COSPAR COLLOQUIA SERIES VOLUME 11
THE O U T E R HELIOSPHERE: THE NEXT F R O N T I E R S
PERGAMON
This Page Intentionally Left Blank
THE OUTER HELIOSPHERE: THE N E X T FRONTIERS Proceedings of COSPAR Colloquium held in Potsdam, Germany 24-28 July 2000
Edited by
Klaus Scherer dat-hex, Katlenburg-Lindau, Germany
Horst Fichtner Institutfiir Theoretische Physik IV, Ruhr-Universiti~tBochum, Bochum, Germany
Hans J0rg Fahr Institut Ff~rAstrophysik und extraterrstrische Forschung, Universittit Bonn, Bonn, Germany
Eckart Marsch Max-Planck-Institutfiir A eronomie, Katlenburg-Lindau, Germany
2001
) PERGAMON An Imprint of Elsevier Science Amsterdam
- London
- New
York
- Oxford
- Paris
- Shannon-
Tokyo
ELSEVIER SCIENCE B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam, The Netherlands
Copyright 9 2001 COSPAR. Published by Elsevier Science B.V.
This work is protected under copyright by Elsevier Science, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier Science Global Rights Department, PO Box 800, Oxford OX5 1DX, UK; phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail:
[email protected]. You may also contact Global Rights directly through Elsevier's home page (http://www.elsevier.nl), by selecting 'Obtaining Permissions'. In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 207 631 5555; fax: (+44) 207 631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of Elsevier Science is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier Science Global Rights Department, at the mail, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made.
First edition 2001 Library of Congress Cataloging in Publication Data A catalog record from the Library of Congress has been applied for.
ISBN:
0 4 4 4 50909 7
@ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.
Preface The 11th COSPAR Colloquium "The Outer Heliosphere: The Next Frontiers" was held in Potsdam, Germany, from July 24 to 28, 2000, and is the second dedicated to this subject after the first one held in Warsaw, Poland in 1989. Roughly a century has passed after the first ideas by Oliver Lodge, George Francis FitzGerald and Kristian Birkeland about particle clouds emanating from the Sun and interacting with the Earth environment. Only a few decades after the formulation of the concepts of a continuous solar corpuscular radiation by Ludwig Biermann and a solar wind by Eugene Parker, heliospheric physics has evolved into an important branch of astrophysical research. Numerous spacecraft missions have increased the knowledge about the heliosphere tremendously. Now, at the beginning of a new millennium it seems possible, by newly developed propulsion technologies to send a spacecraft beyond the boundaries of the heliosphere. Such an Interstellar Probe will start the in-situ exploration of interstellar space and, thus, can be considered as the first true astrophysical spacecraft. The year 2000 appeared to be a highly welcome occasion to review the achievements since the last COSPAR Colloquia 11 years ago, to summarize the present developments and to give new impulses for future activities in heliospheric research.
The conference drew not only scientific but also public interest, which became evident from a variety of related TV and radio presentations as well as a couple of articles in the local press. To cover all aspects of outer heliospheric research and future exploration the conference programme was structured into eleven major sessions, namely: the Large-scale Structure of the Heliosphere, Heliospheric and Interstellar Connections, Messengers from the Outside, Echoes from the Heliopause, the Heliosphere and the Galaxy, Views of Space Agencies and Companies, At the Edge of the Solar System, New Kinetic Aspects of Heliospheric Physics, Modern Heliospheric Spacecraft and Missions, Modern Heliospheric Technologies, and Connections to Earth.
my---
Preface We thank all the participants who have submitted a written version of their presentations and also all referees who helped to improve these manuscripts, and, thus, enabled us to compile this volume of high-quality articles. The remaining presentations are included with their abstracts, and, if published elsewhere, a reference is given. A special thanks go to the conveners for putting together the invited talks. Finally, we are grateful that the Colloquium was held under the auspices of COSPAR.
Klaus Scherer Horst Fichtner Eckart Marsch Hans J. Fahr
--
vi'-
Acknowledgements
-- vii-
Prologue Following the example of the first COSPAR Colloquium of the Outer Heliosphere, also during the second COSPAR Colloquium on the same topic a poll about the distance of the heliospheric termination shock has been taken from the participants. The results of both polls are shown in the figure. On the basis of these "data" we are now able to perform a "non-standard" analysis of the shock dynamics.
Mean S h o c k D i s t a n c e 96.96 [AU]
20.0.
COSPAR
g
~
i0(~
90.0 H e l i o s p h e r i c t e r m i n a t i o n s h o c k d i s t a n c e [AU]
Figure 1. The polls taken at the two COSPAR Colloquia on the Outer Heliosphere in New Hampshire (left panel) and in Potsdam (right panel).
Vshock ~
~ 3.27AU yr -1 11 yrs which is, interestingly enough, about the speeds of the Voyager 1 and 2 spacecraft, of 3.5 and 3.1 AU yr -1. We included an additional question in the Potsdam poll: Will the Voyagers reach the shock at all ? The result was, that 83% of the participants of the poll expected this to happen, 11% did not, and 6% made no guess. From our high-quality data set and the fact that Voyager 1 was at 76 AU at the time of the Colloquium, we can estimate the shock encounter to happen in the year 2091 at a distance of about 395 AU under the (probably wrong) assumption that the shock is continuously moving outward. To check these and other (more serious) predictions made during this second Colloquium on the Outer Heliosphere, we suggest to organize a third one in the year 2011, i.e. again another solar activity cycle later.
--
viii-
Opening Remarks by the COSPAR President
Dear Organizers, Dear Participants, Ladies and Gentlemen,
It is a pleasure for me to convey the best wishes of COSPAR Bureau and Secretariat for a successful Colloquium and at the same time our sincere thanks to the organizers and editors for having undertaken the task to organize this event and publish its proceedings. The great number of attendees tells me that it was timely and worth while to do so. Let me formulate a few thoughts which were stimulated by the theme of your meeting, the outer heliosphere. The scales of the universe are incomprehensible. We, who deal every day with them, have of course developed a way to work with these scales by introducing the parsec as unit and thus shrink the universe into something that can be handled and visualized. We can easily feel at home in this vast space and forget how small the part is that we are occupying. The heliosphere, although small by comparison, is well suited to teach us modesty in two respects. We realize the true size of cosmic distances much better when we think that it takes one day for light to cross the heliosphere, with the nearest star being at a distance of four light years. But it makes us even more humble when we reflect about the completely accidental and by no means distinguished location of the heliospere inside some odd filament of a supernova remnant. If we truely tried to absorb these facts, should we not be equallly shaken as the Portuguese were in 1755 after the earthquake of Lisbon? Under the impression that it struck good and evil people, pious ones and heretics without differentiation, they began to ask questions like: is god not good? does god really exist? Should our questions not be: is our life, are we really the culmination of the universe? are we unique? For you who have assembled here during this week the heliosphere is a place of action, of the solar wind colliding with the interstellar medium, of the interplanetary magnetic field winding up and forming shocks, of interstellar ions being picked up to be accelerated to very high energies, of radio waves bringing back messages from the boundaries, of dust particles telling stories about the chemistry in the expanding clouds of exploding stars, and of many other cosmic processes which you can study here in situ in an exemplary fashion. The communication of your latest insights will keep you away from unhealthy thoughts like the above. You might also consider more practical matters like advancements of the propulsion techniques or the technology of large extended surfaces enabling farther and more efficient space travel. You may wonder whether Europe is going to confine itself for ever to "- iX--
Opening remarks by the COSPAR President
the innermost part of the solar system because of its reluctance to fly RTG's. I think that a meeting like this could also serve as a platform from which messages can go to the agencies and the public about the long-range needs of your science and its benefits in understanding our position in the universe. There is also a task in education. I do not think that scientists have to convey to the public all the deep lying discoveries of their respective disciplines. But teaching them respect for the basic laws of nature and making them aware of the reality of our place in the universe is a duty that we should accept again and again. The imagination of our children is so full of aliens, of UFO's and supermen that they may think that everything is possible and never realize how confined we are to our vulnerable planet. Telling stories about the solar system, the true extent and emptiness of the heliosphere, the efforts and times needed to conquer its distances will not only correct some of the light-heartedly accepted illusions, but also raise curiosity about the wider environment and origin of our Earth. Heliospheric physics is eminently suited to fulfill such an educational role. Finally, I have a few words about the publication of your contributions within the COSPAR Colloquium Series. You and also COSPAR want the book to be purchased and read by a community much wider than what is assembled in this room. Therefore, quality of the papers and reasonably wide interest of the topics are of great importance. Not all of the preceding volumes fulfill these conditions. To achieve the goal implies careful refereeing and editorial judgement. One may also want to fill some gaps by accepting additional papers. You who are involved in this process do a great service to your and the wider space science community. Let me express once more my sincere thanks and wish you an exciting week in Potsdam.
Gerhard Haerendel
--"
X
"-"
Contents S e s s i o n 1: L a r g e S c a l e S t r u c t u r e
Physics of the solar wind interaction with the local interstellar medium G.P. Zank and H.-R. Mfiller Relating models of the heliosphere to Lyman-a absorption observed in Hubble spectra H.-R. Miiller and B.E. Wood
13
Interstellar atoms in the heliospheric interface V. Izmodenov
23
MHD modeling of the outer heliosphere: Numerical aspects N.V. Pogorelov
33
Interaction of the local interstellar medium with the heliosphere: Role of the interior and exterior magnetic field A. Barnes
43
Modeling stellar wind interaction with the ISM: Exploring astrospheres and their Lyman-a absorption H.-R. Mfiller, G.P. Zank, and B.E. Wood
53
Stationary MHD-equilibria of the heliotail flow D. Nickeler and H.J. Fahr
57
Non-stationary magnetic field geometry in the heliosphere I.S. Veselowsky
61
Solar cycle heliospheric interface variations: Influence of neutralized solar wind N.A. Zaitsev and V.V. Izmodenov
65
Time dependent radiation pressure and time dependent, 2D ionisation rate for heliospheric modeling M. Bzowski
69
*Magnetic fields in the heliosphere: Introductory remarks W.I. Axford
73
*Temporal variations of Lyman-alpha-intensities, backscattered from the interstellar hydrogen atoms upstream of the Sun" The ULYSSES-Gas observations M. Witte
73
General Discussion
75
S e s s i o n 2: H e l i o s p h e r i c
and Interstellar Connections
Radiative transfer at Lyman-a in the outer heliosphere E. Quemerais
79
A fluid approach to the heliosphere/VLISM problem R.L. McNutt Jr, M. Wiltberger, J. Lyon, and C.C. Goodrich
89
Possible effects of the interstellar magnetic field on the heliospheric structure and H-atom penetration to the solar system V.B. Baranov
99
*Oral papers and posters which were given at the conference, but for which no manuscripts were submitted - - xi - -
Contents
Interstellar gas flow into the heliosphere E. M5bius, Y. Litvinenko, L. Saul, M. Bzowski, and D. Rucinski
109
Non-stationary transport of neutral atoms in the heliosphere A.I. Khisamutdinov, M.A. Phedorin, and S.A. Ukhinov
121
Pickup ion turbulence: A stochastic growth model G.P. Zank and I.H. Cairns
125
A time-dependent, 3D model of interstellar hydrogen distribution in the inner heliosphere M. Bzowski, T. Summanen, D. Rucinski, and E. Kyr51~i
129
Charge exchange ionization rate of interplanetary hydrogen atoms and Lyman-a intensity pattern in the inflow direction of the interstellar gas T. Summanen, T. M~ikinen, E. Kyr51~i, and W. Schmidt
133
New results derived from Pioneer 10/11 UV data H. Scherer and K. Scherer
137
*Composition and characteristics of the local interstellar cloud and the inner source obtained from pickup ions G. Gloeckler
141
*H atom velocity distribution in the heliospheric interface V.V. Izmodenov, Y.G. Malama, M. Gruntman, and R. Lallement
141
*Estimation of interplanetary hydrogen flow parameters from SWAN Lymanalpha observations E. KyrSl~i, J. Jaatinen, T. Summanen, W. Schmidt, T. M~ikinen, R. Lallement, E. Quemerais, J. Costa, and J.L. Bertaux
142
*The influence of a time variable solar H-Lyman-alpha line profile on LISMparameter deductions from interplanetary glow spectra H. Scherer, H.J. Fahr, M. Bzowski, and D. Rucinski
142
General Discussion
143
Session 3: M e s s e n g e r s from Outside The discovery and early development of the field of anomalous cosmic rays H. Moraal
147
Anomalous cosmic rays: Current and future theoretical developments J.A. le Roux
163
Anomalous cosmic ray observations in the inner and outer heliosphere B. Heber and A. Cummings
173
Cosmic rays as messengers from outside the inner heliosphere M.A. Lee and H. Fichtner
183
Modulation region of galactic cosmic rays in the heliosphere: Estimation of dimension, radial diffusion coefficient, intensity out of region L.I. Dorman
187
Latitudinal gradients and charge sign dependent modulation of galactic cosmic
191
-- xii-
Contents
Antiprotons below 200 MeV in the interstellar medium: Perspectives for observing exotic matter signatures I.V. Moskalenko, E.R. Christian, A.A. Moiseev, J.F. Ormes, A.W. Strong
195
Anomalous cosmic rays outside of the termination shock A. Czechowski, S. Grzedzielski, H. Fichtner, M. Hilchenbach, and K.C. Hsieh
199
Ionizing media and the observed charge states of anomalous cosmic rays A.F. Barghouty, J.R. Jokipii, and R.A. Mewaldt
203
ACR modulation inside a non-spherical modulation boundary S.R. Sreenivasan and H. Fichtner
207
The injection problem for anomalous cosmic rays G.P. Zank, W.K.M. Rice, J.A. le Roux, and W.H. Matthaeus
211
Self-consistent acceleration of pickup ions at the termination shock J.A. le Roux, H. Fichtner, G.P. Zank, and V.S. Ptuskin
215
Energetic neutral helium of heliospheric origin at 1 AU A. Shaw, K.C. Hsieh, M. Hilchenbach, A. Czechowski, D. Hovestadt, B. Klecker, R. Kallenbach, E. M5bius, and P. Bochsler
219
Effects of the heliospheric termination shock on possible local interstellar spectra for cosmic ray electrons and the associated heliospheric modulation S.E.S. Ferreira, M.S. Potgieter, and U.W. Langner
223
*Galactic cosmic rays: Overview M. Forman
227
*Galactic cosmic rays: the outer heliosphere (late paper: see page 513) J.R. Jokipii
227
*Modulation of galactic and anomalous cosmic rays E. Christian, W. Binn, C. Cohen, A. Cummings, J. George, P. Hink, R. Leske, R. Mewaldt, E. Stone, T. von Rosenvinge, M. Wiedenbeck, and N. Yanaska
228
*The spectrum of ACR oxygen and its variations in the outer heliosphere from 1992 to 2000 D.C. Hamilton, M.E. Hill, N.P. Cramer, R.B. Decker, and S.M. Krimigis
228
*On the variability of suprathermal pickup He + at 1 AU B. Klecker, A.T. Bogdanov, M. Hilchenbach, A.B. Galvin, E. M5bius, F.M. Ipavich, and P. Bochsler
229
*Observations of pickup ions in the outer heliosphere by Voyager 1 and 2 and implications on pressure balance S.M. Krimigis and T.B. Decker
229
*Consequences of recently observed galactic synchrotron radio emissions on the local interstellar spectrum for cosmic ray electrons U.W. Langner, O.C. De Jager, and M.S. Potgieter
230
*Variation of the fluxes of energetic He + and He 2+ during the passage of corotating interaction regions D. Morris, E. M5bius, M.A. Popecki, L.M. Kistler, A.B. Galvin, B. Klecker, and A. Bogdanov
230
*The cosmic ray electron to positron ratio in the heliosphere M.S. Potgieter and U.W. Langner
231
General Discussion
233
o . .
--
XIII
---
Contents
Session 4: Echoes from t h e H e l i o p a u s e Energetic neutral atom imaging of the outer heliosphere- LISM interaction H.O. Funsten, D.J. McComas, and M. Gruntman
237
Dual spacecraft measurements as a tool for determining the source of low frequency heliospheric radio emissions W. S. Kurth and D. A. Gurnett
245
Theories for radio emissions from the outer heliosphere I.H. Cairns and G.P. Zank
253
Mapping the heliopause in EUV M. Gruntman
263
Energetic neutral hydrogen of heliospheric origin observed with SOHO/CELIAS at 1 AU M. Hilchenbach, K.C. Hsieh, D. Hovestadt, R. Kallenbach, A. Czechowski, E. MSbius, and P. Bochsler
273
Acceleration of pickup ions at the termination shock in the limit of weak scattering S.V. Chalov and H.J. Fahr
277
Doppler shifted photon emission expected due to reactions of energetic protons with the LISM atoms in the heliosphere M. Hilchenbach, K. C. Hsieh, and A. Czechowski
281
A new diagnosis tool to map the outer heliosphere regions R. Ratkiewicz and L. Ben-Jaffel
285
The Lyman-a echo from the heliospheric bow shock region and its observability from earth H.-J. Fahr, H. Scherer, G. Lay, and M. Bzowski
289
*Energetic neutral atoms in the heliosphere M.A. Gruntman
295
*Radio emission from the outer heliosphere D. Gurnett and W.S. Kurth
295
*Energetic neutral atoms as tracers of the ionization state of the local interstellar medium V.V. Izmodenov and M. Gruntman
296
*Simulation of ENA images of the heliospheric termination shock and interface region E.C. Roelof
296
General Discussion
297
Session 5: T h e H e l i o s p h e r e a n d G a l a x y The solar wind: Probing the heliosphere with multiple spacecraft J.D. Richardson
301
Propagation of the solar wind from the inner to outer heliosphere: Three-fluid model C. Wang and J.D. Richardson
311
Relationships of corotating rarefaction regions outside 40 AU with solar observations: Heliospheric mass loading A. Posner, N.A. Schwadron, and T.H. Zurbuchen
315
-- xiv --
Contents
Recurrent ion events and plasma disturbances at Voyager 2" 5 to 50 AU R.B. Decker, C. Paranicas, S.M. Krimigis, K.I. Paularena, and J.D. Richardson
321
Mapping the detailed structure of the local interstellar medium S. Redfield and J.L. Linsky
325
Effect of different possible interstellar environments on the heliosphere: A numerical study H.-R. Mfiller, G.P. Zank, and P.C. Frisch
329
General Discussion
333
Session 6: View of Space Agencies and Companies Solar sail technology development and application to solar system exploration M. Leipold
337
*Contributions to the Outer Heliospheric Missions in frame of the German Space Program O. RShrig
345
*Deep space missions and technological challenges C. Schalinski and K. Eckardt (Astrium Consortium)
345
General Discussion
347
Session 7: At the E d g e of the Solar System The probable chemical nature of interstellar dust particles detected by CIDA on Stardust J. Kissel, F.R. Kriiger, J. Sil@n, and G. Haerendel
351
In-situ studies of interstellar dust from spacecraft I. Mann and H. Kimura
361
Dynamics of interstellar dust at the heliopause A. Czechowski and I. Mann
365
*Kuiper Belt Objects S.F. Green
369
*Dust in the Outer Heliosphere and Interstellar Dust E. Griin
369
*Comets as a source of heliospheric ions J. Geiss, G. Gloeckler, and K. Altwegg
370
*A Cometary and Interstellar Dust Analyzer for the STARDUST Mission J. Kissel, A. Glasmachers, H. von Hoerner, and H. Henkel
370
*Orbital Evolution of Dust in the Outer Heliosphere under the dust-gas drag force K. Scherer
371
General Discussion
373
Session 8: New Kinetic Aspects of Heliospheric Physics The injection problem J. Giacalone
377
A hydrokinetic description of solar wind electrons using hemispheric distribution functions I.V. Chashei, H.J. Fahr, and G. Lay
387
"--)iV --
Contents
*The solar wind electron velocity distribution O. Lie-Svendson
393
*General aspects of Boltzmann H-theorem for generalized collisions R.A. Treumann
393
*Heating and acceleration of ions by cyclotron- and Landau-resonances E. Marsch and C.-Y. Tu
394
General Discussion
395
S e s s i o n 9: M o d e r n H e l i o s p h e r i c Spacecraft and M i s s i o n s
In-space nuclear power as an enabling technology for exploration of the outer heliopause R. Sackheim, M. Van Dyke, M. Houts, D. Poston, R. Lipinski, J. Polk, and R. Frisbee
399
Interstellar probe using a solar sail: Conceptual design and technological challenges P.C. Liewer, R.A. Mewaldt, J.A. Ayon, C. Garner, S. Gavit, and R.A. Wallace
411
Propulsion for interstellar space exploration G. Genta
421
A realistic interstellar probe R.L. McNutt Jr, G.B. Andrews, J.V. McAdams, R.E. Gold, A.G. Santo, D.A. Ousler, K.J. Heeres, M.E. Fraeman, and B.D. Williams
431
Artificial intelligence techniques for the onboard analysis of space science data P.R. Gazis
435
Sunlensing the cosmic microwave background from 763 AU C. Maccone
439
*Propulsion options for the interstellar probe mission L. Johnson
445
*Solar Orbiter- A high resolution mission to the Sun and the inner heliosphere E. Marsch, B. Fleck, and R. Schwenn
445
General Discussion
447
S e s s i o n 10: M o d e r n H e l i o s p h e r i c T e c h n o l o g i e s
Scientific Payload for an interstellar probe mission R.A. Mewaldt and P.C. Liewer
451
Using multilayer mirrors to detect photons from the heliopause B.R. Sandel
465
A cosmic ray detector for an interstellar probe W. DrSge, B. Heber, M.S. Potgieter, G.P. Zank, and R.A. Mewaldt
471
*Future observations of the outer heliosphere M. Hilchenbach and H. Rosenbauer
475
*State-of-the-art solid state arrays and advanced analog microelectronics for heliospheric physics H.D. Voss
475
General Discussion
477
-- xvi --
Contents
Session 11: C o n n e c t i o n s to E a r t h
The heliosphere as viewed from Earth E.N. Parker
481
The heliosphere, cosmic rays, climate K. Scherer, H. Fichtner, and O. Stawicki
493
*Heliospheric changes in the past: Evidence from cosmogenic isotopes in the polar ice J. Beer
497
*The UV radiation climate on Earth and its impact on the biosphere: past, present and future trends G. Horneck
498
General Discussion
499
S e s s i o n 12: M i s c e l l a n e o u s
Aurora vortex structures as a result of disturbed geomagnetic conditions M.A. Danielides and A. Kozlovsky
503
A non-solar origin of the "SEP" component in lunar soils R.F. Wimmer-Schweingruber and P. Bochsler
507
*Disconnection events in P/Halley's plasma tail M.R. Voelzke and H.J. Fahr
511
Late paper:
Cosmic rays in the outer heliosphere and nearby interstellar medium J.R. Jokipii
513
List of participants
521
~176
--"
XVII
---
This Page Intentionally Left Blank
Session 1: Large Scale S t r u c t u r e
This Page Intentionally Left Blank
Physics of the Solar Wind Interaction with the Local Interstellar Medium G.P. Zank and H.-R. Mtfller Bartol Research Institute, University of Delaware, Newark, DE 19716, USA Plasmas in interstellar and interplanetary space are frequently partially ionized. Thus, the solar wind and stellar winds often interact with an interplanetary medium that is an admixture of protons, electrons, other charged ions, and neutral atoms. For example, the very local interstellar medium surrounding our heliosphere may be less than 50% ionized, with the dominant constituent being neutral hydrogen (H). As a result, the composition of the solar wind in the outer heliosphere beyond some 10 - 15 AU is dominated by neutral interstellar H. Our understanding of the complex physics describing the interaction of the solar wind with the partially ionized local interstellar medium has advanced significantly in the last 5 years with the development of very sophisticated models which treat the coupling of neutral atoms and plasma self-consistently. A number of major predictions have emerged from these models, such as the existence of a large wall of heated neutral hydrogen upstream of the heliosphere. Remarkably, in the ensuing years, this prediction has been confirmed by high resolution Hubble Space Telescope Lyman-~ spectroscopic data. An introductory review of the physics, and associated observations, of the interaction of the solar wind with the interstellar medium is presented for this exciting, rapidly developing field. 1. INTRODUCTION The solar wind forms a bubble, called the heliosphere, in the local interstellar medium (LISM), within which the solar system resides and whose size and properties are determined by the manner in which the solar wind plasma and the partially ionized LISM are coupled. In the last decade, great progress has been made in our understanding of the physical processes thought to describe the outer heliosphere. Fundamental to these advances has been the recognition that the interstellar medium is coupled intimately to the heliosphere itself and that much of outer heliospherie physics cannot be understood independently of the local interstellar medium. With the possibility that the aging spaeeerat~ Voyager 1,2 and Pioneer 10,11 might encounter the heliospherie boundaries in the not too distant future, interest in the far outer reaches of our solar system and the LISM has been rekindled. A convenient, if slightly vague, definition of the outer heliosphere, and one that is adopted here, is that it is that region of the solar wind influenced dynamically by physical processes associated with the LISM. Thus, for example, neutral interstellar hydrogen is the dominant (by mass) constituent of the solar wind beyond an ionization cavity of-~6 - 10 astronomical units (AU) in the upstream direction (the direction anti-parallel to the incident interstellar wind). The neutral hydrogen is coupled weakly to the solar wind plasma via resonant charge exchange - a coupling which leads to the production of pickup ions that come to dominate the internal energy of the solar wind. The solar wind changes then from a small plasma beta ]3sw (the ratio of plasma pressure to magnetic field pressure) environment to one in which [3sw > 3. Figure 1 shows the complex interaction between solar wind plasma and the LISM. While clearly simplistic, it serves to illustrate the primary physical processes. The interstellar plasma may or may not impinge supersonically on the heliosphere. The LISM flow is diverted about the heliospherie obstacle either adiabatically or by a bow shock (BS). The LISM plasma is coupled to the neutral interstellar hydrogen (H) primarily through charge exchange. Charge exchange corresponds to a neutral atom surrendering an electron to a passing ion (typically a proton), so creating a new charged particle and (a hydrogen) atom. Thus, when neutral H charge exchanges with LISM plasma that has been decelerated,
-3-
G.P. Zank and H.-R. Miiller heated and diverted about the heliosphere, the neutral H acquires a subpopulation that too is heated, decelerated and diverted. This neutral hydrogen, together with the LISM H, is called component 1 and is created in region 1 of Figure 1. If the charge exchange mean free path is sufficiently small in the region of decelerated LISM flow, a "wall" of neutral hydrogen will form in the upstream direction (the hydrogen wall). Resonant charge exchange is essentially a scattering rather than an extinction process. A boundary, called the heliopause (HP), separates the heliosphere and the LISM plasma, and is either a contact or a tangential discontinuity. The supersonic solar wind is decelerated, diverted and heated by a reverse shock called the termination shock (TS). The location of the TS is determined by the steady and temporal solar wind pressure exerted on the LISM.
H3-Xk'",, /
HAcR I
+ \
\ d ; . X ~ 2,. I
Figure 1. Schematic of the solar wind - LISM boundary regions which act as neutral H sources whose characteristics are identifiably distinct. The solar wind is enclosed by the termination shock (the curve labeled TS) and is identified as region 3. The shocked solar wind, the heliosheath, is bounded by the termination shock (TS) and the heliopause (HP) and is identified as region 2, while the interstellar medium is found beyond the heliopause. The LISM may or may not form a bow shock (BS) depending on whether the interstellar flow velocity is supersonic. Some sample plasma (H§ and pickup ion (H§ trajectories (dashed) are shown, as well as trajectories of neutral H (solid) which comes into the region from the LISM (Husu) and experiences charge exchange (solid arrow heads) outside the HP (component 1 neutrals, HI), between TS and HP (component 2, H2), and in region 3 (fast component 3, H3).
Neutral interstellar hydrogen that crosses the heliopause into the heliosphere can also experience charge exchange. Neutral H atoms that are created by charge exchange with the very hot subsonic plasma downstream of the heliospheric termination shock possess thermodynamic attributes that correspond to their origin, i.e., very hot (--10 6 K) with a large random velocity and an approximately outwardly directed velocity component. This population of neutral atoms, called component 2, is created in region 2 (Figure 1), and it streams away from the heliosphere and is quite distinguishable from component 1. By means of secondary charge exchange with the LISM plasma, component 2 acts to transport heat from the heliosheath (the region located between the heliospheric termination shock and the heliopause) into the LISM and this leads to an increase in the LISM temperature in the immediate vicinity of the heliopause. The final neutral component, component 3, possesses characteristics quite distinct from the other two components and has its origin in the supersonic solar wind (region 3, Figure 1). Component 3 has large outward radial velocities and relatively low temperatures but is dynamically unimportant owing to its very low energy density. The ions born of charge exchange in the supersonic solar wind, the pickup ions, are, however, of considerable dynamical importance since they remove both momentum and energy from the bulk solar wind flow. The solar wind is decelerated, so reducing the ram pressure and hence the expected size of the heliospheric cavity, and the solar wind acquires a tenuous non-thermalized population of suprathermal ions (typical energies-1 keV). The initial pickup ion population is unstable, generating low frequency magnetic turbulence which can scatter both pickup ions and cosmic rays. Some fraction of the pickup ions is further energized, possibly by shock acceleration, to form the anomalous cosmic ray component, although the precise injection mechanism awaits clarification.
-4-
Physics of the solar wind interaction with the local interstellar medium It should be noted that, historically, LISM conditions were very different in the past than they are presently (and indeed will be in the future) and thus the structure of the heliosphere may have been very different from the models discussed here. 1 Furthermore, since the interstellar medium is frequently partially ionized, we can expect that the interaction of winds of stars whose luminosity is relatively low (comparable to the sun, for example) with their LISM will be similar to that of the heliosphere. Thus, we may expect hydrogen walls to form ahead of the astrospheres of, for example, G-type stars. 2 The material presented here has been reviewed extensively by Holzer and Zank 3 and the reader is referred to these works for more detail and an extended bibliography. 2. THE LOCAL I N T E R S T E L L A R MEDIUM Many of the pertinent physical parameters of the LISM are poorly constrained and so, by implication, the detailed solar w i n d - LISM interaction is not yet well understood. Our best inferences about the LISM flow velocity and temperature have been made from observations of neutral interstellar hydrogen (H) and helium (He) within the heliosphere, using either resonantly scattered solar UV light techniques 4 (since H atoms in the heliosphere are illuminated by the Sun in the HI resonance line i.e., the Lyman-~ line) or in situ spacecraft measurements of the He distribution: The bulk velocity of the LISM flow is-~26 km/s and the plasma temperature is-~8000 K. The H and H + number densities are not well determined, although approximate limits can be derived. For HI (i.e., neutral H only), number densities lie in the range of --4).15 - 0.34 cm -3, while for HII (ionized H), the range is -4).03 - 0.14 cm3, depending on whether the EUV radiation field is included or not. 6 This leads to a limit on the total number density of both neutral and ionized interstellar hydrogen in the range --0.15 - 0 . 3 4 cm 3. Although most models assume that the relative motion of the heliosphere through the interstellar medium is supersonic, it is by no means clear that such an assumption is completely warranted. Our knowledge of both the local interstellar magnetic field strength and orientation and the energy density in interstellar cosmic rays is somewhat rudimentary. Cosmic rays with energies of 300 MeV/nucleon and less may contribute -~ (3 + 2) x 10 -13 dynes cm -2, which can yield an interstellar flow that is either sub-magnetosonic or subsonic. 7 The magnetic field orientation appears to be perpendicular to the interstellar flow velocity vector and estimates for the magnetic field strength range from---1.3 to 2 ktG.S Evidently, our knowledge about important LISM plasma and neutral H parameters is rather poor and this must therefore be taken into account when evaluating models of the global heliosphere. Nonetheless, the basic underlying physical processes appear to be reasonably well understood. 3. GLOBAL H E L I O S P H E R I C STRUCTURE The dynamical or ram pressure (given by p U 2 , 13 the fluid mass density and u the flow speed) and thermal pressure p of the solar wind decrease with increasing heliocentric distance and must reach a value which eventually balances the pressure exerted by the LISM. The relaxation towards pressure equilibrium between the solar and interstellar plasmas is characterized by (i) a transition of the supersonic solar wind flow to a subsonic state, and (ii) a divergence of the interstellar flow about the heliospheric obstacle. The transition of the supersonic solar wind is most likely accomplished by means of a termination shock (TS), and it is anticipated that at least the Voyager spacecraft will encounter this boundary in the early 21 st century. The divergence of the LISM flow about the heliosphere may be accomplished either adiabatically (if the relative motion between the sun and the LISM is subsonic) or by means of a bow shock in the case of supersonic relative motion. If we neglect both the deceleration of the solar wind by resonant charge exchange and temporal variations in the
-5-
G.P. Zank and H..R. Mfiller
solar wind ram pressure, the minimum radius Rrs of the solar wind shock transition can be estimated from 9 Rrs=
~/+3
P0u02
(1)
R0 2(~/+ 1) Po0 ' 2 where PoUo is the solar wind ram pressure at R o = 1AU, y the solar wind adiabatic index, and p~ the total LISM pressure. The LISM pressure term can include the thermal gas, cosmic rays, the interstellar ram pressure, the magnetic field pressure, MHD turbulence, and may be expressed as Poo = pu 2 + Pth + 1"1 + PcR + Paust + P~B2 , where the terms are all evaluated in the LISM and the factor r I is an attempt to include the effects of interstellar magnetic field obliquity. 3' 10.18 Although one can estimate the location of the TS and the heliopause (HP) using simple pressure balance arguments, the interaction of the solar wind with the LISM is fundamentally multidimensional. Thus, the main advances in our understanding of global heliospheric structure since the original analytic models have been more recent and based largely on computer simulations. The initial simulations were of one-fluid gas dynamic models and only now has the inclusion of neutral interstellar hydrogen been considered self-consistently, with the prediction of a hydrogen wall. 11 To model the interaction of the solar wind with a partially ionized LISM, the usual set of 3D MHD equations must be solved. The collisional processes coupling plasma and neutral atoms depend on their relative energies and their relative importance can vary from region to region of a complex plasma system such as the solar wind/LISM interface. The Knudsen number Kn = ~,/L (~, the mean free path of neutral atoms and L a characteristic macroscopic length scale, such as the size of the solar heliosphere,-100 AU), which is a measure of the neutral distribution relaxation distance, is >> 1 inside the heliosphere and-1 in the very local ISM. The introduction of neutral atoms into the magnetized heliospheric plasma with a large Knudsen number implies that the neutral and plasma distributions cannot equilibrate and may possess quite distinct bulk flow speeds and temperatures. Charge exchange between the coupled, non-equilibrated neutral and charged particle populations can therefore introduce distinct new populations of neutral atoms and plasma whose characteristics reflect their parent population. The subsequent interaction and assimilation of the newly created plasma and neutral populations into the existing plasma and neutral distributions may then lead to the substantial modification of the overall partially ionized plasma system. Thus, the total neutral distribution cannot relax to a single Maxwellian distribution, and one must therefore use a multi-component transport or kinetic description for the neutral populations. To work directly with the kinetic description of interstellar and heliospheric neutral hydrogen, one must solve the Boltzmann equation 12 -~t + v . V f +
.V v f=
P-L,
(2)
for neutral H directly. Here, jr(x, v, t) is the neutral hydrogen distribution function expressed in terms of position x, v, and time t. F is the force acting on a particle of mass m, typically gravity and radiation pressure. The terms P and L describe the production and loss of neutral particles at (x, v, t) and both terms are functions of the plasma and neutral distributions. An alternative approach, and one that is much less prohibitive computationally, is to recognize that the heliosphere-LISM environment comprises three thermodynamically distinct regions; the supersonic solar wind (region 3), the very hot subsonic solar wind (region 2), and the LISM itself (region 1 ) Figure 1. Each region acts as a source of neutral H atoms whose distribution reflects that of the plasma
-6-
Physics of the solar wind interaction with the local interstellar medium
distribution in the region. One may identify neutral components 1, 2, and 3 with neutral atoms originating from regions 1, 2, and 3. Each of these three neutral components is approximated by a distinct Maxwellian distribution function appropriate to the characteristics of the source distribution. This observation allows the use of simpler production and loss terms for each neutral component. The complete highly non-Maxwellian H distribution function is then the sum over the three components, i.e., 3
f ( x , v , t ) = ~--'f ( x , v , t ) . i=1
Under the assumption that each of the neutral component distributions is approximated adequately by a Maxwellian, one obtains an isotropic hydrodynamic description for each neutral component, and thus a closed multi-fluid transport model describing selfconsistently the interaction of the solar wind with the partially ionized LISM. ~3
Figure 2. (a) A 2D plot of the global plasma structure for the two-shock model when neutral H is included self-consistently. The contouring refers to Logl0[temperature]. (b) A 2D plot of the corresponding component 1 neutral distribution, where the shading refers to the number density. The hydrogen wall is clearly visible between the bow shock and the heliopause.
Models of the heliosphere that do not include neither interstellar neutrals, interplanetary nor interstellar magnetic fields TM show the basic structure of terminations shock and heliopause. The TS possesses a characteristic bullet-shape, quite unlike the approximately spherically symmetric heliosphere that is assumed typically. The supersonic solar wind velocity is constant until the very strong TS (Mach number-~170). The solar wind temperature decays adiabatically with increasing heliocentric distance. An extended and very hot tail of subsonic solar wind is formed (called the heliotail) in the downstream direction. When neutral H is included, a major effect of charge exchange on the heliospheric interfaces is to decrease the distances to the TS, HP and BS. 15The decrease in distance results primarily from the reduction in solar wind ram pressure, this due to the mediation of the wind by charge exchange. Besides the distance to the various heliospheric boundaries, illustrated in Figure 2a, charge exchange affects the shape of the termination shock, making it more spherical compared to the purely ionized gas dynamic description. This is a consequence of charge exchange in the heliosheath ensuring that the sheath plasma remains subsonic in this region.
Owing to the deposit of interstellar protons in the solar wind when charge exchange is included, the solar wind flow now departs slightly from simple spherical symmetry, and the contribution to the internal energy of the supersonic solar wind by pickup ions is substantial. The tail region is also cooler when the proton fluid and neutral fluid are coupled compared to no-charge exchange models. The very hot heliotail is cooled by charge exchange with cooler component 1 neutrals. To enter the heliosphere, a hydrogen atom has to cross three thermodynamically distinct regions. Through resonant charge exchange, the diverging LISM flow and the shocked solar wind flow act to
-7-
G.P. Zank and H.-R. Mfdler
divert some fraction of the incident interstellar neutral H flux away from the heliosphere. Figure 2b, a 2D plot of the component 1 neutral H density distribution, shows that inflowing component 1 neutrals are decelerated substantially and filtered by charge exchange with the interstellar plasma between the BS and HP in the upstream direction. This leads to the formation of a hydrogen wall with maximum densities-~0.2-0.3 cm -3, column densities .--10TM cm 2, and temperatures ranging from 18 000K to 28 000K (depending on the assumed LISM parameters). The pile-up in the neutral gas results from the deceleration and deflection of the neutral flow by charge exchange with the interstellar plasma, which is itself decelerated and diverted due to the presence of the heliosphere. Note that the charge exchange mean-flee-path is typically less than the separation distance between the HP and BS and so a large part of the incident interstellar neutrals experience charge exchange. Component 2, produced via charge exchange between component 1 and hot shocked solar wind plasma between the TS and HP, streams across the HP into the cooler shocked interstellar gas and heats the plasma through a second charge exchange. This leads to an extended thermal foot abutting the outside edge of the HP. This heating of the plasma by component 2 serves to broaden the region between the BS and HP, as well as to (indirectly) further heat the component 1 interstellar neutrals after subsequent charge exchanges. Some heating of the unshocked LISM also occurs upstream of the BS, thereby marginally reducing the Mach number of the incident interstellar wind. The temperature of component 1 neutrals once inside the heliosphere remains fairly constant in the upstream region, and represents a substantial increase over the assumed LISM temperature. A further increase in the component 1 temperature occurs in the downstream region.
Figure 3a. An example of a kinetic simulation described by Mailer et al. 16for a 2-shock model. One-dimensional profiles of plasma density np and plasma temperature Tp are shown as dashed lines over radial distance R from the sun, and neutral density nH and (averaged) temperature THas solid lines. The profiles are obtained in the upstream direction, antiparallel to the LISM flow. The middle row depicts two-dimensional neutral velocity distribution functions (logarithmic density scale) at various locations on that axis, with prominent HLISM and H~ 26km/s neutrals, and evidence of 100-300 km/s H2 and 400 km/s H3 neutrals.
Figure 3b: Normalized H I radial velocity distribution at a point in the hydrogen wall located 160 AU from the Sun in the upwind direction of the interstellar flow into the heliosphere. The jagged line is the distribution given by the Boltzmann code, and the smooth curve is the distribution resolved by the multifluid code through three-component MaxwellianslS. The total distribution is clearly non-Maxwellian.
The number density of component 1 crossing the TS is approximately half the assumed incident LISM number density ("filtration"). The density varies only weakly between the TS a n d - 1 0 AU from the Sun in the upstream region. In the downstream direction, component 1 densities are lower within the heliosphere. The 1D component 1 density and temperature profiles are plotted in Figure 3a, and the heated hydrogen wall is exhibited clearly. A further effect of filtration is to decelerate the upstream neutral gas from 26 krn/s in the LISM to -~19 km/s at the TS in the region of the nose. Deflection of the
-8-
Physics of the solar wind interaction with the local interstellar medium
flow also reduces the radial velocity component at angles away from the nose (a deceleration which is in accord with observations of the Lycx resonance line scattered by neutral H in the heliosphere). Also plotted in Figure 3 (taken from a kinetic simulation ~6) is the evolution of the neutral H distribution from the interstellar medium into the heliosphere. Components 1 and 2 are not always easily distinguished in 1D omnidirectional plots (Figure 3b) but the departure from a Maxwellian distribution is pronounced. The smoothed curve shows the distribution predicted by a multi-fluid model, which assumes a superposition of Maxwellian distributions corresponding to components 1, 2, and 3. While clearly distinct in detail, the basic aspects of the neutral distribution predicted by kinetic and multifluid models are surprisingly similar. Models which assume a subsonic LISM flow do not have a bow shock. Instead, upstream of the heliopause, the LISM flow is compressed adiabatically. This more gradual compression leads to the formation of a lower amplitude hydrogen wall that is more extended in the radial direction. It is also less extended in the transverse direction because of the localized nature of the adiabatic compression. The maximum density of the wall in the upstream direction is smaller than that of the two-shock counterpart (though still larger than the incident LISM number density). However, because it is wider, the column density is comparable with the two-shock case. The heliosphere is smaller due to the higher assumed LISM pressure. Three dimensional models of the solar wind during solar minimum introduce some interesting differences from the 2D models discussed above. The presence of an anisotropic solar wind with a high velocity, low density, high temperature polar steam acts to divert both the subsonic solar wind and the LISM plasma flow into the ecliptic region of the heliosphere. ~7The presence of a high density band of plasma about the ecliptic plane increases the filtration of neutrals compared to the higher polar regions. With an increased neutral flux of H over the poles, the polar solar wind experiences comparatively greater deceleration than the ecliptic wind, so reducing the degree of anisotropy seen in 3D gas dynamic simulations. In addition, the global distribution of neutral H is anisotropic in heliolatitude. Structurally, the heliosphere at low latitudes can revert to a "bullet" shape, typically seen only in models which neglect to include interstellar neutral hydrogen. It should be borne in mind that these remarks pertain primarily to the period of solar minimum observed by the Ulysses spacecraft. Very few simulations include either the interplanetary or interstellar magnetic fields dynamically and those that do neglect neutral H completely or use an extremely simplified description. TM Nonetheless, it appears that the magnetic field in the heliosheath can acquire a very interesting structure. Along the stagnation line in the heliosheath, the flow velocity decreases approximately as r -2 (r the heliocentric radial distance), leading to an amplification in the azimuthal magnetic field in the region. This, together with the J x B force causes the equatorial current sheet to bend either upward or downward (depending on solar cycle) away from the equatorial plane. This leads to the formation of an asynmaetric 3D magnetic shell or wall in the outer region of the upstream heliosheath with a gap in the neighborhood of the displaced current sheet. The region between the termination shock and the magnetic wall is occupied by solar wind originating from the middle and high latitudes whereas, thanks to the gap in the magnetic wall, the solar wind in the region between the magnetic wall and heliopause originates from the low latitude equatorial neutral sheet region. The ram pressure pu2of the solar wind varies by a factor o f - 2 with solar cycle. Since the location of the termination shock and heliopause is determined by a balance of the ram pressure and the interstellar pressure, temporal variation in the solar wind ram pressure must therefore play an important role in determining the global structure of the heliosphere. The time dependent multi-fluid model 15has been used to extend earlier studies of the effect of solar cycle variability on the global structure of the heliosphere. 19The varying ram pressure of the solar wind leads to large arrhythmic
-9-
G.P. Zank and H.-R. Miiller
excursions of the TS and smaller excursions of the HP. The motion of the TS drives pressure waves into the inner heliosheath which steepen into shocks which propagate into the LISM. For a two-shock model, the dynamical train of shocks propagating into the LISM acts to increase the separation distance between the HP and BS, whereas in the one-shock case, the emitted shocks act as a surrogate bow shock(s) for the heliosphere. In both cases, the emitted shocks in the LISM introduce some variability into the neutral H entering the heliosphere on long time scales. Finally, it has been found that ion-neutral friction can destabilize the HP. 19 4. HELIOSPHERIC STRUCTURE INFERRED FROM L Y - ~ MEASUREMENTS
Until Voyager 2 encounters the termination shock and heliopause, we are forced to rely on remote observations to infer the structure of the heliosphere. The most promising approach presently is to use the Lytz absorption line in the direction of nearby stars) ~ A neutral hydrogen pileup or wall may be detectable in directions where the decelerated H is red-shifted out of the shadow of the interstellar absorption if the interstellar column density is sufficiently low. Red-shifted excess absorption in Hubble Space Telescope GHRS Lyo~ observations towards o~-Cen (seen previously by Copernicus and IUE) have been interpreted as evidence for the existence of the solar hydrogen wall) 1 Shown in Figure 4 are observations towards o~ Cen A (the jagged solid curves), and the wavelength scale is relative to the Lyman-o~ line center in the heliocentric rest frame. The upper solid curve is the assumed intrinsic stellar Lyo~ emission profile. The accurate representation of the intrinsic stellar profile is unimportant since the absorption features of interest vary sharply. The remaining smooth solid curve (the saturated absorption curve) of Figure 4 gives the attenuation of the stellar Lyo: emission by interstellar H. The column density for H was fixed by scaling to the Deuterium column density, assuming the commonly accepted value of D/H = 1.6 x 10 -5 for the local interstellar medium) 2 Figure 5 shows very clearly that additional absorption is required both redward and blueward of the interstellar feature if the fit is to be completed. Furthermore, the fit must be applied preferentially to the redward side, so arbitrarily changing the D/H ratio is unacceptable. The additional redward absorption has been interpreted as evidence for the detection of the hydrogen wall. The blueward absorption suggests the possibility of a hydrogen wall about o~ Cen A and B. Figure 4. The solid jagged curves are GHRS Lyman-O~absorption profiles Since the column depth of the hydrogen wall is three orders of towards O~ Cen A. The uppermost magnitude smaller than the column depth in the LISM toward t~ solid curve is the assumed intrinsic Cen, it may be surprising at first glance that the heliospheric optical stellar Lyman-~ emission profile. depth at +0.1 A is of the same order as the LISM optical depth at The smoothed solid line, which exthat wavelength. The key difference is that the hydrogen wall is hibits saturated absorption, corresponds to the intrinsic stellar emisheated and decelerated, which both broadens and redshifts the sion line after absorption from the heliospheric component away from the -0.07 A centroid of the LISM only. Corresponding theoreLISM absorption and toward the +0.1 A wavelength of interest. tical absorption profiles are plotted22 Figure 4 also shows the Lyo~ absorption at the red edge of the LISM feature computed from models for three assumed LISM Mach numbers (supersonic, subsonic and ~sonic) 22. These synthetic data are compared with the observed heliospheric absorption. The primary and very important conclusion to emerge from such modeling is that the synthetic Lyo~ profiles support the interpretation of the observed additional redward Lyo~ absorption towards o~-Cen as evidence for the remote detection of the hydrogen wall. Thus, it appears that the hydrogen wall has indeed been observed! The unprecedented accuracy of the Hubble GHRS observations suggests further that, with the refinement of solar wind - LISM interaction models, the question of whether the heliosphere has a one- or a two-
-10-
Physics of the solar wind interaction with the local interstellar medium
shock structure may be answered in the foreseeable future by a combination of modeling and observations as discussed further elsewhere 23' 24. 5. CONCLUDING R E M A R K S Evidently, many more physical processes remain to be incorporated in models describing the interaction of the solar wind with the LISM. This will undoubtedly be accomplished in the next decade as computational power increases and the underlying physics is better understood. However, the most important limitation that the field currently faces concerns the paucity of information about the state of the local interstellar environment. Until LISM parameters are better constrained, all modeling efforts describing the solar w i n d - LISM interaction must be viewed as somewhat hypothetical. Nonetheless, with new observational techniques emerging, new missions planned, and the increasing sophistication of theoretical modeling efforts, the outlook for understanding the physics of the outer heliosphere is bright indeed. Furthermore, with the detection and modeling of hydrogen walls about nearby low luminosity stars 24' 25, n e w opportunities for investigating the physics of hitherto undetectable stellar winds and their LISM now exist. Acknowledgements. This work is supported in part by an NSF-DOE grant ATM-0078650, a NASA grant NAGS-6469, a Jet Propulsion Laboratory contract 959167, and a NASA Space College Grant award. REFERENCES
1 G.P. Zank and P.C. Frisch, Astrophys. J. 518, 965 (1999); H.-R. Mtiller, G.P. Zank, and P.C. Frisch, "Effect of different possible interstellar environments on the heliosphere: A numerical study", submitted to the proceeding of the Cospar colloquium "Outer Heliosphere", Potsdam, 2000, edited by K. Scherer (Elsevier, Amsterdam, 2001). 2 P.C. Frisch, Astrophys. J. 407, 198 (1993). 3 T.E. Holzer, in Solar System Plasma Physics, edited by E.N. Parker, C.F. Kennel, and L.J. Lanzerotti (NorthHolland, New York, 1979), p. 101; G.P. Zank, Space Sci. Rev. 89 3-4, 413 (1999). 4 J.-L. Bertaux, R. Lallement, V.G. Kurt, and E. Mironova, Astron. Astrophys. 150, 1 (1985); E. Qu6merais, B.R. Sandel, R. Lallement, and J.-L. Bertaux, Astron. Astrophys. 299, 249 (1995); R. Lallement, Space Sci. Rev. 78, 301 (1996). 5 M. Witte, M. Banaszkiewicz, and H. Rosenbauer, Space Sci. Rev. 78, 289 (1996). 6 j. Vallerga, Space Sci. Rev. 78, 277 (1996); G. Gloeckler, ibid, 335. 7 W._H" Ip and W.I. Axford, Astron. Astrophys. 149, 7 (1985). s j. Tinbergen, Astron. Astrophys. 105, 53 (1982); P.C. Frisch, Science 265, 1423 (1994); Space Sci. Rev. 72, 499 (1995). 9 L.E. Davis, Phys. Rev. 100, 1440 (1955); E.N. Parker, Interplanetary Dynamical Processes (Interscience, New York, 1963). ~0R. Ratldewicz, A. Barnes, and J.R. Spreiter, S.S. Stahara, M. Vinokur, and S. Venkateswaran, Astron. Astrophys. 335, 363 (1998); R. Ratldewicz, A. Barnes, and J.R. Spreiter, "Local interstellar medium and modeling the heliosphere", submitted to J. Geophys. Res. ~ V.B. Baranov, M.G. Lebedev, and Y.G. Malama, Astrophys. J., 375, 347 (1991). 12V.B. Baranov and Y.G. Malama, J. Geophys. Res. 98, 15157 (1993); V.B. Baranov and Y.G. Malama, J. Geophys. Res. 100, 14755 (1995); H.-R. Mtiller, G.P. Zank, and A.S. Lipatov, Proceedings, Solar Wind 9, Nantucket, 1998, edited by S.R. Habal et al. (AIP Conf. Proceed., 1999), Vol. 471, p. 819; J. Geophys. Res. 105, 27,419 (2000); V.V. Izmodenov, J. Geiss, R. Lallement, G. Gloeckler, V.B. Baranov, and Y.G. Malama, J. Geophys. Res. 104, 4731 (1999). 13H.L. Pauls, G.P. Zank, L.L. Williams, and D.T. Hall, Proceedings, Solar Wind 8, edited by D. Winterhalter et al. (AIP Conf. Proceed., New York, 1996), Vol. 382, p. 605; Zank et al.15; L.L. Williams, D.T. Hall, H.L. Pauls, and G.P.Zank, Astrophys. J. 476, 366 (1997); Liewer (1997). 14T. Matsuda, Y. Fujimoto, E. Shima, and T. Inaguchi, Prog. Theor. Phys. 81, 810 (1989); V.B. Baranov and Y.G. Malama, J. Geophys. Res. 98, 15157 (1993); R.S. Steinolfson, J. Geophys. Res. 99, 13307 (1994); R.S.
-11
-
G.P. Zank and H.-R. Miiller
Steinolfson, V.J. Pizzo, and T. Holzer, Geophys. Res. Lett. 21,245 (1994); S.R. Karmesin. P.C. Liewer, and J.U. Brackbill, Geophys. Res. Lett 22, 1153 (1995); H.L. Pauls, G.P. Zank, and L.L. Williams, J. Geophys. Res. 100, 21595 (1995); C. Wang and J.W. Belcher, J. Geophys. Res. 103, 247 (1998). 15G.P. Zank, H.L. Pauls, L.L. Williams, and D.T. Hall, J. Geophys. Res. 101, 21639 (1996). 16H.-R. MUller, G.P. Zank, and A.S. Lipatov, J. Geophys. Res. 105, 27,419 (2000). 17H.L. Pauls and G.P. Zank, J. Geophys. Res. 101, 17081 (1996); H.L. Pauls and G.P. Zank, J. Geophys. Res. 102, 19779 (1997);A. Barnes, J. Geophys. Res. 103, 2015 (1998); T. Tanaka and H. Washimi, J. Geophys. Res. 104, 12,605 (1999). 18H. Washimi and T. Tanaka, Space Sci. Rev. 78, 85 (1996); Ratkiewicz et al.l~ T. Linde, T.I. Gombosi, P.L. Roe, K.G. Powell, D.L. DeZeeuw, J. Geophys. Res. 103, 1889 (1998); N.V. Pogorelov and T. Matsuda, ibid, 237. 19G.P. Zank, Proceedings, Solar Wind 9, Nantucket, 1998, edited by S.R. Habal et al. (AIP Conf. Proceed., 1999), Vol. 471, p. 783. 20B.E. Wood, H.-R. Mtiller, and G.P. Zank, Astrophys. J. 542, 493 (2000). 21 J.L. Linsky and B.E. Wood, Astrophys. J. 463, 254 (1996); Gayley et al.22 22K. Gayley, G.P. Zank, H.L. Pauls, P.C. Frisch, and D.E.Welty, Ap.J. 487, 259 (1997). 23 Gayley et a1.22;B.E. Wood and J.L. Linsky, Astrophys. J, 492, 788 (1998); B.E. Wood, J.L. Linsky, and G.P. Zank, Astrophys. J. 537, 304 (2000). 24B.E. Wood, J.L. Linsky, H.-R. Miiller, and G.P. Zank, "Observational estimates for the mass-loss rates of ct Cen and Proxima Cen using HST Lyman-ct spectra", Ap J. Lett., 547, 49 (2001). 25H.-R. Miiller, G.P. Zank, and B.E. Wood, Ap.J. 551,495 (2001).
-12-
Relating Models of the Heliosphere to Lyman-c~ Absorption Observed in Hubble Spectra H.-R. Mfiller ~ and B. E. W o o d b* ~Bartol Research Institute, University of Delaware, Newark, DE 19716 bjILA, University of Colorado and NIST, Boulder, CO 80309-0440 We model the interaction of the solar wind with the partially ionized local ISM using a self-consistent hybrid model in which the Boltzmann equation for neutral hydrogen is solved with a kinetic particle code. The degree of external contribution to the ISM plasma pressure (e.g. due to cosmic rays or to magnetic pressure) is varied as a model parameter, resulting in a family of heliospheric models ranging from two-shock models to one-shock heliospheres (subsonic ISM). We give an overview of the neutral hydrogen distributions in these models. The column density and temperature of the heliospheric neutral hydrogen have been observed to be large enough to produce detectable absorption signatures in Lyc~ spectra of nearby stars. The heliospheric models can be used to predict the amount of absorption for various lines of sight, and we compare these predictions with Lyc~ observations of six nearby stars obtained by the Hubble Space Telescope, sampling lines of sight ranging from nearly upwind to nearly downwind. We find that the Boltzmann models tend to predict too much absorption in sidewind and downwind directions, especially when we assume high Mach numbers for the interstellar wind. 1. I N T R O D U C T I O N The basic heliospheric structure created by the interaction between the fully ionized solar wind and the partially ionized local interstellar medium (LISM) can to first order be modeled by considering plasma interactions alone [1-3]. However, charge exchange, whereby an interstellar neutral loses its electron to a proton, allows the neutrals to take part in the solar-wind/LISM interaction in important ways (see review by Zank [4]). Heliospheric models that treat the neutral gas and plasma in a self-consistent manner predict somewhat different properties than plasma-only models, such as shorter distances to the termination shock, heliopause, and bow shock [5-9]. These models also suggest that neutral hydrogen in the heliosphere should be very hot, with temperatures of order 20,00040,000 K. The importance of this prediction is that this high temperature gas should *This research was supported by NASA grant NAG5-9041 to the University of Colorado, by an NSF travel grant administered by the American Astronomical Society, and by NASA grant NAG5-6469, NSF-DOE award ATM-0078650,JPL contract 959167, and NASA Delaware Space Grant College award NGT5-40024 to the Bartol Research Institute.
-13-
H.-R. Miiller and B.E. Wood
produce H I Lya absorption broad enough to be separable from the interstellar absorption observed toward nearby stars, meaning hydrogen in the outer heliosphere can potentially be directly detectable. This absorption has in fact been observed using observations by the Hubble Space Telescope (HST). Excess Lya absorption is evident in high-resolution HST spectra of the nearby stars a Cen A and B, and this absorption has been demonstrated to be consistent with a heliospheric origin [10,11]. Models that self-consistently treat the neutrals and plasma are essential to help interpret these observations, but such models are a complex theoretical and computational problem. The fundamental difficulty is that neutrals in the heliosphere are far from equilibrium, and their velocity distribution at a given location can be highly non-Maxwellian. Lipatov et al. [12] adopted a particle-mesh method for solving the neutral Boltzmann equation, which should yield accurate particle distribution functions (see also [13,14]). This creates the opportunity to use heliospheric models created by this Boltzmann code to predict heliospheric Lya absorption profiles that can be compared with the HST observations. In this paper, we create models assuming different Mach numbers for the inflowing LISM to see which best matches the observed amount of absorption. For the a Cen line of sight, a similar analysis has previously been carried out using four-fluid rather than Boltzmann models [11], and assuming different LISM parameters. We compare the absorption predicted by the models with lines of sight observed by HST, including a Cen, 36 Oph, and Sirius lines of sight, which all show evidence for heliospheric absorption [15,16]. 2. M O D E L
DESCRIPTION
AND
RESULTS
We compare the HST observations of absorption of stellar Lya to absorption profiles synthesized from models of neutral hydrogen (H) in the global heliosphere. We use the fully time-dependent hybrid code described by [14] that includes neutral H and its charge exchange interaction with plasma in a self-consistent way. It consists of a 2D gas-dynamical description of the plasma of both solar and interstellar origin coupled to a 2.5D particle description of neutral hydrogen. The plasma code uses the Euler equations of fluid dynamics to solve for the bulk plasma variables of number density np, temperature Tp, and velocity u. This treatment is justified because the mean free path of the protons of the plasma is small in comparison to heliospheric length scales, and the proton distribution function is Maxwellian. The Boltzmann code for neutral H does not restrict the neutral distribution function to a Maxwellian, but admits arbitrary distributions. Transporting and influencing the particles in phase space, the particle code indirectly solves the Boltzmann equation for the time evolution of the neutral hydrogen distribution function. The Boltzmann code thus treats even those neutral components correctly whose mean free path is no longer smaller than typical heliospheric scales (component 2 neutrals and splash component, see below). For the sake of comparing the spatial distributions of the neutral and the ionized species, it is customary to take moments of the neutral distribution functions to arrive at the neutral H number density nil, effective temperature TH, and bulk neutral velocity u s . The effective neutral temperature is misleading in the sense that one often associates it with a single Maxwellian, which is an oversimplification of the intricate shapes of the neutral distribution functions that are encountered in the heliosphere.
-14-
Relating models o f the heliosphere to Lyman-a absorption...
To first order, the dominant interaction between neutral H and plasma is resonant charge exchange. It is incorporated in each piece of the hybrid code through source terms. The source terms of the Euler gasdynamic equations make use of n i l , TH, and UH. Within the particle code, charge exchange is accounted for by loss and production terms in the neutral Boltzmann equation. These terms invoke the bulk plasma parameters as well as an approximation for the Maxwellian plasma distribution function. For the charge exchange cross section, we use an empirical fit [17]. Charge exchange deletes one neutral particle from the neutral distribution, and adds a new neutral that bears the characteristics of the plasma at the site where the exchange took place. The deleted neutral is most likely from the dominant neutral category, i.e., a "component 1 neutral" originally from the LISM neutral population, and its deletion therefore does not significantly change the neutral distribution. The newly born neutral can have very different characteristics, in particular when exchange takes place in the heliosheath or in the heliotail ("region 2") where the plasma is hot and has a low bulk speed, due to the shock heating and deceleration of solar wind plasma at the termination shock. These so-called component 2 neutrals have thermal speeds on the order of 100 km/s, corresponding to 106 K, and travel in essentially random directions. Some of them have trajectories directed towards the inner solar system where they can be detected as energetic neutral (hydrogen) atoms (ENA) in the low energy range below 1 keV [14]. So-called splash neutrals form another distinct component of neutrals, born inside the termination shock ("region 3") through charge exchange with the solar wind plasma that is still supersonic and cold, except close to the Sun. Charge exchange inside the termination shock heats the plasma through the generation of pickup ions, whereas the plasma in the heliosheath between ter~nination shock and heliopause (also in the heliotail) is cooled by charge exchange. In the heliotail, this mechanism works to equilibrate neutral H with the plasma, thereby heating and accelerating neutral H. Some of the hot component 2 neutrals created in the heliosheath deposit their energy upstream of the heliopause through secondary charge exchange, creating an anomalous heat transfer into the oncoming LISM. The resulting plasma temperature gradient in the LISM upstream of the heliopause is communicated to neutral H through charge exchange, creating a neutral temperature gradient as well. This gradient is even more pronounced when the incoming LISM plasma is supersonic. In this case, a bow shock exists where the plasma becomes subsonic, and the associated shock heating of the plasma steepens the plasma temperature gradient and consequently further heats the neutral H between the heliopause and bow shock. The deceleration of neutral H through charge exchange with the decelerated (stagnating) plasma creates what has been termed a hydrogen wall, a density enhanced region of neutral H upstream and sidestream of the heliopause (and bounded on the other side by the bow shock in two-shock models). The filtration of neutral H by the hydrogen wall creates a depletion of neutral H in the heliosphere. For the LISM boundary conditions we choose TH -- Tp - 8000 K, nH -- 0.14 cm -3, np -- 0.10 cm -3, and v - 26 km/s, which are consistent with observations [18-22]. The inner boundary is an inflow of solar wind plasma at 1 AU with T - 105 K, np - 5.0 cm -3, and v - 26 km/s. 7 is set to 5/3 throughout. Figure 1 serves as an illustration of the spatial distribution of neutral density and temperature of a typical model. The Sun is at the origin and the LISM enters from the right. The stagnation axis, which
-15-
H.-R. Miiller and B.E. Wood
cb)t..
+~ -
-
-
b.16,o ~ii.3~I:~ ,~,.c~-o~
9
-,S.; Oo I))} --
~
~..
~0
-,so-j,.,,_.. -soo-
J
."
9 9 0
+
"o
9
Figure 1. Contour plots of (a) neutral H density and (b) effective neutral temperature (for model 5 of Table 1). The Sun is at the origin; the neutral flow field is illustrated by the streamlines in (b). The five density contour levels (in cm -3) and the temperature levels 20,000, 50,000, and 115,000 K are marked in the plots. The small lower panels show 1D profiles along the stagnation axis. Both value axes have a logarithmic scale.
runs parallel to the LISM flow and through the Sun, is a symmetry axis. One can clearly see the density enhancement above the interstellar value of 0.14 cm -3. The 0.16 cm -3 contour approximately follows the hydrogen wall, whose sunward side coincides with the heliopause. The region of maximum density is enclosed by the 0.20 cm -3 contour. The tailward termination shock intersects the stagnation axis at ~100 AU. Downwind of that we see a density depletion and a temperature increase. The neutral temperature also experiences an increase at the wall and in sidestream regions where the plasma has also an increased temperature. Whether or not a bow shock exists upstream of the heliosphere is currently an open question. The LISM plasma temperature and velocity alone would point to the existence of a bow shock (a "two-shock heliosphere") because the Mach number is M - 1.7. However, there is the possibility that mechanisms such as the pressure due to galactic cosmic rays or the pressure of an interstellar magnetic field contribute to the plasma energy, lowering the Mach number of LISM protons to subsonic values and resulting in a one-shock heliosphere. We model such a pressure contribution using a generic factor a to indicate that only a fraction of the total pressure outside of the heliopause is due to LISM protons. The equation of state then reads P - a n v k B T p , where P is the total pressure. The minimal a is 2, expressing the usual assumption that the LISM electrons have the same density and temperature (and therefore the same contribution to the total pressure) as the proton plasma.
-16-
Relating models of the heliosphere to Lyman-a absorption... Table 1 Model parameters. Model # 1 c~ 2.0 Mach# 1.75 Tpl[K] 8,000
2 3.5 1.32 14,000
3 5.0 1.11 20,000
4 7.5 0.90 30,000
5 9.6 0.80 38,000
6 12.5 0.70 50,000
7 18.4 0.58 74,000
We have run seven simulations with differing values for the parameter c~. The values and the corresponding LISM plasma Mach numbers are listed in Table 1. The models range from supersonic without any additional pressure contribution (model 1) to subsonic oneshock models (4-7). The effective plasma temperature Tpl in Table 1 gives an indication what the plasma temperature would be if only LISM protons and electrons accounted for the total pressure. The real temperature for protons, electrons, and neutral H in the LISM is assumed to be 8000 K at the outer boundary in our models.
Table 2 Hydrogen wall results, and c~ M TH,wall [K] 1 2.0 1.75 89,000 2 3.5 1.32 58,000 3 5.0 1 . 1 1 73,000 4 7.5 0.90 78,000 5 9.6 0.80 70,000 6 12.5 0.70 100,000 7 18.4 0.58 80,000
comparison to Lya observations. rtpeak HP Extent Consistency with [cm -3] [ A U ] [AU] 12~ 52 ~ 73 ~ 112 ~ 0.219 104 210 Y N N N 0.229 99 210 Y N N N 0.228 95 210 N Y Y N 0.228 92 250 N Y Y N 0.221 91 260 N Y N N 0.196 88 150 N Y Y Y 0.181 87 180 N Y Y Y
data 139 ~ N N N N N Y Y
148 ~ N N N N N N Y
In Figure 2, we plot one-dimensional profiles of the neutral H density along the stagnation axis for all seven models. The prominent feature in all models is the hydrogen wall, where the densities reach above the 0.14 cm -3 LISM value. The peak density in all the hydrogen walls is the same, rtpeak = 0.22 cm -3 (Table 2), with the exception of models 6 and 7 where it is ,-15% lower. Table 2 contains the locations where the heliopause (HP) intersects the upstream stagnation axis, showing the effect of the increasing interstellar ram pressure for increasing c~. The table also contains estimates for the extent of the hydrogen walls, defined here as the length of the region of the stagnation axis where the neutral density is distinctly larger than the LISM value. This wall extent, too, is constant (~200 AU) for models 1-5 and somewhat thinner for models 6 and 7. The typical effective temperatures of neutral H inside the wall listed in Table 2 don't seem to follow a particular pattern and range from 60,000 K to 100,000 K. Downwind of the hydrogen wall, neutral hydrogen is depleted but hot (~ 105 K, see Fig. l b). In the tail region, the subsonic models have generally lower densities than the two-shock models (see left panel of Figure 2), with the difference between model 1 and model 7 being 20% at 100 AU and
-17-
H.-R. Miiller and B.E. Wood 0.10
I, t~,
Figure 2. 1D density profiles of neutral H upstream (right) and downstream (left) of the Sun, for all models. Note the different scales of the left and right hand panels.
growing to 40% at 1000 AU. In general, higher heliospheric hydrogen temperatures and densities are observed for models with lower values of c~, which is expected since a lower c~ corresponds with a higher LISM Mach number and therefore more heating. Because higher temperatures and column densities will produce broader Lyc~ absorption profiles, models with higher Mach numbers will also generally produce more absorption. The particle Boltzmann code returns the complete phase space information of neutral H. In particular, we obtain the velocity distribution function throughout the heliosphere that we can use to compute the absorption profiles. In Figure 3, radial velocity distributions are displayed for neutral hydrogen particles in the upwind and downwind directions for Model 1. In order to create these distributions, particles are summed within a certain area centered on the locations indicated in the figure. The area measures 104 AU 2 for Figure 3 and 3000 AU 2 for the Lyc~ absorption calculations in Section 3. Poissonian error bars (N ~ are displayed for each velocity bin. The bin size of the histogram is 6 km s -1. Gaussians have been fitted to the distributions (dashed lines), and the poor fits illustrate the non-Maxwellian character of the distributions. Looking upstream, the distribution functions have a broad core at -17 km s -1 which corresponds to moderately heated, decelerated neutral H of LISM origin coming toward the Sun. In the downstream direction, neutral H is moving away from the Sun, with the peak in the distribution at 28 km s -1. In both directions, there is a peak at 400 km s -1 which is the signature of splash component neutrals born through charge exchange in the solar wind. Neutrals with velocities between 100 and 300 km s -1 are component 2 neutrals produced in both the hot heliosheath and the heliotail. 3. C O M P A R I S O N
TO HST DATA
In upwind directions, the heliospheric H I column density is dominated by compressed, heated, and decelerated material just outside the heliopause (the hydrogen wall). For an observer at Earth, the absorption due to LISM H I in the upwind direction is blueshifted relative to the Ly~ rest wavelength. The decelerated heliospheric H I absorption is also blueshifted, but less than that of the LISM. Therefore, the hydrogen wall material accounts
-18-
Relating models of the heliosphere to Lyman-a absorption...
1000.0
-
100.0
10.0
Q) 1.0
o
. ,...~
0.1 150 0 -150 1000.0
100.0
o
10.0
1.0
.........
0.1 100 0
-
)!l JI, ...... h........ 0
-
-100 -200
- 100
0
100
200
300
400
500
( k m s -1) Figure 3. Neutral H radial velocity distributions for two different heliospheric locations based on Model 1, where 0 is the angle relative to the upwind direction of the interstellar flow, and R is the distance from the Sun. The dashed lines are Gaussian fits to the distributions, and residuals of the fits are shown below each panel. From [24].
for most of the non-LISM absorption observed on the red side of the s~turated core of the Lya absorption profile. Component 1 neutral hydrogen develops only small perpendicular velocity components in the hydrogen wall, such that the above holds for all sightlines through the hydrogen wall less than 90 ~ from the upwind direction. At hydrogen walls around other stars, where an observer looks from the outside rather than the inside, this scenario is reversed, meaning that the decelerated material of the stellar hydrogen wall is in fact more blueshifted than the LISM, and the additional absorption shows up at the blue wing of the main (interstellar) absorption feature. Figure 4 shows the observed Lya profile of a Cen B [10], which is 52 ~ from the upwind direction. Excess H I absorption is present on both the blue and red sides of the LISM absorption. The red side excess is due to the heliospheric absorption, while the blue side excess is due to absorption from analogous "astrospheric" material [11]. An additional detection of heliospheric H I absorption only 12~ from the upwind direction was provided by HST observations of 36 Oph [15]. For downwind lines of sight the H I density is much lower than in the hydrogen wall, but the sightline through the heated heliospheric H I is longer, potentially allowing heliospheric Lya absorption to be observed downwind as well [23], again on the red side because neutrals in the tail are accelerated by charge exchange to speeds higher than the LISM speed. A detection of heliospheric absorption along a downwind line of sight toward Sirius has been reported [16].
-19-
H.-R. Miiller and B.E. Wood
2.5 T 2.0
7 O
1.5
~0
o
1.0
" ~ 0.5 N 0.0 1215.0
1215.2
1215.4 1215.6 1215.8 Wavelength
1216.0
1216.2
Figure 4. HST/GHRS Lye spectrum of c~ Cen B, showing broad H I absorption at 1215.6 ~ and D I absorption at 1215.25 ~. The upper solid line is the assumed stellar emission profile and the dashed line is the ISM absorption alone. The excess absorption is due to heliospheric H I (vertical lines) and astrospheric H I (horizontal lines).
We can use the models listed in Table 1 to predict the amount of Lyc~ absorption we expect to see for various lines of sight through the heliosphere [24]. We compare these predictions with the observations of c~ Cen, 36 Oph, and Sirius, which we have already mentioned above as having excess Lyc~ absorption that is presumably heliospheric. In addition to those lines of sight, we also consider three additional lines of sight toward 31 Com, /3 Cas, and c Eri. The HST Lyc~ spectra of these stars show no evidence for excess absorption on the red side of the line that could be heliospheric [25], but these data still provide useful upper limits for the amount of absorption that could be present and they sample different directions through the heliosphere. In Figure 5, the absorption predicted by the seven models in Table 1 is compared with the Lyc~ absorption profiles observed toward the six stars. The 0 values shown in the figure are the angles from the upwind direction, which range from the nearly upwind line of sight toward 36 Oph (0 = 12~ to the nearly downwind line of sight toward c Eri (0 = 148 ~ The predicted Lyc~ absorption is shown after combination with the LISM absorption toward these stars, as determined from previous empirical analyses [10,15,16,25]. An attempt has been made to maximize the agreement between the data and the models by tweaking the assumed stellar emission profiles, as described by [24]. Significant disagreement remains in most cases despite these efforts. Based on Figure 5, we provide in the last six columns of Table 2 our evaluation of which observed stellar lines of sight (labeled by their 0 angles) are inconsistent with which models. None of the models are consistent with every line of sight. The low c~ models do better upwind and the high c~ models do better downwind. In general, the models predict too much absorption, the exception being the 36 Oph line of sight for which most of the models predict too little absorption. Model 7 is the only model that does not predict too much absorption along the downwind line of sight to e Eri.
-20-
Relating models of the heliosphere to Lyman-a absorption...
Cen
1.5 3'6' b ; 'h ' ('0'='1'2'~ . . . . . .
,J o5o 1 0.40
;'1' c;m
i;?;
~ ......
lO 1.o
0.30
8
6
L
0.5
0.20
4 0.10
?
2 ~O
0.00
o
0.0 0
20
40
60
80
o
100
20
40
60
80
20
100
40
60
80
100
~0 Q) e Eri (0=148 ~
'Si; ( 'u; ' 0 ' =' ' l ' 3 ; ~
7
a=2.0 .......... a = 3 . 5
0 v-t v
i/
v.6
N ...... y(/'
--
--
a=9.6 -
/~'1
a = l
0.4
YJ~,, 20
40
60
80
100
0
20
40
60
80
O0
0
20
40
60
80
100
Velocity (km s -1)
Figure 5. The red side of the Lya absorption profiles observed toward six stars, sampling different angles 0 relative to the upwind direction of the interstellar flow into the heliosphere. These data are compared with the heliospheric absorption predicted by the seven models listed in Table 1, which assume different values for a. From [24].
4. C O N C L U S I O N S We have presented models of the heliosphere constructed using a hybrid code combining a fluid treatment for the plasma and a kinetic treatment for the neutral particles. These models are potentially very useful for analyzing the heliospheric H I Lya absorption that has been detected toward a number of stars using HST observations. The models provide detailed information on the spatially-dependent, highly non-Maxwellian neutral H velocity distributions within the heliosphere. In principle, these distributions should allow accurate Lya absorption profiles to be computed for comparison with the HST data. However, we were unable to find a model which successfully reproduced the observed absorption. There are many possible reasons for this lack of success. One is that we simply have not experimented with a large enough range of input parameters. It is only the a parameter that is varied in the set of models in Table 1. It is possible that assuming somewhat different LISM or solar wind parameters, while still forcing them to be within available observational constraints, would improve agreement with the data. Another possibility is that we have neglected one or more physical processes in the models which could be
-21 -
H.-R. Miiller and B.E. Wood
important and affect our results, such as processes involving magnetic fields. One particular physical process that might be a factor is non-charge exchange interactions between the neutrals and the plasma, which are not considered in the current models. Perhaps there are enough neutral-proton collisions to remove a significant number of particles from the far wings of the velocity distributions (see Figure 3), thereby reducing the amount of predicted Lya absorption. This could be especially important for downwind directions where neutrals have traveled longer pathlengths through the heliosphere and have therefore had more opportunity for collisions with protons. And it is particularly in downwind directions where the current models predict way too much absorption. In future analyses, we hope to explore this possible explanation, and others. REFERENCES o
2. 3. 4. 5. 6. 7. o
9. 10. 11. 12. 13. 14. 15. 16. 17.
20. 21. 22. 23. 24. 25.
T. E. Holzer, Ann. Rev. Astr. & Astrophys., 27 (1989) 199. R. S. Steinolfson, V. J. Pizzo, and T. E. Holzer, Geophys. Res. Left., 21 (1994) 245. C. Wang and J. W. Belcher, J. Geophys. Res., 103 (1998) 247. G. P. Zank, Space Sci. Rev., 89 (1999) 413. V. B. Baranov and Y. G. Malama, J. Geophys. Res., 98 (1993) 15157. V. B. Baranov and Y. G. Malama, J. Geophys. Res., 100 (1995) 14755. P. C. Liewer, S. R. Karmesin, and J. U. Brackbill, in Solar Wind 8, ed. D. Winterhalter et al. (New York: AIP, 1996) 613. H. L. Pauls, G. P. Zank, and L. L. Williams, J. Geophys. Res., 100 (1995) 21595. G.P. Zank, H.L. Pauls, L.L. Williams, & D. Hall, J. Geophys. Res., 101 (1996) 21639. J. L. Linsky and B. E. Wood, Astrophys. J., 463 (1996) 254. K. G. Gayley, G. P. Zank, H. L. Pauls, P. C. Frisch, and D. E. Welty, Astrophys. J., 487 (1997)259. A. S. Lipatov, G. P. Zank, and H. L. Pauls, J. Geophys. Res., 103 (1998) 20631. H.-R. Mfiller and G. P. Zank, in Solar Wind 9, ed. S. R. Habbal et al. (New York: AIP, 1999) 819. H.-R. Mfiller, G. P. Zank, and A.S. Lipatov, J. Geophys. Res., 105 (2000) 27419. B. E. Wood, J. L. Linsky, and G. P. Zank, Astrophys. J., 537 (2000) 304. V. V. Izmodenov, R. Lallement, and Y. Malama, Astron. Astrophys., 342 (1999) L13. W. L. Fite, A.C.H. Smith, and R.F. Stebbings, Proc. R. Soc. London Ser. A, 268 (1962) 527. J. L. Linsky, S. Redfield, B. E. Wood, & N. Piskunov, Astrophys. J., 528 (2000) 756. R. Lallement, R. Ferlet, A. M. Lagrange, M. Lemoine, and A. Vidal-Madjar, Astron. Astrophys., 304 (1995) 461. E. Qu~merais, J.-L. Bertaux, B. R. Sandel, and R. Lallement, Astron. Astrophys., 290 (1994) 941. G. Gloeckler, et al., Science, 261 (1993) 70. B. E. Wood and J. L. Linsky, Astrophys. J., 474 (1997) L39. L.L. Williams, D. T. Hall, H. L. Pauls, and G. P. Zank, Astrophys. J., 476 (1997) 366. B. E. Wood, H.-R. Mfiller, and G. P. Zank, Astrophys. J., 542 (2000) 493. A. R. Dring, J. L. Linsky, J. Murthy, R. C. Henry, W. Moos, A. Vidal-Madjar, J. Audouze, and W. Landsman, Astrophys. J., 488 (1997) 760.
-22-
Interstellar atoms in the heliospheric interface V. V. Izmodenov a, aDepartment of Aeromechanics and Gas Dynamics, Faculty of Mechanics and Mathematics, Lomonosov Moscow State University, Vorob'evy Gory, Glavnoe zdanie MGU, Moscow 119899, Russia (
[email protected]) In order to enter the heliosphere the interstellar atoms pass through the heliospheric interface - the region of the solar wind and interstellar plasma interaction. In the interface, the interstellar hydrogen atoms strongly interact with the LISM plasma component by charge exchange. This interaction results in the modification of both plasma and interstellar atom flows. Thus, the atoms penetrate into the heliosphere disturbed. This opens a possibility to use interstellar atoms and their derivatives - pickup ions and anomalous cosmic rays - for remote diagnostics of the heliospheric interface plasma structure. However, the interpretations of remote experiments are critical to accurate theoretical models. In this paper advanced self-consistent models of the heliospheric interface are reviewed. Evolution of the atom velocity distribution in the heliospheric interface is discussed with the emphasis on interpretation of present and future space experiments. 1. I n t r o d u c t i o n At the present time there is no doubt that local interstellar medium (LISM) is partly ionized plasma. Interstellar plasma component interacts with the solar wind (SW) plasma and forms the heliospheric interface (Figure 1). The heliospheric interface is a complex structure, where the solar wind and interstellar plasma, interplanetary and interstellar magnetic fields, interstellar atoms, galactic and anomalous cosmic rays (GCRs and ACRs) and pickup ions play prominent roles. In this paper I will review physical processes connected with interstellar atoms. Interstellar atoms of hydrogen are the most abundant component in the circumsolar local interstellar medium. These atoms penetrate deep into the heliosphere and interact with interstellar and solar wind plasma protons by charge exchange. This interaction influences the structure of the heliospheric plasma interface significantly. Being disturbed in the interface, interstellar H atoms and their derivatives such as pickups and ACRs can serve as remote diagnostics of the heliospheric interface. To make constructive conclusions from space experiments one must use an adequate theoretical model. The main difficulty in the modelling of the heliospheric interface is a kinetic character of the interstellar H *The research described in this publication was made possible in part by Award No.RP1-2248 of the U.S. Civilian Research and Development Foundation for the Independent States of the Former Soviet Union (CRDF), INTAS projects # 97-0512 and #YSF 00-163, RFBR grants # 01-02-17551, 99-02-04025, and International Space Science Institute in Bern.
-23 -
K Izmodenov
Figure 1. Structure of the heliospheric interface. The termination shock (TS), the heliopause (HP), the bow shock (BS) separates the heliospheric interface in four regions. Region 1 is the supersonic solar wind. Region 2 is the heliosheath with compressed solar wind plasma. Region 3 is the region of disturbed interstellar medium. The region is extended upstream BS (see, section 3). Region 4 is undisturbed interstellar medium.
atom flow in the heliospheric interface. Since the effects of elastic H-H, H-p collisions are negligibly small as compared with charge exchange, the latter process determines the character of the H atom flow in the interface (see, e.g., [10] and section 3 of this paper). Atoms newly created by charge exchange have local plasma properties. Since plasma properties are different in four regions shown in figure 1, there are four populations of interstellar atoms in the heliospheric interface. Population 1 is the atoms created in the supersonic solar wind. Population 2 is the atoms created in the heliosheath. Population 3 is the atoms created in the region of disturbed interstellar plasma (region 3 in Figure 1). Population 4 is the original interstellar atoms. The strength of H atom-proton coupling can be estimated through the calculation of mean free paths of H atoms in plasma. Generally, the mean free path of s-particle in t-gas can be calculated by the formulae: L - 5 M s~t / S t " Here w~ is the individual velocity of s-particle, S M u t ~ S t is individual Sparticle momentum transfer rate in t-gas, which is a function of local plasma parameters. Table 1 shows the mean free paths of H atoms with respect of charge exchange with protons. The mean free paths are calculated for typical atoms of different populations at different regions of the interface in the upwind direction. For every population of H atoms there is at least one region in the interface where Knudsen number K n N~ 0 . 5 - 1"0. Therefore kinetic Boltzmann approach must be used to MeanFreePath CharacteristicLength
-24-
Interstellar atoms in the heliospheric interface describe interstellar atoms in the heliospheric interface.
Table 1 Meanfree paths of H-atoms in the heliospheric interface with respect to charge exchange with protons, in AU Population At TS At HP Between HP and BS LISM 110 870 4 (primary interstellar) 150 100 58 190 3 (secondary interstellar) 66 40 110 200 2 (atoms originated in the heliosheath 830 200 240 490 1 (neutralized solar wind) 16000 5!0 , ,
.....
....
2. Heliospheric Interface M o d e l Baranov et al. [1] and Baranov and Malama [2] considered a two-dimensional (2D) axisymmetric model of the SW/LISM interaction. The interstellar and solar wind plasma components were treated as fluids. To describe the flow of interstellar atoms the Boltzmann equation was solved: HPf.(r. wp)dwp 0/H(r , WH) ~ F O/H (r, WH) = --fH(r. WH) If IWH -- Wpla.z WH" 0r mH 0WH -+-fp(r, W H ) [fW ~ i
WH [ HP fH(r, WH)dWH 9 9
(~ §
(1)
r, WH).
Here fH(r, WH) is the distribution function of H atoms; fp(r, wp) is the local distribution function of protons, assumed to be Maxwellian; wp and wH are the individual proton and H atom velocities, respectively; a HP is the charge exchange cross section of a H atom with a proton; fli is the photoionization rate; mH is the atomic mass; ~imp~t is the electron impact ionization rate; and F is the sum of the solar gravitational force and the solar radiation pressure force. The plasma and neutral components interact mainly by charge exchange. However, photoionization, electron impact ionization, solar gravitation, and radiation pressure are also taken into account in equation (1). The interaction of the plasma and neutral components leads to the mutual exchange of mass, momentum and energy. The effect of this interaction can be represented by adding source terms Q -- {ql, q2,z, q2,r, q3, O)T to the right-hand side of the plasma equations. The terms ql, q2 = {q2,~,q2,~}, q3 describe the mass, momentum, and energy sources in the thermal plasma component due to interaction with neutrals. These source terms can be expressed through the integrals of the atom distribution function fH:
q~ - n i l " (~i + ~imp~t), nH = f fH(WH)dWH,
wp) f~ (WH)fp (wp)dw~dwp,
-25-
K Izmodenov 1
Here f l r f uaHP(u)fv(wp)dw v is the charge exchange rate, and u is the relative atomproton velocity. Supersonic boundary conditions are used for the unperturbed interstellar plasma and for the solar wind plasma at the Earth's orbit. The velocity distribution of interstellar atoms is assumed to be Maxwellian in unperturbed interstellar medium. 3. On t h e effect of H-H, H-p elastic collisions Recently Williams et al. [3] have suggested that a population of hot hydrogen atoms is created in the heliosphere through elastic H-H collisions between atoms of population 1 and interstellar atoms of populations 3 and 4. Izmodenov et al. [5] examined the approach used by Williams and argued that two assumptions used by Williams et al. result in significant overestimation of the H-H collision effect. 1. Williams et al. [3] applied the momentum transfer cross-section calculated by Dalgarno [4] for quantum mechanically indistinguishable particles in H-H collisions. Treating the colliding particles as quantum mechanically indistinguishable or classically distinguishable could lead to significant differences in the meaning and values of the cross-sections. Dalgarno's approximation is appropriate for a quantum particle ensemble when velocities of distinct particles are correlated with each other. However Williams et al. applied this like-particles H-H collision momentum transfer cross-section to the description of the collisions between different populations of H atoms that are classically distinguishable. Classical statistics should be used under such conditions. Izmodenov et al. [5] calculated the momentum transfer cross-section for H-H collisions in the 0.01-1000 eV energy range treating colliding particles as classically distinguishable. Figure 2A compares the calculated momentum transfer, Dalgarno, and charge exchange cross sections as functions of relative velocity of the colliding particles. It is seen that the cross-section calculated in [5] is significantly smaller than the Dalgarno cross-section when V > 10 km/s. 2. Williams et al. [3] used a Bhatnagar-Gross-Krook (BGK)-like approximation for the Boltzman's equation collision term. This approximation is based upon on an assumption of a complete randomization of particle velocities in the collisions, without making a distinction between 'strong' and 'weak' collisions. If one calculates momentum transfer term for different H-atom populations using the BGK-like approximation, one would obtain the values that 1-2 order of magnitude larger than for our approximation. For calculations with the Dalgarno's cross section, the overestimation introduced by the BGKlike approximation is much smaller, but still significant. This effect can be explained by different functional dependencies of the cross sections and by the BGK-like collision term. Izmodenov et al. [5] concluded that the influence of elastic H-H, H-p collisions is negligible in the heliospheric interface since the dynamic influence of charge exchange is stronger by several orders of magnitude. 4. Basic results of t h e model: H a t o m s Since the elastic H-H, H-p collisions are not important dynamically, Baranov-Malama model [2] describes the main physical processes correctly. In this section main results of
-26-
Interstellar atoms in the heliospheric interface
~ 10- 6
f
l.qarno [1960i -
~
: E
Figure 2. A. Momentum transfer cross section (in cm 2) as functions of relative velocity (in cm sec-~; B. Comparison of momentum transfer of population 2 due to elastic collisions with other H atom populations (in (cm sec) -2) in the upwind direction.
the model are briefly reviewed. Detailed results of the model reported in Baranov and Malama ([2], [6], [7]), Baranov et al. [8], Izmodenov et al. [9], Izmodenov [10]. One of the main and the most spectacular results of the model is prediction of the hydrogen wall, a density increase at the heliopause (figure 3A). The hydrogen wall is made up by secondary interstellar atoms of population 3. The hydrogen wall was predicted by Baranov et al. [1]. Linsky and Wood [11] were first to demonstrate that observed by HST Ly-a spectra toward Sirius can not be understood without taking into account the absorption produced by heliospheric hydrogen wall. Figure 3C shows the heliospheric absorption toward a-Cen calculated on the basis of Baranov-Malama model. There is still a missing absorption on the blue side. This is probably absorption produced by heated neutral gas around the target star (" astrosphere"), as suggested first by Wood and Linsky. Interpretation of stellar spectra can provide us constraints on both the heliospheric interface and astrospheres. The first attempt to constrain the interface with to help of stellar spectra was done by Gayley et al. [12]. Later, Izmodenov et al. [13] interpreted Ly-c~ spectra toward star Sirius-A (Figure 3D) on the basis of the heliospheric interface model. It has been shown that in this direction (41 ~ from downwind) two populations (populations 3 and 2) of neutral H atoms produce a non-negligible absorption. Figure 3B shows number density of population 2 toward Sirius direction. Despite small number density this population produces non-negligible absorptions due to high temperature. During this COSPAR Colloqium Wood, Muller and Zank reported the study of Ly-a spectra toward six nearby stars observed by HST. Comparing their model calculations with data the authors concluded that their model predict too much absorption in sidewind and downwind directions. More theoretical work has to be done to understand these discrepancies between theory and observations.
-27-
K Izmodenov
Figure 3. A. The hydrogen wall is an increase of number density of population 3 at the heliopause. The distribution is shown for direction toward c~Cen. B.The distribution of population 2 toward Sirius. C. HST spectrum of a Cen-A near Ly-a line center. Also shown is simulated stellar profile prior to interstellar absorption, the same profile after interstellar absorption and the expected profile after the cloud plus the calculated heliospheric H absorption. A missing absorption on the blue side is probably "astrospheric". D. HST spectrum of Sirius-A near Ly-a line center. Simulated profiles are similar of profiles on plot C.
5. Basic results of t h e model: influence on p l a s m a flows Interstellar atoms strongly influence the heliospheric interface structure. In the presence of interstellar neutrals, the heliospheric interface is much closer to the Sun than it would be in a pure gas dynamical case (see, e.g., figure 2 in Izmodenov, 2000 [10]). The termination shock becomes more spherical. The Mach disk and the complicated shock structure in the tail disappear. Interstellar neutral atoms also affect the flow of supersonic interstellar plasma upstream of the bow shock, the flow of the supersonic solar wind upstream of the termination shock, and plasma structure in the heliosheath. The flow of the supersonic solar wind upstream of the termination shock is disturbed by charge exchange with the interstellar neutrals. The new created by charge exchange ions are picked up by the solar wind magnetic field. The Baranov-Malama model assumes immediate assimilation of pickup ions into the solar wind plasma. The effect of charge exchange on the solar wind is significant. By the time the solar wind flow reaches the termination shock it is decelerated (15-30 %), strongly heated (in 5-8 times) and mass
-28-
Interstellar atoms in the heliospheric interface
loaded (20-50 %) by the pickup ion component. The interstellar plasma flow is disturbed upstream of the bow shock by charge exchange with the secondary atoms originated in the solar wind and compressed interstellar plasma. Charge exchange results in the heating (40-70 %) and deceleration (15-30 %) of the interstellar plasma before it reaches the bow shock. The Mach number decreases and for a certain set of interstellar parameters (nH,LISM > > ?~p,LISM)the bow shock may disappear. The plasma structure in the heliosheath is also modified by the interstellar neutrals. In a pure gasdynamic case (without neutrals) the density and temperature of the postshock plasma is nearly constant. However, the charge exchange process leads to a strong increase of the plasma number density and decrease of its temperature. Baranov and Malama (1996) [7] pointed out that the electron impact ionization process may influence the heliosheath plasma flow by increasing the gradient of the plasma density from the termination shock to the heliopause. The effects of interstellar atom influence on the heliosheath plasma flow may be important, in particular, for the interpretations of (1) kHz radio emission detected by Voyager (see, [15], [16]) and (2) possible future heliospheric imaging in energetic neutral atom (ENA) fluxes [17].
6. Interpretations of spacecraft experiments on the basis of the BaranovMalama model The Sun/LISM relative velocity and the LISM temperature are now well constrained. Using the new SWICS pickup ion results and an interstellar HI/HeI ratio of 13 i 1 (the average value of the ratio toward the nearby white dwarfs), Gloeckler et al. (1997) [18] concluded that n L r s M ( H I ) -- 0.2 • 0.03 am -3. However there are no direct ways to measure the circumsolar interstellar electron (or proton) density. Therefore, there is a need for indirect observations (inside the heliosphere) which can bring stringent constraints on the interstellar plasma density and on the shape and size of the interface. Such constraints can been done on the basis of theoretical models of the interface. Izmodenov et al. [9] used the Baranov-Malama model to study the sensitivity of the various types of indirect diagnostics of local interstellar plasma density. The diagnostics are the degree of filtration, the temperature and the velocity of the interstellar H atoms in the outer heliosphere (at the termination shock), the distances to the termination shock, the heliopause, and the bow shock, and the plasma frequencies in the LISM, at the bow shock and in the maximum compression region around the heliopause which constitutes the "barrier" for radio waves formed in the interstellar medium. Izmodenov at al. [9] searched also for a number density of interstellar protons compatible with SWICS/Ulysses pickup ion and Voyager, HST, SOHO observations of backscattered solar Ly c~ and kHz radiations observed on Voyager. Table 2 presents estimates of np,niSM obtained from different types of observations using the Baranov-Malama model. Izmodenov et al. [9] concluded that it is difficult to reconcile these estimates without some modification to the model. Two mutually exclusive solutions have been suggested: (1) It is possible to reconcile the pick-up ions and Ly a measurements with the radio emission time delays if a small additional interstellar (magnetic or low energy cosmic ray) pressure is added to the main plasma pressure. In this case, n p , L I S M - 0.07 cm -3 and n H , L I S M - - 0.23 cm -3 is the favored pair of interstellar densities. However, in
-29-
V. Izmodenov
this case, the low frequency cutoff at 1.8 kHz doesn't correspond to the interstellar plasma density, and one has to search for another explanation. (2) The low frequency cutoff at 1.8 kHz constrains to the interstellar plasma density, i.e., rtp,LISM - - 0 . 0 4 a m - 3 . In this case, the bulk velocity deduced from the Ly a spectral measurement is underestimated by about 30-50% (the deceleration is by 3 km s -1 instead of 5-6 km s-~). Model limitations (e.g. stationary hot model to derive the bulk velocity) or the influence of a strong solar Ly a radiation pressure may play a role. In this case, there would be a need for a significant additional interstellar (magnetic or cosmic ray) pressure as compared with case (1).
Table 2 Intervals of Possible Interstellar Proton Number Densities Type of Heliospheric Interface Diagnostics Range of Interstellar Proton Number Density SWICS/Ulysses pick-up ion 0.02 c m - 3 < rtp,LISM < 0.1 c m - 3 0.09 cm -3 < nH,TS < 0 . 1 4 cm -3 Gloeckler et al., [1997] Ly- a, intensity 0.11 cm -3 < nH,TS < 0.17 cm -3 Qugmerais et al., [1994]
np,LISM '( 0.04 cm -3 or
Ly-a, Doppler shift 18 km s -1 < VH,TS < 21 km s -1 Bertaux et al. [1985], Lallement et al. ,
0 . 0 7 c m - 3 < np,LISM <
0.2 cm -3
0 . 0 8 c m - 3 < rtp,LISM <
0.22
[1996], Cl.rk
np,LISM < 0.07 cm -3 (for nH,LISM - - 0.23 cm -3)
el. [199S]
Voyager kHz emission (events) 110 AU < RAU < 160 AU Gurnett and Kurth [1996] Voyager kHz emission (cutoff) 1.8 kHz Gurnett et al. [1993], Grzedzielski and Lallement [1996]
cm -3
np,LISM - - 0 . 0 4 c m - 3
7. N e x t F r o n t i e r s Several important physical effects are not taken into account in the model. These effects may be: 1) interstellar and heliospheric magnetic fields; 2) heliolatititudional and solar cycle variations of the solar wind; 3) galactic and anomalous cosmic rays; 4) essentially multi-fluid character of the solar wind, when the solar wind protons, pickup protons and electrons must be considered as different fluids. During the last years a big theoretical effort of several groups has been focused on understanding the influence of these (and other) effects on the heliospheric interface. For a recent review of of the theoretical models of the heliospheric interface see Zank (1999)
[19].
-30-
Interstellar atoms in the heliospheric interface
In spite of many interesting and important physical effects studied, many models don't take into account the interstellar neutrals or take them into account in oversimplified approximations. Since the H atoms are probably the most dominant component of LISM, one must be very careful when applying these models in the interpretations of the heliospheric interface observations. The use of inadequate theoretical models can result in incorrect interpretations of spacecraft experiments. This, in turn, may results in wrong conclusions on the interstellar parameters. From our point of view, the future theoretical models of the heliospheric interface must preserve the main advantage of the Baranov-Malama model - the kinetic description of nterstellar atoms- and introduce the new effects one by one. An example of a such study has been done in [21], [20]. These authors considered the effect of the galactic cosmic rays (GCRs) on the heliospheric interface structure. Firstly, the influence of GCRs on the heliospheric interface plasma structure has been studied in the absence of the interstellar neutrals ([21], [20]). Then the interstellar atoms have been added to the model. The results were compared with the results of Baranov-Malama model. In spite of the significant influence of the GCRs on the heliospheric plasma structure in the two component (plasma+ GCRs) model, most of the effects reduce or disappear in case of the three-component model. The influence of GCRs on the plasma flow pattern is negligible as compared to the influence of H atoms everywhere in the heliospheric interface, except near the bow shock, the structure of which can be strongly modified by GCRs. Another example of such study were presented by Vladimir Baranov during this COSPAR Colloquim. He and his co-workers made self-consistent study of the interstellar magnetic field (IMF). It has been shown that in the presence of interstellar atoms the effect of IMF signicantly reduces. Similar studies are needed to account for other physical effects that that have been found to influence the structure of the heliospheric interface in the models without neutrals. Such studies can bring new constraints on the local interstellar parameters and on the structure of the heliospheric interace. Significant progress can be reached if such new theoretical models will be tested by the old and new measurements. One of the promising methods of the heliospheric interface diagnostics is interpretation of absorption spectra measured toward nearby stars. It was briefly discussed in section 4. Another possible heliospheric interface diagnostics is the study of the interstellar minor elements as O, N, Ne, C. The interest in minor species is now growing due to the recent successful detection of pickup and ACR ions, by the Ulysses and Voyager spacecraft, respectively. Oxygen is of particular interest, because it is one of the most perturbed elements due to its large charge exchange cross section with protons. Izmodenov et al. [22] [23] compared OI/HI heliospheric and interstellar ratios. The heliospheric interface filtrations of both hydrogen and oxygen have been computed on the basis of the Baranov-Malama model. A rather good agreement has been found between data and theory. More detail study of the interstellar oxygen (and other elements) in the interface and how their density inside the heliosphere sensive to interstellar parameters is required. Such study can give us additional constraints on the heliospheric interface parameters. The interest to interstellar oxygen in the heliospheric interface increases also, because Gruntman and Fahr ([24], [25]) proposed a mapping of the heliopause in the oxygen ion O + resonance line (83.4 nm).
-31 -
V. Izmodenov A very promising and powerful tool to study the solar wind interaction with the surrounding local interstellar medium is global imaging of the heliosphere in the fluxes of energetic neutral atoms (ENA) ([17]). The ENAs that are produced in charge exchange of the heated plasma and background neutral gas can be readily detected at 1 AU. Global ENA images, the angular and energy dependences of ENA fluxes, are dependent on the solar plasma density, temperature, and velocity in the heliosheath. The size and structure of the heliospheric interface region depend on the parameters of the interstellar plasma and gas. Hence the ENA images should also depend on the LISM parameters (Izmodenov and Gruntman, this conference). Finally, we can conclude, that in spite of our general understanding of the physical processes in the heliospheric interface, a major theoretical and experimential effort is needed to obtain the detailed structure of the heliospheric interface and local interstellar parameters. Acknowledgement. I thank to the referee of this paper for very helpful comments and suggestions. I thank to Baranov, Malama, Lallement, Geiss, Fahr and Gruntman for useful discussions. I also thank to Rosine Lallement for Figures 3C and 3D. REFERENCES
Baranov V.B., Lebedev M.G., Malama Yu.G., Astrophys. J. 375, 347-351, 1991. Baranov V.B., Malama Yu.G., J. Geophys. Res.,98, 15157-15163, 1993. Williams L.L., D.T. Hall, L. Pauls, G. Zank, Astrophys. J., 476, 366 1997. Dalgarno A., Proc. Phys. Soc., 75, 374-377, 1960. Izmodenov V. V., Yu.G. Malama, et al., Astrophys. Space Sci. 274, 71-76, 2000. Baranov V.B., Malama Yu.G., J. Geophys. Res.,100, 14755-14761, 1995. Baranov V.B., Malama Yu.G., Space Sci. Rev., 78, 305-316, 1996. 8. Baranov V.B., Izmodenov V.V., Malama Yu.G., J. Geophys. Res. 103, 9575, 1998. 9. Izmodenov V. V., J. Geiss, R. Lallement, et al., J. Geophys. Res. 104, 4731,1999a. 10. Izmodenov V. V., Astrophys. Space Sci., 274, 55-69, 2000. 11. Linsky, J. and B.Wood, Astrophys. J., 463, 254L, 1996. 12. Gayley, K.G., Zank, G.P., Pauls, H.L., Frisch P.C., Welty D.E., ApJ, 487, 259, 1997. 13. Izmodenov V.V., R. Lallement, Yu. G. Malama, A & A 342, L13-L16, 1999b. 14. Wood B., H. Muller, G. Zank, this issue. 15. Gurnett D. and W. S, Kurth, Space Sci. Rev., 78, 53-66, 1996. 16. Treumann,R. A., Macek, W.M. and V.V. Izmodenov, A & A 336, L45, 1998. 17. Gruntman M., Rev. Space Instr. 68, 3617-3656, 1997. 18. Gloeckler, G., L. Fisk, J. Geiss, Nature, 386, 374-377, 1997. 19. Zank, G., Space Sci. Rev., 89, 413, 1999. 20. Myasnikov, A. V., D.B. Alexashov, et al., J. Geophys. Res. 105, 5167, 2000. 21. Myasnikov, A. V., V.V. Izmodenov, et al., J. Geophys. Res. 105, 5179, 2000.. 22. Izmodenov V.,Lallement R.,Malama Yu.G., Astron. Astrophys., 317, 193-202, 1997. 23. Izmodenov V. V., R. Lallement, J. Geiss, Astron. Astrophys. 344, 317- 321, 1999c. 24. Gruntman M., H. Fahr, GRL 25, 1261-1264, 1998. 25. Gruntman M., H. Fahr, JGR 105, 2000. 26. Izmodenov V. and M. Gruntman, this conference, 2000. .
2. 3. 4. 5. 6. 7.
-32-
MHD modeling of the outer heliosphere: Numerical Aspects Nikolai V. Pogorelov
a,
aInstitute for Problems in Mechanics, Russian Academy of Sciences, 101-1 Vernadskii Avenue, Moscow 117526, Russia
Current achievements and future prospects are discussed in the modeling of the solar wind interaction with the local interstellar medium in magnetohydrodynamic approximation. Farfield boundary conditions are described that make it possible to attain a substantial improvement in the resolution of the complicated flow structure. Special attention is paid to structurally unstable, nonevolutionary MHD shocks which can occur even in three-dimensional cases due to the presence of the symmetry plane.
1. I N T R O D U C T I O N Modern non-linear high-resolution numerical methods have recently become an efficient tool for solving complex multi-shocked gas dynamic flows. This is due to their robustness for strong shock wave computations. Application of linear numerical schemes of an order of accuracy higher than one results in spurious oscillations of distributions in the vicinity of discontinuities [18]. First order of accuracy, however, is insufficient for sharp resolution of discontinuities. For this reason linear schemes of higher order of approximation involve various additional terms called artificial dissipation. Extension of non-linear characteristicallyconsistent numerical methods to magnetohydrodynamics is not straightforward. These methods (Godunov schemes) in one or another way use the solution of the hyperbolic Riemann problem to determine numerical fluxes through the computational cell interfaces [43]. Although the exact solution of the MHD Riemann problem exists [ 19], it is too complicated and multivariant to be used in regular calculations. For this reason different approximate solutions or solutions of the linearized problem are widely applied. Written for the components of the state vector U, the conservative system of ideal MHD has the form (the notation is conventional [22])
0U 0E OF 0G o t + -g-Tx + --~y + Oz
(1)
~ a~
*The work was partially supported by the Russian Foundation of Basic Research under grant No. 98-01-00352.
-33-
N. K Pogorelov
where pv
pu Pu2 + PO
p
pu pv pw e Bx By Bz
U
puv -
,
4re
BxBy
47r BxBz puw 47r (e + po)u - -Bx ~ (v. B)
E-
puw pvw w
F
0 u B y - vBx uBz - wBx
v B x - uBy
0 vBz
pw
G
BxBy 47r By2 Pv2 + PO 47r ByBz pvw 47r By (e + po)v - - ~ (V" B) puv
-
wBy
0
BxBz 47r ByBz 4zc
Pw2 + PO
n
~
m
Bz
47r
x div B.
8Z(v.B)
(e + po)w - - ~
bl V W
w B ~ - uBz wBy - vBz 0
Here we intentionally preserved the source term H which is identically zero, since magnetic charge does not exist and magnetic field is therefore divergence-free. One can omit this term and solve the resulting system. In this case, however, special care must be taken in order to preserve the condition div B = 0 numerically [7]. Different approaches which are frequently used for this purpose include the application of the magnetic field vector potential [17,25], artificial scalar potential [39,41], and staggered grid methods [12,40]. A new method has recently been developed [24] which, in an efficient way, combines the vector potential and staggered grid approaches. Otherwise, one can solve the modified nonhomogeneous system (1) as suggested in [2] and [33]. It seems to allow the numerical magnetic charge to convect away from the computational region. Although this approach became rather popular (see [23,27,36]), it is questionable whether it can be used near stagnation points, in vortex regions, and in genuinely unsteady problems. It is possible to discretize system (1) on the numerical mesh and take advantage of the finite-volume approximation, in which equations are solved expressing the conservation of fundamental properties of the medium in an individual computational cell. To determine the fluxes through the computational cell interfaces, solutions of the MHD Riemann problem are used. The schemes actually differ from each other by the choice of the Riemann problem solver applied. For example, the scheme used in [47,48] is based on the solutions in the form of simple waves (even shocks are approximated by compression waves). This approach represents the extension to MHD of the numerical scheme [6], which is a version of the Osher scheme [22]. In [11] an MHD extension of the PPM (piecewise-parabolic method [9]) was suggested. It is
-34-
MHD modeling of the outer heliosphere: Numerical aspects based on the iterative nonlinear Riemann problem solver in which the relations on discontinuities are exactly satisfied. This method, however, treats expansion waves as expansion shocks. This is a dangerous procedure, since such nonphysical shocks can truly arise in the process of calculation. One of the advantages of this method is that it includes Alfv6n waves in the possible configurations of shocks connecting the states on the opposite sides of a computational cell even in coplanar problems. As will be shown later, this may help to get rid of structurally unstable, nonevolutionary MHD shocks and combination of shocks. The Roe-type method [38] can formally be written out for a MHD system [8,20,32]. This method is based on the solution of the linearized Riemann problem which exactly satisfies, however, the conservation relations on discontinuities. This solution was shown to be nonunique [29,30]. On the other hand, the procedure ensures an exact satisfaction of the Hugoniot relations only in one-dimensional problems. This is also true for pure gas dynamics, but the error in MHD case can be much greater. The main reason lies in the fact that the magnetic field component Bx normal to the cell surface must be continuous in the one-dimensional statement, which is not true as far as multidimensional problems are concerned. The attempts to formulate a Roe-type solver for the one-dimensional version of the nonhomogeneous system (1) turned out to be unsuccessful [34] and this method is now frequently substituted by the Courant-IsaacsonRees type method. In the latter, the numerical flux at the cell interfaces is calculated using the arithmetic mean of the quantities on the both sides of the individual interface [35]. One more possibility is to use an extremely robust TVD Lax-Friedrichs scheme [5].
