BW2003 WORKSHOP
Mathematical, Theoretical and Phenomenological Challenges Beyond the Standard Model Perspectivesof the Balkan Collaborations
This page intentionally left blank
BW2003 WORKSHOP
Mathematical, Theoretical and Phenomenological Challenges Beyond the Standard Model Perspectives of the Balkan Collaborations Vrja&a Banja, Serbia and Montenegro 29 August - 3 September 2003
Editors
G. Djordjevic L. Nesic University of Nis, Serbia and Montenegro
J. Wess Ludwig-Maximillianss University University & Max Planck Institute, Germany
World Scientific NEW JERSEY * LONDON * SINGAPORE BElJlNG * SHANGHAI * HONG KONG - TAIPEI BANGALORE
Published by
World Scientific Publishing Co. Re. Ltd.
5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
British Library Cataloguing-in-PublicationData A catalogue record for this book is available from the British Library
MATHEMATICAL, THEORETICAL AND PHENOMENOLOGICAL CHALLENGES BEYOND THE STANDARD MODEL Perspectives of the Balkan Collaborations Copyright 0 2005 by World Scientific Publishing Co. Re. Ltd
All rights reserved. This book, or parts thereof; may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.
For photocopying of material in this volume, please pay acopying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.
ISBN 981-256-130-7
Printed in Singapore by World Scientific Printers (S)Pte Ltd
V
Preface The Balkan Workshop (BW2003): Mathematical, Theoretical and Phenomenological Challenges Beyond the Standard Model - PERSPECTIVES OF BALKAN COLLABORATIONS - was held from 29 August to 2 September 2003, in VrnjaEka Banja, Serbia, right after the Fifth General Conference of the Balkan Physical Union (BPU5). The main purpose of the workshop was to foster the communication among the researchers in the Balkan region as well as between their international colleagues. This meeting created the opportunity for the researchers working in the general area of high energy physics (HEP) a t different institutions in the Balkan countries to present their work and results. It stands as a logical and natural extension of “Wissenschaftler in globaler Verantwortung (WIGV)” - Scientists in global responsibility initiative - an initiative for the advancement of the scientific contact between Germany and the countries which have emerged from the former Yugoslavia. As expected, a creative and supportive environment facilitated the establishment of new, closer collaborative ties as well as strengthening of the existing regional and interregional collaborations. Attendance of about 50 participants from 17 countries has encouraged closer contacts and cooperation between their faculties, universities and institutions, initiating an exchange of scientific personnel and enabling joint applications for support and participation in international projects. We expect that this scientific integration will have a positive influence on the society in the Balkan/Southeast Europe countries, in general. Between many topics considered during the workshop, we emphasized on: strings and superstrings; supersymmetry and conformal field theory; noncommutative, gauge and string field theories; D-branes and matrix models; cosmology, quantum gravity, extra dimensions, grand unification; particle physics, neutrino physics and various aspect of noncommutative, q-deformed and nonarchimedean models. A few rather poor mathematical topics, but also related to the main scope of the meeting, e.g. integrability of some mechanical systems, were also considered. The invited lectures, ten of which were presented in this volume, gave an excellent review of the “hot” topics in theoretical HEP, Q F T and cosmology. We believe that many readers of this book, as well as the young PhD students who attended the workshop, will benefit a lot from this assembly of excellent papers.
vi
Shorter and more topical papers of other lecturers cover many of all actual problems in theoretical and mathematical problems in HEP. That highlighted the interest of researchers from this region as well as of from about 15 participants from West Europe (mainly from Germany), USA, Russia and South America. In total, more than 80% of the speakers sent their contribution; we would like to thank them for this huge collection. Regrettably, a written document cannot record the stimulating atmosphere and the fruitful informal discussions that took place during lunches and coffebreaks, as well as during a visit to the Serbian orthodox monasteries from Middle Ages, located near VrnjaEka Banja We would like to thank Dragoljub Dimitrijevic for great help in preparing this proceedings. Kind help of Frank Mayer is also warmly acknowledged. We would like to thank also World Scientific Publishing Company (especially to Lance Sucharov, Katie Lydon and Rhaimie Wahap) for their goodwill and interest in publishing this proceedings. Goran Djordjevib LjubiSa NeSi6 Julius Wess NiS, Munich, September 2004
vii
Advisory Committee A. Albrecht (Davis) G. Altarelli (CERN) I. Antoniadis (CERN) E. Arik (Istanbul) TBC C. Bachas (Paris) J . Bagger (Baltimore) S. Bellucci (Frascati) S. Dimopoulos (Stanford) J . Ellis (CERN) B. Guberina (Zagreb) R. Jackiw (MIT) J . Louis (Halle) D. Luest (Berlin) S. Meljanac (Zagreb) L. Mezincescu (Miami) H. Nicolai (Golm) S. Randjbar-Daemi (ICTP) A. Sen (Allahabad) G. Senjanovic (ICTP) E. Sezgin (Texas AM Univ.) I. Todorov (Sofia) J. Wess (Munich) G. Zoupanos (Athens)
Organizing Committee B. Aneva (Sofia) A. Balaz, (Belgrade) G. L. Cardoso (Berlin) G. Djordjevic (Nis, Munich), chairman V. Dragovic (Belgrade, Trieste) L. Moeller (Munich) Lj. Nesic (Nis) A. Nicolaidis (Thessaloniki) C. Sochichiu (Chisinau, F’rascati)
This page intentionally left blank
List of Participants INRNE, Bulgarian Academy of Sciences, 1784 Sofia Bulgaria
[email protected] Gleb Arutyunov Max-Plank-Institut fuer Gravitationphysik Am Muhlenberg 1, D-14476 Golm, Germany
[email protected] Paolo Aschieri Sektion Physik der Ludwig-Maximilians-Universitat Theresienstr. 37, D-80333 Munchen, Germany
[email protected] Borut Bajc Jozef Stefan Institute, Jamova 39, 1001 Ljubljana Slovenia,
[email protected] Antun Balaz Institute of Physics, P.O.Box 57, 11 001 Belgrade Serbia and Montenegro ant
[email protected] .ac.yu Klaus Behrndt Max-Plank-Institut fuer Gravitationphysik Am Muhlenberg 1, D-14476 Golm, Germany
[email protected] Faculty of Physics, University of Belgrade, P.O. Box 368 Maja Buric 11 001 Belgrade, Serbia and Montenegro majabQphy. bg.ac.yu Institut fur Physik, Humboldt University Gabriel Lopes Cardoso Newtonstrasse 15 D-12489 Berlin, Germany
[email protected] Masud Chaichian Helsinki Institut of Physics, P.O. Box 64, FIN-00014 Helsinki, Finland
[email protected] Dragoljub Department of Physics, Faculty of Sciences Dimitrijevic P.O. Box 224, 18000 Nis, Serbia and Montenegro
[email protected] Dragan Djordjevic Department of Mathematics, Faculty of Sciences P.O. Box 224, 18000 Nis, Serbia and Montenegro
[email protected] Goran Djordjevic Department of Physics, Faculty of Sciences P.O. Box 224, 18000 Nis, Serbia and Montenegro gorandj @junis .ni.ac. yu
Boyka Aneva
ix
X
Berkol Dogan
Vladimir Dragovic
Dan Radu Grigore
Sinisa Ignjatovic
Nemanja Kaloper
Ivan Kostov
Dusko Latas
George Lazarides
Katarina Matic
Alejandra Melfo Aleksandar Mikovic
Alexei Morozov
Bogazici University, Department of Physics, Bebek, 80815 Istanbul, Turkey
[email protected] Mathematical Institute Serbian Academy of Sciences and Arts P.O. Box 367, 11001 Beograd, Serbia and Montenegro
[email protected] Department of Theoretical Physics Inst. Atomic Physics, Bucharest-Magurele MG 6, Romania grigoreQtheor1.theory.nipne.ro Faculty of Natural Sciences and Mathematics Mladena Stojanovica 2, Banja Luka Republic of Srpska, Bosnia and Herzegovina
[email protected] Department of Physics, University of California Davis, CA 95616 USA kaloper @physics.ucdavis .edu Service de Physique Theorique, CNRS-URA 2306 C.E.A-Saclay, F-91 191 Gif-Sur-Yvette, F'rance
[email protected] Faculty of Physics, University of Belgrade P.O. Box 368, 11 001 Belgrade, Serbia and Montenegro lat
[email protected] .ac .yu Physics Division, School of Tehnology Aristotle University of Thessaloniki Thessaloniki GR 54124, Greece
[email protected] Faculty of Physics, University of Belgrade P.O. Box 368, 11 001 Belgrade, Serbia and Montenegro
[email protected] CAT, Universidad de Los Andes, Merida, Venezuela
[email protected] Departamento de Matematica e Ciencias de Computacao Universidade Lusofona de Humanidades e Tecnologias Av. do Campo Grande, 376, 749-024 Lisboa, Portugal
[email protected] State Science Center of Russian FederationInstitute of Theoretical and Experimental Physics
xi
B. Cheremushkinskaja, 25 MOSCOW, 117218, Russia
[email protected] Department of Physics, Faculty of Sciences Ljubisa Nesic P.O. Box 224, 18000 Nis, Serbia and Montenegro
[email protected] Argyris Nikolaidis Theoretical Physics Department University of Thessaloniki 54124 Thessaloniki, Greece
[email protected] Bojan Nikolic Faculty of Physics, University of Belgrade P.O. Box 368, 11 001 Belgrade, Serbia and Montenegro
[email protected] .ac.yu Institute for Nuclear Research and Nuclear Energy Todor Popov Bulgarian Academy of Sciences Tsarigradsko Chaussee 72, BG-1784, Sofia, Bulgaria
[email protected]. bg Voja Radovanovic Faculty of Physics, University of Belgrade P.O. Box 368, 11 001 Belgrade, Serbia and Montenegro
[email protected] .ac.yu Riccardo Rattazzi Theoretical Physics Division CERN CH-1211 Geneva 23, Switzerland
[email protected] Branislav Sazdovic Institute of Physics P.O.Box 57, 11 001 Belgrade, Serbia and Montenegro
[email protected] .ac.yu Volker Schomerus Service de Physique Theorique, CEA Saclay F-91191 Gif-sur-Yvette CEDEX, France vschomer @aei-potsdam.mpg .de Goran Senjanovic International Centre for Theoretical Physics Trieste, 34 014, Italy
[email protected] Corneliu Sochichiu INFN - Laboratori Nazionali di Frascati Via E. Fermi 40, 00044 Frascati, Italy Didina Serban
[email protected] Service de Physique Thorique CEA/Saclay - Orme des Merisiers F-91191 Gif-sur-Yvette Cedex, France
[email protected]
xii
Sergei Solodukhin International University Bremen School of Engineering and Science P.O.Box 750561, 28759, Bremen, Germany
[email protected] Jelena Stankovic Department of Physics, Faculty of Sciences P.O. Box 224, 18000 Nis, Serbia and Montenegro
[email protected] Gordan Stanojevic Department of Physics, Faculty of Sciences P.O. Box 224, 18000 Nis, Serbia and Montenegro
[email protected] Harold Steinacket Sektion Physik der Ludwig-Maximilians-Universitat Theresienstr. 37, D-80333 Munchen, Germany
[email protected] Svjetlana Terzic Faculty of Sciences and Mathematics Cetinjski put BB, 81000 Podgorica Serbia and Montenegro
[email protected] Rudjer Boskovic Institute Tomislav Terzic Theoretical Physics Division P.O. Box 180, HR-10002 Zagreb, Croatia
[email protected] Helsinki Institut of Physics, P.O. Box 64, FIN-00014 Anca Tureanu Helsinki, Finland
[email protected] Mihai Visinescu Department of Theoretical Physics National Institute of Physics and Nuclear Engineering Magurele, P.O.Box MG-6 76900 Bucharest, Romania
[email protected],ro Marko Vojinovic Faculty of Physics, University of Belgrade P.O. Box 368, 11 001 Belgrade, Serbia and Montenegro
[email protected]. yu Sektion Physik der Ludwig-Maximilians-Universitat Julius Wess Theresienstr. 37, D-80333 Munchen, Germany
[email protected] Marija Zamaklar International Centre for Theoretical Physics Trieste, 34 014, Italy
[email protected]
...
Xlll
Alexei Zamolodchikov
Laboratoire de Physique Mathematique Universite Montpellier 11, P1. E. Bataillon 34095 Montpellier, France zamolodQLPM .univ-montp2 .fr
Guests Martin Huber
Zvonko Maric
Vladimir Kouzminov
Ilija Savic
Metin Arik
European Physical Society 4 rue des Freres Lumieres F-68200 Mulhouse, France
[email protected] Serbian Academy of Arts and Sciences Knez Mihailova 35, 11001, Belgrade Serbia and Montenegro
[email protected] ROSTE Regional Bureau for Science in Europe Palazzo Zorzi, Castello 4930 30122 Venice, Italy
[email protected] Serbian Physical Society Pregrevica 118, 11080 Zemun Serbia and Montenegro
[email protected] Bogazici University, Department of Physics, Bebek, 80815 Istanbul, Turkey
[email protected]
CONTENTS
Preface Organizers and Committees List of Participants Conference Photo
V
vii ix xiv
I. Invited Lectures Integrable Structures in the Gauge/String Corespondence G. Arutyunov
3
Fluxes in M-theory on 7-manifolds: Gz-, SW(3)- and SU(2)-structures K. Behrndt, C. Jeschek
16
Noncommutative Quantum Field Theory: Review and its Latest Achievements M. Chaichian
32
Shadows of Quantum Black Holes N . Kaloper
47
Yukawa Quasi-Unification and Inflation G. Lazarides, C. Pallis
56
Supersymmetric Grand Unification: The Quest for the Theory A . Melfo, G. SenjanoviC
71
Spin Foam Models of Quantum Gravity A . Mikovic'
88
Riemann-Cartan Space-time in Stringy Geometry B. Sazdovic'
94
xv
xvi
Can Black Hole Relax Unitarily? S. N . Solodukhin
109
Deformed Coordinate Spaces Derivatives J. Wess
122
11. Short Lectures Deformed Coherent State Solution to Multiparticle Stochastic Processes B. Aneva
131
Non-Commutative GUTS, Standard Model and C, P, T Properties from Seiberg-Witten Map P. Aschieri
142
152
Seesaw, Susy and SO(10) B. Bajc On the Dynamics of BMN Operators of Finite Size and the Model of String Bits S. Bellucci, C. Sochichiu Divergencies in &expanded Noncommutative Yang-Mills Theory M. BuriC and V. Radovanovic'
162
SU(2)
Heterotic String Compactifications with Fluxes G. L. Cardoso, G. Curio, G. Dall'Agata and D. Lust Symmetries and Supersymmetries of the Dirac-type operators on Euclidean Taub-NUT space I. I. Cotciescu, M. Visinescu Real and pAdic Aspects of Quantization of Tachyons G. S. Djordjevic', Lj. NeSiC Skew-Symmetric Lax Polynomial Matrices and Integrable Rigid Body Systems V. DragoviC, B. GajiC
171
181
186
197
208
xvii
Supersymmetric Quantum Field Theories
220
D. R. Grigore Parastatistics Algebras and Combinatorics
231
T. Popov Noncommutative D-branes on Group Manifolds
241
J . Pawelczyk, H. Steinacker High-Energy Bounds on the Scattering Amplitude in Noncommutative Quantum Field Theory
247
A . Tureanu Many Faces of D-branes: from Flat Space, via AdS to pp-waves
258
M.Zamaklar Abstracts and Titles of Reports not Included in the Volume
269
Acknowledgements Sponsors Statement of intention Epilogue
273 275 277 279
This page intentionally left blank
I. INVITED LECTURES
This page intentionally left blank
INTEGRABLE STRUCTURES IN THE GAUGE/STRING CORRESPONDENCE
G. ARUTYUNOV Max-Planck-Institut fur Gmvitationsphysik Albert-Einstein-Institut Am Muhlenberg 1, 0-14476Potsdam, Germany E-mail:
[email protected] We discuss the integrable structures both of the classical closed string sigma model with the Ads5 x S5 taget space and of the planar maximally supersymmetricYangMills theory. By using the Biicklund transformations we show that in the sector of highly energetic and spinning strings the integrable structures of string and gauge theories match precisely up to two loops.
