of our ISOGAL SiO targets. For comparison small dots show ISOGALlDENlS and 2MASS associations obtained in several ISOGAL fields. No correction for offset in the photometric zeropoint has been applied. Right Panel: The difference between the 2MASS and DENIS J magnitudes versus the difference of the 2MASS and DENIS K s magnitudes. The crosses indicate 2MASS J upper limits.

4.1

Variability

The 86 GHz SiO maser line intensity is observed to be stronger in oxygen-rich Mira stars. To increase the chance of detecting SiO maser emission, we selected MSX sources with evidence for variability. Unfortunately, the ISOGAL database does not provide information on variability. However, from their position

M. Messineo et al.: 86 GHz maser survey

296

in the ISOGAL/DENIS (Kso - [15]o) versus [15]0 colour-magnitude diagram, it was expected that the ISOGAL SiO targets would be mainly large amplitude variables (Messineo et al. 2002). The J and K s filters used by DENIS and 2MASS are similar, therefore the measurements obtained during the course of the DENIS and 2MASS surveys are directly comparable. For each of the 61 ISOGAL fields containing SiO targets, we cross-correlated the ISOGAL/DENIS and the 2MASS point source positions. For 55% of our ISOGAL stars, the difference between the 2MASS and the DENIS J and the 2MASS and the DENIS K s is larger than 3 times the field dispersion (see Fig. 2). Therefore, our sample contains mostly variable stars. Due to the simultaneity of the J and K s measurements in both the DENIS and 2MASS surveys, a correlation is expected and found between the variation in the J magnitude (4J)and in the K s magnitude ( A K s ) shown , in Fig. 2. A linear least squares fit gives,

AJ

= 1.71(&0.03) x AlCs

+ 0.02 i (0.02).

-

60% of (he amplitude in J band. This results is The relative pulsation amplitude in the K s band is in agreement with the relation between the pulsation amplitudes in J and in K s bands for oxygen-rich Mira stars in the solar neighborhood (Olivier et a]. 2001). A monitoring program of our SiO maser sample will provide pulsation periods and will yield also an estimate of the Galactic center distance through the period-luminosity relation.

5

Longitude-velocity diagram

The line-of-sight velocities of the 86 GHz SiO maser sources range from -274 to 300 km spl, which is consistent with previous stellar maser measurements and with the l2COvelocities toward the inner Galaxy (Fig. 3). An appreciable number of SiO sources are located in a region forbidden for pure circular rotation, at negative velocities between O"< I < 20". Around zero longitude the stellar distribution follows the high velocity gas component of the nuclear disk. Nuclear disk stars are heavily extinct ( A v > 20 mag, Messineo et al. 2003), as seen from Fig. 4, and their line-of-sight velocities range from 150 to -200 similarly to the gaseous nuclear disk line-of-sight velocities (cf. the "CO (1 - ,u) diagram in Fig. km SKI, 4 of Bally et al. 1988). The stellar nuclear disk (Ill < 1.5; Ibl < 0.5; A v >20 mag) rotates very rapidly around the Galactic center: our best-fit slope is 175(+40) km s-l per degree, consistent with the value (180 f 15 km s-l) found for the OHRR stars in the Galactic center (Lindqvist et al. 1992).

-

References Bally, J., Stark, A. A,, Wilson, R. W., and Henkel, C. 1988, ApJ, 324, 223 Bujarrabal, V., Planesas, P., and del Rornero, A. 1987, A&A, 175, 164 Bujarrabal, V. 1994, A&A, 285,953 Dame, T. M., Hartmann, D., and Thaddeus. P. 2001, ApJ, 547, 792 Egan, M. P., Price, S. D., Moshir, M. M., et al. 1999, AFRL-VS-TR-1999, 1522 Fux, R. 1999, A&A, 345,787 Glass, I. S., Matsurnoto, S., Carter, B. S., and Sekiguchi, K. 2001, MNRAS, 321, 77 Habing, H. J. 1996, A&A Rev., 7, 97 Lindqvist, M., Habing, H. J., and Winnberg, A. 1992, A&A, 259, 1 I8 Messineo, M., Habing, H. J., Sjouwerrnan, L. O., Omont, A,, and Menten, K. M. 2002, A&A, 393, 115 Messineo, M., Habing, H. J., Ornont, A., Menten, K. M., and Sjouwennan, L. 0.2003, in preparation Nymdn, L.-A,, Hall, P. I., and Le Bertre, T. 1993, A&A, 280, 551 Olivier, E.-A,,Whitelock, P., and Marang, F. 2001, MNRAS, 326, 490 Omont, A,, Gilrnore, G. F., Alard, C., et al. 2003, A&A, 403,975 Ortiz. R., Blommaert, J. A. D. L., Copet, E., et al. 2002, A&A, 388,279 Schuller, F., Ganesh, S., Messineo, M.. et al. 2003, A&A, 403, 955

-

297

Astron. Nachr./AN 324, No. S 1 (2003)

400

300

e,,,~onnecting arm

200

-

100

i

E 5

0

m 11:

" 2

-100

-200

1

-300

-400

I 40

30

20

10

0

-10

-20

-30

-40

Galactic longitude [degrees]

Fig. 3 Stellar longitude-velocity diagram overlayed on the grayscale CO (1 - v) diagram from Dame et al. (2001). The SiO 86 GHz masers are shown as dots. Gas features are labelled following Fig. I of Fux (1999).

-2

-4

Galactic longitude [degrees]

Fig. 4 Stellar longitude-velocity diagram of our 86 GHz SiO masers. The filled circles indicate sources with visual interstellar extinction larger than 20 mag. Most of those clearly helong to the nuclear disk, fast rolating component. The continuum line indicates our best fit to the nuclear disk component.

This Page Intentionally Left Blank

Astron. NachrIAN 324. No. S1. 299-302 (2003) / DO1 10.1002/asna.200385083

CNO Abundances in the Quintuplet Cluster M Supergiant 5-7 S. V. Rarnirez*’, K. Sellgren2, R. Blum’, and D. M. Terndrup’ ’ Infrared Processing and Analysis Center, CaliforniaInstitute of Technology, Mail Code 10@22,770 South Wilson Avenue, Pasadena, CA 9 I 125, USA Department of Astronomy,The Ohio State University, 140 West 18th Avenue, Columbus, OH 43210 USA NOAO/CTIO, Casilla 603, La Serena, Chile

Key words stars, abundances, Galactic center

Abstract. We present and analyze infrared spectra of the supergiant VR 5-7, in the Quintuplet cluster 30 pc from the Galactic center. Within the uncertainties, the [CIH], “/HI, and [OfH]abundances in this star are equal to those of a Ori, a star which exhibits mixing of CNO processed elements, but are distinct from the abundance patterns in IRS 7.

1 Introduction We have previously published a differential analysis of the iron abundance [Fe/H] in ten cool, luminous stars within 30 pc of the Galactic Center, compared to 1 I stars of similar temperature and luminosity in the solar neighborhood. We found that both samples of stars had a narrow distribution of [Fe/H] centered at the solar value (Ramirez et al. 2000). Carr et al. (2000) also studied [C/H], “/HI, and [O/H] in (1 Ori and in IRS 7, which lies at a prqjected distance of 0.25 pc from Sgr A*. They found that dredge-up of CNO-processed material was present in both M supergiants, as expected theoretically, but that the amount of internal mixing was much stronger in IRS 7, stronger than predicted by current evolutionary models. Our main goal is to investigate whether the strong internal mixing found in IRS 7 is due to the unusual conditions for star formation in the central 100 pc of the Galaxy (Morris 1993, Morris & Serabyn 1996),or is due to some tidal interaction between IRS 7 and the supermassive black hole (2.6 x lo6 Ma: Schoedel et al. 2002, Ghez et al. 2000) at the Galactic Center. Here we present preliminary results on [ C k l ] ,[ N/H], and [O/H] in the M supergiant VR 5-7, which lies in the Quintuplet cluster, 30 pc from the Galactic Center.

2 Observations and Analysis High-resolution (A/AA 25,000) K-band and H-band spectra of VR 5-7 were obtained through Gemini sponsored access to the Keck Telescope using NIRSPEC in June 2001. The spectra were reduced using IRAF, involving Rat fielding, sky subtraction, spectrum extraction, wavelength calibration, and removal of atmospheric absorption features. The observed spectra are of high quality, with a signal to noise ratio above 100, which is needed for detailed abundance analysis. The abundance analysis was done using a current version of the LTE spectral synthesis program MOOG (Sneden 1973). The program requires a line list with atomic and molecular parameters and an input model atmosphere for the effective temperature and surface gravity appropriate for the star. The atomic and molecular parameters (wavelength, excitation potential, gf-value, damping constant, and dissociation constant) were obtained the same way as in Ramirez et al. (2000), and also included C O molecular parameters N

* Corresponding author: e-mail: [email protected],Phone: + I 626395 1919, F d X : +I 626397 7018

@ 2002 WILEY-VCH Veiiag GmhH & Co. KCaA. Weiiihrini

S. Ram'rez et al.: CNO in 5-7

300

from Goorvitch (1994) and OH molecular parameters from Black (private communication). The solar abundance model atmospheres from Plez (1 992) were used for our abundance analysis. The stellar parameters of VR 5-7 were taken from Ramirez et al. (2000): effective temperature T.fi = 3500 K, surface gravity logg = -0.2, microturbulent velocity = 2.9 km s - ' , and macroturbulent velocity = 12.6 km

<

<

S-1.

3 Results and discussion Synthetic spectra were computcd in several regions of the K- and H - spectra, where CO, OH, and CN lines were selected. The resulting synthetic spectra (dotted lines) are plotted in Figures 1 , 2 and 3 together with the observed NIRSPEC spectra (filled squares). We identified four relatively unblended CO lines in the I<-band spectrum, four C O and four OH lines in the H-band spcctrum, and ten CN lines in thc K-band spectrum; thosc suspected of being blended with unidentificd lines or lines with uncertain gf values were not considered in the analysis. Our results of VR 5-7 are listed in Table I , together with the abundances of a Ori and IRS 7 from Carr et al. (2000).

0.

1,

,

,

18300

,

,

,

16305

,

,co,',

,

,

18310

,

,

,

, ,8315

,!, ,

18320

U4

Fig. 1 K-band

synthetic spectrum (dotted line) overplotted with the observed NIRSPEC spectrum (filled squares) for VR 5-7. In these and the following figures, the most prominent absorption features are shown; the ones used in the analysis were selected to be relatively unblended. The g f values for the Sc and HF features are not available.

Fig. 2 H-band synthetic spectrum (dotted line) overplotted with the observed NIRSPEC spectrum (filled squares) for VR 5-7.