2. MHD MODELING OF THE SW-LISM INTERACTION 2.1. Current achievements The presence of the solar and interstellar magnetic fields necessitates solving the MHD equations to model the solar wind (SW) interaction with the local interstellar medium (LISM). The influence of magnetic field is important if the magnetic pressure becomes comparable with the dynamic pressure of the flow. The magnitude and direction of the interstellar magnetic field is uknown. We can only say that a strength of about 1.6 #G is consistent with observations of polarization and pulsar dispersion measurements [ 15]. Estimates of the heliosphere confinement pressure give an upper limit of 3-5 #G (see [10]). Concerning the direction of the magnetic field in the local interstellar cloud, we can only refer to [ 16] which states that if the cloud surface is perpendicular to the direction of the gas flow, the direction of the magnetic field is approximately parallel to the cloud surface. These estimates lead us to the conclusion of the importance of MHD modeling. The first numerical study of the stellar wind interaction with the magnetized LISM was performed in [17,25] by a rather viscous flux-splitting method for the case of B~ II V~, where the subscript oc refers to the LISM parameters. The authors, actually, performed parametric calculations for various ratios of dynamic pressures and stagnation temperatures and some of them were very far from the SW-LISM interaction case. In [3] very accurate shock-fitting calculations of the upwind region of the interaction were performed and some of the results obtained in [17] were criticized. As later turned out [28], this mainly concerned the case of an irregular interaction in which parallel fast MHD shocks were nonevolutionary (structurally unstable), while singular (switch-on) shocks could not exist due to the symmetry restrictions. It was noticed in [45] that the toroidal component of the interplanetary magnetic field (IMF)
-35-
N. V. Pogorelov
Figure 1. General configuration of the heliosphere: density (below the symmetry axis) and total pressure logarithm isolines [28]. The LISM flow is directed from right to left. The Sun is at the origin. can be important. Three-dimensional calculations with the presence of IMF for B ~ ]1 V~ were reported in [46]. Systematic shock-capturing MHD calculations of axisymmetric problems for various magnitudes of the interstellar magnetic field (ISMF) were made in [28,31]. Both regular interaction with a single bow shock and irregular interaction with additional discontinuities were discovered. The general configuration of the flow pattern corresponding to the two-shock model [4] is shown in Fig. 1 for the following SW and LISM parameters: ne = 7 cm -3, Ve = 450 km s -~ , Me = 10, no~ - 0.07 cm -3, Voo = 25 km S-1, and Moo = 2. Here the subscript e refers to the SW parameters at Earth's distance from the Sun. M stands for the Mach number. The dimensionless value of the magnetic field is specified via the Alfv6n number A ~ - V~/~/B2/47rp~. In the presented case A~ = x/~ (see [28]). The abbreviations BS, TS, and HP correspond to the bow shock, the heliospheric termination shock, and the heliopause, respectively. The influence of the ISMF direction on the shape of the global heliopause was studied in [27], [36]. The results turned out to be consistent with the MHD modeling of the three-dimensional heliopause on the basis of the Newtonian approximation [ 14]. The distribution of the streamlines and magnetic field lines in the symmetry plane for the cases with Ao~ = 2 for two different angles between ISMF and LISM velocity are shown in Fig. 2 (see [27]). All previously cited calculations disregarded the charge exchange processes occurring among neutral and charged particles which are extremely important in the SW-LISM interaction. The IMF was also neglected. Three-dimensional calculations for various direction of ISMF, with IMF in the form of the Parker nominal spiral and with the simplified treatment of the charge-exchange processes were performed in [23]. McNutt at al. [26] investigated three-dimensional, both purely hydrodynamic and MHD flows, taking into account the latitudinal variation of the SW velocity and density in accordance with the Ulysses data. The solar cycle dependence of the heliospheric shape was studied in [42] on the basis of
-36-
MItD modeling of the outer heliosphere: Numerical aspects
Figure 2. Streamlines and magnetic field lines for A~ - 2 in the symmetry plane. The angle between B~ and V~ is equal to 90 ~ (left) and 45 ~ (right). a global MHD simulation of the time-dependent SW interaction with the magnetized LISM. Charge exchange processes were, however, neglected. This is not very reasonable, since the one-dimensional results [37] and [44] show that even if magnetic field is added, pick-up ions still remain a dominant factor of the interaction. This is in agreement with the recent results [ 1] where self-consistent axisymmetric MHD calculation were performed with Monte Carlo modeling of the neutral particle motion. The accuracy of numerical calculations can be substantially increased if we reduce the size of the computational region. Otherwise, the resolution of discontinuities will be rather poor [36]. To do this, we must shift the boundary conditions from infinity to the finite distance from the Sun and state nonreflecting boundary conditions at it. Below we consider an efficient approximate procedure created for this purpose.
2.2. Far-field boundary conditions Let us consider a local one-dimensional system 0U 0U 0--7 + a ~ x - 0.
(2)
The x-axis is directed perpendicular to the computational cell interface F. Let us choose the following form of the unknown vector in Eq. (2):
u-
[ p . . . v. w. Ca. 8x. By.
where, in addition to the purely gas dynamic case, the components of the magnetic field strength vector occur as normal (Bx) and tangential (By and Bz) to F (Ca is the acoustic speed of sound). We consider the exit boundary with u > 0. The minimum eigenvalue in this case is ,~ = u - af, where af is the largest of the two magnetosonic speeds af and a~ (af > Ca _> as): af,~- ~-
~
+ x/~
+
c~q 4 ~ r p
-37-
x/~
'
(4)
N. K Pogorelov
IBI 2 - n2x + By2 "-t-"n z2 . Let us fill the artificial cell outside the boundary by the parameters at infinity. If the outflow velocity in the center of the cell adjacent to the boundary is superfast magnetosonic, u0 > af0, we need no boundary conditions. If u0 < af0 but u~ > a f t , we introduce a rarefaction wave accelerating the flow to fast magnetosonic speed at the boundary. If u0 < af0 and uo~ < afo~, we can simply solve the Riemann problem between U0 and U~. The eigenvector corresponding to the chosen eigenvalue is (7 is a specific heat ratio)
r-
[1
1)Ca 0, t~By, ~ B z
af aBy, ceBz, ( 7 '
p'
2p
Caas
vf~
c~ - 2 P v ' T - ~ ( d a - a~) -
'
lT ,
a~(a~ - C2a)sgnBx --ca(By2 ~ nz )-BSS ,
(5)
c 2a 47r(a 2 -- c 2) fl - p(C2a - a~) = By2 + B2z "
If we seek the solution to (2) in the form of a simple wave U(x, t) = U(() with ~ = x / t , we obtain ( A - AI)Ur = 0,
A=(,
where I is the identity matrix. Thus, Ur is proportional to the right eigenvector of A corresponding to A = ~. If we denote the proportionality factor as d, we can rewrite system (2) in the form
afd ur = - - ~ , v( -- ceByd, - c~Bzd, p (Bx)r - O , (By)r - /3Byd, (Bz)r - /3Bzd, u-af- d,
(Ca) r =
('7 - 1)cad , 2p
(6)
~.
In system (6) we now substitute the equation for Ca by the equation for af, which can be easily obtained from Eq. (4) by direct differentiation with respect to ~ and by using Eqs. (6), yielding
tgd (af)( = 7 '
tg-
Oaf
r.
One can easily notice that the eigenvector (5) degenerates for C a - - a~. This happens if 0 and Ca < aA. It is apparent from Eqs. (6) that the derivatives ( B y ) ( and (Bz)r become infinite at degeneration points. One can easily obtain, however, that
By2 + B z2 -
-- ca)by,zpr
,
where by,z - By,z/(B2y + B~) 1/2. The right-hand side of this relation never degenerates if we assume by - bz - 1/x/2 for By2 + B z2 - 0 (see, e. g., [22]). If we take into account the relation
18xl a~af-
Ca2v/_~ ,
the equations for the derivatives of the transverse components of the velocity vector can also be transformed to the form suitable for further application,
vr
Bx ~47rpa~(BY)~'
we-
Bx ~(Bz)r
-38-
M H D modeling of the outer heliosphere: Numerical aspects
Thus, by passing from d to PC, we obtain the reduced system of equations that can be approximated, for example, by finite differencing 0er = (fr -f0)/A~). The system becomes:
(af)r
-
(
u0 + a~
O+af
(Ca)F--(Ca)0+(q/--1)
)
,
0
Pr
-
af)
P0 1 -~ 0 + af 0
vr - vo +
()
Bx 47rpaf
o (Byr - Byo),
(8)
i xl -i xlo,
\~p/
0 (B2)r _ B~~ + 87r[(a 2 _ Ca)by]O(DF2 2 Do),
,
(B2)F --
Wr - Wo +
2 + Srr[(a2 nzo 47rpaf
-
2 ca)b2]o(fiF -- rio),
o
3. NONEVOLUTIONARY MHD SHOCKS 3.1. Concept of evolutionarity The important subject of discussion is related to the fact that certain initial- and boundary-value problems can be solved nonuniquely using different shocks or different combinations of shocks, whereas physically one would expect only unique solutions. The situation in MHD differs from that in pure gas dynamics of perfect gas, where all entropy-increasing solutions are evolutionary and physically admissible. The term "evolutionary" means that the necessary conditions of well-posedness for the linearized problem of the shock interaction with small disturbances are satisfied. In MHD, the condition of entropy increase is necessary, but not sufficient for the shock to be admissible. Only slow and fast MHD shocks were found to be evolutionary, while intermediate (or improper slow) shocks were not and therefore must be excluded in ideal MHD. All MHD shocks are plane-polarized, that is, the magnetic field and the shock normal vectors lie in the same plane both ahead and behind the shock. In contrast to evolutionary shocks, the tangential component of the magnetic field behind a nonevolutionary shock acquires the direction opposite to what it had ahead of the shock. Thus, rotations of the magnetic field vector are possible only on two types of discontinuities" Alfv6n ones and nonevolutionary shocks. If the statement of the problem rules out Alfv6n discontinuities, nonevolutionary shocks can fairly easily occur in the numerical solution [5]. There is a simple rule to determine whether the shock is evolutionary. Fast MHD shocks are always super-Alfv6nic, that is, the magnitude of the velocity component normal to the shock is larger than the Alfv6n velocity a A - - [B~l/2x/-~ both ahead of the shock and behind it. Slow shocks are sub-Alfv6nic. Nonevolutionary shocks are trans-Alfv6nic. 3.2. Nonevolutionary shocks in the SW-LISM interaction If the magnetic field vector is parallel to the shock normal, we have a parallel shock. The behaviour of parallel shocks depends on whether the Alfv6n velocity ahead of them is smaller or larger than the acoustic speed of sound. In the former case, slow MHD shocks do not exist, while fast shocks are evolutionary and admissible for all their intensities. In the latter case, on the contrary, slow shocks are admissible for all their intensities, while fast shocks are admissible only in a certain range of parameters ahead of the shock, even if they correspond to a superAlfv6nic flow. Singular shocks are those for which the tangential component of magnetic field is equal to zero ahead of (behind) the shock and not zero behind (ahead of) it. Such shocks are called switch-on (switch-off), as the tangential component of the magnetic field vector is
-39-
N. V. Pogorelov
switched on (off) at them. Switch-on shocks are always fast while switch-off shocks are always slow, since the tangential component of the magnetic field always increases across fast and decreases across slow MHD shocks. Let us consider what happens if we increase the value of the LISM magnetic field Boo with the rest of the LISM quantities being fixed. We have two dimensionless parameters relating the quantities ahead of the shock, namely, the Mach number M ~ = V~/Caoo and the Alfv6n number A ~ = Voo/aaoo. If C~o~ > aao~ for Mo~ > 1, the forward point of the bow shock corresponds to a fast parallel shock which is always realizable. If we further increase Bo~, sooner or later a a ~ will become larger than Caoo with A~ > 1. In this case the parallel shock, though remaining fast, will be still evolutionary until Boo acquires the value corresponding to the interval
( 7 + 1)M 2
1
(9)
2 + ( 7 - 1 ) M 2"
For A~ from the interval (9), the Alfv6n number behind the shock is smaller than 1, thus resulting in a trans-Alfv6nic shock which is not evolutionary. Occasionally, a singular (fast switch-on) shock becomes admissible exactly in this range of A~. On switch-on shocks B [I n ahead of the shock but B ~/n behind it. That is, a tangential component of the magnetic field must appear at the forward point of the bow shock. On the other hand, this cannot occur due to geometrical reasons, such as the axial symmetry of the flow due to the condition Boo [[ V~. It could be fairly easily expected that the structure of the flow for the mentioned values of Boo must be different from that in the regular case. Let us choose such a strength of the magnetic field Boo for which A~ - x/2, that is, we are within the interval (9). This corresponds to B~ ~ 2.3 #G. The heliosphere pattern is shown in Figs. 1 and 3. In order to see a fine structure of the flow in Fig. 3b we show only the density lines between 1 and 2.6 with the increment 0.04. We see the origin of an additional shock and a high-density layer near the surface of the heliopause. Addition of rotational perturbations makes almost all shocks evolutionary [28], except for the vicinity of the axis, where the obtained configuration of the bow shock can be interpreted as merged switch-on and switch-off shocks. This combination, however, is also known to be structurally unstable [21 ]. As noted on another occasion in [ 13], the obtained solution will be metastable if we introduce a small angle between Voo and B~. The whole pattern rotates under this condition, as shown in Fig. 2. In this case nonevolutionary shocks are expected to appear in the vicinity of the point where the bow shock normal is parallel to the ISMF vector. Due to the absence of axial symmetry, an irregular interaction pattern with the presence of an additional shock can be realized only on one side of this point. This is now due to the presence of the symmetry plane (the mutual plane of Boo and V~). If we introduce a perturbation of the LISM flow which takes the ISMF vector out of this plane, the initial nonevolutionary pattern will be destroyed. It may appear, however, later on in the new symmetry plane. Since various perturbations can be found in the LISM, the resulting solution will be highly unsteady and, due to perturbations of the symmetry plane, nonevolutionary shocks will exist only as time-dependent shock-like structures.
-40-
MHD modeling of the outer heliosphere: Numerical aspects
Figure 3. (a) Streamlines (below the symmetry axis) and magnetic field lines and (b) density (below the symmetry axis) and thermal pressure isolines. 4. CONCLUSIONS AND DISCUSSION Let us try to understand what we might expect if the ISMF magnitude increases. First, assume that ISMF is parallel to the LISM velocity and that parameters for the solar and interstellar winds are adopted as denoted in Section 2.1. Since the Mach number of the LISM flow is greater than 1, the bow shock, if exists, is always fast. If we increase Br soon we obtain a magnetic field dominated flow in which the Alfv6n velocity is greater than the acoustic speed of sound. If Bo~ > 2.152 #G, we reach the irregular region of interaction, and structurally unstable patterns are to be expected. If B~ > 3.253 #G, the Alfv6n number becomes less than 1 and the bow shock disappears. If ISMF is perpendicular to the LISM velocity, from Eq. (4) it is easy to determine that the bow shock disappears for B~ > 2.4 #G. Numerical simulation of hyperbolic problems characterized by the presence of the bow shock can be performed on the basis of the described numerical methods. Otherwise, the stationary problem becomes elliptic and other numerical methods must be developed for its solution. It is clear that the most up-to-date task is to perform regular MHD calculations accompanied by the Monte Carlo treatment of neutrals. In three space dimensions this will require tremendous computer resources. For this reason the application of simple and robust MHD codes is highly favourable.
REFERENCES 1. D.B. Aleksashov, V. B. Baranov, E. V. Barsky, and A. V. Myasnikov, Astron. Lett. 26 (2000) 743. 2. N. Aslan, Computational Investigations of Ideal MHD Plasmas with Discontinuities, Ph.D. Thesis, Nuclear Eng. Dept., University of Michigan, 1993. 3. V.B. Baranov and N.A. Zaitsev, Astron. Astrophys. 304 (1995) 631. 4. V.B. Baranov, K.V. Krasnobaev, A.G. Kulikovskii, Soviet Physics Doklady 15 (1971) 791. 5. A.A. Barmin, A.G. Kulikovskii and N.V. Pogorelov, J. Comput. Phys. 126 (1996) 77. 6. J.B. Bell, P. Colella and J.A. Trangenstein, J. Comput. Phys. 82 (1989) 362.
-41 -
N. V. Pogorelov
7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.
J.U. Brackbill and D.C. Barnes, J. Comput. Phys. 35 (1980) 426. P. Cargo and G. Gallice, C.R. Acad. Sci. Paris 320 (1995), Serie 1, 1269. P. Colella and P. R. Woodward, J. Comput. Phys. 54 (1984), 174. G. Gloeker, L.A. Fisk and J. Geiss, Nature 386 (1997) 374. W. Dai and P.R. Woodward, J. Comput. Phys. 115 (1994) 485. W. Dai and P.R. Woodward, Astrophys. J. 494 (1998) 317. H. De Sterk and S.Poedts, Astron. Astrophys. 343 (1999) 641. H.J. Fahr, S. Grzedzielski and R. Ratkiewicz, Annales Geophysicae 6 (1988) 337. P.C. Frisch, in Physics of the Outer Heliosphere, COSPAR Colloquia Ser. 1 (1991) 19. P.C. Frisch, Space Sci. Rev. 78 (1996) 213. Y. Fujimoto and T. Matsuda, Preprint KUGD91-2, Dept. Eng., Kyoto University, 1991. S.K. Godunov, Math. Sbomik 47 (1959) 271. V.V. Gogosov, J. Appl. Math. Mech. 25 (1961) 148. T. Hanawa, Y. Nakajima and K. Kobuta, Preprint DPNU-94-34, Dept. of Physics, Nagoya University, 1994. A. Jeffrey and T. Taniuti, Nonlinear Wave Propagation, Academic Press, New York, 1964. A.G. Kulikovskii, N.V. Pogorelov, A.Y. Semenov, Mathematical Aspects of Numerical Solution of Hyperbolic Systems, Chapman & Hall/CRC Press, London, Boca Raton, 2001. T. Linde et al., J. Geophys. Res. 103 (1998) A1889. P. Londrillo and L. Del Zanna, Astrophys. J. 530 (2000) 508. T. Matsuda and Y. Fujimoto, Proc. 5th Int. Symp. on Comput. Fluid. Dyn., Sendai, 2 (1993) 186. R.L. McNutt, Jr., et al. Solar Wind Nine, CP471, The American Institute of Physics, 823, 1999. N.V. Pogorelov and T. Matsuda, J. Geophys. Res. 103 (1998) A237. N.V. Pogorelov and T. Matsuda, Astron. Astrophys. 354 (2000) 697. N.V. Pogorelov and A.Yu. Semenov, Numerical Methods in Engineering 96, 1022, John Wiley, Chichester, 1996. N.V. Pogorelov and A.Yu. Semenov, J. Russian Acad. Sci. Physics-Doklady 42 (1997) 37. N.V. Pogorelov and A.Yu. Semenov, Astron. Astrophys. 321 (1997) 330. N.V. Pogorelov et al. Papers 6th Int. Symp. on Comput. Fluid Dyn., Lake Tahoe, 2 (1995) 952. K.G. Powell, ICASE Rept. No. 94-24, ICASE NASA Langley Research Center, Hampton, VA, 1994. K.G. Powell et al., AIAA 12th Comput. Fluid Dyn Conf., San Diego, CA, June 19-22 (1995) 661. K.G. Powell et al., J. Comput. Phys. 154 (1999) 284. R. Ratkiewicz et al., Astron. Astrophys. 335 (1998) 363. W.K. Rice and G.P. Zank, J. Geophys. Res. 104 (1999) A12563. P.L. Roe, J. Comput. Phys. 43 (1981) 357. D. Ryu, T.W. Jones and A. Frank, Astrophys. J. 452 (1995) 785. D. Ryu, E Miniati, T.W. Jones and A. Frank, Astrophys. J. 509 (1998) 244. T. Tanaka, J. Comput. Phys. 111 (1994) 381. T. Tanaka and H. Washimi, J. Geophys. Res. 104 (1999) A12605. E.E Toro, Riemann Solvers and Numerical Methods for Fluid Dynamics, Springer, Berlin, 1997. C. Wang, J.D. Richardson and J.T. Gosling, J. Geophys. Res. 105 (2000) A2337. H. Washimi, Adv. Space Res. 13(6) (1993) 227. H. Washimi and T. Tanaka, Space Sci. Rev. 78 (1996) 85. A. Zachari and P. Colella, J. Comput. Phys. 99 (1992) 341. A. Zachari, A. Malagoli and P. Colella, SIAM J. Sci. Comput. 15 (1994) 263.
-42 -
Interaction of the local interstellar m e d i u m with the heliosphere: role of the interior and exterior magnetic fields Aaron Barnes Space Sciences Division, NASA Ames Research Center, 245-3, Moffett Field, CA 94035-1000, USA
This paper reviews the role that the exterior magnetic field (of interstellar origin) and the interior magnetic field (of solar origin) may play in the heliosphere-LISM interaction.
1. INTRODUCTION When the solar wind reaches some distance where the dynamical pressure of the flow is comparable to pressure that is characteristic of the local interstellar medium (LISM), the flow must adjust to the exterior obstacle and pass through a termination shock. The post-shock heliospheric flow is subsonic, and forms a region known as the heliosheath. The heliosheath is confined by the LISM, and swept back by the exterior flow to form a long tail-like region; the boundary between the heliosheath and LISM is called the heliopause. The LISM itself is flowing past the solar region at a speed of about 26 km/s (probably supersonic), so that most likely the ionized component of the LISM passes through a bow shock as it is deflected around the heliosphere. For a valuable general review of conditions in the local interstellar medium, the reader is referred to Frisch/1/. Table 1 summarizes our current best estimates of interstellar conditions very near the Sun. The neutral and ionized hydrogen densities (Nn)is and (Nn+)~s, the temperature (T)~s, and the flow velocity (V)is (including the direction) are fairly well constrained by observations./2,3,4,5,6,7/. Although there is some observational information about the magnitude and direction large-scale galactic magnetic field in the Sun's neighborhood/8,9/, the local environment may be extremely inhomogeneous and we can say nothing with certainty about the magnitude or direction of the field just outside the heliosphere. Based on what is known about the general galactic field, (B)is--1-3 #G is a plausible guess for the magnitude, and the direction could be anything. Table 1 also shows the external pressures that follow from the values of the observed (or, in the case of the magnetic field, guessed) parameters for the LISM. Taken at face value, these pressures suggest that the location of the termination shock is somewhere in the range 70-150 AU from the Sun, and that the neutral hydrogen component of the LISM has the greatest ef-
-43 -
A. Barnes
Table 1 Parameters for local LISM Observed Parameters
Derived Pressures
(NH)IS
0.10-0.24 cm -3
(PHV2)IS
--1.2-2.8x10 -12 dyne cm -2
(Nn+)is
0.04-0.15 cm -3
(Pn+W2)is
--0.5-1.7•
(T)xs
~7x10 3 K
(NHkT)I S
-~1-2.3x 10-13 dyne
(V)~s
26 km/s
(2NH+kT)Is
-~0.8-2.9•
-13 dyne cm -2
(B)is*
--1-3 #G
(B2[8 ~)is
--0.4-3.6•
-13 dyne cm -2
-12 dyne cm -2 c m -2
*The field direction is unknown. fect in confining the heliosphere. However, the contribution of the ionized component may be appreciable, and other pressures, especially the magnetic pressure, may not be entirely negligible. Moreover, even if the confining pressure of the interstellar magnetic field is modest, its direction can substantially affect the configuration of the heliosphere, especially the location and orientation of the heliospheric discontinuities, as will be discussed later. The densities given in Table 1 correspond to mean free paths of order 100-1000 AU for neutral-neutral and neutral-ion collisions; this may be compared to a mean free path of order 1 AU for ion-ion collisions and an ion gyroradius of order 103 km for any plausible value of the interstellar magnetic field. It may be argued that over regions large in comparison with the heliosphere, to a reasonable approximation, the ionized interstellar medium moves with the neutral component and is in temperature equilibrium with it. On the other hand, although the two components may be strongly coupled for length scales larger than --1000 AU, on the scale of the heliospheric interaction it is probably a fair approximation to regard the interstellar magnetic field as frozen into the ionized interstellar material, which in turn can flow unimpeded through the neutral component. Nevertheless, the neutral component is expected to interact strongly with ionized material inside the heliosphere/10,11,12,13,14,15,16,17/.
2. EFFECTS OF THE INTERIOR HELIOSPHERIC MAGNETIC FIELD As the solar wind flows outward it carries with it magnetic field lines whose footpoints are rooted in the Sun. This field is generally relatively weak.; for heliocentric distances beyond 23 AU the ratio of magnetic to dynamic pressure is typically -- a few times 10-3, so that the field has negligible effect on the unshocked solar wind. During most of the sunspot cycle the geometry of the heliospheric field is fairly simple, comprising two quasi-hemispheres of op-
-44 -
Interaction of the local interstellar medium with the heliosphere:...
positely directed magnetic polarity. The separatrix of polarity is generally tilted with respect to the solar equator; due to the Sun's rotation a periodic alteration of polarity is seen by a stationary observer at low heliographic latitude. As the solar wind crosses the termination shock, the magnetic field will be compressed in direct proportion to the density, because the field is 'frozen in' to the plasma, If the termination shock is a magnetohydrodynamic (or gasdynamic) shock the maximum possible compression is a factor 4. Thus the ratio of magnetic to dynamic pressure in the shocked plasma is at most of order --10%, too small to be of dynamical importance in the immediate post-shock flow. However, it should be noted that the compression ratio may be substantially larger if the shock is non-magnetohydrodynamic (e.g., if it gives up a substantial fraction of the upstream energy into acceleration of the anomalous cosmic rays), and the consequent amplified magnetic field could be dynamically significant/18,19/. Even if the termination shock is magnetohydrodynamic, and therefore the magnetic field is only a minor participant in the immediate post-shock flow, the magnitude of the field farther downstream can be amplified as a natural consequence of the flow process (the CranfillAxford effect/20/). Kinematic models/21,22/predicted that this mechanism would lead to the formation of a "ridge" of magnetic pressure strong enough to be dynamically important in the heliosheath. Fully self-consistent MHD models/23,24,25/seem to confirm this prediction. In particular, in the flow on the heliospheric flanks, the magnetic ridge appears as a separator between materials originating at polar and equatorial solar latitudes/21,22/. In steady MHD models the interior magnetic field introduces a formal asymmetry to the problem, owing to the fact that the solar magnetic field is oppositely directed in the northern and southern hemispheres. It can happen that the interior and external fields near the heliopause are nearly parallel in one hemisphere, and nearly antiparallel in the opposite hemisphere. The resulting asymmetry of the flow patterns is demonstrated clearly in Plates 5 through 10 of the paper of Linde et al./24/. However, this asymmetry requires precise alignment of the heliospheric current sheet with the solar equator, a condition unlikely to arise in practice. In fact, the heliospheric current sheet is generally tilted relative to the equator, and, as Suess and coworkers have pointed out/21,22/, field lines near the heliopause are expected to originate at low heliographic latitudes. Thus, at any given point near the heliopause the field will exhibit alternating polarity due to solar rotation. The solar rotation period is short in comparison with time scales required for establishment of a steady state for the heliosphere, so that the global asymmetry due to the heliospheric field, as found in the models of Linde et al/24/, will not occur in the real heliosphere. Rather, the heliospheric magnetic field near the heliopause will exhibit complex time variability, possibly associated with reconnection and a leaky, diffuse heliopause. Altogether, the interior heliospheric magnetic field may play a role in organizing heliosheath flow, especially near the nose region, and possibly lead to magnetic reconnection and a locally leaky, diffuse heliopause. Apart from these features, however, the interior field probably does not affect the global structure of the heliosphere in a major way.
-45 -
A. Barnes
Table 2 3-D MHD Models of the Outer Heliosphere Model Neutral gas? Ionized gas? Heliospheric B?
Interstellar B?
(B)is oblique?
/23/
No
Yes
Yes
Yes
No
/24/
Yes
Yes
Yes
Yes
No
/39/
No
Yes
No
Yes
Yes
/40/
No
Yes
No
Yes
Yes
/25/
Yes
Yes
Yes
Yes
Yes
/41/
No
Yes
No
Yes
Yes
3. E F F E C T S OF T H E L O C A L I N T E R S T E L L A R M A G N E T I C F I E L D A substantial number of theoretical investigations have focused on various aspects of the interaction between the heliosphere and LISM related to the global magnetic fields, at differing levels of approximation/21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39, 40,41/. All of these, of course, build upon earlier analyses starting with the work of Parker /42/. An excellent thorough review of the subject has recently been given by Zank/17/; valuable earlier reviews have been given by Axford/20/and Holzer/43/. Fully self-consistent three-dimensional steady-state MHD simulations of the global heliosphere have been published by Washimi and Tanaka/23/, Linde et al./24/, Pogorelov and Matsuda /3 9/, Ratkiewicz et al./40/, McNutt et al./25/, and Ratkiewicz et al./41/. Some of these models allow for the neutral LISM and/or internal heliospheric magnetic field, others do not (see Table 2); however, all take account of the interstellar plasma and magnetic field as characterized by parameters consistent with Table 1. Intercomparison of the models shows an overall qualitative consistency, although, because of their differing assumptions, it is not useful in most cases to make detailed quantitative comparisons. In all cases deflection of the interstellar plasma and field involves an interstellar bow shock (the flow is super-Alfv6nic for all models); magnetic field lines in the shocked interstellar plasma are draped around the heliopause. The combined effects of dynamic and magnetic field pressures compress the heliosphere, as well. However, the neutral component of the LISM may be more important for determining the scale of the heliosphere. For example, of all the self-consistent MHD models, that of Linde et a l . / 2 4 / h a s by far the largest external pressure, due primarily to the flow of the interstellar medium (2/3 neutral, 1/3 ionized). The heliosphere is smaller in this model than in the others (this difference is probably due mostly to the large total external pressure rather than to the
-46-
Interaction of the local interstellar medium with the heliosphere:...
Figure 1. Stream lines comparing (a) aligned
(B~sllVis) and (b) transverse (BIs.J_VIs) flow from/41/.
The
models are based on interstellar and solar wind parameters that are identical except for the direction of the interstellar magnetic field: (nH,)XS=0.04 cm 3, (V)~s=26 km/s, (T)~s=7000 K, (B)is = 2 ~tG, (nH§
cm -3 at 1 AU, (V)sw=400 km/s. The two panels are plotted to the same spatial scale; the
interstellar wind flows from-x, and the interstellar magnetic field lies in the x-y plane.
-47-
A. Barnes
inclusion of neutrals per se). Another distinctive feature of this model is that its termination shock is approximately spheroidal in shape, with its center displaced downwind relative to the interstellar flow. This feature is most probably a signature of the interaction of neutral gas with the solar wind, because models without neutral gas give a more 'bullet-shaped' figure for the termination shock. The direction of the interstellar field can have an appreciable effect on the configuration of the solar wind-LISM interaction. First of all, the morphology of the heliosphere-LISM interaction depends sensitively on whether the magnetic field in the interstellar flow is quasialigned (a-Z(V~s,B~s)-0) or quasi-transverse (a-re/2). Figure 1 shows a comparison of the heliospheric configurations for exactly aligned and exactly transverse flows (from/41/), all other solar wind and interstellar parameters being identical (see the figure caption for details). The two panels are drawn to the same spatial scale. The heliopause and termination shock are substantially smaller in the transverse case, owing to the compression of the magnetic field in the nose region (cf. plates 2 and 3 of/41/). Figure 1 shows only streamlines in the plane containing Bis and VIs; for the transverse case there is an asymmetry, i.e., the heliosphere is flattened toward the Bis-Vis plane (e.g., see figure 4 of Ratkiewicz et al./41/). In the example given here the interstellar field provides a substantial fraction of the external confining pressure (Alfv6nic Mach number = 1.2). A reviewer of this paper has pointed out that this choice of parameters, for the field-aligned case, corresponds to the irregular interaction regime described by Pogorelov and Matsuda/441, so that the detailed fine structure of the nose region given in figure la may not be completely resolved. However, the overall morphology is correct, and figure 1 properly illustrates the contrast between quasi-aligned vs. quasi-transverse flow. In fact, the question of quasi-aligned vs. quasi-transverse flow is important even when the magnetic energy is relatively small. For example, the calculations of Linde et al./24/ showed that variation of the interstellar field direction could shift the position of the termination shock by over 10%. Additional asymmetry is introduced if the interstellar field is oblique, i.e., if the angle a=Z(Vis,BIs ) is different from 0 or rt/2. This situation has been investigated in detail by Pogorelov and Matsuda /39/ and Ratkiewicz et al./40,41/. These models exclude the effects of the LISM neutral component, and therefore leave out a process of substantial importance, but do have the advantage of exhibiting magnetic effects in isolation. Also, by good fortune, the independently calculated models can be compared in substantial detail, because they use quite similar parameters to characterize the solar wind and LISM (Table 3). In both models, the shape and structure of the heliospheric boundary region were calculated for various inclination angles a, ranging from 0 to rt/2. The results are in good agreement and show an asymmetry of the interface region. The noses of the heliopause and solar-wind termination shock are shifted away from the upwind direction, so that the direction toward the nose is roughly transverse to the direction of the interstellar magnetic field. The nose of the interstellar bow shock is shifted in the opposite sense. This asymmetry increases with increasing inclination
-48-
Interaction of the local interstellar medium with the heliosphere:...
Table 3 Interior and exterior parameters for two independent, comparable pure MHD models Pogorelov & Matsuda/39/
Ratkiewicz et al./40/
(nn+)~s, c m -3
0.07
0.1
B is, laG
1.5
0.8 to 2.5
V is, km/s T is, K
25 5.7•
26 9.6•
(nn+)sw, c m -3 (1 AU)
7
10
V~w, km/s
450
400
angle a, until it reaches a maximum for some value of a (which depends on the interstellar field strength), and then declines for larger c~, achieving symmetry for a =rt/2. A detailed discussion of the physics associated with this asymmetry, including extensive graphics, is presented in/41/. One final remark is that, at least for situations in which the contribution of the interstellar field to the exterior confining pressure is non-negligible, for most values of a. the heliospheric configuration is more nearly like the case of transverse flow than aligned flow. As an illustration of the point, we may note that figure 2 of/41/shows a quasi-transverse configuration even for a=zr/6. Given that the actual direction of Bis is unknown, the a priori probability that a
4. I M P O R T A N C E OF TIME VARIATIONS Disturbances in the heliosphere will propagate at some characteristic speed Vcharwhose order of magnitude may be estimated as either the flow speed of the unshocked solar wind, or the heliosheath sound speed; Vchar"100 AU/year-480 km/s is a reasonable estimate. Because the characteristic dimensions of heliospheric structures (apart from sharp boundaries) lie in the range from tens to many hundreds of AU, the time scale for readjustments of the heliosphere ranges from months, for the nose region, say, to many years for the entire heliosphere. On the other hand, the solar wind and heliospheric magnetic field exhibit temporal variations on scales ranging from less than the solar rotation time to the time scale of the sunspot cycle, so that the global heliosphere can never be in a truly steady state. In particular, the concept of
-49-
A. Barnes
a quasi-steady heliosphere that expands and contracts with the course of the sunspot cycle is
invalid. The termination shock is unlikely ever to be in equilibrium, and will move inward and outward in response to solar wind variations/18,19,32,45,46,47,48,49,50,51,52,53,54,55,56, 57,58/, and, perhaps, in response to variations in the local interstellar medium/59,60/. Solar wind disturbances generally affect limited ranges of heliographic latitude and longitude, so that at a given moment some portions of the termination shock will move outward while others simultaneously move inward/61/. Abrupt solar wind disturbances will reach the termination shock, propagate outward to the heliopause where they will be reflected, then propagate backward to the termination shock to be reflected again, etc., possibly resulting in a quasiresonant "ringing" of the outer heliospheric boundaries/55/. Considerations of this kind serve as a warning to use care in drawing inferences from the results of steady-state models such as those discussed in this paper. A case in point arises in studies of the role of the (interior) heliospheric magnetic field. As discussed above, field lines near the heliopause are expected to originate at low heliographic latitudes, and will alternate polarity as the Sun rotates/21,22/. Therefore the heliospheric magnetic field near the heliopause will exhibit complex time variability, possibly associated with reconnection and a leaky, diffuse heliopause.
5. SUMMARY This review has focused on the possible roles that the heliospheric and interstellar magnetic fields may play in the heliosphere-LISM interaction. The field in the unshocked solar wind is weak and does not strongly influence the size or shape of the heliosphere. However, the heliosheath field may be of local dynamic importance near the nose of the heliosphere. Near the heliopause the heliospheric field will exhibit alternation of polarity, due to solar rotation, possibly producing intermittent reconnection events. The dominant external pressures of the heliosphere are probably due to the ram pressure of the neutral and ionized components of the LISM. The magnitude and direction of the interstellar field are not known, and therefore we are uncertain about the relative importance of the magnetic field. However, the direction of the interstellar field can exert substantial influence on the structure of the heliosphere, even if the magnetic pressure is somewhat smaller than the ram pressures. For quasi-aligned flow the shock near the nose is essentially a gasdynamic shock, and the exterior magnetic field is convected passively around the heliosphere. In contrast, for quasi-transverse flow the field is highly compressed at the nose of the shock, and actively compresses the heliosphere. Other things being equal, quasi-transverse flow leads to (1) a substantially smaller heliosheath, (2) a termination shock that is smaller and more "bulletshaped" and (3) a weaker, more distant interstellar bow shock. Even for small values of
-50-
Interaction of the local interstellar medium with the heliosphere:... c~-Z(VIs,B~s), the configuration of the heliosphere is much closer to the quasi-transverse than quasi-aligned shape. Oblique alignment of the interstellar field ( a r 0, 7r/2) leads to asymmetries of the heliosphere and shocked interstellar medium. The noses of the heliopause and termination shock are shifted away from the (interstellar) upwind direction, so that the direction to the nose is roughly transverse to the interstellar magnetic field direction. The nose of the interstellar bow shock is shifted in the opposite sense. Finally, it should be kept in mind that time variations in the solar wind plasma and in the heliospheric magnetic field polarity are of essential importance in the global heliosphericLISM interaction.
ACKNOWLEDGMENTS The perspective that I have presented here has evolved and been nurtured in the context of collaborations with R. Ratkiewicz and the late J.R. Spreiter. I also wish to acknowledge helpful discussions and insights from W.I. Axford, V.B. Baranov, H.J. Fahr, R.L. McNutt, E.N. Parker, N.V. Pogorelov, S.T. Suess, and G.P. Zank.
REFERENCES 1. P.C. Frisch, Space Sci. Rev., 72 (1995) 499. 2. M. Witte, H. Rosenbauer, M. Banaszkiewicz, and H.-J. Fahr, Adv. Space Res., 13(6) (1993) 121. 3. G. Gloeckler, Space Sci. Rev., 78 (1996) 335. 4. E. Moebius, Space Sci. Rev., 78 (1996) 375. 5 M. Witte, M. Banaszkiewicz, and H. Rosenbauer, Space Sci. Rev., 78 (1996) 289. 6. R. Lallement, in The Local Bubble and Beyond, D. Breitschwerdt, M.J. Freyberg and J. Truemper (eds.), p. 19, Springer-Verlag. New York, 1998. 7, O. Puyoo, and L. Ben Jaffel, in The Local Bubble and Beyond, D. Breitschwerdt, M.J. Freyberg and J. Truemper (eds.), p. 29, Springer-Verlag. New York, 1998. 8. J. Tinbergen, Astron. Astrophys., 105 (1982) 53. 9. C. Heiles, in The Local Bubble and Beyond, D. Breitschwerdt, M.J. Freyberg and J. Truemper (eds.), p. 229, Springer-Verlag. New York, 1998. 10. V.B. Baranov and Y.G. Malama, J. Geophys. Res., 98 (1993) 15,157. 11. V.B. Baranov and Y.G. Malama, J. Geophys. Res., 100 (1995) 14,755. 12. G. P. Zank, H.L. Pauls, L.L. Williams, and D.T. Hall, J. Geophys. Res., 101 (1996) 21,639. 13. V.B. Baranov, V.V. Izmodenov, and Y.G. Malama, J. Geophys. Res., 103 (1998) 9575. 14. R.L. McNutt, Jr., J. Lyon, and C.C. Goodrich, Simulation of the heliosphere: Model, J. Geophys. Res., 103 (1998) 1905. 15. V.V. Izmodenov, J. Geiss, R. Lallement, G. Gloeckler, V.B. Baranov, and Y.G. Malama, J. Geophys. Res., 104 (1999) 4731. 16. R.L. McNutt, Jr., J. Lyon, and C.C. Goodrich, J. Geophys. Res., 104 (1999) 14,803. 17. G. P. Zank, Space Sci. Rev., 89 (1999) 413. 18. A. Barnes, J. Geophys. Res., 98 (1993) 15,137. 19. A. Barnes, J. Geophys. Res., 99 (1994) 6553. 20. W.I. Axford, The interaction of the solar wind with the interstellar medium, Solar Wind, NASA SP-308, p. 609, 1972.
-51-
A. Barnes
21. S.T. Suess, and S.F. Nerney, Geophys. Res. Lett., 20 (1993) 329. 22. S. Nerney, S.T. Suess, and E.J. Schmahl, J. Geophys. Res., 98 (1993) 15169. 23. H. Washimi, and T. Tanaka, Space Sci. Rev., 78 (1996) 85. 24. T.J. Linde, T.I. Gombosi, P.L. Roe, K.G. Powell, and D.L. DeZeeuw, J. Geophys. Res., 103 (1998) 1889. 25. R.L. McNutt, J. Lyon, C.C. Goodrich, and M. Wiltberger, in Solar Wind Nine, S.R. Habbal, R. Esser, J.V. Hollweg, and P.A. Isenberg (eds.), p. 823, AIP Press, Woodbury, N.Y., 1999. 26. H.J. Fahr, R. Ratkiewicz and S. Grzedzielski, Adv. Space Res., 6(1) (1986) 389. 27. H.J. Fahr, W.J. Neutsch, S. Grzedzielski, W. Macek, and R. Ratkiewicz-Landowska, Space Sci. Rev., 43 (1986) 329. 28. S.T. Suess, and S.F. Nerney, J. Geophys. Res., 95 (1990) 6403. 29. S.T. Suess, and S.F. Nerney, J. Geophys. Res., 96 (1991) 1883. 30. Y. Fujimoto, and T. Matsuda, KUGD91-2, Dep. of Aeronaut. Eng., Kyoto Univ., Kyoto, Japan, 1991. 31. H. Washimi, Adv. Space Res., 13(6) (1993) 227. 32. K. Naidu and A. Barnes, J. Geophys. Res., 99 (1994) 17,673. 33. V.B. Baranov and N.A. Zaitsev, Astron. Astrophys., 304 (1995) 631. 34. J.U. Brackbill and P.C. Liewer, EOS Trans. AGU, 77(46) Fall Meet. Suppl., SH31A-24, 1996. 35. S.T. Suess, and E.J. Smith, Geophys. Res. Lett., 23 (1996) 3267. 36. V.B. Baranov, A.A. Barmin, and E.A. Pushkar', J. Geophys. Res., 101 (1996) 27,465. 37. S.T. Suess, E.J. Smith, J. Phillips, B.E. Goldstein, and S.F. Nerney, Astron. Astrophys., 316 (1996) 304. 38. N.V. Pogorelov, and A.Yu. Semenov, Astron. Astrophys., 321 (1997) 330. 39. N.V. Pogorelov, and T. Matsuda, J. Geophys. Res. 103 (1998) 237. 40. R. Ratkiewicz, A. Barnes, G.A. Molvik, J.R. Spreiter, S.S. Stahara, M. Vinokur, and S. Venkateswaran, Astron. Astrophys., 335 (1998) 363. 41. R. Ratkiewicz, A. Barnes, and J.R. Spreiter, J. Geophys Res., 105 (2000) 25031. 42. E. N. Parker, Interplanetary Dynamical Processes, Interscience, New York, 1963. 43. T.E. Holzer, Ann. Rev. Astron. Astrophys., 27 (1989) 199. 44. N.V. Pogorelov and T. Matsuda, Astron. Astrophys., 354 (2000) 697. 45. J.W. Belcher, A.J. Lazarus, R.L. McNutt, Jr., and G.S. Gordon, Jr., J. Geophys. Res., 98 (1993) 15,177. 46. S. Grzedzielski and A.J. Lazarus, J. Geophys. Res., 98 (1993) 5551. 47. S.T. Suess, J. Geophys. Res., 98 (1993) 15,147. 48. Y.C. Whang and L.F. Burlaga, J. Geophys. Res., 98 (1993) 15,221. 49. K. Naidu and A. Barnes, J. Geophys. Res., 99 (1994) 11,553. 50. R.S. Steinolfson, J. Geophys. Res., 99 (1994) 13,307. 51. S.R. Karmesin, P.C. Liewer, and J.U. Brackbill, Geophys. Res. Lett., 22 (1995) 1153. 52. T.R. Story and G.P. Zank, J. Geophys. Res., 100 (1995) 9489. 53. T.R. Story and G.P. Zank, J. Geophys. Res., 102 (1997) 17,381. 54. R. Ratkiewicz, A. Barnes, and G.A. Molvik, in Solar Wind Eight, D. Winterhalter, J.T. Gosling, S.R. Habbal, W.S. Kurth, and M Neugebauer (eds.), p. 646, AIP Press, 1996. 55. R. Ratkiewicz, A. Barnes, G.A. Molvik, J.R. Spreiter and S.S. Stahara, J. Geophys. Res., 101 (1996) 27,483. 56. V.B. Baranov and N.A. Zaitsev, Geophys. Res. Lett., 25 (1998) 4051. 57. T. Tanaka and H. Washimi, J. Geophys. Res., (1999) (104) 12,605. 58. N.V. Pogorelov, Astrophys. Space Sci., 274 (2000) 115. 59. R. Ratkiewicz, A. Barnes, and J.R. Spreiter, Geophys. Res. Lett., 24 (1997) 1659. 60. G.P. Zank and P.C. Frisch, Astrophys. J., 518 (1999) 965. 61. A. Barnes, Motion of the heliospheric termination shock at high heliographic latitude, The HighLatitude Heliosphere (R.G. Marsden, ed), p. 233, Kluwer, Dordrecht, 1995.
-52-
Modeling stellar wind interaction with the ISM" Exploring astrospheres and their Lyman-c~ absorption H.-R. Mfiller ~, G. P. Zank ~, and B. E. Wood b* ~Bartol Research Institute, University of Delaware, Newark, DE 19716 bjILA, University of Colorado and NIST, Boulder, CO 80309-0440 A numerical study of the interaction of a partially ionized interstellar medium with a stellar wind is presented. Of particular interest are the distributions of neutral hydrogen in the resulting astrospheres. We study e Ind which has a higher relative velocity with respect to its local ISM than the Sun, and a solar-like stellar wind. A second study considers the astrosphere of a star with a mass-loss rate that is different from that of the Sun (~ And). We calculate the additional absorption of the Ly-c~ lines of c Ind and ,~ And due to their respective astrospheric hydrogen walls and compare it with high-resolution UV spectra obtained by the Hubble Space Telescope (HST). Optimizing the fit between data and models leads to estimates of the otherwise elusive stellar wind parameters. 1. I N T R O D U C T I O N The solar wind plasma and the ionized component of the local interstellar medium (LISM) interact with interstellar neutral hydrogen (H I) to create a neutral density enhancement upstream of the heliosphere, the hydrogen wall. This coupling of H I to the plasma through charge exchange heats and decelerates the neutral H, including the population in the hydrogen wall [1]. Since HST high-resolution UV spectra of nearby stars have become available, it has been recognized that, besides the absorption afforded by the column density of H I in the ISM along the line of sight, the H I population of the hydrogen wall contributes to the absorption of stellar Ly-c~ line profiles in upwind directions, such as those toward c~ Cen [2] and 36 Oph [3]. For downwind lines of sight the H I density is much lower than in the hydrogen wall, but the sightline through the heated heliospheric H I is longer, allowing heliospheric Ly-c~ absorption to be observed downwind as well as upwind [4,5]. As was noted subsequently [4,6,7], the analogous creation of an "astrosphere" and a stellar hydrogen wall stemming from the interaction of a stellar wind with its surrounding ISM is possible for other stars as well. While the heliospheric absorption occurs on the red side of the Ly-c~ profile, the astrospheric H I populations give rise to an additional *This research was supported by NASA grant NAGh-9041 to the University of Colorado, by an NSF travel grant administered by the American Astronomical Society, and by NASA grant NAG5-6469, NSF-DOE award ATM-0078650, JPL contract 959167, and NASA Delaware Space Grant College award NGT5-40024 to the Bartol Research Institute.
-53-
H.-R. Miiller, G.P. Zank and B.E. Wood
Table 1 Model parameters. M/Mo
/~ AND 1 AU 5, 10, 40 LISM e INS 1 AU 0.5, 0.8, 1 LISM
nH [cm -3]
0.14 0.14
np [cm -3] 25, 50, 200 0.10 2.5, 5 0.10
v [km/s] 400 53 400 68
T [K] 105 8000 105 8000
Mach 7.6 3.6 7.6 4.6
0E 89 ~ 64 ~
absorption component to stellar Ly-c~ on the blue side [8]. In view of this, we present a numerical study of the interaction between the stellar winds of two stars and their respective local interstellar media, applying a multifluid model [9]. We choose the G8 IVIII star ,~ AND (d - 26 pc) and the K5 V star e IND (d - 3.63 pc) because of the thorough analysis of high-resolution spectral observations of Ly-c~ presented by [7], which suggests the presence of astrospheric absorption toward both stars. 2. M O D E L P A R A M E T E R S
AND RESULTS
Our numerical study [10] assumes boundary parameters for the ISM environment that are similar to those estimated for the vicinity of the heliosphere (the local interstellar cloud). This explicit assumption will be relaxed in future studies which will vary these parameters as well. Here, we only vary stellar wind parameters since they are even less accessible than the LISM values. The LISM parameters are listed in Table 1 (LISM rows), and include the H I density nil, the plasma density np, the temperature T, and the computed bulk velocities v in the astrocentric frame of reference, which are quite large compared to the LIC value for the Sun (26 km/s). Table 1 also shows the effective plasma Mach number and the direction 0E to Earth in the astrocentric frame. The inner radial boundary condition of the model is a prescribed stellar wind plasma; the values at a reference distance of 1 AU are listed in Table 1. Temperature and velocity are assumed to be identical to those of the average solar wind measured at 1 AU. For the star ,~ AND, we assume a mass-loss rate larger than that of the Sun since ,~ AND is classified as being between a subgiant and giant star. We probe parameter space by assuming three different mass loss rates for both ,~ AND and e IND. The stellar wind expands into the ISM until pressure balance prevents further expansion. The supersonic stellar wind terminates at a termination shock (TS), beyond which the decelerated wind flow, now subsonic and very hot, is diverted towards the astrotail. In both stellar systems, this flow becomes supersonic and is shocked again close to a triple point in the tailward TS that defines a bullet shaped astrosphere. The supersonic LISM plasma shocks at a bow shock (BS). A tangential discontinuity (the astropause AP) between the TS and the BS separates the stellar wind and the ISM plasma. As compared to the heliosphere, both the stellar wind ram pressure and the LISM ram pressure are higher for ~ AND. The stationary ~ AND system with 40/1;/| has roughly the same aspect ratio as the heliosphere [10], with distances scaling by a factor of about 1.5 except for the BS. The TS is found at 145 AU upstream (475 AU downstream), the AP at 195 AU, and the BS at 320 AU upstream. Comparable numbers for models of the heliosphere are 95 AU,
-54-
Modeling stellar wind interaction with the ISM: ...
~,~
~. . . . . . . .
10 s
#
. . . . . . . . .
, . . . . . . . .
~
.
.
.
.
.
.
.
.
.
,
-
,,,i
.
.
.
.
.
.
.
.
.
.
.
.
i
.......
IS,....
, .
.
.
,
10 s
7" ~ 1 r 0+
Figure 1. Log-log profiles along the e IND stagnation axis, showing density n (left) and temperature T (right), both for H I (solid) and H II (dashed). Both plots have two panels, the first covering distances from the tail region to the Sun at 0, the second containing the upstream region from the Sun to the LISM flowing in at 1000 AU.
140 AU, and 310 AU, respectively [9]. At e IND with a solar mass loss rate, the TS is found at 32 AU upstream (140 AU downstream), the AP at 41 AU, and the BS at 63 AU upstream (Figure 1). The astrosphere of e IND has an elongated shape in comparison to the heliosphere, owing to the high LISM velocity of 68 km/s. The associated LISM ram pressure dominates the pressure balance, shifting the boundaries closer to the star, and "wrapping" the bow shock and the hydrogen wall around the astropause. The mean free path of H I inside this small astrosphere is long compared to the size of the astrosphere, which calls for a kinetic treatment of the neutrals. We have employed here a multifluid representation of the neutrals [9], which is computationally less demanding yet captures the overall neutral distribution reasonably well. A hydrogen wall of peak density 0.5 cm -3 forms in the ,~ AND case between BS and AP. In the wall, neutral H is decelerated (,,~15 km/s) and heated (from 50,000 K at the BS to 105 K at the AP) through secondary charge exchange. The e IND system has a hydrogen wall with a peak density of 0.58 cm -3 (more than a fourfold enhancement over the LISM value, Fig. 1). The structure of the astrospheres in the direction towards Earth is quite similar to the upstream direction. The hydrogen wall of e INS has a peak density of 0.45 cm -3 and a temperature of ~70,000 K in this direction (,~ AND: 0.38 cm-3; ~60,000 K). 3.
ABSORPTION
RESULTS
In Figure 2, we present a calculation of the Ly-c~ absorption that includes the ISM and the modeled astrospheric component. The histograms are the blue-side wings of highresolution HST/GHRS spectra of the Ly-a line analyzed by [7]. The absorption feature centered at 1215.37 A (for A AND) is due to interstellar deuterium. The dotted line shows the ISM absorption along the line of sight, as determined by [7]. Clearly, there is excess absorption in the spectrum between 1215.4 ~ and 1215.5 .~ where absorption by the ISM alone is less than observed by HST. In principle, the modeling outlined above allows us to constrain the stellar wind parameters of A AND and c IND by varying the stellar wind plasma density, velocity, and temperature, and identifying which model best fits the HST spectra of Fig. 2. Here, we only vary the density (mass loss rate). The dashed and dot-dashed lines in Fig. 2a represent the line profile after both ISM and astrospheric absorption from three ,~ AND models with mass loss rates 5, 10, and 40 times that of the Sun have been included. The calculation of the line profile with both ISM and 40.g/| ~ AND astrospheric absorption included (triple dot-dashed) results in far too much absorption, especially near 1215.45 .t. The increased opacity in this model compared to
-55-
H.-R. Miiller, G.P. Zank and B.E. Wood . . . . . . . . .
i . . . . . . . . .
i . . . . . . . . .
i . . . . . . . . .
~, A n d ..................
-,
2
-
~,~.
ISM
.................. ISM A b s o r p t i o n
Absorption
i ,,u .......... ............
. . . . . . . . . . . I~I=40 l~Ie
'.,, ,~.
?
~ = 1.0 ~
:~
o
~9
1. i1)
,:"
"...
..'
....
"...
;:
...........
/,,"',,,
~}.
,'
",,, ,
-
.
I
1215.2
. . . . . . . . . 1215.3
1215.4 Wavelength
1215.5
12
-
5.6
1215.20
1215.30
1215.40
1215.50
Wavelength
Figure 2. Blue side of the HST spectrum of c IND (a, le•hand panel) and ,~ AND (b, righthand panel) as histograms, together with modeled Ly-c~ profiles at Earth after accounting for ISM absorption alone (dotted), and ISM + astrospheric absorption derived from the models discussed in the text. From [10].
the 5/~omodel lies in the greatly increased extent of the hydrogen wall between the AP and BS while densities remain comparable. The neutral temperature in the 40/1)/r ,~ AND wall is also higher. The ~ AND mass loss rate model of M - 5/1//o fits the data very well. For the c IND absorption (Fig. 2b), we employ three models with mass loss rates 0.5, 0.8, and 1. Two profiles "bracket" the data on the high and on the low side between 1215.4 ~ and 1215.5.1. leading us to the conclusion that the stellar wind of ~ IND has a mass loss rate between 0.5 and 0.8 solar mass loss rates. The results underscore the importance of the astrospheric neutral H distribution for Lyc~ absorption calculations. They also strongly suggest that our model has captured the basic distribution of neutral densities and velocities as well as the basic, effective neutral temperatures that arise at the investigated stars due to the interaction of the LISM with its stellar wind. The reasonably good fit, in turn, supports the principal results, namely the above described morphology of the astrosphere and the estimate of sensible stellar wind parameters. This is the currently most promising method, albeit indirect, to investigate the morphological structure and characteristics of remote astrospheres [10,11]. REFERENCES .
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
G. P. Zank, Space Sci. Rev. 89 (1999) 413. J. L. Linsky & B. E. Wood, ApJ 463 (1996) 254. B. E. Wood, J. L. Linsky, & G. P. Zank, ApJ 537 (2000) 304. V. V. Izmodenov, R. Lallement, & Y. G. Malama, A&A 342 (1999) L13. L. L. Williams, D. T. Hall, H. L. Pauls, & G. P. Zank, ApJ 476 (1997) 366. K. G. Gayley, G. P. Zank, H. L. Pauls, P. C. Frisch, & D. E. Welty, ApJ 487 (1997) 259. B. E. Wood, W. R. Alexander, & J. L. Linsky, ApJ 470 (1996) 1157. H.-R. Miiller & B. E. Wood, this volume (2001). G. P. Zank, H. L. Pauls, L. L. Williams, &: D. T. Hall, J. Geophys. Res. 101 (1996) 21639. H.-R. Miiller, G. P. Zank, ~ B. E. Wood, ApJ 551 (2001) 495. B. E. Wood, J. L. Linsky, H.-R. Miiller, & G. P. Zank, ApJ 547 (2001) L49.
-55-
Stationary MHD-equilibria of the heliotail flow Dieter Nickeler & Hans Fahr ~ ~Institut fiir Astrophysik und extraterrestrische Forschung Auf dem Hiigel 71, 53121 Bonn, Germany
The solar wind, passing through the termination shock, is declined into the heliospheric tail and leaves the solar system as a subalfv~nic tail flow. In this article we present an analytical method to calculate nonlinear and selfconsistent stationary ideal MHD-equilibria with incompressible, field aligned magnetized plasma flow in the heliotail. In this case within the framework of a one fluid approximation it is possible to reduce the governing MHD-equations of the problem for stationary flow states to static MHDequilibria. In magnetohydrostatics (MHS), assuming quasi-independence with respect to one coordinate (-direction) in the plane perpendicular to the tail axis, the equations reduce then to one single nonlinear, elliptic partial differential equation, the Grad-Shafranovequation (GSE) (see e.g. Grad & Rubin [1]). First we calculate static equilibria to determine the geometry and topology of permitted field configurations and map this onto corresponding stationary equilibria with plasma flow to obtain the structure of the electric currents and the magnetic field.
1. I N T R O D U C T I O N Beyond the solar wind termination shock (TS) the plasma of the solar wind is decelerated and the magnetic field is amplified, so that there exists a subalfv~nic plasma flow in the downstream direction. Scherer et al. [2] showed that for small Mach numbers the bulk flow can be assumed to behave incompressible, which even more holds for field aligned flows, where the field lines are acting as quasi-isothermals. As the decelerated solar wind has to adapt to the outer magnetized VLISM (Very Local InterStellar Medium) conditions (e.g. thermal, ram and magnetic pressure) a contact or tangential discontinuity forms, the heliopause (HP), which is stretched into downwind direction. Between the TS and the HP there exists an inner heliosheath extending into the heliotail, similar to the Earth's magnetotail. In the magnetotail the MHD quantities mainly depend on one direction perpendicular to the tail axis, connected with the theta-shaped current system (e.g. Birn [3]). In view of evident similarities we here apply a similar description to the heliotail.
-57-
D. Nickeler and H.J. Fahr 2. S T A T I O N A R Y S T A T E S I N I D E A L M H D The equations for ideal MHD are the conservation of mass, conservation of momentum with isotropic pressure P, Amp~re's law, and the induction equation reading -
•
o
.o5'
- 5" •
- p
e)
(1)
e •
(2) --+
..+
where p is the mass density, B is the magnetic field, g the velocity field, and j the electric current density vector. For incompressible flows with V. g - 0 we can define a flow vector ~ "- v ~ g" These two conditions imply g. V p - 0 and V . ~ - 0. We also introduce the 1--,2 Bernoulli-pressure II "- P + gw . .-+
2.1. Field aligned flows These stationary equilibria should later be tested with respect to their stability and be used to derive dynamical time scales, which may be important with respect to the VLISM-solar wind interaction. In the inconsistent picture given by Suess & Nerney ([4] and references therein) only a kinematic treatment (frozen-in test field approximation) is used. These authors find strong amplification of convected magnetic fields. In addition they identify a cone of 30 degrees, in which their kinematical approach is invalidated. Such orthogonal velocity fields (with respect to components of the magnetic field) have a saddle point structure in linear stability analyses (see Hameiri [5]) and are therefore not suitable to test stability. From this point of view it may be motivated to make the simplifying assumption of a field aligned flow, i.e 9 g - P xMA /~ ' with automatic fulfillment /-fi~ of the induction equation. With all these assumptions, made in section 2., the Euler equation reads VII-
1 (1- M~)(Vx/3)x/3#o
1 / 3 2 V ( 1 - M A 2)
(3)
2/-to
allowing to conclude, that the density p and the Alfv~n Mach number MA on field lines, however, varying perpendicular to them.
are
constant
2.2. Representation by Euler potentials In 1984 Zwingmann [6] had shown the similarity between MHS-equilibria and stationary MHD-equilibria with incompressible, field aligned flows. Later the theory was improved by Gebhardt & Kiessling [7], and used for modelling sunspot magnetic fields with plasma flow by Petrie & Neukirch [8]. In general, magnetic fields can be represented by using Euler potentials. The magnetic fields of MHS-equilibria used in this case are written as BMHS -- V f • Vg, where f and g are scalar functions (Euler potentials) of x, y, z. The equations below show how the magnetohydrostatic fields are mapped to the stationary fields, by performing the noncanonical transformation f = f(a,/5) and g = g(a, ~) with the Poisson bracket 0 < ([f, g],,Z)2 ._ 1 - M~ (valid for subalfv~nic flows).
BMUS
-58-
(4)
Stationary MHD-equilibria of the heliotailflow PMHS - PMHS(f, g)
~
1
P - PMHS- _~pg2
(5)
with the velocity field g - MA BMHS/V/#op (1 -- M~). In these equations f and g are the static potentials, while c~ and/5 are the stationary potentials. Since the post-shock plasma is very hot (see Jgger & Fahr [9]), we have high values of ~-~2, i.e. of the plasma/5, and of the sound velocity c~2 > v~ > V 2 , therefore we are allowed to use these transformations.
2.3. Two-dimensional equilibria Let f = A(x, y); g = z and the transformation c~ = ~(A);/3 = g, where the GSE is written as AA = - # 0 dPMHs/dA , valid under the assumptions adapted here, then the electric current of the stationary equilibrium is calculated by A a - d--~ + ~-~ -#o jz. In the first step we assume, that AA = 0 (potential field), which means that we can find suitable solutions by using a two-dimensional multipole representation. This enables us to find equilibria, e.g. with magnetic neutral points, to get stagnation points in front of the heliosphere, which can be used to receive a Parker-like configuration of flow trajectories/magnetic field lines. We use an extended form of a Laurent series (more in a forthcoming paper). For this general kind of a conformal mapping we exclude the domain around the singularity at (x, y) = (0, 0 ) , representing the inner region of the heliosphere. Symmetric systems have two kinds of boundaries: the axis of symmetry, here the x axis, for which we assume A = 0 and the other three boundaries for which we assume asymptotic boundary conditions, namely lime_~ A = B~ y, where ~ = ~/x2+ y2 and B ~ e~ is the asymptotic VLISM magnetic field, representing the influence of the VLISM. This ensures the homogeneity of the unperturbed VLISM magnetic field and the VLISM plasma flow. Therefore within a second step we have to use suitable transformations. To get three expected current sheets we use the mapping x
oz(A) - 6' A + ~ ak In cosh k=~
/
(6)
dk
where Yk are the locations of the current sheets, i.e. y~ and ya are the heliopause current sheets, and Y2 could be the continuation of the heliospheric or heliotail current sheet (with finite thickness). The other values are normalization constants (dk = thickness of current sheets, C ensures that doz/dA > 1, guaranteeing that the flow is subalfv~nic). 3. D I S C U S S I O N
AND CONCLUSIONS
In the following figures there are shown some representative MHD-values. In MHS, isocontours of the thermal pressure and of the current density are field lines. In Figs. 1 and 2 we can see that this is violated in the case of MA ~ O. If we calculate symmetric potential fields it is also possible to choose transformations to get equilibria with asymmetric field strengths, although the trajectories are symmetric. It is also possible to model a global heliosplaere with asymmetric trajectories, so the stagnation point can lie off the x-axis. The coefficients in equation (6) could be chosen in this way that different current directions are possible for the different current sheets. Therefore this technique permits a great flexibility to construct MHD-flows for stellar magnetospheres far away from the central star, moving through the VLISM.
-59-
D. Nickeler and H.J. Fahr
Figure 1. Symmetrical field/stream lines. The length scale is 100 AU, so the standoff distance of the heliopause is -1.5 (= 150AU).
Figure 2. Isocontours of the current density for a certain transformation, which corresponds to the symmetric field lines of the mapped MHS-equilibrium. The tail field lines are open, while the isocontours are closed. The sun is located in (x, y) = (0, 0).
A c k n o w l e d g e m e n t : We are grateful for financial support granted by the Deutsche Forschungsgemeinschaft within the project Fa 97/23-2. REFERENCES
1. H. Grad, H. Rubin, Proceedings of the Second United Nations International Conference on the Peaceful Uses of Atomic Energy, Geneva 1958, Vol. 31, p.190 ft. 2. K. Scherer, H. Fahr, R. Ratkiewicz, A&A, 287 (1994), p. 219 3. J. Birn, JGR Vol. 92 (1987), p. 11101 4. S.T. Suess, S. Nerney, JGR, Vol. 100 (1995), p. 3463 5. E. Hameiri, Physics of Plasmas, Vol. 5 (1998), p. 3270 6. W. Zwingmann, PhD-thesis, Ruhr-Universit/it-Bochum (1984), Germany, p. 53 ft. 7. U. Gebhardt, M. Kiessling, Physics of Fluids, B 4(7) (1992), p. 1689 8. G. Petrie, T. Neukirch, Geophys. Astrophys. Fluid Dynamics, Vol. 91 (1999), p. 269 9. S. J/iger, H. Fahr, Solar Physics, Vol. 178 (1998), p. 631
-60-
Non-stationary magnetic field geometry in the heliosphere I.S.Veselovsky a ~Institute of Nuclear Physics, Moscow State University, Moscow 119899, Russia Model examples of the 3D time-dependent magnetic fields with the open, semi-open and closed configurations in the heliosphere are presented. 1.
INTRODUCTION
The magnetic field in the heliosphere is dependent on space and time because of the imposed boundary conditions and internal dynamical processes (see, e.g., reviews by Mariani and Neubauer, 1990; Burlaga, 1991). Quasistationary and transient structures coexist in proportions varying with the solar cycle. The purpose of this paper is to present model examples showing magnetic field lines with different geometric properties in the heliosphere. 2.
MODEL ASSUMPTIONS
Let us consider magnetic fields under the ideal conductivity approximation neglecting displacement currents. In this case, the governing Maxwell equations read as follows
-
o.
(2)
If the velocity field g(~', t) is given (kinematic approximation), we have four scalar equations (1,2) for three magnetic field vector components. This means that the system (1,2) is overdetermined as a rule, and nontrivial solutions for B are possible only under specific additional restrictions. Boundary conditions for the magnetic fields can be not prescribed arbitrary, but they are adjusted to the given velocity field. The remaining Maxwell equations can be used for explicit calculations of the electric charges and currents via (V. E ) a n d (V x B ) . Hence, the set (1,2) with some boundary conditions totally determines the magnetic field structure in space and time when g(~', t) is given. This approach is well known in applications to the heliosphere when using the method of characteristics. -+
-+
-61
-
LS. Veselowsky
3.
EXAMPLES
1) In the case when g(~', t) - const and the velocity has only the radial component v, the solution of equation (1) can be written as follows -
--
Bo~
T
Bo -
--
Boo
r
(?)
B~ -
Boo
( (
0,7),t
,
(3)
V
O, ~, t
,
(4)
.
(5)
v
O, ~ , t v
The magnetic field vector at the initial sphere is not an arbitrary function of its arguments, but obeys the additional condition J~0r
0 (sin OBoo) + r0 sin 0 gO
--
OBoe]
o~
(6)
- O.
This restriction is very important. In particular, for the stationary axially symmetric boundary conditions this means Boo - 0 and Bo - O. The field line equation dr rdT)
Br
:
(7)
B~
can be integrated in a straightforward manner with the result showing a family of integral curves starting at the sphere with radius r0. Each field line belongs to the cone defined by 0 = 00 and goes from r0 to infinity along the conical spiral when Bo~ -r 0 or along the radial direction when B0~ =0. The overall geometry of the field lines could be termed as semi-open. 2) For the boundary ..+ conditions which are stationary in the reference frame rotating with the Sun, one has B0 (0, 7)- f~t) at the initial sphere r0. Here ft is the angular velocity of the rotation. In the case when Boo = 0 one obtains from (6) the relation v
f~Bo~ +
r sin 0
Bo~ - 0
(8)
and the solution (3-5) looks as follows
Br
-
(~0) -- ~ Bo~(O,~-f~t'),
(9)
7"
(10)
Bo - 0 , B~-(t--~
B o ~ ( 0 , 7 ) - F i t ' ) sin--------~0v f~r~ '
(11)
where t' - t - ~-r0. Field lines are represented in this case by Archimedian spirals v r sin 0 + a7) - const on the cone, 0 - const, where a - v/f~(O) (Parker, 1958). The spirals
-62
-
Non-stationary magneticfieM geometry in the heliosphere are tighter near the equatorial plane because of the differential rotation f~(0). The real heliospheric magnetic fields follow the model (9-11) rather well in a broad range of heliospheric distances only under the appropriate time averaging because of the non-stationary boundary conditions on the Sun (Mariani and Neubauer, 1990). The strong violation of the assumption Bo = 0 takes place for instantaneous magnetic fields. The measurements show that Bo :/= 0, and that Bo is comparable to Be as a rule for instantaneous fields, and the more general relations (3-5) are better applicable. 3) Let us consider solution (3-6) near the stationary "neutral points", where B~ = 0 in the rotating reference frame. This situation corresponds to the long-living "neutral point" B0~ = 0 in the solar wind source region at r = r0. In this case, the "neutral point" is "projected" from the sphere r0 along the Archimedian spiral into the heliosphere and forms the "neutral line" there. Hence, isolated stationary neutral points cannot exist in the heliosphere and the manifold of neutral points is organized in neutral lines of a spiral shape. The field B(0, Bo, B~) is tangential to the heliocentric spheres at the mentioned neutral lines. In the simplest case when B0~ = B~ - 0 one obtains a'family of similar open or closed field lines on concentric spherical surfaces disconnected from the Sun and from the infinity with two components Bo ~ r -~, B~ ~ r -~, obeying the condition of a vanishing divergence. 4) The potential part of the solar magnetic field can be represented by the multipols varying with the Hale cycle (Hoeksema et al., 1983). The dominat poloidal (r, 0) contributions on the source surface are produced by the lowest harmonics. In addition to this, non-potential fields are present and the toroidal (~) component is not negligible. The magnetic field on the Sun and in the heliosphere is more structured during years of higher solar activity. Higher poloidal and toroidal harmonics are stronger at this time. The number of non-stationary zero points and lines increases at the source surface and in the heliosphere. Accordingly, closed and loop-like structures are more common at the high solar activity years on the Sun and in the heliosphere. 5) A full set of the nonlinear MHD equations should be used instead of Eqs. (1,2) if the velocity field g(~', t) is not given. The structures with laminar and turbulent states of the magnetic field are known when the overall geometry is open, closed or intermittent one. More observations, especially multi-point and tomographic ones, are needeed for a better understanding of the heliospheric structure. Sometimes, the linear analysis is applicable. In this case, "perturbations" are small and a background state is well defined. It could be given by the stationary or evolving solution. The "eigen mode problem" arises for small perturbations. This can be solved only in the simplest cases exemplified by discrete and/or continuous spectra of "waves" and "convective branches". Practically, explicit solutions for the background state and perturbations are not easy to find. Partially because of these difficulties, we do not have as yet even a "common language" to describe observations. A dimensionless scaling approach is pomising in this respect. 6) The degree of openness of the heliospheric field increases with the distance from the Sun in the simplest model, when the solar formation region geometry is represented by a superposition of a dipole and a current sheet. The ratio of the closed field line areas at a given distance r from the Sun decreases with this distance approximately as
-63-
I.S. Veselowsky \
-
1 + ar ) 1 where a is proportional to the length scale #/(~. Here # is the magnetic dipole of the "su n, (I) is the magnetic flux of the polar coronal holes, which is proportional to the heliospheric current sheet strength. This model is semi-quantitatively correct for the solar minimum years. The solar magnetic dipole strength decreases with the activity cycle, when the current sheet strength increases and the number of open field lines should increase because of this. As a consequence, the outer corona looks more radially stretched during the maxima. An opposite tendency is imposed by other, time dependent factors. The corona is more structured and dynamical at maxima. Coronal mass ejections are more numerous. Radially expanding non-stationary elements bring their local fields which could be both open and closed, as well as intermittent. As a result, the overall heliospheric field topology needs additional investigations especially during the high solar activity. 4. C O N C L U S I O N S Non-stationary magnetic fields in the heliosphere are partially open, semi-open or closed in different proportions dependent on time and location. The instantaneous magnetic field lines at distances about 1 AU on average strongly deviate from conical Archimedian spirals and provide the non-stationary magnetic coupling between different latitudes, longitudes and distances inside the corresponding correlation lengths, which are determined mainly by variable boundary conditions near the Sun for B and g as well as by non-local dynamical processes in the heliosphere. Because of the heliospheric curent sheets, the degree of openness of the field in general increases with the distance from the Sun. The heliospheric magnetic field is more fragmented, dynamical and has a larger number of neutral lines, separators and closed elements connected and disconnected from the Sun during the years of high activity, but the overall topology remains to be investigated. 5. A c k n o w l e d g m e n t s This work was supportedd by the RFBR grant 98-02-17660, the Federal Program "Astronomy" project 1.5.6.2, the Federal Program "Universities of Russia" project 990600 and the INTAS/ESA grant 99-00727. The author is grateful to the Organizing Committee of the COSPAR Colloquium in Potsdam for the financial support facilitating the attendance at the meeting and for the help in preparation of the final manuscript.
1. Mariani, F., and F.M. Neubauer. The Interplanetary Magnetic Field, in Physics of the Inner Heliosphere I, eds. R.Schwenn and E.Marsch, pp.183-206, Springer Verlag, Berlin, 1990. 2. Burlaga, L.F.E., Magnetic Clouds, in Physics of the Inner Heliosphere II, eds. R.Schwenn and E.Marsch, pp.1-22, Springer Verlag, Berlin, 1991. 3. Parker, E.N. Dynamics of the Interplanetary Gas and Magnetic Fields, Astrophys. J., 328, 848-855, 1958. 4. Hoeksema, J.T., J.M.Wilcox, and P.H.Scherrer, The Structure of the Heliospheric Current Sheet: 1978-1982, J. Geophys. Res., 88, 9910-9918, 1983.
-
64
-
Solar cycle heliospheric interface variations" influence of neutralized solar wind N. A. Zaitsev a and V. V. Izmodenov b* aKeldysh Institute of Applied Mathematics, Russian Academy of Sciences, [email protected] bDepartment of Aeromechanics and Gas Dynamics, Faculty of Mechanics and Mathematics, Lomonosov Moscow State University, [email protected] New non-stationary self-consistent model of the solar wind interaction with a twocomponent (atoms and plasma) local interstellar cloud is proposed. In this model the primary and secondary interstellar atoms are treated as quasi-stationary kinetic gas. Population of H atoms originated in the supersonic solar wind is considered as zero-pressure fluid. Specific non-stationary effects introduced by the solar cycle fluctuations of the neutralized solar wind are explored. 1. I n t r o d u c t i o n The solar wind (SW) interacts with the local interstellar cloud (LIC). The heliospheric interface is formed in this interaction. The structure of the interface and the interface plasma flow depend on the parameters of the SW and LIC and their variations with time. The LIC velocity with respect to the Sun ( VLIC ~ 26 km s -1) and the LIC temperature (TLIC ~ 7000 K) are reliably established. The sonic velocity, aLIC, corresponding to TLIC, is smaller than VLIC. Therefore, the interstellar flow is supersonic and the LIC Mach number, MLIC = VLIC//aLIC, is larger than unity. Therefore, the interstellar plasma flow is supersonic and two-shock plasma structure is formed in the LIC/SW interaction. Interstellar atoms, galactic and anomalous cosmic rays, interstellar and interplanetary magnetic fields may affect the interface. However, the basic features of the heliospheric plasma interface are the same: 1) the heliopause (HP) separates the SW plasma from interstellar plasma, 2) the termination shock (TS) decelerates, heats and compresses the solar wind plasma and 3) there is the pile-up plasma region - plasma wall- between the heliopause and the bow shock. Interstellar H atom charge exchange with protons and significantly influence the heliospheric interface structure. The mean free path of H-atoms is compatible with the characteristic length of the problem considered. Therefore it is not correct to describe the * T h e research described in this publication was made possible in part by Award No.RP1-2248 of the U.S. Civilian Research and Development Foundation for the Independent States of the Former Soviet Union (CRDF), INTAS projects # 97-0512 and #YSF 00-163, RFBR grants # 01-02-17551, 99-02-04025, and International Space Science Institute in Bern.
-65
-
N.A. Zaitsev and K V. Izmodenov
H atom motion hydrodynamically ([1], [3]). The self-consistent axisymmetric model of the SW interaction with two-component (plasma and H atoms) LIC has been developed by Baranov and Malama [2]. A kinetic approach was used to describe H atoms in the model. However, the model does not take into account non-stationary processes in the SW. Steinolfson [4] has numerically investigated the problem of the heliospheric interface response to 180 day period fluctuations in the solar wind ram pressure in the case of subsonic fully ionized interstellar gas. He found that the variation in the distance to the termination shock is only about 1 AU. Karmesin et al. [5] and Baranov and Zaitsev [6] studied the influence of the 11 year variation of the solar wind pressure onto the heliospheric interface with 2D hydrodynamic models. These authors concluded that the response of the position of the termination shock to the changes of the solar wind parameters is within 8 - 12% (or about 10 AU). The response of the heliopause is smaller and the response of the bow shock is negligible. Baranov and Zaitsev also pointed out that in the supersonic solar wind as well as in the heliosheath the plasma flow has a quasi-stationary behaviour, while in the pile-up region between the bow shock and the heliopause there is a sequence of shocks and rarefaction waves. However, the mean plasma distribution is close to the stationary plasma distribution. In this work we investigate the effect of the solar cycle fluctuations of the fast solar neutrals flux onto the heliospheric interface. These energetic neutrals are created by charge exchange inside the heliopause and have significant influence on the compressed interstellar plasma (see, e.g., [3]). 2. F o r m u l a t i o n of t h e P r o b l e m The model [2] clearly shows, that there are four different types of H atoms in the heliospheric interface: i) the unperturbed interstellar neutral H atoms called Primary Interstellar Atoms or PIAs; ii) the compressed, decelerated, and Heated Interstellar Atoms (HIAs) formed by charge exchange with heated interstellar protons outside the heliopause; iii) the neutralized, decelerated, and Heated Solar Wind Atoms (HSWAs) formed in the heliosheath by charge exchange between the neutral interstellar gas and the hot protons of the decelerated and compressed solar wind, and iv) the neutralized Supersonic Solar Wind Atoms (SSWAs). These types of neutrals have different energy and spatial distributions in the interface. In Baranov-Malama model the influence of neutrals on plasma flow were taken into account as source terms Q in the right parts of ithe Euler equations for plasma component. To study properly the solar cycle influence on the heliospheric interface one should solve simultaneously the time-dependent Euler and Boltzmann equations. While the development of the corresponding Monte-Carlo algorithm is still in progress we would like to understand the basic physical effects of the solar cycle on the time-dependent interface from a simplified model. To study the influence of interstellar atoms we made the following assumptions: 1. Since the positions of the BS and the HP do not change significantly and fluctuations of plasma in this region are around stationary distributions, we assume that there is no influence of the solar cycle on the PIA's and HIA's. Hence the source terms QPIA and
-
66-
Solar cycle heliospheric interface variations: Influence of neutralized solar wind
QHIA into the plasma equations do not depend on the solar cycle and can be taken from the stationary solution. 2. We neglect the changes of HSWA's over the solar cycle. 3. For the number densities npIA and nHIA we assume that nH -- exp
7~H,TS~
T VH,TS
where a~z is the charge exchange cross section; rtE, VE a r e the solar wind proton number density and temperature, respectively; nH,TS VH,TS the number densitrnd velocity of interstellar atoms at the termination shock, r is the heliocentric distance, rE--1 AU is the distance to the Earth. This formula is the zero order approximation of interstellar neutral distributions in the interface, but for our objectives it is sufficient. 4. We assume that SSWA's can be treated as a zero-pressure fluid. The governing equations for this fluid are mass and momentum conservation lawis under the condition that the pressure PSSWA = O. The mutual influence of H-atoms and the plasma flow is modelled by the corresponding source terms in the mass conservation law and the momentum conservation law. In the stationary case this model gives the distribution of the SSWA's number density close to that obtained with the Monte-Carlo simulation. In order to simulate the ll-year solar cycle we changed the plasma velocity at the Earth orbit according to a sinusoidal law so that the ram pressure was varied by a factor of two. The sinusoidal variation of VE is the first harmonic of the real time dependence of the SW parameters. 3. N u m e r i c a l r e s u l t s Numerical results discussed below were obtained on the basis of a statinary solution computed for the following parameters [7]" np~ - 0.07 cm -3, n H ~ -- 0.2 cm -3, V~ = 25 kin~s, T~ = 5672K, nEo = 7 cm -3, VEO = 450 kin~s, TEO = 73507K. Our calculations show that the qualitative features of the non-stationary LIC - SW interaction established in [6] take place in the presence of neutral H-atoms as well. But the effect of the solar activity cycle is quantitatively stronger because the interface is closer to the Sun. For example, the TS excursion during the solar cycle on the axis of symmetry is about 30 AU, i.e. about 30% of its mean solar distance. Figure 1 shows the farthest and the nearest discontinuities positions in the upwind hemisphere parameters (solid lines). In the same figure the steady state positions of the BS, HP and TS are shown (by circles). One can see that the region between the BS and HP becomes wider: the mean location of the BS is farther from the Sun and the mean location of the HP is closer to the Sun than the corresponding steady state positions. The mean distributions of the plasma number density on the axis of symmetry in the region between the BS and HP is shown in Figure 2. The stationary distribution of the plasma number density in the same region is also shown. One can see that the mean density in the non-stationary cases is much less than in the stationary solution. This phenomenon can significantly influence the penetration of neutral H-atoms into the solar system. Figure 2 also shows the presense of a sequence of shocks and rarefaction waves moving from the HP to the BS similar to the one obtained in [6].
-67-
N.A. Zaitsev and V. V. Izmodenov
Figure 1. The geometrical pattern of the discontinuities: solid lines minimum and maximum heliocentric distances, circles the stationary position of the TS.
Figure 2. The interstellar plasma density: stationary solution (solid line), nonstationary solution (diamonds) and the mean density distribution (dashed line).
4. Conclusions The conclusions of presented study can be summarized as following: 1. The solar cycle influence on the variation of the termination shock is stronger in the case with H atoms than it would be in the case without H atom component. 2 Due to the solar cycle variations of the neutralized solar wind, i.e. atoms created in the supersonic solar wind by charge exchange with solar wind protons, the region between the heliopause and the bow shock become wider and mean plasma density in the region become smaller than for the stationary problem. This can be important for interpretations of heliospheric absorption in Ly-c~. REFERENCES
1. Baranov, V. B., V. V. Izmodenov, and Y. G. Malama, J. Geophys. Res., IDS, 95759585, 1998. 2. Baranov, V. B., and Y. G. Malama, J. Geophys. Res., 98, 15157-15163, 1993. 3. Izmodenov, V. V., this issue. 4. Steinolfson, R. S., J. Geophys. Res., 99, no. 7, pp. 13307-13314, 1994. 5. Karmesin, S. R., Liewer, P. C. and Brackbill, J. U. Geophys. Res. Lett., vol. 22, no. 9, pp. 1153- 1156, 1995. 6. Baranov, V. B. and Zaitsev, N. A. Geophys. Res. Lett., Vol. 25, No. 21, pp. 40514054, 1998. 7. Izmodenov V. V., J. Geiss, R. Lallement, et al., J. Geophys. Res. 104, 4731,1999a.
-68-
Time dependent radiation pressure and time dependent, 2D ionisation rate for heliospheric modelling M. Bzowski ~ * ~Space Research Centre PAS, Bartycka 18A, 00-716 Warsaw, Poland
1. M O T I V A T I O N The objective of this research was to develop time dependent, 3D models of the 3 solar parameters affecting the distribution of neutral interstellar hydrogen in the inner heliosphere: Lyman-c~ radiation pressure, charge exchange rate of H-atoms on solar wind protons, and photoionisation rate. The models are intended to be used in a program calculating the hydrogen density (see Bzowski et al., this volume) and while they are based on actual measurements, some simplifying assumptions were a priori made: the modelled quantities depend as 1/r 2 on the heliocentric distance, the periodicities allowed are longer than ~ 1 year, temporal variations of the solar wind occur immediately throughout the heliosphere; furthermore the functional form of the models was chosen to facilitate their use in numerical calculations. 2. R A D I A T I O N
PRESSURE
In lack of reliable data available it was assumed that the solar Lyman-a radiation pressure is spherically symmetric. The time series of the net solar Lyman-a flux from several spacecraft, recalibrated by Tobiska et al. (1997) and covering a time interval from 1977 till 1997.5, was expressed in the units of solar gravity assuming that the flux at the line centre is by number equal to the net flux. Next a periodogram analysis was performed (Press & Rybicki 1986) and based on the periodicities Pj found, a formula J, # (t) - #o + E (PJ cos cejt + qj sin wit), where wj - 2~r/Pj,
(1)
j=l
was fitted. The time series and the fitted model are presented in Fig.1 (left-hand panel) and the numerical values of the parameters in Table 1. 3. I O N I S A T I O N R A T E The net ionisation rate is an arithmetic sum of photoionisation and charge exchange rates, calculated as follows. *M.B. was supported by Polish SCSR grants 2P03C 004 14 and 2P03C 005 19.
-69-
M. Bzowski Table 1 Fitted parameters of radiation pressure and ionisation rates j Radiation pressure Photoionisation Charge exchange #0 = 1.27864 ~ph0 = 11.1862. 10-8s -1 ~chx0 = 55.5692- 10-Ss -1 qj Pj [y] pj 10 .8 qj 10 .8 Pj [y] pj 10 -8 qj 10 -8 Pj[Y] Pj 1 10.653 0.26043 -3.31286 30.007 1.3110 0.4609 27.261 5.1296 -7.5059 26.1185 -0.45974 0.46164 15.125 -0.0292 0.0046 7.5450 1.8444 2.2619 34.8206 -0.04823 0.12659 10.761 5.1934 0.9897 5.2490 -2.5727 -4.6350 43.1192 -0.17331 -0.26181 8.2470 -0.0822 1.3457 B0 = 25.3629 9 10 -s s-1 , b = 3 . 3 5 9 1 9 f o r n = 2 , B 0 = 2 6 . 0 1 9 3 - 1 0 - a s -1, b = 1 7 4 1 . 7 9 f o r n = 8 .....
Lyman-a radiation pressure 2.5 . . . . . . . . . . . . . . . . . . . . . . . .
1.5 1 0.5~
Solar photoionisation rate
'~J : .
1975
.
.
.
.
.
......... .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
~
0.5
1980 1985 1990 1995 2000 time [y]
.
1975
.
.
.
.
.
.
.
.
.
.
1980 1985 1990 1995 2000 time [y]
Figure 1. Solar radiation pressure (left-hand panel) and photoionisation rate (right-hand panel): models defined in Eq.1 vs data.
3.1. Photoionisation In absence of sufficiently long time series of actual data, the rate of photoionisation of hydrogen at 1 AU was calculated from a proxy based on the linear correlation between the solar 10.7 cm radio flux and the solar EUV flux, indicated by Hodges & Tinsley (1981). The formula to convert the 10.7 cm radio flux to the hydrogen photoionisation rate at 1 AU from the Sun was developed following a suggestion by Cummings (private communication), who derived this quantity from the solar EUV flux reported by Torr et al. (1979) for the day 113 of 1974 (solar minimum) and day 50 of 1979 (solar maximum). The photoionisation rates for these dates were equal, correspondingly, to 0 . 5 3 2 . 1 0 -7 s -1 and 1.87.10 -7 s -1 and the 10.7 cm adjusted flux to 66.5 and 214.10-22 W cm -2 s -1. Based on these data, a system of two equations with two variables was solved, yielding the following conversion formula: 9ph --
8.7661.1012 F10.7[W m -2 s -1] - 5.84576.10 -1~
(2)
With the use of this formula, daily averages of the 10.7 cm flux, covering the interval from 1948 till 1998 and publicly available in the Internet ( h t t p : / / w w w . d r a o . n r c . c a / i c a r u s / sol_home.shtml), were converted to a photoionisation rate time series and subjected to a similar analysis as in the case of radiation pressure. The periodicities and model parameters found are presented in Table 1 and a plot of the data and the model curve can be seen in Fig.1 (right-hand panel). Good correlation between the solar Lyman-c~ radiation
-70-
Time dependent radiation pressure and time dependent, 2D ionisation rate for heliospheric modeling pressure and the hydrogen photoionisation rate is striking but the model does not fit perfectly to the data; in particular, a 30% discrepancy about 1996 is visible. While it was possible to reduce this discrepancy by clipping the dataset to the recent solar cycle, this solution was rejected because it was giving unrealistic forward extrapolation and wrong timing of minima and maxima of the earlier solar cycles. Since photoionisation is not the dominant sink of hydrogen in the heliosphere, the 1996 underestimate should not be important for hydrogen distribution modelling.
3.2. Charge exchange Details of derivation of the charge exchange model will be published elsewhere because of the lack of space. In this paper results only will be presented.
Figure 2. Charge exchange rate adjusted to 1 AU in ecliptic (left-hand panel) and from the Ulysses Fast Latitude Scan (right-hand panel).
The rate of charge exchange between hydrogen atoms of the neutral interstellar wind and protons of the supersonic solar wind is calculated from the formula ~ch-x (t, (~) -- O'ch_ x (Vrel) ~tSW (t, r
Vrel
(t, r
/r 2, where Vre 1 -- VSW
(3)
and r is the heliocentric distance in AU, r the heliographic latitude, t time, V~el the relative velocity between the colliding particles, Gch-x the reaction cross section, nsw the solar wind density at 1 AU, and Vsw the solar wind speed, assumed independent on r. The charge exchange cross section used to be calculated from a formula by Fite et al. (1962) or by Maher & Tinsley (1977). But in comparison with experimental data compiled by Barnett et al. (1990), Fite's formula is 30 40% too high and Maher & Tinsley's is good from 120 to 450 km/s. Since the velocity range relevant for heliospheric modelling is from just a few to about 800 km/s, a new formula, spanning the whole range and accurate to at least 2.5%, was fitted to Barnett's data (velocity is expressed in cm/s)" a~h-x (v) - -2.01848.10 -17 In 3 v+1.00136.10 -15 In 2 v-1.71725.10 -~4 In v+1.03807.10-~3(4) Based on the data from Ulysses (Marsden 1996) it was assumed that the solar wind is steady in the polar regions and its variability is limited to an equatorial band. Hence a
-71 -
M. Bzowski functional form of the charge exchange rate model was adopted as r 2/Sch_ x ( r
/~o - Bo + ~ (pj cos wjt + qj sin wit)
t) -- B 0 -t-
exp [-b r
(5)
j=l
To find parameters of this formula, time series of daily averages of charge exchange rate were calculated using Equ.(3) from multispacecraft in-ecliptic solar wind data (time span from 1972 till 1998) and from Ulysses Fast Latitude Scan. The first time series was then converted to monthly averages and subjected to an analysis as for the photioinisation rate and yielded parameters COj - - 27c/Pj, pj, qj, and/50 (Table 1, Fig.2, left-hand panel). The latitudinal profiles were fitted to the second time series and yielded B0 and b for the two models assumed (n = 2, referred to as the gaussian bulge, and n = 8, referred to as the rectangular bulge, see Fig.2, right-hand panel). ,--, 10
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.j
2 o
.
.. .-5..
oo
•I
.
9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
i
.
1991 1992 1993 1994 1995 1996 1997 time [y]
Figure 3. Charge exchange rate adjusted to 1 AU calculated from the whole Ulysses data set from launch till 1997.5 compared with the rectangular bulge model (broken line) and the preferred gaussian bulge model (solid line).
While the rectangular bulge model reproduces better the UFLS data, the gaussian bulge model reproduces better the whole Ulysses data set available (see Fig.3). The model seems to be accurate enough for solar minimum but it needs to be improved to better reproduce the latitudinal behaviour of charge exchange rate during solar maximum. REFERENCES
Barnett, C. F., Hunter, H. T., Kirkpatrick, M. I., Alvarez, I., Cisneros, C., & Phaneuf, R. A. 1990, Atomic data for fusion. Collisions of H, H2, He and Li atoms and ions with atoms and molecules, Vol. ORNL-60S6/V1 (Oak Ridge, Tenn.: Oak Ridge National Laboratories) Fite, W. L., Smith, A. C. S., & Stebbins, R. F. 1962, Proc. R. Soc. London Ser. A, 268, 527 Hodges, R. R. & Tinsley, B. A. 1981, J. Geophys. Res., 86, 7649 Maher, L. J. ~z Tinsley, B. A. 1977, J. Geophys. Res., 82, 689 Marsden, R. 1996, Space Sci. Rev., 78, 67 Press, W. H. & Rybicki, G. B. 1986, Astrophys. J., 338, 277 Tobiska, W. K., Pryor, W. R., & Ajello, J. M. 1997, Geophys. Res. Lett., 24, 1123 Torr, M. R., Torr, D. G., R.A, Ong, & Hinteregger, H. E. 1979, Geophys. Res. Lett., 6, 771
-72-
Oral p a p e r s a n d p o s t e r s
MAGNETIC FIELDS IN THE HELIOSPHERE: INTRODUCTORY REMARKS
W.I. Axford Max-Planck-Institut fiir Aeronomie, Katlenburg-Lindau [email protected]/Fax: +49-5556-979410 In general the heliospheric magnetic field is rather weak and should not have significant effects apart from the region near the stagnation point where it may grow substantially. Nevertheless its topology is interesting and this in turn may have an effect on the entry and escape of low energy cosmic rays and other energetic particles. Reconnection between the internal (solar) magnetic field and the external (interstellar) magnetic field may also be important in this regard. It is also of significance to consider deviations of the field from a simple spiral form either as a result of the random motion of foot-points at the base of the corona or of the differential rotation of the sun.
TEMPORAL VARIATIONS OF LY-c~-INTENSITIES, BACKSCATTERED FROM THE INTERSTELLAR HYDROGEN ATOMS UPSTREAM OF THE SUN: ULYSSES-GAS OBSERVATIONS Manfred Witte Max-Planck-Institut fiir Aeronomie, D-37191 Katlenburg-Lindau, Germany.
While the flow of interstellar neutral hydrogen particles approaches the sun, it becomes increasingly depleted closer to the sun due to charge exchange with the solar wind. The Ly-a intensities emitted by the sun decrease with larger distances, so that the Ly-a intensity/8, scattered from the H-distribution (Is ,'~ nn x/solar) has a maximum at a distance of 1 to 3 AU upstream of the sun (so called maximum emission region, MER). In 1999, this region has been monitored over several solar rotations by the GAS-instrument on ULYSSES, positioned about 5 AU from the sun in crosswind direction, with a time resolution of about 4 days. The observed intensities reveal pronounced temporal and latitudinal (!) variations: 9 there is a longterm trend, lasting several solar rotations 9 there is a pronounced periodic (~ 25 days) variation (solar rotation) 9 and, most noticable, this latter variation is small at equatorial latitudes, but already significant at latitudes of + 20 degrees. The interpretation of these variations will give insight into the variability of a) the solar Ly-c~ radiation, b) the efficiency of the charge exchange process, modifying the H-atom densities in the inner heliosphere. Details of the observations and first interpretations will be presented.
-73-
This Page Intentionally Left Blank
General D i s c u s s i o n Jokipii to Zank: What is the compression ratio of travelling shocks just immediately upstream of where the heliospheric bow shock is? Zank: Those tend to be fairly weak - compression ratios of no more than 1.8 to 2 at most. Dorman to Zank: Did you take into account the non-spherical solar wind? And how did you take into account the nonlinear effect of cosmic rays (CRs)? Zank: In answer to your first question: that was part of the point about the 3-D simulations. There we assumed Ulysses parameters that have been measured, so the velocity of the solar wind over the poles was about 800 km/s and in the ecliptic about 400 km/s. The question about the CRs: we just assume that they contribute to the interstellar pressure. So, we keep galactic CRs in the models but only in a very simple way. We don't actually solve a CR transport equation at this point. Baranov to Mfiller: What parameters do you take into account in your Boltzmann equation? For example for the gravitational and radiation forces? And what is the assumption about the distribution function of protons? Miiller: The last question first: tile protons come out of a gasdynamic simulation, so we assume Maxwellian proton distribution functions. We are further assuming that the radiation pressure is balanced by gravity or vice versa, and we have not included photo ionization yet. Sreenivasan to Izmodenov: How do you treat the right-hand side of your Boltzmann equation? Izmodenov: We use the Monte-Carlo scheme and we just model the effect of charge exchange. We compute the probability of charge exchange and we use the method of splitting of trajectories. The Monte-Carlo code allows us to calculate the distribution functions directly. Baranov to Izmodenov: Could you compare the distribution functions obtained in your model with those obtained with the model by Miiller et al.? Izmodenov: One can probably compare but our calculations are computed in much smaller cells in all regions. Veselovsky to Izmodenov: What can be said about the accuracy of the current cross sections for atomic collisions from experiment and theory? Izmodenov: The cross section that I have shown have been calculated theoretically - experimental data are not available at such low energies, as far as I understand. Jokipii to Nickeler: You use the assumption of field-aligned flows. That seems to be violated rather seriously. Could you tell me the effect of that assumption? Nickeler: It is a simplification, of course. The interesting thing is that, because of the non-linearity of the magnetohydrostatic equations, it is difficult to find analytical solutions. However, we also try to find methods in the same way with a transformation with components of the flow orthogonal to the field lines.
-75-
General Discussion
Fahr to Witte: Is there anything left in your fits for electron impact ionization? Witte: I think to answer that needs more modelling to get the actual shape of the maximum emission region. So far, we are not talking about any values of ionization rates here. Baranov to Axford: Two years ago Linde et al. published a model taking into account the heliospheric as well as the interstellar magnetic field. Are your interesting ideas supported by their results regarding reconnection or the effect of the heliospheric magnetic field near the heliopause? Axford: Reconnection is essentially impossible to stop in these circumstances, it's going to happen all over the place. It means that field is even more complex, and it may be important for low energy electrons to come in. Sreenivasan to Barnes: Astrophysicists generally believe the interstellar magnetic field is of the order of 5-6#G and its direction is supposed to be along the spiral arms with some fluctuating component. Your values are quite different. Do these things make a substantial difference for the direction as well as magnitude of the external field? Barnes: On a scale of things small compared with the Local Bubble, like the heliosphere, any relation between the magnitude and direction of the magnetic field to what it maybe in the general galactic neighbourhood is not at all clearcut. Richardson to Bzowski: The charge exchange ionization rate depends on the solar cycle. Have you modelled that? Bzowski: We have modelled this by a~ssuming that tile rate at the poles is fixed and the ration in the ecliptic is going up and down.
-76-
Session 2: Heliospheric and I n t e r s t e l l a r Connections
This Page Intentionally Left Blank
Radiative Transfer at Lyman cu in the Outer Heliosphere Eric Qu~merais ~ ~Service d'A(~ronomie du CNRS B.P. 3, 91371, Verrihres le Buisson, France eric. quemerais~aerov.j ussieu.fr We present here the results of a model of multiple scattering at Lyman c~ in the interplanetary medium. It uses the formalism of Angle Dependent Partial Frequency Redistribution which allows a computation of the actual line profile for a given hydrogen velocity distribution taking into account actual local velocities and temperatures. This model is applied to prove that multiple scattering effects cannot be neglected anywhere in the heliosphere. We also give results for intensity, apparent velocity and apparent temperature with increasing solar distance. 1. I n t r o d u c t i o n The existence of a strong Lyman c~ background observed in the interplanetary medium has shown the presence of neutral hydrogen atoms backscattering the solar photons. In the outer heliosphere, say outside 40 AU, the hydrogen distribution is not affected by solar EUV photo-ionization or charge exchange with solar wind protons. There, the H atoms are mainly influenced by the interface structure between the expanding solar wind and the ionized component of the interstellar medium. This interface is the object of many speculations concerning its nature, stability and actual position. For a review concerning this problem, see e.g. Baranov [1990]. The neutral H atoms are coupled to the plasma components of the solar wind and of the interstellar medium through charge exchange processes. As a result, the outer heliosphere hydrogen distribution is substantially different from the case of the hot model where the H atom distribution is a simple gaussian distribution with constant number density far away from the sun [Thomas, 1978]. One striking feature of the outer hydrogen atom distribution obtained by theoretical models including the effects of the heliospheric interface is the so-called hydrogen wall. This wall is due to a pile-up of neutral H atoms in the region where the interstellar plasma is strongly heated and decelerated. Charge exchange between slowed down interstellar protons and neutral hydrogen atoms leads to a new neutral hydrogen component. It is characterized by a large temperature and a small bulk velocity in the solar frame [Baranov and Malama, 1993; Zank et al., 1996; Baranov et al., 1998]. Since 1993, a series of observations have been performed by the UV Spectrometers on board the Voyager 1 and Voyager 2 spacecraft to try to observe the signature of the hydrogen wall in the Lyman a pattern [Qu6merais et al., 1995]. Hall et al. [1993] reported that the Lyman c~ intensity measured away from the sun was falling off with distance less quickly
-79-
E. Quemerais than expected from a standard hot model. This result suggested that there was a positive gradient of neutral hydrogen at large distance from the sun and could be explained by the existence of an interface region. However, the actual gradient observed by Hall et al. [1993] was not confirmed by later dedicated observations [Qu~merais et al., 1995]. An alternate explanation is now considered, i.e. that the increase of Lyman c~ intensity in the upwind direction could be partially due to a constant emission from HII regions in the galactic plane [Qu~merais et al., 1995]. This question is still unresolved and the latest data will need to be compared to the best modeling available. In the previous literature, radiative transfer effects have not often been considered very precisely. The main reason is that in the inner heliosphere, the interplanetary medium is supposed to be optically thin at Lyman c~. In such a case, one needs only to integrate the first scattering order over the line of sight to compute the scattered intensity. Unfortunately, this is not quite correct as shown by Keller et al. [1981] and confirmed by Hall et al. [1992] and Qu~merais and Bertaux [1993]. Those three previous models were obtained from Complete Frequency Redistribution hypothesis. A more correct representation of the scattering process was used by Scherer and Fahr [1996] and Qu~merais [2000]. This is called Angle Dependent Partial Frequency Redistribution (A.D.P.F.R.) and was extensively described by Mihalas [1970]. Although, the physical basis used for the computation is very similar, the results obtained by these two groups were different. Scherer and Fahr [1996] estimate that all higher orders of scattering are negligible and that the computation can be stopped at the first scattering in the inner heliosphere. On the contrary, Qu~merais [2000] finds that total intensity computations obtained from ADPFR is very similar to CFR results and confirms the results of Keller et al. [1981], Hall [1992] and Qu~merais and Bertaux [1993], i.e. even in the inner heliosphere, it is necessary to include higher orders of scattering in the computation of ther backscattered intensity. This is even more necessary in the outer heliosphere. The first section tries to give quantitative results on this problem. The second section gives a presentation of the main results obtained for intensity, apparent velocity and apparent temperature derived from line profiles in the outer heliosphere.
2. A r e R a d i a t i v e
Transfer Models
Necessary?
The optical thickness at Lyman c~ in the interplanetary medium varies by 1 for 10 AU. This very rough estimate shows that the medium is optically thick in the outer heliosphere. To illustrate that, we show in figure 1 the profile of the local emissivity corresponding to the first order of scattering for a point in the outer heliosphere. More details are given in Qu~merais [2000]. Keller et al. [1981] were the first authors who carefully computed the effects of multiple scattering for an observer at various distances from the sun and looking radially away from the sun. Their computations were based on a Monte-Carlo approach and used the assumption of Complete Frequency Redistribution. Their conclusions were very important for the study of the interplanetary Lyman c~ background. Indeed, they showed that even at 1 AU from the sun, where the H number density is very small, the optically thin approach does not apply. In the downwind cavity, they found an intensity ratio for multiple scattering over optically thin assumption equal to 1.35. These results were later
- 80-
Radiative transfer at Lymanla in the outer heliosphere
--
/
-
\
t;
a
/
-
/
_
f'
\ ,
/\
_
L
-
l I~] 17~I l~l~l I~l 17~I l~
~~ ~ ~ I~
I
~
J
I
:
Figure 1. Line profile of primary emissivity for a scattering point at 54.2 AU from the sun in the crosswind direction. The dash line shows the radial profile and the dashdot line shows the perpendicular profile. The Monte Carlo results are shown by squares (radial) and diamonds (perpendicular). The hot model parameters in this simulation are NH = 0.15 cm -a, VH = 20 km s -1 and TH = 8000 K.
confirmed independently by Hall [1992] and Qu~merais and Bertaux [1993]. Following this result, various approaches have been adopted. The first one includes a correction to the optically thin computation derived from a comparison with a radiative transfer computation. This was done by Ajello et al. [1987] or Pryor et al. [1992]. Usually this correction is computed for one set of parameters defining the hydrogen distribution assuming that the ratio of intensity from radiative transfer to intensity from optically thin approximation will not change too much with different hydrogen distributions. A second approach used by Bertaux et al. [1985] is based on the following remark. The total intensity can be divided between first order scattering photons and higher orders. It is rather easy to compute the first term which includes extinction between the sun, the scattering point and the observer. However by neglecting extinction between the sun and the scattering point, one will somewhat over estimate the first term thus compensating for the lack of the second term. The main problem with this assumption is that it is not
- 8 t
-
E. Quemerais based on any actual computation. 2.1. C a n M u l t i p l e S c a t t e r i n g be N e g l e c t e d at 1 A U ? The aim of this section is to prove q u a n t i t a t i v e l y that multiple scattering effects cannot be neglected anywhere in the heliosphere, even at 1 AU from the sun, even for low hydrogen number ~t infinity as stated in Scherer and Fahr [1996]. The model used here is a so-called hot model with no interface. The interstellar hydrogen distribution is defined by a bulk velocity of 20 km s -1 and a temperature of 8000 K. We have taken a value for the hydrogen number density equal to 0.05 cm -3 which corresponds to a lower limit of what is currently believed. Higher values of NH will increase the relative importance of multiple scattering. Radiative transfer effects have been computed as described in Qu6merais [2000]. The hydrogen distribution is described locally by its mean velocity in the solar rest frame and temperatures along 3 perpendicular axes. This distribution is used for all scatterings without any simplification.
Table 1 Characteristic Values in the Inner Heliosphere Nu = 0.05 cm -3 UPWIND r in AU 1.00 2.00 5.00 7 2.73 x 10 .3 0.0107 0.0638 T(T) 0.998 0.992 0.956 el/cO 0.0156 0.0361 0.123 0.00 0.00148 0.0480 CROSSWIND r in AU 1.00 2.00 5.00 T 7.12 • 10 - 4 3.66 x 10 .3 0.0314 T(7-) 0.999 0.997 0.978 el/e0 0.0115 0.0265 0.0954 e2/e0 0.00 9.97 10 - 3 0.0456 DOWNWIND r in AU 1.00 2.00 5.00 w 4.17 • 10 .6 5.92 • 10 .5 1.81 • 10 .3 T(~-) 0.999997 0.999958 0.9987 el/C0 0.00640 0.0139 0.0495 e2/eo 0.00 0.00 0.00
10.00 0.207 0.865 0.284 0.178
20.00 0.558 0.682 0.545 0.809
50.00 1.82 0.319 0.785 4.34
10.00 0.128 0.914 0.237 0.183
20.00 0.403 0.757 0.488 0.719
50.00 1.50 0.380 0.818 4.10
10.00 0.0165 0.988 0.134 0.187
20.00 0.0945 0.936 0.321 0.523
50.00 0.629 0.651 0.745 2.83
Table 1 gives the characteristic values of interest here in the case where the interstellar hydrogen number density is equal to 0.05 cm -3 for points at 6 distances from the sun along three directions, upwind, crosswind and downwind. For each point we give the distance from the sun, the optical thickness at line center, the equivalent transmission computed from the Holstein function [Holstein, 1947], and the ratio over the primary term of the emissivity of the two first scattering orders cl and c2.
-82-
Radiative transfer at Lyman-a in the outer heliosphere From this table, we note that for a point at 1 AU from the sun, the secondary emissivity corresponds to roughly less than 2% of the primary term. This would lead us to neglect multiple scattering effects. H o w e v e r this r e a s o n i n g is w r o n g because, in that way one forgets the fact that t h e i n t e n s i t y is o b t a i n e d by i n t e g r a t i o n over t h e line of sight. This line of sight extends over more 10 AU. At 5 AU from the sun in the upwind direction, the secondary term corresponds to more than 25% of the primary term. At 10 AU, it is more than 75%. Although at 1 AU the interplanetary medium is optically thin at Lyman c~, the contribution of the medium beyond 10 AU is still important and at that distance from the sun the interplanetary medium is optically thick at Lyman c~ . Table 2 gives quantitative values for the various components contributing to the total intensity as seen at 1 AU looking radially. The computation has been made for 3 different values of the interstellar hydrogen number density. In the best case, i.e. upwind for 0.05 cm -3, the primary term corresponds to 73% of the total intensity. Photons with order of scattering equal to 2 or higher still represent 15% of the total intensity.
Table 2 Estimates of Various Components of Intensity at 1 AU NH Line of Sight Itot~t Io/[total 0.05 0.05 0.05
I1/Itot~l
I<2/Itotat
UPWIND CROSSWIND DOWNWIND
236 R 164 R 53 R
73% 73% 44%
12% 13% 14%
15% 14% 42%
0.10 cm -3 0.10 cm -3 0.10 cm -3
UPWIND CROSSWIND DOWNWIND
464 R 330 R 118 R
65% 57% 34%
13% 14% 13%
22% 29% 53%
0.15 cm -3 0.15 cm -3 0.15 cm -3
UPWIND CROSSWIND DOWNWIND
684 R 493 R 184 R
60% 52% 29%
14% 14% 12%
26% 34% 59%
cm -3 cm -3 cm -3
3. L y m a n a B a c k g r o u n d in t h e H e l i o s p h e r e This section will detail the results of our Angle Dependent Partial Frequency Redistribution computations for a hot model. This study does not take the interface into account, so the results in the outer heliosphere will be very different from what would be obtained from interface models. In what follows, the observer is looking radially away from the sun. All the computations include the classical phase function established by Bran& and Chamberlain [1959]. The solar parameters are r the ratio of radiation pressure over solar gravitation equal to 1 and the hydrogen lifetime against ionization at 1 AU equal to 1.2 x 106 seconds. These are standard values for an isotropic hot model. The hydrogen distribution depends also on the interstellar parameters. The hydrogen number density distribution is proportional
-83
-
E. Quemerais
~ .
...... 0..
,,'\ -
'\
""-.
X
\
',,\
\,\, ':'\,
B
: -
'\
""-.
~
.......
",,\
, \
.........
..............
", ~ ~ \ X
9
,
\
\ ,
"\,
"\,
"- -..
'\,
\ 9
~ .
-
-~
9
0
_
9
9
_
........
~..