1. Gauge/String Duality
One of the fundamental questions of modern theoretical physics is the connection between gauge and string theories. In 1997 J. Maldacena conjectured a new surprising relation between gauge theories and strings.l According to the AdS/CFT duality conjecture, certain quantum supersymmetric conformal field theories have a dual formulation in terms of a closed superstring theory on the Anti-de-Sitter (Ads) background. To fully appreciate the non-triviality of this statement we recall that closed string theory contains gravity, and now it appears to have an alternative description in terms of a non-gravitational theory! In this lecture we report on a recent exciting progress towards understanding the fundamental example of the gauge/string dual pair, which involves four-dimensional maximally supersymmetric Yang-Mills theory and type IIB superstring propagating in the Ad& x S5 space-time (the product of a five-dimensional Ads space and a five-sphere). Already a first inspection shows that both theories mentioned above possess the same amount of symmetry, which can be taken as an initial evidence that they indeed might relate to each other in a non-trivial way. However, this reasoning is kinematical, and the real question is whether 3
4
these theories also share the same dynamical features. This is much harder to answer. By duality strongly coupled Yang-Mills theory is equivalent to weakly coupled string and vice versa. Because of a lack of adequate theoretical methods we neither have much insight into the strongly coupled regime of the gauge theory, nor on the structure of the spectrum of strings propagating in a curved space-time. Fortunately, due to so far poorly understood reasons, parts of the gauge and the string spectra, both accessible by existing mathematical tools, seem to allow for a direct comparison. The modern development of the gauge/string duality was initiated by Berenstein, Maldacena and Nastase2 (BMN) who noticed that in a certain (Penrose) limit the corresponding string theory becomes solvable and its excitations can be identified with gauge theory operators of a certain type. Later on it became clear that there exists even a larger sector of string state^^^^^^ which is accessible by semiclassical6 methods. In parallel, the recent advances in gauge theory are due to an important observation that the planar JV = 4 super Yang-Mills theory is integrable in the one-loop approximation7t8 and, very likely, at higher loops as well.g 2. Integrable Structure of Gauge Theory
The maximally supersymmetric SYM theory has the following field content: six scalar fields 42, i = 1,.. . ,6, a vector field A,, and four Majorana fermions $,: where T = 1,.. . ,4. All fields are in the adjoint representation of SU(N). The action is
ri)are ten-dimensional Dirac matrices in the Majorana-Weyl Here (P, representation. In what follows it will be convenient to use the ’t Hooft coupling X = g 2 N as a natural loop counting parameter. The theory we consider is known to be finite, i.e. the beta-function vanishes at any loop order. The absence of scale implies that SYM remains conformal even at the quantum level. Supersymmetry together with conformal invariance are combined in a larger superconformal group known as PSU(2,214), which contains the bosonic subgroup S0(4,2) xSO(6): PSU(2,214) 3 Su(2,2)xSU(4)~z S0(4,2)xSO(6)
5
The basic physical quantities of this theory are the local, gauge-invariant composite operators O ( x ) ,i.e. the operators constructed as products of elementary fields. They transform in unitary irreducible representations of the superconformal group. Therefore, to each operator one can associate scaling (conformal) dimension A, the Lorentz spins s1,sa and three additional Dynkin labels ai related to the internal SU(4) symmetry:
[A,s i , ~ 2a ;i, a2, a31 . These are the quantum numbers associated to any highest weight state (also called a superconformal primary state) of a supersymmetry multiplet. An important class of operators we will be concerned in here is
O ( x )= Tr (@pa$@$) + ...,
@1 =
4’
+ i$2,
etc.
The dots indicate arbitrary orderings of the scalars inside the trace. All such operators realize irrep of SU(4) with labels [J2 - J3, J1 - J2, J2 J3]. Note that these operators are holomorphic and their dimension in free theory is
+
A = J1 + J2 + J3.
In general, due to quantum fluctuations, scaling dimensions get shifted from their classical values and acquire an “anomalous” piece. Conformal primary operator 0 is an eigenstate of the dilatation operator D , which is one of the generators of the conformal algebra, and conformal dimension is its eigenvalue
DO = A(& N ) O , k=O
It should be stressed that the existence of anomalous dimensions is one of the most important concepts of conformal field theory as they constitute the spectrum of the theory. At the same time calculation of anomalous dimensions proved to be one of the hardest problems of quantum field theory. Until recently little was known about the general behaviour of anomalous dimensions in super Yang-Mills theory, see however Ref. 10, 11. The major problem arises, even at one loop, due to the complicated mixing between operators differing by the ordering of the scalars inside the trace. However, for the planar case N = 00, a new very important feature arises:’ The problem of diagonalizing the one-loop dilatation operator becomes equivalent to the problem of finding the spectrum of the so-called XXX Heisenberg magnet, the famous integrable short-range spin chain. Moreover, as was conjectured in Ref. 9, the integrability extends to higher loop orders!
6
To convey the basic idea about the spin chain description of gauge theory we restrict ourselves to the simplest case of operators made of two complex scalar fields
... .
n(a:1!@2Jz+
On the space of renormalized fields {Oi} (here index i labels different orderings of scalars inside the trace) the one-loop dilatation operator acts in the non-diagonal fashion
DOi = AijOj. The mixing matrix with elements Aij encodes the Feynman diagrammatics and its eigenvalues are the scaling dimensions of primary operators. The mapping to the XXX spin chain is constructed by identifying the (a1 and $2 fields with ‘hp” and “down” spins (see Fig.1).
Figure 1. Periodic spin chain. The Hamiltonian H acts as 2 J x 2’ matrix, where J = J1 J2 is the length of the chain. We look for the eigenstates of H in a sector with the total spin fixed.
+
Upon this identification the dilatation operator becomes identical to the Hamiltonian of the spin chain. The importance of this observation is difficult to overestimate. For operators of sufficiently small dimensions the mixing matrix can be diagonalized by hand or by a computer. As its rank grows (especially in the thermodynamic limit we are interested in) this becomes rapidly unfeasible. Fortunately integrability saves the day - there is an efficient method, based on existence of local commuting charges, which allows one to determine the spectrum of the Hamiltonian (and simultaneously the one of all the commuting charges). This is so-called algebraic Bethe Ansatz (see Ref. 12 for a comprehensive review). Let us associate to each lattice cite of the periodic spin chain the following Lax operator
7
Here Sf,S3 are the standard spin operators represented by Pauli matrices and cp is a spectral parameter. Matrix L,(cp) is a 2 x 2 matrix in the auxiliary two-dimensional space. The monodromy around the chain is
Using the fundamental commutation relations between the elements of the monodromy operator one can show that the trace of T(cp) is in fact a generating function of commuting charges 5-2
,
[Qk Q m ]
=0.
k=O
Simultaneous eigenstates of Q k can be found by first defining the vacuum state vacuum state 10) (with all spins “up”)
An eigenstate of
Q k
with M = Jz spins down is given by @M = B(cpl)...B(cpM)R
provided the Bethe roots cpj tions
7
= cp(pj) obey the set of algebraic Bethe equa-
where cp(p) = ;cot(ip). The Bethe equations have a beautiful physical interpretation in terms of quasi-particles, called magnons. The state @ M describes an excitation with M magnons, each of them carrying momentum p j . Thus, the Bethe equations can be viewed as quantization condition for the magnon momenta. The total phase gained by the magnon with momentum p j traveling around the chain of length J is equal to the sum of pairwise phase shifts which arise due to its elastic scattering with other M 1magnons. The factorized scattering is one of the remarkable consequences of complete integrability of the model. Further simplification of the Bethe equations occurs in the thermodynamic limit, which amounts to sending the length of the chain J and the number of magnons M = Jz to infinity while keeping the filling fraction
8
9
a= constant. To perform this limit we present the first equation in (1) in the logarithmic form
2nnj + p j =
where the mode (winding) numbers nj define the branch of the log function. Now one can easily recognize on the r.h.s. of this equation the difference analog of derivative. Therefore, in the large J limit the Bethe equations reduce to
Further we assume that the roots c p j condense onto certain smooth contours -+ cp and that
C in the complex plane of the continuous variable cp:
y
y)
the root distribution is described by a density p(cp) = ~~~1 6 (cp normalized as J " d p p(cp) = a. Taking into account that pj -+ 1 in the %. limit we consider we see that Eqs.(2) transform into a singular integral equation
where nC are mode (winding) numbers which are constant on each smooth component of the density support C. From the point of view of soliton theory the density p(cp) describes the continuous spectrum of the model. Thus, in the thermodynamic limit the discrete eigenvalues of the monodromy matrix (the Bethe roots respectively) condense and form the continuous part of the spectrum. Having found the density one can define the generating function of all local commuting charges Qk (resolvent):
In particular, the eigenvalues of the dilatation operator coincide with the value of the second charge Q2. For the simplest ground state solution13 the one-loop gauge theory resolvent was determined in Ref. 14 and it reads as
9
Here and below K(q), E(q) and II(m2,q ) are the standard elliptic integrals of the first, the second and the third kind respectively; the parameters a and b are a = 1/4K(q), b = a/-. Expanding in cp we find explicit form of the first few charges for the ground state solution:
Q6
26 = T K ( ~ )~ 2(8 - 8q
[
+ 3q2)E(q) + (16 - 24q + 18q2 - 5 q 3 ) K ( q ) ],
The gauge theory modulus q is expressed via the filling fraction a through the following transcendental equation
To conclude, the problem of diagonalizing the one-loop dilatation operator is integrable and can be solved by using the algebraic Bethe Ansatz. The spectrum of the theory is encoded in the resolvent whose explicit form for the ground state is Eq.(3). Now we turn our attention to string theory. 3. Integrable Structure of String Theory
String theory we consider can be described by a non-linear two-dimensional sigma model whose target space is a supersymmetric extension of the bosonic Ad& x S5 space-time. Since the corresponding action is highly non-linear and contains fermions, the quantization problem appears to be extremely complicated; at present the full spectrum of the quantum string is beyond our reach. However certain regions of the quantum spectrum can be well approximated by semiclassical string configurations. Typically these configurations are solutions of the classical string equations of motion (supplemented by the Virasoro constraint), which carry “large” energy and spins. Picking one such solution one can approximately determine the string spectrum by performing a semiclassical quantization around it. Thus, we are led to the problem of studying “spinning” strings,i.e. classical strings rotating in the background space-time with large angular momenta. The bosonic sigma model describing propagation of our classical string is a two-dimensional integrable model which can be thought of as a non-trivial
10
matrix generalization of the famous sine-Gordon equation. The action is a sum of S0(2,4) and SO(6) sigma models
\,.
where
Here
X M , M = l , ..., 6 ,
Y M , M = O ,..., 5
are the the embedding coordinates of R6 with the Euclidean metric in L s and with ~ M = N (-1, +1, +I, +1, +1, -1) in LAds respectively; A and i i are the Lagrange multipliers. What are the relevant spinning string configurations? As in the flat space-time the simplest configurations are those corresponding to rigid strings, i.e. to strings whose shape is independent of time. These configurations carry finite energy and can be viewed as solitons (the so-called finite-zone solutions) of the sigma model. A remarkable fact about these solitonic solutions discovered in Ref. 15 is that they are naturally classified in terms of periodic solutions of the Neumann integrable system. This is a finite-dimensional integrable system describing a three-dimensional harmonic oscillator constrained to move on a two-sphere (or a hyperboloid in the non-compact case). Historically this model, discovered by C. Neumann in 1859, is one of the first examples of a completely integrable Hamiltonian system. To see the appearance of the Neumann system we consider the simplest case: string rotates in S5 and is trivially embedded in Ads5 as Y5 iY0 = einT with Yl,...,Y4 = 0. Ansatz for periodic motion with three non-zero angular momentum components Ji reads as:
+
x1+ ix2 = q ( ~ eiwlT, ) ~3 + i xq = z ~ ( GeiwZT ) , ~5 + i& = ~ ( 0eiwzT ) , where
11
Upon substituting this ansatz into the sigma model evolution equations they reduce to the 1-d ( ''mechanical") model (0is time now !)
which is the Neumann integrable system. If the coordinates xi(.) are complex then one gets16 another integrable system known as the NeumannRosochatius model. It is worth noting that the Neumann system inherits its integrable structure from that of the two-dimensional sigma model. Rigid strings appear to be of two types - folded, with the topology of a rod, and circular, with the topology of a circle (see Fig.2).
Figure 2. Folded and circular rigid strings (generically hyperelliptic three-spin solutions). Elliptic two-spin solutions arise when string stretches along the equator in the ("1,121 plane.
The angular momentum components (spins) are
From here one can determine the frequencies wi in terms of spins Ji: wi = Wi(J1,J 2 , J 3 ) .
The general three spin solutions are described in terms of the hyperelliptic functions. Elliptic two spin solutions arise when x3(0) = 0. Explicitly the two-spin solution corresponding to the folded string can be written in
12
terms of Jacobi elliptic functions q ( c r ) =d n ( o 6 , t)
,
22(cr) =
d?sn(cr&,
t) .
(4)
Since we are dealing with closed strings the coordinates xi(.) must obey the periodicity condition: q ( c r 27r) = zi(a). This leads t o the following equation
+
Solving for w’s in terms of spins
the modulus t is then found from
Ji,
2
4
(&) =gX.
J2
(K(t)-E(tJ2-
As was already mentioned in Sect. 1, the rigid strings provide a simple and useful tool to probe matching of the spectra of gauge and string theories. The rigid two-spin solutions are completely described by two integrals of motion which are the spins J1 and J2. However, when we embed a rigid soliton into the two-dimensional sigma model it inherits from its integrable structure an infinite set of local commuting integrals of motion, all of them appear as the non-trivial functions of J1 and 5 2 . How to compute these integrals and compare with the gauge theory charges discussed in the previous section? This problem was completely solved in Ref. 14, 17. An important technical tool which allows to determine the spectrum of commuting integrals is the so-called Backlund transformations. The Backlund transformations transform one solution, X , of the evolution equations into another one, X ( p ) , and they depend continuously on a (spectral parameter) p:
The generating function of the local commuting charges of the sigma model cx)
k=2
can be obtained from the Backlund solution as follows
+
o and = r - (T are the light-cone coordinates. It is a where [ = r very non-trivial problem to solve the Backlund equations (we refrain from presenting them here) and a general solution is unknown. However, the
13
corresponding Backlund solutions X(y) were found14 starting from either two-spin folded (4) or circular string solitons. We will not go into the further details here and refer the reader to the original The final result for an exact generating function of string commuting charges (for the folded two-spin solution) is:14
and z satisfies
Here w1,2 and t are certain functions of 51, J2. Now we are ready to compare the integrable structures of gauge and string theories. 4. Matching the Spectra of Gauge and String Theories
Resolvent (5) provides the “all-loop” result and to compare with our findings in gauge theory one needs to detect from Eq. (5) the “one-loop” contribution. To this end we consider the slightly modified string theory resolvent
where & ( p ) is determined by Eq. (5). Here the string resolvent is viewed as a function of the rescaled spectral parameter ‘p and the so-called BMN coupling constant
We will refer to p and ‘p as to the string and gauge spectral parameters respectively. Clearly, the Backlund resolvent can be expanded in a double series cn k=O
= Q(l’(‘p)
n,k=O
+
+ X 1 2 Q ( 3 ) ( ‘ p ) + O(A’3).
In particular, the leading term Q(l)(‘p) represent the “one-loop” contribution and should be directly compared to our one-loop gauge theory result W3).