Table 1 Abundances of VR 5-7

Star

[CIH]

~~~

[NM]

[O/H]

[(C+N+O)M]

Reference

~

VR 5-7

-0.3

+0.6

4.2

+O. 16

this paper

Ori

-0.3

+0.5

-0.2

-0.09

Carr et al. (2000)

IRS 7

-0.8

+0.9

-0.7

-0.1 I

Carr et al. (2000)

CY

Astron. NachcIAN 324, No. S I (2003)

301

0.8

, 22360

22365

L

3 0.9

0.8

1

,

.

..

L

0.9

I

I

I

22380

1

I

I

I

I '

22385

0.8 22400

I

/

/

I

22405

I

I

,

,

224 10

Fig. 3 K-band synthetic spectrum (dotted line) overplotted with the observed NIRSPEC spectrum (filled squares) for VR 5-1.

The statistical uncertainties arising from the scatter in abundance between individual lines of OH, CO, or CN, are less than 0.1 dex for 5-7. The systematic uncertainty in the abundances, however, from uncertain stellar parameters, is 0.2 dex. Within the uncertainties, VR 5-7 is has comparable CNO composition to c1: Ori, and its CNO composition is distinct from that of IRS 7. The spectral differences between VR 5-7 and cy Ori can be seen in Figure 4, and the differences between VR 5-7 and IRS 7 can he seen in Figure 5 . Current non-rotating evolutionary models predict that a 20 Mo star would have [N/C]=+0.8 dex at the end of He burning, similar to that of cy Ori and VR 5-7, whereas a rotating star with an initial main sequence equatorial velocity of 300 km s-l would have (N/C] = +1.3 dex (Meynet and Maeder 2000). IRS 7 has [N/C] = 1.7 dex, which implies stronger mixing yet. It is possible that IRS 7 is experiencing

S. Ramirez et al.: CNO in 5-7

302

tidal mixing due to its proximity to the super massive black hole (Alexander & Livio 2001, Alexander & Kumar 2000), given that its projected distance t o Sgr A* is 0.25 pc. To fully explore this possible interpretation we need to determine the C N O abundances of more stars in the Central Cluster. We now have similar data taken with NIRSPEC at the Keck Telescope and PHOENIX at Gemini South, for four stars located within 2.5 pc of Sgr A*, and also several solar neighborhood stars, taken with CSHELL at the IRTF. If the black hole is key to mixing in the Galactic Center we might see an effect on the abundance patterns for the Galactic Center stars as their projected distance increases. If the extreme abundances in IRS 7 are related to the star formation process in the inner Galaxy (Morris 1993, Morris & Serabyn I996), then we might expect to see similar patterns in the other Galactic Center stars and VR 5-7, but differences in the mean compared to the solar neighborhood stars. L""''"""""""'i

L " ' " ' ' " " ' ' ' " " ' " i 08

co olico

4

08

04

18275

18286

18280

i I

A

00

08

""

04

0.4 18300

10305

18310

183L5

18310

18300

10305

18315

,8310

MA1

Fig. 4 Comparison between H-band spectra of VR 5-7 (solid line) and ct On (dotted line).

Fig. 5 comparison between H-band spectra of VR 5-7 (solid line) and ct On (dotted line).

Acknowledgements The research described in this poster was partially carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. Support for this research has been provided by NSF grants AST-9619230 and AST-0206331. The authors wish to extend special thanks to those of Hawaiian ancestry on whose sacred mountain we are privileged to he guests. Without their generous hospitality, none of the observationspresented herein would have been possible.

References Alexander, T., & Kumar, P., 2000, ApJ, 549,948 Alexander, T., & Livio, M. 2001, ApJ, 560, L143 Can; J. S., Sellgren, K., & Balachandran, S. C., 2000. ApJ, 530, 307 Ghez, A. M., Moms, M., Becklin, E. E., Tanner, A,, & Kremenek, T. 2000, Nature, 407,349 Goorvitch, D., 1994, ApJS, 95,535 Meynet, G. & Maeder, A., 2000, A&A, 361, 101 Moms, M. 1993, ApJ, 408,496 Morris. M.. & Serabvn. E. 1996. ARA&A, 34.645 Plez, B., 1992, A&& 94,527 Ram'rez, S. V., Sellgren, K., Cam,J. S., Balachandran, S. C., Blum, R., Terndrup, D. M., & Steed, A,, 2000, ApJ, 537,205 Schoedel, R., et al. 2002, Nature, 419,694 Sneden, C . , 1973, Ph.D. thesis, Univ. of Texas-

Astron. Nachr./AN 324, No. S1,303 -307 (2003) / DO1 10.1002/asna.200385038

New results on the Galactic Center Helium stars Thibaut Paumard', Jean-Pierre Maillard', and Susan Stolovy2

' Institut d'astrophysique de Paris (CNRS), 98b Bd. Arago, 75014 Paris, France

* SlRTF Science Center, CalTech, MS 220-6, Pasadena, CA 91 125, USA

Key words infrared: stars, Galaxy: center, stars: early type, stars: Wolf-Rayet, instrumentation: spectrograph, techniques: radial velocities PACS 04A25 The cluster of helium stars around Sgr A* has been re-observed with the BEAR spectro-imager on CFHT, in the 2.06 pm helium line, at a spectral resolution of 52 km s-l and on a field of N 40". This new analysis confirms and completes a previous study at a spectral resolution of 74 km sC1 and on a smaller field of 24", corresponding to the central parsec (Paumard et al. 2001). Nineteen stars are confirmed as helium stars. These observations led to a clear differentiation between two groups of hot stars based on their emission linewidth, their magnitude and their positions relative to Sgr A*. The first class of 6 members is characterized by narrow-line profiles (FWHM -200kms-') and by their brightness. The other, fiinter in K by an average of 2 mag, has a much broader emission component of width rz 1,OOO km s-'. Several of the emission lines show a P Cygni profile. From these results, we propose that the narrow-line group is formed of stars in the LBV phase, while the broad-line group is formed of stars in or near the WR phase. The division into two groups is also shown by their spatial distribution, with the narrow-line stars in a compact central cluster (IRS 16) and the other group distributed at the periphery of the central cluster of hot stars. HST-NICMOS data in Paa (1.87 pm) of the same field reveal a similar association. The identification of the Paa counterpart to the He I stars provides an additional element to characterize the two groups. Bright Paa emitters are found generally associated with the narrow-line class stars while the weak Pact emitters are generally associated with the broad-line stars. A few particular cases are discussed. This confirms the different status of evolution of the two groups of massive, hot stars in the central cluster. As a by-product, about 20 additional candidate emission stars are detected in the central, high-resolution 19" field from the NICMOS data.

1 Findings from BEAR 97 He I 2.06 pm observation With the BEAR spectro-imager, an imaging FTS (Maillard 2000) the central pc of the Galaxy was observed in 1997 at a spatial resolution of 0.5" and spectral resolution of 74 km s-l in the He1 2.058 p m domain, covering a field of 24". The observation provided a homogeneous set of fully resolved line profiles. The spectro-imaging data were associated with Adaptive Optics data from CFHT in the K band (Lai ei (11. 1997) to check the possible confusion of sources. That particular study of the helium emission-line stars in the central parsec of the Galactic Center was published in Paumard et al. (2001). The main results can be summarized as follows: 1. 16 fully resolved P Cygni emission line profiles, cleaned of ISM emission, of purely stellar origin, were extracted. 2. they were found to divide into two distinct classes, with narrow (FWHMrr 200 km spl) and broad-line profiles (FWHM rr 1,000 km SKI). 3. a difference in K of 2 mag between the two classes was measured 4. the spatial distribution of the two groups is different, with the narrow-line objects arranged in a central cluster, and the other class dispersed in a ring beyond a radius of N 0.3 pc from SgrA*.

-

From these findings it was concluded that the group of narrow-line stars can b e considered as formed of stars in the LBV phase, and the other one of stars at the WR stage. @ 2003 WILEY-VCH Verlag GmbH & Co KGdA. Weinhem

304

T. Paumard et al.: Results on the Galactic Center Helium stars

2 Observations New BEAR data in He I 2.06 pm were obtained in June 2000 at higher spectral resolution (52 km s-l). The field, composed of three overlapping circular sub-fields, was wider, and the signal-to-noise ratio higher by a factor of 21 1.6. km s-’. Paa HSTNCMOS observations were taken in 1998 of the central parsec with Camera 1 (Stolovy et al. 1999) and of the central 4 pc with Cameras 2 and 3 (Scoville et al. 2003). Dithered Images were taken in filters F187N centered on the 1.87 pm Paa line and in F190N for the nearby continuum. By subtracting a suitably scaled F190N mosaic image from the F187N mosaic, a map of the stellar and interstellar Paa emission can be obtained. Figure 3 shows the central region of the composite Camera 2 and Camera 3 Paa image, for which the central 19” x 19” has a spatial resolution of 0.”18.

3 He I stars and Paa emission With the new BEAR data, almost all the stars mentioned in Paumard et al. (2001) are confirmed, except the star numbered ‘“6”. The Paa data show a bright, very small ISM feature and no stellar counterpart to this point-like He I emission. Four new broad-line stars are added. Two were out of the previously studied field, and the better signal-to-noise ratio is responsible for the other two new detections. The star “B5” was associated with IRS 13E. Maillard et al. (2003) have shown that there was indeed two emission line stars in the IRS 13E complex, namely IRS 13E2 and IRS 13E4. The broad line clearly detected in He1 belongs to E2 only from Fabry-Perot imaging associated with adaptive optics by Clenet et al. (2003). The line profiles and locations of the 19 stars are shown in Fig. 1 and Fig. 2a. Fig. 2b clearly confirms that the narrow-line stars are generally much brighter in K than the broad-line stars. For B11, the signal-to-noise ratio is just sufficient to claim a detection, hut not to derive reliable line parameters. Table 1 Physical properties of the helium stars: K magnitude, full width at zero intensity (FWZI) of the 2.06pm He1 line ( h s - ’ ) , Paa line flux in units of lop2’ Wcm-2, calibrated from Pacr emission in AF (Najarro et al. 1994). ID 180 is from the photometric list of Ott et al. (1999), He1 N3 is from Paumard et al. (2001).

[ ID I

Name

I

mK

FWZIH,I

470 460

dpaa

1.15 1.12

I

ID Name -

B1 B2 B3 B4 B5 B6 B7 BS B9 BlO B11 B12 B13 -

ID 180 IRS 7E2 IRS 9W IRS 15SW IRS 13E2 IRS 7W AF “W HeIN3 BSD WC9 IRS 29N IRS 15NE IRS 16SE2

12.2 11.8 11.5 11.0 12.0 10.6 11.5 12.7 10.6 9.9 11.2 11.8

1,500 1,900 1,400 1,300 1,600 1,200 1,800 1,600 2,000 1,100 1,500 940

0.28 0.83 0.76 2.27 0.44 4.40 1.05 0.14 0.37 0.07 0.37 1.15

Fig. 3a shows that the He1 stars are associated with the Paa emission stars. The narrow-line stars W cm-’ without taking into (circles) are coincident with bright Paa emitters (mean intensity Y 2 account IRS 34W, Table I), whereas the broad-line stars correspond to fainter Paa emitters (mean intensity Y 0.97 W cm-’). Other Paa emitters are present, which may be also associated with He I emission, but too faint to have been detected with BEAR. A source extraction with the StarFinder procedure (Diolaiti et al. 2000) gives 52 point-like emission features in the high-resolution 19“ central field, of which 43 are emission line stars with a high degree of certainty. The 9 other need further observation to rule out the possibility that these point sources are compact ISM features or incomplete continuum subtraction of stars. However, this result (Fig. 3b) is generally consistent with an independent analysis made by Scoville et al. (2003).

Astron. Nachr./AN 324. No. S 1 (2003)

4

305

Discussion

The central cluster of helium stars is confirmed with a total of 19 members currently identified. The various conclusions on the characteristics of these stars from the first paper, reviewed in the Introduction are confirmed: division into two groups from their linewidths (Fig. 1 and Fig. 2b), from their brightness and from their location (Fig. 2a). The difference of brightness (Fig. 2b) presents few exceptions which were already noticed in the first paper. One of the narrow-line stars (N7, IRS 34W) is weak and one of the broad-line stars (B 1 1, IRS 29N) is brighter than the average of the other stars of the same group. From a long-term photometric study (Ott et al. 1999) IRS 34W is indicated as a variable star. It was weak at the time of our observations (Ott et al., private communication). It was proposed in Paumard et al. (2001) that the group of bright, helium stars was made of LBV-type stars. The high intensity of Pacu (Table I), the variability of IRS 34W confirm that these stars are hot, mass-losing stars, still rich in hydrogen. On the contrary, the weakness of the Paa! emission combined with the very broad helium line are consistent with the other group being more evolved stars. A few sources are exceptions - IRS 13E2, AF, IRS 16SE2 - showing a broad-line He I profile, but strong Pacu emission. This apparent anomaly could certainly be due to the fact that the Paa filter is not perfectly adapted to distinguish between rich and poor hydrogen emitters. Since the Paa line is blended with another significative helium line, He 1(4-3) at 1.869 pm, the intensity detected by the F187N can remain strong even if the hydrogen emission is intrinsically weak. Already mentioned, the weakness of the K magnitude of IRS 34W (Table 1) is due to the star being in a phase of enhanced intrinsic extinction. Naturally, the measured Pact intensity is extremely weak, a factor 14 lower than the mean intensity. However, all these elements confirm the different status of evolution of the two groups of massive, hot stars in the central cluster. Assuming that all these stars were formed in the same star formation event, the differences in evolutionary state would come from the differences in their initial mass. The Pa0 data can help to address the question of whether the identification of emission line stars in the central region is complete or not. Possibly, about twenty new stars, associated with weak Pacu emission are detected in the central parsec (Fig. 3b). With only this indication, it cannot be concluded that they are more WR candidates. A deep, spectroscopic analysis using adaptive optics in the K band is needed. Besides more WRs, some of them could be Be stars, or could belong to the old star population as symbiotic or Mira-type stars in a phase of emission. At any rate, these data represent a new element in the census of spectral type in the central parsecs to better constrain the peculiar star formation conditions in this region of the Milky Way.

References ClCnet, Y.,et a1.2003, these proceedings Diolaiti E., Bendinelli, O., Bonaccini, D., Close, L., Cunie, D., & Pmeggiani, G. 2000, A&AS, 147,335 Lai, O., et al. 1997, In: Optical Telescopes of Today and Tomorrow. A.L. Ardeberg (ed), Proc. SPIE 287 I , 859 Maillard, J.P., 2000, In: Imaging the Universe in 3 Dimensions, E. van Breughel & J. Bland-Hawthorn (eds), ASP Conf. Sene 195,185 Maillard, J.P., Paumard, T., Stolovy, S.R., & Rigaut, F. 2003, these proceedings Najarro, F., Hillier, D.J., Kudritzki, R.P., Krabbe, A,, Genzel, R., Lutz, D., Drapatz, S., & Geballe, T.R. 1994, A&A, 285,573 Ott, T., Eckart, A., & Genzel, R. 1999, ApJ, 523,248 Paumard, T., Maillard, J.P., Morris, M., & Rigaut, F. 2001, A&A, 366, 466 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher, M., & Yusef-Zadeh,F. 2003, ApJ, submitted Stolovy, S.R., McCarthy, D.W., Melia, F., Rieke, G., Rieke, M.J., & Yusef-Zadeh, F., 1999, in The Central Parsecs of the Galaxy, ed. H. Falcke, A. Cotera, W.J. Duschl, F. Melia, M.J. Rieke, ASP Conf. Ser., 186, 39

306

T. Paumard et al.: Results on the Galactic Center Helium stars

B11)

B12)

Fig. 1 'ho upper rows: the 2.06 pm He I narrow-line profiles; five lower rows, the broad-line profiles, all on the same velocity range of -2,000 to +2,M)Okms-l. The intensities are in W m-' an-' (Continuum: a few lo-'' for NI-N5, a few or lo-'' for the other stars). Simple line models are overplotted allowing to determine the full width at zero intensity.

Astron. Nachr./AN 324. No. S 1 (2003)

I

"

'

'

.Bi21

'

'

'

307

I

'

'

'

' .

2600

' / , l 1 I / I 1 , 1 I , ,

1

1

"

1

,

"

,

~

~

,

,

,

'

1

'

'

,

,

'

'

'

"

,

~

om 0-

4

-

2000

+ +

5-

0 Ell

0 02

-~ -

-

o 8s

+

E J

+"

811

+"

+ N3

0 88

-s -

-

I0

I

,

sw -

0

a 0

87

/

+

II En:

u 83 ,

+

0 85

N5

I

+

+

+

1000

0 82 813

-

+ -

+

+

-

Q --

"

N :

1600

,

1 0

/

1

,

,

1 -5

,

1

1

a)

, -10

0-,

B

,,,,,,,,I,, D

~

,

1

,

I

I 10

, ,,

I I I I I I

I

I I I I I I

I

,

I

I

I2

I

,>

I

I

I ,

ib) 13

Fig. 3 a) Pact map of the central 2 pc, with the inner 19" comprised of the high resolution Camera 2 data. Diamonds indicate the locations of the broad-line stars while the narrow-line stars are indicated by circles. All of these stars show emission in this Paa filter. b) All emission line star candidates, in the central region, marked by square boxes. The intensity scale is stretched to show fainter emission.

This Page Intentionally Left Blank

Astron. NachrJAN324, No. SI, 309-313 (2003) / DO1 10.1002/asna.200385085

Ten Thousand Stars Toward the Galactic Center * Frangois Rigaut**’,Robert Blum2,Tim Davidge3,and Angela Cotera4 ’ Gemini Observatory, 670 N A’Ohoku place, Hilo, HI-96720, USA Cerro Tololo Interamerican Observatory, La Serena, Chile Hertzberg Institude for Astrophysics, Victoria, Canada Steward Observatory, Tucson, USA

Key words Astronomical Instrumentation, Stars, Star formation

Abstract. We report on the Galactic center data set obtained at the Gemini Observatory as part of the demonstration science program with the Hokupa’a adaptive optics system. The data set is presented and characterized. Preliminary results on the stellar populations from CO band data and IRS8 are presented.

1 Introduction To check the performance of the telescope and the Hokupa’a+QUIRC instrument, to test observing procedures and data reduction methodologies and train the science staff, the Gemini Board approved in early 2000 a program of “Demonstration Science” for this instrument. After exchanges with the science community through their representatives at the GSC, the Galactic Center was chosen as the most relevant program. A science group was put together, involving members from the Gemini partner countries. Taking advantage of the strength of Hokupa’a (faint guide star, large imaged field) and keeping its weaknesses in mind (lower compensation capabilities than e.g. the Keck system due to the limited number of actuators), the science group converged toward the following science objectives: (1) Stellar Population: Determine the properties of the stars in function of their position in this region, particularly of their distance to Sgr A*. The diagnostics selected were H, K’, CO and CO continuum images. Because of the limitations of the A 0 system at short wavelengths, J was not selected, (2) Variability: Monitor a couple of fields (Central Sgr A* and Arches cluster) for variable stars, (3) Map the extremely structured extinction, and (4) Creatiodenlargement of an astrometric database over a larger field than previous data sets. 2 GB of data were obtained, covering a total of 4000 square arcsec in the vicinity of Sgr A*. These images have been reduced and were released to the international community in November 2000. The data set is briefly described in this paper, together with preliminary analysis of some remarkable features of this region unveiled by these observations (Rigaut et al 2003). A number of papers have been published to date that use this data set (see references).

2 Instrumentation and Observations The University of Hawaii Hokupa’a A 0 system, developed and operated by the UH A 0 group with the support from the NSF, uses curvature sensing and is able to use rather faint guide stars (nominally down * Based on ObSeNatiOnS obtained at the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under a cooperative agreement with the NSF on behalf of the Gemini partnership: the National Science Foundation (United States), the Particle Physics and Astronomy Research Council (United Kingdom), the National Research Council (Canada), CONICYT (Chile), the Australian Research Council (Australia), CNPq (Brazil) and CONICET (Argentina) ** Corresponding author: e-mail: [email protected],Phone: + I 808 9743686, Fax: +1 808 935 9650

@ 2W3 WILEY-VCH Veilag GmbH d To KGaA, Weinheim

Astron. Nachr./AN 324. No. S1 (2003)

311

originally for use on a 4-m class telescope (36 subapertures), and therefore is undersized for use on Gemini in the sense that it has too few actuators for adequate compensation at short IR wavelengths under typical seeing conditions. Reasonable performance in terms of compensated image quality (IQ) can be obtained only under periods of medium to good seeing. The data cover 11 fields of 20” x 20” each. Fields 1 4,12, and 13 are the central field data set. Fields 9 and 10 are control fields to check the properties of the bulge background population. Finally, field I 1 is the Arches Cluster. Fig.1 shows the respective location of the fields. H and K’ images are available for all fields. Dots next to the field number indicate fields for which CO (2.29 pm) and CO continuum (2.26 pm) have been taken. The guide star location for each field is marked with an (orange) star symbol. Guide star magnitude range from 13.5 to 15.1. The data reduction was done in the standard way, by several members of the science team, and images were cross checked to insure the quality and reproducibility of the data reduction process. Depending on the seeing during acquisition, the IQ on the image of this data set varies between 0.085” and 0.18” close to the guide star (median=O.l225”) and 0.11” to 0.23” at the edge of the field (median 0.155”).Crowding in the central part of the cluster goes up to 10 detected stars per square arcsec. Limiting magnitude in this data set varies from K’ x 17 in the most crowded area to K‘ > 21 in the least crowded area.

3 Identification of the young bright stars in the GC core from CO data We report here results from an analysis of the CO data in field 1 and 2 (the field containing Sgr A* and the one immediately to the north of it), followed by a preliminary interpretation of the data. The CO feature is usually a tracer of late-type giant and supergiant stars. By doing these observations, we were trying to discriminate between various stellar population ages. A combination of a CO filter (2.29pm, bandpass=20A) and a K continuum (2.26pm, bandpass=60A), also called CO continuum here, was used. A longer wavelength continuum was not available for these observations. We note that the use of a single continuum filter might affect the estimate of the CO absorption in the case of objects with large color gradients. However, our data indicate that this effect is small compared to the photometric errors. A 0 observations are particularly suited to narrow band photometry, as the observations in the filter and in the immediately adjacent continuum can be taken close enough to each other in time (possibly even interlaced) that the same turbulent conditions apply. The isoplanatic PSF degradation off the guide star is also very similar, considering that the wavelengths are very close. Therefore, even if the error on the absolute magnitude versus the position in the field in the CO filter remains large, for example, the same error is encountered in the continuum filter, thereby compensating for the introduced error. Figure 3 presents the (CO - K continuum = CO index) vs (K continuum) CMD. The CMD for field 2 is shown on the right hand panel. The scatter about a linear fit can be considered as an upper limit estimation of the noise. The scatter on the CO index is 2.5% (0.025 magnitude, 1 sigma) for Kc < 16. To our knowledge, it is one of the first quantitative estimates of the photometric noise in A 0 and, we believe, the lowest. The left hand panel of figure 3 presents the same CO index vs CO continuum, but including field 2 and field 1 -the field containing Sgr A*. The rightmost sequence traces the Asymptotic Giant Branch (as for field 2). But another sequence, with a constant CO index over 5 magnitude range, is apparent. This traces a different stellar population which corresponds to less evolved, younger stars. The occurrence of relatively recent star formation in the galactic center is already known. However, these observations present an additional diagnostic and allow us to trace with higher accuracy the location and concentration of this star formation event (see Fig. 4), although it is doubtful that this data set can address whether these stars have been formed in situ or fell into the core after their formation.

312

F. Rigaut et al.: loo00 Stars toward the GC

Fig. 2 A K‘ image of the central 40” x 40” of the data set. North is up, East to the left. Sgr A* is in the central bright cluster SW of the image. IRS 8 and its bow shock (Rigaut et a1 2003) are north of Sgr A*, West of the high extinction region.

Astron. Nachr./AN 324, No. S1 (2003)

. .

313

,

Fig. 3 CO index (CO - K continuum) vs K continuum for field 2 (right) and field 1 and 2 (left). The two boxes in the left hand panel isolate stellar populations that are only present within N 15" from Sgr A*. Each of the two boxes is associated with a symbol (circle of square) that has been used to mark the objects contained in this box in Fig.4.

Fig. 4 K' image of field I and 2 (20" x 40"). Sgr A* is located by a blue circle. The symbols identify stars of peculiar CO index (see Fig.3).

References Stoke, A., Grebe], E.K., Brandner, W., Figer, D.F. 2003, A&A, in press Figer, D.F. 2002, IAU symposium 212 Yang, Y., Park, H.S., Lee, M.G., Lee, S.-G. 2003, Journal of the Korean Astronomical Society, submitted Rigaut, F., Geballe, T., Roy, J.-R., Draine, B.T. 2003, these proceedings DePoy, D.L., Sellgren, K., Blum, R.D. 2003, BAAS, 198 Figer, D. F. 2003, these proceedings Tanner, A., Ghez, A., Morris, M., Becklin, E. 2003, these proceedings

This Page Intentionally Left Blank

Astron. Nachr./AN 324,No. S1,315-319 (2003) / DO1 10.1002/asna.200385073

Stellar Orbits at the Center of the Milky Way N. Mouawad*’, A. Eckart I , S. Pfalzner I , J. Moultaka I , C. Straubmeier I , R. Spurzem *, R. Schodel 3, and T. Ott

’ LPhysikalisches Institut, Universitat zu Koln, Zulpicher Str.77.50937 Koln, Germany Astronomisches Recheninstitut,Monchhofstr. 12-14,69 120 Heidelberg, Germany

’ Max-Planck-Institut fur extraterrestrischePhysik GiessenbachstraRe,85748 Garching, Germany Key words Galactic Center, kinematics and dynamics, inner cusp, central stellar cluster.

Abstract. During the past ten years, measurements of stellar proper motion, and radial velocities (Eckart et al. 2002, Genzel et al. 2000, Ghez et al. 2000) as well as variable X-ray emission (Baganoffet al. 2001) near the center of the Milky Way have convincingly proven the presence of a super-massive3 million solar masses black hole in the center of our Galaxy. We discuss the possible amount of the unresolved mass present at the inner cusp, and its translation into Newtonian periastron-shifts for stellar orbits in the central cluster. For this purpose we use a 4th-order Hermite integrator. Further calculations provide valuable additional information on the three dimensional distribution and dynamics of the He-Stars. We also discuss how future observations with infrared interferometers (LBT, VLTI, Keck) will help to improve our understanding of the dynamics and distribution of the stars in this region,

1

Introduction

As a result of the rapid development of observation techniques, near-infrared (NIR) high resolution imaging and spectroscopic observations with the speckle and adaptive optics (AO) techniques are now able to probe the inner arcsecond of our Galaxy. Observations were made with the MPE speckle camera SHARP at the ESO NTT and the new adaptive optics CONICA/NAOS at the UT4 of the Very Large Telescope (VLT). In the gravitational potential at the center of the Milky Way, the stars show large orbital velocities. In the central arcsecond those can be observed as proper motions via repeated imaging at the highest possible resolution. Ten year of observations have provide sufficient data determine, for the first time, a unique Keplerian orbit for the star S2 (Schodel ct al 2002) and to measure the enclosed dark mass down to a distance of a mere 0.6 mpc from Sgr A*. With this result many alternative scenarios appeared to be unrealistic, leaving a central super-massive black hole as the only plausible explanation. In addition a central stellar cusp was discovered. In the following we discuss how the presence of an extended mass influences the orbits of the inner most stars. The discussion of the central cusp mass is carried out via several steps: First, we calculate by a direct integration the amount of the stellar mass present within the 0.55” radius of Sgr A* due to the cusp (Genzel et al. 2003 submitted). Next, we briefly discuss how the current data can be used to put an upper limit on the cusp mass and the M/L ratio of the cusp. We study the periastron-shift of the S2 orbit, using a 4th order hermit integrator. Finally, we give a first estimate of the line of sight positions of the He I emission line stars, which are enigmatically present in the central cluster at distances not bigger then about 400 mpc from the center. * Corresponding author: e-mail: nelly@phl mi-koeln.de, Phone: +49 221 470 3495. Fax: +49 221 470 5 162

@ 2003 WILEY-VCH Verlag GmbH & Co KGrA. Wcinheim

N. Mouawad et al.: Stellar Orbits at the Center of the Milky Way

316

2 Mass Estimate of the Inner Cusp 2.1

Stellar Density

With the highest spatial resolution observations presently available in the near-infrared (50 - 60 mas), spatial scales from light hours to a few light years can be probed. The new CONICA/NAOS data were estimated from direct and crowding corrected K, 5 17 and direct ( H S 19) stellar counts in annuli centered on the position of Sgr A*. They clearly confirm the presence of a local stellar concentration, a cusp centered on Sgr A* , indicated earlier by the SHARPNTT and KECK data (Eckart et al. 1995, Alexander 2000). To estimate the cusp mass, we were able to fit the combined SHARP and CONICA stellar count data with a superposition of several Plummer models of the form: p(r)=p(O)[ l+(rR)2]-"/2 (a=S),with different densities p(O), and different core radii. Fig. 1 shows three different fits. The dotted curve gives the fit for the inner cusp with a Plummer model of a core radius R= 0.55" and a spatial density p(r)=4.35x lo7 Ma pcc3. The solid line shows the sum of that initial model with a further inner Plummer model of R= 0.135" and p(r)=6.5x*108 M, p C 3 . The average fit is shown as the large dotted curve. It is similar to the solid curve but with a smaller density of p(r)=3.2Sx 10' Ma pcP3 of the additional R= 0.135" component. Surface density/Distance from SgrA': plummer model fit

h

0 fn (I) u)

e

2

10

s

0u) L

(I)

Q

(I)

8 'v)5 0.1

Distance from SgrA' [arcsecs]

Fig. I A Plummer model fit to the surface density of stars as a function of distance. The grey, filled circles represent the CONICNNAOS data for K,<18. The darker, filled diamonds represent the SHARPNTT data for stars with K,< 15, and scaled upward by a factor of 5 in order to match the fainter CONICA/NAOS counts (Genzel et al. ZOOO). The small dotted curve and the solid curve represent the minimum and maximum fit to the data with our model, respectively. The big dotted middle curve represents the average fit.

LIt

(€002)1 S 'ON 'PZC N V / ' J P N .uoJW

N. Mouawad et al.: Stellar Orbits at the Center of the Milky Way

318

and grey dotted. Similar to Fig. 1 they correspond to maximum, mean, and minimum fit, respectively. Using a fourth order Hermite integrator derived from the one used in high-accuracy N-body simulations (Aarseth 1999, Spurzem 1999, for the first introduction of the Hermite scheme see Makino & Aarseth 1992), and adapting the mid-value of our models, we computed, for an S2 like orbit (Schodel et al. 2002), the trajectory of a star through the extended mass and around the BH .The Hermite scheme allows a fourth order accurate integration based on only two time steps. For that it requires the analytic computation of the time derivative of the gravitational force; therefore the use of Plummer model superpositions as they are used here is very convenient. Further studies with more general density distributions should be undertaken, because they may influence the precise value of the periastron shift. It should also be noted that this integration is purely classical so any relativistic periastron shifts are not taken into account. The resulting retrograde periastron shift amounts to a value of -1.7 arcmins per orbital revolution wich is few times smaller than the relativistic prograde periastron shift (Rubilar and Eckart 2001). Fig. 3 gives us different periastron shift values for different cusp masses, i.e. mass to light ratios.

3

Orbital Fit to the star S2 in the presence of a massive stellar cusp

The gravitational potential in the central 0.5 pc is dominated by a point mass (Eckart and Genzel 1997; Genzel et al. 2000; Ghez et al. 2000; Eckart et al. 2002; Schodel et al. 2002). Currently the most constraining case appears to be the one of the early type star S2 which shows very large velocities in the vicinity of Sgr A* . This star passed through its pericenter in April 2002, at a distance of only 17 light hours from the radio source. At that time had a maximum velocity of 25000 k d s . Schodel et al. (2003, submitted), which we were able to fit by a bound, highly elliptical Keplerian orbit (~=0.87),with an orbital period of 15.73 years, requiring an enclosed point mass of ( 3 . 3 5 0 . 7 ) ~lo6 Ma. Fit tar an 52 like SIW around a 3.3 Millcon Solar Mass BH with 0% cusp

Fa tar 8” 52 like slai wound a 3 7 M W n Solar Mass BH with 10% cusp

2

6

6

$c

4

a

2

g P *

0

I 2 6

4

-2

0 2 r8lahve R Atin mpc]

4

6

Fig. 4 Best fit for a point mass 3.3 million solar masses BH.

-6

-4

-2

2

0 ieilive R.A[m

4

6

mpcl

Fig. 5 Best fit for a point mass 3.3 million solar masses BH plus a 10%of extended mass.

In Figs.4 and 5, respectively, we show a closed Keplerian and an open rosetta shaped orbit due to the presence of an extended mass fraction of about 10% of the total mass. Both orbits fit very well the observed S2 data, and agree within the errorbars with the orbital elements determined by Schodel et al. (2003);both fits result in a 0.86 ellipticity and roughly 15 years orbital period, with a 2002.29 pencenter passage. Investigating the observed data within the framework of possible rosetta shaped orbits allows us to derive a valuable upper limit on the M/L ratio within the cusp. The involved cusp mass lies above the 5500 M a deduced from direct integration. This difference can be explained by the fact that for the direct integration we assumed a value of the mass to light ratio M/L=2 (see paragraph 2.2). This is work in progress and

Astron. Nachr./AN 324.No. S 1 (2003)

319

will be fully presented in a forthcoming paper (Mouawad et al. 2003). With new observational data we will be able t o soon constraint the M/L ratio and the involved cusp mass with a high precision. Data with higher angular resolution, as obtainable with the LBT, VLTI, and KECK interferometry (Eckart et al. 2002, Eckart et al. 2003), will be of great value for such an investigation.

Acknowledgements Support by the SFB494 at the University of Cologne and partial support by the SFB439 at the University of Heidelberg is acknowledged.

References Alexander,T. 1999, ApJ, 527,835 Aarseth,S.J.PASP,lll, 1333 Baganoff, F.K., et al. 2001, Nature, 413.45 Eckart, A,, Genzel, R., Hofmann, R., Sams, B. J.& Tacconi-Garman,L. E. 1995, ApJ, 445 ,L23 Eckart, A. & Genzel, R. 1996, Nature, 383, 415 Eckart,A., Mouawad,N., Krips, M., Straubmeier, C., and Bertram, T. 2002, SPIE 4835-03, Cnf. Proc. of the SPIE Meeting on 'Astronomical Telescopes and Instrumentation', held in Waikoloa, Hawaii, 22-28 August 2002. Eckart, A., Genzel, R., Ott, T. & Schodel, R. 2002, MNRAS, 331,917 Eckart, A., Bertram, T., Mouawad, N., Viehman, T., Straubmeier, C., Zuther, J. 2002, Contribution to the JENAM meeting on: The VLTI - Challenges for the Future, Porto, Portugal, September 4-7,2002 Genzel, R., Pichon, C., Eckart, A,, Gerhard, 0. & Ott, T. 2003, MNRAS, 317, 348 Genzel et al. 2003, submitted Ott, T., Eckart, A,, Genzel, R. 2000, ApJ, 523,248 Ghez, A,, Moms, M., BecMin, E.E., Tanner, A. & Kremenek, T. 2000, Nature, 407,349 Makino, J. & Aarseth, S.J. Proc. Astron. SOC.Japan, 44, 141 Mouawad et al. 2003, in preparation Rubilar, G. F.& Eckart, A. 2001, A&A, 374, 95 Spurzem, R., JI. Comp. Appl. Math, 109,407 Schodel, R., Ott, T., Genzel, R., Hofmann, R.; Lehnert, M.; Eckart, A,, Mouawad, N., Alexander, T., Reid, M. J., Lenzen, R., Hartung, M , Lacombe, F., Rouan, D., Gendron, E., Rousset, G., Lagrange, A,-M., Brandner, W., Ageorges, N., Lidman, C., Moonvood, A. E M., Spyromilio, J., Hubin, N., Menten, K. M. 2002, Nature, 419, 694 Schodel et al. 2003. submitted

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.321-325 (2003)/ DO1 10.1002/asna.200385105

Dynamical Friciton near the Galactic Center

’,

Sungsoo S. Kim* Donald F. Figer’, and Mark Morris2

’ Space Telescope Science Institute, 3700 San Martin Dr. Baltimore, MD 21218, USA

* Division of Astronomy, University of California, Los Angeles, CA 90095-1562, USA Key words Galactic center, numerical simulations, dynamical fricton, star clusters, star fonnation Abstract. Numerical simulations of the dynamical friction suffered by a star cluster near the Galactic center have been performed with a parallelized tree code. Gerhard (2001) has suggested that dynamical friction, which causes a cluster to lose orbital energy and spiral in towards the galactic center, may explain the presence of a cluster of very young stars in the central parsec, where star formation might be prohibitively difficult owing to strong tidal forces. The clusters modeled in our simulations have an initial total mass of 10”-106 M o and initial galactocentric radii of 2.5-30 pc. We have identified a few simulations in which dynamical friction indeed brings a cluster to the central parsec, although this is only possible if the cluster is either very massive (- lo6 Ma), or is formed near the central parsec (5 5 pc). In both cases, the cluster should have an initially very dense core (> lo6 M ~ p c - ~The ) . initial segregation of massive stars into the cluster core can help achieve the requisite density, and can help account for the observed distribution of He I stars in the central parsec.

1 Introduction The central parsec of the Galaxy contains a cluster of very young stars (3-7 Myr old), whose in situ formation is highly problematic owing to strong tidal forces in the central parsec region. Motivated by the 30 pc away from the Galactic fact that similarly young stars have also been found in two clusters at center, Gerhard (2000) proposed that dynamical friction can bring a massive young star cluster into the central parsec during the lifetime of its most massive stars. If the star cluster is massive enough, it will experience a dragging force in the direction opposite to its motion due to the “wave” of background stars, and such dynamical friction will make the cluster sprial into the Galactic center. We have performed numerical simulations for the orbital and structural evolution of star clusters situated in a realistic Galactic potential in order to accurately calculate the timescale of the dynamical friction at the central region of the Galaxy and to predict the final distribution of remnant stars after disruption.

2 The Simulation We have performed 13 simulations using an N-body/SP€I code named GADGET(Springel, Yoshida, & White 2001). Their parameters are listed in Table 1 . The initial cluster parameters are as follows: R i s the galactocentric radius, v,,,~ the orbital velocity, Mcl the total mass, N,l the number of particles, T , the core radius, rt the tidal radius, and p, the core density. The number of particles used for the Galaxy is up to 2 x lo6. More information on the simulations can be found in Kim & Morris (2003). Some of the results are shown in Figure 1 , which displays the evolution of the radial distribution of cluster particles (left panel), and a histogram of the radial distribution at the final simulation step (right panel). * Corresponding author: e-mail: kim @stsci.edu

@ 2003 WILEY-VCH Vrdag GmhH & Co KGaA, Weinhsim

S. Kim et al.: Dvnamical Friction

322

Table 1 Simulation Parameters

Simulation 1 2 3 4 5 6 7 8 9 10 11 12 13

R (pc)

vznzt

(wccvc)

30 30 30 30 5 10 10 5 2.5 2.5 2.5

5 10

1

0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 1 0.5 0.5

Mc1 (Mo) lob lo6 lo6 lo6 lo6 lo6 lo6

lo5 lo5 lo5

lo5

rc Ncl 10' lo5 lo5

lo5 lo5 lo5

lo5 lo4 lo4

lo4

lo5

lo4 LO4

lo5

lo4

(pc) 0.86 0.86 0.86 0.28 0.28 0.43 0.28 0.13 0.080 0.039 0.039 0.039 0.039

rt/rc 6 6 6 19 6 6 9.3 6 6 12 12 20 30

P. Galaxy ( M ~ p c - ~ )Rotation 1.3 x lo5 N 1.3 x lo5 N 1.3 x lo5 Y 3.8 x lo6 N 3.8 x lo6 N 1.0 x lo6 N 3.8 x lo6 N 3.8 x lo6 N 1.6 x 10' N 1.3 X 10' N 1.3 x 10' N 1.3 x 10' N 1.3 x 10' N

Galaxy

Model 1 1 1 1

2 3 3 2 2 2 2 2 3

3 Conclusion We have identified a few simulations that can be regarded as candidates for the origin of the central parsec cluster. However, the required conditions are rather extreme. While it is more probable for a massive cluster (lo6 Ma) to reach the central parsec before disruption, one needs a finely-tuned set of parameters to observe only 16 He I stars out of the whole original mass and to have them be concentiated almost entirely into the central parsec. Less massive clusters (lo5 Ma)might have less problem with the He I star count, but require either an extremely high initial central density (inherently or via relaxation) or a rapid segregation of massive stars in the cluster core. Consideration of internal dynamics and mass spectrum in the cluster will enable the core collapse and mass segregation and thus tend to raise the probability for the cluster core to reach the central parsec intact. These same effects also make the cluster evaporate more quickly and weaken the drag from dynamical friction, thus lowering that probability. The internal dynamics of the cluster are more important for a more compact cluster located at a larger R. We have artificially and partially explored one of the consequences of the internal cluster dynamics by trying some models with extremely high central densities. However, numerical simulations that can handle both the galactic scale (dynamical friction) and the cluster scale (internal cluster dynamics) are still needed to show the exact role of the internal dynamics for the structural evolution of our cluster models for the time intervals considered here. For more on the simulations presented here and our discussion on them, see Kim & Morris (2003). N

References Gerhard, 0.2001, ApJ, 546, L39 Kim, S . S., & Moms, M. 2003, ApJ, submitted Springel, V., Yoshida, N., &White, S. D. M., 2001, New Astronomy, 6,79

Astron. Nachr./AN 324, No. S 1 (2003)

0.0

1.0

2.0

3.0

4.0

323

5.0

Time (My<)

t

0.0

0.0

2.0

10

0.1

3.0 Time ( M y )

4.0

0.2 Time (Myr)

0.3

5.0

0

0.4

Fig. 1 Grey scale map for the temporal evolution of number histogram for the radial distribution of cluster particles (left panels) and the same histogram for the final simulation step (right panels). The grey scale bars next to the maps represent the scales of the density in units of h f 0 p F 3 . In the right panels, thick lines represent all stars in the cluster, and thin lines represent the central 1 % stars at the beginning of the simulation. The horizontal dotted line in the map shows the location of R = 1 pc.

S. Kim et al.: Dvnamical Friction

324

0.0

0.2

2

0

0.6

0.4

4

00

02

04

06

08

0

1

2

0 2

04 Time (Myr)

06

10

(PC)

3

4

5

6

4

5

6

R (PC)

Time (My)

0 0

8

6 R

Time (Myr)

0

1

2

3 R (PC)

Astron. Nachr./AN 324, No. S 1 (2003)

02

0.0

325

0

06

0.4

0.0

02

0.4

0.6

0.8

R

1 0

1.2

1.4

0.5

1.0

1.5

Time ( M y )

20

2.5

3

4

(PC)

0

1

2 R (PC)

3

4

0

1

2 R (PC)

3

4

Time (Myi)

0.0

2

1

Time (Myr)

This Page Intentionally Left Blank

Astron. Ndchr./AN 324. No. S l . 327-33 1 (20031 I DO1 10.1002/asna.200385039

Near-infrared adaptive optics observations of the Galactic Center with NAOSKONICA (ESO) and GriF (CFHT) Y. ClCnet*', D. Rouan, F. Lacombe, E. Gendron, and D. Gratadour

' Observatoire de Paris, LESIA, 5 place Jules Janssen, F-92195 Paris Cedex, France Key words Galaxy: center, stellar content - Infrared: stars - Stars: imaging, emission-line - Instrumentation: adaptive optics, interferometers, spectrographs PACS 04A2.5 We report results obtained with GriF, an infrared 3-D spectrograph coupled to the CFHT adaptive optics system PUEO, and NAOS/CONICA, the ESO VLT adaptive optics system and its associated near-infrared camera. Observations with GriF aimed at surveying, for the first time at the diffraction limit, the so-called "He I stars", currently identified as hot massive stars in transition between the blue supergiant and the WolfRayet phases. NAOS/CONICA has delivered the first thermal infrared images of the Galactic Center region with a spatial resolution down to 0.1". Based on the Science Verification observations of NAOS/CONICA, we discuss here the photometry of S2, the closest star to Sgr A*, the discovery of a new bow-shock star and the putative detection of sources gravitationally amplified by Sgr A*.

1 Introduction In the near-infrared, recent works on the Galactic Center (GC) region were oriented towards three main topics: firstly, the search for the infrared (IR) counterpart of the Sgr A* radio source, secondly the study of the stellar population in relation to the star formation history together with the understanding of the close environment of the black hole, and finally the search for gravitational lensing effects of the central compact source upon the nearest stars. The near-IR is well suited for these studies, mainly because of the much lower interstellar absorption at these wavelengths compared to the visible. In addition, the high spatial resolution provided by adaptive optics (AO) in this wavelength domain is mandatory because of the high stellar density of this region. Based on the poster presented during the "Galactic Center Worshop" in Kona (November 2002), this article reports results obtained with two recent A 0 instruments: GriF at CFHT and NAOSKONICA (hereafter NACO), recently commissioned at ESONLT.

2 Observations 2.1

GriFrun

GriF (ClCnet et al. 2002) is an IR 3-D spectrograph coupled to PUEO, the CFHT adaptive optics system (Rigaut et al. 1998). It was used on 8 May 2001 in its Fabry-Perot mode to scan with a spectral resolution of 2000 the 2.058 pm He I emission in the GC. Images have been acquired with the IR camera KIR (Doyon et al. 1998), with a pixel scale of 0.0348"/pixel and a field of view of 36"x36", centered on Sgr A*. A blocking filter with a spectral resolution around 100 has been used to only transmit the Fabry-Perot order * Corresponding author: e-mail: [email protected], Phone: +33 1450775 48, Fax: +33 145 0779 17 @ 2003 WUEY-VCH Verlag GmbH & Ca. KGaA. Weinheim

Y. ClLnet et al.: NIR A 0 observations of the GC with NACO and GriF

328

scanning the 2.058 pm line. We show on Fig. 1 the resulting spectrum of IRS 13E2 (see Paumard et al. (2001) for the star identification), computed with Starfinder (Diolaiti et al. 2000).

2.2 NACO Science Verification run Observations presented here have been obtained on the VLT UT4 telescope (ESO) during the Science Verification (SV) of NACO (Brandner et al. 2002; Rousset et al. 2000; Lenzen et al. 1998). Images have been acquired with the IR CONICA camera in H , Ks (detector integration time -15s; pixel scale0.01326"/pixel, 29 August 2002) and L' bands (detector integration time - 0.2s; pixel scale - 0.027 I "/pixel, 19 August 2002). They were previously corrected from atmospheric turbulence with the NAOS A 0 system. H (X,=1.66 pm) and Ks band (X,=2.18 pm) images have been corrected with the IR wavefront sensor (WFS) and the A 0 system servoed on the bright K-band source IRS 7 (m~=6.5,about 5.5" to Sgr A*). The 7 . 2 " ~5.7" H and 7.9"x6.1" Ks band images are not shown here since similar images obtained during NACO commissioning runs have already been presented in Schodel et al. (2002). For software reasons, the visible WFS was the only one available for the SV L' band (X,=3.80 pm, Fig. 2) observations and the A 0 reference star was the one usually chosen for visible WFS (m~=13.8,about 25" northeast of Sgr A*). 4

'

'

I

"

.

*

.

.

'

.

I

.

.

'

83

.

.:A -i

LL

B

o

08

-

1

.a. -0 8 8

10

-

1

0

,

.

*

.

.

,

.

I

. . . .

I

.

.

.

a

.

-9 3

-1 0

3 The He I emission-line star in the IRS 13 cluster Till the GriF observations, results on He I stars (at 2.058 pm) with the best spatial (seeing-limited)/spectral resolutions have been obtained by Paumard et al. (2001). In non crowded regions, our results are in good agreement with them but their observations have suffered from a too low spatial resolution in the IRS 13 cluster: their identification of the He I emitter in this complex was indirect and based on the broad-line emission of the emitter and on the classification they have made from other He 1 emitters, associating broad lines with faint stars (identified as late Wolf-Rayet) and narrow lines with bright stars (identified as LBV). Fed by the A 0 system PUEO, GriF had the sufficient spatial resolution to clearly associate the 2.058 pm He I emission in the IRS 13 cluster with the IRS 13 E2 star, the southem-western source of the complex. Paumard et al. (2001) have already identified two exceptions to their classification, observing two faint stars with a narrow 2.058 pm He I line and explaining their relative faintness by dust obscuration. From the GriF observations, IRS 13 E2, as a bright and broad He I emission line star, is a second type of exception. It could be in transition between the LBV and the Wolf-Rayet phases or be an Ofpe/WN9 star, in transition between the blue supergiant and the Wolf-Rayet phases.

Astron. Nachr./AN 324, No. S 1 (2003)

329

4 The IR counterpart to Sgr A* On the 2”x2” L‘ band image of the Sgr A* cluster (Fig. 2), more than 20 stars are detected. Their dereddened colors, assuming the extinction law from Rieke & Lebofsky (1985), a distance to the GC of 8 kpc (Reid 1993) and an extinction value A ~ = 2 . 7(CICnet et al. 2001), are given in Table 1 . Table 1 Dereddened colors of the Sgr A* cluster stars observed in the 2” x2” L’ band image (Fig. 2). IRS 16 stars, saturated in H and Ks bands, are not included. Stars are numbered from left to right, and then from down to up. IDNumber Other name (HKs)~ (KS-II)~

1

2

3

4

5

Sl 0.4 -0.5

0.4 0

0.5 -0.1

0.5 0.3

0.4 -0.2

6 SS 0.4 -0.1

7 07 0

8

9

10

s4

s9 0.5 0.1

s11 0.3 0.1

0.4 -0.1

11

0.5 0.3

12 SIO 0.6 -0.2

13

14

15

s2 0.4 0.8

s1

s12

0.4 -0.3

0.4

0.5

16

17

18

19

20

21

22

0.4 0.2

0.8

0.3 3.7

0.6 -0.1

0.4

0.5 -0.5

0.7 -0.3

2.3

0.2

For these stars, we have found a mean color index value of ((H Ks)o)=0.5, with a standard deviation of 0.1. After removal of the two very red objects ID#17 and ID#18, we have derived a mean value of ((KS-L‘)~)=O,with a standard deviation of 0.3. S2 is the only star with a (Ks-L‘)o index larger than twice this standard deviation. Is this IR excess intrinsic to S2 or due to a contribution from Sgr A* ? The latter hypothesis cannot be directly ruled out since the distance between S2 and Sgr A* during the observations (about 0.04”) was lower than the NACO spatial resolution in L’ band (0.096”), making impossible to separate the S2 emission from the putative emission from Sgr A*. If we assume that S 2 has a (KS-L‘)~index among the highest of its neighbours (say (KS-L‘)~=O.~) and attribute the entire color excess to the L’ band flux (since the S2 (H Ks)o index is similar to the other surrounding stars index), we derive m~f=12.8and a dereddened flux density of F,=7 mJy. This result is higher than the previously published near-IR values (Genzel & Eckart 1999; Hornstein et al. 2002) and could then only correspond to a flare state of Sgr A*, favoring the jet emission model rather than the accretion model (Melia & Falcke 2001). To our knowledge, no simultaneous X-ray observations of Sgr A* have been made to confirm this flare state. May an intrinsic color excess hypothesis also be possible? S 2 colors match the intrinsic colors of a giant (M4 or M4.S) or a supergiant (MI or MI . S ) star, together with an appropriate extinction coefficient (A~=3.6),but a main sequence star seems excluded (Fig. 3). In the supergiant case, the visible absolute , to S2 observed magnitudes of (mH, magnitude taken from Cox (2000) (Mv=-5.6), with A ~ = 3 . 6 leads mKs, mLt)=(9.6, 7.8, 5.6), which are incompatible with the observed magnitudes (15.5, 13.6, 11.5). On the contrary, the giant case is compatible with these magnitudes as the corresponding visible absolute magnitude taken from Cox (2000) (Mv=-0.3 for MS class, the closest found in the appropriate table) leads to the following observed magnitudes: (mH, mKs, m~t)=(14.9,13.1, 10.9). But this giant hypothesis is invalidated by recent Keck A 0 spectroscopic observations (Ghez et al. 2003): no CO absorption is seen in the S2 K-band spectrum. To explain this L’ band excess, Genzel et al. (2003) proposed that dust in the accretion flow onto Sgr A* is heated by S2, considered as an hot 0 main sequence star.

5

A new “bow-shock star”

Several bow-shock stars, interacting with the Northern Arm, may have been discovered by Tanner et al. (2002, 2003) on photometric and radiative transfer modeling basis. Observed with A 0 in the near-IR at Gemini observatory (Cotera et al. 2001, Rigaut et al. 2003), IRS 8 is a bow shock star discovered from its morphology, which seems to result from the interaction between a fast moving star and the Northern Arm. The shape of the Northern Arm is clearly revealed by the NACO L’ hand image (Fig. 2). Less than 2’’ from it, 3.44” north and 2.90” east to Sgr A*, a star shows a bow-shock like structure to the northeast, in the direction of the Northern Arm. As with other how-shock stars, its windenvelope may interact with the gas and dust of the Northern Arm. The bow-shock figure is composed of 3 “hot spots” (with a common L’ hand surface brightness around 8.5 mag arcsec-’), which may indicate structure resulting from this interaction. The star itself shows red color indices, (H K S ) ~ = Oand . ~ (Ks-L‘)o=0.7, but the bow shock structure is much redder since it appears in L’ hand hut in neither the H nor Ks bands.

Y. Clenet el al.: NIR A 0 observations of the GC with NACO and GriF

330

A

B II

Y

-1

0

-1

2

1

3

(Us-L‘),

Fig. 3 Dereddened color-color diagram for the S2 star superimposed on the intrinsic colors of different classes of stars. The line represents the variation of the dereddened S2 colors with respect to the interstellar extinction, from AK=Oto A ~ z . 5 and , crosses on this line are for integer values of AK. Additional symbols correspond to intrinsic colors of stars from Ducati et al. (2001): diamonds are for main sequence stars, triangles for giants and squares for supergiants. Table 2 Amplification factors for the ten most likely image pair candidates of Alexander (2001), computed with the photometry in H , Ks and L’ bands, compared to the value of Alexander (2001). Few stars are not detected in L‘ band, leading to a lower limit value, and one star is out of the field of view in H and Ks band. Errors are computed after derivating the amplification factor formula. One should note there is an uncertainty about the identification of the second image of pair #9 since the corresponding K-band magnitude given in Table 1 of Alexander (2001) is not the same as in Genzel et al. (2000).

Pair number 1

2 3 4 5 6 7 8 9 10

A in Hband

1.12f0.05 1.92f0.64 1.07f0.03 5.94510.72 1.03f0.01 3.24f2.66 1.50f0.27 out of FOV 1.49f0.27 1.0Sf0.02

A in

A in

Ks band 1.04f0.004 1.92f0.16 1.05f0.005 5.94f2.68 1.05f0.004 3.24f0.66 1.66f0.10 out of FOV 1.66f0.10 1.07f0.007

L‘ band 5 1.02f0.007 11.37k42.96 5 1.0Sf0.02 1.77f0.50 F1.10f0.04 2.10f0.85 2.36f1.16 1.17f0.03 1S7f0.33 5 1.14f0.06

A from Alexander (2001)

1.05 3.82 1.03 2.29 1.04 2.15 2.21 1.03 1.03 I .05

6 Gravitational lensing Alexander (2001) has shown the value in looking for the gravitational lensing of a background source by the black hole. The discovery of two corresponding images would allow one to precisely locate the IR position of Sgr A*, providing an alternate method to Menten et al. (1997), who used SiO masers detections to correlate IR and radio coordinates. 1 The amplification factor for the brighter image is given by A = (1 - 10Am/’ where Am is the difference between the magnitudes of the two images. From our SV NACO images, we have computed this factor for the ten most likely image pair candidates that Alexander (2001) has found after Monte-Carlo

‘1

~

Astron. Nachr./AN 324, No. S 1 (2003)

33 1

simulations and Minimal Likelihood analysis (Table 2) . For each pair, results are compared with the value calculated by Alexander (2001) from the angular distances of the images from the lens and the Einstein angle found after the aforementioned simulations. For an image pair, if we accept a maximum difference of three times the error between all our values and the one from Alexander (2001), the remaining candidates for lensing are: #4, #5, #6 and #lo. If we only accept twice the error, only pairs #4 and #6 are still possible candidates.

7 Summary and conclusions We have presented the GriF scan of the 2.058 pm H e I line on the G C field, giving the first diffractionlimited H e I map of the region. Results are in good agreement with previous seeing-limited observations. A notable exception comes from the IRS 13E cluster where the H e I emitter has been wrongly identified till now and is actually IRS 13E2. We have also presented the NACO SV observations of the G C region, giving the near-IR images with the highest spatial resolution to date. The S2 L‘ band excess may come from a Sgr AZkflare though n o Xray data is available at the same period to confirm this hypothesis. In 2003, S2 should be far enough from Sgr A* to know if a L’ band emission from Sgr A* is observed. A new “bow-shock star”, interacting with the gas and dust of the Northern Arm, has been discovered in L‘ band. Its properties appear to be similar to IRS 8, the first GC “bow-shock star” observed in the IR by the Gemini A 0 team: the bow-shock has very red colors, larger than the colors of the star. Finally, a photometric analysis allowed us to eliminate six of the ten most likely image pair candidates for gravitational lensing given in Alexander (2001). This analysis of lensing didn’t account for any spectroscopic identification of the corresponding stars. As for the other points discussed here (nature of S2, properties of the bow-shock), this research will benefit greatly from the next spectroscopic high spatial resolution observations scheduled with NACO. gravitational lensing given in for the spectroscopic identification of the stars and will highly benefit spectroscopic high spatial resolution observations scheduled with NACO.

References Alexander, T. 2001, ApJ, 553, L149 Brandner, W., Rousset, G., Lenzen, R., et a]. 2002, The ESO Messenger, 107, 1 CICnet, Y., Rouan, D., Gendron, E., et al. 2001, A&A, 376, 124 Clenet, Y., le Co&er, E., Joncas, G., et al. 2002, PASP, 114, 563 Cotera, A,, & Rigaut, F. 2001, GCNEWS, 12,4 N. Cox, Springer-Verlag Diolaiti, E., Bendinelli, O., Bonaccini, D., et al. 2000, A&A Suppl., 147, 335 Doyon, R., Nadeau, D., VallCe. P. et al. 1998, Proc. SPIE, 3354, 760 Ducati, J. R., Bevilacqua, C. M., Rembold, S. B., Ribeiro, D. 2001, ApJ, 558, 309 Genzel R., & Eckart, A. 1999, ASP Conf. Ser., 186, The Central Parsecs of the Galaxy, ed. H. Falcke, A. Cotera, W. J. Duschl, F. Media, & M. J. Rieke, 3. Genzel, R., Pichon, C., Eckart, A., et al. 2000, MNRAS, 317, 348 Genzel, R., Schodel, R., Ott, T., et al. 2003, ApJ, in press Ghez, A. M., Duchhe, G., Matthews, K., et al. 2003, ApJ, 586, L127 Hornstein, S. D., Ghez, A. M., Tanner, A., et al. 2002, ApJ, 577, L9 Lenzen, R., Hofmann, R., Bizenberger, P., Tusche, A., et al. 1998. Proc. SPIE, 3354,606 Melia, F., & Falcke, H. 2001, ARA&A, 39, 309 Menten, K. M., Reid, M. J., Eckart, A.,Genzel, R. 1997, ApJ, 475, LI 1 I Paumard, T., Maillard, J.-P., Morris, M., Rigaut, F. 2001, A&A, 366,466 Reid, M. J. 1993, ARA&A, 31,345 Rieke, G . H., & Lebofsky, M. J. 1985, ApJ, 288,618 Rigaut, F., Salmon, D., Arsenault, R., et al. 1998, PASP, 110, 152 Rigaut, F., Geballe, T. R., Roy, J.-R., & Draine, B. T. 2003, these proceedings Rousset, G., Lacombe, F., Puget, P., et al. 2000, Proc. SPIE, 4007, 72 SchZidel, R., Ott, T., Genzel, R., et al. 2002, Nature, 419, 694 Tanner, A., Ghez, A. M., Moms, M., et al. 2002, ApJ, 575, 860 Tanner, A., Ghez, A. M., Moms, M., Becklin, E. 2003, these proceedings

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1,333-336 (2003) / DO1 10.1002/dsna.200385106

Radio Pulsars in the Galactic Center

',

T. Joseph W. Lazio* James M. Cordes2,Cornelia C. Lang3,Eric V. Gotthelf4, and Q. Daniel Wang5

' Naval Research Laboratory, Code 72 13, Washington, DC 20375-535 1 USA ' '

Dept. of Astronomy, Cornell University and National Astronomy &Ionosphere Center, Ithaca, NY 148536801 USA Dept. of Physics & Astronomy, 703 Van Allen Hall, University of Iowa, Iowa City, IA 52242 Columbia Astrophysics Laboratory, Columbia University, Pupin Hall, 550 West 120th St., New York, NY 10027 Dept. of Astronomy, University of Massachusetts, LGRT-B 619E, 710 North Pleasant St., Amherst, MA 01003-9305

Key words Galaxy:nucleus, pulsars, scattering, supernovae Abstract. Radio pulsars in the Galactic center would serve as useful probes of its magnetoionic medium, and possibly its spacetime structure. However, of the roughly 1200 known pulsars, no more than 1 % are within I" of Sgr A* and none have distances consistent with being within or behind the Galactic center. This deficit of pulsars is due to interstellar scattering so severe as to smear together individual pulses. We are engaged in multiwavelength surveys of the GC to identify radio pulsars. Using the VLA at 1.4 GHz we are searching for compact radio sources, with the objective of finding sources on which a periodicity search can be conducted at frequencies high enough to defeat the interstellar pulse broadening (> 10 GHz). With the Chandra X-ray survey of the Galactic center, we have also discovered a number of X-ray point sources with radio counterparts that may he radio pulsars. Our current total census of promising radio pulsars candidates is roughly 10 sources. We are pursuing periodicity searches of these objects on the 100 m Efflesberg and Green Bank Telescopes. The number of candidates also allow us to constrain the supernova rate, and therefore star formation history, in the GC.

1 Introduction The central 300 pc (2" at 8.5 kpc) of the Galaxy has the highest stellar densities in the Galaxy, with a central density near los M o P C - ~ .This region contains clusters of hot, presumably young, massive stars; large-scale features thought to originate from powerful stellar winds and supernovae; and a concentration of X-ray binaries. Consequently, the Galactic center (GC) should also have a large population of radio pulsars, both recently formed and recycled. Discovery and subsequent timing of radio pulsars in the G C would provide opportunities to probe its magnetoionic medium, gravitational potential, and star formation history. However, the pulsar census contains only a handful within a few degrees of the GC, none with distances in (or beyond) the GC. Extreme pulse broadening causes this deficit of radio pulsars. Compact G C sources (Sgr A* and OH masers) display enhanced angular broadening, in some cases exceeding 1" at 1 GHz (van Langevelde et al. 1992; Frail et al. 1994; Yusef-Zadeh et al. 1994). Pulses from a radio pulsar would be scattered as well, with the scattered rays also being delayed in time. The result is pulse broadening or a smearing of the pulse. The pulse broadening implied by the enhanced angular broadening is 3 5 0 ~ 6 ;seconds ~ (Cordes & Lazio 1997; Lazio & Cordes 1998), at a frequency U G H ~GHz. A traditional periodicity search to find * Corresponding author: e-mail: Joseph.Lan'[email protected], Phone: +1202 404 6329, Fax: + I 202 404 8894

@ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

J. Lazio et al.: Radio Pulsars in the Galatic Center

334

radio pulsars would require Y 2 10 GHz in order for the pulse broadening to be comparable to typical pulse periods, and the typically steep radio spectra of pulsars combined with smaller telescope beam sizes at high frequencies means that such a periodicity search would be quite costly with respect to telescope time. We are therefore pursuing alternate strategies €or finding radio pulsars in the GC. The first strategy relies on the fact that sufficiently luminous compact sources (e.g., Sgr A* and the OH masers) can be imaged at much lower frequencies. Any pulsar candidates identified in a radio imaging survey can then be targeted for a high frequency periodicity search. By virtue of being a targeted search, of course, such a search can be much deeper than a general periodicity search survey of the entire region. Cordes & Lazio (1997) estimated that the GC contains 1-1000 sufficiently luminous pulsars, depending upon the vigor of recent starbursts. The second strategy relies on the fact that pulsars can appear as compact X-ray sources. Compact X-ray sources having a compact radio counterpart can then be targeted for a sensitive, high-frequency periodicity search. This paper reports on-going work to find radio pulsar candidates in the GC. In 52 we describe a radio imaging survey using the VLA, and $3 we describe an X-ray survey using the Chandra X-ray satellite. We summarize some of our best radio pulsar candidates in 54 and conclude with some general estimates about the supernova rate and star formation history of the GC in 55.

2 Radio Observations We used the VLA in its A-configuration to survey the inner 1?5 of the GC at 1.4 GHz. In this configuration at this frequency, the angular resolution of the VLA is 1'.'5, well-matched to the expected scattering diameter of a compact source in the GC. In addition, the VLA's field of view is approximately 30' (FWHM). Thus, an overlapping hexagonal grid of a modest number of pointings suffices for the survey. The source detection algorithm in the individual pointings follows closely that described by Lazio & Cordes (1 998). Briefly, sources were found by searching for extra-statistical deviations from the expected noise-only histogram for a source-free image. In practice we found that the intensity histograms had nonGaussian wings and larger standard deviations than would be expected from thermal noise only. These non-Gaussian wings and larger standard deviations are consistent with resulting from the sidelobes from the intense extended emission in the GC direction (particularly from Sgr A). Our resulting catalog contains 250 sources. In order to obtain spectral information on the sources, we cross-correlated our catalog (the 2LC catalog, Lazio & Cordes 2003) with previous radio surveys of the region, most notably the 5 GHz survey of Becker et al. (1994). For a limited number of sources we have conducted additional 5 GHz VLA observations, principally those outside the latitude extent (lbl < 0") of the 5 GHz survey of Becker et al. ( 1 994) or sources that should have been seen in that survey unless their spectral indices are a < -1 ( S rx v").

3 X-ray Observations The Chandra X-ray Observatory has completed a survey of a 2" x I" region centered on the GC (Wang et al. 2002). Approximately 1000 compact sources have been detected from this survey in addition to a significant amount of diffuse extended emission. Based on estimates of the X-ray point source density derived from a Chandra observation of a relatively blank region of the Galactic plane, we expect that no more than roughly 10%of the sources are background sources (i.e., active galactic nuclei, AGN) with the remainder being Galactic. However, the number of X-ray counts for any individual source are sufficiently small that extracting a reliable X-ray spectrum is difficult. We have compared the X-ray catalog to a variety of radio catalogs available in the literature as well as the 2LC described above. Our success rate of identifying radio counterparts to the X-ray point sources has

Astron. Nachr./AN 324,No. S1 (2003)

335

been low ( w 1%).Given the large fraction of Galactic sources expected, this low success rate is somewhat puzzling. Moreover, the hyperstrong scattering region toward the GC implies an angular diameter for AGN possibly exceeding 1 arcminute at frequencies near 1 GHz (Lazio & Cordes 1998; Bower et al. 2001). Many of the radio catalogs used in the cross-correlations are biased againsr finding large diameter sources, so there should be little contamination from AGN in the cross-correlations. We have not yet resolved this apparent paradox.

4 Radio Pulsar Candidates From the radio survey, we identify pulsar candidates as compact, steep-spectrum objects. Based on the diameters of Sgr A* and the OH masers, pulsar diameters should be roughly 1”. The typical spectrum of a radio pulsar is cu = -1.6 (Lorimer et al. 1995); we adopt a selection criterion of a < -1, so as to guard against excluding otherwise acceptable candidates with slightly flatter spectra than the typical pulsar. In addition, the effect of the hyperstrong scattering region should be to bias us against finding AGN in the 2LC, thereby minimizing the possible AGN “contamination” of our sample (53). We have detected (a) 3 sources with steep spectral indices; (b) 3 sources that should have been seen in the 5 GHz survey (Becker et al. 1994) unless their spectral indices are a < -1 ( S c( v”); and (c) 28 sources with the diameters expected for heavily scattered pulsars, but outside the latitude limit, Ibl < 0?4, of the 5 GHz survey. Within the 2LC catalog there are only two sources having X-ray counterparts within the accuracy of the Chundru positions (= 2”). Subsequent 8.4 GHz observations with the VLA have shown that one of these sources 2LC 359.220-0.186 remains unresolved (< 0’12) and has an inverted spectrum. At this frequency the expected contribution from scattering is no more than Cf’01. We have not yet reached any definitive conclusions about the nature of this source: 0

It could be an extremely compact AGN seen through a “hole” in the Galactic center’s scattering screen. As such it would be similar to G359.87f0.18 (Lazio et al. 1999), though that source shows extended, intrinsic structure.

If the source is in the Galactic center, its equivalent linear dimension is of order 5000 AU, comparable to those seen for hypercompact H 11 regions (Tieftrunk et al. 1997). However, there is no infrared counterpart in the MSX survey (21 20” resolution) of the GC (Egan et al. 1999). 0

It could be a pulsar with a high-frequency flattening in its spectrum, which is unusual though not unprecedented (Maron et al. 2000). For instance, PSR 50631+ I 0 is a pulsar with a flat spectrum above 1.4 GHz that was detected initially as an X-ray source (Zepka et al. 1996).

We therefore regard 2LC 359.220-0.186 as a promising source, particularly given its flat spectrum which will make a periodicity search easier at the high frequencies needed to defeat the pulse broadening. Future work involves periodicity searches on this object and the other objects identified in the 2LC catalog. Both the 100-mEffelsberg and the Green Bank Telescope (GBT) are sensitive telescopes capable of operating at high frequencies, and both have or will have shortly pulsar backend systems capable of carrying out such a search. We are also cross-correlating the Chundru X-ray survey with the 330 MHz survey by Nord et al. (2003). Initial indications are that we may find another few candidates from this comparison.

5 Supernovae in the Galactic Center Supernovae are dynamically important in the GC: They may be a source of the lo’ K, X-ray gas filling the central 1” of the GC and their number provides a constraint on the vigor of recent starbursts. Taking

336

J. Lazio et al.: Radio Pulsars in the Galatic Center

into account various selection effects, the number of supernovae is related to the number of pulsars (or candidates, see Cordes & Lazio 1997) by

where fv is the fraction of pulsars with velocities low enough to remain bound to the GC, f~ is the fraction luminous enough to be detected, and f b is the fraction beamed toward Earth. Cordes & Lazio (1997) estimated f v f L f b N 1 0 - ~ - 1 0 - ~ . Using the number of our pulsar candidates, we can constrain N S N .We take Np,, N 10. Clearly the number of pulsars could be smaller; the number could also be larger, though not by more than a factor of a few. We estimate NSN5 104-105 within the past few million years (approximate pulsar lifetime), equivalent to a rate of order 1 0 - 2 4 . 1 yr-’. This estimate is consistent with the number required to produce the X-ray emitting plasma ( w lo3) and number required to produce the 1.8 MeV emission from 26A1(- lo5). Our upper limit is well above the supernova rate (- 10W4 yr-l) estimated by Figer (2003) on the basis of observations of massive stars in the GC. Acknowledgements Radio astronomy research at the NRL is supported by the Office of Naval Research

References Becker, R. H., White, R. L., Helfand, D. J., & Zoonematkermani,S. 1994, ApJS, 9 I , 347 Bower, G. C., Backer, D. C., & Sramek, R. A. 2001, ApJ, 558, 127 Cordes, J. M. & Lazio, T. J. W. 1997, ApJ, 475,557 Egan, M. P., Price, S. D., Moshir, M. M., et al. 1999, “The Midcourse Space Experiment Point Source Catalog, Version 1.2 (June 1999)” Air Force Research Lab. Technical Rep. Am-VS-TR-1999-1522 Figer, D. F. 2003, these proceedings Frail, D. A., Diamond, P. J., Cordes, J. M., & van Langevelde, H. J. 1994, ApJ, 427, L43 Lazio, T. J. W. & Cordes, J. M. 2003, ApJS, in preparation Lazio, T. J. W., Anantharamaiah,K. R., Goss, W. M., Kassim, N. E., & Cordes, J. M. 1999, ApJ, 515, 196 Lazio, T. J. W. & Cordes, J. M. 1998, ApJ, 505, 715 Lorimer, D. R., Yates, J. A,, Lyne, A. G., & Gould, D. M. 1995, MNRAS, 273,411 Maron, O., Kijak, J., Krarner, M., & Wielebinsk, R. 2000, A&AS 147, 195 Nord, M. E., Lazio, T. J. W., Kassim, N. E., Hyman, S., LaRosa, T., & Duric, N. 2003, AJ, in preparation Tieftrunk, A .R., Gaume, R. A,, Claussen, M. J., Wilson, T. L., &Johnston, K. J. 1997 A&A, 318, 931 van Langevelde, H. J., Frail, D. A,, Cordes, J. M., & Diamond, P. J. 1992, ApJ, 396,686 Wang, Q. D., Gotthelf, E. V., & Lang, C. C. 2002, Nature, 415, 168 Yusef-Zadeh, F., Cotton, W., Wardle, M., Melia, F., & Roberts, D. A. 1994, ApJ, 434, L63 Zepka, A., Cordes, J. M., Wasserman, I., & Lundgren, S. C. 1996, ApJ, 456,305

Astron. Nachr./AN 324, No. SI, 337 -341 (2003) / DO1 10.1002/asna.200385069

Review of low-mass X-ray binaries near the Galactic center A. Lutovinov", S. Grebenev I , S. Molkov', and R. Sunyaev'.'

' Space Research Institute, 1 17997 Moscow, Profsoyuznaya str. 84/32,Russia ' Max-Plank Institute for Astrophysics,Garching, Germany Key words X-ray astronomy, Galactic center, neutron stars

Abstract. We briefly review the results of observations of several low mass X-ray binaries (LMXBs) in the Galactic center region carried out with the ART-P telescope on board GRANAT observatory. More than a dozen sources were revealed in this region during five series of observations which were performed with the ART-P telescope in 1990-1992. The investigation into the spectral evolution of persistent emission seen in two X-ray bursters, GX3+1 and KS1731-260, a discussion of QPO and spectral variations detected from the very bright 2-source (3x5-1, and studies of the pulse profile changes of the pulsar GXI +4 are presented

here.

1 Introduction The Galactic Center is one of the most interesting and intensively observed X-ray regions, populated by many sources, including dozens of low-mass X-ray binaries (LMXBs). The ability of the telescope ART-P on board the GRANAT observatory, to simultaneously investigate X-ray emission from several point like sources located within its field of view, predetermined the choice of the Galactic center as a favorable target for observations. The total ART-PIGRANAT exposure time of the Galactic center field observations was -830 ks. Such a long exposure allowed us to build a detailed X-ray map of the region, in order to investigate persistent and transient source emission, leading to the discovery of four new X-ray sources (Pavlinsky et al. 1994). Several years later, this region was observed by the telescopes of the BeppoSAX and RXTE observatories (Sidoly et al. 1999, in't Zand 2001) in approximately the same energy band. Now the INTEGRAL observatory has begun to intensively investigate the Galactic center in gamma rays. In this paper we briefly present results of observations of two X-ray bursters GX3+1 and KS1731-260 carried out with the ART-PIGRANAT telescope in 1990-1992. These objects are regular sources of type I bursts, which are thought to result from thermonuclear flares on the surface of accreting neutron stars. We also present the preliminary results of spectral and timing investigations of the very bright low-mass X-ray binary GX5- 1 and the X-ray pulsar GX 1+4.

2 Results The image of the Galactic center, field reconstructed from the combination of the ART-P observations carried out in the fall of 1990, is presented in Fig.1. Fifteen point sources were revealed; most of them have already been presented in several ART-PIGRANAT papers (see for example Pavlinsky et al. 1992, 1994, Grebenev et al. 1995,1996, Lutovinov et al. 2001, Molkov eta]. 2000,2001 and references therein). Below we focus our attention on the four least studied sources. * e-mail: [email protected],Phone: +7 095 333 2222, Fax: +7 095 333 5377 @ 2003 WUEY-VCH Verlag GrnbH & Co. KGaA. Wnnheirn

A. Lutovinov et al.: Review of LMXB near GC

338

Fig. 1 Galactic center image obtained in the 3-20 keV energy band with the telescope ART-P in 1990. Contours show the significance of source detection at the level 3,4,5,10 .... and more standard deviations.

2.1 GX3+1 This source was observed with ART-P four times during the GRANAT Galactic center field survey in the fall of 1990 with total exposure more than 60 ksec. The analysis of the data led to one interesting finding - a strong X-ray burst was detected from GX3+1 on 14 Oct. An investigation into the penitent source emission on this day shows that its luminosity was 30% less than the luminosity measured in the other erg s-l in the 3-20 keV energy band. Such behavior is in agreement with the results days and 5 x of Makishima et al. (1983). It was only the second case after the HAKUCHO observations, when GX3+1 was found in the bursting state. Now the number of X-ray bursts detected from GX3+1 has increased to 81; a part of these bursts were observed during a source in the high state (den Hartog et al. 2003). The spectrum of the source measured in the high state is shown in Fig.2. It can be well approximated by modelling comptonization of soft photons in the hot electron plasma. However, the obtained values of the model parameters indicate that the model may be not physically correct because it does not take into account effects of free-free absorption. Two more suitable models for description of emission from GX3+1 were discussed in Molkov et al. (1999). The power spectrum of GX3+1 is presented in Fig.3 and doesn’t demonstrate any peculiarities.

-

N

2.2

KS1731-260

The transient X-ray source KS 1731-260was discovered in 1989 with the telescope TTM on board ROENTGEN observatory and was recognized as a burster (Sunyaev et al. 1990). The source was observed many 1.9 ms times with RXTE and BeppoSAX observatories and coherent X-ray oscillations with a period N

339

Astron. Nachr./AN 324, No. S1 (2003)

Fig. 2 Spectrum of GX3+1 measured by ART-P on Oct 11, 1990. Solid line is the best-fit comptonization model with temperature kT = 2.3 keV.

Fig. 3 Power spectrum (without dead-time correction) of the ART-P light curve of GX3+1 obtained on Oct 14, 1990.

were discovered during several bursts (Smith et al. 1997). In addition, a very powerful superburst with duration of several hours was detected from the source (Kuulkers et al. 2002). KS1731-260 was observed with the ART-P telescope four times over one year. Two X-ray bursts with approximately the same time profile were detected from the source (Fig.4). The photon spectra of the source obtained on 23 Aug and 7 Oct 1990 are shown in Fig.5. In both cases the spectra were equally well CRANAT/ART-P

10-1

7 Oct 1990

k'

KS1731-260 I

14h48m34s

40 -

3

-

' 1

U

80

40

-20

20

0

Time,

40

s

Fig. 4 Two X-ray bursts detected by ART-P from KS 1731-260. Gap in the light curve is related to the data transfer from the telescope buffer to main satellite memory.

Energy (keV)

Fig. 5 Spectra of source persistent emission obtained on 23 Aug (open circles) and 7 Oct (dark circles) 1990. The best-fit models are presented by corresponding lines.

A. Lutovinov et al.: Review of LMXB near GC

340

sept 8 , 1991 1

1

"

"

1

"

"

~

"

"

I

"

"

l

I

5 c

U 0

El

58

7.4

72 0

06

I

I6

Pulse phase

2

Pulse phase

a) b) Fig. 6 Observations of two different types of the sourse pulse profile (single asymmetric peak and double peak) and their dependence on the energy. described by the Sunyaev-Titarchuk comptonization or optically thin thermal bremsstrahlung models, but the source luminosity in the 4-20 keV energy band measured on 7 Oct (when the X-ray burst was detected) was 30% lower that ones in another day. Such correlation between source low state and bursting activity was observed for several other bursters. The detailed analysis of the persistent and burst emission of KS1731-260 is in progress. 2.3 GX1+4 The hard X-ray pulsar GX1+4, in a low-mass binary, was observed several times during the GRANAT observatory operations in 1990-1992 (Lutovinov et al. 1994). During the observation on 23 Aug 1990, the source was detected in the soft state where the observed photon spectrum was fitted well by a simple 2. The X-ray luminosity in the 3-20 keV range was equal to powerlaw with a photon index a! 3.6 x erg s-' assuming a distance to the GC of 8.5 kpc. X-ray pulsations from GX1+4, with period of 114.66 f 0.01 s, were detected with a double peak structure i n different energy bands on Oct 7, 1990 (Fig.6a). During observations on Sept 7, 1991, the pulse profile of GX1+4 showed a single broad feature with a period 116.14 f 0.13 s (Fig.6b). This change in the pulse profile from a double to a single pulse structure in about one year indicates either activation of the opposite pole of the neutron star, if the magnetic field is asymmetric, or possibly a change in the beam pattern from a pencil beam to a fan beam.

-

N

2.4 GX5-1 GX5-1 is a second brightest source in X-rays after Sco X-1 . It is called Z-source after the approximate " Z shape describe in an X-ray colour-colour diagram and in an X-ray hardness-intensity diagram (Hasinger, van der Klis 1989). GX5-1 was observed by ART-P many times in 1990-1992 with an intensity in the 3-20 1 Crab. To investigate the source spectral variations we applied to the spectra a keV energy band of few simple models, such as Sunayev-Titarchuk Comptonization of soft photons in a hot electron plasma, N

Astron. Nachr./AN 324, No. S 1 (2003) I

'

'

"

"

'

341 I

I

I

'

' ''''''I

'

"""(

I

'

"""(

'

GX5-'

01

I

10

Frequency. Hz

Energy, keV

Fig. 7 The GX5-1 spectra variations as observed with ART-Pin Sept-Oct 1990. Lines are best-fit models for three different sessions of observations, points are the source spectrum obtained during one of them.

1

"'I

Fig. 8 The power spectrum (without the dead-time correction) of the ART-P light curve of GX5-1 obtained on Oct 20, 1990. The broad QPO feature in the range 13 - 20 Hz is obviously visible.

optically thin thermal bremsstrahlung and the Boltzmann law. Figure 7 illustrates the results, showing the spectra measured o n the autumn of 1990. We also estimated the power spectra of the source by calculating the Fourier amplitudes of the signal for the all sessions of observations. T h e QPO with centroid frequency 16 H z was detected on 20 Oct 1990 (Fig.8). T h e measured frequency correspond to the typical 'horizontal branch' QPO (van d e r Klis 1985). N

Acknowledgements A.L. thanks to the Organizing Committee and the Space Research Institute of the Russian Academy of Sciences for the providing a financial support of the participation in the Conference.

References Grebenev, S., Pavlinsky, M., and Sunyaev, R. 1995, Adv. Space Res., 15, (5)115. Grebenev, S., Pavlinsky, M., and Sunyaev, R. 1996, in: Proceedings of the 2th INTEGRAL Workshop, ESA SP382, p.183. den Hartog, P., in 't Zand, J., Kuulkers, E., Cornelisse, R., Heisse, J., Bazzano, A,, et al. 2003, A&A, 400,633-641. Hasinger, G., and van der Klis, M. 1989, A&A, 225, 79. in't Zand, J. 2001 in: Proceedings of the 4th INTEGRAL Workshop, ESA SP-459, p.463. Kuulkers, E., in 't Zand, J., van Kerkwijk, M., Cornelisse, R., Smith, D., Heisse, J., et al. 2002, A&A, 382, 503. Lutovinov, A., Grebenev, S., Sunyaev, R., and Pavlinsky, M. 1994, Astron. Lett., 20,538. Lutovinov, A., Grebenev, S., Pavlinsky, M., and Sunyaev, R. 2001, Astron. Lett., 27, 501. Makishima, K., Mitsuda, K., Inoue, H., Koyama, K., Matsuoka, M., Murakami, T., et al. 1983, ApJ, 267, 310. Molkov, S., Grebenev, S., Pavlinsky, M., and Sunyaev, R. 1999, Astron. Lett. & Comm., 38, 141. Molkov, S., Grebenev, S., and Lutovinov, A. 2000, A&A, 357, L41. Molkov, S., Grebenev, S., and Lutovinov, A. 2001, Astron. Lett., 27, 363. Pavlinsky, M., Grebenev, S., and Sunyaev, R. 1992, Sov. Astron. Lett., 18, 88. Pavlinsky, M., Grebenev, S., and Sunyaev, R. 1994, ApJ, 425, 110. Pavlinsky, M., Grebenev, S., Lutovinov, A., Sunyaev, R., and Finoguenov, A. 2001, Astron. Lett., 27,297. Sidoli, L., Mereghetti, S., Israel, G., Chiappetti, L., Treves, A,, and Orlandini, M. 1999, ApJ, 525,215. Smith, D., Morgan, E., andBradt, H. 1997, ApJ, 479, L137. Sunyaev, R., Gilfanov, M., Churazov, E., Lomikov, V., Yamburenko, N., Skinner, G., et al. 1990, Sov. Astron. Lett., 16, 59. van der Klis, M., Jansen, F., van Parddijs, J., Lewin, W., van den Heuvel, E., Trumper, J., et al., 1985, Nature, 316, 225.

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1, 343-347 (2003)/ DO1 10.1002/asna.200385052

Neutrons, neutrinos, and gamma-raysfrom the Galactic Center

'

W. Bednarek' ' Department of Experimental Physics, University of todi, Poland

Key words The Galactic Center, supernovae, supernovae remnants, pulsars, cosmic ray origin, cosmic ray sources, y-ray sources.

Abstract. We calculate neutron, neutrino and y-ray fluxes produced by nuclei which are accelerated by a very young pulsar in the Galactic Center region. It is proposed that the neutrons are responsible for the observed excess of cosmic rays at 10'' eV, reported by the AGASA and SUGAR groups from the direction of the Galactic Center. From normalization of the calculated neutron flux to the one observed in the cosmic ray excess, we predict the neutrino and gamma-ray fluxes at Earth and discuss their observability by the I !an2neutrino detector of the IceCube type and future systems of Cherenkov telescopes.

-

1 Introduction Recently, the existence of an extended excesses of cosmic rays (CRs) over a narrow energy range eV, from directions close to the Galactic Center (GC) and the Cygnus region have been reported by the AGASA and SUGAR experiments (Hayashida et al. 1999, Bellido et al. 2001). We propose (Bednarek 2002) that neutrons from the disintegration of iron nuclei, which are accelerated by an energetic pulsar in the GC region, could be responsible for the observed excess of the cosmic rays with energies lot8 eV. 50 pc of the Galaxy is rich in several massive stellar clusters with a few to more than 100 The inner OB stars (e.g. Morris & Serabyn 1996, Blum et al. 2001). These stars should soon explode as supernovae. In fact recent multiple supernova explosions in the GC region (lo3 supernovae in the past lo5 years) are suggested by the observations of the diffuse hot plasma emitting X-rays (Yamauchi et al. 1990). Since it is expected that pulsars are formed in explosions of such massive stars, we can expect that the GC should contain some young pulsars, a number of them being 10' - lo3 yrs old. Motivated by these observational results we may assume that at least one of these young pulsars, formed in a supernova explosion of the type Ib/c, has parameters which allow the acceleration of iron nuclei to energies 1020 eV, and production of 10'' eV neutrons from their disintegration.

-

-

N

2 A pulsar inside a molecular cloud We investigate the scenario in which a very young pulsar is formed in a core collapse of the type Ib/c supernova immersed within a huge molecular cloud (or high density medium), characteristic of the GC. The particles accelerated by the pulsar can escape into the surrounding and diffuse in the magnetic field of the cloud, suffering collisions with the matter from time to time. In our further considerations, we discuss examples for two different media, typical of the GC region in which the pulsar may be immersed. The first is a huge molecular cloud with a radius R, = 10 pc, density n, = lo3 cmP3, magnetic field B, = lop4 G, and a total mass 105M0. The second is an extended high density region inside the GC with R, = 50 pc, n, = 10' cmp", B, = 3 x lop5 G, and a total mass 1O6Ma. N

-

* Corresponding author: e-mail: [email protected]

@ 2003 WILEY-VCH Verlag GmhH B Co. KGaA. Weinhem

344

W. Bednarek Neutrons, neutrinos and gamma-rays from the Galactic Center

Following the recent works of Blasi, Epstein & Olinto (2000) and Beall & Bednarek (2002), we assume that pulsars can accelerate iron nuclei in its wind zone (the mechanism called magnetic slingshot, Gunn & Ostriker 1969). The maximum energies which the nuclei can reach in this model are determined by the magnetic field energy per particle in the pulsar wind zone, and depend on the pulsar parameters as;

where P = 10V3Pm,s is the pulsar period, B = 101’B12 G is the pulsar’s surface magnetic field, 7 1 =~ B ( r ~ ~ ) / ( 2 e c=P3.3 ) x 1011B12P;: cm3 is the Goldreich & Julian (1969) density, TLC = c P / 2 ~= 4.77 x 106Pm, cm, and c is the speed of light. According to the slingshot mechanism, the acceleration of nuclei occurs very fast so that they do not lose energy during this stage. When the envelope becomes transparent, the iron nuclei can be injected into the surrounding of the parent pulsar’s supernova. We obtain the spectrum of iron nuclei injected by the pulsar following the derivation by Beall & Bednarek (2002). That calculation is based on the general prescription given by Blasi, Epstein & Olinto (2000),in which the number of nuclei accelerated to energies ( E )scale as a part (E) of the Goldreich & Julian (1969) density at the light cylinder radius. Beall & Bednarek (2002) modify this approach by noting that the spectrum injected by the pulsar at a fixed time ‘t’ should not be monoenergetic due to the fact that the magnetic energy density, responsible for the acceleration of particles, changes along the pulsar’s light cylinder height. As a result, a pulsar with specific parameters injects nuclei with a spectrum which can be below E F ~and , described by a single power law with the spectral index -1/3. We estimated that these nuclei can escape through the supernova envelope after N 1 yr after the supernova explosion for typical parameters of the supernova, i.e the mass of the envelope in the case of type Ib/c supernovae Men, = 3 M a , and the expansion velocity of the envelope at the inner radius is v,,, = 3 x los cm s-’. The iron nuclei diffuse in the magnetic field of the high density medium in the GC region, i.e. huge molecular clouds. Some of them interact producing neutrons, neutrinos, and y-rays. In order to obtain the equilibrium spectrum of iron nuclei inside the cloud, we have to integrate over the activity time of the pulsar since its parameters evolve in time due to the pulsar’s energy losses. If we assume that the pulsar loses energy only on electromagnetic waves, then its period changes according to the formula P&(tobs) = 1.04 x 10r9t0b,B& Pi,,,,, where Po,msPms(tobs) are the initial and present periods of the pulsar. The equilibrium spectrum of iron nuclei at a specific observation time, to&, is calculated from

+

where t o = 1yr, K gives the part of nuclei produced at the time ’t’ which do not escape from the cloud due to the diffusion and are still present inside the cloud at the time tot,,. X is the mean free path for collision of the iron nuclei with the matter of the cloud. Note that for the parameters considered in this paper the value c(tobs- t)/X, in Eq. 2, is always less than one. Therefore, a significant portion of iron nuclei with energies 10” eV can escape without interaction as postulated by the model of Blasi, Epstein & Olinto (2000). The value of K is estimated from K = (R,/Dd,f)3, where Ddif = (r&/3)lI2 is the diffusion distance of iron nuclei in the magnetic field of the cloud, and T L is the Lannor radius of the iron nuclei with energy E . For the case, Dd,f 5 R,, we take K = 1. The portion of iron nuclei confined within the molecular cloud, interact with a relatively dense medium suffering disintegrations and pion energy losses. The pions decay into neutrinos and y-rays. Applying the equilibrium spectrum of iron nuclei, we calculate the differential spectra of neutrons (from disintegrations of the iron nuclei), muon neutrinos, and y-rays (from inelastic collisions of iron) for the pulsar with the surface magnetic field B12 = 6 (typical of the observed radio pulsars), and the initial period P,, = 2. Such a pulsar is able to accelerate nuclei to such high energies (see Eq. 11, that neutrons from their disintegration fulfill the observational constraint required by the excess of cosmic ray particles at lo7 - los GeV

-

~

345

Astron. Nachr./AN 324, No. SI (2003)

-

h

-3

I

01

-4

E

-5

5 CJ

-6

\

-7

-8 -9

- 10 3

4

5

6 1%

7

8

9

1

0

(E / cev)

Fig. 1 The differential spectra of neutrons (dashed curves) and y-rays (full curves) observed from the Galactic Center on Earth at the time 10, lo2,lo3,3 x lo3, and lo4 yrs (from the thinnest to the thickest curve) after the formation of the pulsar. The thick dotted curve shows schematically the observed cosmic ray spectrum within the 20" circle.

(Hayashida et al. 1999, Bellido et al. 2001). The pulsar born with such parameters slows down as a result of the dipole energy losses to 11 ms after 10' yr, 34 ms (lo3 yr), 60 ms (3 x lo3 yr), and to 106 ms (lo4 yr). So then it is still fast enough to produce pulsed y-rays (Crab and Vela type pulsars).

3 Neutrons, neutrinos, and gamma-rays on Earth The large distance to the GC, DGCN 8.5 kpc, influences the expected fluxes of neutrons and y-rays measured on Earth, due to the decay of unstable neutrons and the absorption of y-rays in the microwave background radiation (MBR). The mean free path for neutrons depends only on their energy, AN = C T N ~ N= S.SEN,+V kpc, where a neutron's lifetime is TN = 918 s, its energy EN = 9.38 x 10W1Oy~in EeV, and YN is the Lorentz factor. The mean free path for the absorption of photons in MBR, A,, was computed in several papers just after the discovery of MBR (e.g. Gould & Schreder 1966). A,, becomes comparable to the distance of the GC at energies of N 1 PeV. Therefore the fluxes of neutrons and y-rays observed on where A y y ; ~= A, or AN respectively. In Fig. 1, we Earth are reduced by the factor e X p ( - D c c / A , , ; N ) , show the fluxes of neutrons and y-rays observed on Earth at different times after the pulsar formation for the case of the pulsar with the previously mentioned initial parameters and two different sets of parameters describing the medium in the GC region mentioned in section 2: a molecular cloud (the left figure), and a a more extended and lower density region around the GC (the right figure). Since the propagation of neutrinos is not influenced by any process, their fluxes can be simply obtained for the known distance to the GC. In Fig. 1 we mark also the observed spectrum of cosmic rays (CR) below 3 x 1018 eV within the 20" circle (the analysis box of the AGASA data). We find that only the pulsars born within the last 3 x lo3 yrs are able to provide fluxes of neutrons which exceed the CR limit, provided that they accelerate nuclei with the efficiency [ = 1. N

4 Discussion and Conclusion The SUGAR group estimates the flux of particles which causes reported excess of the cosmic ray particles - 1018.5eV on (9 f 3) x lop1* mp3 s-l (Bellido et al. 2001). If this excess in the energy range is caused by neutrons produced in the pulsar model discussed here, then the expected flux of neutrons can be compared with the observed one. Based on this normalization, we predict the fluxes of neutrinos and gamma-rays on Earth. This procedure allows us to derive the free parameter of our model (i.e. the

W. Bednarek: Neutrons, neutrinos and gamma-rays from the Galactic Center

346

Table 1 Gamma-rays and neutrinos from the Galactic Center

Model (11

I

I

N,(> 1 TeV) 4.3 x

2.5 x 106.7 x 10-

I N,(> I

10 TeV) 2.2 x 10-12

I

I

N,” I N,”” 23 1 30

1.7 x 10-

<

efficiency of iron acceleration by the pulsar), and limit the age of the pulsar for other fixed parameters, P, B, R,, nc3B,, which are in fact constrained by the observations. We consider five different sets of parameters describing our scenario: model (I) R = 10 pc, n = lo3 cmP3, B, = l o r 4 G, tabs = lo4 yr; (11) tabs = 3 x lo3 yr and other parameters as above; (111) tabs = lo3 yr and other parameters as above; (IV) R = 50 pc, n = 10’ ~ m - B ~,, = 3 x lop5 G, tabs = lo4 yr; and (v) tabs = 3 x lo3 yr and other parameters as in (IV). They all concern two sets of parameters for the medium in which the pulsar is formed, and differ in the pulsar’s age which is not constrained by any observations. In all these models we assume that the pulsar is born with B = 6 x lo1’ G and Po = 2 ms. Normalizing the predicted neutron flux to the observed excess of CR particles we derive the value of the parameter which has to be E x 1 (model I), 0.18 (11), 0.03 (In),0.3 (IV), and 0.09 (V).

<

<

Using the above estimates for we can now predict the expected fluxes of y-rays and muon neutrinos and antineutrinos in the case of every model. The integral spectra of y-rays from the GC region are presented in Fig. 2, together with the sensitivities of the present HEGRA telescope system and the planned next generation of telescopes, i.e. CANGAROO ILI, HESS, VERITAS. In Table 4 we report the y-ray fluxes above 1 TeV and 10 TeV in units cmP2 5 - l . Although the y-ray spectra have a maximum above 10 TeV for all models, the y-ray fluxes in the energy range 1-10 TeV produced in models, (I) 2.1 x lo-’’ cm-2 s-1 , and (11) 2.1 x cmP2 s-’, and probably also in (IV) 8 x cmP2 s-’, should be observed by the future systems of Cherenkov telescopes of the CANGAROO 111, HESS, and VERITAS type. Models (111) and (V) predict fluxes below the sensitivity limit of these Observatories. Only the HEGRA Collaboration observed the Galactic disk including the GC region (Pohlhofer et al. 1999). The upper limit on the possible sources, equals 1/4 of the Crab pulsar in the Galactic plane, which is above the y-ray flux predicted even by the model (I). However, since the GC region can be observed by this experiment only at zenith angles larger than 60°, this limit does not refer to the GC region. N

N

N

In Fig. 3 we show the muon neutrino and anti-neutrino spectra expected in the discussed model. At energies > 10 TeV, these spectra are above the expected flux of atmospheric neutrino background (ANE3) and also above the 3 yr sensitivity limit of the planned large size neutrino detector IceCube (Hill 2001). We estimate the number of muon neutrino events during one year in the IceCube detector basing on the calculations of the likelihood of detecting such neutrinos by a detector with a surface area of 1 km2 obtained by Gaisser & Grillo (1987). The results of these calculations are shown in Table 1. We distinguish the case of neutrinos coming to the neutrino detector from directions close to the horizon, i.e. not absorbed by the Earth (Ny”a),and neutrinos which arrive moving upward from the nadir direction, i.e. partially absorbed by the Earth (N,”)(for absorption coefficients see Gandhi 2000). From Table 1 it is clear that the IceCube detector should detect a few up to several neutrinos per year from the Galactic Center region provided that the excess of cosmic rays at N 101n eV from the GC region is caused by neutrons from disintegrations of iron nuclei, accelerated by a very fast pulsar. The detection of the predicted fluxes of neutrinos from the Galactic Center (or lack) will also place constraints on the recent model of extremely high energy cosmic ray production in the pulsar scenario (Blasi, Epstein & Olinto 2000), since the parent iron nuclei which inject neutrons with energies l(1’’ eV, have to be accelerated to energies N lo2’ eV. N

347

Astron. Nachr./AN 324. No. S I (20031

-71,' -"

-

l

O

'

'

'

'

'

'

'

'

'

'

'

'

'

'

'

'

'

'

1

i

h

W v

2

v

-

M -1

Fig. 2 The integral spectra of y-rays produced in hadronic interactions of iron nuclei with the matter of the molecular cloud at the Galactic Center for the models considered in the text: I (full curve), Il (dotted), III (dot-dashed), and IV (dot-dot-dot-dashed), and V (dashed). The thick-dashed and full curves show the sensitivity of the operating HEGRA System and planned CANGAROO 111, HESS, and VERITAS systems for a 50-hour exposure to a single source (Lorenz 2001, and references therein).

Fig. 3 The differential spectra of muon neutrinos and antineutrinos produced in hadronic interactions of iron nuclei with the matter of a molecular cloud at the Galactic Center. The dashed curves indicate the atmospheric neutrino background, ANB, (Lipari 1993) within a la of the source and the dotted line shows the 3 yr sensitivity of the IceCube detector (Hill 2001).

Acknowledgements This work is supported by the grants: KBN No. 5P03D02521 and the University of Lbdi. References Beall, J.H., Bednarek, W. 2002, ApJ, 569,343 Bednarek, W. 2002, MNRAS, 331,483 Bellido, J.A. et al., 2001, Astropart.Phys., 15, 167 Blasi, P., Epstein, R.I., & Olinto, A.V. 2000, ApJ, 533, L123 Blum, R.D. et al. 2002, in Proc. Hot Star Workshop III: The Earliest Phases of massive Star Birth, ed. P.A. Crowther (Boulder, 2001), p. 283 Gaisser, T.K.,Grillo, A.F. 1987, PhwRev. D, 39, 1481 Gandhi, R. 2000, Nucl.Phys.Suppl.; 91,453 Goldreich, P., & Julian, W.H. 1969, ApJ, 157, 869 Gould, R.J., Schreder, G. 1966, PRL, 16,252 Gunn, J., Ostriker, J. 1969, PRL, 22, 728 Hayashida, N. et al. 1999, Astropart.Phys., 10, 303 Hill, G.C. 2001, Proc. XXXVI Rencontres de Moriond, astro-ph/0106064 Lipari, P. 1993, Astropart.Phys., 1, 195 Morris, M., Serabyn, E. 1996, ARA&A, 34,645 Piihlhofer, G. et al. 1999, Proc. 26th ICRC (Salt Lake City), OG 2.4.1 1 Yamauchi, S. et al. 1990, ApJ, 365,532

This Page Intentionally Left Blank

Astron. NachrJAN 324, No. S1,349 -354 (2003) / DO1 10.1002/asna.200385040

Linear and Circular Polarization from Sagittarius A* G.C. Bower”’

’ Radio Astronomy Lab, UC Berkeley, Berkeley CA 94720

We describe recent results for Sgr A*, M81* and other low luminosity active galactic nuclei. We have conducted linear and circular polarimetry over a frequency range of 1.4 to 230 GHz and detected a variety of phenomena. The polarization properties of the studied sources are substantially different from higher powered AGN. In the case of Sgr A*, we are able to eliminate ADAFs and Bondi-Hoyle flows as possible models based on mmX polarimetry. The detection of linear polarization at 230 GHz with the BIMA array indicates that the accretion rate onto the black hole is < l o p 7 M a ypl. consistent with jet and CDAF models. Our snccess with Sgr A* demonstrates that we can learn about the nature of accretion and outflow in these sources with unprecedented detail. Ultimately, we may develop probes of general relativity in the strong-field limit through polarization imaging.

1 Sagittarius A* We have engaged in a lengthy program exploring the linear and circular polarization properties of Sagittarius A* (Bower et al. 1999a, Bower, Falcke & Backer 1999b, Bower et al. 1999c, Bower et al. 2001, Bower et al. 2002a, Bower et al. 2003). We summarize some of the key results of that program here.

1.1 Circular

<< Linear at cm X

VLA observations between 1.4 and 15 GHz show that circular polarization dominates linear polarization at these frequencies. Linear polarization is not detected to instrumental limits of 0.1% while circular polarization is detected at levels as high as 1%. This is substantially different from what is seen in high luminosity AGN, where linear always dominates circular polarization (Rayner et al. 2000). This property has spawned interest in models that destroy linear polarization and convert linear polarization into circular polarization through Faraday effects (Beckert et al. 2003, Ruszkowski & Begelman 2002). These models give us a handle on the low energy end of the electron spectrum as well as the magnetic field geometry. The fractional circular polarization spectrum is highly variable. In a flaring state, the spectral index is inverted with a maximum value of +1 between 8.4and 15 GHz. This suggests that circular polarization could be detected at very high frequencies but this has not been seen yet. 1.2 Sign of Circular is Constant over 20 years

We have probed the variability of circular polarization on timescales ranging from hours to decades. In one epoch, we see that the circular polarization varied by N 100% in two hours while the total intensity changed by 25%. As noted above, these variations are strongest at the highest observing frequency. This timescale is comparable to the timescale of X-ray flares in Sgr A* (Baganoff et al. 2003). Archival VLA observations allow us to show that the circular polarization at 4.8 GHz did not change significantly over 20 years. In these observations and in shorter term observations, the sign of circular

-

* Corresponding author: e-mail: gbowerQastro.berkeley.edu,Phone: +01 5106424075, Fax: 4 1 5106424075

@ 2001 WILEY-VCH Verlag CmbH & C o KGaA. Weinhem

G . C. Bower: Polarization of S g r A*

350

Polarization Fraction for Sgr A* T

I

25

I I I

I

I I I

-20

8

v

I I 1

S Z

.-0 c

0

LL C .-0

15

c

a

:. 10

-a0 Q

5

-

----- -

-

-

-,

--

-

_ _ - - A

' -- , x_=:.

10 Frequency (GHz)

,v

' / I

/2,/

'X?, . . * r

100

Fig. 1 Linear polarization of Sgr A* as a function of frequency. Triangles are the JCMT (Aitken et al. 2000) data. Squares are the new BIMA data. The inverted triangles are the 3 mm BIMA upper limits. Crosses are VLA upper limits. All error bars and upper limits are plotted at 2u levels. The dashed lines are power law models for the polarization fraction with indices of I .O, 2.0,2.5 and 4.

polarization remains negative. We have no detection at any time of a positive circular polarization from Sgr A*, although the signal sometimes approaches zero. This suggests that the circular polarization originates in a region with a fixed magnetic field orientation. That this field orientation is associated with a flaring component is striking and indicates that it may not have the same character as a shocked region in a higher luminosity jet.

1.3

Linear

> > Circular at mm X

While there is no circular polarization apparent at millimeter wavelengths, we have recently detected linear polarization with the BIMA array at a frequency of 230 GHz (Figure 1). This confirms the JCMT detection at millimeter and submillimeter wavelengths (Aitken et al. 2000). The BIMA observations have an arcsecond beam size, smaller than the JCMT beam by a factor > 100. This allows us to exclude the effects of polarized dust and unpolarized free-free emission with a great degree of certainty. The constant position angle in the upper and lower sideband of the BIMA observations at 230 GHz places a strong upper limit to the rotation measure (RM) of the accretion environment of 2 x lo6 rad m-'. An RM this small excludes a number of models which require large mass accretion rates onto the black Ma y-'. Thus, the low hole. These include ADAF and Bondi-Hoyle models which require luminosity of Sgr A* is due to a low accretion rate rather than to a radiatively inefficient accretion flow.

-

Astron. Nachr./AN 324, No. S 1 (2003)

35 1

Rotation Measure for Sgr A* I

I

I

I

m

40'

' 150

I I

200

I

I

250 300 Frequency (GHz)

350

400

Fig. 2 Position angle as a function of frequency. Triangles are the A00 data. Squares are the BIMA data. The solid line is a fit for the RM excluding the A00 230 GHz result. The best fit is -4.3 f 0.1 x lo5 rad m-2 with a zero-wavelength position angle of 181 & 2 degrees.

-

We have a tentative result of RM 4 x lo5 rad m-2 based on the BIMA and JCMT measurements. However, these measurements are not fully consistent and re-observation is necessary. An actual measurement of the RM will permit us to exclude even lower accretion rate models. We will also be able to probe changes in the accretion environment as a function of time. 10% at 230 The linear polarization spectrum shows a sharp transition from < 1%at 100 GHz to GHz. Bandwidth depolarization is insufficient to account for this transition while beam depolarization is marginally adequate. Beam depolarization requires a fully turbulent medium with a very small outer scale of turbulence. An alternative scenario involves an unpolarized low frequency component and a highly polarized high frequency component with an inverted spectrum. This is consistent with models that characterize the total intensity spectrum as composed of two separate components. The detection of linear polarization in Sgr A* opens up opportunities to study the immediate environment of a black hole in substantial detail. A number of papers have proposed that general relativistic effects may be detected (Broderick & Blandford, these proceedings; Falcke, Melia & Ago1 2000; Bromley & Melia 2001).

-

2 LLAGN The polarization properties of Sgr A* are distinct from those of high powered AGN. In particular, linear polarization dominates circular polarization by typically an order of magnitude in these sources. These powerful AGN are more luminous than Sgr A* by 10 orders of magnitude. To determine whether the

352

G . C. Bower: Polarization of Sgr A*

,,,,............................

, , , I , , , , , , ,

a , , , , , ,

,,,,

+

,,,, ,,,,,,,,,,,,,,

-0.3

0

0.5

1 1.5 Linear Polarization (%)

2

Fig. 3 Linear and circular polarization of low luminosity active galactic nuclei. Dotted lines indicate systematic limits of 0.1%. Most of the LLAGN show no or very little linear or circular polarization.

difference in polarization is an effect of luminosity, we have studied the linear and circular polarization of a sample of nearby LLAGN. In Figure 3 we summarize our observations of 9 LLAGN with the VLA at 8.4 GHz (Bower, Falcke & Mellon 2002). With the exception of M87, all sources showed weak or no linear polarization. Linear

polarization was only detected in two sources at a level of a few tenths of a percent. In NGC 4579 there is a potential detection of RM= 7 x lo4 rad m-’. Circular polarization is only detected convincingly in one source M81*, which we discuss in the next section. The absence of linear polarization for these sources is different from the case of higher luminosity sources and is similar to the case of Sgr A*. The result can be readily explained through Faraday depolarization. An RM- lo5 rad m-’ can produce bandwidth depolarization in these sources at these frequencies. An RM- lo3 rad m-’ can produce beam depolarization. We have demonstrated that both the accretion region and galactic environment of Sgr A* can readily lead to RMs this large. Nevertheless, one cannot exclude the effects of low jet power. If low luminosity jets do not exhibit the same degree of field order as their higher-powered cousins, then they will be weakly polarized. They may lack the powerful shocks that order the magnetic field in high luminosity sources. We note also that there is a clear trend in spectral index and circular polarization strength. Of objects that we have studied, only Sgr A* and M8 1* show circular polarization and only Sgr A* and M8 1 * have inverted spectral indices. This supports the hypothesis that circular polarization is due to an opacity effect in the field. However, only a small number of sources are included in this analysis.

Astron. Nachr./AN 324, No. S1 12003)

-

c

.

m 0

.

L

MRI’

:ed triangles

J1044+719

blue squares

J1053+704

black crosses

-

._ ._

w

353

a

,_,.

L

m

E 1

-r, -___ /.+---

,..,

0.1 : -

-

..__ ---.__ ,

--.__

~~

_/”

__---_-__----

:i

I

i

I

I

0.01

’--,$I

I

I

I

I

I

I

I

Fig. 4 Linear polarization of M8 I * and two calibrators as a function of frequency. M8 1* is unpolarized at all frequencies below 22 GHz.

3 MU* We have explored the linear polarization properties of the LLAGN M8 I * in more depth (Brunthaler, Bower & Falcke 2001). These investigations have shown that the polarization continues to exhibit similarity with Sgr A* as we observe at higher frequencies and study the variability properties. The presence of a jet in M81* suggests that the polarization properties of both sources are dominated by a jet. In Figure 4 we show that linear polarization is absent up to a frequency of 22 GHz. This implies lower limits to the RM greater by a factor of N 7 over those for the 8.4 GHz survey. These RM limits are still not surprising given the expected particle densities and magnetic field strengths near the black hole. The circular polarization properties of M81* are also similar to that of Sgr A* (Figure 5). We see that circular polarization is detected at 4.8, 8.4 and 14.9 GHz with a magnitude as high as 1.5%. The degree of variability increases with frequency. The lightcurves suggest episodic activity that is also characteristic in Sgr A*. The highest point in the circular polarization light curve occurs 10 days after a bright fare in the total intensity. As the flare decays, the high frequency circular polarization disappears.

-

4

Summary

We have presented here a range of observations of polarization in LLAGN, including Sgr A* and M81*. These sources differ markedly from higher luminosity AGN. These differences are consistent with smallscale, low-power jets which see the high density accretion regions andor galactic HI1 regions. Higher frequency observations may reveal a marked transition in the polarization properties of LLAGN other than

G. C. Bower: Polarization of Sgr A*

354

Circular Polarizationin M81 (blue squares) and J1053+704 (black triangles)

_--60

80

100

120

140

160

I

I

180

ZOO

220

240

1.5 1

05 0

60

I

I

4

80

100

120

140 160 Time(days in 2001)

I

I

I

180

200

220

240

Fig. 5 Circular polarization of M81” and a calibrator as a function of time at three frequencies. The circular polarizaiton properties of M81* are similar to those of Sgr A*. Sgr A* as Faraday effects weaken. These observations are at the edge of capability for current millimeter interferometers. Future arrays such as CARMA and ALMA will be able to systematically survey a broad sample of LLAGN. We will be able to determine the nature of their accretion environments, including the role of advection, convection and outflows. We will b e able to explore the stability of magnetic field structures, the presence of black hole spin and general relativistic effects in the vicinity of the black hole. A similar version of this paper was also submitted to the Proceedings of the Amsterdam Workshop on

Circular Polarization, 2003, J.-I! Macquardt and R. Fendel; eds.

References Aitken, D. K., Greaves, J., Chrysostomou, A., Jenness, T., Holland, W., Hough, J. H., Pierce-Price, D., & Richer, J. 2000, ApJ, 534, L173 Bower, G. c.,Backer, D. C., Zhao, J. H., Goss, M., & Falcke, H, 199ga, A ~ J521, , 582 Baganoff, F., et al., 2003, these proceedings Beckert, T., et al., 2003, these proceedings Bower, G. C., Falcke, H., & Backer, D. C. 1999b. ApJL, 523, L29 Bower, G. C., Falcke, H., Sault, R. J., & Backer, D. C. 2002a, ApJ, 571,843 Bower, G. C., Wright, M. C. H., Backer, D. C., & Falcke, H. 1999c, ApJ, 527,851 Bower, G. C., Wright, M. C. H., Falcke, H., & Backer, D. C. 2001, ApJL, 555, L103 Bower, G. C., Wright, M. C. H., Falcke, H., & Backer, D. C. 2003, ApJ, in press Bromley, B. C., Melia, F., & Liu, S. 2001, ApJL, 555, L83 Brunthaler, A., Bower, G. C., Falcke, H., & Mellon, R. R. 2001, ApJL., 560, L123 Rayner, D.P., Nonis, R.P. & Sault, R.J., 2000, MNRAS, 319,484 Ruszkowski, M. & Begelman, M.C., 2002, ApJ, 573,485

Astron. Nachr./AN 324, No. S1.355-361 (20031 / DO1 10.1002/asna.200385096

Intrinsic Radio Variability of Sgr A* Jun-Hui Zhao*' I Harvard-Smithsonian Center for Astrophysics, 60 Garden St., MS 78, Cambridge, MA 02138 Key words Black hole, radio continuum, accretion disk.

Abstract. We review and summarize the results on the variability of Sgr A* based on the recent monitoring programs with the SMA at 1.3 and 0.87 nun, and the VLA at 2, 1.3 and 0.7 cm. We discuss the flares at 1.3 mm and a cross-correlation of the SMA flux density at 1.3 mm with the VLA data at 1.3 cm. We also present a preliminary result on the double quasi-periodic oscillation (DQPO) in flux density of Sgr A* based on our analysis of radio light curves observed with the VLA at 1.3 cm.

1 Introduction Variations in the radio flux density of Sgr A* have been known for more than two decades (Brown & Lo 1982). The nature of the radio variability in flux density appears to be far more complicated than we had thought. At long wavelengths, the flux density of Sgr A* might be modulated by scintillation due to the turbulence in the ISM (Zhao et al. 1989). The early VLA monitoring campaign at 20, 6, 3.6, 2 and 1.3 cm during the period 1990-1993 shows that the fractional amplitude variations increased towards short wavelengths and that the rate of radio flares appeared to he about three per year, suggesting that the intrinsic radio outburst occurs in Sgr A* (Zhao et al., 1992; and Zhao & Goss 1993). The typical time scale of these radio flares is about a month. The observed large amplitude variations in flux densities at 3 mm (Wright & Backer 1993; Tsuboi, Miyazaki & Tsutsumi 1999) are consistent with the wavelength-dependence of the variability as observed at centimeter wavelengths. The monitoring observations at 1.3 mm with the partially completed SMA (Moran 1998) also suggest large amplitude flares. The presence of a 106f10day cycle in the radio variability of Sgr A* was suggested from an analysis of data observed with the VLA in the period of 1977-1999 (Zhao, Bower and Goss 2001). Similar Huctuation in flux density was also observed in the Green Bank Interferometer monitoring data (Falcke, 1999). In this paper, we summarize the result from recent 25 epochs of observations with the partially completed SMA and discuss the preliminary results from a power spectrum density (PSD) analysis of the radio light curve observed with the VLA at 1.3 cm.

2 Flares at 1 mm Observarions of Sgr A* at I .3 mm and 0.87 mm were made using the partially completed SMA with three or four antennas and baselines ranging from 7 to 55 kilo wavelengths at 1.3 mm. A complex, large scale structure in the central region is automatically filtered away. For baselines 20 kX and longer, Sgr A* appears (0 he the dominant source and the confusing flux density from the surrounding medium appears to be less significant at these two short wavelengths. Fig. 1 shows the SMA light curve at 1.3 mm suggesting that Sgi- .A':: varies significantly. Three possible Hares were observed during the monitoring period of 15 I I N - ~ I ~{' ~. ~C. CFig. 1 and Zhao et a1 2003). The time scale of the variation of a month is consistent with

-

* Corresponding author: e-mail: jzhaoQcfa.harvard.edu,Phone: tl617 496 7895, Fax: +1617 496 7554

@ 2003 WILEY-VCH Verlag GmbH & C o KGaA, Weinhem

Jun-Hui Zhao: Radio Variability of Sgr A*

356

t " " ' " " " " " ~ " " ~ " " ~ ' " ' " " ' " ' " " " ' " ' ' " ' ~

*

1;

I -

SMA Light Curve of SgrA* a t 1 3 m m

4

"

?

Flare 3

I

e I

1

*

-~ I

,

,

I

I

I

,

I

I

,

I

I

,

,

l

/

,

.

,

,

,

l

.

,

,

,

,

,

; 1

I

700

750

date number 0 = (2000-06-21)

Fig. 1 The SMA light curve of Sgr A* observed at 1.3 mm in the period between March 2001 and May 2002

2.1 Spectra and Sub-millimeter Component Spectra during a flare and a minimum are plotted (see Fig. 2). The spectral index a ( S , 0: va) appears to be 0.1+0.1 at 100 GHz and below, and 1.5?::: between 232 and 345 GHz, suggesting a break frequency in spectral index at 100 GHz or higher. A flux density excess towards sub-millimeter wavelengths has been observed (Zylka, Mezger, & Lesch 1992; Serabyn et al. 1997; Falcke et al. 1998). The overall spectrum 1(v/v~)~'+S ~ ( Y / V ~Three ) ~ ' .sets of combination of crl can fit two power-law components, i.e., S, = S and a 2 are used in the fitting. First, for c u l = 0 and a2 = 2 (dashed lines in Fig. 2), the sub-millimeter component corresponds to the thermal synchrotron emission either arising from the inner region of the accretion disk (Liu & Melia 2002) or produced from a jet-nozzle (Falcke & Markoff 2000). Secondly, a1 = 0 and a2 = 2.5 (solid lines in Fig. 2), the spectral index of 2.5 suggests that a homogeneous opaque, nonthermal synchrotron source might be present in the inner region of the accretion flow. Such a model appears to be plausible if one considers the non-thermal synchrotron particles to be accelerated inside the compact source, perhaps within a jet nozzle as has been proposed for the case of NGC 4258 (Yuan et al. 2002). Sl(v/v1)O 25ezp(-v/vo), Finally, if the low frequency component has an exponential cut-off, i.e. at vo 7 5 GHz, a smaller value of a 2 1.5 (dashed-dotted lines in Fig. 2 ) for the sub-millimeter component is also consistent with a spectrum produced from the ADAF model. A gradient of T, in the ADAF depresses the rising part of the spectrum (Narayan et al., 1998). The observed spectrum suggests an opaque nature of the sub-millimeter component at 1.3 mm and perhaps also 0.87 mm. Observations at shorter sub-millimeter wavelengths appear to be critical to differentiate between the models. The spectrum in a minimum state is also shown here. The excess at 1.3 mm appears to be less significant. N