20
40
60
Figure 2. Upwind radial intensity as a function of solar distance between 1 and 60 AU. After 30 AU, the optically thin approximation falls off like 1/r. The radiative transfer computation follows 1/r ~ where 1 _< c~ _< 2, depending on the optical thickness of the medium. The other computations decrease very quickly with solar distance.
to NH, the value of the hydrogen number density at large distance from the sun. In this study, we have used 0.12 cm -3 . 3.1.
Lyman
a Background
in t h e O u t e r H e l i o s p h e r e
No approximation can be used to describe the intensity away from the inner heliosphere. Figure 2 is an illustration of this statement. This shows a computation of the intensity for an observer upwind at different distances from the sun and looking radially away from the sun. All approximations, except optically thin, fall down much faster than the multiple scattering value. Since there is no extinction, the optically thin approximation overestimates the correct decrease of intensity. It must be noted that for our model here, the density distribution outside 30 AU is more or less constant. In that case, the optically thin approximation falls off like 1/r and the radiative transfer case falls off like 1/r ~, where a is a coefficient between 1 and 2 [Hall, 1992]. The coefficient a depends on the optical thickness of the medium, i.e. on the value of NH and distance from the sun. When an interface is included the upwind hydrogen number density shows a positive gradient in
- 84-
Radiative transfer at Lyman-a in the outer heliosphere
the outer heliosphere [Baranov and Malama, 1993]. In that case, c~ can be smaller than 1. This is why measuring the fMl-off of the upwind intensity from the Voyager UVS data is important [Hall et al., 1993; Qufimerais et al., 1995].
-- .
~
"'~
.........
_ _2
.
~
2 _
Figure 3. Downwind radial intensity as a function of solar distance between 1 and 60 AU. Here the situation is much more complex than in the upwind direction because even at 60 AU the H number density is not constant. We see also that multiple scattering effects are predominent.
In the downwind direction (Figure 3), the situation is more complex. The hydrogen cavity extends much farther from the sun and even at 60 AU the hydrogen number density has not reached a constant value. The size of the cavity is affected by the solar parameters and its filling by the mean velocity and temperature of the interstellar gas. Study of the downwind interplanetary background as a function of solar distance is then more dependent on the solar and interstellar parameters. It is also strongly affected by the variability of the solar parameters [Rucitlski and Bzowski, 1995]. Figure 4 shows how the upwind profile mean shift (apparent velocity) changes with solar distance. This has been computed for three different values of V/-/, namely 20, 22 and 26 km s -1. We see that the apparent velocity of the line seems to increase by
-85-
E. Quemerais
-18
V_H =
-20
_.. .
.
.
_ . _
E
EL
20 . . . . . . .
km/s,
T_H . . . . .
=
8000
K
9 . . . . . . . . . . . . . . . . . . . . .
_ i
V_H = 22 kin/s, T_H = 10000 K . . . . . . .
- . . . . .
--"
_2.
. . . .
C . _
F--
. . . .
_
,/ ,'/
. . . . . . . . . . .
.
_
-24
I c/')
_
_
z < W
-26
-28
-50
Figure 4. Mean shift of the line profile obtained for an observer looking radially away from the sun in the upwind direction. In the case of the hot model, at large distance from the sun, the shift is equal to the velocity of the interstellar gas. The shifts have been computed for three types of models with VH -20, 22 and 26 km s -1. At one AU in the upwind direction, the apparent velocities found for each model are-22.8 km s -I, -25.1km s -1 and-27.9 km s -1, respectively.
roughly 3 km 8 - 1 in the inner heliosphere. This is caused by the apparent acceleration of the hydrogen atoms due to selection effects (here # = 1, so there is no gravitational acceleration). Slower atoms are more easily ionized because they spend more time near the sun. This results in a selection of faster atoms and an increase of the local mean velocity, i.e. an apparent acceleration. Figure 5 shows the apparent temperature in both the upwind and downwind directions as a function of distance from the sun. On this figure, we show the results for the optically thin approximation (dotted line), the A D P F R radiative transfer (solid line) and the selfabsorbed computation (dashed line). The primary term is not shown here because the extinction near line center makes the estimation of the apparent temperature completely unrealistic. First, we note that the optically thin approximation strongly underestimates the apparent temperature. It goes asymptotically back to the value of the gas in the
-86-
Radiative transfer at Lyman-a in the outer heliosphere
-
"~
~
-.~ _
~.
/
04
11//
/ J
. . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
-..-
o
Figure 5. Apparent temperature of line profiles obtained for an observer looking radially away from the sun. The values are shown in both the upwind and downwind directions. The dotted line gives the result of the optically thin approximation, the solid line the ADPFR result and the dashed line corresponds to the computation of the self-absorbed c&se.
interplanetary medium. Self-absorption gives the value closest to the actual result of ADPFR although the results are systematically underestimated by roughly 1000 Is At very large distance from the sun, the apparent temperature is a complex mixture of coherent and perpendicular scatterings. We know that perpendicular scatterings have an apparent temperature equal to the temperature of the medium. On the other hand, after many scatterings the profile of the photons which are scattered is not white anymore. In an extreme case, let us assume that after n scatterings the source profile for the n + 1 scattering is a doppler profile at the temperature of the gas. Then because we assume a doppler scattering cross section, the coherently scattered profile has an apparent temperature which is half the actual temperature of the gas. The total emissivity will be a complex mixture of these various profiles. The widening of the line is mainly driven by extinction when integrating over the line of sight. A gaussian profile at temperature T attenuated following a gaussian at the same temperature yields an apparent temperature
-87-
E. Quemerais
which is larger than the original one. All these antagonist effects are taken into account in our computation. Finally, the variation of the Lyman a line profile with distance depends strongly on the optical thickness in the interplanetary medium, hence of the parameter
NH. 4. D i s c u s s i o n
The first result we wish to point out here is that our code shows that multiple scattering effects cannot be neglected even in the close vicinity of the sun. This is in contradiction with the results of Scherer and Fahr [1996] and we have given quantitative results to justify this. We have compared systematically the results of our ADPFR multiple scattering code with the different approximations used by various authors. Finally, we have shown typical results for intensity, apparent velocity and apparent temperature both in the inner and the outer heliosphere. We expect that the inclusion of effects of the heliospheric interface on the hydrogen distribution, as obtained by Baranov and Malama [1993] for instance, will modify the line shapes in a dramatic fashion, especially in the outer heliosphere. In that respect, the previous section on outer heliosphere Lyman a background must be taken as a study case. Hall et al. [1993] or Qu~mer~is et al. [1995] have shown that the UVS data in the outer heliosphere do not agree with the results of hot model type computations. REFERENCES o
2. 3. 4. 5. 6. 7. .
9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Ajello J. M., Stewart A. I. F., Thomas G. E., & Graps A., 1987, A p J , 317, 964 Baranov V. B., 1990, Space Science Reviews, 52, 89. Baranov V. B. & Malama Y. G., 1993, J. G e o p h y s . Res. , 98, 15157 Bertaux J. L., Lallement R., Kurt V. G., & Mironova E. N., 1985, A & A , 150, 1. Bran& J.C. and Chamberlain J.W., 1959, A p J , 130, 670. Hall D. T., 1992, Ph. D. Thesis, University of Arizona, Tucson. Hall D. T., Shemansky D. E., Judge D. L., Gangopadhyay P., & Gruntman M. A., 1993, J. G e o p h y s . R e s . , 98, 15185. Holstein T., 1947, Physical Review, 72, 1212. Keller H. U., Richter K., & Thomas G., 1981, A & A , 102, 415. Mihalas D., 1970, in Stellar Atmospheres, W.H. Freeman & Co, San Francisco. Pryor W. R. et al., 1992, A p J , 394, 363 Qu(~merais E. & Bertaux J. L., 1993, A & A , 277, 283 Qu~merais E., Sandel B. R., Lallement R., & Bertaux J. L., 1995, A & A , 299, 249 Qu~merais E., 2000, A & A , 358, 353-367 Rucifiski D. and Bzowsli M., 1995, A & A , 296, 248. Scherer and Fahr, 1996, A & A , 309, 957. Thomas G. E., 1978, Ann. Rev. E a r t h P l a n e t . Sci., 6, 173 Zank et al., 1996, J. Geophys. R e s . , 101, 21.
-88-
A Fluid A p p r o a c h to the H e l i o s p h e r e / V L I S M P r o b l e m R. L. McNutt, Jr. a, M. Wiltberger b, J. Lyon b, and C. C. Goodrich c aThe Johns Hopkins University Applied Physics Laboratory Laurel, MD 20723-6099, U.S.A. bDartmouth College, Hanover, NH 03755, U.S.A. cUniversity of Maryland, College Park, MD 20742, U. S. A. The interaction of the solar wind with the very local interstellar medium (VLISM) depends strongly upon charge exchange between the neutral and ionized hydrogen outside of the heliopause. The resulting structure is of the order of only one mean free path for such collisions, so strictly a kinetic approach is required. However, such an approach is very computationally intensive, while a fluid approach is less so. We consider how far the fluid approach can be pushed in this problem and compare some results with those obtained from kinetic approaches. 1. INTRODUCTION Accurate modeling of the heliosphere-solar-wind system is required to plan for a probe to near-interstellar space. A probe toward the "nose" of the heliosphere minimizes the distance, and hence, flight time into the interstellar medium [ 1-6]. The heliospheric structure has a characteristic scale size of--100 AU; a 10% change in the spatial scale translates into years of probe travel. To plan such a mission, there are both scientific and engineering needs to "get the numbers right." Our approach is to (1) compare fluid and kinetic models using the same sets of parameters to identify similarities and differences, (2) run multiple simulations varying one parameter at a time to test sensitivity to the various input parameters, (3) compare density profiles along the upstream stagnation line as a means of getting a quick look at model differences, and (4) explore fluid models with T e - Tprotonfor comparison with a broad number of kinetic model results. 2. T H E
PROBLEM
The charge-exchange process that diverts the neutral interstellar flow around the heliosphere has a mean free path comparable to the interaction region size. Large mean free paths strictly require a kinetic treatment [7], but a fully time-dependent, six-dimensional kinetic treatment is computationally intensive. Fluid models offer a less computationally-complex approach, but with limitations [8-11 ]. Other approaches, including those with hybrid codes have limitations as well [for a review, see 12].We compare results from the limiting case of no magnetic field and a cylindrically symmetric geometry with results obtained using a kinetic/hydrodynamic approach. We also consider the next order corrections to the hydrodynamic (HD) equations for the neutral component and estimate the size of the errors made in not including these further refinements to the HD approach.
- 89-
R.L. M c N u t t Jr. et al.
2.1. Solution Space The numerical simulation is run in a region between a spherical shell of 20 AU radius out to a cylindrical region stretching from - 6 0 0 AU in the upwind (-x) direction to ~ 1000 AU in the downwind (+x) direction. The cylindrical region is - 8 0 0 AU in radius. The simulations cover roughly 2.72 years every 1000 time steps (-1 day per step). The time steps are adjusted by the code as required: steps are sufficiently small that signals propagate at most across -1/3 cell. The simulations begin at t= 0 with a uniform neutral and plasma interstellar flow pervading the solution space and the solar wind outflow commencing radially outward at the 20 AU inner boundary. Simulations are run to an asymptotic state (--1500 years), and then grid resolution is increased by a factor o f - 2 . The simulation is then run until transients from the regridding damp out (another -- 100 years) [9]. 2.2. Charge Exchange Formulation For a large range of values, the charge-exchange cross section is very accurately represented by a power law. CYcx( g ) - cY0g ~
(1)
where g is the relative speed of two colliding particles. The charge-exchange operators can be determined e x a c t l y for this case. For comparisons with Izmodenov et al. [7] we use the cross section [ 13] CYcx,MT ( g ) - 10 -14 ( 1.64 - 0.0695 In g [cm s -1 ]
)2
[ cm 2]
(2)
A least squares fit in the range from 1 to 1000 km s-~ yields [ 10] Crcx,m = rt(6.6504 a 0 )2
g-0.2398
(3)
corresponding to a repulsive force law -- r -~8(here a 0 is the Bohr radius). Special cases results from taking or, -- cy0 g-1 o r CYcx-- cy0 corresponding to the "hard-sphere model" for purposes of evaluating the momentum-and-energy-exchange terms [8]. We use this latter formulation (v=0) for comparison with [7].
2.3. Analytic Approximations From [ 10] we make use of the approximation
r(2 + n + 89v) y.,,,(a) -
M(_I
_ 1 v , n + ~ , - a 2) = [(88+ n + x1 v) + a 2]89
(4)
r(n+3)
For values of v from -1 to 0, the maximum error occurs for Y0,0 and is less than 2.0%. With a force law F - ~c r -s, we have s = 1 - 4/v so these cases correspond to hard spheres ( s ---> oo ) to Maxwell molecules (s = 5) [14]. This approximation remains fairly good through v - -2 (corresponding to s = 3) with an error of less than -15%.
- 90-
A fluid approach to the heliosphere/VLISMproblem From these results, we define corresponding characteristic speeds *
(5)
2
--
+
(6)
2.4. S i m u l a t i o n E q u a t i o n s
The ion equations are[10]"
a(pu)
Ot
1
+ V 9 (puu) + V P - 7 J x B - QM, H+_~H --
-u)
m proton
OEPl- V* [(Ep + P ) u ] -
J * E - Qe.., .+-~n oa
_U ac~(U~)(2+Tv)(pp, m
a~x(V[.) pp.v;
*
1
,
(7)
puP)+U~u ~cx(U~,M) lpp.(u2-u2) ' m proton 2
m
*
m proton
m
The same methodology can be used analogously to write out explicitly equations for the neutral component. Note that CYcx- const corresponds to v - 0 . 2.5. N e u t r a l C o m p o n e n t E q u a t i o n s to Next O r d e r
The neutral equations become [ 10, 15, 16]" O(PnUn)
Ot
pp.v;
+ V * (puunun + P n ) - - Q M , n+_~u
-u)
m proton
Ep,H
~ +
V - [(Ep,. + Pn)un + q n ] - - Q e
0t
(8)
- -U~ ac~ (U~) (2 + ~ v)(pPn m
....,.n+-~"
*
1
mproton
m
pnP)- U,,M ac~(U~,M) 1 *
mproton
2 ppH(u2H-u2)
where Pn is the pressure tensor and qn the heat flux vector for the neutral component. In the model as implemented qn - 0 and the off-diagonal elements of PH are zero and the diagonal elements of Pn are all equal. Examples of the proton and neutral hydrogen densities using this model and the same parameters as in [9] are shown in Figure 1. The effect of different chargeexchange power laws in shown in Figure 2. 3. C O M P A R I S O N S
Comparisons of the MHD/HD approach with the HD/kinetic approach can be carried out by making the following simplifications to our simulation model: (1) Set B - 0 (2) Use a radially symmetric solar wind (3) Align the interstellar flow with the -X axis of the simulation coordinates.
-91 -
R.L. McNutt Jr. et al.
Figure 1. Logarithm of the proton density (top) and of the neutral hydrogen density (bottom). The case presented here uses the parameters of [9], viz. electron temperature of zero, upstream ion density of 0.30 cm -3, upstream neutral density of 0.10 cm -3, ion and neutral upstream temperatures of 7000K, VLISM velocity of 25.7 km s-1 toward ecliptic longitude 74.9 ~ and latitude of -7.8 ~ A local interstellar magnetic field of 0.14 nT toward galactic longitude 70 ~ and latitude 0 ~ is included in the model. Solar wind parameters are taken from IMP 8 and Ulysses data (details in [9]). Density profiles in the plane of the solar rotational equator are shown.
-
92-
A fluid approach to the heliosphere/VLISMproblem
Figure 2. Differences of the densities of protons (top) and neutral hydrogen atoms (bottom) from using the same run parameters as used in Figure 1 using power-law indices (here labeled as "n" and "nu", cf. Equation 1 of the text) of-1 (Maxwell molecules) and -0.2654; the runs in Figure 1 used indices of 0. Maximum differences are --5% for the proton density and ~ 15% for the neutral density. The index of-0.2654 is slightly different from the best fit to the power law of [7].
-93 -
R.L. McNutt Jr. et al.
These simplifications lead to an axisymmetric model with both the plasma and neutral fluids described by the hydrodynamic equations. 3.1. C o m p a r i s o n P a r a m e t e r s Izmodenov et al. [7] have used a symmetric model with an HD/kinetic implementation. They have included photoionization, electron impact, and radiation pressure on the neutral hydrogen, as well as the effects of gravity. These effects should primarily be noticeable in the inner heliosphere, so we can compare our approach with their runs on a fairly equal basis, even though we have not implemented these effects in our code. The parameters they use are:
7 cm -3 (differs from [17])
nE=
V E "- 450 km S-1
10 (w = 45 km s -1 where 1/2row 2 = kT; differs from [17]) noo ( H ) - 0.2 cm -3 (differs from [17]) noo ( H § ) - 0.3, 0.2, 0.1, 0.07, 0.04 cm -3 (includes case of 0.07 in [ 17]) Vo~ - 25 km s -~ (differs from [17]) Too ( H ) - Too ( H § ) - 5600K (differs from [17]), so that Moo- 2 (woo- 9.615 km s -~ and coo = 12.4 km s-~ for Tio n "- Telectron and y = 5/3. These are lower than currently accepted values but more realistic than the Baranov et al. [17] case) with ~cx - 10 -~4 ( 1.64 - 0.0695 In v )2 (v in cm s-l). We have set both Tio n > > T electronas well a s Tio n -- Telectronas in [7] and compare both cases with the kinetic approach. These are shown in Figure 3. M E -
4. C H A P M A N - E N S K O G E X P A N S I O N To find next-order corrections in collisional hydrodynamics, the Chapman-Enskog approach [15] is used yielding the following q(1) _ _)hV T (9) e/(1)--]-~1
o~j "[- oqxi
- 3
'~
where ( P)~j = P 8..,j+ P..(~,j~and Tr P..,j(~) = 0 by construction. In addition, to carry out fully the nextorder expansion we would need to include the next-order correction terms to the momentumand energy-exchange operators, e.g. _ O(q(1).AU) a n d - O ( AUop~I).AU ). 4.1. R e p u l s i v e P o w e r - L a w Potentials We evaluate the first-order coefficient of viscosity B~ and first-order thermal conductivity ~ using the results of [ 15] for power-law repulsive forces between atoms of monatomic gases, viz. /~'1 ~
q(1)_
15
k
-4-~'/1 u
(10)
2m5/11 V ( 3 k T )
- 94-
A fluid approach to the heliosphere/VLISMproblem
Figure 3. Comparisons of boundary locations using the fluid model described here and the kinetic model of [7]. The top panel shows the comparison for equal proton and electron temperatures and the bottom panel for a negligible electron temperature. The bottom panel also shows one case with a strong magnetic field. The boundary locations are plotted as a function of distance from the Sun along the stagnation line in the flow, i.e., opposite the direction from which the interstellar wind is blowing. Agreement is better for larger values of the interstellar proton density. Note that assumptions on electron temperature have marginal effect on the boundary locations.
-95-
R.L. McNutt Jr. et al.
and (see Chapter 8 of[18])
5~ ]"11-- -8
mw
Al(S)
1 (11)
acx('~W ) A2(s ) F(4 + })
For the hard-sphere model v = 0 and so we use
[.l 1 "-- 0 , 1 9 5 8 3
mw
Oc ( w) kw
'~'1 -" 0 . 7 3 4 3 6
(12)
Oc ( w)
4.2. Steady State Comparison of Terms
In steady state, the equations of motion for the neutral hydrogen are
v.(p.u.)=o V 9 (pnunun) + V P z - - V
*
p~l)_ QM, II+-->H
(13) V * [ ( E z + P.)un ] - - V , (q~' +
u*
p(HI')--QE,o,,t,H+_.>H
and we use these equations to find the fractional corrections introduced by the Chapman-Enskog terms. Note that since the mean free path )~fp~ (ncy)-1 we have P-- pw -- (/fpV) u
(14)
q - nw ~ (lfpV) kT
(15)
To estimate the magnitude of the corrections to the runs as made we can use Equations (9) and (13) to estimate the magnitude of the viscosity term along the stagnation line. This correction term is
xx "-- 3 0 ~
(16)
jUl
For the high-density case the contribution is substantial indicating a lack of self-consistency in the solution. In Figure 4 we show this correction term as well as the other pressure terms. It should be kept in mind that the physically relevant quantities are the gradients of each term as the gradients balance the momentum exchange in Equation (13).
- 96-
A fluid approach to the heliosphere/VLISMproblem Figure 4. Comparison of pressure terms along the upstream stagnation line for the high-density symmetric fluid model. The black line indicates the m a g n i t u d e of the viscosity correction in Equation (16). The solid line indicates the contribution in positive and the dashed line, negative. The lowest-density case has the same qualitative features but at a lower magnitude (by a factor-4).
5. NEED FOR M O D E L S W I T H TIME VARIATION
The recent large solar events once again may produce a global merged interaction region (GMIR) in the outer heliosphere. The speed at 1 AU on 15 July 2000 was -1050 km s-~ (from IMP 8, A. J. Lazarus, private comm.). This suggests that the heliospheric VLF radio noise [ 19] may well again appear in just over a year, if large particle events are seen by both the Voyagers first (signaling a truly "global" event). One can make rough timing predictions [20-22] of the noise onset time, but quantitative models can potentially give us another experimental location for the heliopause interface region [23]. Quasi-similarity models [22] suggest a shock may appear near Voyager 1 in early 2001 with another episode of radio emission --September 2001. It remains to be seen whether this will occur. In any case, in addition to future direct exploration of the VLISM by an interstellar precursor mission, this phenomenon is another reason for the continued investigation of quantitative models of the heliospheric cavity and its dynamics. 6. CONCLUSIONS The MHD/HD simulation run with a no magnetic field compares favorably with the HD/ kinetic approach in a qualitative sense. Quantitatively, the hydrodynamic simulation predicts a higher hydrogen wall in the upstream direction. Furthermore, the wall thickness and height are not as sensitive to the upstream proton density as in the HD/kinetic simulation. The MHD/HD simulation consistently predicts a closer external shock and further heliopause and termination shock than the kinetic approach. The two approaches tend to converge at sufficiently high upstream proton density -0.30 cm-3). Corrections to the lowest order HD equations still need to be explored quantitatively (ChapmanEnskog terms) to see if these may yet yield a moment approach that gives results consistent with the kinetic/HD approach. The variation of the charge-exchange cross section with collision speed can be accurately approximated as a power law and incorporated into the momentum- and energy-exchange operators. To the extent that we can assume the atoms behave as "Maxwell molecules" in the collision process, all moments of the Boltzmann collision operator are independent of the form of the distribution function. Hence, there is justification that an MHD/HD approach may allow a fast, yet accurate, computational treatment and allow for cross checks with kinetic computations.
-97-
R.L. McNutt Jr. et al.
To continue with exploring the possibilities of a fluid approach, the next step is to retain the isotropic-Maxwellian representation for the ions, but adopt the Chapman-Enskog approximation for the neutral distribution. In this case the distribution function has a first-order correction parameterized by the full pressure tensor and the heat flux vector. This distribution can then be used to calculate self-consistently the collision operator for transferring momentum and energy between the two populations. The modified energy and momentum equations can then be used to provide a next-order comparison with the kinetic approach. This is about as far as one can go with such fluid approximations. Further expansion of the fluid equations to next-order terms is possible [24], but typically this yields less-satisfactory results [15]. ACKNOWLEDGMENTS
Support was provided by the Voyager Interstellar Mission under NASA Contract NAG5-4365. REFERENCES
1. 2. 3. 4. 5.
T. E .Holzer, et al. The Interstellar Probe, NASA Publication, 1990. L.D. Jaffe and C. V. Ivie, Icarus 39 (1979) 486. E C. Liewer, R. A. Mewaldt, J. A. Ayon, and R. A. Wallace, STAIF-2000 Proc., 2000. R.A. Mewaldt, J. Kangas, S. J. Kerridge, and M. Neugebauer, Acta Astron., 35 (1995) 267. R.L. McNutt, Jr., R. E. Gold, E. C. Roelof, L. J. Zanetti, E. L. Reynolds, R. W. Farquhar, D. A. Gurnett, and W. S. Kurth, J. Brit. Inter. Soc., 50 (1997) 463. 6. R.L. McNutt, Jr., G. B. Andrews, J. McAdams, R. E. Gold, A. Santo, D. Oursler, K. Heeres, M. Fraeman, and B. Williams, STAIF-2000 Proc., 2000. 7. V.V. Izmodenov, J. Geiss, R. Lallement, G. Gloeckler, V. B. Baranov, and Y. G. Malama, J. Geophys. Res., 104 (1999) 4731. 8. R.L. McNutt, Jr., J. Lyon, and C. C. Goodrich, J. Geophys. Res., 103 (1998) 1905. 9. R.L. McNutt, Jr., J. Lyon, C. C. Goodrich, and M. Wiltberger, AIP Conference Proceedings, CP47 (1999) 823. 10. R. L. McNutt, Jr., J. Lyon, and C. C. Goodrich, J. Geophys. Res., 104 (1999)14,803. 11. G. E Zank, H. L. Pauls, L. L. Williams, and D. T. Hall, J. Geophys. Res., 101 (1996) 21,639. 12. G. E Zank, Space Sci. Rev., 89 (1999) 413. 13. L. J. Maher and B. A. Tinsley, J. Geophys. Res., 82(1977) 689. 14. J. C. Maxwell, Phil. Trans. R. Soc., 157 (1867) 49. 15. S. Chapman, and T. G. Cowling, The Mathematical Theory of Non-Uniform Gases, Third Edition, Cambridge University Press, New York, 1998. 16. S. Harris, An Introduction to the Theory of the Boltzmann Equation, Holt, Rinehart, and Winston, Inc., New York, 1971. 17. V. B. Baranov, V. V. Izmodenov, and Y. G. Malama, J. Geophys. Res., 103 (1998) 9575. 18. J. M. Burgers, Flow Equations for Composite Gases, Academic Press, New York, 1969. 19. W. S. Kurth, D. A. Gurnett, E L. Scarf, and R. L. Poynter, Nature, 312 (1984) 27. 20. R. L. McNutt, Jr., Geophys. Res. Lett., 15 (1988) 1307. 21. R. L. McNutt, Jr., Adv.Sp.Res., 9 (1989) 235. 22. R. L. McNutt, Jr., A. J. Lazarus, J. W. Belcher, J. Lyon, C. C. Goodrich, and R. Kulkarni, Adv. Space Res., 16 (1995) 303. 23. D. A. Gurnett, W. S. Kurth, S. C. Allendorf, and R. L. Poynter, Science, 262 (1993) 199. 23. D. Burnett, Proc. Lond. Math. Soc., 40 (1935) 382
-98-
Possible effects of the interstellar magnetic field on the heliospheric structure and H-atom penetration to the solar system V.B. Baranov a* ~Institute for Problems in Mechanics of the Russian Academy of Sciences, Prospect Vernadskogo 101-1, 117526 Moscow, Russia At present neither direction nor magnitude of the interstellar magnetic field in the vicinity of the solar system are experimentally known. However, there are many papers concerning theoretical problems of its effect on the solar wind interaction with the local interstellar medium (LISM). Critical review of such papers are presented here. Possible effects of the interstellar magnetic field on H-atom penetration to the solar system are discussed. The first results of self-consistent model solution taking into account mutual effect of plasma component, H-atoms and interstellar magnetic field are given in an axisymmetric approximation which is correct in the case when the vector of the LISM velocity is parallel to the vector of the LISM magnetic field intensity. 1. I N T R O D U C T I O N Constructing a quantitative theoretical model for the prediction and explanation of the experimental data is an important goal in various branches of scientific knowledge. However, such a model is useful if it has a reliable, physically correct theoretical basis; otherwise, an interpretation of experiments could be wrong. Let us consider, for example, a problem of the solar wind interaction with the supersonic interstellar gas flow. As it is known [1], four regions of the flow are formed in this case (Figure 1) for the plasma component (electrons and protons) due to the formation of the bow shock (BS), heliopause (HP) and termination shock (TS). Hydrogen atoms of the local interstellar medium (LISM) penetrate the solar system interacting with the plasma component due to processes of the resonance charge exchange. Their parameters are measured on the basis of the interpretation of the scattered solar Lyman-alpha radiation. The results of the interpretation depend critically on a theoretical model considered. Up to the beginning of 80-th the "hot" model was used to interpret experimental data (see, for instance, [2]). A "filter" effect of the interface between the solar wind and the plasma component of the LISM (regions 2 and 3 in Figure 1) is not taken into account in the framework of this model. The first theoretical papers ([3-6]) showed that this effect, predicted first qualitatively in [7], is very important. That is why invalid magnitudes of the LISM hydrogen number density, temperature and bulk velocity were predicted by the "hot" model. *This work was partially supported by CRDF Grant 6096, RFBR Grants 98-01-00955, 99-02-04025.
-
99-
V.B. Baranov
f
H sw (population 2) Moo>
1
H+ e
(4)
H
sw
(population 1 )
Figure 1. Qualitative picture of the solar wind interaction with the LISM without magnetic field. Regions 1 - 4 are the supersonic solar wind (preshock of the TS), the thermalized (in the TS) solar wind, the thermalized (in the BS) LISM plasma and the LISM plasma respectively. Dotted lines are trajectories of different H-atom populations.
The mutual effect of the plasma and hydrogen components was self-consistently taken into account in [4]. These authors used the simplest possible hydrodynamic scenario, where it was assumed that the velocity VH and the temperature TH of neutral hydrogen remain constant throughout the whole computational domain and that the LISM H-atoms are lost only in the regions 1, 2 and 3 (Figure 1) due to resonance charge exchange. This means that they solved only the continuity equation for neutrals. It was impossible to interpret experiments connected with the Lyman-alpha glow on the basis of this simplest scenario. The model suggested in [4] was criticized in [8] where the correct kineticgasdynamic model was suggested. H-atoms are kinetically described by the Monte Carlo method in this model, because the mean free path of the H-atoms is comparable with the characteristic length of the problem (for example, with the size of the heliopause) and the hydrodynamic approximation is not correct all over for the H-atom component. Recently, authors of [9,10] and [11] revisited the hydrodynamic description of H-atom motion in the problem of the solar wind interaction with the LISM although this description is not correct as we have seen above. It was shown in [12] and [13] that the distribution function of the all H-atom populations is not Maxwellian and, therefore, the multi-fluid model in [10] cannot be used to interpret experimental data. The same result is obtained in [29] for subsonic plasma-plasma interface. In Figure 2 (Figures 5c and 7b from [12]) some results of the kinetic-gasdynamic model in [8] and the multi-fluid model in [10] are compared. From Figure 2a we see that the bulk velocity (along the axis of symmetry) of the LISM hydrogen atoms (primary and secondary), which are responsible
-
100-
Possible effects of the interstellar magnetic field...
Figure 2. Comparison of some results obtained on the basis of the kinetic-gasdynamic model by Baranov and Malama (1993) and the multi-fluid model by Zank et al. (1996): (a) the component (along the axis of symmetry) of the H-atom (born in region 1) velocity and (b) the H-atom (born in region 1) temperature as functions on the heliocentric distance.
for the interpretation of the scattered solar Lyman-alpha radiation, changes its sign in the downwind direction for the model in [10]. In this case the continuity equation is not satisfied in the stationary approximation [10] and, therefore, the results of these authors are not physically real. Figure 2b demonstrates another physically unrealistic result obtained in [10]. Namely, the temperature of the energetic solar wind H-atoms, which are born due to the charge exchange between the LISM H-atoms and the solar wind protons (in the region 1), is two orders of magnitude larger (10 ~ K) than the solar wind proton temperature although there are no reasons for their heating. It follows that the model in [10] can also not be used to interpret results of planned experiments connected with the detection of energetic neutral atom (ENA) fluxes with 1 AU [14]. In the last few years several numerical magnetohydrodynamic (MHD) models of the solar wind interaction with the magnetized LISM were published although at present neither direction nor magnitude of the interstellar magnetic field in the vicinity of the solar system are experimentally known. The example of multi-fluid model results mentioned above (Figure 2) shows that caution must be exercised by observers to interpret experimental results on the basis of theoretical models. That is why the part of this presentation will be devoted to the critical review of MHD models of the solar wind interaction with the magnetized LISM (Section 2). In the Section 3 some problems of the interstellar magnetic field effect on H-atom penetration from the LISM to the solar system will be considered.
-101 -
V.B. Baranov
(a)
Hyperbolicregion
(b) Ellipticregion
(c)
Quasihyperbolic region
BS HI'
I-IP
Figure 3. Possible geometrical patterns of the MHD flow in the case when the interstellar magnetic field Boo and the bulk velocity V ~ are parallel. Hyperbolic, elliptic and quasihyperbolic regions are (a),(b) and (c) respectively.
2. A B O U T T H E F U N D A M E N T A L M H D P R O B L E M C O N N E C T E D W I T H THE INTERACTION BETWEEN THE SOLAR WIND AND MAGNETIZED LISM'S PLASMA It is known that a supersonic flow near hard bodies is characterized by a bow shock formation. The wings of the bow shock are weak discontinuities where parameters are continuous but their gradients are not. In doing so the weak discontinuities coincide with characteristics which are lines of small perturbation propagation. Their direction always coincides with a downwind direction in the classic gas dynamics. We have a more complicated situation in the magnetohydrodynamics (MHD) because there are four velocities of small perturbations (entropic, slow and fast magnetosonic and alfvenic) in the presence of the magnetic field. In particular, characteristics can be directed in the upwind direction as well as in the downwind one (see, for example, in [15]). The heliopause HP plays the role of the hard body in the problem of the solar wind interaction with the magnetized interstellar plasma. Possible geometrical patterns of the flow considered are qualitatively shown in Figure 3 in the most simple case when the magnetic field B~ and the bulk velocity V~ of the interstellar gas are parallel (see [16]). Figure 3a demonstrates a possible shape of the bow shock at V~ > max(a0, aA), where a0 and a A - are the sonic and alfvenic velocities of the interstellar plasma respectively, i.e. at the weak interstellar magnetic field. It is the hyperbolic region and analytical results give rise to real characteristics directed to the downwind direction (see wings of the bow shock in Figure 3a). To investigate the shape of the bow shock in the vicinity of the nose region numerical calculations must be used. All published numerical results (see,
-
102-
Possible effects of the interstellar magnetic field...
for example, [16-21]) are obtained in the case presented in Figure 3a (in the hyperbolic region). The geometrical pattern of the flow considered can be significantly changed at an increase of the magnetic field intensity. For example, the bow shock must be absent in the elliptical region which is characterized by inequalities Vo~ < aoaA/(a 2 + a~) -1/2
and
min(a0, aA) < Vo~ < max(a0, aA)
Real characteristics are absent in this region (see Figure 3b). Characteristics are directed in the upwind direction (see Figure 3c) if inequalities
a0a /(a +
<
< min(a0, aA)
are satisfied. It is a quasi-hyperpolic region [15]. There are no published calculations of the flow in the regions presented in Figure 3b and 3c, i.e. at large magnetic field intensities. Let us compare now the numerical results published in [19,21] and [20] for the case presented in Figure 3a. These results are not realistic for real heliospheric structure because they are obtained at n i -- 0 (nil is the number density of hydrogen atoms). However, they are interesting for fundamental problems of MHD. For comparisons we used only the results M ~ = 2 and Alfven Mach number A~ = 1.4 in the interstellar plasma ( M - V/ao, A - V / a A ) . In the Figure 4 the lines p - const are presented, where p is the static pressure. As we see from Figure 4, numerical results of different authors give rise to the different structures of the MHD flow between the bow shock and heliopause (in the nose region of the heliosphere). For example, the bow shock splitting (A-like structure) is observed (Figure 4a) in [19] as well as in [17]. We see a different feature of the bow shock splitting in Figure 4b obtained in [21]. We can not interpret a character of the flow from Figure 4c obtained in [20] due to a not well resolution of these authors numerical method. The pictures presented in Figure 4 are obtained at the use of different versions of the shock capture numerical method. In [16] the shock fitting method was used where relations on the discontinuities (BS, HP and TS) were exactly satisfied. Authors of [16] did not obtain the bow shock splittings, presented in Figure 4a and 4b, and criticized the numerical method used in [17]. However, their calculations were made at A~ - 1.5. From Figure 4 we also see that the complicated structure of the flow in the downwind direction connected with a formation of the Mach disc disappears in the presence of the magnetic field. We interpret different results of different authors, presented in Figure 4, with numerical problems rather than as physical ones. Unfortunately, one can notice that the unpleasant tendency to skip the results of other authors have been formed last time in the heliospheric theoretical community. We tried to correct this tendency. We considered here only the axisymmetric MHD problem to underline difficulties of this flow calculations and their interpretations even for the most simple situation although 3D MHD problems are already considered in literature (see, for example, [20,22]). However, H-atom component and, in particular, a important effect of the resonance charge exchange is not taken into account in these publications. That is why their applications to the problem of the solar wind interaction with the LISM are not physically realistic. Below
-
103-
V.B. Baranov
Ratkiewicz
et al. (1998)
Figure 4. Comparison of the MHD geometrical patterns obtained by different authors in the case when the interstellar magnetic field Boo and the bulk velocity V ~ are parallel (the hyperbolic region)" (a) static pressure (below the symmetry axis) and logarithm density isolines, (b) static pressure (upper half) and density isolines, (c) static pressure isolines.
we will consider some results of the kinetic-MHD model obtained in [23] where H-atom component was self-consistently taken into account for axisymmetric problem. 3. P O S S I B L E E F F E C T OF T H E I N T E R S T E L L A R M A G N E T I C F I E L D ON H - A T O M P E N E T R A T I O N F R O M T H E LISM T O T H E S O L A R S Y S T E M The first 3D MHD numerical model taking self-consistently into account the LISM hydrogen atoms and solar and interstellar magnetic fields was constructed in [24]. Accurate results were obtained due to excellent numerical scheme. However, to simulate the interaction of the H-atom and plasma components via charge exchange processes authors of [24] used the simplest possible scenario that was originally suggested in [4] (see Section 1). Therefore, the model in [24] can not be used for the interpretation of H-atom observations (see below). In particular, the "hydrogen wall" predicted in [25] and first observed in [26] cannot be obtained in the framework of the model in [24] neglecting the secondary LISM
-
104-
Possible effects of the interstellar magnetic fieM...
-
A
oo
=o~
9
..............
6oo
-
A ~
=
1.18
Ao0
=
0.9
t
- -
-I
-
. m
Z,
a.u.
Figure 5. Effect of the interstellar magnetic field on the geometrical pattern of the MHD flow at nH~ --0.2cm -3 in the axisymmetric problem.
H-atoms. Let us consider now the self-consistent axisymmetric MHD model of the solar wind interaction with the magnetized two-component (H-atom and plasma components) LISM's gas calculated in [23]. The model constructed in [8] is extended by these authors to include the interstellar magnetic field. The plasma component is described by MHD equations with source terms taking into account its interaction with H-atoms. Trajectories of Hatoms are calculated by Monte Carlo method suggested in [27] to calculate the source terms in MHD equations. To solve the self-consistent problem as a whole the numerical method suggested in [19] for solving MHD equations and the method of global iterations are applicated. This model and the model in [8] differ by the presence of the interstellar magnetic field, which is parallel to the LISM plasma bulk velocity. The some results of the model in [23] are presented in Figure 5 - 7. The change of the geometrical pattern with increasing the interstellar magnetic field intensity (with decreasing the Alfven Mach number Ao~) in Figure 5 is shown. Comparisons with the MHD results obtained in [16], where the neutral component (nil = 0) was not taken into account, show that the character of the geometrical pattern change is qualitatively retained at n H ~ 0 although a presence of the bow shock at sub-alfvenic interstellar gas velocity (A < 1) is a surprise which authors could not explain physically. It is interesting to note here that the problem of the bow shock splitting (see Section 2) is absent in the presence of the neutral component due to the effect of resonance charge exchange processes. There are 4 populations of H-atoms in the problem considered. Populations 1-4 in the regions 1-4 (Figure 1) are formed respectively. Effect of the interstellar magnetic field on the "hydrogen wall" formed by Populations 3+4 is presented in Figure 6. We see that
-
105-
KB. Baranov
(a) 0.3
0=0
o,
,
. . . . .
A = = 1.18
.......
A ~ = 1.18
Ao~= 0.9
. . . . .
A ~ = 0.9
A
R,
R,
a.u.
a.u.
Figure 6. Effect of the interstellar magnetic field on the "hydrogen wall"" (a) in the upwind direction and (b) in sidestream direction.
this effect is small although it is any lager in the sidestream direction than in the upwind one. Magnitudes of the interstellar magnetic field considered here (from 0 till 3.5.10 -6 G) cannot change an interpretation of the hydrogen wall in [26]. The effect of the interstellar magnetic field on the distribution of the Population 2 number density is presented in Figure 7. We see that this effect is important. Therefore, it can influence the interpretation of the Lyman-alpha absorption spectrum of stars in the downwind direction [28] and the interpretation of the heliosphere ENA image with 1 AU
[14]. 4. C O N C L U S I O N On the basis of the results presented above we can make following conclusions: 1. Examples presented in this paper show that a caution must be exercised by observers at the interpretation of their experimental data on the basis of numerical models. 2. The most published MHD numerical models do not take into account H-atom component. Therefore, they are physically not applicable to the problem of the solar wind interaction with the LISM. In addition, their results are only obtained for weak interstellar magnetic fields (in the hyperbolic region). 3. First results of the self-consistent MHD problem solution (H-atoms, plasma component and the interstellar magnetic field) in axisymmetric approximation show that the effect of the magnetic field on the "hydrogen wall" is not important as the magnetic field is changing from 0 till 3.5-10 -6 gauss. However, this effect can be important for the interpretation of the ENA-fluxes from 1 AU (see [14]).
-
106-
Possible effects of the interstellar magnetic field...
'
o5) O = O
~
i
. . . . . . . .
Ao==
l:.
t
. . . . .
A=o = O.9
'
t.'
'
i!
~
i "1,
1.18
O
=
7"1;/2
i .: [
'
........
Ao== Aoo=
oo 1.18
. . . . .
Aoo=
0.9
i'! i- !
!'i l:i
/ii ~"~r~
-
,~ t _,r."
, ;:
" i - i -
i
;J
-".i t .
S'o.
.-:.- .......
o
i
|
R,
~ ' ~ ' ~ " :
-'1" -" "=" "
:~
|
o
|
a.u.
(o O.O06
""~" " ~ ""
-
O
=
.-""
n
..--
0.005
ooo~ ~ ...... o.ool
o
i
i
. . . . . . . .
A==
1.18
. . . . .
A~
0.9
i
R,
i
i
a.u.
Figure 7. Effect of the interstellar magnetic field on the number density distribution of the Population 2 hydrogen atoms: (a) in the upwind direction, (b) in the sidestream direction and (c) in the downwind direction.
REFERENCES 1. V.B. Baranov, K.V. Krasnobaev, and A.G. Kulikovsky, Sov. Phys. Dokl. 15 (1971) 791. 2. J.-L. Bertaux, R. Lallement, V.G. Kurt, and E.N. Mironova, Astron. Astrophys. 150
(1985) 3. V.B. Baranov, M.G. Lebedev, and M.S. Ruderman, Astrophys. Space Sci. 66 (1979) 441. 4. V.B. Baranov, M.K. Ermakov, and M.G. Lebedev, Sov. Astron Lett. 7 (1981) 206. 5. H. Ripken and H.J. Fahr, Astron. Astrophys. 122 (1983) 183. 6. H.J. Fahr and H.Ripken, Astron. Astrophys. 139 (1984) 551. 7. M. Wallis, Nature, 254 (1975) 207. 8. V.B. Baranov and Yu.G. Malama, J. Geophys. Res. 98 (1993) 15,157. 9. H.L. Pauls, G.P. Zank, and L.L. Williams, J. Geophys. Res. 100 (1995) 21,595. 10. G.P. Zank, H.L. Pauls, L.L. Williams, and D.T. Hall, J. Geophys. Res. 101 (1996) 21,639.
- 107-
V.B. Baranov
11. G . P . Zank and H.L. Pauls, Space Sci. Rev. 78 (1996) 95. 12. V.B. Baranov, V.V. Izmodenov, and Yu.G. Malama, J. Geophys. Res. 103 (1998) 9575. 13. V.V. Izmodenov, M.A. Gruntman and Yu.G. Malama, J. Geophys. Res. (2001) in press. 14. M.A. Gruntman, COSPAR Coll. on: The Outer Heliosphere: The Next Frontiers (Potsdam, Germany, 24- 28 July, 2000), Abstracts (2000) p.16. 15. A.G. Kulikovsky and G.A. Lyubimov, Magnetohydrodynamics (1965), AddisonWesley, Reading. 16. V.B. Baranov and N.A. Zaitsev, Astron. Astrophys. 304 (1995) 631. 17. Y. Fujimoto and T. Matsuda, KUGD91 - 2 (1991). 18. N.V. Pogorelov and A.Y. Semenov, Astron. Astrophys. 321 (1997) 330. 19. A.V. Myasnikov, Preprint IPM RAS (1997) No. 585. 20. R. Ratkiewicz, A. Barnes, G.A. Molvik et al., Astron. Astrophys. 335 (1998) 363. 21. N.V. Pogorelov and T. Matsuda, Astron. Astrophys. 313 (2000) 697. 22. N.V. Pogorelov and T. Matsuda, J. Geophys. Res. 103 (1998) 237. 23. D.B Alexashov, V.B. Baranov, E.V. Barsky, and A.V. Myasnikov A.V., Pis'ma v Astron. Zh. (2000) No. 11 (Astronomy Letters, English translation). 24. T.J. Linde, T.I. Gombosi, P.L. Roe et al., J. Geophys. Res. 103 (1998) 1889. 25. V.B. Baranov, M.G. Lebedev, and Yu.G. Malama, Astrophys. J. 375 (1991) 347. 26. J.L. Linsky and B.E. Wood, Astrophys. J. 463 (1996) 254. 27. Yu.G. Malama, Astrophys. Space Sci. 176 (1991) 21. 28. V.V. Izmodenov, R. Lallement, and Yu.G. Malama, Astron. Astrophys. 342 (1999) L13. 29. R. Osterbart and H.J. Fahr, Astron. Astrophys. 264 (1992) 260.
- 108-
Interstellar Gas F l o w Into the Heliosphere E. M6bius a, Y. Litvinenko a, L. Saul a, M. Bzowski b, D. Rucinski b aDept, of Physics and Space Science Center, University of New Hampshire, Durham, NH03824, USA bSpace Research Centre, Polish Academy of Sciences, Warsaw, Poland Flow speed, temperature, composition of the local interstellar medium and its interaction with the heliosphere, leave their imprint on the gas flow pattern through the inner solar system. Through diagnosing resonant scattering of light by the neutral gas, measuring the distribution function of pickup ions, and imaging directly the gas flow we can tap this information. Recently, even a direct signature of the neutral gas flow has been identified in the pickup ion cutoff that can be used to derive the flow velocity. In particular for helium all these methods complement each other, their consistency can be tested and a firm parameter be derived. All of these methods carry information not only about the spatial distribution of the gas, but also about its original velocity distribution. Helium is not significantly affected by the heliospheric interface and thus provides an almost unbiased view of the nearby interstellar medium. Conversely, hydrogen and oxygen are slowed down and heated by charge exchange in the interface region, thereby producing an altered flow pattern, which allows us to unravel these interaction effects. We will review current attempts to combine different in-situ observations of pickup ions and neutral gas distributions and outline applications to derive the physical processes in the heliospheric interface region. 1. MOTIVATION TO STUDY THE LOCAL INTERSTELLAR GAS The local interstellar medium (LISM) has received increasing attention over the past decade for a variety of reasons. Firstly, the interstellar gas constitutes the raw material of stars, planets and ourselves. This causal connection has sparked our interest in the very nature and composition of interstellar material, which also holds the key to its origin and evolution, generally referred to as nucleosynthesis. All chemical elements and their isotopes were or still are synthesized in three principal environments: 1) Light nuclei, such as H, He, their isotopes, and some 7Li (e.g. [1]), were produced in the Big Bang. 2) Stars synthesize C and all the heavier elements [2], but destroy D [3]. 3) Through spallation, galactic cosmic rays produce very rare elements, e.g. Be and B, in the interstellar medium [4]. The natural sequence of this still ongoing element and isotope formation leads to a continuous increase of the species formed under 2) and 3), which can be traced in the history of the galactic material. In fact, all material that belongs to the solar system has frozen in the state at the sun's birth about 4.5 billion years ago, while a sample of the LISM represents today's galactic material. A detailed comparison of the elemental and isotopic composition of these two disparate samples will provide much needed constraints for the models of galactic matter evolution. Secondly, the LISM is our immediate neighborhood in the galaxy. It imposes the outer boundary conditions on the heliosphere, a region, which is carved out of the interstellar me-
-
109-
E. MObius et al.
dium by the solar wind and its imprinted interplanetary magnetic field (IMF). A balance between the solar wind and the pressure of the LISM determines the size of the heliosphere. Currently, the space environment of the Earth is controlled entirely by the solar wind. The large size of the heliospheric bubble keeps the overwhelming majority of energetic cosmic radiation outside our planetary system and only a small fraction of the most energetic particles reach the Earth. If we consider the relative motion of the Sun and the various local components of the LISM an interesting scenario emerges. Our heliosphere was probably even larger several hundred thousand years ago when the sun was in a very dilute and hot environment, but may shrink substantially a few hundred thousand years from now, when it encounters interstellar clouds of much higher density [5, 6, 7]. Turning our view to our galactic neighborhood, the heliosphere also impacts the surrounding medium by diverting the flow and by filling the vicinity with additional energetic particle populations, generated at its boundaries. In essence our heliosphere is an example - in fact the only example that can be closely examined - o f the more general phenomenon of astrospheres. As an indication for an astrosphere, the diversion and deceleration of the interstellar gas flow appears as an additional Doppler-shifted component in absorption lines of interstellar H in the light of the respective star. Observations of astrospheres in our neighborhood have recently been reported by Linsky [8] and Wood et al. [9]. In order to tackle these important questions, the elemental and isotopic composition as well as the physical parameters, density, temperature, bulk flow and ionization, of key species in the LISM must be determined. To date we have only direct access to local interstellar material that penetrates the boundaries of our system or to line-of-sight integrals that provide an average over several light years. Given these constraints, it is obvious that our current knowledge of the LISM immediately outside the heliosphere is limited and efforts in various directions are needed to improve this situation. Since the heliosphere is impenetrable to interstellar plasma, only the neutral component of the LISM is currently accessible. Even this access is substantially biased by filtering of several species, such as H and O, in the heliospheric interface through charge exchange with the interstellar plasma flow [10, 11, 12, 13]. However, the amount of filtering, a related deceleration and heating of the interstellar material during the passage of the interface, and the degree of ionization of H are causally connected. Therefore, a careful comparison of the observed flow characteristics of species affected by resonant charge exchange (e.g., H and O) with those that are not (e.g., He and Ne) will provide the necessary information. In addition, the comparison of the physical parameters obtained for several species will provide insight whether the LISM is isothermal or non-thermal effects are important. Finally, the LISM may be inhomogeneous on much smaller scales than indicated above. Such inhomogeneities can only be identified by long term observations of the physical parameters. This can be achieved by extension of our data sets into the future and by providing the tools to interpret earlier data sets quantitatively so that a comparison with current results is possible. In this paper we will first summarize the current status of LISM observations and compare different observation methods. We will then describe briefly a coordinated observation campaign aimed at the physical parameters of interstellar He. Finally, we will concentrate on the determination of the LISM flow from pickup ion and neutral gas observations and how this can help to identify interface effects. 2. STATUS OF LISM OBSERVATIONS Of the LISM components, helium is the species for which the neutral parameters can be determined from inside the heliosphere without modification [14, 15]. Helium is also very intriguing because of recent results that it appears to be more highly ionized than hydrogen, in
-110-
Interstellar gas flow into the heliosphere spite of its higher ionization potential. Currently, this has not been reconciled yet with the local radiation environment in the LISM [16]. Over the last fifteen years the tools for the determination of the LISM parameters, density, temperature and bulk flow vector, have been dramatically improved and diversified. The traditional method to observe interstellar neutral He inside the heliosphere by means of resonant scattering of solar UV, first reported by Weller and Meier [ 17], has been augmented by direct particle measurements, using pickup ion distributions [18, 19] and neutral atom imaging [20]. Yet the derivation of the interstellar parameters with all these methods relies on a detailed modeling of the gas flow through the heliosphere and ionization losses on its way. Because of the modeling steps involved in each of the methods and the need for additional information on the sun's radiation and solar wind environment, results of these methods have differed to varying degrees in the past. In addition, the influence of the sun and the heliosphere on the gas distribution changes with the solar cycle, a subject of great interest in itself. A compilation of the physical parameters of He as obtained with the different methods can be found in Table 1. Observationally, these methods can be subdivided into two groups: in-situ methods, i.e. direct detection of the neutral atoms or their pick-up ions in the solar wind, and remote sensing methods, i.e. resonant scattering of solar UV at gas inside the heliosphere or line absorption of stellar UV. Each of these methods provides different raw observations that require substantial physical interpretation and extensive modeling in order to derive the physical parameters. Among these, the use of interstellar UV absorption lines in the spectrum of nearby stars is distinctly different from the other methods. It provides an average temperature and bulk flow velocity along the line-of-sight from the observed line profile over several light years and may be used to assess the homogeneity our neighborhood. From the observations inside the heliosphere one can either use directly the flow velocity information, when available, or the spatial distribution of the neutral gas to derive flow speed and temperature in the LISM. The observed flow velocity distribution in the inner heliosphere allows a direct extrapolation to the flow parameters at the boundary by invoking Keplerian trajectories of the neutral gas in the sun's gravitational field. The flow speed can be obtained by measuring the deflection of the neutral atoms in the sun's gravity [20], from the Doppler shift of scattered UV (e.g. [21 ]), and from the shift of the pickup ion cut-off compared with its nominal value of two times the solar wind speed [22]. So far the particle observations were made for He, while the UV observation has been obtained for H. Similarly, the spatial Table 1" Observations of interstellar helium parameters structure of the He Pick-up Ions Neutral Atoms ln-Situ Pick-up Ions distribution in the ( A M P T E SULEICA) (SWICS) (Ulysses G AS) Methods inner heliosphere r/He (cm -3) 0.009 - 0.012 0.0153+0.002 0.015-0.017 that is formed by the (Ulysses) combined action of 7850+550; 9900+500 5800- 7600 4800- 7200 THe (K) (corrected for diffusion) the sun's gravita25.3 + 0.4 23 - 30.5 VreI (km/s) tional field and ioni(corrected for diffusion) zation by solar UV [25], [38] [20] Reference [39], [42] and solar wind conRemote SensUV Absorption UV Scattering UV Scattering tains the imprint of ing Methods (non-local) the bulk flow and 0.015 0.02 0.0055 0.0145 nile (cm-3) temperature at the 16000 8OO0 7000 • 200 THe(K) boundary. Most pro(shifted UV line) minently, a gravita24 - 30 19-24 25.7 • 1 V~-el(km/s) tional tbcusing cone (shifted UV line) of enhanced He denReference [54] [521 [531
E. MObius et al.
sity is formed on the downwind side of the sun with respect to the flow. From a deconvolution of the UV scattering intensities (e.g. [23]) and from the spatial distribution of pickup ions along the Earth's orbit (e.g. [24]) the density, temperature and bulk flow velocity are derived. Surprisingly, the temperature and bulk flow speed, as derived from direct neutral gas flow observations inside the heliosphere and from absorption line profiles, seem to agree within their respective uncertainties. This might indicate that the conditions at the location of the solar system are well represented by the average in our neighborhood. However, there are still noticeable differences between different methods and even individual observations within a method, if the shape of the gravitational focusing cone is used. These differences call for a careful assessment of observational and modeling uncertainties. Each of these observations and methods has their own strengths and weaknesses. Ideally they could complement each other in the attempt to determine a benchmark set of parameters. Direct observation of neutrals promises the most accurate account of the LISM parameters, but is currently limited in its application to He. Conversely, pickup ions and UV scattering show greater potential towards widespread use among different species and long term observations. Therefore, a successful cross-calibration of all methods will substantially enhance our capabilities. Pickup ions can be observed from a variety of species. In addition to He, the density of H has been derived from pickup H + [19], and the abundance of N, O and Ne in the LISM has been inferred from N +, O +, and Ne + [25, 26]. These results make pickup ions a very versatile tool for the detailed study of the interstellar gas bulk parameters as well as the elemental and isotopic composition, provided the transfer and transport processes from the original interstellar neutral gas population to the measured pickup ion distributions are fully understood. Likewise density, bulk velocity and temperature of H have been derived from observations of scattered Lycz radiation (e.g., [21], [23]). It should be noted here that UV scattering observations are available for H and He since the early 1970' s, i.e. have the longest history. However, there are substantial differences in the results obtained from different data sets. Some of the differences may be due to the quality of the data, which certainly improved over time, but some of them may Table 2: Comparison of LISM observation methods also be connected to intrinsic difficulties of the methods. Strengths Weaknesses The respective strengths and Pickup Ions weaknesses of each method 9Many species (H, He, N, O, Ne, ) ~ Requires knowledge of have been compiled in Table 9 Spatial distributions, Vion "-) Ionization rates 2. As pointed out above, 9 Benchmark He density: He2+ ~ Transport effects pickup ions show the promise 9Long history:1984 - today ~ Unknown variations of fluxes 9 Connection to energetic ions of the most versatile data set in terms of LISM species. Imaging of the Neutral Gas Flow Instruments with large col9 So far only He 9Most precise observations may be extended to O, H, -) Bulk Speed lecting power are possible, and possibly others -') Direction and observations are available 9Minimum S/C- Neutral Temperature that go back to 1984. From velocity 9 Density (good calibration) the pickup ion energy spec~Vionfrom indirect trajectories trum information on the local UV Scattering ionization rate and the spatial 9 Long history: 1970- today ~Very sensitive to solar line distributions can be extracted. ~ Relatively small instruments profiles and variations In addition, a b e n c h m a r k 9 3-D view of spatial distribution 9Only 2 species (H, He) value for the He density has UV A bsorption been derived from He 2+ 9Variety of ISM species 9No Local ISM values pickup ions, because their 9 Connection with Astrophysics
Interstellar gas flow into the heliosphere production by solar wind He 2+ allows observation of both constituents with the same instrument and thus eliminates most calibration uncertainties [27]. However, this method requires independent knowledge of production rates for most species and is subject to a variety of transport effects. Also pickup ion fluxes still show a substantial variation on a wide range of time scales that is not fully understood and is usually taken out by long integration times. The most detailed and direct information on bulk speed, direction and temperature can be deduced from observations of the angular distribution of the interstellar gas flow on the downwind side of the sun. The method makes use of the dependence of the gravitational deflection of the neutrals on their original velocity [20]. In addition, the separation of direct and indirect Keplerian trajectories allows an independent determination of the total ionization rate. However, density determination requires elaborate absolute calibration. The minimum detection energy for atoms limits the application to parts of spacecraft orbits, where the interstellar gas and spacecraft velocities add up favorably. In addition, the method has only been applied to He so far. An extension to O and H appears possible with some development [28, 29]. Yet species, which cannot be easily converted into a negative ion, such as N or Ne, may not be accessible at all. The resonant UV scattering methods have the longest history, rely on relatively small instruments and provide a 3-dimensional view of the spatial distribution [30]. However, the results are very sensitive to the original solar line profiles and their variation. The radiation transport problem for Lycz has recently been treated by Scherer et al. [31]. This also requires a complex deconvolution, including among others careful modeling temporal variations of neutral gas density and velocity within a few AU from the sun [32, 33]. Direct velocity determination via the Doppler effect is only possible for H so far. In addition, only the two main species of the LISM seem to provide enough sensitivity for the method. An extension of this method, i.e. imaging the heliopause in the light of ionized O has been proposed by Gruntman and Fahr [34]. It will become possible with a sensitivity improvement by two orders of magnitude [35]. 3. COORDINATED OBSERVATION CAMPAIGN FOR INTERSTELLAR HELIUM With a unique combination of heliospheric and astronomical spacecraft it has become possible for the first time to employ all methods simultaneously, with the great advantage that they can complement each other with their capabilities. Each year in early December the Earth, and with it all earthbound spacecraft, pass through the interstellar focusing cone, whose structure depends critically on the interstellar parameters and the ionization in the heliosphere. However, the passage in December 2000 provides the opportunity for the most complete coverage. It is also close to maximum solar activity. The combination of UV observations is depicted in Fig. 1. EUVE, still in operation, provides a 3-D view of the cone in the anti-sunward direction [36, 37], while SOHO UVCS scans the cone very close to the sun, where it is most sensitive to the ionization (Raymond, private communication, 2000). Because the LISM flow is inclined by ~ 7 ~ w.r.t, the ecliptic plane the maxima of the scattered light appear below the ecliptic and below the sun, respectively. At the same time ACE, Wind, and Geotail provide a complete and contiguous spatial profile of the pickup ions at 1 AU. SOHO CELIAS SEM observes directly the UV ionization rate [38] and SOHO EIT provides a full disk image of the He II 304 A line, which contributes substantially to the He ionization rate. SOHO CDS and SUMER obtain the disk image and line profile of the He I 584 A line that controls the intensity of the scattered light. In addition, the relative speed of the LISM flow and Ulysses exceeds the observation threshold for the GAS sensor again, and direct neutral observations can be obtained. This combination lends itself to a comprehensive 3-dimensional imaging of the cone from close to the sun to beyond 1 AU with simultaneous in-situ observation of the velocity distribution of the neutrals. The coordinated effort is supported through a Scientific
-113-
E. MObius et al. Team at the International Space Science Institute (ISSI) with contributions by: J.-L. Bertaux, S. Chalov, H. Fahr, G. Gloeckler, R. Kallenbach, R. Lallement, E. MObius (Chair), D. McMullin, H. Noda, J. Raymond, D. Rucinski, W. Thompson, 7". Terasawa, J. Vallerga, R. yon Steiger, M. Witte. How the different data sets complement each other in the comparison with the modeled interstellar gas distribution to deduce the interstellar parameters is shown in Fig. 2. The combination of the neutral gas and pickup ion observations will provide the most accurate results on the He density and flow vector. On the other hand UV observations will add the most accurate account of the ionization processes, because they reach closest to the sun. In particular, with SOHO SEM and EIT the photoionization rate is measured directly and simultaneously, so that the still largely unFig 1: In-ecliptic (top) view and anti-sunward known electron impact ionization can be deview (lower left) of the EUVE pointing, rerived. A main weakness of the UV scattering spectively sunward view (lower right) of the technique has been its strong sensitivity to the UVCS pointing for the coordinated UV scatintensity, location and profile of the relevant tering observations of the gravitational focusing solar line, in this case the He I 584 A line. cone of interstellar He. The extent and location Observations of SOHO CDS and SUMER of the cone is shown schematically by gray will provide the information to remove these shading. uncertainties. Recent observations of the He cone with ACE SWICS [39] provide a substantially improved basis for the derivation of the He flow parameters than the widely spaced observations with AMPTE IRM [24]. Yet transport effects that lead to notably anisotropic pickup ion distributions [40, 41], in particular during times of radial IMF orientation, are likely to modify the spatial distribution of pickup ions in comparison with the original neutral distribution [42]. Because of incomplete pitch angle scattering pickup ions are not immediately convected with the solar wind, leading to a non-radial transport of the interstellar ions. This results in a redistribution of the pickup ions with respect to the neutral profile and thus effectively to an apparent widening of the focusing cone. In addition, substantial variations of the pickup ion fluxes have been observed on a variety of time scales [43], whose origin is not fully understood yet. Therefore, long term averaging has been applied to the data before they are used to derive LISM parameters. While this appears to have a minimal effect in the determination of the LISM density, a systematic effect on the apparent cone width and thus the derived temperature and bulk velocity can be expected. A careful modeling of the pickup ion transport [44, 45, 46] will bridge this gap in understanding, and the comparison of the pickup ion results with the other methods will provide a handle on the transport effects themselves. The neutral gas and He 2+ pickup ion observations with Ulysses provide an independent test for the flow characteristics and absolute density of He, respectively. Reconciliation between the complementary measurement methods raises the expectation that an extended database can be used to study temporal variations of the heliosphere - LISM interaction due to the solar cycle and potential LISM variations. UV scattering observations go
Interstellar gas flow into the heliosphere
Fig 2: Interrelationship of the different complementary observations of the interstellar He parameters together with required auxiliary observations and modeling efforts. back to the 1970's, and those of pickup ions to 1984. It was pointed out recently that also electrostatic analyzers without mass resolution have the capability to distinguish pickup ion distributions (Noda and Terasawa, private communication, 2000) as another welcome addition to the database. 4. OBSERVATION OF THE ISM NEUTRAL SPEED FROM PICKUP IONS
Recently, direct evidence for the influence of the neutral gas flow on the pickup ion distribution has been presented [22]. In contrast to a model in which pickup ions start from neutrals at rest, the cut-off speed is shifted beyond the canonical value of 2V~w in the upwind direction of the LISM flow, most notably for slow solar wind. The reason for this shift is that in the rest frame of the solar wind the injection velocity of pickup ions is the vector sum of the solar wind and the local gas flow. In the upwind direction of the flow these two velocities add, forming a sphere in velocity space whose diameter is increased by the ratio of the neutral and solar wind speeds. Conversely, the sphere is smaller on the downwind side with a cut-off at values lower than 2V~w. Therefore, the cut-off shift varies distinctly along the Earth's orbit. It should be noted that the neutral speed at 1 AU is defined by the original flow speed VBulkin the LISM and the gravitational acceleration to 1 AU. Fig. 3 shows the relative shift of the pickup ion cut-off Av/Vsw as a function of the day of the year (DOY), as obtained with SOHO CELIAS in partially overlapping 30-day intervals. For the analysis all time periods with 300 < Vsw < 350 km/s and an angle between solar wind and IMF of >60 ~ were accumulated. Av/V~w was derived for each interval by modeling the pickup ion spectrum with the appropriate shift to match by eyeball fit the observed cut-off. A suprathermal tail of the pickup ions [22] was ignored, but it was assumed that the steepest part of the cut-off is unaffected by the tail. The lines show the expected cut-off shift of the correct radial flow component at 1 AU for VBulk= 38 km/s at the heliospheric boundary and solar wind speeds of 300, 325 and 350 km/s, with the center line representing the best fit to the current data set. Translating the width of the chosen solar wind
E. MObius et al.
speed interval into an estimated uncertainty of the bulk flow, our result is equivalent to VBulk= 38+7 km/s. The observation agrees qualitatively with the expected variation along the Earth's orbit. However, the value is significantly higher than the reported LISM flow speed of ~25 km/s and contains a relatively large error. A potential reason for the relatively large value of VBulk may be the following. We have ignored any effect of suprathermal ions on the cut-off, and no energy diffusion of pickup ions is included in our model. Both effects would lead to higher cut-off values. In a data set that contains observations on the upstream and the downstream side of the flow these effects can be separated from the bulk flow effect, because the latter switches sign. Therefore, better coverage along the Earth's orbit and a more detailed analysis that includes all transport effects of the pickup ion distribution [46] will probably improve this method significantly over this crude analysis. However, the determination of the flow speed from pickup ions will likely be much less accurate than from direct neutral gas observations.
Fig. 3: Relative cut-off shift Av/Vsw of the pickup ion distribution along the Earth's orbit with SOHO CELIAS in 1996. The horizontal bars indicate the accumulation of intervals. The data points have been obtained as the best fit of the measured distribution at the cut-off with a model distribution that was shifted accordingly ([22]). The lines give the predicted shift for different Vsw with a heliospheric longitude of the flow vector of 75 ~
5. PERSPECTIVES FOR NEUTRAL GAS OBSERVATIONS As pointed out before, the most detailed information about the interstellar flow can be obtained from observations of the velocity distribution of the neutral gas under investigation, in the inner heliosphere. Keplerian trajectories provide a unique transformation for the original velocity distribution after the passage through the interface. Therefore, the distribution at the interface can be constructed from these observations. Current observation methods for neutral atoms can provide precise angular images of the flow, but do not retain adequate energy resolution. This is true for the observation of interstellar helium through sputtering off a LiF surface as used in the Ulysses GAS instrument [20] and for the proposed negative ion conversion instruments for species, such as O and H (e.g. [47], [27]). Therefore, the differential deflection of the interstellar neutral gas flow in the gravitational field of the sun is utilized to convert speed differentials into an angular image. The basic idea is illustrated in Fig. 4. As observed from a specific location, slower gas is deflected stronger, which turns a velocity distribution into an image. The analysis of such images through a deconvolution technique has been described in detail by Banaszkiewicz et al. [48]. The very precise temperature and bulk flow values for LISM He have been derived in this way. If the velocity distribution of H and O were studied with the same scrutiny, this would return decisive information about the filtration of these LISM components, which is needed to derive exact neutral densities and the ionization-state of the LISM. The velocity distribution contains unique information about the related processes, because filtration is accompanied by an effec-
Interstellar gas flow into the heliosphere
tive slowdown and heating of the gases compared with their original state [12]. In detail, neutrals are first lost to the plasma flow by charge exchange and now contribute to a rather hot plasma distribution that is diverted around the heliosphere. In turn ions of this hot flow are converted to neutrals, of which a good fraction are directed towards the inner heliosphere. These neutrals constitute an additional secondary and rather hot gas distribution that is added to the original depleted flow. Because the fraction of these secondaries and their distribution scale with the column density of the heliospheric interface, their observation holds a key to understanding the processes and to constraining the overall effect.
Fig. 4: Schematic view of interstellar neutral trajectories in the inner solar system for different inflow speeds at the heliospheric boundary observed from a spacecraft at 135 ~ from the ISM apex on a 1 x 3 AU orbit. Lower speed leads to stronger deflection in the sun's ~ravitational field.
In order to estimate the observable effect and to define the parameters of the needed instrumentation we have simulated the neutral O distribution in the inner heliosphere and the respective observations at various positions of a spacecraft. We have chosen a 1 x 3 AU orbit with its major axis at 90 ~ w.r.t, the LISM flow, as shown in Fig. 4 and recently proposed for an Interstellar Pathfinder Mission [49]. It was assumed that the original O distribution in the LISM resembles that of He with VBulk= 25 km/s and T = 7000 K as obtained by Witte et al. [20]. According to Izmodenov et al. [ 12], the filtration was adjusted so that the resulting neutral gas flow contains 50% of the primary distribution. 20% of a secondary distribution with a reduced speed of VSec = 21 km/s and T s e c = 10000 K were added. Now the sun's gravitational field is employed to derive flow velocities, from the observed image of the flow in the sky, at locations downstream of the sun. Figure 5 shows simulated data (diamonds) as obtained by a neutral gas imaging instrument with an effective collection area of 1 cm 2 and an angular resolution of 40 in the direction of the scan, utilizing the spacecraft spin. The 1~ spacing of the data points is achieved by oversampling. Such an instrument will achieve the angular resolution by means of a collimator. The neutrals are then converted into negative ions on a conversion surface, which are accelerated into a time-of-flight spectrometer and analyzed for their mass. Overall conversion efficiencies of several % have been reported for O in the expected energy range [28]. The counting statistics of the simulated data reflects the assumption of a conservative overall detection efficiency of 1% and accumulation over 10 days, including the duty cycle due to the spinning spacecraft. The simulated data (diamonds) are compared with an O distribution with the original He parameters (line). The distribution has been normalized such that no negative values are computed after subtracting this distribution from the simulated data up to 10% of the peak value. The squares represent the resulting secondary distribution and its statistical error after this subtraction. The presence of a secondary distribution and its total relative contribution can be determined with ~10% accuracy in the current example. Also its position in angle space and thus its bulk velocity can be determined along with the temperature. This simulated observation applies to a position 135 ~ from the flow direction (c.f. Fig. 4). At a position 225 o from the flow (not shown here) we find a much weaker deviation from the original distribution, because of the
E. MObius et al.
high relative velocity when the spacecraft is moving into the flow. In addition, the absolute flux is much lower, because the O density decreases rapidly towards the sun. Generally, the gravitational deflection is stronger closer to the sun, but the total flux of neutral atoms is greatly reduced. A l x3 AU orbit appears to be an ideal compromise for the attempted neutral gas imaging. With contiguous data along the downwind portion of the orbit, even much smaller secondary O contributions with velocities closer to that of the original LISM flow can be detected and evaluated. Currently, instrumental efforts are unFig. 5: Simulated angular distribution of the interstellar derway to extend the sensitivity of the O flow as observed with an instrument with 5~ resoluneutral to negative ion conversion to tion. In addition to the original primary O component low energy ( 1 0 - 4 0 eV) H atoms, with velocity and temperature as obtained for LISM He a secondary component of 25% is assumed, which is which will then also allow to image produced by charge exchange in the heliospheric interthe H flow of the LISM. As can be seen from this preliminary analysis, a face according to Izmodenov et al. [12]. The error bars indicate the statistical error of the observed distribution comparison of the He, O and H flow (diamonds) and for the derived secondary distribution characteristics in the inner heliosphere (squares) after subtraction of a normalized primary diswill provide powerful tool to deduce tribution. the amount of filtration and its related processes quantitatively from observations in the inner heliosphere. This will be an important task in the near future in order to allow an accurate deduction of the composition of the LISM and its ionization state from observations inside the heliosphere. Any deviation of the flow vector of interface affected species from that of He will also provide constraints on the interstellar magnetic field direction in our neighborhood. The flow pattern of H in the inner heliosphere is also affected by radiation pressure. Imaging neutral H from various locations will provide the most accurate account of this effect and will allow unambiguous separation of interface and radiation pressure related effects. In addition, radiation pressure may also serve as a discriminator between H and D, because the strength of the effect depends on the particle mass. As a result the D image will be separated from the H image in angle, thus giving a neutral imaging mass spectrograph an advantage in the observation of D in the LISM. The D/H ratio in the interstellar medium has been considered an important data point for the evolution of matter [50, 3, 2], because D is the only isotope that is clearly being depleted by reactions inside stars over the primordial amount. Whereas local observations of 3He have been obtained, this task appears to be much more difficult for D [51 ]. 6. CONCLUSIONS AND OUTLOOK Over the past 15 years in-situ observations of interstellar material inside the heliosphere have been established as an important tool to determine the physical parameters and the composition of the LISM just outside the heliosphere. Having a full complement of pickup ion, neutral gas and UV instruments on operating spacecraft provides the unique opportunity to analyze the LISM simultaneously with all complementary in-situ and remote sensing techniques.
-118-
Interstellar gas flow into the heliosphere Therefore, benchmark values of the He component in the LISM can be expected in the near future. From here pickup ion instrumentation and constant monitoring in UV will provide a broad database on LISM composition and the long term response of the heliosphere to variations of the sun and the LISM as drivers. The role of the heliospheric interface in the translation between the data obtained inside and the values in the LISM is on one hand a challenge. It must be overcome on the road towards high accuracy benchmark values for our neighborhood sample of the interstellar medium. On the other hand the variety of observations hold clues themselves about the related interactions in the interface. Future progress in the direct observation of interstellar neutrals beyond He will play a key role in this effort. The effort to extend the composition information to key isotope ratios that provide insight into the various nucleosynthetic processes and the study of the heliospheric interface interactions will require a spacecraft mission, dedicated to the study of the LISM. It requires high collecting power pickup ion instrumentation and an advanced neutral gas imager. Both instrument families are extensions of currently available hardware. These goals can be achieved on broad footing with a Discovery class mission or with some reductions in scope under the NASA Explorer program with a spacecraft on a 1 x 3 AU orbit. A similar mission has been proposed already as Interstellar Pathfinder [49]. Such a mission will be the ideal predecessor for an Interstellar Probe that would explore the interstellar medium itself. ACKNOWLEDGMENTS We wish to thank the International Space Science Institute (ISSI) for its hospitality and all members of the ISSI Team on the interstellar He cone for their efforts. The work benefitted from helpful discussions with P.A. Isenberg and M.A. Lee. This work was supported under NASA Grants NAG 5-2754, NAG5-8733, NAG 5-6912 and 5-4818, NSF Grant ATM9800781, and Polish Committee for Space Research Grant 2P 03C 005 19. REFERENCES: 1. Schramm, D.N., Space Sci. Rev, 84 (1998) 3. 2. Prantzos, N., Space Sci. Rev, 84 (1998) 225. 3. Mullan, D.J., and J.L. Linsky, Astrophys. J., fill (1998) 502. 4. Primas, F., Space Sci. Rev., 84 (1998) 161. 5. Frisch, P.C., Space Sci. Rev., 86 (1998) 107. 6. Frisch, P.C., Am. Scientist, 86 (2000) 52. 7. Zank, G.P., and P.C. Frisch, Astrophys. J., 518 (1999) 965. 8. Linsky, J.L., Space Sci. Rev., 78 (1996) 157. 9. Wood, B.E., J.L. Linsky, and G.P. Zank, Astrophys. J., 537 (2000) 304. 10. Baranov, V.B., Space Sci. Rev., 52, (1990) 89. 11. Fahr, H.J., Space Sci. Rev., 78 (1996) 199. 12. Izmodenov, V., Yu. Malama, and R. Lallement, Astron. Astrophys., 317 (1996) 193. 13. Zank, G.P., and H.L. Pauls, Space Sci. Rev., 78 (1996) 95. 14. Lallement, R., Space Sci. Rev., 78 (1996) 361. l 5. Lallement, R., J.L. Linsky, J. Lequeux, and V.B. Baranov, Space Sci. Rev., 78 (1996) 299. 16. Frisch, P.C., and J.D. Slavin, Space Sci. Rev., 78 (1996) 223. 17. Weller, C.S., and R.R. Meier, Astrophys. J., 193 (1974) 471. 18. M6bius, E., et al.,Nature, 318 (1985) 426. 19. Gloeckler, G., et al., Science, 261 (1993) 70.
-119-
E. MObius et al. Witte, M., M. Banskiewicz, and H. Rosenbauer, Space Sci. Rev., 78 (1996) 289. Quemarais, E., et al., J. Geophys. Res., 104 (1999) 12585. MObius, E., et al., Geophys. Res. Lett, 26 (1999) 3181. Chassefiare, E., J. L. Bertaux, R. Lallement, and V. G. Kurt, Astron. Astrophys., 160 (1986) 229. MObius, E., D. Rucinski, D. Hovestadt, and B. Klecker. Astron. Astrophys., 304 (1995) 505. Geiss, J., et al., Astron. Astrophys., 282 (1994) 924. Gloeckler, G. and J. Geiss, Space Sci. Rev., 86 (1998) 127. Gloeckler, G., L.A. Fisk, and J. Geiss, Nature, 386 (1997) 374. Wurz, P., et al., Opt. Eng., 34 (1995) 2365. Wurz, P., R. Schletti, and M.R. Aellig, Surf Sci., 373 (1997) 56. Lallement, R., Scattering of solar UV on local neutral gases, in: Physics of the Outer Heliosphere, S. Grzedzielski and D.E. Page ed.s., COSPAR Coll. Ser., 1 (1991) 49. 31. Scherer, H., H.-J. Fahr, M. Bzowski and D. Rucinski, Astrophys. Space. Sci., 274 (2000) 133. 32. Bzowski, M., and D. Rucinski, Astron.. Astrophys., 296 (1995) 248. 33. Bzowski, M., H. Fahr, D. Rucinski, and K. Scherer, Astron.. Astrophys., 326 (1997) 396. 34. Gruntman, M., and H.J. Fahr, Geophys. Res. Lett, 25 (1998) 1261. 35. Gruntman, M., and H.J. Fahr J. Geophys. Res., 105 (2000) 5189. 36. Vallerga, J., Space Sci. Rev., 78 (1996) 277. 37. Flynn, B., J. Vallerga, F. Dalaudier, and G.R. Gladstone, J. Geophys. Res., 103 (1998) 6483. 38. Judge, D., et al., Solar Phys., 177 (1997) 161. 39. Gloeckler, G., and J. Geiss, Space Sci. Rev., in press (2001). 40. Gloeckler, G., L.A. Fisk, and N. Schwadron., Geophys. Res. Lett., 24 (1995) 93. 41. MObius, E., D. Rucinski, M.A. Lee, and P.A. Isenberg., J. Geophys. Res., 103 (1998) 257. 42. MObius, E., D. Rucinski, P.A. Isenberg, and M.A. Lee. Ann Geophys., 14 (1996)492. 43. Gloeckler, G., J.R. Jokipii, J. Giacalone, and J. Geiss, Geophys. Res. Lett., 21 (1994) 1565. 44. Isenberg, P.A., J. Geophys. Res., 102 (1997) 4719. 45. Schwadron, N.A., J. Geophys. Res., 103 (1998) 20643. 46. Chalov, S., and H.-J. Fahr, Astrophys. Space Sci., 363 (2000) L21. 47. Ghielmetti, A., et al., Optical Eng., 33 (1994) 362. 48. Banaszkewiecz, M., M. Witte, and H. Rosenbauer, Astron. Astrophys. Suppl. Set., 120 (1996) 587. 49. Gloeckler, G., et al., LOS, Trans. Am. Geophys. Union, 80 (1999) $237. 50. Linsky, J.L., Space Sci. Rev., 84 (1998) 285. 51. Gloeckler, G., and J. Geiss, Nature, 381 (1996) 210. 52. Dalaudier, F., J. L. Bertaux, V. G. Kurt, E. N. Mironova, Astron. Astrophys., 134 (1984) 171. 53. Chassefiare, E., J. L. Bertaux, R. Lallement, B. R. Sandel, and L. Broadfoot, Astron. Astrophys., 199 (1988) 304. 54. Bertin, P., R. Lallement, R. Ferlet, A. Vidal-Madjar, and J.L. Bertaux, J. Geophys. Res., 98 (1993) 15193.
20. 21. 22. 23. 24. 25. 26. 27. 28. 29 30.
-
120-
N o n - s t a t i o n a r y transport of neutral atoms in the H e l i o s p h e r e A.I. Khisamutdinov a M.A. Phedorin b and S.A. Ukhinov b Sobolev Institute of Mathematics of the Siberian branch of Russian Academy of Science, Acad. Koptyug prospect 4, Novosibirsk, 630090, Russia b Novosibirsk State University, Pirogova str. 2, Novosibirsk, 630090, Russia
a
A problem on non-stationary transport of neutral H-atoms in the two-shocks Heliospheric plasma is considered. A point impulse source of H-atoms is situated on the outer boundary of Heliosphere. Some time-spatial-velocity distributions of the atoms are computed by a Monte Carlo methods specially developed. 1. I n t r o d u c t i o n
Interaction between the Local Interstellar Medium (LISM) and the Heliosphere is an object of the scientific interest and research, and this fact is reflected in a lot of publications. One of the problems deals with propagation of neutral particles (arising from LISM) in the Outer Heliosphere (including domains both before and behind the Heliopause) whose parameters are assumed to be independent of the particles. A mathematical model of this physical phenomenon is some Markov jump process, and the linear kinetic transport equation is the "master" equation of this process, in fact. This all makes natural applications of Monte Carlo methods for investigation of neutral atoms' distributions. A few stationary problems on the propagation of H and O atoms were considered in [2-4] with use of the Monte Carlo methods, and interesting results were obtained. Non-stationary problems of the neutral particles transport are more complex for the computation than the stationary ones. As to the most important non-stationary problems, it should be pointed a problem on propagation of the neutral atoms in Heliosphere interface whose parameters are periodic functions (with period of the Solar cycle). In other important problem, parameters of the medium (Heliosphere) are assumed to be stationary but the flux of the neutral atoms from the LISM varies in time. In the present paper we consider a problem of above indicated second type on propagation of H-atoms; the source of H-atoms is impulse point, and its velocity distribution density q~(v) corresponds to the Maxwell distribution in LISM We carry out numerical solving employing Monte Carlo methods, and then analyze some properties of the phenomenon considered. 2. M o d e l of m e d i u m , T r a n s p o r t e q u a t i o n , and functionals c o m p u t e d
Axially symmetric model of the Heliospere is considered, which is chosen according to twoshocks gasdynamical model (see [1, 3]). Heliosphere is assumed to be confined in a bounded spatial domain Va with piecewise smooth boundary c3V~. The scheme of the Heliosphere model is shown in Fig. 1. The origin of the Cartesian coordinate system coincides with Sun center, OZ (axis) is the symmetry axis of the system and is directed towards the flow of interstellar gas. Only the "upwind direction" region is considered in our problem, 0 <_z _<400 au. In order to simplify the assignment of the medium model data, the whole domain V~ was divided into 8 sub-domains (zones): each of four layers between natural boundaries LISM, BS, HP, TS, and the sphere of 1 au radius were cut by the cone of n / 4 angle (see Fig. 1). In Fig. 1 the zones are indicated by numbers from 1 to 8. The distribution of velocities for the plasma protons in Heliosphere is assumed to be Maxwellian with corresponding values of mean velocities and temperatures. The values of plasma parameters in indicated 8 zones were taken according to the Ref. [3] at the points shown in Fig. 1 as "D2 ~ and "D3". It is assumed that the atoms interact with the medium by means of two well-known types of interactions: a charge-exchange scattering and a photoionization (absorption). To express boundary conditions also it is assumed that outside the surface c3V~ an "absolutely absorbing" medium is situated. We introduce the following denotations and assumptions: X=R 6 is a phase space of coordinates r and velocities v; x=(r, v ), x ~ X, V~ -(Vc, ~c?VG) G-Va|
3 GcX
q(t,x) = 6 ( t ) 6 ( r - r0). qv(v ) is the density of H-atoms source, r0 ~ c3V~, r0 = (0,0,z0) , z0=400 au; II q(t, x). dtdx - 1
,/d
- 121 -
A.L Kishamutdinov, M.A. Phedorin and S.A. Ukhinov np(r) is a number density of plasma protons at a point r, Vp is a velocity of a plasma proton, V p E R 3", V(r) is the mean value of velocities of plasma protons at a point r; P m(u[r) , U=Vp-V, is
the density of Maxwellian distribution of protons velocities at r with given mean velocity V(r) and the corresponding temperature T, T=T(r); g=[V-Vp[is the modulus of relative velocity of atom and proton;
Fig. 2. Probability of Remaining in the Heliosphere and Fig. 1. Scheme of the model of the Probability density of the Escaping through its Heliosphere accepted for the current boundaries vs. time for the H-atoms in Heliosphere investigation, resulted from the point 8(0 source situated on the system's axis. croh~(g ) is microscopic cross-section of the "charge exchange" interaction between the atom and proton; wl(g ) - g . Crch.ex ( g ) ; Wohex(g,r) -- n p ( r ) ' w l ( g ) , Vr ~ VG, Vg >_0; C(v -+ v'[Vp) is an indicatrix of charge-exchange scattering of the atom with the velocity v by the proton with the velocity Vp; P0 - {Xt }t~0 is a jump Markov process of the atoms propagation, which is determined by the above given elements; its states are in X ~ {-a , where {a is an absorption state; m t ( x ) - m(t, x), x ~ X w {a , is a phase density of atoms in P0 ; Everywhere above and below the integration domain is not indicated if it coincides with the whole domain (of integration). The phase density m t (x) being considered as some generalized solution satisfies on X to the kinetic transport equation, whose terms of the source and the scattering are q(t,x) and S § respectively, A
S +m(t, x) - II m(t, r, v'). Wch~• ([ V'--Vp I, r)C(v'---~ v [Vp). Pm (U I r)" a v e 9d v ' ,
Vx E a.
In this study we used the values of O'ch.ex ( g ) , C ( v --~ v ' l V p ) and the frequency of photoionization, which are given in [5, 2, 4]. The values of interstellar gas temperature and mean velocity, accepted at assignment of the source, were set as the following: 6700 K and (0, 0 , - 5 . 4 7 7 ) au/year, respectively. To obtain and study the solution m(t,x) different functionals on it are computed: (1) mean densities (including concentrations) and mean fluxes of atoms over given surfaces in V G and over given time intervals for different velocity subsets; (2) mean densities (concentrations) of atoms in given volumes in V~ for given time moments; (3) mean velocities as the ratio of mean fluxes and the corresponding mean densities of atoms. First two types of the means are the linear functionals on m(t,x) of the following form
Is - l d t l d v l d S ' [ ( V , ns)lm(t,x)'Oes(t,x), It,.v- Idx'mt,(x)'oC'v(t*,x), (s)
- 122-
Non-stationa~ transport of neutral atoms in the heliosphere
where (S) is the surface with the normal ns(r) at the point (r), t* is the time moment (t*>0), (V) is the volume, OCs(t,x) and OCv(t*,x ) are the corresponding weight functions. In particular, we have for the mean concentration of atoms over (S) and over the time interval (tl, t2): t2
i d t l d S i d v , l(v, ns)lm(t,r,v)" ~ <s~
1
I(v,n )l-(t2
- t )-II S I1' where IISII is the area size of (S); on the Fig. 3 these functionals are denoted as n. Total numbers of computed functionals were 12"15 of type Is and 1"8 of type It,,v. As volumes (V) the volumes of zones 1 to 8 were used. As a surfaces (S) a flat, perpendicular to OZ axis rings were used; the "thickness" of each ring, i.e. (emax-Rmin), was about 40 au, where Rmax and emin are the maximal and minimal radius of a ring, respectively. Three levels of positions of the surfaces (S) are shown in Fig. 1 as "D 1", "D2", and "D3 ", where markers indicate the positions of the central radius of the rings
(i.e. (Rmax-emin)/2). 3. M o n t e Carlo m e t h o d used
In order to compute all functionals Is and It,,v, trajectories of a random process close to P0 were simulated and corresponding "imitation" estimators were used. When saying about the close process, we mean the fact that the photoionization absorption was taken into account by the known technique "analytical averaging of absorption". It is the important feature that a new method of simulation for joint distribution of random time instant and spatial coordinates of atom's collision and of random velocity of colliding proton under the charge exchange was constructed. In the new method a rejection technique is used, the "rejected" values are treated as the collisions of some complementary type, not changing the phase density. The mean numbers of physical and "fictitious" (complementary type) collisions are given in Table 1. One can see that the method constructed is sufficiently effective. 4. Results and Discussion
For computation a new code was generated. Here are some data about the computational cost and the relative standard errors. The computation time for 50,000,000 trajectories was equal to 9,800 s for the Intel Celeron 500 MHz processor. The errors of all computed functionals were within ( 1-10 %). At first we discuss some common patterns of time behavior, connected with V6 and c3Vc;. These are the probability for atom to stay in V6, the probability density to escape through the "upwind" and "downwind" parts of c3V~, the mean free path, and the mean free flight time between collisions, and .others, that are given in Fig. 2 and Table 1. It is interesting to notice the presence of two maxima on one of the curves on the Fig. 2. The characteristic phenomenon that was established for the stationary problems of H-atoms transport in Heliosphere is the effect of the "hydrogen wall" (see Ref. [3]). This phenomenon is also appearing in the obtained solution of non-stationary problem. As a basic form for presentation of the results obtained we have chosen the time distributions (histograms) of Is. It is adequate for the data analysis to present the phase density of H-atoms as sum of two densities: the scattered and non-scattered atoms; two corresponding summands for any Is turned out to be the functions with the single maximum and vanishing when t -+ 0 and t --~ oo Table 4. Basic parameters characterizing the non-stationary transport of H-atoms in the Heliosphere Time, after which the half (V2) of the H-atoms that initially come into 105 years Heliosphere, abandon it Mean life-time of the H-atom in the Heliosphere 110 years Mean number of the "charge-exchange" collisions 4.8 Mean trajectory length in the Heliosphere 573 au Mean time / mean path among two subsequent collisions 22.9 years / 119.3 au Mean number of the fictitious collisions in the Heliosphere 1.94 Probability of escaping through the plane Z=0 0.96 Probability of absorbing in the Heliosphere _ 0.01 - 123-
A,L Kishamutdinov, M.A. Phedorin and S.A. Ukhinov
One can treat this fact as superposition of some two waves. Naturally, the peaks of non-scattered densities are situated not later the peaks of scattered ones. Everywhere in Fig. 3 and Fig. 4 these two summands are given. It is also important to note that the mean velocities vz(t) of scattered particles for the zones 4,7, and 8 change their sign (see, e.g., zone 4 in Fig. 4).
Fig. 3. Histograms of H-atoms density (a) and velocity module ~) distributions in zone 2 of the accepted model of the Heliosphere.
Fig. 4. Histograms of vz component of the velocity distribution of H-atoms in zones 2 (a) and 4 (b) of the accepted model of the Heliosphere. The characteristic property of all histograms of all different mean velocities for scattered particles is their identical asymptotical behavior under t ~ oo, i.e. the fact that all curves tend to some constants. It should be noticed that absolute values of the corresponding densities and fluxes decrease and vanish under t -4 oo. The reason of this phenomenon as we suppose is the following the phase density of scattered particles mso(t,x) is represented under t - ~ oo as the product of two factors, mso (t, x) => m 1 (t, r). m 2 (v ] r), t-+oo
where the density of conditional distribution of velocity m2(vlr) is time-independent already. 5. A c k n o w l e d g e m e n t s
This work has been supported by INTAS-CNES cooperation project # 97-00512 and International Space Science Institute in Bern. We are also grateful to all participants of the project and especially to R. Lallement, V.B. Baranov, V. Izmodenov, and Yu.G. Malama for useful discussions and for the data presented. References
[1] [2] [3] [4] [5]
Baranov V.B., Space Sci. Rev., 52, 1990, p. 89-120 Malama Yu.G., Astrophys. Space Sci., 176, 1991, 21-46 Baranov V.B. and Malama Yu.G., Space Sci. Rev., 78, 1996, p. 305-316 Izmodenov V., Malama Yu. G., and Lallement R., Astron. Astrophys. 317, 1997, p. 193-202 Maher L. and Tinsley B., J. Geophys. Res., 82, 1977, p. 689-695 - 124-
Pickup Ion Turbulence: A Stochastic Growth Model I
G.P. Zank I and Iver H. Cairns 2 ~Bartol Research Institute, University of Delaware, Newark, DE 19716, USA 2School of Physics, University of Sydney, Sydney, NSW 2006, Australia A new characterization of interplanetary MHD turbulence over the polar regions of the sun, based on a stochastic growth model for waves driven by pickup ion instabilities, is presented. By considering temporal and spatial variations in the local interplanetary magnetic field (IMF), we find that the mean wave growth rate can be very small (and even negative), but the variance can be large, indicating that localized regions experience significant wave growth in a stochastic fashion. This suggests that pickup ion driven turbulence in the outer polar heliosphere has a bursty or intermittent character, occurring in clumps. Thus, individual wave growth events are rarely detected, and persistent beamlike anisotropies in observed pickup ion distributions are expected. 1. INTRODUCTION Virtually all theoretical work addressing interstellar pickup ions in the solar wind over the last ~25 years has assumed that pickup ions generate significant levels of magnetic field fluctuations. The fluctuations were then assumed to scatter the pickup ions rapidly, so ensuring that the pickup ion distribution was essentially isotropic in the solar wind frame and co-moving with the solar wind (see [ 1] for a review). However, concerns about this picture for the pick-up of interstellar ions, wave generation and scattering appeared when a concerted effort by several groups failed to find definitive observational evidence for wave generation by pickup ions in the outer heliosphere. The few events identified as enhancements of magnetic fluctuation spectra near the ion cyclotron frequency and interpreted in terms of pickup ion-driven waves, all occurred during periods when the large scale interplanetary magnetic field IMF was quasi-radial [2,3,4]. During periods when pickup ions were observed, it was difficult to consistently identify enhancements in local IMF spectra that might be associated with pickup ion generated waves, suggesting that the observed wave growth rate Fob s "~ 0 much of the time. Such low wave growth rates would be consistent with the observed anisotropic pickup ion spectra. However, long-time averages of magnetic fluctuations in the outer heliosphere [5] suggest that pickup ions are driving turbulence in the outer heliosphere and at a rate that is consistent with simple scaling estimates [ 1]. The observations are in contrast to the predictions of quasi-linear theory. Here, we describe a newly developed model [6], based on stochastic growth theory (SGT) [7], for the growth of waves driven by the pickup process in the polar solar wind, and offer a resolution to the paradoxical observations of anisotropic pickup ion distributions, the frequent absence of local wave enhancements in magnetic fluctuation spectra, and the significant energy density in pickup ion-generated waves in the outer heliosphere. This work supported in part by NASA grants NAG5-6469, NAG5-7796, an NSF-DOE award ATM0078650, and an NSF award ATM-0072810.
- 125-
G.P. Zank and 1.11. Cairns
Although the large-scale polar IMF is radial, significant small-scale variation is present. We show that variation in field direction engenders spatial and temporal variations of pickup ion distributions that can lead to the wave growth varying in a stochastic manner. 1.1 S u m m a r y of pickup ion bean instabilities
Let 0 denote the angle between the magnetic field B and the radial solar wind flow. For 0 ~ < 0 < 45 ~ the ion-ion right-hand resonant instability dominates, whereas for 0 > 70 ~, the left-hand polarized ion cyclotron instability begins to dominate. The ion-ion righthand resonant instability growth rate is greatest for waves parallel to B and decreases monotonically with increasing obliquity. Simulations show a decreasing level of fluctuations as 0 increases; a result which is consistent with observations made at comet Giacobini-Zinner, which show the absence of large amplitude magnetic fluctuations near the water group ion cyclotron frequency when 0 = 90 ~. See [8] for a review. The basic conclusion that emerges from the above summary is that pickup of interstellar atoms when the IMF is almost radially aligned with the solar wind flow drives the fastest growing waves with the highest saturation level. In contrast, the waves should saturate at a low level in regions where 0 ~ 90 ~. 2. STOCHASTIC G R O W T H OF MHD FLUCTUATIONS At some instant, suppose that the polar interplanetary magnetic field, although radial on a sufficiently large time and spatial scale, has the form depicted in Figure 1. Consider an arbitrary region C with a scale length L and assume that any excited modes are Alfvrnic. Below, we shall use a value of L corresponding to the correlation length for the IMF fluctuations. In choosing a value for L, one should recognize that, in the region of ionization, the magnetic field should be "local." A characteristic interaction time tg of excited Alfvrn waves with C is then tg =- L / VA . C will experience perturbations in the wave growth rate during the time tg as ring-beams created in regions of different magnetic field orientations interact with C. The most important ring-beams, since they have the largest growth rate and highest saturation levels, are the almost pure beam distributions created in nearly radial magnetic fields. Moreover, outside growth sites, the beams tend to reform due to fast particles outrunning slow particles. Physically, we may expect a net balance between particle streaming, which generates waves, and particle scattering caused by wave growth, which attempts to drive the beam towards isotropy. This in turn reduces the wave growth and hence diminishes the scattering and so allows the beam to reform partially. Like all physical systems with instabilities, we expect the waves and particles to relax towards marginal stability [7]. To summarize, if beam-driven waves are most important, then since the driven waves have wave vectors close to parallel to the local field, changes in the background field direction will cause the growth region of wave vector space to move, thereby causing the growth/damping rate of waves with a given wave vector to vary with time and location. (Note that the beam particles remain in the same place in (v H,v;) space, but not in (v x, Vy, Vz) space.) There are thus two reasons for stochastic variations in the growth rate of waves: (i) the "direct" SGT reason of competition between rebuilding and destruction of the beam source of wave growth,
- 126-
Pickup ion turbulence." A stochastic growth model and (ii) changes in the location in wave vector space of locally driven waves due to changes in magnetic field direction. The mean growth rate and its standard deviation, for pickup ion driven turbulence in the outer polar hemisphere, are then [6] ( F ) ~ (Z) and
(Z)(vb)L Z • = u + b / ~
(e}'e
'
- - where F - F - 7
and F / 7 are the wave growth and
'
damping rates, u and b denote fluctuations in velocity and magnetic field (p denotes density), (e) denotes the magnetic fluctuation correlation length, (v~) is a mean beam speed, and P is the probability that a particular section of length (g) is quasi-radial. 3. DISCUSSION, I M P L I C A T I O N S AND CONCLUSIONS Consider the region C for which L ~ (g) and use values Z 2 ~ 2.5 x 10 s m 2 s -2 (g) ~ 2 x 10 9 m, (vb)~ 6 x 10 s m s
-1 .
Then c (F) ~ 5 x 10-' P .... s "1, and (F) = -4 • 10-~ s 1 for the polar regions. The characteristic interaction time for waves in a region with size (g) is tg = 5 x 10 4 s = 14 hours, during which time a number ng((g)) = P ( V b ) / V A ~ 30P of beam
+m
inhomogeneities with characteristic time scale t i = P(e)/(Yb) ~ 1667P s 0.46P hours should pass through. It is important that 1; this
n~((e))>>
o
o
Figure 1. Snapshot of the interplanetary magnetic field, illustrating local departures from the mean IMF (B) and the different newborn pickup ion distributions from region to region. Local newborn velocity space distributions are shown below the magnetic field sketch for regions 1 - 5 and C, showing how distributions range from beam to ring. The horizontal axis in each phase space plot is v, (i.e., particle speed parallel to the local
indicates that the system should indeed evolve to an SGT state close to marginal stability and suggests that the SGT model developed above is self-consistent. The SGT model also predicts the degree ofburstyness of the waves, the characteristic time scale for the SGT inhomogeneities in the pickup ion distributions, and the characteristic size of regions where pickup ion-driven waves might be observable.
The growth rate of the unstable low frequency waves may be estimated [ 1] using F << co ~ (V~/Usw)f~, where f2 is the pickup ion gyrofrequency and co the solar wind frame frequency. At ~3 AU, co r ~ 7 X 10 -4 S-~ , r
-
127-
G.P. Zank and I.H. Cairns
from which F _<10 - 4 S -1 is a reasonable estimate. We use F = 5 x 10 -5s -~ below. Note that o(F) = 5 • 10-~s-' = F, so that this value of F is quite reasonable, as is the value for o(F). Suppose the magnetic field within C is oriented randomly, and, for example, the normalized standard deviation from the mean magnetic field is 30 ~ In this case, P ~ 0.68 and ~ ( F ) / ( F ) ~ 14.5. The number of beam inhomogeneities interacting with C is ng((g)) ~ 20.4, each of which has a characteristic time scale tg ~ 1134s = 0.32 hours. If,
fortuitously, the beams arrive almost successively to interact with C, so giving a total interaction time -~ 2 x 104 s, we can expect large wave amplification within C (since Ft i ~ 1). However, since o ( F ) / ( F ) i s large, substantial wave amplification will be infrequent and wave enhancements will have a bursty character. The conclusions for the I"1= 30 ~ case continue to hold for an IMF with a smaller variance in local departures from the radial mean field direction. However, if the mean IMF is close to radial but the variance is large e.g., 11 = 60 ~ we find that the number of beams ng((g)) is reduced (~11.5) and ti is correspondingly smaller (~637s), so making effective wave amplification more bursty and less likely to be observable (since ~ ( F ) / ( F ) ~ 19.4). In summary, our results support an SGT description of pickup ions and associated MHD waves. The consequent closeness of the system to marginal stability (averaged over time and volume) and the associated arguments for the preservation of beam-like distributions of pickup ions immediately provides a qualitative explanation for the observed pickup ion anisotropies. The primary conclusion to emerge from this study is that the dynamical character of the quasi-radial polar IMF prevents the formation of statistically steady-state pickup ion driven wave enhancements in the magnetic fluctuation spectra. The parameter controlling the frequency of wave enhancements is the variance in the orientation of the fluctuating IMF about the mean radial field. We find that even were the mean field radial, a large standard deviation from the radial direction in the local IMF fluctuations on the length scale of the correlation length would lead to very little effective wave growth. Thus, the SGT model predicts bursty wave growth in polar regions with, first, quasiradial IMF direction and, second, magnetic fluctuations that possess only a small variance in direction from the mean field direction. Such an intermittent and bursty character for pickup ion driven turbulence in the polar solar wind must therefore make the detection of individual wave events difficult. Additionally, the model predicts a bursty, intermittent character for the pickup ion distribution, with preservation of beam-like features and anisotropies in particular. Further details can be found in [6]. REFERENCES 1.
2. 3. 4. 5. 6. 7. 8.
Zank, G.P., Space Sci. Rev., 89 (1999), 413. Smith, E.J., et al., EOS Trans. AGU, 75(16), Spring Meeting Suppl. (1994), $297. Murphy, N., et al., Space Sci. Rev., 72 (1995) 447. Intriligator, D.S., G.L. Siscoe, and W.D. Miller, Geophys. Res. Lett., 23 (1996) 2181. Zank, G.P., Matthaeus, W.H., & Smith, C.W., J. Geophys. Res., 101, (1996) 17 081. Zank, G.P. and I.H. Cairns, Ap.J., 541, (2000), 489. Cairns, I.H., and P.A. Robinson, Phys. Rev. Lett., 82, (1999) 3066. Gary, S.P., Theory of Space Plasma Microinstabilities, Cambridge, 1993.
- 128-
A time-dependent, 3D model of interstellar hydrogen distribution in the inner heliosphere M. Bzowski ~ *, T. Summanen b, D. Rucifiski ~, and E. KyrSls b ~Space Research Centre PAS, Bartycka 18A, 00-716 Warsaw, Poland bFinnish Meteorological Institute, Vuorikatu 15A, 00101 Helsinki, Finland
1. M O D E L The density of interstellar hydrogen in a local point in the inner heliosphere is calculated as a 3D numerical integral of the distribution function composed of a Maxwellian shifted by the bulk velocity vector at infinity, multiplied by the probability of survival of a test atom against 2D, time dependent ionisation during the travel from "infinity" to a local point. The link between the distribution function in the local point and in "infinity" and the probability of survival of the test atom against ionisation are calculated by numerical tracking of the atom travelling in a time- and latitude-dependent ionization field under dynamic influence of the solar gravity and spherically symmetric, time dependent radiation pressure. The concept of the numerical method used was presented by Rucifiski & Bzowski (1995). The models of temporal evolution of the solar Lyman-a radiation pressure and of the time- and latitude dependence of the photoionisation efficiency and of the rate of charge exchange of the hydrogen atoms with the solar wind protons are discussed in the companion paper (Bzowski, this volume). 2. C A L C U L A T I O N S Throughout the paper we used two latitudinal profiles of the ionization rate discussed in the companion paper and two sets of bulk velocity and temperature of the interstellar warm and slow gas) and (VB = 25 km/s, gas at infinity: (VB = 20 km/s, T~ = 12000 K T~ = 7500 K); the ecliptic coordinates of the inflow direction were assumed (AB,/~B) = (252.4~ 7.5~ The calculations were performed in the following three planes: (i) the ecliptic plane; (ii) the polar plane, i.e. the plane perpendicular to the ecliptic and including the inflow direction the scan starts upwind and goes through the north ecliptic pole, downwind direction and the south pole back to the upwind direction; and (iii) the crosswind plane, i.e. the plane perpendicular to the inflow direction and containing the Sun. None of the latter two planes goes through the heliographic poles, the minimum distance is ~ 8~ *M.B.
and D.R. were supported by a Polish State Committee for Scientific Research grant 2P03C 004 14
-
129-
M. Bzowski et al.
3. H Y D R O G E N
DENSITY DISTRIBUTION
The asymmetry of the ionisation rate eliminates the axial symmetry of hydrogen density distribution (see e.g. Summanen 1996). Indeed, a considerable pole-to-ecliptic asymmetry in density extends to about 25 AU from the Sun, where it is still equal to ~ 1.05. Closer to the Sun, the ratio varies from ~ 1.2 to ~ 1.3 at 10 AU and from ~ 1.5 to ~ 2 at 2.5 AU, depending on the phase of solar cycle. Hence also the hydrogen cavity, defined as the region where the gas density is lower than 1/e of density at infinity, is asymmetric. This asymmetry changes with time. The size of the cavity from upwind to downwind, from pole to pole and in-ecliptic varies by ~ 30% during the solar cycle, as shown in Fig.1. and it depends quite strongly on the gas parameters at infinity.
Figure 1. Mean hydrogen cavity (thick ovals) in the three scan planes, superimposed on isocontours of 10% (light), 25% (medium), and 50% (dark) variability amplitude. In the lower-right panel: evolution of the cavity size in time for the warm and slow gas (solid lines) and cool and fast gas (broken lines).
-
130-
A time-dependent, 3D model of interstellar hydrogen distribution...
Crosswind scan, relative intensity profiles
Polar scan, relative intensityprofiles 1.1
1.25
1.05 "~
1.3 rectangular bulge
.~ 1.2
1
"~ 0.95
~ 1.15
~
o.9
~. 1.1
0.85
1.05
0.8 -75 -50 -25 0 25 50 ecliptic latitude/3
-75 -50 -25 0 25 50 ecliptic latitude/3
75
75
Figure 2. Profiles of backscattered intensity in the polar scan (upwind hemisphere) and crosswind scan, normalised to the intensity at solar minimum respectively in the upwind and crosswind in-ecliptic directions, for the warm and slow gas and two different profiles of the ionization rate.
The amplitude of density variations, defined as maximum deviation of the local density from the local density averaged over time, is also presented in Fig.1. In the upwind hemisphere, the mean cavity almost coincides with the region where the amplitude of density variations exceeds 25%. The 50% amplitude region coincides roughly with the positions of peaks of the emissivity function along heliocentric lines of sight and it contains the orbit of Ulysses. 4. E F F E C T S I N I N T E N S I T Y
OF BACKSCATTERED
RADIATION
The intensities of the solar Lyman-a radiation backscattered by the neutral interstellar hydrogen were calculated with the use of the optically thin approximation along hellocentric lines of sight from 1 AU. The calculated intensity patterns vary in time and reproduce qualitatively the groove patterns observed by SWAN on SOHO during the solar minimum (KyrS1/i et al. 1998; Bertaux et al. 1999) (see Fig.2, left-hand panel). The shape of the profile is sensitive to the adopted latitudinal profile of the ionisation rate (Fig.2, right-hand panel). However, during the solar miaxnimum the groove seems to have disappeared from SWAN images, which suggests that contrary to the ionization model adopted, the solar wind is symmetric during that phase. The profiles of the ratio of backscattered intensity at various latitudes to the intensity at crosswind in ecliptic show correlation with the ionisation model. The intensity distribution has some north-to-south asymmetry (see particularly Fig.2, right-hand panel) although the ionisation model used is symmetric about the solar equator. This is caused by the tilt of the gas inflow direction to the solar equator.
-131 -
M. Bzowski et aL Pole-to-ecliptic intensityratio
o
~ 1.4 o
9
"
s-""';'~
.
.
.
1.3
.
.
.
.
.
.
.
.
.
,
.
.
.
.
.
.
.
.
.
.
.
.
.
Intensity at pole and ecliptic 1.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I_pol/ 1.3 1.2 A 1.1
.
/3_ratio
i
"~ 1.2
9 ,...i
1
1.1
..~ . ~...~
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
,
,
i
,
,
,
i
1986 1988 1990 1992 1994 1996 1998 2000 time [y]
0.9 0.8 0.7 1986 1988 1990 1992 1994 1996 1998 2000 time [y]
Figure 3. Correlation of Irati o intensity ratios for warm and slow gas, gaussian bulge model with the/~ratio ionisation efficiency ratio is clearly visible in the left-hand panel (see Equ.(1) for definitions). Ipole/{/pole} and Ie~l/(Iec]) values for 90 ~ off the upwind direction are presented in the right-hand panel. The quantities practically do not depend on the gas parameters at infinity.