14
A remarkable fact14 is that Q(l)(cp) appears precisely the same (after a certain Gauss-Landen modular transformation which relates the string to the gauge moduli) as the one-loop gauge theory resolvent H(cp) generating the Heisenberg charges: Q q P ) = H(cp). (6) Although derived by using the particular gauge/string theory solutions, this formula is universal and does not depend on particular solutions. It exhibits the matching of the spectra of gauge and string theories at the lowest order of perturbation theory. 5. Discussion In this lecture we confined ourselves to demonstrate the matching of the gauge/string spectra at the one-loop approximation. Recently this relationship between gauge and string theories received further spectacular confirmation. Of course, a question of primary importance is to understand what happens at the higher loop level. Very recently the matching of the spectra of gauge and string theories has been extended to two This became possible due to the important observation18 that the two-loop dilatation operator (and the three-loop as well!) can be emulated by the Hamiltonian of another long-range integrable spin chain known as the Inozemtsev spin chain.22 The situation with the three-loop matching is less clear at p r e s e n t l ’ ~and ~ ~ further work is required to clarify it. It is also interesting to analyze occurrence of finite-dimensional integrable systems inside the two-dimensional string sigma model. The theory of finite zone integration allows one to construct solutions of the original equations of motion starting from the stationary periodic solutions of the hierarchy of the evolution equations generated by the higher Hamiltonians. The stationary solutions themselves should be described in terms of certain finite-dimensional integrable systems (presumably generalizing the Neumann model). One could try to identify them explicitly and relate to the theory of Backlund transformations. Perhaps this would lead t o new insights into the gauge/string duality.
Acknowledgments I would like to thank Goran Djordjevic and all other organizers of the B W2003 Workshop in Vrnjacka Banja (Serbia) in August-September 2003 for an inspiring conference, and their warm hospitality.
15
References 1. J. M. Maldacena, Adw. Theor. Math. Phys. 2, 231 (1998), hep-th/9711200. 2. D. Berenstein, J. M. Maldacena and H. Nastase, JHEP 0204, 013 (2002), hepth/0202021. 3. S. Frolov and A. A. Tseytlin, Nucl. Phys. B668, 77 (2003), hep-th/0304255. 4. S. Frolov and A. A. Tseytlin, JHEP 0206, 007 (2002), hepth/0204226. 5. S. F'rolov and A. A. Tseytlin, JHEP 0307, 016 (2003), hepth/0306130. 6. S. S. Gubser, I. R. Klebanov and A. M. Polyakov, Nucl. Phys. B636, 99 (2002), hepth/0204051. 7. J. A. Minahan and K. Zarembo, JHEP 0303, 013 (2003), hep-th/0212208. 8. N. Beisert and M. Staudacher, Nucl. Phys. B670, 439 (2003), hepth/0307042. 9. N. Beisert, C. Kristjansen and M. Staudacher, Nucl. Phys. B664, 131 (2003), hepth/0303060. 10. G. Arutyunov, S. Penati, A. C. Petkou, A. Santambrogio and E. Sokatchev, Nucl. Phys. B643, 49 (2002), hepth/0206020. 11. G. Arutyunov, B. Eden, A. C. Petkou and E. Sokatchev, Nucl. Phys. B620, 380 (2002), hepth/0103230. 12. L. D. Faddeev, hep-th/9605187. 13. N. Beisert, J. A. Minahan, M. Staudacher and K. Zarembo, JHEP 0309, 010 (2003), h e p t h/0306 139. 14. G. Arutyunov and M. Staudacher, JHEP 0403,004 (2004), hep-th/0310182. 15. G. Arutyunov, S. Frolov, J. Russo and A. A. Tseytlin, Nucl. Phys. B671, 3 (2003), hepth/0307191. 16. G. Arutyunov, J. Russo and A. A. Tseytlin, hep-th/0311004. 17. G. Arutyunov and M. Staudacher, Proceedings of the 5th International Workshop on Lie Theory and Its Applications in Physics, Varna, Bulgaria, (2003), hepth/0403077. 18. D. Serban and M. Staudacher, hep-th/0401057. 19. V. A. Kazakov, A. Marshakov, J. A. Minahan and K. Zarembo, h e p th/0402207. 20. M. Kruczenski, hep-th/0311203. 21. M. Kruczenski, A. V. Ryzhov and A. A. Tseytlin, hepth/0403120. 22. V. I. Inozemtsev, Phys. Part. Nucl. 34, 166 (2003); Fiz. Elem. Chast. Atom. Yadra 34, 332 (2003), hepth/0201001.
FLUXES IN M-THEORY ON 7-MANIFOLDS: G2-, SU(3)- AND SU(2)-STRUCTURES
K. BEHRNDT Max-Planck-Institut fur Gravitationsphysik, Albert Einstein Institut A m Muhlenberg 1, 14476 Golm, Germany E-mail: behrndtQaeimpg. de
C. JESCHEK Humboldt Universitat zu Berlin, Institut fur Physik, Newtonstrasse 15, 12489 Berlin, Germany E-mail: jeschekQphysik. hu-berlin. de
We consider compactifications of M-theory on 7-manifolds in the presence of 4form fluxes, which leave at least four supercharges unbroken. Supersymmetric vacua admit G-structures and we discuss the cases of G2-, SU(3)- as well as SU(2)structures. We derive the constraints on the fluxes imposed by supersymmetry and determine the flux components that fix the resulting 4-dimensional cosmological constant (i.e. superpotential).
1. Introduction An essential input in lifting the continuous moduli space might be non-zero fluxes on the internal space. By now one can find a long list of literature about this subject. A starting point was the work by Candelas and Rainel for an un-warped metric which was generalized later in Ref. 2 (for an earlier work on warp compactification see Ref. 3) and the first examples, which preserveN = 1supersymmetry appeared in Ref. 4. The subject was revived around 10 years later by the work of Polchinski and S t r ~ m i n g e rwhere ,~ flux compactifications in type I1 string theory was considered. In the M-theory setting, different aspects are discussed in Refs. 6-14. Fluxes induces a non-trivial back reaction onto the geometry, because for the Killing they appear as specific con-torsion spinor. The resulting spaces are in general non-Kahlerian, which reflects 16
17
the fact that the moduli space is (partly) lifted. In order to see which moduli are fixed, one can deriving the corresponding superpotential as function of the fluxes in a way discussed in Ref. 23, but this approach becomes subtle if the fluxes are not related to closed forms (due to Chern-Simons terms). In this talk we discuss M-theory compactifications in the presence of 4form fluxes, which keep the external 4-d space time maximal symmetric, i.e. either flat or anti deSitter (Ads), where in the latter case the superpotential remains non-zero in the vacuum giving rise to a negative cosmological constant. We start by making the Ansatz for the metric and the 4-form field strength and separate the gravitino variation into an internal and external part. In addition, we have to make an Ansatz for the 11-d Killing spinor, which decomposes into internal 7-d spinors and the external 4-d spinors. In the most general case, the solution will be rather involved and we use G-structure to classify possible vacua (Section 3). These structures are defined by a set of invariant differential forms and are in one to one correspondence to the number of internal spinors, which will enter the 11-d Killing spinor. Using these differential forms, one can formely solve the BPS equations (Section 4), but explicit solutions require the construction of these forms. Note, the case of the G2- and SU(3)-structures have been discussed already before1°-12 and we will be rather short. 2. Warp Compactification in the Presence of Fluxes
In the (flux) vacuum, all Kaluza-Klein scalars and vectors are trivial and hence we consider as Ansatz for the metric and the 4-form field strength
+ habdy"dyb] , P = 5 epLypAdxpA dx" A d x P A d x A + $Fabcd
ds2 = e Z A[ g g d x p d z "
dy" A dyb A dy" A dyd , (1) where A = A ( y ) is a function of the coordinates of the 7-manifold with the metric h a b , m is the Freud-Rubin parameter and the 4-d metric g,$ is either flat or anti deSitter. Unbroken supersymmetry requires the existence of (at least) one Killing spinor 77 yielding a vanishing gravitino variation of 11-dimensional supergravity 0 = dQM = [ a M + 'GRS 4 M rRS where:
+&(rMF-
~ ~ F M ) ] V ,
P E F M N p Q r M N P Q , FM E F M N p Q r N P Q , etC.
Since,
(2)
18
one can bring the variation also in the more common form. Using the convention { F A , r B }= 2vAB with 77 = diag(-, +, + . . . +), we decompose the r-matrices as usual p
L
=y
with p = 0..3, a = 1..7, and .AS&,
- -E,upx^ 1
ra+3 +5
=
= i.$0.$1.$2.$3, 1 = iy1y2y3y4y5y6y7yields €abcdrnnp
ZYY -
(4)
.$58Ya,
abcd
= [a b c d ]
YUPX 7 'YrnnP=Y -7 Y Y Y * (5) 3! 3! The spinors in 11-d supergravity are Majorana and we take all 4-d .$pmatrices are real and T5 as well as the 7-d ya-matrices are purely imaginary and ant isymmet ric. With this notation, we can now split the gravitino variation into an internal and external part. In order to deal with the warp factor, we use 1 ds2 = e2A&2 -+ DM = DM -r i$%vA (6) 2 and find for the external components of the gravitino variation 1 im 1 0 = [V, 8 l F .$,.$5 8 (5 d A -) -e-3A $ ., 8 F ] v , (7) 36 144
+
+
+
7
+
v,
where F = F a b c d Y a b c d , Fa = F a b c d Y b c d , etc. and is the 4-d covariant derivative. In the same way, we get for the internal variation
1 im 1 -ya) - '"j.$'Vp 8 y a - - e-3A .$5 8 Fa]77 , (8) 2 48 4 12 where we eliminated the term N -yaFr] by using Eq. (7). In order to solve these equations, we have to decompose the spinor and introduce the superpotential yielding the negative cosmological constant. The 11-d Majorana spinor can be expanded in all independent spinors as 0 = [W @ (VP' - -aaA
+
N i=l
where and Oi denote the 4- and 7-d spinors, resp. If there are no fluxes, all of these spinors are covariantly constant and N 5 8 gives the resulting extended supersymmetries in 4 dimensions. With non-trivial fluxes one can however impose a relation between the spinors and N does not refer to the number of unbroken supersymmetries (see last Section) , but gives nevertheless a classification of supersymmetric vacua. In fact, with these spinors one can build differential forms that are singlets under a subgroup G c spin(7) and hence define a G-structure, where the number of spinors is directly related to the group G (see next Section). By definition, the spinors
19
are singlets under G and therefore obey certain projector conditions, which annihilate all non-singlet components and, at the same time, can be used to derive simple differential equations for the spinors and constraints on the fluxes (see last Section). If the 4-d spinors are covariantly constant, the resulting vacuum will be a 4-d flat space, but for an anti deSitter vacuum the spinors satisfy
V,€i
N
-i;(W,ij + i T 5 W?) €j .
(9)
Note, the resulting 4-d cosmological constant will be: -IWI2 and we did not take into account a Kahler potential, i.e., our superpotential will not be holomorphic. If there is only a single spinor this equation simplifies to
V,E
N
9, (Wl + i=y5W2)€ ,
and if E is a Weyl spinor it becomes V,E = +,eKI2 WE*with the complex superpotential W = W1 + i W2. If ~i are a set of Weyl spinors, we introduce the superpotential by a 11-d spinor satisfying the equation
[ V , @ F ] q = (T,@W)ij
with:
fj=Wij~i@f?ej*+cc.
(10)
This way of introducing the superpotential might be confusing. Recall, we set constant all 4-d scalars as well as vector potentials and hence the superpotential should just be a number fixing the cosmological constant for the given vacuum. Since we introduced the superpotential in the 11-d Killing spinor equation it will, on the other hand, depend on the fluxes and the warp factor and thus it is in general not constant over the internal space. The correct 4-dimensional superpotential is of course obtained only after a Kaluza-Klein reduction, i.e. after an integration over the internal space and to make this clear we will denote this constant superpotential by W(O).We do not want to discuss issues related to a concrete Kaluza-Klein reduction (over a not Ricci-flat internal space) and want instead determine the flux components that are responsible for a non-zero value of W(O) 3. G-Structures Supersymmetric compactifications on 7-manifolds imply the existence of differential forms, which are singlets under a group G c spin(7) and which define G-structures.a These globally defined differential forms can be constructed as bi-linears of the internal Killing spinors as f?i-Yq...a,f?j,
follow here basically the procedure initiated in the recent discussion by Ref. 17
20
and the group G is fixed by the number of independent spinors 8i which are all singlets under G. E.g. if there is only a single spinor on the 7manifold, it can be chosen as a real G2 singlet; if there are two spinors, one can combine them into a complex SU(3) singlet; three spinors can be written as Sp(2) N SO(5) singlets and four spinors as SU(2) singlets. Of course, all eight spinors cannot be a singlet of a non-trivial subgroup of SO(7) and G is trivial. The 7-dimensional y-matrices are in the Majorana representation and satisfy the relation: ( Y ~ ~ . . . ~=, (-) ) ~ 2 T ~ ~ . . . ~which ,, implies that the differential form is antisymmetric in [i,j ] if n = 1,2,5,6 and otherwise symmetric [we assumed here of course that Bi are commuting spinors and the external spinors are anti-commuting]. This gives the wellknown statement that having only a single spinor, one cannot build a vector or a 2-form, but only a 3-form and its dual 4-form [the 0- and 7-form exist trivially on any spin manifold]. If we have two spinors 6'11p), we can build one vector and one 2-form (and of course its dual 5- and 6-form). Since the spinors are globally well-defined, also the vector field is well defined on X, and it can be used to obtain a foliation of the 7-d space by a 6manifold X 6 . Similarly, having three 7-spinors we can build three vector fields as well as three 2-forms and having four spinors the counting yields six vectors combined with six 2-forms. In addition to these vector fields and 2-forms, one obtains further 3-forms the symmetrized combination of the fermionic bi-linears. We have however to keep in mind, that all these forms are not independent, since Fierz re-arrangements yield relations between the different forms, see Refs. 9, 17 for more details. Using complex notation, we can introduce the following two sets of bilinears [et = (e*)T]:
Ra ,... a k = Oty,,...,,8
and
= 8T
yal...ak8,
where dropped the index i , j which counts the spinors. The associated k-forms becomes now
ak = k!1
and
- ~ ~ ~ . . . ~ ~ e ~ l . " ~ k
1 fik = -f22,1...akea1"'ak. k!
(11)
If the spinors are covariantly constant the group G coincides with the holonomy of the manifold. If the spinors are not covariantly constant, then neither can be these differential forms and the deviation of G from the holonomy group is measured by the intrinsic torsion. In the following we will discuss the different cases in more detail.
21
3.1. G z Strtdures In the simplest case, the Killing spinor is a G2 singlet and reads O = e z6 0 ,
Or
(12)
where is a normalized real spinor. Due to the properties of the 7-d y-matrices (yielding especially OFyaOo = 0), only the following differential forms are non-zero
1 = O,TOo
,
T i f a b c d m n p = 60 TabcdmnpOO .
They are G2-invariant since 00 is a G2 singlet, i.e. it obeys the appropriate projector constraints. Note, the Lie algebra 50(7) is isomorphic to A' and a reduction of the structure group on a general X7 from SO(7) to the subgroup G2 implies the following splitting: SO(7) =
82
@ 8;
.
(14)
This induces a decomposition of the space of 2-forms in the following irreducible G2-modules, R2 = A;
(15)
@
where
A; = {
T X T }= { a E A'I
U J ~ ~E U
Af4 = {Q E A2 I * ('p A-Q) + a = 0)
* ( ~ A - Q-)2 a = 0 } , 82,
with the abbreviation u 1'p = umpmnpand 'p denotes the G2-invariant 3index tensor, which is expressed as fermionic bi-linear in (13). The operator *(pA a ) splits the 2-forms correspondingly to the eigenvalues 2 and -1. These relations serve us to define the orthogonal projections P , onto the k-dimensional spaces:
where .II, = *p. To be concrete, the Gz-singlet spinor satisfies the condition
22
which is equivalent to
,
Tab80 = @abcyc80 Tabcf% = ('@abc TabcdeO =
+ $abcdT d)
(18)
60
( - $abed - 4ip[abcTd])eO 7
where the second and third conditions follow from the first one. These relations can now be used to re-cast the Killing spinor equations into constraints for the fluxes and differential equations for the warp factor as well as the spinor 8. In the generic situation this spinor is not covariantly constant, which reflects the fact that fluxes deform the geometry by the gravitational back reaction. This can be made explicit by rewriting the flux terms as con-torsion termsb v,e = (v, - -41T , ~ " T ~ ~=) oo . Fkom the symmetry it follows that T has 7 x 21 = 7 x (7+ 14) components, but if €J is a G2-singlet the 14 drops out and hence T E A1 @ gi. These components decompose under G2 as
49 = 1 -t- 7
+ 14 + 27 = + + + 7 2 7 , 77
71
714
where ~i are called G2-structures. Since the Killing spinors define p and $, these torsion classes can be obtained from d p and d$ as follows
d p E A4 = A; @ A$ @
d$ E R5 = A$
@
Ai7 , (19)
A:, ,
where the 7 in A$ is the same as in A: up to a multiple. For a general 4-form 0, the different projections are
R(P) =
P
7
W P ) = -&-l P ,
(20)
p27(P)ab = $ ( P c d e { a $ b } C d e ) O
7
where in (-)O we removed the trace. Thus, the different components in the differentials d p can be obtained from dl)
--
where 714 and
1d p
*d$ 727
-
--
,
$(*d$) J ?I,
d7) ,
have to satisfy:
7(27)
93 A
p-1dp,
(dpcde{a$b}cde)O
(21) 7
Az7 = 9 3 A 714 = 0.
bThere is also an ongoing discussion in the mathematical literature, see Ref. 24.