N

N

2.2 Day-to-Day Variability and Intra-day Variability Based on the sparse data, marginal day-to-day variations at a level 2-3u (or 20-30%) were observed during Flare 1 and Flare 2 as well as May 2002. Intra-day variations on short time scales were searched based on the 24 epochs of observations at 1.3 mm (total time of 100 hr). No evidence for significant variations on a time scale of 1 hr has been found, N

357

Astron. Nachr./AN 324, No. S 1 (2003)

a. 2001 -March Flare1

100

1000

I

b. 2001-June Minimum

-h z 0

1000

Frequency (GHz)

Fig. 2 (a) A mean spectrum of Sgr A* made from the observations near the peak of Flare 1. The curves represent the resultant spectrum of two power-law components. (b) A spectrum of Sgr A* during a minimum phase.

[

~

~

~

‘

~

~

~

;

~

~

< 20%, where S,,, and S,,,, are the maximum and minimum i.e. a variability quantity flux densities at 1.3 mm, respectively, in a single observing track of 6 hrs or less. 2.3

Correlation with the VLA data

Fig. 3 shows the SMA light curve at 1.3 mm overlaid on the VLA light curve at 1.3 cm. The SMA data appears to show a correlation with the light curves observed with the VLA. A quantitative analysis of cross-correlation between the light curves at 1.3 mm and 1.3 cm suggests a global delay of &lay > 3d (Zhao et a]. 2003). The global delay is consistent with a model consisting of a flare that occurs from inside out starting from short wavelengths and then continuing to longer wavelengths. Assuming a global delay of &lay > 3d and a source size of 40 R,3,, an expansion velocity, vezp -1200 km spl or < 0.004 c, is inferred, which is far below the escape velocity of 0.1 c at r ,-..40 Rsr. The bulk kinetic energy associated with the flares appears to be too small to power a noticeable collimated jet in Sgr A* during the SMA monitoring period. The inferred small expansion velocity may imply that other processes contribute to the transport of high-energy particles; e.g. diffusion and convection may also play a role in powering Sgr A” at lower radio frequencies.

Jun-Hui Zhao: Radio Variabilitv of Spr A*

358

In Fig. 3, we also mark the X-ray flares observed in Oct 2000 (Baganoff et al. 2001) and May 2002 during the multi-wavelength campaign (Baganoff et al. 2003). The first observed X-ray flare appcars to be 10 d prior to the peak of the quasi-periodic fluctuation observed at I .3 cm. Corresponding to the multiple X-ray flares observed during the last week of May 2002, the flux density measured with SMA at 1.3 mm shows a 20 increase. Unfortunately, no further SMA observations were made following the multiwavelength campaign of May 2002. However, a 2 0 peak was observed with the VLA at 1.3 cm about 4 weeks after the multiple X-ray flares. The observed properties for the flares at 1.3 mm (time scale of a month and amplitude fluctuation of a factor of a few) are different from those of the X-ray flares (time scale of 1 hr and amplitude fluctuation of a factor of 10 or larger). The lack of strong flares on short time scale at 1 mm places a critical constraint on the models of the inverse Compton scattering as has been proposed for the short duration X-ray flares. Considering the opaque nature of the sub-millimeter component at 1.3 mm, the X-ray flares could remain hidden at 1.3 mm due to self-absorption. ",

L w

i

0

200

400

600

800

1000

d a t e number 0 = (2000-06-21)

-

Fig. 3 The SMA light curve at 1.3 mm (squares) is overkdid with the VLA light curve at 1.3 cm (dots). The solid curve is a smoothed 1.3 cm light curve. The longer period component (P2 333 d) of the DQPOs is noticeable. The X-ray flares observed with Chandra in Oct. 2000 and May 2002 are marked.

3 Variability at 1 Centimeter 3.1

New VLA Light Curves

Fig. 4 shows the new weekly sampled radio light curves observed with the VLA at 2, 1.3 and 0.7 cm during the period between June 2000 to the end of 2002 (Herrnstein et al. 2003 in preparation). The spectral indices of a1 3cm/2cm and a0 7cm/l :irm are also calculated. The mean values are a1 .3cm/2cm = 0.3 f0.2 and ~ ~ 0 , 7 3c7rL ~ ~= / 0.12 1 0.19.

*

3.2

Double Quasi-Periodic Oscillation

Fig. 5 shows the light curve at 1.3 cm obtained from the VLA archival data over the period 1990-1999 (upper panel; Zhao, Bower & Goss 2001) with a total number of 92 data points and a mean sampling interval At=39d. The PSD (middle panel) is derived with the Lomb Algorithm from a portion of the light curve observed at 1.3 cm during 1990 March to 1993 October over a 1400-d period with At=22d. We note that for all the PSD plots, no baseline subtraction was applied to the data prior to the application of the Lomb algorithm. In addition to the strong feature at near O.O1d-l (NlOOd),there is a weak periodic signal at 0.004d-' (-240d). The PSD (bottom panel) confirmed the presence of the double oscillation signals at 0.01d-' and 0.004dC1 from all the data observed from 1990 to 1999 at 1.3cm. The double oscillation signals appear to be also present in the densely sampled light curves observed over an interval from June 2000 to December 2002 at 1.3 cm (see Fig. 3) with a mean sampling interval

359

Astron. Nachr./AN 324, No. S1 (2003)

.. . .

Y O .

..

. *

alpho7mn/' 3cm I

-

.

.

.

I

-- -

.

.

.

I

I

)

I

.

.

/

.

I

.

.

I

,

.

.

.

' .

,

.

.

I

.

.

.

.

I

.

.

.

.

I

.

.

.

.

'

.

.

'

'

"

L

SgrAi-2 0 cm

2 0 - -

---

I

TI--

v)

,*i

I

*

JunOO

*+

4

i

BpOO

OecOO

Apml

JuIOl

0ctot

Jan02

Apr02

oct02

Ju102

I

Fig. 4 The densely sampled radio light curves observed with the VLA at 2, 1.3 and 0.7 rnm during June 2000 to the end of 2002 (Hermstein et al. 2003, in preparation).

of 8d. All three PSDs derived from the different observing periods during the course of the monitoring program are dominated by two features at ,-- 0.0075 d-' (133 d) and ,-- 0.003dP1 (333 d), while the power of the shorter period is decreasing in magnitude (see Fig. 6). The high power oscillation feature of P2333 d is also shown as a sinusoidal oscillation in a smoothed light curve ( nearly three completed cycles over 9 14 d, see Fig. 3). The weaker, minor signals (- 2 0 peaks) appear to correspond to P1 133 d mode, which is superimposed on the P2 oscillation mode. The SMA flares at 1.3 mm are likely associated with the PI oscillation mode. N

4

Summary

With nearly 200 epochs over the past 13 years, the variability of Sgr A* at I .3 cm is dominated by two oscillation modes with periods of P1 = l0Od - 133d and P 2 = 250d - 333d based on both white noise and l/f noise analysis. Both the intensity and the period of the oscillation drift in time, but the ratio of the periods maintains a nearly constant ratio of 2.4 -2.5. The DQPO (double quasi-periodic oscillation)

Jun-Hui Zhao: Radio Variability of Sgr A*

360

for 0 to 14006 1 9 p O 3 1-1993.10 1 0

zL

.

$

.

2 , -

ti

8 d \

t i

I( t‘ 0

: ,

r

h\ t

I

!

may well provide crucial information about the activity near the supermassive black hole, which may be triggered by an orbital resonance occurring in the central few thousand Schwartzschild radius (Zhao et al. 2003, in preparation). The SMA flares at 1.3 mm appear to correspond to the P1 oscillation mode revealed from the VLA light curve observed at 1.3 cm, suggesting that Sgr A* may indeed be regularly powered by the activity near the supermassive black hole. The flares at sub-millimeter wavelengths might be a result of collective mass ejections associated with the X-ray flares that originate from the region near the event horizon where the accretion flow is falling into the black hole. The emitting electrons at sub-millimeter and longer radio wavelengths are likely re-processed from the X-flaring plasma through particle acceleration processes, such as shock and magnetic field reconnection. Sgr A* appears to be opaque at 1.3 mm and perhaps at 0.87 mm. Higher frequency observations of Sgr A* at sub-millimeter wavelengths, where the sub-millimeter component may become partially optically thin, are crucial to our understanding of the nature of the flares observed at centimeter, millimeter/sub-millimeter and X-ray wavelengths.

Astron. Nachr./AN 324, No. S1 (2003)

L

361

+

: a

N

0

Fig. 6 The PSDs derived from the new VLA observations at 1.3 cm during 2000-6-21 to 2002-9-10 (see Fig. 1 for the light curve). We break the light curve into three portions with different number of measurements covering different observing periods in order to show the double oscillation features and their drifting behavior in both intensity and frequency as a function of time. Fig. 3 shows the PSDs derived from three different time periods: 1 ) 2000-6-21 to 2001-7-9 ( S O data points), 2) 2000-6-21 to 2002-2-18 (75 data points), 3) 2000-6-21 to 2002-9-10 (100 data points). Acknowledgements The author especially thanks Miller Goss for his encouragement, interest, and many stimulating discussions during the course of this long term collaboration project. This review is based on collaborations with Goss, Bower, Herrnstein, Ho, Lo, Pegg, Tsutsumi and Young. The author is grateful to Jim Moran, Director of the SMA, for encouragement, support and helpful discussion. The author also thanks Barry Clark for scheduling the VLA monitoring program. The Very Large Array (VLA) is operated by the National Radio Astronomy Observatory (NRAO). The NRAO is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.

References Baganoff, F. K., Bautz, M. W., Brandt, W. N. et al. 2001, Nature, 413,45 Baganoff, F. K. et al. 2003, these proceedings Brown, R. L. and Lo, K. Y. 1982, ApJ, 253, 108 Matsuo, H., Teuben, P., Zhao, J.-H., & Zylka, R. 1998, ApJ, 499, 73 1 Falcke, H., & Markoff, S. 2000, AA, 362, 113 Liu, S. and Melia, F. 2002, ApJL, 566, L77 Moran, J. M. 1998, in: Proc. SPIE, Vol. 3357: Advanced Technology MMW, Radio & Terahertz Telescopes, ed. Thomas, G. Phillips, p. 208 Narayan, R., Mahadevan, R., Grindly, J., Popham, R., & Cammie, C. 1998, ApJ, 492,554 Press, W. H., et al. 1992, Numerical Recipes in C, The Art of Scientific Computing, Second Edition, Cambridge University Press Serabyn, E., Carlstrom, J., Lay. O., Lis, D., Hunter, T., & Lacy, J. 1997, ApJL, 490, L77 Tsuboi, M., Miyazaki, A. & Tsutsumi, T. 1999, ASP Conf. Series 186, p105 & Tsuboi, M. 2002, BAAS, #44.09 Wright, M. and Backer, D. C. 1993, ApJ, 417,560 Yuan, F., Markoff, S., & Falcke, H., Biermann, P. L. 2002, AA, 391, 139 Zhao, J.-H., Ekers, R.D., Goss, W.M., Lo, K.Y. & Narayan R. 1989, IAU Symp. 136,535 Zhao, J.-H., Goss, W. M., Lo, K. Y. and Ekers, R. D. 1992, ASP Conf. Series 31,295 Zhao, J:H. and Goss, W. M. 1993, in: Sub-arcsecond Radio Astronomy, R.J. Davis and R. S. Booth, Cambridge University Press, 38 Zhao, J.-H., Bower, G. C., Goss, W. M. 2001, ApJ, 547, L29 Zhao, J.-H., Young, K. H., Herrnstein, R. M., Ho, P.T.P., Tsutsumi, T., Lo, K. Y., Goss, W. M., & Bower, G. C. 2003, ApJL, 586, March 20 issue, in press Zylka, R., Mezger, P. G., & Lesch, H. 1992, AA, 261, 119

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1,363-369 (2003) / DO1 10.1002/asna.200385071

Flares of Sagittarius A* at Short Millimeter Wavelengths Atsushi Miyazaki*',Takahiro Tsutsumi **

2,

and Masato Tsuboi ***

' Nobeyama Radio Observatoryt, National Astronomical Observatory of Japan, Minamimaki, Minamisaku, Nagano 384-1305, Japan National Astronomical Observatory of Japan, 2-21-1 Osawa, Mitaka, Tokyo 181-8588,Japan Institute of Astrophysics and Planetary Science, lbaraki University, 2-1-1 Bunkyou, Mito, Ibaraki 3108512, Japan

Key words galaxies: nuclei, Galaxy: center, radio continuum Abstract. We have performed intensity monitoring observations toward the Galactic center compact nonthermal radio source, Sagittarius A*, at A= 3 and 2 mm (100 and 140 GHz) from 1996 to 2002 using the Nobeyama Millimeter Array (NMA). In 1996, 1997, 1998, and 2000, the observations were performed during a period of one to two months for each year in a single (inteiinediate resolution) array configuration in order to exclude a systematic error caused by sampling different source structures with the different beam sizes. From November 2000 to May 2001, and from November 2001 to May 2002, the observations were carried out over several months in order to investigate the longer term variability. We have detected two flares in March 2000 and April 2002 as well as a flare observed in March 1998 (Tsuboi et al. 1999). The peak flux densities of the flares were 2-3 Jy at 3 mm, while the mean quiescent flux densities was -1 Jy. In particular for the March 2000 flare, the flux densities of Sgr A'at 2 mm had also reached a peak, -4 Jy, on 8 March and increased AS/S- 300%. The flux density decreased by half in a day. We have detected the intra-day variability of Sgr A*. The upper limit for size of the variable component estimated from timescale of the flare is a few tens of AU. For spectra made from our Observations at mm-wavelengths, the variability of flux density increases with frequency. It appears that the variability in the flare propagates from higher to lower frequency. The folded lightcurve with 106 days cycle, which determined from the analysis of the VLA cm-wavelength data, shows distinct high and low activity states. These results provide evidence for quasiperiodic variability of the flux density of Sgr A*at mm-wavelengths.

1 Introduction Sagittarius A* (Sgr A*) is a unique compact nonthermal radio source located at the dynamical center of the Galaxy (e.g. Eckart & Genzel 1996). Sgr A*is suggested to be associated with a supermassive black hole, with about 2.6 x lo6 Ma, at the Galactic center (e.g. Eckart & Genzel 1996; Ghez et al. 1998). The radio emission of Sgr A*is thought to be powered by the gravitational potential energy released by matter as it accretes onto a supermassive black hole. However, the true radiative nature is still not well understood. Because this source is embedded in thick thermal material, it is practically difficult to observe its fine structures with the present VLBI (very long baseline interferometry; Doeleman et al. 2001). Time variability observation is a powerful and alternate tool to reveal the emission mechanism and the structure of Sgr A*. If the time variability is intrinsic to the source, it should be tightly related to the emission * Corresponding author: e-mail: [email protected],Phone: +81267 984381, Fax: +81267 982923 ** Corresponding author: e-mail: [email protected] *** Corresponding author: e-mail: tsuboi @mx.ibaraki.ac.jp Noheyama Radio Observatory (NRO) is a branch of the National Astronomical Observatory, an inter-university research institute operated by the Ministry of Education, Culture, Sports, Science and Technology.

@ 2003 WILEY-VCH Verlag GmbH & Co. KGaA. Weinheirn

364

A. Miyazaki et al.: Flares of Sgr A* at Short mn-Wavelengths

mechanism and the structure of the emitting region. If the origin is not intrinsic, the observation should at least provide the information about thermal materials around Sgr A*. Time Variability in the flux density of Sgr A*at centimeter wavelengths has been studied in the last decade (e.g. Zhao et al. 1992). The variability below 10 GHz is dominated by refractive interstellar scintillation. However, Zhao et al. (2001) reported the presence of a 106 day cycle of variability at centimeter wavelengths based on an analysis of data observed with the Very Large Array over the past 20 years. Wright & Backer (1993) reported that significant flux variations at X=3 mm occurred in a month during the decay of a flare at short centimeter wavelengths in 1990. The millimeter variability is believed to reflect the intrinsic activity of Sgr A*. However, the millimeter variability is not well established because of the lack of systematic monitoring observations. Thus it is important to carry out systematic and multi-epoch monitoring observations at millimeter wavelengths in order to probe the intrinsic variability of Sgr A*. Tsuboi, Miyazaki, & Tsutsumi (1999) had reported the flare of Sgr A’at 3 mm in March of 1998 (Tsuboi et al. 1999; Miyazaki et al. 1999). The flux density of Sgr A*at 100 GHz was flared over AS/S=lOO% in a week and decreased to the mean flux density within two weeks. The flux density at 140 GHz during the flare also increased by more than 100% (AS/S). On the other hand, Zhao et al. (2003) have also detected flares of Sgr A*at 1.3 mm wavelength using the Suhmillimeter Array (SMA). Recently, Baganoff et al. (2001) detected an X-ray flare lasting about 10 ks and with a peak luminosity -50 times higher than the quiescent state by Chandra observations. The relation between X-ray and radio flares is interesting. Sgr A*is a relatively weak (- 1 Jy) compact component embedded in the extended and strong HI1 region of Sgr A West. The contribution of the free-free emission from the extended structure is significant even at millimeter wavelengths. It is necessary to observe with higher angular resolution ( 5 a few arcsec.) to discriminate the compact component from the extended components. We have conducted intensity monitoring experiments toward Sgr A*at millimeter wavelengths using the Nobeyama Millimeter Array (NMA) at the Nobeyama Radio Observatory (NRO).

2 Observations & Calibrations We have performed intensity monitoring observations toward Sgr A*at 100 and 140 GHz band (X=3 and 2 mm) using the NMA, a six-element interferometer at the NRO from 1996 to 2002. In 1996, 1997, 1998, and 2000 the observations were performed during a period of one to two months for each year. Different configurations of the antennas could introduce a bias in the result caused by sampling different source structures with the different beam sizes. While the highest angular resolution can be achieved by the array configuration with the longest baseline, it always requires the best weather conditions and it is not suitable for intensity monitoring. Therefore, we used a single (intermediate resolution) array configuration of the NMA, “C-configuration”, in these period in order to exclude a systematic error as above instance. Each epoch consists of a set of sequent observations of 2-3 days. The epochs of the observations were separated by about 5 days and carried on during a period of one to two months. From November 2000 to May 2001, and from November 2001 to May 2002, the observations were carried out over several months using the various array configurations, which included “AB-configuration” (long baseline), and “D-configuration” (short baseline), in order to investigate the longer term variability as a known variability at centimeter wavelengths. Each epoch consists of a set of sequent observations of 2 days. The epochs were separated by 10 days to 2 weeks. We used double side band (DSB) SIS receivers at the 100 and 140 GHz bands as the front-ends. The data in 1996 were acquired at observing frequencies of 102 and 146 GHz using the FX correlator with 320 MHz bandwidth. The data from 1997 to 2002 were acquired using the Ultra Wide Band Correlator (UWBC; Okumura et al. 2000) with 1 GHz bandwidth. The UWBC is able to observe in the lower and upper side hands separated by 12 GHz simultaneously. The observed frequencies form 1997 to 2002 were 90 and 102 GHz for the 3 mm band and 134 and 146 GHz for the 2 mm band. Because the observations for

Astron. Nachr./AN 324. No. S1 (2003)

365

Sgr A’at 2 mm band requires the best weather conditions and the phase stability, these observations were made less frequently. The instrumental gain and phase were calibrated by alternating observations of Sgr A*and NRAO 530 at about 20 min. intervals. We also observed an additional phase calibrator, 1830-210 of the known QSO, from 2000 to 2002. The flux density scale was established from the observations of Uranus or Neptune. The absolute flux density scales are typical accurate to about 15% for the 100 GHz band and about 20% for the 140 GHz band. To check the validity of flux calibration, Sgr B2(M) is also observed in synthesis mode. We assumed the flux density of Sgr B2(M), which is HIT region, to be constant. Sgr A*

7 Mar. 2000

j_ l

Sgr A*

8 Mar. 2000 146GHz

-

10

5

0

R.A. offset (arcsec.)

m

c

d -5

fi

-10 I

10

5

I

1

0

.

1

’

-5

I

‘

-10

R.A. offset (arcsec.)

Fig. 1 The spatially filtered maps at 102 GHz in 7 March 2000 (Right),and at 146 GHz in 8 March 2000 (Lefr) made from the I/-Vdata by restricting to spatial frequencies larger than 25 kX.

C-configuration of NMA contains U-V distances at 100 and 140 GHz bands of -7-55 kX and -1077 kX, respectively. AB- and D-configuration contains U-V distances at 100 GHz band of -20-100 kX and -4-25 kX, respectively. We made spatially filtered maps from the U-V data by restricting to spatial 2 2 5 k X ) in order to suppress the contamination from the frequencies larger than 25 kX ( ( U z V2)’/’ surrounding extended components around Sgr A*(for the data taken with the compact array configuration, D-configuration, 17kX). We use the peak flux density on the maps to look for variability. Figure 1 shows the spatially filtered maps at 102 GHz in 7 March 2000, and at 146 GHz in 8 March 2000 which are made from the (I-V data by restricting to spatial frequencies larger than 25 kX. The typical size of the synthesized beam with uniform weighting for C-configuration is 3” x 6” at 100 GHz and 2” x 4” at 140 GHz. The resultant peak flux densities are corrected for the effect that the flux density is reduced due to phase errors by atmospheric fluctuations (The fractions of the reduction are 10-20% for 100 GHz and 20-40% for 140 GHz).

+

3 Results & Discussion Figure 2a shows the light curve of Sgr A*at 100 GHz band constructed from all monitor data from 1996 to 2002. Total number of observations is about 60 days. The light curve shows that Sgr A*has distinct active and quiescent phases. There are three active phases in March 1998, March 2000, and April 2002 (Flare I, Flare 11, and Flare 111, respectively, indicated by the arrows in Figure 2a) with duration of, roughly, one month. Mean flux densities of Sgr A*in quiescent phase except for the flare phase are 1.1=k 0.2 Jy and 1.2 k 0.2 Jy at 90 and 102 GHz, respectively. These flare phases are account for 20 % of all data. The increase (AS/S) of the flux density during the flare reached to 200%. The flare in March 1998 (Flare I) has been reported in the previous papers (Tsuboi et al. 1999; Miyazaki et al. 1999). Baganoff et al. (2001) N

366

A. Miyaaaki et a].: Flares of Sgr A* at Short rnrn-Wavelengths MJD 3,5

50875

MJD 3.5

50900

Nov.

1996 Date

2000

OeC.

51625

51600

Jan.

Feb.

Mar.

Date

51650

Apr.

May

51675

Jun

2001

Fig. 2 The light curve of Sgr A'at 100 GHz band from 1996 to 2002. The open and filled circles show flux densities of Sgr A'at 90 and 102 GHz, respectively. (a) The light curve of whole flux density data from 1996 to 2002. There are three active phases in March 1998, March 2000, and April 2002 (Flare I, Flare 11, and Flare EI,respectively, indicated by the arrows). (b) The light curve in 1996, (c) in 1998, (d) in 2000, and (e) from November 2000 to May 2001. The light curve is separated into the active (in 1998, 2000) and the non-active (in 1996,2000-01) phases. The dash line indicates the mean flux density at 100 GHz band.

had detected the X-ray flare of Sgr A*with Chanrlra X-ray Observatory at 26-27 October 2000. However, there are no observations by us around the end of October 2000 (Figure 2e) to make any comparison. In 22 November 2000, our observation detected a flux density of 1.0 Jy at 100 GHz band. N

Astron. NachrJAN 324. No. S1 12003)

361

3.1 Variability Feature in the Flares In the March 2000 Hare (Flare II), the Hux densities of Sgr A*at 3 mm had reached 2 peaks (-3 Jy) on 7 and 21 March and increased AS/Sw 200% for both. The Flare I1 was also observed at 140 GHz band. Figure 3a shows the light curves of Sgr A*at 100 and 140 GHz bands in March 2000. There were several flares in the active phase, most notably a steep peak at 140 GHz band observed on 8 March 2000. The flux densities of Sgr A*at the peak were 3.5 f 0.7 Jy and 3.9 f0.8 Jy, at 134 and 146 GHz, respectively, while the averaged quiescent Hux was about 1 Jy at 140 GHz band. The Hux density then decreased to 2.2 3~ 0.4 Jy at 146 GHz on the subsequent day, 9 March 2000. The half decay timescale of the flare at 146 GHz, was at most 24 hours. The flare amplitude from the mean flux density level (AS/S) reached to about 300% at 146 GHz, which is larger than that at 100 GHz band (- 200%), indicating the variability increases with frequency. We measured the decay timescale of the Hare to be one day (8-9 March 2000). The upper limit of the timescale estimated from time to increase by 100% is about 12 hours. Figure 3b shows the light curves of Hare of Sgr A*at 146 GHz in 8 March 2000. We divided the data set observed in several parts and measured the flux density of Sgr A'from the divided data. The flux density of Sgr A*averaged over 5 to 15 minutes bin at 146 GHz increased from 3.5 to 4.7 Jy between 15h45m UT to 16h 15m UT in 8 March 2000 (Miyazaki et al. 2003). The relative error within one observation session was estimated to be about 5% at 140 GHz band. Thus the 30% Hux increase in 30 minutes is probably real. The time scale that the Hux density increased by 100% is estimated to be about 1.5 hours assuming that the increase has a constant gradient. The increasing time scale, 1.5 hours, provides the physical size (light crossing size) of the Hare region in accretion disk is compact at or below 10 AU (= 200 R,; Schwarzschild radius, R, = 2GM/c2). (Miyazaki et al. 2003). Moreover, there appear also a rapid increase on Flare I11 in April 2002. The Hux density of 1.4 0.2 Jy in 5 April 2002 at 102 GHz had been doubled in only one day and the Hux density was 2.2 0.3 Jy in 6 April. Similar short-term variation at mm-wavelengths has been reported for the galactic nucleus ofM81 (Sakamoto et al. 2001). N

+

*

MJD 51600

51620

51 640

4.0

I7 134GHz

/

102GHz 0 90GHz

5.0

8

7 :eb.

146GHz

8 Mar. 2000 Y 15:30:00

16:OO:OO

16:30:00

U.T. 2000

April

Date Fig. 3 (a) The light curve of Sgr A*at I 0 0 and 140 GHz bands in the observation season of 2000. The flux density at 100 GHz was violently changing (circles).There is a steep peak at 140 GHz band at 8 March 2000 (squares). (b) The light curve of Sgr A*at 146 GHz in the observation on 8 March 2000. This figure shows an intra-day variability of the flux density of Sgr A*.

A. Mivazalu et al.: Flares of Sgr A' at Short mm-Waveleneths

368

3.2 Spectrum of Sgr A*and Long Term Variability Figure 4 shows the spectra of Sgr A'at mm-wavelengths (90, 102, 134, and 146 GHz) in several epochs. The observations of each band (100 and 140 GHz band) have been performed at two consecutive days, respectively. The spectral index of Sgr A*at cm-mm wavelength in the quiescent phase are consistent with 0.3, from the extrapolation of cm-wavelengths ( e g Falcke et al. 1998; the expected spectral index, Krichbaum et al. 1997). We did not find the short-mdsubmm excess in the spectrum of Sgr A*over the expected flux determined from the extrapolation of lower frequency data in the quiescent phase. The excess is presumably caused by the variable component in mm-wavelengths (Tsuboi et al. 1999). Moreover, as a conjecture from the light curve of the Flare I1 (figure 3), the variability of flux density increases with frequency. It appears that the (time sequence of) flares propagate from higher to lower frequency. N

1 /A - 1 Sgr A"

5.0

_p 7-8 Mar

3.0

Flare 13-16Mar.'98

w

'00

14-15Mar '00 (D

0 I02GHz 0 90GHz

13-16 Mar '98 6 Apr '02

.

29 Feb. '00

5

0.4

\ 70 80 90100

200

Frequency (GHz) Fig. 4 Spectra of Sgr A*at mm-wavelengths (90, 102, 134, and 146 GHz). The straight lines indicate the best-fitting power law to the flare and quiescent phases in 1998 compiled the data at 2.25 GHz and 8.3 GHz are from the archival data of the Green Bank Interferometer (GBI) of NRAO. The cm-mm spectra were explained by the power law with the spectral index of 0.55 at time of the flare and 0.34 during the quiescent phase (Tsuboi et al. 1999).

3

0

20

40

60

80

100

Date Fig. 5 The folded lightcurve of the flux density of Sgr A'at 90 and 102 GHz with a 106 days cycle. The open and filled circles show flux densities of Sgr A'at 90 and 102 GHz, respectively. The dash line indicates the mean flux density at 100 GHz band.

Zhao, Bower and Goss (2001) had reported the presence of a quasi-period of 106 days cycle in the variability at centimeter wavelength of Sgr A*based on an analysis of data observed with the VLA over the 20 years. We folded the our NMA light curve with the 106 days determined from the analysis of the VLA centimeter wavelength data. Figure 5 shows the folded lightcurve of the flux density of Sgr A*at 90 and 102 GHz. In Figure 2, the light curve can be separated into the active phase (flares in 1998 (c), 2000 (d)) and the non-active phase (quiescent in 1996 (b), 2000-01 (e)). The folded lightcurve in Figure 5 shows distinct high and low activity states. These results provide evidence for a quasiperiodic variability of the flux density of Sgr A*at mm-wavelengths.

Astron. Nachr./AN 324. No. S1 (2003)

4

369

Summary

We have performed monitoring observations of the flux density of Sgr A*at 3 mm (100 GHz) and 2 mm (140 GHz) bands using the Nobeyama Millimeter Array (NMA) in 1996, 1997,2000,2000-01, and 200102. 1) We detected three flares from Sgr A*in March 1998, March 2000, and April 2002, at 100 and 140 GHz bands. The peak flux densities of the flares were 2-3 Jy at 100 GHz, while the mean quiescent flux densities were -1 Jy. At 140GHz, the flux density increases up to 200% (=AS/S). 2 ) For the March 2000 flare, the flux density of Sgr A*at 2 mm also reached a peak -4 Jy on 8 March and increased AS/S- 300%. The flux was decreased by half in a day. Moreover, we detected the 30% flux increase in 30 minutes on 8 March 2000. This increase demonstrates the intra-day variability of Sgr A*. The upper limit for the size of the variable coniponent estimated from the timescale of the flare is a few tens of AU. 3) For spectra made from our observations at mm-wavelengths, the variability of flux density increases with frequency. It appears that the variability in the flare propagates from higher to lower frequency. 4) The folded lightcurve with the 106 day cycle, determined from the analysis of the VLA cm-wavelength data, shows distinct high and low activity states. These results provide evidence for quasiperiodic variability in the flux density of Sgr A*at mm-wavelengths. Acknowledgements We thdnk the staff of Nh4A group of the Nobeyama Radio Observatory for support in the observation. They also thank T. Kawabata at Bisei Observatory for useful discussions.

References Baganoff, F.K., Bautz, M.W., Brandt, W.N., et al. 2001, Nature, 413, 4.5 Doeleman, S.S., et al. 2001, AJ, 121, 2610 Eckart, A. & Genzel, R. 1996, Nature, 383,415 Falcke, H., et al. 1998, ApJ, 499, 731 Ghez, A.M., Klein, B.L., Moms, M., & Becklin, E.E. 1998, ApJ, 509, 678 Krichbaum, T.P., et al. 1998, A&A, 335, L106 Miyazaki, A.,Tsutsumi, T., & Tsuboi, M., 1999, Advances in Space Research, 23, 977 Miyazaki, A.,Tsutsumi, T., & Tsuboi, M., 2003, in preparation OkUmurd, S.K., et al. 2000, PASJ, 52, 393 Sakamoto, K.,Fukuda, H., Wada, K.,Habe, A. 2001, AJ, 122, 1319 Tsuboi, M., Miyazaki, A., & Tsutsumi, T. 1999, in ASP Conf. Ser. 186, The Central Parsecs of the Galaxy, eds. H. Falcke, A. Cotera, W.J. Duschl, F. Melia, & M.J. Rieke (San Francisco: ASP), p. 105 Wright, M.C.H., & Backer, D.C. 1993, ApJ, 417,560 Zhdo, J.-H., Goss, W.M., Lo, K.Y., & Ekers, R.D. 1992, in Relationships between Active Galactic Nuclei and Starburst Galaxies, ed. A.V. Flilpenko (San Francisco: ASP), p. 295 Zhao, J.-H., Bower, G.C., & Goss, W.M. 2001, ApJL, 547, L29 Zhao, J.-H., et al. 2003, ApJ Letters, in press

This Page Intentionally Left Blank

Astron. Nachr./AN 324,No. S1, 371-376 (2003) / DO1 10.1002/asna.200385062

Limits on the Short Term Variability of Sagittarius A* in the Near-Infrared S. D. Hornstein*',A. M. Ghez's2, A. Tanner', M. Morris', and E. E. Becklin' I

Department of Physics and Astronomy, University of California at Los Angeles, Los Angeles, CA 900951562 Institute for Geophysics and Planetary Physics, University of California, Los Angeles, CA 90095-1567

Key words accretion, accretion disks - black hole physics - Galaxy: center - infrared: galaxies X-rays: galaxies - galaxies: jets

Abstract. The detection of X-ray flares by the Chandra X-ray Observatory and XMM-Newtonhas raised the possibility of enhanced emission over a broad range of wavelengths from Sagittarius A*, the suspected 2.6 x 10' M a black hole at the Galactic center, during a flaring event. We have, therefore, reconstructed 3-4 hr data sets from 2 pm speckle and adaptive optics images (Q,,,, = 50 - 100 mas) obtained with the W. M. Keck 10 m telescopes between 1995 and 2002. The results for 25 of these observations were reported by Hornstein et al. (2002) and an additional 11 observations are presented here. In the 36 separate observations, no evidence of any significant excess emission associated with Sgr A* was detected. The lowest of our detection limits gives an observed limit for the quiescent state of Sgr A* of 0.09*0.005 mJy, or, equivalently, a dereddened value of 2.0f0. I rnJy. Under the assumption that there are random 3 hr flares producing both enhanced X-ray and near-infrared emission, our highest limit constrains the variable state of Sgr A* to 20.8 mJy (observed) or 19 mJy (dereddened). These results suggest that the early model favored by Markoff et al. (2002), in which the flare is produced through local heating of relativistic particles surrounding Sgr A* (e.g., a sudden magnetic reconnection event), is unlikely because it predicts peak 2 prn emission of -300 mJy, well above our detection limit.

1 Introduction The variability of Sagittarius A* at X-ray wavelengths (Baganoff et al., 2001, 2003; Goldwurm et a]., 2003) has bolstered the case for associating this source with the suspected 2.6 x lo6 M o black hole at the center of our Galaxy (Eckart & Genzel, 1997; Genzel et al., 1997,2000; Ghez et al., 1998,2000). The first evidence of X-ray variability was detected by Chundru, during which Sgr A* was seen to flare in intensity over a time scale of -3 hr (Baganoff et al., 2001). While the flare's short duration implied a small region of origin, 5 4 0 0 R, (where R, is the Schwarzschildradius = 2GM./c2), its large amplitude, a factor of 50, has raised the possibility of detecting corresponding intensity enhancements at wavelengths outside the X-ray regime. Later observations with XMM-Newton (Goldwurm et al., 2003) as well as follow-up observations by Chundru (Baganoff et al., 2003) both show similar flaring activity. Existing models for Sgr A*'s flared state make very disparate predictions for the emission at wavelengths between the X-ray and radio regimes (Markoff et al., 2001; Liu & Melia, 2002; Narayan, 2002). The wide differences between these models are a result of assuming different geometries (disk vs. jet) and emission mechanisms for the flaring process (e.g., enhanced accretion rates vs. magnetic reconnection). In some models, the predicted emission in the flared state, at infrared wavelengths, dramatically exceeds that of existing detection limits (Genzel & Eckart, 1999; Stolovy et al., 1999; Morris et al., 2001). For example, * Corresponding author: e-mail: [email protected],Phone: 310-825-4434,Fax: 310-206-2096

@ 2003 WlLEY-VCH Verlag OmbH & Co. KGaA, Weinhem

S. D. Hornstein et al.: Limits on Near-IR Variability of Sgr A*

312

-

the preferred model of Markoff et al. (2001) predicts an observed 2 p m flux density of 13 mJy, or a dereddened flux density of -300 mJy, during the flared state. Unlike the situation at radio wavelengths, where long-term monitoring campaigns have been used to constrain the flared state of Sgr A* (Zhao, Bower, & Goss, 2001), the limited time coverage and spatial resolution of published IR experiments prevent meaningful constraints on the flared state’s IR emission from bcing inferred’ and thus the reported limits are assumed to be associated with Sgr A*’s quiescent state. The W. M. Keck Observatory dynamical study of stars in the central stellar cluster (Ghez et al., 1998, 2000; Gezari et al., 2002) provides a rich source of high angular resolution 2 pm data between 1995 and 2002. In, these proceedings, we present 2 pm flux density limits from maps that were each composed of data from a single night. The elapsed time of 3-4 hours in each map is approximately the same as the time scale of the observed X-ray flares, making this data set ideally suited for possibly detecting a flare of this type. Given the quantity of these observations, our non-detections establish a robust upper limit for the flared state’s 2 pm emission intensity.

2 Observations High-resolution, near-infrared observations of the Galactic center were conducted from 1995 June to 2002 July using both speckle and adaptive optics (AO) imaging techniques on the Keck 10 m telescopes. The speckle observations were obtained in the K band (A, = 2.2 pm, AX=0.4 pm) using the Keck I facility near-infrared camera (NIRC; Matthews & Soifer, 1994) with external reimaging optics. This resulted in a pixel scale of 0‘.’0203 and a field of view (FOV) of 5’.!12x5!’12 (Matthews et al., 1996). During each night of observations, several thousand short exposures (tezp= 0.137 sec) were taken in sets of -200. A more limited set of data was collected using two different science cameras behind the Keck Ll A 0 system (Wizinowich et al., 2000b). The first A 0 data set was collected in the K’ band (A, = 2.1 pm, AA=0.35 pm) in early 1999 with the near-infrared engineering camera (KCAM; Wizinowich et al., 2000a), which had a pixel scale of 0!’0175, an FOV of 4!’4~4!’4. Each image had an exposure time of 5 s. The slit-viewing camera of NIRSPEC (SCAM; McLean et al., 1998) provided a second set of A 0 images for this study. These images, like the speckle images, were made in the K band and had a pixel scale of W0170, a FOV of 4!’4 x4!’4, and an exposure time of 10 s. USNO 0600-28579500 served as the natural guide star for all of these A 0 observations. Since this guide star is both faint (R = 13.2) and distant from the target ( r ~ 3 0 ” ) , the A 0 performance was non-optimal. All observations prior to 2002, with the exception of 1998 Aug and 1998 Oct, are reported by Hornstein et al. (2002) and Table 1 provides a summary of all new observations.

3 Data Analysis & Results Three basic steps constitute the data analysis process in this program. First, high angular resolution maps are generated from the individual short exposure frames (53.1). Second, all point sources in the FOV are identified and a direct detection of Sgr A* is ruled out (53.2). Third, limits for Sgr A* are derived from the residual maps, in which all identified point sources have been removed ($3.3). 3.1 Construction of Images

Image processing proceeds similarly to that carried out for the dynamical experiment, with one exception. Rather than combining all the data from the duration of an observing run, typically 2-3 nights, we synthesize the data over each night to produce 38 maps, each of which is limited to an elapsed time of 3-4 hours.

’

While several papers have reported the possible detection of a variable near-infrared source coincident with Sgr A* (Herbst, Beckwirh, & Shure, 1993; Close, McCarthy, & Melia, 1995; Genzel et al., 1997), subsequent high resolution observations have identified this emission to be from high proper motion sources (Eckart et al., 1995; Eckart & Genzel, 1997; Ghez et a]., 1998, 2003).

313

Astron. Nachr./AN 324. No. S 1 (2003)

Table 1 New Near-Infrared Limits on Sagittarius A*"

PSF Characteristics Limit' Limit" Epoch FWHM,,,, FWHMhalo Eh,,,, (UT) Camera Number of Frames (arcsec) (arcsec) (%) (mag) (mJy) 0.05 0.2 3 16.13 5.1 1998 July4 NIRC 2352 1998 Oct 9 NIRC 2064 0.05 0.3 4 15.40 10.0 0.05 0.3 2 15.55 8.7 2002 Apr 23 NIRC 2880 0.05 0.3 2 16.74 2.9 2002 Apr 24 NIRC 6460 2002 May 23 NIRC 1900 0.04 0.3 3 16.24 4.6 0.05 0.2 5 16.24 4.6 2002 May 24 NIRC 2280 5 16.47 3.7 2002 May 28 NIRC 2660 0.06 0.3 0.05 0.3 3 16.12 5.1 2002 May 29 NIRC 1710 2002 Jun 01 NIRC 5302 0.05 0.2 4 17.09 2.1 0.05 0.3 2 15.61 8.2 2002 J u l l 9 NIRC 1900 2002 Jul20 NIRC 2470 0.06 0.4 3 15.72 7.5 "Used in combination with the results reported previously by Hornstein et al. (2002). bPercentage of total energy contained in the PSF core "Using an A,=30 with Ak/A,=O. 112 and Ak,/A,=O. 1 17 (Melia & Falcke, 2001; Rieke & Lebofsky, 1985), the seventh and eighth columns list observed limits and dereddened values, respectively. Since the details of this method are described elsewhere (Ghez et al., 1998), only a brief summary is provided here. Standard image reduction techniques are applied to all the individual speckle and A 0 frames. For the speckle data, a two stage shift-and-add (SAA; Christou, 1991; Ghez et al., 1998) analysis then produces the final high resolution maps. In the first stage, the 200 frames in each set are combined to form an intermediate SAA image. Then, these multiple intermediate SAA images (from throughout the night) are combined to form one final SAA map. This allows each intermediate image to be examined for seeing quality. In combining the intermediate images, a seeing cut is established so as to exclude those images with the worst seeing from the final map. For the A 0 data, this cut is also carried out on the individual A 0 images before they are registered and averaged together. Figure 1 displays representative final nightly speckle and A 0 maps. 3.2

Point Source Identification & Search for Sgr A*'s near-infrared emission

In all maps, stars are identified using StarFinder, an IDL package developed for astrometry and photometry in crowded stellar fields (Diolaiti et al., 2000). This package iteratively generates estimates of the point spread function (PSF) from a few selected bright stars and then identifies point sources over the entire FOV through cross-correlation of the map with the PSF model. For the PSF extraction, we found that the most reliable PSF models are obtained with a support size of -2", which represents a compromise between needing to accommodate the large PSF halos and yet having a limited FOV. This choice limits our analysis to images with PSF halo sizes of 0!'4 or less, as the PSFs of the remaining two images are poorly characterized by this process. The PSF model is based on four of the five brightest stars in the FOV (IRS 16NE, 16C, 16NW, and 29N; see Figure 1); IRS 16SW is avoided as a PSF model star since it is surrounded by relatively bright stars in its immediate vicinity. For point source identification, a correlation coefficient greater than 0.8 between the PSF model and the actual stellar image is required to avoid spurious detections. This process results in the identification of -100 point sources in each map. The speckle and A 0 images have significantly different PSFs. Nonetheless, both PSFs are composed of a compact core on top of a broader halo. Table 1 provides the characterization of the PSF in each new map based on the radial profile of the PSF model. While the speckle images have PSF core FWHM that are nearly diffraction-limited (-0!'05) and -40% smaller than that of the A 0 images (-U.'08), the A 0 PSF cores contain -30% of the total energy, 12 times more than the typical speckle PSF.

-

S. D. Hornstein et al.: Limits on Near-IR Variability of Sgr A*

374

Fig. 1 Speckle (left) and A 0 (right) images from 1999 May of the Galactic center. The small box indicates a 1”x 1” region centered on Sgr A*, whose approximate position is marked with a cross. Both images are displayed with a histogram equalization stretch to show the fainter stars in the field and are oriented such that north is up and east is to the left.

At this stage in the analysis, it is possible to look for direct detections of Sgr A*. We use proper motion acceleration vectors (Ghez et al., 2000) to pinpoint the central black hole’s position relative to the nominal radio position of Sgr A* (Menten et al., 1997) to within lY.’04 (1 a). Within 0!’08 of this location, 4 sources (13.9 < K < 16.5) are identified, all of which were previously detected in ”monthly” maps made from all data in a single observing run (Ghez et al., 1998,2003) and, furthermore, have significant proper motions. This high stellar density emphasizes the need for improved accuracy in Sgr A*’s position in the IR reference frame in order to measure or constrain its emission. With no stationary source identified in this region, we conclude that Sgr A* has not been detected. 3.3 Flux Density Limits for Sgr A* In order to determine an accurate detection limit at the position of Sgr A*, it is necessary to remove the contaminating seeing halos from nearby sources. A “stars-only” map is created using the PSF model and list of stars generated by StarFinder. This is then subtracted from the original map, producing a residual map. With the residual map, a 3 G point source detection limit for Sgr A* is established based on 3 times the rms of 25 aperture photometry values, which are calculated using -60 mas radius apertures and sky annuli extending from -60 to -90 mas. The 5 x 5 grid of apertures in the residual map corresponds to an area of -0“6x0?6, more than 2 orders of magnitude larger than the uncertainty in the location of Sgr A*. Zero points are obtained through carrying out the same aperture photometry in the original maps (prior to the stars-only subtraction) on all known nonvariable sources brighter than K=lO.S,using the flux densities reported by Blum, Sellgren, & Depoy (1996), and that occur in more than 30% of the frames for each night. The photometric calibration sources used are IRS 16NW, 16C, 16CC, and 16NE, when the FOV allows its inclusion; IRS 29N is omitted since it is found to be marginally variable at the 2 0 level. Typical photometric zero-point 1 a uncertainties of -0.04 mag result from this procedure, Table 1 and Figure 2 contain the resulting 3 a point source detection limits for Sgr A*. The lowest of these upper limits gives an observed limit for the quiescent state of Sgr A* of 0.09~t0.005mJy or, equivalently, a dereddened value of 2.0f0.1 mJy, while the highest limit constrains our analysis of the variable state of Sgr A* to 50.8 mJy (observed) or 19 m l y (dereddened). Although these upper limits

Astron. Nachr./AN 324, No. S 1 (2003)

315

1001

x

-J

E

'

'

'

'

'

'

'

1

'

'

'

'

'

'

'

'

'

'

'

'

'

I

v

zE

'

1

1 &

I

1

1 I .l

l I . . a . . . . . a . . . . . . . . . . . a . . 1996 1998 2000 2002 Epoch (UT)

Fig. 2 The 3 c limiting flux density calculated for Sgr A* for each epoch of observation, including the results reported ) lowest limit of 2.0 mJy and previously by Hornstein et al. (2002) (corrected for reddening by a factor of ~ 2 2 . The the highest limit of 19 mJy constrain Sgr A*'s quiescent and variable states, respectively.

vary significantly from epoch to epoch (because of variable observing conditions and/or a variation in the number of contaminating sources detected and removed), the lowest of them is lower than the best previously reported limit at 2 p m for Sgr A*'s quiescent state prior to Hornstein et al. (2002) (dereddened 4 mJy; Genzel & Eckart, 1999)by a factor of 2.

4 Discussion Using the coverage of the X-ray and near-infraredexperiinents and assuming random 3 hr flaring events that produce both enhanced X-ray and near-infrared emission, we consider the likelihood that a near-infrared flare in excess of our weakest limit occurred over the course of the IR experiment. After an intense X-ray monitoring campaign, the mean rate of factor-of-ten X-ray flares has been established as 0.63~0.3per day (Baganoff et al., 2003). Unfortunately, the flare observed by XMM-Newtonoccurred at the very end of the observations and was not observed in its entirety. Therefore, we do not consider that flare here. Given the duty cycle for the Chandru flares, there was a 10% probability of seeing a flare in any one of our near-infrared observations and a 5% probability of seeing zero flares throughout the entire experiment. We therefore assume that at the 2 (T confidence level, a flare occurred during our experiment, and we use our limits to constrain the variability models. While only a limited amount of modeling of the original 3 hr X-ray flare detected at Sgr A* has been carried out, existing models predict 2 pm dereddened emission as high as 300 mJy in the model preferred by Markoff et al. (2001) but as small as 0.4 mJy by Liu & Melia (2002). The former model explains the elevated X-ray emission, produced by synchrotron self-Compton emission, by an enhanced temperature for the relativistic electron population, as might arise in a magnetic reconnection event. On the other hand, Liu & Melia (2002) present a flare model in which the flare arises as the result of an enhanced accretion rate, and bremsstrahlung emission is dominant at both near-infrared and X-ray wavelengths. Even less modeling has been done of the more recent factor-of-ten X-ray flares (see, e.g., Markoff et al., 2003). For these proceedings, we assume that any factor-of-ten flare produces a detectable near-infrared emission. The lack of a near-infrared detection of Sgr A* makes the Markoff et al. (200 I ) preferred model and any other mechanism that produces flared 2 pm emission in excess of 19 mJy (dereddened) unlikely.

5 Conclusions These proceedings summarize a search for a near-infrared counterpart to Sgr A* in the flared state. From the length of our observations, this search was sensitive to variability on time scales of 3-4 hr. No

376

S. D. Hornstein et al.: Limits on Near-IR Variabilitv of Ser A*

such counterpart was detected. However, by identifying and removing all the stars in the crowded inner -0!’6xU.’6 of the Galactic center, an upper limit for the near-infrared emission from Sgr A* has been inferred for each observation epoch. These limits constrain the quiescent emission from Sgr A* to 50.09 mJy (2.0 mJy, dereddened) and the variable component to 50.8 mJy (19 mJy, dereddened) at the 2 u confidence level. Continued monitoring of Sgr A* at other wavelengths (particularly mid-IR and submm) will provide valuable information about Sgr A*’s flaring properties.

Acknowledgements Data presented herein were obtained at the W.M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W.M. Keck Foundation. This work was supported by the NSF through both grant 9988397 and the Science and Technology Center for Adaptive Optics, managed by the University of California at Santa Cruz under cooperative agreement No. AST - 9876783. AMG also thanks the Packard Foundation for its support.

References Baganoff, F. K. et al. 2001, Nature, 413,45 Baganoff, F. K. et al. 2003, these proceedings Blum, R. D., Sellgren, K., & Depoy, D. L. 1996, ApJ, 470,864 Christou, J. C. 1991, PASP, 103, 1040 Close, L. M., McCarthy, D. W., & Melia, F. 1995, ApJ, 439,682 Diolaiti, E., Bendinelli, 0..Bonaccini, D., Close, L., Cume, D., & Parmeggiani, G. 2000, A&AS, 147, 335 Eckart, A. & Genzel, R. 1997, MNRAS, 284,576 Eckart, A., Genzel, R., Hofmann, R., Sams, B. J., & Tacconi-Garman, L. E. 1995, ApJ, 445, L23 Genzel, R. & Eckart, A. 1999, ASP Conf. Ser. 186: The Central Parsecs of the Galaxy, 3 Genzel, R., Eckart, A,, Ott, T., & Eisenhauer, F. 1997, MNRAS, 291, 219 Genzel, R., Pichon, C., Eckart, A,, Gerhard, 0. E., & Ott, T. 2000, MNRAS, 317,348 Gezari, S., Ghez, A. M., Becklin, E. E., Larkin, J., McLean, I. S., Moms, M. (2002), ApJ, 576, 790 Goldwurm, A,, Brion, E., Goldoni, P., Fernando, P., Daigne, F., Decourchelle, A,, Warwick, R. S., & Predehl, P. 2003, ApJ, 584,751 Ghez, A. M., Salim, S., Hornstein, S. D., Tanner, A,, Moms, M., Becklin, E. E., Duchene, G. 2003, ApJ, submitted (asuo-ph/0306130) Ghez, A. M., Klein, B. L., Morris, M., & Becklin, E. E. 1998, ApJ, 509,678 Ghez, A. M., Moms, M., Becklin, E. E., Tanner, A., & Kremenek, T. 2000, Nature, 407, 349 Herbst, T. M., Beckwith, S. V. W., & Shure, M. 1993, ApJ, 411, L21 Hornstein, S. D., Ghez, A. M., Tanner, A., Moms, M., Becklin, E. E., & Wizinowich, P.2002, ApJ, 577, L9 Liu, S . & Melia, F. 2002, ApJ, 566, L77 Markoff, S., Falcke, H., Yuan, E, & Biermann, P. L. 2001, A&A, 379, L13 Markoff, S. et al. 2003, these proceedings Matthews, K., Ghez, A. M., Weinberger, A. J., & Neugebauer, G. 1996, PASP, 108,615 Matthews, K. & Soifer, B.T. 1994, Infrared Astronomy with Arrays: The Next Generation, I. McLean ed. (Dordrecht: Kluwer Academic Publishers), 239 McLean, I. S. et al. 1998, Roc. SPIE, 3354,566 Melia, F. & Falcke, H. 2001, ARA&A, 39,309 Menten, K. M., Reid, M. J., Eckart, A,, & Genzel, R. 1997, ApJ, 475, L1 I 1 Moms, M., Tanner, A. M., Ghez, A. M., Becklin, E. E., Cotera, A., Werner, M. W., & Ressler, M. E. 2001, American Astronomical Society Meeting, 198,4101 Narayan, R. 2002, Lighthouses of the Universe: The Most Luminous Celestial Objects and Their Use for Cosmology Proceedings of the MPAESO,p. 405,405 (astro-pW0201260) Rieke, G. H. & Lebofsky, M. J. 1985, ApJ, 288,618 Stolovy, S. R., McCarthy, D. W., Melia, F., Rieke, G., Rieke, M. J., & Yusef-Zadeh, F. 1999, ASP Conf. Ser. 186 The Central Parsecs of the Galaxy, 39 Wizinowich, P. L.,Acton, D. S., Lai, O., Garthright, J., Lupton, W., Stomski, P.J., 2000a, Proc. SPIE, 4007,2 Wizinowich, P. L. et al. 2000b, PASP, 112, 315 Zhao, J., Bower, G. C., & Goss, W. M. 2001, ApJ, 547, L29

Astron. Nachr./AN 324, No. S1, 377-382 (2003)/ DO1 10.1002/asna.200385041

A New X-Ray Flare from the Galactic Nucleus Detected with XMM-Newton A. Goldwurm*l, E. Brion2, P. Goldoni', P. Ferrandol, F. Daigne', A. Decourchelle', R. S. Warwick3,and P. Predeh14

' Service d' Astrophysique, DAPNIAIDSMICEA, CE-Saclay, F-91191 Gif-Sur-Yvette,France

* Centre d'Etude NuclCaire de Bordeaux-Gradignan,AllCe du Haut Vigneau, 33175 Gradignan, France Department of Physics and Astronomy, University of Leicester, Leicester LEI 7RH, UK Max-Planck Institut fur Extraterrestrische Physik, Postfach 1312, 85741 Garching, Germany

Key words Accretion, accretion disks - Black hole physics - Galaxy: center - X-rays: general PACS 04A25

The compact radio source Sgr A*, believed to be the counterpart of the massive black hole at the Galactic nucleus, was observed to undergo rapid and intense flaring activity in X-rays with Chandra in October 2000. We report here the detection with XMM-Newton EPIC cameras of the early phase of a similar X-ray flare from this source, which occurred on 2001 September 4. The source 2-10 keV luminosity increased by a factor z 20 to reach a level of 4 erg spl in a time interval of about 900 s, just before the end of the observation. The data indicate that the source spectrum was hard during the flare and can be described by simple power law of slope % 0.7. This XMM-Newton observation confirms the results obtained by Chandra, suggests that, in Sgr A*, rapid and intense X-ray flaring is not a rare event and therefore sets some constraints on the emission mechanism models proposed for this source.

1 Introduction The bright, compact radio source Sgr A" is believed to be the radiative counterpart of the 2.6 lo6 Ma black hole which governs the dynamics of the central pc of our Galaxy (Melia & Falcke 2001). The compelling evidence for such a massive black hole at the Galactic Center (see Schodel et al. 2002 for the most recent results), contrasts remarkably with the weak high-energy activity of this object. In spite of the fact that stellar winds and hot gas probably provide enough material for a moderateflow level of accretion, the total bolometric luminosity of the source amounts to less than lop6 of the estimated accretion power (Melia & Falcke 2001, Goldwurm 2001). This motivated the development of several black hole accretion flow models with low radiative efficiency, some of which have also been applied to binary systems, low luminosity nuclei of external galaxies and low luminosity active galactic nuclei. These models include spherical Bondi accretion in conditions of magnetic field sub-equipartition with a very small Keplerian disk located within the inner 50 Schwarzschild radii (Rs),large hot two-temperature accretion disks dominated by advection (ADAF) or non-thermal emission from the base of a jet of relativistic electrons and pairs, and some other variants or combination of the above models. However any such model still predicts some level of X-ray emission from Sgr A* and determining the properties of such emission would constrain the theories of accretion and outflows in the massive black holes and in general in compact objects. The 20 years search for high energy emission from Sgr A* (Watson et al. 198 1, Predehl & Triimper 1994, Goldwurm et al. 1994) has recently come to a turning point with the remarkable observations made * Corresponding author: e-mail: [email protected],Phone: +3301 6908 2792, Fax: +33 01 6908 6577 @ 2001 WILEY-VCH Verlag GmbH & Ca. KGaA. Weinhem

A. Goldwurm et al.: Sgr A* X-Rav flare seen with XMM-Newton

378

with the ChandraX-ray Observatory in 1999 and in 2000. Baganoff et al. (2001a) first reported the detailed 0.5" resolution images obtained with Chandra in the range 0.5-7 keV, which allowed the detection of weak X-ray emission from the radio source. The derived luminosity in the 2-10 keV band was 2 erg scl, for a distance of 8 kpc with steep power law spectrum (index of 2.7) and some evidence that the source is in part extended on a I" scale. Then, in October 2000, the same source was seen to flare up by a factor of = 45 in a few hours (Baganoff et al. 2001 b). The luminosity increased from the quiescent level measured in 1999 to a value of erg s-'. The flare lasted a total of 10 ks but the shortest variation took place in about 600 s, implying activity on length scales of M 20 Rs, for the above quoted mass of the galactic center black hole. Evidence of spectral hardening during the flare was also reported by the authors who determined a source power law photon index during the event of 1.3 (k0.55). significantly flatter than observed during the quiescent state. These results constrain models of the accretion flow and radiation mechanism for Sgr A*. XMM-Newton, the other large X-ray observatory presently in operation, features three large area Xray telescopes coupled to three CCD photon imaging cameras (EPIC) operating in the 0.1-15 keV range and to two reflection grating spectrometers (RGS) working in the 0.1-2.5 keV band (Jansen et al. 2001). Although its angular resolution (6" FWHM) is insufficient for properly resolving Sgr A* in quiescence, an intense flare such as the one seen by Chandra can be easily detected and studied with XMM-Newton. We report here (see also Goldwurm et al. 2003) the detection of such en event during a 26 ks XMM-Newton observation of the Galactic nucleus performed on 2001 September 4 as part of a large Galactic Center survey program with XMM-Newton (Warwick et al. 2003).

2 Results The EPIC data reduction and analysis of this XMM-Newton observation are described in detail in Goldwurm et al. (2003). The image recorded in the central CCD (1 1' x 1 1' for the MOS) was dominated by the diffuse emission of the Sgr A East region, and in order to search for a variable central source we extracted and analyzed light curves from events collected within a 10" radius region centered on Sgr A*. As shown in Fig. 1, the 2-10 keV count rate from the combined MOS 1 and MOS 2 events selected in this way, is quite stable around an average value of 0.08 cts scl till the last 900 s of the observation. Then the count rate gradually increases to reach a value of about a factor 3 higher in the last bin. The integrated count rate in the last 900 s reaches 7 over the average value measured before the flare and the detected variation has a negligible probability to be a statistical fluctuation. A similar peak (4.3 (T)is observed in the counts extracted from the PN camera which stopped observing about 250 s before the MOS (see Fig. 1 b). Similar light curves, from a larger region far from the source do not show any evidence of such an increase in the counting rate. In Fig. 2 we report a MOS image of the region around the nucleus integrated during the 1000 s before the flare and a similar image integrated during the last 1000 s and fully including the source flare. The brightening we detected in the light curves is clearly due to the brightening of a central source. We compared the data to the instrument point spread function to determine the location of the excess. On the 2-10 keV MOS 1 and MOS 2 image of the last 1000 s, rebinned to have pixel size of 4", we obtained the centroid of the source at R.A. (52000) = 17h 4.Y 39.99' Dec (J2000) = -29" 00' 26.7", with a total error, dominated by residual systematic uncertainties in the XMM-Newton focal plane, of 1S".The derived flare position is therefore compatible with the Sgr A* radio location (Yusef-Zadeh et al. 1999), since it is offset from the latter by only 1.5" and we conclude that the flare detected by XMM-Newton is associated with the galactic nucleus. A first, spectral analysis of the flaring event was performed by computing a simple hardness ratio, defined as the ratio between the measured counts (including background) in the hard band 4.5- 10 keV, and those in the soft band 2-4.5 keV. We found that the hardness ratio increased by 0.32 & 0.13 during the event with respect to the value before the flare. Though the hardening has a modest statistical significance

Astron. Nachr./AN 324. No. S1 (2003)

379

of 2.5 o,it is fairly consistent with the flare trend observed with Chandra. Since the Chandra data showed that the quiescent emission within 10" from Sgr A"s position contains a dominant diffuse emission and a large contribution from close point sources, we had to model in some way these components to study the flare spectrum (see Goldwurm et al. 2003). We extracted MOS and PN count spectra from the 10" radius circular region centered on Sgr A* before the flare and during the flare (last 900 s for MOS and last 700 s for PN). To derive the spectra reported in Fig. 3 we used the spectra extracted before the flare as background components for the flare spectra. After subtraction of the non-flaring count spectrum, the flaring MOS and PN spectra were fitted with a simple absorbed power law with NH fixed to the Chandra measured value of 9.8 10" cm-2 and leaving the MOS and PN normalizations free to vary. Results both without and including dust scattering are reported in Table I . We obtained a best fit photon index of 0.7 f 0.5 (error at 1 D for one interesting parameter) with x; = 0.98 for 20 d.o.f., that is significantly harder than the spectrum measured with Chandra during the quiescent state (2.7 f l.O), and compatible, within uncertainties, to the index measured during the 2000 October flare. This procedure subtracts from the flare spectrum the non flaring component of Sgr A* and therefore assumes that the quiescent emission from Sgr A* is negligible, This is an acceptable approximation since, if at the level measured in 1999 by Chandra, the quiescent emission contributes by only M 5% to the counts of the flare spectrum. On the other hand, this procedure allows to subtract the diffuse emission present in the region of the spectral extraction and the instrumental background in a model-independent way. We remark that the count excess around 6-7 keV in the residuals of the MOS spectrum of Fig. 3 is not significant. Including a narrow gaussian line at 6.4 keV (with fixed centroid and zero width) to the model of the absorbed power law, we can set an upper limit (90 % confidence level in 1 parameter) to an iron emission line of about 1.8 keV equivalent width. The measured absorbed source flux in the 2- 10 keV band, corrected for the fraction of encircled energy at a distance of 10" (60%),is then of (3.3 f 0.6) 10-l' ergs cmp2 spl (1 F errors by fixing best fit erg s-'. This parameters but normalization), equivalent to a 2-10 keV luminosity at 8 kpc of M 4 is the average value in the last 900 s but the last light curve 180 s bin was about a factor 1.4 higher, thus erg s-l. These numbers are subject to large errors due to the the luminosity reached a value of M 6 low statistics available. But the general result which emerges is that the flare we detected presents a harder spectrum than the one measured with Chandra for Sgr A* during the quiescent period.

030.'

"

"

I . '

'

"

"

'

'

'

" '

" '

'

'

'

"

".

0 30

-~ < : 2

020-

0

.

-< 2

0 25 020

2cc 0 1 5 I

5

010

U 0

0 05

000-.. 0

.

.

,

.

.

I

. ,

.......

1x10~

I

. . . . . . , . . ~

2x10~

0 00

3x10~

Time [s]

2 1 x104

2 3x104

2 5x104

2 7x104

Time [s]

a)

b)

Fig. 1 a) Count rate, sampled in bins of 180 s, collected with both MOS cameras from a region within 10" from Sgr A* in the range 2-10 keV (black upper curve). An equivalent light curve collected from a 30" radius region centered about I' East of Sgr A* and rescaled by a factor 0.1 for clarity, is shown for comparison (red lower curve). Dashed lines indicate the average value computed before the flare. b) Zoom of the Sgr A' MOS light curve (black circles) around the period of the flare compared to a similar light curve (count rate within 10" from Sgr A* in the 2-10 keV band in bins of 180 s) from PN data (red crosses)

A. Goldwurm et al.: Sgr A * X-Ray flare seen with XMM-Newton

380

Fig. 2 Images of the 5' x 5' region around the Galactic nucleus in the band 2-10 keV obtained from MOS events integrated in the lo00 s before the flare (a) and in the last 1000 s of the observation including the flare (b).Pixels were rebinned to a size of 5.5'' x 5.5". Sgr A*'s position is right in the middle of the central bright pixel visible in the flare image (b).

3 Discussion The XMM-Newton discovery of a new X-ray flare of Sgr A* in September 2001 confirms the results obtained in the earlier Chandra observations. XMM-Newton observed only the first part of the flare, but the recorded event is fully compatible in intensity, spectrum and time scales with the early phase of the flare seen by Chandra. This detection of another such large X-ray flare from Sgr A* indicates that the event is not rare. In spite of the lack of detection of flares from Sgr A* in another 50 ks XMM-Newton observation performed in February 2002 (Predehl et al. 2003), the daily presence of X-ray flares in Sgr A* has been recently confirmed by further Chandra observations performed in 2002 (Baganoff et al. 2003). The radio source on the other hand has been observed many times and the detected flux variability has never exceeded a factor 2 (Zhao et al. 2001). This implies that it is unlikely that radio emission presents a comparable large increase in flux and this provides some constraints on the models. The X-ray flare from Sgr A* cannot be explained by pure Bondi or ADAF models (Narayan et al. 1998) as in these models the emission is due to thermal bremsstrahlung from the whole accretion flow and arises from an extended region (between lo3 - lo5 Rs) which cannot account for such rapid variability. Models which predict emission from the innermost regions near the black hole involve a mechanism acting either at the base of a jet of relativistic particles (Markoff et al. 2001) or in the hot Keplerian flow present within the circularization radius of a spherical flow (Melia et al. 2001, Liu & Melia 2002). In both cases a magnetic field is present in the flow and the linearly polarized sub-mm radiation is attributed to optically thin synchrotron emission from the inner region, while the X-rays during quiescent period are produced by the synchrotron self-Compton (SSC) mechanism whereby radio to mm photons are boosted to X-ray energies by the same relativistic or subrelativistic electrons that are producing the synchrotron radiation. However large X-ray flux variations produced by a change in accretion rate in these models would imply an equivalent increase in the radio and sub-mm part of the spectrum not compatible with the lower amplitude of radio changes compared to X-rays (Markoff et al. 2001). Not to mention that the X-ray spectrum would remain rather steep. The model of a inner circularized flow however predicts low or anti correlation of the radio emission with the X-rays if the radiation mechanism for the X-ray flare is bremsstrahlung rather than SSC. The sub-mm and far IR domain on the other hand would in this

Astron. Nachr./AN 324.No. S 1 (2003)

381

case show a large correlated increase, but at these frequencies the measurements have not been frequent enough to settle the issue. Though the exact modelling of radiation process depends on viscosity behavior and other uncertain details, the observed hardening of the spectrum during the flare indeed favours the bremsstrahlung emission mechanism in this model rather than the SSC one (Liu & Melia 2002). Another totally different model for the X-ray flares (Nayakshin & Sunyaev 2003) proposes that those result from the interaction of close orbiting stars with a very cold neutral accretion disk around Sgr A*. More compelling constraints on the models will be set when simultaneous observations in radiokub-mm and X-ray wavelengths of such a flare are obtained. We compared the time of the flare to a recent radio light curve of Sgr A* obtained at 1.3 cm and 2 cm with the VLA between 2001 March and November (Yuan & Zhao 2003). The X-Ray flare occurred 1-2 days after a local maximum of the curve, but no radio data points are reported for the day when our XMM-Newton observation took place. It will b e also important to study the shape of the flare spectrum at energies higher than 10 keV to fully understand the radiation mechanism producing the X-rays. In particular by measuring the high energy cut-off of the spectrum one could determine the electron temperature for a thermal emission or the Lorentz factor for non-thermal processes. We estimated that such a flare should be marginally visible in the range 10-60keV with the low energy instruments on board the new gamma-ray mission INTEGRAL,, operating since 2002 October, if the spectrum extends to these energies with the slope observed with Chandra and XMM-Newton. Acknowledgements This paper is based on observations with XMM-Newton, an ESA science mission with instruments and contributions funded by ESA member states and the USA (NASA). References Baganoff, F., et al., 2001a. ApJ, in press (astro-pW0102151) Baganoff, F., et al., 2001b, Nature, 413, 45 Baganoff, F., et al., 2003, these proceedings Goldwurm, A., et al., 1994, Nature, 371,589 Goldwurm, A., 2001, Proc. of the 4th INTEGRAL Workshop, ESA-SP 459,455 Goldwurm, A., et al., 2003, ApJ, 584,751 Jansen, F., et al., 2001, A&A, 365, L1 Liu, S., & Melia, F., 2002, ApJ, 566, L77 Markoff, S.,et al., 2001, A&A, 379, L13 Melia, F. & Falcke, H., 2001, ARAA, 39, 309 Melia, F., Liu, S., Coker, R. F., 2001, ApJ, 553, 146 Narayan, R., et al., 1998, ApJ, 492, 554 Nayakshin, S. & Sunyaev, R., 2003, MNRAS, submitted (astro-pW0302084) Predehl, P. & Triimper, J., 1994, A&A, 290, L29 Predehl, P., et al., 2003, these proceedings Schodel, R., et al., 2002, Nature, 419, 694 Watson, M.G., et al., 1981, ApJ, 2.50, 142 Warwick, R. S., et al., 2003, these proceedings Yuan, F. & Zhao, J., 2003, Chin. J. Astron. Astrophys., in press (astro-pW0203050) Yusef-Zadeh, F., Choate, D., Cotton, W., 1999, ApJ, 51 8, L33 ZhdO, J., Bower, G. C., GOSS, W. M., 2001, ApJ, 547, L29

382

A. Goldwurm et al.: Sgr A* X-Ray flare seen with XMM-Newton

Table 1 Spectral Fit to X-ray Emission from within 10" from Sgr A* during the Flare

Power-law Model

NH [loz2 cm-']

No Dust Scattering

Dust Scattering

9.8

5.3

Photon Index

0.7

0.3 ?::

Norm MOS [10W4 ph cmP2 sP1keV-']

1.3 ?:

0.7 ?::

Norm PN [lop4 ph cm-' s-' keV-']

0.3 ?:.!

0.2;;;:

x: (d.0.f.) 0.98 (20) Notes : Scattering computed for fixed value of Av = 30. Normalization is the flux density at 1 keV. Errors are at 68.3% confidence interval for 1 interesting parameter.

I

1

I

1

I

I

2

0.95 (20)

5

I

10

Energy (keV) Fig. 3 Count spectra from MOS (black data point set) and PN (red data point set) data, extracted from a region of 10" radius around Sgr A* during the flare after subtraction of the non flaring spectra. The spectra are compared to the best fit model of an absorbed power law without dust scattering (see Table I )

Astron. Nachr./AN 324, No. S1,383 -389 (2003) I DO1 10.1002/asna.200385090

Searching for Structural Variability in Sgr A* Zhi-Qiang Shen*',5,6,M. C. Liang'.*, K. Y. Lo3, and M. Miyoshi4

'

Academia Sinica Institute of Astronomy and Astrophysics, PO Box 23-141, Taipei 106 Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91 125 National Radio Astronomy Observatory, 520 Edgemont Road, Charlottesville, VA 22903 National Astronomical Observatory Japan, Osawa 2-21-1, Mitaka, Tokyo 181-8588 Shanghai Astronomical Observatory, Chinese Academy of Sciences, Shanghai 200030 National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012

Key words Galaxy: center, galaxy: individual (Sagittarius A*), techniques: interferometric Abstract. A model fitting procedure for estimating the parameters of an elliptical Gaussian model that describes the radio emission of Sgr A' observed by the VLBA is presented. By (implicitly) using the amplitude closure relation while fitting the amplitude, the procedure can minimize the calibration errors in millimeter wavelength VLBI measurements, which are crucial to our ongoing effort to search for the structural variability in Sgr A*. The preliminary results from the application to seven-epoch A7 mm VLBA observations seem to show the sign of change to the source apparent structure in at least one epoch over the period of 7 years (1994-2001). Because of the large uncertainties in the determination of minor axis caused mainly by the poor spatial resolution along the north-south direction with the VLBA, these results are suggestive but not conclusive. This demonstrates the necessity of adding the NRAO GBT antenna to the future A7 mm VLBA observations, which can greatly improve the resolution in the north-south by a factor O f 3.

1 Introduction Sgr A*, the extremely compact radio source at the Galactic Center, is the best candidate for a single massive black hole from both the observational results and the theoretical models. Recent significant progress on the observations of stellar dynamics of the Galaxy's central stellar cluster has provided new compelling evidence for the existence of a compact dark mass of 3.0 x 10' Mawithin the vicinity of Sgr A*(Schodel et al. 2003; Ghez et al. 2003). Improvement on the determination of the upper limit to the absolute proper lo5 Mo(Reid et motion of Sgr A* has also placed a stringent constraint on its mass t o be greater than al. 2003). The total flux density variation has been puzzling ever since the discovery of Sgr A* almost three decades ago (Balick & Brown 1974). In 1982, Sgr A* was first reported to be variable at A1 1 and 3.7cm radio wavelengths (Brown & Lo 1982). Since then, there have been a lot of intensive monitoring observations using all the available radio interferometers, such as the Very Large Array (VLA) at A20 to 1.3 c m and 7 mm, the Green Bank Interferometer (GBI) at A1 1 and 3.6 cm, the Nobeyama Millimeter Array (NMA) at A3 and 2 m m and the Sub-Millimeter Array (SMA) at X1.3 and 0.87mm. As a result, the total flux density variation has been seen at all the observing wavelengths from centimeter to submillimeter on all time scales from years to days. It was also found that the variation appears to peak at shorter wavelength first with a relatively larger amplitude fluctuation. Recent analysis of radio light curves seems to suggest a quasi-periodic oscillation, or a double quasi-periodic oscillation (Zhao et al. 2003). N

N

* Corresponding author: e-mail: [email protected].!w, Phone: +886-2-3365-2200, Fax:+886-2-2367-7849

@ 2003 WILEY-VCH Verlag GmbH & Co KGaA, Weinhem

Z . Q. Shen et al.: Structrual Variability in S g r A*

3 84

The discovery of a very strong, highly variable and short-lived X-ray flare from Sgr A* by the Chan~ Ma dark masses into a region less than 20 times its dra X-ray Observatory (CXO) places the 3 . 0 lo6 Schwarzschild radius (Baganoff et al. 2001). Intriguingly, 10 days after the X-ray flare, a relatively low amplitude variation was detected by the VLA (Zhao 2002). Nowadays, there is no doubt that Sgr A* is a temporally variable source. The observed intensity variations at both the radio and X-ray bands might be correlated and both might arise from the instabilities in an accretion disk. Thus, it would seem likely that the variability in the flux density of Sgr A* would be accompanied by structural changes in Sgr A*, i.e., Sgr A* should be variable spatially as well.

1 2h

1qh

16'

1gh

20h

- VLBLPT Fig. 1 Uptime plot showing the elevation angle of Sgr A* at different time (in local sidereal time (LST) at PT) seen at the 7 VLBA antenna sites (labelled),from which the fringes are consistently detected. It is clear that the VLBA always looks at Sgr A* at a low elevation, which severely limits the resolution along the north-south direction. LST

Very Long Baseline Interferometry (VLBI) technique is proved to be the most effective and powerful tool for investigating the high-resolution structure of compact objects (Kellermann & Moran 2002). Attempts to measure Sgr A* structure with the VLBI observations, however, have suffered from the angular broadening caused by the diffractive scattering by the turbulent ionized interstellar medium, which dominates the resultant images with a A2-dependence apparent size (Lo et al. 1997). Over the past decade, VLBI experiments have been carried out steadily at millimeter wavelengths (A7 and 3.5mm). Due to the southerly declination of Sgr A* (- - 30"), and the high northern latitudes for most of the existing millimeter VLBI antennas, much of the observational data were taken at low elevation angles where atmospheric effects are substantial (see Fig. I). This fact imposes two limits on the VLBI study of Sgr A*. One is related to the spatial resolution. It would limit the (u,v) sampling in the north-south direction, which happens to be along the minor axis of the scattering structure. Consequently, the spatial resolution in the north-south direction is always inadequate as compared to the scattering size. The other effect is upon the data calibration. The atmospheric absorption due primarily to spectral line transitions of water vapor and oxygen at millimeter wavelengths increases with decreasing elevation angle (larger opacity at lower elevation). The compromised sensitivity or lower signal to noise ratio (SNR), when combined with the short and variable coherence time at millimeter wavelength, results in large calibration uncertainties. As a result, any possible variation in the observed source structure has so far been ascribed to the errors in the calibration of the VLBI data. In order to minimize the calibration errors and thus to improve the accuracy of the measurements, we have developed a model fitting procedure by using the amplitude closure relation. The preliminary results from the application of this method to the existing A7 mm VLBA observations seem to show the sign of

385

Astron. Nachr./AN 324, No. S 1 (2003)

change to the source apparent structure in at least one of seven epochs. However, the significance level of such a change is seriously limited by the poor spatial resolution along the north-south direction with the current VLBA. The situation can be improved by adding the NRAO GBT antenna to the future VLBA observations at X7 mm.

2 Description of the algorithm To improve upon the calibration of current VLBI observations of Sgr A*, we adopted and implemented a model fitting procedure in which the amplitude closure relation is applied. By measuring the closure amplitudes, the knowledge of the absolute flux density is lost, as the absolute position in the case of the closure phases.

2.1

Two basic equations

In general, by ignoring the baseline-dependent effects (in complex gain and noise), the relationship between the observed visibility TbS ( t )and the true source visibility qyueon a baseline i j can be expressed as

where

G, ( t ) G3 , ( t ) is the time-variable complex gain of antenna i and j, respectively, qzJ( t ) is a stochastic thermal noise having a zero mean and variance cz". The best-fit model of source structure can be obtained by minimizing x2, the weighted sum of square4 of the residuals qz3(t),between the observed data l$bs(t) and the predicted data G,(t)G,*(t)V,Y'(t), for all available baselines throughout the whole observation, defined as below,

where ui,,~(t) is the weighting coefficient which is chosen to be the inverse of the

variance of the data due to noise, i.e., 'UIQ= -&, gv

yy"d(t)is the visibility from the source structure model representation. These equations are similar to those widely used in the self-calibration technique for VLBI imaging (Pearson & Readhead 1984). Both are, by definition, consistent with the closure quantities (closure phase and closure amplitude). Both procedures first use the same method to determine the complex antenna gains as a function of time. The difference between two methods lies in the way of obtaining the model of the source sky brightness distribution. Both algorithms should converge to the consistent results with the high enough SNR data. The conventional imaging process has been dominated by the CLEAN deconvolution method, whose biggest drawback is non-uniqueness in the final image, especially when the SNR is poor. The model fitting algorithm, however, can easily search over all the possible models. This could be very effective in fitting to a model that has fewer free parameters as is the case of Sgr A*, whose radio emission, as a first order approximation, can be well represented by an elliptical Gaussian model (3 parameters only). This is because most of the closure phases measured by VLBI observations of Sgr A* are consistent with zero (e.g. Doeleman et al. 2001). Furthermore, the fact that the scattering image of X2-dependent size is always resolved out on the short to intermediate, depending on the observing wavelength, baselines of the VLBA means that we deal with a lot of weak detections ( S N R decreases with increasing projected baseline length).

Z . Q. Shen et al.: Structrual Variability in Sgr A*

386

2.2 Model visibility amplitude From now onwards, our discussion will be restricted to a single Gaussian model whose brightness distribution is symmetric with zero visibility phase. Assuming it is elongated along a position angle (East of North) in degrees with an axial ratio CY and a major axis size of 0 in radians, its normalized visibility amplitude V;jnod also has a Gaussian distribution as

TI)distance of baseline i j in wavelengths, where pij is (u,

paj = 4.2

(ual~

0 - vlJ s szria)2 ~

+ (u,,s z n +~ v,,

c,05+)2

14)

~

here, uzj and ubJ are the East and North components, respectively, of the projected baseline vector seen from the source, in the units of wavelengths.

2.3 The determination of the time-variable antenna-dependent gains As mentioned above, exactly the same self-calibration algorithm is used to solve for the time-variable antenna-dependent complex gains. This usually means an iterative (non-linear) process in order to find gains by minimizing the weighted sum of squares of the residuals v,(t), between the observed data L$‘(t) and the predicted data G,(t)G;(t)Kyod(t), for the available baselines within each integration period instead of the whole observation. For Sgr A* with a symmetrical Gaussian model (zero phase), the antenna-dependent gains should be real and therefore can be expressed as G,(t) = e g z ( t f . By doing so, we can rewrite Eq. ( I ) as a linear function of g2( t )

= +)VV”,’”p‘ ’ ((tt)) N variance ( ~ 2 ’ ~ = ) ~ (+)’v,:. ’ ( t )

where X,,(t)

e, VOb9(t)

and a;,(t)

=

Oa3

1 = _ _ with a corresponding 7i2, y;”(t)

y y ” d ( t ) X“,(t) -

Thus, instead of solving the non-linear least squares problem, we can

easily solve the normal equation (Eq. ( 5 ) )for gi(t) (and gain G,(t))for Sgr A* using the standard matrixinversion method. In matrix notation, regarding time dependence as implicit we have

9 =

( A ~ ~ -lATw’ A ) Y,

where “ + I ” refers to the inverse matrix operator, T to the transpose operator, and g = [gl, 9 2 , ..., g7nlTx,

(m: the number of antennas),

Y = 1nX = [InXlz, InXls, lnXz3,

A is a steering matrix 1 1 0 0 ... 0 0 0 0 1 0 1 0 ... 0 0 0 0

A=

. . . . 0 0 0 0 ... 0 1 0 1 2

x m

387

Astron. Nachr./AN 324. No. S I (2003)

w' is the diagonal matrix of weighting functions as

2.4 Measured visibility amplitude The outputs from the VLBI correlator are complex ones that are integrated over To (typically, 2 seconds for continuum observations of Sgr A*). After global fringe fitting, these are first coherently averaged over T, (10-20 seconds) to obtain a mean visibility amplitude Z k . The baseline denotation i j is now regarded as implicit, for simplicity. Unlike the original complex correlator output, the measured visibility amplitude will have a positive bias with respect to the true visibility amplitude (Thompson, Moran & Swenson 1986). The magnitude of such a bias is a function of SNR with the strongest bias occurring at low SNR. Therefore, in order to use the amplitudes alone to fit a symmetrical Gaussian model to Sgr A* data which have a lot of low SNR ( 5 3 ) detections on long baselines, we must correct for this bias first. Two methods, each involving an incoherent average, were adopted to carry out such a bias-correction, Method one is to, for each observing scan T, (typically, 6-10 minutes) that consists of N(
< z2> =

Ivy + 2

2;

(7)

< z4> = IV14 + 81V122 + 8cr4, where (T is the rms noise level in 21, and assumed to be constant within T,, and < Z 2 > are the arithmetic mean of 2;and Zt,respectively, < Z 2 >= Z; and < Z4 >= Thus, we can obtain

cf,

J V I = [2 < z2>2 - < 2

4

>ill4 ;

(8) and < Z4 >

& CF='=, 2:. (9)

Method two is to solve for an unbiased estimate of the amplitude directly by doing incoherent average within each scan T, (Rogers, Doeleman & Moran 1995; Doeleman et al. 2001)

r.

N

with the variance in the bias corrected amplitude as

here CQ is the noise in 2,which is a constant (inversely proportional to the square root of the product of T, and the bandwidth) in the unit of the correlation coefficient and, the corresponding SNR value sk is defined as 2,/uI,.

2. Q. Shen et al.: Structrual Variability in Sgr A*

388

2.5

The practice of model fitting and error estimation

In practice, only the visibility amplitude information is used for a fitting to the VLBI observations of Sgr A*, i.e., both measured and model visibility vectors are replaced by their amplitudes in solving for the least squares minimization Eq. (2). To find a best fit model that corresponds to the minimal x:~,, a thorough search for x2 in a 3-dimensional space that covers a wide combination of three model parameters (major and minor axes and the position angle) is carried out. In principle, this procedure is applicable to any source whose brightness distribution can be represented by a single Gaussian. should be distributed For the error estimation, we first draw attention to the presumption that &%, as a chi-square distribution with N,l,f degrees of freedom to ensure a credible fit. By definition, this is equivalent to the statement that the best fit model should have reduced chi squares, x ~ = x ~ , , / N d ,equal f, However, in many cases the value for xz at is significantly to unity with a deviation of larger than 1.0 (c.f. Bietenholz et al. 1996). So, we scale up the 68.3% confidence region of parameter + Ax2 with space, as an increase of x2 from xk,, to xLLrL

d m .

xk,,

instead of Ax2= 1. Here, Nd,f is estimated by the summation of the difference between the number of visibilities Nu,, and the number of antennas Nan, over all scans, minus the number of fitting parameters (3 in case of Gaussian model), i.e., N d , f = C,(N,i, - N,,t) - 3. By projecting this confidence contour onto the axis of parameter of interest, we can finally obtain the 1 standard error for that single parameter.

3 Application to the A7 mm VLBA observations of Sgr A* We have applied this model fitting method to seven existing A7 mm VLBA observations of Sgr A* over a period of 7 years (1994 - 2001). The preliminary results show that the apparent source size is dominated by the scattering effect with a position angle (PA) -75". The major axis can be well determined with an error of 0.01 mas, while the minor axis has a quite large error 20.05 mas. Of 7 epochs, however, six epochs showed that the fitted sizes of both major and minor axes are all above the inferred scattering size. On May 31, 1999, a consistent increase in the apparent source size over the scattering size along the minor axis can be seen at all three simultaneously observed A7 mm bands of 39, 43 and 45 GHz with the corresponding significance levels of 3u, 2u and 2a, respectively (see Fig. 2). At another epoch (July 3 1,2001) which was just three weeks after an SMA flare towards Sgr A* (Zhao 2002), the source seems to undergo a structural variation too (Miyoshi et al. 2003), but the uncertainty of 0.12mas is just too large. At this moment, we consider these results to be suggestive.

4 Discussion For two epochs in April and May 1999 when the time-dependent relative gains as a function of elevation could be determined through a least-squares fit of the auto-correlation spectra of the SiO maser towards VX Sgr at different elevation to a well calibrated total power spectrum, we also tried self-calibration imaging and found the consistent results with those from the model fitting. This demonstrates that in principle the difficulties due to the atmosphere can be minimized either by a careful calibration using interleaved maser line observations or by model-fitting using the closure quantities. The resolutions which can be reliably achieved by the VLBA observations of Sgr A* is about I .3 mas by 0.5 mas along PA=10" at A7 mm, compared to the scattering size of 0.37mas by 0.69 mas at PA=8Oo. Obviously, the resolution along the north-south (1.3 mas) was too poor to resolve the scattering size of 0.37 mas. This can explain why the major axis of Sgr A* at a PA-75" can be determined fairly well while the minor axis has a larger uncertainty. Tests on the simulated data also confirm that the lack of

389

Astron. Nachr./AN 324. No. SI (2003)

\

observing frequency u ( G H r )

Fig. 2 The plot of the measured Sgr A* angular size versus the observing frequency. The open circles and filled circles are the sizes of major and minor axes measured at three simultaneously observed A7 mm bands of 39, 43 and 45 GHz by the VLBA on May 31, 1999. The solid line and dashed line represent the scattering dominated sizes as a function of the observing frequency along the major and minor axes with a frequency dependency of 1287/u$~,and 684/uKH, mas (Lo et al. 1998). A consistent deviation in the apparent source size from the minor scattering size at all three bands can be seen.

resolution is an insurmountable barrier to the current VLBA in the determination of minor axis with a good enough accuracy. We have proposed to add the recently commissioned NRAO GBT antenna to the future A7 mm observations. This will improve the resolution along the north-south direction by a factor of 3 by connecting to an isolated, north-south oriented HN-SC baseline. Strong signals have been consistently detected on this single baseline in five of seven epochs data sets (Shen et al. 2003, in preparation). However, the non-detection between either of SC and HN with the rest of the VLBA handicapped the uses of SC-HN detection in the self-calibration process. By introducing the GBT antenna that is available at A7 mm, it is certain that these SC-HN detections can be brought back as the GBT will serve as a “bridge” between SCHN and those western VLBA antennas (e.g., NL/FD).

References Baganoff, F. K., et al. 2001, Nature 413,45 Balick, B, & Brown, R. L. 1974, ApJ 194,265 Bietenholz, M. F., et al. 1996, ApJ 457, 604 Brown, R. L., & Lo, K. Y. 1982, ApJ 253, 108 Doeleman, S. S., et al. 2001, AJ 121, 2610 Ghez, A. M. et al. 2003, these proceedings Kellennann, K. I., & M o m , J. M. 2001, ARAA, 39, 457 Lo, K. Y., et al. 1998, 508, L61 Miyoshi, M., et al. 2003, these proceedings Pearson, T. J., & Readhead, A. C. S. 1984, ARAA, 22,97 Reid, M. J., et al. 2003, these proceedings Rogers, A. E. E., Doelernan, S. S., & Moran, J. M. 1995, AJ 109, 1391 Schodel R., et al. 2002, Nature 419, 694 Thompson, A. R., Moran, J. M., & Swenson, G. W., Jr. 1986, Interferometry and Synthesis in Radio Astronomy. New York: Wiley-Interscience. First (1991) and second (1994) reprintings by Krieger Pub. Co., Malabar (Florida) Zhao, J.-H. 2002, in GCNEWS, Vol. 15,4 Zhao, J.-H., et al. 2003, these proceedings

This Page Intentionally Left Blank

Astron. NachrJAN 324, No. S1,391-395 (2003) / DO1 10.1002/asna.200385113

Observations of the Galactic Centre at 610 MHz with the GMRT Subhashis Roy * and A. Pramesh Rao ' National Centre for Radio Astrophysics (TIFR), Pune University Campus, Post Bag No.3, Ganeshkhind, Pune 41 1 007, India. Abstract. We have observed the central region of the Galaxy at 610 MHz with the Giant Metrewave Radio Telescope (GMRT). We detect emission from Sgr A*, the compact object at the dynamical centre of the Galaxy, and estimate its flux density at 610 MHz to be 0.52~0.1Jy. This is the lowest frequency at which Sgr A* has been detected. Comparison of the 610 MHz and 1.4 GHz map of this region indicates that most parts of the Sgr-A West HII region are optically thick at 610 MHz, having optical depth of 2. However, Sgr A*, which is seen in the same region in projection, shows a flat spectral index between 1.4 GHz and 610 MHz. This is consistent with its high frequency spectral index, which indicates that Sgr A* is located in front of the Sgr-A West complex. An intrinsic turnover in its spectrum between 580 and 330 MHz is likely to be the cause of its non detection at 408 and 330 MHz. N

1 Introduction The Galactic Centre (GC) region has been observed at radio wavelengths with high resolution by the Very Large Array (VLA) at 2 cm (Yusef-Zadeh & Wardle 1993) 6 cm (Ekers et a]. 1983), 20 cm (Pedlar et al. 1989) and 90 cm (Pedlar et al. 1989; Anantharamaiah et al. 1991;LaRosa et al. 2000) and several sources were identified within the central half a degree region of the GC. At the dynamical centre of the Galaxy ~ black hole candidate (Ghez et al. 1998), which coincides with a compact nonthermal is the 2 . 6 ~ 1 0Ma radio source named Sgr A*. Around this point source in a somewhat larger scale in projection, are the three arm spiral configuration of ionised gas and dust known as Sgr-A West (Ekers et al. 1983). Near Sgr-A West, is the supernova remnant Sgr-A East. A 7' halo which is thought to be a mixture of thermal and non-thermal emission (Pedlar et al. 1989) can also be identified in this region. Sgr A* has attracted considerable attention from the time of its discovery (Balick & Brown 1974) since it is associated with the nearest supermassive black hole and could be a prototype for such black holes in extra galactic AGNs. This object has now been studied from radio to the X ray ranges (see Melia & Falcke 2001, and references therein). Though the observational data strongly associates it with the 2 . 6 lo6 ~ Ma black hole at the centre (Genzel et al. 1996; Ghez et al. 1998; Reid et al. 1999), there are several questions that remain unanswered. CornpaTed to AGNs, this object is extremely underluminous at all wavelengths, radiating at x lo-'' times of its Eddington luminosity. It is known to vary at higher radio frequencies, and the flux density variations appears to have a periodicity of 106 days (Zhao et al. 2001). Though, no linear polarisation has been detected at radio frequencies, circular polarisation from this object has been detected (Bower et al. 1999). Sgr A* has not been detected below 1 GHz and observations by Davies et al. 1976 at 408 MHz and Pedlar et al. 1989 at 330 MHz provide upper limits on its flux density. It could have a low frequency turnover below 1 GHz, but the nature of the turnover has never been clarified in detail (Melia & Falcke 200 1 ). In order to estimate the spectrum of the Sgr A* at low radio frequencies, we observed it using the Giant Metrewave Radio Telescope (GMRT) at 620 MHz in Aug & Sep 2001 and at 580 MHz in June 2002. In * Subhashis Roy: e-mail: [email protected]

@ 2003 WREY-VCH Verlag GmbH & Co. KGaA. Wemheim

S. Roy and P. Rao: Galactic Centre at 610 MHz

392

this band, the free-free optical depth of the Sgr A West is expected to be moderate, and it should be possible to identify Sgr A* in the Sgr A West complex. At these frequencies, the GMRT has an angular resolution of about 6'' and field of view 44'. The details of these observations and the data analysis will be described elsewhere (Roy & Rao 2003, in preparation).

2 Results

Fig. 1 The central 15' region of the Galaxy at 610 MHz as observed by the GMRT. The resolution of the image is 1I .4"x 7.6", with the beam position angle of 7'. The RMS noise is about 6.5 m l y beam-'.

Fig. 2 4.8 GHz continuum map of the Sgr A complex (Yusef-Zadeh 1989) in contour overlaid on the 610 MHz gray scale map of the the same region.

2.1 Features in the 610 MHz map The 610 MHz map of the central 15' region of the Galaxy is shown in Fig. 1. The compact source Sgr A* is clearly seen along with other well known sources like Sgr A West, Sgr A East and the 7' halo. The prominent non-thermal filamentary structure, the Radio Arc is clearly seen in a map covering a larger field. An emission feature =30" south of Sgr A* can be seen in the 610 MHz gray scale map. This feature was identified by Pedlar et al. (1989), who suggested that it is associated with Sgr A East. In order to compare the smaller scale features near Sgr A West with what is seen at higher frequency, we have plotted in Fig. 2 the 4.8 GHz VLA map of this region (Yusef-Zadeh 1989) on the gray scale GMRT image with both maps convolved to the same resolution. Sgr A* is clearly seen in both the maps. There is almost one to one correspondence between the higher emission features at 4.8 GHz comprising the Sgr A West region and the relative lower emission features (indicated by white region in the gray scale map) seen at 610 MHz.

393

Astron. Ndchr./AN 324, No. S1 (2003)

BOTW SGRA lP?L 1477.400 MU2

,

I

I

F189"BrrY IHZi

Fig. density ratio Of the map to the GMRT 6 10 MHz map

GHz VLA GC

Fig. 4 The spectrum of Sgr A* from 300 MHz to 20 GHz. Except the 330 Pedlar et al. (1989) and 610 MHz measurements, all the other data points are taken from Melia & Falcke (2001)

2.2 Optical depth of the ionised gas near Sgr A West In order to study the free-free absorption by the ionised gas, we have divided the 1.4 GHz map (made from the archival VLA data acquired and presented by Pedlar et al. 1989) by the 610 MHz map (both the maps were made with same resolution), which is plotted in Fig. 3. In most parts of Fig. 3, this ratio is < 1 confirming that the emission from the 7' halo and Sgr A East is non thermal and probably due to synchrotron emission. However, in the central region comprising Sgr A West and the 3 HI1 regions near the eastern boundary of Sgr A East the ratio is -2 indicative of free free absorption. In the vicinity of Sgr A*, the radio emission has contributions from a non-thermal component mainly due to Sgr A East and a thermal component due to emission from Sgr A West. Since Sgr A West can absorb the background emission as well as its own emission, separating the two components are essential to estimate the optical depth of the free-free absorption. If the 610 MHz emission towards Sgr A West is dominated by thermal emission, then the optical depth of thermal emission can be estimated from the formula T b = Te.[l- eP7]where T b is the brightness temperature, T, is the electron temperature and 7 is the optical depth. Since the flux density at 1.4 GHz near Sgr A* is a factor of two higher than at 610 MHz, the estimated optical depth in the case of only free-free self absorption is ~ 2To. estimate the non-thermal emission at 610 MHz, we use the GC maps made at 3.6 (Roberts & Goss 1993) and 6 cm (Yusef-Zadeh 1989). Emission from the GC region at wavelengths equal to or less than 6 cm is mostly thermal and is believed to be optically thin. Therefore, based on the 3.6 cm map, we constructed a model for the thermal emission from Sgr A West at 6 cm and subtracted it from the 6 cm map. The difference map provides the excess non-thermal emission at 6 cm as compared to 3.6 cm. With an assumed spectral index of - 1.0 for the non-thermal emission, we estimated the non-thermal component at 610 MHz. After considering this non-thermal component along with the thermal emission from Sgr A West, we estimate an optical depth of 2.6 d10.5 for the ionised gas seen towards Sgr A*.

394

S . Roy and P. Rao: Galactic Centre at 610 MHz

2.3

Estimated flux density of Sgr A*

To estimate the flux density of Sgr A*, we partially resolved out the extended emission around it by applying a shorter uv cutoff of 7 ICX while imaging. The flux density estimated from the image plane is about 0.5 10.1 Jy. We note that in an image of this region even after applying a short uv cutoff, there is significant background confusion in a beam of size 7.5” x 4”.This confusion causes an uncertainty of about 0.1 Jy in the estimated flux density. Therefore, we also estimated the flux densities of this object from the uv plane. The estimated flux density after fitting a elliptical Gaussian model to S g r A* is also 0.5 kO.1 Jy. The respective major and minor axis of the Gaussian fit is 3.8”k0.4” and 1.8”10.6/’ with a position angle of 9 3 f 4 ” . We note that the estimated size of Sgr A* at 610 MHz appears to be consistent with what is expected from scatter broadening (Lo et al. 1998) at this frequency.

3 Discussions 3.1 Low frequency spectral index of Sgr A* While the high radio frequency spectrum of Sgr A* is well established, the spectrum below 1.4 GHz is not well determined. At 1.4 GHz, the flux density of Sgr A* (Zhao et al. 2001) is about 0.5 Jy and its spectral index between 1.4 and 8.5 GHz is +0.17 (Melia & Falcke 2001). Davies et al. (1976) found the flux density of Sgr A* to be a factor of 2 less than at 1.6 GHz and suggested that it has a low frequency turnover around 1 GHz. This appeared to be confirmed from their upper limit of 50 mJy to its flux density at 408 MHz and the 100 mJy upper limit set by Pedlar et al. (1989) at 330 MHz. Our measured flux density of 0.5 Jy at 610 MHz raises questions about the earlier measurements. Measurements of the flux density of Sgr A* between 1.4 GHz and 610 MHz which were close in time show that the spectrum between these frequencies is consistent with that between 1.4 and 8.5 GHz. The spectrum of Sgr A* from 300 MHz to 20 GHz is shown Fig. 4. The 610 MHz observations cover a span of nearly an year and no significant propagation effects due to the Inter Stellar Scintillation (ISS) could be detected. Our measurements rule out any turnover at around 1 GHz and indicate that the turnover has to be at frequencies less than 580 MHz. However, the upper limits at 408 and 325 MHz pose problems for this picture since this would require the spectrum to fall steeper than can be explained by thermal free free absorption or synchrotron self-absorption. We have re-examined the 408 MHz upper limit by Davies et al. (1976) and found that their analysis has not taken account for scatter broadening of Sgr A* (8“ at 408 MHz) and the corrected upper limit at 408 MHz is as high as x 2.5 Jy. The 330 MHz upper limit, however is reliable and implies an inverted spectrum with spectral index > 2.5, suggesting that more than one process could be responsible for the turnover. The nature of the low frequency turnover of Sgr A* could occur due to synchrotron self absorption, internaUexternal free-free absorption or Razin effect. It is of great interest to understand the nature of Sgr A* and its environment, and more careful and systematic flux density measurements below 600 MHz are required. These observations should be spread over many years to eliminate the uncertainty due to interstellar scintillation which could have a large time scale in this direction. 3.2

Location of the Sgr A*

At 610 MHz, Sgr A West shows evidence for free free absorption. In the previous section, we have estimated the optical depth of this ionised gas in the Sgr A West region to be x2.6. If Sgr A* would have been located behind Sgr A West, then its flux density would have been attenuated by a factor of 10. However, the spectral index of Sgr A* between 610 MHz and 1.4 GHz is roughly the same as between 1.4 GHz and 8.5 GHz and shows no effect of the free free absorption by Sgr A West. This indicates that S g r A* is located in front of Sgr A West. It is possible to have alternate scenarios like sharp enhancement of the 610 MHz flux density of Sgr A* to compensate for the absorption due to Sgr A West or a hole in Sgr A West along the line of sight to Sgr A*, but these appear to be unlikely. Sgr A* is located slightly north of the junction of the northern and the eastern arm of the Sgr A West. We have inspected the 3.6 cm map

Astron. Nachr./AN 324, No. SI (2003)

395

(resolution x 0.5”) of the region (Roberts & Goss 1993), and w e do see weak emission likely to be from the diffuse ionised gas which becomes optically thick at 610 MHz. Any hole in this ionised gas has to be smaller than 1” which is unlikely. Thus, Sgr A* is either located in front of Sgr A West, or is embedded within the ionised gas such that the optical depth of that region is 20.1 at 610 MHz.

4 Acknowledgements We thank the staff of the GMRT that made these observations possible. GMRT is run by the National Centre for Radio Astrophysics of the Tata Institute of Fundamental Research.

References Anantharamaiah, K. R., Pedlar, A., Ekers, R. D., & Goss, W. M. 1991, MNRAS, 249,262 Batick, B. & Brown, R. L. 1974, ApJ, 194,265 Bower, G. C., Falcke, H., & Backer, D. C . 1999, ApJL, 523, L29 Davies, R. D., Walsh, D., & Booth, R. S. 1976, MNRAS, 177, 319 Ekers, R. D., van Gorkom, J. H., Schwarz, U. J., & Goss, W. M. 1983, A&A, 122, 143 Genzel, R., Thatte, N., Krabbe, A., Kroker, H., & Tacconi-Garman,L. E. 1996, ApJ, 472, 153 Ghez, A. M., Klein, B. L., Moms, M., & Becklin, E. E. 1998, ApJ, 509, 678 LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, AJ, 119, 207 Lo, K. Y., Shen, 2. Q., Zhao, J. H. & Ho, P. T. P. 1998, ApJ, 508, L61 Melia, F. & Falcke, H. 2001, ARA&A, 39,309 Pedlar, A,, Anantharamaiah, K. R., Ekers, R. D., et al. 1989, ApJ, 342, 769 Reid, M. J., Readhead, A. C. S., Vermeulen, R. C., & Treuhaft, R. N. 1999, ApJ, 524,816 Roberts, D. A. & Goss, W. M. 1993, ApJS, 86, 133 Yusef-Zadeh, F. 1989, in IAU Symp. 136: The Center of the Galaxy, 243 Yusef-Zadeh, F. & Wardle, M. 1993, ApJ, 405,584 Zhao, J., Bower, G. C., & Goss, W. M. 2001, ApJL, 547, L29

This Page Intentionally Left Blank

Astron. NachrJAN 324, No. S1,397-401 (2003) I DO1 10.1002/asna.200385055

Closure Amplitude Analysis of 15,22 and 43 GHz VLBA Observations of Sagittarius A*: Size is Consistent with the Scattering Law G.C. Bower*’ I

Radio Astronomy Lab, UC Berkeley, Berkeley CA 94720

Abstract. High frequency very long baseline interferometry is necessary to avoid the effects of strong interstellar scattering when imaging Sagittarius A*. Unfortunately, the reliability of imaging declines with increasing frequency due to amplitude calibration errors for this low declination source. Closure amplitude analysis provides an unbiased method for estimating source parameters. We analyzed several new and several archival VLBA data sets with the closure amplitude technique. Our results indicate that there is no deviation from the scattering law in the minor axis size at 15, 22 and 43 GHz. There is a slight excess (20 3%) over the scattering law in the major axis size at 43 GHz which can be accounted for by a marginal increase in the error estimate, a slight recalibration of the scattering law or an intrinsic source of size 0.2 mas 2 AU 40R,. The absence of an apparent jet or outflow allows us to set an upper limit to the velocity of any ballistic components at 0.001~. N

N

N

1 The Size of Sagittarius A* Measurement of the size of the compact nonthermal radio source in the Galactic Center, Sagittarius A*, has a long and colorful history (Goss 2003). The discovery that the image of Sgr A* is an ellipse with major and minor axes that scale as X2, indicative of interstellar scattering, has pushed observations to short wavelengths where the scattering effects are weakest. The expectation is that intrinsic source structure will dominate the scattering effects at a short enough wavelength. This will allow us to separate the many different models for the emission of Sgr A* based on size and morphology. Lo et al. (1998) showed simultaneous VLBA observations that followed the scattering law from 6 cm to 7 mm wavelength in the major axis, aligned roughly East-West. On the other hand, the minor axis, aligned roughly North-South, showed an apparent increase above the scattering law at 7 mm. Combined with previous measurements by Bower & Backer (1998), Lo et al. inferred a > 40 detection of an intrinsic source with a size of 72R,. Unfortunately, there are numerous technical problems with making this measurement (Bower et al. 1999), principally related to the difficulty of accurately calibrating amplitudes at these wavelengths and at the low elevation of Sgr A* as viewed from North America. Doeleman et al. (2001) showed that the use of closure amplitude could be used to strongly and convincingly constrain the size of Sgr A* with VLBI observations at 3.4 mm. The closure amplitude is an amplitude gain independent quantity calculated for four stations m.,n,p and 4:

(1) where IV,, I is the amplitude of the visibility on the baseline between stations i and j . The closure amplitude is independent of all station-dependent amplitude errors, such as pointing errors, dish deformation, and variable opacity. It does not eliminate baseline-dependent errors such as variable decorrelation (which also influence conventional calibration and imaging techniques). The most significant loss from closure * Correspondingauthor: e-mail: gbowerOastro.berkeley.edu, Phone: 4 1 510 642 4075, Fax: +01 510642 4075

@ 2003 WILEY-VCH Verlag CmhH & Co KGaA. Weinhem

G. Bower: VLBA Observations of Sgr A*

398

Radio Light Curve and New VLBA Observations 1.6-

T

s1.47

v

%

.#.-

r" 1.2-

B X

2

LL

1-

N

I 00.8 m

*

i

0.6 __

. -

'28idO.5

2001

,

2001.5 2002 Time (years)

. .. . . . .

2002.5

Fig. 1 Light curve of Sgr A* at 43 GHz with VLBA observation dates indicated with red vertical bars. The first three observations were at 43 GHz and 22 GHz, while the later observations were at 43 GHz only.

amplitude analysis is the reduction in the number of degrees over freedom relative to a calibrated data set. The number of independent data points for a 7-station VLBA experiment is reduced by a factor 14/21. But this loss is more than offset by the confidence that the result gives through its accurate handling of calibration errors.

2 VLBA Observations and Analysis A number of new VLBA observations were obtained as a counterpart of the VLA radio flux density monitoring program of Sgr A* (Zhao et al. 2003, Figure 1). In addition to these observations a number of past VLBA experiments at frequencies of 15,22 and 43 GHz were reanalyzed, including those of Lo et al. Data were fringe-fit with AIPS first and then transferred to MATLAB and analyzed with proprietary code. The code forms the closure amplitude from the visibilities, averages the closure amplitudes and uses the scatter in the average to determine the error in the closure amplitude, and then uses a nonlinear fitting method for modeling the closure amplitude. Noise is added to the model closure amplitude, avoiding the problem of unbiasing the measured closure amplitude. A typical set of closure amplitude data points and models are shown in Figure 2. Errors in the model were determined by calculating x2 for a grid of models surrounding the solution and fitting constant x2 surfaces (Figure 3). Closure phases were also computed and analyzed. Non-axisymmetric structure is not reliably detected on any triangle, confirming the hypothesis that Sgr A* consists of only a single axisymmetric component. This is in spite of the range of flux densities and activity states observed.

399

Astron. Nachr./AN 324, No. S1 (2003) BR KP LA OV 2.93 2.78

BR KP LA PT 2.07 1.79

BRKPLAPT 1.65 1.58

,0.8

-

$ 0.6 -

a 0.4 0.2 -

I

--

I

-

,

..

3 Closure Amplitude Results We show our results for 3 22-GHz and 7 43-GHz experiments in Figures 4 and 5. At 22 GHz, our mean result is consistent with the measurement of Lo et a1 and with the scattering size. This demonstrates the accuracy of both methods where calibration is more routine. The results at 43 GHz are more precise than at 22 GHz because of the greater fraction of the array contributing to the result: the source is smaller with respect to the array synthesized beam with increasing frequency. The 43 GHz minor axis size in individual measurements and in the mean result is significantly more compact than the extended source measured by Lo et al. The plot includes the data used by Lo et al. and reanalyzed with our procedures. This indicates that the North-South extension previously reported is not present and that intrinsic structure is << 72&. The 43 GHz result appears to differ at the 20-level from the major axis scattering size extrapolated from longer wavelengths. The result is consistent with a constant axial ratio of 0.53, as expected from longer wavelength observations, but is less consistent with a major axis size of 0.69 mas at 43 GHz. The observed size is larger by a factor of 3% at 43 GHz. We are not yet certain whether this is a real difference. However, this suggests that the scattering size law may be slightly miscalibrated due to amplitude calibration errors at longer wavelengths. Alternatively, this may be the result of actual source structure. If the full difference is due to intrinsic structure, then that structure is 2 AU 40R, at 43 GHz. We see no evidence for outflows from Sgr A* in the closure amplitude results and closure phase results. The source remains constant in size at a level of 0.030 mas in the major axis and 0.100 mas in the minor axis during the apparent flaring state in mid-2001. If the flare created an outflow and given the absence of a bright secondary component, then we can we set an upper limit to the outflow velocity of any N

N

N

N

400

0.6

-

._ g0.5'

v

<

0.4

0

.-

fl

0.3-

G. Bower: VLBA Observations of Sgr A*

P i'D

s 12 ' D l

i

:/

\ \ \

i

',\/1

'

100

fc 0 .-c

a

I"

858075 .

j !

i

,

70 -

90

75 I 0.2

.....

1

...... 0.4

0.6 Axial Ratio

0.8

1

Fig. 3 Three cuts through the x2 surface in major axis, minor axis and position angle space for 68%, 90%, 98%, 99.9% and 99.99% confidence intervals of source parameters for one 43 GHz observation analyzed with the closure amplitude method. Cuts are for the best-fit values of each parameter, which are listed above the plot. The lower right corner shows grid points of 98% confidence interval in axial ratio versus position angle space. ballistic components generated by the flare at 0.01~.Note that the thermal distribution of particles required by the presence of a sharp cutoff in the spectrum at infrared wavelengths suggests that we would not see a propagating component, in any case (Yuan, Markoff & Falcke 2002). Higher quality observations can resolve outstanding observations about the size of Sgr A*. The ultimate VLBA experiment to achieve this is possible with the addition of the GBT at 43 GHz. This will I ) increase signal to noise ratio on many station groups and 2) link Hancock and St. Croix stations to the rest of the array, significantly improving North-South resolution. Sgr A* is already detected on the HN-SC baseline. However, no detections are made from these stations to other elements of the array, making them useless for any imaging techniques, including closure amplitude analysis.

References Bower, G.C. & Backer, D.C., 1998, ApJL, 496,97 Bower, G.C. et al., 1999, in "The Central Parsecs of the Galaxy," H. Falcke et al. eds., ASP Conf. #186,80 Doeleman, S. et al., 2001, AJ, 121, 2610 Goss. W.M.. 2003. these oroceedings Lo, K.Y. et &., 1998, Apk, 508,6i Yuan, F., Markoff, S. & Falcke, H., 2002, ABrAp, 383, 854 Zhao, I.-H. et al., 2003, these proceedings

40 1

Astron. Nachr./AN 324, No. S1 (2003)

22.2 GHz Size I

2-

1.8h

"' ,,,.

%

"mr

E

v

(I) .-

1.6-

2 L

0 1.4-

.-c

I 1.2-

1-

+

26

27

28

3

29

Scattering Extrapolation Axial RatwO 53 31

32

Major Axis (mas) Fig. 4 Sizes from closure amplitude analysis of 22 GHz VLBA experiments (blue squares) and the mean (red triangle) compared with the Lo et al. (1998) result (green circle) and with scattering size extrapolated from longer wavelengths (star). The dashed line indicates constant axial ratio at a value fixed by longer wavelength observations.

43.2 GHz Size

I

n

0 15

Axial Ratio=0.53 1

0.65

0.7

0.75

0.8

0.85

0.9

Major Axis (mas) Fig. 5 Sizes from closure amplitude analysis of 43 GHz VLBA experiments (blue squares) and the mean (red triangle) compared with the Lo et al. (1998) result (green circle) and with scattering size extrapolated from longer wavelengths (star). The dashed line indicates constant axial ratio at a value fixed by longer wavelength observations.

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1.403-406 (2003)/ DO1 10.1002/asna.200385072

VLBA observation of a radio intraday flare of Sgr A* Makoto Miyoshi* Hiroshi Imai2, Junichi Nakashima3, Shuji Deguchi4, and Zhi-Qiang

' National Astronomical Observatory Japan, 2-21-1, Osawa Mitaka Tokyo, Japan, 181-8588

* Joint Jnstitute for VLBI in Europe, Posthus 2,7990 AA Dwingeloo, the Netherlands

Department of Astronomy, University of Illinois at Urbana-Champaign 1002 West Green Street, Urbana, IL 61801, USA Nobeyama Radio Observatory, NAOJ, Minamimaki, Minamisaku, Nagano Japan, 384-1305 Academia Sinica Institute of Astronomy and Astrophysics, P.O. Box 23-141, Taipei 106 Shanghai Astronomical Observatory, Chinese Academy of Sciences, Shanghai 200030 National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012

Key words Sgr A*, intraday variation, VLBI, jets. Abstract. We report here the results of a 7-hour VLBA observation of Sgr A* at 43GHz on July 31 2001, when Sgr A* showed an intraday flux variation at mm-wave length. The flux density increased from 1Jy to 3Jy, and the duration is about lo4 seconds which is quite similar to its X-ray flaring that Chandra detected. VLBA snapshot mappings show that Sgr A* flared and jets appeared in opposite two directions (northsouth) with nearIy light velocity. The lengths of jets reached about 15AU at most. The jets were the origin of its rapid increase of radio flux density.

1 Introduction Sgr A* is now the most probable super massive black hole candidate, whose mass is 2.6 - 3.7x106Ma (Ghez et al. 2000, Schodel et al. 2002) and is located at the center of mass of Milky Way (Reid et al. 1998). Since its discovery palick and Brown 1974) Sgr A* has long been recognized to be very quiet though there exists evidence that Sgr A* was active about 300 to 1000 years ago (Murakami et al. 2001). However recent detections of the intraday flaring of its X-ray and radio emission (Baganoff et al. 2001, Miyazaki et al. 2003) and a certain periodic variation of radio flux density in a 106 day cycle (Zhao et al. 2001) are evidence of AGN-like activity of Sgr A*. After the GC02 conference we refined our data reductions, especially carefully at self-calibrations and we applied shorter integration times when we made images of Sgr A*. We found that a radio intraday variation occurred at our observing time and the flux density increase was due to the jet eruption of Sgr A*. Previous VLBA observations already suggested a jet with north-south direction for intrinsic structure of Sgr A* (Lo et al. 1998). The first jet model is proposed by Falcke et a1.(1993,2000). Accretion flow models are proposed by Rees (1982), Melia (1992), and Narayan et a].( 1995). Yuan et a1.(2002) explain the spectrum of Sgr A* using jet and ADAF model.

2 Observations and reduction The purpose of our observation was astrometry of SiO maser sources relative to Sgr A* in the Galactic Center. We observed Sgr A* and SiO masers ( J = 1 - 0, v=l and 2 , 43GHz) at the Galactic Center simultaneously in the single beam using VLBA (NRAO) from OlOOUT to 0800UT on July 31 2001. The * Correspondingauthor: e-mail: rniyoshiOrniz.nao.ac.jp. Phone: +81422 343937, Fax: +81422 34 3635

@ ZM)3 WlLEYVCH Verlag GmbH & Co. KGaA. Weinheirn

M. Mivoshi et al.: VLBA radio intradav flare of Sm A*

404

data of 8 x 4 M H i bands with 2-bit sampling were processed by FX correlator at Socorro NM to produce visibilities with 64cMF in the form of 1.04 sec accumulated fragments. Five correlation processings were done with changing tracking positions for the Sgr A* and 4 SiO maser locations. We used the classic AIPS (NRAO) for data calibration and imaging. For obtaining fringe delays and rates we adopted the FWNG solutions with SNR greater than 4. We put interpolated values for the data points where good solutions could not be obtained. For getting solutions of fine complex gain errors, we repeated CALIB and IMAGR iterations. In this stage we abandoned the data points where good solutions could not be obtained from self calibrations (CALIB). At first we tried to make one image from the whole observing data as usual manner, but could not get a good convergence. This situation is the same when Miyazaki et al. (2003) found the intraday variation of Sgr A* from their observations with the Nobeyama interferometer. The structure of Sgr A* should change rapidly during our observations. Then we gradually shortened the integration times for imagings. For the shortened integration scans we repeated the CALIB and IMAGR iteration individually, and at last we noticed 15 min or 30 min integration times are proper for the imagings. Because we could get the image of SiO maser of IRSlOEE from visibilities to which the same calibrations of delay, rate, phase and amplitude were applied as those of Sgr A*, we judge the calibrations of the whole stations and baselines are sufficiently good for imaging Sgr A*.