The ratio of intensities above the poles to intensities in ecliptic in the crosswind plane is strongly correlated with the ratio of hydrogen ionisation rate in ecliptic to the rate at the pole. In Fig.3 the evolution of quantities Ir~tio and/~ratio is presented, where
/ratio -- /pole (t) / (/pole} and /~ratio -- /~ecl (t) / (/~ecl) /eel (t) / (Iecl} ~pole (t) / (/~pole)
(1)
and (/pole}, (/ecl)and (/3pole),
Bertaux, J.-L., KyrS1/i, E., Qu~merais, E., Lallement, R., Schmidt, W., Summanen, T., Costa, J., & MS&inen, T. 1999, Space Sci. Rev., 87, 129 KyrS1/i, E., Summanen, T., Schmidt, W., Mgkinen, T., Bertaux, J. L., Lallement, R., Qu~merais, E., & Costa, J. 1998, J. Geophys. Res., 103, 14523 Rucifiski, D. & Bzowski, M. 1995, Astron. Astrophys., 296, 248 Summanen, T. 1996, Astron. Astrophys., 314, 663
-132-
L y m a n ct i n t e n s i t y p a t t e r n in t h e inflow d i r e c t i o n of t h e i n t e r s t e l l a r g a s T. Summanen a , T. M~ikinen, E. Kyrhl~i and W. Schmidt aFinnish Meteorological Institute, P.O. Box 503, 00101 Helsinki, Finland Theoretical results by Summanen[1] showed that the local minimum of the interplanetary Lyman ce intensity in the inflow direction of the interstellar gas (wind direction), while observed in the sidewind direction at 1 AU and when high ionization rate is assumed near the heliomagnetic equator, disappears when the tilt angle of the heliomagnetic equator with respect to the heliographic equator is approximately 20 ~ Interplanetary Lyman c~ data measured by SWAN shows that the minimum in the interplanetary Lyman c~ intensities in the inflow direction disappeared in 1998. The maximum extent of the heliosperic current sheet, calculated using the radial and the classic potential field models from the photospheric field observations ([2], [3]) grew from 20 ~ to 60 ~ and from 40 ~ to 75 ~, respectively, in 1998. Thus, results show that the ionization rate which varies according to the heliomagnetic equator is sufficient to explain the disappearance of the local minimum in the interplanetary Lyman ce intensity data measured by SWAN around 1998. 1. I N T R O D U C T I O N The presence of the "groove", the local minimum in the interplanetary Lyman ce intensities near the inflow direction of the interstellar gas during the solar minimum in Prognoz-5 observations, was pointed out by Bertaux et al.[4]. They proposed that the depth of the groove depends on the configuration of the heliospheric current sheet around the heliographic equator. In the observations of OGO 5 there was seen one Lyman c~ intensity maximum in the inflow direction in 1969 and 1970 ([5], [6]). Thus, it was plausible, that a change from the groove to one maximum pattern would be seen also during this solar cycle. In this work we have compared theoretical results by Summanen [1], concerning the Lyman ce intensities in the inflow direction when the ionization rate varies as a function of the heliomagnetic latitude, to the observations of the SWAN (Solar Wind ANisotropies) instrument. 2. O B S E R V A T I O N S O F S W A N SWAN instrument [7] onboard the SOHO satellite has measured interplanetary Lyman c~ intensities in all directions of the sky since January 1996. Figure 1 shows SWAN data in the directions where the ecliptic longitude is 253 ~. The data is shown in two
-
133-
T. Summanen et al.
Figure 1. SWAN observations in the directions of the ecliptic longitude 253 ~.
-
134-
Charge exchange ionization rate of interplanetary hydrogen atoms... parts, the first is measured during Jan 21, 1996- Jun 24, 1998. The second part is measured after the failure and recovery of SOHO during Nov 20, 1998 - Nov 27, 1999. Note that, the sensitivity level of SWAN sensors changed during the data gap. The actual intensity levels of the recent measurements as compared to the earlier measurements are not known, because the Lyman ~ intensity of the Sun also changed during the failure. We have adjusted the level of the intensity to reflect the solar activity increase. In the first part of Figure 1 there is the local minimum present near the ecliptic plane almost all the time shown blue and magenta. In the second part of the figure this minimum is no longer present. It is difficult to say when exactly the local minimum disappears, because the observation direction is obscured by the satellite near the edge of the first part and by the Sun at the beginning of the measurements in the second part. There is also a large data gap in the second part. However, we can say that the local minimum is not visible in the data any more in 1999.
~
9
~
1200
60
:t
" ~,, 'l,J
400 N
0
Figure 2. Theoretically calculated interplanetary Lyman c~ intensities in the inflow direction of the interstellar wind (ecliptic longitude 254.5~ The distance of the observation point from the Sun is 1 AU, its ecliptic longitude is 164.5 ~ and ecliptic latitude is 0 ~
Figure 3. Maximum extents of the heliospheric current sheet calculated using T. Hoeksema's [2], [3] radial (solid line) and classic (dashed line) potential field model.
3. D I S C U S S I O N Summanen [1] calculated the interplanetary Lyman ~ intensities near the inflow direction of the interstellar gas using the hot model for the calculation of the interplanetary hydrogen density and an optically thin model for the calculation of the interplanetary Lyman c~ intensity. In the optically thin model the intensity is calculated as a line-of-sight integral over the density along the line, the phase function and a g-factor corresponding
-135-
T. Summanen et al.
a flat solar Lyman ~ line profile. Details are presented for example in [8]. The calculation of the ionisation factor of the interplanetary hydrogen atoms was modified to take into account the changing ionisation probability of an atom on its path through the solar system, when the ionisation rate of the interplanetary hydrogen atoms is assumed to vary as a function of the heliomagnetic latitude. In the calculations the ionization rate of the interplanetary hydrogen was assumed to vary from 3.05x10 -~ s -1 (approximately from +90 ~ to -4-30~ of heliomagnetic latitude) to 7.0x10 -7 s -1 (at the heliomagnetic equator). The distance of the observation point from the Sun was 1 AU and its ecliptic longitude was 164.5~ and ecliptic latitude was 0 ~ Other parameters used in the calculations are given in [1]. Figure 2 shows (keeping in mind the assumptions of the hot model, the optically thin model and the used ionisation rate) that the local minimum in the interstellar wind direction disappears when the tilt angle of the heliomagnetic equator with respect to the heliographic equator is around 20 ~ Figure 3 shows the maximum extent of the heliospheric current sheet using data from [2]. It shows that the maximum extent of heliospheric current sheet increased from 20 ~ to 60 ~ according to the radial model and from 40 ~ to 75 ~ classic potential field model calculated using photospheric field observations [3] in 1998. The comparison of the results presented in Figures 1, 2 and 3 show that the assumption of the ionisation rate which varies according to the heliomagnetic equator is sufficient to explain the disappearence of the local minimum in the interplanetary Lyman c~ intensity data measured by SWAN around 1998. However, a full analysis of the disappearance of the groove would require a model which take into account the change of the tilt angle as a function of the solar cycle. 4. A C K N O W L E D G E M E N T S The authors thank Todd Hoeksema for the data used in the Figure 3 and Academy of Finland for supporting this work. REFERENCES 1. T. Summanen, Astrophys. and Space Sci. 274, 2000, 143. 2. J. T. Hoeksema, The Wilcox Solar Observatory www-pages, http://quake.stanford.edu/-wso/Tilts.html, visited July, 2000. 3. J . T . Hoeksema, Solar Wind Seven, COSPAR Colloquia Series Vol. 3, eds. E. Marsch and R. Schwenn, Pergamon Press, 191, 1992. 4. J.L. Bertaux, E. Qu~merais and R. Lallement, Geophys. Res. Lett. 23, 1996, 3675. 5. J.L. Bertaux and J. E. Blamont, Astron. Astrophys. 11, 1971, 200. 6. G.E. Thomas and R. F. Krassa, Astron. Astrophys. 11, 1971, 218. 7. J . L . Bertaux, E. KyrSl~i, E. Qu(~merais, R. Pellinen, R. Lallement, W. Schmidt, M. Berth~, E. Dimarellis, J. P. Goutail, C. Taulemesse, C. Bernard, G. Leppelmeier, T. Summanen, H. Hannula, H. Huomo, V. Kelh~i, S. Korpela, K. Lepp~il~i, E. StrSmmer, J. Torsti, K. Viherkanto, J. F. Hochedez, G. Chretiennot, R. Peyroux and T. Holzer, Solar Phys. 162, 1995, 403. 8. T. Summanen, R. Lallement and J.-L. Bertaux, J. Geophys. Res. 98, 1993, 13 215.
- 136-
N e w R e s u l t s d e r i v e d from P i o n e e r 10/11 U V d a t a Horst Scherer ~ and Klaus Scherer b ~Institut ffir Astrophysik und Extraterrestrische Forschung der Universitgt Bonn, Auf dem Hiigel 71, 53121 Bonn, Germany bdat-hex, Torstrafie 3, 37191 Katlenburg-Lindau, Germany In earlier endeavors the Pioneer 10/11 resonantly backscattered hydrogen (1215 /~) and helium (584 ~) UV glow data from the interplanetary medium had been studied in low spatial and time resolution (except at and during planetary encounters). We have now completely reanalyzed the Pioneer 10/11 UV-data from launch towards 1990, which enabled us to study highly resolved data in up to now unknown details. From the combined hydrogen and helium data over the entire period we can show that the optical thin models are good approximations also for the hydrogen. As example for the highly resolved spatial data we show the region of the highest intensities of resonant hydrogen and helium glow close to the direction of the inflowing interstellar medium. 1. T h e d a t a analysis The Pioneer 10/11 UV instruments are mounted with an angular offset onto the spinning spacecraft. It described in detail in Carlson and Judge 1994 [1]. Every two spin periods the data collection switches between the hydrogen and helium channels, which is difficult to handle when the synchronization is lost. We reanalyzed raw data with an improved algorithm, including corrections for the dead currents, and especially, the handling of the data at the boundary between the collection of the hydrogen and helium data. The results are presented in the following figures. 2. T h e i n t e r p l a n e t a r y glow Before 1974 the events in both panels are most probably connected with planetary objects, while the events after 1974 are caused by stars. The helium channel shows an optically thin behavior weakly modulated by the solar cycle. Beyond 1984 the helium channel is contamined by cosmic rays and shows therefore an unusual strong increase [3]. In the hydrogen channel the solar cycle modulation is more prominent. The ratio between the helium and hydrogen intensities are constant until 1984 when the cosmic ray contamination of the helium channel begins. Beside this contamination, it follows from the constant ratio between the hydrogen and helium data that also hydrogen can be fairly well modeled by an optical thin approxi-
-
137-
H. Scherer and K. Scherer
distance (~ 30 AU) over the period 1972 to 1984. In the left panel the glow intensities observed by Pioneer 11 are shown. As it can be seen the hydrogen chanel deviates from all known models (e.g. [2]). This may be caused by a degradation of the instrument beginning at 1977. Nevertheless, the helium data show the same behavior as for the Pioneer 10 data, except that there is no modulation caused by the cosmic rays, because Pioneer 11 has not reached large heliocentric distances during solar minium. Hence, the low energetic cosmic rays could not reach the spacecraft and contamine the data. 3. P l a n e t a r y and i n t e r p l a n e t a r y f e a t u r e s
In left panel of Fig. 2 an one minute clock angle average for the one year period 1972 is shown. It can be seen in the left panel of Fig 2, that there are some objects in the field of view. A precise identification of the objects is at the moment not possible because of problems with the old position data. Nevertheless, a preliminary identification can be made: day
- 138-
New results derived from Pioneer 10/11 UV data
136-138 (see right panel of Fig: 2) is most probably connected with the most emission region, expected in the direction of the inflowing interstellar medium, while during day 152-153 and day 218-225 the Jovian system comes into the field of view. For the days 219-221 in 1972 this is further resolved in Fig 3, where the 5 m i n time averaged spatially resolved H-Lyman-a (left panel) and helium (right panel) data are shown. The x- and y-axes show the heliocentric longitude and latitude in degrees, respectively. The intensity given in Rayleigh is color-coded. The start points of the 5 min. intervals are given by the numbers inside the circles. Gaps are caused by missing data. The intensity increase of the H-Lyman-a glow begins later than the intensity increase in the short-wave channel. This is caused by the fact that the Io-torus is in the ecliptic by a factor 5 more extended than Jupiter. Moreover, Jupiter is more or less a point source, which is indicated by the 'hot spot' in the left figure between 18:55 to 21:55. The short-wave channel does not show such features, but the intensity is completely smeared out through the entire circles, caused by the enlarged structure of the Io-torus. 4. S u m m a r y
-
139-
H. Scherer and K. Scherer
backscattered interplanetary glow can be modeled by a optical thin approximation. Additionally, we have presented highly spatial and time resolved observations of the Jovian system. We found that the intensity of the hydrogen channel is most probably caused by Jupiters glow, while the short-wave channel is dominated by the emission lines of different elements in the Io-torus-plasma. To our knowledge, this is the first time that simultaneous observations in the hydrogen and helium glow were presented, except during planetary encounters. Acknowledgment
We are gratefully to Prof. Dr. Darrell L. Judge, who provided us with the raw data of Pioneer 10/11 UV-instruments. REFERENCES 1. Carlson, R.W., Judge, D.L., J.Geophys. Res. 79, (1974), 3623 2. Scherer, H., in: The Outer Heliosphere: Beyond the Planets (Eds. Scherer,K., Fichtner, H., Marsch, E., Copernicus Gesellschaft, (2000), 91-138 3. Scherer, K., Scherer, H., private communication, 2000
-
140-
Oral papers and posters
C O M P O S I T I O N A N D C H A R A C T E R I S T I C S OF T H E L O C A L INTERSTELLAR CLOUD AND THE INNER SOURCE OBTAINED F R O M P I C K U P IONS G. Gloeckler Department of Physics and IPST, University of Maryland, Department of Oceanic, Atmospheric and Space Sciences, University of Michigan.
Local Interstellar Cloud (LIC) and of the newly discovered "Inner Source". Knowledge gained from this work will be reviewed with an emphasis on LIC characteristics, such as the isotopic and elemental composition of the LIC gas, its density, temperature and ionization state, and limits on the strength of the LIC magnetic field. Future directions for further dramatic advances in pickup ion observations will also be discussed. published in: Space Sci. Rev., in press, 2001
H A T O M V E L O C I T Y D I S T R I B U T I O N S IN H E L I O S P H E R I C INTERFACE V. V. I z m o d e n o v (1), Yu. G. Malama (2), M. Gruntman (3) and R. Lallement (4) (1) Department of Aeromechanics and Gas Dynamics, Faculty of Mechanics and Mathematics, Lomonosov Moscow State University, Vorob'evy Gory, Moscow, 119899, Russia, (2) Institute for Problems in Mechanics, Moscow, Russia, (3) University of Southern California, Los Angeles, California, (4) Service d'A6ronomie, Verrieres-le-Buisson, France. We study evolution of the velocity distribution function of interstellar neutral gas in the heliospheric interface, the region of the solar and interstellar wind interaction. We numerically solve the kinetic equation for neutral gas selfconsistently with hydrodynamical equations for plasma. Neutral and plasma components interact by charge-exchange. This interaction disturbs of atom velocity distribution, which is assumed Maxwellian in the circumsolar interstellar medium. To describe a complex of velocity distribution of interstellar atoms in the heliospheric interface we introduce three atom populations in addition to the primary interstellar neutrals. We present velocity distributions of these populations and discuss their kinetic properties.
-141
-
Oral papers and posters
E S T I M A T I O N OF I N T E R P L A N E T A R Y H Y D R O G E N F L O W PAR A M E T E R S F R O M S W A N L Y M A N c~ O B S E R V A T I O N S E. KyrSl~i (1), J. Jaatinen (1), T. Summanen (1), W. Schmidt (1), T. M~kinen (1), R. Lallement (2), E. Quemerais (2), J. Costa (2) and J . L . Bertaux (2) (1) Finnish Meteorological Institute, Geophysical Research Division, P.O. Box 503, FIN-00101 Helsinki, Finland, (2) Service d'Aeronomie du CNRS, BP. 3, 91371 Verrieres-le-Buisson, France. The SWAN instrument on board SOHO has measured the full sky distribution of Lyman ~ radiation since the beginning of 1996. The full sky coverage and the possibility to obtain spectral measurements by using the hydrogen cells of SWAN make it possible to investigate parameters controlling the flow of interstellar hydrogen through the heliosphere. Here we present results from the fitting of a single-scattering model of the interplanetary Lyman alpha radiation to SWAN spectral data. The new results, that still have considerable uncertainties, show that the inflow direction is 253 deg (longitude) and 11 deg (latitude) that differ only slightly from the earlier results. The same is true for the average inflow velocity, 21 km/s. The unperturbed temperature of the gas, 12000 K, is, however, much larger than previously estimated. The ratio between the radiation pressure and the gravitational attraction was estimated to be 0.99 indicating nearly force-free propagation of H-atoms in the interplanetary space. The data used in this analysis are from measurements obtained during the solar minimum in 1996 and measurement directions are toward areas in the sky that are void of stars.
I N F L U E N C E OF A T I M E V A R I A B L E S O L A R H - L Y M A N ALPHA LINE PROFILE ON LISM-PARAMETER DEDUCTIONS FROM INTERPLANETARY GLOW SPECTRA
THE
H. Scherer (1), H.J. Fahr (1), M. Bzowski (2) and D. Rucifiski (2) (1) Auf dem Hiigel 71, 53121 Bonn, Germany, (2) Space Research Centre of the Polish Academy of Sciences, Bartycka 18 A, PL-00-716 Warsaw, Poland. During a period of 18 months 5 Hubble-Space-Telescope-GHRS interplanetary HLyc~ glow spectra were obtained with different lines of sight from different positions of the earth in its orbit. No common parameter set for density, temperature and velocity of the interstellar hydrogen could be decuced from these Hubble-space-Telescope data. Despite application of a radiation transport model with angle-dependent partial frequency redistribution, selfabsorption by interplanetary hydrogen, realistic solar HLya emission profile, and a time-dependent hydrogen model there still remain some distinct discrepancies between data and the theoretical description, mainly manifest over time scales of the order of a year. These residuals could be explained by possible variations in time of the spectral shape of the solar HLy~ line profile, adopted in earlier attempts of data-interpretation as constant when modeling the radiation pressure and the resonance intensities. The influence of such a variation in time of the solar HLya line profile on the spectral shift of the peak of modeled intensity spectra is tested. Also, the influence on frequency integrated line of sight intensities is checked. Thereby, the possibility to see the effect of such a variation in time of the solar HLya line profile in older data sets (e.g. Pioneer 10/ 11 glow data) is proven.
-
142-
General Discussion Judge to Quemerais: Isn't your measurement depending on the direction of the scan? Quemerais: Yes. So, one thing will be to try directions different from those we had before. Zank to Quemerais: Can you actually look in all directions? Quemerais: No, we don't have the time to make full coverage of the sky. Sreenivasan to McNutt: Have you tried Lattice-Boltzmann techniques? McNutt: No, we haven't used the Lattice-Boltzmann approach. Jokipii to McNutt: I think you are really optimistic on this going to the next order in kinetic theory. There is a magnetic field threading through all of this, you got to have 81 viscosity coefficients to worry about, and there is anisotropic transport. McNutt: I won't argue that point with you. But right now I have just been talking about going to the next order just in the neutral component, so you are not worried about the magnetic field. Jokipii to McNutt: If you are talking about the neutral gas, your mean-free-path is of the size of the system. McNutt: It's interesting to look into the problems of radiation transfer in stellar atmospheres that were solved back a few decades ago. If you look at that treatment of kinetic t h e o r y - it also comes up in neutron kinetic t h e o r y - it turns out that one can actually get rigorous solutions to the kinetic equations that are good down to 1/10 of the meanfree-path. Fahr to Baranov: There is something which you did not take into account, namely the fact that - by producing new ions - you produce a new proton population which doesn't go the way the other protons go. So, you have more or less a two-stream instability situation and you can drive instabilities as well as have a change in the electrical conductivity of the plasma. That may mean you have to apply non-ideal MHD by including an anonalous resistivity. Baranov: Thank you very much for this recommendation. Bzowski to Kyr61g: The radiation pressure value you derived agrees perfectly with the value I derived in my radiation pressure model based on totally independent measurements. What concerns the ionization rate, in the ecliptic I have a slightly higher value in my model, namely 5 x 10 -7 rather than your 4 to 4.5. KyrSlg: Ours was not in the e c l i p t i c - it was at high latitudes. We avoided the ecliptic because of the solar wind anisotropy which poses a more complicated problem. Zank to Gloeckler: You began by showing interstellar pick-up ions as tails of the distributions. Do you see those same tails on the distribution of the inner source ions during quiet periods and do you see similar anisotropies of the inner source? Gloeckler: The inner source is getting pretty weak in the tail, and we are too far away from the inner source. Solar Probe would be good for studying this.
-
143-
General Discussion
Gruntman to M5bius: A lot of pick-up ions are born within 1 AU [and, therefore, are not interstellar]. What about them? MSbius: Yes, this is a good point. We haven't modelled that yet. Lee to Gloeckler: I was trying to think about the really outstanding problems in the transport of interstellar pick-up ions and two came to my mind. One is these very large time variations in the intensities of, in particular, h e l i u m - larger than you would expect from the changes of the UV radiation. The other is these very high energy tails and the puzzling fact that they are not observed in the high-speed streams but in the in-ecliptic low-speed streams. What is the acceleration mechanisms? I know that Len Fisk has suggested transit-time damping but why doesn't it really occur in the high-speed streams? Gloeckler: Regarding the tails, after all, the fast wind is very smooth - the density, temperature and other solar wind parameters, like the ratio of protons and alphas, are very constant with a least amount of fluctuations. So, Len Fisk would probably a~'ee that there is also a different level of turbulence. As to the first p r o b l e m - these puzzling large variations in the pick-up ion fluxes- I don't know the answer to that. Question to Khisamutdinov: Which term is used in your model for charge exchange cross section? Khisamutdinov: In our new method of modelling the general property of a monotonic non-decreasing function was used.
Session 3: Messengers from Outside
This Page Intentionally Left Blank
The d i s c o v e r y a n d early d e v e l o p m e n t of the field of a n o m a l o u s c o s m i c rays H. Moraal School of Physics, Potchefstroom University for CHE, Potchefstroom 2520, South Africa, fskhmC~uknet, puk.ac.za This contribution is the first of three that review the discipline of anomalous cosmic rays. This paper concentrates on the discovery of anomalous cosmic rays in 1973 and the first 20 years of observation, modelling, and the early theoretical understanding of the subject, with the emphasis on the latter. The papers of Heber and Cummings (2001) and le Roux (2001) that follow after this one, respectively review the subsequent experimental and theoretical developments. 1.
DISCOVERY
Anomalous cosmic rays were discovered by Garcia-Munoz et al. (1973), who announced that four points on the Helium spectrum, between 20 and 50 MeWnucleon, observed on the IMP5 satellite during 1972, lay much higher than expected. In Figure 1 these points are marked by the letters A, B, C, and D, and it shows that the He intensity is so high that it exceeds even that of protons. It also deviates strongly from the form proportional to kinetic energy, which is the expected shape in the adiabatic limit of cosmic ray transport.
Figure 1. Proton and helium spectra observed by Garcia-Munoz et al. (1973), in which the data points marked A, B, C, and D show the anomalous character of the Helium spectrum.
:* 9 10
t IO0
~00
The authors drew attention to the tact that this behaviour had not been seen during the previous solar minimum in 1965, and noted that it could not be explained with "previously successful boundary conditions" for the cosmic ray transport equation. In a subsequent paper, Hovestadt et al. (1973) observed a high Oxygen intensity between 1 and 6 MeWnucleon, higher than expected from inferences from 1965 and 1969 measurements, and from the steep spectra of solar energetic particles below this energy range. These points are marked A, B, C in Figure 2 (left). It was uncertain whether the two Carbon measurements, D and E, were also enhanced, because they were upper limits. In this paper, little inference could be drawn about the origin of these particles.
-
147-
H. Moraal
The first definite combined m e a s u r e m e n t and interpretation of this new feature in cosmic ray spectra came from McDonald et al. (1974), who observed that O (and N) spectra were strongly enhanced relative to C and He at 3 to 30 MeV/nucleon. This is shown in Figure 2 (right), with the enhanced O points marked as A, B, C, D, and E. This paper seems to be the first one to use the term "anomalous" for these cosmic rays, a name by which they are still known. This term "anomalous" refers to the unexpected composition rather than to the spectral shape. This is an important distinction, namely that the O and N abundances are anomalously high relative to that of C (and even He). If all species had spectra that strongly deviated from the then well-known ':] ~ T " adiabatic limit, there would have been no story to tell, because one simply would have to search for a different rigidity dependence of the modulation. The true anomaly of these observations was that species like O and N, and to some extent He, showed abnormal abundances, but that others, notably, C and H did not. McDonald et al. (1974) also used observations from Pioneer 10, which was at a radial distance of a few AU, to infer that the radial gradient of these anomalous components was slightly positive. This excluded a solar origin, because then the gradient would have been strongly negative. On the basis of this evidence the authors concluded that "....this is most likely a new extrasolar component of cosmic rays". Figure 2. Spectra observed by Hovestadt et (1973) (left panel) al. which shows high Oxygen intensities marked by A, B, and C, and possibly high Carbon intensities, marked by D and E. The right panel by McDonald et al. (1974) definitely shows anomalously high Oxygen intensities (points A to E), but no high Carbon intensities.
~ '~
?.
.
~
T I
I0"*
........
I
. . . .
. ,..1
........
i
lOt
2.
I N T E R P R E T A T I O N AND T H E D E V E L O P M E N T OF A D I S C I P L I N E
Fisk et al. (1974) almost immediately came forward with an interpretation for the anomalous species, namely that they originated as interstellar neutral atoms, of species with high first ionisation potentials, which became singly ionised in the inner heliosphere, picked up by the heliospheric magnetic field and solar wind, and subsequently accelerated to cosmic ray energies. The acceleration mechanism itself was not specified in this first paper. This proposal of Fisk et al. connected three independent fields of study, which were only weakly developed at the time, namely (a) the properties and transport of neutral particles in the heliosphere, (b) the acceleration of low energy particles to cosmic ray energies, and (c) cosmic ray transport. Of these three, cosmic ray transport was perhaps the best understood, and it took a surprisingly long time to unify these fields to establish the anomalous cosmic ray paradigm. Figure 3 is a timeline of discovery that shows how the fields of anomalous cosmic rays, pick-up ions in the solar wind, shock drift acceleration, and cosmic ray transport theory
-
148-
The discovery and early development of the field of anomalous cosmic rays
ACR
Titneline
Figure 3. Timeline of discoveries in the field of anomalous cosmic rays were integrated over a period of 20 years into this single discipline. This paper is organised according to this diagram. Each of the discoveries marked on it will be discussed. Four of them are highlighted in heavy boxes, because I regard them as landmarks. They are (i) the ACR discovery itself (described in Section 1 above), (ii) the discovery of first order Fermi acceleration, also known as compressive shock acceleration, in 1977, (iii) the first unambiguous observation of pick-up ions, which are the seed particles of ACR's, in the solar wind, in 1985, and (iv) the determination of the charge state in the early 1990's 3.
PICKUP
IONS IN THE SOLAR WIND
The Fisk et al. (1974) hypothesis implies that the ultimate source of anomalous cosmic rays is neutral interstellar gas that is ionised in the inner heliosphere. This gas sweeps relatively freely through the heliosphere, due to the 25 km/s motion of the heliosphere
- 149-
H. Moraal
relative to its interstellar environment. When these particles get near enough to the sun they become singly ionised, mainly due to charge exchange with particles of the solar wind, the density of which falls off as 1/r ~ with radial distance r, as well as through photo-ionisation. In the early 1970's the properties of this singly charged component was studied by authors such as Axford (1972), Bertaux and Blamont (1972), Fahr (1968, 1972), and Thomas (1972). The main motivation for these studies was the effect that this freshly ionised interstellar component might have on the solar wind flow beyond several AU - the so-called mass loading effect. Calculations such as those by Axford (1972) showed that typical mean penetrating distances before ionisation should be 3 to 4 AU for H, O, and N (with first ionisation potentials of the order of 14 eV), while He (24,6 eV) easily penetrates to 0,5 AU, where the probability for photo-ionisation also becomes large. The presence of these pick-up ions could be indirectly inferred from the UV backscatter of radiation from the interstellar neutrals. Vasyliunas and Siscoe (1976), however, calculated the spatial distribution and energy spectrum of these ions, with the view to see whether they could be detected directly. The principle behind the development of the energy spectrum is illustrated in Figure 4. The top left panel shows that when the interstellar neutrals are ionised, they start to spiral about the heliospheric magnetic field with a speed almost equal to the solar wind speed of about 400 km/s (because the interstellar neutral speed is much smaller than this value). According to the top right panel, these particles then lie in a ring distribution relative to the magnetic field, all with the same pitch angle. The irregular component of the field causes pitch angle scattering, therefore the bottom left panel shows that the distribution develops into a shell. Finally, adiabatic and/or other losses will let the distribution develop into a sphere. If this sphere becomes equally filled, the speed distribution is flat, with a cutoff determined by the solar wind speed. For a typical value of = 450 km/s, the kinetic energy cutoff occurs at - 1 keV. This very high energy, compared to typical solar wind energy of a few eV, sets the two populations apart as far as acceleration mechanisms are concerned.
Figure 4b. The first direct observation of He pickup ions by M6bius et al. (1985).
Figure 4a. The evolution of the pickup ion distribution function, described in the text 4.
EARLY P R O P A G A T I O N AND A C C E L E R A T I O N MODELS
The first source mechanism for ACR's, due to Fisk (1976a), is not actually an acceleration mechanism, but rather an investigation of diffusion and trapping of singly ionised particles. If the particles are singly charged, then for Oxygen with energies of the order of 10 MeV/nucleon, the particles should have a very high rigidity, of several GV. Normally, a diffusion coefficient is the product of particle speed with an increasing function of rigidity. Fisk then looked at scattering and trapping situations where there is
-
150-
The discovery and early development of the field of anomalous cosmic rays no (or weak) speed dependence. In such cases, the diffusion coefficient is just as large as that for electrons of a few GV, and this then explains why the modulation may be very weak relative to that of galactic cosmic rays of the same kinetic energy per nucleon. Next, Fisk (1976b) described the process of transit time damping of fluctuations in the heliospheric magnetic field by the particles, as a specific acceleration mechanism. He apparently did this for two reasons, namely to interpret intensity increases in co-rotating interaction regions, as observed by McDonald et al. (1976), as well as to derive an acceleration rate for pick-up ions to anomalous cosmic ray energies. The appropriate momentum diffusion coefficient was derived in this paper, but it was not actually applied to a model calculation on ACR's. This was left for the paper of Keckler (1977) who calculated accelerated (and subsequently modulated) ACR spectra in the heliosphere. His results are shown in Figure 5. Curves I and 2 are calculated spectra for galactic and anomalous species respectively. The solutions contain free parameters and they are approximate in the sense that spectra of the anomalous species are first calculated at a large radial distance by assuming that the spatial diffusion term in the transport equation can be neglected at these high rigidities. Then, standard one-dimensional modulation theory (including the spatial diffusion term) is used to modulate these spectra. Despite these approximations, this very first attempt to fit model ACR spectra, together with galactic ones, obviously produced quite remarkable results.
Figure 5. Model fits by Keckler (1977) to observed anomalous spectra. 5.
THE COSMIC RAY T R A N S P O R T E Q U A T I O N
The acceleration and transport models discussed in the previous section were more or less developed specifically as source mechanisms for ACR's because, at the time of their discovery, there were no obvious acceleration mechanisms available for this species. An exception is, perhaps, the process of shock drift acceleration in quasi-perpendicular shocks (to be described in the next section). The problem with this mechanism, however, was that it only gives a factor of-- 3 acceleration per shock crossing, which is not nearly enough to accelerate pick-up ions form -- 1 keV to several 10's of MeV. Furthermore, the notion of a quasi-perpendicular solar wind termination shock (SWTS), which could produce this acceleration, was almost non-existent at the time, although Jokipii (1968) drew attention to the possibility of acceleration in the termination shock of a stellar wind. This lack of a universal acceleration mechanism changed drastically in 1977 with the discovery of so-called first-order Fermi, or compressive, acceleration. This key discovery eventually led to an integrated understanding of first-order Fermi and shock drift acceleration in the solar wind termination shock as the source of ACR's. This development, however, took surprisingly l o n g - almost 10 years before all the key physical effects were properly understood, and another five years or so before extensive modelling of these effects was being done. The reason for this slow progress is probably the incremental increase in our understanding of the content of the cosmic ray transport equation, which was first
-151 -
H. Moraal derived by Parker (1965) and subsequently by Gleeson and Axford (1967). All the relevant acceleration and transport effects (even stochastic or 2 ~d order Fermi acceleration in principle) are contained in this robust and very general equation which has not been augmented in any way during its 36 years of existence. Its significance was extensively reviewed by Fisk (1999), and because of its central importance I write down the equation in full. It has the form of a continuity equation in configuration and momentum space" ~f / ~t + V.S + p-2 (< p > pZf) _ Q, where f is the directionally averaged cosmic ray distribution function. Its relationship with the differential particle density, U z,, expressed as particles/volume/momentum, is U p - 4 z p Z f . Webb (1976) discussed this form of the equation in detail and showed that the rate of momentum change, , for adiabatic compression/expansion is given by / p - 1 / 3 V . V f / f , (instead of the more generally perceived value of -1/3 V. V ), where V is the plasma flow (solar wind) velocity. The source function on the right hand side is such that Q / 4 z p 2 is the number of injected particles/volume/momentum/time. The differential streaming, or flux density is given by S / 4zp: in particles/area/ momentum/time, where
S - C V f - K . V f - C V f - K S . Vf +S~rir~, and Sdri]t - K r ( B / B ) x V f . Here C - -1/3 (Olnf/Olnp) is the so-called Compton-Getting coefficient, ~r - ~P/3B, where P is particle rigidity, and the diffusion tensor is separated into symmetrical and antisymmetrical parts as K-,,
0
0
0
0
0
;c
0
0
0
I -ll 1 0
where K ~'describes scattering in the field irregularities parallel and perpendicular to the background field, and K T describes gradient and curvature drifts. The flux was first written down by Gleeson and Axford (1968). When these quantities are inserted into the continuity equation, they produce the cosmic ray the transport equation in the two equivalent forms
~f / ~ t - V . ( K S . V f + K r . V f ) + V . V f --1/3(V.V)~f / ~ l n p - Q O f / ~ t - V . ( K S . V f ) + ( V + Vdrilt).Vf - 1/3 ( V . V ) ~ f / O l n p - Q,
and
(la) (lb)
where Vdr;/, - V x K T B / B is the average gradient and curvature drift velocity of the distribution function in the inhomogeneous background field. The form (la) is nearest to the original form of Parker (1965). Jokipii et al. (1977) emphasized the equivalence of the two forms, as well as the subtlety that while the distribution drifts with the velocity, the drift flux of particles is given by the term S drit, above. This means that bulk drift motion of the particles is due to gradients in the field, but the (measurable) drift flux is caused by gradients in the density. The two quantities are related by V. Sdr;/t --Vdrtp. Vf 9 Isenberg and Jokipii (1979), Burger et al. (1985), and Burger (1987) showed that the drift velocity Vj,it, is complete, i.e., that gradient and curvature drift effects are included in this simple expression without any approximation, provided only that the distribution can be expanded in a 0th and 1st order moment only. This near-isotropic approximation is an excellent one for cosmic rays in almost all situations. This comprehensive description of gradient and curvature drift is shown as a separate box in the line of
-
152-
The discovery and early development of the fieM of anomalous cosmic rays discovery on the right h a n d side of Figure 3, because it has played a crucial role in our u n d e r s t a n d i n g of ACR t r a n s p o r t and acceleration. F o r m a n and Gleeson (1975) emphasized t h a t electric field drift is implicitly contained in the convective t e r m of the equation, because the induced electric field in the highly conducting p l a s m a is E = - V x B , so t h a t electric field drift velocity is given by V E - (E x B ) / B 2 - - ( V x B) x B / B 2 - [ B 2 V - ( V . B ) B ] / B 2 -V-V,~-Vs. Thus, the component of the convection velocity perpendicular to the background field is the electric field drift velocity, and this drift should not be added post-hoe. An i m p o r t a n t feature of the t r a n s p o r t equation is t h a t its validity is not limited to gradually varying plasmas, i.e. those with small spatial gradients. Isenberg and Jokipii (1979, 1981) showed t h a t as long as the cosmic ray distribution r e m a i n s nearly isotropic, the equation can be applied to discontinuous changes in plasma flow. This has led to the successful description of cosmic ray modulation in and around the heliospheric neutral sheet, first described by Jokipii et al. (1977), as well as of the process of first order Fermi acceleration in eollisionless astrophysical shocks. 6.
FIRST ORDER FERMI OR COMPRESSIVE
SHOCK ACCELERATION
The discontinuous compression and reduction in flow speed in a shock causes the term 1/3(V.V)Of/Olnp in the t r a n s p o r t equation to represent adiabatic heating, or acceleration. As long as the distribution function remains quasi-isotropic, the t r a n s p o r t equation can be applied to such a situation. Krymski (1977), Axford et al. (1977) and Blandford and Ostriker (1978) therefore showed t h a t in a plane, steady shock, where the field lines are parallel to the normal on the shock front, the solution of the transport equation is t h a t a mono-energetic distribution function in the u p s t r e a m medium is accelerated to a power law of the form f ~ p-3.,/~.,-~, where s - V 2/V~ is the compression ratio of the shock. For a strong shock, this value is s = 4, leading to a distribution function of the form f ~ p-4 or, equivalently, to an intensity spectrum ~ p-2. This solution is the result of matching the intensities on both sides of the shock, and integrating the t r a n s p o r t equation across it, which produces the result [S.n]~ - Q ,
(2)
i.e. t h a t the difference in the s t r e a m i n g vector across the shock equals the source strength. Since this m e c h a n i s m holds for a parallel shock, in which the background magnetic field goes t h r o u g h it continuously, this field has no effect on the solution. The accelerated spectral shape is independent of s:, although the acceleration is caused by the process of pitch angle scattering off the irregularities in the field on both sides of the shock t h a t cycles the particles back across the shock m a n y times before they eventually escape downstream. The reason why ~ is absent from the solution is t h a t the strength of the scattering only determines the rate of acceleration, which can not be derived from the steady state solution. The process is better illustrated by the alternative single-particle, microscopic description by Bell (1978a,b). This explicitly highlights the fact t h a t field irregularities move s times faster from the u p s t r e a m medium into the shock t h a n away from it into the downstream medium. Thus, when viewed from either u p s t r e a m or downstream, the particle sees the irregularities in the other medium always approaching it, i.e., it always makes a head-on collision. Bell showed t h a t (i) this leads to an energy gain per shock crossing of O(V/v), where V and v are flow speed and particle speed, typically with V <<
- 153-
H. Moraal
v, and (ii) repetitive crossings with this energy gain lead to the same power law spectrum as was found from the solution of the transport equation mentioned above. The value of ~; or the strength of scattering, determines the rate of the acceleration. In Drury (1983), and references therein, it is shown that the time scale to establish the above-mentioned power law up to momentum p is given by P
The process is called first order Fermi acceleration to distinguish it form the classical second order Fermi acceleration in randomly moving irregularities, where the acceleration rate is only O ( V / v ) ~, i.e., of second order in small numbers. Because it is a stochastic process that depends on the compression of field irregularities, appropriate alternative names are also compressive or diffusive shock acceleration. First order Fermi acceleration was immediately applied to the blast waves of supernova remnants to demonstrate that they could be the source of cosmic rays up to at least 1015 eV, providing a compelling explanation for the observed power law spectra with spectral index just somewhat softer than o~ p-2 (e.g. Bogdan and V61k, 1983, Moraal and Axford, 1983, Lagage and Cesarsky, 1983). It took a long time, however, to apply this mechanism to the origin of ACR's in the heliosphere. In hindsight, it seems that the main reasons for this delay were that (i) the notion of a permanent, steady solar wind termination shock in the outer heliosphere was only weakly developed at the time, and (ii) even when one thought about such a shock, it was realised that it had a perpendicular magnetic geometry (with the azimuthally directed field lines perpendicular to the normal on the shock, except near the poles), instead of a parallel one. Pitch angle scattering in such a perpendicular shock does not cycle particles back across the shock front repetitively, and it was, therefore, difficult to see how shocks in the outer heliosphere could cause the acceleration. Despite these conceptual problems, Webb et al. (1985) applied the process to the termination shock of a stellar wind, disregarding its magnetic field structure. The analytical solution of the spherically symmetric transport equation produced spectra, but only for a diffusion coefficient that was independent of momentum, which is not realistic. Spherical (or any non-planar) shocks introduce an additional upper limit for the acceleration, namely that the power laws can only be established for energies where ~: /V < r,.,
(4)
i.e., where the diffusive length scale is less than the shock radius. Otherwise, the upstream medium becomes too small to contain the particles that are scattered back into it. It is instructive to note that for an effective radial diffusion mean free path = 1.0P(GV) AU, bearing in mind that ~: =/~v / 3, and for a SWTS with a radius q. - 90 AU, this cutoff in the spectrum occurs at -- 200 MeV for protons, and at 95 and 23 MeV/nucleon, respectively, for singly charged Helium and Oxygen. Before a full and comprehensive application of acceleration models to ACR's developed, however, much more thought had to be applied to the role of the background magnetic and electric fields. These fields produce shock drift and shock drift acceleration, which were known long before first order Fermi acceleration, with the work of, e.g., de Hoffman and Teller (1950), Hudson (1965), and Chen and Armstrong (1975). This acceleration process seemed to be entirely different to that of first order Fermi acceleration. A series of four papers by J.R. Jokipii and his co-workers, however, developed a fully integrated approach to the two mechanisms. The first paper in this series, by Pesses et al. (1981) was exploratory in nature. It discussed (some of the) physical mechanisms without giving
-
154-
The discovery and early development of the field of anomalous cosmic rays detailed results. On the timeline of Figure 3 I show that the second paper of Jokipii (1982) did not apply shock drift acceleration to the SWTS as such, but it emphasized the complementary nature of the two mechanisms. Jokipii (1986) presented the first numerical solutions of the transport equation that included this integrated mechanism, while Jokipii (1987) highlighted that perpendicular shocks may indeed be more efficient accelerators than parallel ones, contrary to belief at the time. I will discuss the developments contained in these four papers in the next two sections, not necessarily in chronological order. 7.
SHOCK DRIFT AND SHOCK DRIFT ACCELERATION
The fundamental point in this mechanism is that the more perpendicular the geometry of a shock, the more acceleration it produces. In what follows I will only refer to the limits of purely parallel and perpendicular shocks. The acceleration does, however, occur for any angle between the shock normal and the magnetic field, and these more general results are written up in the Jokipii papers. Figure 6 summarises the relevant aspects of the problem. The top left panel shows that in a parallel, collisionless shock in a smooth field, a charged particle spirals through a shock without being aware of its presence. The same parallel geometry is drawn in the top right panel, but here the field has irregularities. These irregularities scatter the pitch angle of the particle, so that it can cycle back across the shock many times before it eventually escapes downstream, leading to the process of first order Fermi acceleration, described in the previous section. The bottom left diagram depicts a perpendicular shock. Here the outward pointing, frozen-in field moves through the shock with the flow velocity. Thus, the field in the downstream medium is compressed by a factor s, the shock compression ratio. The moving B field causes an induced electric field, E = - V x B , pointing upwards, and having the same value on both sides of the shock, because the field and flow velocity change by a factor of s and 1/s across the shock, respectively. In this case the particle does not spiral through the shock, but it E-cross-B drifts with the velocity given in section 5. As the particle moves through the shock, it experiences an upward directed gradient drift. The geometric reason for this drift is that the upstream gyroradius is s times larger than the downstream one. It is readily seen that the drift direction is given by Vj,.ii, in Section 5, because in this situation V x B / B = ( V x B ) / B + V ( 1 / B ) x B = V ( 1 / B ) x B , which is, indeed, upward. The magnitude in this case is a Dirac delta, written down by Jokipii (1987), but the virtue of the transport theory is that, as was mentioned in Section 5, this produces a finite drift flux, represented by the term Sjritt. Finally, while the shock drift is due to the gradient in the magnetic field, the particle gains its energy from the electric field, because this field obviously does more "upward directed" work while the particles is in the upstream medium than "downward directed" when it is downstream. A classical result, summarised in, e.g., McLoud and Moraal (1990), is that this process produces an energy gain of-- 3 for a strong shock (compression ratio s = 4). This gain is much larger than the (typically) small gain in the top right hand panel, but it only occurs once, because in this perpendicular geometry, pitch angle scattering seemingly does not cycle the particle back across the shock. It merely transports the particle forwards and backwards along the shock surface, while it E-cross-B drifts away into the downstream medium. Thus, the question mark in the bottom right panel means that it is uncertain how a quasi-perpendicular shock such as the SWTS acts as the source of ACR's.
-155-
H. Moraal
In view of this problem, Pesses et al. (1981) were first to point out that the combined mechanisms of first order Fermi and shock drift acceleration in the SWTS provided a logical source mechanism for ACR's, and they gave the first steps in the development of a full model. Figure 7 shows a meridian cut of the northern half of the heliosphere, with the magnetic field directions drawn in. In this so-called qA>0 state of the heliosphere, the Parker spiral field in the northern hemisphere points outward in the inner heliosphere, and curves out of the paper with increasing radial distance. The field is oppositely directed south of the neutral sheet, which lies in (or near) the ecliptic plane. The field's spiral angle is given by tan~ = ~ ( r - r , . u , ) s i n 0 / V , with 0 the polar angle and g~ the angular velocity of the sun. For typical values of ~ and V, this angle is > 45 ~ if r sin 0 > 1 AU. Thus, in the outer heliosphere this field is quasi-perpendicular, except in a cylinder with a radius of 1 AU over the poles. This means that the shock is quasi-perpendicular almost everywhere, and the physical process of particle transport and acceleration is best described by the bottom left panel of Figure 6. The almost azimuthal magnetic field produces a poleward directed electric field in both hemispheres. When some of these particles reach the poles, the first order Fermi process operates on them in the quasiparallel cylinder. While these particles are accelerated, they penetrate deeper into the heliosphere, and they then drift toward the neutral sheet in (or near) the ecliptic plane, to be observed as anomalous cosmic rays. The cycle may also repeat itself, because the particles E-cross-B drift, or convect outwards again, aided by the so-called neutral sheet drift. From there they land on the shock, drift poleward and gain energy once more.
Figure 7. Half meridian cut of the heliosphere showing drifts in the so-called qA>0 magnetic configuration.
Figure 6. Shock drift (left) and 1 st order Fermi acceleration (right) in parallel (top) and perpendicular (bottom) shocks.
This cycle will only operate if the particles do not escape downstream before they reach the poles. This requirement can be realised through a more refined application of the process of perpendicular diffusion: every time that a particle scatters its pitch angle along the mean field direction, its phase angle also changes abruptly, displacing it approximately one gyroradius perpendicular to the field. This effect scatters a fraction of the particles back across the shock, so that they can repeat the shock drift acceleration of the bottom left panel of Figure 6. This process will, however, only be efficient for particles
-
156-
The discovery and early development of the fieM of anomalous cosmic rays that have already gained a substantial amount of energy relative to the background plasma, because such high energy particles only drift a small fraction of one gyroradius downstream per gyro-orbit. Thus, when they are scattered in pitch and phase angle, they have a high probability to scatter back across the shock. Low energy particles, on the other hand, drift a large fraction of a gyroradius, or even multiple gyroradii per gyroperiod. Consequently, a scattering in phase angle will not be sufficient to cycle them back across the shock. This is the origin of the well-known injection problem for quasiperpendicular shocks, referred to again later. This shock drift mechanism, aided by perpendicular diffusion, enables one to estimate an upper limit for the energy gain in such a perpendicular shock. Jokipii (1982, 1990) pointed out that the electric field causes a potential difference between the ecliptic plane and the poles given by rt/2
rt/2
ArP - ~ Eds = f VB~.dO~ (VB~r~/~f2) f sinOdO ~ 200 MV, 0
(5)
0
which is independent of shock radius, r,., as long as r,. >> re, and where B has been expressed in terms of its value at earth as B - Be(r~ / r) 2 / (~/2cos~), using the numerical values Be = 5 nT, and V =400 km/s. Thus, if the particle manages to stay on the shock from ecliptic to pole, assisted in this by perpendicular scattering, it can gain an energy ZeAO or -- 200 MeV per charge. This number happens to be in the same order of magnitude as the curvature limit Vrs / ~ > 1 of the first order Fermi process in (4) above, and this is no coincidence: Jokipii (1982) showed that the energy gain of the shock drift mechanism in (5) formally has the same structure as that derived form the diffusive matching condition (2). In the shock drift picture dT= ZeEds= ZeVBds, with the important point being that the gain is strictly proportional to the distance travelled along the shock. Jokipii showed that the diffusive matching condition can be put in the same form. In that mechanism, however, the distance that the particle manages to stay on the shock is not only determined by the geometry of the shock, but also by the amount of perpendicular scattering. This equivalence establishes that all the shock drift considerations are properly contained in the diffusive matching condition (2), and that the transport equation contains all the relevant physics to describe shock acceleration. The paper of Jokipii (1986) was the first one to take all of these physical effects into account in a numerical solution of the cosmic transport equation, with the matching condition (2) across the SWTS, to calculate ACR spectra. Subsequently, a similar model was developed by Steenkamp (1995), and since then the numerical modelling of ACR spectra by several groups has become routine. There is one remaining point, namely that Jokipii (1987) showed that the rate of acceleration, calculated from (3), is much faster for a perpendicular shock than a parallel one. I f / ~ = r / r , and if ~• is calculated from the so-called hard sphere scattering model, then ~• / ~,~ = 1 / (1 + r/2) , and the acceleration rate in a perpendicular shock is 1+ 7/2 times as fast as in a parallel one. This scattering model may be oversimplified, but it serves to demonstrate the point. Values of 7/__ 10 are realistic, but there is an upper limit to this parameter. This is set by the requirement that the transport must remain diffusive to keep the anisotropies small, and it is given by 71< v / V , leading to r < ~;• which becomes a stringent condition for low energy particles. In fact, combining this with the curvature limit of (4), establishes that a quasi-perpendicular shock is an efficient accelerator when the diffusive length scale ~ • falls in the range
-
157-
H. Moraal
re < t r 1 7 7 <1",..
(6)
If the lower limit is violated it leads to the well-known, non-diffusive preacceleration/injection problem which remains one of the main research problems in this field, and which is further mentioned in Section 10. Basically, for typical parameters the pick up ions must be accelerated by a factor of = 100 before the perpendicular shock becomes efficient. If the upper limit is exceeded the system becomes too small to contain the particles. Steenberg and Moraal (1999) have pointed out that there is a danger to interpret the order-of-magnitude nature of the upper limits (4) and (5) too quantitatively. Numerical solutions of the transport equation indicate that this cutoff already becomes appreciable at energies at least 5 times lower than this limit. A parallel shock does, of course, not suffer from the lower limit in (6) and can accelerate particles directly from the pickup ion pool. One must also keep in mind that for strong scattering 77--1, ~,l--~• and drift effects in the background field become much less important. In this case the curvature cutoff (4) is just as much determined by ~,~. In fact, from the matching condition (2) it follows that it is always the effective radial value of ~: that determines this cutoff. In the case of weak scattering it is modified strongly by drifts. For strong scattering these drifts have little effect on the acceleration. In this sense the spherically symmetric no-drift solution of Webb et al. (1985) can be interpreted as a strong scattering approximation. The numerical solutions of Steenberg and Moraal (1999) are of the same character. Apart from the kinematics of particle transport discussed above, there are two other equally important theoretical developments, namely that of the back reaction of accelerated particles on the shock structure and on the wave fields that scatter them. These topics formed part of the theoretical development right from the beginning, e.g., in Axford et al. (1977), but they are still actively investigated, and extensively dealt with in the companion paper of le Roux (2001). 8.
DIRECT OBSERVATION
OF PICKUP IONS IN THE SOLAR WIND
The third heavily marked discovery in Figure 3 is the first direct observation of pickup ions. M6bius et al. (1985) observed He + ions of interstellar origin in the solar wind, by measuring their velocity distribution function. Their measurement is shown in Figure 4b, which shows the bump on the mass/charge = 4 channel around 10 keV, cutting off sharply at 4 times the solar wind energy. These measurements were done on the earthbound AMPTE satellite, and for this reason only the He + species could be detected, because its ionisation potential is sufficiently high to become ionised within 1 AU. The connection of this measurement with ACR's in Figure 3 is shown in a dashed line, because the authors did not mention ACR's explicitly, although the implications were certainly quite clear. Detection of the other species, which typically ionise at several AU already, had to await the launch of Ulysses. These first detections are described by Geiss et al. (1994a, 1994b) and Gloeckler et al. (1993, 1994). This detection has been such a success that, as described in Gloeckler et al. (2000), instruments on Ulysses and ACE now provide us with comprehensive elemental and charge state composition measurements in the range from typical solar wind energies (1 keV) to tens of MeV/amu. 9.
THE CHARGE STATE OF ANOMALOUS COSMIC RAYS
The final piece in the ACR puzzle was to unambiguously establish their charge state. The only method available to directly determine this charge state at ACR energies - 158-
The discovery and early development of the field of anomalous cosmic rays
requires the use of the earth's magnetic field as a particle rigidity filter. The basic principle is t h a t for a given energy, the rigidity of a particle decreases with increasing charge state. Thus, when singly charged ions penetrate into the magnetosphere and lose an additional electron through interaction with the upper layers of the atmosphere, their rigidity suddenly drops, trapping them in L-shells of the geomagnetic field to which they would not have had access if they had come in with the higher charge state. The presence of these particles can then be measured by low-orbiting polar satellites. This technique was pioneered on the COSMOS series of satellites in the mid- and late 1980's. With the planned launch of Sampex in 1992, a Russian/American study group on this topic was formed in about 1990. Their joint results were then presented in a highlight session of the 23 rd International Cosmic Ray Conference by Cummings et al. (1994), Panasyuk (1994), and Tylka (1994). Since then, the procedure has become routine, and the results are briefly discussed in the companion paper of Heber and Cummings (2001), and reviewed more extensively by Leske et al. (2000). 10.
FURTHER
DEVELOPMENTS
The experimental and theoretical developments described above established the field of anomalous cosmic rays as a mature discipline in the early 1990's, about 20 years after their discovery. I emphasised the development of the early theory, because without it, one would hardly be able to talk of a true discipline. After this paradigm was established, studies have actively continued during the last decade. These newer developments are described in the companion papers of Heber and Cummings (2001) and le Roux (2001), as well as in two recent, comprehensive reviews by Zank (1999) and Fichtner (2001). These newer developments are mainly about the problems of injection/pre-acceleration, mass composition, and the back reaction of the accelerated particles onto the plasma. All these effects determine the overall efficiency of the process. Parallel to this, numerical solutions of the cosmic ray transport equation continue to be applied to observed ACR spectra to verify to what extent the developing theories fit the newer observations, e.g., Moraal et al. (1999). The detection of multiple charges by Mewaldt et al. (1996) and their explanation, as reviewed by, e.g., Leske et al. (2000) and Jokipii (2000), poses interesting further challenges to this modelling. Modelling of transport and modulation effects in the downstream region of the SWTS is also becoming feasible and important. The SWTS itself has not yet been detected, but Adams and Leising (1991) already concluded that there is no other viable option as an accelerator, because if ACR O*, at a few MeV/nucleon for example, was accelerated outside the heliosphere and had thus travelled more than 0.2 pc, it would have become multiply charged at much lower energies than was subsequently seen. If the source is heliospheric, the SWTS seems to be the only option, because when the realistic numbers ~ = 1.0P(GV) AU and r,. = 90 AU are used in (3), the acceleration time scale is of the order of 5 years. This time scale was directly confirmed by numerical solutions of Steenberg (1998). It is fast enough to avoid excessive multiple ionisation. Energetic Neutral Atoms (ENA's), with energies up to at least 100 keV have been observed and modelled by, respectively, Hshieh and G r u n t m a n n (1993) and Czechowksi et al. (1999), and references therein. These particles originate form the charge exchange of ACR's and pre-accelerated pickup ions with plasma outside the SWTS. This new component implies that there are now at least five distinguishable particle populations in the entire discipline: (i) interstellar neutrals, (ii) pickup ions, (iii) anomalous cosmic rays,
- 159-
H. Moraal (iv) geomagnetically trapped radiation, and (v) energetic neutral atoms. This makes the discipline much richer and much more diverse than that of galactic cosmic rays. In addition, the number and energy density of anomalous cosmic rays inside the heliosphere (or, indeed, in the astrosphere of any star) may exceed that of galactic cosmic rays. These numbers can easily be calculated by integrating under the respective spectra. Integrating from the cutoff point, Vr,./~= 1, upwards in energy (which favours the galactic component) shows that the number density in ACR's is -- 3 times that of GCR's. Similarly, the ACR energy density is -- 1/3 of that of the GCR's. These values are calculated for the outer heliospheric dominate the average over the entire heliosphere due to the volume factor. Thus, anomalous cosmic rays are at least as important and abundant as galactic ones, and it has become anomalous to call them by that name. "Heliospheric cosmic rays" is more appropriate. REFERENCES
Adams, J.H. and Leising, M.D., Proc. 22 nd ICRC., Dublin, 3, 304-306, 1991. Axford, W.I., ed. C.P. Sonnet, et al. NASA Spec. Publ., 308, 609, 1972. Axford, W.I., Leer, E. and Skadron, G., Proc. 15th ICRC., Plovdiv, 11,132-137, 1977. Bertaux, J-L., and Blamont, J.E., Solar Wind, NASA Spec. Publ., 308, 661, 1972. Bell, A., MNRAS, 182, 147, 1978a. Bell, A., MNRAS, 182, 443, 1978b. Blandford, R.D. and Ostriker, J.P., Astrophys. J., 221, L29-L32, 1978. Bogdan, T., and V61k, H.J., Astron. Astrophys., 112, 129, 1983. Burger, R.A., Ph.D. Thesis, Potchefstroom University, 1987. Burger, R.A., Moraal, H., & Webb, G.M., Astrophys. Space Sci., 116, 107-129, 1985. Chen, G, and Armstrong, T.P., Proc. 1 5 th ICRC. 5, 1814, 1975. Cummings. J.R., Cummings, A.C., Mewaldt, R.A.., Selesnick, R.S., Stone, E.C., von Rosenvinge, T.T., Proc. 23 rd ICRC., Invited, Rapporteur, Highlight Vol., 475-482, 1994. Czechowski, A., Fahr, H.J., Fichtner, H., Kausch, T., Proc. 26 th ICRC, 7, 464-467, 1999. De Hoffman, F., and Teller, E., Phys. Rev., 80, 692-703, 1950. Drury, L.O'C., Rep. Prog. Phys., 46, 973, 1983. Fahr, H.J., Astrophys. Space Sci., 2, 496, 1968. Fahr H.J., Report 72-10, Forschungsber. der Astron. Inst. Bonn, 1972. Fichtner, H., Accepted for publication in Space Sci. Rev., 2001. Fisk, L.A., Astrophys. J., 206, 333-341, 1976a. Fisk, L.A., Adv. Space Res., 23, 415-423, 1999. Fisk, L.A., J. Geophys. Res., 81, 4633-4640, 1976b. Fisk, L.A., Gloeckler, G., Zurbuchen, T.H., and Schwadron, N.A., ACE2000 Symposium, ed. R.A. Mewaldt et al, 229-233, 2000 Fisk, L.A., Kozlovsky, B., Ramaty, R., Astrophys. J., 190, L35, 1974. Forman, M.A., and Gleeson, L.J., Astrophys. Space Sci,. 32, 77, 1975. Garcia-Munoz, M., Mason, G.M., Simpson, J.A., Astrophys. J., 184, 967, 1973. Gleeson, L.J., Axford, W.I., Astrophys. J. Lett., 149, Ll15, 1967. Gleeson, L.J., and Axford, W.I., Astrophys. Space Sci., 2, 431-437, 1968. Geiss, J., Gloeckler, G., Mall, U, von Steiger, R., Galvin, A.B., and Ogivlie, K.W., Astron. Astrophys., 282, 924, 1994a. Geiss, J., Gloeckler, G. and Mall, U., Astron. Astrophys. 289, 933, 1994b. Gloeckler, G., Geiss, J., Balsiger, H., Fisk, L.A., Galvin, A.B., Ipavich, F.M., Ogilvie, K.W., von Steiger, R., and Wilken, B., Science, 261, 70-73, 1993.
- 160-
The discovery and early development of the field of anomalous cosmic rays
Gloeckler, G., Geiss, J., Roelof, E.C., Fisk, L.A., Ipavich, F.M., Ogilvie, K.W., Lanzerotti, L.J., von Steiger, R., Wilken, B., J. Geophys. Res. 99, 17637-17643, 1994 Gloeckler, G., Fisk, L.A., Zurbuchen, H., and Schwadron, N.A. ACE2000 Symposium, ed. R.A. Mewaldt et al, 221-228, 2000. Heber, B, and Cummings, A., this volume, 2001. Hovestadt, D.O., Vollmer, O., Gloeckler, G., Fan, C.Y., Phys Rev.Lett., 31,650, 1973 Hsieh, K.C., and Gruntmann, M.A., Adv. Space Res., 13, 131-139, 1993. Hudson, P.D., M.N.R.A.S., 131, 23, 1965. Isenberg, P.A. and Jokipii, J.R., Astrophys. J., 234, 746, 1979. Isenberg, P.A. and Jokipii, J.R., Astrophys. J., 248, 845, 1981. Jokipii, J.R., Astrophys. J., 152, 799-808, 1968. Jokipii, J.R., Astrophys. J., 255, 716, 1982. Jokipii, J.R., J. Geophys. Res., 91, 2929, 1986. Jokipii, J.R., Astrophys. J., 313, 842, 1987. Jokipii, J.R. COSPAR Colloquia Series, 1, 169-178, 1990. Jokipii, J.R., ACE2000 Symposium, ed. R.A. Mewaldt et al, 309-316, 2000. Jokipii, J.R., Levy, H. and Hubbard, W.B., Astrophys. J., 213, 861, 1977. Kecker, B., J. Geophys. Res., 82, 5287-5291, 1977. Krymsky, G.F., Dokl. Akad. Nauk. SSR, 234, 1306, 1977. Lagage, P.O., and Cesarsky, C.J., Astron. Astrophys, 118, 223-228, 1983. Le Roux, J.A., this volume, 2001. Leske, R.A., Mewaldt, R.A., Christian, E.C., Cohen, C.M.S., Cummings, A.C., Slocum, P.L., Stone, E.C., von Rosenvinge, T.T., and Wiedenbeck, M.E. ACE2000 Symposium, ed. R.A. Mewaldt et al, 293-300, 2000. McDonald, F.B., Teegarden, B.J., Trainor, J.H. Webber, W.R., Astrophys. J., 187, L105 L108, 1974. McDonald, F.B., Teegarden, B.J., Trainor, J.H., and Webber, W.R., Astrophys. J., 203, L149, 1976. McLoud, W., and Moraal, H., J. Plasma Phys., 44, 123-136, 1990. Mewaldt, R.A., Selesnick, R.S., Cummings, J.R., Stone, E.C., von Rosenvinge, T.T., Astrophys. J., 466, L43, 1996. MSbius, E., Hovestadt, D., Kecker, B., Scholer, M., Gloeckler, G. and Ipavich, F.M., Nature 318, 426-429, 1985. Moraal, H., and Axford, W.I., Astron. Astrophys., 125, 204-216,1983. Moraal, H., Steenberg, C.D., and Zank, G.P., Adv. Space res., 23, 425-436, 1999. Panasyuk, M.I. Proc. 23 r~ ICRC., Invited, Rapporteur, Highlight Vol., 455-463, 1994. Parker, E.N., Planet. Space Sci., 13, 9, 1965. Pesses, M.E., Jokipii, J.R. and Eichler, D., Astrophys. J., 246, L85, 1981. Steenberg, C.D., Ph.D. thesis, University of Potchefstroom, South Africa, 1998. Steenberg C.D., and Moraal, H., J. Geophys. Res., 104, 24879 - 24884, 1999. Steenkamp, R., Ph.D. thesis, Potchefstroom University, 1995. Thomas, G.E., Solar Wind, NASA Spec. Publ. 308, 668, 1972. Tylka, A. Proc. 23 rd ICRC, Invited, Rapporteur, Highlight Vol., 465-474, 1994. Vasyliunas, V.M. and Siscoe, G.L., J. Geophys. Res., 81, 1247, 1976. Webb, G.M. PhD Thesis, University of Tasmania, Hobart, Tasmania, 1976. Webb, G.M., Forman, M.A., and Axford, W.I. Astrophys. J. 298, 684-709, 1985. Zank, G.P., Space Sci. Rev., 89, 413-688, 1999. Zank, G.O., Rice, W.K.M., le Roux, J.A., Lu, J.Y., and Mfiller, H.R., ACE2000 Symposium, ed. R.A. Mewaldt et al, 317-324, 2000.
-161 -
This Page Intentionally Left Blank
A n o m a l o u s c o s m i c rays: C u r r e n t and future theoretical d e v e l o p m e n t s J. A. le Roux Bartol Research Institute, University of Delaware, Newark, DE 19716, USA
Since the discovery of anomalous cosmic rays during the 1970s much progress has been made in clarifying the origin of these particles. However, the formation and transport of the anomalous component involve a complex chain of events cutting across different disciplines in space physics so that many of the details still need unraveling. A brief review of some of the current and future theoretical efforts in this regard is presented.
1. INTRODUCTION It is currently firmly believed on the basis of theory, simulations and observations that anomalous cosmic rays (ACRs) are formed when mostly interstellar pickup ions (PUIs) experience diffusive shock acceleration at the solar wind termination shock (TS). However, a lot of the complex details of PUI and ACR transport in the solar wind still have to be worked out. For example, it is not well understood (i) how PUIs are pre-accelerated while being transported to the TS by the solar wind, (ii) how PUIs are injected into the process of diffusive shock acceleration at the TS to become ACRs, (iii) how PUIs and ACRs modify the TS, and (iv) how the upstream transport of these particles should be described. A more detailed understanding of the above is essential for explaining why the ACR composition reveals an underabundance of the lighter elements such as hydrogen especially compared to the composition of PUIs [1]. A brief review of the current status and future direction of theoretical work and simulations in these research areas is presented.
2. PICKUP ION TRANSPORT TO THE TERMINATION SHOCK According to current wisdom it is only at a nearly perpendicular TS that the rate of standard diffusive shock acceleration is fast enough to explain the formation of the ACR component as observed [2]. Standard diffusive shock acceleration at a nearly perpendicular TS should be thought of in the context of ACRs undergoing repeatedly shock drift acceleration because scattering by magnetic field fluctuations keep the ACRs close to the TS. This mechanism is described by standard ACR transport theory in the limit of near isotropic ACR distributions [3]. The diffusive acceleration mechanism only works if the ACRs downstream move across magnetic field lines back to the TS which is more difficult than for a parallel TS where the path back to the TS is along field lines. This suggests the existence of an "injection" problem in that unaccelerated PUIs need to be pre-accelerated before crossing the threshold of
-
163-
J.A. le Roux
standard diffusive shock acceleration at a nearly perpendicular TS whereby the formation of the ACR component could be viewed as a two-stage acceleration process. An issue that needs to be resolved in this regard is whether more energetic PUIs are injected directly into diffusive shock acceleration at the TS after pre-acceleration in the upstream solar wind by interplanetary shocks and MHD turbulence, or whether low-energy PUIs injected indirectly into diffusive shock acceleration after local pre-acceleration at the TS itself contribute the most to the formation of ACRs. To resolve this issue a better understanding of PUI preacceleration in the upstream solar wind must be achieved.
2.1 Pickup pre-acceleration by MHD turbulence PUI pre-acceleration in the upstream solar wind itself is currently believed to be a twostage acceleration process with standard diffusive shock acceleration as the second stage occurring at quasi-perpendicular corotating shocks. No consensus exists currently on the dominant first stage acceleration mechanism~ On the one hand, it has been considered that the stochastic acceleration of PUIs by the cyclotron resonant interaction with Alfv6nic turbulence (2nd-order Fermi acceleration) [4] and by the magnetic Landau resonance (transit time damping) with fast mode magnetosonic turbulence [5] are responsible for the initial PUI preacceleration. These mechanisms are believed to be important between the forward and reverse corotating shock pair where there is enhanced turbulence. In contrast, mechanisms that operate at the shock itself, such as stochastic magnetic mirroring acceleration [6], and multiply reflected ion (MRI) acceleration [7] also known as shock surfing acceleration [8], are invoked. Recent observations of PUI spectra close to the Sun suggest that PUIs have accelerated "tails" all the time in the equatorial regions, leading to the speculation that PUIs are accelerated by shocks, but also by MHD turbulence in between and near shocks [9]. It is clear that this issue will continue to receive attention in the future. Regarding the stochastic pre-acceleration of PUIs by MHD turbulence, it was pointed out that fast mode magnetosonic turbulence is more important than Alfv~nic turbulence close to the Sun [5]. This makes sense because close to the Sun Alfv~nic turbulence in the solar wind frame is mainly directed away from the Sun whereby 2nd-order Fermi acceleration is ineffective. However, simulations valid for the quiet background solar wind suggest that Alfv6nic turbulence is more important for PUI pre-acceleration beyond a few AU from the Sun [10], where the amount of inward and outward propagating Alfv~nic fluctuations balances. Another interesting aspect about PUI pre-acceleration in the upstream solar wind is that this may have consequences for the ACR composition. Studies of self-consistent 2nd-order Fermi pre-acceleration of PUIs in the quiet solar wind [11, 12] predict that PUIs damp Alfv6nic turbulence according to an inverse mass dependence at large heliocentric distances so that H + has the largest effect in damping this turbulence. Consequently, because waves that resonantly scatter H + are damped, the efficiency of 2nd-order Fermi acceleration mechanism is biased in favor of the heavier PUI elements so that the pre-accelerated PUI tails represent a potential source for diffusive shock acceleration at the TS with a composition more resembling ACRs than unaccelerated PUIs. A similar trend was also noticed in simulations involving the stochastic pre-acceleration of PUIs due to interaction with large-scale fluctuations in the solar wind speed associated with corotating interaction regions [13]. Observations [14], theory [15] and simulations [16] suggest that MHD turbulence in the solar wind may consist predominantly of two-dimensional (2-D) turbulence with a minor component of Alfv6nic turbulence. However, the effect of 2-D turbulence on the stochastic
-164-
Anomalous cosmic rays: Current and future theoretical developments
pre-acceleration of PUIs still awaits investigation. 2-D incompressible MHD simulations suggest that strong stochastic turbulent electric fields associated with turbulent reconnection processes for a high Reynolds number appropriate for the solar wind can accelerate particles efficiently so that power law spectra result over a certain energy range [17]. This indicates that PUI pre-acceleration by 2-D turbulence in the outer heliosphere might be more significant than provided by Alfv6nic turbulence that typically results in softer exponential spectra.. 2.2 Pickup ion pre-acceleration by corotating shocks The idea to study particle acceleration at quasi-perpendicular interplanetary shocks using the gyrophase-averaged conservation of the magnetic moment across a shock whereby some particles are reflected (magnetic mirroring acceleration) and some are transmitted (shock drift acceleration) is well established [6]. Without scattering to return particles to the shock, however, this approach to first stage acceleration may have difficulty in accelerating particles across the threshold for standard diffusive shock acceleration. Models that combine magnetic moment conservation with allowing particles leaving the shock along magnetic field lines to return to the shock through scattering by magnetic field fluctuations, show substantial additional acceleration [ 18, 19] that might be sufficient for injection. It is only recently that the idea of MRI/shock surfing acceleration has been considered for 1st stage PUI pre-acceleration at quasi-perpendicular interplanetary shocks [7, 8]. According to this approach PUIs with a sufficiently small velocity component along the shock normal are reflected by the cross-shock electric field and energized by the motional electric field in a skimming motion along the shock front with maximum energies reached when their Lorentz force overcomes the cross-shock electric field. Test particle simulations predict that this mechanism should be able to accelerate PUIs up to energies of--1 MeV which should be sufficient to cross the threshold for standard diffusive shock acceleration. Such efficient acceleration requires a narrow shock ramp of the order an electron inertial length. Observations of quasi-perpendicular Earth Bow shock events indicate that the shock ramp width is typically of the order of an ion inertial length, but that there exist strong fluctuations inside the ramp which are much narrower [20]. Simulations including fine structure of the shock ramp show that MRI acceleration is effective [21]. The challenge for continuing research is to study the role and relative importance of the MRI mechanism as a first stage mechanism in comparison with the conservation of magnetic moment approach. Time-dependent MHD simulations of the solar wind show that corotating shocks weaken substantially with increasing distance from the Sun. The inclusion of PUIs in the solar wind MHD simulations raises the fast mode magnetosonic speed somewhat, causing a slightly further reduction in the compression ratio of corotating shocks [22]. This implies that corotating shocks beyond some distance from the Sun may become too weak to reflect and preaccelerate PUIs efficiently in the case of the MRI and the mirroring acceleration mechanisms so that injection into second stage diffusive shock acceleration might cease. The results from recent simulations [23] include an estimate at which heliocentric distance PUI injection into diffusive shock acceleration at corotating shocks ceases. It was found that in order to explain the observed delay between the arrival of interplanetary shocks and the peaks in the ~1 MeV energetic particle fluxes at -47 AU [24], the shock should lose its ability to inject particles into diffusive shock acceleration at --34 AU from the Sun where the corotating shock compression ratio falls below 1.5. This suggests that beyond ~34 AU the preaccelerated PUI spectra will decay due to the dominating effect of solar wind convection and
- 165-
J.A. le Roux adiabatic cooling. A recent interpretation of Voyager observations [25] at large heliocentric distances suggests that this is indeed the case, as the energetic particle fluxes associated with merged corotating interaction regions have a r -3 dependence instead of r -2 associated with convection but no cooling. This raises the question whether enough pre-accelerated PUIs will be injected directly into diffusive shock acceleration at the TS [26] to explain the observed modulated ACR spectra as suggested by ACR drift transport calculations based on preaccelerated source PUI spectra inferred from Voyager observations [27].
3. PICKUP ION INJECTION AT ANOMALOUS C O M P O N E N T
THE
TERMINATION
SHOCK
AND
THE
As mentioned above, it is only at a nearly perpendicular TS that the rate of standard diffusive shock acceleration is deemed fast enough to explain the observed ACR component. Unfortunately, it is at a perpendicular TS that PUIs will face the largest energy threshold for injection into this acceleration process. The pre-acceleration of PUIs in the upstream solar wind might be the solution, but it is not clear whether enough of them will arrive at the TS with sufficiently high energies. Recent work indicates that if the pre-accelerated PUIs observed at ---40 AU are extrapolated to the TS with a r -3 radial dependence [25] and injected into diffusive shock acceleration at the TS with a minimum injection energy of 50 keV, then there would not be enough PUIs to reproduce the observed ACR component [26]. Therefore, a point of considerable debate is whether low-energy PUIs can be pre-accelerated locally at the TS across the threshold for standard diffusive shock acceleration if the TS is nearly perpendicular. 3.1 The injection of low-energy pickup ions PUI magnetic mirroring at a nearly perpendicular TS is not an option for 1 st stage acceleration because only very high energy particles will be reflected [18], and simulations show that shock drift acceleration is only effective in the presence of high amplitude turbulence [28]. However, then shock drift can be interpreted as standard diffusive shock acceleration operating above a low energy threshold [29]. Such high levels of turbulence seem unlikely for the TS [30]. The MRI mechanism, on the other hand, is ideally suited for the pre-acceleration of low energy PUIs at the TS because of the preferential reflection and acceleration of PUIs with a small velocity component along the shock normal [7]. Test particle simulations show that unaccelerated PUIs can be efficiently pre-accelerated at a perpendicular shock to energies of --1 MeV which should be sufficient for injection [7]. This and other work also suggest that hard MRI spectra with an omnidirectional distribution function fMRI (P) ~ p-4 result for uncooled --1 keV PUIs. The implication of these MRI spectra are an associated large MRI pressure of the order of the upstream solar wind ram pressure given that a significant fraction of PUIs are MRI-accelerated as predicted by theory [7]. This suggests that self-consistent calculations need to be done to investigate whether TS mediation by the MRI particles and the ACR source that develops from these particles after injection will not weaken the TS to such a degree that MRI acceleration and injection becomes inefficient. This will be discussed in 3.2. It might be argued that the MRI acceleration mechanism is not the answer for local PUI pre-acceleration at the TS because it predicts the preferential pre-acceleration of the lighter
- 166-
Anomalous cosmic rays: Current and future theoretical developments
PUI elements [7] while the opposite has been concluded from the observed composition of the ACR component [1]. Based on two models for perpendicular diffusion, namely, either a resonant diffusion model using a classical scattering approach, or a postulated model determined by field line random walk, a recent calculation shows that a threshold velocity for injection into standard diffusive shock acceleration can be derived which, when applied to the locally MRI pre-accelerated PUIs at the TS, results in ACRs with a similar composition as observed [31 ]. The possibility also exists that the TS is not nearly perpendicular but somewhat more oblique on average. In this case stochastic magnetic mirroring acceleration becomes a candidate for low-energy PUI pre-acceleration because it operates at much lower energies than for a nearly perpendicular TS. Unaccelerated PUIs, however, do not have enough energy to undergo magnetic mirroring at the TS so that pre-acceleration upstream is needed [ 18, 19]. However, recent simulations show that unaccelerated PUIs can be injected into the magnetic mirroring mechanism at a more oblique TS with the aid of MRI acceleration [19]. This implies that at a more oblique TS low-energy PUIs might locally be pre-accelerated in two stages followed by standard diffusive shock acceleration as the third stage. Another argument is that the TS will only be nearly perpendicular on average, but some part of the time quasi-parallel when the threshold is lower. Diffusive shock acceleration of PUIs during times when the TS is quasi-parallel can then serve to bring PUIs across the threshold. Alternatively, PUIs might be pre-accelerated where the TS is most of the time quasi-parallel as expected at high heliolatitudes and then be transported down toward the equatorial plane along the TS front where the TS is expected to be nearly perpendicular. This would be effective when the polarity of the IMF was directed inward above, and outward below, the HCS, in which case particles would drift from high to low latitudes along the TS in both the northern and southern hemispheres. On the basis of work referenced it seems plausible that both low-energy PUIs locally preaccelerated at the TS as well those pre-accelerated upstream will contribute to the formation of the ACR component. The challenge for current and future research is to find which part of the PUI spectrum makes the dominant contribution. 3.2 How pre-accelerated pickup ions and anomalous cosmic rays affect the termination shock structure
One of the exciting discoveries to be made by the Voyager spacecraft is the solar wind TS structure and position, which will be the first in situ observation of such a structure. In this regard it will be interesting to know whether and to what degree pre-accelerated PUIs and ACRs can mediate the TS. If enough PUIs are injected into diffusive shock acceleration at the TS, simulations show that ACRs have the potential to modify the TS significantly [32]. This follows because ACRs are more or less non-relativistic particles with short enough mean free paths so that a large pressure gradient might develop at the TS. Unfortunately, the injection efficiency is at this stage a poorly constrained parameter due to uncertainties about how PUI injection occurs as discussed above. In recent work [33, 34], further progress was made by going beyond the ad-hoc approaches used to inject PUIs into diffusive shock acceleration at the TS in ACR transport models. This was achieved by using the main features of the theory for MRI acceleration of PUIs [7] to model the physics of injection for a nearly perpendicular TS. It was investigated within the context of a time-dependent, self-consistent model how the local MRI acceleration of unaccelerated PUIs, and ACRs formed by the subsequent injection of the MRI-accelerated
- 167-
J.A. le Roux PUIs into diffusive shock acceleration, affect the TS structure. The results indicate that both MRI-accelerated and A CR protons might significantly mediate the TS. This is due to the fact that a considerable fraction of the PUIs (-- 16%) are MRI accelerated of which a significant amount enters diffusive shock acceleration. Whereas the TS compression ratio is initially--3.1 due the presence of unaccelerated PUIs in the solar wind, MRI-accelerated PUI protons are found to weaken the TS further to a compression ratio of s = 2. Due to this TS modification, the TS structure develops a foreshock or foot region with a scale size of the order of the gyroradius of the highest energy MRI-accelerated PUIs which is --10 -3 AU. Viewed on a larger scale (> 1 AU), the foreshock region of the TS structure cannot be distinguished from the subshock and the compression ratio appears to be the initial value of--3.1. When including the effect of the ACR proton pressure on the TS, the subshock compression ratio appears to be s = 2.3 on this large-scale view of the TS as a large-scale TS precursor due the more energetic ACR protons is formed. This gives an idea of the modification effect of the ACRs on the TS structure. A small-scale view of the TS structure (< 1 AU) shows that the combined effect of MRI and diffusive shock acceleration of PUI protons reduce the subshock of the TS structure to a compression ratio of s = 1.6. It is interesting to note from these simulations that the TS modification by MRI and ACR protons results in a weaker cross-shock electric field so that the maximum energy that MRI acceleration bestows on PUIs is reduced significantly. Initially the maximum energy that the MRI-accelerated PUIs reach is -170 keV for a (sub)shock ramp with width e = 2 (width normalized to an electron inertial length), but after modification of the TS by both MRI and ACR protons the maximum energy is --74 keV. This reduced maximum energy was further enhanced by a factor o f - 2 to ~-148 keV due to adiabatic compression of the MRI-accelerated PUIs when crossing the TS. Based on the assumption that standard diffusive shock acceleration only applies when ACR anisotropies are small, the injection energy was calculated as --50 keV. This threshold allows for enough MRI-accelerated PUIs to be injected into standard diffusive shock acceleration so that the modulated ACR proton spectra upstream compare favorably with observational data recorded by the Voyager spacecraft. Injection is highly time dependent and is interrupted as the TS becomes too strongly modified and the cutoff energy for MRI acceleration falls to far below the threshold. In addition, injection might also be interrupted when the MRI acceleration process itself fails. This happens when the TS becomes so strongly mediated that reflection by the cross-shock electric field ceases. This is expected to happen when the upstream Mach number of the TS drops to below 1-2 [35]. Calculations show that if it is assumed that reflection stops when the Mach number is less than a critical value of 2, injection becomes sporadic as reflection is interrupted most of the time. However, the results discussed here are for an assumed Mach number value of 1.5 whereby reflection occurs continuously. Further simulations underscore the robustness of the model in that comparable results are yielded for e < 10. For a TS (sub)ramp width of less than about an ion inertial length, the MRI acceleration cutoff energy is sufficiently high to allow injection into diffusive shock acceleration. However, intensities of the modulated ACR proton spectra upstream become too small for reproducing observations when the TS (sub)ramp width approaches an ion inertial length (weak MRI acceleration). In summary, these self-consistent simulations indicate that low-energy PUIs pre-accelerated locally at a nearly perpendicular TS by the MRI mechanism might lead to the injection of a sufficient number of MRI-accelerated PUIs into diffusive shock acceleration for an injection
-
168-
Anomalous cosmic rays: Current andfuture theoretical developments
threshold o f - 5 0 keV to reproduce the observed upstream ACR proton spectra. This is possible despite significant mediation of the TS by MRI-accelerated PUIs as well ACRs formed by the diffusive shock acceleration of a fraction of the locally pre-accelerated PUIs. For the same injection energy recent work shows that not enough PUIs pre-accelerated in the upstream solar wind will be injected to explain ACR observations because the flux of preaccelerated PUIs fall strongly off as 1/r3 at large heliocentric distances due to adiabatic cooling [25, 26]. Given modeling to the contrary [27], and the uncertainty in the injection efficiency, these conclusions need further study.
4. ANOMALOUS COSMIC RAY T R A N S P O R T One of the big challenges in the area of ACR transport is to find a theory for the diffusion tensor that will enable modelers to reproduce the observed modulated ACR spectra. Whereas quasi-linear theory for the parallel diffusion coefficient, and classical scattering theory for the antisymmetric drift coefficient in the limit of weak scattering (~Jrg >> 1 where ,~ is the parallel mean free path and rg is the particle gyroradius) appear to be confirmed by test particles simulations in a prescribed background magnetic field with magnetostatic fluctuations, this is not the case for the perpendicular diffusion coefficient [36, 37, 38]. Further progress in the understanding of ACR modulation in the solar wind depends on a better grasp of perpendicular diffusion since all ACR modulation models show that observations can only be reproduced if there is significant levels of perpendicular diffusion at large heliocentric distances which are especially large in the heliolatitudinal direction [38, 39]. Simulations show that there is more perpendicular diffusion than expected from cyclotron resonant wave-particle interactions alone which is attributed to the strong role of field line meandering. Unfortunately, there is no theory for cross-field diffusion which includes largescale field line meandering effects that reproduces the results from simulations except for the case of short time scales (t << v where v is the time scale for macroscopic diffusion to occur) when the particles are tied to and freely streaming along the wandering field lines [38]. In this case the ACRs themselves do not diffuse relative to the wandering field lines but appear to undergo perpendicular diffusion relative to the background field due to the diffusive motion of the large-scale field lines to which they are tied. However, ACR modulation spectra are observed and studied on time scales t >> ~-because ACRs have small anisotropies due to extensive pitch angle scattering that turns them around. On this time scale one expects perpendicular diffusion to be due to a double diffusive effect in that the particles are diffusing along and across the large-scale wandering field lines which are also executing a diffusive motion. Different theories exist for these conditions, one of which is called compound diffusion whereby the particles diffuse along the large-scale meandering field lines but are still tied to the field fluctuations. In this case perpendicular diffusion due to large-scale field line random walk is not sustainable and this is an example of "subdiffusive" behavior. This is not supported by simulations that suggest a diffusive behavior instead [36]. A more sophisticated theoretical description exists that gives rise to sustained perpendicular diffusion called anomalous diffusion [41 ]. In this approach, where particles are allowed to cross field lines so that it includes the contribution from the separation of neighboring large-scale random walking field lines, strong perpendicular diffusion is predicted. Roughly speaking, this theory seem to reproduce the results from the particle simulations qualitatively, but the
-
169-
J.A. le Roux
level of the diffusion in the theory is higher than in the simulations. Such a disagreement is not surprising because the simulations only go down to energies of ~1 MeV while the theory is strictly speaking valid for energies below -1 keV [42]. This theory is promising, however, and needs to be extended. A big stumbling block for finding a diffusion tensor theory that reproduces ACR observations is to describe its spatial dependence correctly throughout the heliosphere given that the heliosphere is observed with only a few spacecraft. A significant advance in this regard is the development of a model for the transport of MHD turbulence in the solar wind that reproduces the observed radial dependence of the energy in the magnetic field fluctuations well in the equatorial regions [30], an important parameter in the diffusion tensor. A further extension of this model should allow one to achieve a good description of the spatial dependence of the turbulence correlation length, which also features in diffusion tensor theory. This MHD turbulence model is now routinely used in cosmic ray modulation models which test diffusion tensor theories against observations [42, 43].
5. THE FUTURE Ever since the realization that ACRs are interstellar PUIs that are accelerated somewhere in the outer heliosphere, attempts have been made to figure out the details of the formation of the ACR component from these PUIs. It has come to be accepted that ACRs are finally formed through standard diffusive shock acceleration at the most powerful accelerator in the solar wind, the TS. The expectation is that PUIs need to be pre-accelerated to enable injection into the standard diffusive shock acceleration mechanism because the ideal site for injection is thought to be where the TS is nearly perpendicular. Since PUI pre-acceleration in the upstream solar wind offers the possibility of direct injection at the TS, efforts are continuing to improve understanding of upstream pre-acceleration. In this regard, trying to clarify the role of stochastic acceleration mechanisms by MHD solar wind turbulence relative to interplanetary shock acceleration mechanisms is of importance. Note also that it is not known how 2-D MHD turbulence will contribute to PUI pre-acceleration despite the fact that it appears to be the dominant magnetic field fluctuation component in the solar wind. However, pre-acceleration of low-energy PUIs can also occur locally at the TS itself due to mechanisms such as MRI or stochastic magnetic mirroring acceleration. It remains to be determined whether low-energy PUIs pre-accelerated locally at the TS or PUIs pre-accelerated in the upstream solar wind are injected the most efficiently into standard diffusive shock acceleration at the TS. As the details of injection become better understood in the future, the prediction by current simulations that the TS will be modified by MRI-accelerated PUIs and ACRs will be reevaluated. Regarding ACR transport and modulation in the solar wind, there will be a continuing emphasis on improving understanding of the role of large-scale field fluctuations (such as found in 2-D MHD solar wind turbulence) in perpendicular diffusion, a crucial modulation parameter. In addition, PUI and ACR transport models need to be extended to deal with a TS that is not necessarily spherical which has implications for ACR modulation levels, and a TS that does not have the same compression ratio everywhere which affects the ACR source spectrum. Further complications are that PUIs themselves have upwind-downwind symmetries and are born with different speeds at high heliolatitudes compared to low latitudes which are all factors in determining the injection efficiency of PUIs at the TS. Such issues
-
170-
Anomalous cosmic rays: Current andfuture theoretical developments
will in the end be best addressed when full 3-D models of the solar wind interaction with the local interstellar medium are integrated with 3-D ACR transport models to allow selfconsistent calculations. The ultimate success of these efforts, however, hinges on the continuing operation of the Voyager spacecraft in the outer heliosphere to pave the way for the Interstellar Probe in order to penetrate through the TS and beyond.
ACKNOWLEDGEMENTS
I thank the organizers for the invitation to present my views on pickup ions and anomalous cosmic rays and acknowledge financial support from NASA grant NAG5-6969, and NSF grants ATM-0072810 and ATM-0072810.
REFERENCES
1. Cummings, A. C., & Stone, E. C. 1996, Space Sci. Rev., 78, 117. 2. Jokipii, J. R. 1992, Astrophys. J., 393, L41. 3. Jones, F. C. 1993, Astrophys. J., 361, 162. 4. Isenberg, P. A. 1987, J. Geophys. Res., 92, 1067. 5. Fisk., L. A. 1976, J. Geophys. Res., 81,4641. 6. Webb, G. M., Axford, W. I., & Terasawa, T. 1983, Astrophys. J., 270, 537. 7. Zank, G. P., Pauls, H. L., Cairns, I. H., & Webb, G. M. 1996, J. Geophys. Res., 101,457. 8. Lee, M. A., Shapiro, V. D., & Sagdeev, R. Z. 1996 J. Geophys. Res., 101, 4777. 9. Gloeckler, G., AIP Conference Proc., in press. 10. Fichtner, H., le Roux, J. A., Mall, U., & Rucinski, D. 1996, Astron. Astrophys., 314, 650 11. Bogdan, T. J., Lee, M. A., & Schneider, P. 1991, J. Geophys. Res., 96, 161 12. le Roux, J. A., & Ptuskin, V. S. 1998, J. Geophys. Res., 103, 4799. 13. Chalov, S., Fahr, H. J., Izmodenov, V. 1997, Astron. Astrophys., 320, 659. 14. B ieber, J. W., Wanner, W., & Matthaeus, W. H. 1996, J. Geophys. Res., 101, 2511 15. Zank, G. P., & Matthaeus, W. H. 1993, Phys. Fluids, A5, 257 16. Shebalin, J. V., Matthaeus, W. H., & Montgomery, D. 1983, J. Plasma Phys., 29, 525. 17. Ambrosiano, J., et al. 1988, J. Geophys. Res., 93, 14383. 18. Chalov, S. V., & Fahr, H., J. 2000, Astron. Astrophys., 360, 381. 19. Rice, W. K. M., Zank G. P., & le Roux, J. A. 2000, Geophys. Res. Lett., submitted. 20. Newbury, J. A., Russell, C. T., & Gedalin, M. 1998, J. Geophys. Res., 103, 29581. 21. Zilbersher, D., & Gedalin, M. 1997, Planet. Space Sci., 45, 693. 22. Rice, W. K. M., & Zank, G. P. 2000, J. Geophys. Res., 105, 5157. 23. Rice, W. K. M., et al. 2000, J. Geophys. Res. Lett., 27, 509. 24. Lazarus, A., Richardson, J. D., & Decker, R. B. 1999, Space Sci. Rev., 89, 53. 25. Decker, R. B., et al. 2000 this volume. 26. Rice, W. K. M., et al.. 2000, Adv. Space Res., in press. 27. Giacalone, J., et al. 1997, Astrophys. J., 486, 471. 28. Giacalone, J., & Jokipii, J. R. 1998, Space Sci. Rev., 83,259. 29. Jokipii, J. R. 1982, Astrophys. J., 255,716. 30. Zank, G. P., Matthaeus, W. H., & Smith, C. W. 1996, J. Geophys. Res., 101, 17093. 31. Zank, G. P., Rice, W. K. M., le Roux, J. A., & Matthaeus, W. H., this volume.
-171 -
J.A. le Roux
32. le Roux, J. A., & Fichtner, H. 1997, J. Geophys. Res., 102, 17365. 33. le Roux, J. A., et al. 2000, Geophys. Res. Lett., in press. 34. le Roux, J. A., et al. 2000, Geophys. Res. Lett., 27, 2873. 35. Edmiston, J. P., & Kennel, C. F. 1984, J. Plasma Phys., 32, 429. 36. Giacalone, J., & Jokipii, J. R. 1999, Astrophys. J., 520, 204. 37. Giacalone, J., et al. 1999, Proc. 26 th Int. Cosmic Ray Conf. (Utah), 7, 37. 38. Mace, R. L., Matthaeus, W. H., and Bieber, J. W. 2000, Astrophys. J., 538, 192. 39. Potgieter, M. S. 2000, J. Geophys. Res., 105, 18295. 40. Jokipii, J. R. 1973, Astrophysical J., 182, 585. 41. Chuvilgin, L. G., & Ptuskin, V. S. 1993, Astron. Astrophys., 279, 278. 42. le Roux, J. A., Zank, G. P., & Ptuskin, V. S. 1999, J. Geophys. Res., 104, 24845. 43. Burger, R. A., & Hattingh, M. 1998, Astrophys. J., 505,244.
-
172-
ANOMALOUS
COSMIC RAY OBSERVATIONS IN THE INNER AND
OUTER HELIOSPHERE B. Heber abc, and A. Cummings d ~Max-Planck-Institut ftir Aeronomie, 37191 Katlenburg-Lindau, Germany bIEAP - Universit~it Kiel, 24098 Kiel, Germany Cnow at FB. Physik, Universit~it Osnabrfick, 49069 Osnabrtick, Germany dCalifornia Institute of Technology, Pasadena, CA, USA Our knowledge on how galactic and anomalous cosmic rays are modulated in the heliosphere has been dramatically enlarged due to measurements of several missions launched in the past years. Among them, Ulysses explored the inner heliosphere near the polar regions during the last solar minimum period and the Solar Anomalous and Magnetospheric Particle Explorer (SAMPEX) uses the Earth's magnetic field to provide insight on the ionic charge composition. A new generation of energetic particle instruments on board the Advanced Composition Explorer (ACE), launched in August, 1997 to the Lagrangian L1 point, provides the possibility to investigate solar modulation with isotopic resolution for elements from Z ~ 2 to Z ~ 30. In the outer heliosphere the two Voyagers are heading towards the termination shock and are still operating, while almost no scientific data is being collected from the Pioneer spacecraft due to the depletion of the power source. We will summarize properties of anomalous cosmic rays observed at different locations in the heliosphere and discuss their elemental and isotopic composition. INTRODUCTION The interaction of the supersonic solar wind with the local interstellar medium defines the space called the heliosphere (see Fig. 1). At the region where the solar wind drops to subsonic speeds, the heliospheric termination shock is formed. Galactic Cosmic Rays (GCRs) entering our heliosphere encounter an outward-flowing solar wind carrying a turbulent magnetic field. These particles are modulated within the solar cycle by various processes, as discussed in Potgieter (1998). Anomalous Cosmic Rays (ACRs) were discovered in the 1970's when GarciaMunoz et al. (1973) found an unexpected shape of the helium spectrum below ~ 100 MeV/n (see Fig. 1 in Moraal, 2001). Fisk et al. (1974) postulated the following mechanism as a source for these particles. The principal ideas were further developed by Vasyliunas and Siscoe (1976), discussed in detail by Moraal (2001) and Le Roux (2000), and are summarized in Fig. 1: Neutral interstellar particles enter the heliosphere and are ionized by the interaction with the solar wind and/or solar radiation and are picked up by the solar wind. Pickup ions are convected out to the heliospheric termination shock and are accelerated to cosmic ray energies. The process
-
173-
B. Heber and A. Cummings
Figure 1. The interaction of the solar wind with the local interstellar medium defines the heliosphere (upper fight panel). Pickup ions are generated from interstellar neutrals by ionization. These pickup ions are accelerated at the heliospheric termination shock to become ACRs (for details see text).
of shock acceleration has been theoretical described by Pesses et al. (1981) and Lee and Fisk (1982). No in situ measurements of ACRs at the heliospheric termination shock are available today. However, interstellar neutral helium and the hydrogen and helium pickup ions were measured with instruments onboard the AMPTE (Moebius et al., 1985) and the Ulysses spacecraft (Witte et al., 1993; Gloeckler et al., 1993). The Voyager 1 and 2 spacecraft, launched in 1977, are at ~ 78 AU and ~ 62 AU (2000 day 207), respectively. Voyager 1 is expected to reach the heliospheric termination shock shock within the next decade. ACRs, like GCRs, are modulated by the turbulent heliospheric magnetic field. The ACR component is different from GCRs in a number of respects. (1) While ACRs should be mostly singly charged, GCRs are fully stripped atoms. (2) ACRs should reflect the elemental and isotopic composition of pickup ions and therefore of the local interstellar neutrals, while the GCR composition is modified during their propagation within the galaxy. (3) The maximum energy of ACRs should be restricted to
-
174-
Anomalous cosmic ray observations in the inner and outer heliosphere
Figure 2. Trajectories of the Ulysses, the two Voyager spacecraft, and Pioneer 10 (Fichtner, 2001). The inset shows the heliographic latitude as function of radial distance from beginning of 1993 to the end of 1997. Solid circles mark the start of each year. The histogram shows the evolution of the Maximum Latitudinal Extent of the heliospheric current sheet c~ (from Heber and Burger, 1999).
several hundred MeV, whereas GCRs are accelerated to much higher energies by presumably much larger shocks. With the instrumentation on board the 1 AU missions it is possible to determine these fundamental properties of ACRs. OBSERVATIONS
IN THE HELIOSPHERE
A unique network for studying the transport of GCRs and ACRs over a vast region of the
heliosphere is given by the combination of Ulysses, the Voyager, Pioneer, and 1 AU spacecraft, as shown in Fig. 2. The range of heliocentric distances and latitudes covered by Ulysses (U), Voyager (V1 and V2), and Pioneer 10 (P10) for the time period from mid 1993 to mid 1997 is displayed in the inset of Fig. 2. In 1994/1995 Ulysses performed a whole latitude scan of 160 ~ within an 11 month period. During that time the radial distance decreased from 2.3 to 1.3 AU close to the equator and again increased to 2.0 AU at 80~ Since then, Ulysses took 29 months to reach the heliographic equator at a distance of 5.3 AU. At the same time V1 at ~ 60 AU and V2 at ~ 40 AU are moving towards the nose of the heliosphere whereas P10 at ~ 60 AU is headed down the tail region. The maximum latitudinal extent of the heliospheric current sheet c~ into the southern and northern hemispheres is shown for time periods when Ulysses is in the southern or northern hemisphere, respectively, c~ can be also used as a proxy for solar activity, indicating that solar activity decreases from a moderate level in 1994 to low levels in 1996/1997. Observations from Earth at solar minimum are from (1) the Solar Anomalous and Magnetospheric Particle Explorer (SAMPEX), which uses the Earth's magnetic field to provide new
- 175-
B. Heber and A. Cummings
Figure 3. Relative intensity profiles as function of SAMPEX invariant latitude for 1216 MeV/n and 20-28 MeV/n O. Curves show calculated profiles for O with charge one, two, and their weighted sum (from Leske et al., 2000; see also Klecker et al., 1995).
Figure 4. Energy spectra of singly charged N, O, and Ne (solid lines) obtained by using the Earth's magnetic field. The dashed lines show the corresponding total energy spectra (from Leske et al., 2000; see also Mewaldt et al., 1996)
insights into the ionic charge composition (Baker et al., 1993), and (2) from a new generation of energetic particle instruments on board the Advanced Composition Explorer (ACE), which allows us to investigate solar modulation with isotopic resolution for elements from Z ~ 2 to Z ~ 30 (Stone et al., 1989). MODULATION AT SOLAR MINIMUM After a quarter of a century since the discovery of ACR helium by Garcia-Munoz et al. (1973), many ACR species are known. The main representatives are H, He, N, O, Ne, and Ar. Other elements like C, Na, S, Si, and Mg have been detected too (Cummings et al., 1999; Reames, 1999). The latter might not be of interstellar but rather of solar origin as discussed by Gloeckler et al. (2000). Crucial for the validity of the paradigm as sketched above is the observational proof that ACRs are singly charged and reflect the elemental as well as the isotopic composition of Pickup ions. Charge state composition of ACRs
Above energies of a few MeV the charge state composition can only be determined by using the Earth's magnetic field as a filter, which requires a spacecraft in a high inclination or polar Earth orbit such as the Solar, Anomalous, and Magnetospheric Particle Explorer (SAMPEX). The geomagnetic filter effect is illustrated in Fig. 3, which shows the normalized 12-16 MeV/n and 20-28 MeV/n oxygen flux as function of SAMPEX invariant latitude. The latter is the magnetic latitude 0 at which a field line on which the observations are made intersects Earth's surface and is related to the magnetic L-shell by cos 2 0 = 1 / L . A s displayed in Fig. 3, singly
- 176-
Anomalous cosmic ray observations in the inner and outer heBosphere
charged oxygen, which has a higher rigidity than multiply charged oxygen, can penetrate towards lower invariant latitudes. Therefore GCR oxygen is excluded for invariant latitudes below 60 ~ and the flux of pure ACR oxygen, as shown in Fig. 4, can be determined by taking into account only the measurements below this geomagnetic threshold. The measured invariant latitude distribution of 12-16 MeV/n oxygen, as displayed in the upper panel of Fig. 3, is in good agreement with the corresponding calculation for singly charged oxygen. At higher ACR energies the distribution differs significantly from the expected one for O + (Klecker et al., 1995). As an example, the 20-28 MeV/n oxygen distribution is displayed in the lower panel of Fig. 3. The dip in the profile at low invariant latitudes is in good agreement with the expected latitude for the O ++ cutoff. The distribution can be fitted if it is assumed that O + and O ++ contribute approximately equally. Applying this method to different energy bands, Klecker et al. (1998) found that the ratio of O + to all charge states is decreasing with increasing energy and by analyzing N and Ne, Klecker et al. (1997) concluded that the total energy better organizes the charge state composition. For total energies above ~350 MeV, more than half of the ACRs are multiply charged. The occurrence of multiply-charged ACRs is due to electron stripping and can be explained within the current paradigm if the typical time to accelerate ACRs to 10 MeV/nuc is of the order of one year (Mewaldt et al., 1996). This value is in agreement with theoretical considerations (Potgieter and Moraal, 1988). Since the acceleration time for diffusive shock acceleration (Jokipii, 1996) is increasing with increasing ACR energy, the ratio of singly charged to all ACRs is expected to decrease with increasing energy.
Elemental composition and the injection efficiency Twenty-two years after the discovery of the ACR component, the flux enhancements for different elements are observed with much improved statistics and resolution. The relative abundances of ACRs are derived by fitting the observations to a modulaton model and comparing intensities at a fixed energy/nucleon at the position of the HTS. To infer the abundances of the neutral interstellar gas that is the ultimate ACR source requires understanding the fractionating processes of ionization and injection/acceleration at the HTS (see Fig. 1). Generally, the interstellar pickup ions, which have now been observed for H, He, N, O, and Ne, are more closely connected to the interstellar neutrals, not having been subjected to the injection and acceleration process. For those five elements, Gloeckler et al. (2000a) have produced a table of densities of neutrals at the HTS and in the LISM based on the pickup ions observations with Ulysses. The imortant question of the injection efficiency of the acceleration process at the HTS can be investigated by comparing the ACR observations with the pickup ions observations (see Fig. 1). From shock acceleration theory, the energy spectra of the ACRs at the HTS at low energies should be power-laws in energy/nucleon. The ACR observations have been made some distance inside the HTS, which results in modulated spectral shapes that do not resemble power laws. However, using a suitable theory of solar modulation, the energy spectra at the shock can be estimated. These same ACR spectra at the shock can be calculated from the pickup ion observations using the theory of Lee (1983) with the injection efficiency parameter adjusted to give the best fit to the spectra deduced from the ACR observations. Cummings and Stone (1996) performed such a comparison and derived the injection efficiencies shown in Fig. 5. The injection efficiency increases with mass which is opposite to what some theories predict but which can be explained as a consequence of competing mass-dependent effects in a two-stage acceleration process (see Zank et al. (2000) for a review of the injection problem).
-177-
B. Heber and A. Cummings
Figure 6. Quiet time energy spectra from 8/97 to 3/98 for N, O, and Ne, from different instruments on the ACE spacecraft. Curves are the sum of a power law spectra (~ ( E / M ) -~ and an exponential spectra through the ACR isotopes, with the ACR relative abundance asFigure 5. Estimated injection efficiencies for sumed to be the same as those found in the soH+, He+, and O+ of pickup ions into the accel- lar system material (from Leske et al., 2000). eration process at the HTS. Derived from Table IV of Cummings and Stone (1996).
Isotopic composition
The new ACE measurements allow us to investigate the isotopic composition with higher precision than previously possible. Fig. 6, from Leske et al. (2000), shows the isotopic spectra for ACRs ( ~< 30 MeV/n) and GCRs. Like the dominant isotopes 160 and 2~ the rare isotopes ~80 and 22Ne also show a clear low-energy ACR enhancement. In contrast to the elemental abundances, the isotopic abundance of ACRs should not be significantly influenced by different injection efficiencies at the heliospheric termination shock and modulation in the heliosphere. It is therefore a much better tool to test if ACRs really reflect pickup ion, neutral particle, and/or solar system abundances. The solid lines in Fig. 6 are calculated by using solar system isotopic abundances. Obviously the oxygen isotopes and the 2~ and 22Ne are consistent with this assumption. The lack of a low energy enhancement in the 15N and 21Ne spectra suggests that the ~SN/laN and 21Ne/2~ ratios can not be more than a factor of 5 and ~10 larger than the standard solar system abundance. It is important to note that the isotopic composition of ACRs and GCRs are quite different. While for GCRs, e.g., 14N and 15N are nearly equally abundant, the ~4N to ~SN ratio increases rapidly with decreasing energy. Modulation of ACRs
Fig. 7 displays the intensities of (a) 155-368 MeV/n helium and (b) 8-27 MeV/n oxygen as measured at Earth in comparison with the intensities measured along the Ulysses (c) and Voyager 2 (d) trajectories over the last A>0 solar magnetic cycle. By comparing GCRs and ACRs (curves (a) and (b)) with each other it is obvious that ACRs are much more sensitive
- 178-
Anomalous cosmic ray observations in the inner and outer heliosphere
Figure 7. Quiet time fluxes of 1 AU 155365 MeV/n helium (a) and 8-27 MeV/n oxygen (b) compared to the 8-27 MeV/n oxygen fluxes along the Ulysses (c) trajectory and 7-17 MeV/n oxygen at the position of Voyager 2 (d) in the outer heliosphere.
Figure 8. a) Fifty-two day averaged energy spectra of H from the Cosmic Ray experiment on Voyager 1. The energy spectrum for the last 52-day period of each year is shown, b) Same as a) except for He. From Stone et al., 1999.
to solar modulation than GCRs. The fact that ACRs immediately recovered in 1992/1993 and GCRs take a longer time period to reach the maximum fluxes has been interpreted by McDonald et al. (2000) as an indication for modulation in the heliosheath. As expected, the fluxes are much larger in the outer heliosphere at the position of Voyager 2 (d) than at Earth in the inner heliosphere (b). Since Voyager 2 is moving outwards, the intensities at Voyager are influenced by temporal as well as the spatial variation. The net result for ~ 10 MeV/nuc ACR oxygen, well above the peak energy of the ACR oxygen spectrum, is an intensity profile that is rather flat from 1993 to 1997 but which gradually rises afterwards until mid-1999. Fig. 8 shows the progression of 52-day averaged spectra from the end of 1994 to the end of 1998 for ACR and GCR H and He. At energies at and below the peak of the ACR energy spectrum, a continual increase in intensity from 1994 through 1998 is observed, whereas above ~50 MeV/n the intensity is nearly constant, suggesting that at these energies the intensity is essentially the same as at the termination shock. The three spacecraft in the outer heliosphere combined with 1 AU baseline and Ulysses, as shown in Fig. 2, can be be combined to yield estimates of the radial and latitudinal gradients of ACRs and GCRs (Cummings and Stone, 1998, McDonald et al., 1997). These authors found that the latitudinal gradient in the inner heliosphere is in good agreement with the one in the outer heliosphere. The radial gradient appears to have a stronger radial dependence in an A<0 solar cycle (as in the 1980's) than in an A>0 solar cycle (as in the 1990's). Ulysses, exploring the "inner" heliosphere, show a remarkable time profile compared to the measurements at Earth and in the outer heliosphere. The heliographic latitude and the spacecraft distance to the Sun are given at the top of Fig. 7. In February 1995 the spacecraft crossed the heliographic equator at a distance of 1.3 AU. At that time Ulysses observations are in good agreement with measurements at Earth. The intensities are then increasing with Ulysses latitude and show a maximum at polar latitudes. Such a behavior is expected from our current understanding of spatial modulation in the heliosphere, and has been intensively investigated
-
179-
B. Heber and A. Cummings
by the Potchefstroom - Ulysses COSPIN/KET collaboration for GCR protons (e.g., Heber and Potgieter, 2000; and Burger et al., 2000) and electrons (e.g., Potgieter et al., 1999). While for GCR the latitudinal gradient was much smaller than anticipated, Fig. 7 clearly demonstrates that these gradients are not negligible for ACR oxygen, leading to the fact that the intensity measured at 2 AU and 80 ~ heliographic latitude corresponds to intensities beyond 40 AU. 1. S u m m a r y
During the minimum phase of the last solar cycle 22 tremendous progress has been made in both measuring and understanding ACRs. The elemental spectra have been determined more precisely than ever before thanks to the advanced instrumentation on Wind and Advanced Composition Explorer at 1 AU. As a result, ACR-like enhancements have been discovered for new species. Measurements of the isotopic composition reflect the corresponding composition of pickup ions and interstellar neutrals, and will also provide insights into the generation as well as acceleration processes of galactic cosmic rays. The current paradigma that ACRs are mainly singly charged was confirmed by measurements of the SAMPEX satellite which uses the geomagnetic field as a filter. Using the geomagnetic filter approach, higher charged ACRs have also been discovered. The charge composition of ACRs can be used to put constraints on the acceleration time scales at the heliospheric termination shock. Another advantage of this approach might be the possibility to track the ACR intensities at 1 AU through the current solar maximum into the next solar minimum around 2003. Voyager 1, at ~78 AU, has not yet reached the heliospheric termination shock. The heliospheric network of space probes available at the 1990's solar minimum showed that the latitudinal gradients in the inner heliosphere are in good agreement with the one in the outer heliosphere. Since these gradients are large compared to the radial gradients, Ulysses at 2 AU 80 ~ observed approximately the same intensity as Voyager 2 beyond 40 AU. ACKNOWLEDGMENT The authors thank E McDonald, R. A. Leske, and R. Marsden for providing IMP helium, 1 AU oxygen and Ulysses oxygen data as well as for helpful discussions. We also thank E. C. Stone for helpful suggestions. This work was partially supported by NASA under Contract NAS7-918. REFERENCES D.N. Baker, G. M. Mason, O. Figueroa, G. Colon, J. G. Watzin, and R. M. Aleman. An overview of the solar, anomalous, and magnetospheric particle explorer (SAMPEX) mission. IEEE transactions on Geos. and remote Sensing, 31:531-541, 1993. 2. R . A . Burger, M. S. Potgieter, and B. Heber. Rigidity dependent of cosmic-ray proton latitudinal gradients measured by the Ulysses spacecraft: Implications for the diffusion tensor. J. Geophys. Res., 2000. in press. 3. A.C. Cummings and E. C. Stone. Composition of anomalous cosmic rays and implications for the heliosphere. Space Science Reviews, 78:117-128, October 1996. 4. A.C. Cummings, E. C. Stone, and C. D. Steenberg. Composition of anomalous cosmic rays 1.