23
3.2. SU(3) Structures Having a G = SU(3), one can find two singlet spinors on X 7 , which are equivalent to the existence of a vector field v. This in turn can be used to combine both spinors into one complex spinor defined as
where the constant spinor 00 is again the Gz singlet and Z is now a complex function. The vector w is globally well-defined and gives a foliation of X7 by a 6-manifold X S and both spinors, 6 and its complex conjugate 8*, are chiral spinors on x6. In this case, we have to distinguish between the forms R and fi as defined in (11) and findlo-12
and all other forms are zero or dual to these ones. The associated 2-form to the almost complex structure on x6 is w and with the projectors (W f i w ) we can introduce (anti) holomorphic indices so that R(370)can be identified as the holomorphic (3,0)-form on XG. There exists a topological reduction from a Gz-structure to a SU(3)-structure (even to a SU(2)-structure). The difficulties arise by formulating the geometrical reduction. Using the vector v the explicit embedding of the given SU(3)-structure in the Gz-structure is:
i
with the compatibility relations e-ZiIm(Z) o ( 3 , O ) A w =
(x+ + i X - ) A w
=0,
24
Now, the projectors (18) for
80
imply for the complex 7-d in ( 2 2 )
+ v a + i'fabcvb'YC)eO , T a b 6 = 5 ( i ( p a b c y c + i(pabcVc + '$abcdvcYd 2 v [ a Y b ] ) e O, 'Yabce = $ ( i ( p a b c + '$abcdYd + 3 i v [ a ( p b c ] d Yd - '$abcdvd 4'&'[abcYd]Vd)eO , 'Yabcde = $(-'$abed - 4'@[abcYd] - 5'$[abcd'Ye]ve 'Yae
=
ez
z ( Y a
-
-
'Yabcdee Yabcdefe
=
5
-4iv[a(pbcd] (-5'$,[abcd'Ye]
= L(-.
ZEabcdefgY'
- 4v[a'$bcd]eYe)e0 - i&abcdefgYgvf
7
- 5v[a'$bcde]
+ EabcdefgvhYj(pghi
- 2 o i v [ a ( p b c d ' Y e ] ) ~ ,O
- i & a b c d e f g v g )00
*
Again, these relations can be used to rewrite the Killing spinor equations in terms of constraint equations- for the fluxes and a differential equation for the warp factor as well as the spinor. The corresponding torsion componentsz5 are now related to the differential equation obeyed by the forms: v,w , R and their dual. As next case one would consider S P ( 2 ) structures implying three (real) singlet spinors. An example is a 7-d 3-Sasaki-space (i.e. the cone yields an 8-d Hyperkahler space with Sp(2) holonomy), with the Aloff-Walach space N1il as the only regular examplesz6 (apart from S7);non-regular examples are in Ref. 27. We leave a detailed discussion of this case for the future and investigate instead the S U ( 2 ) case in more detail.
3.3. SU(2) Structures On any 7-d spin manifold exist three no-where vanishing vector fieldsz8, which implies that one can always define SU(2) structures. The corresponding four (real) spinors can be combined in two complex S U ( 2 ) singlet spinors O 1 p . The three vector fields v,, a = 1 , 2 , 3 can be chosen as Vl = e
1 1
212
= e
2
,
'u3
= (p(Ul,V2),
and they parameterize a fibration over a 4-d base space X,.The embedding of the S U ( 2 ) into the Ga structures is then given by
Since the vector fields are no-where vanishing, we can choose them of unit norm and perpendicular to each other, i.e. ( v , , v ~ ) = hap, and using the 3-form (p, one obtains a cross product of these vectors. One can pick one
25
of these vectors, say v3, to define a foliation by a 6-manifold and on this 6manifold one can introduce an almost complex structure by J = 213 -I ‘p E T * M 6 8 T M 6 . The remaining two vectors, which can be combined into a holomorphic vector‘ v1 iv2 imply that this 6-manifold is a fibration over the base &. On this 4-manifold we can define a basis of anti-selfdual 2-forms whose pullback correspond to the w,. Note, on any general 4-d manifold we have the splitting
+
A~=A:wP., where we can take { w l ,w2, w3) as a basis of A: and this splitting appears in group theory as: 50(4) 2 su(2) @ su(2), which is equivalent to reducing the structure group from SO(4) to SU(2). The 2-forms satisfy the algebraic relations wi2 =2vo1q
wiAwj=O
for,i#j,
and the associating complex structures fulfill the quaternionic algebra (note: the orientation on the 4-fold is negative). We can further split the 2-form bundle into a symplectic 2-form, say w = w3, and the remaining can be combined into complex (2,O)-form. Thus, the subbundle AT decomposes as
A: r X 2 > 0 ~ R ~ . So, besides the symplectic form w , let us introduce the complexified 2-form: x=w1+iw2
which is, with respect to w , a holomorphic (2,O)-form. The SU(2) singlet spinors can again be constructed from the G2 singlet spinor 00 by (29) where v, = vF3;n. Using the expressions from before, it is straightforward to verify the following relations
(v1v2 - iv3)e0= o
,
(vl - iv2)e2= (vl
+ iv2)e1= 0 ,
which imply v,(gff)klel
=ek
,
wek = i e k ,
v,vpek = i € , p x ( a x ) k l e l ,
xek = - i ( a 2 ) k 1 8 ; ,
“Meaning, that it annihilated by the projector: (W - J )
26
where w
= wmnymn, X G Xmnymn and with the Pauli matrices u 3 = ( l0 -1O ) .
(31)
4. BPS Constraints Now we can come back to the BPS equations from Section 2. With the superpotential as introduced before, equation (7) becomes 1 o = f j + [9518( 2-1d A + -)i36m -e-3A ( F IF~) ] q , (33) 144
+
and if: ij = e-+q, equation (8) yields im Ya)ij - i 9 5 y ae - 4 6 o = w 18(vih) +48
-
- e-3A y- 5 B ~ ~ i j (34) . 12
It is useful to decompose the 35 components of the 4-form field strength under Gz as 35 + 1 7 27 with
+ +
where 3(l), F(7)and 3(27) are the projection introduced in Eq. (20). The cases of G2 and SU(3) structures have been discussed already in the literature and we will summarize only the main results.
4.1. G2 Structure In this case, the 11-d spinor is a direct product, i.e. q=E@e,
(36)
and since the 11- and 7-d spinor are Majorana also the 4-d spinor E has to be Majorana (a more detailed discussion is given in Ref. 29). One finds that all internal 4-form components have to vanish
Fabcd = 0 ,
w1 = 0
, m = -36W2.
(37)
27
The Eq. (34) gives a differential equations for the spinor e Z & , which implies
aaz= 0 . The differential equations for 00 fixes the 7-manifold to have a weak G2 holonomy and hence is a Einstein space with the cosmological constant given by the Freud-Rubin ~ a r a m e t e r . ~ ~This > l Oin turn implies, that the 8d space built as a cone over this 7-manifold has Spin(7)holonomy. In fact, after taking into account the vielbeine, this gives the known set of first order 7 differential equations for the spin connection 1-form wab: Wab$9abc = 36 m e b , where m was the Freud-Rubin parameter [note w is here the spin connection and should not be confused with the associated 2-form introduced before]. Using the differential equation for the 7-spinor, it is straightforward to verify that d$=O,
and therefore only dl)is non-zero. The 4-d superpotential is only given by the Freud-Rubin parameter, ie.
which fixes the overall size of the 7-manifold. In the limit of flat 4-d Minkowski vacuum, the Freud-Rubin parameter has to vanish and we get back to the Ricci-flat G2-holonomy manifold. In order to allow for nontrivial fluxes one has to consider SU(3) instead of G2 structures. 4.2. SU(3) Structure
In this case, there is one (complex) 7-d spinor and the 11-d Majorana spinor reads = E 86
+ E*
80*.
(39)
where the 4-d spinors E and E* have opposite chirality (y5e = E ) . More details about this case can be found in Refs. 1 0 , l l . The solution of Eq. (33) read now
w = w1+ i w2= l6e - ( K / 2 + 3 A ) +
[$ 4
(1)
V a a a e 3 ~= + ~ ( 1 ) v a ~ ab ( 2 7 ) v 1b
m=O,
- vaF:;7)vb
+ i v " F p ], (40)
28
and
[the flux components were introduced in (35)]. In addition, one obtains a differential equation for the spinor with the non-trivial torsion components as introduced in Eq. (19)
--
,
dl)
w 2
7A7)
48 w1
-
9 3(l) + $ (PabcvbFi7)+ 27.F;i7)vb . 21,
(43)
To make the set of equations complete, we have to give the differential equations obeyed by the vector field u,which is straightforward if we use the differential equation for the spinor
recall w& = 'p&Uc. Note, unVmun = 0, which is consistent with 1uI2 = 1. Using the decomposition (35) one finds
v [ m u n ]= (drZ64 2 7 1(dd;:
V{,Vn) = --(dmn
+Z
1 + -21' $ " m n a b ) ~ ~ ~ 7 ) ' U C +' U b-'pmn"(d," 4
- V,Vb);Fb(7) ,(45)
1 9) - Zu{m'pn}abu,F~7)
- u(mun})
+ w;w:).F;;7)
- 2dmnF:~7)u"vb. 1
(46)
The first term in the anti-symmetric part is the projector onto the 7,see Eq. (16), and by contracting with 'p and employing Eqs. (41) and (42), one can verify that:" d(e3Av)= 0. One can project the flux components onto X s and using the symplectic 2-form w we can introduce (anti) holomorphic indices. As result, we can define a 3-form H and 4-form G on X S and find for the superpotential
whereas the 4-form has to fulfill the constraint: R 1G = 0 and de3A1w = ~ w ~ H a s w e l l a s u ~ d e ~ ~ = ~ ~ ~ ~ G .
29
4.3. SU(2) Structure Finally, in the SU(2) case we write the 11-d spinor as
7=
Be1 + e 2 B e 2 + c C ,
and we choose chiral4-d spinors with
y~5 E i
'
=Ea.
Eq. (33) gives
1 im 1 W&j* + ( - d A + - + - F ) O j ] . 2 36 144
(49)
If one dos not impose any constraints on the spinors 8 , one finds14
wij with the 4-form
-
eiFej = F
-I
6(4)as derived in (32). Defining the 2-forms:
Gap = V ~ V ~ F m n a b X a b ,
F,p = V z V ; F m n a b W a b ,
we can write Wij as matrix: W ( ~ ~ P Y G , p o ~ ) owith 2 the 0, as Pauli matrices. It would be identical zero if G = 0, but instead we can also impose: ~ ' W i j= 0 so that Wij projects out one of the 4-d spinor as we would need for an N = 1 vacuum. This implies that: det W = 0 which gives one constraint on the complex 2-form G. As next step, the contraction with 6: yields N
m =0,
-
~ A - S2(') I
N
F 1d4),
which implies that: d,A e a p r F p T (with 8, = v,"d,). Finally, one has to contract with Bya as well as with 8tya (with the index a projected onto the base) and if we assume that the dbA = 0 (ie. the warp factor is constant over the 4-d base), we get as further contraints on the fluxes =0,
eyaFe = 0 .
-
These constraints are solved, e.g., if the only non-zero components of the are: v, Avp A w ; ie. are contained in Fap and Gap = 0 (as defined above). These are all constraints on the fluxes, but from the internal variation (34) we get differential equations. Setting, m = 0 and ij = 0, we find
V,&
N
FmnpqynPqei.
If only the components in Fap are non-zero, it is straightforward to further simplify this equation by using the relations in (30). On the other hand,
30
this equation fkes also the corresponding differential equations obeyed by the differential forms.
For the 2-forms eg., our constraints on the fluxes imply that w and X are closed, when projected onto the 4-d base, which is therefore a hyper Kahler space. Unfortunately, we have to leave a detailed analysis of these equaions for the future.
Acknowledgments K. B. is partially supported by a Heisenberg grant of the DFG. C . J. is supported by a Graduiertenkolleg grant of the DFG (The Standard Model of Particle Physics - structure, precision tests and extensions).
References P. Candela and D. J. Raine, Nucl. Phys. B248,415 (1984). B. de Wit, D. J. Smit, and N. D. Hari Dass, Nucl. Phys. B283,165 (1987). B. de Wit and H. Nicolai, Phys. Lett. B148,60 (1984). M. A. Awada, M. J. Duff, and C. N. Pope, Phys. Rev. Lett. 50,294 (1983); B. de Wit, H. Nicolai, and N. P. Warner, Nucl. Phys. B255,29 (1985). 5. J. Polchinski and A. Strominger, Phys. Lett. B388, 736 (1996), hep-
1. 2. 3. 4.
th/9510227. 6. B. Brinne, A. Fayyazuddin, T. Z. Husain, and D. J. Smith, JHEP 03, 052 (2001), hepth/0012194. 7. K. Dasgupta, G. Rajesh and S. Sethi, JHEP 9908, 023 (1999), hepth/9908088; B. S. Acharya and B. Spence, hep-th/0007213. K. Becker and M. Becker, JHEP 11, 029 (2000), hep-th/0010282. 8. P. Kaste, R. Minasian, M. Petrini, and A. Tomasiello, JHEP 09,033 (2002), hep-t h/0206213. 9. J. P. Gauntlett and S. Pakis, JHEP 04, 039 (2003), hep-th/0212008; D. Martelli and J. Sparks, Phys. Rev. D68,085014 (2003), hep-th/0306225; J. P. Gauntlett, D. Martelli, and D. Waldram, Phys. Rev. D69, 086002 (2004), hep-t h/0302158. 10. K. Behrndt and C. Jeschek, JHEP 04, 002 (2003), hep-th/0302047; K. Behrndt and C. Jeschek, hep-th/03llll9. 11. P. Kaste, R. Minasian, and A. Tomasiello, JHEP 07, 004 (2003), hepth/0303127. 12. G. Dall’Agata and N. Prezas, Phys. Rev. D69, 066004 (2004), hepth/0311146. 13. K. Behrndt and M. Cvetic, Nucl. Phys. B676, 149 (2004), hep-th/0308045. K. Behrndt and M. Cvetic, hep-th/0403049.