3 Results and Discussions We show the best cleaned maps of Sgr A* that we obtained with the present reduction technique (Fig. 3). the observing start times (UT) of each cleaned maps are No.1 01:00, No.2 01 :30, No.3 02:00, No.4 02:30, No 5 . 03:00, No.6 03:30, No.7 04:00, No.8 04: 15, No.9 04:30, No.10 04:45, No.11 05:00, No.12 05:15, No.13 05:30,No.1405:45,No.15 06:00, No.1606:15,No.1706:30,No.1806:45,No.1907:00, andNo.20 07:15 respectively. Integration times of No.1-No.6 are 30min, those of No.7-No.20 are 15min. The restoring beams of the cleaned maps are unified to 0.45 mas x 0.15 mas, P A = 0". Except for No.4, No.9, No. 10, and No.11, the size of the restoring beam means what we call super resolution. These cleaned maps show quite different shapes from the shapes of their corresponding dirty beams. These 20 maps indicate that a rapid structural change of Sgr A* occurred at mm-wave emission during 7 hours. From OlOOUT to 0330UT the shape of Sgr A* is roughly ellipsoidally elongated to east-west direction with the flux density of about lJy, which is quite consistent to the shape obtained from previous VLBA observations (Lo et al. 1998, Bower et al. 1998). After 0345UT a jet-like structure emerged towards north and south directions. At the same time the flux density began to increase and reached the peak of about 3.5Jy at 0530UT (Fig.1). Then the flux density began to decrease. The duration o f the flare is about 2.5 hours or lo4 seconds which is similar to the duration of the X-ray flare of Sgr A* (Baganoff et al. 2001). Fig. 2 shows the size variation of Sgr A*. We take the direction of P A = 13" as the jet axis, and measure the projected line length of contour level of 25mJy/B to the jet axis and defined them as jet length. The lengths of jets reached about 15 AU at most, that is about 300 Schwarzschild radii for the GC black hole. The apparent velocity of the jet eruption is 0 . 5 ~ - 1 . 0if~assuming the distance of Sgr A* to be 8kpc. We also measured the width of the emission along the perpendicular direction, namely P A = 103". It is slightly decreased during 7 hours as shown in Fig. 2 (dotted line). If accretion disk elongates perpendicularly to jet, the decrease of the width may indicate the variation of the radius of accretion disk in Sgr A*. Acknowledgements We thank Dr. K. Karneno and Dr. Y. Hagiwara, hk. K. Asada and MI. T. Oyama for helpful

discussions.

References Baganoff, F. K., Bautz, M. W., Brandt, W. N., Chartas, G., Feigelson, E. D., Gannire, G. P., Maeda, Y., Morris, M., Ricker, G. R., Townsley, L. K., Walter, F. 2001, Nature 413,45

405

Astron. Nachr./AN 324. No. S I (20031

40

size variation of SgrA* (2001 -7.31)

flux variation of SgrA* (2001.7.31) I

I

I

45

,.

40 t35

E3n

10 05

05

no

no i

z

3

4

5

6

i

e

U T (hr)

Fig. 1 Flux variation of Sgr A*. The values of flux density were obtained from cleaned images.

I

I 1

2

3

4 5 UT(hr)

6

1

8

Fig. 2 Size variations of Sgr A*. Solid line shows the change of size at P A = 13", the direction of the jets, dotted line shows those at P A = 103", perpendicular to the jet direction.

Balick, B., and Brown, R . C. 1974, ApJ, 194, 265 Bower, G. C., and Backer, D. C. 1998, ApJ. 496, L97 Falcke, H., Mannheim, K., Biermann, P. L. 1993, A&A, 278, LI Falcke, H., Markoff, S. 2000, A&A, 362,113 Ghez, A. M., Morris, M., Becklin, E. E., Tanner, A,, Kremenek, T. 2000, Nature 407, 349 Lo, K. Y., Shen, Zhi-Qiang, Zhao, Jun-Hui, Ho, Paul T. P. 1998, ApJ. 508, L61 Melia, F. 1992, ApJ, 387, L25 Miyazalu,A., Tsutumi,T., and Tsnboi, M. 2003, in these proceedings. Murakami H., K0yama.K. Maeda,Y. 2001, ApJ, 558,687 Narayan, R., Yi, I., Mahadevan, R. 1995, Nature 374,623 Rees, M. J. 1982, in AIP Conf. Proc. 83: The Galactic Center, 166 Reid,M. J., Readhead, A. C. S., Vermeulen, R. C., Treuhaft, R. N. 1999, ApJ, 524, 816 Schodel, R., Ott, T., Genzel, R., Hofmann, R., Lehnert, M., Eckart, A,, Mouawad, N., Alexander, T., Reid, M. J., Lenzen, R., Hartung, M., Lacombe, F., Rouan, D., Gendron, E., Rousset, G., Lagrange, A.-M., Brandner, W., Ageorges, N., Lidman, C., Moonvood, A. F. M., Spyromilio, J., Hubin, N., Menten, K. M. 2002, Nature 419, 694 Yuan, F., Markoff, S., Falcke, 2002, A&A, 383,854 Zhao, J.-H., Bower, G. C. and Goss, W. M. 2001, ApJ. 547, L29

M. Mivoshi et al.: VLBA radio intradav flare of Sm A*

406

2 1 0

-1 -2 2 1 0

-1

-2

2 1

0

-1

-2 2

1 0

-1 -2 1

0

-1

1 0 -1 MilliARC SEC

1

Cont peak flux = 3.4332E-01 JYIBEAM Levs = 1.000E-02 * (2, 3.300, 4.600, 5.900, 7.200, 8.500, 9.800, 11.10, 12.40, 13.70,15, 16.30, 17.60, 18.90, 20.20, 21.50, 22.80, 24.10, 25.40, 26.70, 28, 29.30, 30.60, 31.90, 33.20, 34.50)

Fig. 3 The rapid change of Sgr A* structure. See text for explanations.

0

-1

Astron. Nachr./AN 324, No. S1,407-411 (2003) / DO1 10.l002/asna.200385042

A Chandra View of Diffuse X-Ray Emission in the Central 20 Parsecs of the Galaxy Sangwook Park*I , Frederick K. Baganoff2,MarkW. Bautz2, Gordon P. Garmire‘,Yoshitomo Maeda3,MarkMorris4,and Michael P. Muno2

’ Astronomy and Astrophysics, Penn State University, 525 Davey Lab., University Park, PA. 16802, USA

* Center for Space Research, MIT, Cambridge, MA. 02139, USA

Institute of Space and Astronautical Science, 3-1-1 Yoshinodai, Sagamihara, Kanagawa, 229-85 10, Japan

‘ Physics and Astronomy, UCLA, Los Angeles, CA. 90095, USA

Key words Galaxy: center - ISM: clouds - X-rays: individual (Sagittarius A East) -X-rays: ISM PACS 04A25 Over the last three years, the Galactic center region has been monitored with a series of ChandrdACIS observations. Besides the target object Sgr A*, the surrounding diffuse X-ray emission has been detected within the 17’ x 17’ field of view. As of 2002 June, combining all 12 GTO and GO observations, the total effective exposure reaches -590 ks, which reveals the detail structure of the faint filamentary diffuse X-ray emission with significant photon statistics. We here present preliminary results from the imagingkpectral analyses of these data. The “true-color”X-ray image and the equivalent width (EW) images for the detected elemental species of the Galactic center region indicate that the diffuse X-ray features have complex spatiospectral structures. We detect strong enhancements of He-like Fe within the 1.3’ diameter circular region in the immediate east of Sgr A*, which is most likely emission from the highly ionized Fe associated with the supernova remnant Sgr A East. Enhancements of the low ionization state Fe and highly ionized S and Si EWs in the northeast of Sgr A* may be interpreted as irradiation of photons from external X-ray sources and/or emission from bombardments of high energy particles such as unresolved SN ejecta.

1 Introduction The central -100 pc around the Galactic center (GC) is an extremely complex region with a variety of astrophysical objects: i.e., cold and warm molecular clouds, stellar clusters, supernova remnants (SNRs), HI1 regions etc. Recent X-ray observations with the Chandru X-Ray Observatory revealed over -2300 point sources within -20 pc of the GC (Muno et al. 2003). Besides these point sources, including the supermassive black hole candidate Sgr A*, the Chandra data show complex structure of the diffuse X-ray emitting features within the field of view (FOV). Preliminary results from earlier Chandra observations suggested that these diffuse X-ray emission features present various spectral and morphological characteristics (Bamba et al. 2002). The interpretations of these X-ray features have been highly speculative. We here present the latest deep Chandra observations of the GC. The total exposure has reached -590 ks (Table l), which provides unprecedented rich data with superb angular resolution and decent photon statistics for the diffuse X-ray emission. * Corresponding author: e-mail:

[email protected],Phone: 8148637111, Fax: ,814863 3399 @ 2003 WILEY-VCH Verlag GmbH & Co. KGaA. Weinhem

S. Park et al.: GC Diffuse X-Rays

408

Table 1 ChandrdACIS Observation log.

ObsID

Date

Exp

ObsID

Date

(ks)

Exp

ObsID

Date

(ks)

Exp (ks)

0242

1999-09-21

41

2951

2002-02-19

12

3392

2002-05-25

167

1561a

2000-10-26

36

2952

2002-03-23

12

3393

2002-05-28

158

1561b

2001-07-14

14

2953

2002-04-19

12

3663

2002-05-24

38

2943

2002-05-22

35

2954

2002-05-07

12

3665

2002-06-03

90

2 X-ray Images

Fig. 1 The exposure-corrected true color image of the GC from the Advanced CCD Imaging Spectrometer (ACIS) on board Chandra: red is 1.5 - 4.5 keV, green is 4.5 - 6.0 keV, and blue is 6.0 - 8.0 keV band. The diagonal dashed line is the Galactic plane through Sgr A*. Some Galactic molecular clouds are

schematically presented with dashed ellipses. Each subband image has been adaptively smoothed and the detected point sources have not been removed.

Fig. 1 is a “true-color’’ X-ray image of the GC region. The bright diffuse X-ray emission in the immediate east to Sgr A* is the SNR Sgr A East (Maeda et al. 2002). Some discrete diffuse X-ray features are notable: e.g., No.1 - 6 in Fig 1. The bright hard X-ray knot (No. 5) in the south of Sgr A* is spatially coincident with the bright nonthermal radio knot as a part of the radio SNR (3359.92-0.09 (source E; Ho et al. 1985). The nonthermal X-ray spectrum was detected from this feature with the XMM-Newton (Sakano et al. 2003) and we confirm this with the Chundru data (see $3). X-ray spectra from other features range soft to hard with typical angular size of -1‘. Besides these small, bright knots, we also detect faint, large-scale filamentary or loop-like features (for example, the dotted curves in Fig 1). These large features are soft, presumably with thermal origin. We trace the overall angular distributions of the detected atomic elements by constructing equivalent width (EW) images for the elemental species, Fe, S, and

Si (Fig 2). Sgr A East is undoubtedly detected as a bright source of highly ionized Fe (Fig 2a) near the center of the FOV. The bright hard X-ray knots (No. 1 - 3) are outstanding in low-ionization state or “neutral” Fe line ( E 6.4 keV) EW (Fig 2b). The overall correlations between the large scale enhancements of the neutral Fe line EW and the 50 km s-’ clouds in the northeast of Sgr A* (Tsuboi et al. 1999) may suggest the association of the X-ray emitting cold Fe material with the dense molecular clouds. The nonthermal radio knot (No. 5 ) and the cometary arc-like filament (No. 4) are featureless in the EW images, indicating continuum-dominated spectra. Region 3 is bright in both 6.4 keV Fe EW and He-like S and Si EW maps, suggesting X-ray emission from multiple phase gas.

-

409

Astron. Nachr./AN 324, No. S1 (2003)

400

800

1500

2500

eV

200

400

600

eV

800

b)

a)

200

300

400

600

eV

-

400

c)

500

600

-

800

eV

d)

Fig. 2 False-color EW images of a) Fe Hea ( E 6.7 keV), h) “neutral” Fe K ( E 6.4 keV), c ) S Hecr + Lycv ( E 2.5 keV), and d) Si Hea: ( E N I .9 keV) lines. The underlying continuum has been estimated from higher and lower “shoulders”of the emission lines, and then logarithmically interpolated to the line center energy for each element. The line and continuum images have been adaptively smoothed before calculating the ratios on the pixel by pixel basis.

-

3 X-ray Spectra The spectrum from each source region is presented in Fig 3 and Table 2. The spectra from regions 1 and 2 exhibit a strong 6.4 keV Fe line. The measured Fe EWs are -1 keV with X-ray luminosities (2 - 10 keV), LX = 1 - 2 x ergs s - l . These Fe line characteristics are similar to those from Sgr B2 and Sgr C clouds (Murakami et al. 2000,20Ola, 2001b) which have been interpreted in terms of an X-ray reflection

410

S. Park et al.: GC Diffuse X-Rays

nebula (XRN) model. Assuming the XRN model, if the primary external irradiating source is Sgr A*, the apparent angular distance of these Fe knots from Sgr A* (4' suggests ) an outburst of Sgr A*, with an X-ray luminosity of LX 2 lo3? ergs s-l, -40 yr ago. This is -1% of what might have produced the XRN in Sgr B2 and Sgr C (Murakami et al. 2000, 2001a). Albeit speculative, considering the proximity to an SNR (Sgr A East) and the detected frequent outbursts at various intensity levels of Sgr A* (Baganoff et al. 2003), the suggested outburst of Sgr A* is not unreasonable. Source No. 3 shows the same Fe line characteristics as sources 1 and 2. The absorbing column is lower than the other two, and the soft X-ray emission ( E < 3 keV) is also detected. This soft X-ray emission reveals the presence of a thermal component with a strong S Hea line ( E 2.5 keV). The thermal component indicates an electron temperature of kT 0.3 keV with high abundances. The overabundant thermal emission is likely X-ray emission from SNRs in the intercloud regions. This suggests that the 6.4 keV Fe line emission may also originate from unresolved SN ejecta (Bykov 2002). X-ray spectra from sources 4 and 5 are continuum-dominated with no evidence of strong emission lines. The positional coincidence of source 5 with the radio knot of G359.92-0.09 and the nonthermal Xray spectrum suggest that this bright X-ray knot is the X-ray counterpart of the radio shell of the SNR. The origin of source 4 is uncertain. The arc-like morphology and the nonthermal spectrum may suggest that this is also a part of previously unknown SNR. However, the observed X-ray spectrum (I' 1.2) appears to be harder than synchrotron emission from highly accelerated electrons by blast wave shocks. The source 6 spectrum is typical for X-ray emission from SNR andor stellar wind-blown bubble (kT 1.4 keV with normal abundances). N

N

-

Table 2 Results of Spectral Fittings.

Name

rlkT -I(keV)

NH

FeLine

cmP2) Center (keV)

(

EW

Fluxa

Lxb

x2lu

(keV)

1

1.85??:$$-

34.0?;:8,

6.39:':

1.OX

5.39~k0.06 1.3

68.5153

2

3.12+8:::/-

36.X?:&

6.40':;:;

1.03

5.01f0.05

2.4

33.7144

1.77?;:~~10.31?::~~ 15.8?$;

6.39:::;:

1.29

5.84f0.06

1.5

134.7/121

3 4

1.26f;::;Ii-0.281

6.0'::;

-

1.46f0.02 0.15

61.4148

5

1.71-0.55 -

30.6';:'

-

4.0Xf0.06

1.00

75.3/71

6

41.365:;;;

11.32;:;

-

1.23f0.02

0.40

55.0186

a

ergs s-l crnp2. The 2 - 10 keV band X-ray flux in the units of ergs s-l. The 2 - 10 keV band unabsorbed X-ray luminosity in the units of

Acknowledgements This work has been supported in parts by NASA contract NAS8-01128 for the Chandra X-Ray Observatory.

References Baganoff, F. K., et al. 2003, these proceedings Bamba, A. et al. 2002, Proceedings of the Symposium "New Visions of the X-ray Universe in the XMM-Newton and Chandra era", Noordwijk-NL, Nov. 2001 (astro-ph/0202010) Bykov, A. M. 2002, A&A, 390,327 Ho, P. T., Jackson, J. M., Barrett, A. H., & Armstrong, T. 1985, ApJ, 288, 575 Maeda, Y., Baganoff, F. K., Feigelson, E. D., Moms, M., Bautz, M., Brandt, W. N. et al. 2002, ApJ, 570, 671

Astron. Nachr./AN 324, No. S 1 (2003)

41 1

I

I

1

5

channel energy (kev)

I

L

1 5 chonnsi energy (lev)

I

I 5

2

channel eoargy

(ICY)

Fig. 3 The ChundrulACIS spectrum: a) source No. 1 , b) No. 2, c) No. 3, d) No. 4, e) No. 5, and f) No. 6. Muno, M. P., Baganoff, F. K., Bautz, M., Brandt, W. N., Broos, P. S., Feigelson, E. D. et al. 2003, ApJ, 589, 225 Murakami, H., Koyama, K., Sakano, M., Tsujimoto, M. 2000, ApJ, 534, 283 Murakdmi, H., Koyama, K., Tsujimoto, M., Maeda, Y., & Sakmo, M. 2001a, ApJ, 550. 297 Murakdmi, H., Koyama, K., & Maeda, Y. 2001b, ApJ, 558,687 Sakano, M., Wanvick, R. S., Decourchelle, A., & Predehl, P. 2003, MNRAS, 340, 747 Tsuboi, M., Handa, T., & Ukita, N. 1999, ApJS, 120, 1

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.413-417 (2003)/ DO1 10.1002/asna.200385082

Investigating the subrnillimetre variability of Sagittarius A* with SCUBA Douglas Pierce-Price', Tim Jenness', John Richer2,and Jane Greaves3

' Joint Astronomy Centre, 660 North A'ohokii Place, Hilo, HI 96720, USA

* Cavendish Astrophysics, University of Cambridge, Madingley Road, Cambridge CB3 OHE, UK UK Astronomy Technology Centre, Blackford Hill, Edinburgh EH9 3HJ, UK

Key words Galaxy: center, submillimetre, Sagittarius A*, variability, techniques: image processing. Abstract. We present an analysis of variability in submillimetre continuum observations of Sagittarius A*. The initial observations were made with the Submillimetre Common-User Bolometer Array (SCUBA) on the James Clerk Maxwell Telescope (JCMT) during a wide-field 2.8" x 0.5" survey of the Galactic Centre "Central Molecular Zone" at 850 pm and 450 pm. Our average flux densities for Sagittarius A* were lower than some previous results, such as those of Serabyn et al. (1997). This may be evidence of variability in the submillimetre excess. To investigate this further, we have used additional data from the SCUBA/JCMT archive to calculate flux densities for other dates. We detected a 50% increase in the 850 pm flux density of Sagittarius A* between April 1998 and August 1999. N

1 Introduction This investigation is based on a wider survey of the Galactic Centre region (Pierce-Price et al. 2000; Pierce-Price 2002). In this survey, made with the Submillimetre Common-User Bolometer Array (SCUBA, Holland et al. 1999) from April 1998 to April 2000 on the James Clerk Maxwell Telescope (JCMT), we mapped the 850 pm and 450 pm continuum emission over a 2.8" x 0.5" region covering the "Central Molecular Zone" (CMZ). The submillimetre continuum primarily traces the temperature-weighted column density of dust in the CMZ, except for certain non-thermal sources such as Sagittarius A* (Sgr A*). The ' at 450 pm and 15" at 850 pm, and a sensitivity of 15-20 M a of gas and survey has a resolution of 8 dust per beam. The CMZ images show extremely rich extended structure, including GMCs, filaments, and partial shells throughout the region. We also detected the point source Sgr A* at both wavelengths. The 850 pm flux density was 2.610.3 J y and the 450 pm flux density was 3.2 Jr 0.7 Jy. These results are lower than previous values such as those by Serabyn et al. (1997), and in addition we saw some apparent variation in the flux densities at different epochs during the survey. Sgr A* is known to have a submillimetre excess above its quasi-monoenergetic synchrotron radio spectrum. Related work combined these SCUBA maps with SCUBA polarimetry to confirm that the emission is polarised, and is therefore synchrotron radiation (Aitken et al. 2000). The transition between optically 1mni, highlighting the importance of thin and self-absorbed synchrotron in this component occurs at the submillimetre excess. This is consistent with the excess arising in a separate compact component on scales of a few Schwarzschild radii, as in the core-shell model of Beckert & Duschl(1997). Some models of Sgr A* predict a correlation between variability in the submillimetre excess and that in X-rays, such as the flare detected by Baganoff et al. (2001). N

N

@ 2003 WILEY-VCH Verlag GmhH B Co. KGaA, Weinhem

414

D. Pierce-Price et al.: Submillimetre variability of Ser A*

2 Reduction of multiple-beam SCUBA observations The SCUBA Galactic Centre survey was made in scan-map mode with chopped dual-beam observations to compensate for sky emission and sky noise. The chopping produces images which are convolved with the dual beam response. Thus, sensitivity is reduced to spatial frequencies at the inverse chop throw and its harmonics, and the total power level (zero spatial frequency) is not measured at all. The survey was originally reduced by combining multiple observations made with orthogonal chop throws in a weighted average in Fourier space (Emerson 1995). This approach is simple and quick but has some disadvantages: 0

The resultant maps have unphysical negative regions.

0

Maps can only be correctly mosaiced if they were observed with identical sets of chop throws.

0

The algorithm works on rebinned images and one cannot easily mark part of an observation as noisier than the rest.

We have developed an alternative deconvolution algorithm (Pierce-Price 2002) based on DBMEM by Richer (1992). This code uses a Maximum Entropy approach to finding the optimal reconstructed image given the observed data as constraints. It seeks to minimise the penalty function Q:

where x2 is the misfit between the predicted data derived from a reconstructed image and the actual data, a is a regularising parameter, and S is the entropy of the reconstructed image defined in terms of its pixel values I and a default image ni:

Some of the advantages of this technique are as follows: 0

0

0

Since the entropy is only defined for positive images, the resultant image is guaranteed to be positive. By treating the chopped measurements as a set of constraints it is trivial to co-add or mosaic observations: simply concatenate the data sets. It is also easy to include observations such as jiggle maps where the chop pattern is different, or total power observations.

This flexibility allows us to combine observations of Sgr A* with other observations of its surroundings in a way that is impossible with the previous algorithm. In the observations shown here, the value of a used was zero. This indicates that the maps are not underconstrained, and the method becomes simply a non-negative least-squares fit.

3 Individual SCUBA observations of Sgr A* After seeing some evidence for variability in the individual observations we retrieved other observations of Sgr A* from the JCMT archive for this period. The initial data reduction was done with the automated ORAC-DR pipeline (Jenness & Economou 1999; Jenness et al. 2002). The files were then reduced using the Maximum Entropy-based algorithm described in section 2. Only the 850 pm observations were used as the detections of Sgr A* were too weak in the individual 450 pm observations.

Astron. Nachr./AN 324, No. S 1 (2003)

415

Table 1 Flux densities of Sgr A* at 850 pm for the period 1998-2001. The difference is measured from the average value for April 1998.

UT date 19980407 19980409 1998041 1 19980413 19980521 1 9990212 19990325 19990826 20000318 20000405 20010805

Flux density (Jyheam) 850 pm 4.8 f 0.5 4.5 f 0.5 4.2 f0.4 4.8 f 0.5 6.5 f 0.7 4.0 2~ 0.4 5.0 f 0.5 6.0 f 0.6 5.1 f0.5 5.8 f0.6 4.3 i0.4

Difference (Jyheam) +0.2 -0.1 -0.4 t0.2 +2.0 -0.6 +0.5 +1.4 +0.5 +1.2 -0.3

The individual maps are shown in Figure 1. Some observations, particularly jiggle maps, may suffer from chopping onto source structure due to the extended emission around Sgr A*. To counteract this we combined the datasets. For each UT date, only that night’s data constrained the central 2 arcminute diameter region around Sgr A*. However, for the rest of the area within a diameter of 12 arcminutes, the data from all nights were combined.

4 Results and analysis The results from comparing the set of observations are shown in Figure 2 and Table 1. Note that there is no result for UT date 19990326 as those observations did not actually cover the central 2 arcminute region. The differences quoted are from the mean result for April 1998. The quoted uncertainties reflect a calibration uncertainty of 10%. There is some evidence for an increase in the flux density from April 1998 to August 1999, whilst by late 2001 the flux density appears to have returned to April 1998 levels. The measurement on 19980521 was a very brief map and may not be trustworthy - this requires further analysis. Variations on short timescales could occur - the X-ray flare detected by Baganoff et al. (2001) lasted only 3 hours. The increase in 19990826 is at the 2.3-0 level, the dominant source of error being the calibration uncertainty. The lack of an absolute power measurement is not as important as the background is derived from all the data and is essentially constant. The background emission at the position of Sgr A* is estimated as 2.0 Jyheam from the surrounding structure. This implies that the flux density of Sgr A* itself in April 1998 was 2.6 Jylbeam, rising to 4.0 Jylbeam in August 1999, an increase of about 50%.

5 Conclusions The use of the Maximum Entropy method to deconvolve SCUBA data allows the combination of multiple observations, and wider-field maps can be used to provide information about the surrounding structure where otherwise this extended emission might be chopped out. There is evidence of variability of Sgr A* at 850 pm. However, the uncertainty on this measurement is quite large due to the calibration uncertainties. Taking the ratio of the Sgr A* flux to its immediate

416

D. Pierce-Price et al.: Submifirnetre variability of Sgr A*

- " . " * - -

Fig. 1 Individually reduced 850 pm maps of the region around Sgr A* for the period 1998-2001. Each map is labelled with its UT date, and the units are Jyhearn.

417

Astron. Nachr./AN 324. No. S 1 (2003) Flux density of Sgr A* m

,

.

.

,

(

.

.

,

.

,

.

1

,

.

.

,

.

.

,

.

i

- -------

N

1998

"

"

"

"

"

"

1999

"

"

"

'

2001

2000

2002

UT date of observation

Fig. 2 Flux densities of Sgr A* at 850 pm for the period 1998-2001. The dotted line is the average result for April 1998. The y-axis begins at the estimated 2 Jy background level.

surroundings may give a result with a reduced uncertainty, and we will investigate this further. The observations of Sgr A* are sparse, even when the JCMT archive is included. Future monitoring of Sgr A* in the submillitnetre could provide more further evidence for this variability. Acknowledgements The JCMT is operated by the Joint Astronomy Centre on behalf of PPARC of the UK, the Netherlands OSR, and NRC Canada.

References Aitken, D. K., Greaves, J., Chrysostomou, A,, Jenness, T., Holland, W., Hough, J. H., Pierce-Pnce, D., & Richer, J. 2000, ApJ, 534, L I73 Baganoff, F. K., Bautz, M. W., Brdndt, W. N., Chartas, G . , Feigelson, E. D., Gannire, G. P., Maeda, Y., Morris, M., et al. 2001, Nature, 413, 45 Beckert, T. & Duschl, W. J. 1997, A&A, 328.95 Emerson, D. T. 1995, in ASP Conf. Ser. 75: Multi-Feed Systems for Radio Telescopes, 309+ Holland, W. S., Robson, E. I., Gear, W. K., Cunningham, C. R., Lightfoot, J. F., Jenness, T., Ivison, R. J . , Stevens, J. A., et al. 1999, MNRAS, 303,659 Jenness, T. & Economou, F. 1999, in ASP Conf. Ser. 172: Astronomical Data Analysis Software and Systems VIII, 1714 Jenness, T., Stevens, J. A., Archibald, E. N., Economou, F., Jessop, N. E., & Robson, E. 1.2002, MNRAS, 336, 14 Pierce-Price, D., Richer, J. S., Greaves, J. S . , Holland, W. S., Jenness, T., Lasenby, A. N., White, G. J., Matthews, H. E., et al. 2000, ApJ, 545, L121 Pierce-Price, D. P. I. 2002, Ph.D. Thesis, University of Cambridge Richer, J. S. 1992, MNRAS, 254, 165 Serabyn, E., Carlstrom, J., Lay, 0..Lis, D. C., Hunter, T. R., & Lacy, J. H. 1997, ApJ, 490, L77 I

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.419-423 (2003) / DO1 10.1002/asna.200385091

Near-Infrared Flux Limits for Sgr A* Based on NICMOS Data

’,

Susan Stolovy Fulvio Melia’, Donald McCarthy’, and Farhad Yusef-Zadeh3

’ SIRTF Science Center, CalTech, MS 220-6, Pasadena, CA 91 125, USA

’ Steward Observatory,Tucson, AZ 85721

Northwestern University, Dept. of Physics and Astronomy, Evanston, L 60208

Key words infrared, stars, Sgr A*, Galactic Center, black hole, interstellar medium Abstract. Images of the central arcsec of the Galaxy at near-infrared(1R)wavelengths reveal a tight stellar cluster, however, no point source exactly coincident with the radio source Sgr A* has been clearly detected in the near-IR. “Contaminating”emission due to overlapping point spread functions (PSFs) from this stellar cluster as well as from the bright RS16 sources 1-3” away from Sgr A* makes it difficult to assess directly the near-IR characteristics of Sgr A*. Taking advantage of the stability of the PSF achieved by HSTDJICMOS observations, we employ PSF subtraction techniques in order to investigate the true nature of the near-IR emission from Sgr A*. New limits on the maximum possible flux emanating from a point source coincident with Sgr A* are presented at 1.1, 1.45,1.6, 1.9, and 2.2 pm based on NICMOS data taken at several epochs. These are the faintest reported near-IR flux limits for wavelengths shorter than 2.2 pm . The resulting extinction-correctedfluxes are compared to theoretical models (including both quiescent and flare models) for the spectrum of Sgr A*. We also present NICMOS narrow-band emission line images in Paa where significant stellar PSF residuals have been removed to reveal new diffuse structures in the ionized gas near Sgr A* at two epochs.

1 Motivation We desire to study both the continuum and ionized emission close to Sgr A*. Removal of point source emission from the “S” sources within the central 1” as well as from the very bright IRS16 sources located 1-3’’ away from Sgr A* allows us to place more stringent limits on the near-IR flux from Sgr A* than from the original images. The distribution of ionized gas very close to Sgr A* has never been measured at high spatial resolution; observations of H92a (Roberts & Goss 1993) in the radio suffer from relatively low ( I ” ) spatial resolution, and the overpoweringly bright nonthermal point source emission from Sgr A* itself makes imaging this region impossible in the radio continuum. For bright point sources in the Galactic Center, such as the IRS16 cluster, the NICMOS PSF is detected extending several arcsec from the source. Therefore, removal of these strong PSF signatures from both the continuum and Paa: NICMOS images allows us to study for the the distribution of continuum and ionized emission directly toward Sgr A* with 0.1”-0.2” resolution.

2 Observations and Data Reduction NICMOS observations of the Galactic Center were taken with Camera 1 (0.043” pixels) and Camera 2 (0.076’’ pixels) at wavelengths of 1.1, 1.45, 1.6, 1.9, and 2.2 p m (filters F1 10M, F145M, F160W, F190N, and F222M, respectively) as part of several independent observing programs. The data were re-reduced using improved flats and darks as compared to the original archived data. Individual dithered images were aligned within a relative accuracy of 0.02 pixel to make a mosaic in each filter. @ 2003 WILEY-VCH Verlag GmbH & Co. K G A , Wewheim

S. Stolovy et al.: Near-Infrared Flux Limits for Sgr A*

420

In order to remove stellar PSF artifacts from the NICMOS images, we have used the IDL package “starfinder” (Diolaiti et al. 2000) to detect stars in each image to a high correlation with the PSF. We used “starfinder” to construct a composite PSF for each wavelength (dominated by the brightest object, IRS7) derived from the stars in the NICMOS mosaics. The exception was for the F110M filter, which due to the 11 magnitudes of extinction at 1.I fim toward the Galactic Center, did not have stars with sufficient signalhoise to produce an adequate PSF. In this case the synthetic “Tiny Tim” (Krist & Hook 1997) PSF was adopted. The Paa images, taken at two epochs during March and Oct. 1998, were constructed by adjusting the scale factor between the F187N and F190N 1% filters until the stellar residuals were minimized. In these Paa images, the IRS16 cluster has strong positive residuals in the F187N-Fl90N subtraction due to Paa emission (as well as for some sources, He1 (4-3) emission present within the F187N filter bandpass) in their stellar atmospheres. No significant emission lines are present in the F190N continuum filter. Some stars have negative residuals, due to either Paa in absorption in the stellar atmospheres or to excess local extinction (e.g. IRS7, which is a late type supergiant, would not be expected to have Paa absorption lines and has excess extinction along the line of sight). “Starfinder” was run on the F187N and F190N filters separately, and the reconstructed detected point sources were then subtracted from the original Paa images, effectively removing both positive and negative PSF residuals. This analysis was done for the Patr images spanning the inner parsec of the Galaxy, but we concentrate on the inner 0.1 pc for this paper.

-

N

3 Continuum Flux Limits We found the most sensitive method of determining the possible flux contribution from an unresolved object at Sgr A* was to subtract a scaled point source from the position of Sgr A* and note the level corresponding to the onset of a “hole” in the background flux distribution. The limits quoted here are the maximum possible flux emanating from a point source at the position of Sgr A* that is consistent with the NICMOS data. The artificial excess background due to overlapping PSFs is reduced after subtraction of the detected stars, making a more sensitive measurement possible. The position of the radio source Sgr A* was determined by offsetting from the maser source IRS7 (Menten et al. 1997). This method is illustrated in Figs. 1 and 2 for F11OM and Fig. 3 for F145M. The derived flux limits are listed in Table 1. Care must be taken in interpreting the implications of these flux limits; they typically must be multiplied by a factor of 2-3 to yield a detection of point source at the position of Sgr A*. This has been tested by adding an artificial point source at the position of Sgr A* and then requiring “starfinder” to successfully extract a star with the correct flux and location. This factor can be reduced to close to 1 for a detection if the neighboring stars are subtracted first and then the artificial star is added before running “starfinder”. The signavnoise in the image is much greater than the flux limits for all cases with the exception of F1 IOM, where the flux limit corresponds to a point source with a peak value of 3 LT. We estimate the total error in determining the flux limit via the PSF subtraction method described here to be of order 1520%.

4 Comparisons to Theoretical Models It is of interest to investigate whether or not the near-IR limits for Sgr A* can constrain theoretical models for the emission from Sgr A*. In recent years, Chandra X-ray observations of both quiescent and flared states for the emission from Sgr A* have been made (Baganoff 2001 et al. and Baganoff et al. 2003). Fig 4 shows a new model for the spectrum of Sgr A* (S. Liu, private communication). It is a modification of the bremsstrahlung model of Liu and Melia (2002). This new model allows for non-zero stress at the inner boundary of the accretion disk such that the accreted angular momentum equals zero, providing additional heating which raises the IR flux levels. The dereddened near-IR flux limits from the NICMOS observations are plotted in Fig. 5 as open squares, along with the Keck K-band quiescent and flared limits

Astron. NachrJAN 324,No. S1 (2003)

42 1

Fig. 1 The inner 1.Ox1 .O” (0.04 pc) of the Galaxy at 1.I pm centered on Sgr A* illustrating the flux determination method. All images are displayed at the same stretch. From left to right: 1) original image, 2) reconstructed image of detected stars (using starfinder), 3) original-detected stars, 4) same as 3 ) but with a 1 microJy point source subtracted at Sgr A*, causing a “hole”. The threshold for the onset of the hole was determined to be 0.52 i 0.07 microJy. The peak pixel level of the PSF for this limit also corresponds to roughly 3 0 in this image.

0 10

L

Fig. 2 Illustration of flux limit determination method. I-pixel wide east-west cut through Sgr A* for FI IOM data (Sgr A* is at pixel 10). Solid line: Original data; Dashed line: Residual image (original-detected stars); Dotted lines: PSF subtractions at Sgr A* from residual image for PSF flux densities of 0.5 (upper) and 1 .O microJy (lower). The adopted flux limit is 0.52 microJy.

Fig. 3 Image of the central 4x4” of the Galaxy at 1.45 pm , with a dashed circle centered on Sgr A*. Upper left: original image; upper right: reconstructed image from detected stars; lower left: original-detected stars; lower right: same as lower left but with a 22 microJy point source subtracted at the position of Sgr A*. The grayscale spans a dynamic range of 1OOO:l in the original image and is the same for all images.

derived independently (Hornstein et al. 20021, shown as open triangles. According to this model, the quiescent state of Sgr A* is much too faint to be detected in the near-IR. However, had a similar X-ray flare to the 2000 flare modelled here occurred during the NICMOS or Keck observations, it would have been on the margin of detectability. A significantly brighter flare would would have been necessary for a solid near-IR detection.

422

Filter

S. Stolovy et al.: Near-Infrared Flux Limits for Sgr A*

Observing Date

Original flux limit

Flux limit after PSF

Dereddened flux limit

[JY(mag)]

removal [Jy (mag11

after PSF removal [Jy (mag)]

FllOM

Aug. 1997

7.4e-7 (23.5)

5.2e-7 (23.8)

F145M

Mar. 1998

2.8e-5 (19.1)

2.2e-5 (19.4)

6.7e-3 (13.2)

F160W

Aug. 1998

9.9e-5 (17.6)

7.2e-5 (17.9)

l.le-2 (12.5)

F190N

Mar. 1998

5.le-4 (15.5)

4.k-4 (15.6)

1.9e-2 (1 1.6)

F222M

Oct. 1997

8.8e-4 (14.7)

5.5e-4 (15.2)

l.le-2 (12.0)

2

Fig. 4 Accretion-inducedflared (upper thick curve) and quiescent (lower dotted curve) theoretical models, updated from the model of Liu and Melia (2002) as discussed in the text. Detections of Sgr A* are from VLA radio observations (Falcke et al. 1998) and Chandra X-ray observations (Baganoff et al. 2001). Mid-infrared limits (from MIRLIN on Keck) from Cotera et al. 1999 are also plotted.

1

,

. . . . , . . . . , . . . . ,

Fig. 5 Subsection of previous figure emphasiziig infrared wavelength regime with the dereddened flux limits from NICMOS shown with open squares. The K band quiescent (2mJy) and flared (19 mJx) flux limits from Homstein et al. are shown as triangles.

5 Revealing the Distribution of Ionized Gas Near Sgr A* from Paa Images Fig. 6 shows the distribution of ionized gas in P a a before and after removal of PSF artifacts, from the March 1998 and Oct. 1998 observations. No clear point source enhancement in P a a appears at the position of Sgr A*; subtracting a point source from the position of Sgr A* gives a flux limit in P a a line emission of 1 - 1.3 x erg cm-2s-1 (the range given by measurements at both epochs). For the first tim2, a curious arc-like structure is seen near Sgr Ah, and indications of very high proper motions of the gas (a motion of 1 pixel corresponds to 2700 k d s ! ) are seen over this short interval of 7 months. Additional epochs are needed, however, for robust determination of the proper motion of the ionized gas.

Astron. Nachr./AN 324, No. S1 (2003)

423

Fig. 6 Pacv images of inner 2.5x2.5" (O.lpc), shown in positive grayscale and stretched to show faint emission. A dashed circle is centered on the position of Sgr A*. Upper left: Original March 1998 image; Upper right: Original Oct. 1998 image; In these upper figures, residual emission from the bright Pact emission line stars in the IRS 16 cluster contaminate the image. PSF removal techniques were employed in order to image the true distribution of diffuse ionized emission, as shown in the lower figures; Lower left: March 1998 image after PSF residual removal; Lower right: Oct. 1998 image after PSF removal. Note the discovery of an arc-like diffuse feature and the changes in the distribution of the ionized gas near Sgr A* over this short time interval.

Acknowledgements We thank Siming Liu for assistance in providing the model curves for Figs. 4 and 5.

References Baganoff, F. et al. 2001 Nature, 413,45 Baganoff, F. eta/. 2003, these proceedings. Cotera, A., Moms, M., Ghez, A.M., Becklin, E.E., Tanner, A.M., Werner, M.W. & Stolovy, S.R. 1999, in The Central Parsecs of the Galaxy, ed. H. Falcke et al. ASP Conf. Ser., 186,240 Diolaiti E., Bendinelli, O., Bonaccini, D., Close, L., Cunie, D., & Parmeggiani, G., 2000, A&AS, 147,335 Falcke H. et al. 1998 ApJ, 499,73 1 Hornstein, S.D., Ghez, A. M., Tanner, A., Morris, M, Becklin, E.E. & Wizinowich, P. 2002 ApJ, 577, L9 Krist, J. & Hook, R. The 1997 HST Calibration Workshop with a New Generation of Instruments, 1997, p. 192 Liu, S. & Melia, F. 2002 ApJ, 566.77 Menten, K. M., Reid, M. J., Eckart, A. & Genzel, R. 1997 ApJ, 475, 111 Roberts, D. and Goss, W.M. 1993, ApJS, 86, 133 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher, M., & Yusef-Zadeh, F. 2003, ApJ, submitted Stolovy, S.R., McCarthy, D.W., Melia, F., Rieke, G., Rieke, M.J., & Yusef-Zadeh, F., 1999, in The Central Parsecs of the Galaxy, ed. H. Falcke, ef al. , ASP Conf. Ser., 186,39

This Page Intentionally Left Blank

Astron. Nachr./AN324.No. S1.425-429 (2003) /DO1 10.1002/asna.200385112

The wavelength dependence of Sgr A* size and the unified model of compact radio sources Fedor Prigara * Institute of Microelectronics and Informatics, Russian Academy of Sciences, 21 Universitetskaya,Yaroslavl 150007. Russia

Key words Galaxy: center, radio continuum. Abstract. Using, in particular, the data for Sgr A”, we propose a unified model of compact radio sources, i.e. of pulsars, maser sources, and active galactic nuclei. The unification is based on the wavelength dependence of radio source size. It is shown that the compact sources are characterized by a maser amplification of thermal radio emission. The density, temperature, and magnetic field profiles of compact sources are discussed. The wavelength dependence of Sgr A*’s size is explained by the effect of the gravitational field of a central energy source on a gas flow.

1 Introduction The compact radio source at the Galactic Center, Sgr A*, was discovered as early as 1971 (Lo 1982). In 1976 Davies, Walsh, and Booth examined the wavelength dependenceof Sgr A*’s size and inferred that the observed size of the source is proportional to X2, where X is the wavelength. Later observations confirmed the X2 dependence of the source size at wavelengths of 2.8 cm to 30 cm (Lo 1982). The observed size of Sgr A* is too large to suggest that the X2 dependence of the source size is produced by scattering of interstellar electrons (Lo et al. 1993). Here we show that the X2 dependenceof the source size can be reproduced in a gaseous disk model with thermal emission, provided the effects of stimulated radiation processes are taken into account. This is the unified model of compact radio sources, Sgr A* being a representative of the family. Compact radio sources have small angular dimensions, usually less than 1 mas, and exhibit high brightness temperatures. These properties are common for pulsars (Shklovsky 1984), masers (Bochkarev 1992), and active galactic nuclei (Bower & Backer 1998, Kellermann, Vermeulen, Zensus & Cohen 1998). The brightness temperatures of OH masers have magnitude T b 5 1012K,and those of water masers have the magnitude Tb 5 10°K (Bochkarev 1992). Compact extragalactic sources (AGNs) exhibit brightness temperatures in the range of 10’” K to 1012 K (Bower & Backer 1998, Kellermann et al. 1998), so these temperatures have are of the same order of magnitude as those of OH masers. Another feature, which is common for compact sources, is that their radio emission so far has not received a satisfactory explanation. In particular, it is true for pulsars (Qiao et al. 2000, Vivekanand 2002). In the case of maser sources the modern theory uses some chance coincidences which hardly can maintain in a more profound theory (Bochkarev 1992). At last, it was shown recently that spherical accretion models with the synchrotron mechanism of emission cannot explain the flat or slightly inverted radio spectra of low-luminosity active galactic nuclei (Nagar, Wilson & Falcke 2001, Ulvestad & Ho 2001). Besides, the synchrotron self-absorption produces a change in the polarization position angle across the spectral peak. No such a change was detected in gigahertz-peaked spectrum sources (Mutoh et al. 2002). * Correspondmg author: e-mail: fprigaraQimras.yar.ru, Phone: +7 0852 246552, Fax: +7 0852 246552

@ 2003 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim

F. Prigara: Wavelength dependence on source size

426

The difficulties encounted by plasma mechanisms of radio emission from pulsars clearly show that the radiation is produced by a dense, low-energy medium (Gedalin, Gmman & Melrose 2002). Such a medium is the gaseous disk surrounding the central energy source.

2 The gaseous disk model It was shown recently (Prigara 2001b) that thermal radio emission has a stimulated character. According to this conception thermal radio emission of non-uniform gas is produced by an ensemble of individual emitters. Each of these emitters is a molecular resonator the size of which has an order of magnitude of mean free path 1 of photons

I=-

1 nu

where n is the number density of particles and o is the absorption cross-section. The emission of each molecular resonator is coherent, with wavelength

x = 1,

(2)

and the thermal radio emission of the gaseous layer is the incoherent sum of radiation produced by individual emitters. Condition (2) implies that the radiation with the wavelength X is produced by the gaseous layer with particle number density n . In the gaseous disk model, describing radio emitting gas nebulae (Prigara 2001a), the number density of particles decreases reciprocally with respect to the distance T from the energy center

n K r-I.

(3)

Together with the condition of emission (2) the last equation leads to the wavelength dependence of radio source size:

The relation (4) is indeed observed for sufficiently extended radio sources. For instance, the size of radio core of galaxy M31 is 3.5 arcmin at the frequency 408 MHz and 1 arcmin at the frequency 1407 MHz (Sharov 1982).

3 Extended radio sources The spectral density of flux from an extended radio source is given by the formula

where a is a distance from radio source to the detector of radiation, and the function B, (7') is given by the Rayleigh-Jeans formula

Astron. Nachr./AN 324, No. S1 (2003)

B, = 2 k T u 2 / c 2 ,

421

(6)

where v is the frequency of radiation, k is the Boltzmann constant, and T is the temperature. The extended radio sources may be divided in two classes. Type 1 radio sources are characterized by a stationary convection in the gaseous disk with an approximately uniform distribution of the temperature Txconsi giving the spectrum

F, z const

.

(7)

Type 2 radio sources are characterized by outflows of gas with an approximately uniform distribution of gas pressure P=nkTwxmst.In this case the equation ( 3 ) gives

T cc r ,

(8)

so the radio spectrum, according to the Eq. (9,has the form

Both classes include numerous galactic and extragalactic objects. In particular, edge-brightened supernova remnants (Kulkarni & Frail 1993) belong to the type 2 radio sources in accordance with the relation (8), whereas center-brightened supernova remnants belong to the type 1 radio sources. The typical members of the type 2 family are gigahertz-peaked spectrum sources (GPS), radio emission from GPS sources being produced by the expanding jets (Nagar et al. 2002). The intermediate magnitudes of a spectral index 0 < < 1 (F, cc v-*) can be explained by the joint emission from an outflow (a jet) and a stationary gaseous disk.

4 The wavelength dependence of radio source size For the case of compact radio sources, instead of the relation (4) the relation

is observed (Lo et al. 1993, Lo 1982). This relationship may be explained by the effect of a gravitational field on the motion of gas which changes the Eq. (3) for the equation

The mass conservation in an outflow or inflow of gas gives nvr=const,where 'u is the velocity of flow. In the gravitational field of a central energy source the energy conservation gives

where rs is the Schwarzschild radius. Therefore, at small values of the radius the Eq. (1 1 ) is valid, whereas at the larger radii we obtain the Eq. ( 3 ) .

F. Prigara: Wavelength dependence on source size

428

It is well known (Shklovsky 1984) that the delay of radio pulses from pulsars at low frequencies is proportional to X2. This fact is a mere consequence of Eq.( lo), if we only assume the existence of the radial density wave travelling across the radius with a constant velocity and triggering the pulse radio emission. In this treatment the pulsars also obey the X2 dependence of compact source size. Note that the wavelength dependence of a pulse duration is a similar effect. The spatial distribution of SiO, water, and OH masers (each of which emits in own wavelength) in the maser complexes also is consistent with the X2 dependence of compact source size (Bochkarev 1992, Eisner et al. 2001). To summarize, extended radio sources are characterized by the relation (4),and compact radio sources obey the relation (10).

5 Maser amplification of thermal radio emission The existence of maser sources associated with gas nebulae and galactic nuclei (Miyoshi et al. 1995) is closely connected with the stimulated origin of thermal radio emission. The induced origin of thermal radio emission follows from the relations between Einstein's coefficients for a spontaneous and induced emission of radiation (Prigara 2001b). The high brightness temperatures of compact, flat-spectrum radio sources (Bower & Backer 1998; Nagar, Wilson, & Falcke 2001; Ulvestad & Ho 2001 ) may be explained by a maser amplification of thermal radio emission. A maser mechanism of emission is supported by a rapid variability of total and polarized flux density on timescales less than 2 months (Bower et al. 2001). Such a variability is characteristic for non-saturated maser sources. Note that the spherical accretion models with the synchrotron mechanism of emission are unable to explain the flat or slightly inverted spectra of low-luminosity active galactic nuclei (Nagar et al. 2001; Ulvestad & Ho 2001). The Blandford and Konigl theory used by Nagar et al. (2001) is in some respects similar to the gaseous disk model (Prigara 2001a), the latter being more simple and free from indefinite parameters, such as an empirical spectral index. It is shown by Siodmiak & Tylenda (2001) that the standard theory of thermal radio emission which does not take into account the induced character of emission cannot explain the radio spectra of planetary nebulae at high frequencies without an introduction of indefinite parameters. Thermal radiation in a magnetic field is polarized (see, e.g., Lang 1974). Together with the magnetic field profiles this produces the wavelength dependence of polarization in the form of p o( X-'for extended sources and p o( X-'for compact radio sources (Prigara 2002). Maser amplification in continuum in different versions has been proposed earlier as possible radiation mechanism for pulsars (Lipunov 1987).

6 Conclusions Compact radio sources are characterized by the following properties: 1) small angular dimensions; 2) high brightness temperatures: 3) X2 dependence of radio source size; 4)maser mechanism of radio emission. The emission of compact radio sources differs from the extended ones physically because of the presence of maser emission. The properties of the compact radio sources can be adequately described within the gaseous disk model with a suitable density profile. Acknowledgements The author is grateful to D.A.Kompaneets, Y.Y.Kovalev, V.N.Lukash, V.V.Semenov and B.E.Stern for useful discussions.

Astron. Nachr./AN 324, No. SI (2003)

429

References Bochkarev N.G. 1992, Basic Physics of Interstellar Medium (Moscow: Moscow University Press) Bower G.C., Backer D.C. 1998, ApJ, 507, L117 Bower G.C., Wright M.C.H., Falcke H., Backer D.C. 2001, ApJ, 555, L103 Eisner J.A., Greenhill L.J., Hermstein J.R., Moran J.M., Menten K.M. 2002, ApJ, 569, 334 Gedalin M., Gruman E., Melrose D.B. 2002, Phys. Rev. Lett. (submitted), astro-pW0205069 Kellermann K.I., Vermeulen R.C., Zensus J.A., Cohen M.H. 1998, AJ, 115, 1295 Kulkarni S.R., Frail D.A. 1993, Nature, 365, 33 Lang K.R. 1974, Astrophysical Formulae (Berlin: Springer) Lipunov V.M. 1987, Astrophysics of Neutron Stars (Moscow: Nauka) Lo K.Y. 1982, in AIP Proc. 83: The Galactic Center (eds G.Riegler & R.Blandford) (New York: AIP) Lo K.Y., Backer D.C., Kellermann K.I., Reid M., Zhao J.H., Goss M.H., Moran J.M. 1993, Nature, 361, 38 Miyoshi M. et al. 1995, Nature 373, 127 Mutoh M., Inoue M., Kameno S., Asada K., Fujisawa K., Uchida Y. 2002, astro-pM0201144 Nagar N.M., Wilson AS., Falcke H. 2001, ApJ, 559, L87 Nagar N.M., Wilson A.S., Falcke H., Ulvestad J.S., Mundell C.G. 2002, in Issues in Unification of AGNs, ASP Conf. Series 258, (eds R.Maiolino, A.M,uconi & N.Nagar) (San Francisco: ASP) Prigara F.V. 2001a, astro-ph/OllO399 Prigara F.V. 2001 b, astro-pWOllO483 Prigara F.V. 2003 (in preparation) Qiao G.J., Xu R.X., Liu J.F., Han J.L., Zhang B. 2000, in Pulsar Astronomy- 2000 and Beyond (eds M.Kramer, N.Wex & R.Wielebinski) (San Francisco: ASP) Sharov A.S. 1982, The Andromeda Nebula (Moscow: Nauka) Shklovsky I S . 1984, Stars: Their Birth, Life, and Death (Moscow: Nauka) Siodmiak N., Tylenda R. 2001, A&A, 373,1032 Ulvestad J.S., Ho L.C. 2001, ApJ, 562, L133 Vivekanand M. 2001, MNRAS, 326, L33

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1,431-434 (2003) / DO1 10.1002/asna.200385094

Search for Circular Polarization toward Sagittarius A* at 100 GHz

’,

’,

M. Tsuboi *’, H. Miyahara I , R. Nomura T. Kasuga and A. Miyazaki Ibaraki University, Mito, Ibaraki, 310-8512, JAPAN Hosei University, Tokyo, 184-8584,JAPAN Noheyama Radio Observatory, Minamimaki, Minamisaku, Nagano 384-1305, JAPAN

’

Key words galaxies: active, Galaxy: center, polarization Abstract. We have observed circular polarization of Sgr A* at 100 GHz using the Nobeyama 45-m telescope. The circular polarization of Sgr A* in its quiescent phase was lmcl 5 0.5% after subtraction of the contribution from free-free emission surrounding Sgr A*.

1 Introduction The circular polarization of Sagittarius A* at mm-wavelength has not yet been detected (Bower et al. 2001), although the circular polarization has been detected at 4.8 and 8.4 GHz with the VLA (Bower et al. 1999). It is, moreover, reported that the fractional circular polarization increases with the rise of the observed frequency. These results provide new, important information to understand the structure and the emission mechanism of this source. In observing Sgr A* at mm-wavelength, interferometers have been used because the small beam is important to observe Sgr A’ exclusively, which is a weak point source embedded in relatively strong and extended components. On the other hand, large single-dish telescopes have another advantage that they have lower side lobe levels. This advantage also has benefits for precise polarimetry of such a source because the telescope beam does not observe the additional extended components. Then we could observe circular polarization of Sgr A* at 100 GHz using the Nobeyama 45-m telescope. The good polarization performances of this system should guarantee high-sensitive and high-precision polarimetry.

2 2.1

Observations Circular Polarization

We have observed the circular polarization of Sgr A* at 100 GHz, using the mm-wave reflective-type polarimater (Shinnaga, Tsuboi, and Kasuga. 1999) installed in the Nobeyama 45-m telescope in January 2002. Figure 1 shows a schematic display of the polarimeter. The observation was carried out in a quiescent phase of the 106-day quasi-periodic variation ( Zhao, Bower, and Goss 2001). The telescope has a small beam with a diameter of 16” at 100 GHz (Cf. 20” x 5”, Bower et al. 2001). The side lobe level of the beam was lower than -20 dB. The switching frequency of the beam-switch of the telescope is 15 Hz. The beam-throw of the beam-switch is AAZ = -6’ in the azimuthal direction. The intensity calibration was done by the chopper-wheel method. NGC7027 was also used as the intensity calibrator. The flux density at 100 GHz and the source size were assumed to be 4.8 Jy and 7”, respectively. The receiver front-end is the two orthogonal linear polarization receiver, ”S80/S100”. The system noise temperature during the observations was 230 K. * Corresponding author: e-mail: [email protected],Phone: +8129 228 8362, Fax: +8129 228 8362

@ 2003 WlLEY VCH Verlag GmbH & Co KGaA, Weinhrim

432

M. Tsuboi et al.: Circular Polarization toward Sagittarius A*

45d

135d

SlOO feed $CP<=>RCP 2 %S80 feed $CP<=>LCP

1/4h plate

Fig. 1 Schematic diagram of the mm-wave polarimeter (Shinnaga, Tsuboi & Kasuga, 1999) installed in the beam guide system of the Nobeyama 45-m telescope.

Dual circular polarization (LCP and RCP) can be received simultaneously using the receiver front-end and the mm-wave polarimeter. The polarimeter consists of a rotating 1/4X wave plate and convert the incident circular polarization to the receivable linear polarization. The location of the polarimeter is close to the Coude focus and in front of the SIS receiver. The observation consisted of two alternating cycles. In one cycle, S80 and SlOO receivers observe LCP and RCP, respectively. The telescope beam was pointed sequentially at the first off-source position at AAZ = -3', the on-source position atAAZ = 0' and the second off-source position at AAZ = +3" all for 20 s during each cycle. In another cycle, the S80 and Sl00 receivers observe RCP and LCP, respectively. The instrumental circular polarization of the polarimeter system was completely calibrated with observations of Venus as a non-circularly polarized source at 100 GHz. The calibration was carried out just before the observation of Sgr A*. Table 1 shows the results of observations of Venus (m, = -0.09&0.42% at EL= 30"). The error is statistical, mainly caused by pointing error and scintillation at daytime. We regard the instrumental circular polarization of the telescope to be smaller than 0.1% at least at night. For the system check, we have also observed a strong radio quaser 3C279 (see Table l), for which circular polarization has been detected at cm-wavelength (Wardle et al. 1998; m, = -1.0 & 0.2% at 15 GHz). The fractional circular polarization of 3C279 was m, = -0.16 f 0.31%. The circular polarization of 3C279

Astron. Nachr./AN 324, No. S 1 (2003)

433

was not detected at 100 GHz. Table 1 System Check of the Polarimeter.

Source Name

m.,(%) Am, (%)

Venus

-0.09

1t0.42

3C279

0.16

3~0.31

In the circular polarization observation, Sgr A* was detected with sufficient S N . The observed flux was, however, fluctuating due to pointing error and scintillation. Outputs of the S80 and Sl00 receivers are the sum of dual circular polarizations. The sum corresponds to Stokes I parameter, which should be constant in the condition of no fluctuation by pointing error and scintillation;

f

+ I(I1CP) =const.

= I(LCP)

Thus, we calibrated the data so that the sum becomes a constant value. The residual ambiguity of the circular polarization is 0.1% level after the calibration. 2.2

Linear Polarization

In addition, we have observed the linear polarization of Sgr A* at 100 GHz, using the mm-wave polarimater installed in the Nobeyama 45-m telescope in the same day. When we observed linear polarization at 100 GHz, we tuned the phase-shifter so that it is operated as a 1/2X wave plate at 100 GHz. The 1/2X wave plate converts from the orthogonal pair of the injected linear polarizations to the orthogonal pair which is receivable for the S80/S100 receivers. The rotator is controlled by the timing pulse from the telescope. A polarimetry observation consisted of nine sequences at every 22.5" from 6' = 0" to 180" of the polarimeter. If the signal from the object contains a component of linear polarization, the variation of the outputs must become a sinusoidal curve. Then it is also important for this polarimetry to correct the error by pointing error and scitillation. Because the S80/S100 receiver has two orthogonal linear polarization feeds, the sum of these receiver outputs is correspond to Stokes I parameter. This sum must be constant even at any position angle of the polarimeter . Thus,

f

= f(Q)

+ I ( 0 + 90')

=const.,

where B is position angle of the polarimeter system. We put on the factor at the receiver outputs so that the sum may become a constant value. We also checked the origin of the position angle with observations of 3C279 and 3C273, which are well observed, even at mm-wave range.

3 Results 3.1

Circular Polarization

We have observed Sgr A* twice in 12 January 2002. The two observations were separated by two hours. The fractional circular polarizations within a 16" beam toward Sgr A* were m, = 0.55 i 0.54%, and r n , = -0.61 0.44%, respectively. The statistical error was mainly caused by atomsperic scintillation because of the low elevation of Sgr A*. Although the difference may be significant compared with the statistical error, they are assumed here to be identical within the statistical error. The averaged value of the fractional circular polarization was m, = -0.15 & 0.35%. The flux density toward Sgr A* was I(Bd,, 5 16") = 4.5 Jy. On the other hand, Sgr A* itself is a variable source at millimeter wavelengths (Tsuboi, Miyazaki, and Tsutsusmi, 1999, Wright and Backer,

*

M. Tsuboi et al.: Circular Polarization toward Sagittarius A*

434

1993). The flux density observed by NMA was I(SgrA* itself) = 1.5 Jy within a week of the observation epoch. Then circular polarization of Sgr A* itself can be obtained by subtracting of the free-free emission surrounding Sgr A* within the 16” beam, which is estimated to be 3 Jy. The flux density with NMA shows the observation epoch may be in the quiescent phase of Sgr A*. Thus the observed circular polarization corresponds to an intrinsic circular polarization of Sgr A* itself of m, = -0.5 f 1.0% after the subtraction. This upper limit is consistent with that of the previous interferometric observation for the circular polarization of Sgr A*(mc 5 1.8%). Table 2 Circular Polarization of Sgr A* at 100 GHz.

Source Name

Flux density

m,(%) Am, (%)

Sgr A* ( 5 16”)

4.5

0.55

4~0.54

Sgr A*(< 16”)

4.5

-0.61

f0.44

Sgr A*(< 16”)

4.5

-0.15

~k0.35

Sgr A*

1.5

-0.5

&l.O

3.2

remarks

average

Linear Polarization

The fractional linear polarization within a 16” beam toward Sgr A* in 12 January 2002 was p = 1.7 Z!C 1.0%. Thus we did not detecte linear polarization of Sgr A* in the quiescent phase at a level of 5% after subtraction of the free-free emission contribution surrounding Sgr A*, as mentioned above. Table 3 Linear Polarization of Sgr A* at 100 GHz.

Source Name

Flux density

p(%)

A p (%)

Position angle (deg)

Sgr A* ( 5 16”)

4.5

1.7

k1.0

-20

Sgr A*

1.5

5

+3

-20

References Bower, G. C., Falcke, H. and Backer, D. C. 1999, ApJ, 523, No.2,29. Bower, G . C., Wright, M. C. H., Falcke, H., and Backer, D. C. 2001, ApJ, 555, No.2, 103 Shinnaga, H., Tsuboi, M., and Kasuga, T. 1999, PASJ, 51, 175 Wardle, J.F.C., et al. 1998, Nature, 395, 457 Zhao, J-H., Bower, G. C., and Goss, W. M. 2001, ApJ, 547, No.2,29 M. Tsuboi, M., A. Miyazaki, A., and T. Tsutsumi, T. 1999, in “The Central Parsecs of the Galaxy”, 186,eds H. Falcke, A. Cotera, W. J. Duschl, F. Melia, and M. J. Rieke.(ASP Conference Series), pp. 105 Wright, M. C. H., andBacker, D. C. 1993, ApJ, 417,560

Astron. Nachr./AN 324, No. S1,435-443 (2003) / DO1 10.1002/asna.200385043

Radiatively Inefficient Accretion Flow Models of Sgr A* Eliot Quataert* ' Astronomy Dept., M)1 Campbell Hall, UC Berkeley, Berkeley, CA 94720

I review radiatively inefficient accretion flow models for the= 2.6 x 10 6 M a black hole (BH) in the Galactic Center. I argue for a 'concordance model' of Sgr A*: both theory and observations suggest that hot ambient gas around the BH is accreted at a rate 10P8Ma yrP1, much less than the canonical Bondi rate. I interpret Chandru observations of Sgr A* in the context of such a model: (1) the extended 'quiescent' X-ray emission is due to thermal bremsstrahlung from gas in the vicinity of the Bondi accretion radius, and (2) the lo4 second long X-ray flares are due to synchrotron or Inverse-Compton emission by non-thermal electrons accelerated in the inner N 10 Schwarzschild radii of the accretion flow.

-

-

1 Introduction The case for a M 2.6 x 10611/10 black hole (BH) coincident with the radio source Sagittarius A* in the Galactic Center (GC) is now compelling (e.g., Schodel et al. 2002; Ghez et al. 2003). This only emphasizes the long-standing puzzle that the luminosity from the GC is remarkably low given the presence of a massive black hole. The resolution of this puzzle must lie in how gas from the ambient medium accretes onto the central BH. In these proceedings I review accretion models and their application to Sgr A*. A unique feature of the Galactic Center is our ability to constrain the dynamics of gas quite close to the black hole (relative to other systems), thus providing additional boundary conditions on, and much less freedom for, theoretical models. A canonical formulation of these constraints is the Bondi accretion estimate for the rate at which the BH gravitationally captures surrounding gas (Bondi 1952; see Melia 1992 for an early application to Sgr A*). Given relatively uniformly distributed matter with an ambient density po and an ambient sound speed Q, the sphere of influence of a BH of mass M extends out to R,,, M G M / c ~The . accretion rate of this gas onto the central BH, in the absence of angular momentum and magnetic fields, is then A?B M ~ r R ~ , , p o ~ . Chundru observations of the GC detect extended diffuse emission within 1 - 10" of the BH (Baganoff et al. 2003a). This emission likely arises from hot gas produced when the stellar winds from massive stars in the GC collide and shock (e.g., the He I cluster; Krabbe et al. 1991). Interpreted as such, the inferred gas density and temperature are M 20 cmP3 and M 1.3 keV on 10" scales, and M 100 cmP3 and M 2 keV on M 1" scales (see also Fig. 3). The corresponding Bondi accretion radius is R,,, M 0 . 0 4 ~=~1" and the Bondi accretion rate is & f ~ M 10P5M0 yr-'.' If gas were accreted at this rate onto the BH via a geometrically thin, optically thick accretion disk (Shakura & Sunyaev 1973), a model that has been extensively and successfully applied to luminous accreting sources (e.g., Kortakar & Blaes 1999), the expected luminosity would be L M 0 . l h k ~ c ' 1041 ergs spl, larger than the observed luminosity by lo5. This is the strongest argument against a thin disk in Sgr A*. An additional argument a factor of is the absence of any disk-like blackbody emission component in the spectrum of Sgr A*. If the putative N

N

* e-mail: [email protected]; Phone: 51 0-642-3792

' This is much less than the total mass loss rate from stars in the GC ( M 10-3Ma yr-';

Najarro et al. 1997). implying that

there should also be a global outflow of hot gas from the central parsec (a GC 'wind').

@ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

E. Quataert: Accretion Models for Sgr A*

436

disk were to extend all the way down to the BH, its accretion rate would have to be 5 10-loMo yrpl M l O P 5 A ? ~ to satisfy infrared limits (e.g., Narayan 2002; see his Fig. 2). One possible caveat to the Bondi analysis is that there is far more (by mass) cold molecular gas than hot X-ray emitting gas in the central 1 - 10 parsecs of the GC (e.g., Herrnstein & Ho 2002). It is unclear how close to the BH the molecular gas extends and whether it is important for the dynamics of gas accreting onto the BH. In what follows I ignore this component, but see Nayakshin (2003) for a different view. As emphasized above, the inferred low efficiency of Sgr A* is the strongest argument against accretion proceeding via a thin accretion disk. Instead, the observations favor models in which very little of the gravitational potential energy of the inflowing gas is radiated away. I will refer to such models as radiatively inefficient accretion flows (RIAFs). In the next section (52) I summarize the properties of RIAFs. I then apply these models to the GC (53), emphasizing the interpretation of radio and X-ray observations of Sgr A*. Finally, I conclude with a brief summary ($4).

2 Radiatively Inefficient Accretion Flows RIMSdescribe the dynamics of rotating accretion flows in which L << 0.11dc2, i.e., very little energy generated by accretion is radiated away (e.g., Ichimaru 1977; Rees et al. 1982; Narayan & Yi 1994). Instead, the gravitational potential energy released by turbulent stresses in the accretion flow is stored as thermal energy. As a result, the accreting gas is very hot, with a characteristic thermal energy comparable 1OI2 to its gravitational potential energy; close to the BH this implies T GMmp/3kR 0.1m,c2/k K. At such temperatures, and for gas densities appropriate to systems like the GC, the Coulomb collision time is much longer than the time it takes gas to flow into the BH. The accretion flow then develops a twotemperature structure with the protons likely hotter than the electrons: Tp loi2 K 2 T, 1O’O - 10l2 K. The precise electron temperature is uncertain but important since electrons produce the radiation that we see. The electron temperature depends on how and to what extent they are heated by processes such as shocks, MHD turbulence, and reconnection (see, e.g., Quataert & Gruzinov 1999). Note that because collisions are unimportant one would not expect the electron distribution function to be thermal. Advection-Dominated Accretion Flows (ADAFs) are a simple analytical model for the dynamics of RIAFs; they predict that the structure of the flow is in some ways similar to spherical Bondi accretion, despite the fact that angular momentum and viscosity are important (e.g., Ichimaru 1977; Narayan & Yi 1994). In ADAF models the gas rotates at s1 M 0.3 - 0.5 CIK, where OK = ,/is the Keplerian rotation rate. Because the flow is hot, pressure forces are also important and the inflowing gas is geometrically quite “thick,” with a scale height H M R at every radius. The radial velocity in the flow ~ a is the dimensionless viscosity parameter, c, is the sound speed in is given by U R M acs M a u where the flow and W K = RRK 0: RP1/’. Conservation of mass on spherical shells then implies that the density scales as p oc RP3/’,the characteristic scaling for spherical accretion. ADAF models also predict that, even in the presence of rotation, the rate a? which gas accretes onto the BH from an ambient medium is comparable to the Bondi accretion rate: MADAF M B . Thus ~ in ADAF models the low luminosity of SgrA* is due to a very low radiative eficiency With the advent of global, time-dependent, numerical simulations of accretion flows, it has become possible to numerically simulate RIAFs and test the ADAF predictions. Note that RIAFs are, in one sense, the easiest flows to simulate since (1) no treatment of radiation or radiative transfer is needed and (2) the flow is “thick” with H R, so there is no difficult-to-simulate separation of scales like in a thin disk. On the other hand, for a system like the GC, the proton-electron collision time close to the BH is 6 orders of magnitude longer than the inflow time of the gas. Thus the fluid approximation used by all simulations

-

-

-

-

-

--

-

-

A more accurate estimate might be MADAFN an;r8 (e.g., Narayan 2002). The factor of (Y arises because the inflow velocity of gas at the Bondi accretion radius is N acg in ADAF models while it is % cs in Bondi models. Thus for a fixed density in the ambient medium, the accretion rate in an ADAF will be smaller by a factor of a.

Astron. Nachr./AN 324, No. Sl (2003)

431

to date is suspect (a kinetic treatment should be used; see Quataert et al. 2002). Whether this introduces qualitative or merely quantitative errors in the results is unknown. I suspect the latter. The key result from nearly all simulations to date is that h’ << MB,i.e., very little mass available at large radii uctually accretes onto the black hole (e.g., Stone, Pringle, & Begelman 1999; Igumenshchev & Abramowicz 1999; 2000; Igumenshchev et al. 2000; Stone & Pringle 2001; Hawley & Balbus 2002; Igumenshchev et al. 2003). Another way to state this result is that the radial density profile in the flow is much “flatter” than the ADAF predictions: for a given gas density at large distances from the black hole (e.g., measured by Chandra on I” scales in the GC), the density close to the BH is much less than the ADAF or Bondi predictions. Following a proposal due to Blandford & Begelman (1999), we can parameterize the density profile of RIAFs with a parameter p , where p 0: RP312+P.With this parameterization, the rate where R,, Rs is the inner radius at which gas is actually accreted into the BH is (R,,/ROut)Ph’~, of the flow and Rout R,,, is the outer radius. Values of p M 1/2 - 1, rather than p = 0 predicted by ADAF models, are favored by the simulations. This implies that, within the context of RIAF models, a low , than just a low eficiency, is a major factor in the faintness of SgrA*. accretion rate, ilk << A ~ Brather

-

N

Since the focus of this review is the application of RIAF models to Sgr A*, I will not dwell on why the simulations differ so significantly from the ADAF predictions. A brief discussion is, however, in order. The following material will not be used in later sections so uninterested readers can move directly to 53. In RIAFs, the inflowing gas is heated at a rate M 0.1Akc2, as required by the release of gravitational potential energy in a differentially rotating accretion flow. In ADAF models this energy is stored as thermal energy and carried into the BH. As noted above, the inflowing gas is then very hot with a sound speed at any radius comparable to the escape speed from the BH’s potential well. ADAF models are therefore prone to developing outflows (see Narayan & Yi 1994 or Blandford & Begelman 1999 for a more formal discussion). This led Blandford & Begelman (1999) to propose that, in the absence of radiation, the gravitational binding energy of the accreted gas must be lost through some other (non-radiative) means. Otherwise the inflowing gas is not sufficiently bound to the BH to accrete. The numerical simulations to date are broadly consistent with this hypothesis; e.g., the gas temperature in the simulations is generally a factor of few-5 less than in ADAF models implying that the gas is indeed more strongly bound to the BH. The non-radiative energy loss can take one of two forms: (1) efficient turbulent transport of energy through the accretion flow to large radii or (2) a global outflow (’wind’) that carries away the binding energy of the accreted matter. In hydrodynamic simulations (e.g., Stone et al. 1999; Igumenshchev & Abramowicz 1999) or in MHD simulations with relatively ’weak’ magnetic fields (p 2 10 - 100, where p is the ratio of the gas pressure to the magnetic pressure), convective transport of energy and angular momentum dominates the dynamics of the accretion flow (Narayan et al. 2002; Igumenshchev et al. 2003; see Quataert & Gruzinov 2000a and Narayan et al. 2000 for such ’convection-dominated accretion flow’ models). The convective luminosity through a spherical shell at radius R is IX p$R2 where w, is the turbulent convective velocity. Since the flow is hot, w, c( c, IX OK cx RP1l2. A constant flow of gravitational binding energy from small to large radii therefore requires p 0: RP112;i.e., p = 1 instead of p = 0 as in ADAF modek3 In contrast to the hydrodynamic results, in MHD simulations with strong magnetic fields (/j’S lo), MHD turbulence dominates the flow dynamics and convection is unimportant (e.g., Stone & Pringle 2001; Hawley & Balbus 2002; Igumenshchev et al. 2003). Stone & Pringle (2001) and Hawley & Balbus (2002) find that most of the inflowing gas is lost to a magnetically driven wind. Igumenshchev et al. (2003), on the other hand, find a much more complex flow configuration, though still with a very small accretion rate.

The convective energy flux could launch a thermally driven wind from large radii GcC or from the surface layers of the accretion flow. Thus there i s not necessarily a clean distinction between global energy transport by turbulence and mass outflow as mechanisms for ’non-radiative’ energy loss. Both processes are related and, in particular, the former can drive the latter. N

438

E. Ouataert: Accretion Models for Sm A*

3 RIAF Models Applied to Sgr A* A number of authors have used RIAF models, and in particular ADAF models, to explain the observed properties of Sgr A* (e.g., Narayan et al. 1995, 1998; Manmoto et al. 1997). These calculations have ah?^, can roughly acshown that an ADAF model accreting at the observationally inferred rate, n;l count for the observed luminosity and spectrum of Sgr A*. The key constraint is that the fraction of the turbulent energy that heats the electrons, 6, must be small 2 0.01, so as to not overproduce the observed luminosity. Equivalently the electron temperature close to the BH must be ,$ 10'' K << Tp x lo1' K. An example of such a model in shown by the dotted line in Figure 1. The model roughly reproduces the observed sub-mm emission, satisfies the IR limits, and produces an X-ray luminosity comparable to that seen by Chundru in the quiescent (non-flaring) state (I discuss the Chundru observations in more detail below). It does, however, significantly underproduce the lower frequency radio emission. Since the lower frequency radio emission in Sgr A* is phenomenologically similar to that of other AGN, a natural interpretation is that a jet is present and produces the radio emission (e.g., Falcke & Markoff 2000, Yuan et al. 2001, and references therein). Alternatively, the results in Figure 1 assume purely thermal electrons, for which there is no good justification. As Figure 2 shows, even a small population of nonthermal electrons in the accretion flow can produce the radio emission (e.g., Mahadevan 1998; Ozel et al. 2000).

-

=

37

i

-

. . . , . . . , . . . , . . . , . . .

/

. . , .

SgrA*

f

36

h

F 34 a'

1

2

Y

M

2 32

30

log,,[4WI

Fig. 1 Two 'baseline' RIAF models of Sgr A* that reproduce the quiescent Chandra flux (the Chandra spectrum is discussed in Figs 2 and 3). Dorted line: a RIAF model with p = 0, Ak M k~and 6 = 0.01 (6 = fraction of turbulent energy heating electrons). This is an ADAF-type model. Solid line: a RIAF with p = 0.5 and a net accretion rate into the BH of M % 10-8Ma yr-' << A?B; 6 = 0.5. This Figure is based on Quataert & Narayan (1999).

Fig. 2 Dot-dashed A RIAF with p % 0.4 and 6 % 0.5. There is also a power law distribution of electrons with n ( y ) 0: y-3.5 and % 2% of the electron thermal energy. Power-law electrons produce the low freq. radio emission not accounted for in Fig. 1. Dashed line: Inverse-Compton model for the Chandra X-ray flare (see text for details). Solid line: Total emission. X-ray error bars are (from top to bottom): Oct. 2000 flare, average flare, & quiescent emission. This Figure is based on Yuan et al. (2003).

Quataert & Narayan (1999) showed that a much broader class of RIAF models could also account for the observed properties of Sgr A*. Specifically, the low accretion rate models favored theoretically (52) can reproduce the observations as well. The key requirement is that the electrons must be hotter so as to produce more emission even though k and the gas density are lower. An example of such a model is shown by the solid line in Figure 1: p = 0.5 from Rout = 105Rs down to R,, = R s , implying

Astron. Nachr./AN 324, No. S1 (2003)

439

-

that the accretion rate into the BH is much smaller than the Bondi rate (- lO-*Ma yr-l). The electron temperature close to the BH is 10’l K, rather than N 1O1O K as in ADAF models. To first approximation, the two models in Figure 1 reproduce the observed spectrum of Sgr A* equally “well” (or poorly, depending on one’s vantage point). Howevel; Aitken et al. (2000)and Bower et al.’s (2003) detection of w 10% linear polarization in the sub-mm (230 GHz) emission from Sgr A* argues strongly for low accretion rate models with M 5 10-*MQ yr-’. Beckert & Falcke (2002,2003; see also Ruszkowski & Begelman 2002) present detailed models for the radio polarization of Sgr A* (both linear and circular). Here I summarize the constraints imposed by the A ~ implies B a much higher gas density and magnetic field linear polarization detection: accretion at strength close to the black hole than accretion at << M B . Faraday rotation is therefore much stronger N A?B, the rotation measure is 2 (e.g., Quataert & Gruzinov 2000b; Ago1 2000). In models with 1O1O rad rnp2 in the region of the flow (210-100Rs) where the sub-mm emission is produced. This leads 100 GHz. This large rotation angle implies that intrinsically to a Faraday rotation angle lo5 rad at linearly polarized synchrotron emission would be depolarized propagating to the observer over most of the radio to infrared spectrum. By contrast, Bower et al. (2003) find that R M 5 lo6 rad mp2 in their linear can satisfy this constraint because the density polarization detection at 230 GHz. Models with hk << and magnetic field strength close to the BH are much smaller. For example, for M w 10p8Ma yr-’, the net rotation measure through the accretion flow is R M x lo6 rad m-’, consistent with the observational constraint. Thus the observed detection of linear polarization in the sub-mm emission of Sgr A* argues for a low accretion rate 5 10-*Ma yr-I << A ~ B . One way out of this conclusion is to posit that the magnetic field undergoes so many reversals along the line of sight that the net Faraday rotation is 2 lo5 times smaller than these simple estimates (Ruszkowski & Begelman 2002). I regard this as very unlikely, but am not aware of a direct observational argument against this possibility at the present time. In light of the above arguments the rest of my discussion centers on models with M << these ’concordance’ models are both theoretically favored and satisfy the rotation measure constraint from the linear polarization detection. I focus on interpreting the Chandru X-ray observations of Sgr A*. Much of this material is based on Quataert (2002) and Yuan, Quataert, & Narayan (2003). Chandra observations reveal that there are two components to the X-ray emission coincident with Sgr A* (Baganoff et al. 2001,2003ab): (1) a ’baseline’ X-ray flux with a nearly constant luminosity w 2 x ergs s-’ and a soft spectrum (photon index r w 2.7, where vL, cx Y ~ - ~ This ) . component is clearly extended with a size of w 1” w lo5& and does not vary in time. (2) X-ray ’flares’ occurring at a rate of x 1 day-’ and lasting M lo3 - lo4 s. The luminosity increases by a factor of few-100 during the flare and the spectrum is quite hard (r M 1.2). The flares are not extended; in fact, the observed timescales argue that they arise close to the BH, at 6 10 - 100Rs. N

N

N

MB

MB:

3.1

X-ray flares

The X-ray flares are the most dramatic result from the Chandra observations. Markoff et al. (2001) showed that the flares are probably due to electron heating or acceleration, rather than a change in the accretion rate onto the BH (otherwise there is too much variation in other wavebands). An obvious analogue is magnetic reconnection in solar flares, which one could readily imagine occurring in the inner part of the accretion flow close to the BH. Given an injection of energy into the electrons, there are three emission mechanisms that could, a prior&give rise to flares: (1) bremsstrahlung, (2) synchrotron, and ( 3 )inverse Compton. Bremsstrahlung is attractive because it can naturally explain the very hard spectrum of the flares. The problem is that, to produce a luminosity of L351035ergs spl from a sphere of radius R, the gas density 112 114 cmp3, where T,,lo = T,/1O1OK. For comparison, the ambient must be n M 109L3, Te,,o(R/10Rs)-3~2 ~ ~ . importantly, density in the inner lORs for a model with Ak x 10-*Ma yr-’ is M lo6 ~ 1 1 3 Equally bremsstrahlung emission in RIAFs is dominated by very large radii R,,,, not small radii (e.g., Quataert & Narayan 1999; see below). Thus bremsstrahlung appears unlikely to be responsible for the X-ray flares. N

E. Ouataert: Accretion Models for Ser A*

440

Two possible ways out of this conclusion are (1) to posit that the gas density increases by a factor of lo3 during the flare, and does so preferentially in the very inner part of the accretion flow (e.g., Liu & Melia 2002). It is unclear, however, what would drive such large density changes, particularly since the cooling time is so long that thermally driven instabilities are unlikely to be important, (2) perhaps the accreting gas is a two-phase medium, with a cooler, denser phase giving rise to the X-ray flares (Yuan et a]. 2003). In contrast to bremsstrahlung, synchrotron emission can readily account for the observed flares: if 10% of the electrons (by energy) are accelerated into a power-law tail in the inner N 10Rs, the radio-IR lo5 can produce the X-ray emission is essentially unchanged while electrons with Lorentz factors y emission (e.g., Markoff et al. 2001). Moreover, there is a ’natural’ explanation for the hard X-ray spectrum seen. For lUAF models with M w 10-8Ma yr-’, the magnetic field strength close to the BH is = 20Bzo G. The associated synchrotron cooling time for electrons emitting in the Chandra band is w 20BG3” s. Thus, unless B 5 0.3 G in the emitting region, the cooling time is less than the duration of the flare and there should be a cooling break in the electron distribution function below the Chandra band. For an injected distribution of power-law electrons with p , < 2, where n ( y ) K y-pe, the distribution function for the population of cooled electrons is n(y) cc yp2.This implies a synchrotron spectrum with I’ = 1.5, consistent with the typical flare observed by Chandra (Baganoff et al. 2003b). Synchrotron self-Compton emission can also explain the X-ray flares observed by Chandru. The dashed line in Figure 2 shows a concrete example in which most of the electrons in a M 3Rs volume are accelerated into a power law distribution. This population of electrons produces synchrotron emission and also upscatters synchrotron photons to produce a hard X-ray flare. Note that there is very little change to the radio or IR emission (the dot-dashed line in Fig. 2 shows the baseline non-flaring model in which power-law electrons have only w 2% of the electron thermal energy). Finally, it is important to stress that in the models discussed here, the duration of the flare is set by lORs of the accretion flow. By contrast, there is no a dynamical or viscous timescale in the inner explanation for the mean time between flares, w 1 day. In addition, it is difficult to apply the ideas considered here to the week-long, factor of few, sub-mm ’flares’ observed by Tsuboi et al. (1999) and Zhao et al. (2003). In particular the duration of these ’flares’ is inconsistent with the characteristic timescales in the sub-mm emitting region ( S 10Rs). One possibility is that they are not directly related to the Chandra flares but are instead due to small fluctuations in hl set by dynamics in the accretion flow at larger radii. N

N

N

-

3.2 Quiescent X-ray Emission The steady quiescent emission observed by Chandra is qualitatively different from the flaring emission. In particular, it is softer (r w 2.7) and extended (w1” w 105Rs). The latter fact implies that it is a different emission component since synchrotron and inverse Compton emission are produced at small radii. lUAF models naturally predict that the thermal bremsstrahlung emission is dominated by large radii in the flow (e.g., Quataert & Narayan 1999; Ozel & Di Matteo 2000). Given a density profile of the form p a R-’/’+p and a temperature profile T K R-’ (valid at large radii), the bremsstrahlung luminosity assuming photon energies 5 kT(R).The is dominated by large radii: L a R3p2T-l/’ c( R’p+’/’, resulting spectrum, adding up all radii, is r w 3/2 2p, i.e., vLu K Y ’ / ~ - ~ P . Thus a natural interpretation of the quiescent flux coincident with Sgr A* is that it is bremsstrahlung from hot gas in the outer part of the accretion flow that is resolved by Chandra (e.g.. Yuan et al. 2001; Quataert 2002). This can account for the size of the source, its lack of variability, and the possible presence of a thermal X-ray line (Baganoff et al. 2003b). The above expression for the spectrum of the thermal emission would imply that p w 1/2 is required to explain the spectrum. There is, however, an important problem with this straightforward interpretation: Chandra spectra are extracted in a w 1”region around Sgr A*. This is comparable to the Bondi accretion radius, Rae, ($1). It is thus incorrect to assume that observations of the extended emission directly probe the “accretion flow.” Instead, they probe the complex “transition region” between the accretion flow and the ambient medium.

+

441

Astron. Nachr./AN 324, No. S 1 (2003)

6

1

I

I

-2

-1

,

,

,

,

l

,

0 log,,[R/R,_I

I

I

I

I

1

,

,

I

,

2

-4

-2

-1

0

1

2

log,,[E(kev)l

Fig. 3 Left: Observationally motivated models for n(R)and T ( R )for hot gas around Sgr A*; models are extrapolated to R < R,,, M 1" using Bondi accretion. Right: X-ray spectra based on the density and temperature profiles in the = 1" prediction is consistent with the quiescent X-ray emission coincident with Sgr A*. left panel. The

Figure 3 illustrates the effects of a finite observing beam relative to the Bondi accretion radius; the left panel shows a toy model for the density and temperature as a function of radius around Sgr A* (in units of R,,, M 1").The temperature is measured to be M 1keV at large radii. The density profile on k R,,, scales is adjusted to roughly reproduce the radial variation of the observed diffuse X-ray emission (see Quataert 2002 for details). For R 5 R,,, gas accretes as a Bondi How, with asymptotic ( R < R,,,) scalings p a R-3/2 (i.e., p = 0) and T a Rp1.4 Figure 3b shows the X-ray spectra that would be seen by a telescope with a beam-size Rbeam.For large beams, RbeamM lo", the spectrum is very soft and is dominated by the ambient medium that has T M 1 keV. For small beams, R b e a m G 0.1", the spectrum is dominated by the accretion flow and is quite hard, consistent with the above scalings. For the case applicable to Sgr A*, RbeamM 1" M R,,,, the emission is still relatively soft,consistent with the Chandru observations. This is in spite of the fact that the underlying accretion flow produces a hard X-ray spectrum. The upshot of Figure 3 is that the extended quiescent X-ray source observed by Chundra appears broadly consistent with the emission produced by gas on R,,, M 1" scales. This is gas in the 'transition region' between the ambient medium and the accretion How, rather than the accretion flow itself. Using these results to constrain the dynamics of the accreting gas, e.g., the radial density profilep, will require (1) better theoretical models for the dynamics of the X-ray emitting gas on 1" scales, and (2) tighter observational constraints, such as spectra as a function of radius.

4

Conclusions

The Galactic Center represents a unique opportunity to probe the dynamics of gas accreting onto a massive black hole, from the 'large' scales on which gas is gravitationally captured by the BH to the 'small' scales close to the BH's horizon. In this review, I have tried to argue that radiatively inefficient accretion How (RIAF) models can provide a reasonably coherent picture of accretion onto Sgr A* on all of these scales. 1 chose a Bondi flow for two reasons: (1) Bondi flows predict hard X-ray brernsstrahlung spectra and are thus a good test case for assessing what effect the sop X-ray emitting ambient medium around the BH has on detecting the accretion component, (2) There are no good dynamical models for how rotating RIAFs "match onto an ambient medium outside R,,,.

E. Quataert: Accretion Models for Sgr A*

442

To summarize: hot gas in the central M 1pc of the GC is produced by the (shocked) winds from massive stars. This gas is gravitationally captured by the black hole on scales of R,,, M 1” = 0.04 pc. The Chandru detection of an extended soft X-ray component coincident with Sgr A* may be direct evidence for this gravitationally captured gas (53.2). The net rate at which gas accretes through the BH’s horizon is, however, much less than the canonical Bondi estimate for the rate at which gas is gravitationally captured by M~ 10-5M0 yr-l). The Bondi estimate neglects the angular momentum of the accreting gas, the BH ( k which is likely to be a very poor assumption. Numerical simulations (and analytical models) of rotating RIAFs find that h << h~(52). This conclusion is theoretically attractive because it implies that the radiaas in Bondi and ADAF models (which has always tive efficiency of Sgr A* need not be as low as been difficult to reconcile with the expectation that electron heating and acceleration would be important in the collisionless magnetized plasma close to the BH). A low accretion rate is also strongly suggested by the detection of linear polarization in the sub-mm emission from Sgr A*: Faraday rotation constrains the gas 10W8Ma yr-I << M B , in good density and magnetic field strength close to the BH and argues for n;f 10-sMa yr-’ can agreement with the inference from R I M models. Finally, RIAF models with M explain the basic spectral properties of Sgr A* (see, e.g., Figs. 1 & 2). In particular, the X-ray flares seen by Chandra may be due to synchrotron or Inverse-Compton emission produced by relativistic electrons accelerated in the inner N l0Rs of the accretion flow (53.1). It is important to stress that, although ilk << hi^ is both theoretically and observationally favored, all of the models considered here are still “radiatively inefficient,” and have efficiencies much less than the canonical thin disk value of 10%;e.g., for M 10-’MO yr-’, L N 10-3Akc2.In fact, all of the physics highlighted in $2 that suppresses the accretion rate with respect to the Bondi estimate requires a relatively low efficiency and would not operate in a thin disk. There are several important issues that I have not addressed in this review. To cite two that clearly require further study: (1) are the Chundra flares due to, e.g., reconnection or turbulence in the accretion flow, or are they telling us something more fundamental about the dynamics close to the BH? (2) Both jet and RIAF models can explain the basic spectral properties of Sgr A*. How can we distinguish between these two components, both of which are almost certainly present? E.g., is the linear polarization detection, which requires a relatively coherent magnetic field, consistent with the magnetic field seen in RIAF simulations? Or does it instead require an additional ’jet’ component?

-

-

N

N

Acknowledgements I thank Ramesh Narayan and Feng Yuan for useful discussions. This work is supported by NASA Grant NAG5-12043, NSF Grant AST-0206006, and an Alfred P. Sloan Foundation Fellowship.

References Aitken, D. K. et al., 2000, ApJ, 534, L173 Agol, E., 2000, ApJ, 538, L121 Baganoff, F. et al., 2001, Nature, 413, 45 Baganoff, F. et al., 2003a, ApJ, in press Baganoff, F. et al., 2003b, these proceedings Beckert, T. & Falcke, H., 2002, A&A, 388, 1106 Beckert, T. & Falcke, H., 2003, these proceedings Blandford, R.D. & Begelman, M.C., 1999, MNRAS, 303, L1 Bondi. H., 1952, MNRAS, 112, 195 Bower, G., Wright, M. C. H., Falcke, H., & Backer, D., 2003, ApJ, in press (astro-pW0302227) Falcke, H & Markoff, S., 2000, A&A, 362, 113 Ghez, A. et al. 2003, these proceedings Hawley, J. F. & Balbus, S. A,, 2002, ApJ, 573, 738 Herrnstein, R. M. & Ho, P. T., 2002, ApJ, 579, L83 Ichimaru, S. 1977, ApJ, 214,840 Igumenshchev, I. V., & Abramowicz, M. A. 1999, MNRAS, 303,309 Igumenshchev, I. V., & Abramowicz, M. A. 2000, ApJS, 130,463 Igumenshchev, I. V., Abramowicz, M. A., & Narayan, R. 2000, ApJ, 537, L27 Igumenshchev, I. V., Abramowicz, M. A., & Narayan, R. 2003, ApJ submitted (astro-ph/0301402)

Astron. Nachr./AN 324, No. S1 (2003)

443

Koratkar, A. & Blaes, 0. 1999, PASP, 111, 1 Krahbe, A,, Genzel, R., Drapatz, S., & Rotaciuc, V., 1991, ApJ, 382, L19 Liu, S. & Melia, F., 2002, ApJ, 566, L77 Mahadevan, R., 1998, Nature, 394,651 Manmoto, T., Mineshige, S., & Kusunose, M., 1997, ApJ, 489, 791 Markoff, S., Falcke, H., Yuan, F., & Biermann, P. L., 2001, A&A, 379, L13 Melia, F., 1992, ApJ, 387, L25 Narayan, R. 2002, in Lighthouses of the Universe, ed. M. Gilfanov, & R. Sunyaev (Heidelberg: Springer-Verlag), in press, astro-ph/0201260 Narayan, R., Igumenshchev, I. V., & Abramowicz, M. A. 2000, ApJ, 539,798 Narayan, R., Quataert, E., Igurnenshchev, I. V., & Ahrarnowicz, M. A. 2002, ApJ, 577, 295 Narayan, R. et al., 1998, ApJ,492,554 Narayan, R., Yi, I., & Mahadevan, R., 1995, Nature, 374,623 Narayan, R., & Yi, I. 1994, ApJ, 428, L13 Najarro, F. et al., 1997, A&A, 325,700 Nayakshin, S., 2003, these proceedings Ozel, F. & Di Matteo, T., 2001, ApJ, 563, 276 Ozel, F., Psaltis, D., & Narayan, R., 2000, ApJ, 541, 234 Quataert, E., 2002, ApJ, 575, 855 Quataert, E., Dorland, W., & Hammett, G., 2002, ApJ, 577, 524 Quataert, E. & Gruzinov, A,, 1999. ApJ, 520, 248 Quataert, E. & Gruzinov, A,, 2000a, ApJ, 539.809 Quataert, E. & Gmzinov, A., 2000b, ApJ, 545,842 Quataert, E. & Narayan, R., 1999, ApJ, 520, 298 Rees, M. J., Begelman, M. C., Blandford, R. D., & Phinney, E. S., 1982, Nature, 295, 17 Ruszkowski, M. & Begelman, M. C., 2002, ApJ, 573,485 Schodel, R. et al., 2002, Nature, 419, 694 Shakura, N. I., & Sunyaev, R. A., 1973, A&A, 24,337 Stone, J. M., Pringle, J. E., &Begelman, M. C. 1999, MNRAS, 310, 1002 Stone, J. & Pringle, J. E., 2001, MNRAS, 322,461 Tsuhoi, M., Miyazaki, A., & Tsutsurni, T., 1999, in The Central Parsecs of the Galaxy, ASP Conference Series, ed. Heino Falcke et a]., Vol, 186, p. 105 Yuan, F., Markoff, S., & Falcke. H., 2002, A&A, 383, 854 Yuan, F., Quataert, E., & Narayan, R., 2003, in prep. Zhao, J. et al., 2003, ApJ Letters in press (astro-pM0302062)

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1,445-451 (2003)/ DO1 10.1002/asna.200385070

Jet Models for Flaring in Sgr A* Sera Markor' and Heino Fateke*

' MIT, Center for Space Research, 77 Massachusetts Ave., NE80-6035, Cambridge, MA 02139, USA

* Max-Planck-Institutfur Radioastronomie,Auf dem Hiigel69, D-53 121 Bonn, Germany

Key words Galaxy: center - galaxies: jets - X-rays: galaxies - radiation mechanisms: non-thermal accretion, accretion disks - black hole physics

--

Abstract. Last year, the X-ray mission Chundru observed a factor of 50 flaring in the Galactic Center supermassive black hole candidate Sgr A*. The smallest timescale of 600 seconds argues for an origin within 40 rg of the central engine, which narrows the source down to the inner regions of either the accretion flow or jet outflows. Because the variability in the radio has never shown similar levels of flaring over its observed history, a simplejump of order 50 in the accretion rate is likely not the cause of the fluctuations. We argue that additional electron heating near the base of a jet can account for the flaring which, depending on the underlying physical mechanism, predicts significantly different simultaneous multi-wavelength behavior. For the c a e of direct heating of the thermal electrons by a factor of a few, the flare would be a result of increased synchrotron-self Compton emission. Non-thermal particle acceleration, on the other hand, could account for the flare via high-energy synchrotron emission. We discuss these mechanisms in the context of the jet model, and the respective signatures which would allow us to distinguish between them in future flaring events. We briefly discuss these predictions in light of the newest flare observations. N

1 Introduction Sgr A*, as evidenced by this workshop, is a very well-studied source and thus ideal for testing our models for the physical processes near compact objects. We can now almost certainly associate our Galactic center radio core with a supermassive black hole of 3 x lo6 M a , thanks to the long-term study of its proper motion and the stellar dynamics in the central parsec (see e.g., Eckart et al.; Ghez; Ott et al.; Reid et al.; and Schodel et al. 2003). For as long as we could see them, we have believed these stars to be the source of material fueling Sgr A*'s emission, but estimates of the available amount of matter greatly overpredict how much emission we would see if it all made its way to the central regions (e.g., Falcke & Melia 1997). These predictions became even more constrained with the discovery of Sgr A* as only a weak X-ray source by Chundru (Baganoff et al. 2003a,b). Therefore most accretion models for Sgr A* have invoked low-efficiency, and comprise Bondi-Hoyle accretion (e.g., Melia 1YY2), ADAFs (e.g., Narayan et al. 1998) and variations (e.g., Yuan, Quataert & Narayan 2003), CDAFs (Quataert & Gruzinov 1999), and ADIOS (Blandford & Begelman 1999) models. Alternatively, there are also models which consider the emission of an outflow (e.g., Falcke & Markoff ZOOO), or combinations of both scenarios (Yuan, Markoff & Falcke 2002). For a detailed review of most of these various pictures, see Melia & Falcke (2001). Unfortunately, there is an obvious theoretical degeneracy in that several different models can explain the broadband spectrum fairly well. One way of narrowing things down has recently been provided by the linear polarization (LP) measurements of, e.g., Bower (2003) and Bower et al. (2003), which limit the N

* Corresponding author: e-mail: seraQspace.rnit.edu, supported by an NSF Astronomy & Astrophysics Postdoctoral Fellowship.

@ 2001 WILEY-VCH Veda%GmbH & Co KGaA. Weinhem

S. Markoff and H. Falcke: Jet Models for Flaring in Sgr A*

446

mass accretion rate to be 5 l o p 7 Ma yr-’, suggesting that radiative inefficiency may not be necessary after all, and ruling out Bondi-Hoyle accretion and ADAFs in their standard form. But there are still at least a few models in the running, and we need to look at other ways of discerning between them. The X-ray flares detected by Chandru and XMM-Newton (Baganoff et al. 2001; Baganoff et al. 2003b; Goldwurm et al. 2003a,h) have opened up a whole new face of Sgr A*, which existing models for the quiescent state must now be able to accommodate. We will focus here on the consequences for the jet model of Sgr A*.

2 Why consider a jet? It is worth spending a few paragraphs on the motivation behind a jet model, since to date no jet has actually been resolved in the core using VLBI. There is significant “circumstantial evidence”, however, and we will point out the most compelling here:

1. Recent studies suggest that the radio spectra of low-luminosityAGN (LLAGN), of which Sgr A* is the weakest known, are jet-dominated. The last few years have seen significant work surveying nearby LLAGN with high-precision instruments (e.g., Nagar, Wilson & Falcke 2001; Nagar et al. 2002). They are finding that most radio cores are accretion-powered (i.e., not starburst), with a flat-toinverted spectrum. These are naturally reproduced in jet models (originally described in Blandford & Konigl 1979), but could also result from ADAFs with strong outflow or CDAFs with higher luminosities. However, jets are the favored mechanism for a few reasons. First of all, in the brightest sources the jets are always resolved, and when they are, the jet dominates the unresolved core by at least a factor of three. More generally, but perhaps more importantly, we can argue for applying Occam’s Razor (the simplest solution is likely the right one). Why would the sources with flathnverted spectra which we can resolve have jets, while all the ones we cannot resolve-but which have the exucrly the same spectral characteristics-be accretion related? It seems far more likely that the jets are simply not resolved in these cases. 2. Sgr A* and M81* are almost “twins”, and M81* has a jet. The above point was fairly general. For Sgr A* specifically there is a more interesting line of argumentation. M81* is our nearest LLAGN besides Sgr A*, and resides in the same kind of spiral galaxy as the Milky Way. Its mass has recently been derived from line spectroscopy (using HST; Devereux et al. 2003) to he 7 x lo7 Ma,only 30 times the mass of Sgr A*. But this would really only be enough to call these two sources ‘‘cousins”. What makes them twins is the unique polarization characteristics of their already relatively uniquely inverted radio to mm spectra. These are the only two sources known so far where the circular polarization (CP) conclusively dominates the LP below about 100 GHz, which is the opposite of what is seen in AGN (Brunthaler et al. 2001; Bower 2003). This is especially important in this context, because the source of this polarized emission has been resolved into a jet in M81* (Bietenholz, Bartel & Rupen 2000).

-

-

The M81* jet is one-sided, very small (700-3600 AU depending on the frequency) and variable. Scaling the size linearly by mass would argue for 20-120 AU jets in Sgr A*, but one also has to take into account the scaling from the difference in power. We can estimate this by considering that jet models predict the radio flux scaling as F,, K A@ (e.g., Markoff et al. 2003) and the size scaling z cx &f0.5-0.s, depending on the particulars of the jet model and frequency (e.g., Falcke & Biermann (0 5-0 6)/1 4 F,” 36-0 43. The average flux of M81* at 14 GHz is 200 mJy, 1995). So z K F,, whereas for Sgr A* it is about 1000 mJy, and we must take into account their relative distances of 8 kpc and 3.6 Mpc. Using the largest values gives a maximum expected size for Sgr A*’s jets of 120AU(1000/200) 43(8/3600)2 x AU. The best we can do currently is resolve objects at about 1 AU in the Galactic center, so it is not really surprising that no jets have yet been seen for Sgr A*.

-

-

*

-

Astron. Nachr./AN 324. No. S 1 (2003)

441

Miyazaki, Tsutsumi & Tsuboi (2003) suggest that Sgr A* shows similar mm variability to M81* (e.g., Sakamoto et al. 2001). There are also preliminary detections of variability propagating from higher to lower frequencies, which is expected in a jet model but not a disk. Thus, the data suggest that the mm is indeed related to the radio, and supports a jet origin for the submm emission. 3. Chandru sees an X-ray jet-like feature. As presented in Baganoff et al. (2003b),there is an extended jet-like feature with a hard power-law spectrum pointing almost directly back to Sgr A*. It lines up with the larger scale X-ray and radio "lobes", and is perpendicular to the Galactic plane. While this is speculative, it is certainly intriguing, and may indicate that we are seeing some relics of prior AGN-like activity.

3 Models for the first and largest flare 2

,

'

.

'

,

"

'

lo&,

'

I

'

10

5

0

'

y

(GHz)

Fig. 1 Quiescent state spectrum of Sgr A* showing the jet contribution in the case that the magnetic energy density dominates the electron energy density by a factor of 17.

3.1

Quiescent state

Having presented our case for a jet, we will now discuss models for the first detected flare, in the second Chandra observing cycle presented in Baganoff et al. (2001). In order to talk about the possible role of jets during flares, we should first mention their role during quiescence. The quiescent state is extremely robust; after each episode of flaring the system drops back to the same flux and spectral shape as before. This suggests that the flares could be due to a new process which superimposes itself on top of the stable quiescent state. It could also be that both states originate in the same part of the system, but that the necessarily short-term changes somehow do not affect the ability of the system to "bounce back" consistently each time. When the quiescent X-ray state was first discovered, we (Falcke & Markoff 2000) reconsidered the original radio-only jet model to see what kind of X-ray spectrum it would predict. If one makes the assumption that the nozzle of the jet is at the inner parts of the accretion disk, and then assumes equipartition of the magnetic and particle energy densities, there is a unique solution to fit the submm bump and average radio spectrum. These same parameters also then fix the synchrotron self-Compton emission component, which

S. Markoff and H. Falcke: Jet Models for Flaring in Sgr A*

448

turns out to fall within the Chandra error box with the correct slope. However, the quiescent state has now been established as being dominated by thermal, extended emission from plasma within the Bondi capture radius. Though it is not yet clear how much a non-thermal component could still be contributing, it is feasible that 10-30% of the emission is more variable than what would be expected from the accretion flow alone. The reason this fraction is important is because it will tell us how far out of perfect equipartition (in the direction of the magnetic energy dominating) we have to go in the model to bring the jet SSC contribution down compared to the disk component in the X-rays. While many jet formation scenarios/simulations involve magnetic field domination, it is likely unrealistic to think the magnetic energy would be more than a couple orders of magnitudes greater than the energy of the emitting particles. In Fig. 1 we show the case ~ 17, where UB = B2/8r and Up is the energy density in the electrons. for U B / U = So the scenario as we would propose it is that the quiescent state is comprised of dominant thermal accretion emission, with the non-thermal jet emission lurking not far below. During flares, the jet emission increases to dominate for short periods of time, the length of which is determined by the cooling or expansion times. We will now continue with mechanisms that could bring this about.

3.2 BigFlares

-

When Chandra first detected flaring in Sgr A*, they saw what now seems to be a rather unusually high increase in flux by a factor of 45 (Baganoff et al. 2001). The shortest timescale for significant change was about 600 s, giving a scale for the source region of 5 40r,, where r, = GM/c2. Because there had been no observations at lower frequencies, the theoretical models focused not only on mechanisms, but on finding ways to learn about the relevant physical processes behind the flaring by making predictions for other frequencies. In the context of the jet model, we (Markoff et al. 2001) explored three possible mechanisms which could in principle explain a flare of such magnitude. These models tried to isolate the effects of a single type of change, for predictive purposes, whereas in reality one may see some mixing of these scenarios. The first, which we called the “hi-flare”, was ruled out from the start. It posited that the accretion rate into the jet, and thus the total jet power, suddenly increased. This would have the result of raising both the magnetic and particle energy densities (assumed to be from an increase in particle number density) by the same amount, resulting in “across the board” flaring at all frequencies. To increase the flux by 45 in the X-rays would give almost an order of magnitude increase in the radio, which has never been seen in over 20 years of VLA observations. Also, increasing just the density but not the temperature of the radiating electrons would result in the same steep slope found in quiescence, which is not even in the error box for the hard flared spectrum. The other mechanisms we considered both involved some kind of sudden energizatiodheating of the particles, either by a process in which the electrons retain a quasi-thermal distribution, or a process resulting in a non-thermal distribution, such as acceleration. For the former, one could imagine a magnetic reconnection event which would transfer energy from the magnetic field into the particles, raising their temperature by some fiducial amount. We referred to this as the “T,-flare”, and modeled it by changing only the electron temperature T,, while holding the everything else (density, magnetic field strength) fixed. This does, however, have the effect of changing the partition of energy in the model. A model starting out in equipartition would thus end up with, at least for a short while, the electron energy density dominating the magnetic energy density. Similarly, a model starting out with a dominant magnetic pressure, such as what we find necessary to explain the quiescent state, could “equilibrate” during such an event, driving the system towards equipartition. Fig. 2a shows such a flare, which resulted from a factor of 3 change in the electron temperature. Most notably, this model predicts a very large amount of simultaneous flaring in the MIR, as well as some flaring in the submm and radio bands. Another type of flare is possible, however, which does not predict much change at other frequencies. It invokes a mechanism which seems to be commonplace in the jets of AGN and X-ray binaries, and we see the evidence of it via optically-thin synchrotron emission from a power-law of assumedly accelerated

Astron. Nachr./AN 324, Yo. S 1 (2003)

449

-

,\-

1

~

_,’

-6 -

- l o t ’

I

I’

’

”

I

’

”

1

’

’

‘

I

I I I’

”

1

’

’

‘

I

‘ 1

- l o t ’

”

”

’

”

1

’

”

1

10 log,, u (GHz)

’

”

’

”

‘ 1

”I

10 log,, v (GHz)

Fig. 2 The two flare models presented in Markoff et al. (2001) for the largest (factor of 45) flare seen yet (Baganoff et al. 2001). (a) The “T,-flare”, where a temperature change of N 3 while holding everything else fixed can fit the flares via SSC. This predicts large simultaneous flares in other frequencies, particularly the M R . (b) The “Shock-flare’’in which thermal particles are accelerated into a power-law tail for the duration of the flare.

electrons. The exact form of acceleration is still a matter of debate, and could include Fermi shock acceleration, stochastic acceleration by turbulence or a process initiated by magnetic reconnection. In our paper, we referred to this as the “shock flare”, but really we should call it the “acceleration flare” to be more generic. The idea is that a fraction of the assumedly quasi-thermal particles (we know they cannot be dominantly non-thermal because of the low IR upper limits indicating a drop off where there would otherwise be a power-law) near the base of the jet are accelerated into a power-law tail to create the flare. This is shown in Fig. 2b, where we considered the case of a photon index of 1.8, characteristic of that seen in AGN jets, but which provides only a marginal fit to the data. The cooling time for synchrotron losses is much faster than the lifetime of the flare, so inherent to this model is the assumption that the particles are continuously reaccelerated over the lifetime of the flare. We did not explore this solution in detail, however, because at the time we felt that the “T,-flare” solution was more appealing.

4

New information

Since this first flare was detected, new information has come our way via multiple flare observations by Chandru in the next observing cycle with simultaneous lower frequency observations (Baganoff et al. 2003). We now know that flares are fairly common, although so far with a much smaller amplitude than the first and largest flare discussed in the last section. It seems that flaring by a factor of 5-10 happens almost daily, and so is not likely to be some exotic type of event. We have also learned that, surprisingly, the flares do not seem to correspond to much, if any, lower-frequency activity. No obvious flaring was detected in the radio, and while the case in the mm and submm is not totally clear, it does not seem like there is any kind of smoking gun. At least not for the these smaller, more regular flares for which a statistical argument by Hornstein et al. (2002, 2003) limits the flared IR emission of Sgr A* to 19 mJy at 2.2 pm. This is based on the fact that there is only a 5% probability that their Keck observations could have missed one of these daily-occurring flares, and has a 2u confidence level. While this argument seems compelling for the common flares, we think the jury may still be out for the largest flares, since they may only occur 1% of the time or less. For models of the smaller flares, however, we will consider this limit to be important. For the large factor of 50 flares, we really need another simultaneous observation campaign to see what is going

S. Markoff and H. Falcke: Jet Models for Flaring in Sgr A*

450

2

I

.

.

.

.

,

.

.

.

.

,

I

2

I

.

.

.

I

,

.

.

.

.

,

.

Fig. 3 Two preliminary flare models for the new averaged spectrum of the smaller, almost-daily flares (Baganoff et al. 2003). (a) The “SSC-flare”,which requires a higher temperature and lower magnetic field compared to quiescence. (b) The “Acceleration-flare” in which now 5 1% of the thermal particles are accelerated into a power-law tail for the duration of the flare, and assumes reacceleration during the lifetime of the flare.

on particularly in the submm and IR. Unfortunately, until we manage to get some idea of their frequency of occurrence, it is not easy to convince the telescope time allocation committees. What do these new flares and limits tell us? First of all, any model along the lines of the old T,flare model has to be modified to fit the new lack of lower-frequency variability, which is challenging. A toy model with a simple temperature change can fit the average small flare spectrum marginally, but not perfectly. Therefore we will have to broaden the term to “SSC-flares”, to express that the process (synchrotron self-Compton) is still the same, but that the density and magnetic field may also change at the same time as the temperature. Even with these changes, staying under the Hornstein et al. limit as well as not varying much in the radio-submm is tricky. In Fig. 3a we show a preliminary new “SSC-flare” fit to the averaged smaller flare spectrum presented in Baganoff et al. (2003). In order to suppress the synchrotron compared to the SSC component, to stay under the IR limits, the energy distribution again must move from magnetic domination (quiescence) towards equipartition or particle domination. We have also revisited the “Shock-flare’’model, which we will now call the “acceleration-flare”, shown in Fig. 3b. Because the flares are of smaller amplitude, we can now fit the spectrum with just a tiny fraction of the particles being accelerated, 5 1%.We have required the acceleration region to be within = 40r, of the black hole, and this particular fit assumes that the accelerated particles have a hard distribution, as perhaps could result from other forms of acceleration instead of Fermi acceleration at shocks. It also assumes that the particles are continuously reaccelerated or injected so that a cooling break is not seen. If one assumes a cooling break in the spectrum, the resulting photon index would be more like I‘ 1.5, which would give a more marginal fit. One possible way to avoid this is to have the acceleration region further out in the jet, where the magnetic field is low enough that the cooling break would not effect the Xrays. This would then require some alternative argument for why the assumedly reconnection or turbulent acceleration event lasts only 600s or less.

5 Summary We conclude that heating/acceleration processes in a jet can account for the flares seen on an almost daily basis in Sgr A*. Now that we know more about the characteristics of broad-band flaring, we can construct more realistic models, While we are still uncertain exactly what to expect for flares of a factor of N 50, we now know that at least the smaller X-ray flares do not seem to show up much in the lower

Astron. Nachr./AN 324, No. S1 (2003)

45 1

frequencies. The two models which we originally proposed need to be adapted to fit these criteria, but the mechanisms are the same. The biggest change is that we can no longer consider a simple temperature change alone as a viable mechanism to obtain more SSC emission. Any change in the temperature must also be accompanied by changes in density and magnetic field, but as before, the changes are still consistent with a general movement of energy from the magnetic field into the particles. The acceleration model, which was not our favorite before, may now be the better model. It seems to require less fine-tuning, and requires a much smaller change to the system, which would be consistent with a frequent process that almost “flickers”. A synchrotron model would also predict up to 10% linear polarization, as seen in the submm bump. Interestingly, Yuan, Quataert & Narayan (2003) are proposing very similar mechanisms for the flaring in the context of the inner accretion flow, while other models include Liu & Melia (2002, 2003) and Nayakshin, Cuadra & Sunyaev (2003). Future multi-wavelength observations, polarization measurements and a better understanding of the modeling and simulations should help us discern between them.

References Baganoff, F.K., Bautz, M.W., Brandt, W.N., et al. 2001, Nature, 413,45 Baganoff, F.K., Maeda, Y., Morris, M., et al. 2003a, ApJ, in press Baganoff, F.K., et al. 2003b. this workshop Bietenholz, M.F., Bartel, N. & Rupen, M.P. 2000, ApJ, 532, 895 Blandford, R.D. & Begelman, M. 1999, MNRAS, 3030, L1 Blandford, R.D. & Konigl, A. 1979, ApJ, 232, 34 Bower, G.C. 2003, these proceedings Bower, G.C., Wright, M.C.H., Falcke, H. &Backer, D.C. 2003, ApJ, in press (astro-ph/0302227) Brunthaler, A., Bower, G.C., Falcke, H. & Mellon, R.R. 2001, ApJ, 560, L123 Devereux, N., Ford, H., Tsvetanov, Z., & Jacoby, G. 2003, AJ, 125, 1226 Eckan, A,, et al. 2003, this workshop Falcke, H. & Biermann, P.L. 1995, A&A, 293,665 Falcke, H. & Markoff, S. 2000, A&A, 362, 113 Falcke, H. & Melia, F. 1997, ApJ, 479, 740 Ghez, A. 2003, these proceedings Goldwurm, A., Brion, E., Goldoni, P., et al. 2003a, ApJ, 584, 751 Goldwurm, A., Brion, E., Goldoni, P., et al. 2003h, these proceedings Hornstein, S.D., Ghez, A.M., Tanner, A,, et al. 2002, ApJ, 577, L9 Hornstein, S.D., Ghez, A.M., Tanner, A,, et al. 2003, this workshop Liu, S. & Melia, F. 2002, ApJ, 566, L77 Markoff, S., Falcke, H., Yuan, F. & Biermann, P.L. 2001, A&A, 379, L13 Markoff, S., Nowak, M., Corbel, S. et al. 2003, A&A, 397, 645 Melia, F. 1992, ApJ, 387, L25 Melia, F. & Falcke, H. 2001, ARA&A, 39, 309 Miyazdki, A., Tsutsumi, T. & Tsuboi, M. 2003, these proceedings Nagar, N.M., Wilson, A S . & Falcke, H. 2001, ApJ, 559, L87 Nagar, N.M., Falcke, H., Wilson, A.S. & Ulvestad, J.S. 2002, A&A, 392, 53 Narayan, R., Mahadevan, R., Grindlay, J.E., et al. 1998, ApJ, 492, 554 Nayakshin, S., Cuadra, J. & Sunyaev, R. 2003, A&A, submitted (astro-ph/0304126) Ott. T., Genzel, R., Eckart, A. & Schodel, R. 2003, these proceedings Qudtaert, E. & Gruzinov, A. 1999, ApJ, 520,248 Reid, M., Menten, K.M., Genzel, R., et al. 2003, these proceedings Sakamoto, K., Fukuda, H., Wada, K. & Hahe, A. 2001, AJ, 122, 1319 Schodel, R., Eckart, A., Genzel, R. & Ott, T. 2003, these proceedings Yuan, F., Markoff, S. & Falcke, H. 2002, A&A, 383, 854 Yuan, F., Quataert, E. & Narayan, R. 2003, ApJ, submitted (astro-pW0304125)

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S l , 453 -458 (2003) / DO1 10.1002/asna.200385044

A Jet-ADAF Model for Sgr A*

',

F. Yuan* S. Markoff *2, and H. Falcke3

' Harvard-SmithsonianCenter for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA MIT Center for Space Research, 77 Massachusetts Ave., NE80-6035, Cambridge, MA 02139, USA Max-Planck lnstitut fur Radioastronomie, Auf dern Hugel 69,53 121 BOM, Germany