- 180-
Anomalous cosmic ray observations in the inner and outer heliosphere and other ions from voyager observations. In 26th International Cosmic Ray Conference, volume 7, page 531, 1999. 5. H. Fichtner. Anomalous cosmic rays" Messengers from the outer heliosphere. Space Sci. Rev., 2001. 6. L.A. Fisk, B. Koslovsky, and R. Ramaty. An interpretation of the observed oxygen and nitrogen enhancements in low-energy cosmic rays. Astrophys. J. Lett., 190:L35, 1974. 7. M. Garcia-Munoz, G. M. Mason, and J. A. Simpson. A new test for solar modulation theory: the 1972 May-July low-energy galactic cosmic ray proton and helium spectra. Astrophys. J. Lett., 182:L81, 1973. 8. G. Gloeckler, L. A. Fisk, T. H. Zurbuchen, and N. A. Schwadron. Sources, injection, and acceleration of heliospheric ion populations. In AlP Conf. Proc. 528." ACE 2000 Symposium, volume 528, pages 221 +, 2000. 9. G. Gloeckler, J. Geiss, H. Balsiger, L. A. Fisk, A. B. Galvin, F. M. Ipavich, K. W. Ogilvie, R. von Steiger, and B. Wilken. Detection of interstellar pick-up hydrogen in the solar system. Science, 261:70-73, July 1993. 10. G. Gloeckler, J. Geiss, and L. A. Fisk. Heliospheric and Interstellar Phenomena Revealed from Observations of Pickup Ions. Springer-Praxis, 2000. in press. 11. B. Heber and R. A. Burger. Modulation of galactic cosmic rays at solar minimum. Space Sci. Rev., 89" 125-138, 1999. 12. B. Heber and M. S. Potgieter. Galactic cosmic ray observations at different heliospheric latitudes. Adv. Space Res., 26(5):839-852, 2000. 13. J. R. Jokipii. Theory of multiply charged anomalous cosmic rays. Astrophys. J. Lett., 466:L47-+, July 1996. 14. B. Klecker, M. C. McNab, J. B. Blake, D. C. Hamilton, D. Hovestadt, H. Kaestle, M. D. Looper, G. M. Mason, J. E. Mazur, and M. Scholer. Charge state of anomalous cosmic-ray nitrogen, oxygen, and neon: Sampex observations. Astrophys. J. Lett., 442:L69-L72, April 1995. 15. B. Klecker, R.A. Mewaldt, J.~TV. Bieber, A.(~. Cummings, L. Drury, J. Giacalone, J.l~. Jokipii, F.(~. Jones, M.t~. Krainev, M.A. Lee, J.A. Le Roux, R.G. Marsden, F.13. McDonald, R.13. McKibben, C.I). Steenberg, M.G. Bating, D.(~. Ellison, L.J. Lanzerotti, R.A. Leske, J.l~. Mazur, H. Moraal, M. Oetliker, V.S. Ptuskin, R.~,. Selesnick, and K.J. Trattner. Anomalous cosmic rays. Space Science Reviews, 83:259-308, 1998. 16. B. Klecker, M. Oetliker, J.B. Blake, D. Hovestadt, G.M. Mason, J.E. Mazur, and M.C. M c N a b . . In Proc. 25. Inter. Cosmic Ray Conf., volume 7, pages 333-336, 1997. 17. M. A. Lee and L. A. Fisk. Shock acceleration of energetic particles in the heliosphere. Space Sci. Rev., 32:205-228, 1982. 18. R. A. Leske. Anomalous cosmic ray composition from ace. In AlP Conf. Proc. 516." 26th International Cosmic Ray Conference, ICRC XXVI, volume 26, pages 274+, 2000. 19. R. A. Leske, R. A. Mewaldt, E.R. Christian, C/,M.S. Cohen, A. C. Cummings, EL. Slocum, E.C. Stone, T.T. von Rosenvinge, and M.E. Wiedenbeck. Observations of anomalous cosmic ray at 1 au. In AlP Conf. Proc. 528." ACE 2000 Symposium, volume 528, pages 293+, 2000. 20. F.B. McDonald, E Ferrando, B. Heber, H. Kunow, R. McGuire, R. Mtiller-Mellin, C. Paizis, A. Raviart, and G. Wibberenz. A comparative study of cosmic ray radial and latitudinal gradients in the inner and outer heliosphere. J. Geophys. Res., 102:4643-4652, March
-181
-
B. Heber and A. Cummings 1997. 21. E B. McDonald, B. Heikkila, N. Lal, and E. C. Stone. The relative recovery of galactic and anomalous cosmic rays in the distant heliosphere: Evidence for modulation in the heliosheath. J. Geophys. Res., 105:1-8, January 2000. 22. R.A. Mewaldt, J.R. Cummings, R.A. Leske, R.S. Selesnick, E.C. Stone, and T.T. von Rosenvinge. A study of the composition and energy spectra of anomalous cosmic rays using the geomagnetic field. Geophys. Res. Let., 23:617-620, 1996. 23. E. Moebius, D. Hovestadt, B. Klecker, M. Scholer, and G. Gloeckler. Direct observation of he(+) pick-up ions of interstellar origin in the solar wind. Nature, 318:426-429, 1985. 24. H. Moraal. The discovery and early development of the field of anomalous cosmic rays. Space Sci. Rev., 2001. 25. M. E. Pesses, J. R. Jokipii, and D. Eichler. Cosmic ray drift shock-wave acceleration, and the anomalous component of cosmic rays. Astrophys. J. Lett., 246:L85-L88, 1981. 26. M. S. Potgieter. The modulation of galactic cosmic rays in the heliosphere: Theory and models. Space Science Reviews, 83:147-158, January 1998. 27. M. S. Potgieter, S. E. S. Ferreira, B. Heber, E Ferrando, and A. Raviart. Implications of the heliospheric modulation of cosmic ray electrons observed by ulysses. Advances in Space Research, 23:467-470, 1999. 28. M. S. Potgieter and H. Moraal. Acceleration of cosmic rays in the solar wind termination shock, i - a steady state technique in a spherically symmetric model. Ap. J., 330:445-455, July 1988. 29. D. V. Reames. Quiet-time spectra and abundances of energetic particles during the 1996 solar minimum. Ap. J., 518:473+, 1999. 30. J. A. Le Roux. Anomalous cosmic rays: current and future theoretical developments. Space Sci. Rev., 2001. 31. E. C. Stone, A. C. Cummings, D. C. Hamilton, M. E. Hill, and S.M. Krimigis. Voyager Observations of Anomalous and Galactic Cosmic Rays During 1998. In Proc. 26th International Cosmic Ray Conference, Salt Lake City, Utah, USA, August 17-25, 1999, volume 7, page 551, 1999. 32. E.C. Stone, L.F. Burlaga, A.C. Cummings, W.C. Feldman, W.E. Frain, J. Geiss, G. Gloeckler, R. Gold, D. Hovestadt, S.M. Krimigis, G.M. Mason, D. McComas, R.A. Mewaldt, J.A. Simpson, T.T. Rosenvinge, and M. Wiedenbeck. the Advanced Composition Explorer, volume 203, page 48. AIP Conference Proceedings, 1989. 33. V. M. Vasyliunas and G. L. Siscoe. On the flux and the energy spectrum of interstellar ions in the solar system. J. Geophys. Res., 81:1247-1252, March 1976. 34. M. Witte, H. Rosenbauer, M. Banaszkiewicz, and H. Fahr. The ulysses neutral gas experiment - determination of the velocity and temperature of the interstellar neutral helium. Advances in Space Research, 13:121-130, June 1993. 35. G. E Zank, W. K. M. Rice, J. A. le Roux, J. Y. Lu, and H. R. Mtiller. The Injection Problem for Anomalous Cosmic Rays. In R. A. Mewaldt, J. R. Jokipii, M. A. Lee, E. M6bius, and T. H. Zurbuchen, editors, AlP ConfProc.: Acceleration and Transport of Energetic Particles Observed in the Heliosphere: ACE 2000 Symposium, volume 528, pages 317324, 2000.
-
182-
Cosmic Rays as Messengers from Outside the Inner Heliosphere M.A. Lee a and H. Fichtner b aSpace Science Center, EOS, Morse Hall, University of New Hampshire, Durham, NH 03824, USA blnstitut fur Theoretische Physik, Ruhr-Universit~it, 44780 Bochum, Germany The two populations of cosmic rays in the outer heliosphere, galactic and anomalous, are introduced briefly. Their coupling to the solar wind in the process known as solar modulation is described and the transport equation used to describe that process is introduced. Finally new results from the Voyager spacecraft and Interstellar Probe are anticipated, and open questions and challenges are identified. 1. I N T R O D U C T I O N There are two populations of cosmic rays in the outer heliosphere: galactic cosmic rays and anomalous cosmic rays. Both interact strongly with the solar wind. The two populations are shown in Figure 1 which is a schematic diagram of the heliosphere including the Sun, a radially outflowing solar wind with embedded magnetic irregularities, the solar wind termination shock where the solar wind is deflected toward the heliospheric tail (solid circle), the heliopause separating solar and interstellar plasma (inner dashed line), and a possible bow shock where the interstellar plasma flow is decelerated to flow around the heliopause (outer dashed line). The galactic cosmic rays (GCR) are denoted by small dots mostly outside the heliopause (low-energy GCRs with energies _<300 MeV/nucleon) and large dots throughout space (highenergy GCRs with energies > 300 MeV/nucleon). Outside the heliopause the intensity of lowenergy GCRs should dominate that of higher-energy GCRs since the GCR differential intensity in the several-GeV energy range is proportional to E -2.7. The GCRs originate from outside the heliosphere at supernova shock waves throughout the Galaxy, and also possibly at stellar wind termination shocks, pulsars, or other more exotic objects. They form a sea of energetic particles throughout the Galaxy with a pressure comparable to that of the interstellar plasma and magnetic field. They are partially expelled from the heliosphere by the "sweeping" action of the solar wind in the process known as solar modulation. As shown in Figure 1, the low energy GCRs barely penetrate the heliosheath, that region of space between the termination shock and the heliopause, due to modulation effects. The high energy GCRs penetrate all the way to Earth's orbit and bombard the upper atmosphere, creating the "showers" of nuclear interactions and secondary particles which led to the discovery of cosmic rays and the existence of a modulating environment about the Sun now known as the heliosphere. The so-called anomalous cosmic rays (ACR) are denoted by small dots mostly within the heliopause in Figure 1. The ACRs should perhaps better be called heliospheric cosmic rays since they owe their existence to the heliosphere. The ACRs originate at the solar wind
-
183-
M.A. Lee and H. Fichmer
Figure 1. A schematic diagram of the heliosphere showing the solar wind termination shock (solid line), heliopause (inner dashed line), a possible interstellar bow shock (outer dashed line), anomalous cosmic rays (small dots extending out to the heliopause), high-energy galactic cosmic rays (large dots), and low-energy galactic cosmic rays (small dots outside the heliopause). termination shock in the process of diffusive shock acceleration. They are swept into the heliosheath by the subsonic solar wind flow and may leak across the heliospheric magnetic field at the heliopause into interstellar space. Due to their lower energies of <300 MeV/ion, in spite of their being mostly singly charged ions, they are strongly modulated but do reach Earth orbit during periods of minimum solar activity. Prior to acceleration they arrive in the heliosphere as interstellar neutral gas which flows into the inner heliosphere (shown in Figure 1 as dashed horizontal lines) and is ionized to form interstellar pickup ions, which are advected with the solar wind to the termination shock. At the termination shock the pickup ions, with their suprathermal halo distribution in velocity space, are favored in their injection into the process of shock acceleration. This paper serves as an Introduction to the Sessions on Galactic and Anomalous Cosmic Rays: Messengers from Outside the Heliosphere.
-
184-
Cosmic rays as messengers from outside the inner heliosphere 2. C O S M I C R A Y T R A N S P O R T IN T H E S O L A R W I N D
Both populations of cosmic rays are strongly coupled to the solar wind plasma by scattering on solar wind magnetic irregularities, and by curvature and gradient drift in the large-scale heliospheric magnetic field. Thus, the plasma structure of the heliosphere must be known to extrapolate from the cosmic ray energy spectrum observed in the heliosphere to the energy spectrum of high-energy GCRs in interstellar space or of ACRs at the termination shock. Conversely, since cosmic rays observed at Earth have already sampled much of the heliosphere, the GCRs and ACRs may be used as probes of heliospheric structure. The time variations in the GCR intensity provide information on the solar-cycle variation of the heliosphere and the propagation of shocks throughout the heliosphere. The spectral shape of ACRs in the outer heliosphere currently provides the best estimate of the distance to the termination shock. The interaction between cosmic rays and the heliosphere also provides insights into the interaction of cosmic rays with stellar winds or the plasmaspheres of more exotic objects not accessible for direct study. Fortunately the transport equation appropriate to the study of cosmic ray modulation and acceleration in the heliosphere is well established (Parker, 1965; Gleeson and Axford, 1967; Jokipii et al., 1977): ~f~t+ (V + V D). V f - V . (K. V f ) - 8 9 V . Vp ~ :
Q
(1)
where f (p, r, t) is the omnidirectional distribution function of a specific species of cosmic rays, p is momentum magnitude, V is the solar wind velocity, V D is the drift velocity due to the inhomogeneous average magnetic field, K is the symmetric part of the spatial diffusion tensor, and Q is a source term necessary to describe pickup ion injection at the termination shock to create the ACRs. Equation (1) requires that the particle speed v satisfy v >> IVI and that the cosmic ray anisotropy be small, conditions which are readily satisfied with the exception of very low energy ACRs at the termination shock. According to equation (1), f (p, r, t) is insensitive to structures smaller than the scattering mean free path, on the order of several particle gyroradii. The most controversial quantity in equation (1), because it is so difficult to specify, is K, which depends crucially on the behavior of the magnetic field and its fluctuations. Magnetic field-line "braiding" and "random walk" affect the components of K perpendicular to the average field but are only understood qualitatively. The drift velocity VD depends on cosmic ray charge and produces a circulation of cosmic rays in latitude through the heliosphere and along the termination shock. Generally the modulation of electrons and ions show the expected symmetry due to drift, but puzzles remain. The most helpful measurement in this regard would be a comparison of the modulation of electrons and positrons which differ only in the sign of their charge. Terms in equation (1) describing stochastic acceleration by solar wind turbulence and cosmic ray viscosity are small and have been neglected. Although equation (1) is linear in its simplest form in which the solar wind is not modified by cosmic rays, realistic geometries require numerical solutions, while analytical approximations such as the "force-field" solution provide guides for intuition. 3. C O S M I C R A Y S A N D T H E H E L I O S P H E R E : F U T U R E R E S E A R C H
In spite of having a trustworthy transport equation at our disposal, many puzzles remain concerning the transport of cosmic rays, particularly in the outer heliosphere. Voyager and Interstellar Probe should resolve many of these as they traverse previously unexplored regions of space. When the Voyager spacecraft encounters the termination shock in the next few years
- 185-
M.A. Lee and H. Fichmer
it may reveal how pickup ions are injected into the process of shock acceleration, and whether any solar wind ions are injected. Voyager should observe whether the pressure of ACRs modifies the plasma structure of the termination shock. It should also determine whether the nature of modulation is different in the heliosheath where V 9V -- 0 and VD may be reduced. The Interstellar Probe will reveal the nature of the heliopause: What is the nature of the magnetic field adjacent to the heliopause? Do the ACRs escape across the heliopause into the interstellar medium? How effectively do the low-energy GCRs penetrate the heliosheath? Does the pressure gradient of either species modify the flows adjacent to the heliopause? Can ACRs and GCRs be distinguished in this region? Beyond the heliopause is a new frontier: What is the energy spectrum and composition of low-energy GCRs? What is the role of these low-energy cosmic rays in the state of the local interstellar medium? Do escaping ACRs "pollute" this spectrum? What is the interstellar magnetic field strength and does its direction agree with that inferred from the cosmic ray anisotropy observed at -105 GeV? Can low-energy GCR nuclides be detected which reveal new information on the cosmic-ray life-cycle? A half century ago cosmic rays first revealed the existence of a huge volume of plasma surrounding, and controlled by, the Sun, later called the heliosphere. Now we are on a journey to the edges of the heliosphere to reveal for the first time the low-energy cosmic rays in interstellar space. The authors wish to thank the speakers and poster presenters in the cosmic ray session, and the Chair, I. Lerche, for an exciting session which stimulated much discussion. In particular we are grateful to R. Leske for agreeing to present an invited talk on short notice. M.A.L. is very grateful to the organizers for excellent organization of the Colloquium and an outstanding choice of venue in Potsdam. M.A.L.'s portion of this work was supported, in part, by NSF Grant ATM-0091527.
Acknowledgments.
REFERENCES 1. L.J. Gleeson and W.I. Axford, Cosmic rays in the interplanetary medium, Astrophys. J.,
149 (1967) Ll15. 2. J.R. Jokipii, E.H. Levy, and W.B. Hubbard, Effects of particle drift on cosmic ray transport, 1, General properties, applications to solar modulation, Astrophys. J., 213 (1977) 861. 3. E.N. Parker, The passage of energetic charged particles through interplanetary space, Planet. Space Sci., 13 (1965) 9.
- 186-
Modulation region of galactic cosmic rays in the heliosphere" Estimation of dimension, radial diffusion coefficient, intensity out of region L.I. Dorman 1,2 1israel Cosmic Ray Center and Emilio Segre' Observatory, affiliated to Tel Aviv University, Technion and Israel Space Agency; P.O. Box 2217, Qazrin 12900, ISRAEL 2 Cosmic Ray Department of IZMIRAN, Troitsk 142092, Moscow region, RUSSIA ABSTRACT On the basis of neutron monitor data as well as solar activity data for four solar cycles 19-22 the hysteresis phenomenon is investigated by taking into account convection-diffusion and drifts, and estimated the dimension of modulation region, radial diffusion coefficient, expected cosmic ray intensity in the interstellar medium and absolute level of galactic cosmic ray modulation observed on the Earth's orbit. It was estimated that the average time of particle propagation from the boundary of modulation region to the Earth's orbit is much smaller than characteristic time-lag in hysteresis phenomenon, what supports the using in this research of quasi-stationary model of high-energy particle modulation in the Heliosphere. INTRODUCTION A short historical introduction to the research of the hysteresis phenomenon between longterm variations of cosmic ray (CR) intensity observed at Earth with solar activity (SA) is given in Dorman et al. (2001). Analysis made by Dorman (2001) leads to the conclusion that observed long-term CR modulation is caused by two processes: a convection-diffusion mechanism (see discussion and references in Dorman et al., 2001) that does not depend on the sign of the solar magnetic field (SMF), and a drift mechanism (e.g. Burger and Potgieter, 1999) what gives opposite effects with the changing sign of the SMF. In the present paper we try to determine the relative role of convection-diffusion and drifts for neutron monitor (NM) data in solar cycles 19-22. We will determine radial diffusion coefficient and transport path in radial direction, expected CR intensity in the interstellar medium and absolute level of galactic CR modulation. COSMIC RAY L O N G - T E R M VARIATION CAUSED BY C O N V E C T I O N - D I F F U S I O N Because the basic quasi-stationary model of CR-SA hysteresis phenomenon was described in details in Dorman et al. (2001), we give here only the final equations what will be used in our research. According to this model the expected CR long-term modulation at the Earth's orbit is:
ln(n(R,Xo,rE,t)exp)= A-Bx F(t,Xo,W(t-XIXXE), where
-
187-
(1)
L.I. Dorman 1 2(1 W(t_X)/Wmax) X -+( X~ )= ?(W(t-X)/Wmax) 3 3 dX, (2) F t, Xo,W(t- X I X E XE X = r / U , X E = I A U / u , X o=r o/u (X o is in units of av. month =(365.25/12) days
=2.628x106 s),
n(R, Xo,rE,t)exp is expected galactic CR density on the Earth's orbit in
dependence of the value of parameter
X o . Here r is the distance from the Sun, ro is the
expected radius of modulation region, R is the effective rigidity of detected CR particles; W is the monthly sunspot number (or some other parameter of SA), and Wmax is sunspot number in the maximum of SA. Coefficient A determines the CR intensity out of the modulation region, and coefficient B determines radial diffusion coefficient in period of maximum SA. COSMIC RAY L O N G - T E R M VARIATION CAUSED BY DRIFTS According to the main results of the drift mechanism approach to the CR long-term variation (e.g. Burger and Potgieter, 1999), we assume that the drifts are determined mainly by tilt angle T as parabola with 0 points at 15 ~ and 90 ~ and changed sign during periods of the SMF polarity reversal. We used data on tilt-angles from Internet for the period from May 1976 up to June 2000. On the basis of these data we determined the correlation between T and W for 11 month-smoothed data as T : 0.354W + 14.94 ~ (3) with correlation coefficient 0.938. For the period January 1953-April 1976 we determined tilt angle T approximately by Eq. (3) on the basis of 11 month-smoothed W data. Information on SMF polarity reversal periods we used from ftp://ftp.ngdc.noaa.gov/STP/SOLAR_DATA) ANALYSIS ACCORDING TO CLIMAX AND H U A N C A Y O / H A L E A K A L A NM DATA According to the procedure described above we correct the l 1-month-smoothed data of Climax NM on the drift effect for different values of Adr from 0% (no drift effect) up to 4%. The dependence of the correlation coefficient on the value of each value of
Adr
in Figure 1,
Xomax(Adr) can
X o is shown in Figure 1. For
be easy determined when the correlation
coefficient reaches a maximum value Rmax . The functions Rmax (Adr) and
X o max (Adr) are
shown in Figure 2. The function Rmax (Adr) can be approximated with correlation coefficient 0.9994__.0.0002 by parabola
Rmax (Adr )= aA2r + bAdr + c ,
(4)
where a = 0.00472__.0.00006, b = -0.00241+_0.0003, and c = -0.919__.0.011. From Eq. (4) we can determine Adr max when Rmax reaches the biggest value:
Adr max = - b / 2 a = (2.55 _ 0.06)%. The function X o max (Adr) can be
(5) approximated
with
correlation
coefficient
0.99982+__0.00007 by Xo max (Adr)- -(0.367 +_O.OO2)Adr + (16.35 __.0.11) av.month,
(6)
what gives X o max (Adr max )= 15.42 __.0.20 av.month.
(7)
-
188-
Modulation region of galactic cosmic rays in the heliosphere... For Huancayo NM (12S, 75W; height 3.4 km, cut-off rigidity 12.92 GV)/Haleakala NM (20N, 156W; 3.03 km, 12.91 GV) data we found by the same procedure that Adr max = (0.41+_ 0.01)%, X o max (Adr max )= 15.39 _ 0.19 av.month (8)
5w
/ Figure 1. Dependences of the correlation coefficient from X o for different Adr from 0% (no drifts) up to 4% according to Climax NM data (N39, W106; height 3.4 km, cut-off rigidity 2.99 GV) in cycles 19-22.
16
'1
~-
I,LI
e'-
-0.94
E
"~ I11 n,' -0.925 O O
Figure 2. Functions Rmax (Adr) and X o max (Adr) for 19-22 cycles according to Climax NM data.
-
189-
L.L Dorman MAIN RESULTS AND DISCUSSION According to Eq. (7) and (8) the time of solar wind moving from the Sun to the boundary of modulation region is 15.4+_0.2 av. month what gives for the dimension of modulation region ro = u x X o max = 119.0 t 1.5 A U (9) (according to direct measurements on space probes the average solar wind speed for the period 1965-1990 was u=4.41x107cm/s, therefore av. month corresponds to 7.73 A U). Determination of parameters A and B in Eq. (1) gives important possibility to estimate CR intensity out of the modulation region (A = ln(n o (R))) and the radial diffusion coefficient in the maximum of SA (DrmaX(R)= 1.5u2/B ). We can determine regression coefficients A and B only for integer values of X o (because we used monthly data). Therefore, for example, for Climax NM we determine A and B at Adr--2.5% according to Eq. (5), for Xo=15 and 16, and then
by
extrapolation
for
Xomax =15.4__.0.2 what
gives
A=8.37084+0.00009,
B=(0.0132+_0.0002) (av. month) -1 . It means that for Climax NM (10-15 GV primary CR)
In(no(R)) = 8.37084 +_0.00009, Drmax = (5.81__. 0.09)x 1023
cm2/s.
(10)
The transport path and time diffusion of galactic CR 10-15 GV from the boundary of modulation region to the Earth's orbit in maximum SA can be estimated approximately as max (R) ~" r 2 / d D max (R) ~ 0.35 av.month 9 -'rAmax(R)= (5.8 +_0.1)x 1013 cm, r die
(11)
Estimated TdmFC(R) is much smaller than time-lag between CR and SA variations (this supports the using for NM energies of quasi-stationary model of CR modulation in the Heliosphere). By the same way for Huancayo/Haleakala NM data (primary CR 30-40 GV) at Adr--0.4% according to Eq. (8), we obtain: ln(n o (R))= 7.46635 -+ 0.00004, Drmax = (2.19 + max = 0.092 av. month (12) - 0.06)1024 cm2/s ' Tdif The change of radial diffusion coefficient with SA can be described according to Eq. (2) by
D r (R,W)= Drmax (R)x (W/Wma x )-(1/3)-(2/3)(1-W/Wmax ) .
(13)
Now we can determine the absolute CR modulation relative to the intensity in the interstellar space. For example, for Climax NM the absolute decreases were on 23% in February 1958 (cycle 19), on 18% in May 1969 (cycle 20), on 25% in August 1979 (cycle 21), and on 34% in June 1991 (cycle 22). The relative role of drifts for cycles 19-22 for 10-15 GV galactic CR was 2x2.5%/25% - 0.2 and role of convection-diffusion about 0.8. For 30-40 GV galactic CR the relative role of drifts was about two times smaller: 2x0.4%/8% ~- 0.1 and role of convectiondiffusion about 0.9. R E F E R E N C E S
1. R. A., Burger, and M. S. Potgieter, The effect of large heliospheric current sheet tilt angles in numerical modulation models: A theoretical assessment, Proc. 26th Inter. Cosmic Ray Conf., 7, 13-16, 1999. 2. L. I. Dorman, Cosmic ray long-term variation: even-odd cycle effect, role of drifts, and the onset of cycle 23,Adv. Space Res., Paper D1.1-0037, 2001 (in press). 3. L. I. Dorman, N. Iucci, and G. Villoresi, Time lag between cosmic rays and solar activity in solar minimum 1994-1996 and the residual modulation, Adv. Space Res., Paper D1.1-0034, 2001 (in press).
-
190-
Latitudinal gradients and charge sign dependent modulation of galactic cosmic rays B. Heber ~, R Ferrando b, C. Paizis c R. Mtiller-Mellin d, H. Kunow d, M.S. Potgieter e, S.E.S. Ferreira ~, and R.A. Burger e ~Max-Planck-Institut ftir Aeronomie, Germany bCEA/Saclay, Service d'Astrophysique, France ~Universita di Milano, Italy aIEAP, University of Kiel, Germany ePotchefstroom University for CHE, South Africa Ulysses, launched in October 1990, began in December 1997 its second out-of-ecliptic orbit. The solar activity is rising to maximum conditions for this second orbit, whereas it was declining to low activity for the first out-of-ecliptic orbit. According to drift-dominated modulation models, the intensity of galactic cosmic rays depends on the latitudinal extension of the heliospheric current sheet (HCS). The latitudinal gradient as well as the charge sign dependent variation of 2.5 GV protons and electrons observed during the previous Ulysses orbit can be described by such models. In this paper we investigate these two parameters along the Ulysses maximum orbit. The electron-to-proton ratio and the latitudinal gradient are qualitatively in agreement with the model predictions up to fall 1999. Although Ulysses was then moving to higher southern latitudes, the latitudinal gradient is significantly reduced, and the electron-toproton ratio does not depend on the latitudinal extension of the HCS until June 2000, when the electron-to-proton ratio increased again. In both cases, we suggest that these changes are correlated with the reconfiguration of the heliospheric magnetic field. 1. INTRODUCTION The intensity of galactic cosmic rays (GCRs), energetic charged particles entering the heliosphere, is modulated as they traverse the turbulent magnetic field embedded in the solar wind. These particles are scattered by irregularities in the interplanetary magnetic field and undergo convection and adiabatic deceleration in the expanding solar wind. The large-scale heliospheric magnetic field (HMF), approximated by an Archimedean spiral (Parker, 1965), [7] leads to gradient and curvature drifts of cosmic rays in the interplanetary medium (Jokipii et al., 1977) [6]. The latter causes the GCR flux to vary with the 22-year solar magnetic cycle: 1. In an A<0 (the heliospheric magnetic field is pointing inward in the northern hemisphere) the time profile of positively charged particles is peaked, whereas it is more or less flat in A>0 solar magnetic cycle. Electrons are expected to behave oppositely. 2. Latitudinal and radial gradients are ex-
-191 -
B. Heber et al.
r [AU]
1.4
. . . .
t
. . . .
F . . . .
i
. . . .
~ . . . .
i
. . . .
f . . . .
i . . . .
i . . . .
o [o]
i -5
I -20
I -35
I i - ~ - 7 0
li 7060
I 30
i 0
I -15
i
I
-30-45
A+ 1.2
e/p at 1 GV
r
0.8
0.6
0.4
. . . .
i
10
. . . .
i
20
. . . .
i
30
. . . .
I
40
. . . .
i
. . . .
50
i . . . .
60
J
70
. . . .
i , , , ,
80
10
90
1990 1992 1994 1996 1998 2000
et [degree]
Year
Figure 1. Calculated electron-to-proton ratio at 1 GV as function of the maximum latitudinal extent (oz) of the heliospheric current sheet.
Figure 2. 52-day averaged quiet time count rates of > 100 MeV protons at Ulysses (a), at Earth (c) and 2.5 GV electrons from KET on Ulysses (b).
pected to be different for oppositely charged particles. Both the charge sign dependent time profiles as well as the positive and negative latitudinal gradients for GCR nuclei have been observed around solar minimum (Heber and Potgieter, 2000 [3], and references therein). Recently Burger and Potgieter (1999) [1] investigated the proton and electron intensities as a function of the maximum latitudinal extent, c~, of the Heliospheric Current Sheet (HCS). The electron-to-proton ratio as a function of o~ for 1 GV particles during a full A>0 solar cycle is shown in Fig. 1. Burger and Potgieter (1999) [1] found three regimes in the electronto-proton ratio: 1. Below 40 ~ the electron-to-proton ratio is decreasing with increasing oz, 2. between 40 ~ and 80 ~ it is nearly constant, and 3. it is increasing towards the non drift value between 80 ~ and 90 ~ Heber et al. (2000) [4] suggest that drifts are globally important as long as the HMF has a basic, ordered configuration during increased solar activity. In this paper we extend their investigation to later time periods, including the period when the solar magnetic field polarity is reversing. 2. INSTRUMENTATION AND OBSERVATIONS Ulysses, launched on October 6, 1990, encountered in February 1992 the planet Jupiter, and, using a Jupiter gravity assist, began its journey out of the ecliptic plane. The observations presented here were made with the Kiel Electron Telescope (KET) on board Ulysses and the University of Chicago (UoC) instrument on board IMP8 at 1 AU. Fig. 2 displays the 52-day averaged quiet time count rate of > 100 MeV protons measured by KET (a) and the UoC IMP (c) particle sensor. Quiet time fluxes have been defined by analyzing the ~50 MeV proton channels, as described in Heber et al. (1995) [5]. These two measurements allow us to separate spatial from temporal variations in the GCR flux. The intensity time profiles at both locations are very similar. Differences in these profiles due to Ulysses' trajectory can be seen in the figure around solar minimum from 1994 to late 1996; for a detailed discussion see Heber and Potgieter (2000) [3]. Curve (b) in Fig. 2 displays
-
192-
Lat#udinal gradients and charge sign dependent modulation of galactic cosmic rays
Figure 3. Heliographic equator equivalent electron-toproton ratio as a function of the time shifted c~.
Figure 4. Source surface fields in April 1999 (a) and June 2000 (c) (for details see text).
the count rate of 2.5 GV electrons from KET on Ulysses. Comparing curves (a) and (b) it was found that GCR protons and electrons have different latitudinal gradients, in agreement with modulation models. Since the gradient of 2.5 GV electrons is consistent with zero (for a detailed discussion see Heber et al., 1999, [2]) and the > 100 MeV protons have nearly the same rigidity (~1.7 GV), it follows by comparing curves (b) and (c) in 1996 and 1997 that electrons respond differently to changes in c~ than protons (Heber, et al. 1999) [2]. In what follows we discuss the temporal evolution of the electron-to-proton ratio from 1998 on. Since no new IMP data are available, we will refer to the results given in Heber et al. (2000) [4]. 3. D I S C U S S I O N AND C O N C L U S I O N S During the present approach to solar maximum and with Ulysses at high southern heliographic latitudes, it is possible to determine simultaneously the electron and proton count rate ratio and the latitudinal gradient for GCR protons. Since cosmic rays do not respond immediately to a change in c~, but with a certain delay, we shift the calculated c~ at the source surface by 5 solar rotations to later times (cts). Fig. 3 displays the (heliographic equator equivalent) electron-to-proton ratio as a function of c~s, as described in Heber et al. (2000) [4]. The (heliographic equator equivalent) electron-to-proton ratio is calculated from the measured electron and proton count rates by correcting the protons only for the observed latitudinal variation. The corresponding correction factor has been obtained from the latitudinal variation during the fast latitude scan in 1994 and 1995. In contrast to protons electrons do not show any significant latitudinal gradient (see Heber and Potgieter, 2000, and references therein) [3], so that no corrections to the electrons have been applied. The filled and open dots are the ratios during the decreasing and rising phase of the solar cycle. (a) and (c) correspond to two time periods where the source surface maps are shown in Fig. 4. These two maps in April 1999 and June 2000 were obtained from http://quake.stanford.edu/~wso/and
-
193-
B. Heber et al. calculated by using the "radial" boundary conditions at 3.25 solar radii. Within these 10 months the source field changed from a two-sector structure with the northern and southern polarity in the northern and southern hemisphere to a structure with a predominantly southern and northern polarity in the northern and southern hemisphere. The structure in summer 1999 during the time period (b) in Fig. 3, not shown here, was characterized by a more complex behaviour. Therefore the HMF may rapidly change from a relatively ordered (dipole-dominated) configuration to one where higher order components become equally important (Wang et al., 2000) [8] during increased solar activity. As discussed in Heber et al. (2000) [4], the latitudinal gradient before fall 1999 (a) is consistent with a mean latitudinal gradient of Go ~ 0.23 %/AU at solar minimum and drops to low values during period (b). The fact that in April 1999 the observed latitudinal gradient is still consistent with the one measured around solar minimum, suggests that drifts are still important. During the second period (b) the HMF pattern becomes much less regular. At the same time the latitudinal gradient drops and the electron-to-proton ratio reaches a constant value for 10 months. Given that drift motions have a net resulting effect on modulation when extending over large (global) scales, one can argue that when the HMF configuration progressively develops from a well ordered (dipole-like) configuration to a "fragmented" pattern, global drift motions are becoming progressively less important (phase out). It is important to note that this increase in the electron-to-proton ratio is not necessarily indicative of large drift effects, but rather an indication of how drifts phase out for electrons and protons. The newest data (c) indicates a further increase in the electron-to-proton ratio. Such an increase is expected in our "model" when a better ordered HMF configuration has established itself during an A < 0 epoch. The lower panel in Fig. 4 indicates that during this period the solar magnetic field is indeed reversing. Simultaneously, a change from a positive to a negative latitudinal gradient for protons should also be measured, i.e., the GCR proton flux should become lower at high latitudes than at Earth. However, it is difficult to predict how long the magnetic field structure will keep this pattern during extreme solar activity. It might have changed again with the extensive activity in July 2000. Within the following year, with Ulysses at high heliographic latitudes and the possibility to determine the intensity time profiles for protons and electrons we will make progress on how the global configuration of the HMF influences the modulation of GCRs. REFERENCES 1.
2.
3. 4. 5. 6. 7. 8.
R.A. Burger and M. S. Potgieter. In Proc. 26th ICRC, 7, 13, 1999. B. Heber, E Ferrando, A. Raviart, G. Wibberenz, R. Mtiller-Mellin, H. Kunow, H. Sierks, V. Bothmer, A. Posner, C. Paizis, and M. S. Potgieter. Geophys. Res. Lett., 26(14):2133, 1999. B. Heber, and M. S. Potgieter. Adv. Space Res., 26(5):839, 2000. B. Heber, M.S. Potgieter, R. A. Burger, G. Wibberenz, R. Mtiller-Mellin, H. Kunow, E Ferrando, A. Raviart, C. Paizis, and C. Lopate. J. Geophys. Res., 2000. submitted. B. Heber, A. Raviart, C. Paizis, W. Dr6ge, R. Ducros, E Ferrando, H. Kunow, R. MtillerMellin, C. Rastoin, K. R6hrs, H. Sierks, and G. Wibberenz. Adv. Space Res., 16, 1995. J.R. Jokipii, E. H. Levy, and W. B. Hubbard. Ap. J., 213:861-868, 1977. E.N. Parker. Planet. Space Sci., 13:9-49, 1965. Y.-M. Wang, N. R. Sheeley, and N. B. Rich. Geophys. Res. Lett., 27(2): 149-152, 2000.
- 194-
Antiprotons below 200 MeV in the interstellar medium" perspectives for observing exotic matter signatures I.V. Moskalenko a-t, E.R. Christian ~, A.A. Moiseev ~, J.F. Ormes a, and A.W. Strong b aNASA/Goddard Space Flight Center, Greenbelt, MD 20771, USA bMax-Planck-Institut ffir extraterrestrische Physik, 85741 Garching, Germany
Most cosmic ray antiprotons observed near the Earth are secondaries produced in collisions of energetic cosmic ray (CR) particles with interstellar gas. The spectrum of secondary antiprotons is expected to peak at ~ 2 GeV and decrease sharply at lower energies. This leaves a low energy window in which to look for signatures of exotic processes such as evaporation of primordial black holes or dark matter annihilation. In the inner heliosphere, however, modulation of CRs by the solar wind makes analysis difficult. Detecting these antiprotons outside the heliosphere on an interstellar probe removes most of the complications of modulation. We present a new calculation of the expected secondary antiproton flux (the background) as well as a preliminary design of a light-weight, low-power instrument for the interstellar probe to make such measurements.
1. I N T R O D U C T I O N The nature and properties of the dark matter that may constitute up to 70% of the mass of the Universe has puzzled scientists for decades, e.g. see [1]. It may be in the form of weakly interacting massive particles (WIMPs), cold baryonic matter, or primordial black holes (PBHs). It is widely believed that the dark matter may manifest itself through annihilation (WIMPs) or evaporation (PBHs) into well-known stable particles. The problem, however, arises from the fact that a weak signal should be discriminated from an enormous cosmic background, including a flux of all known nuclei, electrons, and v-rays. Antiproton measurements in the interstellar space could provide an opportunity to detect a signature of such dark matter (see [2] and references therein). High energy collisions of CR particles with interstellar gas are believed to be the mechanism producing the majority of CR antiprotons. Due to the kinematics of the process they are created with a nonzero momentum providing a low-energy "window" where exotic signals can be found. It is therefore important to know accurately the background flux of interstellar secondary antiprotons and to make such measurements outside the heliosphere to avoid any uncertainties due to solar modulation. * N R C Senior Research Associate talso Institute of Nuclear Physics, M.V.Lomonosov Moscow State University, Moscow 119899, Russia
-
195-
I. V. Moskalenko et al.
Figure 1. L e f t : Calculated proton interstellar spectrum (LIS) and modulated spectrum (~ = 750 MV) with data [8-11]. R i g h t : Calculated i3 interstellar spectrum (LIS) and modulated spectrum (~ = 400 MV) with data [12,13]. The top curves are the total background. The "tertiary" components (LIS and modulated) are shown separately. Also shown is an example exotic signal [14] extended to lower energies.
2. C A L C U L A T I O N S
OF T H E A N T I P R O T O N
BACKGROUND
made a new calculation of the CR/3 flux in our model (GALPROP) which aims to reproduce observational data of many different kinds: direct measurements of nuclei, p's, e+'s, "),-rays and synchrotron radiation [3-5]. The model was significantly improved, and entirely rewritten in C + + . The improvements involve various optimizations relative to our older FORTRAN version, plus treatment of full reaction networks, an extensive cross-section database and associated fitting functions, and the optional extension to propagation on a full 3D grid. For this calculation, we used a cylindrically symmetrical geometry with parameters that have been tuned to reproduce observational data [5]. The propagation parameters including diffusive reacceleration have been fixed using boron/carbon and l~ ratios. The injection spectrum was chosen to reproduce local CR measurements, ~ flp-2.3s where/3 is the particle speed, and p is the rigidity. The parameters used" the diffusion coefficient, D ~ - 4.6 • 1028/3(p/3 GV) ~/3 cm 2 S-1, Alfven speed, VA -- 24 km s -1, and the halo size, Zh -- 4 kpc. We calculate iv production and propagation using the basic formalism described in [3]. To this we have added i0 annihilation and treated inelastically scattered/3's as a separate "tertiary" component (see [6] for the cross sections). The i0 production by nuclei with Z > 2 is calculated in two ways: employing scaling factors [3], and using effective factors given by Simon et al. [7], who make use of the DTUNUC code, and which appears to be more accurate than simple scaling. (The use of Simon et al. factors is consistent since We
- 196-
Antiprotons below 200 Me V in the interstellar medium: ...
their adopted proton spectrum resembles our spectrum above the f production threshold.) The effect on the p flux at low energies is however small, and the two approaches differ by about 15%. We believe our calculation is the most accurate so far since we used a self-consistent propagation model and the most accurate production cross sections [7]. The results are shown in Figure 1. The upper curves are the local interstellar flux (LIS) and the lower are modulated using the force-field approximation. The two lowest curves in Figure 1 (right) show separately the contribution of "tertiary" ffs, which is the dominant component at low energies. The adopted nucleon injection spectrum, after propagation, matches the local one. There remains some excess of ffs. The excess for the lowest energy points is at the 1 a level. Many new and accurate data on CR nuclei, diffuse gamma rays, and Galactic structure have appeared in the last decade; this allows us to constrain propagation parameters so that the limiting factor now becomes the isotopic and particle production cross sections. At this point we cannot see how to increase the predicted intensity unless we adopt a harder nucleon spectrum at the source in contradiction with constraints from high energy p data [3,5]. More details will be given in a subsequent paper. 3. A N T I P R O T O N
DETECTOR
Very limited weight and power will be available for any experiment on board an interstellar probe [15]. We therefore propose a simple instrument which is designed to satisfy these strict constraints [2]. We base our design (Figure 2 left) on a cube of heavy scintillator (bismuth germanium oxide [BGO]), with mass of the order of 1.5 kg. The cube, 42 g cm -2 thick, will stop antiprotons and protons of energy below 250 MeV. A time-of-flight (TOF) system is used to select low-energy particles. The particles with energy less than 50 MeV will not penetrate to the BGO crystal through the T O F counters, setting the low-energy limit. The separation of antiprotons from protons is the most challenging aspect of the design. A low-energy proton (below 250 MeV) that would pass the T O F selections cannot deposit more than its total kinetic energy in the block. Therefore an event will be required to deposit > 300 MeV to be considered an antiproton. A proton can deposit comparable energy in this amount of material only through hadronic interaction, which our Monte Carlo simulations show requires a proton with energy > 500 MeV. The T O F system can effectively separate low-energy particles (< 250 MeV) from such protons and heavier nuclei. As a conservative estimate, we assume that all protons with energy > 500 MeV have the potential to create a background of "p-like" events and their integral flux in interstellar space is somewhat uncertain but would be ~ 1 cm -2 s -~ sr -~. The exotic p signal, to be significantly detected above the background, should be of the order of 10 -6 cm -2 s -~ sr -1 in the energy interval 50-200 MeV, which corresponds to an expected signal/p-background ratio of ~ 10. We thus can allow only one false antiproton in 107 protons. Simulations to date indicate that the current design will have rejection power of 2 • 106. We expect to get the next factor of five by fine-tuning the design and selections. The efficiency of antiproton selection is shown in Figure 2 (right). The antiproton rate will be 0.1-1 particle per day, and the statistical accuracy will be ~ 10% after 3 years of
-
197-
I.V. Moskalenko et aL
~TOF1
Incident
O
.,..~
antiproton
o
0.9
...............................................................................................................................
,.. . . . . . .
0.8
...............................................................................................................................
i ...........
~=
0.7
................................................................................
oo
0.6
,.-w
TOF2 I
'1 . . . . . .
I
~l
+
~
+
~ ............. i .......................... i........... i
TOF3 .~..~
-~
~
0.5
+
~.~
+
.o TOF4
~
. . . . .
0.3
.....................................................................................................................................................
0.2
......................................................................................................... i .............. !-...........................
m
B I N ) crystal 0.l
......................................................................................................
0
. . . . . . . . . . . . . .
0
0.025
0.05
'
. . . . . . . . .
0.075
0.1
i
. . . . . . . . .
0.125
0.15
i+
. .......................................
;
. . . .
0.175
;
0.2
. . . .
i
. . . .
0.225
0.25
Energy of incident antiproton, GeV
5 cm
Figure 2. Left: Preliminary design of the detector. detection after selections applied.
Right: efficiency for antiproton
observation. I.V.M. acknowledges support from NAS/NRC Senior Research Associateship Program. REFERENCES ,
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
Trimble, Ann. Rev. Astron. Astrophys. 25 (1987) 525 J.D. Wells, A. Moiseev, J.F. Ormes, Astrophys. J. 518 (1999) 570 I.V. Moskalenko, A.W. Strong, O. Reimer, Astron. Astrophys. 338 (1998) L75 A.W. Strong, I.V. Moskalenko, Astrophys. J. 509 (1998) 212 A.W. Strong, I.V. Moskalenko, O. Reimer, Astrophys. J. 537 (2000) 763 L.C. Tan, L.K. Ng, J. Phys. G: Nucl. Part. Phys. 9 (1983) 227 M. Simon, A. Molnar, S. Roesler, Astrophys. J. 499 (1998) 250 W. Menn et al., Astrophys. J. 533 (2000) 281 M. Boezio et al., Astrophys. J. 518 (2000) 457 T. Sanuki et al., Astrophys. J. 545 (2000) 1135 I.P. Ivanenko et al., Proc. 23rd ICRC (Calgary) 2 (1993) 17 S. Orito et al., Phys. Rev. Lett. 84 (2000) 1078 S.J. Stochaj et al., Astrophys. J. (2001) in press L. BergstrSm, J. EdsjS, P. Ullio, Astrophys. J. 526 (1999) 215 R.A. Mewaldt, P.C. Liewer, These Proceedings
V.
- 198-
Anomalous Cosmic Rays outside of the Termination Shock A. Czechowski ~ S. Grzedzielski ~' b H. Fichtner c M. Hilchenbach d and K.C. Hsieh e ~Space Research Centre, Polish Academy of Sciences, Bartycka 18A, PL 00-716 Warsaw, Poland bService d'Aeronomie du CNRS, B.P. No 3, Verrieres le Buisson, F-91371, France CInstitut fiir Theoretische Physik IV, Ruhr-Universit/it Bochum, 44780 Bochum, Germany dMax-Planck-Institut fiir Aeronomie, D-37191 Katlenburg-Lindau, Germany ~Physics Department, University of Arizona, Tucson AZ 85721, U.S.A. Anomalous Cosmic Rays (ACR) are most probably generated at the solar wind termination shock. When observed in the region inside the shock, the ACR energy spectrum is significantly modulated. The region outside the shock is so far inaccessible to observations. However, the low energy (~ 100 keV) part of the spectrum beyond the shock could be sampled by means of energetic neutral atoms (ENA), into which the ACR ions convert by charge-exchange with the neutral atoms. As the distribution of ACR reflects the large-scale structure of the heliosphere, this offers a possibility of imaging the outer heliosphere by ENA from this source. 1. I N T R O D U C T I O N The possibility of observing energetic neutral atoms (ENA) of heliospheric origin was first discussed by Hsieh et al. [1]. Their discussion included also the ACR ENA, the anomalous cosmic ray (ACR) ions which become neutralized by charge-exchange with the atoms of the background. As the low energy ACR ions (with large charge-exchange cross section) cannot penetrate upstream of the solar wind termination shock, the main source of ACR ENA must be the region beyond the shock. Grzedzielski [2] first observed that the ACR distribution in the outer heliosphere should reflect the asymmetry of the heliospheric plasma flow, with the stagnation point and extended heliotail. Assuming simple models of the heliospheric plasma flow (Parker [3]), the ACR spatial distribution beyond the termination shock was calculated ([4], [5], [6]). Because of convection by plasma flow and the diffusion across the heliopause, the ACR were found to concentrate in the region of the heliotail, so that the ACR ENA flux would have a maximum from the heliotail (anti-apex of the LISM, the local interstellar medium) direction. The observations by CELIAS/HSTOF on SOHO during 1996 and 1997 [7] were consistent with this conclusion: the flux of mass=l 55-80 keV neutral particles detected by the instrument peaked during the periods when the instrument field-of-view included the LISM anti-apex.
-
199-
A. C z e c h o w s k i et aL
'
'
'
'
'
'
'
'
'
'
"
' .....
kousch' Porker
'
'
'
....... ""'"'
o"
".
" ~ "
o*
~E l'Oxl 0 - 4 o
~
Figure 1. ACR density for different models of the heliosphere: Kausch's, Parker's and Kausch's with nH set to 0.1 cm -a as used in Parker's model. The density profiles shown are for 0 = 0 ~ (apex)104 ~ 133 ~ 162 ~ 180 ~
2. A C R D I S T R I B U T I O N
BEYOND
Figure 2. ACR ENA hydrogen flux at 63 keV for the models illustrated in Fig. 1: Kausch's and Parker's. The flux is shown as a function of direction (0 = 0 ~ is the apex direction).
THE SHOCK
The next step ([8],[9]) was to use a more realistic model, based on a numerical solution of the gas-dynamical equations obtained by Kausch [10]. In this model the termination shock is nonspherical (with nonuniform parameters), the flow has nonzero divergence, and the neutral hydrogen from the LISM has the density inside the heliopause reduced (by a factor of 3-4) due to interaction with plasma. All these features affect the transport of ACR. As the simple models used before, Kausch's model is effectively two-dimensional (axially symmetric with respect to the LISM apex-antiapex axis). The ACR distribution function f(x, p, t) is calculated by solving numerically the transport equation (Parker [11]) in the region downstream from the termination shock: O t f -- V . ec. V f - V .
V f nt
1
Of ~ V
301np
. V-
flf
(1)
where V is the velocity of the plasma flow, ~ the particle diffusion tensor, p the absolute value of the particle momentum, and # the loss rate (at ~ 100 keV it is predominantly the charge-exchange loss term #~). The ACR particles are treated in a test-particle approximation. The boundary conditions at the shock prescribe the ACR flux as a function of position and energy. Figures 1 to 4 correspond to the ACR spectrum uniform over the shock (flux ~ E -~42, see [12]). The nonuniform spectrum determined by the local shock parameters was also considered [13]. Most calculations were restricted to the timeindependent case. Also, the solution was assumed to be axially symmetric with respect to the LISM apex-antiapex axis. The diffusion tensor was replaced by a scalar coefficient, equal to ec2 outside and to ~c~ inside the heliopause. As the magnetic field downstream of the shock is expected to have a very complicated form, this assumption, corresponding
-
2 0 0
-
Anomalous cosmic rays outside of the termination shock
> "
0_ 4
m
,,
1
I
Figure 3. Evolution of the ACE energy spectrum downstream from the shock. The dotted lines correspond to omitting the V . V term.
Figure 4. The ENA hydrogen energy spectrum in the anti-apex and apex direction. The ACR spectrum assumed at the shock is also shown.
to the disordered field, may be reasonable as a lowest order approximation between the shock and the heliopause. In most calculations ec2 >> eel was assumed. ~1 was estimated by eel - (1/3)ec I with ec I given by the formula of le Roux et al. [14]. In Fig. 1 the ACR distributions following from the simulations based on the Kausch and Parker models are compared. The density profiles start at the termination shock and have a sharp change in slope at the heliopause (ec2 - 102ec~). With the exception of the heliotail region, the density falls rapidly towards the heliopause, which is approximately the free escape surface. The difference between the models is partly due to different assumptions about the neutral gas density (0.1 cm -a in the calculations based on the Parker model, while Kausch's solution corresponds to 0.02-0.03 cm -a inside the heliopause) but also to the difference in the plasma flows. The difference can be seen both in the ACR density distributions and in the corresponding ACR ENA flux (Fig. 2). The ACR ENA flux has a peak from the heliotail direction (0 - 180~ The shape of the peak is clearly model dependent. Similar results were obtained in the case of nonuniform ACR shock spectrum Figure 3 shows the evolution of the ACR energy spectrum downstream of the shock in the heliotail direction [9]. The spectra at the distances of 187 AU (shock), 275 AU, 380 AU, 600 AU and 990 AU from the Sun are shown. The dotted lines are obtained by neglecting the ~ V - V term in the transport equation. The flow divergence apparently does not affect very strongly the shape of the spectra (the effect is even weaker in the apex direction). The main contribution to modulation of the spectrum is due to chargeexchange processes which reduce the density at low energy. Figure 4 presents the ACR ENA energy spectrum. An important conclusion is that the spectrum, particularly in the heliotail direction, is not as steep as expected for a product of the ACR energy spectrum by the charge-exchange cross section: this is because the fall in the charge-exchange rate reduces the loss rate of ACR, so that the ACR density decreases less rapidly with distance.
-201 -
A. Czechowski et al.
3. D I S C U S S I O N The ACR-generated ENA flux should have a maximum from the heliotail direction (which may be shifted from the anti-apex due to interstellar field, see [15]). The detailed shape of the peak in the ENA flux as a function of direction depends on the model of the heliosphere (Fig. 2). The ACR ENA flux intensity scale is determined by the ACR flux intensity at the source, which (in the test particle approximation) is simply proportional to the ACR flux intensity at the termination shock, a parameter of the model. The estimations of its value rely on the observations of a highly modulated (no low energy part) ACR energy spectra in the solar wind [12]. Extrapolation of the result to the energy range of interest to us (~ 100 keV) is a source of large uncertainty, also because the shape of the ACR spectrum at low energy may deviate from simple power law. The results of model simulations were compared with the CELIAS/HSTOF observations in [8]. One of the problems was the difference in ENA flux intensity between the 1996 and 1997 peaks. Re-evaluation of the CELIAS/HSTOF data (Hilchenbach et al. this conference) changed the situation: the peaks are now of similar height. On the other hand, the new calibration suggests that the absolute ENA intensity may be an order of magnitude higher than reported before, suggesting that the extrapolation of the ACR spectrum used in the ACR ENA calculations underestimates the low energy spectrum at the shock. A.C. acknowledges support from the project KBN 2 P03C 004 14. R E F E R E N C E S
1.
2. 3. 4. 5. 6. 7. 8.
9.
10. 11. 12. 13. 14. 15.
K.C. Hsieh, K.L. Shih, J.R. Jokipii and S. Grzedzielski 1992: ApJ 393, 756 S. Grzedzielski 1993: Adv. Space Res. 13, (6)147 E.N. Parker 1963: Interplanetary Dynamical Processes, Interscience, New York S. Grzedzielski, A. Czechowski and I. Mostafa 1993: Adv. Space Res. 13, (6)261-(6)264 A. Czechowski, S. Grzedzielski and I. Mostafa 1995: A&A 297, 892 A. Czechowski and S. Grzedzielski 1997: in: Proceedings of the 25th International Cosmic Ray Conference, Durban, July 30- August 6 1997, Vol. 2, p. 237 M. Hilchenbach, K.C. Hsieh, D. Hovestadt, B. Klecker, H. Griinwaldt et al. 1998: ApJ 503, 916 A. Czechowski, H. Fichtner, S. Grzedzielski, M. Hilchenbach, K.C. Hsieh, J.R. Jokipii, T. Kausch, J. Kota and A. Shaw 1999: in: Proceedings of the 26th International Cosmic Ray Conference, Salt Lake City, Utah, August 17-25 1999, Vol.7, p. 589-592 A. Czechowski, H. Fichtner and T. Kausch 1999: in: Proceedings of the 26th International Cosmic Ray Conference, Salt Lake City, Utah, August 17-25 1999, Vol.7, p. 523-525 H.J. Fahr, T. Kausch and H. Scherer 2000: A&A 357, 268-282 E.N. Parker 1965: Planet. Space Sci. 13, 9-49 E.C. Stone, A.C. Cummings and W.R. Webber 1996: J. Geophys. Res. 101, 11017 A. Czechowski, H. Fichtner, S. Grzedzielski, M. Hilchenbach, K.C. Hsieh, J.R. Jokipii, T. Kausch, J. Kota and A. Shaw 2001, A&A 368, 622 J.A. le Roux, M.S. Potgieter and V.S. Ptuskin 1996: J. Geophys. Res. 101, 4791 A. Czechowski and S. Grzedzielski 1998: Geophys. Res. Lett. 25, 1855-1858
- 202
-
IONIZING MEDIA AND THE OBSERVED CHARGE STATES OF ANOMALOUS COSMIC RAYS A.F. Barghouty 1, J.R. Jokipii 2, and R.A. Mewaldt 1 1California Institute of Technology, Pasadena, CA 91125, USA 2 University of Arizona, Tucson, AZ 85721, USA ABSTRACT Singly-charged anomalous cosmic rays (ACRs) can give rise to multiply-charged ACR ions when they suffer further ionization during acceleration at or near the solar-wind termination shock. Measurements at 1 AU by SAMPEX have shown that above ~ 25 MeV/nucleon ACR nitrogen, oxygen, and neon ions are multiply charged. These observations have also established that the transition from mostly singly-charged to mostly multiply-charged ACRs occurs at a total kinetic energy of ~ 350 MeV. Recent simulations for ACR oxygen using ambient hydrogen as the only ionizing medium at or near the termination shock are able to successfully model this transition. The simulated oxygen intensity, however, appears deficient at high energies, where multiply-charged ACRs dominate. This paper presents further simulations that now include neutral helium as part of the ionizing medium in addition to the ambient, neutral hydrogen. The inclusion of helium helps reduce the deficiency, but appears to fall short of accounting fully for the observed spectrum. To that end, including heavier neutrals, e.g., oxygen, as well as taking multi-electron stripping into account, are suggested for more realistic modeling of the observed charge states of ACRs.
INTRODUCTION According to the accepted theory of ACRs (Fisk et al., 1974) these cosmic-ray ions are believed to originate as interstellar neutrals that penetrate the heliosphere before getting ionized -either by solar radiation or by charge-exchange collisions with solar-wind protons- to become singly-charged pickup ions. The pickup ions are then convected by the solar wind to the solar-wind termination shock (SWTS) where they are accelerated up to tens of MeV/nucleon via the process of diffusive shock-drift acceleration (Pesses et al., 1981). Unlike galactic cosmic rays or solar energetic particles observed at 1 AU, ACRs are expected to be predominantly singly-charged ions. However, recent observational evidence from SAMPEX (Mewaldt et al., 1996a; Klecker et al., 1998) have shown that ACR nitrogen, oxygen, and neon above ~ 25 MeV/nucleon are multiply charged, with ionic charge states of 2, 3, and higher. At energies below .~ 20 MeV/nucleon most of the observed ACRs are singly charged. SAMPEX observations (Klecker et al., 1998) have further established that the transition from mostly singly-charged to mostly multiply-charged ACRs occurs at a total kinetic energy of ~ 350 MeV. The theory of diffusive shock-drift acceleration (Jokipii, 1996) gives the amount of energy A E an ACR ion gains in drifting along the SWTS (from the equator towards the poles during the 1990s) to be directly proportional to its ionic charge q, i.e., A E ~ 240 (MeV)• As a result, multiply-charged ACRs are accelerated to higher energy per nucleon than the more abundant singly-charged ions. Thus, at energies well above 240 MeV, the theory predicts that multiply-charged ACR ions will dominate if there is a source of such ions to accelerate. The predominance of multiply-charged ACRs at high energy has been interpreted as evidence that some singly-charged ACR ions suffer additional ionization during their acceleration at or near the SWTS (Mewaldt et al., 1996b; Jokipii 1996). Recent simulations (Barghouty et al., 2000) of ACR oxygen using ambient hydrogen as the only ionizing medium at or near the termination shock and a refined set of hydrogen-impact ionization cross-sections (Barghouty 2000) are able to successfully model the transition energy from singly to multiply-charged
- 203 -
A.F. Barghouty, J.R. Jokipii and R.A. Mewaldt
1.0 O
0.8
He/H=0 la; 0/H=0.005
0.6
He: Z ' x 0.13
O
~) 0.4
-
> .,.~ .,m
5
-
-~ 0.2 c~
0.0 10 1
10 2 10 3 10 4 K.E. ( k e Y / n u c l e o n )
10 5
oxygen. The simulated total, i.e., sum over charge states, ACR-oxygen spectrum, however, is found to be deficient at high energies (> 50 MeV/nucleon) where most of the oxygen ions are multiply charged. This deficiency suggests that further ionization during acceleration is still needed so as to provide more multiplycharged ions to be accelerated to higher energies. This paper examines the potential role of helium and heavier neutrals as ionizing media, in addition to hydrogen, at or near the SWTS. Multi-electron stripping in these processes, i.e., heavy atom-impact ionization, can also play a role, an effect that will be explored in detail in a forthcoming publication. AND HEAVY ATOM-IMPACT IONIZATION CROSS-SECTIONS At energies E > few hundreds of keV/nucleon and for ionizing media with nuclear charge Z > 1, the needed ionization cross-sections can be estimated using the scaling relation (Gillespie, 1983): HELIUM
3
Jokipii (1996) and more recently in Jokipii (2000). Very briefly, the model solves the time-dependent Parker heliospheric transport equation in 2 spatial dimensions with drift terms:
Ofq 0 ( Ofq ~ Ofq _ Vd,iOfq 10Vw,i Ofq ~-sourcesq Ot = Oxi _NiJ~ j - Vw,iOxi ~ -~ 3 0 x i Oln p
- sinksq
(2)
where fq(~',p, t) is the distribution function of the ACR ion with charge q = 1, 2, 3,..., Z. From left to right, the RHS terms in the equation depict spatial diffusion, convection, drift, energy gains and losses, and particle sources and sinks due to ionization. The salient transport parameters used for this study are tabulated below. [In Table 1 R is the ion's rigidity, ~ its Lorentz factor, and B the strength of the heliospheric magnetic field at the heliospheric point ~.] Figure 2 illustrates a sample simulation for ACR oxygen at 1 AU. Here we show simulations with and without taking the ionizing contribution of helium into account. Figure 2 suggests that taking helium into
-
204 -
Ionizing media and the observed charge states of anomalous cosmic rays
"6 -5 z 10 > 10 - 6 9
(~'..
!
o
r
~....
~, 10-8 h-,
.,.w
10-9 1
10 K.E. ( M e V / n u c l e o n )
100
account as an ionizing medium, in addition to the ambient hydrogen, does help reduce the deficiency in the simulated intensity at high energies. This addition, however, still seems to fall short of accounting fully for the observed spectrum at high energies. Varying one or more of the salient transport parameters in Table 1 can, in principle, yield a somewhat better fit. In this work, however, they are kept at their nominal values so as to focus on the ionization media and processes at or near the SWTS. Below we discuss the potential contribution and role of multi-electron loss channels in both hydrogen and heavy-atom impact ionization. M u l t i - E l e c t r o n Loss P r o c e s s e s While data remain scarce for proton or hydrogen-impact ionization cross-sections, multi-electron loss cross-sections have been shown to be insignificant compared to single-electron loss in ion-electron collisions at energies relevant to ACR studies (e.g., Krishnakumar and Srivastava, 1992; Deutch et al., 1999). For hydrogen-impact collisions at such energies, multi-electron removal is also expected to be insignificant compared to single-electron removal. For example, the cross section for the two-electron removal process p+He-+p + He +2+2e at 1 MeV/nucleon is found both experimentally (Shah and Gilbody, 1985) as well as in Monte Carlo studies (e.g., McKenzie and Olson, 1987) to be about 2.5 orders of magnitude smaller than that for the single-electron removal process p + H e - + p + H e +1 + l e (see Figure 3). However, as can been seen from Figure 3, for heavy atom-impact ionization, double-electron loss can be ,.o few percent of the single-electron loss cross-section at energies ~ MeV/nucleon. This small, but perhaps not insignificant contribution, will be explored in a future work.
1. Solar wind velocity SWTS radius SWTS strength Heliospheric boundary Neutral-H density Injection energy Heliospheric B-field
350 (equator)- 700 km/s (poles) 100 AU 2.5 160 AU 0.1 cm -3 60 keV/nucleon Parker's+polar modification+ fiat current sheet+ qA > 0 conditions
R1/~ ~/B(~
Parallel diffusion coeff. Perp. diffusion coeff.
0.03• Parallel
- 205 -
A.F. Barghouty, J.R. Jokipii and R.A. Mewaldt ,
.
.
.
.
,
~" 1 0 - 1 5 s
o -16 "~ 10
i
n
g
l
~
0
'~ -17 -~ 10 m 10 - 1 8 m 0
~
lO
-19 .
4,bodY 1 MeV/nucleon
...-~
1 10 N u c l e a r C h a r g e of I o n i z i n g M e d i u m
1
CONCLUSIONS While the simulations presented here for the charge states of ACR oxygen at 1 AU, which are based on the theory of diffusive shock-drift acceleration, appear to capture the essential physics of the charge-changing ionization processes taking place at or near the termination shock the simulated spectrum at high energies remains deficient even after taking helium as an additional ionizing medium into account. This suggests that further ionization during acceleration is still needed. To that end, including heavier neutrals, e.g., oxygen, as well as taking multi-electron stripping into account, are suggested for more realistic modeling of the observed charge states of ACRs. A CKN OWLED G EMENT S Work is supported by NSF grant 9810653 and NASA-JOVE NAG8-1208 (A.F.B.) and by NASA NAS530704 and NAG5-6912 at Caltech. A.F.B. thanks Ed Stone, Alan Cummings, Rick Leske, Conrad Steenberg (Caltech), and Mark Weidenbeck (JPL) for stimulating discussions and insightful comments. REFERENCES Barghouty, A.F., J.R. Jokipii, and R.A. Mewaldt, in Acceleration and Transport of Energetic Particles Observed in the Heliosphere, eds. R.A. Mewaldt et al., pp. 337, AIP ~528, Washington, DC, 2000. Barghouty, A.F., Phys. Rev. A, 61, 052702, 2000. Deutch, H., et al., Int'l. J. Mass Spect., 192, 1, 1999. Fisk, L.A., B. Koslovsky, and R. Ramaty, Astrophys. J. Lett., 190, L35, 1974. Gillespie, G.H., Phys. Lett., 93A, 327, 1983. Gloeckler, G., J. Geiss, and L.A. Fisk, in The Heliosphere near Solar Minimum" the Ulysses Perspectives, eds. A. Balogh et al., Springer-Praxis, Berlin, 2000. Jokipii, J.R., in Acceleration and Transport of Energetic Particles..., ibid, pp. 309, 2000. Jokipii, J.R., Astrophys. J. Lett., 466, L47, 1996. Klecker, B., et al., Space Sci. Rev., 83,259, 1998. Krishnakumar, E., and S.K. Srivastava, Int'I. J. Mass Spect. ~ Ion Proc., 113, 1, 1992. Leske, R.A., et al., in Acceleration and Transport of Energetic Particles..., ibid, pp. 293, 2000. McKenzie, M.L., and R.E. Olson, Phys. Rev. A, 35, 2863, 1987. Mewaldt, R.A., et al., Geophys. Res. Lett., 23, 617, 1996a. Mewaldt, R.A., et al., Astrophys. J. Lett., 466, L43, 1996b. Pesses, M.E., J.R. Jokipii, and D. Eichler, Astrophys. J. Lett., 246, L85, 1981. Shah, M.B., and H.B. Gilbody, J. Phys. B, 18, 899, 1985.
- 206 -
To achieve a self-consistent description of the heliosphere, a fusion of the models addressing its large-scale structure with those of the transport of energetic particles is desirable. We report here on a study of the potential for a thorough analysis of 4-D phase space distributions of anomalous cosmic rays, investigating their energy spectra in further detail. While the computation of such 4-D distributions, thus far, was limited to a spherical, heliocentric shock, we attempt here an investigation of the significance of more realistic shock geometries. The analysis is limited to periods around maximum solar activity for drifts are not taken into account. 1. I N T R O D U C T I O N The overwhelming majority of studies of galactic and anomalous cosmic rays (GCRs and ACRs) within the heliosphere (for a recent review see [1]) is based on the assumption that their phase space distributions do not possess any longitudinal structure. While this assumption appears to be well-justified for heliocentric distances smaller than about 50 AU, it is questionable for the outer heliosphere. There, longitudinal gradients should exist at least for ACRs because: 9 the particle population from which the ACRs originate, i.e. the pick-up ions (PUIs), have longitudinal flux variations (e.g., [2,3]); 9 the injection efficiency of PUIs into the process of diffusive acceleration at the heliospheric shock is likely to be a function of ecliptic longitude, because both the local shock structure (precursor, foot, ramp; e.g., [4,5]) as well as the orientation of the heliospheric magnetic field relative to the shock surface, are likely to change from upwind to downwind within the ecliptic; 9 according to recent models of the large-scale structure of the heliosphere (e.g., [6,7]) the distance to the heliospheric shock increases systematically from the upwind to the downwind direction. We present here results of an extension of our earlier modelling of 3-D [8] and 4-D ACR phase space distributions [9] by considering explicitly a non-spherical heliospheric shock.
-207-
S.R. Sreenivasan and H. Fichtner 2. T H E M O D E L The ACR transport is described on the basis of the time-independent Parker equation + 1
g(V'g~~
Of = 0 ; Olnp
~(~',p)=
(~j_ 0 0
0 ~. 0
0 ) 0 e;ll
(1)
with the distribution function f - f(~', p) in the 4-D phase space defined by heliocentric position f" and momentum p. The solar wind velocity is taken to be radial and constant, +-~ i.e. v~w = 400 km/s g~, and the diagonal elements of the spatial diffusion tensor e; are ~11 - 1022cm2/s (w/c)(P/1 GV)(B/B1Au) and e;• - 0.03~11, where w and P are particle speed and rigidity, respectively. Particle drifts in the heliospheric magnetic field are not included, so that the analysis is limited to periods close to maximum solar activity. The boundary conditions are a vanishing radial gradient of f at the inner boundary at r - to, and prescribed shock spectra at the outer boundary, i.e. the heliospheric shock. They are the same as in [9], i.e. are based on the studies [10] and [11], and depicted as the uppermost lines in the two panels of Fig. 2. 3. T H E L A R G E - S C A L E
ACR DISTRIBUTIONS
Fig. 1 shows the spatial distribution of anomalous hydrogen with a kinetic energy of 31 MeV for a non-spherical heliospheric shock (outermost dashed line) in an equatorial plane (left) perpendicular to the symmetry axis of the heliospheric magnetic field but containing the upwind-downwind axis (horizontal solid line) and in a meridional plane (right) containing both the symmetry axis of the heliospheric magnetic field and the upwind-downwind axis. The shock is elongated in the polar and the downwind direction by factors of 1.3 and 1.5, respectively, which are found in (M)HD studies of the large-scale structure of the heliosphere (see, e.g., [7]). 4. T H E A C R S P E C T R A
AND THEIR COMPARISON
Fig. 2 shows the spectral distribution of anomalous hydrogen for a spherical (left) and a non-spherical heliospheric shock (right), respectively. Since the non-spherical heliosphere is elongated downwind by a factor of 1.5, the boundary spectrum is located at 120 AU in that direction. Therefore, in the right panel of Fig. 2 there are two additional dash-dotted lines giving the spectra at 100 and 120 AU. Obviously, there are significant differences between the outer heliospheric downwind spectra for the spherical and the non-spherical heliosphere. In order to visualize these differences more clearly, Fig. 3 shows the ratio of the downwind spectra R - jsphericat/jnon-spherical for the three distances 66, 78 and 80 AU in the energy interval 1 MeV to 1 GeV. The longitudinal structure is restricted to the outer downwind heliosphere, i.e. to the region r > 50 AU, see left panel in Fig. 1. There, however, the differential flux at a given position is evidently influenced by the large-scale heliospheric structure. While at 66 AU the flux ratio is still rather moderate (R < 5) for all energies above 1 MeV, we find for the interval 1 - 10 MeV 60 > R > 3 at 78 AU and even 110 > R > 4 at 80 AU. For the region beyond the ratio is even higher but since this is downstream of the spherical shock,
-208-
A CR modulation inside a non-spherical modulation boundary
6 "\'~...., T ~..
:
.... ..
,, ! l: ii"
-
2 0 9
-
st
~
S.R. Sreenivasan and It. Fichmer
5. S U M M A R Y
AND CONCLUSION
We confirm our earlier findings about the existence and the order of magnitude of (nonlocal) longitudinal gradients in the phase space distributions of ACRs. In particular we find that" if the heliosphere is non-spherical (see, e.g., Fig. 1), there exists a small longio;.o .... , ................. tudinal gradient in the outer heliosphere (r > 50 AU), shown here as an example for ACR H at a kinetic energy of 31 MeV; 9 this longitudinal structure is most clearly visible from a comparison of upwind and downind spectra (Fig. 2). The effect of heliospheric geometry is probably stronger than that of the variation in the source strength across the shock surface;
Figure 3" The ratio of the differential fluxes R - j~ph~i~g/j~o~-~ph~i~l in the outer downwind heliosphere (see Fig. 2) for the energy interval of interest for ACR H.
9 the polar elongation reduces the absolute flux as well as the latitudinal gradient at a given location in the heliosphere (Fig. 1); 9 the in-ecliptic spectra in the upwind direction (Fig. 2, solid lines) are basically unaffected by the geometry of the heliosphere;
9 the in-ecliptic spectra in the downwind direction (Fig. 2, dashed lines, and Fig. 3) show significant differences between a spherical and a non-spherical heliosphere, but the effect is limited to the outer heliosphere and to relatively low energies (< 10 MeV). In conclusion we find that, at least for anomalous hydrogen, there is a clear signature of the large-scale structure of the heliosphere in its spectra but that it is not likely to be observed in-situ with the presently active deep space probes: the effect is too small in the upwind heliosphere which is the region that is accessible to direct observation with the two Voyager spacecraft. REFERENCES ,
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
H. Fichtner, Space Sci. Rev. 95 (2001) 639. S.V. Chalov, H.J. Fahr and V. Izmodenov, Astron. Astrophys. 320 (1997) 659. U. Mall, H. Fichtner, E. Kirsch, D.C. Hamilton and D. Rucinski, Planet. Space Sci. 46 (1998) 1375. J.A. le Roux, H. Fichtner, G.P. Zank and V.S. Ptuskin, JGR 105 (2000) 12557. J.A. le Roux, G.P. Zank, H. Fichtner and V.S. Ptuskin, GRL 27 (2000) 2873. H.J. Fahr, T. Kausch and H. Scherer, Astron. Astrophys. 357 (2000) 268. G.P. Zank, this issue. H. Fichtner, H. de Bruijn and S.R. Sreenivasan, Geophys. Res. Lett. 23 (1996) 1705. H. Fichtner and S.R. Sreenivasan, Adv. Space Res. 23 (1999) 535. J.A. le Roux, M.S. Potgieter, and V. Ptuskin, JGR 101 (1996) 4791. C.D. Steenberg, Ph.D. Thesis, Univ. of Potchefstroom, South Africa (1998).
-210-
The Injection Problem for Anomalous Cosmic Rays I G.P. Zank, W.K.M. Rice, J.A. le Roux, and W.H. Matthaeus Bartol Research Institute, University of Delaware, Newark, DE 19716, USA 1. I N T R O D U C T I O N That interstellar pickup ions are related to the anomalous cosmic ray (ACR) component is now well established. However, pickup I-I+ has typical energies of 1-3 keV, sufficiently energetic to form a suprathermal, energetically important plasma component in the outer heliosphere, but well below ACR energies which extend to several 100 MeV/nuc. How some fraction of the interstellar pickup ion population becomes preferentially energized up to such high energies remains a central question in the physics of cosmic ray acceleration. It is generally accepted that for sufficiently energetic pickup ions, first-order Fermi acceleration, otherwise known as diffusive shock acceleration, at the heliospheric termination shock can account for the accelerated ACR spectrum. However, the precise mechanism by which some pickup ions are selected and possibly pre-energized up to energies sufficiently large that they can be Fermi accelerated remains a puzzle. This is the so-called "injection problem" for ACRs. Observations [ 1-3] at interplanetary shocks suggest that (i) accelerated ions emerge directly from the thermal pool; (ii) the injection efficiency appears to depend on the nature of the underlying particle distribution function, and (iii) the observed pickup ion injection efficiency appears to correlate inversely with mass. However, Cummings and Stone [4] used anomalous cosmic ray (ACR) energy spectra measured by Voyager 1 and 2 during 1994 to determine an injection efficiency for ACRs. This approach depends on uncertainties associated with the diffusion tensor model (and associated turbulence and drifts), termination shock location, neutral and pickup ion fluxes, shock strength, and the possible role of ions pre-accelerated by interplanetary shocks. Nonetheless, a basic conclusion is that, even if the estimated injection efficiencies are possibly inaccurate, heavier ions are injected preferentially. Furthermore, besides the ACR "injection efficiencies" [4] being in the opposite sense of those presented by Gloeckler et al. [2], the "injection efficiencies" of the latter authors are considerably higher than those found for ACRs. Finally, it appears [ 12] that heavy ion diffusive acceleration times can be longer than proton diffusive acceleration times. Another important result related to the injection and acceleration of ACRs at the termination shock is that the diffusive shock acceleration time scale for ions must be rapid [5]. To ensure that most ACRs remain singly ionized, one is obliged to assume a weak hard-sphere scattering model at a quasi-perpendicular termination shock [5]. The requirement that the scattering be weak (in the hard sphere sense) implies that the pickup ions already be energetic for them to be Fermi accelerated. If r/~ (defined below in 3) is a measure of the scattering strength, so that r/~ >> 1 corresponds to weak scattering, then for a particle to be scattered multiple times as a field line convects through a shock, the particle velocity v must satisfy v >> v,~o~, where V~his the shock speed [5,6]. Using reasonable This work supported in part by NASA grants NAG5-6469, NAG5-7796, an NSF-DOE award ATM-0078650, and an NSF award ATM-0072810.
-211 -
G.P. Zank et al. parameters [5] requires that pickup ions to have energies close to ~ 1 MeV [7]. As a consequence, it appears that some pre-energization of pickup ions is necessary. A successful resolution of the "injection problem" at the quasi-perpendicular termination shock must relate the above sets of observations consistently. 2. MULTIPLY REFLECTED ION ACCELERATION Since part of the pickup ion distribution function in the shock frame has a very small normal velocity component at the shock interface, these particles are reflected at the electrostatic cross shock potential ~ [7,8]. For a pickup ion shell ahead of the shock, the fraction of the distribution Rref that is incapable of surmounting the cross shock potential barrier is found to be [7]
Zm Rr~ --
2
M 2MI, ( r - 1)
,
where m refers to the proton mass, M and Z to the mass and charge of the particle of interest (pick-up I-F, He +, etc.). These reflected ions are capable of being accelerated to large energies by experiencing multiple reflections at the electrostatic barrier. If we interpret reflection efficiency as injection efficiency, then (i) heavier pickup ions, i.e., with M > m, are less efficiently injected, and (ii) injection efficiency increases with increasing charge. The downstream or transmitted pickup ion distribution is a power law in energy that is extremely hard (v-a) [7]. The balancing of the particle Lorentz force against the electrostatic potential gradient shows [7,8] that the maximum energy gain is proportional to the ratio of an ion gyroradius (whose velocity is that of the solar wind) to the smallest characteristic electrostatic shock potential length scale Lramp. If Lro~p is the thermal solar wind ion gyroradius, the initially very low velocity pickup ions will be accelerated up to no more than the ambient solar wind speed. However, our current understanding of the micro-structure of quasi-perpendicular shocks is that fine structure in the shock potential can be on the order of electron inertial scales [9], so yielding pickup ion energies of several 100 keV at even weak interplanetary shocks. 3. A SYNTHESIS
Although pickup ions are energized by the MRI mechanism, not all the MRI accelerated ions are sufficiently energetic to be further accelerated by a second-stage diffusive shock acceleration process. For diffusion theory to be applicable at a perpendicular shock, particles downstream of the shock must be capable of diffusing upstream. This requires the cosmic ray anisotropy be small. For a perpendicular shock, this implies that the particle velocity v 1/2r , where V,~ is the upstream flow speed in satisfy the condition [10] v >> (3V,~/r)(1 + 772) the stationary shock frame and r the shock compression ratio. For the present, we assume hard-sphere scattering to describe the transport of diffusive particles. Thus, if ;t n is the parallel mean free path ~ll/r_L = l+ 0 z, rlc = 3Xll/(vrg)= ~,ll/rg. Here rg = pc/(QB)is the particle gyroradius (p is particle momentum, Q the charge, c the speed of light, and B the magnetic field strength). In the hard sphere scattering model, r/, is a measure of the strength of scattering: rt~ small implies strong scattering whereas r/, large corresponds to weak scattering. For resonant scattering, Au oc R 1/3 where R = p c / Q is the particle rigidity. It
-212-
The injection problem for anomalous cosmic rays follows then that r/c ,,~ R -2/3, or
0 c rlcp(m/M) 2/3, 77cp -(2ll/rg)proton.
important
implication is that even if scattering is weak for protons, i.e., r/cp >> 1, the inverse dependence of 17C on M implies that 0c can be much smaller for heavy ions. Thus, heavy ions are accelerated diffusively at a much lower threshold velocity than lighter ions. Consequently, a larger fraction of heavy MRI accelerated ions will enter a second-stage diffusive shock acceleration process than light MRI accelerated ions. It remains to determine whether the two competing mass dependence effects conspire to satisfy the Cummings and Stone [4] ACR injection results. We consider two cases; (i) a highly peaked distribution such as a shell, and (ii) a power law distribution such as that obtained by Vasyliunas and Siscoe [1 1]. This gives (i) grey =
Rref
=
Rr~y(H +)[m/M] 1/2, and (ii)
R~/(H+)[m/M] 9 Simulations suggest that for (i), we use a (V/Vo) -4 accelerated spectrum, and for (ii) a
(v/v o)-5
accelerated spectrum. We suspect that the correct model lies between these two extremes. The differential intensity of the ith species ji(m-Zs-'sr-lMeV -1) at the lID
termination shock is given by j~ = pZfa(p,M). In order to calculate the acceleration efficiency in the same way as [4], we rewrite the differential intensity in their form,
10"
q e iF~ -~oiP(q-4 )/ZE(-q+ Z)/Z , w h e r e 47r F~(cm -2s -l ) is the flux of the ith pickup ion
i.e.,Ji = Fig. 1. Inverse acceleration efficiencies for I-I+, He +, O +, and Ne +, plotted as a function of ion mass M, and normalized to He +. The acceleration efficiencies for O +, and Ne + are very similar and greater than those of He + and IV. We would predict this to be true of C + and N + too. The injection/acceleration efficiency for I-I+ is anomalously low. The filled triangles correspond to a pickup ion shell distribution with the reflected 9 --4 MRI accelerated spectrum proportional to v , and the open circles to a Vasyliunus and Siscoe pickup ion distribution and a softer reflected MRI accelerated spectrum proportional to v -5. The derived observations are given by the open squares [4].
species at the termination shock, E0~ the
injection energy of 5.2 x 10 -3 MeV/nuc. used by Cummings and Stone, and e~ is the acceleration efficiency that is to be computed. Note our distinct use of the term "acceleration efficiency" for e~, which is referred to as "injection efficiency" by [4]. We use the former term to distinguish our use of injection efficiency in the context of MRI injection. The injection energies needed to render an ion diffusive range f r o m - 3 3 0 keV for pickup H + to about either ~ 185 (shell) or --75 (power law) keV for C +, N +, O +, and Ne +. Our predicted acceleration efficiency and that inferred by Cummings and Stone can now be compared directly since the same parameters -1 are used. In Fig. 1, we plot, following [4], the inverse acceleration efficiency e~ , normalized to He+, as a function of ion mass. The solid triangles correspond to an in initial pickup ion shell distribution, the open circles to the VS power law case, and the open squares to the
-213-
G.P. Zank et al.
inferred observations presented in [4]. Two points stand out from Fig. 1. The first is the obviously close agreement between the model results and the observed results. The second is that the extreme cases considered here, the initial shell and the VS distributions, yield acceleration efficiencies that are almost identical, suggesting that a more realistic pickup ion distribution at the termination shock is unlikely to alter our conclusions significantly. In the limit of strong scattering, the injection ~0"threshold for ions to be viewed as diffusive is pl virtually identical for all ion species. 10 ~ Consequently, no difference exists between the ~ ~0' injection efficiencies of different mass ion 10' species in the case of strong scattering. This is an important point since it relates for the first lOS time the differential injection efficiency of pickup ions of different masses to the particle t- ~0-'scattering strength. ~o~,0~. One can use the diffusive part of the MRI Energy (MeV) spectrum for pickup H + as the explicit source for ACRs at the termination shock, and hence Fig. 2. Fluxes of pickup I-I+, MRI determine the ACR spectrum at the shock and accelerated I-I+, and I-I+ ACRs at the termination shock, together with the the modulated spectrum at any point within the resulting modulated ACR flux at 57 AU heliosphere. The combined pickup ion, MRI (dashed line), accelerated and ACR spectrum at the termination shock, assumed to be located at 80 AU with parameters given by [4], is illustrated in Fig. 2. The flux of ACRs shown in Fig. 2 is in accord with the expected ACR source spectrum used to model the observed modulated ACR flux within the heliosphere. 4. C O N C L U D I N G R E M A R K S MRI acceleration or shock surfing [7,8] can explain naturally the injection characteristics of pickup ions at interplanetary shocks [2,3] as well as the inferred injection efficiency at the termination shock for ACRs [4,5]. Weak scattering yields an anomalously low injection efficiency for I-I+ compared to He +, C +, N +, O +, and Ne +. Computed ACR termination shock and modulated fluxes compare well to those observed. Strong scattering eliminates the mass dependence of the ACR injection efficiency. See [ 10] for further details and discussion. REFERENCES IGosling, J.T., et al., J. Geophys. Res., 86, (1981) 547. 2Gloeclder, G., et al., J. Geophys. Res., 99, (1994) 17,637. 3Fr/tnz et al., Geophys. Res. Lett., 26, (1999) 17. 4Cummings, A.C., and Stone, E.C., Space Sci. Rev., 78, (1996) 117. 5jokipii, J.R., ApJ, 313, (1987) 842. 6Webb, G.M., Zank, G.P., Ko, C.M., & Donohue, D.J., ApJ, 453, (1995) 178. 7 Zank, G.P., Pauls, H.L., Cairns, I.H., & Webb, G., J. Geophys. Res., 101, (1996)457. 8Lee, M.A., Shapiro, V.D., & Sagdeev, R.Z., J. Geophys. Res., 101, (1996) 4777. 9Newbury, J.A., Russell, C.T., & Gedalin, M., J. Geophys. Res., 103, (1998) 29,581. l~ G.P., Rice, W.K.M., le Roux, J.A., Matthaeus, W.H., ApJ, (2000) in press. llVasyliunas, V.M., and Siscoe, G.L., J. Geophys. Res., 81, (1976) 1247. ~2Scherer, H. et al., J. Geophys. Res., 103, (1998) 2105.
-214-
Self-consistent acceleration of pickup ions at the termination s h o c k J. A. le Roux a, H. Fichtner b, G. P. Zank a, and V. S. Ptuskin c aBartol Research Institute, University of Delaware, Newark, DE 19716, USA bTheoretische Physik IV, Weltraum- und Astrophysik, Ruhr-Universit~it Bochum, Bochum, Germany Clnstitute for Terrestrial Magnetism, Ionosphere, and Radiowave Propagation (IZMIRAN), Troitsk Moscow District, Russia
Self-consistent simulations indicate that when low-energy pickup protons are preaccelerated locally at a nearly perpendicular solar wind termination shock (TS) by the multiply reflected ion (MRI) acceleration mechanism, this might lead to the injection of a sufficient number of pickup protons into standard diffusive shock acceleration to enable reproduction of the observed upstream anomalous cosmic ray (ACR) proton spectra. This is possible despite a significant mediation of the TS by the MRI-accelerated PUIs as well as ACRs that weakens MRI acceleration.
1. INTRODUCTION It is generally accepted that the ACR component is formed when interstellar pickup ions (PUIs) undergo standard diffusive shock acceleration at the nearly perpendicular solar wind TS. At a nearly perpendicular TS, however, a significant energy threshold needs to be overcome by the PUIs before the standard theory applies. It is not known what fraction of PUIs can overcome the threshold by virtue of preacceleration in the upstream solar wind [ 1] or what fraction are injected due to local preacceleration at the TS. Here we investigate the self-consistent, local preacceleration of unaccelerated pickup protons, their injection into, and acceleration by standard diffusive shock acceleration at the TS. By basing the preacceleration and injection of PUIs on MRI acceleration theory [2, 3], an improvement on previous ad-hoc approaches to injection, it is investigated how MRI-accelerated PUI and ACR protons mediate the TS structure, how the mediation affects the efficiency of the two acceleration mechanisms, and what consequences this has for upstream modulation of ACR protons.
2. THE MODEL
The model is time dependent and spherically symmetric with its parameters chosen to reflect conditions in the upwind, equatorial regions. To determine the TS modification by
-215-
J.A. le Roux et aL
MRIs and ACRs, a set of three-fluid equations is solved numerically whereby the fluids are the solar wind plus unaccelarated pickup, MRI-accelerated, and ACR protons with the latter two fluids taken to good approximation to have a low mass density. The fluid equations are closed by calculating the MRI-accelerated pickup proton spectrum and pressure for unaccelerated PUIs, and by solving numerically the standard ACR transport equation to determine the ACR spectrum and pressure. To determine the MRI spectrum and pressure, we first calculate the unaccelerated PUI spectrum at the TS from the analytical solution of the standard PUI transport equation (the standard ACR transport equation without diffusion but with an appropriate source term for the production of pickup protons due to charge exchange). MRI theory is applied to the PUI spectrum to determine the fraction of PUIs involved in MRI acceleration, their expected maximum energy, and the form of the MRI spectrum. The idea behind MRI acceleration is that PUIs with small velocity components normal to the TS are unable to penetrate the crossshock electric field. These particles skim along the TS surface while gaining energy from the motional electric field. Their maximum energy is determined when their Lorentz force overcomes the cross-shock electric field force. The MRI spectrum is updated in the model as the TS is mediated and the cross-shock electric field is weakened under the dynamic influence of the MRIs and ACRs.
\\
84,8
84,9
....
R
85,0
Radial distance (AU) Figure 1. Solar wind speed normalized to Uo = 400 km/s as a function of heliocentric distance in astronomical units (AU). The dashed curves denote the case where only MRI-accelerated pickup protons mediate the TS while the solid curves also include the modifying effects of diffusively TS accelerated ACR protons. The left panel provides a close-up view of the TS shown in the right panel. A portion of the MRI spectrum is diffusively shock accelerated at the TS provided particle speeds satisfy v > Vinj = 4au2(~lJ~• ]/2 where a' = 2, u2 is the downstream solar wind speed and ~,,, is the parallel, perpendicular diffusion coefficient [4]. The radial diffusion coefficient 1err is determined on the basis of quasi-linear theory for parallel diffusion, a theory for the random walk of field lines applied to perpendicular diffusion, and a transport theory for MHD turbulence in the solar wind [5]. For more details about the model see [6, 7].
-216-
Self-consistent acceleration of pickup ions at the termination shock 3. RESULTS The results indicate that both MRI-accelerated pickup and A CR protons might significantly mediate the TS. This is because a considerable fraction of the PUIs (--16%) are MRI accelerated of which a significant amount enters diffusive shock acceleration. Whereas the TS compression ratio is initially -3.1 due the presence of PUIs in the solar wind, MRIaccelerated PUI protons alone is found to weaken the TS compression ratio to s = 2 (dashed curve in left panel of Figure 1). Viewed on a larger scale (20 AU) appropriate for detecting TS modification by ACRs, the TS modification by MRI-accelerated PUIs becomes almost invisible (dashed curve in right panel of Figure 1). When including the effect of the ACR proton pressure on the TS, the compression ratio appears to be s --- 2.3 on this large-scale view of the TS as a large-scale TS precursor due the ACR protons is formed (solid curve in right panel of Figure 1). This gives an idea of the effect of the ACRs on the TS structure. A smallscale view of the TS structure (< 1 AU) shows that the combined effect of MRI and diffusive shock acceleration of PUI protons reduce the TS compression ratio to s = 1.6 (solid curve in Figure 1). The TS modification by MRI and ACR protons results in a weaker cross-shock electric field so that the maximum energy that MRI acceleration provides to PUIs is reduced significantly. Initially, the ~- 1o maximum energy that the MRI-accelerated lo~ pickup protons reach is -170 keV for a 18 1! ~ ~o o (sub)shock ramp with width e = 2 (width ,8 normalized to an electron inertial length) i5 o [8], but after modification of the TS by both 10: 10' MRI and ACR protons the maximum 10-a 10-7 10-0 10-s 10 4 10-3 10-2 10-1 10 0 101 10 2 energy is --74 keV. The maximum energy was further enhanced by a factor of--2 due ,~
0
Figure 2. Differential intensity in particles m -2 s -1 s r -1 MeV -] as a function of kinetic energy in GeV. The top solid curve labeled rsh depicts the simulated combined unaccelered PUI, MRI-accelerated PUI, and ACR proton spectrum downstream of the TS, while the modulated ACR proton spectra upstream are shown at heliocentric distances 1, 23, 42 and 65 AU (lower solid curves from bottom to top). The dashed curve above (below) the top solid curve is the PUI (MRI) spectrum upstream. The filled and open circles denote a reproduction of Voyager 1 data at --65 AU and Voyager 2 data observed a t - - 5 2 AU during 1997, respectively [ 13, 14].
to adiabatic compression of the MRIaccelerated PUIs when crossing the TS. The injection energy was calculated as -50 keV. This threshold allows for enough MRIaccelerated PUIs to be injected into standard diffusive shock acceleration so that the modulated ACR proton spectra upstream compare favorably with observational data recorded by the Voyager spacecraft (see Figure 2). Even though the injection occurs most of the time, it is highly time dependent and sometimes is interrupted as the TS becomes too strongly modified and the cutoff energy for MRI acceleration falls to far below the threshold. In addition, injection might also be interrupted when MRI acceleration itself
-217-
J.A. le Roux et al.
fails. This happens when the TS becomes so strongly mediated that reflection by the crossshock electric field ceases. This is expected too happen when the upstream Mach number of the TS drops to between 1 and 2 [9]. Calculations show that if it is assumed that reflection stops when this critical Mach number is less than 2, then injection becomes sporadic as reflection is interrupted most of the time. The results shown here are for an assumed value of 1.5 whereby reflection occurs continuously. Further simulations underscore the robustness of the model in that comparable results are yielded for e < 10. For all TS (sub)ramp widths of less than about an ion inertial length, the MRI acceleration cutoff energy is sufficiently high to allow injection into diffusive shock acceleration. However, intensities of the modulated ACR proton spectra upstream become too small for reproducing observations when the TS (sub)ramp width approaches an ion inertial length (weak MRI acceleration).
4. SUMMARY AND CONCLUSIONS
In summary, these self-consistent simulations indicate that low-energy pickup protons preaccelerated locally at a nearly perpendicular TS by the MRI mechanism might lead to the injection of a sufficient number of MRI-accelerated PUIs into diffusive shock acceleration for an injection threshold o f - 5 0 keV to enable reproduction of the observed upstream ACR proton spectra. This is possible despite significant mediation of the TS by MRI-accelerated PUIs as well ACRs formed by the diffusive shock acceleration of a fraction of the locally preaccelerated PUIs. For the same injection energy recent work shows that not enough PUIs preaccelerated in the upstream solar wind will be injected to explain ACR observations because the flux of preaccelerated PUIs fall strongly off as 1/r 3 at large heliocentric distances due to adiabatic cooling [10, 11]. Given modeling to the contrary [12], and the uncertainty in the injection efficiency, these conclusions still need further study. REFERENCES
Chalov S. V., & Fahr, H., J. 2000, Astron. Astrophys., 360, 381. Zank, G. P., Pauls, H. L., Cairns, I. H., & Webb, G. M. 1996, J. Geophys. Res., 101,457. Lee, M. A., Shapiro, V. D., & Sagdeev, R. Z. 1996 J. Geophys. Res., 101, 4777. Webb, G. M., Zank, G. P., Ko, C. M., & Donohue, D. J. 1995, Astrophys. J., 453, 178. Zank, G. P., et al. 1998, H., J. Geophys. Res., 103, 2085. le Roux, J. A., Fichtner, H., Zank, G. P. & Ptuskin, V. S. 2000, J. Geophys., Res., 105, 12557. le Roux, J. A., Zank, G. P. Fichtner, H., & Ptuskin, V. S. 2000, Geophys. Res. Lett., 27, 2873. Newbury, J. A., Russell, C. T., & Gedalin, M. 1998, J. Geophys. Res., 103, 29581. 9. Edmiston, J. P., & Kennel, C. F. 1984, J. Plasma Phys., 32, 429. 10. Decker, R. B., et al. 2001 this volume. 11. Rice, W. K. M., Zank, G. P., le Roux, J. A., & Matthaeus, W. H., 2001 Adv. Space Res., in press. 12. Giacalone, J., et al. 1997, Astrophys. J., 486, 471. 13. Cummings, A. C., & Stone, E. C., 1998, Space Sci. Rev., 83, 51. 14. McDonald, F. B., Space Sci. Rev., 83, 33, 1998. ~
2. 3. 4. 5. 6. .
~
-218-
Energetic Neutral Helium of Heliospheric Origin at 1 AU A. Shaw ~, K. C. Hsieh ~ *, M. Hilchenbach b, A. CzechowskF, D. Hovestadt a, B. Klecker a, R. Kallenbach ", E. M6bius f P. Bochslerg ~Department of Physics, University of Arizona, Tucson, AZ 85721, USA bMax-Planck-Institut fiir Aeronomie, D-37189 Katlenburg-Lindau, Germany r
for Space Research, Polish Academy of Sciences, PL-00716 Warsaw, Poland
dMax-Planck-Institut fiir Extraterrestrische Physik, D-85740 Garching, Germany eInternational Space Science Institute, CH-3012 Bern, Switzerland fEOS, University of New Hampshire, Durham, NH 03824, USA gPhysikalisches Institut der Universits Bern, CH-3012 Bern, Switzerland Using measurements from the HSTOF (High-Energy Suprathermal Time-of-Flight sensor) instrument on SOHO (Solar and Heliospheric Observatory) at 1AU, we report on a possible detection of Energetic Neutral Helium Atoms of probable heliospheric origin. Observations made under quiet interplanetary conditions showed an extra component in the helium spectrum between 85 and 141 keV, here we examine the nature of this component and the basis upon which we tentatively identify it as neutral. 1. I N T R O D U C T I O N A singly-charged energetic ion in a space plasma, such as a proton or He + of the anomalous cosmic ray (ACR) population, can become an energetic neutral atom (ENA) by charge-exchanging on an atom of the ambient neutral gas, e.g. H or He of the local interstellar medium (LISM). The resulting ENAs, unmodulated by magnetic fields, can reach regions of space normally inaccessible to the original charged population. This enables us to sample space plasmas in currently inaccessible regions; e.g., ACR in the outer heliosphere. The use of ENAs to probe the ACR population in the outer heliosphere was first discussed by Hsieh et al.[1]. The instrument HSTOF of the Charge, Element and Isotope Analysis System (CELIAS) on SOHO is the first instrument in interplanetary space capable of detecting ENAs, Hovestadt et al.[2], enabling Hilchenbach et al.[3] to measure the intensity of energetic hydrogen atoms (EHA) between 55 and 80 keV at 1 AU under quiet time (QT) interplanetary conditions for the first time. *The work at the University of Arizona is supported in part by NASA grant NAG5-7966 and NSF grant ATM9727080. K.C. Hsieh's travel is partly funded by the University of Arizona's Foreign Travel Grant.
-219-
A. S h a w e t al.
2.
SEARCHING
FOR
ENERGETIC
NEUTRAL
HELIUM
Figure 2. He + and He ++ extraction
Figure 1. HSTOF data 1996 t o 1999
The TOF vs. E' analysis of HSTOF separates the elements into tracks according to mass (Fig. 1). It also effectively reduces noise in the data, requiring each genuine particle to trigger all three detectors: Start, Stop, and PSSD (Pixelated Solid State Detector). These signals from a real event fall in sequence within extremely short time intervals. Noise due to accidental coincidence can be estimated and removed by examining events registered in a region of the TOF vs. E ~plane that do not correspond to genuine particles. Helium exists in three charge states: He ~ (neutral), He +, and He ++. The parallelplate E / Q filter of HSTOF has no effect on He ~ but excludes low energy ions. Fig. 2 displays the energy profile of He + and He ++ during a solar-related event. We use eight such events, during which no significant He ~ flux is expected, to show HSTOF's distinct response to He + and He ++ (Fig. 3). This instrument response set 85 < E < 141 keV as the best energy range for the detection of energetic neutral helium (ENHe), also providing a means to estimate and remove the probable background in that energy range, caused by He + and He ++ during both quiet and non-quiet interplanetary conditions. The periods of quiet interplanetary conditions used by Hilchenbach et al.[3] have been extended from 1997 to late 1999. The total helium flux as a function of kinetic energy, E, accumulated over all these periods - 506 days along the SOHO orbit (1 AU) from 1996 to 1999 - is shown in Fig. 4. The qualitative difference between the energy profiles of the two distinct populations of helium shown in Fig. 2 and Fig. 4 seems to indicate the presence of ENHe in the assigned energy range, 85-141 keV, during the QT periods (~ 100 counts). More analysis is however required to completely rule out possible contamination. We shall now try to determine the direction of arrival of these assumed ENHe.
- 220-
Energetic neutral helium of heliospheric origin at 1 A U
Figure 4. HSTOF Quiet Time (QT) Flux
Figure 3. HSTOF He + and He ++ response
Figure 5. HSTOF Look Direction
3. A R R I V A L D I R E C T I O N
Figure 6. Helium Cone[6]
OF ENHe FLUX
HSTOF has a bore-sight lying in the ecliptic plane, 37 ~ west of the SOHO-sun line (Fig. 5). The field-of-view of HSTOF covers 2 ~ either side of the bore-sight in the ecliptic plane, and +17 ~ normal to the ecliptic. Since SOHO moves with the Earth-Sun L1 Lagrangian point, it's daily location can be referenced by Earth's Day of Year (DOY). The direction of the Sun's motion with respect to the LISM is 254 ~ this is the Apex (upwind) direction. HSTOF looks towards the apex around DOY 12, and in the opposite direction (Anti-Apex or downwind) towards the heliotail around DOY 195. The QT ENHe flux as a function of DOY is shown in Fig. 7. Each data point represents 10 days of data. The flux level is higher and more scattered in 1996 than in 1997. The 1997 data suggests a peak ENHe flux at around DOY 200, possibly originating from interactions between He + accelerated in co-rotating interaction regions (CIRs) and the
-221 -
A. Shaw et al.
O:
7
ii:, -
t !: t
,
ii
Figure 7. HSTOF Quiet Time Flux vs. Time
LISM helium cone (Fig. 6), as is the case for EHA[4]. 4. C O N C L U S I O N Though the final absolute helium flux calibration is pending, data from HSTOF/CELIAS/SOHO suggests the first detection of energetic neutral helium (ENHe) in interplanetary space. The arrival direction of the peak flux in 1997 suggests an origin in the interaction between CIRs and the LISM helium cone. If the observed flux is confirmed as ENHe then careful comparison of our results with the observed EHA flux[3], models of EHA production from anomalous cosmic rays (ACR)[5] and CIR accelerated ions[4] must be performed before any definitive conclusions as to the origin of this flux can be established. REFERENCES
1. 2. 3. 4. 5. 6.
K.C. Hsieh, K.L. Shih, K.R. Jokipii, and S. Grzedzielski, Astophys. J. 393 (1992) 756. D. Hovestadt et al., Solar Phys. 162 (1995) 441. M. Hilchenbach, K.C. Hsieh, D. Hovestadt et al., Astophys. J. 503 (1998) 916. J. Kota, et al. submitted J. Geophys. Res. (2000) A. Czechowski et al. submitted Astron. & Astrophys. (2000) E. M6bius, SSRv. 78 (1996) 375.
- 222 -
Effects of the heliospheric termination shock on possible local interstellar spectra for cosmic ray electrons and the associated heliospheric m o d u l a t i o n S. E. S. Ferreira, M. S. Potgieter and U. W. Langner School of Physics, Potchefstroom University for CHE, 2520 Potchefstroom. South Africa
The 'local' interstellar spectrum (IS) for cosmic ray electrons at energies of interest to heliospheric modulation studies is still basically unknown. Recently, new computations of the IS based on advanced modeling of cosmic-ray propagation in the Galaxy [1], and observations including diffuse galactic gamma rays, indicated that the electron IS may be considerably lower at energies below-100 MeV than previously assumed. For this work different scenarios for the electron 'local' IS, and their subsequent modulation in the heliosphere, are studied using a shock-drift modulation model. The effects of the heliospheric termination shock (TS) on each of these scenarios are illustrated, together with the subsequent effects on their modulation in the heliosphere. We find that the computed effect of the TS on galactic electron intensities at 16 MeV is relatively small, in general, but more pronounced if the TS is positioned at 80 AU, than at 90 or 100 AU. The larger the 'local' IS value is, the larger the effect of the TS on electron modulation at this energy becomes. 1. I N T R O D U C T I O N The study of the modulation of cosmic ray electrons in the heliosphere is an important and useful tool in understanding various aspects of heliospheric modulation. Modulated electron intensities in the lower-MeV range give a direct indication of the average parallel and perpendicular mean free paths in contrast to protons that experience adiabatic energy changes below-300 MeV (e.g. [2]). Gradient and curvature drifts become less important for electron modulation at lower energies, with almost no effect below 100 MeV (e.g. [3]). The Pioneer 10 radial-intensity-profiles f o r - 1 6 MeV electrons ([4],[6]) indicate almost no radial gradients out to -70 AU, which put serious constraints on the diffusion tensor. New computations of the interstellar spectra (IS) [1], indicate that the electron IS may be considerably lower at energies below -100 MeV than previously assumed (e.g. [11]). For this work, two different (extreme) scenarios for the 'local' IS for cosmic ray electrons ([1],[5]), and their subsequent modulation in the heliosphere are studied using a shock-drift-modulation model. Satisfying the constraints imposed on the diffusion tensor by the Pioneer 10 electron data in the outer heliosphere, the effects of the location of the heliospheric termination shock (TS) on each of the IS scenarios are illustrated, together with the subsequent effects on their modulation. 2. M O D U L A T I O N
MODEL AND PARAMETERS
The model is based on the numerical solution of the transport equation (TPE) [7]:
- 223 -
S.E.S. Ferreira, M.S. Potgieter and U W. Langner
3fot = - ( V
I(V.V) + (v i) ) ) . V f + V . ( K s . V f ) + ~-
3f +J 3 ln-~
,
(1)
source
where f (r~,t) is the CR distribution function; R is rigidity, r is position, and t is time, with V the solar wind velocity. Terms on the right-hand side represent convection, gradient and curvature drifts, diffusion, adiabatic energy changes and a source function, respectively. The symmetric part of the tensor Ks consists of a parallel diffusion coefficient (Kll) and a perpendicular diffusion coefficient (K• The anti-symmetric element KA describes gradient and curvature drifts in the large scale HMF. J~our~e can be any source, e.g. Jovian electrons, pick-up ions, and/or the local interstellar spectra (IS). Here, we concentrate on the IS for electrons, neglecting all other local sources. The TPE was solved using a two dimensional TSdrift model developed by le Roux et al. [8], and expanded by Haasbroek [9]. The outer modulation boundary is assumed at 120 AU. A TS with a compression ratio of 3.2 < s < 4.0, and scale length of L = 1.2 AU was assumed at rs = 80 AU. The solar wind speed V was assumed to change from 400 km.s 1 in the equatorial plane (0 = 90 ~ to a maximum of 800 km.s ~ when 0 < 60 ~ At the shock, V decreases from the upstream value of V1 = 400 km.s -1 in the equatorial plane according to the relationship given by le Roux et al. [8]: V(r)= VI(S+ 1)_ V I ( S - 1 ) t a n h ( r - r~'] 2s 2s [, L )
(2)
For the diffusion coefficients that describe diffusion parallel, Kjl, and perpendicular, K• to the average heliospheric magnetic field (HMF) as well as the asymmetric coefficient KA, which describes gradient and curvature drifts in the background HMF, we assumed: Be", Kii = K ofif ( R ) --ff-
K_l_r = a 1 + (k,KI]ll/rg
" )2 '
K 2.0
Kll )5," = b 1 + (All/rg
/3n KA = (KA)o ~
(3)
Here/3 is the ratio of the speed of the cosmic ray particles to the speed of light; f(R) gives the rigidity dependence (in GV) with f(R)=R when R > 0.4 GV, and f(R)=0.4 when R ~ 0.4 GV; B is the HMF magnitude modified [12] in qualitatively agreement with Ulysses observations [13]; K0 is a constant in units of 6.0 x l 0 2~ c m 2 S"1", a is a constant which determines the value of K• which contributes to perpendicular diffusion in the radial direction, and b is a constant that determines K• which contributes to perpendicular diffusion in the polar direction. Diffusion perpendicular to the HMF is therefore enhanced in the polar direction by assuming b > a. ([3],[10],[11]). The ratio of the scattering mean free path to the particle gyroradius, )~ll /rg, is larger than unity in view of the Bohm limit, ~'11= rg. The TPE was solved in a spherical coordinate system with the current sheet "tilt angle" (x = 10 ~ for so-called A > 0 epochs (-1990 to present) when electrons primarily drift inward through the equatorial regions of the heliosphere. 3. R E S U L T S
AND DISCUSSION
Figure l a shows the radial profiles of computed 16 MeV galactic electron intensities with an outer boundary at rB -- 120 AU, with the IS of Strong et al. [5] assumed to be the local electron spectrum. Three different radial profiles are shown; the solid line corresponding to a TS at rs = 80 AU, the dashed line to a TS at rs = 90 AU, and the dotted line to a TS at rs = 100 AU. The electron data from Pioneer 10 are presented as shaded areas for radial distances up to
- 224 -
Effects o f the heliospheric termination shock on ...