31
14. K. Behrndt and C. Jeschek, Class. Quant. Gruv. 21, S1533 (2004), hepth/0401019. 15. A. Strominger, Nucl. Phys. B274,253 (1986). 16. K. Becker, M. Becker, K. Dasgupta and P. S. Green, JHEPO304, 007 (2003), hep-th/0301161; K. Becker, M. Becker, K. Dasgupta and S. Prokushkin, Nucl. Phys. B666,144 (2003), hep-th/0304001; K. Becker, M. Becker, P. S. Green, K. Dasgupta, and E. Sharpe, hep-th/0310058. 17. 3. P. Gauntlett, D. Martelli, S. Pakis, and D. Waldram, Commun. Math. Phys. 247,421 (2004), hep-th/0205050. 18. T. Friedrich and S. Ivanov, math-dg/0112201. 19. J. Louis and A. Micu, Nucl. Phys. B626,26 (2002), hep-th/0110187; S. Gurrieri, J. Louis, A. Micu, and D. Waldram, Nucl. Phys. B654, 61 (2003), hep-th/O211102. 20. G. L. Cardoso et al., Nucl. Phys. B652, 5 (2003), hep-th/0211118; G. L. Cardoso, G. Curio, G. Dall’Agata, and D. Lust, JHEP 10,004 (2003), hepth/0306088. 21. G. Dall’Agata, hep-th/0403220. 22. A. R. Frey, hep-th/0404107. 23. S. Gukov, Nucl. Phys. B574,169 (2000), hep-th/9911011; C. Beasley and E. Witten, JHEP 07,046 (2002), hep-th/0203061; G. Curio, JHEP 03,024 (2003), hepth/0212211. 24. T. Friedrich, I. Kath, A. Moroianu, and U. Semmelmann, Journal of Geometry and Physics 2 3 , 4 1 (1997); T. Friedrich and S. Ivanov, [math-dg/0102142]; I. Agricola and T. Friedrich, Class. Quant. Graw. 20,4707 (2003), [mathdg/0307360]; S. Karigiannis, math-dg/0301218. 25. S. Chiossi and S. Salamon, math-dq/0202282. 26. T. Friedrich and I. Kath, Commun. Math. Phys. 133,543 (1990). 27. K. G. C.P. Boyer and B. Mann, Journ. Reine u. Angew. Math. 455, 183 (1994). 28. E. Thomas, Bull. Americ. Math SOC.75,643 (1969). 29. A. Bilal, J.-P. Derendinger, and K. Sfetsos, Nucl. Phys. B628, 112 (2002), hep-th/Oll1274.
NONCOMMUTATIVE QUANTUM FIELD THEORY: REVIEW AND ITS LATEST ACHIEVEMENTS
M. CHAICHIAN
P.
Department of Physical Sciences, and Helsinki Institute of Physics, 0. Box 64, 00014 University of Helsinki, Helsinki, Finland E-mail:
[email protected]
Some properties of quantum field theories on noncommutative space-time are reviewed. Studying the general structure of the noncommutative (NC) local groups, we present a no-go theorem for NC gauge theories. This no-go theorem imposes strong restrictions on the NC version of the Standard Model (SM) and in resolving the standing problem of charge quantization in noncommutative QED. We also consider the phenomenological implications of noncommutative y on the spectrum of the H-atom and derive a bound on the noncommutativity parameter 8. Finally, in the framework of noncommutative quantum field theories (NC QFT), we show the general validity of the CPT and spin-statistics theorems, with the exception of some peculiar situations in the latter case.
1. Introduction It is generally believed that the notion of space-time as a continuous manifold should break down at very short distances of the order of the Planck length Xp M 1.6 x 10-33cm. This would arise, e.g. from the process of measurement of space-time points based on quantum mechanics and gravity arguments. Arguments for noncommutativity arise also from string theory with a constant antisymmetric background field, whose low-energy limit, in some cases, turns up to be a noncommutative quantum field theory (NC QFT).2 This in turn implies that our classical geometrical concepts may not be well suited for the description of physical phenomena at very small distances. One such direction is to try to formulate physics on some noncommutative space-time. 1-3 If the concepts of noncommutative geometry are used, the notion of point as elementary geometrical entity is lost and one first expectation is that an ultraviolet cutt-off appears. In Ref. 4 this expectation was shown not to be fulfilled in general. Instead, a peculiar 32
33
UV/IR mixing appear^.^ The usual way of constructing a noncommutative theory is through the Weyl-Moyal correspondence: in a NC space-time the coordinate operators satisfy the commutation relation:
where 8’”” is a constant antisymmetric matrix of dimension (length)2. In QFT the operator character of the space-time coordinates (1) requires that the product of any two field operators be replaced by their *-product, or Weyl-Moyal product. The *-product compatible with the associativity of field products is given by:
An important step in constructing a physical noncommutative model is t o develop the concept of local gauge symmetry. Intuitively, because of the inherent nonlocality of noncommutative field theories, the notion of local symmetry in the noncommutative case should be handled with special care. As a result, the pure noncommutative U(1) theory behaves similarly to the usual non-Abelian gauge theories, but now the structure constants depend on the momenta of the fields.6 This feature induces a charge quantization problem,’ in the sense that the electric charges in the noncommutative quantum electrodynamics (NC QED) based on NC U(l) group are quantized only to f 1 , O . The solution of this problem was sought in the construction of a noncommutative version of the Standard Model (NC SM),8 based on a no-go t h e ~ r e m and , ~ is discussed in Section 11. In Section 111 phenomenological implications of the noncommutativity are also addressed on a concrete model of the H-atom, for which we present the noncommutative corrections to the spectrum and, using the data for the Lamb shift, we find a bound on the noncommutativity parameter 0.” In Section IV, we show that a breaking of the spin-statistics relation in NC QFT could occur only in the case of theories with NC time. We also present in Section V a general proof that the CPT theorem remains valid in NC field theories, for general form of noncommutativity, although the individual symmetries C,T and P are broken.”
34
2. Noncommutative Gauge Groups. A No-go Theorem 2.1. Charge Quantization Problem in NC QED In Ref. 7 it was shown that in NC QED based on the NC U(l) group, one can encounter only fields with charge +1:
$74= V ( X ) * $(). 9 Dp$ = a,$ - iA, * 4,
$(XI
+
(3)
fields with charge -1: $(XI
+
$'(.I D,$
= $(XI = a,$
*u - w
+ i$
1
* A,,
(4)
and fields with charge 0:
Xb)
-+
X'(X)
=W
X:)
*X(X) *U-W ,
D,x = a,x
+ i[x,A,]* .
(5)
This immediately raises the question about other known charges, i.e. the fractional charges of quarks. The simple extension
D p ~ ( n=)
- inA,
* $(n) ,
(6)
with $(4--+ $44 = U*"* $'"'
(7)
for the field $ with integral multiple n of a (conventional) unit charge fails to transform covariantly. In conclusion in NC QED, charge is quantized only to 0,fl. A possible way out from this situation is to construct a NC version of the Standard Model, to which end we have to choose the gauge groups and their representations and also define the direct product of group factors. 2.2. A No-go Theorem The following result was partially obtained in Ref. 12 in the framework of noncommutative gauge groups and extended to a no-go theorem in Ref. 9. In general, as discussed in Ref. 13, it is not trivial to define the noncommutative version of usual simple local groups, as the *-product will destroy the closure condition. Consequently, the only group which admits a minimal noncommutative extension is U ( n ) (we will denote its extention by U*(n)). However, the NCSO and U S p algebras have been constructed in a more involved way. l3
35
To define the pure NC U * ( n ) gauge theory we take as generators of the u*(n) algebra: T”, a = ,n2 - 1 ( n x n su(n) generators) and T o = -&lnxn. The u*(n) Lie-algebra is defined with the star-matrix bracket : 1 , e . a
[f,9l* = f
*9
-
9 *f
,
f , 9 E u*(n>.
(8)
The U,(n) gauge theory is described by the u,(n) valued gauge fields: n2-1
G, =
C
G$(x)TA,
(9)
A=O
with the infinitesimal gauge transformation
G,
4
GG = G,
+ i8,X + g[X,G,],
, X E u*(n).
(10)
Under the above tranformation, the field strength G,u = d[,Gv]
+ i9[G,, GI*,
(11)
transforms covariantly:
leaving invariant the action of the pure U*(n) gauge theory:
4n
dDa: n(G,, *G’””)
One peculiar feature to be noticed in the case of the pure U*(n)gauge theory is that, fixing the number of gauge field degrees of freedom (which is n2) the dimension of the matrix representation is automatically fixed, i.e. the gauge fields must be in the n x n matrix form. The main physical implication is that the matter fields coupled to the U * ( n ) gauge theory can only be in fundamental, antifundamental, adjoint and singlet states. Another nontrivial point in the noncommutative gauge theories is to define the direct product of NC gauge groups. In the commutative case, if GI and G2 are gauge groups, then G = G1 x GZ is defined through:
9 = 91 x 92 19) = 9‘1 x 9: 1 g i ,9: E Gi , 9 , g t E G , 9 . 9 ) = (91 x 92) * (9; x 9;) = (91 .9;) x (92 * 9:) *
(14)
In the noncommutative case, let G1 = U*(n) and G2 = U*(m). But now, the group products involve the *-product so that the group elements can not be re-arranged. As a result, the definition of direct product cannot be straightforwardly generalized to the NC case and consequently the matter fields cannot be in fundamental representations of both U,(n) and U*(m).
36
The only possibility left is for a matter field to be in the fundamental representation of a gauge group and the antifundamental representation of another:
9 -+ 9' = u 9 * V - l ,
U E U,(n), V E ~,(rn).
(15)
In the general case of n gauge groups N
the matterfields can be charged under at most two of the U,(ni) factors. 2.3. N C Standard Model. A Solution to the Charge Quantization Problem Based on the above no-go theorem, we have built a noncommutative version of the Standard Model.8 The model is based on the gauge group U,(3)x U,(2) x U,(1)(the general elements of the respective group factors will be denoted by U E U,(3), V E U,(2), II E U,(l)) and comprises: one gauge field, B, , valued in u,(1), four gauge fields of u,(2) :
where d ,i = 1,2,3 are the Pauli matrices and fields of u,(3):
go = 1 z X 2
and nine gauge
R
A=O
where T", a = 1,2,. . , 8 are the Gell-Mann matrices and To = 1 3 x 3 . This choice of the gauge group is due to the fact that there i s no straightforward noncommutative extension of the S U ( n ) groups. However, compared t o the commutative Standard Model, two additional gauge fields have appeared, corresponding to the extra U(1) factors. The reduction of the extra U(1) factors is achieved through a Higgs-type of mechanism, in two stages. First the mechanism is run with the symmetry-reducing scalar field with U1 E U(1)C U,(3)and V1 E U(1)c U,(2).In the second stage, the symmetry is reduced eventually to that of hyper-charge, throught the scalar particle @z(.)
+
s(.)@2.-l()
1
(20)
37
with sEU(l)residual and v ~ U * ( l ) After . the symmetry reduction, two of the gauge fields become massive (Go and W o )and the gauge field corresponding to the residual U ( 1 ) symmetry will be the (masless) hyper-photon Y . When coupling matter fields to the U*(3) x U*(2) x U,(1) theory, we have to keep in mind that, according to Ref. 9, the fields can be only in the fundamental and/or anti-fundamental representation of the group factors. It is interesting to note that the no-go theorem allows six different types of charged particles in the case of three simple group factors and the matter content of the original Standard Model (including the Higgs particle) exhausts those possible types of charges. By properly taking the representations of the matter fields and performing the U(1) symmetry reduction introduced earlier, it is straightforward to show that the couplings of all matter fields to the hyper-photon Y, are realized through the usual hypercharges.' Moreover, after performing the spontaneous symmetry breaking of the original Standard Model, all particles will couple to the photon A, through the usual electric charges, i.e. 1, -1, 0, -I/?, 2/3, so this model provides a solution to the NC charge quantization problem. Another proposal for a noncommutative version of the Standard Model is based on the Seiberg-Witten (SW) map,14 which assigns to commutative gauge configurations the noncommutative equivalent configurations, linked by field-dependent noncommutative gauge transformations. This version of the NC SM is constructed from NC fields realized by SW map as a tower of commutative fields, transforming under G = U(1) x S U ( 2 ) x SU(3). There are no additional U(1) gauge fields, so there is no need for the U(1) factor reduction. The gauge symmetry is considered on Lie algebra level and not Lie group level. Consequently, arbitrary (fractional) U (1) charges are admissible. However, this last point can be considered as a disadvantage: in the NC SM based on the no-go theorem, the U(1) factor reduction fixes the correct (hyper) charges for all SM particles.
3. Lamb Shift in NC QED. Bounds on 8 In this section we focus on the hydrogen atom and, using the non-relativistic limit of NC QED results, we propose the Hamiltonian describing the NC H-atom. Given the Hamiltonian and assuming that the noncommutativity parameter ( & j ) is small, we study the spectrum of H-atom. We show that because of noncommutativity, even at field theory tree level, we have some corrections to the Lamb shift (2+, -+ 2&/2 transition)." Hereafter, we shall consider the electron of the H-atom moving in the
38
external field of the proton. However, similar results (up to a numerical factor) would be obtained by treating the proton as a composite particle, e.g., in the naive quark m0de1.l~The latter analysis infirms the treatment of Ref. 16, where the proton is taken as an elementary particle, thereby obtaining no noncommutative corrections for the H-atom spectrum at tree level. To start with, we propose the following Hamiltonian for the noncommutative H-atom:
with
p i ,3y = ieij , pi,fijl = itisij , bi16j] =0.
(22)
The NC Coulomb potential
Ze2 e Zer V ( r )= --r - -(e 4ti x P) . (-$
+ o(e2),
with e i j = i f i j k e k can be obtained either as the nonrelativistic limit from the NC photon exchange diagram or from the change of variables:
xi = xi+ -eijfij, 2ti
(24)
Pi = Pi 7
where the new variables satisfy the usual canonical commutation relations: [Zi,Zj]
=0
,
[Pi,Pj] =0
,
(25)
[Zi,Pj] =m i j .
Using the usual perturbation theory, the leading corrections t o the energy levels due to noncommutativity, i.e. first order perturbation and in field theory tree level, are:
1 f o r j = l f i a n d f n , l = n34+3)(1+1). The case of our interest, the 2P112--f 2S1p transition (Lamb shift), for the noncommutative H-atom, besides the usual loop effects, depends on the j , quantum number (only for the 2P112 level, as the levels with 1 = 0 are not affected) and is there, even in the field theory tree level. Hence we call it polarized Lamb shift. New transition channels are opened (notation nlp), i.e. 2PG;l2 4 2P://2"and a split of
the usual Lamb shift occurs: 2Pt;:
+ 2S1p
and 2Pz;l2 --$
25'112.
39
One can use the data on the Lamb shift to impose some bounds on the value of the noncommutativity parameter 8. Of course, to do it, we only need to consider the classical (tree level) results, Eq. (26). Comparing these results, the contribution of (26) should be of the order of smaller than the usual one loop result and hence,
The same bound is obtained also from the violation of Lorentz invariance, based on the clock-comparison experiments, which monitor the difference between two atomic hyperfine or Zeeman transition frequencies, searching for variations as the Earth rotates.17
4. Noncommutative Quantum Field Theory and Spin-statistics Theorem Pauli’s spin-statistics relationls is responsible for the entire structure of the matter and for its stability. Experimentally, the relation has been verified to high accuracy. Theoretically up to now there has been no compelling argument or logical motivation for its breaking. However, the violation of Lorentz invariance, as well as the intrinsic nonlocality of noncommutative field theories, may suggest that the (presumably very small, of the order of lPVlm2)breaking of this fundamental theorem, as well as of the CPT theorem, might be possible. Pauli demonstrated18 the spin-statistics relation based on the following requirements: (i) The vacuum is the state of lowest energy; (ii) Physical quantities (observables) commute with each other in two space-time points with a space-like distance; (iii) The metric in the physical Hilbert space is positive definite. In the noncommutative case the physical quantities (observables) which are in general products of several field operators, are no more local quantities and could therefore fail to fulfil the above requirement ii). For instance, taking the normally ordered product : qb2(x) : for a real scalar field with mass rn, its noncommutative version : 4(x) 4(z) : could give a nonvanishing equal-time commutation relation (ETCR). In particular, the matrix element between vacuum and a two-particle state, on a d-dimensional space, when Bose statistics is used, is:”
*
40
-
d
2i
1 (e--ip'x--ipy
+
e--ipx--ip'y
)
G
where wk = ko = and = (kl, ..., kd). The r.h.s. of (28) is nonzero only when Ooi # 0. This holds for observables expressed as any power of both bosonic fields and their derivatives, with *-product analogous to (28), and spinor fields and their derivatives, with anti-commutation relation used in the latter case. The study of NC QFT also showed a violation of both causalitylg and unitarity" conditions, for theories with noncommutative time (Ooi # 0). Indeed, while the low-energy limit of string theory in a constant antisymmetric background field B"", which exhibits noncommutativity, reduces to field theory with the *-product when Ooi = 0, for the case Ooi # 0 there is no corresponding low-energy field theory limit. The field theories with light-like noncommutativity, OpVOp, = 0, i.e. Ooi = --eli, become very interesting from this point of view as they preserve unitarity.21 In this case, however, the microcausality in the sense of ETCR (28) is still violated.ll If the field theory with light-like noncommutativity is indeed the lowenergy limit of string theory, as stated in Ref. 21, it is then intriguing that the theory is unitary but acausal (as it is known that a low-energy effective theory should not necessarily be unitary, as is the case, e.g., for the Fermi four-spinor interaction). 5. CPT Theorem in NC Field Theories
The CPT t h e ~ r e m (see ~ ~ also t ~ ~Ref. 24 for a review) is of a universal nature in that it is valid in all the known field theories. Here we shall recapitulate essential features of the CPT transformation and then extend the CPT theorem to noncommutative field theories. First, we shall summarize the common properties of anti-unitary transformations, including time reversal and CPT transformation. An antiunitary transformation denoted hereafter by 4 is a generalization of complex conjugation and satisfies
(a' ,a')
= (a
,a).