Key words accretion, accretion disks - black hole physics - galaxies: active - galaxies: nuclei - Galaxy: center - hydrodynamics PACS 04A25 We briefly review the observations and theoretical models for Sgr A*, the supermassive black hole in the center of our Galaxy, emphasizing the coupled jet plus accretion disk model for this source. In this model, the accretion flow is described as an ADAF fed by hot plasma in the Galactic Center region. At the innermost region of the accretion flow, a small fraction of the accretion flow is ejected out and forms a jet. Because the accretion flow is supersonic at the innermost region, during this process a standing shock occurs. As a result, the electron temperature increases to- 2 x 10 "K, which is about 10 times higher than the highest temperature attained in the ADAF, therefore leading to strong radiation via synchrotron and the Compton scattering processes. The emergent spectrum of Sgr A* is the sum of the emission from jet and underlying ADAF. Specifically, the submm-bump is produced by the base of the jet, while the radio spectrum is due to the outer part of the jet. The X-ray flux in the quiescent state is the sum of the bremsstrahlung emission from the underlying accretion flow and SSC from the jet. We also briefly discuss the application of our model to another low-luminosity AGN, NGC42.58.

1 Introduction The energetic radio source Sgr A* located at the center of our Galaxy is now widely believed to b e the signature of a massive black hole with mass M = 2.6 x 106Ma (Melia & Falcke 2001). Due to its proximity, there is a plethora of observational data for this source, ranging from radio, submm, IR, to Xray. Especially, the recent Chandra and radio polarization observations put new and strict constraints on its theoretical models. Thus, Sgr A* offers us a good chance to understand the physics of a class of sources.

2 Observational results Spectral results: From a few GHz to 86 GHz, the spectrum is a power-law with spectral index of Q % 0.3. Above 86 GHz, up to 1000GHz, there is a millimeter/submillimeter excess - the so-called submm bump, which is the most prominent feature in the spectrum of Sgr A*. In the infrared waveband, Sgr A* has never been firmly detected, only some upper limits are available there. Two X-ray states were found by Chandra for Sgr A*, namely the quiescent and flare states (Baganoff et al. 2003; 2001). In the quiescent state, the spectrum is very soft, with spectral index of cy. = 1.2t::;. In the flare state, the spectrum is very hard, with 1 0 p 8 L ~ d d . On the other spectral index of Q = 0.3?:::. The bolometric luminosity is very low,

-

* Corresponding author: e-mail: [email protected], Phone: +01 617 384 7695 * * NSF Astronomy & Astrophysics Postdoctoral Fellow

@ 2003 WILEY-VCH Verlag GmbH & Cu.KGA. Weinhem

F. Yuan et al.: A Jet-ADAF model for Sgr A*

454

hand, Chandra detects the density and temperature of the plasma near the Bondi radius from which we can calculate the accretion rate assuming Bondi accretion, which is 10-6Ma yr-l. Assuming that the L E ~ ~ . accretion efficiency is of the standard value of 0.1, the emergent luminosity would be 6 x So the accretion flow must be radiatively inefficient. Variability: A periodic variability around lcm with period around 106 days was detected (Zhao et al. 2001; 2003). At 1.3 mm, the SMA observed flares with timescale of days (Zhao et al. 2003). At X-ray, the source is rapidly variable on short timescale of 1hr. A flux drop on timescale of 10 min was detected in the flare state. On the other hand, comparison between two observations indicates the steady X-ray flux remains almost constant during an interval of one year. Combining with the rapid variability, this indicates that there are two components responsible for the X-ray emission. Polarization results: At centimeter wavelengths, no linear polarization was detected but strong and variable circular polarization was detected (Bower et al. 1999; 2003) At submm wavelengths, JCMT and BIMA detected high linear polarization but no circular polarization (Aitken et al. 2000; Bower et al. 2003). The observation of BIMA also put an upper limit on the magnitude of the rotation measure of 2 x 10‘rad m-’. This argues for an accretion rate of 10-*M0 yr-l at the region where submm-bump comes from. N

N

N

N

3 Brief review of accretion and jet models for Sgr A* 3.1

Accretion models

Melia (1992; see also Melia, Liu, & Coker 2000) proposed a spherical accretion model for Sgr A*. In this model the accretion flow is assumed to free-fall until a Keplerian disk is formed within a small “circularization” radius. The main contributors to the radio and X-ray spectra are synchrotron radiation and bremsstrahlung, respectively, from the roughly free-fall flow beyond the small disk. However, spherical accretion is likely to be an over-simplification, since the accretion flow still possesses some angular momentum. An advection-dominated accretion flow (ADAF) model therefore is more dynamically exact in this sense (Narayan et al. 1998; Narayan 2002). The most attractive feature of the ADAF model is its ability to explain the unusual low-luminosity of Sgr A* given the relatively abundant accretion material. This is because most of the viscously dissipated energy is stored in the flow and advected beyond the event horizon rather than radiated away. The bremsstrahlung emission prediction for the quiescent state seems to be confirmed by the observations. One problem for the ADAF, and actually for all the general accretion models, is that it under-predicts the low-frequency radio emission of Sgr A* by over an order of magnitude. Another problem for the canonical ADAF model is that it can not explain the observed high linear polarization at submm-bump. This is because it assumes the accretion rate is independent of radius. Thus, the density at the innermost region, where submm radiation comes from, is so high that the implied rotation measure is well above the observationally implied upper limit. 3.2 Jet model Following the initial paper by Reynolds & McKee (1980), Falcke et al. (1993) proposed that it is the jet stemming from the disk rather than the disk itself which is responsible for the radio spectrum of Sgr A*. In this model, the submm bump is produced by the acceleration zone of the jet, called the “nozzle”, while the low-frequency radio spectrum comes from the part of the jet beyond the nozzle (Falcke & Markoff 2000). The nozzle is of order a few Rs long and forms from the disk at a radius of 2&. This model gives an excellent fit to the radio spectrum of Sgr A*, including the low-frequency spectrum below the break and the submm bump. But the jet model can’t explain the resolved extended X-ray emission in the quiescent state. More importantly, it is not clear in the model why the parameters of the jet, such as the electrons temperature, possess the required values, particularly in reference to the inferred underlying accretion disk. N

Astron. Nachr./AN 324, No. S1 (2003)

455

Therefore, it is cruciai to consider the jet and accretion flow as a coupled system in Sgr A", and to consider what are their respective roles if both are truly present in Sgr A * .

4

Jet-ADAF model

Both accretion and jet models possess attractive features. It is our task in our jet-ADAF model to consider the jet and accretion flow as a coupled system in Sgr A', and we will see that we can supply a satisfactory explanation to the otherwise puzzling parameters of the jet. We will concentrate on the spectral interpretation in the present paper. Specifically we will emphasize the quiescent state of Sgr A*. Markoff (2003; see also Markoff et al. 2001) and Beckert (2003; see also Beckert & Falcke 2002) will concentrate on the interpretation to the flare state and the polarization of Sgr A* in the context of jet, respectively.

Sgr A*

Fig. 1 The schematic diagram ofjet-disk coupling model for Sgr A*

The schematic diagram of our model is presented in Fig. I . The accretion disk is described as a radiatively inefficient accretion flow such as an ADAF or other modified versions of ADAFs. At a certain radius of the innermost region of the disk, a small fraction of the accretion material will be ejected out of the disk and form a jet due to some physical mechanism. Because the radial velocity of the accretion flow at small radii is supersonic, when the accretion flow is transferred into the jet where the plasma velocity is in vertical direction, a standing shock should occur due to the bending. We can calculate the physical quantities such as the electron temperature and density in the nozzle, from the shock jump conditions in the hydrodynamic approximation,

456

F. Yuan et al.: A Jet-ADAF model for Sgr A*

We then calculate the physical quantities following the propagation of the jet plasma along the jet. In summary, the parameters describing the jet are: The location of jet formation from the disk, the jet radius T O ; The fraction of accretion material going into the jet, which will determine the density in the jet, qj

= rYijet/M;

The ratio of temperature between electrons and ions in the post-shock plasma, T,/Ti in the post-shock plasma; the angle between the jet axis and the line of sight, 8. The values of the above parameters in our fitting presented in Fig. 2 are as follows: TO M 1.7R,(< rSonic).qj M 0.5% if we assume that the accretion rate is determined by the Bondi accretion formula and remain constant with radii. Note that qj will increase if there are strong mass loss due to, say, outflow. T,/Ti M 0.1. This value is typical from the investigation to the ISM shocks in the context of SNR. 6 = 35". The parameters of the ADAF are as usual (see Yuan, Markoff & Falcke for details). We calculate the emergent spectrum from both the ADAF and the jet and sum them to fit the observed spectrum of Sgr A*. The results are as follows: The ADAF is the reason for the low intrinsic luminosity of Sgr A*. Synchrotron emission from the nozzle is responsible for the submm bump. The self-absorbed synchrotron emission from the outer part of the jet fit the flat radio spectrum of Sgr A*. The bremsstrahlung emission from the ADAF at around the Bondi radius can explain the resolved extended and constant X-ray emission of the quiescent state of Sgr A*. In our calculation, we didn't take into account the possible existence of strong outflows from the ADAF. This will not affect our results because outflow has little effect on the bremsstrahlung emission arisen at large radii around the Bondi radius. SSC from the nozzle may also contribute some fraction which may be one of the reasons why rapid variability was detected by Chandra. The results are shown in Fig. 2. We see that the emission of Sgr A* is basically dominated by the jet rather than the underlying disk although only a very small fraction of accretion material is transfered into the jet. The reasons are that 1) the ADAF is extremely radiatively inefficient; and 2) the bending shock at the base of the jet efficiently transfers the gravitational energy into the thermal energy of electrons and the radiative efficiency of the high-T, electrons in the jet is very high.

5 Application of jet-ADAF model for NGC4258 We applied the same jet-ADAF model to another low-luminosity AGN, NGC4258 (Yuan et al. 2002). This source was modeled by an ADAF alone previously (Gammie, Narayan & Blandford 1999). However, the theoretical prediction of the model is not consistent with the later IR observation (Chary et al. 2000)~ . ADAF predicts a fv c( v1I3 hard spectrum, while observation give a quiet soft power-law, f,,0: Y - ~ . In fact, a spectrum like NGC4258's seems to be very popular in the low-luminosity society (Ho 1999). We propose that the steep IR spectrum is produced by the power-law electrons in the jet nozzle, due to the acceleration in the bending shock (see Fig. 1 and Yuan et al. 2002). The synchrotron emission

Astron. Nachr./AN 324, No. S 1 (2003)

457

Fig. 2 The jet-disk spectral model for Sgr A*. The dotted line is for the ADAF contribution, the dashed line is for the jet emission, and the solid line shows their sum.

from these power-law electrons can fit the IR spectrum very well. Inverse Compton will produce the X-ray spectrum. The electrons in the outer part of the jet will get accelerated further and are responsible for the radio spectrum. So for NGC4258, all the spectrum ranging from radio, IR, to X-ray are dominated by the emission from the jet, while the underlying ADAF contributes little, as shown by Fig. 3.

6 Discussion: current status and future work Sgr A* is unique in many senses, therefore it is worthwhile to investigate it in great detail, both theoretically and observationally. Theoretically, the jet-ADAF model currently supplies us a with a good framework and a new direction for further research. On the other hand, however, it is still necessary to speculate other possibilities, especially considering that the evidence for the jet in Sgr A* is tentative (Lo et al. 1998). For example, up to now most of the accretion disk models assume the electrons are in thermal distribution. One project is therefore to investigate the role of non-thermal electrons which possibly exist in accretion flow and may be responsible for the low-frequency radio spectrum of Sgr A*. There are some work in this aspect (e.g., Ozel, F., Psaltis, D., & Narayan ZOOO), but they obviously need to be revisited giving the new observational constrains. Research work along this line is being done (Yuan, Quataert, & Narayan 2003). Initial results indicate that accretion model with hybrid thermal and power-law electrons can also explain the radio spectrum. Then an interesting question is what are the different predictions of jet-ADAF model and accretion disk model and whether we can rule one of them out via further observations such as multiwavelength campaign.

References Aitken, D.K., Greaves, J.S., Chyrsostomou, A. et al. 2000, ApJ, 534, L173 Baganoff, F. K., Maedd, Y., Moms, M. et al. 2003, ApJ, in press (astro-pN0102151)

F. Yuan et al.: A Jet-ADAF model for SD A*

458 43 42 41 40 39 38 37 36 35

log[v(Hz)l The jet-disk spectral model for NGC4258. The long-dashed and the dotted lines are for the synchrotron and SSC emission from the jet nozzle; the thick-solid line is their sum. The thin solid line is for the outer part of the jet. The short-dashed and dot-dashed lines are for the standard thin disk and ADAF, respectively. They contribute little to the observed spectrum. Baganoff, F. K., Bautz, M. W., Brandt, W. N. et al. 2001, Nature, 413,45 Beckert, T., & Falcke, H. 2002, A&A, 388, 1106 Beckert, T. 2003, these proceedings Bower, G. C., Falcke, H., & Backer, D. C. 1999, ApJ, 523, L29 Bower, G. C. 2003, these proceedings Chary, R., et al. 2000, ApJ, 531,756 Falcke, H., Mannheim, K., and Biermann, P. L. 1993, A&A, 278, L1 Falcke, H. & Markoff, S. 2000, A&A, 362, 113 Falcke, H., Goss, W. M., Matsuo, H., Teuben, P., Zhao, J. H., & Zylka, R. 1998, ApJ, 499,731 Gammie, C., Narayan, R., & Blandford, R. 1999, ApJ, 516, 177 Ho, L. C. 1999, ApJ, 516,672 Lo, K.L., Shen, Z.Q., Zhao, J.H., & Ho, P.T.P. 1998, ApJ, 508, L61 Markoff, S., Falcke, H., Yuan, F., & Biermann, P. 2001, A&A, 379, L13 Markoff, S. 2003, these proceedings Melia, F. 1992, ApJ, 387, L25 Melia, F., & Falcke, H. 2001, ARA&A, 39, 309 Melia,F., Liu, S., & Coker, R. 2001, ApJ, 553, 146 Narayan, R., Mahadevan, R., Grindly, J.E., Popham, R., & Gammie, C. 1998, ApJ, 492,554 Narayan, R. 2002, in Lighthouses of the Universe: The Most Luminous Celestial Objects and Their Use for Cosmology, Proceedings of the MF'AESO, p. 405 Ozel, F., Psaltis, D., & Narayan, R. 2000, ApJ, 541, 234 Reynolds, S.P. & McKee, C.F., 1980, ApJ, 239, 893 Yuan, F., Markoff, S., & Falcke, H. 2002, A&A, 383,854 Yuan, F., Markoff, S., Falcke, H., & Biermann, P. 2002, A&A, 391, 139 Yaun, F., Quataert, E., & Narayan, R. 2003, in preparation Zhao, J., Bower, G.C., Goss, W.M., 2001, ApJ, 547, L29 Zhao, J. 2003, these proceedings

Astron. NachrJAN 324, No. S1,459-465 (2003) / DO1 10.1002/asna.200385051

A model for polarised radio emission from Sgr A* T. Beckert"' I Max-Planck-Institutfur Radioastronomie, Auf dem Hugel 69, D-53 121 Bonn, Germany

Key words Plasmas, radiative transfer, polarization - Galaxy: center Abstract. The detection of circular polarisation in compact synchrotron sources provides new insights into magnetic field configurations and the low-energy population of electrons in relativistic jets. Conversion of linear to circular polarisation can be driven by Faraday rotation or turbulence in the source itself. A detailed model for the properties of Sgr A* in the galactic centre is presented.

1 Introduction The enigmatic radio and X-ray source Sgr A* in the Galactic Centre is associated with a dark mass (black hole) (Schodel et al. 2002) of about 2.6 lo6 Ma.The source is highly variable in X-rays and less so in radio to sub-rnm bands. Sgr A* shows two unusual polarisation properties, namely stronger circular (CP) than linear polarisation (LP) below 100 GHz and a jump in LP to about 10-15% above 100 GHz. It is generally assumed that the black hole accretes interstellar material, which gets within its Bondi radius. While the mass accretion rate is uncertain the accretion flow will be optically thin above the synchrotron self-absorption frequency and emission will be dominated by synchrotron radiation, Bremsstrahlung, and inverse Compton processes. While the sub-mm bump might be produced by the accretion How, the Hat radio spectrum is likely to be formed in an outflow or jet generated by accretion into the black hole. Without assuming a particular feeding process, I present a turbulent outflow model, which can account for the radio to sub-mm spectrum and the polarisation properties of Sgr A*.

2 Representation of the radiation field The polarisation properties of a radiation field are usually described in terms of Stokes parameters (intensities) I , Q, U ,V ,which carry no phase information of the wave field. A rotation of coordinates by x = 180" in the plane of the sky leaves all Stokes unchanged. This is reflected in the definition (e.g. Chandrasekhar 1960) of

Q U

= Pcos2/3cos2x , = P cos 2 p sin 2% ,

where the combination P cos 2/3 is the linearly polarised intensity. Two linearly polarised waves, represented by I = Q and I = -Q have electric vectors perpendicular to each other, while waves with I = Q and I = U have electric vectors rotated by 45" with respect to each other. The electric vector of incoherent synchrotron emission is perpendicular to the static magnetic field and in this paper a local suitable orientation of coordinates is used so that the electric vector associated with Q is perpendicular to the local magnetic field projected onto the sky. * Corresponding author: e-mail: tbeckertarnpifr-bonn.rnpg.de,Phone: +0049525 189, Fax: +0049525437

@ 2003 WILEY-VCH Verla~GmbH & Co. KGaA, Weinhem

T. Beckert: A model for polarised radio emission of Sgr A*

460

3 Outflow Model Many synchrotron sources where circular and linear polarisation has been found are compact radio sources of large brightness temperatures associated with relativistic jets or outflows. The single dish radio fluxes reveal flat to moderately inverted radio spectra, which are best explained by self absorbed, inhomogeneous jets. If the jets are over-pressured relative to their surroundings close to the jet launching point, the outflow will have a conical shape and the radiating particles will suffer almost no adiabatic losses, which is one possibility to produce the observed flat spectrum of radio cores. In contrast the hot gas in advection- or convection-dominated accretion flow models is adiabatically compressed and viscously heated, which leads to highly inverted radio spectra. In the outflow model sub-adiabatic energy losses of relativistic particles and a constant outflow velocity is assumed. In the simplest case the local radius R of the conical flow grows linearly with distance z and the particle density decreases correspondingly (neglecting pair creation or annihilation)

R=sin$z

,

n,0:RP2,

(3)

where $I is the half-opening angle of the outflow. 3.1 Magnetic Fields The poloidal magnetic field of the accretion flow (along the rotation axis) will connect the accreted with the out-flowing gas. Rapid rotation of the foot points of this B-field component from the disk twists the field lines when the outflow expands laterally. Together with a possible toroidal B-field component from the disk that can be carried with the gas, when escaping into the outflow, it leads to a spiral structure of the magnetic field in the outflow or jet. If magnetic flux through jet cross sections is conserved, this implies for the B-field along the jet:

B, cx RP2 .

(4)

For the toroidal component Bb in the jet it is expected that the field energy is conserved along the flow leading to

B4 cx R-'

.

(5)

The different behaviour of poloidal and toroidal field produces a spiral of field lines, which winds stronger and stronger as the gas moves outward. Because synchrotron emission is sensitive to the magnetic field orientation along the line of sight, the averaged mean l?-field of the spiral changes with optical depth. For constant particle density and constant B, and Bb in jet cross sections Fig. ?? illustrates the change of the averaged magnetic field orientation, weighted by the optical depth. The weighted mean field is

( B )= ( R Z n ) - ' l X d 2 L x d $

(

Bb cos q3 -B;inq3)

zexp(-T)

where 7 is the optical depth for synchrotron self-absorption and KI the absorption coefficient. While the %-componentof l? in Eq. ?? vanishes due to symmetry, the transversal B-component vanishes only in the optical thin case. For equally strong B, and Bb, Fig. ?? shows that the ratio of transversal to z-component increases with optical depth and reaches a limit at large optical depth corresponding to a rotation of 38"

46 1

Astron. Nachr./AN 324, No. S1 (2003)

Frequency [Hz]

Fig. 1 Weighted B-field components as function of TO = K I R. The dashed line shows the weighted (B,)/ B z .The decrease with TO shows the reduces effective area from which radiation reaches the observer. The dash-dotted line gives the transversal (Bd) /Bd, which shows a characteristic peak at TO = 1. The solid line gives the ratio (B+) / ( B , ) for equally strong B, and Bm. The rapid change at about TO = 1 produces a change in the effective B-field orientation seen.

from the jet axis. In addition to radiative transfer effects discussed below, the spiral structure by itself produces a rotation of the polarisation position angle with optical depth. Variations in brightness in resolved images of radio jets (Lobanov & Zensus 2001) can be interpreted as Kelvin-Helmholtz-instabilities in the jet flow on large scales. These instabilities may drive MHDturbulence, when the free energy can be cascaded down to smaller length scales. At a characteristic length scale L energy will be fed into the inertial range of turbulence with an energy power spectrum E ( k ) 0; !d. This produces an additional turbulent, tangled magnetic field component & not associated with magnetic flux along the jet, which is fed energetically by the kinetic energy of the outflow. In a steep turbulent cascade /3 M 2, field reversals due to turbulence, if they happen at all, will occur on the driving length scale L. The mean number of field reversals N = L / R determines the reduction of polarisation. In contrast to piecewise homogeneous models (Jones 1988, Ruszkowski & Begelman 2002), where N is the number of the smallest cells, a superposition of Alfven waves with amplitudes drawn from the energy spectrum and random phases and orientations, which get de-correlated within a few wavelength, is used for the radiative transfer simulations presented here. Instabilities are expected to grow with a characteristic growth length ZMHDalong the outflow. Downstream of ZMHDradiative transfer will strongly depend on the properties ( Land amplitude) of turbulence.

4 Radiative Transfer Effects The propagation of electromagnetic waves in a magnetised plasma is influenced by the motion of charged particles in the combined static magnetic field and the field of the wave. The reaction of particles modify the normal modes for wave propagation in the plasma. These effects are stronger for small inertia (ymo) of the particles and for long wavelength. Propagation of radio waves, even in dilute cosmic plasma like in jets, is thus heavily affected. The radiative transfer in terms of Stokes parameters (Eq. ??) involves emissivities and absorption coefficients for Stokes I , Q, V and two rotation coefficients, which couple Stokes Q with U

T. Beckert: A model for polarised radio emission of Sgr A*

462

and U with V. With an isotropic particle distribution and a local coordinate system rotated appropriately, so that emission and absorption of Stokes U vanish, the change of Stokes parameters along the line of sight s is given by

The absorption coefficients K I , K Q , K V connect Stokes Q and V to the total intensity I . This leads to the well known rotation of polarisation position angle (P.A.) by 90" in a homogeneous source, when changing from optically thin to optically thick emission. The Faraday rotation coefficient K F rotates the linear polarisation between Q and U and therefore the PA. in a electron-ion plasma. This is understood when considering the propagation of normal modes along the magnetic field direction. Normal modes are circularly polarised waves, which travel with different phase velocities, whenever particles of a particular charge dominate the reaction in the plasma to the waves. Recombining the circularly polarised modes after leaving the plasma results in a rotated P.A. In a pair plasma or in a region with vanishing net magnetic flux along the line of sight no Faraday rotation occurs. 4.1 Conversion from linear to circular polarisation In contrast to Faraday rotation, which depends on the projected B-field along the line of sight, conversion from linear to circular polarisation and back again can happen whenever there is a B-field component perpendicular to the line of sight. Because particles are free to move along the magnetic field lines, but not transverse to it, the independent normal modes are linearly polarised with the electric vector oriented perpendicular (Stokes Q = I ) or along the B-field (Stokes Q = - I ) . The phase velocities for these two normal modes are different and two linear polarised waves, which are out of phase, represent a partly circularly polarised wave. The polarisation of an incoming LP radiation with a Stokes-U-component and a corresponding electric vector in the wave at an oblique angle to the static B-field is now changed to an elliptical polarisation. A cyclic process transforming U -+ V -+ -U -+ -V -+ U . . . is started. This conversion process does not depend on the sign of the charge nor on the orientation of the B-field, but on the strength of its transverse component. It is the sign of Stokes U to start with and the 'conversion depth' rc = KCL,which determines the sign and strength of the resulting circular polarisation.

5 Two drivers for source intrinsic conversion Circular polarised radiation is generated by conversion within a synchrotron source if the radiation has a non-vanishing Stokes U locally, which is not provided by the local emission process. Either Faraday rotation along the line of sight or a change in the B-field orientation in the plane perpendicular to the line of sight is required to start conversion. In real sources these are competing processes and which ever comes first determines the path to conversion. The critical number is the 'Faraday depth' TF = K F Lwithin the typical length scale for changes of the B-field L.

5.1 Strong rotativity (FD-Model) Faraday rotation as the driver for conversion can only be important in an electron-ion plasma with a significant population of low energy electrons. For a power-law distribution in energy the rotation coefficient is

Astron. Nachr./AN 324, No. S1 (2003)

463

where vg is the cyclotron frequency, n, the electron density, T , the classical electron radius, p the spectral index of the power-law cx y-p, and ymin the low energy end of the power-law distribution. 0 is the angle between magnetic field and the line of sight and includes the dependence on the polarity of the magnetic field. When averaged over the source volume, this corresponds to the net magnetic flux towards the observer. A large number of low energy electrons is necessary for the strong rotativity case. In Beckert & Falcke (2002) we described a model for Sgr A" based on Faraday driven conversion and this cannot be discussed in detail here. The model requires a very low ymin = 5 and predicts a kinetic power, which is lo5 times larger than the radiated luminosity. Within the strong rotativity model the observed long term stability of the sign of Stokes V signals the persistent net magnetic flux along the outflow. But the large particle density along the whole jet efficiently destroys the linear polarisation and cannot account for the observed strong sub-mm linear polarisation (Aitken et al. 2000). 5.2

Turbulence Driven Conversion (TD-Model)

In a situation where TF < 1 the change of B-field orientations drives conversion, either on a scale comparable to the source size, like the global spiral in the outflow model above, or on a smaller scale L set by the driving length scale of turbulence. In isotropic turbulence the mean resulting CP is zero without the help of the global spiral or any other gradual change of the mean B-field. It is the conversion depth TC and the ratio N = R / L ( Rthe jet radius) together with possible symmetries of the global B-field which determine the residuals (CP) and ( LP) in the turbulence driven model. The growth of instabilities and the development of turbulence to a level of saturation along the outflow will give rise to a characteristic length Zturb along the jet. Closer to the jet starting point turbulence is unimportant and will grow towards Zturb. Downstream of Zturbturbulence will saturate and may be damped further out. The fact that the frequency, where the emission becomes optically thin, changes with Y , , ~ cc 2-l for conical outflows with a flat spectrum, maps the evolution of turbulence along the jet into the polarisation spectrum (see Fig. ?? and Fig ??) The sign of the residual Stokes V depends on the handedness of the magnetic spiral and therefore on the rotation of the underlying accretion disk andor the black hole. This has recently been discussed in paper by Ensslin (2003).

6 Results - Application to Sgr A* The model for turbulence driven conversion (Section ??) is able to explain most of the radio data including the measurements for CP (Bower et a]. 1999b, Sault et al. 1999),its variability (Bower et al. 2002), and LP (Aitken et al. 2000, Bower et al. 2001, Bower et al. 2003). The previous model described in Beckert & Falcke (2002) was based on the strong rotativity case and is unable to reproduce the observed LP. The model spectra for the radio fluxes from Sgr A* shown in Fig. ?? (left) and the percentage of polarisation in Fig. ?? are results from simulations solving the radiative transfer problem (Eq. ??) along many lines of sight spread over a grid, which covers the visible surface of the outflow. Turbulence is mimicked by a superposition of waves in the B-field as described above. The model parameters for the radio spectra with turbulence driven conversion in an outflow are given 1-surface. For in Table ??. Here ~~~d is the y of electrons, which dominate the emission from the T conical, flat-spectrum jets yrad is constant along the outflow. Ri, is the width of the outflow at the start given in units of the Schwarzschild radius ( M B H= 2.6 lo6 M a , Rs = 7.8 lo1' cm). B, and B, are equally strong at z = 5Rs, which leads to the sub-mm bump due to the strong evolution of B, along the outflow. The steep jump in LP at N 100 GHz (Fig. ??) is due to the onset of turbulence at Z t u r b = 5.5Rs. Related to the start of turbulence and the sub-mm bump is a change of the effective optical depth from which emission dominates the radiation. This produces a jump in the P.A. of linear polarisation at about X 2 mm. This rapid change of P.A. roughly mimics a Faraday rotation measure RM = -2 lo5 rad/m2 in the simulations as seen in Fig. ?? (right). Its position in frequency is very sensitive to Zturb and the N

-

464

T. Beckert: A model for polarised radio emission of Sgr A*

100.00

-

10.00

6\"

u

a > C

+ .o_

1 .oo

0

.-L

N

U -

0

a

0.10

0.01

I o9

10'1 Frequency [Hz]

10'2

Fig. 2 Percentage of polarisation from the TD-model ( C P diamonds, LP: triangles) together with upper limits on LP (arrows) and confirmed LP measurements in the mm - sub-mm range (Aitken et al. 2000). The measured variability range of CP is given with circular symbols at the upper and lower bound. The expected range of variability due to turbulence from the TD-model is shown below 10'' Hz (shaded areas), where turbulence drives conversion and reduces LP.

strength of the sub-rnm bump. It might therefore be related to changes of the time variable sub-rnm flux. The viewing angle 6' is constrained to values around 45" in this particular model. For e being smaller The Table 1 Parameters used in the turbulence driven (TD-)model. The values for n, and Bin are those at the starting point with width Ri,. 18 50

Ymin

%ad Ymax

P Rin Bin

half-opening angle $) viewing angle 6' ne ufc Lkinetic Zturb

[Rsl [Gauss] [%I [degl [cmp3 at Ri,] [ergls]

tRsl

250 2.5 2.1 29 16 42 3.5 105 0.4 5.6 5.5

465

Astron. Nachr./AN 324, No. S1 (2003)

-.

"Oi 100

I

0.1 Frequency [Hz]

1.o Wavelength [mm]

10.0

Fig. 3 Left: Measured radio to IR spectrum of Sgr A" together with synchrotron outflow TD-model (solid line) in total intensity ( I ) . The variability range of the measured V is shown with circular symbols at both ends. The model results for V give the correct level of circular polarisation (diamonds). The smoothed linear polarisation P from the model are shown as a dashed line. Right: Rotation of the position angle of linear polarisation relative to the P.A. of optically thin emission beyond lo1' Hz. The model results (squares) are compared to the effect of an external Faraday screen with R M = -2 lo5 radm'.

change in LP (see Fig. ??) gets to slow for smaller 0, while for larger values of 0 the levels of L P and C P become almost equal in the radio bands. The kinetic luminosity Lkinetic is given for comparison with the radiation luminosity. The required mass accretion rate for feeding the outflow by accretion is = fpl 9 lo-'' Ma/yr with f being the 0.1 the required accretion efficiency for redirecting the particle flux into the outflow. For efficiencies f rate is comparable to the Bondi accretion rate derived from X-ray observations. The radiation efficiency of the outflow is t lop1 and therefore larger than expected for accretion in ADAF models (e.g. Beckert & Duschl 2002), which justifies the neglect of radiation from the accretion flow in the model presented here.

h'

N

N

7 Conclusions The comparison of the turbulence driven conversion model with the observations of the unique source Sgr A* in the centre of our galaxy, which is the best studied (extreme-)low-luminosity AGN, shows that the L P + C P conversion is a likely cause for CP in Sgr A* itself. Turbulence driven conversion can explain the 100 GHz when turbulence develops fast along the outflow. The stability of the observed jump in LP at sign of Stokes V signals the stable handedness of the magnetic spiral and reflects the spin of the black hole or the accretion flow. Both accretion and outflow in Sgr A" are radiatively inefficient. N

References Aitken, D. K.,Greaves, J., Chrysostomou, A,, et al. 2000, ApJ, 534, L173 Beckert, T., Duschl, W. J. 2002, A&A, 387,422 Beckert, T.,Falcke, H. 2002, A&A, 388, 1106 Bower, G. C., Falcke, H., Backer, D. C. 1999, ApJ, 523, L29 Bower, G. C., Wright, M. C. H., Falcke, H., Backer, D. C . 2001, ApJ, 555, L103 Bower, G. C., Falcke, H., Sault, R. J., Backer, D. C. 2002, ApJ, 571, 843 Bower, G. C., et al. 2003, these proceedings Chandrdsekhar, S., Radiative transfer, Dover, New York, 1960 Ensslin, T. 2003, A&A, in press, astro-ph/0212387 Jones, T. W. 1988, ApJ, 332,678 Lobanov, A. P., Zensus, J. A. 2001, Science, 294, 128 Ruszkowski, M., Begelman, M. C. 2002, ApJ, 573,485 Sault, R. J., Macquart, J.-P. 1999, ApJ, 526, L85 Schodel, R., T. Ott, T., Genzel, R., et al. 2002, Nature, 419, 694

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 467-473 (2003)/ DO1 10.1002/asna.200385081

On the Chandra Detection of Diffuse X-Ray Emission from Sgr A* M. E. Pessah*‘ and F. Melia**

‘3’

’ Steward Observatory, The University of Arizona, Tucson, AZ 85721 Physics Department, The University of Arizona, Tucson, AZ 85721

Key words Galaxy center, black hole physics, gravitation, accretion, relativity, non-thermal radiation mechanisms Abstract. Kinematic studies of the stellar motions near Sgr A* have revealed the presence of several million solar masses of dark matter enclosed within 0.015 parsecs of the Galactic Center. However, it is not yet clear what fraction of this material is contained within a single point-like object, as opposed to an extended distribution of orbiting matter (e.g., in the form of neutron stars). Recent Chandra observations suggest that the X-ray emission from this source is partially diffuse. This result provides an important clue that can be used to set some constraints on the mass distribution surrounding the black hole. Here, we develop a simple model in which the diffuse emission is produced by a halo of neutron stars accreting from the gas falling toward the center. We discuss the various accretion mechanisms that are likely to contribute significantly to the X-ray flux, and show that a highly magnetized fraction of old neutron stars may account for the diffuse high-energy source. If this picture is correct, the upper bound to the mass of the central black hole is zz 2.2 x lo6 M a . The core radius of the dark cluster must then be N 0.06 pc. We also discuss the sensitivity of our results to the various assumptions made in our calculations.

1 Introduction The large gas velocities and stellar proper motions at the Galactic Center point to the gravitational influence of a highly concentrated mass, which presumably includes a supermassive black hole within 0.1’’ of the non-thermal radio source Sgr A* (Haller et al. 1996; Genzel et al. 1996, 1997; Ghez et al. 1998, Melia & Falcke 2001). The enclosed dark mass is apparently constant from 1 pc down to almost 0.015 pc, providing a best fit value of CY 2.6 x lo6 Ma within this region. The idea that most of this mass should be concentrated within a single object is supported by arguments concerning the lack of stability of a stellar cluster that could otherwise mimic the presence of a point-like source (Maoz 1998). However, when considering the combined effects of the peculiar initial mass function in the inner parsec (apparently biased toward high masses; Morris 1993) and central migration due to mass segregation, one could reasonably expect a configuration in which the putative black hole is surrounded by a halo of orbiting dark stars. One wonders, therefore, how the dark mass is actually partitioned between the black hole and a possibly distributed component. This question deserves re-assessment in light of recent observations with the Chandru X-ray observatory. A high resolution imaging analysis of the central parsec of the Galaxy with Chandra (Baganoff et al. 2002) has apparently shown the emergence of a diffuse X-ray component associated with Sgr A*. After accounting for possible problems with the aspect solution of the instrument, Baganoff et al. (2002) concluded that Sgr A* is probably extended in the X-band with an intrinsic size of about 1”-2” in diameter * e-mail: rnpessahQas.arizona.edu,Phone: +015206216540, Fax: +01520621 1532 *%

e-mail: [email protected],Phone: 4 1 520621 9651, Fax: 4 1 520621 4721

@ 2003 WILEY-VCH Verlag GmhH & Co. KGaA, Weinheirn

M. Pessah and F. Melia: Diffuse X-ravs from Sm A*

468

(1" Y 0.04 pc). Of the various plausible mechanisms suggested to account for a diffuse X-ray component at the Galactic nucleus, only one seems to be viable (see, e.g., Baganoff et al. 2002). Assuming that BondiHoyle accretion initiates the accretion process toward the supermassive black hole (Melia 1992,1994), one may estimate, based on the conditions at the Galactic Center (see, e.g., Ruffert & Melia 1994; Coker & Melia 1997), that matter may be captured at a rate as high as N 1021gs-' by the accreting object on a spatial scale of N 0.06 pc. If this hot plasma radiates via optically-thin bremsstrahlung emission, it may contribute significantly to the diffuse X-ray source (Quataert, 2002). Here, we show that an intermediate dark mass distribution, in which the black hole has a halo of accreting neutron stars, may also account for the extended X-ray emission.

2 The Dark Cluster Model 2.1 Cluster structure and dynamics The possibility that a distribution of stellar remnants (with a total mass as large as = lo6 Ma)may fill the inner few tenths of the central parsec was suggested by Morris (1993). He envisaged a net inward migration of the more massive objects due to dynamical friction. Haller et al. (1996) also suggested that the X-ray flux from the central few parsecs is consistent with a total mass of a few times lo6 M a in stellar remnants. Based on very simple physical arguments, Maoz (1998) estimated that even under the most favorable conditions, a cluster of dark objects could not be dynamically stable for more than about 10' yr. This result is in agreement with numerical studies of the dynamical evolution of dense nuclear stellar systems which suggest that they are likely to be unstable to the formation of moderate-mass (lo3 M a ) black holes (Quinlan & Shapiro 1990). Detailed simulations following the time evolution of such systems (spanning a large range of initial core masses) have been carried out by Murphy et al. (1991). These calculations confirmed that a seed black hole of lo4 M a would increase its mass up to a value comparable to that of the cluster over a Hubble Time (i.e., roughly 15 Gyr). With the previous arguments as a starting point, we consider a scenario in which some fraction of the enclosed mass in the inner parsec of our Galaxy has not yet coalesced into the central black hole. Studying the emission produced by accretion onto this cluster would require the proper evaluation of its stellar phase space density. However, it is possible to propose a simple distribution function that captures the main features revealed by the sophisticated numerical models. The low-mass, low-density models considered by Murphy et al. (1991) have characteristics similar to those found in the central region of the Milky Way (Mc N 106-107Ma)and are therefore suitable for our present purposes. We can briefly summarize their main characteristics as follows. The overall density profiles present after a Hubble time are described by broken power-laws, with internal slopes of -1/2 that increase roughly to -7/4 at larger radii. The break-radius generally decreases with decreasing core mass and range from 10-1 pc to lop3 pc. The corresponding relaxation times are of order 108-109 yr. An appropriate density profile to match the results described above can be provided by a member of the family of 7-models introduced by Tremaine et al. (1994). In their most general form, these models allow for the presence of a central point mass. An internal slope of -1/2 can be obtained by setting the parameter 11 = 2.5. For a cluster of total mass M , and core radius T O , the density profile is given by p(r) = p 0 ( 5 / 6 ) / [ ( ~ / ~ 0 ) ~ / ~T(/ 1T o ) ~ / ' ] , with po = 3ML/47rr03. Furthermore, the relatively short relaxation time scales found in the simulations ensure that the cluster constitutes a well relaxed system at the present epoch. This means that encounters between stars have had sufficient time to efficiently set up a Maxwellian distribution in velocity space (Binney & Tremaine 1987). Based on these arguments, we adopt the following functional form for the normalized distribution function,

+

Astron. Nachr./AN 324. No. S1 (2003)

469

+

The velocity dispersion has two contributions, ~ ' ( r=) c&,,,(r) O ; ~ ( T ) .Each of these terms can be computed following the prescription described by Tremaine et al. (1994). It can be seen that the usual divergence in P near the black hole is present in all 7-models; for small T , we have u2 E 1/(4 - 7 ) r . An order of magnitude analysis of the dynamical state of the proposed system can be easily outlined using the expressions for the relaxation, collisional and evaporation time derived by Binney & Tremaine (1987). For simplicity, we assume that the cluster contains only neutron stars, that its size is about r, = 0.1 pc and that the mass of the black hole is roughly Mbh n/ 2 x lo6 Mo (we assume that the total enclosed dark mass in the inner pc is 2.6 x lo6 Ma ). We will not enforce these two conditions later on, but they are useful in providing an estimate of the time scales associated with this black holeldark cluster model. The = 100 km s-'. Defining the characteristic crossing time velocity scale is set roughly by u, = as t,,,,, = r,/7), we can obtain an estimate for the relaxation time tTel = O.l(N/ 1nN) x t,.,,, = lo7 yr. For the collisional time we have t c o ll Y 8&nGM*R,/u = 1014 yr where we have used n, ru lo8 P C - ~ . Thus, the inferred dynamical state is such that the cluster is well relaxed, while collisions (taken as destructive events) have not played an important role. We can estimate the evaporation time for the cluster as teuap= 300 trel. This value, suggestively close to a Hubble time, implies that the density of neutron stars in the cluster may be set by evaporation over that period.

d-

2.2

Accretion onto each compact object

Having established the global properties of the cluster, we now explore different aspects of the accretion of matter by each compact object. The accretion rate onto a given star at a distance r from the black hole is given by Ak* = 7rriccp,(r) [uzeE(r) c 2 { r ) ] where r,,, is the accretion radius, p g ( r ) is the local gas density fixed by the black-hole accretion rate A&, = 47rr2pg(r)ug(r),and v,,l(r) = [ U ; ( T ) 71, 2vg(r)uC O S @ ] ' / ~ is the relative velocity between the star and the gas falling into the black hole (with 0 being the angle subtended between v and the radial direction from the black hole). Finally c(r) is the local sound speed. To compute u g ( r ) ,we assume that the gas is falling freely in the potential of the compound system formed by the cluster and the black hole. For c ( r ) , we further assume an adiabatic inward flow with an ideal equation of state. In order to calculate the X-ray luminosity emerging from each neutron star, we need to investigate the dominant mechanism for accretion under the conditions present at the Galactic Center. Let's assume for the moment that all the neutron stars are unmagnetized and that they accrete via the classical Bondi(for Hoyle mechanism. The accretion radius for a given star will then be given by rgrav = c2)]1/2= 1000 km neutron stars M* = 1.4 Mo). Due to the high effective velocities ( u e f f = [(& s-') present close to the black hole, the average achievable accretion rate would be of order 1O6-1Os g s-' (corresponding to a total luminosity around 1026-1028erg s-'). These low rates translate into low effective temperatures (- 1-10 eV), and matching these results to the X-ray luminosity observed from Sgr A* ( L x = 2 x los3 erg s-l) would then require a number of accreting objects whose total mass will exceed the measured enclosed mass in the inner parsec. In a more realistic scenario, the neutron star population would have a distribution of magnetic field strengths ranging up to supercritical values of 1014-1015G. There is now circumstantial evidence for the presence of such strong magnetic fields at least in a fraction of recently born, rapidly rotating neutron stars (Duncan & Thompson 1992). These objects are known as magnetars and their birthrate has been estimated to be Y 10% of all isolated pulsars (Kouveliotou et al. 1994; Kouveliotou et al. 1998). The release of stored magnetic energy in magnetars seems to be the origin of the bursting activity observed in Soft Gamma-ray Repeaters (SGRs) and the quiescent emission of both SGRs and Anomalous X-ray Pulsars (AXPs). If the quiescent emission is powered by magnetic field decay, the lifetime of magnetars as bright X-ray sources (L: = 10"-1036 erg s-l) is estimated to be of order 105-10Gyr (see, for example, Hey1 & Kulkarni 1998).

+

'',,

+ +

+

M.Pessah and F. Melia: Diffuse X-ravs from S e r A*

470

In the context of strong magnetic fields, a recent investigation by Rutledge (2001) points to a different accretion mode for fast moving old neutron stars whose magnetic field has not decayed below 1014G over a Hubble time. This value for B is the lower bound acceptable for the neutron star to go through the propeller phase, and thus be able to accrete efficiently, in less than 10 Gyr. These magnetically accreting objects (called MAGACs) present an effective cross section whose value is determined by the length scale of the magnetosphere of a non-rotating neutron star, rmag = [p2/(47rpvzff)]1/6,with a magnetic moment p = BR2/2. Unlike their Bondi-Hoyle counterparts, for which the amount of accreted matter is a decreasing 0: v);, MAGACs become more efficient accretors the faster function of the effective velocity (A&,,,,

Mmag/ugrav

they move (ilkmagcx vi$y). For a given magnetic field and flow density, the ratio scales as (vejf/lOO km s-*)"/~ (see Rutledge 2001). Therefore, for the typical velocities we are considering here, old strongly magnetized objects are more likely to contribute significantly to the observed X-ray emission.

2.3 Emergent Flux We now develop the theoretical framework needed to determine whether parameter values may be found (primarily Mbh and ro) that provide an acceptable match to the Chandra observations. The flux distribution in the X-band (as a function of projected distance) may be calculated as follows. The energy per unit time per unit volume being liberated in the range of frequencies between u and v dv, at the position r away from the black hole can be written as:

+

The factor S represents the fraction of objects that are efficient magnetic accretors. To set the scale, and as a first approximation, we assume that they constitute 10% of the total population of neutron stars and thus 6 = 0.1. In order to calculate this emissivity we need to know how the emerging spectrum from each star varies as a function of the accretion rate. In the present case, even with the high velocities and densities involved, the accretion rates are several orders of magnitude (between 6 and 7) lower than the corresponding Eddington limit for neutron stars. Calculations by Zampieri et al. (1995) showed that the expected spectrum for low luminosity (i.e., low accreting) neutron stars is slightly diluted and harder than the corresponding blackbody spectrum with the same luminosity. As a first approximation we assume that the emergent spectrum from each star is a diluted blackbody and thus L*,(r,v) = 47r2R~y-4B,(Tc,1(r, v)). where R, = 10 km, B, is the Planck function and Tcol= YTejf.The hardening ratio y is of order 2.5 for the accretion rates quoted earlier (see Zampieri et al. 1995). The effective temperature for each star will be fixed by the condition L$ = 47rR:~T$~ = (GM*/&)&f,(r, v). Assuming isotropy we can obtain the total specific intensity emerging at the projected distance R in the plane of the sky by integrating the emissivity per unit solid angle jV/(47r) along the line of sight

The total intensity, I x (2-10 keV) can then be obtained by integration over the corresponding frequencies. To compute the flux distribution we integrated I X over the successive solid angles.

3 Results The physical scenario suggested by observations of the inner parsec is such that wind velocities are in the range vw N 500-1000 km spl while number densities are of the order of lzw N 103-104crnp3. Under these circumstances the classical Bondi-Hoyle mechanism leads to an accretion rate &fbh N 1021-1022g s-' (Coker & MeIia 1997). The parameters defining the accretion rate onto the black hole in our model

47 1

Astron. Nachr./AN 324. No. S1 12003)

10

" " " " " " " " " ' " ' " " ' ~ '

:!

T 0 Theoretical prediction

0 Chondro observation

1

0-

1

I . O 0

L

,

,

,

,

,

,

,

,i,, j,, ,

2

1

8

,

,

,

,!

1 3

"'1

Fig. 1 Surface brightness of the diffuse component versus viewing angle from S g r A* (indicated as corrected flux measured on earth). Open circles indicate values inferred from Chandra data. The scaled profiles of two nearby sources (CXOGC 5174538.0-

290022 and CXOGC 5174540.9-290014), were subtracted from Sgr A* profile to obtain the diffuse component. Filled circles represent our best fit model With Mbh = 2.2 X lo6 and T g = 0.06 pC.

-2.

I

-

-. -

-6 -

-6

-4

-2

0

2

4

6

Fig. 2 Two dimensional spatial rendering of the psfconvolved calculated diffuse emission, coded on a linear scale of shading. Black corresponds to 6.5 x erg s-' cm-2 and the lightest shade of gray erg s-' cm-'. In order to get this to 4 x image the original bidimensional flux map was convolved with a PSF obtained from observations of two nearby sources, CXOGC 3174538.0-290022and CXOGC 5174540.9-290014, (FWHM N 1").

were set to &,f, 'v 10" g s C 1 CJ Ma yr-' and ugas(cm) 'v 1100 km s-l (which implies a medium number density n(m)N 3 x lo3 cmp3). The magnetic field for 10% of the total neutron star population G. These values were set as reasonable upper limits in order to get an estimate was fixed at B = 5 x of a lower bound for the number of stars needed to match the observed luminosity. A comparison of our best fit model with the Chandra data is shown in Figure 1, and Figure 2 gives the corresponding two-dimensional spatial rendering of the psf-convolved calculated diffuse emission. The best fit values obtained were Mbh = 2.2 x 10' Ma and TO = 0.06 pc implying a total X-ray luminosity LX N 3 x erg s-'. The typical accretion rates found among the magnetically accreting neutron stars were of order lo1'- lo1' g s-', implying L$ 'v 1031-1032erg s-l and therefore Tcolru 10-100 eV. When the neutron stars' magnetic field was set instead to B = 5 x 1014 G we obtained Mbf, = 10' A& and TO = 0.03 pc. In this case, the typical accretion rates were of order 10l1 g s-', with L; 'v 1031 erg s-l and thus TcolN 10 eV. Although the emergent flux was similar to the previous one, this model disagrees with the measured enclosed mass function. Figure 3 shows, together with the data for the measured enclosed mass in the inner parsec, the two 77-models corresponding to the two sets of best fit parameters. For B ru l O I 5 G the inferred enclosed mass is clearly in good agreement with the data, while that corresponding to B N 1014 G is almost 3cr away from the innermost data point.

4 Discussion and Conclusions Once the best fit parameters were determined in each case, we examined the sensitivity of the resulting flux profile and luminosity by exploring the two regions shown in Figure 4. Denoting the best fit values with bf, the following general behavior was found. For values of TO < rif the flux profile was peaked and the luminosity increased beyond erg s-', while for T O > rEf the values of the luminosity were acceptable but the flux was too extended. For Mbh < M,",f the flux profile was similar to the one in Figure 1, but the luminosity was of order the luminosity was again of order erg s-'. Finally, for Mbh >

Mil

M. Pessah and F. Melia: Diffuse X-rays from Sgr A*

472

0

-

-1

d

I

0 m

-

-2

-3 0.001

0.1 1 Distance from Sgr A‘ [pc]

0.01

10

Fig. 3 Plot of the enclosed mass versus distance from Sgr A* for two extreme models considered in this paper. The dashed curve is a Plummer model for the visible matter. Open and filled circles correspond to total mass and inferred dark matter. The thick solid curve corresponds to the best fit model shown in Figure 1 ( B = 5 x lo1‘ G). The dotted curve corresponds to a similar model except for the lower bound adopted for the magnetic field ( B= 5 x l O I 4 G).

0.5

1 .o

1.5 Mbh

2.0

2.5

[ 1 06Ma1

Fig. 4 Contour plot for the enclosed mass versus radius on the r o - h f b h plane. Each curve is labeled by its specific x2 value, ranging from 0.1 to 3.0. In addition, the two points corresponding to the theoretical curves in Figure 3 are indicated by the circle (the best fit model) and the square (correspondingto the lowest value of h f b h for which a reasonable fit to the diffuse emission could be obtained).

1032-1033erg scl and the flux profile was lower on average. Note that the value of X2/dof is about 0.2 for the case with B 2i 1015 G, while it is around 1 for B N 1014 G . These results have several important implications for our understanding of the black-hole nature of Sgr A*. Clearly, this model does not provide an alternative to the black-hole scenario. On the contrary, it requires the presence of a massive point source to create a deep potential and provide the physical conditions for efficient accretion. However, it is no longer clear that Mbh = 2.6 x lo6 Ma.In fact, if the diffuse X-rays are produced by the cluster described here, our results point to an upper bound for the black-hole mass of 2.2 x lo6 M a . This value is in agreement with independent lower values obtained by numerical simulations of the dynamics of a massive black hole under the gravitational influence of a dense cluster. For example, Chatterjee et al. (2002) derived a lower bound of 1.1 x lo6 M o for the black hole at the Galactic Center. In this regard, it may be worth recalculating the emissivity of Sgr A* in models that depend rather sensitively on the black-hole mass. We have assumed that the cluster contains only neutron stars and have neglected any possible representation from white dwarfs and solar mass black holes. While these are unlikely to contribute significantly to the X-ray emission (see Haller et al. 1996), they may be important dynamically. We have also assumed that a fraction (1 0 %) of the neutron stars have strong magnetic fields that have not decayed below 1014 G over a Hubble time. Deviations from these assumptions should be reflected in the inferred mass of the dark cluster through the factor 6 in Eq. (2). It is also worth noting that our model implies that roughly 4 x lo3 strongly magnetized neutron stars are present within 1” of the supermassive black hole. If any of these erg s-’), the inferred luminosity were young neutron stars powered by magnetic field decay ( L x > would be inconsistent with current observations. However, star formation at the Galactic Center occurs in bursts, the latest of which peaked some 10-100 million years ago. It is therefore reasonable to assume that the last generation of magnetars faded long ago. Considering the simple model proposed here for the dynamical structure of the cluster, the overall agreement with the observed diffuse emission is reasonably good. The model appears to be viable and warrants additional, detailed study. Theoretically, the dynamics of such a cluster needs to be better understood, particularly with regard to its stability. In this work we set the parameter in the 7 model to 2.5. This value

Astron. NachrJAN 324, No. SI (2003)

473

is physically motivated because it produces the internal slope of -1/2 in the density profile observed in earlier numerical simulations. Nonetheless, to test the sensitivity of our results on 77, we repeated our calculations with 71 = 2 (internal slope of -1). The results showed that it is possible to obtain an acceptable description of the actual flux distribution but the fit was not as good as the one shown in Figure 1. The mass of the cluster suggested by the best fit model for 77 = 2 was similar to the one found for 17 = 2.5, but the core radius was this time E 0.1 pc. Clearly, obtaining a more refined stellar density profile will reflect on the ultimate distribution of the diffuse emission. Also, the emergent spectrum from each star needs to be taken into account with greater care, since this will modify the number of stars required to match the observed luminosity. Finally, on the observational side, it would be desirable to acquire the spectral component of the diffuse emission separately from the point-source emission. Acknowledgements We would like to thank Daniel Eisenstein, Dennis Zaritsky and Scott Tremaine for useful discussions. We are also very grateful to Fred Baganoff and Mark Moms for providing the data prior to publication. This research was partially supported by NASA under grants NAG54239 and NAG5-9205, and has made use of NASA's Astrophysics Data System Abstract Service. FM is grateful to the University of Melbourne for its support (through a Miegunyah Fellowship).

References Binney J. & Tremaine S. 1987, Galactic Dynamics (Princeton Univ. Press, Princeton). Baganoff F. K., Maeda Y., Moms M. et al. 2002, ApJ, in press (astro-ph/0102151) Chatterjee P., Hernquist L. & Loeb A. 2002, ApJ, 572.37 1 Coker, R.F. & Melia, F. 1997, ApJ, 488, L149 Duncan, R. &Thompson, C. 1992, ApJ, 392, L9 Genzel R., Thatte N., Krabbe A,, Kroker H. & Tacconi-Garman L. 1996, ApJ, 472, 153 Genzel R., Eckart A,, Ott T. & Eisenhauer F. 1997, MNRAS, 291, 219 Ghez A,, Klein B., Morris M. & Becklin E. 1998, ApJ 509,678 Haller J., Rieke M., Rieke G., Tamblyn P., Close L. & Melia F. 1996, ApJ, 456, 194 Hey1 J.S. & Kulkarni S.R. 1998, ApJ, 506, L61 Kouveliotou, C. et al. 1994, in: AIP Conf. Proc. 307, Second Huntsville Gamma-Ray Burst Workshop, edited by G. Fishman, J. Brainerd & K. Hurley (New York ATP), 167 Kouveliotou, C. et al. 1998, Nature, 393,235 Maoz E., 1998, ApJ, 494, L181 Melia, F. 1992, ApJ, 387, L2.5 Melia, F. 1994, ApJ ,577, 426 Melia, F. & Falcke, H. 2001, ARA&A, 39, 309 Morris M. 1993, ApJ, 408,496 Murphy B. W., Cohn H. N. & Durisen R. H. 1991, ApJ, 360,60 Quataert E. 2002, ApJ, 575, 855 Quinlan G. & Shapiro S. 1990, ApJ, 356,483 Ruffert, M. & Melia, F. 1994, A&A, 288, L29 Rutledge, R. 2001, ApJ, 553,796 Tremaine S., Richstone D. O., Byun Y.I., Dressler A., Faber S. M., Grillmair C., Kormendi J. & Lauer T. R. 1994, AJ, 107,634 Zampieri, L., Turolla, R., Zane, S. & Treves A. 1995, ApJ 439, 849

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S I , 475 -48 I (20031 / DO1 10.1002/asna.200385045

A Relativistic Disk in Sagittarius A*

’

Siming Liu * and Fulvio Melia’ * 2

’ Center for Space Science and Astrophysics, Stanford University, Stanford, CA, 94305-4060

’ Department of Physics and Steward Observatory, The University of Arizona, Tucson, AZ,85721 Key words Supermassive Black Hole, Galactic Center, Radio Astronomy. PACS 04A2.5 The detection of a mm/Sub-mm “bump” in Sgr A*’s radio spectrum suggests that at least a portion of its overall emission is produced within a compact accretion disk. This inference is strengthened by observations of strong Linear polarization (at the 10 percent level) within this bump. No linear polarization has been detected yet at other wavelengths. Given that radiation from this source is produced on progressively smaller spatial scales with increasing frequency, the mm/Sub-mm bump apparently arises within a mere handful of Schwarzschild radii of the black hole. We have found that a small (10-Schwarzschild-radii) magnetized accretion disk can not only account for the spectral bump via thermal synchrotron processes, but that it can also reproduce the corresponding polarimetric results. In addition, the quiescent X-ray emission appears to be associated with synchrotron self-Comptonization, while X-ray flares detected from Sgr A* may be induced by a sudden enhancement of accretion through this disk. The hardening of the flare-state X-ray spectrum appears to favor thermal bremsstrahlung as the dominant X-ray emission mechanism during the transient event. This picture predicts correlations among the mm, IR, and X-ray flux densities, that appear to be consistent with recent multi-wavelength observations. Further evidence for such a disk in Sgr A* is provided by its radio variability. Recent monitoring of Sgr A* at cm and mm wavelengths suggests that a spectral break is manifested at 3 mm during cm/Sub-mm flares. The flat cm spectrum, combined with a weak X-ray flux in the quiescent state, rules out models in which the radio emission is produced by thermal synchrotron process in a bounded plasma. One possibility is that nonthermal particles may be produced when the large scale quasi-spherical inflow circularizes and settles down into the small accretion disk. Dissipation of kinetic energy associated with radial motion may lead to particle acceleration in shocks or via magnetic reconnection. On the other hand, the identification of a 106-day cycle in S g r A*’s radio variability may signal a precession of the disk around a spinning black hole. The disk‘s characteristics imply rigid-body rotation, so the long precession period is indicative of a small black-hole $pin with a spin parameter a / M around 0.1. It is interesting to note that such a small value of a / M would be favored if the nonthermal portion of Sgr A*’s spectrum is powered by a BlandfordZnajek type of process; in this situation, the observed luminosity would correspond to an outer disk radius of about 30 Schwarzschild radii. This disk structure is consistent with earlier hydrodynamical and recent MHD simulations and is implied by Sgr A*’s mm/Sub-mm spectral and polarimetric characteristics. For the disk to precess with such a long (106-day) period, the angular momentum flux flowing through it must be sufficiently small that any modulation of the total angular momentum is mostly due to its coupling with the black-hole spin. This requires that the torque exerted on the inner boundary of the disk via magnetic stresses is close to the angular momentum accretion rate associated with the infalling gas. Significant heating at the inner edge of the disk then leaves the gas marginally bounded near the black hole. A strong wind from the central region may ensue and produce a scaled down version of relativistic (possibly magnetized) jets in AGNs.

* Corresponding author: e-mail:

[email protected], Phone: +01650 723 01 12, Fax: 4 1 650 723 4840

* * Miegunyah Fellow

@ 2003 WILEY-VCH Verlag GmhH & Co K G A Weinheim

S. Liu and F. Melia: Relativistic Disk

476

1 Introduction The compact radio source, Sgr A*, at the dynamical center of our Milky Way Galaxy, is believed to be associated with a supermassive black hole (Melia & Falcke 2001). Evidence in support of this is quite compelling, especially with the detection of a 3 hour X-ray flare (Baganoff et al. 2001) from the direction of Sgr A* and the recent monitoring of a star orbiting within light-days of the black hole, which points to a central mass of 3.7 f 1.5 x lo6 Ma (Schodel et a1 2002), consistent with an earlier measurements of 2.6 x lo6 A4, (Eckart & Genzel 1996; Ghez et al. 1998). The nature of Sgr A* thus bears critically on our understanding of black-hole physics. Several mechanisms have been proposed over the past decade to explain its broad band spectrum and polarization, among them Bondi-Hoyle accretion (Melia 1992); a two-temperature, viscous disk (Narayan, Yi & Mahadevan 1995); a relativistic nozzle (Falcke & Markoff 2000) and a recent combination of advection-dominated disk with a nozzle (Yuan, Markoff & Falcke 2002). Over the past decade or so, our group has adopted a theoretically-motivated phenomenological approach (Melia, Liu & Coker 2000,2001 j, in which the observations play a crucial role in constraining the theoretical picture. The detection and confirmation of a mm/Sub-mm bump in Sgr A*'s spectrum suggests an emission component different from that responsible for the cm radio emission (Zylka, Mezger & Lesch 1992; Falcke et al. 1998). This emission component is also implied by Sgr A*'s variability. Radio Observations show that Sgr A*'s fluctuation amplitude increases toward high frequency (Zhao & Goss 1993) and there is a spectral break at 3 mm during radio flares (Zhao et al. 2003). Since high-frequency radio emission is produced by relatively more energetic particles, located deeper in the gravitational well of the black hole (Melia, Jokipii & Narayanan 1992), the mm/Sub-mm emitting gas should be very close to the black hole's event horizon. The detection of linear polarization in this spectral bump enhances this inference further and sets severe constraints on possible explanations for this component (Aitken et al. 2000). No significant linear polarization has yet been detected at frequencies lower than 112 GHz, though relatively strong circular polarization persists in the cm band (Bower et al. 2002). The flip of the position angle of the polarization vector by about 90" between 230 GHz and 350 GHz favors a scenario where the mm/Sub-mm emission is produced within a small, optically thin, magnetized accretion disk (Melia et al. 2000). No other model so far can explain this linear polarization characteristic (In the empirical model of Ago1 (2000), the frequency where the position angle flips by 90" is much lower than the frequency corresponding to the spectral peak of the flux density, which is not in line with the observations). The existence of a small disk is also motivated theoretically. Earlier hydrodynamical simulations suggested that black-hole accretion from stellar winds, as is the case for Sgr A*, is characterized by a small angular momentum of the captured gas (Coker & Melia 1997). This accreted angular momentum is too small for the gas to settle onto a large disk, as required by the ADAF model (Narayan et al. 1995). The captured angular momentum is instead barely sufficient to circularize the gas just before it falls across the black hole's event horizon. Detailed MHD simulations have provided an indication of the structure for such a disk (Hawley & Balbus 2002). The fact that the Magneto-Rotational Instability (Balbus & Hawley 1991) can induce a MHD dynamo in the disk provides a straightforward explanation for the mm/Sub-mm bump as the result of synchrotron process. The magnetic field also provides an anomalous viscosity and, given its strength, couples the electrons and ions via reconnection, so that a single temperature fluid is maintained (Melia et al. 2001j. X-ray observations of Sgr A* have provided additional means of learning about its nature. It turns out erg s-l that Sgr A* is an extremely weak X-ray source with a quiescent X-ray luminosity of 2.2rE x (Baganoff et al. 2001j. Interestingly, electrons responsible for the mm/Sub-mm emission also Comptonize the radio photons into the X-ray band. It is notable that the physical conditions required to produce the mm/Sub-mm spectrum can also account for Sgr A*'s quiescent X-ray emission. Chundvu also detected a strong X-ray flare from the direction of Sgr A*. The flare lasted about 3 hours and featured a variation on a 10 minute time scale, suggesting an emission region no bigger than 20 T S , N

477

Astron. Nachr./AN 324.No. SI (2003)

where T S = 7.7 x 1011 cm is the Schwarzschild radius for the 2.6 x lo6 111,supermassive black hole associated with Sgr A*. The peak flux density for this flare is 50 times higher than that in the quiescent state. Recent X-ray observations have shown that this type of X-ray flare is common to Sgr A*, occurring about once per day (Baganoff et al. 2003; Goldwurm et al. 2003). The fact that the flares are very strong, are variable on a 10 minute time scale and have a flat spectrum, pose telling theoretical challenges which, at the same time, also create a valuable opportunity for constraining the physical conditions near the black hole’s event horizon. Our study has shown that an enhancement of the mass accretion rate through the disk can not only account for these flares, but can also induce strong Sub-mm/Far-IR flares that should occur simultaneously with the X-ray flares (Liu & Melia 2002a). Recent observations of Sgr A*’s mm/Sub-mm variability has indicated that the Sub-mm spectral index increases significantly during radio flares (Zhao et al. 2003), consistent with our prediction. However, radio flares usually last for several days, which is much longer than the duration of an X-ray flare, suggesting more complicated physical processes. Nevertheless, the nondetection of an IR flare (Hornstein et al. 2002) seems to favor this model, where X-ray flares are produced via thermal bremsstrahlung processes, over the nozzle model, in which synchrotron self-Comptonization is introduced to account for the X-ray flare emission (Markoff et al. 2001). Moreover, the low quiescent X-ray flux also delimits hot gas content around Sgr A*. When this constraint is combined with the flat radio spectrum, one can show that the cm radio emission from Sgr A* cannot be produced by a bounded, thermal synchrotron source (Liu & Melia 2001). One possibility is that the radio emission is produced via nonthermal synchrotron processes in the region where the large scale quasi-spherical inflow circularizes to form the small accretion disk responsible for the mm/Sub-mm and X-ray emission. Energetic, nonthermal electrons can in principle be produced by the dissipation of kinetic energy associated with the radial motion of the infalling gas in shocks or magnetic reconnection. Assuming that a fixed fraction of particles is accelerated in this way, one can obtain a good fit to the radio spectrum. The circular polarization properties may then be associated with the turbulent nature of the gas in this region (see, e.g., Beckert & Falcke 2002; Ruszkowski & Begelman 2002). On the other hand, a 106 day period in Sgr A* radio variability recently reported by Zhao et al. (2001) appears to be associated with the precession of a small hot disk under the influence of a spinning black hole (Liu & Melia 2002b). The physical characteristics of the disk indicate that it will precess as a rigid-body. However, for the disk to survive longer than the observed period, the net angular momentum flux through the disk must be extremely small, which requires that the inward angular momentum flux associated with the accreting gas must be cancelled almost completely by the outward angular momentum induced by torque associated with the magnetic stresses. A nonzero torque at the inner edge of an accretion disk has been discussed extensively (Krolik 1999; Gammie 1999; Ago1 & Krolik 2000) during the past few years. Recent MHD simulations have also confirmed several of these theoretical speculations (Hawley & Balbus 2002). Should this picture be correct, it should be noted that a small black hole spin of 0.1 M , where M is the mass of the black hole, would be favored if the nonthermal portion of Sgr A*’s spectrum is instead powered by energy extracted from the black hole via a Blandford-Znajek type of process. The 3Ors, consistent with the general precession period then requires that the disk has an outer radius of picture described above. The power extracted from the black hole also heats up the gas near the event horizon and unbinds it. The ensuing wind is not unlike the relativistic jets observed in AGNs. Further exploration of this idea may eventually reveal a more refined view of the processes hidden in the central engine of these sources.

-

N

2 A Relativistic Disk Model for the mm/Sub-mm Emission from Sgr A* The model of a hot, magnetized, small accretion disk in Sgr A* has been developed fully in the paper by Melia et al. (2001) where, prior to the availability of all the observational constraints described above, the inner boundary condition was chosen to have zero torque. In this instance, the temperature at the outer boundary of the Keplerian region is the primary free parameter. The disk structure is determined once

47 8

S. Liu and F. Melia: Relativistic Disk

one specifies the inner (rJ and outer ( r o )radii, the magnetic (&) and viscous (By)parameters, the mass accretion rate hk and the inclination angle of the disk. The best fit to the mm/Sub-mm polarization and spectral data is shown in Figure 1 (Melia et al. 2000). Note that here a negative percentage means that the position angle of the polarization vector is parallel to the angular momentum vector of the disk, while positive polarization means that the polarization vector flips by 90" with respect to negative polarization. 3.3 x 10l1 Hz, which is higher than the peak The frequency at which the polarization vector flips is frequency of the flux density, 2.1 x lo1' Hz. This is a unique feature of our relativistic disk model, that is apparently not yet matched by alternative scenarios (cf. Ago1 2002). It is straightforward to understand these polarization characteristics. At mm wavelengths, the red shift side of the disk becomes optically thin first. At this point, the emission is mostly from the front and back of the black hole, where it is polarized in the direction parallel to the disk's spin axis due to the influence of the very strong toroidal field within the disk. At Sub-mm frequencies, even the gas to the front and back of the black hole becomes optically thin, and the emission from the blue shifted side of the disk dominates; the polarization vector thus flips by 90". Faraday rotation by the intervening plasma will make the observed flip of the polarization vector different from go", which can reconcile the slight difference between the theoretical prediction and the observational results. Due to the relatively poor angular resolution of JCMT (22" at 220 GHz), the corresponding error bars are. quite big, as can be seen from Figure 1 (Aitken et al. 2000). However, the detection of strong linear polarization is quite obvious. Recent high resolution (3.6" x 0.9") BIMA observations have confirmed strong linear polarization at 220 GHz (Bower et al. 2003), adding some confidence to the model. N

N

4

1

I

,

,

,

I

.

,

,

.

.

,

.

RIUA JCW

0

-.,

-2

\

r '

I

-4

5 -6

-8

-10

Fig. 1 Best fit to the linear polarization of radio emission from Sgr A*. Here the model parameters are as follows: M = 4.1 x 1OI6g s-', PP = 0.02, Py = 0.2, ri = 1.8rs and r , = 8.5 r s . The inclination angle of the disk is 30". At the outer boundary, the gas temperature is fixed by the assumption that the thermal energy of the gas equals 7% of its dissipated gravitational energy.

Fig. 2 Best fit to Sgr A*'s quiescent spectrum. The model parameters are shown in the figure. Here the disk has an inclination angle of 45". The thermal energy of the gas is assumed to equal 80% of its dissipated gravitational energy at ro. The dashed line here denotes emission from the small disk. The dotted line gives emission produced by nonthermal particles in the circularization zone.

3 X-ray Emission from the Relativistic Disk Chandra observations indicate that quiescent X-ray emission from Sgr A* is very weak and soft, with an X-ray luminosity 2.2-t::; x erg s-l and a spectral index 1.5:$, which is not consistent with an ADAF (Baganoff et al. 2001). Given the fact that the plasma is so hot in the disk that electrons are relativistic

Astron. Nachr./AN 324, No. S 1 (2003)

419

(Melia et al. 2001), one is motivated to investigate the effects of synchrotron self-Comptonization (SSC) in this medium. Radio variability observations and recent theoretical developments (e.g., Liu & Melia 2002b) favor a nonzero stress at the inner edge of the accretion disk. We will henceforth adopt a zero angular = -P,&R,T/prR. The momentum flux condition. The radial velocity of the accretion flow is then IJ, other equations derived in Melia et al. (2001) are still applicable and we won't reproduce them here. In Figure 2, we provide the best fit to Sgr A*'s broadband spectrum. SSC evidently accounts for Sgr A*'s quiescent X-ray spectrum very well.

2

,x

0

-2

\

-.'

-zs

-4

-6

-a 2

--. -

0

-2

2

-4

-b

-6

0

-8 -10

I

...

I . . . I . . . I . . . l . . . -

Fig. 3 An accretion induced X-ray flare from Sgr A*. The righthand panels give the temperature (thin lines) and density profiles (thick lines) for the disk ( T ~= r s / 2 ) .The left panels show the corresponding disk spectra. Note that the model prediction is consistent with the IR upper linuts (Hornstein et al. 2002).

However, Sgr A*'s X-ray emission during a flare is much more complicated than that during quiescence. Although the short variation time scale of 10 minutes is consistent with the flare being induced by an accretion process in the disk, the fact that the X-ray flux density can increase by a factor of 50 suggests dramatic changes in the disk's structure. Moreover, the hardening of the X-ray spectral index also rules out SSC as the dominant X-ray emitting mechanism during the flare (Liu & Melia 2002a). We note that the hot disk described here is not stable when the mass accretion rate is large. When A? increases, bremsstrahlung cooling becomes more and more important and can be the dominant cooling mechanism. The X-ray flare may in fact be associated with enhanced thermal bremsstrahlung emission. Figure 3 depicts such a scenario. Here the disk has an outer radius of 9 rs and we assume that the enhancement of accretion through it can suppress the MHD dynamo. A justification for this is that cooling becomes more efficient with increasing Ak, and this decreases the gas temperature. We infer that & is thus anti-correlated with A?. During the X-ray flare, the gas density can be as high as 10'' cmP3 and the magnetic field reaches 100 Gauss. The corresponding synchrotron cooling time is a few hours, consistent with the general picture outlined above (Petrosian 1985). Recent multi-wavelength observations have indicated that there is no obvious flux change at 3 mm during the X-ray flare (Baganoff et al. 2003). According to our model, the disk is optically thick at 3 mm, so the flux density is not expected to change significantly (see Figure 3).

4 The Nature of Radio Emission from Sgr A* In Figure 2 we shiiwcd a fit tii thc cm cmissiun from Sgr A* under the a s s u r n p h lhat this radiation is produccd via nonthcrinal synchrotrrin prrmsscs in thc circularir,ation zonc (the Inodcl ddails can he f'uund in I,iu & Mclia 2001 ). The dintritutirin ( i f nonthennal Iurticles i s givcn hy N(E.r.) : 1 . 7 x 10 I:' - " ' 7 ~ ( r )whcrc % E is the electron encrgy, and 71,(r)i s thc cleclron density at radius T . Althuugh magnetic rcctinnection at smaller radii of the disk can also induuc pitrticlc accelcraliun (adding to the contribufiun made by (he nonihcmal parliclcs in thc circul;ir,zation zone), the fact h a t the gas tempr.raturc is as hiph as 10 M c v therc appcars to make thc thcrrnal proccss dominarit. .UevcrhlcPs, ii complete

treatmcnt of this pruhlctn i n c o q i o r k n g particle acceleraticlrr via magnetic 1,econncctiori is warranled. Vie can understarid the nnnthcrinal natiIrc uf cin radio emission using Ihc following argument. Hccausc quiesccnt X-+ly emission from Sgr A" i s w a k (sec Fiigurc 2)- H'C can constrain thc hut gas contcrit in Sgr A + via its hremsslrahlung emissivity. If u'c a.wrIic That thc gas is bounded, h e gas lcmpcraturc musL he lower than its virial value. Cornhining thcsc two upper limits, one can show that to produce the 1.36:[.;Hi: flux from Sgr A * rhc miignctic field energy density musl he moxc than ten times higgcr than thc thermal cncrgy densit) of the hoi pas. Such ii configuration is no1 physical iT the rrisgnctic field is intrinsic to thc hot gas. Of coursc, it is iilso prissihlc lhal the radio einisqion is produccd by m n c unhonitdcd plaxma, as prtiposcd iri the Jet model hy Markoff arid Palckc (2000). Then t h e origin of the jet becomes !he largc unknown i n thc rririrlcl. Tlic dctecliun o f a 106 d;iy radio cycle is intriguing hecauye i L is intrinsic lu Sgr A" (%hat> ti al. 20(11) and rcccnt V I A o1)aervatiuris in& catc [hat ihe erriission is prnduccd x i t h i n 1,40r,! (Hoacr ct al 20132). Thc dynainical t i m e scalt. within such a sniall region is muck sburlcr than this period, suggesting il inay he as.m%led with ail iniriii
Astron. Nachr./AN 324, Yo. SI (2003)

48 I

fraction of the gas is accelerated (nonthermally), either via energy extracted from a spinning black hole, or by energy liberated below the marginal stable orbit. These energetic particles then diffuse to larger radii. The flat radio spectrum is produced by these particles. The precession of the disk around the spin of the black hole can lead to a corresponding modulation of the outflow perceived in projection as well, inducing a periodic variation of the radio flux density.

5

Conclusions

Our study has shown that a hot, magnetized, relativistic accretion disk plays an essential role in revealing the nature of Sgr A*. Future observations of this source at 690 GHz (private communication with J. H . Zhao), combined with current X-ray observations, will help us to understand the interaction between the disk and the black hole, which is believed to be a key element of all AGNs. A comprehensive investigation of the supermassive black hole at the Galactic Center may eventually help us to unravel the mysterious inner workings of the most powerful engines at the nuclei of active galaxies.

Acknowledgements S. Liu thanks the organizing committee of the 2002 GC conference for financial assistance. This research was supported by NASA grants NAG5-8239 and NAG5-9205 at Arizona, and by the Center for Space Science and Astrophysics at Stanford.

References Agol, E. 2000, ApJ, 538, 121L Agol, E., & Krolik, J. 2000, ApJ, 528, 161 Aitken, D.K. et al., 2000, ApJ, 534, L173 Baganoff, F. et al., 2001, Nature, 413, 45 Baganoff, F. et al. these proceedings eds. Falcke, H. et al. 2003 Balbus, S.A., & Hawley, J.F. 1991, ApJ, 376,214 Bardeen, J.M., & Petterson, J.A. 1975, ApJ, 195, L65 Beckert, T., & Falcke, H. 2002, A&A, 388, 1106 Blandford, R., & Znajek, R. 1977, MNRAS, 179,433 Bower, G., Falcke, H., Sault, R.J., &Backer, D.C. 2002, ApJ, 571, 843 Bower, G., Falcke, H., Sault, R.J., &Backer, D.C. 2003, ApJ, 588, 331 Coker, R.F., & Melia, F. 1997, ApJ, 488, L49 Eckart, A,, & Genzel, R. 1996, Nature, 383,415 Falcke, H., & Markoff, S. 2000, A&A, 362, 113 Falcke, H., Goss, W.M., Matsuo, H., Teuben, P., Zhao, J.H., & Zylka, R. 1998, ApJ, 499,73 1 Gammie, C. 1999, ApJ, 522, L57 Ghez, A.M. et al., 1998, ApJ, 509,678 Goldwurm, A. et al., 2003, ApJ, 584, 751 Hawley, J.F., & Balbus, S.A. 2002, ApJ,573, 738 Hornstein, S.D. et al., 2002, ApJ, 577, L9 Krolik, J. 1999, ApJ, 515, L73 Liu, S., & Melia, F. 2001, ApJ, 561, L77 Liu, S., & Melia, F. 2002a, ApJ, 566, L77 Liu, S., & Melia, F. 2002b, ApJ, 573, L23 Melia, F. 1992, ApJ, 387, L25 Melia, F., & Falcke, H. 2001, ARABA 39,309 Melia, F., Jokipii, R., & Narayanan, A. 1992, ApJ, 395, L87 Melia, F., Liu, S., & Coker, R. 2000, ApJ, 545, L117 Melia, F., Liu, S., & Coker, R. 2001, ApJ, 5.53, 146 Narayan, R., Yi, I., & Mahadevan, R. 1995, Nature, 374,623 Petrosian, V 1985, ApJ, 299,987 Ruszkowski, M., & Begelman, M.C. 2002, ApJ, 581,223 Schodel, R. et al., 2002, Nature, 419, 694 Yuan, F., Markoff, S., & Falcke, H. 2002, A&A, 383, 854 Zhao, J.H., & Goss, W.M. in: Sub-arcsecond radio astronomy (Cambridge: Cambridge Univ. Press, 1993), p.38 Zhao, J.H., Bower, G.C., & Goss, W.M. 200 I , ApJ, 547, L29 Zhao, J.H. et al., 2003 ApJ, 586, 29 Zylka, R., Mezger, P., & Lesch, H. 1992, A&A, 261, 1 19

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 483-489 (2003) / DO1 10.1002/asna.200385076

The frozen (inactive) disk in Sgr A*: freezing the accretion of the hot gas too?

’

Sergei Nayakshin*

’ Max Planck Institute for Astrophysics, Garching, Germany. Abstract. The black hole (BH) in our Galactic Center (GC) is extremely underluminous for the amount of hot gas available for the BH consumption. Theoretical understanding of this fact rests on a likely but not entirely certain assumption that the electrons in the accreting gas are much cooler than the protons. In this case the hot gas as a whole is too hot to accrete, and is too tenuous to radiate away its gravitational energy. Here we propose a drastically different picture of the accretion process in Sgr A* not based on the unchecked two-temperature assumption. Namely, we argue that there should exist a very cold inactive disk - a remnant of a past stronger accretion activity in Sgr A*. Such a disk would be a very efficient cooling surface for the hot flow. We show that under certain conditions the cooling due to thermal conduction cannot be balanced by the viscous heating in the hot flow. Along with the heat, the hot flow loses its viscosity and thus ability to accrete. It settles (condenses)onto the cold disk slightly inside of the circularization radius. If the latter is very large, then the liberated energy, and the luminosity emitted, is orders of magnitude less than naively expected. We build a simple analytical model for this flow and calculate the expected spectra that appear to be in a very reasonable agreement with observations. Strong additional support for the presence of the inactive disk comes from the recent observations of X-ray flares in Sgr A * , The properties of these flares are very similar to those produced by stars passing through a cold disk.

1 Introduction Currently, the pitiful luminosity of Sgr A* is most commonly explained in the framework of Non-Radiative Accretion Flows ( N W ) , a generalization of solutions discussed in greatest detail by Narayan & Yi (1994; NY94 hereafter). These solutions are valid when the rate at which the protons pass their gravitational energy to the electrons is not significantly higher than that due to Coulomb interactions only. This assumption is unfortunately prohibitively difficult to test (see references in Narayan 2002). However, provided it is valid, the electrons radiate only a tiny fraction of the total energy (“94, Narayan et al. 1995) in stark contrast to the standard disks (Shakura & Sunyaev 1973). The second important feature of NRAFs was pointed out by Blandford & Begelman (1999; BB99 hereafter) who showed that these flows should produce powerful thermally driven winds (see also Quataert 2003). In this paper we would like to point out a likely and essential element of the accretion picture in Sgr A* that has so far escaped (except for Falcke & Melia 1997; FM97) the attention it deserves. Sgr A* is believed to be closely related to the Low Luminosity AGN (LLAGN; e.g. Ho 1999). Most if not all of these sources seem to have cold neutral and inactive disks that often can be seen only through water maser emission (e.g. Miyoshi et al. 1995) arising in a range of radii where the gas temperature is 200 - 1000 K. The continuous SED spectra of these objects also support the existence of cold disks (Quataert et al. 1999). Because of the extremely low ionization level of these cold disks, the disk viscosity may be nearly zero (e.g. see Menou & Quataert 2001). As such these “frozen” disks arc not accretion disks (e.g. Siemiginowska et al. 1996). Falcke & Melia (1997) studied the evolution of such a disk on very long time scales in Sgr A’ and concluded that unless the stellar wind (which is the current source of gas for the hot flow) has a very * e-mail: sergarnpa-garching.rnpg.de,Phone: +049 089 30000 2258

Q 2003 WLEY-VCH Verlag GmbH & Co. KGaA. Weinheirn

S. Navakshin: Inactive Disk in Ser A*

484

large angular momentum, it would have violated the near infra-red limits. In 54 we argue that due to some more recent data (e.g. Genzel2000). the wind does have a substantial angular momentum and hence the problem needs to be re-considered. Consider now the implications of an inactive disk presence for the flow of the hot gas in Sgr A*. The cool disk is razor-thin but is much more massive than the hot flow (Nayakshin et al. 2003). For temperatures as high as lo7 K, thermal conduction is usually very important. For the conditions at hand, the time to cool off via thermal conduction, tcond, is much shorter than that by radiation. Moreover, this time can even be shorter than the flow viscous time. In this case the hot flow gets “frozen” by thermal conduction before it can accrete onto the BH. As the flow loses its thermal energy, it also loses its vertical pressure support, and therefore it has to settle down (condensate) onto the inactive disk. Becoming a part of the inactive disk, the gas loses its viscosity and its ability to accrete, and can stay in the same location essentially indefinitely (the cold disk viscous time is as long as lo6 - 10’ years at large radii). The accretion process in this picture would be “delayed” because the gas is currently piled up in the cold disk. The expected 0.1fic2(few R,/R,) where R, is the bolometric luminosity of this “frozen flow” is roughly Lhol / cgravitational ~ radius. If R,.R, 2 lo4, then circularization radius of the hot flow and R, = ~ G M B H is the low luminosity of Sgr A* could be understood naturally: the hot gas does not penetrate very deep into the BH potential well, thus gaining little energy as it settles onto the inactive disk. During the workshop we presented 2D hydrodynamical simulations of the hot flow above a cold inactive disk. Using an insight from our simulations, we have recently discovered a simple idealized analytical solution that provides a much easier understanding of the numerical results. We present this solution below. The spectra resulting from this condensing flow appear to agree quite well with the observational constraints if condensation radius is R, 2 3 x 104R,, the disk is highly inclined, i s . z 2 75”, and the 3 x lop6Ma/year. accretion rate in the hot flow is at its “nominal” value, A& In addition, during the meeting we realized that the observed X-ray flares may well be due to stars passing through the inactive disk. In Nayakshin & Sunyaev (2003; NS03 hereafter) and in Nayakshin et al. (2003) we calculated the expected rate of flares, duration of a typical flare, X-ray spectra, luminosities, plus flare radio and NIR luminosities. All of these quantities closely resemble the observational picture reported by Baganoff et al. (2001) & Baganoff et al. (2003). It appears to us that both quiescent and flare spectra of Sgr A* can be explained if we accept existence of an inactive frozen disk. N

N

2 A simple analytical model A disk with temperature T d 2 100 K and with outer radius R d few x 104R, could be very hard to detect in Sgr A* with any current telescopes (NS03) if the disk is also highly inclined. Since the disk is razor-thin, its viscous time scale is very large ( 2 lo5 years), and we can consider its structure as given on shorter time scales. We thus introduce the disk only through the boundary conditions for the hot flow. Two-component accretion flows are too complex to be studied analytically except for very simplified special cases (which can however be most insightful - see BB99 for an example). Therefore, we will restrict ourselves to vertically averaged equations for a Keplerian hot flow. We consider radii R < R, where R, is the circularization radius of the hot gas. For the low densities concerned, the radiative cooling of the hot gas is negligible, while the thermal conduction is at its “best” - at the maximum or the “saturated” value. Namely, the heat flux is given by Fsat = 54Pc,, where 4 5 1 is the saturation parameter (Cowie & McKee 1977), and P is the vertically averaged pressure, P = pc; ( p and c, are the gas density and the isothermal sound speed, respectively). Note that if 4 1 then the energy flow due to thermal conduction is effectively supersonic. This is because the flux is carried by the electrons whose thermal speed is much higher than that of the protons in a one-temperature plasma (for more on this see Cowie & McKee 1977). On the other hand, the flow of energy in the radial direction will be proportional to the radial velocity that is normally much smaller than the sound speed (see below). Thus the energy losses due to thermal N

485

Astron. Nachr./AN 324. No. S1 (2003)

conduction in the limit of large (p are much larger than the rate at which the energy can be gained by sinking in the BH potential well. That is why condensation should ensue in this situation. Our equations are best understood through comparison with those of the well known ADAF solution (NY94; see their eqs. 1-4). The stationary mass conservation equation takes into account exchange of mass in the vertical direction:

where ‘ORis the radial flow velocity ( U R > 0 for accretion), pv, is the mass flow density for condensation (71, < 0 ) or evaporation (v, > 0). The vertical scale height H is introduced through the hydrostatic balance as H = cs/O where 0 is the angular velocity. Since the latter is assumed to be Keplerian, the radial momentum equation (eq. 2 in NY94) is trivially satisfied. Since the cold disk is also Keplerian, there is no exchange of specific angular momentum between the two flows and the angular momentum conservation equation (3 in NY94) is unaltered. With at = O K we get

it?

127ra d [pc2R2H] , RRKd R

= 4rRHp.u~ = __-

where Q is the viscosity parameter (Shakura & Sunyaev 1973). The entropy equation should include (in addition to the usual terms) the thermal conduction flux, F,,, and the hydrodynamical flux of energy in the vertical direction. To derive this equation, we follow the formalism of Meyer & Meyer-Hofmeister ( 1 994; see their equation 8), with the following exceptions. We neglect winds here because we are interested in cooler, condensing solutions. Thus their sideways term (their eq. 5 ) is not included. In addition, the radial entropy flow term (the “advective cooling”; NY94) is designated Qadv. Following NY94 we set Qadv = fadv&+, where fad,, 5 1 is a parameter. Further, for subsonic flows << c:, and we obtain flu;

where Q+ = (9/2)apc%is the viscous heating rate. The factor of 512 in eq. (3) is y/(y - 1) for the = 513 gas that we consider here. = ~c;/2, we find that the condensation velocity is After some simple algebra and using G M B H H ’ / ~ R

where b is introduced for convenience. The equation (4) is crucial for the rest of the paper. If thermal conduction is vigorous, i.e. b > 0, then the hot flow is condensing onto the cold disk. In the opposite case of a small (p and a “large” a, b < 0 , viscous heating prevails. Thermal conduction then serves to evaporate the inactive disk. We should also note that the regime of large a ( w 0.3) was already studied by F. Meyer and collaborators in many papers. Their solutions are for higher accretion rates and therefore they are in the non-saturated regime, which is roughly speaking equivalent to the 4 << 1 case. As they found, the evaporation is a very strong function of Q (K a3). Our results (eq. 4 in particular) are thus in a complete agreement with that of Meyer & colleagues, and we essentially extend their work for the case 4 >> a. Inserting now 11, = -bc, into equation (1) and also substituting U R for its value found from equation (2), we arrive at a second order differential equation that contains two variables, p and c,:

bpRc, = 3 a -

a dR

1 8 {[pc:$]} ROK dR -

.

This equation cannot be solved in a general case. By introducing ZJ, # 0 we added an extra variable to the accretion flow equations, and the number of independent equations is now smaller than the number of

S. Nayakshin: Inactive Disk in Sgr A*

486

unknowns. (In particular, both v, and c, are to be found from the single energy equation 3). This situation is well known in analytical ADIOS wind solutions (BB99). In the latter case one has to introduce three free parameters that describe the mass, energy and angular momentum carried away by the wind. The most natural way to proceed here is to suggest that the temperature is a power-law function of radius. For example, for AJlm, T ( R )0: R-‘ in a broad range of radii. On the other hand, thermal conduction tends to smooth out temperature gradients (for example within supernova remnants), and hence in the other extreme T ( R )N const (we in fact observed this nearly constant T ( R )in our numerical simulations). Thus, cs = co(R,/R)’, where 0 5 X 5 1/2 and co is the sound speed at R,. If we define u = then equation (5) can be re-written as

a,

Finally, defining i j a2fi

-=-au2

= pu7-6x,we obtain

1 12

fi y4(1-’)

(7)

’

Here 1, is the dimensionless “condensation length”:

If there were no thermal conduction losses, the internal energy of the hot gas would be of order its gravitational energy. If the inactive disk extends to R > R,, the thermal conduction will reduce the gas thermal energy. Hence we expect that czR,/GMBH 5 1. We are most interested in the case 4 >> a and therefore we shall only explore the X = 0 case below.’ a/q5 << 1. This circumstance In addition, with a << $, condensation length is small, i.e. 1: < a/b facilitates finding an approximate analytical solution for equation (6): N

i j = const exp

[-&I

.

(9)

+

Indeed, d2fi/du2 = 6 (-2/Zcu3 1/1zu4) N ij/l:u4 due to the fact that l/l,u >> 1. Let us now explicitly write down the results from this approximate solution:

Here p, is gas density at R,, and A& is the accretion rate at that point. Note that equation (1 1) shows that the radial velocity is constant and is substantially smaller than the sound speed. Further, for 1, < 1/2 (recall that we assumed 1, << 1).the accretion rate increases with R, as it should for a condensing flow.

’ Nevertheless we note that scale-freesolution.

(I)

approximate solutions may be obtained for a general value of A, and (ii) for X = 1/2 there is a

487

Astron. Nachr./AN 324, No. S1 (2003)

3 Sample results To illustrate the analytical solution presented above, we plot the gas density and the mass accretion rate profiles (equations 10 & 12) in Figure 1 for 4 = 0.2 and two different values of a. These are chosen to represent the two opposite extremes. In the case of larger a, condensation length is I , = 0.38 and this is barely satisfies the condition 1, << 1 under which our approximate solution is valid. The accretion rate changes slowly with radius in this case, meaning that condensation is relatively slow. In the case of very small LY the hot flow collapses onto the disk far quicker: the mass flow is reduced by 100 times already at R ci R,/2. This is easily understood by noting that ? l , / P i ~ = N 4 for the larger a value, whereas for a = 0.004 this ratio is v,/UR N 16. The large value of the ratio U,/UR in our solutions justifies the statement that we made in the introduction: “condensation time”, tcond = H/w,, is much shorter than the viscous time, tvisc= R/uR for our solutions.

a

t’

N

i

1000

2

10000

Radius, R,

a)

1016 frequency

10’8

Fig. 1 (a) -Radial density (solid) and accretion rate (dashed) profiles for 4 = 0.2, R, = 3 x 104R,, and two values of viscosity parameter a = 0.05 (upper curves) and 0.004 (lower curves). (b) - Spectra corresponding to the radial profiles shown in (a). The dotted and solid curves are for a = 0.05 and a = 0.004, respectively. For both cases M B year-’. The time-averaged spectrum of Sgr A* inclination angle i = 75” and accretion rate h i 0 = 3 x is shown with triangles (radio data; detections), upper limits (infra-red) and the bow-tie (X-rays; Chandra data). The lower frequency data are from references given in Melia & Falcke (2001), except for the 2.2 pm point that is from Homstein et al. (2003), and for Chandra data that are from Baganoff et al. (2003). For the X-ray part of the data the spectrum should only be lower than the Chandra observations because the latter are probably dominated by the emission of hot gas at R RB > R,, which is not included in the model.

-

It is useful to define the “apparent” bolometric efficiency coefficient of the frozen accretion flows, &bol in the standard way:

-

where Lbol is the bolometric luminosity of the source. One should recall that for standard accretion flows, 0.1 for a non-rotating BH. The energy deposition into the cold disk is given by the second and the third terms in equation (3). Integrating this expression over the disk surface area from 3R, to Re,one can amveat&bol= 1 . 5 ( 1 + ~ ) [ 1 + 2 z , + 2 Z ~ ] ( R , / R e ) , w h e r e (Cl -~O . Y c u / $ ) ~ ’ . F o r $ > > a w e h a v e C = l and 1, << 1,so &bol = 3R,/Re. This expression of course only applies at R, >> 3R, because we neglected any relativistic corrections to the gravitational potential. In addition, the derived expression is only a rough R (which estimate since we assumed vertically averaged equations which are clearly inaccurate for H in our simple model occurs at R = R,). One can in fact show that realistically &hol should be smaller by

488

S . Nayakshin: Inactive Disk in Sgr A*

a factor of at least few. The point here is that our simplistic solution is not bound at R = Rc because its thermal energy exceeds gravitational energy at that point. So either a wind takes away the excess energy (as BB99 argued for ADAF solutions) or more likely this simply means that we over-estimated the gas temperature at Re. We should also explicitly insert the disk inclination angle into the definition of the apparent efficiency since the observed luminosity of the thin disk is approximately proportional to cos i. Variation of the predicted spectrum with R, is shown in Figure (3). Summarizing, the apparent bolometric efficiency of the freezing flow is

4 Discussion We have suggested here that there exists an inactive disk around Sgr A*, a remnant of past powerful accretion (and probably star forming) activity. The disk may be quite light compared with both the BH and M a of the hot gas present in the the star cluster (Nayakshin et al. 2003), yet it easily out-weighs the region interior to the Bondi radius. The disk then serves as a very efficient cooling surface for the hot flow. The flow essentially gets frozen (stopped), and its energy is radiated as thermal emission at frequencies much below the X-ray band. The X-ray emitting flow thus simply disappears from “the radar screen”. This in our opinion may be the explanation of the exceptional apparent X-ray radiative inefficiency of Sgr A*. Further, since the hot flow does not penetrate very deep into the BH potential well, its total bolometric few R,/R, times smaller than that expected if the gas made it all the way into the BH luminosity is and was radiatively efficient. Thus the flow also appears radiatively inefficient in the bolometric sense. As we explained in the paper, we believe that the accretion of the winds from the hot stars is simply delayed in time and it is by no means radiatively inefficient in the long run. Falcke & Melia (1997; FM97) have assumed the hot wind infall as given and studied viscous evolution of the “fossil” disk on long time scales, whereas we concentrate on much shorter time scales on which the 106ff-l years for Tdl& = 100 K and R = lO4RY). structure of the disk does not change (i.e t < t,,,, Our study is therefore complimentary to that of FM97. N

N

-

489

Astron. Nachr./AN 324, No. SI (2003)

Our results concerning the conditions under which the wind-disk (or the hot flow-disk) interactions will not violate the tight NIR limits are quite similar to that of FM97. In particular, FM97 note that “the BondiHoyle wind must be accreting with a very high specific angular momentum to prevent it from circularizing in the inner disk region where its impact would be most noticeable”. We find that the circularization radius should be 2 3 x 104Rg,implying a very large angular momentum indeed. Further, we suggested that the disk and the hot flow angular momenta are at least approximately aligned or else there would be a substantial heating due to friction between the two, a heating not included in our analysis. It remains to be seen whether results will be qualitatively similar if the disk and the hot flow rotation axes are misaligned. FM97 considered a ‘‘large’’ value of R, being rather unlikely. We however note that according to recent data, the stars from which the hot wind originates appear to be on tangential orbits counter-rotating the few arcsecond few x 105Rgoff Sgr A* (e.g. Genzel 2000).There is thus Galactic rotation and are no deficit of angular momentum at these distances. Finally, we only studied here the region of the flow interior to R,. However the exchange of the angular momentum between the hot flow and the disk should take place at R > R, where the hot gas is sub-Keplerian. This should enrich the hot flow with the angular momentum directed as that of the the disk and hence the circularization of the hot flow should occur even if it had zero angular momentum at infinity. Thus the requirement of a large angular momentum in the wind may be relaxed, although this effect remains to be quantified with future calculations.

-

5

N

Conclusion

In this paper we suggested that there exists a very cold inactive disk in Sgr A*, and that its role in the accretion picture is significant. While the hot gas is very tenuous and cannot radiate its energy away, it can easily transfer its energy into the cold disk via thermal conduction. The cold disk is much denser and much more massive than the hot flow and can serve as a very powerful freezer (or radiator) for the hot flow. As the hot flow looses its energy, it also looses its viscosity and “sticks” to the cold disk. The accretion flow is thus quenched by this seemingly “rzon-radiative”cooling. One can easily check that neither internal viscous dissipation nor the heat input from the hot flow in Sgr A* are sufficient to overcome the radiative cooling and restart the accretion in the inactive disk. It appears that only arrival of a new large supply of low angular momentum material could revive the inactive disk in Sgr A* now. We thank H. Falcke, C. McKee, F. Meyer, R. Narayan, R. Sunyaev and H. Spruit for discussions.

References Baganoff F. K., et al., 2001, Nature, 413, 45 Baganoff F. K., et al. 2003, these proceedings Blandford, R., & Begelman, M.C., 1999, MNRAS, 303, LI Cowie, L.L., & McKee, C.F. 1977, ApJ, 21 I , 135 Falcke, H, & Melia, F. 1997, ApJ, 479,740 Genzel, R. 2000, in the Proceedings of the Star2000 Meeting, editor R. Spurzem (astro-pW0008119) Hornstein, S.D. et al. 2003, these proceedings. Melia, F., & Falcke, H. 2001, ARA&A, 39, 309 Menou, K., & Quatdert, E. 2001, ApJ, 552,204 Meyer, F., & Meyer-Hofmeister, E. 1994, A&A, 288, 175 Miyoshi, M., et a]. 1995, Nature, 373, 127 Nardyan, R., & Yi, I. 1994, ApJ, 428, L13 Narayan, R., Yi, I., & Mahadevan, R. 1995, Nature, 374,623 Narayan R., 2002, p. 405, in “Lighthouses of the Universe”, Springer 2002, editors Gilfanov, M., Sunyaev, R., & Churazov E. Nayakshin, S., & Sunyaev, R. 2003, submitted to MNRAS (astro-ph/0302084) Nayakshin, S., et al. 2003, in preparation. Quataert, E., et al. 1999, ApJL, 525, L89 Quataert, E., 2003, these proceedings Schodel R., et al. 2002, Nature, 419, 694 Shakura, NJ., & Sunyaev, R.A. 1973, A&A, 24,337 Siemiginowska, A., Czerny, B., & Kostyunin, V. 1996, ApJ, 458,491

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1,491-495 (2003) / DO1 10.1002/asna.200385046

Gamma-ray emission from an ADAF around a Kerr black hole Kazutaka Oka*' and Tadahiro Manmoto**2

' Department of Earth and Planetary Sciences, Kobe University, Kobe 657-8501, Japan

* Department of Physics, Chiba University, Chiba 263-8522, Japan

Key words accretion, accretion discs - black hole physics - Galaxy: center - gamma rays: theory PACS 04A25 We investigate the gamma-ray spectrum emitted from an ADAF and its dependence on the spin parameter of a central Kerf black hole, in order to examine whether the spectrum can be used to probe the spin parameter of black holes. We consider that the gamma-rays are produced through the decay of neutral pions created by proton-proton collisions in the vicinity of the central black hole. Since the energy distribution of the ion particles in an ADAF is not known, we consider two types of proton energy distributions: a thermal distribution and a power-law distribution. In the thermal model, we find that changes in the spin parameter from-0.95 to 0.95 can enhance the gamma-ray intensity by orders of magnitude. Thus, if the proton gas in an ADAF has a thermal distribution, the gamma-ray spectrum can be used as a probe to investigate the spin parameter of the central black hole. In the nonthermal model, on the other hand, the gamma-ray intensity is much less sensitive to the changes in the spin parameter than in the thermal model, and it would he difficult to estimate the spin parameter from the gamma-ray spectrum. We apply our model to the Galactic Center, Sgr A'. The unidentified gamma-ray source 3EG J17462851 is observed towards Sgr A* by EGRET. Our results show that the gamma-ray intensities predicted from our models are much lower than observations and we cannot find the spin parameter. We, however, consider that this is not a serious problem against our model since it is unclear whether the observed gammarays are from a point or a diffuse source at the Galactic Center. In order to investigate the spin parameter via the gamma-rays from the Galactic Center instruments with higher angular resolution is needed such as GLAST.

1 Introduction The dimensionless spin parameter a is an important physical quantity representing the black hole spin. If we could determine the spin parameter from observations, it provides not only the confirmation of the accretion theories, but also some insight into the accretion history of supermassive black holes in active galactic nuclei (AGNs). Then, how can we investigate the spin parameter of a black hole? In the present study we consider a Kerr black hole surrounded by an advection-dominatedaccretion flow (ADAF), which is a geometrically thick, optically thin hot accretion flow and has low radiative efficiency (Ichimaru 1977; Narayan & Yi 1994,1995a,b; Abramowicz et al. 1995). Emissions in radio to X-ray bands are determined by the cooling processes of electrons such as synchrotron, bremsstrahlung, and Compton processes. In the gamma-ray band, on the other hand, Mahadevan, Narayan, & Krolik (1997) pointed out that the ion temperature in an ADAF close to a black hole is high enough (Tt lo1' K) to produce gammarays through the decay of neutral pions, which are created by proton-proton (p-p) collisions. Their method is very interesting, because once the spectrum produced by the electrons is fixed, which means that all the N

' Corresponding author: e-mad: kazutakaakobe-u.ac.jp, Phone: +81 78 803 5754, Fax: +8178 803 5747 * * Corresponding author: e-mail: manmotoQ astroschiba-u.ac.jp @ 2003 WILEY-VCH Verlag GmbH & Ca KGaA, Weinhem

K. Oka and T. Manmoto: Gamma-rays from a Ken black hole

492

parameters in an ADAF are determined, no additional parameters are required to calculate the gamma-ray spectrum. Thus one can calculate the gamma-ray spectrum uniquely. They, however, considered only the case of Schwarzschild black hole, and thus did not investigate the case of rotating black hole. Manmoto (ZOOO), on the other hand, studied the ADAF spectrum around a Kerr black hole by taking into account only the cooling processes of electrons. Manmoto (2000) showed that dependences of ion temperature and electron temperature on the spin parameter are different. Thus, if we combine the gamma-ray spectrum, which contains information on the ion temperature, and the spectrum in other bands, which contains information on the electron temperature, the spin parameter of a black hole could be determined. In the present study we examine whether the gamma-ray spectrum can be a probe to study the spin parameter. So far the ADAF model successfully explained spectra of several objects including the Galactic Center, Sgr A*. Among them, the Galactic Center is the only object whose gamma-ray intensity is above the detection threshold of EGRET (Mahadevan et al. 1997). In the present study we apply our model to the Galactic Center.

2 Description of Model 2.1 Gamma-ray Emission Mechanism In the ADAF model, the dissipated energy is stored in the accretion flow and advected inward. Since ions hardly radiate, they are heated almost up to the virial temperature. The ion temperature in the vicinity of a central black hole is so high (T, 10°K) that a p-p collision produces a neutral pion, T O , which then decays into two gamma-ray photons (Mahadevan et al. 1997). The emergent shape and luminosity of the gamma-ray spectrum from an ADAF dramatically depend on the proton energy distribution, which is in turn determined by the mechanism of viscous heating. However, the detailed mechanism of the heating is not well understood at present, and therefore it is not known whether the viscous heating leads to a thermal or a nonthermal distribution of proton energies (see Mahadevan & Quataert 1997). In order to calculate the gamma-ray spectrum we assume two different proton energy distributions: a thermal distribution (eg., Dermer 1986; Mahadevan et al. 1997) and a power-law distribution (Mahadevan et al. 1997; Mahadevan 1999). The details of the calculations of the gamma-ray spectrum are described in Oka & Manmoto (2003). N

2.2 Calculation of ADAF The structure of an ADAF is determined by the following parameters: the viscous parameter a, the ratio of the gas pressure to the total pressure &, the mass of the central black hole M , the mass accretion rate M , the fraction of the viscous heating that goes into electrons 6, and the spin parameter a(-1 < a < l), where a positive (negative) a means that the black hole corotates (counterrotates) with the accretion flow. In the calculation of the ADAF structure, we assume the ion particles to be thermalized, and therefore we obtain the ion temperature at each radius. For the nonthermal model we redistribute the energy of the ion particles with the total energy at each radius fixed. We assume that the electrons are always thermalized by action of self-absorbed synchrotron photons (Mahadevan & Quataert 1997). The details of the calculations of the ADAF structures are described in Manmoto (2000).

3 Results 3.1 Typical AGN-mass black hole We show the results obtained for the case of a typical AGN-mass (A4 = lO*M,) black hole. Three spin parameters of a = 0.95,0, and -0.95 are investigated. We assume that the spectra due to electron cooling processes have the same intensity at an X-ray point (1 keV). To do this, we adjust the mass accretion rate = 1 . 0 310-3A&, ~ h;r = 10-3&'c, and 5 . 8 10-*&fC, ~ for the models with a = -0.95,0, and 0.95 to be

Astron. Nachr./AN 324, No. S 1 (2003)

493

Fig. 1 Lef: Spectra in the thermal model for a = 0.95 (solid lines), 0 (dashed lines), and -0.95 (dotted lines). ~ For a = -0.95 (dotted Gamma-ray spectrum due to the pion-decay is located at frequencies higher than 0 - 1 0 ~Hz. lines), the actual gamma-ray part of the spectrum is 10 times lower than shown. Righr: Spectra in the power-law model with s = 2.75. Gamma-ray spectrum for a=0.95 (solid line) and a=O (dashed line) are very similar, and thus overlapped. In both figures the black dots show the X-ray point (1 keV).

respectively. Here Ak, is the Eddington mass accretion rate. Other parameters are set to be (Y = 0.1 and /?, = 0.5. We also assume that almost all the dissipated energy heats the ions by setting b = 0.001, although the determination of the value of 6 is still a controversial issue (see e.g., Bisnovatyi-Kogan & Lovelace 1997). Using the ADAF+Kerr model (Manmoto 2000), we obtain the structure of the flow, and then calculate the gamma-ray spectrum. Here the Doppler and gravitational shifts, or the bending of the photons path are not taken into account. The left panel of Figure 1 shows the spectra from the radio hand to the gamma-ray band for a = -0.95, 0, and 0.95 in the thermal model. The spectrum due to the pion-decay is located at frequencies loz1 Hz, while the spectrum due to the electron cooling processes such as synchrotron, higher than v bremsstrahlung, and Compton processes appears at frequencies lower than v N lo2' Hz. We can see that the gamma-ray intensity is enhanced by orders of magnitude when the spin parameter changes from -0.95 to 0.95. Based on the calculation, we can conclude that if proton gas in an ADAF has a thermal distribution, the gamma-ray spectrum can give a constraint on the spin parameter of the central black hole. The right panel of Figure 1 shows the spectra in the nonthermal model. We set the power-law index s of the power-law distribution to be 2.75. Other parameters are the same as in the case of the thermal model. We find that the gamma-ray intensity for each spin parameter has almost the same magnitude. Thus, it would be difficult to estimate the spin parameter from the gamma-ray spectrum. N

3.2

Application to Sgr A*

We apply our model to the Galactic Center, Sgr A*. The spectrum of Sgr A* has been explained naturally with the ADAF model (Narayan, Yi & Mahadevan 1995; Manmoto, Mineshige, & Kusunose 1997; Narayan et al. 1998; Manmoto 2000; Narayan 2002). Although there are several objects whose spectra are explained by the ADAF model, Sgr A' is the only object whose gamma-ray intensity is above the detection threshold of EGRET (Mahadevan et al. 1997). Our models are adjusted so that the predicted emission agrees with the X-ray flux in quiescence measured by Chandra (Baganoff et al. 2001). We adopt both the thermal and the power-law distributions of proton energies. The model prediction is compared with the gamma-ray source 3EG 51746-2851 observed towards Sgr A* by EGRET. The ADAFparameters are as follows. The mass of Sgr A* is M = 2.5x106Ma,

K. Oka and T. Manmoto: Gamma-rays from a Ken black hole

494

8

10

12

14

18

18

20

22

24

Fig. 2 Radio to gamma-ray spectrum for Sgr A*. Both thermal and power-law distributions of proton energies are adopted. In the gamma-ray band, the spin parameters are a = 0, -0.95, and 0.95 (from top to bottom) for the powerlaw model, and a = 0.95, 0, and -0.95 for the thermal model. For the thermal model with a = -0.95, the actual gamma-ray intensity is 100 times lower than shown. The gamma-ray data corresponds to 3EG 517462851 observed by EGRET (Hartman et al. 1999).

a is 0.1, pp is 0.5, and 6 is 0.1. Power-law index is 2.75. Spin parameters and mass accretion rates are (a, A?>= ( 0 . 9 5 , 1 . 7 4 ~ 1 0 ~ ~ A( k0 ~, 8) ., 6 ~ 1 0 - ~ A ? and ~ ) , (-0.95,l x10W5hj,). Figure 2 shows the radio to gamma-ray spectrum for Sgr A*. All the predicted gamma-ray intensities are lower than the EGRET data by more than two orders of magnitude, and therefore we cannot decide the spin parameter for Sgr A*. However, we do not consider that it is a serious problem against our model; our model suggests that most of the gamma-ray flux obtained by EGRET does not originate in Sgr A*. Since the angular resolution lo,it would have contamination from other sources such as radio arc and Sgr A" East of EGRET is (see e.g., Pohl 1997; Melia et al. 1998a,b; Markoff, Melia, & Sarcevic 1999). In order to remove such contamination and evaluate the gamma-ray intensity from the Galactic Center, instruments with higher angular resolution is needed such as the next generation gamma-ray telescope, Gamma-Ray Large Area Space Telescope (GLAST). N