Figure l a" Computed radial profiles of 16 MeV galactic electron intensities with an outer boundary at rB = 120 AU, and with the IS of Strong et al.[5] assumed to be electron local spectra. The solid line corresponds to a TS at rs = 80 AU, the dashed line to a TS at rs = 90 AU, and the dotted line to a TS at rs - 100 AU. The electron data from Pioneer 10 are presented as shaded areas, lb" The same as in la, but with the IS from the recent calculations of Strong et al. [1]. --50 AU [4], and at --70 AU [6]. The height of the shaded areas incorporate the error bars present in the data, as well as a small time dependent effect due to the solar modulation of the electrons over the period --1972 to --1991. The data are assumed to be of galactic origin, with the Jovian component dominating only for r < 25 AU [4]. The computed intensities are compatible with both data sets. We assumed Ko = 0.5, a - 0.25 and b = 0.6. The effect of the shock is visible in all the radial profiles, indicating an increase in the radial gradient upstream of the shock but a decrease beyond the shock. For a TS at rs -- 80 AU, the effect of the shock on the radial gradients upstream and downstream of the shock is significantly larger than for the two other scenarios. Although the shock is more effective for larger radial distances due to the larger shock radius, the radial diffusion coefficient, due to its radial dependence oc r, is also larger, leading to a smaller effect for the more distant shocks. Figure l b shows the case when the more recently calculated IS of Strong et al. [1] is assumed as the local electron spectrum. Evidently, it is considerably lower than the IS used in Figure la. All three scenarios for the shock locations are again compatible with the observed Pioneer 10 data, but in order to obtain that, we had to assume Ko = 58, a=18 and b=24 in Equation 3. These larger diffusion coefficients are needed in order to produce less modulation between the outer boundary and the data a t - 7 0 AU. The effect of the different TS locations on the computed electron intensity profiles is much less pronounced, with almost no difference between the three scenarios, in contrast to Figure la. These larger diffusion coefficients clearly decrease the effectiveness of the shock.
- 225 -
S.E.S. Ferreira, M.S. Potgieter and U.W. Langner
4. SUMMARY AND CONCLUSIONS Two different scenarios for the electron IS ([1],[5]), and their subsequent modulation in the heliosphere have been studied using a shock-drift modulation model. The effects of the heliospheric termination shock (TS) on each of these scenarios were also studied. At 16 MeV the two electron IS differ by a factor of--100. They were assumed as the local IS at 120 AU, as an outer boundary of the heliosphere. Compatibility between the observed Pioneer 10 ([4],[6]) radial profiles and the model simulations were strictly required for the two IS cases, and the three different TS shock positions, as shown in Figure 1. For the highest IS of Strong et al. [5], very large radial gradients (-~10%/AU) were found between 70 AU, the shock position rs, and the outer boundary. The effect of the shock on the radial intensity profile is more pronounced with rs = 80 AU, than for rs = 90 AU or 100 AU. The radial gradients are found to be larger upstream of the shock than downstream. Although the TS should more effective for larger radial distances due to the larger shock radius, the diffusion coefficients are also larger, leading to a smaller modulation effect for the more distant TS positions. For the lowest IS of Strong et al. [1], the radial gradients beyond 70 AU were significantly less, with the effect of the different TS locations on the computed electron intensities far less pronounced than in the first case. The reason is that larger diffusion coefficients are needed to provide compatibility with the data. These larger diffusion coefficients not only decrease the total modulation in the outer heliosphere, but also the effectiveness of the shock. For the interstellar spectra given in Figure 1, the computed radial intensity profiles for--16 MeV electrons indicate that the detection of the crossing of the TS by a spacecraft registering only these low energy electrons would be ambiguous. Some other observations have to be utilized (e.g. magnetic field or solar wind speed changes) in order to provide information on when a spacecraft actually crosses the TS. REFERENCES ~
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.
A.W. Strong et al., Astrophys. J. 537 (2000) 763. L. J. Haasbroek et al., Proc. 24th ICRC 4 (1995) 706. S. E. S. Ferreira et al., J. Geophys. Res. 105 (2000) 18305 C. Lopate, Proc. 22nd ICRC (Dublin) 2 (1991) 149. A. W. Strong et al., A&A 292 (1994) 82. C. Lopate, private communication (2000). E. N. Parker, Planet. & Space Sci. 13 (1965) 9. J. A. le Roux et al., J. Geophys. Res., 101 (1996) 4791. L. J. Haasbroek, Ph.D. thesis, Potchefstroom University, South Africa (1997). J. K6ta and J.R. Jokipii, Proc. 24th ICRC (Rome) 4 (1995) 680. M. S. Potgieter, J. Geophys. Res. 101 (1996) 24411. J. R. Jokipii and J. K6ta, Geophys. Res. Lett. 16 (1989) 1. A. Balogh et al., Science. 268 (1995) 1007.
- 226-
Oral papers and posters
G A L A C T I C C O S M I C RAYS: O V E R V I E W M.A. Forman Department of Physics and Astronomy, State University of New York at Stony Brook, USA. Cosmic rays from the galaxy bring us clues about energetic processes in stars and in the interstellar medium in our galaxy. They are also one of the important probes of the magnetic conditions all over the heliosphere, and its connection to the interstellar medium. They have been there. Over 50 years ago, before the space age, cosmic ray detectors on Earth revealed that the cosmic rays were positively charged and had energies from at least 5 10 8 electron volts to over 10 18 . 11-year solar modulation of cosmic rays (with energies less than about 10 10 eV) was discovered, and more types of cosmic ray decrease also associated with increased solar activity on time scales of the solar cycle, the solar rotation and after transient solar events. Clearly the Sun was doing some powerful electromagnetic work out there, all the time, in a huge volume of space extending at least to the orbit of Earth, and most likely well beyond. Heliospheric cosmic ray research was one of the first scientific fields to benefit from access to space, because of its obvious need to be in space and because its visionary early experimentalists made rugged instruments ready. We have a pretty good idea of what causes the modulation: large scale moving magnetic fields and cosmic-ray coupling to the ambient plasma through magnetic turbulence. We have a good idea of the distribution of cosmic rays inside 100 AU. We should remember that galactic cosmic rays have already visited the furthest parts of the heliosphere, on all possible routes and tell us about the heliosphere and beyond.
G A L A C T I C C O S M I C RAYS: T H E O U T E R H E L I O S P H E R E J. R. Jokipii University of Arizona, Tucson, AZ 85721 USA.
However, we are still not at the point where we can reliably extrapolate to the interstellar spectrum at energies less than several hundred MeV. A significant problem is our lack of understanding of the outer heliosphere and the interface with the interstellar gas. The use of cosmic rays as remote probes of this region is very important. The current status of our understanding and the results of recent modelling efforts will be discussed, with emphasis on processes ocurring in the outer heliosphere.
- 227 -
Oral papers and posters
MODULATION RAYS
OF G A L A C T I C
AND ANOMALOUS
COSMIC
E. R. C h r i s t i a n (1), W. R. Binns (2), C. M. S. Cohen (3), A. C. Cummings (3), . S. George (3), P. L. Hink (4), R. A. Leske (5), R. A. Mewaldt (5), E. C. Stone (5), T. T. von Rosenvinge (6), M. E. Wiedenbeck (7) and N. Yanasak (7) (1) NASA GSFC Code 661, (2) Washington U., St. Louis, (3) Caltech, (4) Washington U., St. Louis, (5) Caltech, (6) NASA GSFC Code 661, (7) PL. The temporal history of cosmic ray intensities at 1 AU is an important component to the understanding of solar modulation. The large collecting power and high resolution of the Cosmic Ray Isotope Spectrometer (CRIS) and the Solar Isotope Spectrometer (SIS) instruments on the Advanced Composition Explorer (ACE) allow us to investigate the changing modulation on short time scales and over a wide range of rigidities. With these data, we will present the di erences between the short term and long term e ects and the correlation of these e ects with magnetic field, current sheet tilt angle, and other phenomena. The data span the period from the launch of ACE in August 1997 to the present.
T H E S P E C T R U M O F A C R O X Y G E N A N D ITS V A R I A T I O N S I N T H E O U T E R H E L I O S P H E R E F R O M 1992 T O 2000 D.C. H a m i l t o n (1), M.E. Hill (1), N.P. Cramer (1), R.B. Decker (2) and S.M. Krimigis (2) (1) Department of Physics, University of Maryland, (2) Johns Hopkins Applied Physics Laboratory. The Voyager 1/2 LECP instruments are measuring the anomalous cosmic ray oxygen spectrum (0.3 - 40 MeV/nuc) in the outer heliosphere and have observed large intensity variations, particularly in the low energy portion of the spectrum. At Voyager 1, the peak ACR oxygen flux was observed at a nearly constant energy of 1.3 MeV/nucleon during the years 1993 to 1999 as V1 traveled from 53 AU to 76 AU. The flux at that energy increased by a factor of about 100 from 1992 to 1999, reflecting a large decline in modulation. In fact the ACR flux at V1 and V2 continued to increase through 1997, 1998, and into 1999, well after the nominal minimum of the 11-year solar activity cycle in 1996. At Voyager 2 (at 60 AU near the beginning of 2000) it appears that the oxygen flux near the 1.3 MeV/nucleon ACR peak has begun to decrease, declining by about a factor of three from its maximum in mid-1999 to February 2000. At Voyager 1 the picture is less clear. Although the flux has decreased since September 1999, this decrease appears to be in phase with a series of quasi-periodic variations observed over the last three years. These variations at V1 have a period of about one third year and an amplitude of about 50%. It will soon become clear whether this latest downturn is part of that sequence or is the beginning of an increase in long-term modulation. Differences in ACR behavior at V1 and V2 are interesting because V1, at 34 degrees north heliolatitude, is beyond the current sheet while V2, at 21 degrees south, is still within but near the limit of current sheet excursions.
-228-
Oral papers and posters
O N T H E V A R I A B I L I T Y OF S U P R A T H E R M A L 1 AU
P I C K U P H E + AT
B. Klecker (1), A.T. Bogdanov (1), M. Hilchenbach (2), A.T. Galvin (3), E. MSbius (4), F.M. Ipavich (4) and P. Bochsler (5) (1) Max-Planck-Institut fr extraterrestrische Physik, D-85740 Garching, Germany, (2) Max-Planck-Institut fr Aeronomie, D-37819 Katlenburg-Lindau, Germany, (3) University of Maryland, College Park, Md, USA, (4) University of New Hampshire, Durham, NH, USA, (5) University of Bern, CH-3012 Bern, Switzerland. Using data from the STOF experiment onboard SOHO we investigate the variation of suprathermal He+/He 2+ abundances in the energy range 85-280 keV during the years 1997 to 1999. We observe a large variability of the He + abundances ranging from He+/He 2+ < 5% to ,.~ 1. The very large abundances are closely related with the passage of interplanetary shocks. Combining the data from STICS/WIND and STOF/SOHO we are able to identify in these events a pickup He + distribution with the typical cutoff energy at twice the solar wind velocity and a suprathermal tail extending to a few 100 keV. We correlate daily averages of the He + abundances of the suprathermal tail for all days with significant He + flux with solar wind parameters and find a general anticorrelation of He + abundances with solar wind velocity and solar wind thermal velocity. We discuss possible causes of this variability, in particular variations of the source strength of pickup ions and solar wind alphas and variations of the acceleration efficiency for He + and He 2+.
OBSERVATIONS OF PICK-UP IONS IN THE OUTER HELIOSPHERE BY VOYAGERS 1 AND 2 AND IMPLICATIONS ON PRESSURE BALANCE S.M. Krimigis and R.B. Decker Applied Physics Laboratory, Johns Hopkins University, Laurel, MD 20723. Observations from the Low Energy Charged Particle (LECP) instrument on the Voyager 1 (V1) spacecraft at NTOAU during the last solar minimum revealed an excess of counts from the sunward direction in the nominal proton energy range ~40 to ~ 140keV. This, above background, response was also seen by Voyager 2 at ~55AU and was present during the previous solar minimum at V1 at ~28AU. The angular distribution is inconsistent with these counts being due to hot protons convected into the sunward-viewing sector only. We have examined possible sources for the observed counts and have considered that they may be due to singly-ionized heavy (A_>16) ions picked up by the solar wind and convected into the sunward sector of the detector. The spectrum is steep (dj/dE c< E-6), and exhibits a cutoff at ,-~25keV/nuc. The intensities are within range of predictions of stochastic pre-acceleration or phase-space diffusion of interstellar pickup ions in the solar wind (e.g. Le Roux and Ptuskin, 1998; Chalov and Fahr, 1998). Preliminary estimates show that pickup oxygen pressure is c( 10-13 dynes cm -2 at ~70AU, comparable to the extrapolated pickup H+ pressure at 80AU (Whang et al, 1999), at the putative location of the termination shock. The observations suggest that pickup oxygen is present beyond 5AU and, along with other interstellar species, could be the pre-accelerated seed population for anomalous cosmic rays (ACR). LeRoux and Ptuskin, JGR, 103, 4799, 1998. Chalov and Fahr, Astron. Astrophys., 335, 746, 1998. Whang et al, JGR, 104, 28255, 1999.
- 229-
Oral papers and posters
CONSEQUENCES OF RECENTLY OBSERVED GALACTIC SYNCHROTRON RADIO EMISSIONS ON THE LOCAL INTERSTELLAR SPECTRUM FOR COSMIC RAY ELECTRONS U.W. Langner, O.C. de Jager and M.S. P o t g i e t e r School of Physics, Potchefstroom University, Potchefstroom. South Africa. Propagation models for galactic electrons give a synchrotron spectral index which is larger than the recently measured radio index between 22 to 408 MHz. Diffuse gamma-ray data appear to be "contaminated" by Crab-like point sources, so that it is difficult to derive a consistent local interstellar spectrum (IS) for electrons in the 1 to 30 MeV range. Using a phenomological approach, we show that the synchrotron spectral indices calculated from the best-fit IS of Strong and Moskalenko (Proc. 5th Comp. Symp., 1999) - for a 3 to 5 mG field agree well with the spectral indices calculated from their full propagation model in the frequency range of interest. This allowed us to introduce an adjusted local IS, such that the model radio spectral index agrees with observations above 20 MHz. By adding the constraints expected from galactic modulation, we find that the local IS at ,.~4 MeV is marginally above the lower limit for a local IS set by the Pioneer 10 electron data observed in the outer heliosphere
V A R I A T I O N O F T H E F L U X E S O F E N E R G E T I C H E + A N D H E 2+ DURING THE PASSAGE OF CO-ROTATING INTERACTION REGIONS D. Morris (1), E. MSbius (1), M.A. Popecki (1), L.M. Kistler (1), A.B. Galvin (1), B. Klecker (2) and A. Bogdanov (2) (1) (Dept. of Physics and Space Science Center, University of New Hampshire, Durham, NH), (2) (Max-Planck-Institut fr extraterrestrische Physik, Postfach 1603, D-85740 Garching, Germany). With the ACE SEPICA instrument detailed observations of the variation of the energetic He + and He 2+ fluxes have been obtained during the passage of several co-rotating interaction regions (CIR) in 1999 and 2000. For all CIRs under investigation the He+/He 2+ ratio increases consistently from the start of the event towards the end. However, the absolute flux of the energetic ions usually reaches a maximum close to the beginning of the event. Due to the solar rotation and related motion of the spacecraft relative to the CIR, the spacecraft is magnetically connected to the compression region and related shocks between the two different solar wind streams at distances from the sun, which increase with time. Therefore, the increasing He +/He 2+ ratio can be interpreted as an increase of the relative importance of the interstellar gas over the solar wind as a source for the observed energetic ions. The fact that the maximum flux of the energetic ions is observed close to the beginning of the event, supports the recent finding by Chotoo et al. (2000) that efficient particle acceleration also occurs in a region, where no shock has formed yet. Chotoo, K., et al., J. Geophys. Res., in press, 2000.
-230-
Oral papers and posters
THE COSMIC RAY ELECTRON TO POSITRON RATIOS IN THE HELIOSPHERE M.S. P o t g i e t e r and U.W. Langner School of Physics, Potchefstroom University, Potchefstroom. South Africa. The heliospheric modulation of cosmic rays disguises the true spectral form of the local interstellar spectra (LIS) of all cosmic ray species below about 10 GeV. The lower the energy, the more uncertain they seem to become which is especially true for cosmic ray electrons. Recent modeling of the propagation of cosmic rays through the Galaxy by Moskalenko and Strong (Ap. 493, 694, 1998) indicates that the LIS for positrons may be known more reliably than before. Using this information, and the di erent scenarios for the electron LIS, the electron to positron ratios, as modulated through the heliosphere, are computed with a comprehensive numerical drift model. These results can be of use for future missions to the outer heliosphere and beyond, and may assist in establishing the true electron LIS.
-231
-
This Page Intentionally Left Blank
General Discussion Forman to Moraal: I would like to hear some more ideas about the low energy cosmic ray spectrum, what you might learn from that and what it might be. And also the anomalous cosmic rays - do they have higher intensity than the low energy cosmic rays or not? Moraal: As far as the low energies are concerned, if you believe anything about the acceleration mechanism, then you cannot make a distinction between anomalous and galactic cosmic r a y s - anything at low energy should sort of go up like p-4, whether you should call it anomalous or galactic cannot be said. Fahr to Jokipii: What could bc thc rcason, in your vicw, for thc filamcntcd and vcry much disordered structure of the magnetic fields downstream of the heliospheric shock? Jokipii: Remember that the whole structure of the inner heliosphere now is quite consistent with the picture of braided magnetic fields caused by the supergranular motion on the surface of the Sun. If Len Fisk's very interesting field is there, it would also do that and, also, reconnection can occur. Those fluctuations grow in relative importance with radius because of the transverse expansion. And you can show from the number that we think of pretty firm that by the time you get to 100 AU the field line has meandered essentially over 7r radians. And this gets carried out into the heliosheath. Lee to Dorman: How is the distance to the modulation boundary determined? You need to know the diffusion coefficients. Dorman: No, we used only the average velocity. Gruntman to Moskalenko: What is the origin of these anti-protons? Moskalenko: We calculated the anti-proton flux from the interaction of galactic cosmic rays with interstellar gas. So, it's galactic origin on a large scale, the scale of the Galaxy. Marsch: Maybe a general question concerning the diffusion picture. The impression I got from Randy Jokipii's talk is that the Parker equation essentially describes it all. But is this diffusion picture really describing everything that one should look at? I think that there is a lot of convected structure in the flow on mesoscales where usually the features are not easily describable in terms of fluctuations and waves. I am not talking about the convection by the wind in general, there is structure in the wind itself, all sorts of transient features. Lerche: Maybe Gene [Parker] wants to talk about his own equation? Parker: I quite agree [to the existence of convected structures]. That's why there is a convective term in the equation. Also the velocities of structures in the solar wind are in the equation. Jokipii: The real question is the scale of the variation. As soon as you are talking about structures like corotating interaction regions which have scales greater than a fraction of an AU, than you can describe them very well with this transport equation. If we get to very small scales and long mean-free-paths near the S u n - to solar flare p a r t i c l e s - then you have to be a little bit more careful. But for the galactic and anomalous cosmic rays there is no real difficulty with using the equation. And the convcctcd structures arc all
- 233 -
General Discussion
in there. Veselowsky: From the discussion we understand that the equations are sound and reliable. Also the inner boundary conditions near the Sun are under control. But we understand that this is different for the external boundary conditions at the so-called modulation boundary, this mysterious surface. And what we need there is a real understanding of the space-time structure of the LISM. Cummings: Yes, and one question that comes to my mind is: how far do we need to go? Based on what I saw today there seems to be different opinions about how far out you need to get to the undisturbed interstellar medium. Forman: A general question regarding the modelling: Why didn't many observers adopt more sophisticated models including drifts but are rather using simple ones like force field or spherically-symmetric models? Doesn't it make any difference? Wiedenbeck: I think the observers are simply not up to speed on any of these sophisticated codes. I am not sure that they have been put into a form where one can readily use them for interpreting data. Lee: It's just puzzling here, that that we don't see any changes in the nature of modulation. I would have thought that as we get to the termination shock it changes dramatically because the flow beyond the shock is divergence-free. And so, a force-field approximation doesn't apply there. Lerche: Well, or the shock is much further out than you think it is. Veselowsky: The region you are talking about is very complicated in structure because it's keeping the memory of previous solar events. Giacalone to le Roux: A shock would have a foot and a ramp, perhaps an overshoot and some steady oscillations downstream. You discussed a fine structure of the termination shock, and I am not sure what you mean by that? le Roux: From bow shock observations, where they have really perpendicular cases [comparable to the termination shock], it looks that the shock ramp has details. Although the ramp thickness was found to be equal to about an ion inertial length, there are fluctuations insidc thc ramp of much narrowcr lcngth providing a finc structurc. Zank to Czechowski: Is there a sensitivity of the flux of anomalous cosmic rays to the actual structure of the assumed heliosphere? Czechowski: Yes, it is quite sensitive to it.
- 234-
Session 4:
Echoes from the Heliopause
This Page Intentionally Left Blank
Energetic Neutral A t o m I m a g i n g o f the Outer H e l i o s p h e r e - L I S M Interaction Region H.O. Funsten a, D.J. McComas b, and M. Gruntman c a Los Alamos National Laboratory, MS D466, Los Alamos, NM 87545, USA* b Southwest Research Institute, P.O. Drawer 28510, San Antonio, TX 78228, USA c University of Southern California, Los Angeles, CA 90089-1191, USA
Energetic neutral atom (ENA) emission from the region where the heliosphere and the local interstellar medium (LISM) interact will enable remote observation of this region's structure and dynamics. Imaging of these ENAs will allow us to understand the complex physics of the interaction and will help distinguish between competing models of the region. In this paper, we describe the imaging technique of ENA ionization by an ultrathin transmission foil followed by electrostatic deflection and coincidence detection. We also show new laboratory results that demonstrate the ability of this technique to detect ENAs over the energy range of approximately 0.2-6 keV that is critical for understanding the physics of the processes that govern the interaction region. This technique has a large intrinsic geometric factor that will enable imaging of the extremely dim ENA emissions and will allow quantification of the important characteristics of the outer heliosphere-LISM interaction region.
1. INTRODUCTION The interaction of the heliosphere with the local interstellar medium (LISM) is believed to generate energetic neutral atoms (ENAs) that result from charge exchange of hot plasma ions, predominantly hydrogen, with cold interstellar neutrals [1-4]. Once created, these neutral atoms follow ballistic trajectories from the source region and can be remotely detected within the inner heliosphere. These ENAs carry important information about processes occurring at their source region. Models of the interaction region (see [4] and references therein) predict predominant ENA emission at several hundred eV, although at very low fluxes. With increasing energy, the ENA flux generally decreases at a rate that is strongly dependent on the LISM parameters and the physics of the interaction. Imaging of ENAs with an energy Eo greater than-200 eV will provide a sensitive measure for determining the physics of the interaction region, including
* This work was performed under the auspices of the United States Department of Energy. The authors gratefully thank Ed Roelof (APL/JHU) and Hans Fahr (U. Bonn) for numerous insightful discussions on ENA emission from the outer heliosphere.
- 237-
H.O. Funsten, D.J. McComas and M. Gruntman
the type of shock and the evolution of pickup ions, and will distinguish between the various competing models. The primary challenge of ENA imaging is accurate measurement of ENA trajectory and energy against large EUV/UV and charged particle backgrounds. Several ENA imaging techniques covering different energy ranges have been developed to overcome this challenge [5,6]. On the recently successful IMAGE mission, three different imaging techniques were used to measure the ENA emission [7] from the terrestrial magneto-sphere at high time resolution across three different energy ranges: the High Energy Neutral Atom (HENA) Imager utilized a thick foil to block UV, allowing ENAs over 15 keV to pass to the detector section [8]; the Medium Energy Neutral Atom (MENA) Imager utilized freestanding transmission gratings to block UV while allowing a fraction of ENAs to pass to the detector section [9,10]; and the Low Energy Neutral Atom (LENA) Imager employed a reflection surface that ionized a fraction of ENAs followed by electrostatic deflection of ionized ENAs into the detector section [ 11]. These instruments have demonstrated the powerful technique of ENA imaging for viewing global morphology and dynamics of the magnetosphere. However, due primarily to the scientific thrust of the IMAGE mission on the terrestrial magnetosphere, these imagers were not designed to focus on the energy range of several hundered eV to several keV that is critical for imaging of the heliosphere-LISM interaction region. The ENA imagers on IMAGE were designed to detect the comparatively copious ENA fluxes from the magnetosphere at high time resolution and would need to be substantially larger to measure the extremely dim ENA emission from the heliosphere-LISM interaction region. In this paper, we investigate the feasibility of detecting ENAs that are emitted from the heliosphere-LISM interaction region over an energy range of 0.2 to 6 keV. The technique described here of ENA ionization via transmission through an ultrathin foil has certain advantages over other techniques at this energy range, including a geometric factor that is large enough to measure the extremely small ENA flux from the heliosphere-LISM interaction region. This technique has been investigated extensively at somewhat higher energies [ 12-15] but not below several keV.
2. NEUTRAL ATOM IMAGER FOR VIEWING THE INTERACTION REGION The measurement objectives for imaging ENAs from the heliosphere-LISM interaction region include: a large enough geometric factor for statistically significant measurements in one day, an energy resolution AE/E < 1, an imaging resolution in the range of 5o to 10~ and extremely high signal-to-noise ratio for clear identification of ENAs. These objectives are met with a single-pixel imager based on the technique of ENA ionization using an ultrathin foil [12,15]. An imager of this type is shown schematically in Fig. 1. An ENA enters the first set of collimators that consists of a set of alternately biased plates to reject solar wind ions up to an energy-per-charge of several 10s of keV/e from the ENA measurement. ENAs then pass through a second set of collimators that collimate in a direction orthogonal to the first set of collimators. The two collimators result in a single pixel field-of-view (e.g., 7~176 A fraction of the ENAs that pass through the collimator and transit the first ultrathin carbon foil, labeled F1, would be ionized. These ionized ENAs then transit a hemispherical electrostatic energy analyzer (ESA) if their exit energy lies within the selected ESA energy passband. In addition to its energy resolving capabilities, the ESA also serves to prevent UV
- 238-
Energetic neutral atom imaging of the outer heliosphere... light from entering the detector section. The ionized ENAs then enter the detector section that consists of a second ultrathin carbon foil (F2) for generation of secondary electrons that are detected separately from the ENAs for a coincidence measurement. For the energy range 0.2-6 keV, a single pixel instrument is superior to a true imaging instrument having multiple pixels due to significant ENA scattering in foil F1 that results in loss of information of the incident ENA trajectory. The scattering half-angle aPl/2 of ionized ENAs emerging from this foil is related to the incident energy according to apl/2 ~- 13 [keV-deg] / E0 Fig. 1. The ENA imager shown schematically in the [14]. For example, at 1 keV this figure is optimized to measure the faint ENA emission corresponds to an angular FWHM of from the heliosphere-LISM interaction region. The 26 ~ which is far larger than the desired ENA ionization foil can span enormous aperture areas, angular imaging resolution o f - 7 o. This enabling a large geometric factor. As an example, this loss of knowledge of the incident ENA imager would have an aperture area of 64 cm2, a singletrajectory precludes use of a true pixel FOV of 7~ ~ and six contiguous energy passbands covering 0.2-6 keV. imaging instrument such as the ENA imagers on IMAGE. The optimum implementation therefore uses a single pixel instrument in which the field of view is collimated and does not depend on scattering in the foil. The ENA imager could be placed so that its single pixel view is orthogonal to the spin axis on a sun-pointed spinning spacecraft, resulting in a 7~ ~ view each spin. The full heliosphere would be completely viewed twice over each orbit of the spacecraft around the Sun. Broader coverage could be obtained using several single pixel instruments placed at different viewing angles relative to the spacecraft spin axis. An important feature of the imager is that the foil F1 used for ENA ionization can be biased to a high positive potential VVl. ENAs that exit the foil as positive ions (e.g., H +) are accelerated perpendicular to the foil, resulting in three important advantages for enhanced ENA imaging. First, the large angular scattering at low energies, which would normally result in the loss of ionized ENAs by scattering out of the angular acceptance passband of the ESA, is mitigated by this "proximity focusing" effect: the final trajectories after acceleration are nearly perpendicular to the foil, and so these scattered ions are pushed back into the angular acceptance of the ESA and are transmitted through the ESA. Second, the accelerated ENAs transit the detector section at a much higher energy so that (a) the secondary electron yield at the second foil is higher and (b) the ENA detection probability is high. These two features combine to greatly increase the probability of coincidence detection (and therefore increase the geometric factor). Third, the ionized ENAs are accelerated to a higher energy at which the ESA energy passband is broader. This results in a larger apparent energy resolution AE/Eo of the measurement and a proportional increase in the geometric factor.
- 239-
H.O. Funsten, D.J. McComas and M. Gruntman
We illustrate this last effect by considering a spherical ESA having an intrinsic energy resolution AE/Ec = k where Ec is the central energy of an energy passband of width AE and k is a constant dependent only on the ESA geometry. ENAs with incident energy E0 are ionized by the foil, accelerated due to the foil bias, and enter the ESA at energy E0+ q VF1 (q is the ion charge). By equating Ec = E0+ q VF~, the apparent energy resolution AE/Eo from the perspective of the ENAs is
AEEo= k ( l + q V .FE1o)
(1)
The apparent energy resolution is clearly dependent on qVF1/Eo. If qVF1/Eo << 1, minimal effect is observed. However, when q VF1 becomes a substantial fraction of Eo, AE/Eo significantly increases. This results in an equivalent increase in the energy geometric factor, which is proportional to AE/Eo.
3. DETECTION OF ENAS IN THE ENERGY RANGE 0.2-6 KEV The imaging technique of ENA ionization using an ultrathin foil followed by electrostatic deflection is generally superior to other techniques for ENA imaging of the heliosphere-LISM interaction region due to the ability of ultrathin foils to span large aperture areas, their robustness, their extensive flight heritage, and their consistent performance under "dirty" vacuum environments typical of an outgassing spacecraft that might otherwise contaminate an atomically clean surface. We will not discuss the technique of ENA ionization by surface reflection, although we note that new results obtained with more exotic reflection surfaces than the single crystal tungsten used on LENA/IMAGE show enhanced ionization of H under ultrahigh vacuum conditions at energies greater than 200 eV [ 16-18] and may provide another promising approach to ENA imaging of the heliosphere-LISM interaction region. A key factor governing the efficiency of the instrument shown in Fig. 1 is the ionized fraction f ( H § or f ( H ) of hydrogen that exits the foil. We have investigated both the fraction of H § [14,15] and H- [19] exiting an ultrathin carbon foil at energies greater than several keV. At the same energy, the fraction of H exiting a carbon foil is much less than the fraction of H § so the sensitivity of the ENA imager is maximized based on conversion to H+ rather than conversion to H-. Our previous measurements o f f ( H +) have been obtained only for energies down to 1 keV, at which the conversion efficiency was f ( H § ~ 0.06. While Overbury et al. [20] measured f ( H § down to 500 eV, their foils were --2.4 times thicker than those used in this study, and the much larger energy loss in these thick foils precludes comparison of their data with our results. Using the same experimental apparatus and method described in detail in Ref. [14], we have measured the exit fraction f ( H § down to an incident energy E0 - 330 eV for a nominal carbon foil thickness of 0.5 ~tg/cm2 that is typically measured to be ~1.1 ~tg/cm2 [14]. As shown in Fig. 2(a),f(H § increases with increasing incident energy E0 from 1.3% at 330 eV to 6.9% at 1 keV and to 13.1% at 6 keV. This energy dependence is expected since, with increasing energy, the mean charge state in the bulk foil is driven toward a value of +1 while the charge exchange processes that act to neutralize an ion at the exit surface become less important [21 ].
- 240 -
Energetic neutral atom imaging of the outer heliosphere...
Incident Energy, E 0 [keV] Fig. 2. The foil that converts ENAs to positive ions is a critical component that governs instrument performance. Panel (a) shows laboratory data of the fractionf(H +) of hydrogen that exits a nominal 0.5 ~tg/cm2 carbon foil as H +. Panels (b)-(d) show the results of SRIM [22] simulations of hydrogen through a 1.1 carbon foil: (b) probability of transmission, (c) energy loss, and (d) the ratio AEF/EFof the FWHM of the exit energy distribution to the exit energy. An interesting feature of these measurements is an apparent break point at E0 = 0.9 eV. Below and above this break point, f (H +) can be well-represented by the following linear equations:
f ( H +) = -0.015 + 0.088 Eo, f ( H +) = 0.054 + 0.013 Eo,
0.3 keV < E o < 0.9 keV 0.9 keV < E o < 6 keV.
(2)
(3)
The break point likely indicates a change in the fundamental charge exchange mechanism at the exit surface that defines the final, measured exit charge state of a hydrogen atom.
-241 -
H.O. Funsten, D.J. McComas and M. Gruntman
The sensitivity of the ENA imager shown in Fig. 1 is also dependent on the probability of ENA transmission through the foils. Figures 2(b)-2(d) show the results of SRIM 2000 Monte Carlo simulations [22] of hydrogen transiting a 1.14 ~tg/cm2 carbon foil. The probability of hydrogen transmission through the foil, shown in Fig. 2(b), is 50% a t - 2 7 0 eV and increases rapidly with increasing energy. The instrument geometric factor is proportional to the transmission of through foil F1, whereas ions transiting foil F2 have been accelerated to an energy Eo+qVvl and transit foil F2 with almost 100% transmission. We note that the decrease in the transmission (and therefore geometric factor) at energies less than -500 eV corresponds to energies at which the ENA fluxes from the interaction region are expected to be the highest (see Ref. [4] and Fig. 3 of this paper). Figure 2(c) shows the mean energy loss of ENAs in the foil as a function of incident energy. The energy loss slowly increases with increasing energy, nearly as log(E0), from 0.15 keV at E0 = 0.2 keV to 0.51 keV at E0 = 10 keV. The circle in this figure corresponds to a measurement from Funsten et al. [15] of the mean energy loss by 5 keV H incident on a 1.1 ~tg/cm2 carbon foil. This measured value is 20% higher than the simulation result. One result of energy loss is that the ENAs exiting foil F1 will transit the ESA at a lower energy at which the energy passband of the ESA is correspondingly narrower. This lowering of the ENA energy distribution by energy loss in the foil would normally act to lower the geometric factor. However, we compensate for this by the accelerating ionized ENAs after exiting foil F 1 to much higher energies as discussed in the previous section. Using the SRIM simulations and energy loss data, we have also computed the ratio of the full-width at half-maximum (AEF) of the energy distribution to the mean final energy EF of the ions, which is shown in Fig. 2(d). The data was fit to the equation AEF/EF = 0.28/E0 which is shown as the solid line in the figure. The circle is a 5 keV measurement from Ref. [15] and agrees extremely well with the simulation. The combination of AEF/EF and the energy resolution of the electrostatic energy analyzer govern the energy resolution of the imager. From the figure, AEF/EF < 1 for E0 a 280 eV, and AEF/EF < 0.5 for E0 a 560 eV. Since ionized ENAs transit foil F2 at a considerably higher energy than foil F1, the impact of energy loss in foil F2 on the energy resolution of the imager is negligible.
4. DISCUSSION AND SUMMARY Calculation of the geometric factor of the instrument shown in Fig. 1 includes the following components: a 7~ ~ (0.015 sr) field-of-view, 64 cm 2 aperture area (ESA plate diameters of 16 cm and 24 cm), ENA ionization probability f ( H +) shown in Fig. 2(a), ENA transmission probability from Fig. 2(b), collimator transmission of 0.76, grid transmissions totaling 0.24, energy dependent secondary electron emission yields [23], and microchannel plate detection efficiencies. The resulting energy geometric factors G for a 7~ ~ field-of-view are approximately 0.0005 cm 2sr at 0.3 keV, 0.0025 cmZsr at 0.9 keV, and 0.01 cmZsr at 6 keV. The full energy range of 0.2 - 6 keV could be covered in six contiguous energy passbands. To demonstrate the potential for an ENA imager to distinguish between different models of the heliosphere-LISM interaction, Fig. 3 shows calculated instrument count rates from ENAs as a function of the measurement angle 0 relative to the upwind direction of the heliosphere in the LISM (assuming a cylindrical symmetry of the heliospheric boundary). The counts are based on ENA fluxes calculated for three types of termination shocks: a gasdynamical strong
-242-
Energet& neutral atom imaging oft he outer heliosphere... shock, a weak shock with thermalized pickup protons, and a weak shock with non-thermalized protons [4]. The figure shows the number of detected ENAs per day in the energy passband 0.58 < E < 1.19 keV and in the 7 o band around the axis of cylindrical symmetry centered at an angle 0. The counts were obtained by multiplying the computed ENA flux [4] by an energy geometric factor of 0.0025 cm2sr. Based on this figure, instrument scanning of the celestial sphere with the geometric factor of this instrument will yield statistically significant ENA measurements. Note also the enormous variation in the projected count rates between the different models and also as a function of angle O. The physics of the interaction region is therefore distinguishable even in the presence of significant errors in the counting statistics or in knowledge of the absolute geometric factor of the instrument. The data acquired using this imaging technique should provide the first opportunity to clearly study the physics of the interaction region.
106
_
I
l
[
I
I
[
I
I
[
I
I
l
I
J
[
I
I
; _ m
] 0.58 keV < E < 1.19 keV] _
105
Strong gas-dynamical shock
_--
~
-~
\...
r ~ 10, _ r/3 *~ wO 103 f~
i
1o~ _
_
10~ 0
Pickup-ion modulated, n o n - t h e r m a l i z e d f
J
I 30
i
i
I 60
I
I
I 90
i
J
I 120
J
I
I
I
-
J
150
1 10
Angle Relative to Upwind Direction, 0 [Deg]
Fig. 3. ENA fluxes at Earth orbit have been computed for three different models assume a cylindrically symmetric heliosphere-LISM interaction region around the axis defined by the Sun and the upwind direction of the Heliosphere in the LISM [4]. The figure shows the calculated ENA counts per day in a 7~ band around the axis of cylindrical symmetry at an angle 0 relative to the upwind direction. Spatial and spectral measurements of the ENA flux should readily distinguish between the competing models of the heliosphere-LISM interaction region.
REFERENCES [1] Gruntman, M.A., V.B. Leonas, and S. Grzedzielski, Neutral Solar Wind Experiment, in Physics o.f the Outer Heliosphere, ed. S. Grzedzielski and D.E. Page, pp. 355-358, Pergamon Press, New York, NY, 1990. [2] Gruntman, M. A., Anisotropy of the Energetic Neutral Atom Flux in the Heliosphere, Planet. Space Sci., 40, 439-445, 1992. [3] Hsieh, K. C. and M. A. Gruntman, Viewing the Outer Heliosphere in Energetic Neutral Atoms, Adv. Space Res., 13, 131-139, 1993. [4] Gruntman, M.A., E.C. Roelof, D.G. Mitchell, H.-J. Fahr, H.O. Funsten, and D.J. McComas, Energetic Neutral Atom Imaging of the Heliospheric Boundary Region, J. Geophys. Res., 2001, in press. [5] Gruntman, M.A., Energetic Neutral Atom Imaging of Space Plasmas, Rev. Sci. Instrum., 68, 3617-3656, 1997. [6] Wurz, P., Detection of Energetic Neutral Particles, in The Outer Heliosphere: Beyond the Planets, eds. K. Scherer, H. Fichtner, and E. Marsch, pp. 251-288, Copernicus Gesellschafl e.V., Katlenburg-Lindau, Germany, 2000. [7] Williams, D.J., E.C. Roelof, and D.G. Mitchell, Global Magnetospheric Imaging, Rev. Geophys., 30, 182-208, 1992.
-243 -
H.O. Funsten, D.J. McComas and M. Gruntman
[8] Mitchell, D.G., S.E. Jaskulek, C.E. Schlemm, E.P. Keath, R.E. Thompson, B.E. Tossman, J.D. Boldt, J.R. Hayes, G.B. Andrews, N. Paschalidis, D.C. Hamilton, R.A. Lundgren, E.O. Tums, P. Wilson IV, H.D. Voss, D. Prentice, K.C. Hsieh, C.C. Curtis, F.R. Powell, High Energy Neutral Atom (HENA) Imager for the IMAGE Mission, Space Science Reviews, 91, 67-112, 2000. [9] McComas, D.J., H.O. Funsten, and E.E. Scime, Advances in Low Energy Neutral Atom Imaging, in AGU Monograph 103: Measurement Techniques for Space Plasmas (Fields), eds. R. Pfaff, J. Borovsky, and D.T. Young, pp. 275-280, American Geophysical Union, Washington, DC, 1998. [10] Pollock, C.J., K. Asamura, J. Baldonado, M.M. Balkey, P. Barker, J.L. Burch, E.J. Korpela, J. Cravens, G. Dirks, M.-C. Fok, H.O. Funsten, M. Grande, M. Gruntman, J. Hanley, J.-M. Jahn, M. Jenkins, M. Lampton, M. Marckwordt, D.J. McComas, T. Mukai, G. Penegor, S. Pope, S. Ritzau, M.L. Schattenburg, E. Scime, R. Skoug, W. Spurgeon, T. Stecklein, S. Storms, C. Urdiales, P. Valek, J.T.M. van Beek, S.E. Weidner, M. Wriest, M.K. Young, C. Zinsmeyer, Medium Energy Neutral Atom (MENA) Imager for the IMAGE Mission, Space Science Reviews, 91, 113-154, 2000. [11] Moore, T.E., D.J. Chornay, M.R. Collier, F.A. Herrero, J. Johnson, M.A. Johnson, J.W. Keller, J.F. Laudadio, J.F. Lobell, K.W. Ogilvie, P. Rozmarynowski, S.A. Fuselier, A.G. Ghielmetti, E. Hertzberg, D.C. Hamilton, R. Lundgren, P. Wilson, P. Walpole, T.M. Stephen, B.L. Peko, B. Van Zyl, P. Wurz, J.M. Quinn, G.R. Wilson, The Low-Energy Neutral Atom Imager for IMAGE, Space Science Reviews, 91, 155-195, 2000. [12] Bernstein, W., R.L. Wax, N.L. Sanders, and G.T. Inouye, An Energy Spectrometer for Energetic (1-25 keV) Neutral Hydrogen Atoms, in Small Rocket Instrumentation Techniques, ed. K.-I. Maeda, pp. 224-231, North Holland, Amsterdam, 1969. [13]McComas, D.J., B.L. Barraclough, R.C. Elphic, H.O. Funsten, and M.F. Thomsen, Magnetospheric Imaging with Low-Energy Neutral Atoms, Proc. Nat. Acad. Sci. USA, 88, 95989602, 1991. [14] Funsten, H.O., D.J. McComas, and B.L. Barraclough, Ultrathin Foils Used for Low Energy Neutral Atom Imaging of Planetary Magnetospheres, Opt. Eng., 32, 3090-3095, 1993. [15] Funsten, H.O., D.J. McComas, and E.E. Scime, Low Energy Neutral Atom Imaging for Remote Observations of the Magnetosphere, J. Spacecraft and Rockets, 32, 899-904, 1995b. [16] Wurz, P., R. Schletti, and M.R. Aellig, Hydrogen and Oxygen Negative Ion Production by Surface Ionization Using Diamond Surfaces, Surface Science, 373, 56-66, 1997. [ 17] Jans, S., P. Wurz, R. Schletti, T. Fr6hlich, E. Hertzberg, and S. Fuselier, Negative Ion Production by Surface Ionization at Aluminum-Nitride Surfaces, J. Appl. Phys., 87, 2587-2592, 2000. [18] Jans, S., P. Wurz, R. Schletti, K. Brtining, K. Sekar, W. Heiland, J. Quinn, and R.E. Leuchtner, Scattering of Atoms and Molecules off a Barium Zirconate Surface, Nucl. Instrum. Meth. B, in press, 2001. [ 19] Funsten, H.O., Formation and Survival of H-and C Ions Transiting Ultrathin Carbon Foils at keV Energies, Phys. Rev. B, 52,R8703-R8706, 1995a. [20] Overbury, S.H., P.F. Dittner, S. Datz, and R.S. Thoe, Energy Loss, Angular Distributions and Charge Fractions of Low Energy Hydrogen Transmitted Through Thin Carbon Foils, Radiat. Eff, 41, 219-227, 1979. [21 ] Los, J., and J.J.C. Geerlings, Charge Exchange in Atom-Surface Collisions, Phys. Rep., 190, 133190, 1990. [22] Ziegler, J.F., J.P. Biersack, and U. Littemark, The Stopping and Range of Ions in Solid..s., Vol. 1, Pergamon, New York, NY, 1985. [23] Ritzau, S.M., and R.A. Baragiola, Electron Emission from Carbon Foils Induced by keV Ions, Phys. Rev. B, 58, 2529-2538, 1998.
low-frequency heliospheric radio emissions W. S. Kurth and D. A. Gurnett ~*
Individual spectral elements of the low frequency heliospheric radio emissions are often observed to have different intensities as determined by the two widely-spaced Voyager spacecraft. In principle, these intensity differences allow the determination of a locus of possible source locations, assuming the source emits uniformly in all directions. We take into account the dipole antenna patterns of the plasma wave antenna system on each spacecraft. The results of this relative intensity analysis are then coupled with the results of direction-finding measurements using the rotating dipole technique with Voyager 1. Near the beginning of the major radio emission event observed from mid-1992 through 1994, the results of this analysis are consistent with a source near the nose of the heliosphere. 1. I N T R O D U C T I O N Kurth et al. (1) reported the discovery of low-frequency radio waves observed by the two Voyager spacecraft in the outer heliosphere. Gurnett et al. (2) reported a second major radio emission event and suggested that the emissions were generated as a result of a global merged interaction region and associated shock interacting with the interstellar medium just beyond the heliopause. See Kurth and Gurnett (3) for a review of the low-frequency heliospheric radio emissions. Gurnett et al. (4) used the rotation of the Voyager plasma wave antennas performed a few times per year for calibration purposes to determine the plane containing the source of the radio emission at several times during the most recent radio event (1992 - 1994) and concluded that the source direction was consistent with the direction to the nose of the heliosphere early in the event, but moved away from the nose at later times. In this paper we take advantage of the fact that the two Voyager plasma wave receivers observe components of the low-frequency emissions at slightly different intensities, differing by up to a few dB. This difference in received power is most likely due to differences in the distances between the source and the two spacecraft, which are separated by more than 40 AU during this time interval. Another source of differing received powers, however, is the likelihood that the antennas on the two spacecraft present substantially different aspects with respect to the source direction. Since the dipole antenna response *We are grateful to support provided by S. Allendorf, L. Granroth, and R. Poynter. This research is supported by NASA by Contract 959193 through the Jet Propulsion Laboratory.
-245-
W.S. Kurth and D.A. Gurnett
2. S O U R C E D E T E R M I N A T I O N BY TWO OBSERVERS
USING RELATIVE POWER
DETECTED
In principle, we can use the relative power detected by two widely separated observers such as the two Voyager spacecraft to determine a locus of possible source locations of a radio emission source. If we assume omnidirectional sensors, then the basic equation is (r~)2 _ P1 r~ P2
(1)
where r~ and r2 are the vectors to the source from Voyager 1 and 2, respectively, and P~ and P2 are the observed power fluxes at the two spacecraft. The geometry is illustrated in Figure 1 and detailed information is given in Table 1 in heliocentric distance, ecliptic latitude/3, and longitude A. For any given ratio of received power, one can determine a surface upon which the source must lie. In three dimensions, each of these surfaces is a sphere surrounding the spacecraft receiving the most power except for the limiting case of no difference (0 dB) which gives a plane orthogonal to and bisecting the line segment between the two spacecraft.
rl
..
Figure 1. The geometry for relative intensity measurements.
-
246
-
Dual spacecraft measurements as a tool for determining...
Table 1 Geometry for August 7, Day 220, 1992 Vector R (AU) 49.4 37.9
136 138
fl(deg) 33.5 -10.5 -31.3 -34.8 5 8 75
A(deg) 245.0 283.0 309.7 19.7 254 249 219
*Ajello et al. (6)
a1(01) r2 2 P1 a2(02)(~) = P2
3. C O M P A R I S O N
(2)
OF V O Y A G E R 1 A N D V O Y A G E R 2 A M P L I T U D E S
-247-
Kurth and D.A. Gurnett
Figure 2. Frequency-time spectrograms for Voyager 1 and 2 showing the subtle differences in the intensities of the low frequency heliospheric emissions, especially above 2.4 kHz. The arrows indicate the time for the source location determination presented herein.
-248-
Dual spacecraft measurements as a tool for determining...
Voyager PWS '
'
'
'
I
.
.
.
.
I
.
.
.
.
I
.
.
.
.
, I
.
.
.
.
I
.
.
.
.
Voyafler-1 1992 224 02:41:58 Voyafler-2 1992 224 04:10:16
.#
o~
-3
Figure 3. A detailed comparison of the spectrum observed at the two spacecraft at nearly the same time. Notice that the 2-kHz component is nearly identical in intensity, but there is a substantial difference in the high-frequency component.
detected power of order 1 dB. A line labeled "nose direction" indicates the projection of the nose of the heliosphere into the plane of analysis, with an assumed distance of 150 AU. Notice that this direction is very close to all three contours for an extended distance. Gurnett et al. (4) found the source to be consistent with the direction to the nose, so it is comforting to see the contours lie close to the nose position, as well. Because the contours extend for a large distance beyond the nose, the relative amplitude source location method does not, in this case, provide a useful restriction in the distance to the source. Note that both the rotating dipole technique and the relative amplitude method have ambiguities; e.g. there is no a priori way of eliminating the 3-dB contour in the upper half of the plane as the true source direction. We note that the feature centered at 2.7 kHz in Figure 3 does not coincide with the frequency channel used by Gurnett et al. for the rotating dipole technique. However, the response of that channel does include the high-frequency wing of the 2.7-kHz line and we assume that the entire band is generated in the same general location. Furthermore, there is no emission apparent at higher frequencies to act as a source of confusion. The cross-hatched regions in Figure 4 represent locations in the source plane where the source would seem to be ruled out by the modulation index m of 0.29 reported by Gurnett et al. (4). The modulation index is basically the amplitude of modulation observed in the received power as the antennas are rotated. There are three factors which determine the modulation index: (1) the elevation angle of the source out of the spacecraft spin plane (perpendicular to the roll axis), (2) the angular size of the source, and (3) scattering of the radio waves between the source and the spacecraft. In Figure 4, the spin plane is parallel to the Y' axis and perpendicular to the plane of the illustration. If the source were directly in the spacecraft spin plane, was a point source, and there was no scattering,
-249-
W.S. Kurth and D.A. Gurnett
I
ss
L 0
I
S
" 100
200
x' (AU)
-250-
300
Dual spacecraft measurements as a toolfor determining... time-of-flight determination of Gurnett and Kurth (10) designated by the crosses in the figure and transform back to the ecliptic coordinate system. Note that this entire distance range along the 3-dB contour falls within the region which is not cross-hatched in Figure 4. Not surprisingly, one of the solutions shown in Table 1 is within a few degrees of the direction to the nose; the other solution is well separated from the nose. Errors from the direction-finding analysis of Gurnett et al. (4) are of order a few degrees and those from the analysis of the intensity differences are at least as large. The uncertainty in the distance determined from time-of-flight considerations (10) is large as mentioned above. 4. C O N C L U S I O N S We have developed a technique for defining the locus of possible source locations for heliospheric radio emissions based on the relative power detected at the two Voyager spacecraft. We have combined the relative power technique with the results of the rotating dipole technique by presenting the loci of possible source positions given by the relative power technique in the plane of the source as determined by the rotating dipole technique. We further restrict the source location by excluding a portion of this plane on the basis of the modulation index reported by Gurnett et al. (4). The results are consistent with a source near the nose of the heliosphere early in the 1992-1994 heliospheric radio emission event. REFERENCES
1. W.S. Kurth, D.A. Gurnett, F.L. Scarf, and R.L. Poynter, Nature 312, 27-31 (1984). 2. D.A. Gurnett, W.S. Kurth, S.C. Allendorf, and R.L. Poynter, Science 262, "199-203 (1993). 3. W.S. Kurth, and D.A. Gurnett, in Cosmic Winds and the Heliosphere, eds. J. R. Jokipii, C. P. Sonett, and M. S. Giampapa, University of Arizona Press, Tucson, pp. 793-832 (1997). 4. D.A. Gurnett, S.C. Allendorf, and W.S. Kurth, Geophys. Res. Lett. 25, 4433-4436 (1998). 5. F.L. Scarf and D.A. Gurnett,Space Sci. Rev. 21, 289-308 (1977). 6. J.M. Ajello, A.I. Stewart, G.E. Thomas, and A. Garps, Ap. J. 317, 964-986 (1987). 7. I.H. Cairns, Geophys. Res. Lett. 22, 3433-3436 (1995). 8. I.H. Cairns, in Solar Wind Eight, AIP Conf. Proc. 382,ed. by D. Winterhalter et al., pp. 582-585 (1996). 9. J.W. Armstrong, W.A. Coles, and B.J. Rickett, J. Geophys. Res. 105, 5149-5156 (2000). 10. D.A. Gurnett, and W.S. Kurth, Adv. Space Sci. 16, (9)279 (1995).
-251 -
This Page Intentionally Left Blank
Theories for Radio Emissions From the Outer Heliosphere Iver H. Cairns a and G. P. Zank b aSchool of Physics, University of Sydney, NSW 2006, Australia bBartol Research Institute, University of Delaware, Newark, De 19716, USA The Voyager spacecraft observe episodic bursts of radio emission generated in the outer heliosphere, arguably when shock waves driven by global merged interaction regions (GMIRs) pass beyond the heliopause. Theories for the source region, generation, and propagation of the radiation are reviewed. Special foci are the successes and problems of the current G MIR model and a new theory which explains the turn-on and generation of the radiation in the outer heliosheath and its propagation into the heliosphere. 1. I N T R O D U C T I O N Radio emissions detected beyond Saturn's orbit at ~ 2 - 3 . 5 kHz by plasma wave instruments on the Voyager spacecraft [1-7] have excited considerable interest. The radiation comes from the outskirts of the heliosphere, most likely from beyond the heliopause, and is arguably the most powerful radio emission associated with the solar system. Questions of interest include: where and how are the emissions generated, how are they associated with solar activity and the very local interstellar medium (VLISM), can the emissions be used to constrain the location of the heliopause and the plasma characteristics of the source region, and are there more emissions to be discovered? The purpose of this paper is to review theories and empirical models for the observed radio emissions [4,8-16]; particular attention is devoted to explaining why particular theoretical ideas are or are not viable and to summarizing a new theory [16] which provides the first detailed theoretical basis for Gurnett et al.'s empirical model [4] for the radiation. Attention is also given to potential tests of theories, specifically pointing to the theoretical difficulties posed by the existence of two classes of radio emissions and their frequency fine structures. Before reviewing theories it is vital to summarize the observational constraints that must be met by a successful theory, - see [7] for an observational review. These are: (1) Radiation frequencies f ,,~ 1 . 8 - 3.6 kHz [1-5]. (2) The observed intensities and power fluxes, which arguably correspond to the most powerful known solar system radio source [4], with a power > 1013 W. (3) The radiation is produced in rare outbursts associated with global merged interaction regions (GMIRs) moving outwards from the Sun [4,5], with one major outburst approximately once per solar cycle. (4) The radiation turns on in the outer heliosphere or beyond since (i) the Voyager 1 and 2 dynamic spectra are
-253 -
I.H. Cairns and G.P. Zank
essentially identical despite spacecraft separations ~ 2 0 - 50 AU [1,4], (ii) the time delay between a GMIR passing Voyager and the radio onset implies that the source is at a heliocentric distance R ,,~ 120-140 AU [4], not inconsistent with the heliopause, and (iii) the Voyagers observe no local radio emission near the GMIRs [4]. (5) Triggers are essential for observable radiation to be produced [8,16]. This is implied by the discreteness and rarity of the radio events and the lack of emission in the solar wind near GMIRs. (6) Two classes of narrowband ( A f / f < 0.3) emissions exist [2,12,4]. First, "transient emissions", which increase almost linearly in frequency from ,,~ 2 to ,,~ 3.5 kHz in about 180 days before disappearing. Often these appear to repeat with larger starting frequencies and slower drift rates. Second, the "2 kHz component", which is very long lasting (~ 500 days), remains almost constant in frequency with slow changes in amplitude. (7) Both transient emissions and the 2 kHz component have frequency fine structure, often splitting into several non-harmonic bands separated by a few hundred Hz. This last point is not widely recognised but is clearly visible in published dynamic spectra [e.g., 2,4]. A successful theory must explain why and how the radiation has these observed properties. Accordingly the theory must specify, link together, and explain quantitatively the (i) emission mechanism, (ii) source location and plasma characteristics, (iii) radiation frequencies, intensities, and time variations, (iv) triggers for radiation outbursts, and (v) propagation of the radiation from the source to the observer. As well as explaining a fascinating phenomenon that links astrophysics and space physics, a viable theory will provide a means to remotely infer the characteristics, location, temporal and spatial variations, and physics of the source plasma using the observed radiation. The structure of the paper is as follows. Section 2 reviews emission mechanisms, finding that generation of radiation at multiples of the electron plasma frequency fp in a foreshock region upstream of a shock is most plausible. Theories involving fp and/or 2fp radiation generated near the termination shock [10-12] or in the inner heliosheath [13,14] are reviewed in Section 3. Section 4 reviews the qualitatively compelling GMIR model of Gurnett et al. [4-6] and describes the theoretical questions left unanswered. A new theory [16] which provides a first detailed theoretical basis for the GMIR model is then summarized and partially explained in Section 5. Directions for future research are listed in Section 6, focusing on theoretical difficulties in explaining the observed frequency fine structures. Section 7 contains the conclusions. 2. E M I S S I O N M E C H A N I S M S Potential emission mechanisms include the following. First, gyro-synchrotron or synchrotron emission at harmonics of the electron cyclotron frequency fee from relativistic electrons. This mechanism is familiar from many astrophysical sources (e.g., radio galaxies and associated jets), the Jovian radiation belts, and perhaps type IV solar radio bursts [17,18]. Second, cyclotron maser emission at harmonics of f~e, due to microinstabilities driven by gradients Of/i:gv of the electron distribution function f(v) [19,20]. This mechanism is believed to produce the dominant radio emissions from Earth, Jupiter and other planets [20], and solar decimetric spike bursts. Third, mode conversion processes similar to that producing continuum radiation in planetary magnetospheres [21-23] Finally, radiation at fp and 2fp produced from Langmuir waves driven by electron beams, as observed
- 254-
Theories for radio emissions from the outer heliosphere for many solar and interplanetary emissions [17,18,24-29]. Several other mechanisms for astrophysical sources, such as pulsars, have already been eliminated [1]. Synchrotron (and gyro-synchrotron) emission is very implausible. First, while the observed radiation is narrowband, the mechanism produces broadband radiation centered around a high harmonic s ~ (7 sin a) 3 of f~, where 7 is the electron Lorentz factor and a is the emission angle relative to the ambient magnetic field B. Second, current estimates are B ..~ 0.1 nT [30], so that fc~ ~ 3 Hz, and f / f ~ ,-,., 103. This implies 7 > 10. What is the source of these relativistic electrons and why have they not yet been detected directly? Finally, even if B greatly exceeds 0.1 nT the mechanism should produce multiple harmonics, inconsistent with the data. Cyclotron maser emission is implausible for similar reasons. First, theory implies a necessary condition f~ > lOfp [19,20], which is strongly inconsistent with current estimates for B and fp [30]. Second, these estimates for B would require emission at very high harmonics of fee where the mechanism is broadband, but the radiation is narrowband. Finally, while the mechanism is indeed narrowband at low harmonics of f~, the required values f~ > 1 kHz are very implausible. Fahr et al. [21] speculated that conversion of electrostatic waves at the heliopause might produce radio emissions, similar to conversion of upper hybrid waves into continuum radiation in planetary magnetospheres [22,23]. Their discussion of wave modes, instabilities, and conversion processes lacks detail, is outdated, and should be revised. Currently, however, the extremely small thickness (ideally a few ion gyroradii) predicted for the heliopause, the small range of frequencies observed (cf. the range ~ 0.1 - 3 kHz predicted for fp at the heliopause), and the two classes of emission observed make the mechanism and source location unattractive. Radiation produced at fp and/or 2fp is the favored theoretical interpretation for the Voyager radio emissions [1,4,10-16]. Reasons include the following. First, this mechanism is well accepted for type II and III solar radio bursts in the corona and solar wind [17,18,2426,33,34] and radiation from Earth's foreshock [27-29] Common characteristics of these phenomena are electron beams, Langmuir-like waves near fp, and conversion of these waves into fp and/or 2fp radiation in plasmas where fp > 0.2f~. Second, a well developed and successful theory exists for fp and 2fp emission based on these characteristics [17, 29,23,33,34]. Third, Earth's foreshock radiation and type II bursts are associated with a shock/moving plasma system, strongly analogous to the association between GMIRs and the Voyager emissions. Finally, the above solar and interplanetary emissions vary in frequency when fp varies but are otherwise constant, as anticipated in most interpretations of the two classes of Voyager emissions. The basic model is thus that the Voyager radio emissions are fp and/or 2fp emission produced upstream of a shock wave (Figure 1): electron beams are produced naturally by time-of-flight effects in the foreshock upstream of the shock [31,32], Langmuir waves are driven by an electron beam instability [31,32], and fp and/or 2fp radiation is produced from the Langmuir waves by either nonlinear processes or linear mode conversion [17,18,23, 31,33,34]. The time-of-flight effect results from all electrons undergoing an E • B plasma drift perpendicular to B so that dispersion in parallel speed leads to spatial dispersion and so the development of a beam in the electron distribution. Emission downstream from the shock is not favored since no observations exist, nor a theoretical mechanism, for
-255-
I.H. Cairns and G.P. Zank
Flow x_E
v -
I\ v [\:ExB
f & 2f radiation P P
t3 f & 2f r a d i a t i o n ~
~
B -
~
-
!
electronbeams Langmuirwaves(L)
/ Foreshock/
Foresh?ck \ / '
Figure 1. Schematic of a foreshock source of fp and 2fp region; the foreshock is upstream of the shock but downstream from the tangent magnetic field line.
the formation there of electron beams with speeds significantly greater than the electron thermal speed V~. Instead, with the exception of one Ulysses observation of radiation at harmonics of the downstream fp [35], all the radiation, Langmuir waves and electron beams observed thus far are in the foreshock region upstream of the shock. 3. T E R M I N A T I O N
SHOCK AND INNER HELIOSPHERE
The first detailed theory for the 2-3 kHz radiation involved fp and 2fp radiation generated in the foreshock region sunward (upstream) of the termination shock [10-12]. This region is predicted to contain electron beams and Langmuir waves due to reflection and acceleration at the shock [10]. Moreover, using known analytic theory for the radiation processes and extrapolations of the Langmuir fields in planetary foreshocks to large R ~ 5 0 - 150 AU, the observed radiation levels can be explained semi-quantitatively [1012]. However, three reasons exist why the theory is implausible for the observed 2-3 kHz radiation. First, with Voyager 1 yet to cross the termination shock, RTS > 75 AU and the observation h ( R ) ,~ 25(R/1AU) -~ kHz [43] implies that f p ( R T s ) < 300 Hz is much less than the observed radiation frequencies. Second, radiation produced at such low values of fp cannot propagate into the inner heliosphere, since radiation is reflected by regions with fp > f. These difficulties can be lessened, in principle, by generating the radiation when transient solar wind regions with greatly enhanced density (by factors > 1 0 - 25) enter the foreshock [12]. However, such large density increases are rarely if ever observed by Voyager [43]. Third, the observed association between radiation outbursts and GMIRs [4] with only modest increases in density (factors < 5) cannot be simply explained. This theory thus cannot explain the observed 2-3 kHz radiation. However, radiation produced in this way might be observable sufficiently close to the termination shock [12,15,16]. Figure 2(a) summarizes Whang and Surlaga's theory [13] for the radiation: the GMIR shock reaches the heliopause and produces a reflected (R) and transmitted (T) shock which move into the inner and outer heliosheaths, respectively; the 2 kHz component is -256-
Theories for radio emissions from the outer heliosphere
Figure 2. Schematics of theories" (a) Whang and Burlaga [13] and (b) Zank et al. [14].
produced downstream from shock R, ' while transient emissions are produced downstream of shock T as it propagates up a density ramp beyond the heliopause. The theory accounts approximately for the radiation frequencies and predicts two different classes of emission. However, it is very implausible since it relies on downstream emission, inconsistent with available theory and almost all observations. Figure 2(5) shows Zank et al.'s theory [14]. The theory is based on simulations of a cosmic ray-modified termination shock (CRMTS), which markedly increases the compression ratio and downstream fp, and two triggers: a solar wind density hump interacts with the CRMTS to produce density ramps in the inner heliosheath, a subsequent shock rides up an induced ramp and produces a drifting "transient emission" in its foreshock, while the 2 kHz component is produced ahead of the shock in the undisturbed inner heliosheath. This theory can thus account plausibly for the observed frequencies and two classes of radiation. However, it is incomplete since (1) it cannot explain why the radiation turns on beyond the termination shock and (2) MHD and charge-exchange effects are neglected (only GD and cosmic ray effects are included). The model is thus not widely favored. 4. T H E G M I R M O D E L Gurnett et al.'s "GMIR" model [4] posits that the radiation is generated after the GMIR shock enters the outer heliosheath beyond the heliopause. Figure 3 summarizes the model, which assumes the radiation is produced at fp and/or 2fp by the standard electron beam/Langmuir wave/coupling mechanism [10-12]. The increasing frequency of transient emissions is attributed to a hypothesized density ramp (by a factor of ~ 3) beyond the heliopause nose, while the 2 kHz component comes from outer heliosheath regions where the density ramp is minimal. The model assumes that radiation propagates across the GMIR shock into the inner heliosphere; however, a strong GMIR shock (density jump = 4) would reflect the fp or 2fp radiation back into the outer heliosheath. The GMIR model is qualitatively compelling. Its successes include (i) explaining the observed association between GMIRs and radiation outbursts, (ii) implying a plausible -257-
I.H. Cairns and G.P. Zank
Figure 3. Schematic of the GMIR model [4].
distance (~ 120-140 AU) to the heliopause from the time delay between a GMIR passing the Voyagers and the radio onset, (iii) predicting radiation frequencies ,,~ 1 - 1 0 kHz using the model and independent estimates [30] of the VLISM plasma density (~ that in the outer heliosheath-), and (iv) qualitative explanations of the two classes of radio events. Major theoretical issues or problems exist for the GMIR model. In particular: (I1) Why does the observed radiation turn on? (I2) Why are no emissions observed in the solar wind or from the inner heliosheath? (I3) Where exactly does the observed radiation turn on? (I4) Why is the radiation at frequencies f of order expected values of fp in the outer heliosheath (based on global simulations and remote sensing measurements [30])? (I5) Where is the 2fp radiation expected from the emission mechanism and inner solar system observations? (I6) How does the radiation surmount the shock density ramp and propagate to the Voyager spacecraft? (I7) What is the origin of the density ramp (by a factor of 3) beyond the heliopause nose that is assumed in the model for transient emissions? (I8) Why do the transient emissions and 2 kHz component have such strikingly distinct dynamic spectra and not a smooth transition between two limiting forms if the sources are both in a smoothly varying outer heliosheath? (I9) How does one explain frequency fine-structures in both the 2 kHz component and transient emissions? The next section summarizes a new theory [16] that resolves issues (I1) - (I6) above, provides the first detailed underlying theoretical basis for the G MIR model, and preserves the qualitative and observational successes of the model. Subsequently issues (I7) - (I9) are discussed in more detail in Section 6. 5. N E W T H E O R Y F O R T H E R A D I A T I O N The theory involves the following major components [16]: (i) The radiation turns on when the GMIR shock enters a region primed with an enhanced superthermal electron tail beyond but near (within ~ 50 AU) the heliopause nose. (ii) The priming mechanism involves electrons being resonantly accelerated by wave-particle interactions with lower hybrid waves driven by a ring-beam instability of pick-up ions. This mechanism is well-258-
Theories for radio emissions from the outer heliosphere known in laboratory and cometary plasma physics as "lower hybrid drive" [36-38]. Here the pick-up ions result from charge-exchange of "component 2" neutrals that originate in the inner heliosheath [30,39]. (iii) Constraints on the instability and thermal damping of the lower hybrid waves localize the electron tail to the outer heliosheath. The priming process does not occur in the solar wind or inner heliosheath. (iv) The radiation is produced in the G MIR foreshock when in the primed region, due to reflection and acceleration of the tail electrons leading to dramatically enhanced electron beams, Langmuir wave fields, and emission of fp and 2fp radiation. (v) Combining the predicted tail characteristics with shock-drift acceleration at the shock and analytic plasma theory for specific nonlinear Langmuir processes [34] implies that both fp and 2fp radiation are produced effectively. (vi) Propagation effects, scattering by density irregularities, and reflection by regions with fp > f are vital in determining what radiation reaches the inner heliosphere. (vii) The fp radiation is strongly scattered and trapped near the GMIR shock, but diffuses around the shock until sunwards of the heliopause and able to enter the inner heliosphere. (viii) In contrast, most 2fp radiation is lost into the VLISM and outer heliosheath since it is scattered relatively weakly. Calculations of the predicted electron tail characteristics [16] predict that the tail should extend up to parallel speeds ~ (10-20)Ve and contain ~ (10-~-10 -4) of the total electron density, dramatically enhancing the number of superthermal electrons over the predictions for a Maxwellian. Localization of the priming mechanism to the magnetic draping region in the outer heliosheath (likely near the heliopause nose) is due to the following [16]. First, strong lower hybrid damping prevents effective growth of lower hybrid waves and the electron tail in the inner heliosheath since the ambient ion thermal speed is > the ring speed v~i~g (which is ~ the perpendicular phase velocity of the waves). Second, the lower hybrid instability and tail formation proceeds only when the Alfven speed VA obeys Vring/VA < 5 [36-38]; this condition prevents the priming process proceeding in the solar wind (v,.i~g = v~o and Vsw/VA ~ 8) or in the VLISM and outer heliosheath when far from the draping region (vri,~g/VA ~ 8 too). However, in the outer heliosheath's draping region v,.ing/VA ~ 0.8 -- 2 because of the predicted factor of 4 - 10 amplification of B, due to slowing and draping of the flow in the Axford-Cranfill effect, predicted by existing simulations [40,41]. Thus, these arguments and calculations predict that the priming mechanism is effective only in the outer heliosheath within ~ 50 AU of the draping region. In combination with acceleration of the tail by the GMIR shock, these results semiquantitatively explain why and where the radiation turns on, thereby removing problems (I1)- (I4) of the GMIR model. Discussion of the radiation processes is omitted here (but see [10-12]) in favor of propagation and scattering effects, which remove problems (14) - (I6) above for the GMIR model. The above theory predicts efficient generation of both fp and 2fp radiation, yet no harmonic structures are observed. However, radiation is reflected by dense regions with fp >__ f and angular scattering by density irregularities increases as f ~ fp with a rate or ( 1 - f~/f2)-2 [42]. Figure 4 shows the spatial variations in h ( X , 0) from recent 2-D global simulations [39], where X is along the Sun-nose axis (the symmetry axis) and 0 is the associated polar angle. Density increases at the termination shock, heliopause and outer bow shock are clear. Now superpose a strong GMIR shock (density jump of 4) onto the fp(X, 0 - 0) profile. Foreshock fp radiation is then clearly trapped (by reflection and
-259-
1.H. Cairns and G.P. Zank
-" .
:
"--..
| ~
*%.,.
......
f i
i
1
Figure 4. Profiles fp(X, O) from 2-D, cylindrically-symmetric global simulations [16,39].
scattering) between the shock and the slow density increase at larger R, remaining within a few AU of the shock. However, the slow decrease in fp(R, O) as 0 increases means that the radiation diffuses to larger 0 around the side of the shock and to smaller R closer to the heliopause. Eventually the radiation reaches locations where the shock is sunwards of the heliopause, so that fp << f and the radiation can then propagate freely into the heliosphere. In contrast, 2fp radiation is relatively weakly scattered but is reflected at the GMIR shock: accordingly almost all 2fp radiation propagates outwards into the outer heliosheath and VLISM and is lost there. Only a small fraction should diffuse around the side of the shock to where f < fp and then enter the heliosphere; it is suggested that this 2fp radiation is below current detection thresholds. These ideas explain problems (I4) - (I6) for the GMIR model but also allow prediction of an annular source shape for the radiation when viewed interior to the heliopause and the existence of 2fp radiation below current detection thresholds. The other primary ideas of the theory (e.g., the pickup ions, lower hybrid waves, electron tail etc.) are also predictions suitable for observational testing by the Voyagers and future interstellar missions. 6. O U T S T A N D I N G P R O B L E M S A N D F U T U R E W O R K The most critical areas for future research involve the detailed frequency spectra and fine structures of the radiation (issues (I7)- (I9) of the GMIR model), quantitative studies of the lower hybrid drive, radiation processes, and scattering/propagation aspects of the new theory [16], and quantitative comparisons of theory with data. Attention is focused on frequency spectra and fine structures since these arguably require new ideas. The clearly different dynamic spectra of the 2 kHz component and transient emissions are interpretable in terms of two different source regions and/or two different mechanisms. With only one mechanism appearing plausible, fp and 2fp emission, this means two different source regions. Calculations of dynamic spectra for this mechanism [15], using 3-D density profiles obtained by unwrapping the azimuthal symmetry of Figure 4's simulations, show that fp emission from the outer heliosheath closely resembles the 2 - 260 -
Theories for radio emissions from the outer heliosphere
kHz component. However, these calculations do not yield drifting emissions (even from the heliopause ramp) with the time scales or frequency drifts of transient emissions [15]. Previously transient emissions were interpreted in terms of Fermi upshifting [9]. Drawbacks are, however, that all radiation should then drift upwards but the 2 kHz component does not [12], and the inferred cavity efficiencies are too large [4]. A density ramp thus remains the most attractive option to explain transient emissions. However, the steadystate ramps in Fig. 4 and current simulations [39,41] are not large enough to produce emissions from 2 to 3.6 kHz, which requires a factor ~ 3 increase in density. Time-varying density ramps due to solar cycle effects [44] or coupling of transient solar wind phenomena into the outer heliosheath appear attractive. Further global simulations are required to investigate this possibility. Frequency fine structures, principally two or more similar components split by ~ 200 Hz, exist in published dynamic spectra of both classes of radio event [cf. 2,4,6] but have received no attention until now. These could result from the shock moving up a large-scale density ramp at different locations (for transient events only), as inferred for "multiple lane" type IX solar bursts [17] and in calculated dynamic spectra [15]. Perhaps a more attractive idea, since it can apply more easily to the 2 kHz component as well, is that the fine structure is intrinsic to the emission mechanism. Analogies are then to "split-band" type II bursts and to splitting at f ~ / 2 in Earth's foreshock [45]. Since the latter analogy implies B > 10 nT, well above current estimates, these ideas require further consideration but are potentially very important. 7. C O N C L U S I O N S The GMIR model for the Voyager radio emissions [4], based on observational data, is qualitatively compelling. A new theory [16] provides an underlying theoretical basis for the GMIR model: using the lower hybrid drive mechanism, known radiation processes, scattering, and propagation effects, it removes many previously unexplained theoretical issues with the GMIR model. These include why and where the radiation turns on in the outer heliosheath, the frequencies of the observed radiation, and how radiation reaches the inner heliosphere. The theory provides multiple predictions suitable for observational testing but is not complete. In particular, explanations for the two classes of emission and frequency fine structures remain primitive although several promising research directions exist. Certain previous theories involving emission just sunward of the termination shock or in the inner heliosheath are not viable for the observed emissions but may be relevant to presently undetected emissions that exceed detection thresholds closer to their sources. Financial support from the Australian Research Council, NASA grants NAG5-7390 and NAG5-7796, and JPL contract 959167 is gratefully acknowledged. REFERENCES
1. W.S. Kurth et al., Nature 312 (1984) 27. 2. W.S. Kurth et al., Geophys. Res. Lett. 14 (1987) 49. 3. W.S. Kurth, in Proc. Sixth Int. Solar Wind Conference, Vol. II, NCAR/TN-306+Proc (1988) 667. 4. D.A. Gurnett et al., Science 262 (1993) 199.
-261 -
I.H. Cairns and G.P. Zank
D.A. Gurnett and W.S. Kurth, Adv. Space Res., 16(9) (1995) 279. D.A. Gurnett, S.C. Allendorf, and W.S. Kurth, Geophys. Res. Lett., 25 (1998) 4433. D.A. Gurnett, this volume (2001). McNutt, R.L., Jr., Geophys. Res. Lett., 11 (1988) 1307. A. Czechowski and S. Grzedzielski, Nature 344 (1990) 640. W.M. Macek et al., Geophys. Res. Lett. 18 (1991) 357. I.H. Cairns and D.A. Gurnett, J. Geophys. Res. 97 (1992) 6235. I.H. Cairns, W.S. Kurth, and D.A. Gurnett, J. Geophys. Res. 97 (1992) 6245. Y.C. Whang and L.F. Burlaga, J. Geophys. Res. 99 (1994) 21,457. G.P. Zank et al., J. Geophys. Res. 98 (1994) 14,729. I.H. Cairns and G.P. Zank, Geophys. Res. Lett. 26 (1999) 2605. I.H. Cairns and G.P. Zank, Astrophys. J. submitted (2001). N.R. Labrum and D.J. McLean (eds), Solar Radiophysics, Cambridge U. Press, Cambridge, 1985. 18. D.B. Melrose, Plasma Astrophysics, Gordon & Breach, New York, 1980. 19. C.S. Wu and L.C. Lee, Astrophys. J. 230 (1979) 621. 20. P. Zarka, J. Geophys. Res. 103 (1998) 20,159. 21. H.J. Fahr et al., Space Sci. Rev. 43 (1986) 329. 22. K.G. Budden, Radio Waves in the Ionosphere, Cambridge U. Press, Cambridge, 1961. 23. I.H. Cairns and P.A. Robinson, in Radio Astronomy at Long Wavelengths, Geophys. Monograph 119, AGU, Washington (2000) 27. 24. J.P. Wild, Aust. J. Sci. Res., A3 (1950) 541. 25. M.J. Reiner and M.L. Kaiser, J. Geophys. Res. 104 (1999) 16,979. 26. S.D. Bale et al., Geophys. Res. Lett. 26 (1999) 1573. 27. D.A. Gurnett, J. Geophys. Res. 80 (1975) 2751. 28. S. Hoang et al., J. Geophys. Res., 86 (1981) 4531. 29. I.H. Cairns, J. Geophys. Res. 93 (1988) 3958. 30. G.P. Zank Space Sci. Rev. 89(3/4) (1999) 1. 31. P.J. Filbert and P.J. Kellogg, J. Geophys. Res. 84 (1979) 1369. 32. I.H. Cairns, J. Geophys. Res. 92 (1987) 2315. 33. P.A. Robinson and I.H. Cairns, Astrophys. J. 418 (1993) 506. 34. P.A. Robinson and I.H. Cairns, Sol. Phys. 181 (1998) 395. 35. S. Hoang et al., in Solar Wind 7, Pergamon Press, Oxford (1992) 465. 36. J.B. McBride et al., Phys. Fluids 15 (1972) 2367. 37. Y.A. Omelchenko et al., Sov. J. Plasma Phys. 15 (1989) 427. 38. V.D. Shapiro et al., Physica Scr. T75 (1998) 39. 39. G.P. Zank et al., J. Geophys. Res. 101 (1996) 21,639. 40. H.L. Pauls and G.P. Zank, J. Geophys. Res. 101 (1997) 17081. 41. T. Linde et al., J. Geophys. Res. 103 (1998) 1889. 42. J.-L. Steinberg, S. Hoang, and C. Lacombe, Astron. Astrophys. 7 (1985) 151. 43. J.W. Belcher et al., J. Geophys. Res. 98 (1993) 15,177. 44. G.P. Zank, in Solar Wind 9, AIP Conf. Proc. 471 (1999) 783. 45. I.H. Cairns, J. Geophys. Res. 99 (1994) 23,505. o
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
- 262-
M a p p i n g the H e l i o p a u s e in E U V Mike Gruntman Department of Aerospace and Mechanical Engineering, MC-1191 University of Southern California, Los Angeles, CA 90089-1191
We know very little about the heliopause, a boundary that separates the solar wind and the galactic plasma of the local interstellar medium (LISM), with the direct experimental data next to nonexistent. We propose to explore the heliopause remotely, from 1 AU by an observer outside of the geocorona. Interstellar plasma ions beyond the heliopause would glow under solar extreme-ultraviolet (EUV) radiation in the resonance lines of oxygen (83.4 nm) and helium (30.4 nm). The measurements of this glow would map the heliopause. Heliopause mapping in EUV is a way to remotely explore the heliospheric interface region and the LISM ionization state and to probe the asymmetry of the interstellar magnetic field. 1. H E L I O P A U S E
The interaction of our star, the Sun, with the surrounding local interstellar medium (LISM) leads to the buildup of the heliosphere, the region where the Sun controls the state and behavior of the plasma environment. The heliosphere is a complicated phenomenon where solar wind and interstellar plasmas, neutral interstellar gas, magnetic fields, anomalous and galactic cosmic rays, and energetic neutral atoms play prominent roles (Figure 1). Experimental data on the sun-LISM interaction region are exceptionally scarce, Figure 1. Possible solar wind interaction (twomostly indirect and often ambiguous. We shock model) with the LISM: TS - termination know very little about the heliopause, a shock; HP-heliopause; BS - bow shock; CRboundary that separates the solar wind cosmic rays; ISP(G) -interstellar plasma (gas); plasma and the interstellar plasma. 1 B - magnetic field. Interstellar plasma flows Many important questions remain outside the heliopause (gray area). Angle 0 is unanswered, for example" What is the counted from the upwind direction. distance to and shape of the heliopause? What is the ionization state of interstellar gas in the LISM? What is the direction and magnitude of the interstellar magnetic field? Is the interstellar wind subsonic or supersonic? Is a bow shock formed in front of the heliosphere?
- 263 -
M. Gruntman
Voyager 1, now approaching 80 AU, is expected to cross the termination shock and explore the heliospheric sheath properties in one point-direction. The spacecraft may not however reach the heliopause by the end of the mission in ---2025. The interstellar wind, solar wind latitudinal variations, and interstellar magnetic field make 10E+01 1 TS 'P ' the heliosphere asymmetric. Only remote ~ ]1 H ] l techniques, complemented by the "ground truth" ~ ~ I~ Voyager and future Interstellar Probe in-situ o can provide a global view of the ~ time-varying three-dimensional heliosphere on a ~ ' Iv continuous basis. ~ ~ ,~ The heliopause separates the interstellar and ~ 1.05-03 ~ ., solar wind plasmas with the number densities 2 orders of magnitude different. Figure 2 illustrates 1.05-04 ............................ ~.................. the dependence of the plasma number density 2 0 100 200 300 400 50c on the heliocentric distance in the approximately heliocentric distance, AU upwind direction (with respect to the interstellar Figure 2. Typical plasma number density wind). The plasma density rapidly decreases with in the upwind direction. The arrows the expansion of the solar wind. Interstellar indicate the positions of the termination plasma cannot cross the heliopause and flows shock (TS), heliopause (HP), and bow around it. One can imagine the sun surrounded shock(BS). by an interstellar ion "wall" beyond the empty cavity, the "heliopause moat," limited by the heliopause boundary. Is the heliopause stable under such conditions? This heliopause moat suggests a way of remote, from 1 AU, mapping of the heliopause. 3'4 Singly charged interstellar ions (He +, O +) would scatter the corresponding solar extremeultraviolet (EUV) line emissions. Measurements of this scattered radiation, the LISM plasma glow, would open an access to the heliopause and the region beyond. Heliopause imaging would map the heliopause and provide an important insight into the LISM ionization state and the asymmetry of the interstellar magnetic field. Heliopause EUV mapping, combined with the heliosheath plasma imaging in energetic neutral atoms, 5'6'7'8 will explore in detail the three-dimensional time-varying region of the sun-LISM interaction. r
2. HELIOPAUSE MAPPING
Interstellar neutral atoms are unsuitable for heliopause mapping because they penetrate deep into the heliosphere. Most of the glow of heliospheric neutrals, as seen by an observer near 1 AU, would originate within the 10-AU region. Interstellar protons cannot be imaged optically at all. Interstellar helium (He +) and oxygen (O +) ions are ideally suited for heliopause mapping. Helium is the most abundant interstellar gas constituent (---10%) after hydrogen, with the exceptionally bright corresponding solar line (30.4 nm). Measurement of the ionization state of interstellar helium will provide an important insight into heating and cooling of the LISM. Oxygen is the most abundant interstellar gas minor constituent (~0.1%). The ionization states of hydrogen and oxygen are tightly coupled by efficient charge exchange. Therefore, the experimental determination of the ionization state of oxygen will establish the ionization state of hydrogen.
- 264 -
Mapping the heliopause in EUV
For an observer at 1 AU looking in the antisolar direction, the radiance (photon sr -1) F(O) is an integral along the line of sight
c m -2 s -1
oo
1 R~ ~N, + F ( O) - --4-~
(R,O)gI(R)dR
where the g-factor (scattering rate per ion per second) gl(R) depends on the distance from the sun, NI+(R,0) is the local ion number density, RE = 1 AU, and the scattering phase function is assumed isotropic. For a simplified case of 1) vanishing ion number density inside the heliopause (R < RHp) and 2) interstellar plasma (beyond the heliopause) at rest with a uniform number density and constant temperature, the radiance would be F(O) =
NI+ gI.E R2E 4~r
1 Rzp
where gx.E is the g-factor at 1 AU. The sky brightness is thus inversely proportional to the distance to the heliopause in the direction of observation. Measuring the directional dependence (imaging) of the interstellar plasma glow is a way to establish the size and shape of the heliopause. The detailed temperature, velocity and number density flow fields of the LISM plasma are needed for accurate treatment of the problem, and the g-factors should be calculted for the specific local velocity distribution functions of the ions. The plasma flow field was calculated for this work by Vladimir Baranov and co-workers in the Russian Academy of Sciences, Moscow using their two-shock sun-LISM interaction model 2 with the following LISM parameters (at infinity)" velocity, 25 km s-l; temperature, 5672 K; electron (proton) number density, n~ = 0.07 cm-3; neutral hydrogen number density, nH = 0.14 cm -3. The solar wind was assumed to flow spherically symmetric with a velocity of 450 km s -1 and a number density of 7 cm -3 at 1 AU. Interstellar helium and oxygen abundances were assumed 0.1 and 7x 10-4, respectively, by the number of atoms relative to hydrogen.
3. SOLAR LINES, BACKGROUND, AND FOREGROUND The solar emissions in the He + and O + resonance lines are well k n o w n . 9'10 Figure 3 shows the line profiles used in this work. The total solar flux in the helium line (30.4 nm) was assumed to be 6.0• 109 cm -2 s-1 at 1 AU and the line FWHM 0.01 nm (0.1 A). 9 An oxygen ion O + has a triplet transition at 83.2754, 83.3326, and 83.4462 nm. Both the solar emission O + and the nearby O 2+ multiplet could excite moving interstellar and heliospheric O + ions. The total solar flux l~ in all the lines was assumed to be 5.3• c m "2 s -1. The continuum c m -2 S "1 nlTl "1 for the helium and oxygen contribution is 1.8x108 c m -2 S "1 nlI1-1 and 4.0• lines, respectively. This contribution is important for the glow of oxygen ions, but unimportant for helium. There are two other major sources of radiation at 30.4 and 83.4 nm, the solar wind produced foreground and the galactic background. The feasibility of heliopause mapping critically depends on spectral properties and relative strength (brightness) of this interfering radiation. One requires the LISM plasma glow (the "heliopause glow") to be brighter and/or spectrally separated from the background and foreground radiation.
- 265 -
M. Gruntman
0
2.E+11
0t.--
Q- ].E+I]
Figure 3. Solar emissions in the He+ (30.4 nm) and O+ (83.4 nm) resonance lines used in this work. Arrows mark the OII triplet lines; other nearby lines are of the OIII multiplet. The continuum is important for the oxygen irradiance, but unimportant for helium. The glow of the He + and O + ions in the solar wind would produce the foreground radiation. These ions are produced by ionization of interstellar neutrals penetrating the heliosphere. The newly formed ions are picked up by the solar wind flow and carried to the termination shock as the pickup ions. 11'12 The pickup ions are singly charged and characterized by a spherical shell velocity distribution function. The glow of the pickup ions is the line radiation spectrally similar to the LISM plasma glow. We calculated the number densities of the pickup ions describing the inflowing interstellar neutral helium and oxygen by the hot model. Interstellar helium is only slightly affected by the crossing of the heliospheric intreface region. In contrast, the properties of interstellar oxygen are significantly modified by the crossing, which requires use of modified gas parameters at infinity. The details of the calculations can be found elsewhere. 4'13 The glow of the pickup ions was calculated similarly to that of the LISM plasma beyond the heliopause. We used the spherical shell velocity distribution function in calculations of the gfactors. The observed asymmetry 14'15 of the ion distribution function would only slightly modify the expected pickup ion glow and was disregarded in this work. At 30.4 nm, there is another important source of the foreground, viz. the emissions produced in charge exchange between the solar wind alpha-particles (He 2+) and heliospheric atomic hydrogen. The existence of this emission was known for some time, 16'17 but only recently it was analyzed in detail. 18 This emission is spectrally separated from the glow of the LISM and pickup ions and would not interfere with heliopause mapping (see below). We also note here that the all-sky images in the charge-exchange emissions will remotely reveal the three-dimensional solar wind flow properties everywhere in the heliosphere, including in the regions over the sun's poles and on the other (from the observer) side of the sun. 18 Two main sources of the EUV background are the radiation emitted by hot interstellar plasmas (diffuse galactic background) and by the stars (stellar radiation field). The stellar
- 266-
Mapping the heliopause in EUV radiation field is a continuum with negligible line emissions, while hot plasmas emit a continuum with prominent line emissions. The stellar radiation dominates the EUV continuum background at wavelengths > 20 r i m . 19'20 The total (at least 90% complete) continuum radiation field is ~13 photon c m -2 s -1 nm -1 and ~9.3 photon cm -2 s 1 nrn -1 at 30.4 nm and 83.4 nm, respectively, z~ The stellar EUV Figure 4. Projections on the ecliptic plane of the most important background radiation is stellar sources (Adhara, G191-BZB, and Feige 24) of the highly anisotropic. 2~ A background radiation at 30.4 nm and 83.4 nm; X and [3 are the single bright star, ecliptic longitude and latitude, respectively. Also shown are the Canis Majoris (Adhara interstellar wind, solar apex, and the closest star, a Centauri. or Adara), produces most of the radiation at 83.4 nm. Two white dwarfs, Feige 24 and G 191-B2B dominate the background at 30.4 n m . 20 For isotropic background, the estimated stellar radiation 2~ field translates into the ~l.3x 10 .2 mR/nm and ~9.3x10 -3 mR/nm at 30.4 nm and 83.4 nm, respectively. (1 Rayleigh = 1 R = l0 s mR = 106 ~tR = 106/(4 ~) phot cm 2 s~ sr-1.) For observations in the directions other than toward these bright sources (Figure 4), the stellar continuum background would be significantly smaller. Hot (~106 K) interstellar plamas efficiently emit EUV radiation that includes both the line emissions and continuum. The sun is embedded in a relatively small, a few parsec long, and dense (~0.1 cm 3) local interstellar cloud. This local interstellar cloud (LIC), our LISM, is too cold (~7000 K) to emit EUV radiation. LIC is positioned in the center of a region, the Local Bubble, filled with hot and dilute plasmas. These plasmas would emit the EUV radiation that after partial absorption in the LIC would reach the sun. The hot plasma EUV background (line emissions and continuum) was calculated using a standard plasma emission model 21 with the temperaure 106 K, emission measure 0.0006 cm -6 pc, and 1018 cm -2 column density of the absorbing LIC hydrogen. 18 Charge exchange of the solar wind alpha-particles would form, with some probability, single charged helium ions in the metastable state. 18 The metastable ions would decay by a two-photon emission process contributing to the background for X>30.4 nm. Actually, this two-photon decay would dominate the continuum background in most directions in the 3 5-90
- 267-
M. Gruntman
nm wavelength range. 18 The contributions from various sources to the background EUV continuum are summarized in Table 1. Table 1 Background EUV continuum spectral radiance, mR/nm Wavelength
4.
Stellar field (isotropic)
Local Buble plasma emission
30.4 nm
1.3
10-2
6.7
10 -4
83.4 nm
9.3
10-3
4.3
10 -6
RADIANCE
Solar Wind charge exchange emission
(1.0-2.8)
10-3
AT 30.4 NM AND 83.4 NM
Figure 5 shows the calculated radiance of the glow of the LISM plasma beyond the heliopause and the glow of the pickup ions. The helium glow is much brighter (milliRayleighs) as compared to the oxygen glow (micro-Rayleighs). The initial slight increase of the LISM plasma brightness with the angle 0 is due to the Doppler effect as the velocity radial component diminishes in the plasma turning around the heliopause.The glow falls as the heliopause moves away from the sun and the Doppler effect reduces the g-factor. If the plasma had the constant number density, velocity and temperature beyond the heliopause, then the glow angular dependence would have been inversely proportional to the distance to the heliopause. This latter inverse radial dependence is shown by the solid curves. The difference between the solid curves and the calculated radiance (empty circles) illustrates the sensitivity of heliopause mapping to the plasma flow field beyond the heliopause. 18
16
/
14
E o C-
8
~
L
-ul
6
Figure 5. Glow at 30.4 nm and 83.4 nm of the LISM plasma beyond the heliopause and the solar wind pickup ions. The solid curves are the function 1/R~ normalized at 0 = 0, R ~ is the distance to the heliopause.
- 268 -
Mapping the heliopause in EUV The assumed ion-to-neutral ratios were 0.3125 and 0.5 for helium and oxygen, respectively. The brightness of the LISM plasma glow is roughly proportional to the number density of the interstellar gas ionized component, while the pickup ion glow is roughly proportional to the number density of the neutral component. Figure 6 illustrates this effect for oxygen, 4 showing the angular dependence of the radiance for three different ion-to-neutral ratios, 1"2, 1:1, and 2:1. Helium glow properties exhibit a similar dependence on the ionization state. By measuring the upwind-to-downwind brightness ratio one would establish the ionization state of helium and oxygen. 5. SPECTRAL RADIANCE
10
t
~
t
~
2 "1 ne = 0.07 cm^(-3) -
Figure 6. Sky radiance directional dependence at 83.4 nm for various ionization states of the LISM. Interstellar oxygen is assumed ionized similarly to hydrogen. Total LISM number density is 0.21 cm3, and the oxygen relative abundance is 7• 10-4.
The expected spectral radiance at 30.4 nm is shown in Figure 7. The LISM plasma and pickup ion glows are in practically the same spectral range. Most of the adjacent plasma line emissions (from the Local Bubble) are produced by interstellar OIII, AIIX, and SiXI ions. The continua of the plasma emission and the stellar radiation field are negligible. The
Figure 7. Typical spectral radiance at 30.4 nm. The glow of the LISM plasma beyond the heliopause is shown as white (empty) bars. Also shown are the pickup ion glow (light gray bars), Local Bubble interstellar emissions (black bars), and the solar wind charge-exchange emission (dark gray bars; the two peaks correspond to the fast/polar and slow/ecliptic solar wind).
- 269-
M. Gruntman
Doppler shifted solar wind chargeexchange emissions 18 are clearly spectrally separated from other emissions. Measurements of the heliopause glow with the spectral resolution-~0.025 nm would allow one to eliminate the interfering contributions of the solar wind emissions. In order to efficiently remove the contributions of the interstellar lines, one would require the resolution of 0.005 nm. Figure 8 shows the spectral radiance at 83.4 nm. The glows of the LISM plasma and the solar wind pickup ions are of comparable total radiance of several milliRayleighs. However, the spectral radiance of the LISM plasma glow is about one order of magnitude brighter since it is concentrated in the narrower spectral range (Figure 8a, b). The continuum radiation due to two-photon decay of the metastable ions in the solar wind 18 and stellar radiation field 2~ are not negligible (Table 1). The combined spectral radiance from all sources is shown in Figure 8c for a typical continuum 0.16 ~R/(0.01nm). The glow measurements with a 0.01nm spectral resolution would allow identification of the contributions from the LISM plasma and from the pickup ions. Much modest resolution of~0.1 nm, should provide the combined glow radiance. The recently developed EUV spectrometers EURD advanced the sensitivity of the diffuse radiation detection to 1 mR. 22 The new space mission CHIPS, presently under preparation, will study the diffuse galactic radiation with a sensitivity of ~1 mR/line at X< 26 nm. The proposed heliopause mapping requires, however, significantly higher (a factor of 100) spectral resolution at 30.4 nm and significantly higher sensitivity (a factor of 100) at 83.4 nm. Development of diffuse EUV radiation spectrometers with such advanced performance characteristics is a challenging but not impossible task.
8
E r-
o o --- 0 . 2
F
d O t--
._~ "~
0.1
I-k. O
O~
0.0
---
E t-
.c,-
o
2.0
~
1.5
do 1.0 t~
Figure 8. Spectral radiance at 83.4 nm. a) LISM plasma glow; b) pickup ions glow; c) combined spectral radiance including the continuum.
-270-
Mapping the heliopause in EUV
Mapping the all-sky radiance at 30.4 nm and 83.4 nm will establish the distance to and shape of the heliopause. The ionization states of interstellar helium and oxygen (and correspondingly hydrogen) will be obtained from such measurements. The asymmetry of the insterstellar magnetic field would also be pronounced and identified in the all-sky maps. The geocorona is bright at 30.4 nm and 83.4 nm, and too many photons would reach the night side through multiple scattering. 23 The heliopause mapping experiment can be performed only from a spacecraft outside the geocorona. REFERENCES:
1. S.T. Suess, Rev. Geophys., 28, 97-115, 1990. 2. V.B. Baranov and Yu. G. Malama, J. Geophys. Res., 98, 15157-15163, 1993. 3. M. Gruntman and H.J. Fahr, 25, 1261-1264, 1998. 4. M. Gruntman and H.J. Fahr, J. Geophys. Res., 105, 5189-5200, 2000. 5. M.A. Gruntman, Planet. Space Sci., 40, 439-445, 1992. 6. M. Gruntman, Rev. Sci. Instrum., 68, 3617-3656, 1997. 7. E.C. Roelof, this volume. 8. H.O. Funsten et al., this volume. 9. G.A. Doschek,, W.E. Behring, and U. Feldman, Astrophys. J., 190, L 141-L 142, 1974. 10. R.R. Meier, Geophys. Res. Lett., 17, 1613-1616, 1990. 11. E. Moebius et al., Nature, 318, 426-429, 1985. 12. G. Gloeckler et al., Science, 261, 70-73, 1993. 13. M.A. Gruntman, J. Geophys. Res., 99, 19213-19227, 1994. 14. L.A. Fisk et al., Geophys. Res. Lett., 24, 93-96, 1997 15. E. M6bius et al., J. Geophys. Res., 103,257-265, 1998. 16. F. Paresce et al., J. Geophys. Res., 86, 10038-10048, 1983. 17. M.A. Gruntman, Geophys. Res. Lett., 19, 1323-1326, 1992. 18. M. Gruntman, J. Geophys. Res., 106, N.A5, in press, 2001. 19. K.-P. Cheng and F.C. Bruhweiler, Astrophys. J., 364, 573-581, 1990. 20. J. Vallerga, Astrophys. J., 497, 921-927, 1998. 21. M. Landini and B.C. Monsignori Fossi, Astron. Astrophys. Suppl. Ser., 82, 229-260, 1990. 22. S. Bowyer, J. Edelstein, and M. Lampton, Astrophys. J., 485,523-532, 1997. 23. R.R. Meier, R.R., Space Sci. Rev., 58, 1-185, 1991.
-271 -
This Page Intentionally Left Blank
E n e r g e t i c N e u t r a l H y d r o g e n of Heliospheric O r i g i n O b s e r v e d w i t h SOHO/CELIAS
at 1 AU
M. Hilchenbach a K.C. Hsieh b D.Hovestadt c R.Kallenbach d A. Czechowskie E.MSbius f and P.Bochsler g aMax-Planck-Institut fiir Aeronomie, D-37191 Katlenburg-Lindau, Germany bphysics Department, University of Arizona, Tucson AZ 85721, U.S.A. CMax-Planck-Institut fiir Extraterrestrische Physik, D-85740, Garching, Germany dInternational Space Science Institute, Hallerstr. 6, CH-3012, Bern, Switzerland eSpace Research Centre, Polish Academy of Sciences, Bartycka 18A, PL 00-716 Warsaw, Poland flnstitute for the Study of Earth, Oceans, and Space, University of New Hampshire, Durham, NH 03824 gPhysikalisches Institut, University of Bern, CH-3012, Bern, Switzerland
The High-Energy Suprathermal Time-of-Flight sensor (HSTOF) of the Charge, Element and Isotope Analysis System (CELIAS) on the Solar and Heliospheric Observatory (SOHO) near the Lagrangian point L1 is capable of observing energetic hydrogen atoms (EHAs). The EHAs are accumulated from 1996 to 2000 under quiet interplanetary conditions and the observed EHA flux level, their energy spectrum and their anisotropy will be discussed. 1. I N T R O D U C T I O N In 1992 Hsieh et al. [1] proposed to study the acceleration and propagation of the anomalous cosmic rays (ACRs) in and out of the heliosphere via energetic neutral atoms (ENA) originating from ACRs charge exchange with atoms of the local instellar medium (LISM). In 1995, Czechowski et al.[2] modelled the expected ENA flux originating from the ACRs in the outer heliosphere and predicted an apex- tail flux anisotropy. In 1998, Hilchenbach et al. [3] observed energetic hydrogen atom (EHA) fluxes during quiet-time interplanetary conditions in 1996 and 1997 with a high flux anisotropy coming from the approximate tail direction of the heliosphere, e.g. as predicted for ACR produced EHAs. In this follow-up study we address the in-flight calibration of the HSTOF instrument, the selection of interplanetary quiet-time periods and the energy spectrum of the EHAs.
-273-
M. Hilchenbach et al.
Figure 1. In-flight cross calibration of SOHO/CELIAS/HSTOF with ACE/ULEIS and ACE/EPAM [4].
Figure 2. Proton and EHA flux for 'less quiet' times (826 days). About a third of the low energetic particle events can be attributed to transmitted protons.
2. M e t h o d s and C a l i b r a t i o n
To determine the flux of EHAs requires accurate knowledge of the HSTOF detection efficiency in order to subtract the low-energy protons from the particle spectrum measured with HSTOF. The remaining events are either due to EHAs or accidental coincidence events. More details about this method and its rationale as well as the instrument are described in Hilchenbach et al. [3]. For energetic-particle events we compared proton fluxes measured with HSTOF with fluxes measured by instruments on the Advanced Composition Explorer (ACE, Level 2 Data, time interval DOY 301/1998 to DOY 99/2000 [4] ). ACE is, as SOHO, on a halo orbit near the Lagrangian point L1. With this comparison we could determine the inflight efficiency of HSTOF for protons. It turned out to be ~ 10 times lower than expected from pre-flight calibrations (Fig. 1). The origin of this large discrepancy is currently under investigation. The HSTOF efficiency did not vary systematically beyond a level of about ~ 15% since launch (data not shown). 3. R e s u l t s For the present study, we analysed HSTOF data in the time interval from DOY 43/1996 to DOY 99/2000. We selected interplanetary quiet-time periods according to the registered events in the mass = 1 and in the 5 0 - 6 6 0 k e V energy range: < 12 cts/day for 'very quiet', < 20 cts/day for 'quiet' and < 100 cts/day for 'less quiet' time periods. In the 'less quiet' times, we found that about a third of the particles in the 5 8 - 8 8 k e V energy range could be attributed to transmitted quiet-time protons (Fig. 2). For the quiet-time protons
-274-
Energetic neutral hydrogen of heliospheric origin... .
--*--
:
.
.
EHA
.
i
.
.
.
.
] r~
:
,4
~5 1 0 .4
(keV)
Figure 3. Energetic particle spectrum at 'quiet' times (555 days). Compared to Fig. 2 the proton flux is reduced due to the lowered quiet-time selection threshold.
Figure 4. Particle spectrum for 'very quiet' times (401 days). Compared to Fig. 2 and Fig. 3, the EHA flux is consistent.
we found a power law with 7 = -2.5 which agrees with previous observations (Richardson et al. [5]). Using only data from quiet and very quiet-times reduces the proton flux below the ~ 10% level. The EHA flux in the energy range 58 - 88keV remains unaffected (Fig. 3 and Fig. 4). Again, the quiet-times reveal an anisotropic flux arising from the tail region of the heliosphere (Fig. 5). For the newer data reported here, the quiet-time EHA observations were hampered by the 3-month loss of SOHO in 1998 as well as the return of enhanced solar activity. The energy spectrum of the EHAs during very quiet-times originating from the apex and tail regions is shown in Fig. 6 (tail region DOY 170-220 and apex region DOY 1-169, 221-365(366)). Calculations based on Kausch's heliospheric model (Fahr et al. [6], multiplied by factor of 10 for the tail and 40 for the apex region) are shown for comparison (dotted lines in Fig. 6). 4. D I S C U S S I O N We cross calibrated the HSTOF instrument in flight with instruments on the ACE satellite. The reduced sensitivity of HSTOF transforms into an EHA flux which is higher than assumed previously [3]. The cause is still under consideration, e.g. part of the timeof-flight foils could have been altered during the SOHO launch. We checked the quiet-time threshold criteria for the EHA analysis. We found that the proton flux is a function of the selection threshold and the EHA flux is independent. The EHA flux observed by HSTOF is about 10 times higher than previously explained by model calculations (e.g. Czechowski et al.[7]). Furthermore, recent modelling of another EHA source, originating from accelerated protons in co-rotating interaction regions (CIRs), show a significant
-275-
M. Hilchenbach et al.
Figure 5. Quiet-time 5 8 - 8 8 k e V EHA flux as observed by HSTOF (left panel), the background level plotted for comparison (right panel) and the heliocentric ecliptic longitude coordinate system (top labels, heliotail at 74 ~
Figure 6. EHA energy spectrum for 'very quiet times' (tail and apex regions).
EHA flux anisotropy from the direction of the LISM helium focusing cone, overlapping the heliospheric tail region (Kdta et al., [8]). We acknowledge the work of the people involved in the ACE ULEIS and EPAM instruments who made the CELIAS/HSTOF in-flight calibration study possible [4]. We thank the referee, R.F. Wimmer-Schweingruber, for his helpful comments. REFERENCES
1. 2. 3. 4. 5. 6. 7. 8.
K.C. Hsieh, K.L. Shih, J.R. Jokipii and S. Grzedzielski 1992: ApJ, 393, 756 A. Czechowski, S. Grzedzielski and I. Mostafa 1995: A&A, 297, 892 M. Hilchenbach, K.C. Hsieh, D. Hovestadt et al. 1998: ApJ, 503, 916 ACE Science Center Level 2 Data, http://www.srl.caltech.edu/ACE/ASC/ I.G. Richardson,D. V. Reames,K.-P.Wenzel,J.Rodriguez-Pacheco 1990: ApJ, 363L, L9 H.J. Fahr, T. Kausch and H. Scherer 2000: A&A, 357, 268-282 A.Czechowski, H.Fichtner, S.Grzedzielski et al. 2001: A&A, 368, 622 J. Kdta et al. 2000" JGR, submitted
-276-
Acceleration of pick-up ions at the termination shock in the limit of weak scattering S.V. Chalov ~* and H.J. Fahr bt ~Institute for Problems in Mechanics of the Russian Academy of Sciences, Prospect Vernadskogo 101-1, 117526 Moscow, Russia UInstitut fiir Astrophysik und Extraterrestrische Forschung der Universitgt Bonn, Auf dem Hiigel 71, D-53121 Bonn, Germany It is generally agreed that interstellar pick-up ions constitute a seed population for anomalous cosmic rays (ACRs) originating at the solar wind termination shock (TS) through the shock-drift acceleration and first-order Fermi mechanism. Acceleration of ACRs at the TS is usually described in the limit of strong scattering when the cosmic-ray velocity distribution function is almost isotropic and continuous through the shock front. We consider here the opposite case of weak scattering taking into account anisotropy of the velocity distribution which can be of great consequence for the acceleration process at least at low energies. In order to describe the motion of pick-up ions in the upstream and downstream parts of the solar wind flow near the TS the Fokker-Planck transport equation for anisotropic velocity distribution function is used, while through the shock front the conservation of the magnetic moment of particles is assumed. It is shown that downstream spectra of accelerated pick-up ions close to the ecliptic plane depend strongly on longitude due to longitudinal dependence of the angle between the shock normal and heliospheric magnetic field connected with departure of the TS from sphericity. 1. I N T R O D U C T I O N
Acceleration of ACRs at the TS is often considered in the diffusive approximation which is valid only in the case of small pitch-angle anisotropy. However, the velocity distribution of particles in the vicinity of the TS can be essentially anisotropic if pitchangle scattering is not too strong, so that the application of the diffusive theory is highly problematic at least for low and moderate energy particles. As an example we refer to the recent paper [1]. In the paper the differential energy spectra of 50-keV to 20 MeV protons accelerated at forward and reverse shocks at corotating interaction regions observed at Ulysses during 1992 and 1993 have been compared with the predictions of two models based on diffusive shock acceleration theory [2,3]. It has been shown in [1] that the *SVC was partially supported by Avard No. RPI-2248 U.S. Civilian Research and Development Foundation, RFBR Grant 99-02-04025, and DFG Grant 436 RUS 113/110/6-2. tHJF was partially supported by DFG Grant Fa-97/24-2 and RFBR Grant 99-02-04025.
-277-
S. V. Chalov and H.J. Fahr
observations are inconsistent with the main predictions of the theory. Specifically, the observed spectra are regularly much harder than predicted. In the present paper we present results concerning acceleration of interstellar pick-up protons at the TS up to energies of ACRs taking into account anisotropy in their velocity distribution in the vicinity of the shock. An essential point of our model is preacceleration of pick-up protons by solar wind turbulence. 2. S T O C H A S T I C
AND REGULAR
ACCELERATION
OF P I C K - U P
IONS
The transport of pick-up ions in the heliosphere up to the TS is described by the well-known transport equation for the isotropic velocity distribution function including continuous production of pick-up ions, their convection with the solar wind velocity U, adiabatic deceleration, and energy diffusion. It should be pointed out that isotropy of the velocity distribution is assumed here only for particles which did not suffer action of the TS. Solar wind turbulence is considered as consisting of small-scale Alfv(~nic turbulence and nonlinear large-scale fluctuations in the solar wind velocity and magnetic field. Thus the energy diffusion coefficient can be written in the form: D - DA + DL, where DA and DL depend on the intensities of corresponding fluctuations ~AE - - ( ( ~ B ~ E ) / S ~ and Cbs - (aU~s)~/2/Us at 1 AU (for more details, see [4]). Figure 1 shows fluxes (in the solar wind rest frame) of accelerated pick-up protons in front of the TS (rsh -- 90 AU in the upwind direction). The solar wind and LISM velocities and number densities are adopted to be Us - 450 km s -1, n~,s - 7 cm -a, cm -a, respectively. Curve 1 corresponds -- //H,LISM -- 0.14 V L I S M - - 26 km S - 1 , • e , L I S M to the case when only Alfv(~nic turbulence is taken into account" ~AE - - 0 . 4 , (~LE - - 0. Curves 2 and 3 show fluxes in the case when the level of Alfv(~nic turbulence is lower (Cas - 0.2) but acceleration of large-scale fluctuations is included (~LS -- 0.3 and 0.5 for curves 2 and a, respectively). One can see that the large-scale fluctuations have a dominant role in formation of high energy tails in spectra of pick-up ions accelerated by solar wind turbulence. The preaccelerated pick-up ions experience further acceleration at the TS. We consider here the TS as a discontinuity neglecting the cross-shock potential. Then in the case of weak scattering considered here one can use the assumption of conservation of the energy of pick-up ions in the de Hoffmann-Teller frame and the assumption that the magnetic moment of a particle is the same before and after the encounter with the shock. From these two assumptions one can obtain conditions under which particles are transmitted through the shock or reflected from the shock due to abrupt change of the magnetic field (see [5]). Due to the reflection particles gain energy and move along magnetic field lines away from the shock front. However, the interaction of the particles with upstream Alfv~nic turbulence leads to their pitch-angle scattering, and the result is multiple reflection from the TS. In order to describe the process of the multiple reflection the following transport equation for anisotropic velocity distribution function of pick-up ions f(t, x, v, #) in the vicinity of the TS is considered here"
Of
at +
(Ux +
Of _
x)b-7
(1)
f + O(x, v),
-278-
Acceleration of p&kup ions at the termination shock...