(29)
The transformation of state vectors corresponds to the Schrodinger picture and we can also attribute the same transformation to operators correspond-
ing t o the Heisenberg picture by
(9' ,Q@')
= (a , Q ' 9 ) .
In what follows we shall mainly discuss the latter approach. a) The transformation of operators obeys the following rules: (CIA
+ cgb)'
= clA' (AB)' = B'A',
+ czB'
(linearity) ,
where c1 and c g are c-number coefficients. b) Let us assume that
Q' = EQ,( E = f l ) and that 9 is an eigenstate of Q with the eigenvalue q,
&@==a; then ' 9 is also an eigenstate of Q and QQ' = Eq9'.
5.1. The C P T h n s f o r m a t i o n of Local Elementary Fields
+
In what follows we shall use the symbol exclusively for the C P T transformation and we shall first define it for local elementary fields. Let $Ia, Ga and $ A ~ . . . A ~be local elementary fields representing spinors and tensors, respectively; then the C P T transformation is specified by:25 $I%)
= (iY5)cu&d--2)
7
= $p(--S)(iYdpa
7
' $34
4 A 1 , . . A n (-2)
= (-')n$A1...An(-z)
.
(35)
This set of rules completely specifies the transformation of any local elementary field carrying definite spinor and/or tensor indices. Then the C P T theorem for local field theories can be formulated in the following form: 5 . 2 . C P T Theorem for Local Fields Let $Ia, $a and $A~.,.A,, be local but composite fields representing spinors and tensors, respectively; then they are transformed exactly in the same form as Eq. (35) for local elementary fields. In what follows we shall clarify the significance of this theorem.
42
1) Let us consider local composite scalar fields of which free and interaction Lagrangian densities, as well as interaction Hamiltonian densities, are typical members and we have:
LfI(.
, L,?,,>.(
= Lf(-).
= Lint (-.)
,
HL&) = H i n t ( - . ) .
(36) (37)
In Ref. 25, Eq. (37) has been referred to as the CPT theorem and its proof has been given there so that we skip it. When asymptotic conditions are valid, the CPT invariance of the S matrix follows from it:
s' =s.
(38) 2) Next, let @ A be a local composite vector field and q 5 a~ local elementary vector field, respectively; then a composite scalar field @ = q 5 ~ @is ~ transformed as (36) or (37) and q 5 ~as (35). From the above information we deduce:
'
= -@A(-)
(39) and similarly we can prove Eq. (35) for spinors and tensors. As an example of local composite vector fields we choose the electric current density j ~ ( z ) ; then the conserved electric charge Q transforms as: @A(.)
Q' =
/
d 3 2 j t ( 2 )= -
/
7
d3zjo(-2) = -Q.
3) The energy-momentum vector PA can be expressed as the space integral of the energy-momentum tensor of the second rank. Therefore, we immediately conclude
P!
= PA.
(41) 4) The generators of the Lorentz transformation Mpu can be expressed as the space integral of a tensor of the third rank, so that we have:
M:u
-Adpu. (42) This indicates that the spin of a particle defined in terms of the PauliLubanski operator should reverse its direction under CPT. In general, the CPT transformation of an operator is determined by the tensorial rank of its density. 5 ) We assume the validity of the LSZ asymptotic conditions;26then on the basis of their definition of the asymptotic fields, it is shown straightforward that the CPT transformation turns incoming fields into outgoing fields and vice versa. =
43
5.3. C P T Theorem for Noncommutative Fields
The validity of CPT theorem for noncommutative QED has been discussed in Ref. 27, where it was concluded that CPT is accidentally preserved, although the charge conjugation and time reversal symmetries are broken due to noncommutativity. However, in Ref. 27 the specific version of NC QED of Ref. 7 was studied, where the photon couples only to particles with the electric charges +1,-1 and 0. The latter is usually referred to as the "charge quantization problem". In Ref. 28 the CPT invariance of the noncommutative Yang-Mills theories has been shown using the SeibergWitten map. In the following, we shall show the general validity of the CPT theorem for any noncommutative quantum field theory of the type described in Sect. 11, without reference to any specific model or to the Seiberg-Witten map. Let H ( z ) be the Weyl-Moyal product (2) of field operators representing the interaction Hamiltonian in a noncommutative field theory. It is understood that H ( z ) stands for a normal product in the interaction representation. The CPT theorem is given by
In order to prove it we shall choose as an illustration a n-linear form for H ( z ) ,namely,
H(z)=
c c
fal...an+il,(z>*...*+~n(z)
il.. .in
= eD
. ...2,.
fi ,...in+:,
(21)...~~n(2,)111= ...=In~a: 1
(44)
2,
where ij with j = 1,...,n stand for spinorial or tensorial indices and the coefficients fil...in are so chosen as to make H ( z ) a scalar under proper Lorentz transformations, in the local limit. D stands for the differential operator of the form
with general W. Then the CPT transform of H ( z ) is given by:
44
where f' is given by f!2 1 ...2,. = (-1)W".
21
...2,
(47)
7
and F stands for the number of the Fermi fields involved in H ( z ) . When we reverse the order of multiplication back to the original one in (44), we obtain:
H'(.)
c c ...in4:,
=e
D
fz l... z,4~l(-z~)..,4~,(-z~)121 =...=2,=2
il
=
. ...2.,
...in fz,
(-z)
* ... * 4;,
(-z) = H ().-
.
(48)
21
Thus the CPT theorem is valid not only in local field theories but also in noncommutative field theories. This can be also seen from the fact that, when we expand the interaction Hamiltonian density in powers of 8, the first term is the local limit of the Hamiltonian expressed in terms of the Weyl-Moyal product. It is a local but composite scalar density. The coefficients of other terms are local but composite tensor fields of even ranks obtained by differentiating the fields involved in the first term, an even number of times. Therefore, they transform in the same way as the first term under CPT. From this point of view it is intuitively clear that the Hamiltonian density expressed in terms of the Weyl-Moyal product transforms in the same way as the local ones under CPT. As seen from the proof presented above, the CPT theorem is valid for any form of noncommutativity, including the case Oo2 # 0. Individual discrete transformations P, C and T The individual transformations P, C and T are violated in many cases and we shall comment on them only by comparison with the local (commutative) limit of the noncommutative field theory in question. In the case of only space-space noncommutativity (8'2 = 0), the parity of a noncommutative field theory is the same as for its commutative limit, while charge conjugation and time reversal are broken, even if they hold for the commutative limit. This is due to the fact that C and T imply a complex conjugation, that would change the sign of the phase in (45). In the case of a space-time noncommutative theory (8" # 0) - whose commutative limit is P, C and T invariant - all these discrete transformations are violated, as in the NC QED case.27
6. Conclusions In the framework of noncommutative gauge theories, we present a no-go theorem according to which the closure condition of the gauge algebra im-
45
plies that: 1) the local NC u ( n ) algebra only admits the irreducible n x n matrix-representation. Hence the gauge fields are in nxn matrix form, while the matter fields can only be in fundamental, adjoint or singlet states; 2) for any gauge group consisting of several simple-group factors, the matter fields can transform nontrivially under at m o s t two NC group factors. In other words, the matter fields cannot carry more than two NC gauge group charges. This no-go theorem imposes strong restrictions on the NC version of the Standard Model and in resolving the standing problem of charge quantization in noncommutative QED. Elaborating on the phenomenological implications of noncommutativity we have calculated the noncommutative corrections to the spectrum of the H-atom and obtained a bound on 0 from the data on the Lamb shift. We have found that the CPT theorem is generally valid in NC FT, irrespective of the form of the noncommutativity parameter O p v involved, although Lorentz invariance is violated. The spin-statistics theorem holds in the case of field theories with space-space noncommutativity, which can be obtained as a low-energy limit from the string theory. A violation of the spin-statistics relation in the case of NC time can not be justified, given the pathological character of such theories. The case of light-like noncommutativity (compatible with unitarity) deserves, however, more attention. In conclusion, it is of importance to study further the light-like case, as to determine whether it can indeed be obtained as a low-energy limit of string theory. Questions concerning a possible breaking of the spin-statistics relation are of utmost importance, since such a violation, no matter how small, would have a crucial impact on the structure and the stability of matter in the Universe. The issue, on the other hand, is of fundamental interest by itself, since up to now no theoretical argument or motivation for such a breaking has been presented.
Acknowledgments The financial support of the Academy of Finland under the Project no. 54023 is acknowledged.
References 1. S. Doplicher, K. Fredenhagen and J. E. Roberts, Phys. Lett B331,39 (1994); Comm. Math. Phys. 172, 187 (1995). 2. N. Seiberg and E. Witten, JHEP 9909, 31 (1999), hepth/9908142. 3. A. Connes, Noncommutative Geometry, Academic Press, New York (1994).
46
4. T. Filk, Phys. Lett. B376, 53 (1996); M. Chaichian, A. Demichev and P. Prehajder, Nucl. Phys. B567, 360 (ZOOO), hep-th/9812180. 5. S. Minwalla, M. Van Raamsdonk and N. Seiberg, JHEP 9906, 020 (2000), hepth/9912072. 6. M. M. Sheikh-Jabbari, JHEP 9906, 015 (1999), hep-th/9903107. 7. M. Hayakawa, Phys. Lett. B478, 394 (ZOOO), hep-th/9912094. 8. M. Chaichian, P. Presnajder, M. M. Sheikh-Jabbari and A. Tureanu, Eur. Phys. J. C29, 413 (2003), hepth/0107055. 9. M. Chaichian, P. PreSnajder, M. M. Sheikh-Jabbari and A. Tureanu, Phys. Lett. B526, 132 (2002), hep-th/0107037. 10. M. Chaichian, M. M. Sheikh-Jabbari and A. Tureanu, Phys. Rev. Lett. 86, 2716 (2001), hep-th/0010175. 11. M. Chaichian, K. Nishijima and A. Tureanu, Phys. Lett. B568, 146 (2003), h e p t h/0209008. 12. S. Terashima, Phys. Lett. B482, 276 (2000), hep-th/0002119. 13. L. Bonora, M. Schnabl, M. M. Sheikh-Jabbari and A. Tomasiello, Nucl. Phys. B589, 461 (2000), hep-th/0006091. 14. X. Calmet, B. JurEo, P. Schupp, J. Wess and M. Wohlgennant, Eur. Phys. J.C23, 363 (2002), hepph/Olllll5. 15. M. Chaichian, M. M. Sheikh-Jabbari and A. Tureanu, hep-th/0212259. 16. P.-M. Ho and H.-C. Kao, Phys. Rev. Lett. 88,151602 (2002), hep-th/0110191. 17. S. M. Carroll, J. A. Harvey, V. A. Kostelecky, C. D. Lane and T. Okamoto, Phys. Rev. Lett. 87, 141601 (2001), hep-th/0105082. 18. W. Pauli, Phys. Rev. 58, 716 (1940); Progr. Theor. Phys. (Kyoto) 5, 516 (1950). 19. N. Seiberg, L. Susskind and N. Toumbas, JHEP 0006, 044 (2000), h e p th/0005015; L. Alvarez-GaumB and J. L. F. Barbon, Int. J. Mod. Phys A16, 1123 (2001), hep-th/0006209. 20. J. Gomis and T. Mehen, Nucl. Phys. B591, 265 (ZOOO), hep-th/0005129. 21. 0. Aharony, J. Gomis and T. Mehen, JHEP 0009, 023 (2000), hepth/0006236. 22. G. Luders, Dansk. Mat. Fys. Medd. 28, 5 (1954). 23. W . Pauli, Niels Bohr and the Development of Physics, W. Pauli (ed.), Pergamon Press, New York (1955). 24. R. F. Streater and A. S. Wightman, CPT, Spin, Statistics and All That, W. A. Benjamin, Inc., New York (1964), and references therein. 25. K. Nishijima, Fundamental Particles, W. A. Benjamin, Inc., New York (1963). 26. H. Lehmann, K. Symanzik and W. Zimmermann, Nuovo Cimento 1, 1425 (1955); 6, 319 (1957). 27. M. M. Sheikh-Jabbari, Phys. Rev. Lett. 84, 5265 (2000), hep-th/0001167. 28. P. Aschieri, B. JurEo, P. Schupp and J. Wess, Nucl. Phys. B651, 45 (2003), h e p t h/02052 14.
SHADOWS OF QUANTUM BLACK HOLES
N.KALOPER Department of Physics, University of California Davis, C A 95616, USA E-mail:
[email protected]
We discuss our recent conjecture that black holes localized on a brane in AdSD+I should be interpreted as quantum-corrected D-dimensional black holes, rather than classical ones, in the dual CFT coupled to gravity. Thus in 4D they include the corrections from Hawking radiation.