4 Summary We investigated the dependence of the gamma-ray spectrum from an ADAF on the black hole rotation and examine whether the gamma-ray spectrum can be a probe to investigate the spin parameter. We also applied our model to Sgr A*. We can summarize our results as follows. If the proton gas has a thermal distribution of proton energies, changes in the spin parameter from -0.95 to 0.95 enhance the gamma-ray intensity by orders of magnitude. Therefore in the thermal case, we can estimate the spin parameter from the gamma-ray spectrum using the multi-wavelength observations. If the proton gas forms a power-law distribution of proton energies, the gamma-ray spectrum is not sensitive to changes in the spin parameter, and thus it is not easy to estimate the spin parameter from the gamma-ray spectrum. We applied our model to the Galactic Center. We found that the expected gamma-ray intensities are much lower than the observed value and thus could not find the value of spin parameter. Our model suggests that most of the gamma-ray flux observed by EGRET does not originate in Sgr A*. Acknowledgements K.O. would Like to thank the organizing committees of Galactic Center 2002 and Kobe University for their financial supports.

Astron. Nachr./AN 324. No. S1 12003)

References Abramowicz, M. A., Chen, X., Kato, S., Lasota, J.-P., Regev, 0. 1995, ApJ, 438, L37 Baganoff, F. K., et al. 2001, ApJ, submitted (astro-ph 0102151) Bisnovatyi-Kogan, G. S., Lovelace, R. V. E. 1997, ApJ, 486, L43 Dermer, C. D. 1986, ApJ, 307,47 Hartman, R. C., et al. 1999, ApJS, 123,79 Ichimaru, S. 1977, ApJ, 214, 840 Mahadevan, R. 1999, MNRAS, 304,501 Mahadevan, R., Narayan, R., Krolik, L. 1997, ApJ, 486, 268 Mahadevan, R., Quataert, E. 1997, ApJ, 490,605 Manmoto, T., 2000, ApJ, 534, 734 Manmoto, T., Mneshige, S., Kusunose, M. 1997, ApJ, 489, 791 Markoff, S., Melia, F., Sarcevic, I. 1999, ApJ., 522, 870 Melia, F., Fatuzzo, M., Yusef-Zadeh, F., Markoff, S. 1998a, 508, L65 Melia, F., Yusef-Zadeh, F., Fatuzzo, M. 1998b, ApJ, 508. 676 Narayan, R. 2002, in Lighthouses of the Universe, eds. Gilfanov, Sunyaev et al. Springer-Verlag, p 405 Narayan, R., Mahadevan, R., Grindlay, J. E., Popham, R. G., Gammie, C. 1998, ApJ, 492,554 Narayan, R., Yi, I. 1994, ApJ, 428, L13 Narayan, R., Yi, 1. 1995a. ApJ, 444,231 Narayan, R., Yi, I. 1995b, ApJ, 452, 710 Narayan, R., Yi, I., Mahadevan, R. 1995, Nature, 374,623 Oka, K., Manmoto, T. 2003, MNRAS, 340,543 Pohl, M. 1997, A&A, 317,441

495

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1.497-504 (2003) / DO1 10.1002/asna.200385047

The Discovery of Sgr A* W.M. Goss*I, Robert L. Brown'g2,and K.Y. Lo'

' National Radio Astronomy Observatory

' National Astronomy and Ionospheric Center Key words Ga1axy:center-galaxies:individual (Sagittarius)-techniques: interferometer The galactic center compact radio source Sgr A* was discovered on 13 and 15 February 1974 by Bruce Balick and Robert L. Brown using the Green Bank 35 km radio link interferometer (Balick & Brown 1974). We discuss other observationsof this source in the years 1965-1985. Early VLBI observations are described. The name Sgr A* was first used by Robert L. Brown (1982) and has become the accepted name for the compact source at the center of the Milky Way.

1 Introduction The discovery of Sagittarius A as a radio source coincident with the center of the galaxy has been discussed by Goss & McGee (1996). Almost 20 years after the recognition that the center of the galaxy could be associated with Sgr A by Piddington & Minnett (1951) and McGee & Bolton (1954) , Sgr A* was discovered in February 1974 by Bruce Balick and Robert L. Brown in Green Bank, West Virginia. This discovery is certainly one of the more important galactic radio astronomy discoveries of the 1970's and has had wide ramifications during the last 30 years. As an example, the recognition that the radio source Sgr A* is the dim radio source associated with a 2.6 x lo6 Ma black hole has represented a fundamental advance in our understanding of the nuclei of galaxies. The participants in the complex story of the discovery of Sgr A* are numerous: B. Clark, D. Hogg, G. Miley, B. Turner, C. Heiles, R. Ekers, D. Lynden-Bell, D. Downes, A. Martin, B. Balick, R. Brown, M. Goss, K.Y. Lo, U. Schwarz, D. Rogstad and others. (Most of these are still active astronomers.) We present a short history of the discovery process (section 2 ) and provide some details on the naming of Sgr A* in section 3. In section 4, we provide a short summary of the determinations of the secular parallax of Sgr A*. Goss (2003) has also presented a summary of the discovery of Sgr A*. The process of discovery of Sgr A* was the result of the application of a "matched filter" in angular resolution to the properties of Sgr A"; of course, the construction of this filter can only be understood with an a posteriori knowledge of the properties of Sgr A*.

2 The years 1965-1985 In 1966, Clark & Hogg ( 1 966) used the newly completed 2 element Green Bank interferometer at 1 1 cm to investigate the small scale structure of a number of radio sources at I1 cm with a resolution of 10". The source Sgr A was found to have a compact feature with a flux density of 0.3 f.u. (Jy); with this resolution the confusion from Sgr A West is a dominant effect. But these observers were "close" to the discovery of Sgr A". We now know that a resolution of 3" is required at this frequency.

-

-

N

* Corresponding author: e-mail: mgossQnrao.edu, Phone: +01505 835 7267, Fax: +01 505 835 7027

@ 2003 WILEY-VCH Verlag GmbH & Co KGaA. Wcinheim

W. M. Goss et al.: Discovery of Sgr A*

498

The key observation that led to the discovery of the compact source at the center of the galaxy was an observation by Miley, Turner, Balick, & Heiles (1970). With a baseline of 35 km at 11 cm, these authors discovered a compact source in the H 11 region W5 1 with a Tb > lo5 K. The 42 foot telescope , located at the time of the Miley et al. observations at Huntersville, West Virginia, is shown in Fig. 1. Note the limited tracking capability of this antenna. Although this brightness component in W51 has not been confirmed, the result did set off a string of circumstances that led to the discovery of Sgr A* only four years later. The theoretical framework for the search for evidence of the presence of a compact object at the galactic center was provided by Lynden-Bell (1969) and Lynden-Bell & Rees (1971) , who made the analogy between quasars and the high energy phenomena at the center of the Milky Way: the latter authors propose four tests for the possible detection of a massive object at the center of the galaxy. The second test is :“Very ”. If so Long baseline interferometry may soon be possible with ....as weak as 0.5 f.u. to diameters of it may be possible to determine the size of any central black hole that there may be in our galaxy. However H II may render the central source opaque with a greater angular size.” In the course of 1970 (January and May), Ekers & Lynden-Bell(197 1)used the newly constructed 40 m antenna at the Owens Valley Radio Observatory (Caltech) along with one of the original 90 foot antennas to look for the signature of a black hole in the galactic center. At 6 cm the resolution was 6” by 18”. Ekers & Lynden-Bell detected fine scale structure in the Sgr A West H 11 region. “Although stimulated by the black hole idea our observations are thus more simply explained in terms of young stars and giant H 11 regions.” Again if the resolution had been a factor of about two more favorable , Sgr A* would have been detected. (Goss has pointed out in several lectures in Australia that Ron Ekers has mispelled his name in the acknowledgements to the Ekers & Lynden-Bell paper!) Ekers & Lynden-Bell also performed one of the first interferometer searches for radio recombination lines; they used the 90 foot antenna interferometer at 6 cm. They searched for broad recombination lines from Sgr A (Ekers, private communication); the negative result was not reported in the Ekers & Lynden-Bell paper. This test had also been suggested by Lynden-Bell & Rees (see above) to search for the existence of a massive object at the galactic center. The next step in the quest for compact sources at the galactic center was the result of investigations by Downes & Martin (1971) using the Cambridge One Mile Telescope at 11 and 6 cm with resolutions in RA of 11” and 6”. They describe the overall one dimensional structure (SgrA West and East) with a determination of the spectral indices of the various components. They mention the presence of structures < 10‘‘ in size with flux densities < 1 Jy. Again the discovery of Sgr A* was just over the “resolution horizon”. The discovery of Sgr A* did occur on 13 and 15 February 1974 by Bruce Balick and one of us ( R.L. Brown) using the Green Bank interferometer with an 45 foot antenna at the Huntersville West Virginia site at a distance of about 35 km. This site was the same as the one used in the earlier Miley et al. (1970) observations of W5 1but an improved antenna was now used . This antenna at the Huntersville site is shown in Fig.2; the antenna had a wider range of sky coverage and was operated at the dual frequencies of 11 and 3.7 cm. This interferometer was constructed to serve as a prototype of the Very Large Array which was under construction at the time. The publication of “Intense Sub-arcsecond Structure in the Galactic Center” was published in December, 1974 (Balick &Brown 1974). The resolution at 11 and 3.7 cm was 0.7” and 0.3”, respectively. With this resolution and uv coverage (the three simultaneous baselines from the Green Bank 3 element interferomter and the single antenna at Huntersvile), the extended confusion from the Sgr A West (flux density of 25Jy and size 4U‘)complex was resolved out. Balick & Brown write: “The unusual structure of the sub-arcsecond structure and its positional coincidence with the inner 1-pc core of the galactic nucleus strongly suggests that this structure is physically associated with the galactic center (in fact, defines the galactic center).” These authors compare the compact source with energetic nuclei of other galaxies and even suggest that variations in the radio flux density might be observed. Bob Brown has unearthed a number of fascinating letters from Bruce Balick written during the analysis period from mid March 1974 to 2 May 1974. No copies of Bob’s letters to Bruce were saved. There are no dates on Bruce’s letters. In these letters, Bruce gives in some detail his analysis of the data and their possible interpretations. A number of possible models for the observed visibilities are proposed. Reading N

N

N

N

N

-

Astron. Nachr./AN 324, No. Sl (2003)

499

these letters today, we are impressed with the meticulous attention to detail in the interpretation of this difficult observation. Here are a few amusing quotes from the letters with an emphasis on a fear (in retrospect somewhat unfounded) of competition. The time frame is toward the end of the period March-May 1974. Bruce writes: “Here are a few thoughts on the 45 foot Sgr A observations. Fred Lo re-analysed some of his VLB observations of Sgr A based on a new position I gave him and found 0.3 f.u. at X 6 cm [see later for a description of this October 1973 observation. The detection was only 2-3 (TI. I think his baseline was Green Bank-Haystack [in fact, Maryland Point]. We’d better publish fast if we want to heat him into print. I haven’t heard from Goss or Downes. Could you call Dave Hogg [then Green Bank site director] and ask if he’s heard anything?”. The following letter expresses some aprehension about IR competetion: “Dave Rank and I [at the University of California at Santa Cruz] are going to try to detect these sources in the IR. Please keep the positions kind of quiet, cause Becklin and co. can wipe us out if they want to. So can Rieke. Your faithful collaborator ...” In 1975, Ekers, Goss, Schwarz, Downes, & Rogstad (1975) combined Westerbork (WSRT) data with Owens Valley Radio Observatory data at 6 cm and made a 2-dimensional image of Sgr A with a resolution of 6“ x 18”. The image was called the WORST image - the Westerbork Owens Valley Radio Synthesis Telescope. In fact the beam shape looked like a sausage - “worst” in Dutch. Sgr A” was just visible at the longest spacings of the interferometer and Sgr A East, the non-thermal source that may be a luminous supernova remnant, was clearly detected as well as hints of the mini-spiral structure of Sgr A West ,an H 11 region associated with the center of the Milky Way. The first VLBI attempt to detect a compact source at the center of the Milky Way was carried out by K.Y. Lo and collaborators in October, 1973. The observation is described by Lo (1974) in his MIT Phd thesis of August 1974: “Interstellar Microwave Radiation and Early Stellar Evolution”. Lo was following up on the Miley et al. W5 1 observation of 1970, to try to confirm the detection of compact structure in HI1 regions, beyond what had been expected theoretically. The observations were at 6 cm between the Green Bank 140 foot (see above) and the Naval Research Labortory 85 foot Maryland Point radio telescope. The GB-MP baseline is mainly E-W with a length of 228 km. A source with a diameter less than 26 mas (EW) 0.1 Jy. As we now know, the size of would be unresolved and detectable down to a flux density of Sgr A* at 6 cm is broadened by interstellar scattering to an EW by NS size of 51 x 27 mas (Lo et al. 1998; Davies, Walsh & Booth 1974). The orientation of the baseline was therefore quite unfavorable for a detection. So, while there were hints of a signal in the visibility amplitude, the detection was not definitive. If the baseline had been oriented in a roughly N-S direction, the source would have been detected. For one of us (Goss), an amusing and somewhat embarrassing episode occurred in the years 1972- 1974. On 2 June 1972, D.Downes and Goss (both working at the Max Planck Institute for Radio Astronomy in Bonn, Germany) submitted proposal D43 to NRAO for an observation of the galactic center with the Green Bank 35 km radio link interferometer. The propsosal was sent to D.Heeschen, D.Hogg and W.E.Howard. We have been able to reconstruct all these events based on the extensive paper archive preserved (in 2003) by Dennis Downes from these pre-email days. The proposal included two positions in Sgr A and three in Sgr B2. A few key sentences from the proposal follow: “In view of the increasing interest in highly collapsed nuclear objects as probable sources of the energy in QSO’s and radio galaxies, it is of paramount importance to pursue investigations of compact structure in Sgr A. Although the center of the galaxy is relatively quiescent, it is so close that we can observe details on a much finer linear scale than is possible in external galaxies, even with VLB techniques. .... We regard this project as an experiment which may he a useful guide to future observations by the VLA These observations might be used in future programs to investigate short-term variability in the galactic center.” Although these ideas were relevant, Goss and Downes were not able to come to the US in 1973-1974 due to problems obtaining travel funds. Also initial observations with the 45 foot telescope were somewhat delayed from late 1972 to mid 1973. Early in 1973, Goss had moved to the Netherlands in a visiting position at the University of Groningen to work on WSRT projects. In addition, Downes was quite busy with early observations with the 100 m Effelsberg N

W. M. Goss et al.: Discovery of Sgr A*

500

telescope. With these pressures, the urgency to complete the Downes-Goss proposal with the Green Bank interferometer decreased. D. Hogg had been in continual contact with Downes and Goss about scheduling. As shown above, Balick and Brown had an earlier NRAO proposal to observe small scale structure (W51 type components) in H 11 regions and Sgr A and Sgr B2 were included. Dave Hogg became aware of the proposal conflict in early 1974 and wrote Downes a letter on 15 February 1974 (note the precise discovery date) proposmg several ways to resolve this conflict. However, Goss and Downes seemed to have lost interest at this point. Of course, the significant result is that Balick and Brown did discover Sgr A* in early 1974 - in fact on 13 and 15 February. The first successful VLBI detection of Sgr A* was made the following year (19 May 1975) by Lo et al. (1975) using the OVRO 40 m and the NASA Goldstone 64-m Mars antenna at 3.7 cm. K.Y. Lo had become a postdoctoral fellow at OVRO after his MIT thesis, but he was interested in following up on the tantalizing hints of detection of Sgr A* on the GB-MP experiment at 6 cm in 1973. It is interesting to recall that after some persuasion by Lo, the observation of Sgr A* was added to the program of his colleagues R. Schillizzi and M. Cohen to study compact symmetric double radio sources. From the California baseline, the inferred size was 20 mas. At an URSI meeting in Boulder, probably in January 1976, after Lo had reported the detection of Sgr Ah at 3.7cm on the OVRO-Mars baseline, Don Backer asked the interesting probing question of how one can be sure that Sgr A* was not a background compact radio source. Interestingly enough, as indicated below, Don Backer answered his own question some years later when he and Dick Sramek detected the secular parallax of Sgr A* due to the rotation of the Sun about the Galactic Center. In the period 10 June 1974 to 10 September 1975, Sgr A * was observed with the early MERLIN array at 0.408,0.96 and 1.66 GHz with baselines of 24 and 127 km. The detections at the latter two frequencies suggested the angular size scales as A’, originating in a turbulent electron distribution along the line of sight (Davies, Walsh & Booth 1974). A number of groups worked on the subsequent VLBI observations of Sgr A* (Kellermann et al. 1977, Lo et al. 1977, Lo et al. 1981, Lo et al. 1985, & Lo et al. 1993). In the 1985 publication of Lo et al., this group determined for the first time that the scattering size of Sgr A* at 3.6 cm was asymmetrical with an axial ratio of 0.55, and at 1.35 cm the limit to the angular size was 20 AU or 2.5 mas. The Green Bank 35 km interferometer was used to determine that the radio spectrum of Sgr A” (Brown, Lo & Johnston 1978) is inverted. Brown & Lo (1982) carried out a ground breaking investigation of the variability of Sgr A* at 11 and 3.7 cm over a time interval of 3 years with 25 epochs (Brown & Lo 1982); time variations were detected over all time scales from days to years. This ground breaking project became the basis for future detailed VLA studies of the various scales in the time variations of Sgr A* (Zhao et al. 2001).

-

N

3 The naming of Sgr A* As far as we can ascertain, the only credit attributed to the naming of Sgr A’ by Brown (1982) is in the Annual Reviews article by Melia & Falcke (2001). The first attempt at a convenient name of the galactic center compact source is by Reynolds & McKee (1980) in a paper entitled : “The Compact Radio Source at the Galactic Center”. This publication presents a model of relativistic outflows, with either a spherical or jet geometry. The fact that the luminosity is 100 times greater than a pulsar but much less than other galactic nuclei was a puzzle. Reynolds & McKee suggest the name GCCRS- the galactic center compact radio source. This name has not survived. Brown & Lo (1982) discuss the variability of Sgr A* (see above) : ”Throughout this paper we use the name Sgr A to refer only and specially to the compact radio source. When necessary, we distinguish this from the more extended radio structure at the galactic center.” In 1982, Backer & Sramek (1982) presented the initial results of the secular motions of Sgr A* using the Green Bank radio link interferometer. The motions were found to be consistent with an object at rest in

-

Astron. Nachr./AN 324, No. S 1 (2003)

50 1

the center of the Milky Way. Backer & Sramek propose the name: SgrA(cn) from “compact non-thermal” object in the galactic center. Again this name has had no staying power. Eight years after the discovery, one of us (Brown) invented the name Sgr A* to distinguish the compact source from the other components in the galactic center and to emphasize the unique nature of this source. Brown (1982) proposed a model of Sgr A* consisting of twin precessingjets with a period of 2300 years. The model has not stood the test of time but the name immediately was accepted. As an example, the VLBI results discussed by Lo et al. (1985) uses the name Sgr A*; the review article by Lo (1987) also uses this nomenclature. Bob Brown provides the following rationale for the name: “Scratching on a yellow pad one morning I tried a lot of possible names. When I began thinking of the radio source as the ‘exciting source’ for the cluster of H IT regions seen in the VLA maps, the name Sgr A* occurred to me by analogy brought to mind by my Phd dissertation, which is in atomic physics and where the nomenclature for excited state atoms is He*, or Fe”, etc.”

4

The motions of Sgr A*

The physical association of Sgr A* with the mass centroid of the nuclear region of the Milky Way remained circumstantial until the observation of the secular motion of Sgr A’ by Backer & Sramek, noted above, was made in 1982 with the Green Bank radio link interferomter. Even in 1983, Martin Rees wrote to Robert L. Brown to report that at the IAU symposium held in June 1983 in Groningen (Netherlands, IAU Symposium No 106) that Jan Oort was worried that the lack of formaldehyde absorption toward Sgr A* could have implied that the radio source is located nearer than the true galactic center. Such concerns were again soundly put to rest based on two companion papers, which were published in the Astrophysical Journal issue of 20 October 1999 (Backer & Sramek 1999; Reid et al. 1999), that summarize the VLA and VLBA determinations of the motions of Sgr A*. Don Backer has pointed out to us that the initial Green Bank 35 km radio link interferometer observation of the 1970’s was inspired by a lunch time conversation with Rick Fisher in about 1975 (see the acknowledgement in Backer & Sramek 1999). The secular parallax due to the motion of the Sun around the center of the galaxy, of course, establishes that Sgr A* is in the galactic center, but more importantly can be used to set a lower limit on the mass of the black hole of a few thousand Ma. In addition, a number of constraints on galactic rotation constants can be determined. The long range goal of the VLBA program (Reid et a1 1999) is the determination of a parallax distance to the galactic center.

5

Summary

The observations of the last 30 years have provided a wealth of information about the source Sgr A* and the environs of the black hole at the center of the Milky Way. Many puzzles remain. We can only imagine the contents of a possible conference on the center of the galaxy that might be held at the time of the 60th celebration of the discovery of Sgr A* in 2034. In March 2004, The National Radio Astronomy Observatory will host a conference in Green Bank, West Virginia, to honor the discovery of Sgr A* exactly 30 years previously and to discuss recent results on this fascinating radio and X-ray source. Acknowledgements The National Radio Astronomy Observatory is a facility of the National Science Foundation operated by Associted Universities, Inc. under cooperative agreement. We thank Dennis Downes, Bruce Balick, Dave Hogg, Don Backer, Ron Ekers and Barry Clark for helpful comments. Much of the work on Sgr A* by Lo was done before his joining the NRAO; he appreciates the important support provided at the critical times in the early years by B. Burke, K. Johnston, J. Moran, A. Rogers, M. Cohen, R. Schillizzi, D. Backer and J. Welch. We thank Pat Smiley for assistance with the Green Bank interferometer photo archive. M. Goss thanks Tony Beasley for his hospilality at the Owens Valley Radio Observatory during February, 2003, while this paper was being written. The late J.H. Oort

SO2

W. M. Goss et al.: Discoverv of Sm A'

engendered the enthusiasm that initiated the fascination of Goss, Ekers and Schwarz in the 1970's for galactic center research.

References Backer,D.C., & Sramek,R.A. 1982, ApJ, 260,512 Backer,D.C., & Sramek,R.A. 1999, ApJ, 524,805 Basart,J.P., Clark,B.G., &Kramer,J.S. 1968, PASP,80,273 Basart,J.P., Miley,G.K., & C1ark.B.G. 1970, Proc.EEE Trans. Antennas & Propogation, AP-18,375 Balick,B. & Brown,R.L. 1974, ApJ, 194,265 Brown,R.L. 1982, ApJ, 262, 110 Brown,R.L., Lo,K.Y., & Johnston, K.J 1978,AJ, 83, 1594 Brown,R.L., & L0,K.Y. 1982, ApJ, 253, 108 C1arkB.G. & Hogg,D.E. 1966, ApJ, 145,21 Downes,D., & Martin,A.H.M. 1971, Nature, 233,112 Davies,R.D., Walsh,D., &Booth,R.S. 1976, MNRAS, 177,319 Ekers,R.D., Lynden-Bel1,D. 1971, Astrophys. Let., 9, 189 Ekers, R.D., Goss,W.M., Schwarz,U.J., Downes,D., & Rogstad,D.H. 1975, A&A, 43, 159 Fomalont,E.B.2000, in National Radio Astronomy Observatory Workshop Number 27, Radio Interferometry: The Saga and the Science, Proceedings of a Symposium to honor Bany Clark at 60, ed. D. Finley & W.M. Goss (N-kA0),41 Goss,W.M. 2003, in ASP Conf.Ser Vol, Radio Astronomy at the Fringe, ed. J.A. Zensus, M.H.Cohen, & E.Ros (San Francisco: ASP) Goss,W.M., & McGee,R.X. 1996, in ASP Conf. Ser. Vol. 102,The Galactic Center, 4th ESOlCTIO Workshop, ed.R. Gredel (SanFrancisco; ASP), 369 Kellermann,K.I., Shaffer,D.B., Clark,B.G., & Geldzahler,B.J. 1977, ApJ, 214, L61 Lynden-Bel1,D. 1969, Nature, 223,690 Lynden-Bell,D., & Rees,M.J. 1971, MNRAS, 152,461 Lo.K.Y., 1974, Phd thesis, MIT L0,K.Y. 1987, in AIP Conf. Proc No. 155, The Galactic Center, Proc. of the symposiumhonoring C.H. Townes,ed. D.C. Backer (New York AIP), 30 Lo,K.Y., Schilizzi,R.T., Cohen,M.H., & Ross,H.N. 1975, ApJ, 202, L63 Lo.K.Y., Cohen,M.H., Schilizzi,R.T.,& Ross,H.N. 1977, ApJ, 218, 668 Lo,K.Y., Cohen,M.H., Readhead, A.S.C., & Backer, D.C. 1981,249,504 Lo,K.Y., Backer,D.C., Ekers,R.D., Kellermann,K.I., Reid,M.,& Moran, J.M. 1985, Nature, 315, 124 Lo,K.Y., Backer,D.C., Kellermann,K.I., Reid,M., Zhao,J.-H., Goss, W.M., &Moran,J.M. 1993, Nature, 362, 38 Lo.K.Y., Shen,Z.-Q., Zhao,J.-H., & H0,P.T.P. 1998, ApJ, 508, L61 Melia,F. & Falcke,H. 2001, ARA&A, 39,309 McGee,R.X., Bo1ton.J.G. 1954, Nature, 173,985 Miley,G.K., Tumer,B.E., Balick,B., & Heiles,C. 1970,ApJ. 160, L119 Piddington,J.H. & Minnett,H.C. 1951, Aus J Sc. Res., A4,495 Reynolds,S.P. & McKee,C.F. 1980, ApJ, 239,893 Reid,M.J.,Readhead,A.C.S., Vermeulen,R.C., &Treuhaft,R.N. 1999, ApJ, 524,816 Zhao,J.-H., Bower, G.C. & Goss, W.M. 2001, A@, 547, L29

Astron. Nachr./AN 324, No. S1 (2003)

503

Fig. 1 The 42 Foot antenna. The initial location of this antenna was at Spencer's Ridge - 11 km NE of the Green Bank interferometer. The operation was at 1 1 cm. Basart et al. 1968 and Basart et al. 1970 describe initial observations with a two element interferometer consisting of this antenna combined with one of the 85 Foot antennas at Green Bank. The sky coverage was limited to declinations from 0 'to +66 'and hour angles within 2h40"'n of the meridian. For the observations of W51 described by Miley et al. 1970, the antenna had moved to the Huntersville site - 35 km to the SW of Green Bank. These observations were a partial inspiration for the Sgr A* observations of 1974. Later the 42 Foot antenna was replaced by the fully steerable antenna shown in Fig.2 at the Huntersville site. Fomalont (2000) summarizes the development of the Green Bank radio link interferometer in the years 1966 to 1978.

SO4

W. M. Goss et al.: Discoverv of S m A*

Fig. 2 The 45 Foot telescope used at Huntersville, West Virginia as the 3.5 km outstation for the NRAO Green Bank radio link interferometer. The discovery of Sgr A* was made in February, I974 by Bruce Balick and Robert L. Brown using this instrument (Balick & Brown 1974). The dual frequency instrument operated at 11 and 3.7 cm and was used as a prototype for the VLA in the planning stages during the early 1970’s (see Fomalont 2000). The 45 Foot telescope was the fourth element of the interferometer ; correlations were performed with all three of the 8.5 Foot antennas at Green Bank. This smaller antenna is now at the Green Bank site and has had an illustrious career as a component of the tracking stations for the HALCA VLBI spacecraft. The two antennas for telemetry are pointing to the Green Bank site to the northeast; the local oscillator signal was transmitted by a two way link at 1.3GHz while the 18GNz link was used for telemetry and IF transmission. During the summer there was no clear line of sight to the main interferometer site due to leaves in intervening trees. A passive reflector was used on a hill behind the main site to overcome this problem. Much of the development work for the radio link was done by N.G.V. Sarma who was on a sabbatical from the Tata Institute for Fundamental Research (Ooty) in In&a

Astron. Nachr./AN 324, No. S1.505 -51 1 (2003)/ DO1 10.1002/asna.200385084

The Position, Motion, and Mass of Sgr A* Mark J. Reid * I , Karl M. Menten’, Reinhard Genze13, Thomas Ott3, Rainer Schode13, and Andreas Brunthaler’ ’ Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138, U.S.A. Max-Planck-Insititut fur Radioastronomie,Auf dem Hugel 69, D-53121 Bonn, Germany Max-Planck-Insititut fur extraterrestriche Physik, Giessenbachstrasse,D85748 Garching, Germany

Key words Sgr A*, black holes, proper motions, SiO masers

Abstract. We report progress on measuring the position of Sgr A* on infrared images, placing limits on the motion of the central star cluster relative to Sgr A*, and measuring the proper motion of Sgr A* itself. The position of Sgr A* has been determined to within 10 mas on infrared images. To this accuracy, the gravitational source (sensed by stellar orbits) and the radiative source (Sgr A*) are coincident. Proper motions of four stars measured both in the infrared and radio indicate that the central star cluster moves with Sgr A* to within 70 km sC1 . Finally, combining stellar orbital information with an upper limit of 8 km s-’ for the intrinsic proper motion of Sgr A* (perpendicular to the Galactic plane), we place a lower limit on the mass of Sgr A* of 4 x lo5 Mo.

1 Introduction The precise position and proper motion of Sgr A* are of fundamental importance in order to understand the nature of the super-massive black hole (SMBH) candidate and its environment. Unfortunately, Sgr A* lies behind about 30 mag of visual extinction, and currently it can only be detected in the radio, infrared, and x-ray bands. While its radio emission is easily detected, the same cannot be said for its infrared and x-ray emission. In both of these wavebands, emissions from nearby (in angle) stars make it difficult to 10 milli-arcseconds isolate and measure the emission from Sgr A*. Only with positions accurate to (mas) can one confidently separate Sgr A* from confusing stellar sources and determine its spectral energy distribution and time variations. Stellar proper motions, accelerations, and even orbits are now being determined to high accuracy at infrared wavelengths, and the position of the central gravitational source (presumably Sgr A*) can be measured to mas accuracy. If Sgr A* is indeed a SMBH, then the gravitational source, inferred from 10 stellar orbits, and the radiative source, directly seen in the radio band, should coincide to within Schwarzschild radii (10R,,h E 0.08 mas M 1013 cm for Sgr A*). Thus measuring the position of Sgr A* in the infrared to sub-mas levels is of fundamental importance in testing the SMBH paradigm. The apparent proper motion of Sgr A* directly determines the sum of the angular rotation speed of the and any peculiar motion of Sgr A* ( V s g T ~with * ) respect Sun about the Galactic center, ( 0 0 VO)/Ro, to the dynamical center of the Galaxy. Thus, Sgr A*’s proper motion can provide a direct measurement of Galactic rotation. In addition, the combination of stellar motions and an upper limit on the motion of Sgr A* itself can yield a strong lower limit to the mass of the SMBH candidate. This paper reports recent progress on locating Sgr A* on infrared images, measuring its proper motion, and placing a lower limit on its mass. N

N

+

* Corresponding author: e-mail. [email protected],Phone: +1617495 7470, Fax: + I 617495 7345 0 ZW3WILEY-VCH Verlag GmbH & Co

KGaA, Weinhem

M. J. Reid et al.: Position. Motion & Mass of Sm A*

506

Fig. 1 Location of Sgr A* on a July 1995 2pm wavelength image of the inner 2 arcsec of the Galactic center, adapted from Menten et al. (1997) by Reid et al. (2003). The circle centered at the position of Sgr A* has a radius of 15 mas, corresponding to a la position uncertainty.

2 Previous Results Menten et al. (1997) detected SiO maser emission from red giant and supergiant stars within 12 arcsec of Sgr A*. Since the maser emissions originate from within about five stellar radii of the host star, they can serve to precisely locate the star at radio wavelengths relative to the strong radio source Sgr A*. Also, the SiO maser stars are very bright at infrared wavelengths, and one can use the radio positions of two or more stars to calibrate the infrared plate scale and rotation and then align the infrared image with the radio image containing Sgr A*. Following this method, Menten et al. located Sgr A* on a 2 pm wavelength image to an accuracy of 30 mas (lo).No source of emission was seen at the position (see Fig. 1) of Sgr A*, and an upper limit of 9 mJy (de-reddened) was established. Reid et al. (1999) and Backer & Sramek (1999) published observations of the apparent proper motion of Sgr A*. Both papers show that Sgr A* appears to move toward the south-west along the Galactic plane at about 6 mas yr-l. This is consistent with the angular rotation rate of the Sun in its 220 Myr period about

Astron. Nachr./AN 324, No. S 1 (2003)

SO7

the Galaxy. Removing the effects of the Sun's orbit yields an upper limit to the peculiar motion of Sgr A* of about 20 km s-l . Reid et al. interpreted this upper limit to indicate that the mass of Sgr A* exceeds lo3 Ma, ruling out any stellar source.

-

3 Recent Advances 3.1 The Infrared Reference System Recently, a wide-format infrared camera (CONICA; Lenzen et al. 1998) with an adaptive optics assisted imager (NAOS; Rousset et al. 2000) was installed on one of the ESO 8.2-m VLT telescopes. This has produced excellent data for diffraction-limited imaging. Very deep images of the Galactic center with a field of view of 28 arcsec were taken with these instruments at 2 pm wavelength early in 2002. These images proved to be of excellent quality. Over the past seven years, radio frequency observations of SiO maser sources in the Galactic center have been conducted with the NRAO VLBA and VLA. Both telescopes have been used to measure positions and proper motions of SiO maser stars with accuracies of about 1 mas and 1 mas yr-', respectively, relative to Sgr A*. Seven maser stars within 15 arcsec of Sgr A* have now been measured to these accuracies (Reid et al. 2003). By combining the radio positions of seven maser stars with their apparent positions on the new VLT images, we could determine the infrared plate scales and rotations with high accuracy. After aligning the radio and corrected infrared images, we found the residual differences in the maser star positions to be about 6 mas. This verified that the new infrared images have very small distortions across the field of view and allowed us to determine the position of Sgr A* with a 10 mas (la)uncertainty (Reid et al. ). Figure 2 shows a 2 p m wavelength image taken on 2 May 2002, with the position of Sgr A* and two nearby stars indicated. At this time, the fast moving star S2 was near pericenter in its orbit about Sgr A* (Schodel et al. 2002). The proximity of S2 to Sgr A* at this time (16 mas) precludes any significant measurement of the flux density of Sgr A*. The location for Sgr A* is within 10 mas of the gravitational source inferred from orbital solutions for the star S2 (Schodel et al. ; Ghez et al. 2003). Thus, the radiative source (Sgr A*) and the gravitational source of the SMBH candidate are co-located to within about 1000 Rsrh. Previous infrared proper motions have been relutive motions, with the motion reference defined by setting the average of large numbers of stellar motions to zero. We have compared the radio and infrared proper motions directly in order to transfer the infrared motions to a reference frame tied to Sgr A* (Reid et al. 2003). In principle, one can make this reference frame transfer using a single star with well determined motions. However, we chose to average the results from the four SiO maser stars within 10 arcsec of Sgr A* that have measured proper motions both in the radio and infrared. The unweighted mean difference (and standard error of the mean) of these stars is 0.8410.85 mas yr-' toward the east and -0.2510.96 mas yr-l toward the north. Since I mas yr-' corresponds approximately to 38 km s-I (for a distance to the Galactic center of 8.0 kpc), we conclude that the central star cluster moves with Sgr A* to within about 40 km sP1 per coordinate axis, or within about 70 km s-l for a 3-dimensional motion. 3.2

Proper Motion of Sgr A*

The apparent proper motion of Sgr A*, with respect to extragalactic radio sources, was measured with the VLBA by Reid et al. (1999). In that program, Sgr A* was used as a phase reference to calibrate the interferometer phases for two compact radio sources. The location of these sources is shown in Figure 3. Relative to an extragalactic source, Sgr A* would be expected to move toward the south-west, mostly along the Galactic plane (as depicted in Fig. 3), owing to the orbit of the Sun about the Galactic center. Figure 4 shows the position residuals of Sgr A* relative to the compact extragalactic source 51745283. The positions for 1995 through 1997 are from Reid et al. (1999). Those for 1998 through 2000 are new measurements. As one can see, the apparent motion of Sgr A* continues along the Galactic plane. The dominant term in the apparent motion of Sgr A* comes from the orbit of the Sun. This can be decomposed

508

M. J. Reid et al.: Position, Motion & Mass of Sgr A*

Fig. 2 Location of Sgr A* on 2 May 2002 on an infrared (2 p) image of the inner 2 arcsec of the Galactic center, adapted from Reid et al. (2003). The circle centered at the position of Sgr A* has a radius of 10 mas, corresponding to a 117position uncertainty. The star S2 was near pencenter in its orbit about Sgr A* when this image was taken (Schodel et al2002; Ghez et al. 2003)

into a circular motion of the local standard of rest, Oo/Ro = 220 k m sC1/8.O kpc (see Reid et al. 1999for details), and the peculiar motion of the Sun, Vo/& M 20 km s-'/8.0 kpc. Removing these terms from the observed proper motion, yields estimates of the peculiar motion of Sgr A*.

+

While we currently do not know the component of 00 Va in the plane of the Galaxy to better than about 10 to 20 km s-l , we do know the component perpendicular to the Galactic plane to better than 1 km s-l . Since the circular motion of the LSR is, by definition, entirely in the plane of the Galaxy, the only contribution to the apparent motion of Sgr A* perpendicular to the plane of the Galaxy is the Z-component of the Sun's peculiar motions, V z o . This component can be estimated by averaging the motions of very large numbers of stars in the Solar neighborhood, which should directly indicate -Vzo. An estimate, using the Hipparcos database, indicates Vza = 7.16 i 0.38 km s-l toward the north Galactic pole (Dehnen & Binney 1998). After removing this contribution to the apparent motion of Sgr A*

Astron. Nachr./AN 324,No. S1 (2003)

509

Fig. 3 Positions of Sgr A* and two compact extragalactic radio sources superposed on a 90 cm wavelength image of the Galactic center region made with the VLA by LaRosa et al. (2000). The expected motion of Sgr A*, owing to the orbit of the Sun about the Galactic center, is indicated by the arrow.

perpendicular to the plane of the Galaxy, we arrive at an estimate of 5 i3 km s-l for this component of Sgr A*’s peculiar motion. This result significantly improves on the limits given by Reid et al. (1999) and Backer & Sramek (1 999).

4

The Mass of Sgr A*

From infrared observations of stellar orbits (Schodel et al. 2002, Ghez et al. 2003), we know that a mass of M 3 x lo6 Ma is contained within a radius of = 100 AU. With this information, and an upper limit on the Z-component of the velocity of Sgr A*, V,, for which we adopt 8 km s-’ , one can estimate a lower limit to the mass of Sgr A*. The basic parameters of the problem are the total enclosed mass, M p n r ( R )including , a possible SMBH and stars with typical individual mass, m, that are enclosed within a radius, R, and an upper limit on the Zcomponent of the velocity, V,, of a “test” object (Sgr A* in our case) of mass M . In the past, two limiting cases of mass estimators have been discussed for this problem: equipartition of kinetic energy (see Backer

M. J. Reid et at.: Position, Motion & Mass of Sgr A*

510 I

"

"

I

"

"

I

"

"

5

Sgr A* - 51745-283

0

A

i

v

-5

\ \

4 a,

m

2

-10

1998

0

\

5 k

0

Z

-15

1999

-20

-25 5

0

-5

-10

-15

East Offset (mas) Fig. 4 Position residuals, with lu error bars, of Sgr A* relative to J1745-283 on the plane of the sky. Each measurement is indicated with an ellipse, approximating the apparent scatter broadened size of Sgr A* at 43 GHz. The dashed line is the variance-weighted best-fit proper motion, and the solid line gives the orientation of the Galactic plane. The expected position of Sgr A* at the beginning of each calendar year is indicated.

& Sramek 1999) and momentum (see Reid et al. 1999). Equipartition of kinetic energy implies that

MV2

N

mu2 ,

(1)

where v2 NN GM,,,(R)/R is a characteristic stellar velocity at radius R (which must be great enough so that the mass in stars exceeds that of the the test object, i.e., Menc(R)2 2M). Equipartition of energy is both theoretically and observationally well founded for the case of stellar clusters. However, for the case of a dominant central mass, which could greatly exceed the total mass of stars (within a given radius), Reid et al. argued that one would be dealing with true orbits and that equipartition of momentum would then be appropriate: MVwmu. (2)

Astron. Nachr./AN 324, No. S1 (2003)

51 I

It turns out that both estimators are correct, but for answering different que3tions. If one asks what is the expected velocity of a SMBH that is perturbed by close passages of stars which orbit it, then the momentum equation (2) applies. This is almost surely the case for Sgr A* and nearby stars such as S 1 and S2. For star S2, which has a mass rn of x 15 Mo (Ghez et al. 2003), during pericenter passage t i M 6500 km s-l and Eq. (2) implies that one would expect a 3 x lo6 M a SMBH’s peculiar motion to be V N 0.03 kni s-l . Following this approach one can calculate an extremely conservative lower limit for Sgr A*’s mass. While this is valid, it is not an optimum estimate. Alternatively, if we ask for the minimum mass of a central object which does not totally dominate the within a given radius R, and which complies with the observed velocity limit, enclosed mass, hfenC(R), then we have a different case. For this case, where the enclosed mass in stars within R is comparable to or exceeds the mass of Sgr A*, M , equipartition of kinetic energy should apply. When evaluating Eq. (1) one must use velocities for stars with radii near R. Conceptually, as the velocity limit for Sgr A* improves, the estimated mass limit increases quadratically in V. This continues until the estimated mass dominates over the stellar component and our assumption is violated. At this point, however, one has already ascribed most of the enclosed gravitational mass to Sgr A*. A recent paper by Chatterjee, Hernquist & Loeb (2002) analyzes our mass estimation problem in a manner similar to that described above. They assume a black hole at the center of a stellar cluster, which is distributed in space according to a Plummer profile with a characteristic scale a. The rnininiurn black hole mass occurs for a approximately equal to the radius, R, within which the enclosed mass is measured. In this case, the mass estimator (their equation 42) can be simplified to the following:

provided V,” > Gm/R,which is met for V, = 8 km s-’ , 711 = 1 Ma,and R = 100 AU. Then for the observed M,,,(R) = 3 x lo6 Ma, Eq. 3 gives a lower limit to the mass of Sgr A*of M 2 . 4 x los M a . Our lower limit to the mass of Sgr A* is now within about a factor of 10 of the total mass required by recent IR data. Since the uncertainty in the proper motion (cv)can decrease with the spanned observing ’7 ; (until the limit approaches time ( T )as ov c( T-”’, and the lower limit to the mass of Sgr A* scales as ~ the total enclosed mass), we can expect improvement in the limit for A4 cx T 3 over the next few years. When we reach a motion limit of 1 to 2 km s-l for Sgr A*, then essentially all of the mass sensed gravitationally by stellar orbits must come from Sgr A* itself. Should future VLBI measurements at 5 1inm wavelength show that the intrinsic size of Sgr A* is 5 0.1 AU, then we may be in a position to conclude that for Sgr A* most of the mass required for a SMBH is contained within a few Rsch!

References Backer, D. C. & Sramek, R. A. 1999, ApJ, 524, 805 Chatterjee, P., Hernquist, L., & Loeb, A. 2002, ApJ, 572,37 1 Dehnen, W. & Binney, J. J. 1998, MNRAS, 387 Ghez, A. et al. 2003, to appear in ApJ(Lett.) LaRosa, T. N., Kassim, N. E., Lazio, T. J. W., & Hyman, S. D. 2000, AJ, 119,207 Lenzen, R., Hofmann, R., Bizenberger, P., & Tusche, A. 1998, Proc. SPIE, 3354, 606 Menten, K. M., Reid, M. J., Eckart, A,, & Genzel, R. 1997, ApJ(Lett.),475, L111 Reid, M. J., Readhead, A. C. S., Verrneulen, R. C., & Treuhaft, R. N. 1999, ApJ, 524, 816 Reid, M. J., Menten, K. M., Genzel, R., Ott, T., Schodel, R. & Eckart, A. 2003, to appear in ApJ, 587 Rousset, G. et al. 2000, Proc. SPIE, 4007, 72 Schodel, R. et al. 2002, Nature, 419, 694

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1.513-519 (2003)/ DO1 10.l002/asna.200385098

Tidal processes very near the black hole in the Galactic Center Tal Alexander Faculty of Physics, The Weizmann Institute of Science, POB 26, Rehovot 76100, Israel

Key words Galaxy: center, galaxies: nuclei, Black hole, stars: kinematics Abstract. The accessibility of the 3 x loGA40 massive black hole (MBH) in the Galactic Center (GC) offers a unique opportunity to probe a variety of strong tidal interactions of stars with a MBH or with other stars in the high density cusp around a MBH. We show that such interactions can affect a significant fraction of the stellar population within the MBH radius of influence. We consider three processes that could possibly modify stellar structure and evolution there. (1) Tidal spin-up by hyperbolic star-star encounters. (2) Tidal scattering of stars on the MBH. (3) Tidal heating of tidally captured inspiraling stars-"squeezars". We discuss the implications for stellar populations near MBHs and for the growth of MBHs by tidal disruption, and the possible signatures of such processes in the GC. We compare the event rates of prompt tidal N

encounters (tidal disruption and tidal scattering) with slow inspiral events (squeezars / tidal capture), and find that, contrary to what was assumed in past studies, tidal capture in the presence of scattering is an order of magnitude less efficient than prompt disruption and so does not contribute significantly to the growth of the MBH.

1 Introduction Strong tidal interactions involving stars are expected to occur frequently near a MBH in a galactic center. First, the MBH is a mass sink, which drives a flow of stars from the MBH radius of influence T h to the center, to replace those it has destroyed. An inevitable consequence of this flow is that some stars are deflected into orbits whose periapse r p lies just outside the critical radius for destruction. We focus here on MBHs like the one in the GC, whose mass m is small enough so that the event horizon T , lies inside the where M, and R, are the stellar mass and radius 108Mo tidal disruption radius T t R*(T~/M*)'/~, for a solar type star). Such stars will suffer an extreme tidal impulse, but will not be destroyed, at least not on their first peri-passage. There are two possible outcomes: that the star is ultimately disrupted, or that it avoids subsequent encounters with the MBH. Both will be considered here in detail. Second, a variety of formation scenarios predict that MBHs should lie in the center of a high density stellar cusp (e.g. Bahcall & Wolf 1977; Young 1980). The diverging stellar density implies that there must be some volume around the MBH where close tidal encounters occur on timescales significantly shorter than the typical stellar lifetime. Such encounters will have a very different outcome from those that occur in globular clusters that do not contain a MBH. In most cases the encounters will not lead to tidal capture. Instead the two stars will continue on their separate ways after experiencing a brief strong tidal impulse. Extreme tidal interactions, which transfer energy and angular momentum from the orbit to the star, can affect its structure and subsequent evolution by heating it, spinning it up, mixing it, or ejecting some of its mass. This is interesting in view of the observed presence of unusual stellar populations near MBHs: the blue nuclear cluster in the inner 0.02 pc of the GC (Genzel et al. 1997), and around the MBH in M3 1 (Lauer et al. 1998); evidence for anomalously strong rotational dredge-up in an M supergiant near the MBH in the GC (Can; Sellgren & Balachandran 2000), but not in a high density nuclear cluster without a MBH (Ramirez et al. 2000); the unusually high concentration of very rare extreme blue He supergiants around the Galactic MBH (Krabbe et al. 1991; Najarro et al. 1994). The observable signature of an extreme tidal interaction on a star cannot be predicted with certainty at this time, although some reasonable conjectures can be made (Alexander & Livio 2001; Alexander &

-

(ms

-

@ 2003 WILEY-VCH Verlq GmbH & Co KGaA. Weinhem

T. Alexander: Tidal processes near Black Hole

5 14

Morris 2003). It is clear however, as shown below, that the tidal energy deposited in the star can reach a significant fraction of its binding energy, and that the angular momentum extracted from the orbit can spin it up to a significant fraction of its break-up velocity. We therefore proceed on the assumption that the effects can be observationally interesting and explore the dynamical processes that give rise to such interactions. Furthermore, tidal disruption and collisional stellar mass loss are important channels for supplying mass to a low-mass MBH (Murphy, Cohn & Durisen 1991), and so extreme tidal processes may provide observable links between the properties of the stellar population near the MBH and its evolutionary history. At a distance of 8 kpc (Reid 1993), the low mass 3 x lo6 Ma MBH in the GC (Ghez et al. 2000; Schodel et al. 2002) is the nearest and observationally most accessible MBH. Although heavily reddened, deep high resolution astrometric, photometric and spectroscopic IR observations of thousands of stars near the MBH provide information on their luminosity, colors and orbits (e.g. Eckart, Ott & Genzel 1999; Figer et al. 2000; Gezari et al. 2002). The GC thus offers a unique opportunity to probe a variety of close tidal interactions of stars with a MBH or with other stars in the high density cusp around it.

-

-

2 Tidal Spin-up by star-star encounters Stars moving rapidly around a MBH in a dense stellar cusp will suffer numerous hyperbolic tidal encounters over their lifetimes. Although such encounters transfer some orbital energy and angular momentum to the colliding stars, they rarely remove enough energy for tidal capture. This is in marked contrast to collisions in high density cores of globular clusters without a MBH, where the colliding stars are on nearly zero-energy orbits and close collisions result in tidal binary formation. The effects of hyperbolic encounters on the stars are mostly transient. The stellar dynamical and thermal timescales are very short compared to the mean time between collisions, and so apart from some mass-loss in very close collisions, the star is largely unaffected. It is however more difficult for the star to shed the excess angular momentum, since magnetic breaking operates on timescales of the order of the stellar lifetime (Gray 1992). High rotation is therefore the longest lasting dynamical after-effect of a close encounter. Over time, the stellar angular momentum will grow in a random walk fashion due to successive, randomly oriented tidal encounters. We consider the tides raised by an impactor star of mass m on a target star of mass M, and radius R, as m follows an unbound orbit with periapse rP from M,. - We will use the tilde symbol to express quantities in - units where G = M* = R, = 1. In these units, 0 = 1 is the centrifugal break-up angular frequency, Eb = -1 is the stellar binding energy, up to a factor of order unity, and Ft = 6 * / is3 the tidal radius. The orbital energy A E and angular momentum AJ”that are transferred from the orbit to the star by an impulsive tidal encounter are related by AE= fipAY, where fip is the relative angular velocity at periapse. Assuming for simplicity rigid body rotation, the change in the stellar angular velocity due to one parabolic encounter is given to leading order in the linear multipole expansion by (Press & Teukolsky 1977)

where i i s the stellar moment of inertia (assumed to remain constant), Tz is the C=2 tidal coupling coefficient (calculated numerically for a given stellar structure model), and q= [
515

Astron. Nachr./AN 324. No. S1 (2003)

0 05

01

0 15

02

0 25

03

r (PC) Fig. 1 Tidal spin-up in the Galactic Center (Alexander & Kumar 1999). The mean spin-up of a solar type star after 10 Gyr, in terms of the breakup angular velocity, is plotted as function of distance from the MBH.

type star after 10 Gyr in the GC, as function of distance from the MBH, in a high density, n* oc T - ~ . cusp of a continuously star forming population that is deduced to exist there (Alexander & Sternberg 1999; Alexander 1999). 66 reaches values as high as 0.3 within 0.03 pc, (- 60 times higher than is typical . falls off only slowly with distance in the field) and decreases to 0.1 at 0.3 pc (- 0.1 to 0.2 of ~ h ) 66 from the MBH because the increased tidal coupling in slower collisions far from the MBH compensates for the decrease in the stellar density there. Thus, long-lived main sequence stars with inefficient magnetic breaking are expected to rotate on average at a significant fraction of their centrifugal breakup velocity in a large volume of the dense stellar cusp around a MBH . N

N

3 Tidal scattering by the central black hole Tidal disruption is important for feeding low-mass MBHs that accrete from a low density cusp where collisional mass-loss is low. Numeric models of the growth of a MBH in a stellar cluster suggest that tidal disruption supplies between f -0.15 (Murphy, Cohn & Durisen 1991) to f -0.65 (Freitag & Benz 2002) of the MBH mass, depending on model assumptions. We adopt here a representative value of f =0.25. Dynamical analyses of 2-body scattering, which deflects stars into "loss-cone'' orbits that bring them within Ft of the MBH (Frank & Rees 1976; Frank 1978; Lightman & Shapiro 1977; Magorrian & Tremaine 1999; Syer & Ulmer 1999) show that tidally disrupted stars typically originate from the MBH radius of influence, Fh = F i / Z z , where Z is the 1D velocity dispersion far from the MBH. The stellar mass within f;, is comparable to EL,and so the stars are on slightly unbound orbits relative to the MBH. The cross-section for such stars to pass within Fp of the MBH scales as Fp' (Hills 1975; 55). It then follows that for every star on an orbit with 0 5 Fp 5 Ft that is disrupted promptly, there is a star that skirts the tidal disruption zone on an orbit with Ft 5 Fp 5 2Ft. Such "tidally scattered" stars narrowly escape disruption on their first peri-passage after being subjected to extreme tidal distortion, spin-up, mixing and mass-loss (Fig. 2). We will now argue that such stars avoid subsequent total disruption with high probability, either by being deflected off their orbit or by missing the MBH due to its Brownian motion (Alexander & Livio 2001). First, the scattering timescale is shorter than the dynamical one, and so stars wander in and out of the loss-cone several times during one orbital period. After the first peri-passage, the stars are on very eccentric N

' This holds for isotropically scattered $tars on parabolic orbits. The loss-cone replenishment rate peaks sharply at where the empty loss-cone (dfisive, or small angle scattering) regime changes to the full loss-cone ("pinhole", or large angle scattering) Fh,

regime. Large angle deflection relative to the loss cone opening angle leads to an isotropic redistribution of the velocity.

~

516

T. Alexander: Tidal processes near Black Hole

1 1

1:

t

t

I step

74500

Fig. 2 A Snapshot from an SPH simulation of a star tidally scattered by a MBH, shortly after pen-passage (F-& = 8). Left: the star and the MBH trajectory in the star’s rest frame. Right: a zoomed image of the star. The star is on a parabolic orbit with Fp = 1.5Ft and appears to be on the verge of breaking in two. However, by the end of the

simulation the two fragments coalesce, leaving a distorted, mixed and rapidly rotating bound object.

orbits with apoastron 5 27h, and so there is a considerable chance that they will be scattered again out of the loss-cone before their next close passage, and avoid eventual orbital decay and disruption.

-

Second, the MBH Brownian motion relative to the stellar system further increases the survival probability. A low mass MBH evolving in an initially constant density core of radius Fh, has Brownian fluctuations with an amplitude that is much larger than the tidal radius (Bahcall & Wolf 1976),

The timescale of the Brownian motion, like the orbital period of the scattered stars, is the core’s dynamical time. The orbits of the Etan take them out of Fa,where they are affected only by thc center of mass of the nucleus, and not by the relative shift between the MBH and the stellar mass. Therefore, on re-entry into the volume of influence, their orbit will not bring them to the same peri-distance from the MBH.

-

These qualitative arguments are verified by detailed analysis (see $ 5 ) . which shows that the survival 0.8-0.9. This implies that the mass fraction of surviving probability of tidally scattered stars is P, tidally scattered stars within Fh is comparable to the mass fraction of the MBH supplied by tidal disruption. Depending on the definition of what constitutes an extreme tidal interaction, there exists some limiting periapse, %, parameterized by b’ = %IFt, which corresponds to sufficiently strong tidal interactions. For example, for a solar type star, 6’ = 1.25 corresponds to a minimal tidal energy deposition of A E = 0.02 (Eq. 3). Since the angular velocity at periapse is f&= (2/b’3)1/2 = 1.01 and the solar moment of inertia 0.07, this translates to a minimal angular spin deposition of AT= 0.02 and a minimal spin-up of is A6=0.28. Over time, the mass fraction of tidally scattered stars within NFh will rise to f(b’-l)Ps -0.05 (for f = 0.25, b’ = 1.25 and P, = 0.8). Tidally scattered stars thus constitute a non-negligible fraction of the stellar population in the MBH radius of influence, and will remain there as relics of the early stages of the MBH evolution even after its mass grows above the tidal disruption limit, possibly detectable by correlations between unusual spectral properties and highly eccentric orbits.

r=

Astron. Nachr./AN 324. No. S I (2003)

517

4 Squeezars: Tidally powered stars A small fraction of the stars deflected into Fp 2 Ft orbits will be tidally captured by the MBH and spiral into an ever tighter orbit as orbital energy is gradually extracted each peri-passage. The orbital energy a star has to lose to circularize from an E = 0 orbit exceeds its binding energy by orders of magnitude, E, = 6 2 // 23b >> 1 where b = ;jlp/Ft. A tidally heated star-a “squeezar”-will ultimately be disrupted by expanding beyond its Roche lobe or by radiating above its Eddington luminosity. The orbital and internal evolution of an inspiraling squeezar depends on its initial structure and on the changes in its mechanical and thermal properties in response to the tidal heating. One approach to the challenging problem of modeling squeezar evolution is to consider two simplified cases that likely bracket the range of possible responses (Alexander & Morris 1993): (1) Surface heating and radiative cooling (“hot squeezar”), where the tidal oscillations dissipate in a very thin surface layer that expands moderately and radiates at a significantly increased effective temperature (McMillan, McDermott & Taam 1987). (2) Bulk heating and adiabatic expansion (“cold squeezar”), where the tidal oscillations dissipate in the stellar bulk and cause a large, quasi self-similar expansion at a constant effective temperature (Podsiadlowski 1996). Given these prescriptions, the evolution of the squeezar orbit, size, luminosity and temperature can be derived from the tidal energy deposition equation ~

A E = T2(b3I21 2/ b 6

(6 >> 1)>

(3)

where is the expanded stellar radius in terms of the original radius, and the Keplerian orbital equations for the semi-major axis a, the period p and eccentricity e, in terms of the orbital energy E are

Figure 3 shows the evolution of a 1Ma hot squeezar, with the tidal coupling coefficient calculated for a solar model (Alexander & Kumar 2001). The mean number of squeezars at any given time is E = % J i , where t o is the mean inspiral time, and r Lis the inspiral event rate. In the GC, 5 0.1-1 (55), assuming conservatively that the loss-cone is replenished only by two body scattering in a spherical system. One way of expressing the observable consequences of the existence of A squeezars, on average, near the MBH is by the properties of the “leading squeezar” (the one with the shortest period). The leading squeezar has, on average, completed :/to = E / ( f i + l ) of its inspiral. Figure 3 presents the mean properties of the leading squeezar, as function of W. It is evident that the effects of tidal heating on the leading squeezar can be quite pronounced even if W is small, and that the properties of the leading squeezar can be much more extreme if n was under-estimated (see discussion in Alexander & Hopman 2003). N

5 Prompt disruption .vs. slow inspiral The orbital decay of a tidally heated star is just one case of a dissipative interaction that can lead to orbital inspiral. Other possibilities include gravitational wave (GW) emission (Hils & Bender 1995; Sigurdsson & Rees 1997; Freitag 2001, 2003) or drag against a massive accretion disk (Ostriker 1983; Syer, Clarke & Rccs 1991; Vilkoviskij & Czemy 2002). Unlike prompt disruption or tidal scattcring, where the star reaches the MBH directly in less than the initial orbital period Po, slow inspiral proceeds gradually over a timescale TO>> PO,which is typically a steeply rising function of the periapse. For the extracted orbital energy to power a high luminosity of gravity waves, tidal heat, or mechanical energy in the disk, as the case may be, the star has first to decay into a short period orbit. The time available for inspiral is limited by two-body collisions similar to those that deflected the star into its eccentric orbit in the first place, since this poses a they can deflect it again to a wider orbit where the dissipation is inefficient. Because t o >> much more severe constraint for an inspiraling star than for a promptly disrupted star.

PO,

T. Alexander: Tidal processes near Black Hole

518

0.1

0

0.25

0.2

Mean number of squeezars 0.5 0.75 1 1.5 2

0.4

0.6

'

3 45

0.8

10

1

tdwupt

Fig. 3 Hot squeezar evolution in the GC (Alexander & Moms 2003). The 1 Ma star is on an orbit with b = 1.5 and PO= 1 . 4 ~ 1 yr 0 ~(to =4.9x105 yr) and is disrupted when k = b, after 3 . 7 lo5 ~ yr. At disruption the tidal luminosity 640 times larger than the intrinsic one, but the orbit is still almost radial with eccentricity e = 1 - 2.3 x The average of properties of the leading squeezar as function of the mean number of squeezars can be read off the top axis.

-

-

Novikov et al. (1992) estimated that tidal capture' by a MBH occurs for orbits with b < b, 3. This means that disruption due to tidal capture is b, - 1 2 more frequent than disruption due to prompt infall. These considerations led to the suggestion (Frank & Rees 1976; Novikov et al. 1992; Magonian & Tremaine 1999) that slow tidal inspiral may be at least as important as prompt disruption for feeding the MBH and for producing observable tidal flares (Frank & Rees 1976). This implies that the already large contribution of prompt tidally disrupted stars to the mass budget of a low-mass MBH ($3) should be further scaled by the ratio of the cross-sections of tidal capture and prompt disruption, b, - 1. If the relative contribution of inspiraling stars were indeed so high, the implications would be far-reaching: stars could supply most or even all of the MBH mass, thereby establishing a direct link between 6 and stellar dynamics on a scale of r h . However, a small initial periapse does not in itself guarantee inspiral and ultimate disruption. The star must also have enough time to complete its orbital decay. This time constraint can be taken into account correctly by considering only stars that are scattered from a volume that is close enough to the MBH so that the inspiral can be completed in time (Alexander & Hopman 2003). The inspiral time increases with F,. Therefore, the volume from which a scattered star can inspiral faster than the time it takes for two body relaxation to significantly deflect it, decreases with r p . The maximal possible periapse corresponds to the point where the available volume shrinks to zero (The truncation of the cusp near the MBH by destructive stellar collisions also limits the available volume). We find that for the GC, the rate, mean number of inspiraling stars and the maximal periapse for inspiral yr-', A= 0.2 and max rp = 2 . 1 for ~ 1MDhot squeezars; rZ= 4 x yr-', 5i= 0.2 are I?, = 3 x and maxr, = 2 . 8 ~for 1Ma cold squeezar; and = 2 x 10V7yr-*, 5L= 0.04 and mar, = 25r, for gravitational wave emission by 0.6 Ma white dwarfs (WD) that comprise 10%of an old stellar population. The rates were estimated for a spherical single-mass Keplerian stellar cusp, n, K T - ~', normalized to contain a total stellarmass 2 . 6 ~ 1 0M~a (=m)within T h = 1.8pc (after Schodel et al. 2002). The predicted prompt disruption rate for this simple model is I?, =9 x lop5 yr-', in general agreement with independent yr-' estimates from previous studies, rp= 5 x yr-I (Syer & Ulmer 1999) and rp few x (Alexander 1999). The rate of WD inspiral derived here, rz 2 x lOP7yrp1, is also consistent with the N

N

N

Based on the requirement that after the first pi-passage the apoapse should decrease to ZG 5 Fo,where Fo is the distance from which the star was scattered into the loss-cone. This tidal capture criterion does not include timescale considerations.

Astron. Nachr./AN 324,No. S1 (2003)

519

estimates of Sigurdsson & Rees (1997) and Freitag (2003). The survival probability of tidally scattered stars can also be calculated by the formalism of Alexander & Hopman (2003),and it is found to be close to unity, P, 0.8-0.9, as anticipated by general arguments (53). However, we find that the tidal inspiral rate is only 0.05 of the prompt disruption rate, and not 2-3 times larger, as was assumed by previous studies. We conclude that the contribution of tidal capture to the MBH mass budget and to the tidal flaring rate from galactic nuclei is negligible compared to prompt disruption. Past studies, which assigned similar weights to prompt disruption and tidal capture, overestimated the contribution of tidal disruption to the growth of the MBH by at least a factor of two. N

N

6 Summary We show that strong tidal interactions of stars with the MBH or with other stars in the high density cusp around it can deposit large amounts of orbital energy and angular momentum in a significant fraction of the stellar population within a large volume around the MBH. We propose that such interactions could alter the evolution and appearance of stars in galactic centers and thereby probe the evolution of the MBH and the stellar system around it. We explored tidal spin-up by star-star encounters, tidal scattering by a MBH and tidal inspiral into a MBH. We showed that tidal capture is inefficient in the presence of two body scattering.

References Alexander, T. 1999,ApJ, 520, 137 Alexander, T.& Kumar, P. 2001,ApJ, 549,948 Alexander, T., & Livio, M. 2001,ApJ, 560, L 143 Alexander, T., & Moms, M. 2003,ApJL, submitted Alexander, T., & Hopman, C. 2003,ApJL, submitted Alexander, T., & Sternberg, A. 1999,ApJ, 520,137 Bahcall, J. N., & Wolf, R. A. 1976,ApJ, 209,214 Bahcall, J. N., &Wolf, R. A. 1977,ApJ, 216, 883 Can; J. S.,Sellgren, K., & Balachandran, S. C., 2000,ApJ, 530,307 Eckart A., Ott, T., & Genzel, R. 1999,A&A, 352,L22 Figer, D. F., et al. 2000,ApJ, 553, LA9 Frank, J. 1978,MNRAS, 184,87 Frank, J. & Rees, M. J. 1976,MNRAS, 176,633 Freitag, M. 2001,Class. Quant. Grav., 18, 4033 Freitag, M. 2003,ApJ, 583,L21 Freitag, M., & Benz, W. 2002,A&A, 394,345 Genzel, R., Eckart, A., Ott, T., & Eisenhauer, F. 1997,MNRAS, 291 ,219 Gezari, S.,Ghez, A.M., Becklin, E.E., Larkin, J., Mchan, I.S., Morris, M. 2002,ApJ, 576,790 Ghez, A. M., Moms, M., Becklin, E. E., Tanner, A. & Kremenek T. 2000,Nature, 407,349 Gray D.F. 1992 in The Observation and Analysis of Stellar Photospheres (2nd Ed.; Cambridge: CUP), 386 Hils, D., & Bender, P. L. 1997,ApJ, 445,L7 Hills, J. G. 1975,Nature, 254,295 Krabbe, A., Genzel, R., Drapatz, S., & Rotaciuc, V., 1991,ApJ, 382,L19 Lauer, T. R., Faber, S. M., Ajhar, E. A., Grillmair, C. J., & Scowen, P. A,, 1998,AJ, 116,2263 Lightman, A. P., & Shapiro, S. L. 1977,ApJ, 21 I , 244 Magoman, J, & Tremaine, S. 1999,MNRAS, 309,447 McMillan, L. W., McDermott, P. N., & T a m , R. E. 1987 ApJ, 318,261 Murphy, B.W., Cohn, H. N., & Durisen, R. H.1991,ApJ, 370.60 Najarro, F., et al., 1994,ABA, 285,573 Novikov, I. D.,Petchik, C. J., & Polnarev, A. G. 1992,MNRAS, 255,276 Ostriker, J. P. 1983,ApJ, 273.99 Podsiadlowski, P. 1996,MNRAS, 279,1104 Press, W.H., & Teukolsky, S. A. 1977,ApJ, 213, 183 Ramirez. S.V., et al. 2000,ApJ, 537,205 Reid, M.J. 1993,ARA&A, 31, 345 Schodel, R., et al. 2002,Nature, 419,694 Sigurdsson, S., & Rees, M. J. 1997,MNRAS, 284,318 Syer, D.,Clarke, C. J., & Rees, M. J. 1991,MNRAS, 250,505 Syer, D., Ulmer, A. 1999,MNRAS, 306,35 Vilkoviskij, E.Y.,& Czemy, B. 2002,A&A, 387,804 Young, P. 1980,ApJ, 242,1232

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.521-526 (2003) / DO1 10.1002/asna.200385059

New MIR Excess Sources north of the IRS 13 Complex *. A. Eckart**', J. Moultaka I, T. Viehmann I , C. Straubmeier I, N. Mouawad 2, T. Ott 2, and R. Schodel

', R. Genzel

' LPhysikalisches Institut, Universitat zu Kbln, Zulpicher Str.77,50937 Koln, Germany Max-Planck-Institut fur extraterrestrische Physik Giessenbachstralle, 85748 Garching, Germany

Key words Galactic Center, star formation, dust, extinction, pre-main-sequence.

Abstract. The ISAAC and NAOS/CONICA systems on the ESO VLT UTI and UT4 telescopes have been used to image the central stellar cluster of the Milky Way in the MIR L and M band (3.3 pm and 4.6 j m ) in order to obtain new information on the stellar populations within the central cluster. The NAOS/CONICA adaptive optics images of unprecedented sharpness and sensitivity have resulted in the detection of a small cluster (<0.13 light years) of 7 compact sources with a strong infrared excess. Our MIR imaging and spectroscopy data clearly show that this excess is due to the contribution of warm T<500 K dust. The nature of these newly found sources is currently unclear. Three possible explanations are mentioned and briefly discussed 1) The L band excess sources north of the IRS 13 complex may be heavily extincted luminous stars; 2) They may be hot stars that heat the more ambient environment of the local mini-spiral; 3) Finally the newly found objects may even be young stars with luminous accretion disks.

1

Introduction

The entire central parsec of our Galaxy is powered by a cluster of massive stars (Blum et al. 1988, Krabbe et al. 1995, Genzel et al. 1996, Eckart et al. 1999, ClCnet et al. 2001, Paumard et al. 2001, and Maillard et al. in these proceedings). Within that cluster the 7 most luminous (L> lo5 75 La ), moderately hot (T<104.5 K), blue supergiants contribute half of the ionizing luminosity of that region (Najarro et al. 1997, Krabbe et al. 1995, Blum et al. 1995). Such massive and hot stars have also been found in dense clusters within the Galactic bulge, i.e. the Arches cluster (Cotera et al. 1992, see also Figer et al. 2002 and references therein) and the Quintuplet cluster (e.g. Figer et al. 1997). The presence of young massive stars a s well as a few dust embedded, spectrally featureless objects within the central parsec of the Milky Way has frequently raised the question of whether or not active star formation can be sustained against the deep potential of the central stellar cluster with a 3 x lo6 M a black hole at the position of the radio source Sgr A* (Eckart & Genzel 1996, Genzel et al. 1997, Ghez et al. 1998, Genzel et al. 2000, Ghez et al. 2000, Eckart et al. 2002, Schodel et al. 2002; also contributions by Ghez et al., Schodel et al., Ott et al., and Reid et al. in these proceedings). Here we report on recent MIR L and M band observations that contribute to the investigation of the star formation process in the central stellar cluster. In particular we report o n the discovery of a small cluster of infrared excess sources which are likely to be dusty, embedded, and possibly young stars. * based on observatlons at the Very Large Telescope (VLT) of the European Southern Observatory (ESO) on Paranal in Chile ** Corresponding author: e-mail: [email protected],Phone: +49221470 3546, Fax: +492214705162

@ 2003 WlLEY VCH Verlag GmbH & Co KGaA, Weinhem

522

A. Eckart et al.: MIR VLT observations

2

Observations and data reduction

The infrared data were obtained with the new NAOSKONICA adaptive optics assisted imager/spectrometer at the UT4 (Yepun) as well as with ISAAC on UT1 (Antu), at the ESO VLT (Lenzen, et al. 1998, Rousset et al. 1998, Brandner et al. 2002). During the commissioning and observatory preparation of NAOSKONICA the instrument was used to obtain diffraction limited images in the H , K,, and L band centered at wavelengths of 1.6, 2.1, and 3.3 pm, with a FWHM of 43, 56, and 88 mas, respectively (at 0.0132”/pixel). Further details are given in Schodel et at. (2002) and Genzel et al. (2003). Our recent ISAAC L band imaging and spectroscopic monitoring program of Sgr A* (Baganoff et al. 2001 and these proceedings, Eckart et al. 2003) which has been carried out simultaneously to Chandra observations has lead to high quality L band spectra of several prominent sources within the central cluster that fell into slit settings across S g r A*. For spectroscopy and imaging we used a 1024x 1024 SBRC Aladdin array. Spectra were taken through a 0.6” wide slit across Sgr A* providing a resolution of R=600 across the L band from 2.8 to 3.6 pm. The wavelength calibration was done using Xe/Ar lamps. Separate 18’’ and random nodding within 2” along the observations were carried out using chopper throws of slit. The spectral response was calibrated via the reference stars HD 148703 and HIP 100881. A final correction for the foreground extinction was carried out via a comparison of a 3000 K black body spectrum to the spectrum of a late type star 7” east of the center, well off the region contaminated by thermal dust emission due to the mini-spiral in that area. As expected, this calibration step resulted in Rayleigh Jeans type spectra for the hot stars that are not located on the mini-spiral (see IRS 16NE and IRS16C in Fig.1).

-

-

3

MIR spectra of bright sources in the central cluster

Dust-rich regions show deep and broad HzO absorptions centered at w3pm wavelength. In addition they show absorption features just longward of 3.4 p m and 3.5 pm (e.g. IRS 21). Compared to the red and dust-rich objects the blue He-stars have bluer L band spectra and weaker or no HzO absorption features (Moultaka et al. 2003) since they are not embedded in dust shells. The residual absorptions are probably mostly due to material of the mini-spiral at the center. The spectra and red colors of the objects IRS 2 I , IRS 1W, and IRS 3 are most likely strongly influenced by the fact that they are interacting with gas and dust in the mini-spiral (Tanner et al. 2002, Genzel et al. 2003). The hot and dust free sources like IRS16NE, NW, C and CC are all well clustered in a HKL two color diagram. After taking into account the 30 magnitudes of visual extinction towards the Galactic Center, these objects fall close to the location of hot and massive stars near the main sequence. The fact that all the sources with similar spectroscopic properties also show similar colors indicates that the variation in foreground extinction with position is small and that its correction gives consistent results for all sources in the central few arcseconds. The recent observations of the GC using NAOSKONICA at the UT4 have produced 3 pm images of that region at unprecedented sharpness and sensitivity. A close comparison of the H , K,, and L band images reveals a cluster of previously unknown, weak (mL
Astron. Nachr./AN 324, No. S l (2003)

I

"

"

"

'

,

'

'

I

1

Fig. 1 ISAAC spectra of selected sources within the central stellar cluster.

523

A. Eckart et al.: MIR VLT observations

524

Fig. 2 L band images of the IRS 13 stellar complex obtained with CONICA/NAOS at the UT4. In (b) we show a high pass filtered L band image giving the location o f the individual stars (Greek letters) in the newly found cluster of young stars IRS 13N.

IRS 13N x 1.4

Fig. 3 ISAAC spectra of the sources IRS16C, the IRS13 complex, and the group of young stellar objects IRS13N located 0.5" north of the IRS13 complex.

IRS 13 complex and clearly demonstrates that the strong L band color excess is due to the emission of warm T d 0 0 K dust. The proper motion results by Ott et al. (2003) and Genzel et a1 (2003) indicate that the dynamical time scale for the combined motion of the entire IRS 13 complex and the mini-spiral gas is of the order of w104 years. The nature of these newly found L band excess sources is currently unclear. There are at least three possibilities that have to be taken into account:

Possibility No. 1: The stars could be bright young stars that are deeply embedded in the gas and dust of the mini-spiral close to the location of the IRS 13 complex. Alternatively they may be located behind a dense clump of gas and dust. In order to explain their K-L colors one would require an extinction that

Astron. Nachr./AN 324, No. S 1 (2003)

525

exceeds the average value towards the central region by a few 10 visual magnitudes. This would then require that the small cluster 0.5” north of the IRS13 complex consist of objects that are possibly as bright as the luminous He-stars in the overall central stellar cluster.

Possibility No. 2: The IRS 13N sources could be hot 0-stars that heat the gas and dust in their immediate vicinity within the mini-spiral. After correction of the 30 magnitudes of visual extinction the intrinsically redder colors (compared to the hot stars) of the new L band sources are most likely due to the presence of appreciable amounts of warm (500-1000 K) dust in their circumstellar environment. Detailed high resolution MIR imaging has to be carried out in order to determine how much of this dust is in the immediate environment of the stars (i.e. in accretion disks) and how much in a more ambient medium [see 12.5pm Keck image of that region in Tanner et al. 2002) in which the heating of the dust is provided by several stars. Here shocks and the nearby IRS I3 cluster of hot stars could be of importance in addition to some accretion through disks. Possibility No. 3: Finally - and this is the most exciting and extraordinary explanation - the IRS 13N sources may be a cluster of young stars. Here the K-L>3 color as well as the L, K, and H band flux densities at the positions of the IRS 13N sources (Eckart et al. 2003) are consitent with the colors and luminosities of young Herbig AeBe (Herbig group 11) stars or dust embedded young stellar objects reddened by 30 magnitudes. With lo3 L, the newly found L band excess sources are well within the expected luminosity range of 10’ to lo4 La for 2 to 8 Mo Herbig group I1 sources and young stellar objects (e.g. Fuente et al. 2002). In this scenario the L band excess may be due to warm dust located in their circumstellar disks. The dynamical time scale for the combined motion of the entire IRS 13 complex and the mini-spiral gas, on the order of -lo4 years, is small compared to the plausible ages of the massive young stars. With ages that probably range between 0. I and only a few Myrs (Hillenbrand et al. 1992, Fuente et al. 2002, Ishii et al. 1998) they must have been formed within 10 parsecs of the central cluster. The age range would imply that they could be 10 to 100 times younger than the hot He-stars in the central parsec, i.e. much less than lo7 years, while their minimum age is only a factor of 10 times larger than the above mentioned dynamical time scale of 1O4 years. However, the high degree of gas and dust compression expected within the mini-spiral streamers in combination with the extreme tidal fields due to the super-massive black hole (e.g. Sanders 1992, Gerhard 2001), may result in an accelerated formation and evolution of dust enshrouded young stellar objects. This indicates that the newly found IR-excess sources may indeed have been formed well within the central few light years, possibly even within the mini-spiral, N

-

It is not straightforward to decide which of these three possibilities is most suitable to explain the properties of the newly found L band excess sources. The explanations No. 1 and No. 2 require the presence of a small cluster of sources. On its way into the central stellar cluster this association of stars should already have sheared into a spatially broader distribution of objects. If the stars have been formed recently (potentially within the mini-spiral) then one has to explain the origin of the reservoir of gas and dust that is required for the (as yet unknown) formation process as well as the associated IME Explanation No. 1 requires exceptionally high extinction towards a cluister of probably very bright sources which appears to be an unlikely scenario. In the scenario implied by explanation No. 2, the detailed modeling of the sources based on highest angular resolution MIR imaging and spectroscopy (using A 0 or interferometry) will have to determine how the stars are associated exactly with the gas and dust that is heated by them. These investigations then have to determine how well the observable physical properties of the sources can be related to their age.

Acknowledgements This work was supported in part by the Deutsche Forschungsgemeinschaft (DFG) via grant SFB 494. We are grateful to M. Moms for valuable comments to an earlier version of the manuscript. We are also grateful to all members of the NAOSKONICA team from MPIAMPE, Meudon/Grenoble Observatories, ONERA, ESO. In particular, we thank N.Ageorges, K.Bickert, W.Brandner, Y.Cltnet, E.Gendron, M.Hartung, N.Hubin, C.Lidman, A.M. Lagrange, A.F.M. Moorwood, C.Rohrle, G.Rousset and J.Spyromilio.

526

A. Eckart et al.: MIR VLT observations

References Baganoff, EK. et al. 2001, Nature, 413, 45 Blum, R. D.; Sellgren, K.; Depoy, D. L. 1996, AJ, 112, 1988 Brandner, W. et al. 2002, The ESO Messenger 107, 1 Clhet, Y., Rouan, D., Gendron, E., Montri, J., Rigaut, F., Lfia, P., Lacombe, F. 2001, A&A, 376, 124 Cotera, A. S., Erickson, E. F., Colgan, S. W. J., Simpson, J. P., Allen, D. A., & Burton, M. G. 1996, ApJ, 461, 750 Draine, B.T. 1989, Proc. 22nd ESLAB Symp. on W Spectroscopyin Astron., Kaldeich, B.H. (ed.), ESA SP290, p. 93 Eckart, A. & Genzef, R. 1996, Nature, 383,415 Eckart, A., Ott, T., Genzel, R., & Lutz, D. 1999, in Proc. of IAU Symp. No.193 on 'Wolf-Rayet Phenomena in Massive Stars and Starburst Galaxies' Puerto Valarta, Mexico, November 3-7, 1998, van der Hucht, K.A., Koenigsberger, G., Enens, P.R.J. (eds.), Kluewer, pp.449 Eckart, A., Genzel, R., Ott, T. and Schoedel, R. 2002, MNRAS, 331,917 Eckart, A., et al. 2003, in preparation Figer, D.F., Najarro, F. et al. 2002, ApJ, 581,258 Figer, D. F., Najarro, F., McLean, I. S., Moms, M., Geballe, Th. R. 1997, in Luminous Blue Variables: Massive Stars in Transition. ASP Cod. Series; Vol. 120, p. 196, ed. A. Nota and H. Lamers Fuente, A,, Martin-Pintado,J., Bachiller, R., Rodriguez-Franco,A., Palla, F. 2002, A&A, 387,977 Genzel, R., matte, N., Krabbe, A., Kroker, H., Tacconi-Garman, L. E. 1996, ApJ, 472, 153 Genzel, T., Eckart, A., Ott, T.& Eisenhauer, F. 1997, MNRAS, 291,219 Genzel, R., Pichon, C., Eckart, A., Gerhard, 0. & Ott, T. 2000, MNRAS, 317,348 Genzel et al. 2003, ApJ, submitted Gerhard, 0.2001, ApJ, 546, L39 Ghez, A,, Klein, B.L., Moms, M. & Becklin, E.E. 1998, ApJ, 509,678 Ghez, A., Moms, M ,Becklin, E.E., Tanner, A. & Kremenek, T. 2000, Nature, 407,349 Hillenbrand, L.A., Strom, S.E., Vrba, F.J., Keene, J. 1992, ApJ, 397,613 Hillenbrand, L. A., Strom, S. E., Vrba, F. J., Keene, J. 1992, ApJ, 397,613 Hornstein, S.D., Ghez, A.M., Tanner, A,, Moms, M., Becklin, E.E., Wizinowicb, P. 2002, ApJL, 577, L9 Ishii, M., Nagata, T., Sato, S., Watanabe, M., Y., Yongqiang, J., Terry J. 1998, AJ, 116, 868 Krabbe, A. et al. 1995, ApJL, 447, L95 Lenzen, R., Hofmann, R., Bizenberger, P. & Tusche, A. 1998, Proc. SPIE, IR Astronomical Instrum. (A.M.Fowler ed.) 3354,606 Liu, S., & Melia, F. 2002, ApJL, 566, L77 Markoff, S., Falcke, H., Yuan, F., Biermann, P. L. 2001, A&A, 379, L13 Moultaka et al. 2003, in prep. Najarro, F., Krabbe, A., Genzel, R., Lutz, D., Kudritzlu, R. P.,Hilher, D. J. 1997, A&A, 325,700 Ott et al. 2003, ApJ, submitted Paumard, T., Maillard, J. P., Moms, M., & Rigaut, F. 2001, A&A, 366,466 Rousset, G., et al. 1998, Proc.SPIE Adaptive Optics Technology (D.Bonaccini,R.K.Tyson eds) 3353,508 Scbodel, R., et al. 2002, Nature, 419,694 Tanner, A,, Ghez, A. M., Moms, M., Becklin, E. E., Cotera, A., Ressler, M., Werner, M., Wizinowich, P. 2002, ApJ, 575,860 Yusef-Zadeh, F., Roberts, D. A., Biretta, J. 1998, ApJL, 499, L159

Astron. Nachr./AN 324, No. S1.527 - 533 (2003) / DO1 10.1002/asna.200385103

Full Three Dimensional Orbits For Multiple Stars on Close Approaches to the Central Supermassive Black Hole A. M. Ghez*', E. Becklin', G. DuchZne', S. Hornstein', M. Morris', S. Salim', and A. Tanner'

' Department of Physics and Astronomy, University of California, Los Angeles, CA 90095-162 Key words Black hole, Orbits, Adaptive Optics, Proper Motions, Spectroscopy

Abstract. With the advent of adaptive optics on the W. M. Keck 10m telescope, two significant steps forward have been taken in building the case for a supermassive black hole at the center of the Milky Way and understanding the black hole's effect on its environment. Using adaptive optics and speckle imaging to study the motions of stars in the plane of sky with f - 2 mas precision over the past 7 years, we have obtained the first simultaneous orbital solution for multiple stars. Among the included stars, three are newly identified (SO-16, SO-19, SO-20). The most dramatic orbit is that of the newly identified star SO-16, which passed a mere 60 AU from the central dark mass at a velocity of 9,000 km/s in 1999. The orbital analysis results in a new central dark mass estimate of 3.6(f0.4)x 106(&)3Ma. This dramatically strengthens the case for a black hole at the center of our Galaxy, by confining the dark matter to within a radius of 0.0003 pc or 1,000 R,h and thereby increasing the inferred dark mass density by four orders of magnitude compared to earlier estimates. With the introduction of an adaptive-optics-fed spectrometer, we have obtained the first detection of spectral absorption lines in one of the high-velocity stars, SO-2, one month after its closest approach to the Galaxy's central supermassive black hole. Both Br y (2.1661 pm) and He I (2.1126 pm) are seen in absorption with equivalent widths and an inferred stellar rotational velocity that are consistent with that of an 08-BO dwarf, which suggests that SO-2 is a massive (-15 M a ) , young (
1 Introduction While the Milky Way was neither the first nor the most obvious place to search for a supermassive black hole, the case for one at the center of the Galaxy is quickly becoming the most iron clad. The first hint of a central concentration of dark matter came from radial velocity measurements of ionized gas located in a t h r e e - m e d structure known as the mini-spiral, which extends from the center out to about 1-2 pc (Lacy et al. 1980). Concerns that the gases' motion are not tracing the gravitational potential were quickly allayed by radial velocity measurements of stars, which are not susceptible to non-gravitational forces (McGinn et al. 1989; Haller et al. 1996; Genzel et al. 1997). These early dynamical measurements of the gas and stars suggested the presence of 3 x 106Mo of dark matter and confined it to within a radius of -0.1 pc; the implied dark matter density was not sufficiently high to definitively claim this as evidence for a single supermassive black hole, since the measurements only imposed a lifetime for clusters of dark objects of 5 x lo9 yrs, which is not significantly shorter than the age of the Galaxy (Maoz et al. 1998). To make * Corresponding author: e-mail: ghez Oastro.ucla.edu, Phone: 3 10-206-0420, Fax: 3 10-206-2096

@ 2003 WLLEY-VCH Verlag GmbH & Co K G d . Wemherm

A. M. Ghez et al.: Stellar Orbits

528

further progress in understanding the underlying source of dark matter at the center of the Galaxy, it was necessary to use techniques that compensated for the distorting effects of the Earth’s atmosphere, which had restricted the earlier studies of the dark matter distribution to radii of 0.1 pc or larger. In the early- to mid-l990’s, two independent groups initiated 2 pm high spatial resolution imaging studies of the central stellar cluster to measure the motions of stars in the plane of the sky. While the ESO team began their program using speckle imaging at the 3.6 m NTT and this year have moved to the adaptive optics system on the 8 m VLT, the Keck team initiated their program using speckle imaging on the 10 m Keck I telescope and began using adaptive optics on the 10 m Keck I1 telescope in 1999. The first phase of these experiments yielded proper motion velocities, which increased the implied dark matter density by 3 orders of magnitude to 1012M0/pc3 (Eckart & Genzel 1997; Ghez et al. 1998). This eliminated a cluster of dark objects, such as neutron stars or stellar mass black holes, as a possible explanation of the Galaxy’s central dark mass concentration (Maoz et al. 1998) and left only the fermion ball hypothesis (e.g., Tsiklauri & Viollier 1998, Munyaneza & Viollier 2002) as an alternative to a single supermassive black hole. The velocity dispersion measurements also localized the dark matter to f l 0 0 mas (4 milli-pc) at a position consistent with the nominal location of the unusual radio source Sgr A* (Ghez et al. 1998), whose emission is posited to arise from accretion onto a central supermassive black hole (e.g., Lo et al. 1985). The proper motion experiments proceeded to strengthen both the case for a supermassive black hole and its association with Sgr A* with the detection of acceleration for three stars - SO-1, SO-2, and SO-4, which increased the dark matter density to 1013M0/pc3 and positional accuracy to f 3 0 mas (Ghez et al. 2000; Eckart et al. 2002). These experiments also revealed that the orbital periods for SO-2 and SO-1 could be as short as 15 and 35 years, respectively, which would open a new arena for dynamical studies of the central stellar cluster. This paper summarizes the recent progress that has been made in this field on two fronts with the W. M. Keck telescope. The first, with approximately a decade of proper motion measurements, is the derivation of complete 3-dimensional orbits for multiple stars that are making close approaches to the supermassive black hole at the center of the Milky Way Galaxy. The second is the measurement of spectral lines in one of these high velocity stars. These two steps forward make the strongest case yet for the presence of a supermassive black hole at the center of the Galaxy and, for the first time, allow us to take an in-depth look at the question of where these stars formed.

2 Observations & Results 2.1 Proper Motions Beginning in 1995, K[2.2 pm]-band diffraction-limited images have been obtained with the W. M. Keck I 10 m telescope to achieve an angular resolution of 50 milli-arcsec and a positional accuracy of -2 milliarcsec on stars located in the central 5”x5” of our Galaxy. Since our original reporting of stars in this region (Ghez et al. 1998), the sensitivity to high-velocity stars has been significantly improved. Two primary factors contribute to the increased number of recognized stars in this region. First, the onginal approach was conservative in an effort to avoid falsely identifying high velocity stars on the basis of the three maps separated by a full year. With multiple observations each year beginning in 1998, it was clear that many real sources were not being identified, leading to a decrease in the threshold used to identify sources. Second, in 1998 it became possible to calibrate the alignment of the Keck telescope’s mirror segments on NIRC, the Keck I facility infrared camera, allowing the telescope to be calibrated during the observing run at the elevation of the Galactic Center, thereby significantly improving the image quality of our speckle maps. motion, they Figure 1 shows the three new proper motion stars (SO-16, SO-19, SO-20) and the original proper motion stars (SO- 1, SO-2, SO-4) whose motions can now be modeled with Keplerian orbits. Of particular note among the new sources (those with a label larger than 15) is SO-19. Its high proper motion has caused it to be misidentified in earlier papers. While it was detected by us in 1995 as a K=14.0

Astron. Nacht/AN 324, No. S1 (2003)

529

source, two possible counterparts were identified in 1996. With limited time coverage, it was not possible to definitively identify either as the correct counterpart and it was not included in the proper motion sample (see discussion in Ghez et al. 1998). The same source was reported as 53 (K=15) moving Westward by Eckart & Genzel(l997) and Genzel et al. (2000). Genzel et al. (1997) report the detection of a new source, S12 (K-15),in their 1996.43, located -0."1 of a source labeled S3 and proposed it as the best candidate for the infrared counterpart of the compact radio source Sgr A*. This has been used in several recent papers to constrain models of Sgr A*'s flared state (e.g., Narayan et al. 2002). In our analysis, it is now clear that S12 is simply a high velocity star that was coincident with Sgr A* in 1996. It is labeled SO-19 in this paper and it should be associated with the 1992-1995 detection of S3. The discrepancy in magnitudes arises from the difficulties of carrying out accurate photometry in such a crowded region. This source illustrates the challenges associated with making a definitive detection of infrared emission associated with Sgr A*, given the high stellar densities and velocities and modest stellar intensity variations in this region.

Fig. 1 (a) A 1" x 1" cleaned image centered on the nominal position of Sgr A* showing the 2001 positions of some of the stars that have been followed over the course of the Keck proper motion study. Three of the newly identified stars are SO-16, SO-19, and SO-20. (b) The annual positions of some of these stars with orbital solutions. Each star is labeled by its first measurement.

2.2

Spectral Lines

In 2002, the high proper motion star, SO-2, was observed with NIRC2, the facility near-infrared adaptive optics instrument (Matthews et al. 2003) in a mode that achieved a spectral resolution of R-4000 (-75 k d s ) . The resulting spectrum of SO-2, shown in Figure 2, has two identifiable spectral lines. These are both seen in absorption and are identified as the H I (4-7) or Bry line at 2.166 pm and the He I triplet at 2.1 126 pm ( 3 p 3P0 - 4s 3S),which is a blend of three transitions at 2.1 1274,2.11267, and 2.1 1258 pm. The detailed properties of these two lines, which are obtained by fitting the background continuum over the whole spectrum with a low-order polynomial and fitting the lines with a Gaussian profile are reported in Ghez et al. (2003a).

530

A. M. Ghez et al.: Stellar Orbits

180

160

-

140

t

Fe

120

v

..-

100

80

60

2.05

2.10

2.15

2.20 A

2.25

(w)

2.30

0.1

0.05

0 -0.05 -0.1 bRA from Sgr A* (arcsec)

-0.15

Fig. 2 In the left panel is the first spectrum of SO-2 to show detectable photospheric absorption lines (Bry and He I (2.1126 pm)). The final spectrum (middle) is the raw spectrum (top; with only an instrumental background removed) minus a local sky (bottom). The horizontal dimension has been re-binned by a factor of two for display purposes only. The vertical lines are drawn at 2.10899 and 2.16240 pm, which correspond to the locations of Bry and He I for a VLSR of -513 Ws.This spectrum was obtained in 2000 June at the same time as one of the proper motion measurements reported by Ghez et al. (2003) and shown in the right panel (filled circles). The crosses mark January 1 of each year between 1995 and 2004 for the best fit orbit solution (solid line), wluch is based on both the radial velocity and proper motions. The dotted line is the line of nodes, which reveals SO-2 to be behind the black hole for a mere -0.5 years out of its 15 year orbit. Adapted from Ghez et al. (2003)

3 Discussion & Conclusions 3.1 Dynamics The strong deviations from linear motions on the plane of the sky for stars within O.”2 of Sgr A* along with the radial velocity measurements for SO-2 provide new and powerful constraints on their orbital parameters, which are presented in Ghez et al. (2003a, b) and Ghez (2003). In the orbital fits, we assume that ( I ) the stellar masses are insignificant compared to a central point source, (2) the central point source has no significant velocity with respect to the Galaxy, which is supported by the lack of motion detected for Sgr A* by Reid et al. (1999) and Reid (2003), and (3) the central point source has a distance of 8.0 kpc (Reid yo), Period (P), Semi-Major Axis 1993). This leaves the following 9 unknowns: Center of Attraction (zo, (A), Eccentricity (e), Time of periapse passage (To), Angle of nodes to periapse (w), Angle of the line of nodes (Q), and Inclination ( i). We begin by fitting each star independently. Each of these stars has recently gone through periapse (1999 - 2002). In the most extreme case, SO-16 passed within a mere 60 AU of the central dark mass, while traveling at a velocity of 9,000 km/sec. The solution for SO-2 is consistent with that reported by Schodel et al. (2002) and, despite the two additional free parameters introduced by fitting for the center of attraction, the uncertainties on the orbital parameters are reduced by a factor of 2-3. Since the independent centers of attraction are consistent with one another, we proceed to fit the orbital motion for all the stars simultaneously with a common center of attraction. This first orbital estimate of the Galaxy’s dynarnical center is not only consistent with the nominal infrared position of Sgr A* to within the uncertainties on the latter (Reid et al. 2003), but is also a factor of 7 more precise (*1.5 milli-arcsec). Furthermore, the agreement between the masses inferred from the simultaneous Keplerian orbit fits for multiple stars (see Figure 3 and Table 1) suggests that the central dark mass potential is well modeled by a point source with mass 3.6(*0.4) x 106(D/8Fcpc)3M~, consistent at the ~ 2 level u with earlier estimates based on velocity dispersion measurements. These measurement increase the dark mass density by four

53 1

Astron. NachdAN 324. No. S1 (2003)

orders of magnitude, ruling out Fermion balls as an alternative hypothesis for all supermassive black holes (Viollier 2003). Table 1 Estimates of the

central dark mass from fits to the stellar orbital motion. The reported values come from fits that solve for a common center of attraction and assume a distance of 8 kpc. Only solutions with fractional uncertainties less than 30% are listed here.

Fig. 3 Enclosed mass as a function of radius. The masses from the individual star’s orbital motion agree both with one another and the earlier estimates based on velocity dispersion measurements. The solid line shows the best fit black hole plus luminous cluster model based on the earlier measurements. The new orbital masses increases the central dark mass density by 4 orders of magnitude, dramatically strengthening the case for a central supermassive

Star

Mass (106M,)

SO-2

4.1 f0.6

SO-16

3.0 f0.7

SO-19

3.4 f 0.9

Average

3.6 & 0.4

black hole.

For SO-2,the addition of radial velocity measurements also breaks the ambiguity in the inclination angle, i. With the proper motion data alone, only the absolute value of the inclination angle can be determined, leaving the questions of the direction of revolution and whether the star is located behind the black hole at periapse (closest approach) unresolved. Our radial velocity measurements indicate a negative inclination angle and consequently that SO-2 is both counter-revolving against the Galaxy and behind the black hole at the time of periapse. The improved location of the center of attraction from the orbital analysis results in a minimum offset of SO-2 from the black hole in the plane of the sky of 14 f 2 milli-arsec, which is significantly larger than the expected Einstein radius (RE = 0.42 milli-arcsec for an assumed distance behind the black hole of 100 AU) and therefore makes gravitational lensing a negligible effect (Wardle & Yusef-Zadeh 1992;Alexander & Loeb 2001). In principle, the addition of radial velocities to the study of SO-2’s dynamics allows the distance to the Galactic Center, R,, to be a free parameter in the orbital fits (Salim & Gould 1999). The measurements, however, were obtained just 30 days after the star’s closest approach to the black hole when the radial velocity was changing very rapidly (see Figure 4). While the current radial velocity and proper motion data set constrains M / R O 3very effectively (-15% uncertainty), it does not yet produce a meaningful measurement of R,. Nonetheless, as Figure 4 shows, the radial velocities from the currently allowed orbits quickly diverge, producing a spread of a few hundred km/s in one year. Within the next few years, the orbital fits based on both proper motions and additional radial velocity measurements should provide the most direct and precise estimate of the distance to the Galactic Center, making it a fundamental rung in the cosmic distance ladder. N

3.2

Stellar Astrophysics

The detection of absorption lines in SO-2 allows us to sort out the spectral classification ambiguities present when only photometric information is available and to determine if this star’s photosphere has been altered as a result of its close proximity to the central black hole (Ghez et al. 2003a). The average brightness at 2.2 ,urn for SO-2 is K 13.9 mag and there is no evidence of brightening after periapse passage (Ghez et al. 2003b). With a distance of 8.0 kpc and K-band extinction of 3.3 mag (Rieke, Rieke, & Paul 1989), the

-

532

A. M. Ghez et al.: Stellar Orhits

. . .

I

'

1

.

.

'

I

,

'

'

.

I

Date (years)

Fig. 4 The measured radial velocity along with the predicted radial velocities. The solid curve comes from the best fit orbit and the dotted curves display the range for the orbital solutions allowed with the present data sets. Adapted from Ghez et al. (2003a)

2.2 pm brightness of SO-2 implies that, if it is an ordinary star unaltered by its environment, it could either be an 0 9 main-sequence star or a K5 giant star; all supergiants are ruled out as they are too bright by at least 2 magnitudes in the K bandpass. Kleinmann and Hall (1986) provide a 2.0 - 2.5 pm spectral atlas of late-type stars that demonstrates that if SO-2 is a K5 giant star, then it should have deep CO absorption lines, which definitively were not detected in either this experiment or our earlier experiment reported by Gezari et al. (2002). In contrast, the spectral atlas of 180 0 and B stars constructed by Hanson, Conti and Rieke (1996) shows that an 0 9 main sequence star both lacks the CO absorption and has Br 7 and He I (2.1126 pm) consistent with the observed values. Furthermore, stars earlier than 0 8 in this comparison sample show NIII (2.115 pm) in emission and He I1 (2.1885 pm) in absorption above our 3 CT thresholds; the lack of photospheric He I (2.058 pm) absorption does not provide any additional constraints. Similarly, dwarf B-type stars later than BO have absorption-line equivalent widths that are too large. Together, the photometry and absorption line-equivalent widths permit dwarf spectral types ranging from 0 8 to BO. Likewise, the rotational velocity of 224 km/s is reasonable for this range (Gatheier, Lamers, & Snow 1981). SO-2, therefore, appears to have a spectral type, and hence effective temperature (-30,000 K), as well as luminosity (-lo3 L a ) that are consistent with a main sequence star having a mass of -15 M , and an age < 10 Myr. It is challenging to explain the presence of such a young star in close proximity to a supermassive black hole. Assuming that the black hole has not significantly affected SO-2's appearance or evolution, SO-2 must be younger than 10 Myr and thus formed relatively recently. If it has not experienced significant orbital evolution, its apoapse distance of 1900 AU implies that star formation is possible in spite of the tremendous tidal forces presented by the black hole, which is highly unlikely. If the star formed at larger distances from the black hole and migrated inward, then the migration would have to be through a very efficient process. Current understanding of the distribution of stars, however, does not permit such efficient migration. This problem is similar to that raised by the He I emission-line stars (e.g., Sanders 1992,1998; Morris 1993, Moms et al. 1999; Gerhard 2001; Kim & Moms 2002), which are also counter-revolving against the Galaxy (Genzel et a]. 1997), but amplifies it with a distance from the black hole that is an order of magnitude smaller. An alternative explanation for SO-2's hot photosphere is that it may be significantly altered by its environment. While its periapse passage is too large for it to be tidally heated by the black hole as explored by Alexander & Morris (2003), it may be affected by the high stellar densities found in this region. On the one hand, the high stellar densities might allow SO-2 to be an older giant star that has had its outer atmosphere stripped through collisions; however, to generate the necessary luminosity, significant external heating is required (Alexander 1999). On the other hand, high stellar densities might lead an unliely capture of a component in a massive binary star system (Gould & Quillen 2003) or a

Astron. Nachr./AN 324, No. S1 (2003)

533

cascade of merger events (Lee 1996), which which would allow SO-2’s formation process to have begun more than 10 Myr ago. However a large number of collisions would have had to occur to provide the necessary lifetime to bring it in from sufficiently large radii. More exotically, it could be a ”reborn” star, which occurs as the product of a merger of a stellar remnant with a normal star. None of these possibilities is altogether satisfactory, leaving the Sgr A* cluster stars as a paradox of apparent youth in the vicinity of a supermassive black hole. Acknowledgements This work has been supported by the National Science Foundation through the individual grant AST99-88397 and the Science and Technology Center for Adaptive Optics, managed by the University of California at Santa Cruz under Cooperative Agreement No. AST - 9876783. The W.M. Keck Observatory is operated as a scientific partnership among the California Institute of Technology, the University of California and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W.M. Keck Foundation.

References Alexander, T. 1999, ApJ, 527,835 Alexander, T., & Loeb, A. 2001, ApJ, 551,223 Alexander, T., & Moms, M. 2002, in prep Eckart, A,, & Genzel, R. 1997, MNRAS, 284,576 Eckart, A., Genzel, R., Ott, T., & Schodel, R. 2002, MNRAS, 331,917 Genzel, R., Eckart, A., Ott, T., & Eisenhauer, F. 1997, MNRAS, 291,219 Genzel, R., Pichon, C., Eckart, A., Gerhard, 0. E., Ott, T., 2000, MRAS, 317,348 Gerhard, 0.2001, ApJ, 546, L39 Gezari, S., Ghez, A. M., Becklin, E. E., Larkin, J., McLean, I. S., Moms, M. 2002, ApJ, 576,790 Ghez, A. M. 2003, Carnegie Observatories Astrophysics Series, Vol. 1: Coevolution of Black Holes and Galaxies, ed. L. C. Ho (Cambridge: Cambridge Univ. Press) Ghez, A. M. et al. 2003a, ApJLett, in press (astro-pW030229). Ghez, A. M., Hornstein, S., Salim, S., Tanner, A., Morris, M., and Becklin, E. E. 2003b. in prep Ghez, A. M., Klein, B. C., Moms, M., & Becklin, E. E. 1998, ApJ, 509, 678 Ghez, A. M., Morris, M., Becklin, E. E., Tanner, A., & Kremenek, T. 2000, Nature, 407, 349 Gould, A,, & Quillen, A. 2003, ApJ, submitted (astro-pW0302437) Hanson, M. M., Conti, P. S., & Rieke, M. J. 1996, ApJS, 107,281 Kim, S. S., & Moms, M. 2002, ApJ, in press Kleinmann, S. G., & Hall, D. N. B 1986, ApJS, 62, 501 Lee, H. M., 1996, IAU 169,215 Lo, K. Y., Backer, D. C., Ekers, R. D., Kellermann, K. I., Reid, M., &Moran, J. M. 1985, Nature, 315, 124 Maoz, E. 1998, ApJ, 494, 181L Matthews, K. et al. 2003, PASP, in prep Morris, M., 1993, ApJ, 408,496 Morris, M., Ghez, A. M., Becklin, E. E. 1999, Adv. Spa. Res., 23,959 Munyaneza, F., Viollier, R. D. 2002, ApJ, 564, 274 Reid, M. J. 1993, ARA&A, 31, 345 Ried, M. J. 2003, this proceedings Reid, M. J., Readhead, A. C. S., Vermeulen, R. C., Treuhaft, R. N. 1999, ApJ, 524,816 Reid, M. J., Menten, K. M., Genzel, R., Ott, T., Schodel, R., & Eckart, A. 2003, ApJ, submitted Rieke, G. H., Rieke, M. J., & Paul, A. E. 1989, ApJ, 336,752 Salim, S., & Gould, A. 1999, ApJ, 523,633 Sanders, R. H. 1992, Nature, 359, 131 Sanders, R. H. 1998, MNRAS, 294,35 Schodel, R. et al. 2002, Nature, 419,694 Tsiklauri, D., Viollier, R. D. 1998, ApJ, 500, 591 Wardle, M. & Yusef-Zadeh, F. 1992, ApJ, 387, L65

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 535-541 (2003)/ DO1 10.1002/asna.200385048

The Galactic Center stellar cluster: The central arcsecond* R. Schodel**I , R. Genzel'.2,T, Ott', and A. Eckart3 Max-Planck-Institutfur extraterrestrische Physik, GiessenbachstraBe,Garching, Germany Department of Physics, University of California, Berkeley, CA 94720 LPhysikalischesInstitut, Universitat zu Koln, Ziilpicher StraRe , Koln, Germany

Key words Galactic Center, Sagittarius A*, stellar dynamics, black hole PACS 04A25

With 10 years of high-resolution imaging data now available on the stellar cluster in the Galactic Center, we analyze the dynamics of the stars at projected distances5 1.2 " from the central black hole candidate Sagittarius A* (Sgr A*). We find evidence for radial anisotropy of the cluster of stars surrounding Sgr A*. We conlidfind accelerated motion for 6 stars, with 4 stars having passed the pericenter of their orbits during the observed time span. We calculatedkonstrained the orbital parameters of these stars. All orbits have moderate to high eccentricities. The center of acceleration coincides with the radio position of Sgr A*. From the orbit of the star S2, currently the most tightly constrained one, we determine the mass of Sgr A* to 3.3 & 0.7 x 106Ma and its position to 2.0 & 2.4 mas East and 2.7 =k 4.5 mas South of the nominal radio position. The data provide compelling evidence that Sgr A* is a single supermassive black hole.

1 Introduction Because of its proximity, the center of the Milky Way offers the unique opportunity to study phenomena in detail (on scales << 1 pc) that are generally thought to occur in galactic nuclei. With the enigmatic radio and X-ray source Sagittarius A* (Sgr A*), our galactic center (GC) harbours a prime candidate for a supermassive black hole (for reviews on the GC see e.g. Genzel, Hollenbach, and Townes 1994; Morris and Serabyn 1996; Melia and Falcke 200 1 ). Near-infrared (NIR) imaging observations with speckle or adaptive optics (AO) techniques allow resolving the central stellar cluster with subarcsecond resolution. Since 1991, such observations were regularly carried out with the MPE SHARP (Hofmann et al. 1995) NIR speckle camera at the ESO NTT in La Silla, Chile (e.g. Eckart et al. 1992; Eckart et al. 1995). Stars are ideal test particles for measuring the gravitational potential because they are not subject to forces such as winds or magnetic fields. The first proper motion measurements on the stars in the central few arcseconds by Eckart and Genzel(l997) showed that the gravitational potential in the GC is indeed dominated by a point mass of -3 million solar masses. These results were confirmed with an independent proper motion study by Ghez et al. (1998), who used the 10m-class Keck telescope. Subsequently, the first accelerations of stars near Sgr A* were found by Ghez et al. (2000) and Eckart et al. (2002). These observations set even tighter demands on the density of the central dark mass than the previously measured velocity dispersions and allowed us to constrain the position of the dark mass via the projected acceleration vectors. * Based on observations at the Very Large Telescope (VLT) of the European Southern Observatory (ESO) on Paranal in Chile * * Corresponding author: e-mail:

[email protected]: 4 9 89 3oooO 3837, Fax: 4 9 89 30000 3490 @ 2003 WILEY-VCH Verlag GmbH 81 Ca. KGaA. Weinheim

R.Schodel et al.: Dynamics in the central arcsecond

536

-

Here we report on an extensive analysis of the proper motions of stars within 1.2” of Sgr A*, based on ten years of high resolution observations. We put special emphasis on the bright (K- 14) star S2 close to S g r A*. Observations made in spring/summer 2002 during commissioning and science verification of the new NIR camera and adaptive optics system CONICALVAOS (“NACO’) at the ESO VLT on El Paranal, Chile, allowed determining a unique keplerian orbit for that star. The orbital parameters put very tight constraints on the position and concentration of the central dark mass.

2 Observations and data reduction In spring 2002, the new NIR camera and adaptive optics system CONICA/NAOS (“NACO) was commissioned at the unit telescope 4 (Yepun) of the ESO 8m-class VLT on El Paranal, Chile. With its unique near-infrared wavefront sensor, this instrument is ideally suited for adaptive optics observations of the Galactic Center. While optical wavefront sensing can only be performed on a relatively weak guiding star 30” away from Sgr A*, the bright K 6.5mag supergiant IRS 7 at 5.5” from Sgr A* can be used in a straightforward manner for wavefront corrections with NAOS. Reaching Strehl ratios of up to 50%, the NACO commissioninglscience verification observations of the GC provide the deepest images of that region up to now. They also bring two extremely valuable contribitions to our proper motion program: First, the large field of view of CONICA enable us to use 7 SiO maser stars for establishing an accurate astrometry relative to Sgr A* (Reid et al. 2003). Second, the observations cover the pericenter passage of the star S2 around Sgr A* with a tightly sampled time series. For the present work we compiled GC observations from three different data sets: The NACO commissioningkience verification data from 2002, the Gemini North observatory Galactic Center Demonstration Science Data Set from the year 2000, and observations with the MPE-built NIR speckle imaging camera SHARP at the ESO NTT in La Silla, Chile, carried out between 1992 and 2002 (e.g. Eckart et al. 1995; Eckart et al. 2002). Standard data reduction procedures, i.e. sky subtraction, deadhad pixel masking, and flat-fielding, were applied to all the imaging data. From the SHARP speckle imaging data we selected a few hundred frames of highest quality for each epoch, by a mixed automatic/manual process. The selection criterion was that the first diffraction ring around the dominant speckles of the brightest stars must be clearly visible in the speckle frames. Combining these frames of highest quality resulted in SSA images with Strehl ratios > 30%. Iterative blind deconvolution (IBD, Jeffries and Christou 1993) was applied to the selected SHARP imaging data. As implementation of IBD, we used the publicly available IDAC program code, developed at Steward Observatory by Matt Chesalka and Keith Hege (based on the earlier Fortran Blind Deconvolution code - IDA - developed by Stuart Jefferies and Julian Christou). Gemini and NACO images were deconvolved with a Lucy-Richardson deconvolution. We obtained our final maps after restoring the deconvolved images with a beam of -100 mas (for SHARP and Gemini images) and w 60 mas (for NACO images) FWHM. For more details on data reduction and observations see Schodel et al. (2003) and Genzel et al. (2003).

-

-

-

3 Astrometry and proper motions Reid et al. (2003) combined several epochs of VLANLBA data to measure the positions and proper motions of SiO maser stars in the GC. They determined the position of Sgr A* in NIR NACO images with an accuracy of 10 mas. Ott et al. (2003) used their results to obtain precise positions and proper motions for -1000 stars within -10” of Sgr A*. For the present work we established the astrometry via the positions and proper motions of 9 sources from the Ott et al. (2003) list. Errors on the stellar positions were determined by quadratically combining the error from measuring the (pixel) position of the stars in our maps with the error from the transformation into the radio astrometric system.

Astron. Nachr./AN 324, No. S 1 (2003)

537

We determined proper motions by linear least square fits to the time series of stellar positions. Errors on the measured projected stellar velocities were generally < 15%. We found 6 stars with significant acceleration (deviation of > 3a from linear motion). Three of them, S l , S2, and S8 are well known from previous publications (Ghez et al. 2000; Eckart et al. 2002; Schodel et al, 2002). The other three sources, S12, S13, and S14, are fainter ( K 2 15) sources, the proper motions of which could only be disentangled from the high confusion central stellar cluster with a sufficiently large data base such as presented here (see also Schodel et al. 2003).

4 Radial anisotropy We examined the Sgr A* stellar cluster using Y T R = (v? - v g ) / w zas anisotropy estimator, where v is the proper motion velocity of a star, and VT and U R its projected tangential and radial components. A value of +1 signifies projected tangential motion, -1 projected radial motion of a star. The properties of the anisotropy parameter T T R are discussed in detail in Genzel et al. (2000). They show that an intrinsic threedimensional radiavtangential anisotropy is reflected in the properties of the two-dimensional anisotropy estimator Y T R . In Figure 1 we show a histogram of the parameter Y T R for different sub-samples (distance to Sgr A* 5 0.6", 5 LO", and 5 1.2") of our proper motion data for the epoch 2002.7. The projected velocities of the accelerated stars at this epoch were estimated by linear fits to sufficiently short parts of their trajectories. Repeating the anisotropy analysis for the epoch 1995.5 (where some of the stars had significantly different positions and velocities) does not change the distribution of counts in the histrograms significantly. The number of stars on projected radial orbits is 2 - 3a above the number of stars on projected tangential orbits (taking Poisson errors). The number of stars on projected tangential orbits decreases significantly with decreasing distance to Sgr A*. More proper motion data are needed in order to settle the question of anisotropy, but the present analysis presents a very intriguing result. Should the radial anisotropy of the Sgr A* cluster indeed be proven to be true with the larger proper motion samples expected from future observations, theoretical and modeling efforts will be needed to understand this property of the Sgr A* stellar cluster. As a bottom line, we want to point out that the general distribution of the anisotropy parameter definitely excludes a tangentially anisotropic cluster. A significant tangential anisotropy would be expected in systems with a binary black hole, where stars on radial orbits would be ejected or destroyed preferentially (see e.g. Gebhardt et al. 2002).

5

Stellar orbits

The NACO GC observations in springhmmer 2002 covered the pericenter passage of the star S2 around Sgr A* in a tightly sampled time series. Combining these observations with SHARP imaging data since 1992 (taken from Ott et al. 2003), Schodel et al. (2002) determined a unique keplerian orbit for S2. In the left panel of Figure 2 we compare the orbit of S2 of Schodel et al. (2002) with the orbit of S2 as determined in the present work. There are three important differences between the two analyses (see Schodel et al. 2003): ( 1 ) Here, the SHARP positions were obtained with different data reduction and analysis techniques (from a comparison with Ott et al. 2003 we estimated an overall systematic error of -3 mas). (2) Schodel et al. (2002) measured the positions of S2 from one final shift-and-add image for each NACO epoch and estimated the errors conservatively. Here, the S2 position for each NACO observing epoch results from measurements on several tens of individual short-exposure NACO images, with the standard deviation taken as error. (3) We treated the projected position of the focus of the elliptical orbit as a free parameter in the fit (see Schodel et al. 2003). The two analyses compare very well, with the determined orbital parameters agreeing within the errors. In our present analysis, we obtain a central mass of 3.3 f 0.7 x 106Ma. The position of the acceleration