. . . . . . . .
~"
1E+2
1E+1
1E+t
1E+0
|
. . . . . . . .
!
. . . . . . . .
O
.-
t/J tg)
le-1 E ._o_
E
1E-1
O
1E-2
.,..,
1E-3
~-
1E-4
\
1E-4
III
"-a
x ii
1E-6 |
1E-7 1E-1
. . . . . . .
i
. . . . . . . .
i
. . . . . . . .
1E+1 KINETIC E N E R G Y (keV)
Figure 1. Fluxes of accelerated by solar wind turbulence pick-up protons in front of the TS. 1 - r = 0.4, ~LE = 0; 2 - r = 0.2, (Ls = 0.3; 3 - ~ , s = 0.2, (Ls = 0.5.
Figure 2. Downstream fluxes of accelerated pick-up protons. 1 - ~b - 70 ~ 2 - ~b - 80 ~ 3 - ~b - 85 ~ , 4 - ~b - 89 ~ .
where the z-axis is directed perpendicular to the shock which is considered to be planar, v is the velocity of pick-up ions in the solar wind rest frame, # is the cosine of the pitchangle of ions, X is the cosine of the shock normal angle ~b, the source Q corresponds to pick-up ions arriving at the TS (see Figure 1), and S f is the scattering operator including pitch-angle scattering and energy diffusion (see [6]). In the following, we consider regions of the heliosphere close to the ecliptic plane. Since the shape of the TS is not spherically symmetric due to the interaction of the solar wind with the moving interstellar medium the shock normal angle ~b is a function of longitude measured from the upwind direction. In the upwind and downwind direction ~b = 90 ~ while at the flanks of the TS ~b can reach 65 ~ [7]. Figure 2 shows downstream fluxes of pick-up protons accelerated at the TS for different shock normal angles. The initial upstream flux is given by curve 3 in Figure 1. The level of Alfv~nic turbulence in front of the TS is adopted to be CA = 0.01 (weak scattering). The key distinction of fluxes in Figure 2 from fluxes predicted by the diffusive theory is their nonmonotonic behaviour. In the case of weak scattering the downstream fluxes can be considered as consisting of two parts. The low energy parts are formed by protons which were transmitted through the shock and did not experience multiple reflections. The high energy parts are formed by protons which experienced multiple reflections at the shock. The strong dependence of the fluxes on the shock normal angle and, therefore, on longitude can be seen in Figure 2. Downstream fluxes at the flanks of the TS exceed fluxes at its nose part in the energy range from about 100 keV to several MeV. It is connected with increasing reflection efficiency at the shock with decreasing ~b.
- 279-
S. K Chalov and H.J. Fahr
,
,
|
,
,
|
,
,
|
,
,
,
|
,
,
1E-4 <
E "
"
"
1 E-5 ,,.., i_ txl
100 iii .,,_..
1 E-6
-.l
X 2 3
--
1 E-7
i
,
,
i
,
,
i
,
,
i
,
,
|
,
,
LI_
kONGITUDE
(degrees)
Figure 3. Fluxes of energetic neutral hydrogen at 1 AU as functions of longitude counted from the upwind direction. The numbers signify the energy in keV. The squares show differential fluxes of neutral hydrogen in the energy range from 55 to 80 keV detected by CELIAS/HSTOF on SOHO.
3.
ENERGETIC
NEUTRAL
HYDROGEN
FLUXES
Charge exchange between neutral interstellar hydrogen and accelerated pick-up protons in the region between the TS and heliopause results in formation of energetic atoms which can be detected at the Earth's orbit. Figure 3 shows calculated fluxes of energetic neutral hydrogen at 1 AU. The calculations are based on a simplified Parker model of the plasma flow in the heliosheath region. Spatial diffusion of pick-up protons is not taken into account and their velocity distribution function is assumed to be conserved along the streamlines of the flow. The shape of the TS has been taken from [7]. The labels at the curves signify the energy in keV. Pronounced longitudinal anisotropy of fluxes which tend to increase toward the downwind direction can see in Figure 3. The squares show differential fluxes of neutral hydrogen in the energy range from 55 to 80 keV detected by CELIAS/HSTOF on SOHO [8]. It is clear that our simplified model of production of energetic atoms in the heliosheath region can display the behaviour of the observed fluxes only qualitatively. Specifically, spatial diffusion along the magnetic field lines will result in some longitudinal fl~ttening of the calculated fluxes. REFERENCES 1.
2. 3. 4. 5. 6. 7. 8.
M.I. Desai, R.G. Marsden, T.R. Sanderson, et al., J. Geophys. Res. 104 (1999) 6705. L.A. Fisk and M.A. Lee, Astrophys. J. 237 (1980) 620. F.C. Jones and D.C. Ellison, Space Sci. Rev. 58 (1991) 259. S.V. Chalov, H.J. Fahr, and V. Izmodenov, Astron. Astrophys. 320 (1997) 659. R.B. Decker, Space Sci. Rev. 48 (1988) 195. S.V. Chalov and H.J. Fahr, Astron. Astrophys. 360 (2000) 381. V.B. Baranov and Yu.G. Malama, J. Geophys. Res. 98 (1993) 15157. M. Hilchenbach, K.C. Hsieh, D. Hovestadt, et al., Astrophys. J. 503 (1998) 916.
- 280 -
Doppler Shifted Photon Emission expected due to Reactions of Energetic Protons with the LISM Atoms in the Heliosphere M. Hilchenbach a K.C. Hsieh b and A. Czechowski c aMax-Planck-Institut fiir Aeronomie, D-37191 Katlenburg-Lindau, Germany bphysics Department, University of Arizona, Tucson AZ 85721, U.S.A. CSpace Research Centre, Polish Academy of Sciences, Bartycka 18A, PL 00-716 Warsaw, Poland The anomalous cosmic ray (ACR) protons interact with the atoms of the local interstellar medium (LISM) in the outer heliosphere. As a function of the reaction cross-section and particle velocities, the ACR protons can capture an electron from the neutral LISM atom into an excited state. Photons, Doppler-shifted relative to the observer due to the velocity of the excited energetic hydrogen, will be emitted. We estimate the observational significance of these elemental atomic processes as a tool to observe the solar wind plasma parameters beyond the termination shock. The faint Doppler-shifted Lyman-c~ flux is at about 50 photons per .~ per day with a space-borne 2-meter telescope at 1 AU. 1. I N T R O D U C T I O N The possibility of observing the outer heliosphere via neutralized anomalous cosmicray (ACR) ions was first discussed by Hsieh et al. [1]. The low energy ACR ions cannot penetrate upstream of the solar wind termination shock due to the interplanetary magnetic field. However, energetic neutral atoms (ENA) originating from ACR ions neutralized by charge-exchange with the atoms of the local instellar medium (LISM) in the outer heliosphere are detectable in the inner heliosphere, e.g. at 1 AU ( Czechowski et al.[2], Hilchenbach et al. [3]). In the course of the atomic collision processes either one of the collision partners may become excited either by direct or by charge exchange excitation processes. For sufficiently large incident energies, no restrictions with respect to the transfer of excitation energy are imposed and any, e.g. hydrogen, state may result during the collision. The atomic processes can lead to the formation of excited target and/or, via capture, excited projectile states (Hippler [4], Brendan et al. [5]]), and the observation of the photons has been proposed for solar flare studies (Orrall and Zirker [6]) and Lyman-c~ polarisation due to this atomic process has been observed (Emslie et al. [7]). We note that Meinel [8] discovered the presence of energetic H in aurorae by the detection of Doppler shifted Hc~ line from the ground before the Space Age. The outer heliosphere might faintly glow due to the photons of the Lyman or even the Balmer series of the collisionally excited hydrogen atoms. As for the ENA, these photons could give the possibility to probe remotely the
-
281
-
M. Hilchenbach, K. C. Hsieh and A. Czechowski
Figure 1. Angular distribution of ACR proton flux of different energies in the ecliptic. The flux is modelled according to Kausch's numerical heliospheric model (with LISM hydrogen density nH set to 0.1 cm -a)
Figure 2. Emission of hydrogen lines generated through charge exchange and capture into an excited state. The energetic ACR ions gyrate around the interstellar magnetic field lines. The photons are Dopplershifted relative to the observer.
plasma and structure of the outer heliosphere. We will estimate the photon flux due to this source and discuss the possibility of observations. 2. M e t h o d s The flux of the ACR protons and the density of the LISM is modelled along the lines of Kausch's heliospheric model (Fig. 1), which is based on the numerical solution of the gas-dynamical equations obtained by Fahr et al. [9]. In this model the termination shock is nonspherical, the flow has nonzero divergence, and the density of the neutral hydrogen from the LISM inside the heliopause is reduced by a factor of 3-4 due to interaction with the solar wind plasma. Kausch's model is basically two-dimensional and axially symmetric to the LISM apex-antiapex axis. The non-thermal emission of hydrogen lines generated through charge exchange by ACR protons gyrating around the interstellar magnetic field is computed. The atomic process rates are modelled along the Chebyshev fit parameters given by Barnett et al. [10]. The magnetic field is assumed to have no preferential orientation. The LISM velocity is at 26 km/s. The plasma velocity just outside the shock is 80 km/s (apex) to 120 km/s (tail). In the heliotail the velocity drops to 50 km/s at 200 AU from the shock and increases again to 80 km/s at 1000 AU. The ACR velocities are assumed to be isotropic in the frame co-moving with the plasma. Computations have been carried out for observations made near 1 AU. In such conditions, the photons emitted through charge exchange are Doppler shifted relative to the observer in both wings of the emission line (Fig. 2). We did not consider the polarization of the emission lines.
-
2 8 2
-
Doppler shifted photon emission expected due to...
-~'
H" 10
,
4
-,'/ 10.6
"-::2.,.,~,
,,,
....
,
....
; ....
; . . . . . . . . .
,
....
i
....
i
....
=
:
...:.x.',
.fi...
T''
Figure 3. Photon emissions from the Apex direction give raise to a faint photon flux shifted off the Lyman-c~ emission line.
Figure 4. As Fig 3, but for heliospheric Tail direction.
3. Results The convolution of the photon emissions from the outer heliospheric regions in the Apex and Tail direction give rise to a faint photon flux shifted off the Lyman-c~ emission line. We calculated the contributions of the H(2p) as well as the metastable H(2s) states (the later has a lifetime of about 0.1 sec). The expected flux is an order of magnitude larger from the Tail region than from the Apex region of the outer heliosphere (Fig. 3 and Fig. 4, for Apex and Tail regions, respectively). Due to the plasma velocity beyond the termination shock, the Doppler-spread emission is not symmetric for the red and blue wings of the photon emission. The intensity ratio is about 0.8 to 0.9, (Fig. 5) and a function of the plasma velocity. The red Balmer Hc~ line is visible and might therefore be observed with earth borne telescopes, we calculated the very faint photon flux expected due to the charge exchange and excitation of the ACR protons with He in the LISM (Fig. 6). 4. D I S C U S S I O N The faint glow of the outer heliosphere due to charge exchange reactions into excited states is a direct observational tool to determine the parameters describing the ACR protons as well as the LISM (for example, plasma velocity beyond the termination shock, density and velocity distribution of ACR protons). The blue and red shift off the Lyman-c~ line from the center is 4 to 20 A. Therefore, the interference with the scattered solar Lyman-c~ emission, for example, is marginal. However, the expected photon flux is very faint. A space-borne telescope of about 1 m
- 283 -
M. Hilchenbach, K.C. Hsieh and A. Czechowski
Figure 5. Due to the plasma velocity beyond the termination shock, the Dopplerspread emission is not symmetric for the red and blue wings of the emission (Fig. 4).
Figure 6. Balmer Ha line in the visible regime of the spectrum (red). The calculated photon flux expected due to the charge exchange and excitation of the ACR protons with He in the LISM is very faint.
radius would just be able to collect about 10 to 100 photons per ~ per day shifted off the Lyman-c~ line. For the Balmer line, even with a large telescope on earth, one would expect less than one photon per ~ and hour. REFERENCES
.
2. 3. 4. 5. .
7. 8. 9. 10.
K.C. Hsieh, K.L. Shih, J.R. Jokipii and S. Grzedzielski 1992: ApJ 393, 756 A. Czechowski, S. Grzedzielski and I. Mostafa 1995: A&A 297, 892 M. Hilchenbach, K.C. Hsieh, D. Hovestadt et al. 1998: ApJ 503, 916 R. Hippler 1993: I. Phys. B: At. Mol. Opt. Phys. 26 1-42 M. Brendan, T.G. McLaughlin, G. Winter and J. F. McCann 1997: J. Phys. B: At. Mol. Opt. Phys. 30, 1043 F.Q. Orrall and J.B. Zirker 1976: ApJ, 208, 618 A.G. Emslie, J. Miller, E. Vogt , J. Henoux, S. Sahal-Brechot 2000: ApJ, 542,513 A.B.Meinel 1951: ApJ, 113, 50 H.J. Fahr, T. Kausch and H. Scherer 2000: A&A 357, 268-282 C.F. Barnett et al. 1990: Atomic Data for Fusion, Oak Ridge National Laboratory, ORNL-6086
- 284
-
A new diagnosis tool to map the outer heliosphere regions Romana Ratkiewicz ~ and Lotfi Ben-Jaffel b ~Space Research Center, Bartycka18A, 00-716 Warsaw, Poland, bInstitut d'Astrophysique de Paris, CNRS 98 bis, Blvd Arago, F-75014 Paris, France This paper is an extended abstract summarizing the conclusions of the papers by [1] and [2]. Sophisticated 3D MHD modeling of the solar wind (SW) interaction with the local interstellar medium (LISM) including a general orientation of the interstellar magnetic field and charge exchange between protons and H neutrals, is a necessary tool to interpret properly combined data sets from both deep space observatories (Voyager 1 & 2) and Earth-based high resolution spectroscopy (HST) in the far UV. This approach to analyze the data should provide a new diagnosis tool that would help to map the 3D shape of the interface region between the SW and the LISM in a proper way, using only particular properties of the Ly-c~ emission line of the medium. We try to show how the orientation of the interstellar magnetic field could then be derived from that mapping. Consequently this will provide the key parameters to interpret distribution of neutral (H, He, ENA, etc) and ion (pickup ions, ACR, etc.) components measured by in situ spacecrafts in the inner heliosphere. In order to achieve this goal the pure 3D MHD model ([3], [4]) has been enriched by neutral particles influence on the interaction between the solar wind and the magnetized interstellar plasma ([2]). The interaction with neutral hydrogen appears via charge exchange between protons and H atoms in each region of the heliospheric interface and inside the termination shock. In a first attempt, the neutral hydrogen parameters such as the number density, velocity and temperature are assumed to be constant throughout the simulation. The charge exchange cross-section is also taken constant everywhere. It has been shown ([2]) that the interaction of the spherically symmetric solar wind with the magnetized interstellar plasma in the presence of a constant neutral hydrogen flux leads to the following phenomena: the supersonic solar wind is heated by the inclusion of pickup ions created through charge exchange with hot neutrals, and the supersonic solar wind decelerates. The heating and deceleration imply that the sound speed increases, and the hydrodynamic Mach number decreases with increasing heliocentric distance. The shape and size of the termination shock change, the distances to the boundaries as termination shock (TS), heliopause (HP), and the bow shock (BS) are greatly reduced. The main features of asymmetries introduced by the interstellar magnetic field are the same as in the no charge exchange case. [1] have made the first step to correlate the data from Voyager 1 (in situ) and HST (remote). By combining the Baranov model ([5], [6]) with the Newtonian approximation
-285-
R. Ratkiewicz and L. Ben-daffel
approach ([7]), and the sky background measurements they proposed how to derive the inclination angle of the interstellar magnetic field. Voyager 1 UVS recorded the L y - c~ brightness distribution shown in Figure 4 ([1]) in a plane that contained both the upwind direction and the direction defined by (~ = 284 ~ /3 = -14~ This distribution shows a 20 ~ deviation from upwind for the maximum of the intensity. The comparison of the Fermi intensity model described by [1] to the UVS emission excess reveals that a Mach number M~ 1.8 bow shock exists for ~27% of the H neutrals, and that the nose of the heliopause makes an angle of 12~ from upwind. According to the ([6]) model, this 27% percentage of H neutrals that exchanged their charge with shocked interstellar protons is consistent with the initial unperturbed interstellar proton density of ~ 0.043 cm -a. [1] have translated the 12 ~ deviation of the heliopause nose from the upwind axis to an inclination of the magnetic field using the Newtonian approximation ([7]). For that purpose, they have calculated the Alfv~n velocity assuming a 1.8 p G magnetic field in the range proposed so far for the local interstellar medium ([8]; [9]), and a 15 % ionization rate that corresponds to a proton density of 0.043 c m - 3 ([10]). Using Equation (67) of [7], they have found that the interstellar magnetic field is inclined almost ~ 40 ~ with respect to the upwind direction. It has been stressed that the 3D MHD calculations ([3], [4]) of the plasma interaction between the solar wind and the LISM, including a general orientation of the magnetic field in space, have shown that the fine structure of the interface region is much more complex than depicted by the Newtonian approximation. At present the new 3D MHD model which includes the influence of the neutrals ([2]) can be used to properly translate the drop-off velocity distribution revealed in the frame of the Fermi model to accurate intrinsic properties of the interstellar flow. To go one step further and improve the diagnostic quality regarding the strength and the orientation of the interstellar magnetic field, we propose to use the new 3D MHD+N model ([2]) in place of the axisymmetric model of Baranov, and follow the same procedure as described by [1] but now with orientation of magnetic field included in the model in a self-consistent way. Qualitatively, the procedure could be described as follows: 9 (step 1) derive the radial velocity of plasma (versus B , n v , n i l , etc.) in the same conditions of observational geometry as obtained by Voyager UVS, 9 (step 2) describe the neutral particles velocity as unperturbed one plus a contribution from plasma through charge exchange process (from step 1), 9 (step 3) use the Fermi model as proposed by [1] to derive the corresponding Fermi light curve in the same geometry of observation as obtained by UVS, 9 (step 4) compare to UVS data and adjust the plasma parameters (go to step 1). Because the final product of this new diagnostic tool will provide the interstellar magnetic field strength and orientation in a self-consistent way, the scenario of the interaction of the interstellar flow with the solar wind as proposed by [1] could be revisited and updated to provide a better estimate of the magnetic field effects on the intrinsic properties of the different regions from the outer heliosphere down to the inner solar system.
-
2 8 6
-
A new diagnosis tool to map the outer heliosphere regions Acknowledgments. Authors acknowledge support from the Centre National de la Recherche Scientifique of France (CNRS) and Polish Academy of Sciences (PAN) under Program 5037 (CNRS-PAN protocol) and program Jumelage Pologne. RR acknowledges the support in part by the KBN Grants No. 2P03C 004 14 and 2P03C 005 19, and by the Max-Planck-Institut fiir Aeronomie, Katlenburg-Lindau, Germany. REFERENCES
L. Ben-Jaffel, O. Puyoo and R. Ratkiewicz, Ap.J. 533 (2000) 924. 2. R. Ratkiewicz, A. Barnes, H.-R. Miiller, G.P. Zank and G.M. Webb, Adv. Space Res. in press (2000b) R. Ratkiewicz, A. Barnes, G.A. Molvik, J.R. Spreiter, S.S. Stahara, M. Vinokur and S. Venkateswaran, Astronom. Astrophys. 335 (1998) 363 R. Ratkiewicz, A. Barnes and J.R. Spreiter, J. Geophys. Res. 105 (2000a) 25,021 5. V. B. Baranov and Y.G. Malama, J. Geophys. Res. 98 (1993) 15,157 6. V. B. Baranov and Y.G. Malama, J. Geophys. Res. 100 (1995) 14,755 7. H. J. Fahr, S. Grzedzielski and R. Ratkiewicz, Ann. Geophys. 6 (1988) 337 8. G. Gloeckler, L.A. Fisk and J. Geiss, Nature 386 (1997) 374 9. A. G. Lyne and F.G. Smith, MNRAS 237 (1989) 533 10. O. Puyoo and L. Ben-Jaffel, Lecture Notes in Physics 506 (1998) 29 .
o
.
-287-
This Page Intentionally Left Blank
The Lyman-Alpha echo from the heliospheric bow shock region and its observability from earth H.J.Fahr, H.Scherer, G. Lay a, and M. Bzowski b ~Institut fiir Astrophysik, Universit/it Bonn, Auf dem Hiigel 71, D-53121 Bonn (Germany) bSpace Research Centre of Polish Academy of Sciences, Bartycka 18A, 00-716 Warsaw (Poland) In first line the solar system is adapting to the counterflowing interstellar medium on the basis a magnetohydrodynamic plasma-plasma interaction. In second line then the neutral interstellar H-atoms are dynamically coupling to this resulting plasma configuration by charge exchange collisions with protons and ,in the region ahead of the heliopause, are piling up in density to form an H-atom wall due to the charge-exchange-induced friction with the stagnating plasma flow. Here we investigate the spectral intensity distribution of solar H-Lyman-Alpha photons resonantly backscattered from this upwind H-atom structure to the inner heliosphere. We consider the change of the solar emission profile with increasing optical thickness at growing solar distance and the thermodynamical properties of H-atoms in the wall structure in order to answer the question whether or not the Hubble Space Telescope spectrometer GHRS would be able to detect this spectral feature from the earth's orbit. 1. D i a g n o s t i c s of the c i r c u m s o l a r H - d i s t r i b u t i o n by d e t e c t i o n of its L y m a n A l p h a glow
Interstellar H-atoms passing over the solar system are resonantly excited by the solar Lyman-Alpha line emission and re-emit resonantly scattered photons into all directions according to a phase probability function. Looking with a Ly-c~ detector into a specific direction one collects photons which have been scattered from H-Ly-c~ sources along that specific line of sight. If the detector spectrally resolves the arriving Ly-c~ photons, then information is obtained not only on density, but in addition on velocity and temperature of H-atoms along the line-of-sight (LOS). Such Ly-c~ spectra were obtained in the recent past with the GHRS spectrometer of the Hubble Space Telescope (Clarke et al., 1995,199813][4]) and were interpreted by theoretically modelled spectra calculated by Scherer et al. (1997, 1999, 2000)[10][11][12]. The interpretation of the five different spectra taken from different positions and different directions did all show a dominant geocoronal Ly-c~ peak and a secondary interplanetary peak which was either redshifted or blueshifted, dependent on the LOS-direction. The simultaneous interpretation of the five Doppler-shifted interplanetary peaks led to the determination of a set of bestfitting in-
- 289-
H.-J. Fahr et al.
terstellar H-atom gas parameters with some residuals left unexplained. Recently Scherer et al. (2000)[12] have shown that careful attention to the variability in the solar line emission profile will partly reduce these residuals, though not complete. Most recently now Ben-Jaffel et al. (2000)[2] have reprocessed the spectral data of GHRS/HST which had led to the two earlier upwind profiles obtained by Clarke et al. (1998)[4] and so have surprisingly found that instead of two they can identify even three spectral peaks in the upwind Ly-c~ spectra, one being very inferior and blueshifted from the earlier interplanetary one. The interpretation given by these authors with concern to this additional, third peak is connected with an expected Ly-c~ echo from the H-atom wall structure outside of the heliopause. Disregarded the reality of this third peak we shall study in the following what spectral emission feature can be expected as the Ly-a echo from the H-atom wall. 2. L y m a n - A l p h a p h o t o n s b a c k s c a t t e r e d f r o m t h e H - a t o m wall In our calculations we shall make use of the kinetic description of the heliospheric Hatom distribution given by Osterbart and Fahr (1992)[8] or Fahr et al. (1993)[5] as solution of an underlying Boltzmann integro-differential equation. From this model description one obtains local values of density, bulk velocity, and temperature of H-atoms within a 1000 AU sphere around the sun. The resonant excitation of H-atoms in this sphere is caused by the solar Ly-a line emission with a self-inverted Gaussian profile given by:
where according to OSO-8 data we have used the following parameters" a -7.96; Aa = 0.351 ~; b - 5.69; Ab - 0.2 ~. The above profile due to optical depths effects is not simply reduced in spectral intensity by the geometrical factor ( l / r ) 2 but is also modulated by spectral absorption of H-atoms in between the sun and some space point ~ . This absorption needs to be taken into account since we can prove that at space points with distances larger than 40 AU the H-column between ~ and the sun produces optically thick conditions with ~-(A)>I around the line center. Thus in order to do the radiation transport correctly in this 1000 AU H-atom sphere one has to solve the radiation transport equation up to higher scattering orders. In the following we solve the radiation transport problem expanding the integro-differential radiation transport equation into scattering orders and cutting off this expansion with the second order (see Scherer and Fahr, 199619]). The resonant scattering process is described as angle-dependent partial frequency redistribution process. We compare two different H-atom distributions with equal LISM H-atom densities of nH,lism 0.1 cm -a but different proton densities n p , l i s m , one with a typical H-atom wall structure for n p , l i s m - - 0.1cm -a , and one without that structure for n p , l i s m - - 0.05 cm -a. For the underlying theory in the modelling the reader should study Fahr et al. (1995)[6]. We remind the reader that due to the assumption of incompressibility in these modellings the wall density enhancement is only by a factor of 1.4 instead of 2 or 2.5 in consistent multifluid modellings as, for instance, by Baranov and Malama (1993)[1] or Fahr et al. (2000)[7]. In Figs. 1,2,3 we are showing upwind cuts through these two alternative models for density, temperature and bulk velocity, respectively. Associated with these models we =
-
290-
The Lyman-a echo f r o m the heliospheric bow shock region...
Figure 1. The density is shown as function of the upwind distance r for two alternative H-atom models with proton densities np,li~m - 0.05 cm -3 (full) and np,mm - 0.1 cm -3 (dashed), but identical H - a t o m density nH,ti~m -- 0.1 cm -3. For modellings see et (1995)[6]
Figure 2. The t e m p e r a t u r e is shown as function of the upwind distance r for two alternative H - a t o m models with proton densities np,li~n = 0.05 cm -3 (full) and np,li~,~ - 0.1 cm -3 (dashed), but identical H - a t o m density nH,l~m ---- 0.1 cm -3. For modellings see Fahr et al., (1995)[6]
Figure 3. The bulk velocity is shown as function of the upwind distance r for two alternative H - a t o m models with proton densities np,li~,~ - 0.05 cm -a (full) and np,li~m - 0.1 cm -3 (dashed), but identical H-atom density nil,libra -- 0.1 cm -3. For modellings see Fahr et al., (1995)[6]
Figure 4. The L y - a glow spectrum seen from earth in upwind direction including sources up to 80 AU (dashed curve) and up to 1000 AU (full curve), n-w: np,z~m - 0.05cm-3; w: np,zi~m - O . l c m - 3 9
-291 -
H.-J. Fahr et al.
Figure 5. Upwind Ly-c~ glow spectra seen from 40 AU for the two alternative Hatom models shown in Figs. 1,2,3. All Ly-c~ sources up to a distance of 1000 AU are taken into account.
Figure 6. Upwind Ly-c~ glow spectra seen from 80 AU for the two alternative Hatom models shown in Figs. 1,2,3. All Ly-c~ sources up to a distance of 1000 AU are taken into account.
then show in Fig. 4 the resulting alternative spectral Ly-c~ intensities seen when looking with a detector at earth into the upwind direction. Curves labeled "n-w" show the Ly-c~ resonance glow from the non-wall-model, curves labeled "w" the one from the wall- model. In both cases we are giving two curves, one where we have integrated over emission sources only up to an upwind distance of 80 AU (dashed), and one were all sources up to 1000 AU including photons scattered from the wall structure have been included (full line). As one can clearly see the differences between these two curves in both cases are hardly recognizable and if at all showing up in the redshifted wing of the line profile. This clearly can confirm that no direct spectral Ly-c~ features can be seen from earth connected with photons resonantly scattered at distances larger than 80 AU. Nevertheless one should notice that the presence of a wall like H-atom structure is indirectly reflected by the absolute spectral intensities of the heliospheric Ly-c~ glow which are different for the two alternative models though the LISM H-atom density in both cases is identical, i.e. nH,lism =0.1 cm -a. This fact is explained by the different filtering effects of the sheath plasma ahead of the heliopause. On the other hand it is also important to note that the presence of an H-atom wall structure also leads to clear qualitative changes in the upwind Ly-c~ glow spectrum when seen with a detector from large upwind distances. With detectors at upwind distances r>80 AU the H-wall structure can clearly be identified in typical spectral features as demonstrated in Figs. 5 and 6, however, appearing on the red wing of the line while Ben-Jaffel et al. (2000)[2] were expecting it to occur on the blue wing. 3. C o n c l u s i o n s The H-wall structure expected to be present in the region ahead of the heliopause does not lead to a specific Ly-c~ spectral feature which could be observed from the earth. It would , however, be detectable with a spectrally resolving Ly-c~ detector at upwind positions larger than 80 AU. On the other hand, no Fermi-l-blueshifted line is produced
-
2 9 2
-
The Lyman-a echo from the heliospheric bow shock region...
by multiply scattered and rescattered photons between H-atom populations in the Hwall and upstream of it, with different bulk velocities. The latter is connected with the fact of a very moderate optical thickness of the H-wall with T _< 1.2 and results since resonant scattering processes are frequency-coherent scattering processes in the atomic frame of the scatterer with no complete frequency redistribution occurring. Thus even many consecutive scattering and rescattering processes due to their frequency coherence do not lead to a systematic frequency blue shift. 4. A c k n o w l e d g e m e n t The authors are grateful for the financial support of this work by the DLR within the frame of the project TWINS-Lyman-Alpha.
V.B. Baranov and Y. Malama, J.Geophys.Res. 98 (1993) 15157 2. L. Ben-Jaffel, O. Puyoo and R. Ratkiewicz, Astropys.J., 533 (2000) 924 3. J.T. Clarke, R. Lallement, J.L. Bertaux and E. Quemerais, Astrophys.J. 448 (1995) 893 J.T. Clarke, R. Lallement, J.K. Bertaux, H.J. Fahr, E. Quemerais and H. Scherer, Astrophys.J. 499 (1998) 482 H.J. Fahr, R. Osterbart and D. Rucinski, Astron.Astrophys. 274 (1993) 612 6. H.J. Fahr, R. Osterbart and D. Rucinski, Astron.Astrophys. 294 (1995) 587 7. H.J. Fahr, T. Kausch and H. Scherer, Astron.Astrophys. 357 (2000) 268 8. R. Osterbart and H.J. Fahr, Astron.Astrophys., 264 (1992) 260 9. H. Scherer and H.J. Fahr, Astron.Astrophys. 309 (1996) 957 10. H. Scherer, H.J. Fahr and J.T. Clarke, Astron.Astrophys. 325 (1997) 745 11. H. Scherer, M. Bzowski, H.J. Fahr and D. Rucinski, Astron.Astrophys. 342 (1999) 601 12. H. Scherer, H.J. Fahr, M. Bzowski and D. Rucinski, Astrophys.Space Sci. 274 (2000) 133 .
.
,
-
293 -
This Page Intentionally Left Blank
Oral papers and posters
ENERGETIC
NEUTRAL ATOMS IN THE HELIOPSHERE
Mike G r u n t m a n University of Southern California, Los Angeles. Energetic neutral atoms (ENAs) are born in the heliosphere in charge exchange collisions between energetic ions and background neutral gas. The origin and properties of the heliospheric ENAs will be reviewed. Recording ENA uxes as a function of observational direction, one can reconstruct a global image of the heliosphere and study its asymmetric three-dimensional structure. Global ENA images will provide a powerful tool to study the solar wind interaction with the surrounding local interstellar medium (LISM)
RADIO EMISSIONS FROM THE OUTER HELIOSPHERE D. A. G u r n e t t and W. S. Kurth Dept. of Physics and Astronomy, University of Iowa, Iowa City, IA 52242 USA. For more than fifteen years the Voyager 1 and 2 spacecraft have been detecting radio emissions from the outer heliosphere in the frequency range from about 1.8 to 3.6 kHz. Two particularly strong events have been observed, the first in 1983-84 and the second in 1992-93. In both cases the onset of the radio emission started about 400 days after a period of intense solar activity, the first of which occurred in mid-July 1982, and the second of which occurred in May-June 1991. The radio emissions are believed to have been produced when a system of strong shocks and associated plasma disturbances from this solar activity interacted with one of the outer boundaries of the heliosphere, most likely the heliopause. The radio emission frequencies are consistent with the electron plasma frequency in the vicinity of the heliopause. From the 400-day travel time and the speed of the interplanetary shock, which is known to be in the range from about 600 to 800 km//s, the distance to the heliopause can be estimated, and is about 110 to 160 AU. From various fluid dynamic simulations it is believed that the termination shock should be at about 75 percent of the distance to the heliopause. Based on these estimates the distance to the termination shock should be in the range from about 82 to 120 AU, well within the distances that can be reached by of the Voyager 1 and 2 spacecraft, which are now at 78 and 61 AU, respectively.
- 295 -
Oral papers and posters
E N E R G E T I C N E U T R A L A T O M S AS T R A C E R S O F T H E I O N I Z A T I O N S T A T E OF T H E L O C A L I N T E R S T E L L A R M E D I U M V.V. I z m o d e n o v (1) and M. Gruntman (2) (1) Department of Aeromechanics and Gas Dynamics, Faculty of Mechanics and Mathematics, Lomonosov Moscow State University, Vorob evy Gory, Moscow, 119899, Russia, (2) Department of Aerospace Engineering, University of Southern California, Los Angeles, California, U.S.A.. Global images of the heliosphere in energetic neutral atom (ENA) fluxes are a powerful tool to study the solar wind interaction with the surrounding local interstellar medium (LISM). We explore the sensitivity of the ENA images to the ionization state of the LISM. Solar wind plasma is compressed and heated in the termination shock transition. The shocked solar plasma is convected toward the heliospheric tail in the heliosheath, the region between the termination shock and the heliopause. The ENAs that are produced in charge exchange of the heated plasma and background neutral gas can be readily detected at 1 AU. Global ENA images depend on the solar wind plasma density, temperature, and velocity in the heliosheath. The size and structure of the heliospheric interface region depend on the parameters of the interstellar plasma and gas. Hence, the ENA images would also depend on the LISM parameters. We explore in this work the sensitivity of the ENA images to the ionization state of the LISM. We use an axis-symmetric model of the solar wind/LISM interaction with the self-consistent treatment of plasma-gas coupling and Monte Carlo simulations of the neutral gas distribution
S I M U L A T I O N OF E N A I M A G E S O F T H E H E L I O S P H E R I M I N A T I O N SHO A N D I N T E R F A E R E G I O N
TER
Edmond . Roelof Johns Hopkins University/Applied Physics Laboratory, Laurel, Maryland 20723, USA. Energetic neutral hydrogen atoms (ENAH) are emitted from the region of the interface between the local interstellar cloud and the outer boundaries of the heliosphere. ENAH can be imaged from vantagepoints in the inner heliosphere to give all-sky pictures that reveal the nature of the interaction. The solar wind plasma becomes subsonic at the termination shock (TS) that forms an off-center cavity (radius> 100 AU) containing the supersonic solar wind. It then interacts with the ionized component of the interstellar gas at a heliopause within the interface region. ENAH are produced when energetic H + ions undergo chargeexchange collisions with the cold interstellar neutral atoms flowing through the interface region. Solar wind thermal protons are heated by gas--dynamical processes at the termination shock. Most of these H + ions have energies <1 keV. Higher energy ENAH will be generated from pick-up protons created from interstellar H-atoms that are ionized in the inner heliosphere and are then convected outward by the solar wind to the TS where they are accelerated to energies > 1 keV (becoming the seed population for anomalous cosmic-ray hydrogen). Both thermal and supra-thermal H + populations can be imaged with an ENAH detector utilizing electrostatic rejection of ambient ions, conversion of ENAH to H-ions, and electrostatic analysis in the energy range 0.3-6.0 keV. Simulated ENAH images and spectra are presented to illustrate those structural details of the heliospheric TS and interface region that can be remotely sensed from near 1 AU.
- 296 -
General Discussion Funsten to Gruntman: I presume the IMAGE-ENA instrument is not sensitive enough to see the heliosphere. How much increase in sensitivity do you need? Gruntman: I would like to defer this question to the next couple of talks because they are more about implementation of that kind of measurement. Only one comment: even if we see something on IMAGE we have to be very careful because it may take one year of accumulation of data to get a composite image of the heliosphere. And if you see anything on the image it could be hard to prove that it is not from some stray magnetospheric ions doing a charge-exchange. Hsieh to Roelof: You said you have 2 hours of accumulation time. Do you think it's possible to extend that? Roelof: Thank you for mentioning that point. As long as we have apogees during quiet times all of these data can be superimposed. We have 2 apogees per day, so if we would have 5 days of quiet data, we go up by a factor of 10 in the number of counts. Fahr to Funsten: Is there a chance to follow variations in time with your instrument? Funsten: You might be able to follow time variations on the order of days. Marsch to Funsten: Could you give us a rough idea about the power consumption and weight of such a typical energetic neutral atom instrument? I am asking that because we consider to fly such an instrument on the Solar Orbiter if it becomes reality. Funsten: Something like 5 kg and a comparable number of Watts. Zank: After these three talks I actually become extremely favourably exposed towards energetic neutral atoms which I hadn't been before. One of the things I showed on one of my transparencies yesterday is an order of magnitude difference between the [neutral] fluxes from the l-shock and the 2-shock model at the low energy end. So, there maybe a very intersting possibility of actually trying to determine the full global structure of the heliosphere from such fluxes of energetic neutral atoms. Gruntman: Yes, this is a very interesting energy range. But note that the modelling result will depend on the structure of the shocks. Fahr to Hilchenbach: How do corotating interaction regions produce the particles you are observing? Is it a population you get within the corotating interaction regions and part of which are recombining? Hilchenbach: Yes, you trans-charge part of that. And because in the helium cone the density is 5 times higher you get more neutrals from there. Lee to Chalov: Is the acceleration process you discussed the same as shock-surfing or multiply-reflected ions? Chalov: No, it's not shock surfing. Of course, shock surfing is also an operating process but here we consider acceleration of particles with sufficiently large velocities, so we assume that gyro-radii of particles are larger than the shock thickness, and it's simply shock drift acceleration. -297-
General Discussion
Lee to Chalov: How is an ion reflected from the shock? From the shock potential? Chalov: No, the reflection is adiabtaic reflection due to the change in the strength of the magnetic field. Lee to Gurnett: I'm confused. If you excite the waves at the plasma frequency (fp) upstream of this forward propagating shock, those waves can't get back? Gurnett: The radio emission at two times the plasma frequency (2 fp) could get back. The radio emission at fp could not. McNutt to Gurnett: Just a comment on the distance calculation. If you look at most ideas of how blast waves move out in the solar wind, as well as looking at some of the simulations, of course, you see that shocks always entrain some material, so that they inherently slow down. That will bring the estimation of the distance down by something like 30 AU. Gurnett: I agree, the shock wave does slow down. If you take the lowest shock velocity we see, you might get the termination shock inside of 100 AU, but I personally doubt it. We have backed up our estimates with simulations confirming numbers in the range I am talking about. Cummings to Cairns: How do you know that your are getting 2 fp or fp? Cairns: Probably the strongest argument I could say for being able to identify a particular frequency as being the frequency of the local interstellar medium (LISM) is this low frequency cut-off in the 2 kHz component which Don Gurnett mentioned. I think that almost certainly is a propagation cut-off and, therefore, is most likely the plasma frequency fp of the LISM or that of the heliosheath. Cummings to Cairns: You said that the radio emission is coming from within 50 AU of the nose. Is that laterally or is it possible that it comes from further out than the heliopause? Cairns: Both directions, but the 50 AU is just an estimate, of course. Lallement to Fahr: Just a remark about the paper by Ben-Jaffel et al. you were mentioning. The location of this Fermi-line is exactly at the expected location of the deuterium Lyman-a emission from Earth. Why is not the entire Fermi-line just the deuterium Lyman-a emission from the Earth? Fahr: I agree that this another question which is still open to the authors. Jokipii to Fahr: I did see that there was an enhancement on the blue side of the line in the Ben-Jaffel et al. figure you showed. What is that due to? Fahr: Well, I am not even ready to take it serious, because we did not have any signature like that in our Hubble Space Telescope data.
-
298-
Session 5: The Heliosphere and Galax~
This Page Intentionally Left Blank
The Solar Wind: Probing the Heliosphere with Multiple Spacecraft John D. Richardson Center for Space Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA The combination of the Voyager spacecraft in the outer heliosphere, Ulysses at high latitudes, and multiple solar wind monitors near Earth provides a unique opportunity to study the global structure and evolution of the Solar wind. From solar minimum to solar maximum the latitudinal gradients of density and speed reverse so that at solar maximum speeds are higher near the solar equator, but solar cycle changes in the dynamic pressure occur at all solar latitudes. The merged interaction region (MIR) frequency increases during the ascending phase of the solar cycle and effects the cosmic ray intensities. One example shows a MIR that is clearly driven by a coronal mass ejection (CME). Ulysses and Voyager observations are combined to quantify the speed decrease in the solar wind with distance and estimate the density of the local interstellar medium. 1. I N T R O D U C T I O N The spacecraft fleet in place throughout the heliosphere provides an opportunity to understand the structure of the heliosphere as we move from solar minimum to solar maximum. The Voyager 1 and 2 spacecraft are at 75 and 60 AU and 22 ~ S and 34 ~ N latitude, respectively. Ulysses is moving southward, starting its second polar orbit, this time near solar maximum (see [1] for a review of the first polar orbit of Ulysses). WIND, ACE, IMP 8, and other spacecraft monitor the solar wind near Earth. This paper will focus on several topics best addressed with multiple spacecraft. Solar cycle variations have long been known to occur in solar wind parameters. Initially these variations were observed only near the ecliptic so it was problematic whether these changes occurred because of the changing streamer belt configuration or resulted from a global solar change in the solar wind source. The recent exploration of higher latitudes by the Voyager and Ulysses spacecraft helps to address this question. Similarly, the question of the latitude profiles of solar wind parameters could previously only be addressed by remote sensing methods (eg., interplanetary scintillation). The combination of spacecraft at different latitudes enables us to put together profiles of solar parameters at solar cycle maximum and solar minimum. In the outer heliosphere, the ascending phase of the solar cycle portends an increase in merged interaction regions (MIRs) and their concomitant exclusion of cosmic rays. We show evidence of these events in the current solar cycle. The outer heliosphere is effected by the neutral H of the local interstellar medium which penetrates into the heliosphere, where it is ionized and picked up by the solar wind. Observations from radially separated
-301 -
J.D. Richardson
spacecraft can be used to determine the slowdown of the solar wind due to these pickup ions and thereby estimate the density of the interstellar H. The temperature may also be influenced by these interstellar pickup ions; comparison with spacecraft in the inner heliosphere allows us to determine the amount of heating required to explain the observations and thus the rate of energy transfer between the pickup and thermal ion populations.
2. S O L A R C Y C L E V A R I A T I O N S 2.1.
PRESSURE
CHANGES
Solar wind conditions have long been known to vary significantly over the solar cycle [see review by Gazis [2] and references therein]. In particular, the solar wind dynamic pressure varies by a factor of 2 [3] with the minimum dynamic pressure at solar maximum followed by a rapid increase in pressure over the next 1-2 years and then a slow decrease until the next solar maximum [4]. Since these initial results were based on data from near the ecliptic plane, it was difficult to ascertain whether these variations were intrinsic to the sun or a result of the changing configuration of the high and low speed wind regions and the solar surface magnetic field. This difference is very important for the dynamics of the outer heliosplJere, since a change of the dynamic pressure by a factor of 2 would result in a roughly 13-15 AU change in the position of the termination shock over the course of the solar cycle [5,6]. Figure 1 shows the solar wind speed, density normalized to 1 AU, and normalized dynamic pressure observed by Voyager 2, IMP 8 and Ulysses as well as the latitudes of each spacecraft. The profiles are 2 solar rotation running averages. The Ulysses and Voyager 2 data are time-shifted back to 1 AU using the observed speed at each spacecraft. IMP 8 moves between 7.25 ~ N and S latitude as Earth revolves about the Sun, Voyager 2 moves from 0~ to 20 ~ S heliolatitude, and Ulysses started out of the ecliptic in 1992 and samples latitudes as high as 80 ~. The dip in Ulysses speed and spike in density in 1994 correspond to the Ulysses fast latitude scan crossing of the solar equator [7]. All three speed traces match fairly well until 1993, when Ulysses began seeing larger speeds as it moved to higher latitudes. The speeds observed by Ulysses at heliolatitudes above 40 ~ are over 700 km/s during the first polar orbit which was centered on solar minimum. Voyager 2 and IMP 8 see similar speeds until 1996, when the latitudinal speed gradients increase near solar minimum. The speeds converge in 1998 as we enter the ascending phase of the solar cycle. The important point to note is that from 1993 to 1997 Ulysses is in the high speed wind while IMP 8 and Voyager are in lower speed wind, and that from 1995 to 1998 IMP 8 is in low speed wind while Voyager at intermediate latitudes sees intermediate speed wind. The second panel shows the densities normalized to 1 AU. We note that the IMP 8 densities are systematically about 20% higher than those from the other spacecraft. The density profiles are essentially the inverse of the speed profiles, with lower densities at higher latitudes near solar minimum. These top two panels show very clearly that we are sampling times when these spacecraft are in very different solar wind regimes. The third panel shows the solar wind dynamic pressure. Despite the vastly divergent speed and density traces, the pressure profiles look essentially identical. Each shows a slow decrease in pressure from 1992 into the year 2000. These times span the solar minimum
- 302-
The solar wind: Probing the heliosphere with multiple spacecraft
Figure 1. Speed, density, pressure, and heliolatitude of IMP 8, Voyager 2 and Ulysses.
period with the strong latitudinal speed and density gradients through the ascending phase of the approaching solar maximum. Thus we conclude that the dynamic pressure changes of the solar wind are a global phenomena with similar changes in pressure occurring at all solar latitudes. This conclusion justifies the assumptions which lead to the results that the heliosphere will expand and contract by 15 AU over the solar cycle [5,6]. In the middle of 2000 the termination shock was probably nearing the end of a long inward movement; this contraction should end soon after solar maximum when the pressure is expected, based on past solar cycles, to rapidly increase. No clear signature of this solar cycle's dynamic pressure increase has yet been observed. 2.2. S P E E D A N D D E N S I T Y P R O F I L E S During solar minimum, the speed and density decrease rapidly away from the solar equator. The slow speed region is narrow enough that Earth's 7.25 ~ inclination produces
-303-
J.D. Richardson
significant speed effects [8-10] Figure 1 shows that large differences in speed are observed near solar minimum at different latitudes. Figure 2 shows an enlarged scale of the transition of speed and density from solar minimum to solar maximum for Ulysses, IMP 8, and Voyager from 1997-2001. In this case the IMP 8 densities are calibrated to the Ulysses and Voyager densities by multiplying by 0.8. When Ulysses is at low latitudes in 1997 and 1998, the Ulysses and IMP 8 densities and speeds match well as would be expected. As Ulysses continues to higher southerly latitudes and solar maximum approaches, the Ulysses densities are systematically higher and speeds lower than at IMP 8 near the ecliptic. Indeed, Ulysses observations from July 2000 show some speeds below 300 km/s [11]. Voyager speeds are also less than those seen at IMP 8, but it is difficult to deconvolve latitudinal gradients from the slowdown due to pickup ions. Density structures at Voyager have evolved into a series of merged interaction regions, again making direct comparison with the other spacecraft diMcult.
Figure 2. Speed and density of IMP 8, Voyager 2 and Ulysses.
Data from the previous solar cycle are used to test this result. Figure 3 shows speed profiles for IMP 8, Pioneer 11, and Voyager from 1986-1992. In 1986-1987 the speeds increase from the spacecraft with the lowest average latitude to that with the highest (Pioneer 11). From the end of 1988 to the end of 1991, near solar maximum, the Pioneer 11 speeds are lower than those observed by the lower-latitude spacecraft. Thus we observe this effect, lower speeds at higher latitudes, in two solar cycles. Figure 4 shows a schematic diagram of the solar wind speed and density profiles as a function of latitude at solar minimum and solar maximum. At solar minimum, low speeds and high densities are found only near the equator in a band with half-width of order 10 ~ with a several degree transition region to the fast, low density wind which persists up to
-
304 -
The solar wind." Probing the heliosphere with multiple spacecraft
Figure 3. Speed and lieliolatitude of IMP 8, Voyager 2 and Pioneer 11.
high latitudes [9]. At solar maximum, the highest speeds and lowest densities are at the equator, with a slow decrease in speed and increase in density with latitude up to at least 60~ depending on whether the polar coronal hole persists through solar maximum or not the speed may increase, and density decrease, very near the pole. This decrease of speed with latitude might seem unexpected; the closed field regions are thought of as producing similar speed wind (leading to a flat latitude gradient). One possible driver of the driver of these gradients may be coronal mass ejections (CMEs), which occur more often at low latitudes and could cause the higher speeds observed at the equator [12]. 3. M E R G E D
INTERACTION
REGIONS
The Voyager 2 density profiles in figure 2 show periods with enhanced density. Figure 5 shows daily averages of the density observed by Voyager 2 and the >70 MeV/nuc counts measured by the cosmic ray subsystem (CRS) on Voyager 2. Four clear density enhancements are seen, one in early 1997, one in late 1998, and one each in early 1999 and 2000. Burlaga and Ness [13] identified the 1998 event as a MIR based on the enhanced IMF during this time period. The other events have very similar plasma signatures and we suggest these are also MIRs. The frequency is increasing towards solar maximum, as expected. The large IMF during these events acts as a barrier to the inward penetration of cosmic rays. After each MIR identified here a decrease in the CRS counts is observed, although CRS decreases sometimes occur when no density enhancement is present. In past solar cycles, the CRS counts increased towards solar maximum, with minor decreases triggered by MIRs followed by recoveries, until a large event causes a major decrease in flux which persists until the next solar cycle [14]. The Voyager 1 CRS counts (gray) look
-305 -
J.D. Richardson
Figure 4. Diagram of speed profiles versus latitude for solar minimum and maximum.
almost identical to those observed by Voyager 2; thus we know that the effects of these MIRs cover a latitude range of at least 50 ~ and a longitude range of at least 45 ~ longitude.
Figure 5. Solar wind density (black), CRS counts (gray), and CME ejecta (gray line).
The driving mechanism for large MIRs may be large CMEs, or groups of CMEs, occurring on the Sun. Signatures of CMEs near Earth are low temperatures, smooth field rotations (magnetic clouds), counter streaming electrons, and enhanced helium abundances [12]. The signature most likely to survive until the outer heliosphere is the helium abundance. Wang and Richardson [15] recently completed a survey of helium abiundance enhancements (HAEs) observed by Voyager, defined as periods where the He++/H + ratio is greater than 10%. This condition seems sufficient to identify some CMEs, but other CMEs do not have He++/H + ratios of this magnitude. The only multiple day HAE event in the time period covered by Figure 5 is shown by the vertical line just before 1999.5; the He++/H + ratio was enhanced over a 6 day period probably resulting from a series of CMEs. The location of the HAEs on the trailing edge of the MIR is a clear indication
-
306-
The solar wind: Probing the heliosphere with multiple spacecraft that the MIR is driven by the CMEs in this case. More work should be done for the other MIRs, looking for He++/H + ratios which are below the 10% threshold but are still enhanced to see if CME drivers can be identified. 4. S O L A R W I N D C H A N G E S
WITH DISTANCE
4.1. Solar W i n d S l o w d o w n
One difficulty in understanding solar wind observations is decoupling changes due radial distance, temperature, and solar cycle effects. Figure 2 shows that the Voyager 2 speeds in 2000 are well below those observed by IMP 8 and Ulysses. This speed difference is at least partially due to the addition of interstellar neutrals to the solar wind; when these neutrals are ionized they are accelerated to the solar wind speed by the Lorenz force. The energy for this acceleration comes from the kinetic energy of the solar wind. Quantifying the speed decrease provides estimates of the density of H in the local interstellar medium. The first attempt to quantify the speed decrease [16,17] found a decrease of about 30 km/s near 30 AU giving a density for interstellar neutrals of 0.05 cm -3, but this result was controversial [18]. Isenberg [19] argued that the slowdown in this period was greater than 30 km/s and thus not inconsistent with values of the density of the local interstellar medium derived by Gloeckler et al. [20] of 0.115 cm -3 at the termination shock and 0.22 cm-3 in the undisturbed interstellar medium. Wang et al. [21,22] took advantage of two time periods when Ulysses and Voyager 2 were at similar latitudes to look for the speed slowdown when the spacecraft were separated in radius by 30 and 55 AU. They used the Ulysses data as input to a 1-D MHD model and propagated the solar wind out to the radial distance of Voyager 2. Wang and Richardson [23] show histograms of speeds observed at Voyager 2, speeds predicted for no density of the local interstellar medium, and speeds for a density of the local interstellar medium density of 0.05 cm -3 at the termination shock. This value for the density of the local interstellar medium gives good agreement with the observations, but is lower than results from observations of pickup ions and from UV measurements. These observations give densities of about 0.22 in the density of the local interstellar medium of which roughly half is lost before the termination shock [20]. Thus the values derived from the speed decrease are about half of other determinations of this value. Doubling the density of the local interstellar medium in our model would roughly double the observed speed decrease, which would not be consistent with the observations. We note that the Gloeckler results are for a limited set of solar wind conditions, whereas the slowing of the wind is an integral process occurring over 60 AU, which could explain the apparent discrepancy. As Voyager gets further from the Sun the speed decrease will grow even larger and perhaps provide the opportunity for more determinations of the slowdown. 4.2. T E M P E R A T U R E
Many competing effects determine the temperature of the thermal component of the solar wind. The plasma cools as it moves outwards due to adiabatic expansion. But it can also be heated by numerous effects. The velocity difference between different solar wind streams results in the formation of shocks which transfer energy from the solar wind flow into the internal plasma energy. Dissipation of speed shear and magnetic turbulence results in heating of the plasma [24]. Pickup ions are created with energies equal to the
- 307-
J.D. Richardson
flow energy, about 1 keV, and this energy may be transferred to the thermal protons. Figure 6 shows a comparison of proton temperatures observed at Voyager 2 and IMP 8. The temperature decreases from Voyager launch to somewhere between 1986 and 1990, when Voyager was at 20-30 AU. The minimum in 1986 is probably a solar minimum effect due to Voyager 2 being in very slow solar wind flow. The decrease in temperature is much less than the R -4/3 adiabatic expansion would predict, falling as R - 5 - R -7 from 1-30 AU [8,16]. After 1990 the temperature increases to about 1997, then begins to decrease. Comparison with the IMP 8 temperature profile shows a very good correspondence for feature with scales on order of a year out to almost 60 AU. After 1998, the temperatures at IMP 8 increase while those at Voyager 2 decrease.
Figure 6. Temperatures (solid lines) and speeds (dotted lines) at IMP 8 and Voyager 2.
Also plotted on Figure 6 are scaled speed profiles measured by each spacecraft. At 1 AU the speeds and temperatures show very good correspondence. For Voyager 2, speed and temperature features generally match quite well. The surprising feature of this plot is the very good agreement between the magnitudes of the traces after 1983 when Voyager was at about 15 AU. The decrease in temperature relative to the speed inside 15 AU is due to the adiabatic cooling lessened by the effects described above. The similarity of the traces outside 15 AU suggests that the heating which occurs compensates almost exactly for the expected adiabatic cooling over a distance of 45 AU. The remaining fluctuations are then explained quite well by the speed variations. 5. S U M M A R Y Solar wind parameters vary with changes in the solar cycle, with changes in heliolatitude, and with changes in radial distance. Even with multiple spacecraft it can be difficult
-308-
The solar wind: Probing the heliosphere with multiple spacecraft to deconvolve these various effects. We find that the solar wind dynamic pressure varies by a factor of 2 over the solar cycle at all latitudes and distances. The speed and density (which determine the dynamic pressure) have dramatically different latitude profiles at solar maximum and solar minimum. At solar minimum, a narrow, 20-30 ~ latitude wedge of slow, dense solar wind at low latitudes is surrounded by fast tenuous solar wind at high latitudes, whereas at solar maximum the highest speeds and lowest densities are at low latitudes with a slow decrease in speed towards higher latitudes. The speed of the solar wind decreases with radial distance due to the pickup of interstellar neutrals; at 60 AU this slowdown is about 40 km/s. The solar wind is heated as it moves outwards so it does not cool adiabatically. The observed temperatures in the outer heliosphere are are still very strongly correlated with the solar wind speed, a relation initially imposed at the solar source. 6. A c k n o w l e d g m e n t s I thank D. McComas for the Ulysses data used in this paper and E. Stone for the Voyager CRS data. This work was supported under NASA contract 959203 from the Jet Propulsion Laboratory to the Massachusetts Institute of Technology and by Heliospheric GI grant NAGW-6473. R E F E R E N C E S
1. D . J . McComas et al., J. Geophys. Res., 105 (2000) 10419. 2. P.R. Gazis, Rev. Geophys., 34 (1996) 379. 3. A . J . Lazarus and R. L. McNutt, Jr., in Physics of the Outer Heliosphere, edited by S. Grzedzielski and D. E. Page, Pergamon Press, New York (1990) 229. J. D. Richardson, K. I. Paularena, A. J. Lazarus, and J. W. Belcher, Geophys. Res. Lett., 22 (1995) 1469. S. R. Karmesin, P. C. Liewer, and J. U. Brackbill, Geophys. Res. Lett., 22, (1995) 1153. C. Wang, and J. W. Belcher, J. Geophys. Res., 104 (1999) 549. 7. J. L. Phillips et al., Solar Wind 8, edited by D Winterhalter et al., (1996) 416. 8. P. R. Gazis, J. Geophys. Res., 98 (1993) 9391. 9. J. D. Richardson and K. I. Paularena, Geophys. Res. Lett., 24 (1997) 1435. 10. Miyake, W., et al., Plan. Sp. Sci. 36 (1998) 1329. 11. E. J. Smith, A. Balogh, R. J. Forsyth, D.J. McComas, COSPAR 2000, E2.2-0037, Warsaw (2000). 12. J. T. Gosling, Coronal Mass Ejections, edited by N. Crooker, J. A. Joselyn, and J. Feynman, Geophysical Monograph 89, American Geophysical Union (1997) 9. 13. L. F. Burlaga and N. F. Ness, J. Geophys. Res., 105 (2000) 5141. 14. F. B. McDonald, Sp. Sci. Rev., 83 (1998) 33. 15. C. Wang and J. D. Richardson, J. Geophys. Res. (2000), in press. 16. J. D. Richardson, Physics of Space Plasmas (1995), No. 14, edited by T. Chang and J. R. Jasperse (1996) 431. 17. J. D. Richardson, K. I. Paularena, A. J. Lazarus, and J. W. Belcher, Geophys. R es. Lett., 22 (1995) 325. .
.
o
-
309-
J.D. Richardson
18. P. R. Gazis, Geophys. Res. Lett., 22 (1995) 2441. 19. P. A. Isenberg, Solar Wind 9, edited by S. R. Habbal, R. Esser, J. V. Hollweg, and P. A. Isenberg, (1999) 189. 20. G. Gloeckler, L. A. Fisk, and 3. Geiss, Nature, 386 (1997) 374. 21. C. Wang, J. D. Richardson and a. T. Gosling, J. Geophys. Res. 105 (2000) 2337. 22. C. Wang, J. D. Richardson and J. T. Gosling, Geophys. Res. Lett. 26 (2000) 2429. 23. C. Wang and J. D. Richardson, this volume. 24. W. H. Matthaeus, G. P. Zank, C. W. Smith, and S. Oughton, Phys. Rev. Lett., 82 (1999) 3444.
-310-
Propagation of the solar wind from the inner to outer Heliosphere" Three-fluid model C. Wang* and J. D. Richardson Center for Space Research, Massachusette Institute of Technology, Cambridge, MA 02139, USA We take advantage of two time periods (1991 and the beginning of 1999) when Ulysses and Voyager 2 were at about the same latitude to study the propagation of the solar wind from the inner to outer heliosphere (up to 60 AU). A three-fluid model of the solar wind which consists of solar wind protons, pickup ions and electrons is employed to examine the effect of pickup ions. As expected, the pickup ions play an important role in the evolution of the solar wind in the distant heliosphere, especially beyond 40 AU. We find a decrease of about 40 km/s or 10% in the radial velocity near 60 AU. This speed decrease implies an interstellar neutral density at the termination shock of 0.05 cm -3. 1. I N T R O D U C T I O N The propagation of the solar wind in the outer heliosphere, especially shock evolution and interaction, have been studied intensively [1,2]. However these studies could not take advantage of recent spacecraft observations out to 70 AU and did not appreciate the importance of pickup ions in the distant heliosphere. Only recently have investigations of the effect of pickup ions on dynamical processes in the solar wind begun appearing in the literature [3-6]. However, almost all these pickup models assume that wave-particle interactions proceed sufficiently quickly that pickup ions are soon assimilated into the solar wind, becoming indistinguishable from solar wind protons. A substantial increase in the solar wind temperature with increasing heliocentric distance beyond ~ 5 AU is thus predicted. Such a temperature increase is, however, not observed in the outer heliosphere. A model which distinguishes the pickup ions from the solar wind ions is appropriate and a simple version was developed by Isenberg [7]. We will extend the Isenberg three-fluid model in this study. Voyager 2 continues to explore the outer heliosphere as Ulysses studies the latitudinal dependence of the solar wind. The trajectories of Ulysses and Voyager 2 are shown in Figure 1. During the year 1991 these spacecraft were within 2~ latitude and their radial separation was larger than 30 AU. They were at about the same latitude again at the beginning of 1999 and their radial separation was as large as 55 AU. This latitudinal proximity provides a good opportunity to study the evolution of the solar wind from the inner to outer heliosphere and to investigate the effect of pickup ions. *Also at the Lab. for Space Weather, Chinese Academy of Sciences, Bejing, China
-311 -
C. Wang and J.D. Richardson
Figure 1. Ulysses.
Trajectories of Voyager 2 and Figure 2. Steady-state Heliosphere.
2. T H R E E - F L U I D
MODEL
Isenberg [7] developed a three-fluid model which consists of co-moving thermal populations of solar wind protons, pickup ions and electrons. The steady-state solution yields a solar wind with a "core" proton distribution which cools, essentially, adiabatically. This simplified three-fluid model, which does not consider the interaction of the solar wind protons and pickup ions and assumes that all the energy of the pickup process remains with the pickup ions, does not reproduce the observed temperature profile. We extend this three-fluid model by depositing part of the energy from the ionization and pickup process into the solar wind protons. We suggest an adjustable parameter e(t, r, v, n, ...) to represent the ratio of this portion of the energy to the total thermal energy generated from the ionization and pickup process. That is E~ CEtotal. For simplicity, we assume c to be constant. With this constraint, we find that c = 20V0 gives the best fit to the observations, as illustrated in the third panel in Figure 2. The cold neutral density distribution for the interstellar neutrals, nil(r), is taken as [8] =
(1)
e -AIr
where A = 4 AU and UH~ = 20 km/s. The subscript infinity refers to the boundary values (the boundary being the termination shock). The source terms for charge exchange were -312-
Propagation o f the solar wind from the inner to outer heliosphere: ....
summarized in our previous work [9]. The solar wind speed, proton density, proton temperature, pickup ion density, and pickup temperature as functions of heliocentric distance are shown in Figure 2 from the top to bottom panels, respectively.
Figure 3. Evolution of the solar wind (1991 Figure 4. Comparison of the model results data). with Voyager 2 observations.
3. E V O L U T I O N
OF T H E S O L A R W I N D
In order to study solar wind propagation from the inner to outer heliosphere, we feed the five streams from day 140 to 210 in 1991 from the Ulysses data into our three-fluid model as an input. We follow the evolution of the solar wind structure to the location of Voyager 2. Figure 3 shows the theoretical development of these streams as if they were "observed" at successive distances (10, 20 AU and the location of Voyager 2 (~35 AU)). The interaction and evolution of the streams observed by Ulysses produce a quite different solar wind structure at Voyager 2. The structure loses the traces of the original stream structure and form two strong forward shocks at Voyager 2, which agrees roughly with the observations. 4. S L O W D O W N
OF THE SOLAR WIND
During 1998 and 1999, it took about 6 months for the solar wind to travel from the location of Ulysses to Voyager 2. So the solar wind observed by Ulysses during the time period from 1998.5 to 1999.5, when the latitudinal difference between these two spacecraft was less then 10 ~ reached Voyager 2 during 1999. We carry out numerical calculations
-313-
C. Wang and,I.D. Richardson
g
0.2 0.0 300
0.0
350
400
450
500
550
600
300
550
400
450
5.00
550
600
. . . .
300
350
L 1 ] .............
400
450
500
550
600
s
Figure 5. Histograms of the solar wind speed.
using two different models, namely an MHD model and an MHD model which mimics the effect of the interstellar neutral hydrogen (MHD PI model). Figure 4 shows comparisons of the observed flow speed at Voyager 2 (solid lines) with those simulated by the MHD model (without pickup ions, dotted line in the top panel) and the MHD PI model (with interstellar neutral density at the termination shock n/zoo - 0.05 cm -a, dotted line in the bottom panel). Overall, the speed predicted by both MHD and MHD PI models roughly lies in the range of the observations; however, neither of them reproduce the fine structure of the speed profile observed by Voyager 2. As shown in Figure 4, the MHD model in general predicts higher speeds than are observed at Voyager 2. As expected, we need to incorporate pickup ions into our model to decelerate the solar wind speed sufficiently to match the observations. Figure 5 shows histograms of the frequency of occurrence of various speeds for both the Voyager 2 observations in 1999 and the model results using the Ulysses observations as input. From the left to the right panels, the figure shows the Voyager 2 observations, the MHD and the MHD PI model speed predictions (with n~oo - 0.05 cm-a), respectively. Both the MHD and MHD PI models reproduce the observed histogram shape. The average speed of the observations for 1999 is 392 km s -~. However, the MHD model gives a higher average speed of 432 km s -~. By contrast, the MHD PI model with n~oo - 0.05 cm -3 gives an average speed of 389 km s -1, very close to the observed value. These results are similar to our previous finding using a one-fluid MHD model [6]. REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9.
A . J . Hundhausen, AGU Monograph, 34 (1985) 37. Y. C. Whang, Space. Sci. Rev. 57 (1991) 339. G.P. Zank, and H. L. Pauls, J. Geophy. Res. 102 (1997) 7037. W . K . M . Rice, and G. P. Zank, J. Geophy. Res. 104 (1999) 12563. C. Wang, J. D. Richardson and J. T. Gosling, J. Geophys. Res. 105 (2000) 2337. C. Wang, J. D. Richardson and J. T. Gosling, Geophys. Res. Lett. 26 (2000) 2429. P.A. Isenberg, J. Geophys. Res. 91 (1986) 9965. V.M. Vasyliunas, and G. L. Siscoe, J. Geophys. Res. 81 (1976) 1247. C. Wang, and J. W. Belcher, J. Geophys. Res., 104 (1999) 549.
-314-
Relationships of corotating rarefaction regions outside 40 AU with solar observations: Heliospheric mass loading A. Posner, ~ N.A. Schwadron, ~ and T.H. Zurbuchen ~ * ~Department of Atmospheric, Oceanic and Space Sciences, University of Michigan, 2455 Hayward, Ann Arbor, MI 48109-2143, USA In mid-1994 the solar wind instrument on the Voyager 2 spacecraft observed a recurrent pattern of compression and rarefaction regions on the timescale of the solar rotation period. The spacecraft location at that time was at 12 ~ southern heliographic latitude. The time period of the release of the large scale structures at the Sun was coincident with observations of recurrent solar wind structures by IMP 8 in Earth orbit. Additionally, the Yohkoh/SXT soft x-ray telescope made synoptic observations of the upper solar corona. With the data available we identified the coronal sources of solar wind structures observed by Voyager 2 at ~43 AU in the Yohkoh/SXT observations of the Sun at the time the solar wind was released. Using the Sun's rotation as a clock a significant slowdown of the solar wind can be derived. A numerical solution for the mass loading of the solar wind with pickup ions is used to infer the interstellar neutral hydrogen density as n~ = 0.127+0.019. 1. I N T R O D U C T I O N Bryant [1] discovered a periodicity in the intensity of heliospheric energetic protons that resembled the synodic solar rotation period. Barnes and Simpson [2] associated these with quasi-stationary solar wind structures in the heliosphere. Krieger et al. [3] discovered coronal holes as the source of high speed streams in the ecliptic plane by ballistic backmapping of the radially propagating solar wind. Due to solar rotation fast streams interact with ambient slow solar wind to form sets of Corotating Interaction Regions (CIRs) and Rarefaction Regions (CRRs) (see sketch in [4]). Solar sources for CIRs and CRRs are the western and eastern boundary regions of coronal holes, respectively. Outside 10 AU CIRs start to merge into MIRs (Merged Interaction Regions) [5]. In the literature, values from 12 AU [6] to 45 AU [7] are found for the maximum distance that a CIR can exist without merging with CIR structures from another rotation. Tilted dipole 3-D MHD simulations [8], however, suggest that forward and reverse shocks pass each other at different latitudes [9], leaving behind undisturbed rarefaction regions in the outer heliosphere. The ballistic backmapping of CIRs is inaccurate due to the interaction of the converging streams, resulting in acceleration and deflection of the bulk solar wind flow. Ballistic backmapping that focuses on corotating rarefaction regions avoids the effects of these shocks. It is therefore the most efficient and easiest method to relate solar *This work was supported, in part, by NASA contracts NAG5-2810 and NAG5-7111.
-315-
A. Posner, N.A. Schwadron and T.H. Zurbuchen
wind observations to solar source regions. 2. O B S E R V A T I O N S
AND DISCUSSION
Figure la shows one hour averages of the solar wind speed measured by Voyager 2/PLS in 1994 at 43 AU from the Sun, at 12~ S heliographic latitude. The long-term solar wind speed increase and the recurrent structure on the basis of solar rotations (shading) are caused by a stable southern polar coronal hole extension observed by Yohkoh/SXT and its associated fast solar wind stream observed by IMP 8 in Earth orbit in early 1994. In Carrington rotation (CR) 1880 and 1881 an equilibrium is reached with the beginning of the decay of the coronal hole extension. Here ballistic backmapping should most accurately map the rarefaction region back to its coronal source longitude. Figure l b shows backmapped solar wind data of Voyager 2 and IMP 8 along with synoptic maps of the solar corona by Yohkoh/SXT for two consecutive CRs. The origin of the rarefaction is found at ~230 ~ Carrington longitude (CL) in the synoptic map. IMP 8 observations reveal a dwell in backmapped solar wind data at these CLs in CR 1880 (data gap in 1881). For Voyager 2 observations this dwell is shifted from 230 ~ CL in CR 1881 to 40 ~ CL in CR 1880. Streamlines, which, according to the Parker model, are associated with the magnetic field Archimedian spirals, would map back to their coronal source. If the solar wind is slowing down, for example by the interaction with pickup ions, the backmapping based on a Parker spiral associated with the observed solar wind speed is more tightly wound than to the actual stream line. Therefore the backmapping gives a deviation to the west of the actual source of the solar wind. We quantify the slowdown of the solar wind to about 10% for the 600 km/s solar wind observed in the rarefaction. The original speed in the inner heliosphere was ~667 km/s. 3. D E R I V I N G
THE INTERSTELLAR
NEUTRAL
DENSITY
Given the slowdown in solar wind speed (~10%), presumably due to mass loading from interstellar pickup ions, we may solve a simple model to infer the interstellar density. The equations below describe the mass, momentum, and energy flow in the solar wind: r 2 d r ( r 2 p u ) - ~3prop
1 r 2 dr
dP r
(2)
+ d-7 = --~Tnnp~t2
5p) r2
(1)
nn
_
1
7
O)
[12]. According to [13] the photoionization rate is given by 3p - 0.9.10-7/s. After charge
-316-
Relationships of corotating rarefaction regions outside 40 A U...
Figure 1. a) Voyager 2/PLS observations at 43 AU, 12~ S heliographic latitude during 1994. From top to bottom: (top) one hour averages of the solar wind speed, shaded according to the consecutive association to Carrington rotation of solar wind release, (middle) 50 day running mean of the speed (light) and proton density (dark), and (bottom) the solar wind proton density shaded as in the top panel. Carrington rotation numbers are given on top. b) Yohkoh/SXT synoptic maps and backmapped solar wind data for Carrington rotations 1880 and 1881. The middle panel shows the solar wind speed (thick black lines) and density (thin gray lines) observations by IMP 8 in Earth orbit. The dominant magnetic field polarity is indicated according to IMP 8 in situ observations. The fast solar wind stream for these two consecutive rotations tracks the longitudes of the southern polar coronal hole extension with a rarefaction region at ~240 ~ CL in CR 1880 (data gap in CR 1881). The lower panel shows backmapped Voyager 2 solar wind speeds (thick black lines) and densities (thin gray lines) from a distance of 43 AU. Two vertical lines are connected with a horizontal. The left vertical line indicates the approximate source longitude of the rarefaction region identified in the Yohkoh/SXT synoptic map. The right vertical line indicates the dwell longitude of the rarefaction region solar wind observed at Voyager 2. The horizontal line itself indicates the difference in the longitudes that is equivalent to the time delay At = Ak/co e.
-317-
A. Posner, N.A. Schwadron and T.H. Zurbuchen
Figure 2. The interstellar neutral density depending on the slowdown of the solar wind from 667 km/s at 43.5 AU distance from the Sun. The error bars indicate the statistical uncertainty of our method to quantify the solar wind slowdown and the tolerance of the parameters used for the numerical integration.
exchange or photoionization the ions of interstellar origin are picked up by the frozen-in solar magnetic field. The momentum of these ions is transferred to the solar wind, which leads to a slowdown of the bulk solar wind. The location of Voyager 2 in 1994 is ~30 ~ off the interstellar upstream direction. We derived the slowdown of the solar wind to ~10% (667 km/s to 600 kin/s) based on the observations of corotating rarefaction regions at 43 AU combined with the ballistic backmapping technique. Numerical integration of the equation system leads to the dependency of the interstellar neutral density after passage through the termination shock on the slowdown of the solar wind at 43 AU shown in Fig. 2. With our method we derive the interstellar neutral hydrogen density, as the major contribution for the momentum transfer to the solar wind, to 0.127 cm-3+ 0.019 cm -3. The given error refers to the statistical error in the observed speed. Our density estimate is consistent with the result of Gloeckler et al. [14] of 0.115 cm-3+ 0.025 cm -3. 4. C O N C L U S I O N S An attempt was made to associate the recurrent structures in the solar wind speed profile at Voyager 2 with coronal observations. The solar wind appeared to be slowed down. We interpret this slowdown to be according to mass loading of the solar wind with interstellar pickup ions. By comparison of the backmapped structures with the source of these structures in the corona we derived a slowdown in the order of 10% of the original solar wind speed. A numerical integration of the solar wind mass, momentum, and energy equations led to an interstellar neutral density of 0.127 cm-3+ 0.019 cm -3, which is close to the result of Gloeckler et al. [14] based on observation of pickup ions with Ulysses/SWICS. Other methods, like observations of resonant back-scattering of solar UV radiation, give results ranging from 0.03 to 0.3 cm -3 [15], [16] and suffer from
-318-
Relationships of corotating rarefaction regions outside 40 A U... the dependence on complex back-scattering models with poorly known parameters as well as on the uncertainty of a galactic UV emission background. In our method to determine the slowdown of the solar wind we cannot entirely exclude the possibility that other causes, for example merging shocks, made a significant contribution. For the following reason it appears unlikely that MHD effects may have affected the results significantly. The contribution of CIR-associated shocks on the high speed solar wind (667 km/s) could only lead to a further slowdown of the solar wind, therefore our method would still give an upper limit to the pickup ion momentum transfer and interstellar neutral density. However, attempts are planned to constrain the slowdown of the solar wind for Voyager 2 observations with improved and more sophisticated MHD codes [17] in the near future. A more detailed version of this paper can be found online at http://www-p ersonal, umich, edu/~ ap o/isdens.html.
~
.
3. 4.
7. 8. 9. 10. 11. 12.
17.
D. A. Bryant and T. L. Cline and U. D. Desai and F. B. McDonald, Phys. Rev. Lett., 14 (1965), 481. C.W. Barnes and J.A. Simpson, Astrophys. J. Lett.,210 (1976), L91. A. S. Krieger and A. F. Timothy und E. C. Roelof, Solar Phys., 29 (1973), 505. I. G. Richardson and L. M. Barbier and D. V. Reames and T. T. van Rosenvinge, J. Geophys. Res., 98 (1993), 13. P. R. Gazis and F. B. McDonald and R. A. Burger and S. Chalov and R. B. Decker and J. Dwyer and D. S. Intrilligator and J. R. Jokipii and A. J. Lazarus and G. M. Mason and V. J. Pizzo and M. S. Potgieter and I. G. Richardson and L. J. Lanzerotti, Space Sci. Rev., 89 (1999), 269. Z. K. Smith and M. Dryer and R. S. Steinolfsen, J. Geophys. Res., 90 (1985), 217. Y. C. Whang, J. Geophys. Res., 103 (1988), 17,419. V.J. Pizzo, J. Geophys. Res., 99 (1994), 4173. V.J. Pizzo, personal communication (2001). L. F. Burlaga and N. F. Ness, J. Geoph. Res. 101 (1996), 13,473. P. Isenberg, J. Geophys. Res., 91 (1986), 9965. C. F. Barnett, Atomic Data for Fusion, Tech. Rep. ORNL-6086, Vol. 1 Oak Ridge National Laboratories, TN, 1990. D. Rucifiski and A. C. Cummings and G. Gloeckler and A. J. Lazarus and E. MSbius and M. Witte, Space Sci. Rev., 78 (1996), 73. G. Gloeckler and L. A. Fisk and J. Geiss, Nature, 386 (1997), 374. D.P. Cox and R.J. Reynolds, Annu. Rev. Astron. Astrophys., 25 (1987), 303. E. Quemerais and J.-L. Bertaux and B.R. Sandel and R. Lallement, Astron. Astrophys., 290 (1996), 941. P. Riley, personal communication (2000).
-319-
This Page Intentionally Left Blank
R. B. Decker a, C. Paranicas a, S. M. Krimigis a, K. I. Paularena b, and J. D. Richardson b ~Applied Physics Laboratory, Laurel, MD 20723, U.S.A. bMassachusetts Institute of Technology, Cambridge, MA 02139, U.S.A. Models of heliospheric shock propagation predict that pickup ions dominate the pressure in the outer heliosphere and play a key role in weakening CIR-associated shocks beyond a few tens of AU. To test this prediction, we used Voyager 2 PLS and LECP data to examine the temporal relationship between recurrent (~ 26-day) intensity increases of energetic ions and shocks or shock remnants. We focused on the periods 1983-1985 (11-19 AU) and 1992-1994 (36-45 AU). We found that ion intensity peaks lag further behind shocks or shock remnants during the later period, consistent with model predictions. 1. I n t r o d u c t i o n Forward and reverse shocks bounding corotating interaction regions (CIRs) accelerate ions to energies of at least 10 MeV/nucleon [1]. Between ~ 3-20 AU, intensity peaks of these recurrent (~ 26-day) ion events correlate with the passage of shocks, as inferred from plasma and magnetic field data. Surveys of data from the Pioneer and Voyager spacecraft show that beyond ~ 15-20 AU, the occurrence rate and strength of CIR shocks decrease [2,3]. In previous work, we used Voyager 2 data beyond 30 AU to show that CIRs or their remnants are still associated with well-defined peaks in quasi-recurrent ion intensity increases. However, the temporal association between the measured abrupt increases in plasma flow speed and ion intensity peaks becomes less clear farther from the Sun [4]. Models of shock propagation in the heliosphere show that beyond a few tens of AU, pickup ions dominate the pressure, thereby weakening CIR-associated shocks [5]. This weakening reduces a shock's efficiency for injecting and accelerating ions. A shock with speed Vs will outrun the intensity peak of accelerated ions (formed when the shock was relatively strong) that convects outward at the post-shock (downstream) solar wind speed Vd < Vs. Thus, peak ion intensities should lag the shock or shock remnant in the outer heliosphere when the contribution to the total pressure from pickup ions dominates [6]. 2. Observations The Voyager 2 LECP data we discuss herein are 6-hour averaged intensities of 0.521.45 MeV protons. These data have been carefully corrected by removing background counts due to penetrating cosmic rays. We have also examined intensities of eight Z>_1 ion channels, 0.040-4.0 MeV. These data corroborate our conclusions based on analysis of the proton channel. To determine if the predicted lag is consistent with observations, we
-321 -
R.B. Decker et al.
Figure 1. Quasi-recurrent (~ 26 day) increases in 0.52-1.45 MeV proton intensities during" 1983-86 (top) and 1992-95 (bottom).
analyzed Voyager 2 data during two time periods, one solar activity cycle apart. These periods are 1983-1985 and 1992-1994, when Voyager 2 was between 11-19 AU and 36-45 AU, and lie within the declining phases of the solar activity cycles during the peak CIR "season." The intensity of 0.52-1.45 MeV protons during the two periods of interest is shown in Figure 1. Vertical lines indicate arbitrary 26-day intervals. If the model is correct, the time lag between observed ion intensity peaks and signatures of shocks or shock remnants should increase with radial distance from the Sun. For each time period, we selected candidate shocks from the Voyager 2 data set and estimated the lag time between that measurement and the measured ion intensity peak. Only single peaks were chosen for this study and the times used were located at the mean along the FWHM of the peak. Representative data are presented in Figure 2. CIR-associated shocks weaken in the outer heliosphere [4], so changes in solar wind parameters become less discrete at larger helioradii. We are thus comparing shocks in the first period with shock remnants in the second period. As Figure 2 reveals, the size of the flow speed changes associated with a shock remnant are often comparable to those associated with a shock. For simplicity, we will refer to both kinds of discontinuities in the solar wind parameters as "shocks" in this paper. Shocks identified for this study originated from a candidate list of discontinuities in solar wind parameters. The P LS team compiled this list by identifying changes in solar wind speed, total density, thermal ion temperature, and, when available, magnetic field strength from the MAG instrument. The specificity of these criteria decline with time due to changes in the nature of the shocks, the solar wind parameters, and the sensitivities of the various instruments. The list of candidate discontinuities was therefore more comprehensive for the earlier study period, 1983-1985. For this reason, we used energetic particle data to expand the list of shocks in the outer heliosphere. The presence of sig-
- 322-
Recurrent ion events and plasma disturbances et Voyager 2: ...
Figure 2. Top to bottom" Solar wind speed, energetic proton rates, and delay time during 1984 (left) and 1994 (right). Vertical lines show shock or shock remnant passage. Dots in lower panels show delay times (ordinate) and times of particle intensity peaks (abscissa).
nificant intensity peaks occurring nearly simultaneously with changes in the solar wind parameters were regarded as shocks. We used the criteria that proton increases must have single peaks and that the peak count rates exceed 0.02 c s -~. This also allowed us to identify shocks where the plasma instrument had data gaps, but where the criteria for a shock was satisfied on both sides of the gap. The modified candidate list produced 17 (13) events in the 1983-1985 (1992-1994) period. We also assessed the variation of peak proton intensity with heliospheric distance. These results are shown in Figure 3(a). Only proton peaks associated with shocks were selected. Figure 3(a) shows that between these two study periods, the peak proton intensity decreased with helioradius as r -2"9. Time delays are shown in Figure 3(b). 3. S u m m a r y and conclusions During 1983-1985, the average delay time between the shock and 0.52-1.45 MeV proton intensity peak was ~ 0.6 days (5= 0.9 days) whereas during 1992-1994, it was ~ 3.4 days (+ 2.8 days). This radial increase in time lag, which is consistent with model predictions [6], and the c< r -3 peak intensity decrease have implications for the processes of shock evolution and particle transport in the outer heliosphere and for the expected intensity of CIR-associated seed ions incident on the heliospheric termination shock. Energetic
- 323 -
R.B. Decker et aL
Figure 3. Peak intensity (a) and delay time (b) of 0.52-1.45 MeV protons versus helioradius for quasi-recurrent events observed during 1983-1985 and 1992-1994.
ions are accelerated closer to the Sun, before the CIR shocks have dissipated, and then convect outward with the solar wind, undergoing spatial diffusion, adiabatic deceleration, and possible reacceleration by solar wind turbulence and other interplanetary shocks. When they reach the termination shock, a small fraction of these superthermal ions will be accelerated to ACR (anomalous cosmic ray) energies (~ 10-100 MeV/nucleon). This work was supported at APL by NASA Grant NAG5-4365 and at MIT by NASA Grant NAG-6473 and NASA Contract 959167. REFERENCES
1. 2. 3. 4. 5. 6.
G.M. Mason, et al., Space Sci. Rev. 89 (1999) 327. P.R. Gazis, et al., Space Sci. Rev. 89 (1999) 269. P.R. Gazis, J. Geophys. Res. 105 (2000) 219. A.J. Lazarus, et al., Space Sci. Rev. 89 (1999) 53. G.P. Zank and H.L. Pauls, J. Geophys. Res. 102 (1997) 7037. W.K. Rice, et al., Geophys. Res. Lett. 27 (2000) 509.
-324-
Mapping the Detailed Structure of the Local Interstellar Medium S. Redfield ~ and J.L. Linsky b ~JILA, University of Colorado Boulder, CO 80309-0440, USA bjILA, University of Colorado and NIST Boulder, CO 80309-0440, USA High resolution Hubble Space Telescope (HST) absorption spectra are used to construct a model of the three-dimensional shape of the Local Interstellar Cloud (LIC). We use spherical harmonics as our basis functions and apply a least-x 2 fitting routine to determine the best fit. This technique provides a powerful means for determining the true structure of the Local Interstellar Medium (LISM) by providing predictions of the amount of LIC material for any arbitrary line of sight. As a first attempt, our model is successful at modeling the LIC with a quasi-spherical shape, having a volume of ~ 93 pc 3 and a mass of ~ 0.32 3//o. An axis of symmetry in the direction of 1 ~ 315 ~ may indicate that the shape of the LIC is influenced by the flow of hot gas from the Scorpius-Centaurus Association. Our model places the Sun near the edge of the LIC at a distance of _< 0.19 pc, and the Sun should encounter the boundary of the LIC in _< 5000 years. Implications for applying this technique to other nearby clouds are also discussed. Physical parameters of the LIC are listed. The homogeneity of the LIC is discussed with reference to a sample of Hyades stars observed with the Space Telescope Imaging Spectrograph (STIS) aboard HST. 1. L I C M o d e l We develop a three-dimensional model of the LIC by calculating distances to the edge of the LIC from high resolution HST spectra, and using spherical harmonics to fit all available observations [1]. We adopt the technique described in [2] in which we determine the distance to the edge of the LIC, dedge(LIC), along a given line of sight from the hydrogen column density (inferred from the deuterium column density in HST spectra). We assume that the interstellar gas moving with the LIC speed has a constant density, rtHI = 0.10 cm -a, and the LIC extends from the heliosphere to an edge determined by the value of Nm(LIC) along each line of sight. We fit spherical harmonics to 32 lines of sight for which we have values of d e d g e ( L I C ) o r upper limits from HST, EUVE, and Ca II data. The LIC model is clearly not a long thin filamentary structure such as those seen in optical images of some interstellar clouds (e.g., reflection nebulae in the Pleiades), but neither is it spherical in shape. As seen from the North Galactic Pole, the LIC has an axis of symmetry that points in the direction 1 ~ 315 ~ (see Figure 2 in [1]). Since the
-
325
-
S. Redfield and J.L. Linsky
Figure 1. The location of the HST, EUVE, and Ca II stars used in this analysis are shown in Galactic coordinates. The symbols identifying the stars are listed in the first columns of Tables 1 and 2 of [1]. The shadings indicate values of NHI in units of 10 is cm -2 from the Sun to the edge of the LIC, based on Data Set B. From darkest to lightest, the shadings designate > 2.0, 1.0-2.0, 0.5-1.0, 0.25-0.50, 0.10-0.25, 0.05-0.10, and < 0.05 in these units. Figure from [1].
direction of the center of the Sco-Cen Association is 1 - 320, the shape of the LIC could be determined by the flow of hot gas from Sco-Cen. The LIC model shows that the Sun is located just inside its edge in the direction of the Galactic Center and toward the North Galactic Pole. The absence of Mg II absorption at the LIC velocity toward a Cen indicates that the distance to the edge of the LIC in this direction is < 0.19 pc and the Sun should cross the boundary between the LIC and the G cloud in less than 5,000 yr [3,4]. The LIC model can be seen at the Colorado Model of the Local Interstellar Medium web site http://casa.colorado.edu/~sredfiel/ColoradoLIC.html. The input data, prescription for computing the model, and a tool for calculating the hydrogen column density in any direction are also at this web site. As new data appear and we can model other warm clouds, updated versions of the LISM model will be placed in this web site.
-
326-
Mapping the detailed structure of the local interstellar medium
Table 1 Physical and Morphological Properties of the LIC Value Property 7000 + 1000 K Temperature 0.10 cm -3 Neutral hydrogen density (nHi) 0 911+0-12 Electron density (he) ~-0.06 cm -3 0.52-+-0.18 Hydrogen ionization (X(H) = np/(nHi + rip)) +1280 Gas Pressure (P/k) 1620_630 cm - 3 K -1.1 + 0.2 Log (depletion of Mg), D(Mg) -1.27 Log (depletion of Fe), D(Fe) -o.25 Log (depletion of O), D(O) 6.2pc Maximum dimension 4.7 pc Minimum dimension 1.9 x 1018 cm -2 Maximum NHI 1.5 x cm -2 Minimum NHI
Reference [5] [2] [6]
[1] [1] [5] [7]
[7] [1] [1] [1] [1]
1.1. P h y s i c a l P r o p e r t i e s
Table 1 lists our adopted empirical properties of the LIC and the references from which these data were taken. The physical parameters and hydrogen column density of the LIC are roughly consistent with the warm ISM models that assume pressure and ionization equilibrium [8], but the empirical hydrogen ionization is much higher and the gas temperature lower than the theoretical models predict. The high ionization is naturally explained by the LIC gas being in a recombining phase following shock ionization from a nearby supernova as proposed by [9]. The higher ionization increases the gas cooling, which can explain why the gas is 2400 K cooler than the ionization equilibrium models predict. Computed and observed temperatures are remarkably in agreement for a theoretical model with the observed LIC electron density.
1.2. H o m o g e n e i t y "
the Hyades Sample
The Space Telescope Imaging Spectrograph (STIS) instrument aboard HST observed 18 members of the Hyades star cluster. This dataset was acquired under observing program 7389 with principal investigator E. Bohm-Vitense. Only the Mg II h and k lines were observed. Because all 18 stars are members of the Hyades cluster, their lines of sight are very closely spaced on the sky. Thus, they sample a very small region of the LIC, and therefore offer a unique opportunity to study the homogeneity of the LIC. Figure 1.2 shows the difference in Mg II column density as a function of angular distance. The Mg II column density difference does not consistently exceed a factor of two for stars less than 8 ~ away. The hydrogen column density calculated by our LIC model [1] successfully predicts the gradient of the Mg II column density in the Hyades sample [10].
-327-
X RedfieM and J.L. Linsky
Figure 2. The difference in Mg II column density as a function of angular distance. With 18 target stars, there are 153 baselines, and therefore many angular distances to compare. The Mg II column density difference does not consistently exceed a factor of two for stars less than 8~ away. Figure from [10].
REFERENCES 1.
2. 3. 4. 5. 6. 7. 8. 9. 10.
Redfield, S., & Linsky, J.L. 2000, ApJ, 534, 825 Linsky, J.L., Redfield, S., Wood, B.E., & Piskunov, N. 2000, ApJ, 528, 756 Linsky, J.L., Wood, B.E. 1996, ApJ, 463, 254 Wood, B.E., Linsky, J.L., & Zank, G.P. 2000, ApJ, 537, 304 Piskunov, N., Wood, B.E., Linsky, J.L., Dempsey, R.C., & Ayres, T.R. 1997, ApJ, 474, 315 Wood, B.E., & Linsky, J.L. 1997, ApJ, 474, L39 Linsky, J.L., Diplas, A., Wood, B.E., Brown, A., Ayres, T.R., & Savage, B.D. 1995, ApJ, 451,335 Wolfire, M.G., Hollenbach, D., McKee, C.F., Tielens, A.G.G.M, & Bakes, E.L.O. 1995, ApJ, 443, 152 Lyu, C.-H., & Bruhweiler, F.C. 1996, ApJ, 459, 216 Redfield, S., et al., in progress
- 328-
Effect of Different Possible Interstellar Environments on the Heliosphere: A Numerical Study H.-R. Miiller ~, G. P. Zank ~, and P. C. Frisch b* ~Bartol Research Institute, University of Delawar(i, Newark, DE 19716 bDepartment of Astronomy and Astrophysics, University of Chicago, Chicago, IL 60637 At present, the heliosphere is embedded in the local interstellar cloud (LIC) which is warm (8000 K) and moderately dense (~0.2 cm-3). The sun will leave the LIC in perhaps about 10,000 years. This suggests that during its lifetime, the heliosphere has been, and will be, exposed to different interstellar environments. By means of a multi-fluid model, the interaction of the solar wind with a variety of partially ionized interstellar media is investigated. We assume that the basic solar wind parameters remain/were as they are today but we consider a range of ISM parameters (hot and cold, low density and ultrahigh density). In response to different interstellar boundary conditions, the heliospheric structure changes, as does the abundance of neutrals in the inner heliosphere, with possible implications for pickup ions and cosmic rays in the vicinity of Earth. 1. I N T R O D U C T I O N The Sun and its heliosphere is currently embedded in a warm interstellar cloud that has a relative Sun-cloud velocity of 26 km s -1. The main components of this cloud are neutral hydrogen (H) atoms and a plasma consisting of protons and electrons. The interaction of the local interstellar medium (LISM) with the fully ionized solar wind gives rise to the heliospheric morphology which includes the heliopause HP (a contact discontinuity separating solar wind and LISM), the termination shock TS (where the solar wind becomes subsonic and is diverted downstream to form a heliotail), and, when the LISM plasma is supersonic, a bow shock upstream of the heliopause. These general boundaries are created by the plasma interaction, yet the presence of neutral H and its coupling to the plasma protons via charge exchange greatly influences the details of the heliospheric morphology and the location of its boundaries (see [1] for a review). By studying the absorption of stellar profiles of nearby stars (within ..o100 pc) at various wavelengths along many lines of sight, it is possible to explore the fine structure of the interstellar medium surrounding the solar system [2]. There is evidence for different clouds and cloudlets, ranging in scale from fewer than 1 pc to tens of pc [3]. Some of these clouds possess characteristics that are quite different from those found in the local interstellar cloud (LIC). The clouds' inferred velocities can be different, their temperatures can be *This research was supported by NASA grant NAG5-6469, NSF-DOE award ATM-0078650,JPL contract 959167, and NASA Delaware Space Grant College award NGT5-40024.
- 329-
H.-R. Miiller, G.P. Zank and P.C. Frisch
higher or lower, and their densities and ionization states can span a wide range of values. Moreover, it is conceivable that the interstellar magnetic field, which is difficult to infer, differs from region to region. The Sun entered the cloud system of the so-called Scorpius-Centaurus association some 105 years ago [3]. There are indications that the Sun will leave our current immediate LIC in under 104 years from now [2]. This suggests that throughout Earth's history the solar system has moved between different clouds (or, depending on the viewpoint, different clouds have flowed past the solar system). Different interstellar environments should be capable of producing noticeable changes in the environment at 1 AU, as indicated by the amount of neutral H, anomalous (ACR), and galactic cosmic rays (GCR) at 1 AU. It has been suggested [4] that lunar soils contain an archive of elemental abundances that are different from the particle environment of the present era. Antarctic ice cores show signatures that may be interpreted as cosmic ray background variations at Earth [5]. These possibilities have motivated us to study how the global heliosphere would behave under LISM conditions that are different from today's. We undertake a numerical study, probing the parameter space of LISM density, temperature, and velocity. The parameter space is huge, of course, and we present the results of ten choices that are somewhat arbitrary, but are in a range suggested by interstellar clouds elsewhere or by other star's different velocity with respect to their LISM.
2. M O D E L S
We use the multifluid model developed by [6] to generate ten heliospheric models. The multifluid code tracks the plasma component and three thermodynamically distinct populations of neutral H. The models are essentially three-dimensional with spherical geometry; however, we assume azimuthal symmetry about the stagnation axis (the axis parallel to the LISM flow that contains the Sun) which reduces the dimensionality to two. The inner radial boundary condition is a prescribed solar wind plasma that corresponds at 1 AU to an averaged plasma density of 5.0 cm -3, a temperature of 10~ K, and a radial velocity of 400 km s -1. Table 1 lists the varied LISM boundary conditions. The first three models represent the effect of increasing relative LISM velocities, ranging from 26 to 100 km s -1. The remaining seven models are ordered in decreasing heliospheric size, where the upstream termination shock location ranges from 180 to 9 AU. The LISM parameters have been inspired by real astrophysical systems. Model 1 corresponds to the e Ind system (see Mfiller, Zank, and Wood contribution in this volume), model 6 to c~ Cen. Models 0, 4, and 5 represent variants of the contemporary heliosphere. Model 2 is an example of a superbubble shell, a possible LISM environment that is the result of a nearby supernova explosion sweeping up the outflow of a starforming region. Models 3 and 7-9 correspond to several different types of clouds in the Orion association that have been observed in the ISM in the direction of 23 Orionis [7]. All ten models have a supersonic LISM plasma, which leads to the formation of a bow shock where the plasma is heated and decelerated. All models have a termination shock, a heliopause, and a hydrogen wall. As an example, Figure 1 shows data from model 9, with boundaries marked in the plot. In upstream directions, the hydrogen wall is typically confined between the HP and the bow shock (BS). In downstream directions,
-330-
Effect of differentpossible interstellar environments on the heliosphere:...
Table 1 Model parameters and results. LISM parameters X nH np T v cm -3 K km/s
6 0.4 0.14 0.10 7 0.5 0.10 0.10 8 ~0 11.00 0.15 9 ~0 15.00 0.20 Ionization fraction ) / =
Morphology TS HP BS AU AU AU
5650 25 98 8000 26 86 100 26 18 3000 26 9.5 np/(nH +np); nil, np,
Neutrals/Pickups nH(TS) npI(TS) nH(1AU) cm -3 cm -3 cm -3
140 280 0.04 120 245 0.04 32 130 1.10 15 35 1.70 npi: number density H,
0.00007 0.0009 0.00007 0.001 0.008 0.08 0.02 0.27 plasma, pickup ions.
the enhancement in neutral H typically does not follow the HP. The TS either has a bullet shape (models 1, 2, 4, and 7) or is spherical (models 0, 3, 5, 6, 8, and 9). The spherical case occurs when the LISM neutral H density is high (with the exception of model 1 where the high LISM speed creates a bullet shape), and is characterized by a heliosheath and heliotail plasma that is subsonic throughout. In contrast, in the bullet shape cases the initially subsonic plasma of the nose of the heliosheath accelerates to supersonic speeds like in a nozzle, shocks to subsonic speeds further downstream, and meets heliotail plasma at a contact discontinuity. There is a characteristic triple point where heliosheath shock, termination shock, and the contact discontinuity meet. 3. R E S U L T S
An obvious result of the comparison between the ten models is the variation in sizes of the heliosphere, as expressed in the location of upstream TS, HP, and BS in Table 1. A shrinking heliosphere is caused by the strength of the LISM ram pressure (dominated by the speed in cases 1 and 2) or similarly through the weakening of the solar wind ram pressure by an increased rate of charge exchange inside the heliosphere owing to an increased neutral number density (models 5, 8, and 9). Models 3 and 4 are especially large due to a lower LISM plasma ram pressure, and both a lower LISM temperature or a lower neutral density lets the heliosphere expand more. The responses of the heliosphere to these parameters are non-linear, and the inclusion of ACR acceleration and its feedback on the TS would further increase this non-linearity. There are several models listed in Table 1 whose heliosphere is so small that some of the outer planets find themselves beyond the TS in the hot heliosheath, at least for parts of their orbits (e.g. Fig. 1). The upstream distances from the Sun to the TS range from 9 to 180 AU, those to the HP range from 12 to 240 AU. The bow shock of a cool, tenuous LISM (model 3) is as far as 700 AU away from the Sun.
-331 -
H.-R. Miiller, G.P. Zank and P.C. Frisch
. . . . . . . .
i
. . . . . . . .
i
. . . . . . . .
:
"
'
. . . . . . .
i
. . . . . . . .
i
. . . . . . . .
T
10 s ++
101
nH
P + "
TH ,
"'.:'l'IP
TS
p
,
el__
101
H
m
if) i
i
i
I
i
i iii
i
i
i
i
i
i iii
i
i
I
i
. . . . .
i
i
i
i
i llI
i
i
i
i
....
|
i
i
i
i
i
i
i
i
Figure 1. Profiles of density (cm -3) and temperature (K) of model 9 along the upstream stagnation axis.
The density of neutral hydrogen that enters through the termination shock at the upstream stagnation axis varies over a wide range in the ten models; however the value of today's heliosphere (model 0, 0.04 cm -3) is reached or surpassed in all but the lowest density model 4. This finding even occurs at the fixed distance of 1 AU (last column of Table 1), suggesting that if the LISM hydrogen has been only slightly denser or faster in the past, there was a larger background of neutral H arriving at Earth with potential consequences for the terrestrial atmosphere [8,9]. We try to assess the consequences of different ISM environments for the ACR density at Earth by using the density of pickup ions at the TS as a proxy, calculated with a Vasyliunas-Siscoe type model [10]. The ACR and the GCR environment will influence planetary magnetospheres, terrestrial climate, atmosphere, and biology [11]. Again, most models have a higher value than that of the model of the current heliosphere. These values, however, are only a rough indicator of the terrestrial cosmic ray background. The shrinking distance to the heliospheric boundaries alone (models 7-9) will increase the GCR background [11], and the high neutral and pickup density in models 8 and 9 leads to a strong modification of the TS through the ACR acceleration processes. REFERENCES .
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
G. P. Zank, Space Sci. Rev. 89 (1999) 413. J. L. Linsky, S. Redfield, B. E. Wood, & N. Piskunov, ApJ 528 (2000) 756. P. C. Frisch, Space Sci. Rev. 72 (1995) 499. R. F. Wimmer-Schweingruber & P. Bochsler, AIP Conference Proceedings 528 (2000). G. M. Raisbeck et al., Nature 326 (1987) 273. G. P. Zank, H. L. Pauls, L. L. Williams, & D. T. Hall, J. Geophys. Res. 101 (1996) 21639. D. E. Welty, L. M. Hobbs, J. T. Lauroesch, et al., ApJ Supp. 124 (1999) 465. M. Bzowski, H. J. Fahr, & D. Rucinski, Icarus 124 (1996) 209. G. P. Zank & P. C. Frisch, ApJ 518 (1999) 965. V. M. Vasyliunas & G. L. Siscoe, J. Geophys. Res. 81 (1976) 1247. K. Scherer, in: The Outer Heliosphere: Beyond the Planets Copernicus Gesellschaft (2000).
-332-
General Discussion Baranov to Richardson: Are pick-up ions included in your observations? Richardson: No, their fluxes are too low and they are too hot for us to see. Jokipii to Richardson: Given that the model does not even come close to represent the data, would it be fair to say that there is no real evidence for the wind slowing down? Richardson: The wind is definitely slowing down. Wc arc definitely seeing the 40 km/s decrease. We are just not seeing the 80 km/s decrease one would expect with George Gloeckler's numbers. Barnes to Richardson: With respect to the temperature plateau you see: isn't it true that there is a variation in energy in various forms between what's left of the stream structure, pressure imbalances and so on - isn't there sufficient cnerg3z to account for the heating? Richardson: That's certainly possible. But I would expect more heating near solar maximum rather than less. Marsch to Richardson: Did you look at the coronal conditions during the time period when the temperature drops? Because it's generally known that the slow solar wind has a lower tempertaure to start with, so there is a natural variation of the temperature in the source region of the solar wind. Richardson: Yes, it is possible that it is the difference between slow and fast wind temperatures. Fahr to Lallement: You gave us the view that we are sitting in one of these cloudlets and we are just moving along the periphery. But these cloudlets are co-existing with the hot medium around them, and you may suspect that there is a transition region. Is it possible that we move through such a transition region? Lallement: Yes, but we don't know. It is also possible that the cloudlets are touching each other, so that no hot medium is in between. One of the biggest problems in the physics of the Local Bubble is that nobody knows about the interfaces between the cloudlets and the hot gas. Cummings to Lallement: Do you assume a 40% ionization of helium in the local interstellar medium, or where do you get it from? Lallement: That is a consistent result from line-of-sight measurements to white dwarfs. But we cannot exclude a local variation. Dorman to Lallement: Do you have some information on the magnetic field in the cloudlets? Lallement: If there is a pressure equilibrium between the cool and the hot gas (and given the age of the local bubble there should be) the only plausible way to produce the equilibrium is to have a strong magnetic field at the boundaries of the cloud.
-333-
This Page Intentionally Left Blank
Session 6: View of Space Agencies and Companies
This Page Intentionally Left Blank
-337-
M. LeipoM
-338-
Solar sail technology development and application...
-339-
M. Leipold
-
3 4 0
-
Solar sail technology development and application...
-341
-
M. LeipoM
-342-
Solar sail technology development and application...
-
343
-
M.. Leipold
-
3 4 4
-
Oral papers and posters
CONTRIBUTIONS TO OUTER HELIOSPHERE MISSIONS IN FRAME OF THE GERMAN SPACE SCIENCE PROGRAMME O. R S h r i g German Aerospace Center, Directorate Space Programmes, D-53 183 Bonn. A short overview of the German Space Science Programme will be presented with emphysis on the past, present and future German engagements in heliospheric missions. Special attention will be given to the highlights, starting with the German/US project HELIOS, and the strong participation in the ESA projects ULYSSES and SOHO. German experiments have flown on the US Pioneer and Voyager Probes. But also ion release experiments and ground based experiments on German initiative have contributed to the picture of the sun which we have today. An outlook into the future will deal with the German intention to participate, e.g. in the missions STEREO, Solar Probe, Solar Orbiter. At last, solar sail technology may become a promising tool for future outer heliospheric missions.
DEEP SPACE LENGES
MISSIONS
AND
TECHNOLOGICAL
CHAL-
C. Schalinski and K. Eckardt Astrium GmbH, Friedrichshafen. Current and future missions require advanced technologies for instruments and spacecraft, e.g. formation flying of separated satellites with a high degree of onboard autonomy for interferometric missions like LISA or DARWIN. Astrium GmbH (former Dornier Satellite Systems) has acquired leadership among European Space companies as main industrial contractor for cornerstone missions in ESA's science program since the successful launch of ULYSSES. We will present progress on current projects and study results.
- 345 -
This Page Intentionally Left Blank
General Discussion Mewaldt to Leipold: What are the goals of the deployment [of a solar sail] in space? Are you also going to use the photon pressure to accelerate or navigate? Leipold: The objective is to keep it low cost, so we have to limit it to a pure deployment mission. Gruntman to Leipold: What are the physical limits on the minimum thickness of a solar sail? Leipold: The thinnest film samples that have been coated and that I am aware of are of the order of 1 micron. Genta to Leipold: What is the largest sail you think you can built with your boom design technique? Leipold: Right now it's 40 by 40 m. The problem with larger sails will be the load on the boom root. Marsch to Leipold: Since you expose the sail to a plasma: did you look at the problems of electrostatic charging? Leipold: Not yet. We have actually designed some experiments that will be performed this year. Sackheim to Leipold: Has anybody looked at a hybrid approach- i.e. an ion propulsion followed by a sail mode? Leipold: We have not looked at that at this point.
-347-
This Page Intentionally Left Blank
(1)
t~ C/3 O
or) Imio
~b
0 --,1
0
O0 ('b
This Page Intentionally Left Blank
The Probable Chemical Nature of Interstellar Dust Particles Detected by "CIDA" Onboard "STARDUST" J. Kissel a, F.R. Kruegerb, J. Silen c, and G. Haerendel a a Max-Planck-Institut fuer Extraterrestrische Physik, Giessenbachstrasse, Postfach 1312, D-85741 Garching, Germany b lngenieurbureau Dr. Krueger, Messelerstr.24, D-64291 Darmstadt, Germany c Finnish Meteorological Institute, Vuorikatu 24, SF-00101 Helsinki, Finland
The dust impact mass spectrometer CIDA onboard the NASA spacecraft STARDUST has detected five interstellar particles and recorded mass spectra during its first measuring period. Due to their unexpected complexity the analysis is an arduous task and requires new methods. However, the dominating substance class - namely, polymeric heterocyclic aromates and aliphates - and the particle masses could be determined. Both together are consistent with optical properties of those interstellar particles being able to reach the inner solar system.
1. INTRODUCTION On the 7th of February 1999 the spacecraft STARDUST had been launched from Cape Canaveral/F1. to its long journey to comet p/Wild-2, and further back to earth. Its main task in January 2004 is sampling cometary dust by aerogel plates made from silicon dioxide of very low density (approx. 0.03 g/ml), bringing them back to earth in January 2006. However, it is expected that the organic component of the dust may be altered or even be destroyed during the collection by the aerogel. As a result the information gathered will be limited to the mineral component, and to the isotopic distribution of all elements contained in the dust, except for silicon and oxygen, of course. In order to perform an in-situ chemical analysis of the cometary dust during encounter we implemented a dust particle impact mass spectrometer onboard, named CIDA (Cometary Impact Dust Analyzer). Not only will CIDA analyze the molecular composition of the dust of comet p/Wild-2, but also that of interstellar dust intercepted during its flight from earth to the comet and back to earth again. Due to the fact that the trajectories of the interstellar dust streams are often perpendicular or at least inclined relative to those of the interplanetary dust, there are periods of some months each during which the instrument is sensitive to interstellar
-351 -
Jr. Kissel et aL
dust exclusively. One of those periods is already completed, and the results are reported here. Before encounter with p/Wild-2 there are two more such periods, and, if all goes well, also during the spacecrafts's journey back home.