1. Black Holes in AdS/CFT with a Cutoff: the Role of Tunneling The semi-infinite Randall Sundrum modell is based on a bulk geometry of AdSD+I space ending on a D - 1-dimensional domain wall, or brane. A prototype is the RS2 model where Ads5 ends on a 3-brane, which should model our 3+1 dimensional world. It is natural to ask what is a suitable description of a black hole in this scenario. The attempts to find exact, static, asymptotically flat black hole solutions localized on the brane in AdSD+l>4, with regular horizons both on and off the brane, have been marred with very serious It has even been suggested that static, asymptotically flat, spherical black holes on the brane might not altogether exist in the RS2 model. In contrast, exact static solutions localized on a 2-brane in Ad& have been found in Refs. 7, 8. This is surprising in two ways. First of all, the RS2 model in 5 0 is believed to have a dual description in terms of purely 4 0 physics, as a strongly coupled CFT with a UV cutoff, coupled to weak gravity and perhaps Standard Model-like excitations. One would expect that such a theory should admit the conventional black holes, which are no more difficult t o construct than is usual. On the other hand, the same argument applied to a 3 0 CFT coupled to gravity would suggest that there should not be any black holes to begin with, since as is well known,g there are no black hole solutions of 2 1 Einstein gravity in asymptotically flat spaces. So why do
+
47
48
such objects appear in the variant of RS2 in the 4 0 bulk? In order to resolve this, in the recent work with Emparan and Fabbri," we have proposed a connection between the bulk and dual CFT+gravity interpretation of black holes in RS2, based on a modification of AdS/CFT correspondence" for the RS2 m ~ d e l . ~ ~Our - ~main ' result was the following conjecture:
The black hole solutions localized o n the brane in the AdSD+1 braneworld which are found by solving the classical bulk equations in AdSD+1 with the brane boundary conditions, correspond to quantum-corrected black holes in D dimensions. rather than classical ones. This comes about as follows: according to AdS/CFT, the classical dynamics in the AdSD+1 bulk encodes the quantum dynamics of the dual D-dimensional conformal field theory (CFT), in the planar limit of a large N expansion. Cutting the bulk with a brane introduces a normalizable D-dimensional graviton while on the dual side this same Ddimensional gravity mode is merely added to the CFT, which is also cutoff in the ultraviolet. Then, solving the classical D 1-dimensional equations in the bulk is equivalent to solving the D-dimensional Einstein equations G,, = 87rG~(T,v)CFT,where the CFT stress-energy tensor incorporates the quantum effects of all planar diagrams. These include particle production in the presence of a black hole, and possibly other vacuum polarization effects. We have shown that the conjecture is fully consistent with the existence of black holes in 2+1 CFT+gravity, which emerge solely due to the quantum corrections in the CFT sector of the the0ry.l' In fact, this result provides for an interesting spinoff, showing that in 2 1 dimensions, the quantum dynamics of a CFT naturally serves as a cosmic censor, regulating the (6function) naked singularities, and dressing them in a horizon. Thus the black holes in 2 1 dimensions are generic, once quantum corrections are included. This is true independently of whether the CFT is strongly or weakly coupled, and is more efficient when there is more CFT degrees of freedom. We have also found a consistent reinterpretation of black holes in the physically more relevant case of a 3-brane in AdS5.l' The main point is that the CFT+gravity dual allows us to reinterpret the alleged obstruction for finding a static black hole3 as a manifestation of the backreaction from Hawking effects. As long as the bulk is asymptotically Ads, the conformal
+
+
+
49
symmetry of the dual CFT is valid in the infrared, and so there is no mass gap. Thus any black hole at a finite temperature will emit CFT modes as a thermal spectrum of Hawking radiation, which on the bulk side is captured by a deformation of the bulk geometry close to the brane. Since in the asymptotically flat space Hawking radiation escapes t o infinity, carrying away the black hole energy, the black hole mass must be time-dependent, and hence the geometry is not stationary. The mechanism of the tunneling suppression plays an important role in the determination of the Hawking effects from the bulk side, and we will review this below. We begin with several key aspects of the AdS/CFT d i c t i ~ n a r y . ~ ~ ~ ~ ~ Since we want to discriminate between classical and quantum effects, we must keep ti in our formulas, while setting c = 1. Then, the 4-dimensional Newton’s constant G4, Planck length l 4 , and Planck mass M4 are related to each other as
ti M4 = -
-e4
G4 = -
M4 l
e4
In Ads braneworlds the 5 0 bulk Newton’s constant and the bulk cosmological constant A5 = -10/L2 together determine the Newton’s constant induced on the D-dimensional brane as 1
G4=-G5.
L
(2)
In light of our discussion from previous sections, this means that we are taking the volume of the Calabi-Yau space to be negligible compared to the volume of the brane throat, which is the opposite limit t o that we worked in previously. The precise details of the dual CFT depend on the specifics of the string/M-theory construction that yield the Ads background. For our purposes here it is enough to determine the effective number of degrees of freedom of the CFT, g*. For D = 4, the dual pair are IIB string theory on Ads5 x S5 of radius L e1,(g,N)1/4 and N = 4 S U ( N ) super Yang-Mills theory, leading to N
where we have used Eq. (2) to get the final expressions, 9%is taken to be a large number, in order to keep small quantum corrections to the supergravity approximation to string/M-theory. For the CFT, this is a large N limit where planar diagrams give the leading contribution. The Planck brane that cuts off the Ads bulk denotes that very high energy states of the dual CFT are integrated out, breaking the conformal invariance of the
50
theory in the UV. However, the breaking washes into the low energy theory only through irrelevant operators, generated by integrating out the heavy CFT states at the scale p u v h/L. In the IR, at energies E < p u v , the effects of the conformal symmetry breaking are suppressed by powers of E / p u v , meaning that the bulk geometry far from the brane is Ads. Cutting off the bulk yields also a normalizable graviton zero mode localized on the brane; this same D-dimensional gravity mode is added to the dual theory. However, note that the CFT cutoff p u v is not equal to the'induced D-dimensional Planck mass. Instead,
-
which is much smaller than the Planck mass on the brane, and is what we would expect for the single throat limit of the "octopus", as discussed previously.
1.1. Resolving the Mystery of the Missing 3 -/- 1 Black Hole As we have said above, as long as the bulk is Ads5 far from the brane, the dual CFT is conformal in the IR, without a mass gap separating the CFT modes from the vacuum. Hence any black hole at a finite temperature will have unrestricted access t o a large number of light CFT modes, and will them with a thermal spectrum, which is precisely the Hawking radiation.a On the bulk side, this must be described by a deformation of the bulk geometry near the brane, which arises because the black hole appears as a source in the classical bulk gravity equations. Computing these effects entails the usual complications involving the choice of the vacuum for a quantum field theory in a black hole background, with the possibilities being (1) the Hartle-Hawking state, (2) the Unruh state, and (3) the Boulware state. Then by our conjecture, the black hole on the RS2 brane must correspond to one of these choices, with the corrections from the backreaction included. This immediately shows why the search for a static, "We should note that a step in this direction for the case of RS2 in Ads5 was made by T. Tanaka,20 and, simultaneously, by R. Maartens and the author, in order t o explain the results of Ref. 3. A naive argument that the bulk dynamics encodes the backreaction from Hawking radiation would lead one to expect that all asymptotically flat branelocalized black holes are time-dependent. This would be in conflict with the exact static 2 1 solutions of Refs. 7, 8. Our conjecture that the classical bulk dynamics encodes all quantum corrections at the level of planar diagrams completely resolves this conflict. These exact solutions in fact strongly support the conjecture as presented in here."
+
51
asymptotically flat black hole solution on the brane has failed so far: the state (1) is not asymptotically flat, ( 2 ) is not static, and (3) does not have a regular horizon. This leads us to considering a radiative solution as the leading-order description of the exterior of a black hole localized on the brane. The detailed description of this geometry on the bulk side would require either the exact bulk solution, which has been missing so far, or a much better approximation than the existing ones. On the side of the 3 f l CFT+gravity, a description at the same level of rigor would require a careful backreaction analysis, where we should start with a classical Schwarzschild black hole and perturb it by means of the (Tpy) in the Unruh state evaluated in the classical background geometry. The far-field outgoing metric encodes the flux of Hawking radiation pouring out of the black hole, which is described by the stress-energy tensor
where u is the retarded null coordinate and L(u) is the flux luminosity. The perturbed geometry is
ds2=-
(1 - 2G4y(u)) du2 2drdu + r2dS22, -
where = -L(u). To check our conjecture, we should recover the relation between L and M from leading-order corrections to the black hole geometry induced from the bulk. The precise calculation would require the detailed matching of the far-field solution (6) to a near horizon one, which should then be matched onto the interior. In order to circumvent these details, we have considered the radiative collapse of a very massive dust cloud, forming a black hole of large mass. This collapsing cloud of dust, whose interior is described by the bulk dynamics encompassing leading order quantum CFT corrections, determined in Ref. 3 can be matched to an outgoing Vaidya metric (6), following the work of Ref. 21. The quantum corrections propagate through the matching regions, and this relates the outgoing flux of radiation to the subleading correction in the interior star geometry, which is c( ( G ~ M J ~ ) as ~ calcu/R~, lated in Ref. 3, r.h.s. of their Eq. (6) (we only consider the limit Q = A = 0 of this equation, which is sufficient for our purposes). Comparing to Eq. (5) we find L G 4 ( M L ) 2 / R t hg,(G4M)2/R:, where Ro is the radius of the matching surface. For a large collapsing mass, this will be near 2 G 4 M , so L hg*/(G4M)2. This corresponds to a flux of Hawking radiation of N
N
-
52
-
g* degrees of freedom of the CFT, at a temperature TH ti/(G4M), as required. Replacing M ( u ) by M is consistent since L c( ti and we are working in an expansion in ti. While this does not reproduce a detailed formula with accurate numerical coefficients, it does give the correct scalings with the black hole and CFT parameters, in complete accord with our conjecture. We should still check the consistency of the matching of geometries across the horizon. A simple way to check this is to compare the quantum trace anomalies of the backreacted states in the exterior and interior. The trace anomaly of the quantum stress tensor is a local geometric quantity independent of which vacuum the field is in.22>23 It has been studied in detail in the AdS/CFT context,24and in particular in the case of Ads braneworlds in Refs. 25-27. It gives us further insight into our problem, in that it provides a simple leading-order consistency check, which a configuration must pass in order to be described by the leading-order effects in the duality pair. In the case of D = 4 N = 4 SU(N) SYM at large N N
Note the absence of the term RPVapRFYap.Ref. 24 showed how this anomaly is precisely reproduced from a computation in the AdSs bulk. This result is perturbatively identical to the familiar quadratic stress-energy correction terms that appear in the effective long distance 3 1 gravity equations in Ads braneworlds,28 which can be checked explicitly recalling g* N2.“J6 The matching to the far-field Vaidya metric (6), is consistent with this form of the anomaly, because the tracelessness of the radiation stress-energy implies R P , , = 0, and so the anomaly vanishes, with no contributions from the R P y a p R P V a p terms. Although this argument by itself does not fully guarantee that the bulk will be free from singularities, it passes the anomaly check with only minimal assumptions which are physically well-motivated. It is now straightforward to correctly interpret the ”no-go theorem” of Ref. 3. The anomaly matching requires that the exterior is radiative, leading to a time-dependent evaporating black hole (6). Therefore, the classical bulk dynamics does require braneworld black holes to be time-dependent. However, this is simply a natural consequence of black hole quantum mechanics, and is generated as a leading order quantum correction. Understanding this picture from the point of view of the full bulk Ads5 spacetime, and in particular the details of the dual description of the Hawking radiation as a classical bulk process represents a very
+
-
53
interesting challenge. In Ref. 10 we have touched upon some aspects of this picture. The mechanism of tunneling ~ u p p r e s s i o nplays ~ ~ a very important role for understanding the difference between the rapidly evaporating large black holes and slowly evaporating small ones. For the small black holes, the description in terms of a 3 1 theory of gravity+CFT breaks down. A black hole of size r H << L is approximated near the horizon by a five-dimensional, static Skhwarzschild solution. Classical radiation into the bulk, and therefore 3 1 Hawking radiation into CFT modes, is suppressed for such light black holes. Namely, while large black holes are shaped like pancakes around the brane, they extend to distances larger than the Ads radius L. Thus they couple to all the CFT modes, including the lightest ones, with M4 couplings, without any suppression. On the other hand, while the small black holes are bulging away from the brane, they are much smaller than the Ads radius, and from the perturbative point of view, they live inside the RS2 "volcano". Hence their classical couplings to all bulk graviton modes are tunneling-suppressed in the sense of Ref. 29, and are exponentially weaker than M4. Thus the radiation rate must go down significantly.b Hence the light black holes evaporate, although more slowly, only via bulk Hawking radiation, where they mostly emit along the brane.30 Since this picture for the evolution of an evaporating black hole is based on specific properties of the UV extension provided by the bulk theory, there is no reason why it should apply to situations that do not have an AdS/CFT dual description.
+
+
2. Conclusions We have reviewed here our recent proposal concerning black holes on branes in Ads space." On the bulk side, this is realized by putting the black hole on a brane in the cutoff AdSD+I bulk, which localizes dynamical gravity. These black holes provide for a dual description of quantum-corrected Ddimensional black holes in CFT+gravity. On the dual side, the quantum Hawking radiation of CFT modes is described as the classical evolution in bThe same effect occurs for a large object on the brane, of mass A4 >> M4 but lower density than a black hole, such as a star. Even if the star had accelerated by being stuck to the brane, the bulk deformation it would cause would have been confined to distances less than L , so its emission would have been tunneling-suppressed. The reason why a black hole radiates in the bulk whereas a star does not is also dual to the problem of the different choices of CFT vacua and boundary conditions for the radiation.
54
the bulk, and the classical bulk point of view may lead to a better understanding of quantum black hole evaporation. In a later we have considered the phenomenological implications of these effects. We expect that many of our results should naturally extend t o any CFT+gravity theory, even if a dual bulk description along the lines of RS2 does not exist. We hope to return t o these questions in the future.
Acknowledgements The author wishes t o express his gratitude t o the organizers of the “BW2003 Workshop” for putting together such a stimulating meeting and for their assistance during the workshop. Particular thanks go t o Goran Djordjevib and Ljubisa Nesib. It is a great pleasure t o thank R. Emparan, A. Fabbri and J. Garcia-Bellido for collaborations which led t o the results presented here and which helped shape author’s thinking on these subjects. Further thanks go to R. Maartens, L. Susskind and T. Tanaka for very useful discussions. The work of NK was supported in part by the DOE Grant DE-FG03-91ER40674, in part by the NSF Grant PHY-0332258 and in part by a Research Innovation Award from the Research Corporation.
References 1. 2. 3. 4. 5. 6. 7. 8. 9.
10. 11.
12. 13.
L. Randall and R. Sundrum, Phys. Rev. Lett. 83, 4690 (1999). A. Chamblin, S. W. Hawking and H. S. Reall, Phys. Rev. D61,065007 (2000). M. Bruni, C. Germani and R. Maartens, Phys. Rev. Lett. 87, 231302 (2001). N. Deruelle, Stars on branes: The view from the brane, gr-qc/0111065. P. Kanti and K. Tamvakis, Phys. Rev. D65, 084010 (2002). R. Casadio, A. Fabbri and L. Mazzacurati, Phys. Rev. D65, 084040 (2002). R. Emparan, G. T. Horowitz and R. C. Myers, JHEP 0001, 007 (2000). R. Emparan, G. T. Horowitz and R. C. Myers, JHEP 0001, 021 (2000). S. Deser, R. Jackiw and G.’t Hooft, Annals Phys. 152, 220 (1984); S. Giddings, J. Abbott and K. Kuchar, Gen. Rel. Grau. 16, 751 (1984); S. Carlip, Quantum Gravity in 2+l Dimensions, Cambridge University Press, Cambridge (1998). R. Emparan, A. Fabbri and N. Kaloper, Quantum black holes as holograms in A d s braneworlds, hepth/0206155. J. M. Maldacena, Adu. Theor. Math. Phys. 2, 231 (1998); S. S. Gubser, I. R. Klebanov and A. M. Polyakov, Phys. Lett. B428, 105 (1998); E. Witten, Adv. Theor. Math. Phys. 2, 253 (1998); for a review see 0. Aharony, S. S. Gubser, J. M. Maldacena, H. Ooguri and Y. Oz, Phys. Rept. 323,183 (2000). N. Arkani-Hamed, S. Dimopoulos, G. R. Dvali and N. Kaloper, Phys. Rev. Lett. 84, 586 (2000). J. Maldacena, unpublished; E. Witten, unpublished.
55
14. 15. 16. 17. 18. 19. 20.
H. Verlinde, Nucl. Phys. B580, 264 (2000). S. S. Gubser, Phys. Rev. D63, 084017 (2001). S. B. Giddings, E. Katz and L. Randall, JHEP 0003, 023 (2000). M. J. Duff and J. T. Liu, Phys. Rev. Lett. 85, 2052 (2000). S. B. Giddings and E. Katz, J. Math. Phys. 42, 3082 (2001). N. Arkani-Hamed, M. Porrati and L. Randall, JHEP 0108, 017 (2001). T. Tanaka, Classical black hole evaporation in Randall-Sundrum infinite
braneworld, gr-qc/0203082. 21. N. 0. Santos, MNRAS 216, 403 (1985); A. K. G. de Oliveira, N. 0. Santos and C. Kolassis, MNRAS 216, 1001 (1985). 22. D. M. Capper and M. J. Duff, Phys. Lett. A 5 3 , 361 (1975). 23. S. M. Christensen and S. A. Fulling, Phys. Rev. D15, 2088 (1977). 24. M. Henningson and K. Skenderis, JHEP 9807, 023 (1998). 25. S. de Haro, K. Skenderis and S. N. Solodukhin, Class. Quant. Grav. 18, 3171 (2001). 26. T. Shiromizu and D. Ida, Phys. Rev. D64, 044015 (2001). 27. S. Kanno and J. Soda, Brane world effective action at low energies and AdS/CFT, hep-t h/0205 188. 28. T. Shiromizu, K. I. Maeda and M. Sasaki, Phys. Rev. D62, 024012 (2000). 29. S. Dimopoulos, S. Kachru, N. Kaloper, A. E. Lawrence and E. Silverstein, Phys. Rev. D64, 121702 (2001); Generating small numbers b y tunneling in multi-throat compactifications, hepth/0106128. 30. R. Emparan, G. T. Horowitz and R. C. Myers, Phys. Rev. Lett. 85, 499 (2000). 31. R. Emparan, J. Garcia-Bellido and N. Kaloper, JHEP 0301, 079 (2003).