0

I

5-

;: L 0 O - . D

5

“7-

-1 5

center is offset a mere 2.0 f2.4 mas East and 2.7 k 4.5 mas South of the nominal radio position of Sgr A*, i.e. clearly within the error circle of the radio measurement. This strongly supports the assumption that the dark mass is indeed coincident with Sgr A*. The orbit has an eccentricity of 0.87 f0.02, an inclination of 45.7 f 2.6 degrees, a period of 15.7 f 0.74 years, a semi-major axis of 4.54 f0.27 mpc, and a pericenter distance of 0.59 f0.10 mpc. Significant sections of the orbits were observed as well in the case of the stars S12 (pericenter passage in 1995.3) and S14 (pericenter passage in 1999.9). However, the constraints on their orbits from our data are not very tight. S14 is identical with the source SO-16 of A. Ghez et al. (priv. c o r n . ) , who first determined an orbit for this source. S14 is on an extremely eccentric (e = 0.97 f0.05) and highly inclined (i > SOdeg) orbit and approaches Sgr A* to within -0.4 mpc (S2: 0.6 mpc). In principle, its orbit would allow to constrain the central mass distribution even tighter than S2. Unfortunately, the uncertainty in the orbital parameters of S14 resulting from our data is too high for this purpose. The orbital segments observed for the stars S1 (pericenter passage around 1999/2000), S8, and S13 are too small for determining a unique set of parameters for them, but we constrained them by using fixed values for the inclination angle (see Schodel et al. 2003). Approximate values of the inclination of the orbital planes could be estimated from the measured acceleration of the stars and the well known mass of the central dark object (see Figure 3). We plot all analyzed six orbits in the right hand panel of Figure 2. All analyzed orbits have moderate to high eccentricities. Future measurements of orbital eccentricities of more stars near Sgr A* will allow testing for anisotropy of the central cluster (see Schodel et al. 2003).

6 Nature of the enclosed mass With the measured proper motions within 1.2” of Sgr A* we calculated Leonard-Merritt (LM, Leonard and Merritt 1989) estimates of the enclosed mass. In order to take the strongly variable velocity of the 6 stars with significant acceleration into account, we produced various velocity lists for the analysis, where we estimated the projected velocity of these stars at different epochs. From the diffrent lists, we obtain an average LM mass of 3.4 f 0.5 x 1O6M@(for details see Schodel et al. 2003). This agrees well with the mass estimate from the orbit of S2.

Astron. Nachr./AN 324, No. S1 (2003)

i

539

1997.54

0

0.5

Right Ascension

["I

Fig. 2 Left panel: The orbit of S2 (black) as determined by Schodel et al. (2002) compared with the orbit as determined in the present work (green). The black cross and circle denote the radio position of Sgr A* and its errors. The red cross designates the position of the focus of the orbit and its error resulting from the present analysis. Right panel: Currently, we can determinekonstrain the orbits of 6 stars near Sgr A*: S1, S2, S8, S12, ,513, and S14. all orbits have

moderate to high eccentricities. Figure 3 is a plot of the measured enclosed mass against distance from Sgr A*, in close analogy to Figure 17 of Genzel et al. (2000) and Figure 3 of Schodel et al. (2002). The data show that the central mass distribution is remarkably well described by the potential of a point mass over 3 orders of magnitude in spatial scale, from 0.8 light days to 2 light years. The contribution of the extended stellar cluster around Sgr A* to the total mass cannot be more than mostly a few hundred solar masses within the peri-center distance of S2 (Mouawad et al. 2003). Fitting a model composed of a point mass plus the visible outer stellar cluster with a core radius of 0.34 pc and a power-law slope of a = 1.8 to the data gives a value of 2.9 0.2 x 106Ma for the central dark mass. This agrees within the errors with the LM mass estimate of the innermost stars and with the masses calculated from the orbital parameters of S2 and S12. It is higher than the 2.6 0.2 x 106Ma given by Schodel et al. (2002), but the two values agree within their errors. The main differences of the present analysis to Schodel et al. (2002) are: (1) The error of the mass estimate from the orbit of S2 has been reduced by taking the position of the orbital focus explicitly into account. (2) The innermost LM mass estimate of Schodel et al. (2002) was based on the Ott et al. (2003) data. It has been replaced by the LM mass estimate from the present work, which is based on a more abundant 1" of Sgr A*. (3) The LM mass estimates in Figure 3 of Schodel et al. data base in the region within (2002) and Figure 17 of Genzel et al. (2000) were corrected downward by 510% because they assumed a power-law slope of a = 1.8 for the stellar cluster in the innermost few arcseconds. Here, we use a power-law slope of a M 1.4 for the stellar cusp around Sgr A* (Genzel et al. 2003). This means that the LM mass estimates have previously been underestimated by 10%. The orbit of S2 places very tight constraints on the distribution of the central dark mass: If the central point mass were replaced by a Plummer model cluster of dark astrophysical objects, its central mass density would have to exceed 2 . 2 ~ 1 0 ' ~ M ~ p calmost - ~ , 5 orders of magnitude greater than previous estimates

+

+

-

-

540

R. Schodel et al.: Dvnamics in the central arcsecond

(Ghez et al. 1998,2000; Genzel et al. 2000). The lifetime of such a hypothetical cluster would be < 10' years (Maoz 1998). An alternative model to supermassive black holes in galactic nuclei are balls of heavy, degenerate neutrinos (Tsiklauri and Viollier 1998; Munyaneza and Viollier 2002). In order to explain the whole mass range of dark central objects in galaxies with such a model, the neutrino mass cannot be higher than 17keV (Melia and Falcke 2001). However, the orbital parameters of S2 would demand a neutrino mass of > 5OkeV in the case of the dark mass in the Galactic Center. The only dark particle matter explanation that cannot b e ruled out by the present data is a ball of bosons (Torres et al. 2000). However, it would be hard to understand how the bosons first manage to reach such a high concentration, and then avoid forming a black hole by accretion of the abundant gas and dust in the GC. We therefore conclude that the most probable form of the dark mass at the center of the Milky Way is a single, supermassive black hole.

Acknowledgements We like to thank the ESO NTT team for their help and support during ten years of observations with the SHARP guest instrument. We thank the NAOS and CONICA team members for their hard work, as well as the staff of El Paranal and the Garching Data Management Division for their support of the commissioning and science verification. Based on observations obtained at the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under a cooperative agreement with the NSF on behalf of the Gemini partnership: the National Science Foundation (United States), the Particle Physics and Astronomy Research Council (United Kingdom), the National Research Council (Canada), CONICYT (Chile), the Australian Research Council (Australia),CNPq (Brazil) and CONICET (Argentina).

References Chakrabarty, D. and Saha, P. 2001, AJ, 122,232 Eckart, A., Genzel, R., Krabbe, A., Hofmann, R., van der Werf, P.P., Drapatz, S. 1992, Nature, 355,526 Eckart, A,, Genzel, R., Hofmann, R., Sams, B.J., and Tacconi-Garman L.E. 1995, ApJ, 445, L23 Eckart, A. and Genzel, R. 1997, MNRAS, 284,576 Eckart, A., Genzel, R., Ott, T., and Schodel, R. 2002, MNRAS, 331,917 Gebhardt, K., Richstone, D., Tremaine, S., Lauer, T.R., Bender, R. et al. 2002,astro-ph/O209483 Genzel, R. and Townes, C.H. 1987, ARA&A, 25,377 Genzel, R., Hollenbach, D., and Townes, C.H. 1994, Rep. Prog. Phys., 57,417 Genzel, R., Thatte, N., Krabbe, A,, Kroker, H., and Tacconi-Garman, L.E. 1996, ApJ, 472, 153 Genzel, R., Pichon, C., Eckart, A., Gerhard, O.E., and Ott, T. 2000, MNRAS, 317, 348 Genzel, R., Hofmann, R., Lehnert, M., Ott, T., Schijdel, R., Eckatt, A,, Alexander, T. et al. 2003, in press Ghez, A., Klein, B.L., Moms, M., and Beckhn, E.E. 1998, ApJ, 509,678 Ghez, A., Morris, M., Becklin, E.E., Tanner, A., and Kremenek, T. 2000, Nature, 407,349 Hofmann, R., Brandl, B., Eckart, A,, Eisenhauer, F., Tacconi-Garman, L.E. 1995, Proc. SPIE, 2475, 192 Jeffries, S.M. and Christou, J.C. 1993, ApJ, 415,862 Maoz, E. 1998, ApJ, 494, L181 Melia, F. and Falcke, H. 2001, ARA&A, 39,309 Morris, M. and Serabyn, E. 1996, ARA&A, 34,645 Mouawad, N., Eckart, A., Pfalzner, S., Straubmeier, C., Spurzem, R., Genzel, R., Ott, T.,and Schodel, R. 2003, in preparation. Munyaneza, F. and Viollier, R.D. 2002, ApJ, 564,274 Ott, T., Genzel, R., Scbdel, R., and Eckart, A. 2003, in preparation Reid, M.J., Menten, K.M., Genzel, R., Ott, T., Schodel, R., and Eckart, A. 2003, ApJ, in press Schodel, R., Ott, T., Genzel, R., Hofmann, R., Lehnert, M., Eckart, A,, et al. 2002, Nature, 419,694 Schodel, R., Genzel, R., Ott,T., Eckart, A., Mouawad, N., and Alexander, T. 2003, m press Tsiklaun, D. and Viollier, R.D. 1998, ApJ, 500,591

Astron. Nachr./AN 324, No. S 1 (2003)

54 1

radius (light hours) 1o2

10'

1o3

1o5

1o4

n

cn

2.87(+0.15) xi06 M(sun) point mass plus visible star cluster

a,

cn cn

!!

1 o7

kj

0

cn cn cn

W

! ! U a, cn 0

1o6

dark cluster

7

0

c

a,

0.001

0.01

0.1

1

10

radius (parsec) Fig. 3 Mass distribution in the Galactic Center assuming an 8 kpc distance (Reid et al. 2003). The filled black circle denotes the mass derived from the orbit of S2, the red filled circle the mass derived from the orbit of S12, and the purple circle the mass derived from the orbit of S14. Filled dark green triangles denote Leonard-Merritt projected mass estimators from the present work (at 0.025 pc) and from a new NTT proper motion data set by Ott et al. (2003), separating late and early type stars, and correcting for the volume bias determined from Monte Carlo modeling of theoretical clusters and assuming a central density profile with a power-law slope of a = 1.37 (Genzel et al. 2003). An open down-pointing triangle denotes the Bahcall-Tremaine mass estimate obtained from Keck proper motions (Ghez et al. 1998). Light-blue, filled rectangles are mass estimates from a parameterized kanS-equdtiOn model, including anisotropy and distinguishing between late and early type stars (Genzel et al. 2000). Open circles are mass estimates from a parameterized Jeans-equation model of the radial velocities of late type stars, assuming isotropy (Genzel et al. 1996). Open red rectangles denote mass estimates from a non-parametric, maximum likelihood model, assuming isotropy and combining late and early type stars (Chakrabarty and Saha 2001). The different statistical estimates (in part using the same or similar data) agree within their uncertainties but the variations show the sensitivity to the input assumptions. In contrast, the orbital technique for S2/S12 and S14 is much simpler and less affected by the assumptions. Green letter "G" points denote mass estimates obtained from Doppler motions of gas (Genzel and Townes 1987). The blue continuous curve is the overall best fit model to all data. It is the sum of a 2.87 & 0.15 x l o 6 Ma point mass, plus the visible outer stellar cluster of central density 3.6 x 106Mopc-3, core radius 0.34 pc and power-law index a = 1.8. The grey long dash-short dash curve shows the same stellar cluster separately, but for a infinitely small core (i.e. a 'cusp'). The red dashed curve is the sum of the stellar cluster, plus a Plummer model of a hypothetical very compact (core radius -0.00019 pc) dark cluster of central density 2.2 x 10*7Mopcp3.

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S1.543-549 (2003) / DO1 10.1002/asna.200385080

Stellar Dynamics in the Galactic Center: 1000 Stars in 100 Nights Thomas Ott”, Reinhard Gemel’, Andreas Eckart2, and Rainer Schodel’ Max-Planck-Institut fkr extraterrestrische Physik, Giessenbachstr., 85748 Garching, Germany 1 . Physikalisches Institut, Universitat zu Koln, Ziilpicher Str. 77,50937 Koln, Germany

’

Key words galaxy: center, stars: infrared, stars: dynamics

Abstract. We present the results of a largely automatized re-analysis of all near-infrared imaging data obtained with the MPE speckle camera SHARP I at the NTT. We were able to increase the number of measured proper motions in the Galactic Center stellar cluster by about an order of magnitude in comparison to

previous work done on this issue. We have made astrometric positions and projected sky velocities available for about 1000 stars in the central parsec of our galaxy. We also present radial velocity measurements of 100 Stars which were obtained with the integral field spectrograph 3D at the 2.2m ESO/MPG telescope located on La Silla, Chile. Finally, using the Gemini North science verification narrow band data enables us to distinguish the stellar populations to fainter magnitudes. Using this new and extensive dataset, we present an analysis of the dynamical properties of the late-type and early-type stellar populations. While the old population is relaxed and their movements isotropic, the young stars experience noticeable anisotropy with mainly tangential proper motions.

1 Introduction Because of its proximity (distance 8 kpc, Reid 1993), the center of the Milky Way is a unique laboratory for studying the physical processes in galactic nuclei. In particular, the Galactic Center offers the unique opportunity for investigating stars and gas in the immediate vicinity of a supermassive black hole, at a level of detail that will not be accessible in any other galactic nucleus in the foreseeable future. There are several different stellar populations/components in the central parsec (for a review see Genzel 2001). The stellar mass and the near-IR light at K> 13 is dominated by red giants in the old (1-10 Gyrj component of the nuclear star cluster. A group of about two dozen luminous, blue supergiants (‘He1 emission line stars’) strongly affects the near-IR maps at the bright end (K- 9- 12j, and probably indicates recent formation of massive stars within the last 2-7 Myrs (Forrest et al. 1987, Allen, Hyland & Hillier 1990, Krabbe et al. 1991, 1995, Tamblyn et al. 1996, Blum et al. 1996, Paumard et al. 2001). A number of bright (K- 10 - 12) asymptotic giant branch (AGB) stars sample an intermediate mass, intermediate age component (2100 Myr, Lebofsky & Rieke 1987, Krabbe et al. 1995, Blum et al. 1996). Finally there is a group of dust embedded stars with near-featureless near-IR spectra (Becklin et al. 1978, Krabbe et al. 1995, Genzel et al. 1996), many of which are associated with the gaseous mini-spiral. Their nature is uncertain. The mean stellar velocities (or velocity dispersions) follow a KepIer law ((v ’) 0; R -l) from 0.1” to 2 20”, and provide compelling evidence for the presence of a central compact mass (Genzel et al. 1996, 97, 2000; Eckart and Genzel 1996,97; Ghez et al. 1998). Overall the stellar velocities are consistent with an isotropic velocity field but the He1 emission line stars appear to be preferentially on tangential orbits (Genzel et al. 2000). N

* Corresponding author: e-mail: ott@mpe,mpg.de,Phone: 4 9 89 3oooO 3276, Fax: 4 9 89 3oooO 3390 @ 2001 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheirn

T. Ott et al.: Stellar Dvnamics in the Galactic Center

544

2 Observations and data reduction The observations presented in this contribution were carried out using a variety of telescopes and instrumentation: The proper motion study used speckle data obtained with the MPE speckle camera SHARP I at the New Technology Telescope (NTT), which is located at La Silla, Chile and operated by the European Southern Observatory (ESO). Imaging spectroscopy using the MPE 3D spectrograph was carried out at the ESOMPG 2.2m telescope on La Silla. These data provide stellar classification and radial velocities for the brightest 100 stars in the GC. Initial astrometry and stellar identifications were obtained using NAOSlCONICA at the VLT. These data were recorded during science verification of the instrument. The public Gemini SV dataset provided stellar classification down to lower magnitudes than has been possible using the 3D data. The data reduction was carried out using the MPE dpuser software package. All data were sky subtracted, flat-fielded and deadhot pixels were corrected for by interpolation of usable neighboring pixels. In the case of the SHARP I speckle data, the final diffraction limited image was created using the Simple Shiftand-Add algorithm (see Eckart & Genzel, 1997). The Gemini and NACO data needed no further treatment. In order to calibrate the wavelength scale of the 3D data, at the beginning or the end of each observing night exposures of spectral lamps (argon lamps in our case) were done. These emit a known line-spectrum and were used to measure the dispersion of the spectrograph. The wavelength scale of the instrument was then stretched such as to get a linear relationship between detector element and wavelength. Spectral calibrator stars with a known spectrum were observed at similar airmass as the galactic center. These standard stars were divided by a spectrum of the same stellar type (Kleinmann & Hall 1986) in order to remove stellar features resulting in an atmospheric transmission spectrum. The source data were then divided by this spectrum.

3 Astrometry The task of astrometry is to determine the positions of stars for each epoch of observations. Our approach was to create a “master-list” of stars using one high-quality observation, which serves to re-identify the stars in all other observations. The measured positions then have to be transformed to a common coordinate system so they can be compared and allow the determination of proper motions. 3.1 Master-list For the creation of the master-list it is desirable to have one single observation which includes the full field-of-view of the SHARP observations, in order to eliminate systematic effects due to image distortions. 20” x 20”. Since we observed the galactic center as a mosaic, the complete field covers a field of Fortunately, in May 2002 the first observations with NAOS/CONICA at the VLT were available which cover a field of 40” x 40”, which is large enough to cover all SHARP pointings. We used the NAOSICONICA data to create a first list of stars using package “SExtractor” (Bertin & Amouts, 1996). Although SExtractor was able to identify a lot of stars, omissions in the list were common in regions of high stellar density. It was necessary to create an image in which each identified star was marked and then to manually edit the list to add sources and to remove some spurious misidentifications. This initial list of stars consists of stellar positions in pixel coordinates of the CONICA frame. It is desirable, though, to transform this to physical astrometric coordinates (relative to the radio N

N

Astron. Nachr./AN 324, Yo. S1 (2003)

545

source SgrA*) in order to later derive proper motions in units of "/year or (knowing the distance to the galactic center) in k d s .

Fig. 1 Identification of masex sources in the near infrared imaging

This can be done using radio interferometry, since SgrA* is known to be a radio point source. These measurements can be done to an accuracy of a few milli-arcseconds by comparing the positions with extragalactic sources (as quasars), which is far more accurate than the positioning achievable with infrared direct imaging. In the CONICA frame, we are lucky to observe several maser sources which are both observable in the NIR and the radio (Menten et al., 1997,Reid et al., 2002). Here we used six sources (see Table 1) to calculate the plate constants of our infrared imaging. Table 1 Positions and proper motions of the maser sources relative to SgrA*

IRS 9

5.6650 f 0.0019

-6.3433 f 0.0030

IRS 7

0.0326 f 0.0030

5.5353 f 0.011

IRS 2

-3.2574 f0.0012

-6.8980 f0.0015

* 1.22 -0.91 * 0.23

SiO-B

10.4697 f0.0026

-5.8024

-0.08 f 1.30

IRS l0EE

7.6854 f 0.0010

4.2067 f0.0010

IRS 15NE

1.2197 f0.0011

11.2948

* 0.0053

* 0.0019

3.61 f 0.53

-1.13

1.72 f 0.88 -2.90

* 2.90

-2.73 f 0.28 -3.61

* 2.27

0.36 i 0.23

-2.13 3z 0.24

-1.68 f0.24

-6.04 i0.35

In order to transform the initial list of stars to this astrometric reference frame, we have calculated the centroids of the maser sources in the CONICA frame. When using six radio positions, we are able to determine image distortions up to second order. The transformation matrix was then applied to all stars in the initial list, resulting in an astrometric (relative to SgrA*) master-list of stars at the epoch of the CONICA observation.

T. Ott et al.: Stellar Dynamics in the Galactic Center

546

3.2 Identification of stars In order to determine the positions of a certain star in a given image, it is first necessary to identify this star in the image. This can be done by transforming the star’s astrometric position (2,y) (given by the masterlist) to the coordinate system (z’, y’) in the image frame. The transformation equations for the linear case are: 2’ = a2

y’ = dx

+ by + c

+ ey + f

These take into account translation, rotation, and image scale in both axes. Since it is necessary to determine six unknowns, one needs at least the positions of three stars. The greater the distance between these stars, the more reliable the solution will be. Since the SHARP I imaging was done using four mosaic settings, we needed four different triples of stars. These have been identified manually in each frame. Using the stars to Equation (1) it is then possible to transform the master-list to the current image, identifying all the stars. Of course, these positions are only accurate to within certain limits. In order to determine their positions more accurately, we have therefore searched for the maximum pixel intensity within a search radius of a few pixels. This works very well for most of the stars, only the central cluster within the central arcsecond with stars moving in excess of 500 km/s needed special treatment. Here we identified the stars manually by comparing the images of several epochs and identifying by similarity. Then we calculated the centroid ( X ,Y )of the stars in the image I ( s ,y) as follows:

As a result we then get a list in the same ordering as the master-list with the positions of the stars in the current image. Since each exposure has a different limiting magnitude G at which a reliable determination of the centroid is possible, it is in addition necessary to determine the magnitude of the faintest usable star, else the method mentioned above centers on the brightest noise peak, not resulting in a meaningful measurement of the stellar position. A stable criterion for G turned out that the brightest pixel of a star must be at least G > medzan(1mage) 0 . meddev(1mage). A reliable value for 0 was 150.

+

3.3 Transformation to astrometric coordinates When applying the steps described above for all images used in this analysis, we end up with a List of stellar positions in the detector coordinate system for each image. In order to transform them back into astrometric coordinates, a reverse transformation according to equation 1 is necessary. To do this reverse transformation more accurately and take into account image distortions of higher order, we also allow quadratic and mixed terms in the transformation matrix, which then looks like: 2’ = az

TI’= gz

+ b y + cxy

+ dz2 + ey2 + f

+ hy + i z y +j22 + ky2 + 2

(3)

This set of equations now has 12 unknowns, making it necessary to use at least six stars to solve for the unknowns. Using the six maser sources mentioned above is not possible for our data, since the SHARP I frames only have a maximum of two stars common to a single frame. Therefore we have to reference the images to the master-list given by the CONICA-observations. Our ultimate goal, though, is to derive proper motions for all the stars in our master-list. If we were to use only six stars, we would have to know

Astron. Nachr./AN 324, No. S1 (2003)

541

either their proper motions beforehand, or assume that their proper motions are close to zero. Since neither of these are known, it was necessary to use a different approach: When using a large (between 50 and 200) of stars to solve for the unknowns in Eq. (3), the set of equations is, of course, over-determined. It can be solved using a least-squares approach. The only assumption about the motions of the stars then is that their average motions are close to zero, an assumption which is justified by the isotropy of the motions of most of the stars (Genzel et al., 2000). This explains also the minimum number of 50 stars necessary, since when using less stars their average motions do not cancel out using a random selection of stars. The maximum number of stars of 200 is given by the quality of our data. If using more stars, we start using measurements of centroids of faint stars with low signal-to-noise, which introduces noise in the solution of Equation (3). The results of this approach can be improved by the following method: For each star, one can estimate a transformation error by comparing its transformed position with the position given by the master-list. Stars with large errors are caused by either spurious noisy measurements, misidentifications, or extremely fast moving stars. Equation (3) can then be solved for again neglecting those stars. From the now known astrometric positions of the stars at different epochs (ranging from March 1992 to June 2002) it is straightforward to derive proper motions. For all but two stars (S1 and S2 in the central cluster, see Schodel et al., 2002) a straight line fit was sufficient to describe their motions. 3.4

Determination of radial velocities

The stellar spectra which could be extracted from the imaging spectroscopy data constitute of two different types: Stars with line emission and stars exhibiting CO bandhead absorption. In order to derive radial velocities from these stars, we used two different approaches: 0

When using stars exhibiting line-emission, one (or several in case of P-Cygni type absorption) gaussians were fit to the emission lines giving a direct measure of the redshift of this star

Stars showing CO absorption were cross-correlated with a template spectrum and the redshifts were determined from the maximum in this cross-correlation In the former case, each emission line results in an independent measure of the star’s radial velocity. These were then averaged, taking their standard deviation as an estimate for the error. In the latter case, we used the star HD 78647 as a template, taking into account its radial velocity of 9 k d s . The overall accuracy for which radial velocities could be determined ranges between 100 k d s for faint stars and better than 30 k d s for bright stars. From these measured radial velocities, we subtracted a value of 38 k d s to account for the Earth’s movement in its orbit around the sun and the movement of the solar system with respect to the GC (local frame of rest).

4 The dynamical properties of the central cluster We are considering here the normalized angular momentum along the line of sight, JZ/JZ(max),which we define as JZ/JZ(max) = (33jy

-

(4)

YV,)/PV,,

where v,, vy and up are the R.A.-, Dec.- and total proper motion velocities of a star at (x,y) on the sky -1, 0 and +1, depending on whether the stellar orbit and at projected radius p. JZ/JZ(max)is projected on the sky is mainly counter-clockwise tangential, radial or clockwise tangential with respect to the projected radius vector from the star to SgrA*. Figure 4 shows the projected radial distribution of JZ/Jz(max)for late and early type stars, as identified from the narrow-band CO-index (Figure 2). The N

N

N

T.Ott et al.: Stellar Dvnamics in the Galactic Center

548

-0.4

9

10

11

12

13

14

15

K Fig. 2 CO-index (m(C0) = m(2.29)- m(2.26) ) as a function of K-magnitude for those stars in the proper motion sample that also have Gemini science demonstration data, narrow band maps. Stars marked with filled rectangles denote early type stars, and stars marked with filled circles denote late type stars confirmed by the 3D spectroscopic data. For K 5 15 stars with m ( C 0 )2 0.04 are identified as late type stars, while stars with m ( C 0 ) < 0.04 are identified as early type stars.

early type stars show a preponderance of tangential orbits (Figure 4 right panel). Within the central 3" 51(f10) % and 18(&6) % of the early type stars are on clockwise, and counter-clockwise, tangential orbits, in marked contrast to the random pattern of the late type stars (Figure 4 left panel). Early type stars with clockwise, tangential orbits dominate within a few arcseconds of SgrA*, and bunch up in the IRS16 complex WSE of SgrA".

. F

.' * I

laletypestarsK455

I-

g o >"

0

2

4

6

8

10

PmA.(-)

Fig. 3 Distribution of z-velocities of spectroscopic late type stars (squares) and early type stars (filled circles) as a function of Dee.-offset from SgrA*.

Fig. 4 Normalized angular momentum along the line of sight (J,/J,(max)) as a function of projected separation from SgrA' for the K< 15.5 late type stars ( m ( C 0 ) 2 0.04, left panel), and for the K514.7 (green) and K<12 (dark blue) early type stars ( m ( C 0 )< 0.04).

A similar pattern emerges from the line-of sight velocities of late and early type stars (Figure 3). Apart from an average blue-shift, the late type stars show a random (relaxed) distribution of line-of-sight velocities. In contrast, early type stars north of SgrA* are almost all blue-shifted, while stars south of SgrA* are almost all red-shifted. This is indicative of a coherent rotation pattern, with a direction opposite (counter)

Astron. NachrJAN 324,No. S1 (2003)

549

to that of the overall Galactic rotation. With the exception of 4 early type stars N 9” S W of SgrA*, the phase space distribution of 13 He1 emission line stars with 1 v, 12 80 k d s can be well fit by a disk in clockwise and counter-Galactic rotation, with its major axis along p.a. -20 to -60” and inclination 40”. We conclude that a large fraction of the K S 15 stars within a few arcseconds of SgrA* are early type stars with an unrelaxed distribution of angular momenta. The most obvious interpretation is that these stars, like the brighter He1 emission line stars, are young, massive OD3 and Wolf-Rayet stars, and that they are members of the 2-8 M y r ‘starburst’ component. stars in the ‘SgrA* cluster’ 0, 5 0.5”) show a different angular momentum The early type (‘S-) distribution (Figure 4, right panel). Keeping in mind that a few of these stars may only be projected to lie in the central arcsecond, the ‘SgrA* cluster’ stars do not show an excess of tangential orbits. On the contrary, a more detailed analysis shows an excess of radial orbits (Genzel et al. 2000, Schodel et al. 2003). 54 (+14) % of the 35 observed stars in the ‘SgrA* cluster’ are on radial orbits. These stars are probably observed near the apoapse of their orbit (roughly twice the semi-major axis), where they spend most of their time. It is therefore likely that the radial stars are tightly bound to the black hole, with semi-major axes of order 20 mpc. This statistical inference, together with the direct derivation of semi-major axes of < 10 mpc for the 6 stars in the inner cusp whose orbits were solved (Schodel et a1 2003), suggests that the SgrA* cluster as a whole is tightly bound to the black hole.

References Allen, D.A., Hyland, A.R. & Hillier, D.J. 1990 MNRAS, 244, 706 Becklin, E.E., Mathews, K., Neugebauer, G. & Willner, S.P. 1978, ApJ, 219, 121 Bertin, E. & Amouts, S. 1996, A & AS, 117, 393 Blum, R.D., Sellgren, K. & DePoy, D.L. 1996, AJ, 112, 1988 Eckart, A. & Genzel, R. 1996, Nature, 383,415 Eckart, A., & Genzel, R. 1997, MNRAS, 284,576 Forrest, W.J., Shure, M.A., Pipher, J.L. & Woodward, C.A. 1987, in The Galactic Center, AIP Conf.Proc. 155, ed.D.AmetInst.of Phys.: New York, 153 Genzel, R., Thatte, N., Krabbe, A., Kmker H. & Tacconi-Garman,L. E. 1996, ApJ, 472, 153 Genzel, R., Eckart, A,, Ott, T.& Eisenhauer, F. 1997, MNRAS, 291,219 Genzel, R., Pichon, C., Eckart, A., Gerhard, 0. & Ott, T. 2000, MNRAS, 317, 348 Genzel, R. 2001, in Dynamics of Star Clusters and the Milky Way, ASP Conf.Ser. 228, eds. S.Deiters, B.Fuchs, R.Spurzem, A h s t & R.Wielen, Astr.Soc.of the Pacific:San Francisco, 291 Ghez, A. M., Klein, B. L., Moms, M., & Becklin, E. E. 1998, ApJ, 509,678 Kleinmann, S . G., &Hall, D. N. B. 1986, ApJS, 62,501 Krabbe, A., Genzel, R., Drapatz, S. & Rotaciuc, V. 1991, ApJm 382, L19 Krabbe, A., et al. 1995, ApJ, 447, L95 Lebofsky, M.J. & Rieke, G.H. 1987, in The Galactic Center, AIP 155, ed. D.Backer, Amer.Inst.Phys: New York, 79 Menten, K., Reid, M., Eckart, A,, & Genzel, R. 1997, ApJL, 475, 1 11 Paumard, T., Maillard, J. P., Moms,M. & Rigaut, F. 2001, A&A, 366,466 Reid, M. J. 1993, ARA&A, 31,345 Reid, M. J., Menten, K. M., Genzel, R., Ott, T., Schodel, R. & Eckart, A. 2003, ApJ, submitted Schodel, R.et al. 2002, Nature, 41 9, 694 Schodel, R., Ott, T., Genzel, R. et al. 2003, ApJ, submitted Tamblyn, P., Rieke, G. H., Hanson, M. M., Close, L. M.,McCarthy, D. W. & Rieke, M. J. 1996, ApJ, 456,206

This Page Intentionally Left Blank

Astron. NachdAN324, No. S1.551-555 (2003) / DO1 10.1002/asna.200385101

A Bow Shock of Heated Dust Surrounding IRS 8 F. Rigaut I, T. R.Geballe", J.-R. Roy', and B. T. Draine2

' Gemini Observatory, 670 N. A'ohoku Pl., Hilo, HI 96720, USA

* Princeton University Observatory, Princeton, NJ 08544 USA

Key words Galaxy: center, ISM kinematics and dynamics, stars: individual (IRS 8), stars: mass-loss Abstract. High resolution images in the H and K' hands obtained by the Gemini North Telescope of the peculiar Galactic center source, IRS 8, reveal a central pointlike object enveloped in a remarkable bowshock, whose apex is located 0.2" to the northeast. The H - K' color of the how shock is considerably redder than that of the central star. A UKIRT K band spectrum reveals that the combined spectrum of the point source and how shock is nearly featureless and that no shocked line emission (e.g., from Hz) is physically associated with the bow. We interpret the how as resulting from the interaction of the envelope or wind of the central star of IRS 8 with the extension of the northern arm of Sgr A West andor the molecular ring, and its emission as coming from radiatively- and possibly shock-heated dust.

1 Introduction The Galactic center source, IRS 8, has long been one of the mystery objects of the central infrared cluster. Located 30 arcsec (1.2 pc assuming a galactocentric distance of 8 kpc) north of Sgr A* (Becklin & Neugebauer 1975), the source is well removed from the other bright 10 pm infrared sources, which are all well within the central parsec. However, it is probably still bathed in the intense and high temperature (-35,000 K) radiation field mainly produced by the cluster of hot stars within the central parsec. IRS 8 stands out at 2.2 pm and, like the ridge sources and IRS 3, it is much brighter at 10 pm than at short infrared wavelengths, even after correcting for extinction. This trait indicates that its infrared radiation arises predominantly from heated dust. Unlike IRS 3 and the ridge sources that were detected at 2.2 pm, IRS 8 was spatially resolved at 2.2 pm in a 2.5" aperture by Becklin & Neugebauer (1975). This marked it as a unique object in the Galactic center. The luminosity of the source has been estimated to be -1 x l o 5 La (Becklin et al. 1978). Lacy et al. (1979,1980) found that the bright 12.8 pm Ne II fine structure line at the position of IRS 8 has two velocity components at YLSR of f l l O lcm s-l and -10 km s-'. The redshifted component, whose intensity is enhanced near the position of IRS 8, has been interpreted by Lacy et al. (1980)as an extension of the redshifted northern arm of dust and ionized gas, whereas the narrow - 10 km s-l component is more localized at IRS 8. This suggests that the LSR velocity of the gas associated with IRS 8 is -10 km s-'. Although near diffraction-limited and diffraction-limited images of the central several arcseconds of the Galaxy have been obtained by a number of research teams, the fields that have been imaged do not include IRS 8. Thus, the large area H and K' band mapping of the Galactic center carried out at the Gemini North telescope in 2000 using the University of Hawaii's Hokupa'a adaptive optics system and QUIRC camera (as part of the Gemini science verification observations, see Rigaut et al. 2003), contains the first near diffraction-limited observations of the IRS 8 region obtained by any of the new 8-10 m class telescopes. li

Corresponding author: e-mail: [email protected],Phone:4 1 808 9742519, Fax: 4 1 808 935 9650

@ 2003 WREY-VCH Verlag GmbH & Co. KGaA, Weinheim

F. Rigaut et al.: Bow Shock Around IRS 8

552

-2

Fig. 1 Gemini A 0 image of the IRS 8 region at 2.3 pm. The field of view is approximately 7”x.S’. North is up and east to the left.

-1

0 arcsec

1

2

Fig. 2 Contour plot of the IRS 8 region at 2.3 pm with stars removed. The cross corresponds to the location of the point source in IRS 8.

2 Observations and Results Figure 1 is an image of the IRS 8 region in a narrow band filter centered at 2.3 pm. A bow shock is seen at IRS 8 surrounding a bright pointlike object, presumably a luminous star. The apex of the bow is located 0.22” (-0.01 pc) to the northeast of the central object. The central pointlike object accounts for approximately 10 percent of the flux from IRS 8 at K’. Radio images of the same region (e.g., Yusef-Zadeh, Moms, & Ekers 1990) show a bright and extended clump of ionized gas at IRS 8, strongly suggesting that IRS 8 contains a hot star. Broad band images at H and K’ also show the bow shock. The H - K‘ color of the bow is much redder than both the central star of IRS 8 and other nearby stars, indicating that its emission is from much cooler material. Figure 2 is a contour map of IRS 8 and its surroundings, generated from the narrow band 2.3pm image. In addition to the bow shock, a very faint arc of emission can be seen just outside of the eastern part of the bow. Additional emission can be seen approximately 1.5’‘ to the west. We have recently obtained K band spectra of IRS 8 and its surroundings at the United Kingdom Infrared Telescope (UKIRT), in order to search for line emission specifically associated with the bow-like feature. The data were obtained with the facility spectrograph CGS4 using a 0.61” wide slit (oriented EW) and 0.61” square pixels, both of which are too large to isolate the point source from the bow. Spectra of IRS 8 (the point source and the bow shock) are shown in Fig. 3. The raw sky-subtracted spectrum contains considerable extended continuum and line emission. When this is removed by subtracting the mean spectrum immediately adjacent to IRS 8 to better isolate the combined spectrum of the bow and stellar object, the result is a nearly featureless and very red continuum. Weak Br y (2.166 pm)and and He I (2.058 pm) emission lines are present in the subtracted spectrum, suggesting a slight enhancement of them at the location of the bow shock and point source (compared to a strong enhancement of the continuum). It is possible that these lines are residuals from the subtraction process described above, due to small nonuniformities in the extended emission. However, the ratio of the He 1 to Br y intensities is greater in the subtracted spectrum. As ionization of helium requires higher energy U V photons than ionization of hydrogen, this enhancement suggests the slight line enhancements are real and the presence of an internal source of UV photons. No obvious indicators of shocked gas, such as the lines of Hz that are prominent in shocked molecular clouds, are present in the subtracted spectrum.

553

Astron. Nachr./AN 324, No. S 1 (2003) l

I

l

"

~

'

l

.

Extended (0.6" x 1.8")Region

5 -

0

"

IRS 8

6 -

'*

"

4-

v 3

3Bowshock a n d Point Source

-

1

"

' 2

~

"

" ~ " " ~ " 2.1 2.2 Wavelength (pm)

"

' 2.3

~

'

~ 24

1'

~

Fig. 3 UKIRT K band spectrum of the IRS 8 region, obtained in a 0 . 6 " ~1.8" (NSxEW) aperture and at R-800. The upper trace is the measured spectrum. The lower trace has subtracted from it the average spectrum from the immediately adjacent apertures. Wavelengths of He I , Hz, and H I Br y lines are indicated. The H2 absorption line in the unsubtracted spectrum is due to line emission at the sky position, located 2' east of IRS 8.

3 Discussion The characteristic shape of the extended emission surrounding IRS 8 clearly demonstrates that the H and K band emission originates in a bow shock. As JRS 8 is superposed on an extension of the northern arm of dust and gas, as well as on the molecular ring, the cause of the shock appears to be the collision of the wind or envelope of the starlike object in IRS 8 with material in one or the other of these, at a velocity exceeding the difference in radial velocities of the two Ne 11 components, 120 km sP1 (see Fig. 4). In such a situation the gas and dust densities in the post-shock material would be highly enhanced. Both would be collisionally heated as well. The dust will also be warmed by the radiation from the central star of IRS 8 and by the ambient Galactic center radiation field. One would therefore expect such a bow shock to be a prominent feature in infrared images. Another bow-shock-like structure associated with the Galactic center source IRS 7, seen in Ne II and radio continuum observations (Yusef-Zadeh &Morris 1991, Serabyn, Lacy & Achtermann 1991), has been interpreted by Yusef-Zadeh and Melia (1992) as due to the collision between a wind from the IRS 16 cluster and a wind from IRS 7. The IRS 7 bow shock is not observed in the Gemini images, due to saturation. IRS 7 is located only 0.3 pc from Sgr A* and the IRS 16 cluster and the bow shock is located on the side of IRS 7 facing the center. Both the distance from the center and the orientation of the bow shock are quite different for IRS 8. 3.1

Shape and Dynamics

Wind bow shocks are formed when stars with highly supersonic space velocities and powerful winds interact with the interstellar medium through which they are travelling (Baranov et al. 1971; Comer6n &

F. Rigaut et al.: Bow Shock Around IRS 8

554 Sgr A West IRS 8 [Ne Ill

~""""""""'""~

h

I

l

l

-400

1 I I I I I I I I I

-300

I

-200

I

r

( I , ,, , I , , , , l , , 1 , 1 1 , * , 1 , 1 , ~ i -100

V,

0

100

200

300

(kds)

Fig. 4 Spectrum of the 12.8 p n fine structure line of Ne TI at IRS 8 (data from Lacy et al. 1991). The apeme is ZA"x4.0" and the resolution is 33 km s-l

Kaper 1998). The structure and size of a wind bow shock are determined by the balance between the ram pressure of the moving wind bubble and that of the ambient medium, according to

where pa is the mass density of the ambient medium, v* is the space velocity of the star, d, the distance between the star and the apex (or standoff point) of the bow shock, and M, and v, the mass loss rate and terminal velocity of the wind. This may be reexpressed as follows (Huthoff & Kaper 2002).

(v, /170km s-l) = (&fw,-cv,,B/n3) (.OOgpc/d,)

(2)

where M,,-6 = MW/l0-'MMoyr-', V,,B = v,/lOOO km SKI, and n3 = n ~ / 1 ~0 m ~ -~. The apex of the bow shock is 0.009 pc from the star. The density of the ambient medium is inaccurately known. Lacy et al. (1979) estimated that the densities of typical cloud clumps in the galactic center are lo4 cmP3, and one might expect that intercloud densities are -lo3 ~ m - Using ~ . n3 = 1and assuming normal wind properties for the presumed hot star in IRS 8 (e.g., M,,s x 1, 21,,8 x l),we find v* M 150 km s-'. A relative velocity of 150km s-l would only result in the destruction of a small fraction of the dust (Draine & Salpeter 1979), so the shocked material would emit continuum radiation. Since, as discussed earlier, it seems most plausible that IRS 8 is associated with the narrow Ne II velocity component seen at -10km s-* (see Fig. 4), most of the motion of IRS 8 is in the plane of the sky, which is consistent with the apparent geometry of the bow shock. The orientation of the bow shock indicates that IRS 8 is moving away from the Galactic center.

3.2 Continuum and Line Emission Typical grains of sizes -0.1 pm located at the distance of the bow shock from IRS 8 and heated solely by its radiation field would achieve steady state temperatures of a few hundred Kelvins. Smaller particles (-0.Olpm) would be heated occasionally by single photons to higher temperatures. The shock-heated gas

Astron. Nachr./AN 324. No. S 1 (2003)

555

will have electrons and ions with kinetic energies of about 100 eV, and stochastic collisional heating can raise grains with radii N 0.001pm to temperatures approaching 1000 K (Dwek 1986). The population of grains with temperatures of 500-1000 K (some heated by photons, others by the shock) could account for some of the spatially extended emission seen in the H and K bands, while lower temperature grains, warmed by the radiation field, would account for virtually all of the mid-infared continuum emission. Shock-heating in molecular clouds usually results in strong line emission from Hz, which has either been collisionally excited to high vibrational states or has reformed in excited states following dissociation. The lack of such line emission in the IRS 8 bow shock is probably due to the lack of molecular gas owing either to its destruction in the shock or to the generally hostile conditions close to the center.

4 Conclusion Among the bright mid-infrared sources in the central cluster, IRS 8 stands out as being the most isolated from the rest of the infrared cluster. Assuming a velocity in the plane of the sky of 150km s-l, IRS 8 could have been in the vicinity of the IRS 16 cluster and Sgr A* about lo4 years ago. It is possible that IRS 8 was flung out of the central cluster by gravitational interactions at that time. Another possibility is that it is an interloper from well outside of the center, perhaps a runaway star as the result of a distant supernova. The central parsec of the Galaxy contains a number of objects with infrared characteristics similar to IRS 8. Their emission may also be the result of shock-heating or shock-compression of dust. Tanner et al. (2003)report that 2.2 p m morphologies of several such objects in the northern arm are reminiscent of bow shocks. Thus, IRS 8 may be the most conspicuous member of a class of objects in the Galactic center. The type of star embedded in IRS 8 is unknown. To constrain its properties will require near-infrared spectroscopy at much higher angular resolution than provided here, in order that the spectrum of the central point source not be so diluted by emission from dust. Acknowledgements We thank J.H. Lacy for making available a portion of his Ne II data. The Gemini Observatory is operated by the Association of Universities for Research in Astronomy, Inc., on behalf of the international Gemini partnership of Argentina, Australia, Brazil, Canada, Chile, the United Kingdom and the United States of America. UKRT is operated by the Joint Astronomy Centre on behalf of the U.K. Particle Physics and Astronomy Research Council. BTD was supported in part by NSF grant AST-9988126.

References Baranov, V. B., Krasnobaev, K. V. & Kulikovskii, A. G. 1971, Sov. Phys. Dok., 15,791 Becklin, E.E. & Neugebauer, G. 1975, ApJ, 200, L71 Becklin, E.E., Matthews, K., Neugebauer, G. & Willner, S.P. 1978, ApJ, 219, 121 Comerh, E, & Kaper, L. 1998, A&A, 338,273 Draine, B.T., & Salpeter, E.E. 1979, ApJ, 231,438 Dwek. E. 1986, ApJ, 320,363 Huthoff, F., & Kaper, L. 2002, A&A, 383,999 Lacy, J.H., Achtermann, J.M. & Serabyn, E. 1991, ApJ, 380, L71 Lacy, J.H., Baas, F., Townes, C.H. & Geballe, T.R. 1979, ApJ, 227, L17 Lacy, J.H., Townes, C.H., Geballe, T.R. &Hollenbach 1980, ApJ, 241, 132 Rigaut, F., Geballe, T.R., Roy, J.-R., Blum, R.D., Davidge, T.J. & Cotera, A. 2003, these proceedings Serabyn, E., Lacy, J.H. & Achtermann, J.M. 1991, ApJ, 378, 557 Tanner, A,, Ghez, A., Morris, M. & Becklin, E. 2003, these proceedings Yusef-Zadeh, F. & Melia, F. 1992, ApJ, 385, LA1 Yusef-Zadeh, F. & Moms, M. 1991, ApJ, 371, L59 Yusef-Zadeh, F., Moms, M. & Ekers, R.D. 1990, Nature, 348,45

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.557-561 (2003)/ DO1 10.1002/asna.200385058

Monitoring Sagittarius A* in the MIR with the VLT *. A. Eckart**’, J. Moultaka I , T. Viehmann I , C. Straubmeier I , N. Mouawad 2, T. Ott R. Schodel ’, F.K. Baganoff 3, and M.R. Morris

’,

’, R. Genzel

’ I.Physikalisches Institut, Universitat zu Koln, Ziilpicher Str.77,50937Koln, Germany

Max-Planck-Institutfur extraterrestrische Physik GiessenbachstraRe,85748 Garching, Germany Massachusetts Institute of Technology,Center for Space Research, 77 Massachusetts Avenue, Cambridge, MA 021 39-4307, USA Division of Astronomy,Box 951562, UCLA, Los Angeles, CA 90095-1562, USA

Key words Galactic Center, Sgittarius A*, accretion processes

Abstract. The ISAAC and NAOS/CONICA systems on the ESO VLT UTl and UT4 telescopes have been used to monitor the MIR L and M band (3.3 pm and 4.6 pm) flux density at the position of Sgr A* and to obtain new information on the stellar populations within the central cluster. The monitoring program was carried out simultaneously to Chandru observations and resulted in upper limits on the flux density of Sgr A* during a newly detected out-burst of the X-ray source at the position of Sgr A*.

1

Introduction

Near-infrared diffraction limited imaging over the past I0 years (Eckart & Genzel 1996, Genzel et al. 1997, Ghez et al. 1998, Genzel et al. 2000, Ghez et al. 2000, Eckart et al. 2002, Schodel et al. 2002) has resulted in convincing evidence for a 3 x lo6 Ma black hole at the center of the Milky Way (see also contributions by Ghez et al., Schodel et al., Ott et al., and Reid et al. in these proceedings). This finding is supported by the discovery of a variable X-ray source at the position of Sgr A* (Baganoff et al. 2001, and Baganoff et al. in these proceedings). Here we report on recent MIR L and M band observations that contribute to the investigation of the variable X-ray source at the position of Sgr A*. The recent May/June 2002 Chandru ACIS-I (see Baganoff et al. in these proceedings) observation of Sgr A* detected three X-ray flares with amplitudes in excess of a factor of 10 and several flares of more than a factor of five with durations ranging from one half to several hours and with rise and fall times similar to the X-ray flare discovered with Chandru in 2000 (Baganoff et al. 2001). During the Chandra observations numerous ground-based telescopes were employed for simultaneous observations from the radio cm- to the nearinfrared wavelength domain. The VLT took part in this campaign observing the Sgr A* position in the L and M band.

2

Observations and data reduction

The infrared data were obtained with the new NAOS/CONICA adaptive optics assisted imager/spectrometer at the UT4 (Yepun) as well as with ISAAC on UTI (Antu), at the ESO VLT (Lenzen, et al. 1998, Rousset et al. 1998, Brandner et al. 2002). During the commissioning and observatory preparation of NAOS/CONICA the instrument was used to obtain diffraction limited images in the H , K,, and L band * based on observations at the Very Large Telescope (VLT) of the European Southern Observatory (ESO) on Paranal in Chile * * Corresponding author: e-mail: eckartQphl.mi-koeln.de,Phone: 449221 4703546, Fax:+492214705162 @ 2003 WILEY-VCH Verlag GmbH & Co KCaA. Weinhem

558

A. Eckart et al.: VLT monitoring of Sm A*

Fig. 1 High resolution adaptive optics image as taken on August 19, 2002 in the L band with NAOSKONICA on the VLT UT4 during the science verification phase. At the position of Sgr A* the L band flux density is dominated by the star S2. For relative positioning of Sgr A* with respect to S2 see Schiidel et al. 2002 and Reid et al. in these

proceedings.

centered at wavelengths of 1.6, 2.1, and 3.3 pm, with a EWHM of 43, 56, and 88 mas, respectively (at 0.0132”lpixel). Further details are given in Schodel et al. (2002) and Genzel et al. (2003). The paired flat fielded images at different chopper throws (18” =k 2”) and chopping position angles (0” to 180”)were subtracted from each other, resulting in frames containing a positive and a (shifted) negative image. The frames were then shifted to a common reference point that coincides with a positive image of a source. Subsequently, frames belonging to the same batch, i.e. taken sequentially with identical or different chopper throws andor chopping angles at a given date were combined by calculating a median. Since the images were moved to a common reference point this procedure eliminates the negative “shadows” generated by the subtractions. This procedure also effectively removes cosmic rays and bad pixel structures. Such a batch typically consists of up to 40 images, and the resulting combined image covers an integration time of approximately 35 minutes. The resulting median images with the best seeing were then selected as reference images for the differential flux measurements. A comparison of our seeing limited ISAAC images to an image taken with the adaptive optics system NAOS/CONICA at the VLT UT4 (see Fig. 1) shows that the flux density at the position of Sgr A” (see Schodel et al. 2003, Schodel et al. in these proceedings, and Genzel et al. 2003) is dominated by the high velocity star S2. Therefore differential images are the appropriate way to estimate possible flux density contributions from Sgr A* itself. In order to obtain differential images from our ISAAC data, the reference images were convolved with a Gaussian of appropriate FWHM to match the seeing conditions and subsequently subtracted from the other median images. The differential images were flux calibrated relative to IRS 16NE and used to derive

559

Astron. Nachr./AN 324, No. S 1 (2003) 0.00003

0.00002

I

I

I

I

I

2

-

4a

,

,

,

,

Lo

;="

-

I

,

,

,

I

,

E

E

&-

m

5

2

E

-

.-c

-

-

0 >

-t?

.d

c

rd

-

0 -

-0.00001

'

8

I

'

a

I

30000

I

I

35000

a

I

'

1

'

I

40000

Fig. 2 L band spectrum at the position of Sgr A* taken through a 0.6' wide slit using ISAAC at the VLT UTI. The locations of hydrogen lines are marked.

a mean upper flux density limit of a region to within ~k0.2"of the Sgr A* position. The L and M band fluxes were finally dereddened assuming A ~ w 1 . 6(CICnet et al. 2001) and Ax o( X-1.75 (Draine 1989). Our recent ISAAC L band imaging and spectroscopic monitoring program of Sgr A* (Baganoff et al. 2001 and these proceedings, Eckart et al. 2003), which was carried out simultaneously with Chundru observations, has lead to high quality L band spectra of several prominent sources within the central cluster that fell into slit settings across Sgr A*. Data reduction and calibration of the spectroscopic data is described in Eckart et al. in these proceedings. A spectrum at that position is shown in Fig. 2. Here the weak hydrogen recombination lines are most likely due to emission from the mini-spiral. The Rayleigh Jeans shape of the continuum spectrum compares well to that of other hot stars in the central stellar cluster.

3

MIR monitoring of Sgr A*

One X-ray burst was well covered by our MIR observations (see Fig. 3). On 29 May 2002 this burst lasted from about 0 5 4 0 to 08:20 and peaked at approximately 06: 10 UTC. The VLT MIR imaging times cover intervals between 04:26 - 05:41 UTC AND 08:49 - 10:03 UTC. MIR spectroscopy data with the slit positioned on Sgr A* were taken between 06:28 - 07:03 and 07:43 - 08:17 UTC. No MIR burst activity was recorded during these times. The absence of MIR variations is in good agreement with the results of Hornstein et al. (2002). Dereddened upper limits in the differential flux density that could be attributed to a variable source is of the order of 10 mJy for the imaging in typical seeing conditions of 0.8". Compared to the imaging data, the spectroscopic exposures averaged seeing variations over longer integration times of 20 to 30 minutes. In addition the source structure was averaged across the spectrometer slit. The observations were of course also taken through the collimator optics of the spectrograph. All of these effects resulted in an angular resolution of only about 1 arcescond. The corresponding differential flux density limit that could be attributed to a variable source is of the order of 100 mJy for the spectroscopic exposures.

560

A. Eckart et al.: VLT monitoring of Sgr A* x

5

400

MIR

E!

x

2

200

3 =

o

8

2

0.06

I

X-ray

g

U

% .

"

0.04

5

2

3 002 0

I

10000

20000

30000

rim? tn s r m n d \

Fig. 3 Sgr A* X-ray burst monitored simultaneously in the mid-infrared with ISAAC at the VLT UT1. The midinfrared date were obtained in the M band and L band at 3.3pm and 4.6pm, respectively. In addition the flux density limits that were obtain from spectroscopic measurements are labeled Ms and Ls. The time axis starts on 29 May 2002 at 01:2448 UTC.

In contrast to theoretical predictions (e.g. Markoff et al. 2001), the X-ray flares were not accompanied by large variations in the radio andlor infrared domain. This makes models in which the X-ray bursts are explained via an increased accretion rate (i.e. increase in magnetic field strength andlor electron density) or an increase in electron temperature (i.e. a local heating in the accretion flow), less preferable. The increased accretion rate model by Liu & Melia (2002) with X-ray emission from synchrotron self-Comptonization is consistent with the observations but results in a predicted X-ray slope that is too soft. However, nonthermal shock heating models as mentioned by Markoff et al. (2001) are still fully consistent with the results of our most recent multi-wavelength campaign. Acknowledgements This work was supported in part by the Deutsche Forschungsgemelnschaft (DFG) via grant SFB 494. We are grateful to all members of the NAOWCONICA team from MPIA/MPE, MeudodGrenoble Observatories, ONERA, ESO. In particular, we thank N.Ageorges, K.Bickert, W.Brandner, Y.Cltnet, EGendron, M.Hartung, N.Hubin, C.Lidman, A.-M. Lagrange, A.F.M. Moonvood, C.Rohrle, G.Rousset and JSpyromilio.

References Baganoff, F.K. et al. 2001, Nature,413, 45 Brandner, W. et al. 2002, The ESO Messenger 107, 1 CJBnet, Y., Rouan, D., Gendron, E., Montri, J., Rigaut, F., Lfia, P., Lacombe, F. 2001, A&A, 376, 124 Draine, B.T. 1989, Proc. 22nd ESLAB Symp. on IR Spectroscopy in Astron , Kaldeich, B.H. (ed.), ESA SP290, p. 93 Eckart, A. & Genzel, R. 1996, Nature, 383,415 Eckart, A., Ott, T., Genzel, R., & Lutz, D. 1999, in Proc. of MU Symp. No.193 on 'Wolf-Rayet Phenomena in Massive Stars and Starburst Galaxies' Puetrto Valarta, Mexico, November 3-7, 1998, van der Hucht, K.A., Koenigsberger, G., Enens, P.R.J. (eds.), Kluewer, p.449 Eckart, A., Genzel, R., Ott, T. and Schoedel, R. 2002, MNRAS, 331,917-934 Genzel, R.,Thatte, N., Krabbe, A., Kroker, H., Tacconi-Garman,L. E. 1996, ApJ, 472, 153

Astron. Nachr./AN 324. No. SI (20031

561

Genzel, T., Eckart, A., Ott, T. & Eisenhauer, F. 1997, MNRAS, 291, 219 Genzel, R., Pichon, C., Eckart, A,, Gerhard, 0. & Ott, T. 2000, MNRAS, 317. 348-374 Genzel et al. 2003, in preparation Ghez, A,, Klein, B.L., Moms, M. & Becklin, E.E. 1998, ApJ, 509,678-686 Ghez, A,, Moms, M., Becklin, E.E., Tanner, A. & Krernenek, T. 2000, Nature, 407, 349 Hornstein, S.D., Ghez, A.M., Tanner, A,, Moms, M., Becklin, E.E., Wizinowich, P. 2002, ApJL, 577, L9 Lenzen, R., Hofmann, R., Bizenberger, P. & Tusche, A. 1998, Proc. SPIE, IR Astronomical Instrum. (A.M.Fowler ed.), 3354, 606 Liu, S., & Melia, F. 2002, ApK, 566, L77 Markoff, S., Falcke, H., Yuan, F., Biermann, P. L. 2001, A&A, 379, L13 Rousset, G. et al. 1998, Proc.SPIE Adaptive Optics Technology (D.Bonaccini, R.K.Tyson eds), 3353, 508 Schodel et al., 2002, Nature, 419, 694

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.563 -566 (2003')/ DO1 10.1002/asna.200385104

The magnetic field in the central parsec A. C. H. Glasse'', D. K. Aitken***,and P. F. R o ~ h e * * * ~

' Joint Astronomy Centre, 660 N. A'Ohuku Place, Hilo, HI 96720, U S A .

* University of Hertfordshire, College Lane, Hatfield, Hertfordshire, ALlO 9AB, U.K. University of Oxford, Astrophysics, Keble Road, Oxford, OX1 3RH, U.K.

Key words polarimetry, infrared, imaging, galactic centre Abstract. We present imaging polarimetry of Sgr A at a wavelength of 12.5 pm, obtained with Michelle on UKIRT early in 2002. With considerably more sensitivity than previously available we have concentrated on the faint region bounded by the northern arm north of IRSl, IRS8, and the western arc of the molecular

ring, to probe the magnetic field directions in this region.

1 Introduction Polarimetry in the mid-infrared gives information on magnetic field directions. This is because, in equilibrium with ambient gas, dust grains spin with rotational frequency in the MHz range, and as a consequence develop a magnetic moment along their spin axes which then precess about the ambient field direction. In itself this does not produce alignment but any departure from isotropy will produce an averaged spin direction either along or at right angles to the field. In practice most disturbance mechanisms align the spin axes, which are normal to a grain long dimension, towards the field and the radiation emitted by a warm grain is then polarized normal to this direction. The Galactic Centre polarisation contains an interstellar absorptive component which, at 12.5 pm, is well represented by a uniform amplitude over the ionised filaments of 1.8% at zero degrees (Aitken, Smith, Moore, & Roche 1998, hereafter ASMR). Mid-infrared imaging polarimetry of the central parsec can therefore be corrected to provide an accurate measure of the component of the magnetic field direction in the plane of sky wherever the measured polarisation fraction is greater than a few percent.

2 Observations The observations shown in Figure 1 were made on the night of 27th April 2002 at UKIRT using the Michelle mid-infrared imager and spectrometer. A 1 pm passband filter with a central wavelength of 12.5 pm was used in series with a cryogenically cooled wire grid analyser to measure the linearly polarised intensity of the target region, with the plane of polarisation modulated by rotating a warm, externally mounted half-waveplate. The observing sequence then involved coadding 50ms exposures in on-target and off-target chop positions while the telescope was chopped in an east-west direction at a frequency of 6.3 Hertz and with an amplitude of 20 arcseconds on the sky. After 20 seconds, the difference of the two coadded frames was saved, and the half-waveplate advanced to the next rotation angle in the sequence 0, 45, 22.5 and 67.5 degrees, as required for full sampling of the sinusoidally varying polarised intensity. * e-mail: a.glasseQjach.ac.uk, Phone: +44 131 6688100, Fax: +00999999999 ** e-mail: [email protected],Phone: +44 131 6688100, Fax: +00999999999 * * * e-mail [email protected]

@ 2003 WILEY-VCH Verldg GmbH & Co KGaA. Weinhem

564

A. Glasse et al.: The magnetic field in the central D X S ~ C l

"

"

l

"

"

l

"

"

l

,IRS8

20

10

0

-10

RA. O f f s e t ( a r c s e c o n d s) Fig. 1 Imaging polarimetry of the galactic centre. The vectors have been plotted in the direction normal to the direction of polarisation to show the magnetic field direction. A constant 0.018 has been subtracted from the ' Q component of the fractional polarisation to correct for interstellar absorption. The contours increase by a factor of ,/2, with the lowest plotted at 250 mJy/arcsecond*. The hatched area masks off regions contmnated by the 20 arcsecond chop throw or vignetted by the waveplate. The positions of infrared sources are shown, with IRS21 marked by an

additional '+' symbol.

At the end of the 67.5 degree coadd, the telescope was offset to blank sky 3 arcminutes to the west of the target position, and the sequence repeated. This pattern was then used for a higher level offsetting sequence of target, sky, sky, then target position, with 180 degrees added to the waveplate rotation angle at each repetition. The unpolarised standard star E Sgr was used to flux calibrate the images by assigning it a 12.5 p m flux of 140.1 Jy, based on its IRAS 12 pm flux of 213.7 Jy, colour corrected for a 3500 K black body. The polarisation of E Sgr was measured to be < O h % , consistent with the instrumental polarisation being negligibly low.

565

Astron. Nachr./AN 324, No. S1 (2003)

The total elapsed time to take the observation was 102 minutes, during which the array was integrating on the target for 24 minutes and on the sky for 24 minutes. The remaining 54 minutes were taken up with reading out the detector, offsetting the telescope and moving the waveplate. The problem of confusion between the chop beams caused by using a 20 arcsecond chop throw on such an extended target was addressed by taking an additional set of observations in which the telescope was offset by a distance close to double the chop throw, such that the N-S bar appeared in the east in one offset position, and in the west in the other position. Subtraction of the sky fields taken in the first dataset from the target fields taken in this extra dataset allowed us to form an uncontaminated (but also poorly sky subtracted) image which was used as a guide in generating masks for the selection of regions uncontaminated by the small chop throw. Only the first data set was used to form the polarisation images presented here.

10

5

0

-5

RA. Offset (arcseconds)

Fig. 2 Detail of the region around Sgr A*, which is located at the origin. The contours increase by a factor of with the lowest plotted at I Jylarcsec'. The hatching markes the edge of the waveplate, as in Figure 1.

A,

3 Discussion The region to the west of the northern arm shows faint filamentary structure (shown in Figure 1, with the locations of the infrared sources marked), curving away from the general direction of the arm towards the circumnuclear disk (CND); the arm itself becomes faint and indistinct in the direction towards IRS8. The details of mid-infrared structure follow closely the 6cm radio images of Yusef-Zadeh and Wardle (1993), and implying that the dust and ionized gas are well mixed. Figure I also displays the observed polarization fraction and the inferred magnetic field direction.

566

A. Glasse et al.: The magnetic field in the central parsec

In the brighter parts of Figure 1 the field directions and polanzed fractions reported earlier (ASMR) are reaffirmed, and the convergence of field lines from IRSl to a region south of Sgr A* is very evident, suggesting orbital compression. This region is shown in greater detail in Figure 2. Figure 1 also shows the field directions in the fainter filaments and near IRS8 for the first time. The general features are that the field directions follow the local directions of the northern arm and also those of the fainter filaments and structures to the west and the extension towards IRSI. The polarization fraction varies along the northern arm,reaching a maximum of 7-8% close to IRSl, almost vanishing south of IRS7, and reducing through IRS5 and further north. Alignment is thought to be almost uniform in Sgr A (ASMR) and the variation in polarization has been attributed to the angle, 4, the field makes with the plane of the sky:

P = Pocos2f$, a relationship which suggests a method for investigating 4. Here, for instance, the polarization amplitudes in the faint filaments are very similar to those in the northern arm at the same declination, 3-4%, suggesting a similar out-of-sky-plane angle ( w 45 degrees) but with a sky position angle of about 150 degrees, out of the Galactic Plane. The field pattern also shows a slight convergence at IRS8 which, while possibly associated with the bow-shock structure observed by Rigault et a1 (poster, this workshop), seems counter to expectations, and an inflexion appears in the field direction at IRS21, another source suspected of disturbing the northern arm flow (Tanner et al. 2002, and these proceedings). Higher spatial resolution is needed to clarify these interactions between fields, winds and flows. Two faint short radio filaments appear to extend the northern arm towards IRS8, albeit with an abrupt change of direction. While these structures are only hinted at in the MIR the indicated field direction is similar, about 135 degrees. At IRS8 itself there is a change in the field direction back to north-south and a significant increase in polarization fraction, suggesting a change in direction relative to the line of sight. It has been suggested that the northern arm arises from the inner rim of the CND in the region near IRSS, though Latvakoski et al. (1999) have detected structures at 30 to 40 pm wavelengths outside the CND which may be dynamically related to the arm.The field directions in the filaments are probably due to shearing motions within them and will reflect their general direction of motion. In the radio images of Roberts and Goss (1 993) more filaments are seen bridging the space between the CND and the northern arm and the impression is that they are tributaries feeding the northern arm.However, since these filaments cross the western arc they probably do not originate in it. Similarly the situation near IRSS is left ambiguous; the sudden changes of direction of structure merely confuses whatever association there may be with the molecular ring. We look forward to using the 30-fold speed increase and doubling of spatial resolution that is expected when Michelle is moved from UKIRT to Gemini North, to trace the magnetic field direction in the northern arm and its associated filaments where they encounter and cross the CND and are suspected to be interacting with outflows from stars.

References Aitken, D. K., Smith, C., Moore, T. J., & Roche, P.F. 1998, MNRAS 299,743 Latvakoski, H. M., Stacey, G. J., Gull, G. E., Hayward, T. L. 1999, ApJ, 511,761 Roberts, D. A,, & Goss, W. M. 1993, ApJS, 86,133 Tanner, Ghez, A. M., Moms, M., Becklin, E. E., Cotera, A., Ressler, M., Werner, M., & Wizinowich, P. 2002, ApJ, 515,860 Yusef-Zadeh,F., & Wardle, M. 1993, ApJ, 405,584

Astron. Nachr./AN 324, No. S1.567-571 (2003)/ DO1 10.1002/asna.200385079

Mid-Infrared Imaging and Spectroscopic Observations of the Galactic Center with SubarulCOMICS

’,

Y. Okada* T. Onaka’,T. Miyata2,H. Kataza3, Y. K. Okamoto4,S. Sako17’,M. Hondal.’, T. Yamashitas.’,and T. Fujiyoshi’

’ Department of Astronomy, School of Science, University of Tokyo, Japan

’

Institute of Astronomy, School of Science, University of Tokyo, Japan The Institute of Space and Astronautical Science, Japan Institute of Physics, Center for Natural Science, Kitasato University, Japan Subaru Telescope, National Astronomical Observatory of Japan, Hawaii, U.S.A.

Key words Galactic center, mid-infrared, spectroscopy, dust extinction, silicate.

Abstract. We report the results of mid-infrared (7.8 pm-13.2 pm) high-spatial resolution imaging and spectroscopic observations of the Galactic center region with the Cooled Mid-Infrared Camera and Spectrometer (COMICS)on the Subaru telescope. The images clearly show bright infrared sources and small structures in the diffuseemission. The spectra of all the observed positions show the 9.7 pm silicate absorption feature. After corrected for the empirically-derivedextinction, the intrinsic spectra of the infrared sources show either strong silicateemission or absorption, while the intrinsic diffuse emission has a power-law type spectrum. This difference indicates a possibility of dust processing due to the interaction between the infrared sources and their surrounding medium or a different origin of the dust grains surrounding the sources from those in the diffuse region.

1 Introduction High-resolution mid-infrared (MIR) imaging in the central parsecs of our Galaxy reveals significant morphological details for both the diffuse emission and numerous compact infrared sources. Gezari et al. (1996) used MIR imaging observations of 8 bands to show that compact infrared sources are warm spots in the dense Galactic center (GC) dust complex, containing embedded luminous stars. They also suggested from MIR color maps that diffuse dust clouds are heated by Her stars distributed in the GC region. Cotera et al. (1999) indicated from imaging observations with Keck II/MTRLIN that the heat source for the diffuse emission in the northern arm is the IRS 16 cluster, a hot-star cluster near Sgr A’. On the other hand, MIR spectroscopic observations give an insight into the silicate dust properties in the GC region and along the lines of sight. Roche and Aitken (1985) observed the central parsecs with the beam size of 2”-4” in MIR spectroscopy and found that most infrared sources indicate an interstellar extinction of ~ ( 9 . 7 = ) 3.6 +C 0.3, except for IRS 3 that may suffer local silicate absorption. In the present paper we report the results of imaging and spectroscopic observations of the GC region with the Cooled Mid-Infrared Camera and Spectrometer (COMICS; Kataza et al. 2000; Okamoto et al. 2003) on the 8.2 m Subaru telescope. The present observations achieved the highest spatial resolution ever made for the GC region in MIR, allowing us to study detailed structures around compact sources and of the diffuse emission. * Corresponding author: e-mail: okadaC0astron.s.u-tokyo.ac.jp,Phone: +81358414268, Fax: +813 5841 7644 @ 2003 WILEY-VCH Verlag GmbH & Ca KGaA, Weinham

Y. Okada et al.: Mid-infrared observation of the Galactic Center

568

2 Observations We observed the GC region on 2002 July 30 with Subaru/COMICS. The chopping throw was 60” in the north-east to south-west direction. Imaging observations were made for the area of 42” x 32’’ around Sgr A* with the spatial resolution of about 0.3”. The pixel scale was 0.13’’pix-’. Images were taken with 3 band filters of 8.8 pm, I 1.7 pm, and 12.4 pm. Each filter had about 1 pm width. The on-source integration time was 26 sec, 15 sec, and 22 sec, respectively. After subtracting the off-beam image and dividing by the flat image, we co-added frames of the same band with the accuracy of 1 pix. Spectroscopic observations were carried out for 7.8-13.2 pm with the spectral resolution of 250 (N-low spectroscopy mode of the COMICS with the slit of 0.3” wide x 36” long). We had two slit positions. One intersected IRS 1 and the bar region just south of IRS 6, and the other was on the line that connects IRS 7 and IRS 10.

Fig. 1 High-resolution 12.4pm map with SubadCOMICS. The positions of the infrared sources are indicated by arrows. The box indicates the position of Sgr A*. The asterisk shows the position where the spectrum of Fig. 3d was taken. North is up and east is to the left.

3 Results and Discussion Fig. 1 shows the high-resolution 12.4pm map. Small structures in the diffuse emission of the northern arm and the bar regions are clearly seen. Sgr A* does not show any intensity enhancement in all the observed bands, which is compatible with a small enhancement of 25 mJy at 8.7 pm at the position of Sgr A* suggested by Stolovy et al. (1996). Fig. 2 is the 8.7 pd12.4pm color map, showing clearly that individual infrared sources are at local temperature maxima and that the western side of the northern arm facing to Sgr A* has higher temperatures than other diffuse regions. These facts indicate that the heating sources are embedded in the individual sources and that the hot-star cluster near Sgr A* heats the dust grains in the diffuse region around it as pointed out by Cotera et al. (1999). Fig. 3 shows the spectra of the IRS sources and the diffuse bar region about 10 arcsec west from IRS I indicated by the asterisk in Fig. 1. The silicate absorption feature centered at 9.7 p m and the [Ne 111 12.8 pm

Astron. Nachr./AN 324, No. S 1 (2003)

569

Fig. 2 8.7 pd12.4 pm color map. White color corresponds to high temperature. Pixels with the flux less than the 5 u noise level are masked and indicated by black.

emission line are clearly seen at all the observed positions. Other forbidden lines, such as [S IV] 10.51 pm and [ArIII] 8.99 pm, and the unidentified infrared (UIR) bands at 8.6 and 11.2 pm are not detected. To investigate the effect of interstellar extinction on the intrinsic spectra, we model optically-thin hot dust emission, attenuated by cold dust along the line of sight. We fit the observed spectra with the following equation:

where &(A) is the emissivity (including the beam filling factor) of the emitting hot dust component, Bx is the Planck function, T d is the hot dust temperature, and T(X) is the optical depth of the cold dust component. For the spectral dependence of the emissivity .(A), we try three cases, the silicate emissivity of Ossenkopf et al. (1992), and power-law emissivities of X - l and XP2. For most positions, the silicate emissivity gives the best fit. The derived 7(9.7) is about 4 at IRS 1 and IRS 10, increasing up to about 6 along ) 6 is significantly larger the line from IRS 1 towards the diffuse bar region. An optical depth of ~ ( 9 . 7 = than that derived from compact sources by Roche & Aitken (1985). The difference may be interpreted as local extinction, caused by unevenly distributed cold dust grains, which does not effect the infrared sources IRS 1 and IRS 10 as they have swept away these grains by their influence. This possibility cannot be ruled out a priori, although previous observations suggest rather uniform extinction over the GC region using both spectra of several compact sources (Roche and Aitken 1985) and diffuse emission (Chan et al. 1997). Chan et al. (1997) derived ~ ( 1 9 . 7 ) 2.1-2.2, which is compatible with r(9.7) 3.6. If the cold dust optical depth does not differ appreciably from ~ ( 9 . 7 ) 3-4 for both the compact sources and diffuse emission, the intrinsic spectrum of diffuse emission has to be different from the silicate emission model. ) 2 lower than for The fit for the diffuse region with a X-‘ or X-’ emissivity gives a ~ ( 9 . 7 approximately the case of the silicate emissivity. No significant difference is found between the X-’ and XP2 emissivity fit. In the following, only the case of the A-’ emissivity is discussed. The average -49.7) of the diffuse bar region over a 6.2’’ x 0.3” area is 3.13. Then we assume the X-’ emissivity for the intrinsic emission from

-

-

N

Y. Okada et al.: Mid-infrared observation of the Galactic Center

570

6

8

9

10 11 12 Wavelength [ rrn]

13

.

8

9

10

11

12

13

Wavelength [pm]

Fig. 3 Observed spectra of (a) IRS1, (b) IRS 7, (c) IRS 10, and (d) the diffuse bar region indicated by the asterisk in Fig. 1. The 9.3-9.9 pm part of the spectra is not plotted because of the large uncertainty due to the ozone emission from the terrestrial atmosphere.

the diffuse bar region and empirically derive .(A) (Fig. 4a). With the derived .(A) and ~ ( 9 . 7 = ) 3.13 we estimate the intrinsic spectra at all the observed position, including the infrared sources. Figs, 4b-e plot spectra of the intrinsic emission for the hot dust component after extinction correction. The extinctioncorrected spectra of IRS 1 and IRS 10 reveal the strong silicate emission feature. A sharp feature around 10pm in IRS 1 and IRS 10 is a result of the empirically-derived extinction and may be spurious. Even at positions 2” away from IRS 1, the strong silicate emission feature is still seen. The IRS 7 spectrum shows an extra silicate absorption feature (Fig. 4c). The spectrum of IRS 3 has a low S/N but also indicates a similar absorption feature (not shown; see also Roche and Aitken 1985). As expected the extinctioncorrected spectrum of the diffuse region does not indicate the silicate emission and is rather flat. The present observations provide MIR N-band spectra of the GC region with the spatial resolution of 0.3” for the first time and indicate that the intrinsic spectra of infrared sources and diffuse emission may differ if the extinction over the 30” area is uniform and ~ ( 9 . 7 )N 3-4 as previously suggested. This indicates that the dust properties of the sources are different from those in the diffuse region. The spectrum of the emission surrounding the compact sources, however, is similar to the sources themselves, suggesting that the dust around them originates from stellar mass-loss. The surrounding emission can extend to over ten thousands AU. Tanner et al. (2002) found that the constant mass flow model is not suitable for the observed infrared intensity distribution around IRS 21, and suggested that the extended structure surrounding the sources is a result of the interaction between the star and gas, such as shocks generated by stellar wind. In this case, the dust properties around the sources and diffuse emission are expected to be the same. The difference in the intrinsic spectra thus suggests that there might be dust processing triggered by the interaction between the source and surrounding dust. Alternatively the dust grains surrounding the sources may have a different origin from those of the diffuse region, such that they are supplied by individual sources as a result of extraordinarily strong mass-loss.

Astron. Nachr./AN 324, No. S1 (2003)

57 1

0.5

8

10

9

11

12

13

~ l o ~ ~ o

I

E

20

;

0

0 8

9

10

11

12

13

8

Wavelength[ pm]

9

10

11

12

13

Wavelength[ pm]

Fig. 4 (a): Empirically-derived extinction emissivity (solid line; see text) and silicate emissivity by Ossenkopf et al. (1992) (dashed line), and intrinsic spectra after empirically-derivedextinction correction of (b) IRSl, (c) IRS 7, (d) IRS 10, and (e) the diffuse bar region. A sharp feature around 10pm in (b) and (d) may be spurious due to the extinction correction.

Acknowledgements We would like to thank all the staff members of the Subaru Telescope for their support during the observations and development of the instrument.

References Chan, K.-W., et al. 1997, ApJ, 483,798 Cotera, A,, et al. 1999, ASP Conf. Series, 186, 240 Gezari, D., et al. 1996, International Astronomical Union, 169, 23 1 Kataza, H., et al. 2000, Proc. SPIE, 4008, 1144 Okamoto, Y. K., et al. 2003, Proc. SPIE, in press Ossenkopf, V., et al. 1992, A&A, 261,567 Roche, P. F., and Aitken, D. K. 1985, MNRAS, 215,425 Stolovy, S. R., et al. 1996, ApJL, 470, L45 Tanner, A,, et al. 2002, ApJ, 575,860

This Page Intentionally Left Blank

Astron. Nachr./AN 324,No. S 1, 573 -576 (2003) / DO1 l0.1002/asna.200385102

Physical Conditions in the Central Parsec Modeled from MidInfrared Imaging Photometry

',

Dan Gezari* Eli Dwek', and Frank Varosi2 ' NASNGoddard Space Flight Center, Infrared Astrophysics Branch, Code 685, Greenbelt, MD 20771 USA

University of Florida, Department of Astronomy, Gainsville, FL 3261 1 USA

Key words infrared, imaging, galactic center Abstract. We have made array camera images of the central parsec with 1 arcsec resolution at eight midinfrared wavelengths between 4.8 and 20.0 pm. The images are used to model the temperature, opacity and bolometric luminosity distributions of the emitting dust in the central parsec, as well as the dust extinction in the line of sight. Several new results emerge from the model calculations: 1) The compact IRS sources are all local peaks in the emitting dust temperature distribution. 2) The IRS source positions are local minima in the emitting dust opacity distribution. 3) The opacity and temperature distributions are generally complementary. 4) The compact IRS sources are very luminous, particularly the IRS3 and IRSl (L 3x lo5 Lo). and they make a significant contribution to the energy budget of infrared complex.

-

1 Introduction Two schools of thought have emerged to explain the high infrared luminosity and an apparent large mass concentration within the central parsec of the Galaxy. One contends that a recent episode of massive 0 star formation has occured which would account for the observed infrared luminosity and ionizing radiation from Sgr A West (Rieke and Lebofsky 1982, Allen et al. 1990). The competing view is that the high total infrared luminosity and high velocities observed in the central parsec are both evidence for a "central engine", possibly a massive black hole surrounded by an accretion disk (Lynden-Bell and Rees 1971, Rees 1982, Melia 1992). The exotic object might also give rise to the non-thermal emission observed from bright radio continuum point source Sgr A*. Three fundamental questions remain to be addressed: 1) What is the origin of the bright infrared emission? 2) What is the source of the high luminosity of the central parsec? 3) What is the nature of Sgr A*? Possible answers include visible/UV starlight reprocessed in compact circumstellar shells (either a superposition of numerous individual sources or a few very luminous objects), diffuse dust clouds heated by the local ambient ionizing radiation field, or a dominant "central engine". We have made array camera images of the central parsec with 1 arcsec resolution at eight mid-infrared wavelengths between 4.8 and 20.0 pm. The images are used to model the temperature, opacity and bolometric luminosity distributions of the emitting dust in the central parsec, as well as the dust extinction in the line of sight. Our array camera system uses a 58 x 62 pixel gallium doped silicon (Si:Ga) array sensitive from 5 - 17 pm and six fixed interference filters (0.1) between 7.8 and 12.4 pin, and one at 18.1 pm (as well as a 5 - 14 pm circular variable filter (CVF). The array camera system optics and electronics are presented in further detail by Gezari et al. (1992). The Galactic Center images were made at the 3.0-meter NASA Infrared Telescope Facility (IRTF) at Mauna Kea in 15 March 1988. * Corresponding author: e-mail: [email protected]: +01301286 3432, Fax: +01301286 1617

@ 2003 WILEY-VCH Verlag GmbH & Co. KGaA. Weinhem

D. Gezari et al.: Phvsical Conditions

574

A complete set of array camera images at eight wavelengths between 4.8 and 20.0 p m has been obtained for the central 15 arcsec array field of view (24 arcsec = 1 parsec at 8.5 kpc), each with typically 5 minutes of integration time. A large-scale 12.4 pm mosaic image of the Galactic Center, covering an area of 75 x 105 arcsec, was assembled from 50 individual array camera exposures using the MOSAIC image processing software we developed (Varosi and Gezari 1993, http://astrolab.gsfc.nasa.gov/mosaic).

2 Dust Emission-Extinction Model The observed infrared intensity is calculated from each source element with source opacity total line of sight (LOS) extinction. Dust opacities for the emitting and absorbing grains were calculated for grain radii a ranging from 2 x p m to 5 x 10-1 pm, assuming a grain size number distribution proportional to a diameter power law of -3.5 and characterized by optical constants from Draine and Lee (1984) and Draine (1985). However, the source dust composition is unknown a priori, and could be either siliconor carbon-rich. The abundance ratio therefore must be chosen to represent a range of mixtures between all-silicate and all-graphite grains, and the parameters are varied to achieve the best fit to the observed spectra. N

N

2.1 Model Results Several new results emerge from the model calculations: i) The compact IRS sources are all local peaks in the emitting dust temperature distribution. Essentially all the compact IRS sources in the central parsec coincide with local temperature enhancements in the model temperature distribution results. Temperatures in the field range from about 200 - 1300 K. ii) The IRS source positions are local minima in the emitting dust opacity distribution. The emitting dust opacity differs dramatically from the temperature distribution. The two significant peaks in dust emission opacity occur between the strong compact sources on the northern arm ridge, not coincident with them. The ridge of the opacity distribution coincides with the northern arm,however, there is no well-defined opacity feature corresponding to the east-west bar. iii) The opacity and temperature distributions are generally complementary. iv) The extinction opacity distribution of cool absorbing grains in the line-of-sight shows more structure than the directly observed 9.8 bm silicate feature strength distribution. Typical extinction opacity 7 2 at 12.4 pm across the extended ridge, peaking at 2.5 at IRS3. v) The model luminosity distribution is very similar to the mid-infrared brightness distribution. The compact IRS sources are very luminous, particularly the IRS3 and IRSl (L 3 x 105La), and they make a significant contribution to the energy budget of infrared complex. The luminosity along the extended ridge is generally about 5 x lo4 L a arcsec-' . The total luminosity integrated over the 25 arcsec (1 parsec) region modeled (corrected for extinction) is about N 1.4~ lo8 La. This represents roughly half the total luminosity of the entire Sgr A West complex. vi) No enhancement in mid-infrared emission or temperature is seen at the position of Sgr A*, setting a luminosity upper limit of L 1x lo3 La arcsec-2. N

N

N

N

3 Discussion If an extraordinary central, compact object located near Sgr A* was the dominant source of the observed infrared luminosity of the complex, the dust temperature distribution should show evidence for a centralized radiation field, and the temperature distribution would be expected to decrease radially at least as r-' . A radial temperature gradient would also be expected across the Sgr A West infrared ridge. The influence of a centrally concentrated luminosity source should be seen in the model results if such a phenomenon exists. The modeled temperature distribution shows a significant radial temperature gradient across the ridge, with the inner rim of the ridge (closest to Sgr A*) being significantly hotter than the outer rim. However, there is no evidence that Sgr A* perturbs the extended mid-infrared or radio emission surrounding it (e.g., the east-west bar), or interacts with any of the compact IRS sources The general sense is that the

Astron. Nachr./AN 324. No. SI (2003)

575

temperature decreases radially from a position nearer to IRS3-IRS7 than to Sgr A*. The influence of the luminous sources IRS3 and IRS7 has not figured prominently in earlier interpretations of the observations. Alternately, if embedded stellar luminosity sources were distributed within the dust ridge they would be expected to produce local temperature “hot spots” at the stellar source positions. The temperature distribution is also well correlated with the mid-infrared brightness distribution, especially near the compact IRS sources, and the luminosity distribution and the mid-infrared brightness distributions are nearly identical. These results suggest the presence of imbedded, luminous objects at the IRS source positions, which have created warm cocoons in the extended dust clouds surrounding them. The fact that the density of the emitting dust is anti-correlated with the mid-infrared emission at the IRS source positions (suggesting that the stars have eroded the nearby dust ridge) which also supports this view. There is a strong correlation between the mid-infrared and radio continuum distributions in the Galactic Center. The strong correlation suggests that the dust and gas are well-mixed, and that the radiation responsible for heating the dust and ionizing the gas is of common origin (Gezari and Yusef-Zadeh 1990). The temperature and opacity distributions are consistent with the presence of self-luminous objects imbedded at prominent the IRS source positions. However, temperatures are highest along the inner edge of the central cavity, while the dust opacity peaks radially further out. The higher temperatures along the inner ridge suggest heating by centrally located concentrated luminous sources, including IRS3 and IRS7. There is some evidence for physical interaction between the warm emitting dust and clusters of luminous stars, including dozens of hot He I emission line stars (Krabbe et al. 1991) and B[e] stars (Allen and Burton 1994).

4

Conclusions

There are presently no observations which tie the S g r A* radio source to the other infrared or radio emission in Sgr A West. The logical but unappealing conclusion that follows is that Sgr A* could be located anywhere along the line-of-sight within tens of parsecs of the Galactic Center. Yet the chance alignment of such a unique source with the core of Sgr A West is unlikely. Thus it is clear that either Sgr A* is quite benign (having no observable influence on its neighbors), or it is well separated along the line-of-sight from the extended dust and gas clouds surrounding it in the Sgr A West source complex. From the mid-infrared array data and the model results we conclude that: 1) the compact IRS sources are all local hot spots in the Galactic Center dust clouds, 2) the IRS sources are also local luminosity peaks; 3 ) the local dust density peaks are not correlated with the IRS sources, suggesting that 4) the IRS sources contain imbedded luminous stars; 5) the cluster of luminous He I emission line and B[e] stars are generally distributed in the central parsec, and do not coincide with the extended dust clouds in Sgr A West or the IRS sources, 6) the He I stars, B[e] stars, and the sources imbedded in the IRS objects can together account for the observed infrared and submillimeter luminosity of the Galactic Center.

References Allen,D.&Burton,M.1994,P.A.S.A.,11,191 Allen, D. A., Hyland, A. R. & Hillier, D. J. 1990, M. N. R. A. S., 244,706 Draine, B. T. & Lee, H. M. 1984, Ap. J., 285,89 Gezari, D. Y., W. Folz, L. Woods & Varosi, F. 1992, P. A. S. P., 104, 191 Gezari, D. Y. & Yusef-Zadeh, F. 1990, ”Astrophysics with Infrared Arrays”, A. I. P. Conference Series, 13,214 Krabbe, A., Genzel, R., Drapatz, S. & Rotaciuc, V. 1991, Ap. J. (Letters), 382, L19 Lynden-Bell, D. & Rees, M. J. 1971, M. N. R. A. S., 152,416 Rees, M. J. 1982, ”The Galactic Center”, ed. G. Riegler and R. Blanford, American Institute of Physics, Conf. Series, 83, 166 Rieke, G. H. and Lebofsky, M. J . 1982, ”The Galactic Center”, ed. G. Riegler and R. Blanford, American Institute of Physics, Conf. Series, 83, 194 Varosi, F. & Gezari, D. Y. 1993, “Astronomical Data Analysis Software and Systems 11”, P. A. S. P.Conf. Series, 52,393

516

D. Gezari et al.: Physical Conditions

B

LOS Opacity- 12.4 um

Source Luminosity 6

D Fig. 1 Model results for physical conditions in the central parsec. a) Dust temperature - Peaks occur at the positions of the prominent compact sources. IRS1, IRS5 and IRSlO are all about T=300 K, IRS3 is 600 K and IRS7 is 800 K. A general temperature gradient is evident across the northern arm and east-west bar. with warmer regions toward the inside of the central cavity. b) Source Opacity at 12.4 um - Source opacity at 12.4 um ranges from T = 0.003 at IRS3 to 7- = 0.040 on the ridge between the strong compact IRS sources and radially outward, away from the central cavity. c) Extinction opacity at 12.4 pm - The LOS extinction opacity distribution is smooth and typically about T 2 across the field, suggestion that the principal contribution is from interstellar grains along the line of sight, except at IRS3 where a circumstellar contribution is evident. d) Source Luminosity - Luminosity of warm grains (L&rcsec*) corrected for total LOS extinction, ranging from about 4 x lo4 to 2 x lo5 for IRS sources on the ridge and 3 x 10' for IRSJ. The total luminosity over the central parsec is about 1.4 x 10'. No enhancements are seen at the position of Sgr A*, about 5.7" south of IRS7 (0,O). N

Astron. Nachr./AN 324, No. S1,577-581 (2003) / DO1 10.1002/asna.200385092

-

LINCNRVANA The LBT Near-Infrared Interferometric Camera C. Straubmeier", A. Eckart', T. Bertram', and T. Herbst' I

I. Physikalisches Institut, University of Cologne, Zulpicher Strak 77,50937 Cologne, Germany Max-Planck-Institut fur Astronomie. Konigsstuhl 17,69117 Heidelberg, Germany

Key words astronomical instrumentation, interferometry

Abstract. The development of the interferometric near-infrared camera LINCNRVANA for the Large Binocular Telescope (LBT) is an international collaboration of the Max-Planck-Institute for Astronomy in Heidelberg (PI institute), the Osservatorio Astrofisico di Arcetri and the I. Physikalische Institut of the University of Cologne. LINCNIRVANA will allow for interferometric imaging in the near-infrared JHK bands with high angular resolution (about 9, 12 & 17 mas at respective wavelengths) over a wide field of view (FOV % 100 urcsec2). This FOV is sufficiently large to observe the entire central stellar cluster of the Galactic Center with one single exposure [Rubilar 20011, a unique possibility that makes LINCiNIRVANA the detector system of choice for scientific investigations of respective stellar populations and dynamics. In this article we summarise the implicationsof the Fizeau-type optical layout of the LBT for interferometric imaging and present the principle of operation and current design concept of the LINCiNIRVANA camera.

1 The Large Binocular Telescope (LBT) The Large Binocular Telescope (LBT) represents a revolutionary and fascinating new type of astronomical telescope in terms of optical principle of operation and physical dimensions. In contrast to all other largescale interferometric telescopes (either already operational or currently under construction) the optical layout of the LET resembles a Fizeau interferometer, with both primary mirrors sharing a common mount (Hill & Salinari 1998, Angel et al. 1998, see also Fig. 2 & Fig. 3). This special optical design has two important consequences for interferometric operation: In contrast to non-Fizeau interferometers like VLTI' (von der Liihe et al. 1997, Eckart et al. 1997, Glindemann & Uv&que2000) or Keck I & I1 (Booth et al. 1999, Swanson et al. 1997), where the interferometric baselines are defined by the relative geographical locations of the individual telescopes, the baseline of a Fizeau interferometer is not fixed with respect to the ground, but with respect to the movable common mount of the two mirrors. Consequently, the geometrical projection of the baseline onto the plane of the incoming wavefront (i.e. perpendicular to the line of sight) remains constant in time (at 14.4 meters in the case of LBT), while the pointing of the telescope follows the apparent nightly motion of an observed astronomical source. Since the angular resolution of an interferometer is a linear function of its baseline this parameter remains constant in time as well, which allows for direct interferometric imaging of an astronomical target without the need of complicated optical delay lines to compensate changing baseline projections. Furthermore, the fixed baseline allows for homothetic mapping, i.e. an optical camera design where the spatial pupil configuration is preserved up to the exit pupil. This special arrangement provides an exceptionally large (in terms of optical * Corresponding author: e-mail: cstraubmQph1.uni-koeln.de,Phone: +492214703552, Fax: 4 9 221 4705162

' VLTI - Very Large Telescope Interferometer

@ 2003 WILEY-VCH Verlag GmhH & Cu KGaA. Weinhem

578

C. Straubmeier et al.: LINCLNIRVANA - The LBT Near-Infrared Interferometric Camera

Fig. 1 Enclosure of the Large Binocular Telescope on Mt. Graham near Tucson, Arizona (USA). This picture was taken in June 2002, at a time when the enclosure reached its full operabihty, i.e. provides a rotable

and weather-proof housing with movable observing doors. interferometers) field of view of several arcminutes, which is mainly limited by the performance of the Adaptive Optics systems and the affordable cost to cover the diffraction limited focal plane with astronomical detector arrays. The linear separation of 14.4 meters between the centres of the two primary mirrors is small compared to their respective diameters of 8.4 meters. While the large diameters of the primary mirrors ensure a large light-collecting area of about 110 m2 and therefore a high sensitivity of the telescope for faint sources, the separation of the two mirrors defines the interferometric angular resolution. At wavelengths of the near-infrared J band, the baseline of 14.4 meters corresponds to an optical resolution of about 9 milliarcseconds. In comparison to the high angular resolutions achievable with non-Fizeau interferometers like the VLTI or Keck I & I1 (in the range of microarcseconds) this might be only a moderate value. However, with respect to the wide field of view provided by the LBT and the availability of large NIR detector arrays, this moderate resolution makes it possible to cover almost 100 arcseconds' on a standard 2k x 2k pixels' detector array (assuming Nyquist sampling). To summarise the statements listed above, the interferometric capabilities of the Large Binocular Telescope will offer an exceptional and so far unprecedented combination of wide field of view (10 to 120 arcseconds depending on the used detector array, the observed wavelength and the performance of the Adaptive Optics Systems), high angular resolution (approx. 9 milliarcseconds at X = 1.25 ,urn) and large collecting area (about 110 m2>.While other astronormcal interferometers (e.g. VLTI or Keck I & 11) will be able to reach much higher angular resolutions, their field of view will be less than one arcsecond wide and therefore too small to help answer many interesting astrophysical questions. Based on a multinational collaboration of several universities and astrophysical research institutes in the USA, Germany and Italy, the Large Binocular Telescope is currently under construction on Mt. Graham near Tucson, Arizona (USA). After the voluminous enclosure was finished successfully in 2002 (see Fig. 1) the LBT construction work is focussed on the integration of the telescope structure and mechanics onto the observing platform.

519

Astron. Nachr./AN 324, No. S1 (2003)

Fig. 3 Test-erection of the mechanical structure of the LBT in June 2001 at Ansaldo Inc. at Milan, Italy. To finalise the manufacturing process, the structure and mechanics of the telescope system were assembled to full large pentagonal box at the centre of the image (orange operation for testing and verification tasks. At this phase color in a colored reproduction). Furthermore the rays the telescope structure was steerable in both axes and the of incoming light are shown in transparent blue and a proper function of the system has been checked successsketched person is included for scale comparison. fully.

Fig. 2 The sketch above shows a raytracing model of the telescope structure of the LBT. The position and the approximate size and shape of the LINCNIRVANA camera system are represented by the

With respect to astronomical observations of the Galactic Center, the geographical location of the LBT (= 33"N) allows for about 6 hours of observing time per night and a respective rotation of the interfero-

metric image axis of more than 60". While still showing an appreciable ellipticity of the interferometric beam shape this rotation will be sufficient for an effective two dimensional image restoration. In parallel with the on-site work on Mt. Graham several groups of the participating institutes and universities are developing a variety of astronomical camera systems for the LBT at their local workshops. Due to the nature of the LBT twin-telescope, the scientific interest is concentrated to a high degree on interferometric detector systems. However, because of the high complexity and cost of such devices, there will be only two interferometric camera systems installed at the LBT the LBT nulling beam combiner LBTI (built under the leadership of the University of Arizona (Hinz 2001) and the LINCNIRVANA system described in this article. The non-interferometric first light of the LBT using only one primary mirror is scheduled for summer 2004. The first interferometric observations with both primary mirrors installed are planned to start one year later in summer 2005.

2 LINCNIRVANA

- The interferometric near-infrared imaging camera

for the LBT The LINC2/NIRVANA3system will allow for scientific imaging observations at near-infrared JHK wavelengths with an unprecedented combination of high angular resolution (= 9 mas at X = 1.25 pm and M 17 mas at X = 2.4 p m ) and wide field of view (up to 10 x 10 arcsec2 limited mainIy by the physical size of the used HAWAII I1 detector array). As explained before, the Fizeau-type interferometric layout of the LBT removes the need of optical delay lines and the wavefronts of the two single-eye telescopes can be superimposed directly at a common focal plane. However, for achieving time-stable interferometric performance both optical arms of the

* LBT hterferometric Camera Near-InfraredVisible Adaptive lnterferometa for Astronomy

580

C. Straubmeier et al.: LINCNIRVANA - The LBT Near-Infrared Interferometric Camera

Fig. 4 Results obtained by numerical simulations of the interferometric PSF in the combined focal plane of LINC/NIRVANA. For the graphical results presented above a perfect operation of the MCAO systems has been assumed, i.e. the incoming wavefronts of the single-eye telescopes were considered to be absolutely flat. Furthermore, only monochromatic light with an infinite coherence length has been considered. Under this simplifying assumption no chromatic dispersion effectsare present and the horizontal fringe pattern extends up to infinitely high interferometnc orders, i.e. the interferometric intensity modulation is detectable even far from the central maximum. However, considering physically plausible filter bandwidths the observable fringe contrast will decrease with increasing piston. Fig. 4a) presents a colour coded contour plot of the simulated two-dimensional PSF. Fig. 4b) is based on the same data but shows the intensity distribution along the interferomtnc baseline of LINCNIRVANA, together with the results of two numerical fits to the data. Red colour corresponds to the numerical fit including the interferometric intensity modulation. The blue coloured fit represents the Airy distribution of a single-eye 8.4m aperture. Both pictures use logarithmic scaling on the intensity axis.

beam-combiner have to be operated at the diffraction limit of the 8.4 m primary mirrors, and the pistonic phase difference between the two incoming wavefronts has to be eliminated by active optics as well. To maximise sky coverage and field of view (FOV) LINCMRVANA uses Multi Conjugate Adaptive Optics (MCAO)(Ragazzoni et al. 2002, Berkefeld et al. 2001, Diolarti et al. 2001) for the correction of the single-eye wavefronts. So far, the Adaptive Optics (AO) systems of current telescopes (e.g. ALFA (Eckart et al. 2000, Davies et al. ZOOO), NAOS (Brandner et al. 1998, Lenzen et al. 1998) are analysing and correcting the aberrations of the incoming wavefront for only one reference object and one only conjugated atmospherical layer (i.e. the ground layer). Using only one bright star to drive the A 0 loop, the diffraction limited FOV around this reference star is limited by the size of the isoplanatic patch of the atmosphere (about 20 arcsec in J band). With respect to the celestial distribution of suitably bright reference stars the usage of standard A 0 systems is limited to about 1 % of the sky. MCAO overcomes these limitations by using multiple reference stars (up to 20 for the case of LING/ NIRVANA) and by analysing and correcting the incoming wavefront for several conjugated atmospheric layers (up to three for the case of LINCNIRVANA). Therefore, the limiting magnitudes of suitable individual reference stars are raised significantly and the angular size of the diffraction limited FOV can be extended up to a few arcminutes. Using MCAO the resulting sky coverage of LINC/NIRVANA is higher than 86 % for galactic latitudes b < 20" and still amounts to about 12 % at the Northern Galactic Pole. Fig. 4 shows the result of numerical simulations of the interferometric point spread function (PSF) at the combined focal plane of LINCNIRVANA. Due to the Fizeau type optical layout of the LBT the twodimensional PSF (left picture) exhibits different angular resolutions in parallel and perpendicular to the interferometric baseline of the telescope, i.e. the connection line of the centres of the two primary mirrors. Perpendicular to this baseline (vertical axis in Fig. 4a and blue line in Fig. 4b) the angular resolution still resembles the Airy distribution of the 8.4 m aperture of the primary mirrors (= 37 mas in J and w 72 mas

Astron. NachrJAN 324,No. S1 (2003)

58 I

in K-band). Nevertheless, in parallel to the baseline (horizontal axis in Fig. 4a and red trace in Fig. 4b) the angular resolution is defined by the interferometric intensity modulation arising from the 14.4 m separation of the two primary mirrors (% 9 mus in J and 17 mas in K-band). However, despite the axial asymmetry in angular resolution of a single exposure with LINUNIRVANA, it is possible to generate images of an astronomical source showing the maximum interferometric angular resolution along all axes. Since the orientation of the interferometric baseline (and thus the image axis of improved angular resolution) is not fixed with respect to the reference frame of the observed sky, but follows the daily rotation of the earth, the baseline rotates by 180 degrees while the telescope is tracking an astronomical source during 12 hours of observing time. Using the data of several exposures (which have been taken at different times of the night) and a suitable numerical algorithm it is therefore possible to generate a reconstructed image, which shows the improved interferometric angular resolution along both axes and all angles in between.

=

Acknowledgements The contribution of the I. PhysikalischeInstitut to LINCDJIRVANA is supported in parts by the Deutsche Forschungsgemeinschaft (DFG) via grants SFB 494, HBFG #I 11-519 & #111-520 and Verbundforschung #IDlCUlYQ.

References Angel, R., Hill, J., Strittmatter,P., Salinari, P. & Weigelt, G. 1998, SPIE, 3350, 881 Berkefeld, T., Glindemann, A. & Hippler, S. 2001, ExA, 1 1 , I , 1 Booth, A. J., Colavita, M. M., Shao, M.,Swanson, P. N., van Belle, G. T, et al. 1999, ASPConf, 194,256 Braldner, W., Rousset, G., Lenzen, R., Huhin, N., Lacomhe, F., et al. 2002, Msngr, 107, I Davies, S., Eckart, A., Hackenberg, W., Ott, T., Butler, D., et al. 2000, ExA, 10, 1, 103 Diolaiti, E., Ragazzoni, R. &Tordi, M. 2001, A&A, 372,710 Eckart, A., Genzel, R., Hofmann, R., Drapatz, S., Katterloher,R., et al. 1997, SVLTwork, 259 Eckart, A., Hippler, S., Glindemann, A., Hackenberg, W., Quirrenhach, et al. 2000, ExA, 10, 1, 1 Glindemann, A,, LCvkque, S. 2000, FEPCconf, 468 Hill, J. & Salinari, P. 1998, SPIE, 4004, 36 Hinz P. 2001, AAS, 198,5105 Lenzen, R., Hofmann, R., Bizenherger, P. & Tusche, A. 1998, SPIE, 3354,606 von der Liihe, O., Derie, F., Koehler, B., Leveque, S., Paresce, F., Verola, M. 1997, SPIE, 2871,498 Ragazzoni, R., Diolaiti, E., Farinato, J., Fedrigo, E., Marchetti, E., Tordi, M., Kirkman, D. 2002, SHE, 4494, 52 Rubilar, G. & Eckart, A. 2001, A&A, 374,95 Swanson, P., Colavita, M., Boden, A., van Belle, G., Shao, M. 1997, AAS, 191,0912

This Page Intentionally Left Blank

Astron. NachdAN 324, No. S1.583-589 (2003) / DO1 10.l002/asna.200385049

Hot Molecular Gas in the Central 10 Parsecs of the Galaxy R. M. Herrstein”’ and P. T. P. Ho’ MS-I0,60 Garden St., Cambridge, MA 02138

Key words Galactic Center, ISM, radio lines PACS 04A25 We present results from observations of NH3(1,1), (2,2), (3,3), and (6,6) with the Very Large Array. The data sample the inner 10 pc (4‘) of the Galaxy and have a velocity coverage of -140 to +130 !an S K I . The velocity-integrated NH3(3,3) image shows that the Sgr A East supernova remnant is impacting the 50 km s-’ GMC in the northeast, the northern ridge in the north, and the western streamer in the west. These results imply that the Sgr A East has a large effect on the molecular environment near Sgr A* and may be pushing much of the molecular gas away from Sgr A*. The physical properties of the western streamer and its relation to Sgr A East are discussed in detail. We also summarize the detection of hot molecular gas less than 2 pc from Sgr A* in projected distance. This gas is seen only in NH3(6,6) and has line widths of 75-85 km s p l , indicating that it is physically close to the nucleus.

1 Introduction At a distance of only 8.0 k 0.5 kpc (Reid 1993), the Galactic Center provides a unique opportunity to study in detail the environment around a supermassive black hole. It is now generally accepted that a black hole of 2.6 x 1O6MDis located at the dynamical center of the Galaxy (Eckart & Genzel 1997; Ghez et al. 1998, 2000; Schodel et al. 2002). In the radio, emission from the inner region of the accretion flow is observed as the strong (- 1 Jy) source, Sgr A*. Sgr A* is surrounded by arcs of ionized gas (Sgr A West) that appear to be feeding the nucleus (Lo & Claussen 1983; Roberts & Goss 1993). These arcs are, in turn, surrounded by an apparent “ring” of molecular material at a radius of 2 pc from Sgr A* called the circumnuclear disk (CND, Giisten et al. 1987). Sgr A*, Sgr A West, and the CND appear to be located near the front edge of the expanding supernova remnant (SNR),Sgr A East, but the exact position of the features along the line-of-sight is very difficult to determine (Pedlar et al 1989; Maeda et al. 2002). For the past two decades, the origin of the clouds in the CND and the mini-spiral has remained unclear. Many attempts have been made to detect connections between two nearby giant molecular clouds (GMCs) and the CND. Okumura et al. (1989), Ho et al. (1991), and Coil & Ho (1999, 2000) detect a long filamentary “streamer” in NH3(l,1) and (2,2)emission that connects the “20 km s-l cloud” (M-0.130.08; Gusten, Walmsley, & Pauls 1981) to the southeastern edge of the CND. A small velocity gradient along this “southern streamer” as well as heating and increased line widths as the streamer approaches the Galactic Center indicate that gas may be flowing from the 20 km s-l GMC towards the circumnuclear region. This connection has also been observed in HCN(3-2) (Marshall et al. 1995), 13CO(2-l) (Zylka, Mezger , & Wink 1990), and 1.1 mm dust (Dent et al. 1993). Other candidates for connections between the GMCs and the CND include a possible connection between the northeastern edge of the CND and the N

-

* Corresponding author: e-mail: [email protected], Phone: 617 495 4142, Fax: 617 4967554

@ 2003 WILEY-VCH Verlag GmbH & Co KGaA, Weinhem

R. M. Herrnstein and P. T. P.Ho: Hot Molecular Gas

584

“50 km s-l cloud” (M-0.03-0.07; Ho 1993) and a second connection between the 20 km s-l GMC and the southwest lobe of the CND (Coil & Ho 1999,2000). Spectral line observations of the Galactic center are inherently difficult due to the large range of velocities in the region. In order to detect all of the emission from the CND, a velocity coverage of at least ill0 km s-l is necessary. Near Sgr A*, clouds with velocities as high as -185 km s-l have been detected (Zhao, Goss, & Ho 1995). Previous NH3 observations by Coil & Ho (1999,2000) focused on the kinematics of the 20 and 50 km s-l GMCs using a velocity window of 75 km s-’ and a velocity resolution of 4.9 km s-‘. Pointings were made towards Sgr A* as well as in the direction of these GMCs in the south and east. In order to produce a more complete picture of the molecular environment at the Galactic center, we have imaged the central 10pc of the Galaxy in NH3(1,1), (2,2), (3,3), and (6,6) with the Very Large Array’ (VLA). These data fully sample the inner 10 pc (4’) of the Galaxy and have a velocity coverage of -140 to +130 km s-l. In this article, we summarize two of the most important results of this project, but we also refer the reader to McGary, Coil, & Ho (2001), McGary & Ho (2002), and Herrnstein & Ho (2002,2003) for more detailed discussions. The observations and data reduction are summarized in Section 2. Section 3 focuses on the velocity-integrated NH3(3,3) image, which indicates that Sgr A East has a large effect on the molecular gas in the region. In Section 4, we present the results of our observations of NH3(6,6) at the Galactic Center and the detection of hot molecular gas less than 2 pc in projected distance from Sgr A*.

2 Observations and Data Reduction The metastable (J=K) NH3(J,K) rotation inversion transitions at N 23 GHz have proven to be useful probes of dense ( 104-105 cmU3)molecular material near the Galactic center. They tend to have a low optical depth and a high excitation temperature at the Galactic center, making them almost impervious to absorption effects (Ho & Townes 1983). Satellite hyperfine lines separated by 10-30 km s-l on either side of the main line enable a direct calculation of the optical depth of NH3, although the large line widths at the Galactic center make it necessary to model effects due to blending of the Line profiles (Herrnstein & Ho 2002). In addition, line ratios of different transitions can be used to calculate the rotational temperature, TR,of the gas. NH3(l,1), (2,2), and (3,3) were observed with the VLA in 1999 March. Observations were made in the D north-C array, which provides the most circular beam at the low elevation of the Galactic Center. With a maximum projected baseline of 1 km, this smallest configuration of the VLA provides the highest sensitivity to extended features such as long, filamentary streamers. A five-pointing mosaic was centered 6~~0 =0-29”00’26”.6), with the remaining four pointings offset by on Sgr A* (a2000 = 17h45m40s.0, N 1’ to the northeast, northwest, southeast and southwest. The resulting data fully sample the central 4’ (10 pc) of the Galaxy. By using a velocity resolution of 9.8 km s-’, we were also able to obtain a velocity coverage of -140 to +130 km s-’, including almost all of the velocities observed near the nucleus. With these data, we can probe the morphology and kinematics of the entire CND as well as the surrounding molecular material. Each pointing was calibrated separately using AIPS. (A detailed discussion of the data reduction can be found in McGary, Coil, & Ho 2001.) The data were then combined in the uv-plane and deconvolved using the Maximum Entropy Method in MIRZAD. A Gaussian taper was applied to the data to aid in the detection of extended features. The final beam size for all three transitions is roughly 15” x 13” with a PA of 0”. For the NH3(3,3) velocity-integrated image, the lo noise level (calculated assuming line emission typically appears in seven channels) is 633 = 0.33 Jy beam-’ km s-l (McGary, Coil, & Ho 2001). Observations of NH3(6,6) were made on 2001 October 1 and November 16 with a setup and spatial coverage identical to our previous N H 3 data. Observations of this line, which has a frequency of 25 GHz,

-

The National Radio Astronomy Observatory is a facility of the National Science Fonndatlon operated under cooperative agreement by Associated Universities, Inc.

585

Astron. Nachr./AN 324, No. S 1 (2003)

1w

0

-100

RA onset ( w c ~ ~ c JZOOO) :

Fig. 1 a) Velocity-integrated NH3(3,3) emission in contours in steps of 4a33 overlaid on a 6 cm continuum emission image made by Yusef-Zadeh & Moms (1987). The position of Sgr A* (6, = 0", 66 = 0") is marked by a star. The northern ridge, western streamer, and 50 km s-' GMC lie along the edge of Sgr A East (McGary, Coil, & Ho 2001). b) Velocity-integrated NHs(6.6) emission in contours in steps of 3066 overlaid on the same continuum image. The NH3(6,6) image is dominated by emission within I .5 pc (40") of Sgr A*, interior to the circumnuclear disk (Herrnstein & Ho 2002).

only recently became possible at the VLA after the upgrade of the 23 GHz receivers. These data represent the first successful observation of this line with the VLA. The data calibration and image deconvolution also used the same method as our original NH3 observations. The final beam size after application of a Gaussian taper (FWHM=lO") to the uv data is 12.0" x 9.2" with a PA of -1.52'. For the velocityintegrated image, the l a noise level is a 6 6 = 0.23 Jy beam-' km s-'.

3 The Effect of Sgr A East on the Molecular Environment Figure l a shows the velocity-integrated NH3(3,3) image in contours overlaid on a 6 cm continuum image (Yusef-Zadeh & Morris 1987). The contours are in steps of 4033. In the continuum image, the point source Sgr A* (Acu = 0", A6 = 0") is the brightest feature and is labeled with a star. The Sgr A East shell appears as faint, extended emission with a roughly circular shape and centered slightly to the east of Sgr A*. NH3(3,3) is detected throughout most of the central 10 pc and the major features discussed in this paper are labeled in Figure la. The gain of the telescope goes to zero at the edge of the mosaic, roughly 5 pc from Sgr A*. The 50 km sP1 GMC extends beyond the mosaic to the northeast while the 20 km s-l cloud is located almost entirely outside our mosaic, to the south of the southern streamer. However, the northern ridge and western streamer are located within the edge of our mosaic and their filamentary morphologies are real. Some of the clouds in our image, including SEI and the southern streamer, appear to be kinematically connected to gas near the nucleus (McGary, Coil, & Ho 2001). However, much of the gas appears to lie along the edge of the expanding Sgr A East supernova remnant (SNR). In the northeast, Sgr A East is impacting the 50 km s-l GMC. Originally, it was believed that the molecular cloud had been pushed away from the nucleus by the expanding shell (Mezger et al. 1989; Pedlar et al. 1989). However, with a mass of 6 x lo4 Ma (Gusten, Walmsley, & Pauls 1981; Mezger et al. 1989), it seems unlikely that Sgr A N

R. M. Herrnstein and P. T. P. Ho: Hot Molecular Gas

586

100

50

0 Velocity (km s-')

-50

-100

Fig. 2 a) Velocity-integrated NH3(3,3) image showing the location of the position-velocity cut through the western streamer. b) Position velocity diagram showing a I km s-l gradient along the entire length of the western streamer. The 0" position is at the southern end of the cut. Emission at 0 km sC1 at a position of 30" is associated with the northern edge of the 20 km s-l GMC.

East has a large effect on the 50 km s-' GMC. The physical properties of the NH3 along the edge of the 50 km s-' GMC appear to support this theory. The intrinsic line width of gas in the 50 km s-l cloud is 15 km s-', roughly equal to the mean line width throughout the inner 10 pc. In addition, one would expect the gas to be heated if it is being moved by the expanding shell. The rotational temperature of the gas can be calculated from our NH3( 1,l) and (2,2) data by