2. THE MEASURING PRINCIPLE OF "CIDA" Already as early as 1986 three dust impact mass spectrometers (PIA on board GIOTTO, and PUMA-1 and -2 onboard VEGA-1 and -2) encountered comet p/Halley. With those we were able to determine the elementary composition of the cometary dust. Molecular information, however, was only marginal. This was due to the fact that the relative velocities between each spacecraft and the coma of this retrograde comet, were extremely high (GIOTTO: 69 km/s; VEGA: 78 km/s). Because p/Wild-2 is a prograde comet, the velocity relative to STARDUST is only about 6 km/s. The velocities relative to the interstellar dust are estimated to be in the order of 25 km/s, but may be even lower due to the effect of the radiation pressure on these particles. In any case, with these velocity regimes CIDA is more sensitive to molecular rather than atomic information. Depending on the position and pointing of the CIDA instrument, interstellar, interplanetary, or cometary particles may impact on the about 150 cm 2 circular silver target (Fig.l). During impact they instantaneously evaporate forming molecular and atomic fragments (their yields depending on the impact speed, as already mentioned), which are in part electrically charged, either positively or negatively. By a voltage of U=+/-1 kV applied to the target either positive or negative ions are accelerated through the grounded acceleration grid. Due to the fact that the main chemical information is "contained" in the positive ions, only those are measured until encounter with the comet. Namely, the detection of negative ions is considered more risky due to the higher voltage required for the instrument's ion detector (which in any case measures secondary electrons after cathodic conversion). After encounter also negative ions can be measured during some flight periods. As a result special information about the oxygenand sulfur-chemistry of some organic functional groups (or mineralic parts) may be gained.
-352-
The probable chemical nature of interstellar dust particles
Fig. 1" CIDA-sensor schematics, for explanation see text. After acceleration, all ions of different molecular (or atomic) masses m (in dalton) possess the same kinetic energy E = 1/2 m v 2 = e U . Consequently, their velocity v in the subsequent drift space is inversely proportional to the square-root of their mass m, respectively (empirical evidence suggests ions with one single elementary charge e exclusively). However, a certain velocity distribution width is due to different initial energies of the ions produced. This uncertainty is compensated in first order by a reflector. Namely, those ions of the same mass m, travelling "too fast" due to their higher initial energy, will fly deeper into the electrostatic reflector than the others. Due to their thus longer flight path they arrive at the same time at the detector as the slower ones of the same mass m. Thus the total travel time (from the target all the way to the detector) t of each ion of mass m is directly proportional to the square-root of m. Thus yields t = a*sqrt(m) + b ; with "a" being a function of the flight path effective length and the voltages applied; in principle, "b" is given by the electronic signal travel times relative to the time t=0 for (an imaginary) mass m=0. The time spectrum of the ion current at the detector can thus be converted into a mass spectrum, exact knowledge of a and b provided. (If the impact time is u n k n o w n - which indeed is the case with three of the five impact events there are great mathematical difficulties to overcome in order to determine b.) The mass resolution m/delta-m is about 200. Thus the mass line to m=200 is still fairly resolved from that to m=201. The quality of the above transformation "time into mass" is best if there are
-353-
J. Kissel et al.
only spectral peaks due to INTEGER mass numbers AND a and b are in agreement with the actual parameters of the instrument. This transformation has been performed semiautomatically as follows. In the mathematical space ( a , b ) - due to the above transformation equation t=a*sqr(m)+b the fraction of those ion signals which corresponds to an interval of any integer mass number +- 0.1 daltons was calculated. If the time signals of the ions would have been randomly distributed over the entire time/mass regime, trivially about 20% of the ion flight times would be after transformation be found within the m+-0.1 windows. A good mass spectrum, however, would typically contain about 90% of all flight times in these mass windows. Indeed, near "canonical" (calculated by flight paths, voltages values, and pulse travel times) of a and b such transformations had been found with the impact mass spectra as discussed here. As a consequence, the quality of the transformation is good. The chemical information about the dust particle impacted is then "hidden" in the intensity distribution of the detector signals over the entire mass regime. In order to gain this information some other mathematical routines had been used, as described in the next chapter.
3. THE ANALYSIS OF THE MASS SPECTRA When analyzing substance mixtures in the laboratory, generally those mixtures are first (e.g., chromatographically) separated, and subsequently each single substance is analyzed mass spectrometrically. Naturally, any pre-separation is impossible with such fast dust particles. The mass spectrum generated by a particle impacting on the CIDA target thus represents the mixture within the particle as a whole for large impact velocities. With low impact velocities the mass spectra are more sensitive to the surface rather than to the interior of the particle. Consequently, one can only hope that at least an analysis of substance CLASSES majorizing the particle's matter is possible. Then it comes as a necessary condition for class analysis that the mixture is not too heterogeneous chemically. However, even in a favorable case this class type analysis challenges a completely new problem. ( - In this very context it appeared as a great help to us, that we, together with a group of chemists in Vienna, are just developing a chemometric method into the direction of analyzing functional groups in substance mixtures (1). As early as 1987 we made use of a principally similar, but very simple, method in analyzing the molecular ion contribution in p/Halley's dust impact mass spectra (2). However, a further methodological development is needed for our experiment COSIMA (COmetary Secondary Ion Mass Analyzer) onboard the ESA-spacecraft ROSETTA to comet Wirtanen, which will be launched in 2003. In rendezvous with that comet we will softly collect cometary dust on targets from metal black, and subsequently analyze it by Secondary Ion Mass Spectrometry (SIMS). As we know from laboratory experiments the chemical process of ion formation in SIMS is very much comparable to that in dust impact in the lower velocity (<25 km/s) regime). The chemometric method of analysis of substance classes, as used here, is actually a type of auto-correlation analysis. Just the mathematically most important (and used by us) one should be described: Let Y(m) be the ion yield of mass number m (within the +- 0.1 window, for instance). As m are taken as integer mass numbers, mass differences Am are also integer. Then the autocorrelation probability P(Am) is given by (c being a normalization constant) : 00
P(Am) = 1/c Z Y(m)" Y(m+Am) 1
- 354-
The probable chemical nature of interstellar dust particles If there is neither correlation nor anti-correlation in a mass spectrum the function P is a monotonuously decreasing series of values, because (1) the summation actually is not taken up to infinity but only to a maximal rn in the mass spectrum, and (2) the yields meanwise decrease with larger m. However, if there is are masses m correlated with masses, say, m+14, the value of P(14) will be generally larger than P(13), i.e.: P(Am) > P(Am-1). Let us take the alkanes as an example, like propane, butane, pentane, hexane, octane, and paraffines being higher alkanes. All linear alkanes are expressed by the general chemical formula C n H2n+2 (n=1,2,3,...). With the mechanisms of ion formation from the condensed phase (SIMS, Impact Ionization, et al.) they (among the majority of organic substance classes) most frequently form non-radical (i.e. even-electron) ions of the common form C n H2n+l + directly as a quasi-molecular ion. However, decomposition by elimination of CH2-groups leads to the same type of ions. A mass spectrum of such a mixture of alkanes is characterized by a strong abundance of ions corresponding to mass lines which molecular mass differ by a fixed Am = 14. The mass line basis is m=15, i.e., n=l (CH3+). - Pure carbon, as a further example, would show up with characteristic mass differences Am = 12 with the mass basis 12 : C +, C2 +, C3 + .... An over-statistical representation of Am = 48 (difference of 4 C - atoms) indicates graphite or large PAH's (Poly-Aromatic Hydrocarbons) with little contribution of hydrogen in ion formation. Diamonds ionize with an expressive even-odd pattern of the numbers of carbon atoms in the ion; thus in this case, Am = 24 is more representative than Am = 12 is. If, however, Am = 24, 48 is not statistically over-represented in comparison to Am = 12, this would indicate amorphous carbon, which by the way does not tend to form large molecular ions, in contrary to the other modifications of carbon. To state it already now: With our interstellar particles we did not find indications for these substance classes just mentioned. There are some additional chemical classificators than modulo-functions found by auto-correlations, however, being not treated here.
4. THE DOMINANT SUBSTANCE CLASS OF THE I N T E R S T E L L A R PARTICLES An important question is whether silicates may form the dominant substance class. Silicates mainly contain the elements O, Mg, Si, Ca, and Fe. Other minerals may contain C (carbides, carbonates) or S (sulfides), additionally or alternatively. As it is, their integer atomic mass numbers are multiples of '4'. Thus the mass numbers m of atomic and molecular ions thereof are also multiples of 4. Consequently, in this case we expect their mass spectra to be strongly autocorrelated with Am = 4 with a basis mass number dividable by 4, too. This is not found in any of the mass spectra! Silicondioxide, silicates, as well as iron minerals, can thus be excluded as dominant components of our interstellar particles. However, a large P(2) value was always found, which prooves the substance classes at least to be hydrocarbons- due to the "even-electron-rule" (or nonradical-ion-rule) valid in this type of ion formation. Examining the first two impact events was a great surprise. Up to mass numbers beyond m=1000 ions were present, even quite abundant. From the total ion number and its distribution
-355-
J. Kissel et al.
over m we were able to estimate the total mass of the impacting particles. It was, with the two first particles some nanogram each. Furtheron, their high ion masses strongly indicated a highly polymerized material, perhaps similar to plastics which are built from various copolymers. The next three particles having impacted, apparently were lighter, about 10 picograms. Their ion masses were distributed up to about m=400. However, this does not mean that they are less polymerized. Namely, due to their lower total mass, the detector ion current was lower, just dropping under threshold in the higher ion mass regime (fig.2). A further surprise was that we did not see many smaller particles impacting: CIDA's detection threshold is well below 0.1 picogram. Usual particle size distributions show more abundance with lower masses! The solution of this enigma arrived from totally different arguments (other than from mass spectrometry), and will be discussed in the last chapter.
Fig.2 : Top panel:the original time-of-flight data for the fifth particle 'Quintus' together with the target signal, indicating the time-of-impact. Lower panel: The massspectrum derived from the time-of-flight data, using known instrument parameter a (see text).
-356-
The probable chemical nature of interstellar dust particles
What are the functional groups of the giant molecules we found in the interstellar dust, and how did we determine them? After having excluded the dominance of several (the rockforming) elements, there are only 5 elements left which could have built up the majority of these polymers, namely: Hydrogen (H), carbon (C), nitrogen (N), oxygen (O), and (unsure) sulfur (S). The integer mass numbers of stable molecules built from these elements except nitrogen (!) are even! (This is due to the even valence numbers of C, O, and S and the even mass number of their main isotopes.) Abundant non-radical positive ions (formed by protonization, negatively charged by deprotonization) thus have odd mass numbers! This is still tree, if the molecule (or ion, respectively) contains an even number of N atoms, as can be easily verified. Thus, Am = 2 is the strongest auto-correlation, as being well established in all 4 impact mass spectra analyzed yet. The main base mass numbers are thus odd, especially with the 3rd ("Tritus") and 5th ("Quintus") impact: with these particles nitrogen apparently plays a minor role only (see fig.3). The furthermore found correlations and base masses there indicated an unsaturated-carbon chemistry with some oxygen contribution. Furanes or dioxines may be subgroups of those polymers; their main constituents, however, are more or less reduced (hydrogenized) homocyclic aromates which may also be tight together via oxygen bridge bonds (ether-type binding). It is important to note here that those (only in part aromatic) polymers are steric. I.e., their conformation deviates strongly and erratically from the planar geometry PAH's or graphite. They are probably more similar to tars, rather than crystallized matter The contribution of nitrogen is only low with the 3rd and 5th particle, however, it seems to be larger with the 1st ("Primunculus") and 2nd ("Dromedary") particle. This is recognized by the dominance of odd delta-m, esp. 15 and 27 (difference due to an HN- or an HCN-group or eliminant). An elementary structure may result from chinoline or carbazole groups, for instance (fig.3). But let us now focus onto condensed aromates - which apparently is the main substance class found -, with heterocyclic components: Formally we depart from a homocyclic aromate like coronene (C24H12). An exchange of carbon by oxygen can only take place with the outer ring atoms (or we may have keto-groups there). However, this is breaking locally the mesomeric electronic structure - that molecule would not be planar any more. An inner Catom cannot be replaced by an O-atom, because it has only two bonds; just an O-bridge (ethertype) is possible being also not planar. An N-atom, however, could well replace an outer CHgroup isoelectronically without breaking the planar geometry. An inner C-atom, however, can only be replaced by an N-atom disturbing electronics and conformation, because it causes a steric sp3-hybridic bonding which breaks the mesomeric electronic structure as well. As a result, steric ring annealation is then possible (as we know them, for instance, from alkaloids like sparteine, these being more reduced, i.e., hydrogenated), however. Dehydrogenated (oxidized) ring structures, some via N sterically annealed, and some bound via O-bridges, are the main characteristics of the interstellar dust matter found by CIDA. Those structures can be produced, e.g., by heavy irradiation of frozen mixtures of PAH's, water, and ammonia.
-357-
J. Kissel et al.
Fig. 3: The probable chemical structure of the organic material identified in the interstellar particles found with CIDA. Case a-c are for the 'big' particles (not discussed here), case d is for the 'small' particles.
5. I N T E R S T E L L A R DUST IN T H E SOLAR SYSTEM Of course, we do not and we will not have a representative sample of interstellar dust grains impacting CIDA, but we will get only those which survive the approach to the sun down to within 1 - 2 a.u. In contrary to the cometary dust - which also is said to be of interstellar origin - the free interstellar dust lacks the cooling by the cometary nucleus. Which free particles we meet depends on their gravitative, electromagnetic, optical, and chemical-thermodynamic properties. Particles smaller than some tenths of a micron can hardly absorb the sun light, and thus may travel into the inner solar system without being hindered. However, particles smaller than .2 um do not trigger a CIDA signal, thus we do not see them. Particles larger than some microns absorb the sun light more or less, depending on its constitution; some of them (the larger and the more refractive ones) may thus as well reach the inner solar system. Medium sized particles (which CIDA would see if being present) absorbing in the visible and nearultraviolet would either be reflected already far out by radiation pressure or evaporate (thus becoming smaller), respectively. Indeed, Landgraf et al. (3) think of having found such a particle size gap by their dust detectors onboard GALILEO and ULYSSES. In any case, CIDA did not see "small" particles between 10 femto-grams and 1 pico-gram, although it is sensitive in this regime. Moreover, if "normal" dust size distribution would be valid, the probability to find many of them in the measuring period would have been high. Namely, in this case smaller particles should be much more abundant than larger ones, of which CIDA had seen
-358-
The probable chemical nature of interstellar dust particles
even five pieces: By means of optical calculations (via the complex refraction index) Landgraf was able to show that those particle being detected by his instruments could not have been dominated by graphite or silicon dioxide. This is in agreement with our results. Let us repeat it again: We can only see those fraction of interstellar dust being intercepted by our instrument in the 1 - 2 a.u. distance regime from the sun. Less polymer organic substances than those found can survive in grains only in the coldness of far space. They evaporate in the solar system. Graphite grains or those constituted by very large PAH's, if at all present, would be hindered to enter the inner solar system by radiation pressure. Nevertheless, some speculation may be allowed about the formation and constitution of that solid interstellar matter for which entry to our habitat regime is prohibited: Looking to the chemistry of original interstellar gas (!) surrounding far stars, a fraction of highly unsaturated carbon chain molecules and the contribution of nitrogen and oxygen to them is remarkable. Hydrogen interacts by far less than its abundance there might let us think. These unsaturated molecules, however, are highly reactive. They can grow even by simple two-body collisions due to their many inner degrees of freedom. Thus they polymerize and cyclisize in the cooler parts (<1000 K) of the interstellar space. Their aggregation to dust particles, inter alia those as we have seen, is thus easily possible. Also some patterns of the "diffuse clouds" may have optical properties compatible with homologues of such matter. By the way, the similarity of the first two particle's composition to Halley's dust lead us to check if the STARDUST spacecraft had flown through the dust trail in a comet's orbit. A comparison of the trajectory planes of some thousand comets relative to STARDUST's trajectory around the impact times did not show such an event. We like to estimate this calculation having been performed by our colleague Rita Schulz at ESTEC (4). Moreover, the sensitive direction of CIDA was not in the relative direction of the streams of interplanetary particles. Needless to say that a lot of work is still necessary in the near future. Right now we are lucky to have for the first time undisturbed matter from other stars "in our hands", to compare it to the dust which may have formed our solar system billons of years ago.
REFERENCES 1. K. Varmuza, W. Werther, F.R. Krueger, J. Kissel, E.D. Schmid, Int. Joum. of Mass Spectrometry, 189, 79-92, 1999 2. J. Kissel and F.R. Krueger, Nature Vol.326, 755 - 760, (1987) 3. M. Landgraf, W. J. Baggeley, and E. Grtin, JGR, 105, No. A5, 10, 343, 2000 4. R. Schulz (ESTEC), Private communication (Noordwijk 2000).
-359-
This Page Intentionally Left Blank
In-situ studies of interstellar dust from spacecraft Ingrid Mann a* and Hiroshi Kimura b aESA Space Science Department at ESTEC, Noordwijk, The Netherlands bMax-Planck-Institut f'tir Aeronomie, Katlenburg-Lindau, Germany Comparison of the present in-situ measurements of interstellar (IS) dust within the solar system to dust models based on astronomical observations can give clues about the evolution of IS dust. Hence, further in-situ measurements in the solar system are expected to provide a better understanding of IS dust properties. Nevertheless, measurements within the solar system cover the large size end of the IS dust distribution, while the majority of the very small IS dust can only be measured from spacecraft that cross the heliopause. 1. INTRODUCTION The study of interstellar (IS) dust particles is important for understanding elemental and isotope abundances in the interstellar medium (ISM) as well as for understanding the formation and evolution of grains. Astronomical observations reveal average optical properties of IS dust along the line of sight, but are biased to the sizes that yield the main contribution to the observed quantities, such as small dust particles determining the extinction of the ISM. The presence of IS dust in the solar system, on the other hand, provides the opportunity of retrieving ISM conditions with in-situ measurements. So far the in-situ measurements provide data of the mass density, mass distribution and flux rate of IS dust entering the solar system [1], which allow crude estimates of the basic properties. 2. EXPERIMENTAL RESULTS IS dust particles have been identified with measurements aboard Ulysses [2], Galileo [3] and Hiten [4] as well as radar meteor observations indicate larger particles to enter the solar system from interstellar space [5]. The flux of IS dust at solar distance 1.8 < r < 5.4 AU was derived to be 1.5 x 10 -4 m -2 s -1 from the Ulysses measurements between 1992 and 1995 [2]. The derived mass density of 2.8 x 10 -23 kg m -3 agrees with densities derived for the average ISM but is beyond the densities derived for the local ISM [1 ]. The mass distribution of IS dust differs for measurements made within and beyond a distance of 3 AU from the Sun, measurements at r < 3 AU showing a dip in the distribution around m ~ 10 - 1 7 kg. The dust impact rate for small masses has decreased since the beginning of 1996 [6] as a result of deflection in the solar magnetic field. While the different in-situ measurements of IS dust mentioned before are comparable within the detection limits, measurements beyond 5 AU on Pioneer [7] and on *Also at Instimt f'tirPlanetologie, Westf~ilischeWilhelms-Universit~itMttnster, Germany.
-361 -
I. Mann and H. Kimura Voyager [8] are controversial and may indicate a component of dust that is variable in space and time. Moreover, the detected grains cannot be clearly identified as interstellar [1 ], but may stem from comets or from dust production related to the Kuiper belt objects [9]. This points to the problem of separating IS dust from other dust components measured within the solar system. 3. DYNAMICS AND SEPARATION OF INTERSTELLAR DUST The separation of IS dust particles within a data set is based on their impact speed and direction which is different from solar system dust. The uncertainties of the speed and the direction measurements, as well as the dynamical effects that deflect the grains from their initially monodirectional stream hamper the separation. The identification based on statistical arguments and based on the comparison to other dust components seems, however, reliable. Particles at the large end of the size spectrum are influenced by solar gravitational force Fgrav and solar radiation pressure force Frad. If gravity is the dominant force, i.e., the ratio - - F r a d / F g r a v < 1, particles are in hyperbolic orbits focusing in interstellar downwind direction. This is the case for masses m > 10 -15 kg, for smaller grains with [3 > 1, particles are repelled in hyperbolic orbits. Particles with masses m < 10 -18 kg are deflected at the heliopause [10], while particles with m < 10 -17 kg can enter the solar system but are deflected from their original orbits in the solar magnetic field [11 ]. The mass distribution derived from Ulysses measurements between 1992 and 1995 is shown in Figure 1 as a histogram with error bars derived from Poisson statistics (see [ 12]). Two effects possibly reduce the amount of IS dust identified within the data at the small size part of the distribution compared to the actual dust flux: For one, the detection efficiency of the instrument is reduced for impacting dust with small masses. And secondly small grains that are deflected from the initial stream are not identified as interstellar when applying the velocity criterion. The measured distribution is assumed to better reflect the flux of IS dust within the solar system for masses m > 2.5 x 10-17 kg, shown in a hatched histogram in Figure 1 which allows a discussion of the properties of these large IS dust particles. 4. SIZE DISTRIBUTION AND MODELS OF IS DUST Comparing the results to ISM studies, the detected grains are clearly larger than grains of "typical" ISM dust models which describe the size distribution as dn/da , ~ a -3"5 with a dropo f f b e y o n d 10 -16 kg (radius a, a = 0.25 #m). The Ulysses results, in contrast, are approximated with a size distribution dn/da ~ a -2"65 for compact grains of constant density [1,12]. The derived size interval covers particle radii, a, 0.015 < a < 4.1 #m, when assuming spherical grains. Also IS dust particles found in meteorites indicate the existence of larger grains: RowanRobinson [13] could explain the observed millimetre emission from cold dust near the galactic plane with grains of 30 #m radius. Moreover composite grains of 0.003 to 3 #m in size are suggested by a recent model to account for revised abundances of the heavy elements in the ISM [14]. The large dust particles can be interpreted as results of the coagulation of grains in the ISM. The suggested size distribution is shown in Figure 1. Although it is different from the distribution derived from Ulysses measurements, both results point to the existence of large grains in the ISM and the in-situ measurements provide the opportunity of estimating particle properties for this part of the size distribution. Large particles predominantly influenced by solar
- 362 -
In-sire studies of interstellar dust from spacecraft
Figure 1. The number density of IS dust as function of mass: Ulysses measurements [12] (a histogram with error bars); composite grain model [14] (solid curve).
Figure 2. Calculated ~ ratios for large composite grains (solid curve) and for coremantle particles (dotted curve from [15]). Insitu results require condition 1.4 < [3 < 3.1 for m ,,~ 10 -17 kg.
gravity and radiation pressure force can be studied for instance with detailed measurements near 1 AU. Already the present data allow for estimating the acting radiation pressure force. 5. MODELS OF THE RADIATION PRESSURE FORCE We estimate the radiation pressure force that could explain the measured mass distribution of grains and its gap by repulsion of in-falling IS dust (see [1,6,12]). The [3 ratios required to lead to this repulsion are 1.4 < ~max < 3.1 for m ~ 10 -17 kg. We compare this to calculations of the [3 ratio for different model assumptions of IS dust as shown in Figure 2. Assuming compact particles with a silicate core and a mantle of an ice-dust mixture produced by surface condensation, the model calculation leads to a maximum 13value of 0.8 at a mass of 10 -16 kg [15]. These particular calculations [ 15] for core-mantle particles is based on the assumption that the mantle consists of water ice. The lifetime of the ice mantles, however, is limited by sublimation, so that these model assumptions are not suitable to describe the detected dust particles. We further assume properties of dust produced by coagulation growth [14] as suggested in the above mentioned model for large dust in the ISM. The model assumptions for composite grains meet the conditions in terms of the maximum [3 value as well as in terms of the mass at which the [3 value maximizes, if the grains are assumed to be of low porosity (45%) as shown in the figure. 6. DISCUSSION We conclude that at least a fraction of IS dust is accessible to in-situ studies in the solar system. In-situ measurements, even only of the mass distribution and the flux of IS dust, allow for some estimates of the particle properties. The conditions of the entry into the solar system are however depending on the properties of IS dust, which has to be taken into account for an analysis. Dedicated in-situ measurements would provide information about the size, structure
- 363 -
1. Mann and H. Kimura
and composition of some IS dust particles. The data lead towards a better understanding of the physics of dust in the ISM, but require a discussion as to how representative they are for the average ISM dust properties. Measurements would certainly benefit from improved detector capabilities, such as a better measurement of the dust velocities as well as from measurements of the dust composition. Since experimental data obtained in the solar system will cover a selected range of particle sizes, compositions and properties, measurements on spacecraft that leave the heliosphere are a further step. Such measurements could provide a larger sample of different IS grains and also reveal the properties and the behaviour of smaller dust particles that are expected to be most abundant in the ISM and prevented from entering the heliosphere. Moreover, the data would allow for a study of the entry conditions and hence would also increase the scientific return of the measurements carried out within the solar system.
REFERENCES 1. 2. 3. 4.
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
I. Mann and H. Kimura, J. Geophys. Res. 105 10317 (2000). E. Grtin, B. Gustafson, I. Mann, et al., Astron. Astrophys. 286 915 (1994). E. Grtin, P. Staubach, M. Baguhl, et al., Icarus 129 270 (1997). H. Svedhem, R. Mtinzenmayer and H. Iglseder, In Physics, Chemistry, and Dynamics of Interplanetary Dust, B.A..S. Gustafson and M.S. Hanner (eds.), pp. 27, ASP, San Francisco (1996). A.D. Taylor, W.J. Baggaley and D.I. Steel, Nature 380 323 (1996). M. Landgraf, J. Geophys. Res. 105 10303 (2000). D.H. Humes, J. Geophys. Res. 85 5841 (1980). D.A. Gurnett, J.A. Ansher, W.S. Kurth and L.J. Granroth, Geophys. Res. Lett. 24 3125 (1997). S. Yamamoto and T. Mukai, Astron. Astrophys. 329 785 (1998). H. Kimura and I. Mann, Astrophys. J. 499 454 (1998). G.E. Morrill and E. Grtin, Planet. Space Sci. 27 1283 (1997). H. Kimura, I. Mann and A. Wehry, Astrophys. Space Sci. 264 213 (1998). M. Rowan-Robinson, Mon. Not. R. Astron. Soc. 258 787 (1992). J.S. Mathis, Astrophys. J. 472 643 (1996). M. Wilck and I. Mann, Planet. Space Sci. 44 493 (1996).
- 364 -
Dynamics of Interstellar Dust at the Heliopause Andrzej Czechowski ~ and Ingrid Mann b ~Space Research Centre, Polish Academy of Sciences, Bartycka 18A, PL 00-716 Warsaw, Poland bSpace Science Department, ESA, ESTEC SCI-SO, 220 AG Noordwijk, The Netherlands We present preliminary results of a Monte-Carlo approach to the interstellar dust dynamics near the heliopause. An enhancement in number density of small (0.01 #m) grains at the flanks of the heliopause is found. 1. I N T R O D U C T I O N The entry of interstellar dust into the solar system as a result of the motion of the Sun relative to the interstellar medium (ISM), is accompanied by several mechanisms which lead to the depletion of the ISM dust flow (see Mann and Kimura [1]). In the case of small (<0.1pm) grains, the charge-to-mass ratio is relatively large ([2]) and the dynamics in the interface region between the heliosphere and the ISM is determined by the electromagnetic forces. Our aim is to estimate the distribution of these small dust grains around the heliosphere. Here we describe our approach and present first results. 2. M O D E L C A L C U L A T I O N S We treat the dust particles as "test particles" so that the influence of the grains on the plasma is assumed to be negligible. We determine the grains trajectories by solving numerically (Runge-Kutta 5th order) the equation of motion for charged dust grains dv/dt-~
(Qdust(t)/?Ttc)((v -- Vp) x B) + Fgrav/?TZ
(1)
where v is the grain velocity, vp the plasma flow velocity, E = -(1/c)(Vp x B) the electric field induced by the plasma flow, m the mass of the dust grain and Fgrav the solar gravity force. Qe~t(t), vp and B are calculated simultaneously by the code. The electric charge of the grain is obtained by solving the charging equation
aO s /at- Z
(2)
i
where Ji are the charging rates associated with the different charging mechanisms: the main ones are the photoelectron emission, sticking and recombination with plasma particles and secondary electron emission. The secondary electron emission increases as a result of high plasma temperature between the heliopause and the termination shock, leading to an increasing surface charge of the grains compared to the surface charge in
-365-
A. Czechowski and L Mann
the LISM. The method used to estimate the charging rates is based on Kimura and Mann [2]. For the smallest grains (,,~ 0.001 pm), instead of using Eq. (2) we treat the charging processes stochastically, which allows us to take charge quantization into account. In the present calculations we use a simple model of the heliosphere and the magnetic field ([3], [4]) which we extended to include the interstellar field. The plasma flow, axially symmetric respective to the ISM apex-antiapex axis and incompressible, is given by a simple analytical expression. The influence of the magnetic field on the flow is neglected. Although oversimplified, the model is useful because it doesn't require using a spatialgrid. In result, the flow and the field can, within the model, be calculated with desired accuracy, avoiding errors of interpolation. This is important if we want to describe the motion of small charged grains, with small Larmor radius, particularly in the regions close to the discontinuity surfaces (like the heliopause). In interpreting the results one must of course be aware of the model inaccuracies, in particular the assumption of incompressible flow and too high magnetic field intensity near the heliopause (which would tend to suppress crossing the heliopause by small grains). The model can incorporate a time variation of the magnetic field, so that the effect of solar magnetic field reversals during a solar cycle can be included. The code employs our modifications of the routines from "Numerical Recipes", in particular Runge-Kutta 5th order and the "stiff' routine, the latter needed for very small (~ 0.001 pro) grains. Our code is able to deal with the case of an arbitrary magnetic field and plasma flow defined on a spatial grid, although, as explained above we do not use this option here.
3.
FIRST
RESULTS
We present our results in the coordinate system with the origin at the Sun. The z - axis points in the direction of the motion of the LISM, y = 0 is the ecliptic plane (which we assume to contain the LISM apex, neglecting the ~ 7 o shift) and we assume for trial calculations that the interstellar magnetic field is perpendicular to the ecliptic with IBLISMI--- 5 # G , in y-direction. The corresponding spherical coordinates are (r, 0, r where 0 = 0 ~ is the ISM apex direction and 0 = r = 90 ~ the direction of the interstellar field. The radius of the termination shock (which is spherical in the model) is 85 AU and the distance to the stagnation point 127.5 AU. The particles start on a paraboloid approximating the surface of the bow shock with a closest distance of 195 AU from the Sun towards the apex. Their initial speed is 26 k m / s along the LISM flow direction. The plasma is assumed to slow down by a factor of 2 at the bow shock, so that the initial difference in speed between dust and plasma is 13 km/s. We followed 10000 particles each of 0.1 and 0.01 #m size, using the results to derive the density distributions in a Monte Carlo approach. Some of the trajectories are shown in the figures. Figures 1 to 2 show a sample of 100 trajectories for 0.01 #m silicate dust grains, projected onto respectively x-y and z-y planes (with the heliopause and the termination shock as a background). Figure 1 illustrates the deflection of particles towards the interstellar magnetic field direction. In Figure 2 one can see particles which leave the region near the heliopause producing a region of enhanced dust number density at the flanks
-
366-
Dynamics of interstellar dust at the heBopause
Figure 1. 0.01 #m grains, (x,y) plane
Figure 2. 0.01 #m grains, (z,y) plane
of the heliopause towards interstellar downstream direction. Neither of those trajectories crossed the heliopause. A similar sample of trajectories for 0.1 #m grains is shown in Figures 3, 4. The solar cycle variation is not included in this trial run and trajectories inside the termination shock are not calculated. The trajectories of particles which cross the heliopause show a tendency to bunch in respectively y > 0 and y < 0 halves of the heliosphere. The sample density distributions outside and inside the enhancement region for 0.01 #m grain size are presented in Figures 5, 6. Number density of the grains (relative units) is plotted as a function of the angle qS. The ranges of 0 and r (coresponding to the bins used in statistical analysis of the results) are shown in the figures. The enhanced density region appears at some distance from the heliopause in the x=0 plane (q5 = 90 ~ 270~
Figure 3. 0.1 #m grains, (x,y) plane
Figure 4. 0.1 #m grains, (z,y) plane
-367-
A. Czechowski and 1. Mann
Figure 5. The number density (relative units) of the 0.01#m grains as a function of r at the heliopause and for r = 2 1 4 - 238 AU; 0 = 39 ~ 47 ~.
Figure 6. The number density (relative units) of the 0.01#m grains as a function of r at the heliopause and for r = 2 8 7 - 307 AU; 0 = 113 ~ 123 ~.
4. D I S C U S S I O N Although our statistics is not sufficient to derive the density distributions in detail, from this first model calculation we find the region of modified dust distribution around the heliosphere for small size (0.01 #m) grains. The 0.01 #m particles are forming a region of enhanced dust number density on the flanks of the heliosphere close to the x=0 (VL~SM,BLISM) plane. The excess density increases with 0 (the angle counted from apex direction) and at 0 = 1200 the density in the plane parallel to BLISM (x=0 plane) is higher by a factor of 5 than in the plane perpendicular to BLISM (y=0 plane). Because of their larger Larmor radius, 0.1 #m particles outside the heliopause move approximately in straight lines. Inside the heliopause their charge increases and the effective magnetic force for the trajectories becomes apparent. Our model is easily generalized to other values of the heliospheric plasma parameters and dust properties. Although this application is beyond the scope of this paper we note that, compared to Linde and Gombosi [5], who estimate the amount of ISM grains that enter the inner heliosphere in a model of the heliosphere based on a stationary MHD solution, the present model also allows to study the influence of the solar cycle variations. REFERENCES 1. 2. 3. 4. 5.
Mann, I. and Kimura, H., 2000: J. Geophys. Res., 105, 10317-10328 Kimura, H. and Mann, I. 1998: Astrophys. J., 499, 454-462. Suess, S.T.; Nerney, S., 1990: J. Geophys. Res., 95, 6403-6412 Nerney, S.; Suess, S. T.; Schmahl, E. J, 1991: Astron. & Astrophys., 250, 556-564 Linde, T.J. and Gombosi, T.I., 2000: J. Geophys. Res., 105, 10411
-368-
Oral papers and posters
KUIPER BELT OBJECTS S.F. G r e e n Planetary and Space Science Research Institute, The Open University, Walton Hall, Milton Keynes, MK7 6AA, U.K.. The presence of a large population of icy planetesimals beyond Neptune had long been postulated, as remnants of the primordial solar nebula and as a source for short period comets, before the first detection (except for Pluto) was made in 1992. Since then almost 300 Kuiper Belt Objects (KBOs) have been discovered with a range of orbital and physical characteristics. Centaurs, with unstable orbits crossing those of the outer planets, may be a transitionary class between the Kuiper Belt and short period comets. Due to their relatively small size (up to a few hundred km) and large distance, the majority of these objects are at the extreme range of large telescopes so determination of physical properties and composition is a challenging task. KBOs exhibit a wide range of optical colours interpreted as variable resurfacing of dark, red cosmic ray irradiation mantles by subsurface primordial ices through some as yet undetermined mechanism. There are ,~ 105 objects larger than 100kin (the asteroid belt has only 230 objects this size) and ,-~ 109 of the size of a typical cometary nucleus (>5kin) in the inner Kuiper Belt (within 50AU), amounting to about 0.1 Earth mass. This is more than 200 times smaller than the mass required for formation through planetesimal growth and our current understanding of the formation and evolution of the Kuiper Belt is incomplete. I will review what we have learnt so far about the dynamical and physical properties of KBOs, describe plans for future detection and observation campaigns, and the possibilities for in-situ exploration.
DUST IN THE OUTER HELIOSPHERE DUST
AND INTERSTELLAR
E. Grfin Dust measurements in the outer solar system are reviewed. Only the plasma wave instrument on board Voyagers 1 and 2 recorded impacts in the Edgeworth-Kuiper belt (EKB). Pioneers 10 and 11 measured a constant dust flux of 10-micron-sized particles out to 20 AU. Dust detectors on board Ulysses and Galileo uniquely identified micron-sized interstellar grains passing through the planetary system. Impacts of interstellar dust grains onto big EKB objects generate at least about a ton per second of micron-sized secondaries that are distributed by Poynting-Robertson effect and Lorentz force. We conclude that impacts of interstellar particles are also responsible for the loss of dust grains at the inner edge of the EKB. While new dust measurements in the EKB are in an early planning stage, several missions (Cassini and STARDUST) are en route to analyze interstellar dust in much more detail. published in: Minor Bodies in the Outer Solar System, ESO Astrophys. Symp., Springer-Verlag, 99-108, 2000
- 369-
Oral papers and posters
C O M E T S AS A S O U R C E O F H E L I O S P H E R I C
IONS
J. Geiss (1), G. Gloeckler (2) and K. Altwegg (3) (1) ISSI, Bern, (2) U Michigan, (3) U Bern. The ion composition measured by SWICS Ulysses in the very distant tail of C Hyakutake is strikingly similar to the composition obtained by Giotto IMS in the outermost coma of P Halley, demonstrating that ions of cometary origin carry a specific and unique signature. The detectability of unseen comets as a function of encounter distance and comet gas production will be discussed for SWICS Ulysses and for the mass spectrometer on the proposed Interstellar Pathfinder.
CIDA-A COMETARY AND INTERSTELLAR DUST ANALYZER FOR THE STARDUST MISSION J. Kissel, A. Glasmachers, H. von Hoerner and H. Henkel CIDA is a time-of-flight mass spectrometer to analyze ions generated upon impact of cometary or interstellar dust particles. The target area is about. 100 cm 2 (projected), and biased to + - 1 kV, for positive or negative ion extraction, respectively. The mass resolution m dm is > 200 (FWHM) due to its ion reflector, which compensates for a wide range of initial ion energies. The instrument is under software control, providing autonomy to operate during the cometary fly-by (Wild-2) in 2004 without interference from ground. CIDA was launched onboard STARDUST on Feb. 7th, 1999. CIDA mass is 10.6 kg, it uses less than 15 W power.
-370-
Oral papers and posters
ORBITAL EVOLUTION OF DUST IN THE OUTER SPHERE UNDER THE DUST-GAS DRAG FORCE
HELIO-
K.Scherer dat-hex, Torstr.3, Katlenburg-Lindau, Germany. In the outer region of the heliosphere (> 20 AU) the neutral gas density becomes larger than the solar wind plasma density. The neutral hydrogen gas exerts a drag force on the dust particles similar to the plasma Poynting-Roberston effect. However, the monodirectional velocity of the interstellar gas, connected with the inflow direction of the interstellar material, induces an effect which is very much different from that induced by the radial solar wind velocity. This causes an asymmetric force acting on the dust particles, which forces the eccentricity and semimajor axis to increase rapidly. Therefore, the lifetime of dust grains in the Edgeworth-Kuiper Belt is not determined by the electromagnetic or plasma Poynting-Robertson effect but by the drag of the neutral gas leading to lifetimes of the order of half a million years for a 10-#m-sized particle. published in: J. Geophys. Res. 105, 10329-10341, 2001
-371
-
This Page Intentionally Left Blank
General Discussion Scherer to Green: A few years ago, we did some tests with the Pioneer 10 spacecraft and figured out that the masses in the Edgeworth-Kuiper belt are by a factor of hundred too low to be detected with it. With your new significantly higher numbers for the masses one should go back to the data. With such high values one should also consider the possibility of a close encounter of a Kuiper belt object (KBO) with Pioneer 10. Green: Various different methods 'all give reasonably self-consistent masses. The big question is where the mass is in the smaller objects. And, of course, if the mass is fairly evenly distributed, it will be much harder to find from spacecraft data. Mewaldt to Green: The intensity of energetic particles is probably dominated by anomalous cosmic rays (ACRs) in the region of the termination shock. Is it possible that many of the KBOs cannot be seen because they are so dark because they are bombarded with ACRs? Green: Even if the KBOs are dark we should have seen them. So, I don't think it's a reason for not finding the objects further out. But it might be an explanation for a difference in distribution of colors and albedos as a function of heliocentric distance. Jokipii to Griin: You didn't mention the electric field of the charged dust. Is that because you are concentrating on the larger sizes? Griin: I did not mention it explicitly but in our modelling we are taking it into account. Fahr to Griin: You said you start seeing the interstellar dust particles when you go to larger sizes. So you are missing the smaller ones. Could it mean that the dust is also plasma charged outside [the heliopause]? And could one get an estimate of the magnetic field in the local interstellar medium (LISM) from the loss of dust particles? Griin: We are not that far to derive a magnetic field from considering the effect of such filtering, but it is compatible with the results from models for the magnetic field at the bow shock. Gruntman to Kissel: Could you comment on the current status of the capability of reconstructing the original molecules from what you observe? Kissel: This is a very critical question that we have put to ourselves, of course. We currently can only say that we have a trend of seeing nitrogen and oxygen contributions in our spectra, and once we invest more work and can really establish a more solid mass scale, we can identify some of these individual lines that we see. Then, of course, we can make more reliable statements of what the exact nature of these molecules is. Fahr to Geiss and Gloeckler: Did you also get the velocity spectra for these cometary ions? Ge'~s: Yes, we have them. Gloeckler: Note, however, that we are, of course, quite far away from the coma of the nucleus, so the spectra we see are very much adiabatically cooled. Gruntman to Scherer: If I understand correctly, the drag by gas is much more important
-373-
General Discussion
than gravity beyond 27000 AU. If this is correct, then I would expect that any dust grain in the Oort cloud instead of escaping would rather be accommodated in the inflowing interstellar gas flow, and that would mean that all dust that we call interstellar is actually a combination of truly interstellar dust grains and the dust grains produced in the O f t cloud. I think that's a very important result. Scherer: Yes, you have no chance to say whether a particle is coming from the Oort cloud or the interstellar medium, except if you look at the composition of the dust grains. Geiss: There are estimates that if all dust grains would come from the Oort cloud, it would disappear within 105 or 106 years. So I think there is not enough mass in it.
- 374-
Session 8: New K i n e t i c Aspects of Heliospheric Physics
This Page Intentionally Left Blank
The Injection Problem Joe Giacalone University of Arizona, Tucson, AZ, 85721 The physics of the injection problem associated with charged-particle acceleration at collisionless shocks is discussed. The limit of diffusive shock acceleration is considered for three special cases: weak scattering, parallel shock, and perpendicular shock. We find that the diffusive approximation is more easily met at parallel shocks than perpendicular shocks. Yet, we also find that, in general, perpendicula, r shocks may also be effective at accelerating low-energy particles. Implications for the acceleration of pickup ions at the termination shock to anomalous cosmic-ray energies is also discussed. Finally, we present an alternative mechanism for accelerating charged particles by compression regions in the solar wind. We conclude that compression regions, such as corotating interaction regions prior to the formation of the associated forward and reverse shocks, may accelerate pickup ions up to 1-10 MeV. 1. I N T R O D U C T I O N
There is currently no widely-accepted theory that explains which, and how many particles of a thermal population incident on a collisionless shock will be accelerated to high energies. Diffusive shock acceleration, for instance, is a well-known theory which describes the acceleration of particles at shocks; however the key assumption in this theory is that the particle distribution must be isotropic to first order. This is not usually met by particles whose energy is near that of the thermal peak. Consequently, the diffusive theory is not adequate to describe how the particles are injected. The injection problem is important in a number of applications. Anomalous cosmic rays (ACR) are a good example. Observational constraints strongly suggest that ACRs are the result of the acceleration of ionized interstellar atoms at the termination shock of the solar wind. However, the argument can be made that pickup ions which encounter the termination shock without having been previously accelerated cannot be accelerated to high energies by the termination shock. Some mechanisms have been suggested[i-3]; however none of these have been widely accepted. 2. L I M I T O F D I F F U S I V E
SHOCK
ACCELERATION
It is well known that the energy spectrum of charged-particles which can be described by the diffusive transport equation, downstream of a collisionless shock is a power law in the absence of energy sinks (such as adiabatic losses or free-escape boundaries). Since this theory does not address injection, we consider the limits under which the theory is valid. Since the key assumption of the theory is that, the pitch-angle distribution is
-377-
J. Giacalone
quasi-isotropic, we require that the diffusive-streaming anisotropy in the plasma frame of reference to be small. The general form of this condition is given by[4]
T -- 3Ulv[1 +
(KA/KII) 2 sin 2 0B~ + (1 -- tc•
2 sin 2 0B~ cos 2 0B~ 1/2 +
<<1
(1)
where U1 is the flow speed upstream of the shock, v is the plasma-frame particle velocity, ~A is the antisymmetric component of the diffusion tensor, and ~• are the components perpendicular and parallel to the average magnetic field. Shown in Figure 1 is a distribution function showing the thermal peak and a high-energy tail. wi~j is a characteristic speed at which the process is diffusive (i.e. wi~j is the value of v in Equation (1) for which T = 1). Our interpretation of the injection problem is that the physics which describes the suprathermal tail, i.e. its velocity dependence and density is not known. There are a number of processes described in the literature which address this physics, but there is currently no consensus theory. Examples of some important factors contributing to the formation of suprathermal tails are the microphysics of the shock layer[5,1,6] and also statistical mechanisms such as transit-time damping[7].
/'k f (w)
W " Winj Diffusive W < W.,i Non-Diffusive]
//
t\
/
t ~uper-thermal tail
Maxwellian
\
,,
i
wi,q
w
Fig. 1: A distribution function showing a thermal peak and suprathermal tail. The distribution is power-iaw at high energies (w > w~j). T in Equation (1) is related to the injection emciency which can be understood in terms of the speed at which the process becomes diffusive (i.e. the speed at which (1) is satisfied). The intensity for w > wi~j will depend critically on wi~j. Consequently, large values of T imply that the injection is likely to be very inefficient, whereas, smaller values indicate that the injection will be more efficient. It is useful to consider special cases of Equation (1). These are treated separately below. 2.1. T h e L i m i t of W e a k S c a t t e r i n g For weak scattering, we have ~11 >> ~A, ~ l 1 and equation (1) becomes.
T -
3u~ sec OBn
(2)
1This is true for the case of classical scattering where the assumption that the mean-free path is much larger than the particle gyroradius implies that ~11 >> m.L. However, this is not a general result.
-378-
The injectionproblem U1 sec OB~ is the minimum speed which a particle must have in order to stream ahead of the shock. This result is similar to the scatter-free result derived by Decker[8]. The strong dependence on OB~ is evident. Nearly perpendicular shocks, for this reason, have long been thought to be unable to accelerate low-energy charged particles. In fact, observations at the Earth's bow shock are suggestive of a strong OBn dependence of the intensity of backstreaming particles[9]. Additionally, the simulations by Baring et al.[10] also show a strong dependence of the injection efficiency on OBj. The injection process at nearly perpendicular shocks is discussed further below. 2.2. P a r a l l e l S h o c k For a parallel shock we take the limit OB~ --+ 0 in (1) which gives T =
3U1
(3)
1)
This result shows that parallel shocks are efficient accelerators of particles whose speed is on the order of the plasma flow speed. Unless the temperature of the plasma is very low, there are a significant number of particles within the thermal pool which can easily be accelerated by the parallel shock. Self-consistent hybrid simulations confirm this [11]. These simulations also show that the injection mechanism is a coherent process similar to cyclotron acceleration. The low energy particles have gyroradii which are comparable to the thickness of the shock. Hence, the electromagnetic structure of the shock has an important influence on the acceleration of these particles. The higher energy particles are accelerated via first-order Fermi acceleration. Note that (3) is independent of the diffusion coefficient. Hence, parallel shocks are capable of accelerating low-energy particles regardless of their mean-free path. Some scattering is expected for other reasons however; the formation of the parallel shock is often described in terms of the firehose instability [12] which would lead to magnetic fluctuations which will scatter the particles. 2.3. P e r p e n d i c u l a r S h o c k For a perpendicular shock we take the limit 0B~ -+ 90 ~ in (1) which gives T=
3U~
1+
~
(4)
V
In the case of classical scattering, ~l and ~A can be written in terms of the ratio of the scattering mean-free path, All to the particle gyroradius, T" For this case, Equation (1) becomes,
T-
3~[1 1 -JF
(5)
Equation (5) confirms that the stronger the scattering (and hence, the smaller the mean-free path) the more efficient the particle injection. This is in agreement with the Monte-Carlo simulations by Baring et al.[10] On the other hand, there is no reason to believe that classical scattering theory gives a valid description of the cross-field diffusion coefficient. In fact, Giacalone and Jokipii[4]
- 379-
.1. Giacalone showed that for certain forms of the magnetic turbulence spectra, ~• does not agree with the theory of classical-scattering. In the absence of a consensus theory for ~;z, we can consider other cases which seem reasonable. For instance, assuming ec• to be a constant, which may be reasonable in the case of large-scale magnetic fluctuations (those with a typical scale size which far exceeds the particle gyroradii), we find T-
)211,2
3U1
(6)
v
where we use etA -- (1/3)vrg (as we did for Equation (5)), which is a widely accepted representation.
Tnon~K_~_4_l............... :":
d,ffuxive
0.1 )~11/rg
Fig. 2: T determined from Equations (5) (soiid 111)e)&nd (6) (dashed line)as a function of the ratio of the particie mean-free path to gyroradius, v - 10U1 was used t'or both cases, and ~• 0.02 was used in Equation (6). Figure 2 shows the value of T as a function of the ratio of the mean-Dee path to particle gyrore~dius for the two cases described above (Equations (5) and (6)). Note that contrary to the case of classical scattering (solid line.) in which the injection efficiency decreases with increasing scattering mean-free path, the case of a constant ~:• , shows that weaker scattering (longer mean-free path) is required in order to have more efficient injection. This may apply to the acceleration of interstellar pickup ions at the termination shock of the solar wind. 3. I M P L I C A T I O N S
FOR HELIOSPHERIC
PHYSICS
3.1. P i c k u p I o n s a n d A n o m a l o u s C o s m i c R a y s Anomalous cosmic rays (ACR) are an important component of cosmic rays which are accelerated locally, presumably at the termination shock of the solar wind. ACR start out as freshly-ionized interstellar neutral atoms, as originally suggested by Fisk et ed.[13] The acceleration to the highest energies is currently thought to occur at the termination shock of the solar wind[14,15]. The process of acceleration from the ~ 10 keV energies of pickup ions to the >_ 1 GeV energies of the highest-energy ACR observed involves a very, large parameter range and the whole process cannot yet be modeled by one technique.
- 380-
The injection problem
Giacalone et al.[16] proposed a model for anomalous cosmic rays in which interstellar pickup ions are initially accelerated in the inner heliosphere. They are then transported into the outer heliosphere where they are accelerated further at the termination shock. Interplanetary shocks, or turbulent regions within corotating interaction regions can more readily accelerate pickup ions [17,7]. If Equation 1 is satisfied for these preaccelerated pickup ions upon encountering the termination shock, then the results of Giacalone et a1.[16] show that there are enough low-energy ions observed at 40 AU to be further accelerated at the termination shock and account for the observed levels of anomalous cosmic rays. They used a widely accepted model for the global transport and acceleration of cosmic rays[15]. They assumed that the preaccelerated pickup ions could be adequately described by the diffusive transport equation[18]. If in fact the Voyager 2 observations of CIR-related particles[19] are accelerated interstellar pickup ions, as suggested by Gloeckler et al.[17], then the acceleration of pickup ions in the inner heliosphere is an important stage in the formation of e--~malous cosmic rays. 10~
I
........
I
L
,
........
,
........
,,,
,
Voyager2 LECP (January - July, 1993)
' ~ ',,
~40AU
102 "...\
g
Voyager1 CRS-HET (DOY 157- 262, 1994)
~
57AU -\
\
,0~
~ ,..9 \o~, . . /
..o,. \
'\ /
Model calculation
1 0 -4
........
I
,
, ......
I
........
I
,
,,
Energy (MeV)
Fig. 3: Simulated anomalous cosmic-ray proton spectra at 40 A U and 57 A U using a source located in the inner heliosphere (10 A U). The star symbols are Voyager 2 observations of iow-energy ions when Voyager 2 was at 40 A U and the square symbols are Voyager i observations of anoma]ous cosmic-ray protons when Voyager 1 was at 57 A U. This figure is adapted from Giacalone et al.[16J
Figure 3 is adapted from Giacalone et a1.[16] It shows the simulation results (curves) normalized using Voyager 2 data (stars) and compared to anomalous cosmic-ray data (squares). The simulations used reasonable solar wind and termination shock parameters and diffusion coefficients which were based on consensus values. The source was located at 10 AU. Note that the low-energy portion of the spectrum decreases with increasing radial distance, while the peak at anomalous cosmic-ray energies increases with radial
-381
-
.i.. Giacalone
distance. These are consistent with observations reported by Hamilton et al.[20] 3.2.
Particle
acceleration
at compression
regions
The spatial distribution of accelerated particles at corotating interaction regions (CIRs) provides constraints on possible acceleration mechanisms. CIRs are regions of compressed solar wind which corotate with the sun. At large heliocentric distances (>1-2 AU), there are forward and reverse shocks at the leading and trailing edges of the compression. The association of energetic nuclei with (CIRs) has been recognized for more than 20 years [21,22]. Currently, it is thought that the particles are accelerated to several MeV energies at these forward and reverse shock waves. However, Gloeckler et al.[17] and Schwadron et al.[7] have also suggested that the acceleration of the low-energy pickup ions from ~ 1 keV to ~ 50 keV occurs via the mechanism of transit-time damping of magnetic fluctuations within the CIR, rather than at the shocks. It is important to understand the evolution of a CIR as the fast solar wind overtakes the slow solar wind and how charged particles can be affected by this formation. If particles diffuse through the velocity gradient associated with the CIR formation, they will be efficiently accelerated due to the fact that this gradient is very large (about a factor of 2) prior to the formation of the shocks. Consequently, considerable acceleration can occur before the shocks form (< 1-2 AU) and this can lead to a peak in the intensity within the CIR. This might explain recent observations by Mason[23]. In order for acceleration to occur at a compression region, we require that the diffusive length scale Ld be larger than the compression region width L. This gives, (7)
L < Ld = ~ / v
where ~ is the diffusion coefficient along the direction of the propagation direction of the compression region (assumed to be radial) and v is the particle speed. Consider the acceleration of interstellar pickup ions by a solar wind compression region which corotates with the sun. It is known that there will be a forward and reverse shock at distances beyond ~ 2 AU. Near 1 AU, however, there is a smooth transition froln the fast to slow solar wind which occurs over a scale which is much larger than the gyroradius of the pickup ions. We further assume that the gradient associated with the plasma compression is aligned with the radial direction. Using these considerations, it is straightforward to show that (7) leads to L < All, where All is the pickup-ion parallel mean-free path. If, however, All is too large, the particles will not scatter downstream (the same side as the sun) of the compression region and will instead be mirrored in the strong magnetic field near the sun. This will likely not lead to any significant acceleration. Consequently, we require both that the mean-free path is larger than the width of the compression, but smaller than about 1 AU. Pickup ions probably satisfy these criteria and, therefore, we expect them to be efficiently accelerated at solar wind compression regions. The mechanism is similar to diffusive shock acceleration in that the particles are accelerated as they scatter off of converging scattering centers. The acceleration can occur for any particles provided that their mean-free path is large enough to sample the velocity gradient and small enough to be scattered downstream of the compression region. The highest energy attainable, therefore, can be estimated by setting the mean-free path to the heliocentric distance of the compression region. A
- 382 -
The injectionproblem compilation of observations of particle mean-free paths by Palmer[24] has shown that even 1-10 MeV particles can have mean-free paths which are smaller than, or are comparable to 1 AU. Consequently, we conclude that pickup ions can be accelerated by corotating compression regions (at distances smaller than where the forward and reverse shocks form) up to energies of about 1-10 MeV. 800 km/s
Plasma
Speed
~
Grad"-ilentLength,L Pickup-iongyroradius
400 km/s Distance
~>
Fig. 4: Cartoon sketch of the expected solar wind speed for a compression region. 4. C O N C L U S I O N S
We have discussed the injection problem arising from the inability of analytic theory to explain the entire distribution function, from energies typical of the thermal peak to high energies, of charged particles which have encounted a collisionless shock. Although the injection problem is evident in the theories of shock acceleration, shocks are observed to accelerate particles. So, there really is not an injection problem in nature. Diffusive shock acceleration, which assumes that the particle distribution function is isotropic to first order, naturally explains high-energy tails (e.g. power law, exponential etc.) in the energy distribution. We have considered the validity of this theory for three special cases: weak scattering, parallel shocks, and perpendicular shocks. On the one hand, we have shown that the diffusive approximation is more easily met at parallel shocks than perpendicular shocks. On the other hand, we have also shown that, in general, perpendicular shocks can be effective at accelerating low-energy particles. The propagation of the charged particles along large-scale meandering of magnetic field lines can enhance the diffusion normal to the shock front which enables particles to remain near the shock and be accelerated. An important consequence of the injection problem is understanding the anomalous component of cosmic rays. Because of the charge state of this component (predominantly singly charged) and radial gradient, among other things, they are certainly accelerated interstellar pickup ions at the termination shock of the solar wind. However, there is no widely accepted analytic theory explaining how low-energy pickup ions can be accelerated by the nearly perpendicular termination shock. One possibility is that the particles are accelerated at propagating shocks much nearer to the sun (1-10 AU) where the geometry and initial acceleration is more favorable. They are then transported to the termination shock with high enough energy to be in the diffusive regime. Their continued acceleration
-383-
J.. Giacalone
and transport is well understood. We have discussed an alternate view of particle acceleration at compression regions, rather than shocks. This process is very much like shock acceleration in that the energy gain arises from scattering between converging scattering centers. However, in order for low-energy particles to be accelerated, they must have a long enough mean-free path to sample the gradual velocity gradient associated with the compression. Spacecraft observations indicate that pickup ions, and other low-energy particles, have much longer mean-free paths than one would expect from theory. Thus, compression regions may accelerate pickup ions to high energies (1-10 MeV). 5.
ACKNOWLEDGEMENTS
This work was supported by NASA under grants NAGS-2251 and NAG5-6620. REFERENCES
,
,
,
6. 7. 8. ,
10. 11.
12. 13.
14. 15.
Zank, G. P., H. L. Pauls, I. H. Cairns, and G. M. Webb, J. Geophys. Res. 101 (1996) 457. Giacalone, J., J. R. Jokipii, and J. Kdta, Ion injection and acceleration at quasiperpendicular shocks, J. Geophys. Res. 99 (1994) 19351. Kucharek, H., and M. Scholer, Injection and acceleration of interstellar pickup ions at the heliospheric termination shock J. Geophys. Res. 100 (1995) 1745. Giacalone J., and J. R. Jokipii, The transport of cosmic rays across a turbulent magnetic field, Astrophys. J. 520 (1999) 204. Giacalone, J., T. P. Armstrong, and R. B. Decker, J. Geophys. Res. 96 (1991) 3621. Lee, M. A., V. D. Shapiro, and R. Z. Sagdeev, J. Geophys. Res. 101 (1996) 4777. Schwadron, N. A., L. A. Fisk, and G. Gloeckler, Geophys. Res. Lett. 21 (1996) 2871. Decker, R. B., Computer modeling of test particle acceleration at oblique shocks, Space Sci. Rev. 48 (1988) 19,5. Burgess, D., Simulations of backstreaming ion beams formed at oblique shocks by direct reflection, Ann. Geophys. 5 (1987) 133. Baring, M. G., D. C. Ellison, and F. C. Jones, Monte-Carlo simulations of particleacceleration at oblique shocks, Astrophys. J. Suppl. Ser. 90 (1994) 547. Giacalone, J., D. Burgess, S. J. Schwartz, and D.C. Ellison, Hybrid simulations of protons strongly accelerated by a parallel collisionless shock, Geophys. Res. Lett 19 (1992) 433. Parker, E. N., A quasi-linear model of plasma shock structure in a longitudinal magnetic field, J. Nucl. Energy C2 (1961) 146. Fisk, L. A., B. Koslovsky, and R. Ramaty, An interpretation of the observed oxygen and nitrogen enhancements in low-energy cosmic rays, Astrophys. J. Lett. 190 (1974) L35. Pesses, M. E., J. R. Jokipii, and D. Eichler, Cosmic ray drift, shock wave acceleration, and the anomalous component of cosmic rays, Astrophys. J. Lett. 246 (1981) L85. Jokipii, J. R., Particle acceleration at a. termination shock 1. Application to the solar wind and the anomalous component, J. Geophys. Res. 91 (1986) 2929.
- 384-
The injection problem 16. Giacalone, J., J. R. Jokipii, R. B. Decker, S. M. Krimigis, M. Scholer, and H. Kucharek, Pre-acceleration of anomalous cosmic rays in the inner heliosphere, Astrophys. J., 486 (1997) 471. 17. Gloeckler, G., J. Geiss, E. C. Roelof, L. A. Fisk; F. M. Ipavich, K. W. Ogilvie, L. J. Lanzerotti, R. yon SteJger, and B. Wilken, Acceleration of interstellar pickup ions in the disturbed solar wind observed on Ulysses, J. Geophys. Res. 99 (1994) 17,637. 18. Parker, E. N., The passage of energetic charged particles through interplanetary space, Planet. Space Sci. 13 (1965) 9. 19. Decker, R. B., S. M. Krimigis, S. M., and M. Kane, Proc. 24th Int. Cosmic Ray
Co f. (Rom )4 ( 995)42 . 20. Hamilton, D. C., M. E. Hill, R. B. Decker, and S. M. Krimigis, Proc. 25th Int. Cosmic Ray Conf. (Durban)2 (1997)261. 21. Barnes, O. W., and J. A. Simpson, Evidence for interplanetary acceleration of nucleons in corotating interaction regions, Astrophys. J. 210 (1976) L91. 22. McDonald, F. B., B. J. Teegarden, J. H. Trainor, and T. T. yon Rosenvinge, The interplanetary acceleration of energetic nucleons, Astrophys. J. 203 (1975) L149. 23. Mason, G. M., Composition and energy spectra of ions accelerated in corotating interaction regions, in Acceleration and Transport of Energetic Particles Observed in the Heliosphere , edited by R. A. Mewaldt, et al., American Institute of Physics, New York, (2000). 24. Palmer, I. D., Transport coefficients of low-energy cosmic rays in interplanetary space, Rev. Geophys. 20 (1982) 335.
-385-
This Page Intentionally Left Blank
A hydrokinetic description of solar wind electrons using hemispheric distribution functions I.V.Chashei ~ , H.J.Fahr and G. Lay b ~P.N.Lebedev Physics Institute, Russian Academy of Sciences, Leninskii Prospect 53, 117924 Moscow (Russia) bInstitut fiir Astrophysik und Extraterrestrische Forschung der Universit/it Bonn, Auf dem Hiigel 71, D-53121 Bonn (Germany) Above the coronal exobase no solar wind electrons with superescape velocities should appear in the sunward hemisphere of velocity space. For a better solar wind understanding it is nevertheless relevant to describe how this hemisphere is repopulated with electrons. Thus we study the influence of superescape electron depletions in terms of modified velocity moments of bi-hemispheric Maxwellians. We show that then all higher moments of the distribution function can be generated simply based on knowledge of the three lowest moments. Using solar wind data on electron density, drift, and temperature, we so derive an expression for the electron heat flow which perfectly fits the ULYSSES heatflow measurements both by its absolute magnitude and by its radial gradient. To justify this bi-hemispheric Maxwellian approach of the distribution function we claim that by quasilinear interaction of electrons with whistler wave turbulences a fast pitchangle diffusion is operating, however, leading to a resonance gap at # = 0, i.e. between the two hemispheres. 1. A n e x o s p h e r i c view to t h e c o r o n a already early exospheric theories of the solar wind by Aamont and Case (1962)[1], Brandt and Cassinelli (1966)[4] revealed that above the collision-dominated solar corona Maxwellian distributions cannot be expected neither for electrons nor ions. In these theories the lower corona is the only particle source, and hence in the collisionless regime above the corona no particles with super-escape velocities should populate the sunward hemisphere of the velocity space. Observed features of electron distributions are not explainable by collisionless exospheric concepts (see e.g. Olbert, 1983112]) since the validity of magnetodynamic invariants would generate much too anisotropic distribution functions. Coulomb collisions cannot impede the electron distributions from degeneration into highly anisotropic functions. Only wave-particle interactions by electron Whistler waves (see Dum et al., 198017], Gary et al., 1994111]), by fast-mode MHD waves (Summers and Ma, 2000116]) or perhaps excited plasma instabilities of the fire-hose type (see e.g. Fahr and Shizgal, 198318]) may help as a remedy at larger distances. In our view one should favor electron-wave interaction mechanisms for the reshaping of the anisotropizing
-
387-
L V. Chasheg H.J. Fahr and G. Lay electron distribution function (see Chashei and Fahr, 2000[5]) based on quasilinear interactions of the electrons with preexisting whistler wave turbulences. Connected with such turbulences specific Fokker Planck diffusion coefficients can be evaluated which describe wave-induced electron pitch-angle diffusion as the process operating with the highest rate(see e.g. Denskat et al.,198316]). This process is described by a pitch-angle diffusion coefficient D..(v,#) as derived by Achatz et a1.(1993)[2]. It turns out that both for negative and positive values of # - cos t9 pitch-angle diffusion operates quite efficient and tends to rapidly isotropize the distribution function in the two hemispheres p <_ 0 and p >_ 0, respectively, whereas due to a resonance gap in the cyclotron interaction of electrons with whistler waves (e.g. Schlickeiser et al., 1991114]) around pitch-angles with p ~_ 0 the pitch-angle diffusion between these two hemispheres is strongly impeded. 2. E l e c t r o n h y d r o k i n e t i c s b a s e d on b i - h e m i s p h e r i c M a x w e l l i a n s We assume that by wave-induced relaxation operating above the exobase the genuine Liouvillean velocity distribution is converted into one describing a macroscopic drift and the appearance of trapped particles on the basis of quasi-Maxwellians. We assume that the local electron distribution function f~(~, g) can be approximated by a Maxwellian truncated at the escape velocity in the sunward hemisphere (# _< 0). Then with a density normalization n(r) and a local thermal electron velocity spread C(r) the distribution is given by:
L(-"~,'-~)
d3v -- /'~ (r) [71-C2 (/,)]-3/2 [H (Vma x - v) -11-H (cos 0) H (v - Vmax (r))] x
C 2 (r) v 2dv sin OdOdr
x exp
(1)
v, ~9, r are polar coordinates with the magnetic field as polar axis. ~9 then represents the electron pitch-angle with # - cos 0. The step-functions H(X) take care of the appropriate local truncation suppressing electrons in the sunward hemisphere with superescape velocities v >__Vmax, with' Vmax2- 2cA(I) (r)/m~, A(I) being the potential step to infinity. Density, bulk velocity, temperature etc. have to be obtained as velocity moments of the function given by Equ.(1). As shown elsewhere (Chashei and Fahr, 2000[5]) the following relation between n(r) and the actual electron density n~ (r) is obtained" n~ (r) - n (r) ( 4 / v ~ ) [$2 (0) - 0.5S2 (~X)]
(2)
Here the function S2(x) is defined, for j - 2, by the following integral function:
SN ( z ) - f ~ cJ exp ( - c 2) dc,
(3)
The quantity A - A(r) in Equ.(2) has the following definition" 1
- [eAff ( r ) / m ~ C (r)2] ~ .
(4)
The electron bulk velocity U(r) is obtained by" -
( )cos { ss
(s)
-
388-
A hydrokinetic description of solar wind electrons using...
with S3(x) defined by Equ.(3) for j = 3. ~ is the angle between the radial direction and the magnetic field. The electron pressure analogously is found by:
Re(~)-[~ m~ (~)n (r)
[$4 (0) - 0.5S4 (A)] +
C (r)
[s2 (0)- o.5s2 (A)]]
(6)
with S4(x) calculated with Equ.(3) for j = 4. Of greatest interest here is the electron heat conduction flow q~ which on the basis of Equ. (1) is attained in the following form: 1 q~(r) - -2n~m~ c ~ ( ~ ) 9 (~)
(7)
with
(r) -
ss(a) $3 (A)[$4 (0) - 0.5s4 (A)] s3 ~ (A) ] 4 [$2 (0) - 0.5S2 (A)] + ~2 [s2 (0) - 0.5s2 (A)] ~ + 64 [$2 (0) - 0.5s2 (A)] 3 (8)
One is able to represent the heat conduction flow q~ as a functional of the lowest three velocity moments of this distribution, namely n~(r), P~(r) and U(r), and thus to reach a closed hydrodynamic system of differential equations. Here we want to check the validity of these newly derived relationships. To evaluate expression (7) we need an expression for A connected with the consistent electric polarization potential which can be obtained from the equation of electron motion given in the following form (Fahr et al., 197719], Chashei and Fahr, 2000[5]): 0 - en~dap/dz - dP~l I d z - (llB)-d~z(Pe• - P~II),
where z is the space coordinate parallel to the field ~ , and where P~• and P~ are the electron pressure tensor elements perpendicular and parallel to ~ which on the basis of Equ. (1) are given by:
(lO)
/:'ell = (1/3)Pe and: P~_t_= (2/3)Pr The electric potential then is obtained from the following integral: ~xe(~) - ( K / 3 ~ )
[[T,(~) -
which can be solved adopting the electron temperature profile obtained by Scime et al. (1994)[15] given as a power law with respect to r with the power index a - 0.85. With Equ. (11) we then find: A(r) -
4 + a [1 - Tr (too)/Tr (r)] 3a
]1/2
.
(12)
Assuming that at ro~ - 5 A U the asymptotic level of the electric potential is already reached, we then can evaluate qe(r) given by Equ.(7) using C2(r) ~- 2 K T e ( r ) / m e . With
-
389-
I. K Chashei, H.J. Fahr and G. Lay
~
0.5
o
-0.5
-2
I
[
Figure 1. Shown is the electron heat flow qe(r) in units of [ # W / m 2] as function of the solar distance r for various values of r~(i.e 5, 6, 7AU).
n~(r - 1AU) - 8cm -3 we then find the absolute value of the heat conduction flow at r - 1AU to be given by" (13)
q e r ( r - 1AU) - 7 . 7 6 p W / m 2,
which is surprisingly close to the value found by Scime et al. (1004)[15] (i.e. 8 . 8 # W / m 2) and also agrees very well with values of between 5.0 to 8 . 0 # W / m 2 given by Feldman et al. (1975)[10] and Pilipp et al. (1990)[13]. In addition here we are interested in the study of the radial gradient of the electron heat flow which is also measured by Scime et al. (1994)[15] onboard ULYSSES. On its in-ecliptic itinerary to Jupiter ULYSSES was predominantly exposed to the low speed solar wind (see Bame et al., 199313]) with an average speed of U = Ue = 4 0 0 k m / s and an average density of ne(r = 1AU) = 8cm -3. Evaluating Eq.(37) for these above conditions and using Parker's spiral field formula we obtain the function qer(r) which is displayed as function of the solar distance r in Fig. 1. As one can see from the linear curve appearing in the double-logarithmic plot of this figure the heat flow q,~(r) behaves exactly like a power law in the radial coordinate r given by qer(r) '~ q e r ( r - - 1AU)(r/rE)-7~,where the exponent % evaluates to % = 3.08. This again is a very nice result since it nearly exactly fits the result derived from ULYSSES solar wind electron observations (see Scime et al., 1994115]) yielding q~(r) ~- 8.8(r/rE)-3[pW/m]. Hereby, as evident from the additional curves given in Fig. 1 it can be recognized that a variation of the value r ~ plays a very inferior role for the result. Thus it seems as if with our parametrized solar wind electron distribution function we do solve two outstanding problems in the thermodynamic behavior of solar wind electrons at larger distances; we
- 390-
A hydrokinetic description of solar wind electrons using... obtain both the correct magnitude and the correct radial gradient of the electron heat conduction flow. 3. A c k n o w l e d g e m e n t The authors gratefully acknowledge financial support of this work by a collaborative grant of NATO. REFERENCES
-391 -
This Page Intentionally Left Blank
Oral papers and posters
THE SOLAR WIND ELECTRON
VELOCITY DISTRIBUTION
O. L i e - S v e n d s e n Norwegian Defence Research Establishment, Div. for electronics, P.O. Box 25, NO-2027 Kjeller, Norway. 0 y s t e i n . L i e - S v e n d s e n ~ f f i . n o / F a x : [+47] 63807212 The solar wind electrons carry information about conditions in the solar corona and about plasma processes taking place in the solar wind. They therefore serve as a valuable "probe" of regions in which we cannot make in situ measurements. But before we can extract such information from measured velocity distributions we need to understand what effect basic processes, such as Coulomb collisions and large-scale electric and magnetic fields, will have on the distributions. Based on numerical solutions to the kinetic transport equation we discuss how these processes influence the electron velocity distribution and can actually explain the basic features of observed distributions. Due to their small mass and therefore high mobility, electrons may also, through heat conduction, supply a large fraction of the energy needed to support the wind. Kinetic calculations show that, as far as energy transport is concerned, the electrons behave almost like a collision-dominated gas. To a good approximation the energy transport may therefore be described using classical transport theory.
GENERAL ASPECTS OF BOLTZMANN'S H-THEOREM THE PHYSICS OF THE OUTER HELIOSPHERE
AND
R.A. Treumann The outer Heliosphere is an ideal laboratory to test the behaviour of an ever more with distance diluting and cooling plasma. In such conditions one can study over long distances the way a collisionless plasma behaves. Neglecting the changing composition of the plasma which is due to the mixing with neutral gas and cosmic ray intrusions from the extra-heliospheric environment
- 393 -
Oral papers and posters
HEATING AND ACCELERATION OF IONS BY CYCLOTRON AND L A N D A U R E S O N A N C E W I T H P L A S M A WAVES E. M a r s c h (1) and C.-Y. Tu (2) (1) Max-Planck-Institut fr Aeronomie, Max-Planck-Str. 2, D-37191 Katlenburg-Lindau, Germany, (2) Department of Geophysics, Peking University, Beijing, 100871, China. The general heating and acceleration rates for ions being in cyclotron or Landau resonance with plasma waves are derived within quasilinear theory, which assumes interactions between the ions and a broad-band spectrum of waves with arbitrary propagation directions. The special kinetic theory of resonant interactions of the ions with field-aligned ion-cyclotron waves is then used to study the coronal and solar wind ion heating and acceleration, and to explain some of the observations made recently in the solar corona by the SUMER and UVCS instruments on SOHO. We consider a plasma composed of three species: protons, He+2 and 0+5. The anisotropic multi- uid equations can adequately describe the evolution of the ion drift velocities and temperatures with altitude in the corona. The evolution equation used for the spectrum includes both the large-scale MHD effects and small-scale resonance damping. The evolutions of the uid parameters and the wave spectrum are calculated self-consistently. It is found that the ion plasma betas and drift velocities can be modelled in accord with the plasma parameters found in coronas holes and the associated fast solar wind.
- 394-
General Discussion Baranov to Treumann: Is there a connection between your equation and the Klimontovich equation for the one-particle-distribution function if the Landau collision integral is equal to zero? Treumann: Of course, there is a connection because the Klimontovich representation is only a different kind of representation. Tile Landau collision term has nothing to do with my collision term. The Landau collision term, however, is the perturbation expansion of the Boltzmann collision term by assuming that the interacting potential is a Coulomb potential. Only if the statistics I derived here is going to be Coulomb statistics, then it's exactly the Landau collision term which follows. Sreenivasan to Treumann: Your model equations are valid for steady states. But you are applying them to situations which are far from equlibrium. Treumann: It is indeed far from equilibrium. However, the distribution functions here describe the thermodyanmics or statistical mechanics of a situation where you have no equilibrium but stable states. MoraM to Giacalone: Your result that the acceleration is proportional to the size of the simulation b o x - is that formaly equivalent to the interpretation that the perpendicular diffusion seems to be small in solar particle events but large in cosmic ray events or galactic cosmic rays in general? Giacalone: My interpretation is simply that I'm having longer wavelength waves and, since these particles have gyroradii which are smaller than that, their diffusion is going to be enhanced because of the large-scale meanderings of the field. So, perpendicular diffusion is actually increased as you increase the size of the box. Cairns to Lie-Svendson: How did you initialize the electrons? Did you initialize them at, say, one solar radius, at about the photosphere? Lie-Svendson: The lower boundary is roughly at a million degrees. We have to put it so low in order to be certain that there are enough collisions, so that we can justify having Maxwellians. Barnes to Lie-Svendson: Could you just quickly compare your results with those by Scudder and Olbert? Lie-Svendson: We do disagree with them. What Scudder is doing now is that he is not starting with a Maxwellian but with something like a ~-distribution. Also, we do get an electric field which is the reason why we don't have any velocity filtration. Jokipii to Lie-Svendson: Would you say that your results would not be in agreement with theirs if you would start with co-distributions? Lie-Svendson: No. The only thing I'm saying is that the observations do not show a need for x-distributions, they are rather consistent with Maxwellians. Marsch to Fahr: How do you reconcile your Coulomb collision-free case with that by LieSvendson where Coulomb collisions are important?
-395-
General Discussion
Fahr: I started at 1 A U - there Coulomb collisions are not very important and the distribution functions have already reacted to the interaction with the turbulences. And what the turbulences do, in my view, is that they cause pitch-angle scattering, and I think this is is reasonably modelled with these truncated Maxwellians. And taking their moments into account we also come to reasonable agreement with observations. Lee to Marsch: Do you accurately take into account in your fluid modelling the rcsoaaanc,e cut-off due to the fact that you can no longer resonate with the ion-cyclotron waves? As you increase the velocity outward, you disconnect the ions from resonance with ioncyclotron waves. Marsch: That's all taken cart of. It's a thermal finite-beta plasma dispersion relation that goes into that and the resonance kinematics is fully included. The model distributions are, however, only flexible in terms of responding as bi-Maxwellians and in terms of drifts. That might be a serious limitation, of course.
- 396-
Session 9: M o d e r n Heliospheric Spacecraft and Missions
This Page Intentionally Left Blank
In-space nuclear power as an enabling technology for exploration of the outer heliopause Robert Sackhe~m, Melissa Van Dyke, a Mike Houts, a David Poston, b Ron Llpmsk~," 9 .c Jay Polk, d and Robert Frisbee d 9
a
aNASA/MSFC, Huntsville, AL bLos Alamos National Laboratory, Los Alamos, NM %andia National Laboratories, Albuquerque, NM dJet Propulsion Laboratory, California Institute of Technology, Pasadena, CA Deep space missions, both for scientific and human exploration and development, are as weight limited today as they were 35 yr ago. Right behind the weight constraints is the nearly equally important mission limitation of cost. Launch vehicles, upper stages, and in-space propulsion systems also cost about the same today with the same efficiency as they have had for many years (excluding impact of inflation). These dual mission constraints combine to force either very expensive, megasystem missions or very light weight, but high risk/low margin planetary spacecraft designs, such as the recent unsuccessful attempts for an extremely low cost mission to Mars during the 1998-99 opportunity (Mars Climate Orbiter and Mars Polar Lander). When one considers spacecraft missions to the heliopause, the enormous weight and cost constraints will impose even more daunting concerns for mission cost, risk, and the ability to establish adequate mission margins for success. This paper will address the benefits of using a safe in-space nuclear reactor as the basis for providing both sufficient electric energy and high performance space propulsion that will greatly reduce mission risk and significantly increase weight initial mass in low-Earth orbit (IMLEO) and cost margins. Weight and cost margins are increased by enabling much higher payload fractions and redundant design features for a given launch vehicle (higher payload fraction of IMLEO). This paper will also discuss and summarize the recent advances in nuclear reactor technology and safety of modem reactor designs and operating practices and experience, as well as advances in reactor control coupled with high power generation and high performance nuclear thermal and electric propulsion technologies. It will be shown that these nuclear propulsion and power technologies are major enabling capabilities for higher reliability, higher margin and lower cost deep space missions designed to reliably reach the heliopause for scientific exploration. 1. I N T R O D U C T I O N
1.1. Nuclear Power is Enabling for Deep Space Missions Tremendous amounts of energy are required for exploring deep space (see Figure 1). Traditionally, a significant fraction of this energy has been provided by gravity assists, as utilized by Pioneer, Voyager, Galileo, Cassini, and other missions. However, reliance on gravity assists can result in long transit times (e.g., Galileo, Cassini) and is not practical for certain missions. Slowing down at a destination can require tremendous amounts of energy, which must
- 399-
1~ Sackheim et al.
Figure 1. Maximum distance from Sun as a function of Total AV, 15-yr mission. be supplied propulsively if aerobraking is not practical. Transmitting real-time video or other high volume data streams from the destination also requires significant amounts of energy. A safe, reliable and affordable high-density energy source is required for enabling advanced space missions. Fissionable material stored in a volume no larger than a 12-ounce German beer can (uranium metal) yields as much energy as that contained in 50 Space Shuttle external tanks. The use of fission can thus in theory completely remove energy-related constraints for a high performance, deep space mission enabling propulsion system. Fission systems can be designed to be safe and reliable, to have long life, and to have mission-enabling performance (Isp >800 s for high thrust nuclear thermal systems, and Isp >>3000 s for continuous impulse, "in-space systems"). 1'2 Fission systems scale extremely well to high power levels. One measure of this scaling is specific mass, which is the ratio of system mass to power produced. For continuous impulse applications, fission systems are predicted to have specific masses ranging from 20 kg/kWe at a power level of 100 kWe to less than 1 kg/kWe at power levels exceeding 50 MWe. For applications requiting short (few hr), high thrust (>10 kN), moderate specific impulse (>800 s) bums, specific masses below 0.01 kg/kW should be attainable. Unlike proposed high energy density systems that have not yet proven feasible (e.g., fusion), generating a self-sustaining fission reaction is quite straightforward. All that is required is for the fight materials to be placed in the fight geometry--no extreme operating conditions or reaction drivers (e.g., lasers, magnetic fields) are needed. As an example of enabling performance, consider the velocities required by interstellar precursor missions such as a Kuiper object rendezvous (Figure 2). If these types of missions are to be completed in an acceptable time (e.g., <20 yr) velocity increments greater than 100 km/s are typically required. Such a mission requires both a high specific impulse and an adequately high acceleration (or thrust-to-mass ratio). The best chemical propellant system available today is liquid oxygen/liquid hydrogen which has a maximum delivered specific impulse of 460-480s To achieve a total AV of 100 km/s with a payload, power system, engines, and other inert masses ("non-fuel mass") of 5000 kg would require 21 billion tons of fuel.
In-space nuclear power as an enabling technology for exploration...
Figure 2. Mission profile for Kuiper object rendezvous mission, 23 kg/kWe propellant energy source. Clearly, today's best chemical propulsion system could not be a candidate, as conventional non-combined cycle chemical systems have reached their performance limit. Both NASA and industry (for commercial space applications) have recently develoP~3d very reliable electric propulsion accelerators with a demonstrated specific impulse of 3300 s. An even higher Isp can be achieved either by using a larger voltage on the grids or by using a lighter gas. NASA is also developing a variable specific impulse magnetoplasma rocket (VASIMR) electric thruster which is projected to achieve up to 40,000 s Isp. 4 To reach a total AV of 100 km/s with a specific impulse of 10,000 s and 5000 kg of payload and structure requires only 8.9 tons of propellant. While the electric thruster could be driven by solar cells, it is evident that once the spacecraft goes beyond Mars, the capability of the system becomes limited due to the 1/r2 decrease in solar radiation intensity (Figure 3).
Figure 3. Solar intensity versus distance from Sun (natural fusion energy source).
-401 -
R. Sackheim et al.
Fission systems have much higher performance potential. Fissioning 1 kg of uranium provides as much energy as the kinetic energy in 5 million ~ounds of propellant exhausted with an Isp of 850 s. The specific energy of fissile fuel is 8x 10 J/kg, over six orders of magnitude greater than the best chemical fuels and greater than the charged particle specific energy in D-T (6.8x1013) or D-D fusion (7.8x1013). Fissile fuel is virtually non-radioactive (a factor of 100 million less than tritium) and is readily available (helium-3 and tritium for potential fusion machines are extremely scarce). Terrestrial fission systems have demonstrated very high fissile fuel burnup fractions, showing that the practical fuel energy density in an operating system is close to the theoretical energy density. Space fission systems have a high technology readiness level (TRL) compared to other high energy density options (Figure 4) and should be considered as the most viable nuclear candidate for near-term deep space missions. The development and use of fission propulsion systems is a critical first step towards developing other highperformance propulsion systems based on fusion, matter annihilation, or other processes. Predicted performance attributes of various propulsion systems are shown in Figure 5 and Figure 6 shows the benefits of nuclear-electric propulsion for future crewed Mars missions. 1.2. Historical Lessons The discovery of fission was reported in February 1939. Rapid progress was made in understanding the phenomena, and on December 2, 1942 the world's first self-sustaining fission chain was realized at the University of Chicago. Fission reactors have since been used extensively by the Navy (powering submarines and surface ships) and the commercial power industry (20% of the U.S.' electricity is provided by fission reactors, 70% of France's electricity is produced by sfission reactors). Operating experience in the U.S. alone totals thousands of reactor-years.
Figure 4. Specific energy and TRL of various propellant energy source options.
- 402 -
In-space nuclear power as an enabling technology for exploration...
Figure 5. Performance potential of various propulsion system options.
Figure 6. Potential applicability of advanced nuclear-electric propulsion systems to crewed Mars mission
-403 -
R. Sackheim et al.
2. BACKGROUND 2.1. Technology Maturity The potential for using space fission systems to open the solar system to extensive exploration, development, and settlement has been recognized for decades. Space fission propulsion technology has been extensively developed over the past 50 years; however, despite numerous U.S. programs aimed at developing and utilizing fission systems, the only U.S. flight of a fission system occurred over three decades ago on April 3, 1965 (SNAP 10-A). The SNAP 10-A, developed over a four-year period, operated successfully for 43 days. It was a UZrH fueled, NaK-cooled system that employed thermoelectric conversion systems, developed 533 W of electric power, and had a mass of 436 kg. The former Soviet Union successfully utilized 34 fission systems in space. As will be discussed later in this paper, the recommended approach for developing and implementing use of in-space nuclear power should be based upon an affordable, safe, logical, and straightforward three-phase program. Nuclear fuels applicable to phase 1 systems have flown in space. High efficiency electric thrusters have flown in space. Deployable radiators have flown in space. High efficiency energy conversion is now being developed for space.
2.2. Space Fission Propulsion System Safety The use of space fission systems must not adversely affect mission safety. Space fission propulsion systems produce from ten to ten thousand times less power than commercial reactors, depending on the mission application. Safety features can be straightforward and easily explained to regulators and the public. Using current technology it is possible to fabricate reactors for in-space propulsion that are virtually non-radioactive at launch. The radiological inventory in the fresh fuel for these systems is about the same as a truckload of uranium ore, with which uranium miners work safely in nations around the world. The nuclear characteristics of the space reactor can be fully verified prior to launch without any significant inventory buildup through the use of low-energy critical experiments. The primary concern for safety will be to ensure that the reactor does not inadvertently turn on during handling or any conceivable launch accident scenario. This can be accomplished through proper choice of materials and geometry, in-core shutdown mechanisms, or launch with a portion of the fuel removed. Safety during operation can be addressed through proper radiation shielding, passive removal of decay heat, and redundant and diverse shutdown mechanisms. For example, the core can be designed with negative thermal feedback, allowing the reactor power to naturally follow variations in load without needing adjustment of the control elements. Finally, steps can be taken to ensure that the reactor does not re-enter the earth's biosphere before radioactivity generated by high-power operation has decayed back to acceptable levels.
2.3. Fission Systems Differ from Radioisotope Systems Both radioisotope and fission systems can be made extremely safe. However, fission systems are completely different from the radioisotope systems currently used by NASA (over 20 have been launched by the U.S.). While radioisotope thermoelectric generators (RTGs) use the alpha decay heat of Pu-238 as their source of energy, fission systems use the heat generated from fissioning U-235. Unlike radioisotope systems, fission systems are virtually non-radioactive at launch. For example, the general purpose heat source-radioisotope thermoelectric generator contains 133,000 curies of Pu-238, whereas a fission reactor (launched cold) has a total radiological inventory of <10 curies at launch. The radioisotope heater units on the Mars Pathfinder's Sojourner Rover contained nearly 100 curies at launch. Fission systems can have a much higher energy density and energy production rate than radioisotope systems, and are suitable for propelling large payloads throughout the solar system.
In-space nuclear power as an enabling technology for exploration...
2.4. Fission System Launch Approval Process Well Defined There is a precedent for operating small research-sized reactors in space. There are presently over 30 shut-down nuclear reactors orbiting Earth. All but one of these are Russian reactors from Rorsat high-power radar satellites or other missions. The one U.S. reactor is the SNAP-10A, launched in 1965. Every U.S launch of a payload involvi0ng nuclear material must be reviewed by an Interagency Nuclear Safety Review Panel (INSRP). The INSRP reviews the sponsor's assessment of the risk and reports to the Office of Science and Technology Policy under the Executive branch. The President or his designee (usually the Science Advisor) then decides whether to grant launch approval. This process has been followed for 25 launches of nuclear materials over the past 40 yr and approval has always been granted. All but one of these launches have involved radioisotopic power sources (the other launch being that of the SNAP-10A reactor), and it is expected that a modem space reactor will follow the same process. 3. PRODUCING THRUST FROM FISSION ENERGY 3.1. Continuous Impulse Systems (Electric Propulsion) Energy supplied by the fission system can be used to drive several types of continuous impulse (electric propulsion) systems. This section describes a few candidates that could be considered for a phase 1 system. Electrostatic or ion propulsion systems generate thrust by accelerating charged particles (ions) through electrostatic fields. The exhaust velocity of the ions depends on the accelerating potential, the charge of the particles, and their mass. The major ion production techniques are electron bombardment, contact/surface ionization, and radio frequency (RF) ionization. The VASIMR is being developed in the Advanced Space Propulsion Laboratory at Johnson Space Center. The VASIMR engine creates plasma by ionizing a gas using a helicon injector and heats the plasma by the process of ion cyclotron resonance. Radio frequency (RF) electromagnetic waves are used in both the ionizing and heating processes. An axisymmetric magnetic field is used to confine the plasma radially and to direct and accelerate the plasma axially by the action of a magnetic nozzle. The plasma released through the magnetic nozzle provides the thrust. Energy is delivered to the VASIMR module via a 1330 VAC three-phase bus. The energy is converted to 50 VDC by a lightweight transformer/rectifier. It is then converted to useful RF energy by solid-state RF amplifiers contained in the VASIMR system. Input to these amplifiers will be 50 V and 120 A. Since the radiators used by the fission reactor will operate at temperatures well above the operating temperature of many of the components in the VASIMR system, the VASIMR will require separate radiators. The 60kWe VASIMR system will need to radiate 30 kW at acceptable temperatures. This will be accomplished by radiators attached to the RF module. The PIT is a high thrust electromagnetic propulsion system that can operate at high efficiency over a wide range of specific impulse values. In its basic form, the PIT consists of a fiat spiral coil covered by a thin dielectric plate. A pulsed gas injection nozzle distributes a thin layer of gas propellant across the plate surface at the same time that a pulsed high current discharge is sent through the coil. The rising current creates a time varying magnetic field, which in turn induces a strong azimuthal electric field above the coil. The electric field ionizes the gas propellant and generates an azimuthal current flow in the resulting plasma. The current in the plasma and the current in the coil flow in opposite directions, providing a mutual repulsion that rapidly blows the ionized propellant away from the plate to provide thrust. The thrust and specific impulse can be tailored by adjusting the discharge power, pulse repetition rate, and propellant mass flow, and there is minimal if any erosion due to the electrodeless nature of the discharge. The PIT operates at very high power density and would be a good candidate for high thrust electric thrusting missions. Electric-to-jet efficiencies of as high as 80% appear to be achievable. In addition, a wide range of propellants can be utilized by the PIT.
- 405 -
R. Sackheim et al.
Systems for conversion to electricity (for the electric thrusters) include thermoelectric fiE), thermionic, Rankine, Stirling, and recuperated Brayton cycles. Thermoelectrics have been the most used in space and typically result in only 5% conversion efficiency. Thermionics are a bit more efficient but require a higher temperature. Brayton, Rankine, and Stirling achieve typically 20-30% efficiency or more but require moving parts. Higher efficiencies can be achieved if a large radiator is available to provide a low-temperature heat sink for the cycle. There has been extensive development of all of these conversion systems on Earth, but no dynamic conversion system has been used in space to date. The Department of Energy and Jet Propulsion Laboratory developed operational hardware for a 25-kW toluene Rankine system driven by solar thermal energy. The program also generated several conceptual designs for 100 kWe systems. There is a very extensive industrial data base and fabrication experience for open-cycle Brayton units: they form the basis for commercial and military jet engines as well as helicopter engines. The Brayton conversion system could be designed to have only one moving part: a single shaft connected to the turbine, the electrical generator rotor, and the compressor. In closed systems, this single shaft floats on a gas bearing bled off from the main gas flow and returned to it. Space Brayton units would operate at constant output in a weightless environment, so they are expected to last many years without maintenance. A 52,000-hr ground test of a 10.7-kWe closed Brayton unit at NASA/LeRC in the 1960s supports these expectations. Stirling Technology Company is currently developing Stirling engines for space applications. Units capable of producing 3 kWe each are expected to be available in CY01. 3.2. Direct Thermal Thrust Production In nuclear thermal propulsion, the fission reactor operates as a compact heat exchanger, transferring fission energy into hydrogen propellant. This enables a high thrust-to-weight ratio. In the U.S., the Nuclear Engine for Rocket Vehicle Application program tested 21 nuclear reactors, which demonstrated as much as (in different engines) 845 s Isp, 930 kN thrust, 109 min accumulated time at full power, and 4.1 GWth power. Though considered a technical success, the program was canceled in 1973 when plans for human missions to Mars were abandoned. The Space Nuclear Thermal Propulsion program tested five fuel elements and a critical assembly. Before the program was terminated in 1993, an environmental impact statement was generated. Advanced solid fuels (ternary carbides, cermets) may enable missionaveraged Isp's of >_900 s. Extensive Russian experience also indicates that solid core nuclear thermal propulsion systems with specific impulses >900 s may be feasible. Liquid fissile fuels may enable Isp's >1200 s, and gas/plasma fissile fuels may enable Isp's >2500 s. ~3 3.3 Phase 1 system capable of providing 300 kWt currently under investigation One candidate for a phase 1 space fission propulsion system is the Safe Affordable Fission Engine (SAFE)-300. The SAFE-300 is a potential near-term, low-cost space fission propulsion system capable of producing 300 kWt even following multiple failures. The system consists of 284 fuel pins and 103 heat pipes. The outer diameter of both the fuel pins and the heat pipes is 1.27 cm. The fuel pin inner diameter is 1.1 cm. The dimensions of the hexagonal core are 25.4 cm across the fiats by 40 cm long. The system is fueled with uranium dioxide with an average smear density of 92% in the ground-fueled zone and 85% in the space-fueled zone. Uranium enrichment in both zones is 97%. A cross section of the SAFE-300 is shown in Figure 7. A picture of a SAFE-30 full-core primary heat transport test is shown in Figure 8. The SAFE-300 builds on experience gained from the heat pipe bimodal system effort that has been ongoing at Los Alamos National Laboratory (LANL) since 1995.
- 406 -
In-space nuclear power as an enabling technology for exploration...
Figure 7. Cross section of the SAFE-300.
Figure 8. Full core primary heat transport test of SAFE-30.
-407-
R. Sackheim et al.
The SAFE design is optimized to ensure that safety, cost, and schedule goals are met. The nuclear portion of the SAFE is virtually non-radioactive at launch (<10 curies), remains subcritical during all credible launch accidents, has no single-point failures (capable of delivering full thermal power even after multiple failures), operates within established limits on fuel burnup and radiation damage, and is designed to allow extensive, realistic, and affordable testing. The SAFE design approach is applicable to moderate power levels (<2 MWt), and takes advantage of design simplifications that are possible in that power range. The SAFE is applicable primarily to propelling and providing energy to robotic missions, although extremely close derivatives of the SAFE could provide an energy-rich surface environment for crewed Mars missions. Nuclear system components (fuel, control systems, radiation shields, etc.) developed and utilized by the SAFE are on the development path towards second or third generation nuclear propulsion systems that can be used to propel humans around the solar system. No new nuclear technology development is required for full confidence in the SAFE. Development challenges (if any) are kept in the area of thermal-hydraulics, structures, and "balance-of-plant." A key to the affordability of the SAFE-300 is that it can be developed primarily at a nonnuclear facility. Development at a non-nuclear facility is viable because the SAFE is designed to operate well within established limits on nuclear fuel burnup and radiation damage, and is designed to allow realistic full-system tests with resistance heaters simulating heat from fission. Because all components operate well within established limits on radiation damage and nuclear fuel bumup, radiation damage is not an anticipated cause of failure. Significantly more likely causes of failure (thermal and stress-related) can be resolved using resistance heaters in place of heat from fission. The SAFE-300 is designed to be resistant to all potential failure mechanisms, and resistance to the most significant potential failure mechanisms can again be demonstrated at a non-nuclear facility. The design allows for extensive testing of both development units and actual flight units to ensure that the system can withstand launch loads and both normal and offnormal operating environments. Realistic, timely, and affordable testing of the SAFE system will be performed at a non-nuclear facility. On completion of thermal and stress testing, minimal operations are required to remove the heaters, fuel the system, and ready the system for launch. Data required to verify the predicted nuclear safety and operating characteristics of the system are obtained from low-energy criticality experiments. These experiments can be performed at existing Department of Energy laboratory facilities (e.g., LANL TA-18, Sandia Area-V). The baseline SAFE-300 would operate for over 20 yr without exceeding demonstrated nuclear fuel burnup capability. In addition, the core structural materials used by the SAFE-300, and liquid metal heat pipes have withstood fast neutron fluences several times those that would be received during a 20-yr mission. Because adequate radiation tolerance has been demonstrated, required in-reactor testing of SAFE-300 components is minimized. If in-reactor tests are required, it is important to note that they are typically two orders of magnitude less expensive than full-power ground nuclear tests, and can be accomplished quickly. Approximately 110 kg of highly-enriched uranium (HEU) dioxide fuel is required by the SAFE. This fuel can be fabricated using existing facilities at LANL's TA-55. The LANL facility has previously fabricated highly enriched space reactor fuel at a rate of 10 kg/week. For producing the SAFE fuel, glove boxes would first be cleaned, then fresh HEU utilized in the fuel fabrication. LANL has demonstrated capability in finely controlling fuel stoichiometry, fuel dimensions, fuel density, and other parameters required for ensuring excellent fuel performance. The estimated total cost for the SAFE highly enriched U O 2 fuel is $8M. Estimated delivery time is 24 mo from receipt of funding at LANL. Non-nuclear testing of refractory metal SAFE modules has already been completed with a high degree of success. Specifically, a module temperature >1750 K was achieved as well as an isothermal heat pipe temperature of 1450 K. Resistance-heated testing of a full 30 kWt stainless steel core is scheduled to begin in August 2000 at MSFC.
-
408 -
In-space nuclear power as an enabling technologyfor exploration... 4. Conclusion
High power fission propulsion systems have the potential for delivering tremendous in-space performance. Developing that performance to its full potential has been hindered by the perception that fission propulsion systems require excessive time and money to utilize. A simple, programmatically straight forward, low risk and safe three phase approach to successful utilization of fission propulsion systems has been devised. If successfully developed and implemented, in-space nuclear power is an enabling technology for deep space exploration. Safe affordable nuclear reactors operating reliably in deep space are crucial to providing any hope of meaningful scientific and/or human exploration missions with adequate margins for success. Additionally, nuclear reactors provide the solution for generating sufficient energy in deep space and/or on planetary surfaces to do required scientific and exploratory investigations. With these overwhelming imperatives, combined with our desire to explore, in-depth, our solar system out to the heliopause and beyond, plus the technical maturity and inherent safety features of today's nuclear reactors; it is time to embark on a program to flight demonstrate an in-space fission system for the global benefit and sustained economic growth of all mankind. REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
11.
12. 13.
La Point, M.R. and John M. Sankovic, "High Power Nuclear Electromagnetic Propulsion Research at the NASA Glenn Research Center," Proceedings of STAIF 2000, ISBN 1-56396-920-3, 2000, pp. 1538-1543. Koenig, D.R., "Experience Gained from the Space Nuclear Rocket Program (Rover)," LA-10062-H, Los Alamos National Laboratory, Los Alamos, NM 87545. Brophy, J.R., "Ion Propulsion System Design for the Comet Nucleus Sample Return Mission," AIAA 2000-3414, 36th AIAA Joint Propulsion Conference, July 2000, Huntsville, AL. Chang-Diaz, F. et al., 'VASIMR Workshop,' Advanced Space Propulsion Laboratory, NASA, Johnson Space Center, Houston, TX, 22-24 March 2000. Shepherd, L.R. and A.V. Cleaver, "The Atomic Rocket-l," J. Brit. Int. Soc., 7, 185-194, 1948a. Shepherd, L.R. and A.V. Cleaver, "The Atomic Rocket-2," J. Brit. Int. Soc., 7, 234-241, 1948b. Shepherd, L.R. and A.V. Cleaver, "The Atomic Rocket-3," J. Brit. Int. Soc., 8, 23-27, 1949. Bussard, R.W., and R.D. DeLauer, Fundamentals of Nuclear Flight, McGraw-Hill Book Company, New York, 1965. Stuhlinger, E., Ion Propulsion for Space Flight, McGraw-Hill Book Company, New York, 1964. Sholtis, Jr., J. A., Connell, L. W., Brown, N. W., Mims, J. E., and Potter, A., "U. S. Space Nuclear Safety: Past, Present, and Future," in A Critical Review of Space Nuclear Power and Propulsion 1984-1993, M. S. E1-Genk, (ed.), U. of New Mexico Institute for Space Nuclear Power Studies, Albuquerque, NM, AIP Press, New York, NY, pp. 269-304, 1994. Borowski, S. K., L. A. Dudzinski and M. L. McGuire, "Bimodal NTR and LANTR Propulsion for Human Missions to Mars/Phobos," in Proceedings of the Space Technology and Applications Forum, AIP Conference Proceedings 458, American Institute of Physics, pp. 1261-1268, 1999. Anghaie, Samim, Personal Communication, University of Florida, June, 2000. Angelo, J.A. and D. Buden, Space Nuclear Power, pp 179, Orbit Book Company, Inc., Malabar, FL, 1985.
- 409 -
This Page Intentionally Left Blank
Interstellar Probe using a Solar Sail: Conceptual Design and Technological Challenges P. C. Liewera, R. A. Mewaldtb, J. A. Ayona, C. Garnera, S. Gavita, and R. A. Wallace a ajet Propulsion Laboratory, Pasadena, CA 91109 USA bCalifomia Institute of Technology, Pasadena, CA 91125 USA
NASA's Imerstellar Probe, conceived to travel to 200-400 AU, will be the first spacecraft designed specifically to explore the unknown regions beyond the solar system and directly sample the dust, neutrals and plasma of the surrounding imerstellar material. Here we presem the mission concept developed by NASA's Interstellar Probe Science and Technology Definition Team in 1999 and discuss the technological challenges it presents. The Team selected a solar sail concept in which the spacecraft reaches 200 AU in 15 years. This rapid passage is made by using a 400-m diameter solar sail and heading first inward to--0.25 AU to increase the radiation pressure. The Probe then heads out in the interstellar upwind direction at--14/AU year, about 5 times the speed of the Voyager 1 and 2 spacecraft. Advanced lightweight instruments and spacecraft systems, as well as solar sail propulsion, are needed to achieve these high speeds. The spacecraft coasts to 200-400 AU, exploring the Kuiper Belt, the boundaries of the heliosphere, and the nearby imerstellar medium. 1. INTRODUCTION Little is known about what lies beyond the solar system. NASA's Interstellar Probe mission, which will travel to 200-400 AU, will be the first spacecraft designed to travel through the outer reaches of the solar system and sample the nearby interstellar material beyond the influence of our Sun. On its way, Interstellar Probe will explore the outer solar system and the boundaries of the heliosphere -- the bubble blown in the imerstellar medium by the supersonic solar wind. Its unique trajectory allows the mission to address key questions about the nature of the primordial solar nebula, the structure and dynamics of our heliosphere, the properties of organic material in the outer solar system and interstellar medium, the nature of other stellar systems that may also harbor planets, the chemical evolution of our galaxy, and the origins of matter in the earliest days of the universe. As envisioned by NASA's Imerstellar Probe Science and Technology Definition Team, a 400-m solar sail would be used to carry the Probe to 200 AU in 15 years, with sufficient consumables (power, fuel) to last to 400 AU (30 years). The most critical advanced technology needed for this ambitious mission is solar sail propulsion, but advances in spacecraft subsystems and instruments are also necessary. This paper will present the Interstellar Probe concept developed by the Team and discuss some of the advanced technologies needed.
-411-
P.C. Liewer et al.
Table 1 NASA's Interstellar Probe Science and Technology Definition Team Members Chairman Richard Mewaldt, Emma Bakes, NASA Ames Research Center Califomia Institute of Technology Priscilla Frisch, University of Chicago Study Scientist Paulett Liewer, Jet Propulsion Laboratory Herbert Funsten, Los Alamos National Laboratory Program Manager Sarah Gavit, Mike Gruntman, University of Southern Jet Propulsion Laboratory California Program Scientist Vernon Jones, Les Johnson, Marshall Space Flight Center NASA Headquarters Deputy Program Scientist James C. Ling, J. R. Jokipii, University of Arizona NASA Headquarters William Kurth, University of Iowa Program Executive Glenn H. Mucklow, Jeffrey Linsky, University of Colorado NASA Headquarters Renu Malhotra, Lunar and Planetary Institute NASA Transportation Dave Stone, NASA Ingrid Mann, California Institute of Interagency Representatives Technology Dave Goodwin, DOE Ralph McNutt, John Hopkins University, Eugene Loh, NSF Applied Physics Laboratory Foreign Guest Participants Wolfgang Droege, University of Kiel, Eberhard Moebius, University of New Germany Hampshire Bemd Heber, Max Planck Institute for William Reach, California Institute of Aeronomy, Germany Technology Claudio Maccone, Torino, Italy Steven T. Suess, Marshall Space Flight Jet Propulsion Laboratory Support Center Juan Ayon Adam Szabo, Goddard Space Flight Center Charles Budney Jim Trainor, Goddard Space Flight Center, Eric De Jong Retired Neil Murphy Gary Zank, University of Delaware Richard Wallace Thomas Zurbuchen, University of Michigan The Interstellar Probe Science and Technology Definition Team (Table 1) met during the spring and summer of 1999 under sponsorship of the NASA Office of Space Science. The Team was charged with defining the science requirements and developing a concept for an interstellar probe mission for the Sun-Earth-Connection Roadmap, a part of NASA's strategic planning activities. The resulting concept builds on several previous studies. In a 1990 study by Holzer et al., a 1000 kg spacecraft was to acquire data out to -~200 AU, exiting the solar system at-~10 AU/year using chemical propulsion coupled with impulsive maneuvers near the Sun. In a 1995 study of a smaller interstellar probe (Mewaldt et al., 1995), a-~200 kg spacecraft was to reach exit velocities of-~6 to 14 AU/year, depending on launch vehicle and trajectory, using chemical propulsion with planetary gravity assists or impulsive maneuvers near the Sun. Recent technological advances, notably lighter reflective sail materials (Garner et al., 1999; Garner and Leipold, 2000) and lighter spacecraft designs, now make it feasible to accomplish essentially the same mission using a solar sail to accelerate a 150 kg spacecraft to -~15 AU/year, allowing the mission to reach-~200 AU in ~15 years and ~400 AU in-~30 years.
-412 -
Interstellar probe using a solar sail: ...
2. SCIENCE OBJECTIVES AND SCIENTIFIC PAYLOAD The Sun is thought to be located near the edge of a local interstellar cloud (LIC) of low density (~0.3 /cc) material blowing from the direction of star-forming regions in the constellations Scorpius and Centaurus. The solar wind, a continual low-density flux of charged particles, streams outward from the corona and expands supersonically throughout and beyond the solar system. The solar wind and the interstellar medium interact to create the global heliosphere, shown schematically in Fig. 1. The interaction between the solar wind, flowing radially outward at 400-800 km/sec, and the local interstellar material, flowing at -~25 km/sec, creates a complex structure extending perhaps 200-300 AU in the upstream (towards the local interstellar flow) direction and thousands of AU tailward. At present, there are no direct measurements of the size and structure of the heliosphere and our present
Figure 1. The global heliosphere is created by the supersonic solar wind diverting the interstellar flow around the Sun. The interstellar ions and neutrals flow at 25 km/s relative to the Sun. The solar wind makes a transition to subsonic flow at the termination shock. Beyond this, the solar wind is turned toward the heliotail. The heliopause separates solar material and magnetic fields from interstellar material and fields. understanding is based on theory and modeling, constrained by a few key measurements. Voyager 1&2, now at approximately 78 and 61 AU respectively, should soon reach the first boundary in this complex structure, the solar wind termination shock, where the solar wind makes a transition from supersonic to subsonic flow. Beyond the termination shock lies the heliopause, which is the boundary between solar wind and interstellar plasma. Several recent
-413 -
P.C. Liewer et aL
estimates place the distance to the termination shock at ~80 to 100 AU, with the heliopause at ~120 to 150 AU. The heliosphere shields the solar system from the plasma, energetic particles, small dust, and fields of the interstellar medium. The local interstellar medium is thought to be younger material than that which formed the presolar nebulae and is expected to have a different elemental composition. Our present knowledge of the interstellar medium surrounding our heliosphere comes either from astronomical observations, measurements of sunlight resonantly scattered back towards us by interstellar H and He, or in situ measurements of the dust and neutral gas that penetrate the heliosphere. To observe these directly it is necessary to go beyond the heliopause. The Interstellar Probe Mission would be designed to travel beyond the region strongly influenced by the solar wind and make a significant penetration into nearby interstellar space. Interstellar Probe's unique voyage from Earth to beyond 200 AU will enable the first comprehensive measurements of plasma, neutrals, dust, magnetic fields, energetic particles, cosmic rays, and infrared emission from the outer solar system, through the boundaries of the heliosphere, and on into the ISM. This will allow the mission to address key questions about the distribution of matter in the outer solar system, the processes by which the Sun interacts with the galaxy, and the nature and properties of the nearby galactic medium. The principal scientific objectives of the Interstellar Probe mission as defined by the Team are to 9Explore the nature of the interstellar medium and its implications for the origin and evolution of matter in our Galaxy and the Universe; 9Explore the influence of the interstellar medium on the solar system, its dynamics, and its evolution; 9Explore the impact of the solar system on the interstellar medium as an example of the interaction of a stellar system with its environment; 9Explore the outer solar system in search of clues to its origin, and to the nature of other planetary systems. To achieve these broad, interdisciplinary objectives, the strawman scientific payload includes an advanced set of miniaturized, low-power instruments specifically designed to make comprehensive, in situ studies of the plasma, energetic particles, fields, and dust in the outer heliosphere and nearby ISM. These instruments will have capabilities that are generally far superior to those of the Voyagers. The wide variety of thermal and flow regimes to be encountered by Interstellar Probe will be explored by a comprehensive suite of neutral and charged particle instruments, including a solar wind and interstellar ion and electron detector, a spectrometer to measure the elemental and isotopic composition of pickup and interstellar ions, an interstellar neutral atom spectrometer, and a detector for suprathermal ions and electrons. Two cosmic ray instruments are included, one for H, He, electrons and positrons, and one to measure the energy spectra and composition of heavier anomalous and galactic cosmic rays. The magnetometer will make the first direct measurements of the magnetic fields in the ISM, and the plasma and radio wave detector will measure fluctuations in the electric and magnetic fields created by plasma processes and by interactions and instabilities in the heliospheric boundaries and beyond. As 