-

where is the ratio of the main hyperfine line of NH3 (2,2) and ( 1,l) and r, ( 1 , l ) is the opacity of the NH3( 1,l) main hyperfine line (Ho & Townes 1983). This equation assumes equal excitation temperatures and beam filling factors for both transitions. For the 50 km s-l cloud, we calculate a temperature of 20 K (Herrnstein & Ho 2003), corresponding to a kinetic temperature of 25 K (Walmsley & Ungerechts 1983; Danby et al. 1988). With a slightly elevated rotational temperature of 25 K, the northern ridge may have been heated by the impact of Sgr A East. However, it is the western streamer that appears to be most strongly affected by the expansion of the SNR. The western streamer has a striking velocity gradient of 1 km s-l arcsec-' (25 km s-' pc-l) along its entire length of 150" (6 pc) (See Figure 2). The velocity gradient can be explained by a ridge of gas moving outwards with the expansion of Sgr A East and highly inclined to the line-ofsight. This scenario would place the southern part of the streamer on the front side of the shell. The western streamer also shows the largest rotational temperature of any feature in our map, with T R ZM~ 50 K, or a kinetic temperature near 80 K (Hermstein & Ho 2003). Assuming an abundance of NHx relative to Hz (X(NH3)) of lo-' (Townes et al. 1983; Harju, Walmsley, & Wouterloot 1993), the total mass of the western streamer is 10' Ma, more than two orders of magnitude less than the 50 km s-' GMC (Herrnstein & Ho 2003). Observations of hot NH3 cores indicate that X(NH3) is elevated in warm environments and can (Hiittemeister et al. 1993). The mass estimate for the western streamer is therefore be as high as assumed to be an upper limit. It is not surprising that this less massive feature shows much more evidence for interaction with S g r A East than the 50 km s-' GMC. N

N

N

587

Astron. Nachr./AN 324, No. S1 (2003)

0

M

i

B

?

RA

O!l&

: .c r a (

noaa)

Fig. 3 Left: Velocity-integrated NH3(6,6) in contours overlaid on velocity-integrated HCN( 1-0) emission (Wright et al. 2001). Sgr A* is located at (6, = Of’, 66 = O f f ) and is marked by a small circle. Much of the NH3(6,6) is located interior to the clumpy ring of the CND seen in the HCN(1-0) image. Right: Spectra at the three positions labeled in Figure 3a showing the large line widths of gas interior to the CND.

In McGary, Coil, & Ho (2001), we detected evidence for a connection between the northern ridge and the northeastern lobe of the CND. A physical connection between these two features must imply that Sgr A East is close to the CND. The large impact of Sgr A East on molecular gas in the central 10 pc has led us to question whether observed connections between the GMCs and CND represent infalling material. It is possible that the Sgr A East shock front recently passed through Sgr A*, and is now pushing material away from the supermassive black hole. This scenario has also been used to explain other features at the Galactic center including an ionized gas halo surrounding Sgr A East (Maeda et al. 2002) and high negative velocity features seen near Sgr A* (Yusef-Zadeh, Melia, & Wardle 2000). Although more data are necessary to confirm or refute this theory, it is clear that a detailed understanding of Sgr A East may be necessary before we can fully understand Sgr A* and the central few parsecs of the Galaxy.

4 Hot Molecular Gas near Sgr A* Molecular gas is expected to be heated as it approaches the nucleus. An increased rotational temperature in the part of the southern streamer closest to the CND has been used to argue that the cloud is physically close to Sgr A* (Coil& Ho 1999). However, the reality of this effect has remained in doubt because the emission also becomes faint near Sgr A*. In Figure I a, the NH3(3,3) emission weakens and almost disappears near Sgr A*. The lack of emission in the inner 2 pc of the Galaxy could signal the existence of a molecular hole near Sgr A*. We suspected that temperatures become so high near Sgr A* that even the NH3(3,3) transition is no longer well-populated. In order to detect hot molecular gas near Sgr A*, we observed the central 4’ (10 pc) of the Galaxy in NH3(6,6) using the new 23 GHz receivers at the VLA. At 412 K above ground, NH3(6,6) bas more than three times the equivalent energy of NH3(3,3). Figure 1b shows velocity-integrated NH3(6,6) emission in contours in steps of 3 ~ 6 6overlaid on the same 6 cm continuum image. NH3(6,6) is detected in many of the features seen in lower NHs transitions, including the western streamer, but the image is dominated by emission less than 1.5 pc (40”) in projected distance from the nucleus. Figure 3a overlays the velocity-integrated NH3(6,6) image on a grey-scale image of velocity-integrated HCN(1-0) (Wright et al. 2001). The HCN(1-0) shows the bright clumps that

R. M. Hermstein and P. T. P. Ho: Hot Molecular Gas

588

form the inner ring of the CND. Much of the NH3(6,6) emission comes from a region interior to the CND in projection. Spectra taken at three positions interior to the CND are plotted in Figure 3b. The large line widths of 75-85 km s-’ in the central 1.5 pc indicate that this gas is physically close to the nucleus. Spectra are well-fitted by Gaussian profiles. Although the shape of the line profiles are similar to those from the CND for other molecules (e.g. Wright et al. 2001), the emission appears to be kinematically independent of material in the CND (Hemstein & Ho 2002).

50

-

-

-z -

-

0-

0

--

-

-50-

a

-

-100

-

-

3.4

-150 -

Fig. 4 Ratio of m a n line NHs(6.6) and (3,3) emission in grey scale overlaid on contours of velocity-integrated NH3(6,6). The velocity channel for the main line is chosen using the NH3(6,6) image cube. Line ratios are calculated for every point with (6.6) emission > 3033. For pixels with faint (3,3) emission, 3033 is used to estimate a lower limit for the line ratio. Sgr A* is located at (0,O).

The detection of NH3(6,6) emission from a cloud is a strong indication that the cloud is very warm. In our data, clouds in which NH3(6,6) is detected tend to have high (2,2)-to-( 1,l) rotational temperatures. Figure 4 shows the line ratio of NH3(6,6) to (3,3). The NH3(6,6) has been convolved to the resolution of the NH3(3,3) data. Because NH3(3,3) is so faint near the nucleus, we use 303~as the NH3(3,3) flux density to calculate a lower limit of the line ratio for those pixels with faint (3,3) emission. Interior to the CND, line ratios of NH3(6,6) to (3,3) exceed the theoretical limit of 2.3 (Herrnstein & Ho 2002). These large line ratios may the the result of a larger filling factor for NH3(6,6), or the dynamic range of the NH3(3,3) data may be limited by nearby bright emission (McGary, Coil, & Ho 2001). It is unlikely that the gas is out of thermal equilibrium because the equilibration time is lo3 s. Line widths are quite large in the region (50 - 80 km s-l) making it unlikely for the NH3(6,6) population to be inverted. The large line widths also make absorption of NH3(3,3) by an un-associated, cool foreground cloud unlikely. N

If the NH3(6,6) emission originates in a radiatively heated cloud, then it is possible that the NH3(3,3) is absorbed by cooler material in the same cloud (with the same line width) that has been shielded from the radiation. The NH3(6,6) would be unaffected by absorption because the cooler gas would contain almost no NH3(6,6). This shielded layer of cool gas must be located between the heated layer of the cloud and the observer. Therefore, if the clouds with line ratios greater than 2.3 are heated by photons emanating from the nucleus, then they must be located in front of the nucleus along the line-of-sight.

Astron. Nachr./AN 324. No. S1 (2003)

5

589

Summary

Using the VLA, we have observed NH3(1,1), (2,2),(3,3), and (6,6) to investigate the physical properties of molecular gas in the central 10 pc of the Galaxy. These data fully sample a 4' field centered on Sgr A* with a velocity coverage of -140 to +130 km s-'. Much of the NH3 emission originates along the edge of Sgr A East, indicating that this expanding shell is greatly affecting the physical environment near Sgr A*. In addition, a hot component of molecular gas appears to be located less than 2 pc from Sgr A*. This gas has line widths of more than 70 k m s-l and high (6,6)-to-(3,3) line ratios, indicating that it is physically close to the nucleus.

References Coil, A. L. & Ho, P. T. P. 1999, ApJ, 513,752 Coil, A. L. & Ho, P. T. P. 2000, ApJ, 533, 245 Danby, G., Flower, D. R., Valiron, P., Schilke, P., & Walmsley, C. M. 1988, MNRAS, 235, 229 Dent, W. R. F., Matthews, H. E., Wade, R. , & Duncan, W. D. 1993, ApJ, 410,650 Eckart, A. & Genzel, R. 1997, MNRAS, 284,576 Ghez, A. M., Klein, B. L., Morris, M. , & Becklin, E. E. 1998, ApJ, 509,678 Ghez, A. M., Moms, M., Becklin, E. E., Tanner, A., & Kremenek, T. 2000, Nature, 407,349 Giisten, R., Genzel, R., Wright, M. C. H., Jaffe, D. T., Stutzki, J. , &Harris, A. I. 1987, ApJ, 318, 124 Giisten, R., Walmsley, C. M. ,& Pauls, T. 1981, A&A, 103, 197 Harju, J., Walmsley, C. M., & Wouterloot, J. G. A. 1993, A&AS, 98, 51 Hermstein, R. M. & Ho, P. T. P. 2002, ApJL, 579,83 Hermstein, R. M. & Ho, P. T. P., in pre aration Ho, P. T. P. 1993, in Proc. of the 2"'Cologne-Zermatt Symposium, The Physics and Chemistry of Interstellar Molecular Clouds, ed. G. Winnewisser & G. Pelz (New York Springer), 33 Ho, P. T. P., Ho, L. C., Szczepanski, J. C., Jackson, J. M., Armstrong, J. T. , & Barrett, A. H. 1991, Nature, 350, 309 Ha, P. T. P. & Townes, C. H. 1983, ARA&A, 21,239 Huttemeister, S., Wilson, T. L., Henkel, C., & Mauersberger, R. 1993, A&A, 276, 445 Lo, K. Y. & Claussen, M. J. 1983, Nature, 306, 647 Maeda, Y. et al. 2002, ApJ, 570, 671 Marshall, J., Lasenby, A. N. , & Harris, A. I. 1995, MNRAS, 277, 594 McGary, R. S., Coil, A. L., & Ho, P. T. P. 2001, ApJ, 559,326 McGary, R. S. & Ho, P. T. P. 2002, ApJ, 577, 757 Mezger, P. G., Zylka, R. Salter, C. J., Wink, J. E., Chini, R. Kreysa, E . , & Tuffs, R. 1989, A&A, 209,337 Okumura, S. K., et al. 1989, ApJ, 347, 240 Pedlar, A., Anantharamaiah, K. R., Ekers, R. D., Goss, W. M., van Gorkom, J. H., Schwarz, U. J., Zhao, J.-H. 1989, ApJ 342,769 Reid, M. J. 1993, ARA&A, 31,345 Roberts, D. A. & Goss, W. M. 1993, ApJS, 86, 133 Schodel, R. et al. 2002, Nature, 419, 694 Townes, C. H., Genzel, R., Watson, D. M., & Storey, J. W. V. 1983, ApJL, 269, 11 Walmsley, C. M. & Ungerechts, H. 1983, 122, 164 Wright, M. C. H., Coil, A. L., McGary, R. S., Ho, P. T. P., & Harris, A. I. 2001, ApJ, 551, 254 Yusef-Zddeh, T., Melia, F., Wardle, M. 2000, Science, 287, 85 Yusef-Zddeh, E & Moms, M. 1987, ApJ, 320,545 Zhao, J.-H., Goss, W. M., & Ho, P.T. P. 1995, 450, 122 Zylka, R. Mezger, P. G., &Wink, P. E. 1990, A&A, 234, 133

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.591-596 (2003)/ DO1 10.1002/asna.200385089

The ISM and Stellar Distributions Near Sgr A* Nick Scoville”, Susan R. Stolovy2,and Micol Christopher’ I Astronomy 105-24, Caltech, Pasadena, CA 9 1125, USA * SRTF Science Center, MS 220-6, Caltech, Pasadena, CA 91125 Key words Galaxy: center,dust, extinction, infrared ISM and ionized hydrogen Abstract. HST-NICMOS imaging in the 1.87pmPa line and I .90pm continuum and OVRO interferometry of the 3mm HCN line provide maps of the ionized and molecular gas at 0.2 and 3” resolution, respectively. Comparison of the observed Pa emission with the 6-cm radio continuum yields a map of the foreground extinction. High apparent extinction is seen along the periphery of the ionized disk and is well-correlated with the molecular torus. The median extinction is Av = 31.1 mag. This extinction distribution is then used to correct the observed P a emission map to yield an ’extinction-corrected’ image of the ionized gas at 0.2”resolution. The stellar distribution near SgrA’ is also analyzed from the 1.9pm continuum image -

indicating a drop in the surface brightness distribution of star light within 1” radius of SPA*. This could be the result of stellar collisions which deplete giant stars within this radius or indicate the presence of a secondary lower mass black hole.

1 Introduction Within the central few parsecs of our Galaxy, radio, infrared and x-ray observations have revealed a massive black hole lying within a region of ionized ’spiral’ arms, encompassed by a ring of dense molecular clouds. The black hole is identified with the non-thermal radio source Sgr A*, and its mass is estimated to be 2 . 6 ~ 1 0Ma ~ from the motions of nearby stars and the neighboring ionized gas (Eckart & Genzel 1996, Genzel et al. 2000, Ghez et al. 2000). At a projected radius of approximately 2 pc from Sgr A*, lies the circumnuclear disMring (CND) of dense molecular gas and dust clouds (Genzel et al. 1985, Gusten et al. 1987), and within this ring, ionized gas streams extend down to a few arcsec from Sgr A* (Lo & Claussen 1983; Lacy et al. 1991). The molecular cloud rindarms can be modelled as a torus (0.5 pc thick) inclined at 70” to the line of sight (Jackson et al. 1993). Although clearly much less energetic and massive than the luminous AGN seen in many galactic nuclei, this region probably provides our best opportunity to observe in detail the processes associated with buildup of massive black holes and their accretion processes. We have recently undertaken two investigations aimed at providing the highest resolution imaging of both the ionized and molecular gas in the circumnuclear disk. The molecular gas clouds were imaged in HCN at 3” resolution by Christopher et al. (2002) using the Owens Valley mm-Array; here we present HST-NICMOS imaging at 0.2” resolution of the 1.87pm Pa (HI) emission line.

2 Molecular (HCN) and Ionized (Pa)Gas The reduction of the NICMOS images are described in detail in Scoville et nl. (2003). The major difficulty in imaging the Pa line was the subtraction of the stellar continuum -in the 1% bandwidth filters available on NICMOS, there are over 1500 stars with continuum emission exceeding the brightest pixel in Pa. Moreover, the reddening over this region is spatially variable and the intrinsic stellar colors are not all * Corresponding author: e-mail: [email protected],Phone: +OO626395 4979, Fax: +00 626568 9352

@ 2003 WILEY-VCH Verlag GmbH & Ca. KGaA, Weinheim

592

N. Scoville et al.: Pa Imaging

Fig. 1 Image of the Pa emission obtained by a scaled subtraction of the 1.90pm continuum image from the 1.87pm image (see text). Due to variable reddening and stellar colors it is necessary to adopt scale factors for the continuum subtraction which vary from star to star.

the same. This is illustrated in Fig. 1 where we show the result of straightforward subtraction of a scaled (0.91) version of the 1.9pm image from the 1.87pm image. Numerous stellar residuals (both negative and positive) are still present, indicating that a single scale factor can not be used. The best approach was to derive scale factors between 1.9pm and 1.87pm local to each bright star and then for pixels not having a bright star, we used the average scale factor derived from all the stars (see Scoville et al. 2003 for more details). emisThe extinction in front of the Pa emission was then estimated from the ratio of the observed PLY sion to the X = 6 cm radio map of Yusef-Zadeh & Wardle (1993). The extinctions range from Av 20 to 50 mag and the median pixel extinction is A v = 31.1 mag. These extinction values are in excellent agreement with those derived from the stellar colors (Blum et al. 1996, Cotera et al. 2000). The extinction distribution is shown in Fig. 2 together with contours showing the HCN emission distribution at 3” resolution (Christopher et al. 2003). the extinction exhibits large-scale gradients across the region (see Fig. 2) and low extinctions (20-25 mag) are seen in front of the East-West bar and the Northern arm and in the direction of the IRS 16 cluster. The generally low extinctions seen along the NE-SW swath south of Sgr A* is consistent with the low extinction values derived for the stellar continuum in IRS 16 (Blum et al. 1996). High extinctions (35-45 mag) are seen in the Western arm and in general along the outskirts of the Pa: N

593

Astron. Nachr./AN 324, No. S1 (2003)

.-

HCN w i t h A,

40

20

0

-20

-40

40

20

0

-20

arcsec

Fig. 2 The extinction distribution derived from the observed ratio of Pa to 6-cm radio continuum emission. The Pa emission was convolved to the same spatial resolution as the radio and we assume the radio continuum is entirely free-free except on Sgr A*. The contours are for the HCN emission at 3” resolution (Christopher et al. 2003). A broad minimum in the extinction is seen in the IRS 16 cluster and the extinctions increase at the periphery of the ionized gas in the molecular features. (The dashed contours in the HCN indicate locations where the HCN appears in absorption of the radio free-free (from the ionized gas) and nonthermal (Sgr A*) continuum.) The apparently high extinction on Sgr A*is due to the strong non-thermal contribution which yields spuriously large extinction estimates.

emission. This may be due to the increased extinction associated with dust in the molecular gas along the periphery of the ionized region. The apparently high extinctions shown on Sgr A* and a few arcsec to the north are spurious, due to the prescense of non-thermal radio emission there; the Av derived from the H92a line (Roberts & Goss 1993) shows no anomaly on Sgr A* . On the emission line stars the extinctions are also likely to be underestimated. The molecular gas, shown as contours of HCN (1-0) emission in Fig. 2 - 3, encircles the ionized gas at 20 - 40” radius from Sgr A*. With the high resolution OVRO observations, the HCN emission is resolved into over 20 clouds or clumps with masses of a 3 x lo2- 2 x lo3 Mo . The mean internal densities of these - ~ et al. 2003). These densities are approximately equal to the Roche clouds are 3 x lo6 ~ r n (Christopher critical densities for tidal stability in the presence of the central black hole and the interior mass of stars. Fig. 3 shows the extinction-corrected Pa emission within 40” of Sgr A* together with contours of 3mm HCN emission mapped with the OVRO array at 3” resolution (Christopher et al. 2003). The prominent N

N. Scoville et al.: P a Imaging

594

HCN w i t h unextincted P a

450

407

364

32 1

278

u a,

235

07 V

L

0

192

-2

149

106

-4 63

20

40

20

0

-20

arcsec

Fig. 3 The Pa emission corrected for extinction using the extinction derived from the ratio of Pa! to radio continuum emission . The color scale units are lo-’‘ ergs emp2 sec-’ per 0.0753” pixel and spatial offsets are relative to Sgr A*. The contours are for the HCN emission at 3” resolution (Christopher et al. 2003).

ionized emission features in Fig. 3 correspond to: the east-west ’bar’ structure approximately 7” south of Sgr A*, the northern arm which extends from -2” to over 25” north of Sgr A*, and the western arm approximately 14’’ west of Sgr A*. With the high resolution and sensitivity of the Pa images these largescale features break up into numerous filamentdarcs which are presumably density enhancements in the ionized gas or individual ionization fronts at the edges of neutral clouds. Smaller scale emission peaks are seen on the 38 emission line stars identified in Scoville et al. (2003). The ’mini-cavity’ appears as a clearly evacuated spherical hole in the emission of diameter 2.7” at -1.7” , -2.8’’ from Sg A* (Yusef-Zadeh, Morris & Ekers 1990). The total flux (within 40” radius of Sgr A*) in the extinction-corrected Pa image is 2,4x lo-’ ergs cm-’ ~ Pa photons per sec. The total Lyman sec-l; the luminosity is therefore 1 . 8 ~ ergs sec-l or 1 . 7 lo4’ ’ or 2 . 7 0 ~ 1 0La ~ (assuming continuum emission rate required for the Pa emission is 3 . 9 4 ~ 1 0 ~secpl 1.2 Rydbergs per Lyman continuum photon). Within 20” radius of Sgr A*, the Pa flux is 50% of that given above; thus 50% of the ionizing photons are absorbed within 20’’ or 0.8 pc. All of the emission line stars are within this region. The derived Lyman continuum production rates are minimum estimates since we assume Case B recombination with no Lyman continuum escaping the region. Given the apparent geometry of a tilted ring in the neutral gas, it is, in fact, very likely that the actual production rate could be

Astron. Nachr./AN 324, No. S1 (2003)

595

several times higher. The eight emission line stars studied by Najarro et al. (1997) have a combined Lyman continuum emission rate of 3 x 10’” sec-’ and thus they can probably account for the ionization without any extra input from Sgr A*. The Lyman continuum emission rate obtained from P a is consistent with that derived directly from the radio continuum ( w 3 x lo5’ sec-l, Genzel et al. 1994) and avoids uncertainties with respect to the radio non-thermal contributions. Correlation (or lack of correlation) between the molecular gas mapped by the HCN contours in Fig. 2 and the derived extinctions in front of the ionized gas might be used to infer the relative placement of the Hz and HI1 gas along the line of sight. The strong increase in extinction at the edge of the Western arm suggests that some of the ionized gas is behind or mixed with the neutral gas and dust in the molecular Western arm. Increased extinction in front of the ionized gas is also apparent where two filaments of molecular gas (traced by the HCN contours in Fig. 2) extend southward into the northern boundary of the ionized gas (Am = -5” , A6 = 15” , see Fig. 2). However, Christopher et al. (2003) estimate volume densities nHz 2 lo6 cm-3 for the HCN molecular ’ ~ then imply column densities 7 x loz3 H2 cmp2 or clouds. Their typical sizes (- 5” or 7 ~ 1 0 cm) AV 700 mag for a standard Galactic dust-to-gas ratio. Clearly, the extinctions in front of the the Western arm and northern Pa emission are not nearly this high. We therefore conclude that PLYemission on the west and northern border of the region (Aa = -10” , A6 = 10” j must be on the front face of the molecular gas (unless the dust is severely depleted or the molecular gas is extremely clumpy on scales << 1” ).

-

N

3 Stellar Distribution The mean surface brightness of the diffuse stellar continuum near Sgr A* is shown in Fig. 4. Here to avoid the strong effects of bright individual stars, the surface brightness (both mean and median pixel brightness) were computed from the pixels in the lower 80% of the brightness distribution at each radius. (The median also attenuates the effects of a few very bright pixels.) Both the observed and extinction-corrected brightness distributions increase at smaller radii but exhibit a sharp drop at 0.8” radius, corresponding to 0.03 pc ~ To evaluate whether this drop is statistically significant, we independently evaluated the or I . O X ~ O ’cm. mean and medians for the four quadrants (NE, SE, SW, NW) from Sgr A’and then estimated the uncertainties from the standard deviation of these independent samples - these are the vertical error bars shown in Fig. 4. The central drop is significant given these uncertainty estimates. The estimate we derive from ratioing P a to a radio recombination line at the position of Sgr A*is Av 30 mag. The number density of discrete stars (as opposed to the total light distribution) has been derived as a function of projected radius by Genzel et al, (1996). They also found a decrease in the projected surface density of star counts. For stars with K 5 12 mag, the drop occurs at 3” radius; for brighter star with K 5 10.5 mag, the decrease occurs inside 5” radius. Thus our results differ significantly from theirs, showing the drop occuring at 0.8” . Genzel et al. (1996) attribute the decrease at small radii as being due to collisions of the giant stars with main sequence stars which then deplete the giant stars because of their relatively large cross-sections. To test whether the drop in the surface brightness distribution inside 1” is due to a lack of a small number of bright giant stars inside 0.8” , we also show in Fig. 4, the median pixel surface brightness. This median flux should be virtually unaffected by the bright stars which dominate a relatively small fraction of the pixels; it still exhibits a modest decrease at small radii, although not as large as that seen in the mean surface brightness. The drop or leveling off of surface brightness inside 1” could have several explanations: it might be the core radius of the nuclear stellar distribution or it might be due to the depletion of late type stars with high mass-to-light ratios (Phinney 1989; Bailey & Davies 1999). Gravitational lensing would produce a drop inside the Einstein radius which is on the scale 0.01” (Wardle & Yusef-Zadeh 1992) and is therefore not a likely explanation. Sellgren et al. (1990) found a diminishing depth of the 2.3pm CO stellar absorption feature in the central few arcsecs, possibly indication a depletion of giant stars at the center. An alternative N

-

N. Scoville et al.: P a Imaging

596 Surface Brightness

1 .ooo

--

0.1 00

U

VI

? % v 3

rnedion

01

1.o Radius (arcsec)

10.0

Fig. 4 The surface brightness distribution for the observed and extinction-corrected 1.9 pm radiation is shown. Discrete, bright sources (such as IRS 7) were excluded by removing the brightest 20% of the pixels in each radial bin. Because the noise is very non-Gaussian (dominated by the stellar brightness distribution), the vertical error bars were estimated from the standard deviation of values calculated in the 4 quadrants relative to Sgr A*. possibility is ejection of stars from the central 1” by a past or present binary black hole (Milosavljevic & Phinney - private communication).

References Bailey, V. C., & Davits, M. B. 1999, MNRAS, 308,257 Blum, R.D., Sellgren, K., & DePoy, D. L. 1996, ApJ, 470,864 Cotera, A. S., Simpson, J. P., Erickson, E. F., Colgan, S. W. J., Burton. M. G., &Allen, D. A. ZOOO, ApJS, 129, 123 Christopher,M. H., Scoville, N. Z., Stolovy, S. R., Coker, R., & Yun, M. S. 2003, in preparation Eckart, A,, & Genzel, R. 1996, Nature383,415 Genzel, R., Crawford, M. K., Townes, C. H., Watson, D. M. 1985,ApJ, 297,766 Genzel, R.,Hollenbach, D. J., & Townes, C. H. 1994, Rep. F’rogr. Phys., 57,417 Genzel, R. Thatte, N., Krabbe, A , Kroker, H., & Tacconi-Gman, L. E. 1996, ApJ, 472, 153 Genzel, R., Pichon, C., Eckart, A,, Gerhard, 0. E., & Ott, T. 2000, MNRAS, 317,348 Ghez, A. M., Moms, M., Becklin, E. E., Tanner, A,, & Kremenek, T. 2000, Nature, 407,349 Gusten, R., Genzel, R.,Wnght, M. C. H., Jaffe, D. T., Stutzki, J., &Harris, A. I. 1987, ApJ, 318, 124 Jackson, J. M., Geis, N., Genzel, R., Harris, A. I., Madden, S., Poglitsch, A, Stacey, G. J., & Townes, C. H. 1993, ApJ, 402,173 Lacy, J. H., Achtermann, J. M., & Serahyn, E. 1991, ApJ, 380, L71 Lo, K. Y. & Claussen, M. J. 1983, Nature, 306,647 Najarro, F., Krabbe, A., Genzel, R.,Lutz, D., Kudritzki, R. P., & Hillier, D. J. 1997, A&A, 325,700 Phinney, E. S. 1989, in ”The Center of the Galaxy”, ed. M. Morris (Dordrecht:Kluwer),543 Roberts, D. A,, &Goss, W. M. 1993, ApJS, 86, 133 Scoville, N. Z., Stolovy, S. R., Rieke, M., Christopher, M. & Yusef-Zadeh, F. 2003, ApJ, submitted Sellgren, K., McGinn, M. T., Becklin, E. E., & Hall, D. N. B. 1990, ApJ, 359, 112 Wardle, M. & Yusef-Zadeh, F. 1992, ApJ, 387, L65 Yusef-Zadeh, F. Moms, M., & Ekes, R. 1990, Nature, 348,45

Astron. Nachr./AN 324, No. S I, 597-603 (2003)/ DO1 10.1002/asna.200385093

Resolving The Northern Arm Sources at the Galactic Center Angelle M. Tanner*I , A. M. Ghezl.', M. Morris', and E. E. Becklin'

' Department of Physics and Astronomy,University of California at Los Angeles, Los Angeles, CA 900951562 Institute for Geophysics and Planetary Physics, University of California, Los Angeles, CA 90095-1567

Key words Galaxy: center -infrared: stars Abstract. Diffraction limited images obtained with the W. M. Keck telescopes have spatially resolved the cool luminous Galactic Center sources IRS 21, IW, 2, 5 , and 1OW at wavelengths ranging from 2 to 25 prn. Their gaussian convolved sizes (-2000 AU or o"25 in diameter), along with their mid-infrared color temperatures, favor the hypothesis that they are centrally heated stellar sources rather than externally heated dust clumps. The near-infrared Keck speckle and A 0 images as well as 2.2 pm Gemini A 0 images of IRS 8 reveal asymmetric structures indicative of bow shock structures around all the Northern Arm sources. The presence of such large bow shocks around these objects suggests that their central heating sources have large winds requiring them to be young, massive stars like those observed in the nearby IRS 16 cluster. This number of windy stars is expected considering the a 10% volume coverage of tbe Northern Arm and the number of windy stars detected over a smaller region of the central parsec by Paumard et al (2001). This increases the total population of such stars to fifteen, which is comparable to the number of similar stars found in the nearby Quintuplet cluster.

1 Introduction Within the Galaxy's central cluster there are a number of enigmatic sources (IRS 1, 2, 5 , 8, 10 and 21) that have eluded classification for almost three decades (Rieke & Low 1973; Becklin & Neugebauer 1975). These objects have a number of properties in common. They are all spatially coincident with the Northern Arm, a tidal stream of dust and gas that is infalling and orbiting around the Galaxy's central supermassive black hole. Other common characteristics include spectral energy distributions that peak near 10 pm, high luminosities, and, in the cases of IRS 1 and 21, nearly featureless near-infrared spectra (Krabbe et al. 1995) and polarization properties that cannot be accounted for by the ISM (Eckart et al. 1995; Ott et al. 1999). Based on these observed properties, a number of investigators have suggested that these Northern Arm sources are dust enshrouded young stellar objects (Krabbe et al. 1995; Blum et al. 1996; Ott et al. 1999). If these sources are indeed young, then that would suggest that star formation is ongoing despite the lack of molecular cloud material and the extreme tidal forces induced by the 3 x lo6 M a central black hole, (see discussion in Sanders 1992). Also, if they are young then they would have to have formed or have begun forming, prior to the infall of the Northern Arm. In this case one would not expect the stars to still be embedded in the gas of the Northern Arm given that the newborn stars would have immediately started following ballistic orbits while the gas in which they formed is subject to strong, additional, nongravitational forces. Recent diffraction-limited 2-25 pm images of IRS 21, obtained by Tanner et al. (2002) on the W. M. Keck 10m telescopes, spatially resolved this source ( R N ~ R650 AU and E M I R 1600 AU). Using a simple radiative transfer code to simultaneously model the near- and mid-infrared photometry and intensity profiles, we argued that IRS 21, and by analogy all the Northern Arm sources, are massive stars

-

N

* Corresponding author: e-mail: [email protected],Phone: 310-794-5582, Fax: 310-206-2096 @ 2003 WILEY-VCH Verlag GmbH & Co KGaA. Weinhem

A. Tanner et al.: Northem Arm Sources

598

experiencing rapid mass loss. As the Northern Arm flows past these stars, the interactions between the material in the Northern Arm and the intrinsic stellar winds generate dusty bow shocks, which produce the large mid-infrared luminosities and, in the case of IRS 21, the observed spatially resolved emission. While the resolved structure for IRS 21 lacked the characteristic horseshoe morphology, this could simply be a projection effect if the relative velocity between the outflow source and the Northern Arm is near the line of sight. The bow shock model is appealing because it explains the association of these sources and the Northern Arm without requiring that the stars form within the Northern Arm before or during the infall toward the central black hole. A bow shock model instead allows for the much more plausible circumstance that the sources are simply in the path of the infalling Northern Arm gas and dust. In this contribution, we present new near- and mid-infrared observations for the remaining Northern Arm sources, IRS lW, 2, 5, 1OW and 8, in order to test the bow-shock hypothesis presented by Tanner et al. (2002). Section 2 describes the data sets that were obtained for this work, Section 3 presents the derived spatial extents, morphology and proper motions for the surveyed sources, and Section 4 discusses the implications of these observations.

2 Summary of Observations We have obtained high-resolution images of the Northern Arm sources, including 2.2 pm speckle observations with KeckI/NIRC, 3.8 pm A 0 observations with KeckWNIRC2 and 8.8, 12.5,20.8, and 24.5 prn direct imaging observations with KeckIIMirlin. In addition, we incorporate 2.2 prn A 0 observations with Gemini/Hokupa'a+QUIRC1 (see Rigaut et a1 2003). Figure 1 shows contour plots of the Northern Arm sources observed at near-infrared wavelengths. Our mid-infrared observations are presented in Tanner 2003. PSF

IRS 21

rRs 1w

IRS2

IRSS

IRS 1OW

IRSS

ffl

2 0 0

@

0.2 0.0

ffl

-0.2

1 @ ---t--@ t__

ffl

2

0.2

0

0

0.c

ffl

? -0.2

6

t c

: 0.2 ffl

0 0

0.c

$ Y -0.2 <

I

.

.

.

Fig. 1 Contour plots at 1.65 (top), 2.2 (middle) and 3.8 (bottom) pm of the O'tSxO'tS region around all the Northern Arm sources showing their asymmetric structure. The contours plotted represent 90.40% of the peak value in 10% intervals. North is up and East is to the left.

Based on observations obtained at the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under a cooperative agreement with the NSF on behalf of the Gemini partnership: the National Science Foundation (United States), the Particle Physics and Astronomy Research Council (United Kingdom), the National Research Council (Canada), CONICYT (Chile), the Australian Research Council (Australia), CNF'q (Brazil) and CONICET (Argentina).

Astron. Nachr./AN 324, No. S1 (2003)

599

3 Results 3.1

Near- and Mid-Infrared Sizes

Resolved structure around the Northern Arm sources is evident in Figure I, which contrasts near-infrared images of IRS 21,2, low, lW, 5 and 8 to a known point source. For simplicity, our analysis of their intrinsic size is performed in one dimension using the azimuthally-averaged intensity profile of each Northern Arm source and that of the corresponding point source. To determine their intrinsic sizes, the radial profile of a gaussian convolved PSF is compared to that of the Northern Arm source though x2 minimization as outlined in Tanner et al. (2002). The near-infrared radii (HWHM of the best fitting gaussian+PSF) include 470&30 AU for IRS 2,750+10 AU for IRS 1W, 5 1 0 f 4 0 AU for low, 7 6 0 ~ t 2 0for IRS 5, and 1800f140 AU for IRS 8. The mid-infrared radii range from 1000 to 2000 AU with no significant variation with wavelength. All of the radii are an order of magnitude larger than a stellar photosphere (4 AU for an M supergiant, Drilling & Landolt 2000) implying we are observing radiation from a dust distribution.

3.2

Morphology of the Northern Arm Sources

In addition to being resolved, all of the Northern Arm sources have asymmetric structures clearly visible in Figure 1, warranting further study of their morphology. Despite a crowded stellar field, we are able to extract 0!’5 x U’5 regions around each extended source and an isolated PSF from within the same image. We compare the results from the different deconvolution algorithms and conclude the blind deconvolution (IDAC, Jefferies & Christou 1992) routine provides consistent results given different PSFs, number of iterations and epochs. Figure 2 shows the deconvolved versions of the same Northern Arm sources shown in Figure 1 using the IDAC deconvolution routine. Also shown is the deconvolution of two separate PSFs. Deconvolution distributes the energy from the halo to the core of the PSF enhancing the asymmetric structure, and allowing us to compare their morphology to IRS 8. At 2.2 pm IRS 5 has a horseshoe morphology similar to IRS 8 but on a smaller spatial scale, while the structure around IRS 2 is also similar but with a significantly different position angle (PA) (see Rigaut et al. 2003 for an image of IRS 8). The morphology of the resolved material is not as apparent in deconvolved results from the 2.2 pm data for IRS 1W and 21. However, it is evident in the IDAC deconvolved 3.8 pm images. Finally, IRS 1OW does not show a horseshoe structure but does have a resolved feature to the southeast. We can estimate the position angles representing the angle of the symmetry axis of the resolved structure estimated from the deconvolved images. The resulting PAS and near-infrared radii can be compared to those expected from bowshocks created by the interaction of the Northern Arm material with the intrinsic stellar winds of the embedded sources, as was proposed for IRS 21. Figure 2 shows the observed PAS overplotted on the deconvolved images.

3.3 Proper Motions of the Northern Arm Sources In order to investigate the dynamics between the embedded sources and the Northern Arm, the proper motions of IRS 21, low, 1W and 2 are measured from the 2.2 pm speckle data. The position of IRS 2 is estimated by aligning the images from three epochs of different data sets including the 1999 July KeckI/NIRC 2.2 p m speckle map, the July 2000 Gemini North 2.2 pm A 0 map, and the May 2002 KecMVNIRC2 3.8 pm A 0 map. The proper motions are estimated from a weighted linear least squares fit through their position as a function of time. Since IRS 21 is located within the same field-of-view as Sgr A*, it has been repeatedly observed as part of the proper motion study by Ghez et al. (1998, 2000, 2003) and, therefore, has much smaller uncertainties (f50k d s ) . In contrast, the other sources require mosaicing, a process which began in 1998 resulting in larger uncertainties (+loo-200 Ws). Nonetheless, these values permit the measurement of the proper motions of these sources with respect to the gas in the Northern Arm.

A. Tanner et al.: Northern Arm Sources

600 PSF

IRS 21

IRSlW

IRSlOW

IRS 2

IRs5

Fig. 2 Contour 2.2 pm images (top) and 3.8 pm images (bottom) of the 0'.'5x0"5 region around the Northern Arm sources and a point source deconvolved using the IDAC deconvolution algorithm (Jefferies & Christou (1992)). The contours plotted represent 90-20% of the peak value in 10% intervals. Also shown are arrows depicting the observed PA (dashed line) and the PA of the relative velocity between the stars and Northern Arm gas (solid line).

Fig. 3 Plots of the a and 6 positions of IRS 1W, l o w , 21 and 2 as a function of time along with the weighted least square fit.

4 Discussion The bowshock hypothesis can now be further tested through comparison of (1) the resolved sizes and expected standoff distances and ( 2 ) the PA of the resolved asymmetry and the PA of the relative velocity vector between the stars and the gas.

4.1 Proper Motion of the Northern Arm Material The dynamics between the Northern Arm and embedded sources are further determined by the velocity and direction of the material flowing along the Northern Arm itself. Several models have been proposed for the motion of the gas in the Northern Arm, including a one-armed linear spiral, a hyperbola, and an

Astron. Nachr./AN 324, No. S1 (2003)

60 1

ellipse based on the radial velocity of the ionized gas as estimated from near-infrared ([Ne 111 and Br a ) and radio (H92 C Y ) emission line studies (Lacy et al. 1991; Roberts, Yusef-Zadeh & Goss 1996; Herbst et al. 1993). We do not consider the hyperbolic model since it assumes the material is moving in a direction opposite that suggested by the proper motion study of the gas by Yusef-Zadeh et al. (1998). Figure 4 shows the Gemini 2.2 pm A 0 image with overplots of the ellipse and linear spiral models as presented in the literature. No velocity information is available at the position of IRS 2 from the spiral model since it does not reach that source. Also plotted are velocity vectors showing the proper motion of the material at the position of the Northern Arm sources. 30

"

I ' / "

I

. 4'

"

'

" "

I

..

","

I

L 20 -

8

:

:

10-

0-

'

'

'

30

I

1

20

2

!&

10

0

1

10

5

0 *rE.eConds

-5

-10

-15

-10

-5

0

5

10

15

lvcIecond8

Fig. 4 Greyscale 2.2 pm Gemini A 0 image of the central parsec with an overplot of the ellipse and linear spiral models of Herbst et al. (1993) and Lacy et al. (1991), respectively, which provide proper motions of the Northern Arm material. The arrows plotted show the the relative proper motion between the gas and star at the position of each Northern Arm source

4.2

Relative Motion of the Embedded Sources and the Northern Arm Material

With the proper motions of both the embedded sources and the Northern Arm material, we calculate the resultant velocity vectors between these two components and compare these model PAS to the observed bowshock PAS. Figure 2 plots both the model PAS and observed PAS for IRS 21, lW, and 1OW. The PA's are consistent within errors for all the sources except IRS 1OW. Although we do not have a proper motion for IRS 5, the PA of the velocity of just the Northern Arm material is comparable to its observed PA. Using the relative velocity between the Northern Arm material and each embedded source, we can estimate a standoff distance of the bowshock to compare to the source's observed 2.2 pm size. Given the large luminosity of these sources (L=104 L g , Becklin et al. 1978), the two most likely scenarios for the identities of the central sources include AGB stars (v,,,d=40 km/s and ri2 = LOp5 Ma/yr) and 0 supergianWolfRayet (WR) stars (v,,,~=lOOO km/s and m = lop4 Mglyr) (Tanner et al. 2002 and references within). Only those standoff distances expected for an O/WR star are consistent with the observed 2.2 pm radii, thus ruling out the possibility that the embedded sources are AGB stars. The discrepancy of the observed morphology of IRS 1OW with that expected given its proper motion with respect to the Northern Arm could be explained by weaker stellar winds or its location within a less dense portion of the clumpy Northern Arm.

By successfully modeling the asymmetric dust morphology of the Northern Arm sources as bowshocks from stars with wind velocities on the order of 1000 k d s , we can add IRS 8, 5, lW, low, 21, and 2 to the population of massive, windy stars within the central parsec. Nine such stars have recently been

A. Tanner et al.: Northem Arm Sources

602

confirmed as windy, massive stars through their broad He I emission lines (Paumard et al. 2003, Paumard et al. 2001). When asking whether the Northern Arm sources are part of the same population or represent a unique group of sources, we can compare the observed number of luminous stars which are expected to be coincident with the Northern Arm assuming a 10% volume coverage of the Northern Arm within this region. To do this we use the Gemini 2.2 p m image over the same region of sky studied by Paumard. Within this region, there are 20 detected stars with magnitudes greater than or equal to the faintest Northern Arm source, IRS 2, which has a K magnitude of 10.6 (Blum et al. 1996). The 10% volume coverage then implies an expectation value of two stars associated with the Northern Arm. Thus, finding the (IRS 21, 1 and 2) luminous stars coincident with the Northern Arm within this same region is consistent with that expectation.

5

Conclusion

The addition of the six Northern Arm sources brings the total number of luminous, windy stars to fifteen which is comparable to the fifteen WCL stars found in the nearby Quintuplet cluster (Figer et al. 1999). The locations of the Northern Arm sources agree with the observations that the broad-lined He1 stars lie in an annulus around the narrow-line stars as would be expected if the stars formed at the same time but within distinct annuli centered around Sgr A* (Paumard et al. 2002). There are other prominent infrared sources in the central parsec, IRS 3 and 29N, and in the nearby Quintuplet cluster (Moneti et al. 2001; Figer et al. 1999) that also have featureless K-band spectra but are not as extended in the near-infrared. These sources are not coincident with the Northern Arm or any other diffuse dust feature and have been identified as either dusty Wolf-Rayet stars or protostars (Figer et al. 1999; Zinnecker et al. 1996). A complete understanding of the star formation history within the central parsec is still a work in progress. Deeper spectroscopic observations with high spatial resolution in the near-infrared would be worthwhile in order to look for the obscured spectroscopic signature of the stellar photosphere or wind. For example, Chiar et al. (2003) observe the 6.2 p m PAH feature toward the Quintuplet sources which is known to be observed in intrinsic circumstellar dust shells around late type WR stars (DWCLs). Such observations would distinguish whether these sources are 0 supergiants or their WR progenitors. Acknowledgements Data presented herein were obtained at the W M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California and the National Aeronauhcs and Space Administration. The Observatory was made possible by the generous financial support of the W.M. Keck Foundation. This work was supported by the NSF through both grant 9988397 and the Science and Technology Center for Adaptive Optics, managed by the University of California at Santa Cruz under cooperative agreement No. AST - 9876783. AMG also thanks the Packard Foundation for Its support.

References Becklin, E. E. & Neugebauer, G. 1975, ApJL, 200, L71 Becklin, E., E., Matthews, K., Neugebauer, G., & Willner, S. P., 1978, ApJ, 220,831 Blum, R. D., Sellgren, K., & DePoy, D. L. 1996, ApJ, 470,597 Chiar, J., et al. 2003, these proceedings Cotera, A. S., Eriekson, E. F., CoIgan, S. W. J., Simpson, J. P., Allen, D. A., & Button, M. G. 1996, ApJ, 461,750 Drilling, J.S., & Landolt, A.,U., 2000, in Allen’s Astrophysical Quantities, 4th edition, ed. A. Cox, AW Press, 381 Eckart, A., Genzel, R.,Hofman, R., Sarns, B. J., Tacconi-Garman,L.E.1995, ApJ, 445, L23 Figer, D.F., McLean, I.S., &Moms, M. 1999, ApJ, 514,202 Ghez, A., et al. 2003, in prep Ghez, A., Moms, M., & BecWin, E. E., Tanner, A., Kremenek, T. 2000, Nature, 407,349 Ghez, A,, Klein, B. L., Moms, M., & Becklin, E. E. 1998, ApJ, 509,678 Herbst, T., Beckwith, S., Shure, M. 1993, ApJ, 41 1,21 Jeffenes, S. M. & Christou, J. C. 1992, American Astronomical Society Meeting, 181, 1308 Krabbe et al., 1995 ApJ, 447, L95 Lacy, J. H., Achterrnann, 3. M., & Serabyn, E., 1991, ApJ, 380, L71 Moneti, A., Stolovy, S., Blommaert, J., A., D., L., Figer, D. F., Najarro, F. 2001, AAP, 366, 106

Astron. Nachr./AN 324. No. S 1 (2003) Ott, T., Eckart, A., & Genzel, R. 1999, ApJ, 523, 248 Paumard, T., et al. 2003, these proceedings Paumard, T., Maillard, J. P., Morris, M., & Rigaut, F. 2001, AAP, 366,466 Rieke, G. H. &Low, F. J. 1973, ApJ, 184,415 Rigaut, F., et al. 2003, these proceedings Roberts, D. A., & Goss, W. M. 1993, ApJS, 86, 133 Sanders, R. H. 1992, Nature, 359, 131 Tanner, A. et al., 2003, in prep Tanner, A., 2003, PhD Thesis, University of California, Los Angeles Tanner, A. et al., 2002, ApJ, 575, 860 Yusef-Zadeh, F., Roberts, D. A., & Biretta, J. 1998, ApJL, 499, L159 Zinnecker, H., Stanke, T., Kaufl, H., 1996, The Messenger, 84, 18

603

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. SI, 605-612 (2003) / DO1 10.1002/asna.200385050

Structural analysis of the Minispiral from high-resolution Bry data Thibaut Paumard',Jean-Pierre Maillard', and Mark Morris2

' Institut d'astrophysique de Paris (CNRS), 98b Bd. Arago, 75014 Paris, France

' University of California,Los Angeles, Div. of Astronomy, Dept. of Physics and Astronomy, Los Angeles, CA 90095-1562, USA

Key words dynamics, ionized gas, Sgr A West, Galaxy: Center, infrared, spectro-imaging, FTS PACS MA25 Integral field spectroscopy of a roughly 40" x40" region about the Galactic Center was obtained at 2.16 ,urn (Bry) using BEAR, an imaging Fourier Transform Spectrometer, at a spectral resolution of 21.3 k m s - l , and a spatial resolution of N 0.5". The analysis of the data was focused on the kinematics of the gas flows concentrated in the neighborhood of SgrA*, traditionally called the "Minispiral". From the decomposition into several velocity components (up to four) of the line profile extracted at each point of the field, velocity features were identified. Nine distinguishable structures are described: the standard Northern Arm, Eastern A n n , Bar, Western Arc, as well as five additional moving patches of gas. From this analysis, the Northern Arm appears not limited, as usually thought, to the bright north-south lane seen on intensity images, but consists instead of a continuous, weakly-emitting, triangular-shaped surface having a bright western rim, and narrowed at its forward apex in the vicinity of SgrA' where a strong velocity gradient is observed. The gravitational field of the central Black Hole can account for both the strong acceleration in this region and the tidal compression of the forward tip of the Northern Arm. Keplerian orbits can be fitted to the velocity field of the bright lane, which can be interpreted as formed by the bending of the western edge of the flowing surface. These results raise questions regarding the formation of the Sgr A West gas structures.

1 Introduction Within the inner 2 pc of the Galactic Center (GC) lies the Sgr A West region, dominated by ionized gas which has been detected in the infrared and at radio wavelengths because of high obscuration along the line of sight. The gas distribution has been observed in infrared and radio emission lines, as well as in radio continuum (Lacy et al. 1991, Lo & Claussen 1983, Roberts & Goss 1993, Zhao & Goss 1988), and proper motion of the bright blobs have been derived (Zhao & Goss 1988). All these data show that the ionized gas in the inner few parsecs of Galactic Center is organized into a spiral-like feature (Fig. 1) with a number of "arms", which led to the name "Minispiral" for the entire pattern. The various features give a spiral appearance primarily because of the way they are superposedon each other. However, a new analysis of Lacy's data was conducted by Vollmer & Duschl (2000) to re-examine the kinematic structure of the ionized gas. Using a three-dimensional representation, they confirm the standard features, but with a more complex structure, including two features for the Eastern Arm, a vertical finger of high density, a large ribbon extending to the east of SgrA", and two distinctly different components in the Bar. Data in different lines, acquired with better spectral and spatial resolution, warrant an independent kinematic analysis. In the present paper, the gas content in the inner region of the G C is presented and analyzed from high spectral resolution (21.3 km s-l) data cubes on the Bry line, obtained with BEAR, an imaging FI'S @ 2003 WILEY-VCH Verlrg GmbH & Co KGaA, Wemheirn

T. Paumard et al.: Structural analysis of the Minispiral

606

Fig. 1 Bry intensity integrated between -400 and t400 krns-’ toward the two mosaicked fields of Sgr A West observed with BEAR. The standard bright features, Northern and Eastern Arms, Bar, and the

mini-cavity, are indicated. Also, a few emission line stars show up as bright points in the image. (Moms & Maillard 2000) (Maillard 1995,2000) at the f/35 infrared focus of the 3.6-m CFH Telescope; the same work on He I data is ongoing. A preliminary analysis of these data have been presented in Moms & Maillard (2000). In Paumard et QZ. (2001, hereafter Paper I), where the helium stars were studied, the He I ISM emission was identified and described for the first time. The Bry data cover a field of view of 40” x 28” at a seeinglimited resolution of 0.6”. The HST-NICMOS Pact data (Scoville et al. 2003), used in Paumard et QZ. (2003). are used here for high-resolution morphological considerations. A multi-component line fitting procedure applied to the emission-line profiles in each point of the field is described in Sect. 2. From this decomposition, the identification of defined gas structures constituting the whole Sgr A West ionized region is presented in Sect. 3. Attempts to adjust Keplerian orbits to the flowing gas are presented. This is followed in Sect. 4 by a discussion of the implication of these identifications on the formation and the lifetime of the inner ionized gas.

-

2 Structure identification In each point of the field (Fig. 1 ) the Bry emission profile appears complex, clearly showing that along each line of sight, several ISM clouds or flows are present. So the study of these ISM features required the development of a multi-component line fitting procedure able to work on 3D data. From a coarse examination of the datacube the fitting by a maximum of four components seemed adequate. MivilleDeschgne (private communication) provided us with such software, that we adapted and developed. By comparison of the velocity components from one line of sight to the next, and by assuming continuity, it is possible to perform a reconstruction of the large-scale velocity structures. In the end, it might be possible to conclude whether these structures are independent, or continuous spiraling flows. Thus, the process is split into two main parts: first, the line profile decomposition of all the points of the field, and then the structure identification. The natural line shape of a single velocity component of the emission line from the ISM has been assumed to be Gaussian, defined by three parameters (central velocity, amplitude, and line width). The fitting procedure then adjusts four independent Gaussian lines, convolved with the instrumental PSF - which is a sine cardinat to each spectrum. Hence the fitting function depends upon twelve independent parameters for each point of the field. The consecutive steps are the following:

Preparation: These twelve parameters are first manually determined for a few starting points, evenly distributed over the field, chosen to be representative of the most obvious features, and located on spots where four components are clearly present.

Astron. Nachr./AN 324. No. S 1 (2003)

607

Step 1: From the starting points, a first procedure attempts to fit a four-component line shape function to each spectrum. For each new spectrum, the initial guess is determined from the results found for the neighboring points. The spectra are studied sequentially in parallel spiral-mode scannings arround each starting point. Except for the initial guess, the fitting of a spectrum is independent of all the others. Step 2: For the brightest point of the field, the neighbors are examined, and searched for a component such that the velocity gradient between the point of interest and this neighbor is less than a certain amount, which is set by the user at runtime. The procedure is iterative, and once a few neighbors have been selected into a structure, their neighbors are in turn examined for possible selection. The procedure stops when every component of every point of the field has been assigned to exactly one spatial structure. This procedure allows only one component of a given point to be selected into a given structure. Structures so warped that they overlap themselves spatially, thus causing two velocity components on the same line of sight, cannot be directly detected as such: they artificially split into several structures. Step 3: This procedure requires that the detected structures be manually inspected. The user has then the possibility to add some more common sense heuristics in the structure identification, a little difficult to implement but easy to apply manually. Several probleins can occur: during step 1, the fitting procedure might fit only one component where two blended components are indeed more appropriate, during step 2, the procedure can falsely cross-connect two structures, i.e. reconstruct two structures, each one being made of one part of one physical structure, and one part of the other one. Step 4: Then, these manually correctedresults are used to perform a second fit on each point of the field; at this point, 2D information is entirely included in the initial guess provided to the fitting procedure. Iteration: Steps 2 , 3 and 4 must be iterated a number of times, until a stable set of plausible structures is reached. “Plausible structures” means only that the structures are more extended than the spatial resolution, and brighter than the detection limit of the instrument.

3 Results 3.1 General description of the results The analysis described above leads to a vision of the Minispiral more complex than usually thought, one which is consistent with, but more detailed than the vision proposed by Vollmer & Duschl(2000). After a careful examination we identify 9 components of various sizes (Fig. 2 ) . Two types of velocity map appear, some with an overall velocity gradient, others without any large-scale velocity gradient. The deviation from mean motion, defined as the local difference between the velocity measured at one point and the mean value for the neighboring points, and divided by the uncertainty, ranges from roughly one tenth to ten for all the features, which means that every velocity structure shows significant (over 3 a ) local features. The areal size of the structures (Table l), expressed in terms of solid angle covered on the sky, ranges from 1 7 arcsec’ to 300 arcsec’ for the part of the Northern Arm that is visible in the BEAR field of view. The surface area of each structure must be considered as a lower limit because BEAR may not detect the weakest parts nor parts where blending with a brighter structure in the spectral domain is possible, and because the field of view does not cover the entire Minispiral. 3.2 Morphology of the ionized gas in Sgr A West A brief description follows for each identified velocity structure, whose velocity maps are given in Fig. 2. Table 1 gives the surface coverage on the sky (within the BEAR field of view), and the maximum and minimum velocity within the given structure.

T. Paumard et al.: Structural analysis of the Minispiral

608

Fig. 2 Velocity maps for the nine de-

ID

Featurename

a b c d

Northern Arm Bar Eastern Arm Eastern Bridge Westernkc Western Bridge Tip Northern Arm Chunk Bar Overlav

e f g h i

s 2414 1389 833 670 47 1 327 207 185 I36

s (arcsec2)

vmin

vmax

300’8 173.1 103.8 83.5 58.7 40.7 25.8 23.1 16.9

-286.9 -21 132.9 34.9 -37.1 -121.1 222.9 14.9 -267.0

188.9 196.9 242.9 182.9 74.9 100.9 339.0 74.9 -7.1

Table 1 Feature identifications,

with surface areas (pixels and square arcseconds), and minimum and maximum radial velocities (kms-l)

Northern Arm: Contrary to its standard description, the Northern Arm is not seen here as a bright N-S lane, but as an extended, triangular surface. One edge of this triangle is the bright rim generally noticed, but it extends over to the Eastern Arm. The third edge of the triangle is the edge of the field, so viewing this feature on a larger field may yield a slightly different description. Its kinematics will be thoroughly described in the next section. Bar: The Bar is the most complex region, where at least three components are superimposed. The most important feature is very extended, from the Eastern Arm (c) to the Western Arc (e), very straight and shows a smooth overall velocity gradient. Vollmer & Duschl(2000) mention two complementary components of the Bar, which they call Bar 1 and Bar 2, though their description is not sufficient to determine precisely the positions of these two components. We see two additional features, which we propose to call Western Bridge (9 and Bar Overlay (i). Parts of the Bar are also superimposed on almost every other structure, including the Eastern Arm (c), the Tip (g), the Eastern Bridge (d) and the Northern Arm (a). Eastern Arm: The region is split into two parts: the Ann itself and a ZIp (8). The velocity gradient of the Arm is directed along the minor axis of the structure, not along its major axis as expected for a flow. Eastern Bridge: A structure of medium size extends from the Eastern Arm to the bright rim of the Northern Arm. It does not show any large-scale velocity gradient, and its shape does not show any principal axis that would indicate a flow. It is superimposed on the faint regions of the Northern Arm, and partly superimposed on the Eastern Arm, the Bar and the Tip. The Paa map (Fig.5) shows that this feature may extend outside our field-of-view into a elongated feature parallel to the Eastern Arm.

Astron. Nachr./AN 324, No. S1 (2003)

609

Fig. 3 Microcavity feature in the region of IRS 9W, which would be approxi-

mately located in the upper right comer. The center of this field is located 5.4” east and 8.1” south of SgrA*.

WesternArc: The Western Arc lays just at the edge of the field, so we have access only to its innermost part. It is seen as a rather simple feature, with large scale velocity gradient. Westem Bridge: The Western Bridge is a tenuous, elongated feature oriented east-west and extending from the Bar to the Western Arc. Tip: The Tip is, in projection, a very concentrated and relatively small object with the most redward velocity in the region (Y300 km s-’). The Tip has already been noticed by Vollmer & Duschl(2000) only on a morphological basis, as a finger-like feature of the Eastern Arm in their three dimensional data. Here, we see that the Eastern Arm and the Tip are two distinct features, superimposed on the line of sight, thus we do not adopt the representation-dependent denomination “Finger”. At the elbow between the Eastern Arm and the Tip, in the IRS 9W region, is a bubble-like feature, or a Microcavity (radius N l”),with a rather bright rim (Fig. 3), which appears at a specific velocity (250 km s-’). Northern A m Chunk: A small tenuous structure is seen superimposed on the Northern Arm, a few arcseconds north of IRS 7. It lays at the edge of our field, so it could extend further out; however the Pacv image shows a small, horizontal bar at his location, crossing the bright rim of the Northern Arm, and that does not seem to be much extended. Bar Overtuy: The Bar Overlay looks like a small cloud that is superimposed upon the western region of the Bar and that shows a velocity gradient similar to the one of the main Bar at the same location, with an offset of N -40km s-’. It may indicate that these two features are closely related. They could, for example, be the two faces of a single cloud. 3.3 Keplerian orbit fitting velocity - maps - show a view of the features very different from the usual flux maps which, by themselves, can be misleading. For instance the morphology of the Northern Arm with its typical bright rim may lead one to think of it as the true path for most of the material. On the other hand, the velocity map shows no peculiar feature at the location of the rim. This is particularly intriguing for the location where it bends abruptly, just a few arcsecond north of IRS 1 and east of IRS 7 (Fig. 1). Thus we are led to the idea that the kinematics of the Northern Arm should be studied independently of its intensity distribution. As a first attempt at exploiting the information contained in the new tools that are the velocity maps, we tried to analyse the Northern Arm as a Keplerian system. For a first, simple approach, we created a dedicated IDL graphical package called GuiMapOverlay (Fig. 4). With this tool, the user can easily adjust one Keplerian orbit over a velocity map. A Keplerian orbit in 3D is defined by five orbital parameters: the eccentricity, two angles defining the orientation of the orbital plane, the periapse (distance of closest approach to the center of motion), and a third angle defining the position of the periapse. Once the user is almost satisfied with the orbital parameters he has found by trial and error, an automatic fitting procedure can be called. It is possible to fix parameters (at least two must remain free), and the orbit can be forced to go through a selected constraint point by tying the periapse to the other parameters. After a few experiments with this tool, we are led to some general conclusions:

T. Paumard et al.: Structural analvsis of the Minisuiral

610

Fig. 4 GuiMapOverlay: given a set of orbital parameters, this tool draws the path of the corresponding Keplerian orbit on a

velocity map, extracts the observed velocities along this path, and plots both the observed and computed velocity curves. The < x2> is also computed and shown. 0

a good agreement can be found between observed and calculated velocities, except in the region of the Minicavity; this model alone is not sufficient to decide whether the orbits are bound or not, or whether the data are compatible with elliptic, parabolic, or hyperbolic motion.

The second point mentioned above is not satisfactory, as one of the most interesting questions is to decide whether the gas is bound. Nevertheless, the first remark encouraged us to persevere in the direction of Keplerian modeling, so we attempted to model the Northern Arm with several orbits instead of only one. Each orbit is bound to pass through a different point, by tying its periapse to the other four parameters, and all these constraint points are aligned across the gas lane, close to the line of zero radial velocity, thus ensuring that they are indeed on separate trajectories. To ensure a smooth model - we are interested only in the global motion - the four functions that map each constraint point to one of the parameters, have been chosen to be described as spline functions, defined by their value at four points. Thus, having four functions (one for each of the orbital parameters), each of them being defined by four values, the model is dependent on sixteen parameters. We designed a fitting procedure to adjust this model based on the observed velocity map by minimizing the reduced < x2>. In order to study a certain hypothesis, it is possible to either fix some parameters, or to force them to have the same value for each orbit. This way, for example, it is possible to really check whether the observed velocity map is consistent with coplanar orbits or with uniform eccentricity. To avoid studying only local minima in the parameter space, it is also important to use several different initial guesses. A tool quite similar to GuiMapOverlay has been designed to easily study whether the data are consistent with an homothetic' set of orbits, which is the simpler model. 3.3.1 Results With the homothetic set of orbits hypothesis, the eccentricity still cannot be well constrained. Bound orbits seem to be prefened, but the agreement is as good with circular orbits and very elliptic orbits, close to parabolic. The residuals map has always the same shape: the observed velocities are always smaller than the computed ones along the inner edge of the bundle of orbits, and higher along the outer edge. The agreement is always poor, with < x2>'/'cz 70. A few of these homothetic models have been chosen as initial guesses for other adjustments, with released constraints. It is first interesting to check the coplanar hypothesis, by keeping uniform only the two parameters that define the orbital plane, and the uniform

' Two orbits are said to be "homothetic" when they are identical except for their scale, i.e. when they share the same orbital parameters, except the periapse.

Astron. Nachr./AN 324. No. S1 (2003)

61 1

eccentricity hypothesis. The agreement is much better when making either the eccentricity or the orbital plane free, but it is still better making both free. Even with the most general situation, the parameters are still not constrained enough to decide whether the orbits are all bound or not, to extrapolate the model outside the field of view, nor even to derive reliably the direction of proper motion. However the models share a few characteristics that we judge to be robust because of their repeatability: 1. the orbital planes are close to that of the CND; 2. the orbits are not quite coplanar; the two angles that define the orbital plane vary over a N 10”range; 3. the eccentricity varies from one orbit to another, beeing close to parabolic or above for the innermost orbits, and closer to circular (below N 0.5) for the outermost. The variations of the orbital parameters induce a particular shape for the Northem Arm (Fig. 5): for all the non-coplanar models, the Northern Arm looks like a warped surface, and this warp induces a crowding of orbits that closely follows the bright rim of the structure. That suggests that the Northern Arm is either a warped disk, or the ionized surface of a neutral cloud. The bright rim itself is not due only to the stronger UV field and a real local enhancement of the density, but also to an enhancement of the column-density due to the warp. An interesting point is that, in some models, no orbit follows the bright rim, which emphasizes that it is really important to consider the dynamics independently from the morphology of the Northern Arm. Another characteristic present in all the models is that the period of the orbits ranges from a few lo4 years to a few lo5 years, which implies that the Northern h would have a completely different shape in a few lo4 years, and cannot be much older than that timescale. Since the agreement in radial velocity is now rather good, it makes sense to look at the deviations from global motion by looking at the extended features on the residual velocity map (Fig. 5): A) the flow shows a rather significant deviation in the region just southwest of the embedded star, IRS 1W; this perturbation could be due to the interaction with this star’s wind; B) the region of this model closest to the Minicavity is perturbed; C) another deviation is seen at the precise location where the bright rim bends a lot, just east of IRS 7E2; D) finally, an elongated feature is seen on the fainter rim coming from IRS 1W towards the northeast.

4 Discussion The presence of at least three isolated gas patches (the Western Bridge, the Northern Arm Chunk and the Bar Overlay, but also possibly the Eastern Bridge, which may or may not extend ouside the field) in addition to the standard large flows has been demonstrated. In addition to that, the Microcavity at the elbow between the Eastern Arm and the Tip is a new example of an interaction between an ISM feature and a stellar wind, similar to the Minicavity, as is the deviation from Keplerian motion detected close to IRS 1W. In this context, it makes sense to ask what is the influence of the large number of mass losing stars present in the central parsec? These massive, hot stars of the central cluster, named “helium stars” from their strong 2.06 pm He I emission line, presumably LBV-type and W stars, being particularly concentrated in two clusters, IRS 16 (Krabbe et al. 1991) and IRS 13E (Maillard et al. 2003), must be a major source of helium in their environment. Therefore the following question arises: what happens to this helium enriched material? Could it form or enrich the gas patches that we see? Comparing the helium and hydrogen distribution in the central parsecs, and their abundance in the different structures and in the CND is guaranteed to help us better understand the origin of the different ISM structures. The geometry of the Northern Arm has been studied from its velocity map, leading to the conclusion that it may not be a planar structure, but rather a three-dimensional structure. Fig. 5a is quite compatible with the Northern Arm indeed being the ionized surface of a neutral cloud, which could come from the CND. in accordance with the standard formation scenario.

T. Paumard et al.: Structural analysis of the Minispiral

612

20

10

0

-10

-20

20

10 da

0 -10 cos(6) ( Arcsec )

- 20

20

0 -10 da . cos(6) ( Arcsec )

10

-20

Fig. 5 On this Paa map (Scoville et al. 2002), one of the Keplerian models is overplotted (left panel). This one is quite consistent with the Northern Arm and the Western Arc being related structures. On the right panel, the most significant deviations from Keplerian motion discussed in the text are labeled A to D, and indicated as filled contour.

References Krabbe, A,, Genzel, R., Drapatz, S., & Rotaciuc, V. 1991, ApJ, 382, L19 Lacy, J.H., Achtemann, J.M., & Serabyn, E. 1991, ApJ, 380, L71 Lo, KY., & Clausen, M.J. 1983, Nature, 306,647 Maillard, J.P. 1995, In: Tridimensional Optical Spectroscopic Methods in Astrophysics, L4U Col. 149, G . Comte & M. Marcelin (eds), ASP Conf. Series, 71,316 Maillard, J.P. 2000, in Waging the Universe in Three Dimensions, E. van Breughel & J. Bland-Hawthorn (eds), ASP Conf. Serie 195, 185 Maillard, J.P., Paumard, T., Stolovy, S.R., & Rigaut, F. 2003, these proceedings Moms, M. & Maillard, J.P., 2000, in Imaging the Universe in Three Dimensions, E. van Breughel & J. BlandHawthorn (eds), ASP Conf. Sene 195, 196 Paumard, T., Maillard, J.P., Moms, M., & Rigaut, F. 2001, A&A, 366,466 (Paper I) Paumard, T., Maillard, J.P., Stolovy, S.R., & Rigaut, F. 2003, thesepmceedings Roberts, D.A. & Goss, W. M. 1993, ApJS, 86, 133 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher,M., & Yusef-Zadeh, F., 2002, submitted to ApJ Vollmer, B., & Duschl, W.S. 2000, New Astronomy, 4,581 Yusef-Zadeh, F., Roberts, D.A., & Biretta, J. 1998, ApJ, 499, L159 Yusef-Zadeh, F., Stolovy, S. R., Burton, M. Wardle, M., &Ashley, M. C. B. 2001, Apl, 560,749 Zhao, J.H. & Goss, W.M. 1988, ApJ, 499, L163

Astron. Nachr./AN 324, No. SI, 613-619 (2003)/ DO1 10.1002/asna.200385095

Gas physics and dynamics in the central 50 pc of the Galaxy B. VoUmer*', W.J. Duschl***.',and R.Z ~ l k a * * ' ~

' Max-Planck-Institutfur Radioastronomie, Auf dem Hugel 69,53121 Bonn, Germany

* Institut fur Theoretische Astrophysik der Universitat Heidelberg, TiergartenstraBe 15, 69 121 Heidelberg, Germany IRAM, 300, rue de la piscine, 38406 Saint Martin dHeres, France

Key words Gas physics, gas dynamics, theory, numerical modelling Abstract. We present models for the gas physics and dynamics of the inner 50 pc of the Galaxy. In a first step the gas properties of an isolated clumpy circumnuclear disk were analytically investigated. We took the external UV radiation field, the gravitational potenQal, and the observed gas temperature into account. The model includes a demiption of the properties of individual gas clumps on small scales, and a treatment of the circumnuclear disk as a quasi-continuous accretion disk on large scales. In a second step the dynamics of an isolated circumnuclear disk were investigated with the help of a collisional N-body code. The environment of the disk is taken into account in a third step, where we calculated a pro- and a retrograde encounter of an infalling gas cloud with a pre-existing circumnuclear disk In order to constrain the dynamical model, we used the NIR absorption of the giant molecular clouds located within the inner 50 pc of the Galaxy to reconstruct their line-of-sight distribution.

1 Introduction During the discussion led by R. Narayan at this conference it became clear that the mass accretion rate onto the central black hole in the Galactic Centre at a radius of <1 pc is lop8 Ma yr-'. The central black hole is thus extremely sub-Eddington. In order to understand the fueling mechanisms of the central engine, there is an inevitable need for understanding the gas physics and dynamics in the inner -50 pc of the Galaxy. Gas that flows radially into the Galactic Centre has to pass several bamers. At large scales (
* Corresponding author: e-mail: bvollrnerOrnpifr-bonn.rnpg.de,Phone: NO49228 525315, Fax: NO49 228 525436 * * email: [email protected] *** email: [email protected] @ 2003 WILEY-VCH Verlag GmbH & Co KGaA. Wemheim

B. Vollmer et al.: Gas Dhvsics and dvnamics in the central 50 DC of the Galaxv

614

2 The analytical model (Vollmer & Duschl2001a, 2001b) In the inner 20 pc of the galaxy the gas is very clumpy with a volume filling factor of -1%. These clumps have masses of -30 Ma and sizes of -0.1 pc (see e.g. Jackson et al. 1993). They are illuminated by the UV radiation field of the central He1 star cluster and form a ring-like structure that is known as the Cicumnuclear Disk (CND) (Giisten et al. 1987). Fig. 1 illustrates the situation. The whole CND has a total gas mass of several lo4 Ma.The inner edge is located at a radius of -2 pc where the density of the neutral gas drops by more than an order of magnitude. The gas in the central 4 pc is ionized and forms the giant HII region Sgr A West. Sgr A Wesf

Hd stax

rJ

0 0

0

0

":*?

wo

'b*

0

D

0

4pc 15F

Fig. 1 Sketch of the inner 15 pc of the Galaxy. The Circumnuclear Disk (CND) consists of small, dense gas clumps that are illuminated by the central He1 star cluster. The dotted line delineates approximately the extension of the HII region Sgr A West.

Our analytical model to describe the gas properties consists of two parts: (i) Small scale: the gas clumps are described as isothermal spheres that are partially ionized by the UV radiation field (Fig. 2). The gas temperature of the clouds is assumed to be proportional to R-4. (ii) Large scale: the properties of clouds located at different distances from the Galactic Centre are determined. The clumpy gas distribution is then smeared out to obtain a quasi-continuous disk structure that is described by the standard set of accretion disk equations with a modified viscosity prescription taking the radiative energy dissipation during clumpclump collisions into account. Both models are connected via the central density of the clumps that is proportional to the central density of the disk: pCl = a;' Pdisk, where @V = 0.01 is the volume filling factor. In addition, the clump properties influence the disk viscosity via the local energy dissipation rate. The main results of this modelling are: (i) There are two solutions for our set of equations that correspond to two stable clump regimes: the observed heavy clumps -10 M a and the stripped cores of the heavy clouds with masses between Ma. and (ii) Within the disk, the number of collisions between clouds is very low (RC,ll N yr-' for about 500 clouds). Mo yr-'. (iii) The inferred mass accretion rate for the isolated CND is 2 N (iv) The CND is much more stable and has a much longer lifetime (- lo7 yr) than previously assumed. The disk clouds are exposed to strong stellar winds, the central UV radiation field, and strong tidal shear. We show that stellar winds shape the clouds, but do not affect their other properties. In order to resist the strong shear due to differential rotation each cloud's central densities must increase with decreasing distance to the Galactic Centre. Its outer radius is determined by the UV radiation field at the illuminated side and by the external gas pressure at the shadowed side opposite to the direction of the Galactic Centre. We show analytically that for a temperature gradient of the form T 0: R-i the radius of the massive clouds are constant irrespective of their location in the disk. Thus, clouds that can resist

Astron. Nachr./M 324, No. S I (2003)

615

W radia~on

lllllllll

Jo

neutral candensation

Fig. 2 Illustration of the partially ionized globule model described in the text. The UV radiation comes from the direction of the Galactic Centre. The boundary opposite to the Galactic Cenue is determined by the gas pressure of the surrounding ionized low density gas.

tidal shear will become more and more massive when approaching the Galactic Centre. Finally, they will collapse when their masses exceed the Jeans limit. We suggest that the cloud distribution within the CND reflects two selection effects. First, the clouds have to be dense enough to resist tidal shear and second, the clouds that are too massive collapse and will form stars. Magnetic fields and rotation stabilize otherwise gravitationally unstable clouds. This mechanism naturally explains the existence of the inner edge of the CND: Fig. 3 shows the central density of the heavy clouds versus the distance from the Galactic Centre. The limits for gravitational collapse and tidal disruption cross at R -2 pc. The dashed surface represents the range of densities where clouds are gravitationally and tidally stable.

0

2

4

6

radius (pc)

Fig. 3 Central density of the heavy clouds versus the distance to the Galactic Centre. Dashed line: maximum central density above which gravitational collapse occurs. Solid line: minimum density in order to resist tidal shear. Dashed surface: range of densities where clouds are gravitationally and tidally stable.

616

B. Vollmer et al.: Gas physics and dynamics in the central 50 pc of the Galaxy

Fig. 4 The evolution of the cloud infall into the Galactic Centre as the observer would see it from the Earth, Left side: prograde encounter. Right side: retrograde encounter. The elapsed time is plotted on the top of each frame.

3 The numerical model (Vollmer & Duschl2002) In a next step, we use our knowledge acquired with the analytical model to build a realistic dynamical model in order to investigate the gas dynamics in the inner 50 pc of the Galaxy. Since a circumnuclear disk (CND) consisting of small clumps with a tiny volume-filling factor can be long-lived (several Myr, see Sect. 2), we study the scenario where a part of a giant molecular cloud falls onto a pre-existing CND. We use a collisional N-body code where each panicle represents a gas clump with a certain mass and radius. The infalling gas cloud also consists of a number of small subclumps. These clumps have masses around the observed value of 30 Ma. When a clump approaches the Galactic Centre closer than 2 pc, it is assumed to be destroyed by tidal forces or by gravitational collapse (see Sect. 2). These clumps are counted as accreted. The collisional N-body code yields a realistic cloud collision rate that depends on the cloud radius, density, and dispersion velocity. A realistic simulation of an isolated CND shows the observed disk structure. 2 Myr and a mass accretion rate This simulation yields a mean collision time scale of a cloud of tcoll M a yr-' 5 Ak 5 10V3 M a yr-'. The infalling cloud, which has a mass of several lo4 Ma, of is assumed to be on (i) a prograde orbit and (ii) a retrograde orbit with respect to this CND. We study the resulting cloud-cloud collision rate and the mass accretion rate using different loss rates of the kinetic energy during a collision. Fig. 4 shows a prograde (left side) and a retrograde (right side) encounter with a loss of 10%of the kinetic energy of the clumps during a collision. The main difference between the two simulations is that the CND is destroyed 3-4 Myr after the first encounter with the external cloud in the

-

Astron. Nachr./AN 324, No. S1 (2003)

617

case of a retrograde encounter. Its mass is mainly accreted onto the Galactic Centre. At the end of the simulation a second CND has formed that has approximately the angular momentum of the infalling cloud, but there is still a counter-rotating core visible, which represents the remnant of the former CND. In the case of a prograde encounter, the infalling mass is partly added to the pre-existing one. The outcome of this simulation is a warped CND that is more massive than the pre-existing CND. Within our scenario of an encounter between an infalling molecular cloud and a CND, it is possible that the observed He11 star cluster in the Galactic Centre has been formed by a retrograde encounter of a cloud with the CND -7 Myr ago. The cloud that formed the Hen star cluster has been destroyed by tidal forces and can presently no longer be distinguished as a single kinematical entity.

4

The LOS distribution of the GMCs (Vollmer et al. 2003)

The fueling of the central black hole in the future depends mainly on the present dynamics of the gas in the inner 50 pc of the Galaxy. Within this region the gas is heavily clumped and mainly in the form of giant molecular clouds. Following Zylka et al. (1990) three main giant molecular cloud complexes can be distinguished: (i) Sgr A East Core, a compact giant molecular cloud with a gas mass of several lo5 Ma located north-east of Sgr A". (ii) The giant molecular cloud M-0.02-0.07 located to the east of Sgr A*. Since its mean radial velocity is -50 km s-', it is also called the 50 km s-' cloud. (iii) The GMC complex M-0.13-0.08 located south of Sgr A*. Since its mean radial velocity is -20 kms-l, it is also called the 20 kms-' cloud. The Sgr A East core is part of the 50 kms-l cloud complex, thus we will treat these features as a single structure. Fig. 5 shows a sketch of the inner 30 pc of the Galaxy, where the main features are indicated (Minispiral, CND, 20 km s-l cloud, 50 km s-l cloud). If one wants to understand

Fig. 5 Sketch of the inner 30 pc of the Galaxy. The central black dot represents the He1 star cluster surrounded by the Minispiral and the CND (where only the high velocity lobes are shown). Most of the mass is located at negative b.

the gas dynamics in the central 50 pc of the Galaxy, it is of crucial importance to know the line-of-sight distribution o f the giant molecular clouds (20 and 50 km s-l clouds) located in this region. We reconstruct the line-of-sight distribution assuming (i) an axis-symmetric stellar distribution and (ii) that the clouds are optically thick and have an area filling factor 1, i.e. that they entirely block the light from the stars located behind them. Fig. 6 shows the reconstructedLOS distribution of the giant molecular clouds in colors. The IRAM 30m 1.2 mm observations of Zylka et al. (1998) are overlayed as contours. Due to the method of reconstruction, LOS distances close to Sgr A* (<-I0 pc) have a small uncertainty, whereas larger LOS distances might be located up to a factor 2 farther away from Sgr A*. The relative distances are robust results. We found that: All structures seen in the 1.2 mm observations (Zylka et al. 1998) and CS(2-1) observations (Gusten et al. in prep.) are present in absorption. The 50 km s-l cloud complex is located between 0 pc and -5 pc, i.e. in front of Sgr A*. It has a small LOS distance gradient. The 20 km s-l cloud complex is located in front of the 50 km s-l cloud complex. The subclump of N

B. Vollmer et al.: Gas Dhvsics and dvnamics in the central 50 DC of the Gdaxv

618

reconstructed LOS distribution k 1.2mm emission

40

200 20

m

d 3 3

0

0 4

a

-

-20 -200

..

400

200

0

-200

galactic longitude offset [arc sec]

Fig. 6 Contours: IRAM 30m 1.2 mm observations of Zylka et al. (1998). Colors: LOS distance distribution filtered with a median filter of 11 pixel size.

strongest absorption has a LOS distance between -40 pc and -20 pc. The CND is not seen in absorption. This gives an upper limit to the cloud sizes within the CND of -0.06 pc. The combination of the LOS distribution of the gas and its kinematics will help to unravel the dynamics of the gas in the inner 50 pc of the Galaxy, which represents the future fueling of the central black hole.

5 Outlook From the present modeling of the gas in the inner 50 pc of the Galaxy we have learned about crucial aspects of the gas physics and dynamics. We found that the mass accretion rate into the central parsec is highly variable and a period of almost no mass accretion is conceivable. This period of starvation might last about lo4 yr, which is the cloud-cloud collision time within the CND. We are now in the position to use the acquired knowledge to determine the temporal behaviour of the mass accretion rate in the past and how it might change in the future. Remaining questions are: Can we find clear signs of past interactions of an external cloud with a pre-existing circumnuclear disk? Will the accreted mass exclusively come from the CND in the near future? Will there be a major accretion event in the near future, when a part of a massive giant molecular cloud interacts with the CND?We already have the keys in our hand to unravel the exciting history and future of the gas dynamics in the Galactic Centre. Acknowledgements This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis CentedCalifomia Institute

Astron. Nachr./AN 324, No. S1 (2003)

619

of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation. m travel ej~~ funding. ch~~ BV thanks the Deufsche ~ o r ~ c h ~ ~ g ~ g e for

References Giisten R., Genzel R., Wright M.C.H. et al., 1987, ApJ 318, 124 Jackson J.M., Geis N., Genzel R. et al., 1993, ApJ 402, 173 Launhardt R., Zylka R., & Mezger P.G. 2002, A&A, 384, 1 12 Vollmer B., Duschl W. J. 2001a, A&A, 367,72 Vollmer B., Duschl W. J. 2001b, A&A, 377, 1016 Vollmer B., Duschl W. J. 2002, A&A, 388, 128 Vollmer B., Zylka R., & Duschl W.J. 2003, A&A, submitted Zylka R., Mezger P.G., &Wink J.E. 1990, A&A, 234, 133 Zylka R., Philipp S., Duschl W.J., Mezger P.G., Herbst T., & Tuffs R. 1998, in: The central regions of the Galaxy and galaxies, Proceedings of the 184th symposium of the International Astronomical Union, held in Kyoto, Japan, August 18-22, 1997. Edited by Yoshiaki Sofue. Publisher: Dordrecht: Kluwer, p.291

This Page Intentionally Left Blank

Astron. Nachr./AN 324. No. S1.621-627 (2003)/DO1 10.1002/asna.200385086

The First Measurement of Radial Acceleration of Ionized Gas Near Sagittarius A* Doug Roberts*

'x2

and Farhad Yusef-Zadeh***

' Adler Planetarium & Astronomy Museum, 1300 S. Lake Shore Drive, Chicago, IL. 60605 Department of Physics and Astronomy, Northwestern University, Evanston, IL. 60208

Key words Galaxies:nuclei, Galaxies: The Galaxy, Radio sources: Lines, Interferometry

Abstract. Motivated by the presence of a massive black hole coincident with Sgr A* at the center of our Galaxy, a number of observers have successfully detected proper motion of ionized gas and stars at the Galactic center. Here we compare the radial velocities of ionized gas in the inner parsec of the Galaxy in three epochs (1993, 1999 & 2002). VLA observations were carried out in the H92a radio recombination line from Sgr A West; each observation had identical velocity coverage as well as velocity and spatial resolutions. Accurate velocity fields were derived from Gaussian fitting the data from the three epochs; radial acceleration was determined by comparing the fitted velocity fields. In the Northern Arm of Sgr A West, we report the the detection of radial acceleration ( 4 . 2 1 k m - ' yr-') consistent with motion dominated by gravity. In the minicavity, the measured acceleration is too large to be the result of gravity alone and suggest that non-gravitational effects are significant. Additionally, we detect a single region that shows peculiar motion of +4.1 kms-'yr-l southeast of Sgr A*. The first evidence of the acceleration of ionized gas in the Galactic center is presented.

1 Introduction There have been many observations of gas motions near the center of the Galaxy over the past decades. As the observations became more precise, the effect of a large central mass concentration at the position of the radio point source Sgr A* became clearer. Currently, the presence of a massive black hole at the center of our Galaxy coincident with the Sgr A* is consistent with a number of measurements that have successfully detected proper motion of ionized gas (Yusef-Zadeh, Roberts & Biretta 1998; Zhao & Goss 1998) and stars (Eckart & Genzel 1997; Ghetz et al. 1998) at the Galactic center. This project further investigates the role of gravitational and non-gravitational effects in the center of the Galaxy by comparing high quality recombination line data as a function of time to measure the radial acceleration of gas near Sgr A*.

2 Observations We used the Very Large Array (VLA) of the National Radio Astronomy Observatory' to observe the H92a radio recombination line emission from Sagittarius A West (Sgr A West) in three observations spread throughout a decade starting in 1993. In all observations, the bright (1 Jy) continuum emission from Sgr A* was used to self calibrate the complex gains of the spectral line data set. Table 1 shows the relevant observational information. * Corresponding author: e-mal: [email protected], Phone: +OO 847 467 6889, Fax: +00 847 467 7729 * * Corresponding author: e-mail: [email protected],Phone: +OO 847 491 8147, Fax: +00 847 467 9982

' The National Radio Astronomy Observatory is a facility of the National Science Foundation, operated under a cooperative agreement by Associated Universities, Inc.

@ 2003 WUEY-VCH Verlag GmbH & Ca. KGaA, Weinhelm

622

D. A. Roberts and F. Yusef-Zadeh: Radial Acceleration of Ionized Gas near Sm A*

Table 1 Observational Parameters

Observing Parameter

Value

VLA Configuration Primary beam Pointing center (J2000) Synthesized beam FWHM Line Rest frequency Number of channels Polarization Total Bandwidth Velocity Coverage Velocity Resolution Primary Gain Calibrator Complex Gain Calibrator Bandpass Calibrator Epoch 1 Date Total Observing Time r.m.s. noise Epoch 2 Date Total Observing Time r.m.s. noise Epoch 3 Date Total Observing Time r.m.s. noise

BnA 5' 17h45m40?230-29"00'27!'069 1!'25 x 1!'0 (6 x a ) H92a 8,309.383 MHz 32 1 (RCP) 12-5MHz (450 km-') -420 to +20 kms-' 14 km-' (390.625 kHz) Hanning smoothed 3C286 NRAO 530 3C84,3C286 & NRAO 530 February 18,1993 2x8hr 0.16 mJy Jybeam-' October 18, 1999 1.4 x 8 hr 0.21 mJy Jybeam-' February 18,2002 2x8hr 0.16 mJv Jvbem-'

Radio observations of the center of the Galaxy present a unique observational challenge, as radial velocities span a large range from almost -400 to +200 kms-l; additionally, structure in the line necessitates high spectral resolution. Specifically, near Sgr A* the recombination line emission from Sgr A West roughly falls into two broad velocity ranges. Most of the emission occurs at -200 to +200 k m s - ' and southwest of Sgr A*, there is gas at high negative velocity at -300 to -100 kms-'. As the velocities of the line emission fill the entire bandwidth, it is difficult to properly remove the continuum, especially in the presence of strong continuum emission from Sgr A West, including the very bright, compact source Sgr A*. We are fortunate, however, in the fact that the two velocity components are separated spatially (i.e., high negative and velocities around zero are not emitted from the same place in the image). This spatial separation allows us to remove the continuum by first making two image cubes. For the first cube, the line emission is assumed to be at negative velocities (centered around -300 to -100 kms-l); a linear continuum is fit to the corresponding continuum channels in two ranges at the edge of the band (-420 < V L ~
Astron. Nachr./AN 324. No. S 1 (2003)

623

would have less (even negative) line emission than a properly subtracted continuum. In this manner the two cubes were merged to create a single cube that represented only the line emission. For each epoch, we then fit a single Gaussian component to the line emission at every pixel in the image where the line was detected at 3-a above the noise. Gaussian fits were considered valid where the signalto-noise in all parameters (line intensity, width and center velocity) were above 3. The three parameters and their associated errors were recorded for the data cube obtained at each of the three epochs. At each point where there were valid fits in all three epochs, a least squares fit of a linear change of radial velocity as a function of time (i.e., radial acceleration) was carried using the center velocity and errors of all three epochs. An image showing the distribution of radial acceleration is shown in Fig. I .

21"

24" n

0 0 0 01 -3

27"

W

.-0 4

0

30"

33"

3 6" -2 !goo1'39" 17h45m41?04058 4056 4054 4032 40?0 3958 3956 3954 Right Ascension (JZOOO) Fig. 1 Image of radial acceleration from H92a line emission observed at three epochs (1993, 1999 & 2002). The color of the image represents radial acceleration toward (bright) and away (dark)from us. The four circles are averages of acceleration over representative areas. The fBsymbol indicates acceleration away from us and the 0 indicates acceleration toward us; the size of the circle is proportional to the magnitude of acceleration. The star (A) marks the position of Sgr A*. The triangle (A) marks the position of the peculiar acceleration discussed in Section 3.3. The magnitudes of acceleration at the indicated positions are listed in Table 2.

Figure 1 shows significant detections of radial acceleration in regions with 10" of Sgr A*. The color in Fig. I represents the acceleration from dark to light (corresponding to acceleration toward and away from us). In order to compare the acceleration results with information from other images, we have averaged the accelerations (which are determined on a pixel-by-pixel basis in Fig. 1) over four representative areas. The @ symbol indicates acceleration away from us and the 0 indicates acceleration toward us; the size of

D. A. Roberts and F. Yusef-Zadeh: Radial Acceleration of Ionized Gas near Sgr A*

624

Table 2 Derived Radial Acceleration

Positiona

aradial

rproj

1 2

+0.21 kms-lyr-l -0.52 kms-lyr-l -0.27 kms-lyr-' -0.26 kms-lyr-l

0.14~~ 0.20 pc 0.21 pc 0.28 pc

3 4 a: Positions labeled

in Fig. 1

the circle is proportional to the magnitude of acceleration. The star (*) marks the position of Sgr A* and the triangle (A) marks the position of the peculiar acceleration discussed in Section 3.3. The magnitudes of acceferation and the projected distance of the indicated positions from Sgr A* are listed in Table2. For spatial reference, symbols (0and €B) representing averaged accelerations and are overlaid on Figs. 2 and 3. Figure 2 shows a greyscale image of the 2 ern continuum radio emission from Sgr A West with vectors showing the direction and magnitude of gas proper motion averaged over the respective regions (shown as boxes centered on the tails of the vectors) from Yusef-Zadeh, Roberts & Biretta 1998.

n

0 0 0

C!

27'*

I=

.-0

3 30" C

.-

0 Q)

D

33" 36"

-2 !9'00'39'' 17h45m41?04058 4096 4054 4052 4050 3958 3956 3954 Right Ascension (J2000) Fig. 2 Greyscale image of radio continuum emission at 2 cm observed in 1990. The vectors show the direction and magnitude of gas proper motion averaged over the respective regions (shown as boxes centered on the tails of the vectors) from Yusef-Zadeh, Roberts & Biretta 1998. The four circles are averages of radial acceleration (see Fig 1 for more additional information) over representative areas. The star (*)marks the position of Sgr A*.

Astron. Nachr./AN 324. No. S1 (2003)

625

Figure 3 shows the radial acceleration overlaid on an color image representing the radial velocities of the gas near Sgr A* (image from Roberts, Yusef-Zadeh & Goss 1996). The colors of radial velocities that are shown represent velocities (from light to dark) of +20 kms-l to -420 kms-l. The overlaid vectors are the same ones that are shown in Fig. 2.

21 "

24" h

-0. 0 0

27"

30" u 0, n

33"

36"

-2 :goo'0'39" 17h45m415040?8 40s6 4054 4052 4050 3958 3956 39?4 Right Ascension (J2000) Fig. 3 Image of radial velocity from H92a line emission observed in 1993 (first epoch of the acceleration project).

The vectors show the direction and magnitude of gas proper motion averaged over the respective regions (shown as boxes centered on the tails of the vectors) from Yusef-Zadeh, Roberts &Biretta 1998. The color of the image represents radial velocities between +20 (light) and -420 (dark). The four circles are averages of radial acceleration (see Fig. 1 for more additional information) over representative areas. The star (*)marks the position of S g r A*.

3 Discussion Our ultimate goal of this investigation is to determine the 3-dimensional kinematics of the ionized gas observed near Sgr A*. These new acceleration data provide an additional constraint to the gas motion. We are in the process of creating a more thorough model of the motions. For these proceedings, our primary objective is to determine whether the observed radial acceleration is consistent with Keplarian orbits or whether the measurements require another non-gravitational source to explain the gas motions. We carried out this comparison in each of the four identified regions.

626

D. A. Roberts and F. Yusef-Zadeh: Radial Acceleration of Ionized Gas near Sm A*

Since our primary goal is to determine whether the motion is dominated by the gravitational potential of Sgr A*, we begin by determining the role of gravity on the gas motion. The gravitational acceleration (total) is given by atotaz =

GM, r2

-

These observations are only able to measure the radial component of acceleration (aradzal), which is atotalsinO, where 0 is the angle between the total acceleration vector and the plane of the sky. For gravitational acceleration from a point mass, the angle (0) is directed from the gas toward Sgr A*. Thus, the magnitude of the radial acceleration is GM,sinB aradaal =

~

r2

We only know the distance between the gas of interest and Sgr A* only in projection on the plane of the sky (rProJ= rcos8). Here 0 is the same angle between the plane of the sky and the radial vector between the gas and Sgr A*. For each of the four positions, we use the observed radial acceleration for the ionized gas and projected distance between the gas and Sgr A* to solve for an angle between the vector joining the gas to Sgr A* to the line of sight. We assume that the enclosed mass includes both Sgr A* (2.5 x lo6 Ma) and a stellar component 4.25 x lo6 A&, r1 where r is in parsec.

',

3.1 Systematic Motion of the Northern Arm We are able to determine 0 for the radial acceleration measured in the Northern Arm (position 1 in Fig. 1) to be 25", which implies an inclination of the orbital plane to be 65". This is consistent with the geometry of gas in an orbit around Sgr A* and supports earlier modeling of the orbit of ionized gas in Sgr A West based only on radial velocity measurements (see Roberts, Yusef-Zadeh & Goss 1996, Vollmer & Duschl 2000), as well as determinations including both radial velocity and proper motion results (see Yusef-Zadeh, Roberts & Biretta 1998). The angle (0) and the positive sign of the radial acceleration is consistent with a model in which the northern part of Northern Arm is infiont of Sgr A*. Thus, we conclude that the motion of the northern part of Northern Arm is influenced primarily by gravity.

3.2 The Minicavity The same analysis was carried out for the two positions in the minicavity (positions 2 & 3 in Fig. 1). For these two positions, the magnitude of measured radial acceleration is much too large to be explained by gravity. Even though the observed magnitude of the acceleration cannot be explained by gravitational acceleration, we can assume that gravity contributes significantly to observed motion. Under this assumption, we can investigate whether the sign of the radial acceleration supports current kinematic models. The negative sign of the acceleration of the minicavity is consistent with it being behind Sgr A*. Thus, it is likely that the motion of the gas in the minicavity is influenced by both gravitational and non-gravitational effects, although additional modeling is needed to determine the origin of the acceleration. Stolovy (this proceeding) have reported the presence of mass-losing stars in the minicavity. We consider whether the observed radial acceleration could be due the ram pressure from stellar winds. The pressure of n;l winds from hot stars (&f- lop5 Mayr-l, ZIwind M 100 kms-') would be P W & ~ , where p = 4nR2v,,,d. In order for the winds to be responsible for the acceleration, the mass losing stars would have to be very close (- 0.025 pc) to the ionized gas.

Astron. Nachr./AN 324, No. S 1 (2003) 3.3

627

Peculiar Motion

There is one case of a peculiarly high radial acceleration (+4.1 I!Z 0.6 k m - ’ yr-l). This is observed southeast of Sgr Ah (shown as the A symbol in Fig. 1). This acceleration is higher by a factor of two than any other position measured. At this position, the change in radial velocity is readily visible by comparing spectra taken at different times. Figure 4 shows profiles at this position from the 1993 and 2002 epochs. This source is likely a result of large local acceleration. It maybe related to sources like the “Bullet” (see Yusef-Zadeh, Roberts & Biretta 1998). which are fast moving objects in the inner few parsecs of the Galaxy. 1993 and 2002 Profiles

1 -a

3 0

t . . . . ~ . . . . ’ . . . . V’ . . . . ~ ~ -300

-100

-400

-200

E L 0 (km/s)

Fig. 4 Plot of H92a spectra toward the location where peculiar high radial accelerations are observed. The two spectra show the line observed in 1993 (thick line) and 2002 (thin fine). The spectra were taken from a location southeast of Sgr A*, which is indicated by the triangle (A) in Fig. 1. The derived acceleration at this position (obtained with data from 1993, 1999 & 2002) is +4.1 0.6 h s C 1 y r - ’ .

References Fhetz, A.M., Klein, B.L., Moms, M. & Becklin E.E. 1998, ApJ, 509,678 Eckart, A. & Genzel, R. 1997, MNRAS, 284,576 Roberts, D.A. & Goss, W.M. 1993, ApJS, 86, 133 Roberts, D.A., Yusef-Zadeh, E &Goss, W.M. 1996, ApJ 459,627 Stolovy, S . 2003, these proceedings Vollmer, B. & Duschl, W.J. 2000, New Astronomy 4, 581 Yusef-Zadeh, F., Roberts D.A., &Biretta, J. 1998, ApJ, 499, L159 Zhao, J.-H. & Goss, W.M. 1998, ApJ, 499, L163

This Page Intentionally Left Blank

Astron. Nachr./AN 324, No. S 1,629-634 (2003) / DO1 10.1002/asna.200385057

Simple hydrodynamical Simulations of the Circumnuclear Disk Robert F. Coker*’,Michol H. Christopher2,Susan R. Stolovy3,and Nick Z. Scoville2

’ Los Alamos National Laboratory, T-087, Los Alamos, NM 87545 Caltech, Department of Astronomy, 105-24, Pasadena, CA 91 125 SIRTF Science Center, Caltech, 314-6, Pasadena, CA 91 125

Key words GMC, simulation, Sgr A Abstract. The “circumnuclear disk” (CND) is a dense, clumpy, asymmetric ring-like feature centered on Sgr A*. The outer edge of the CND is not distinct but the disk extends for more than 7 pc: the distinct inner edge, at a radius of N 1.5 pc, surrounds the “mini-spiral” of the HII region, Sgr A West. We present simple 3D hydrodynamical models of the formation and evolution of the CND from multiple selfgravitating infalling clouds and compare the results with recent observations. We assume the clouds are initially Bonner-Ebert spheres, in equilibrium with a hot confining inter-cloud medium. We include the gravitational potential due to the point-mass of Sgr A* as well as the extended mass distribution of the underlying stellar population. We also include the effects of the ram pressure due to the stellar winds from the central cluster of early-type stars. A single spherically symmetric cloud cannot reproduce the clumpy morphology of the CND; multiple clouds on diverse trajectories are required so that cloud-cloud collisions can circularize the clouds’ orbits while maintaining a clumpy morphology. Collisions also serve to compress the clouds, delaying tidal disruption while potentially hastening gravitational collapse. Low density clumps are disrupted before reaching the inner CND radius, forming short-lived arcs. The outer parts of more massive clumps get tidally stripped, forming long-lived low-density wide-angle arcs, while their cores potentially undergo gravitational collapse. The fine balance between resisting tidal disruption and preventing gravitational collapse implies that most if not all clumps are not stable for much more than an orbit. Thus, in order for the CND to be a long-lived clumpy object, it must be continually fed by additional in-falling clouds. Clouds that survive to small radii are likely to be the sites of present or future star formation. However, within a few parsecs of Sgr A*, the stellar winds decelerate any in-falling cloud so that the wind-cloud interface becomes Rayleigh-Taylor unstable, potentially disrupting the cloud and inhibiting star formation.

1 Introduction The proper motions of the young, bright stars in the central parsec of our Galaxy strongly suggest (Genzel et al. 2000; Ghez et al. 2000) that Sgr A*, the stationary, compact, non-thermal radio source (Backer & Sramek 1999; Melia & Falcke 2001) located at the very heart of our Galactic Center (GC), is a 2.6 0.2 x 1 O 6 M dark ~ mass. It is probable that the emission we see is associated with the accretion of matter onto a super-massive black hole (Genzel & Eckart 1999). The 7000 K HI1 region Sgr A West, known as the Mini-spiral, surrounds Sgr A* and is about 1.5 pc in radius (Roberts & Goss 1993; Ekers et al. 1975). The western edge of the Mini-spiral is coincident with the inner edge of a ring of molecular and atomic gas. This ring is centered on Sgr A* and is known as the Circumnuclear Disk (CND). Many observations of the CND have been made over the years (e.g., Gatley et al. 1986, Serabyn et al. 1986, Jackson et al. 1993, Wright et al. 2001, Christopher et al. 2002). The CND is asymmetric and very 1.5 pc; its outer edge is less distinct but it extends for clumpy with a distinct inner edge at a radius of more than 7 pc. The mass of the CND is estimated to be 1 0 4 - 5 Ma with a filling factor of about 0.01. Individual clouds have masses in the range of 1 to 1 0 3 M a , sizes of 0.1 - 0.4 pc, temperatures of more . is evidence that some of the CND clouds are than 100 K, and number densities of 105-6 ~ m - ~There

*

N

‘ Corresponding author: e-mail: robcQlanl.gov,Phone: +I-505-665-1245, Fax: +1-505-665-1231 @ 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

R. F. Coker et al.: Simulations of the Circumnuclear Disk

630

on Keplerian trajectories and are stretched perpendicular to their radii towards Sgr A* (Christopher et al. 2002). It has been estimated that the collisional time-scale of individual CND clouds is more than 1 Myr (Vollmer & Duschl2001b) but the lifetime of the CND itself may be much shorter since many clouds are tidally unstable and may be disrupted in less than lo5 yrs (Gusten et al. 1987; Christopher et al. 2002). For comparison, if the clouds in the CND are pressureless and uniform, their gravitational free-fall collapse time would be Y lo5 yrs. Also, it appears that some of the Giant Molecular Clouds (GMCs) in the GC are interacting with the outer edge of the CND (Coil & Ho 2000). Clearly, the CND plays a key role in how gas reaches the central pc and Sgr A*; it may also play a critical role in star formation in the GC since many of the young stars of the central parsecs are associated with it. Thus, we have set out to model the formation and evolution of the CND. Previous investigations of the formation and evolution of the CND have utilized collisional N-body (Vollmer & Duschl2002) and sticky particle (Sanders 1998) codes. In this work, we use a version of the finite-difference Eulerian code ZEUS3D (Norman 1994) with self-gravity to construct initial simple 3D hydrodynamical models of the CND.

2 The Simple Hydrodynamical Model The premise of the model is that the CND is formed and maintained by GMCs falling down the Galactic potential well towards Sgr A*. This potential is assumed to be due to a spherical distribution of mass such that within a radius T (in pc) from Sgr A*, the total enclosed mass is given by (see, e.g., Genzel et al. 2000, Fig. 17):

men,-(^) = 2.6 x lo6 + 1.6 x 106~1.25Ma. The first term in Eq. 1 is due to the point mass of Sgr A* and the second is due to the stellar potential. Interestingly, the stellar potential begins to dominate that of the point mass at a radius of 1.5 pc, the inner radius of the CND. Also, due to the increasing mass of the stellar cluster with radius, there is a shallow 4.5 pc, very close to the average radius of minimum in the Keplerian velocity of a circular orbit at the CND. Thus, one would naturally expect material to collect around 4.5 pc and be potentially disrupted within 1.5pc. Each GMC is initially a non-isothermal version of a Bonner-Ebert sphere (Bonner 1956; Ebert 1955); that is, they are in gravitational hydrostatic equilibrium with a hot, confining ISM. We use a canonical adiabatic index y = 1.1since GMCs are generally nearly isothermal. Since in reality there are supersonic motions in GMCs that are on the order of 10s of km s-l, we use an effective central temperature T, = 1000 K. Finally, choosing a central density pc determines the size of the GMC. Note that clouds which have an initial non-zero velocity will not strictly be in equilibrium and, given time, would be expected to collapse = lo6 K thus perpendicular to their direction of motion. The ISM is assumed to have a temperature TISM the ISM number density, TIISM, is 0.001 times the surface density of the GMC. In addition to self-gravity and the gravity of the point mass and underlying stellar population, we have included the stellar winds from the central star cluster, IRS 16. To model the winds, we inject a total of 3 x 10-3M,3yr-1 into cells within 0.25 pc of Sgr A*. This gas is given a radially outward velocity of 700 km s-' and a temperature of lo6 K. We ignore the UV radiation field from the central cluster; the ionization of the illuminated side of a GMC would result in gas from that side being ejected into the ISM and thus a smaller cloud (Vollmer & Duschl 2001a). N

N

3 Specific Simulations We have constructed two general types of models: single massive GMCs and multiple smaller GMCs. These models have been run with various orbits, resolutions, and cloud sizes. The small GMCs have radii

Astron. Nachr./AN 324, No. S l (2003)

-

63 1

-

of 0.2 pc, central densities of lo6 cmP3, and masses of 450Ma. The large GMCs have radii of 2.0 pc, central densities of lo4 cm-3, and masses of 45OOMa . Most simulations have 2563 cells.

3.1 Single Cloud Results A single GMC with a large impact parameter, b, compared to its radius, gets tidally stripped into largeangle arcs within a few orbits. An example of this is shown in Fig. 1, where a single GMC has entered from the lower left with an upward trajectory. In the figure, the cloud has moved less than an orbit; the core of the cloud is still visible but after a few orbits the entire cloud is smeared out over most of the volume of the simulation. Some of the arcs fall onto Sgr A*, potentially resulting in a mass accretion rate onto the putative black hole of up to 10p3M0yr-’. At no point do repeated collisions of parts of the cloud result in regions denser than the initial cloud’s central density (lo4 cmP3). This is in contrast to sticky particle calculations (Sanders 1998) which show arcs of the disrupted GMC colliding with itself, resulting in regions of significantly enhanced density. The IRS 16 winds have a ram pressure of = 9 x l0-’/r2 dyne cmP2, where T is the distance from Sgr A* in pc. The large GMCs have a gravitational binding energy density of E 2 x lo-’ erg cmP3. Thus, the winds are only dominant within two pc or so of Sgr A* and have little effect on clouds with large impact parameters. A GMC with an impact parameter smaller than its radius will produce a ring of ejected material. This is shown in Fig. 2, where a cloud with b = 0 has been entirely disrupted. Again, large-angle arcs form but no clumps and no regions of high density. After a few lo5 yrs, the whole structure smoothes out. Interestingly, the energy of the ejected material is comparable to that of a supernova. Since a cloud with b = 0 approaches the singularity of the point mass, the detailed structure and evolution depends on the physical resolution of the simulations (the 3D calculations are not converged in resolution). Nonetheless, it appears that the clumpy nature of the CND cannot be reproduced by the disruption of a single GMC. However, turning on the IRS 16 winds in the case of b = 0 results in an unstable interface between the cloud and the wind. As shown in Fig. 3, this Rayleigh-Taylor unstable interface clumps up and could potentially result in small dense clouds which match CND observations; this simulation did not have sufficient resolution to follow the development of the instability any further than is shown in Fig. 3.

-

3.2

Multiple Cloud Results

The multiple GMC models included 24 clouds, each on a Keplerian orbit. In order to mimic a velocity dispersion in the group of clouds, the orbits were perturbed by a Gaussian distribution of random velocity . models were intended to investigate how clumps in components with a magnitude of 0.25 V K ~ ~These an existing CND would evolve. As expected, some of the small GMCs merge with each other, some fall inwards towards Sgr A*, some escape the system entirely, and some are directly disrupted into short arcs of gas. This is shown in Fig. 4. Merged GMCs generally have densities which are temporarily enhanced in the azimuthal direction since the chosen initial orbits rarely produce radial collisions. However, after a few orbits ( w lo5 yrs), all of the clouds are disrupted and dispersed. Before disruption though, these small merged GMCs match the size (- 0.1 - 0.4 pc). central number density ( w lo6 cmP3), and morphology (stretched azimuthally) of the observations. Note that with the increased binding energy of the smaller clouds relative to the larger GMCs (E,,, cx p;’ for an EB sphere), the IRS 16 winds are not likely to significantly affect the clouds outside of a pc.

4 Discussion & Conclusions The preliminary, simple calculations presented here do not include radiative heating or cooling, magnetic fields, phase transitions or possible stellar sources within the GMCs. Nor do they consider rotating or inhomogeneous clouds. Nonetheless, they are the first attempt at 3D hydrodynamical models of the evolution of GMCs in the environment of the central pcs of the Galaxy. In the future, we will run higher resolution multi-material simulations which include radiative heating and cooling. We will also consider non-hydrostatic and rotating clouds.

632

R. F. Coker et al.: Simulations of the Circumnuclear Disk

There is clearly a fine balance between gravitational collapse and tidal disruption in the GC. Only those clouds dense enough to be stable against tidal disruption can approach Sgr A* while clouds which are too dense undergo gravitational collapse before they reach the central parsecs. Even rotating clouds can only get to within N 2 pc (Vollmer & Duschl2001b) of Sgr A* before their critical density to resist tidal disruption exceeds their critical density for gravitational collapse. Combined with the observation that the clouds now in the CND are not tidally stable (Christopher et al. 2002), this implies that individual clumps within the CND are short-lived (- lo5 yrs) objects and if the CND as a whole is a long-lived clumpy object ( 2 lo6 yrs), it must be continually fed by in-falling clouds. Only multiple GMCs can reproduce the characteristics of the CND and even then only for a short time; a single large GMC cannot, even through multiple self-collisions, reproduce the CND’s morphology. If the GMCs presently interacting with the CND (Coil & Ho 2000; Zylka et al. 1990) are any indication, the GMCs that formed the CND came from 2 100 pc out and took 2 lo6 yrs to fall into the central parsecs. Thus, they could have formed stars while falling into the GC and the clumps we see today in the CND could be the remnants of their cores. If so, there would likely have been recent star-formation going on in the CND clumps. If these GMCs produced the hot, massive stars such as IRS 16 that pervade the central parsecs, this would resolve the difficulty of how so many young stars exist so near a super-massive black hole. The role of the stellar winds from the IRS 16 cluster is not clear. Any interface between the hot winds and the GMCs will be unstable; however, it is not clear if the instability would induce collapse and subsequent star formation or if it would completely disrupt the GMC. Higher resolution simulations will help address this issue. Also, the in-falling streamers of stripped GMCs (e.g. Fig. 4) will be ionized by the IRS 16 stars, perhaps producing the Sgr A West HI1 region. Acknowledgements This work was supported in part by UK PPARC and DOE. Some of the calculations reported here were carried out using UKAFF, a high-performance computing facility funded by UK PPARC.

References Backer, D. C. & Sramek, R. A. 1999, ApJ, 524,805 Bonner, W. B. 1956, MNRAS, 116,351 Christopher,M., Scoville, N., Stolovy, S., Yun, M., & Coker, R. 2002, AGN: from Central Engine to Host Galaxy, Eds.: S. Collin, F. Combes and I. Shlosman. , ASP Conference Senes, p49 Ebert,R. 1955.Z. fiirAstr.,37,217 Ekers, R. D., Goss, W. M., Schwarz, U. J., Downes, D., & Rogstad, D. H. 1975, A&A, 43, 159 Gatley, I., Jones, T. J., Hyland, A. R., Wade, R., & Geballe, T. R. 1986, MNRAS, 222,299 Genzel, R. & Eckart, A. 1999, ASP Conf. Ser. 186: The Central Parsecs of the Galaxy, 3 Genzel, R., Pichon, C., Eckart, A., Gerhard 0. E., & On, T. 2000, MNRAS, 317,348 Ghez, A. M., Moms, M., BecMin, E. E., Tanner, A., & Kremenek, T.2000, Nature, 407,349 Gusten, R., Genzel, R., Wright, M. C. H., Jaffe, D. T., Stutzki, J., &Harris, A. I. 1987, ApJ, 318, 124 Jackson, J. M., Geis, N., Genzel, R., Harris, A. I., et al., 1993, ApJ, 402, 173 Melia, F. & Falcke, H. 2001, ARA&A, 39,309 Roberts, D. A. & Goss, W. M. 1993, ApJS, 86, 133 Sanders, R. H. 1998, MNRAS, 294,35 Serabyn, E., Gusten, R., Walmsley, J. E., Wink, J. E., & Zylka, R. 1986, A&A, 169,85 Vollmer, B. & Duschl, W. J. 2001a, A&A, 367,72 Vollmer, B. & Duschl, W. J. 2001b, A&A, 377, 1016 Vollmer, B. & Duschl, W. J. 2002, A&A, 388, 128 Wright, M. C. H., Coil, A. L., McGary, R. S., Ho, P. T. P., & Harris, A. I. 2001, ApJ, 551,254 Zylka, R., Mezger, P. G., & Wink, J. E. 1990, A&A, 234, 133

Astron. Nachr./AN 324. No. S1 (2003)

633

Fig. 1 A log plot of density in the z-plane showing a disrupted large cloud at t CY lo5 yrs which had an initial velocity of 100 km s-' and b = 3 pc, and was falling towards Sgr A*, located at the center, from the lower left in the figure. The . figure is 15 pc on a side. The IRS 16 winds Black corresponds to n 5 lop3 cm-3 and white to n ,? 10' ~ m - ~ were not turned on for this simulation.

Fig. 2 A log plot of density in the z-plane showing a disrupted large cloud at t CY 8 x lo4 yrs whlch had a small initial velocity, b = 0, and was falling towards S g r A*, located at the center, from below in the figure. Black corresponds to n5 ~ r n and - ~ white ton 2 lo3 ~ r n - ~The . figure is 15 pc on a side. The IRS 16 winds were not turned on for this simulation.

634

R.F. Coker et al.: Simulations of the CircumnuclearDisk

Fig. 3 A log plot of density in the z-plane showing a disrupted large cloud like that in Fig. 2. Black corresponds to n 5 10-1 cmP3 and white to n ,?, lo3 ~ m - ~The . figure is 15 pc on a side. The IRS 16 winds were turned on for this simulation.

Fig. 4 A log plot of density in the z-plane for 24 small orbitlng clouds with perturbed orbits at t N lo5 yrs. Black . figure is 15 pc on a side. The IRS 16 winds were corresponds to n 5 lo-' cmP3 and white to n ,?, lo3 ~ r n - ~The not turned on for this simulation.

This Page Intentionally Left Blank