YUKAWA QUASI-UNIFICATION A N D INFLATION
G. LAZARIDES AND C. PALLIS Physics Division, School of Technology, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece E-mail:
[email protected],
[email protected] We review the construction of a concrete supersymmetric grand unified model, which naturally leads to a moderate violation of 'ksymptotic'' Yukawa unification and, thus, can allow an acceptable b-quark mass within the constrained minimal supersymmetric standard model with p > 0. The model possesses a wide and natural range of parameters which is consistent with the data on the cold dark matter abundance in the universe, b + sy,the muon anomalous magnetic moment and the Higgs boson masses. Also, it automatically leads to a new version of shifted hybrid inflation, which avoids overproduction of monopoles at the end of inflation by using only renormalizable terms.
1. Introduction The most restrictive version of the minimal supersymmetric standard model (MSSM) with gauge coupling unification, radiative electroweak breaking and universal boundary conditions from gravity-mediated soft supersymmetry (SUSY) breaking, known as constrained MSSM (CMSSM)l, can be made even more predictive, if we impose Yukawa unification (W),1.e. ' assume that the three third generation Yukawa coupling constants unify at the SUSY grand unified theory (GUT) scale, MGTJT.The requirement of W can be achieved by embedding the MSSM in a SUSY GUT with a gauge group containing SU(4), and SU(2),. Indeed, assuming that the electroweak Higgs superfields hy", hzw and the third family right handed quark superfields t C ,b" form SU(2), doublets, we obtain2 the 'asymptotic' Yukawa coupling relation ht = h b and, hence, large t a n 0 mt/mb. Moreover, if the third generation quark and lepton SU(2), doublets [singlets] 43 and l3 [b" and T " ] form a SU(4), 4-plet [a-plet] and the Higgs doublet QW which couples to them is a SU(4), singlet, we get h b = h, and the 'asymptotic' relation m b = m7 follows. The simplest GUT gauge group which contains both SU(4), and SU(2), is the Pati-Salam (PS) group
-
56
57
GPS = SU(4), x SU(2), x SU(2), and we will use it in our analysis. However, applying W in the context of the CMSSM and given the experimental values of the top-quark and tau-lepton masses (which naturally restrict t a n p N 50), the resulting value of the b-quark mass turns out to be unacceptable. This is due to the fact that, in the large t a n p regime, the tree-level b-quark mass receives sizeable SUSY correction^^^^^^*^ (about 20%), which have the sign of p (with the standard sign convention7) and drive, for p > [
2.684 GeV 5 mb(MZ) 5 3.092 GeV with a , ( M z ) = 0.1185.
(1)
This is derived by appropriately' evolving the corresponding range of m b ( m b ) in the M S scheme (i.e. 3.95 - 4.55 GeV) up to M z in accordance with Ref. 9. We see that, for both signs of p, W leads to an unacceptable b-quark mass with the p < 0 case being less disfavored. A way out of this m b problem can be found8 without abandoning the CMSSM (in contrast to the usual strategy6~10~11~12) or YU altogether. We can rather modestly correct W by including an extra SU(4)c non-singlet Higgs superfield with Yukawa couplings to the quarks and leptons. The Higgs SU(2), doublets contained in this superfield can naturally develop13 subdominant vacuum expectation values (VEVs) and mix with the main electroweak doublets, which are assumed to be SU(4), singlets and form a SU(2), doublet. This mixing can, in general, violate SU(2),. Consequently, the resulting electroweak Higgs doublets h;w, do not form a su(2)R doublet and also break SU(4),. The required deviation from YU is expected to be more pronounced for p > 0. Despite this, we will study here this case, since the p < 0 case has been excluded14 by combining the Wilkinson microwave anisotropy probe (WMAP) re~trictions'~ on the cold dark matter (CDM) in the universe with the recent experimental results16 on the inclusive branching ratio BR(b 4 sy). The same SUSY GUT model which, for p > 0 and universal boundary conditions, remedies the m b problem leads to a new version17 of shifted hybrid inflation", which avoids monopole overproduction at the end of inflation and, in contrast to the older version1', is based only on renormalizable interactions. In Sec. 2, we review the construction of a SUSY GUT model which modestly violates W, yielding an appropriate Yukawa quasi-unification condition (YQUC), which is derived in Sec. 3. We then describe the resulting CMSSM in Sec. 4 and introduce the various cosmological and phenomenclogical requirements which restrict its parameter space in Sec. 5. In Sec. 6,
58
we delineate the allowed range of parameters for p > 0 and, in Sec. 7, we outline the corresponding inflationary scenario. Finally, in Sec. 8, we summarize our conclusions. 2. The SUSY GUT Model We will take the SUSY GUT model of shifted hybrid inflation" (see also Ref. 19) as our starting point. It is based on Gps, which is the simplest gauge group that can lead to W. The representations under GPS and the global charges of the various matter and Higgs superfields contained in this model are presented in Table 1 , which also contains the extra Higgs superfields required for accommodating an adequate violation of W (see below). The matter superfields are Fi and FF (i = 1 , 2 , 3 ) , while the electroweak Higgs doublets belong to the superfield h. So, all the requirements for exact W are fulfilled. The breaking of GPS down to the standard model (SM) gauge group GSM is achieved by the superheavy VEVs (N MGUT)of the right handed neutrino type components of a conjugate pair of Higgs superfields H", I?. The model also contains a gauge singlet S which triggers the breaking of GPS, a SU(4), 6-plet G which gives2' masses to the right handed down quark type components of H", H " , and a pair of gauge sinfor solving21 the p problem of the MSSM via a Peccei-Quinn glets N , (PQ) symmetry. In addition to Gps, the model possesses two global U ( l ) symmetries, namely a R and a PQ symmetry, as well as a discrete Z:* symmetry ('matter parity'). A moderate violation of exact W can be naturally accommodated in this model by adding a new Higgs superfield h' with Yukawa couplings FF'h'. Actually, (15,2,2)is the only representation, besides (1,2,2),which possesses such couplings to the fermions. In order to give superheavy masses to the color non-singlet components of h', we need to include one more Higgs superfield h' with the superpotential coupling h'h', whose coefficient is of the order of MGUT. After the breaking of Gps to GSM,the two color singlet SU(2), doublets h i , hi contained in h' can mix with the corresponding doublets h l , h2 in h. This is mainly due to the terms h'h' and H'n'h'h. Actually, since
w
H c H c = ( 4 , 1 , 2 ) ( 4 , 1 , 2 )= ( 1 5 , 1 , 1 + 3 ) + . . -, h'h = (15,2,2)(1,2,2)= ( 1 5 , 1 , 1 + 3) + . * * , there are two independent couplings of the type H"A"L'h (both suppressed by the string scale MS M 5 x lo" GeV, being non-renormalizable). One of
59
Table 1. Superfield Content of the Model Superfields
Representations
Global
under Gps ~~
Symmetries
R
PQ
Z;"
0
~
Matter Fields
Higgs Fields
h
(1,212)
0
1
N
(1,1,1)
1/2
-1
0
N
( 1 , L 1)
0
1
0
Extra Higgs Fields
them is between the su(2)R singlets in H'H" and h'h, and the other between the SU(2)R-triplets in these combinations. So, we obtain two bilinear terms h',hl and hhh2 with different coefficients, which are suppressed by M G U T / M ~These . terms together with the terms hih', and hhhh from h'h', which have equal coefficients, generate different mixings between hl , h', and h2, hh. Consequently, the resulting electroweak doublets h;", h5w contain SU(4), violating components suppressed by MGUT/MSand fail t o form a s u ( 2 ) R doublet by an equally suppressed amount. So, YU is moderately violated. Unfortunately, as it turns out, this violation is not adequate for correcting the b-quark mass within the CMSSM for p > 0. In order to allow for a more sizable violation of YU, we further extend the model by including q5 with the coupling q5h'h. To give superheavy masses t o the color non-singlets in 4, we introduce one more superfield 4 with the coupling 64, whose coefficient is of order MGUT.
60
The terms $4 and $H"B" imply that, after the breaking of Gps t o GSM, 4 acquires a superheavy VEV of order MGUT.The coupling &'h then generates SU(2)R violating unsuppressed bilinear terms between the doublets in h' and h. These terms can certainly overshadow the corresponding ones from the non-renormalizable term HcHcK'h. The resulting SU(2), violating mixing of the doublets in h and h' is then unsuppressed and we can obtain stronger violation of W.
3. The Yukawa Quasi-Unification Condition
To further analyze the mixing of the doublets in h and h', observe that the part of the superpotential corresponding to the symbolic couplings i'h', &'h is properly written as mtr (h't7L'c)
+ ptr (~ ' ~ 4 7 L,t )
(2)
where EE is the antisymmetric 2 x 2 matrix with €12 = +1, t r denotes trace taken with respect t o the SU(4), and SU(2), indices and tilde denotes the transpose of a matrix. After the breaking of Gps t o GSM,4 acquires a VEV (4) MGUT. Substituting it by this VEV in the above couplings, we obtain N
tr(h'&'E)
= ?lchi
+ hl,eh; + . .. ,
- = -tr(h'EashE) (4) - tr(h'e4he) d
(3)
(4) = = -(h:ehz
Jz
- k1&!J
,
(4)
where the ellipsis in Eq. (3) contains the colored components of h', h' and a3 = diag(1, -1). Inserting Eqs. (3) and (4) into Eq. (2), we obtain
m%E(hh - alhz) + m($
+ a1i1)eG
with
a1
= -p($)/z/2m.
(5)
So, we get two pairs of superheavy doublets with mass m. They are predominantly given by
The orthogonal combinations of h l , h', and hz, hh constitute the electroweak doublets
The superheavy doublets in Eq. (6) must have vanishing VEVs, which readily implies that ( h i ) = - a l ( h l ) , (ha) = al(h2). Equation (7) then
61
+
gives (h;") = (1 la112)1/2(h1), (h;") = (1+ l a 1 1 ~ ) l / ~ ( hFrom 2 ) . the third generation Yukawa couplings y33F3hFi1 2yi3F3h'Fi, we obtain
where p = yi3/y33. From Eqs. ( 8 ) and (9), we see that YU is now replaced by the YQUC,
h t : h b : h, = ( l + c ) : (1 - c ) : (1+3c), with 0 < c = p a l / & <
1 . (10)
For simplicity, we restricted ourselves to real values of c only which lie between zero and unity. 4. The resulting CMSSM
Below MGUT,the particle content of our model reduces t o this of MSSM (modulo SM singlets). We assume universal soft SUSY breaking scalar masses mo, gaugino masses M1/2and trilinear scalar couplings A0 at MGUT. Therefore, the resulting MSSM is the so-called CMSSM' with p > 0 and supplemented by Eq. (10). With these initial conditions, we run the MSSM renormalization group equations ( R G E s ) ~ between ~ MGUTand a common SUSY threshold MSUSYN (m;,m;2)1/2 ( i 1 , 2 are the stop mass eigenstates) determined in consistency with the SUSY spectrum. At MSUSY,we impose radiative electroweak symmetry breaking, evaluate the SUSY spectrum and incorporate the SUSY correction^^>^>^ to the b and T masses. Note that the corrections to the T mass (almost 4%) lead14 to a small reduction of tanp. From MSUSYto M z , the running of gauge and Yukawa coupling constants is continued using the SM RGEs. For presentation purposes, M1/2 and mo can be replaced22 by the lightest SUSY particle (LSP) mass, m ~ s p and , the relative mass splitting between this particle and the lightest stau ?2, A, = (mq - ~ L S P ) / ~ L S P . For simplicity, we restrict this presentation to the A0 = 0 case (for A0 # 0 see Refs. 8 and 23). So, our input parameters are mLSp and A,. For any given m b ( M Z ) in the range in Eq. (1) and with fixed mt(mt)= 166 GeV and m,(Mz) = 1.746 GeV, we can determine the parameters c and tan@ at MSUSYso that the YQUC in Eq. (10) is satisfied.
62
5. Cosmological and Phenomenological Constraints Restrictions on the parameters of our model can be derived by imposing a number of cosmological and phenomenological requirements (for similar recent analyses, see Refs. 11, 12 and 24). These constraints result from: Cold dark matter considerations. In the context of CMSSM, the LSP can be the lightest neutralino. It naturally arises25 as a CDM candidate. We require its relic abundance, &ph2, not t o exceed the 95% c.1. upper bound on the CDM abundance derived15 by WMAP: RCDMh2
5 0.13.
(11)
We calculate S2Lsph2 using micrOMEGAs26, which is certainly one of the most complete publicly available codes. It includes all possible coannihilation processes27 and one-loop QCD corrections2' to the Higgs decay widths and couplings to fermions. Branching ratio of b + sy. Taking into account the recent experimental results16 on this ratio, BR(b --f sy), and combining' appropriately the experimental and theoretical errors involved, we obtain the following 95% c.1. range: 1.9 x
5 BR(b + sr) 5 4.6 x l o p 4 .
(12)
We compute BR(b --+ sy) by using an updated version of the relevant calculation contained in the micrOMEGAs package26. In this code, the SM contribution is calculated following Ref. 29. The charged Higgs ( H * ) contribution is evaluated by including the next-to-leading order (NLO) QCD corrections3' and tan ,B enhanced contribution^^^. The dominant SUSY contribution includes resummed NLO SUSY QCD corrections3', which hold for large tanp. Muon anomalous magnetic moment. The deviation, Sap, of the measured value of a, from its predicted value in the SM, aEM,can be attributed to SUSY contributions, calculated by using the micrOMEGAs routine31. The calculation of aEMis not yet stabilized mainly because of the instability of the hadronic vacuum polarization contribution. According to the most up-to-date e ~ a l u a t i o n there ~ ~ , is still a considerable discrepancy between the findings based on the e+e- annihilation data and the ones based on the r-decay data. Taking into account these results and the experimental m e a ~ u r e r n e n tof~ ~a,, we get the following 95% c.1. ranges: -0.53 x 10-l'
-13.6 x 10-l'
5 Sap 5 44.7 x 10-l' , 5 Sap 5 28.4 x 10-l' ,
e+e--based ; -r-based.
(13) (14)
63
2.0
1.6
$3.0
0.6
0.0 300
200
300
400
600
600
700
Figure 1. The various restrictions on the m ~ -~Afzp plane for p > 0, A0 = 0 and a S ( M z )= 0.1185. From left to right, the solid lines depict the lower bounds on m L S p from baP < 44.7 x 10-lo, BR(b + sy) > 1.9 x lop4 and m h > 114.4 GeV and the upper bound on m ~ s from p R ~ s p h<~0.13 for m b ( M z ) = 2.888 GeV. The dashed [dotted] line depicts the bound on m L S p from R ~ s p < h ~ 0.13 for m b ( M z ) = 2.684 [3.092] GeV. The allowed area for m b ( M z ) = 2.888 GeV is shaded.
Following the common practice24, we adopt the restrictions to parameters induced by Eq. (13), since Eq. (14) is considered as quite oracular, due to poor r-decay data. 0 Collider bounds. Here, the only relevant collider bound is the 95% c.1. LEP bound34 on the mass of the lightest CP-even neutral Higgs boson h: rnh
2 114.4 GeV.
(15)
The SUSY corrections to mh are calculated at two loops by using the FeynHiggsFast program35 included in the micrOMEGAs code26. 6. The Allowed Parameter Space
We will now try to delineate the parameter space of our model with p > 0 which is consistent with the constraints in Sec. 5. The restrictions on the mLSp plane, for Ao = 0 and the central values of a , ( M z ) = 0.1185 and rnb(Mz) = 2.888 GeV, are indicated in Fig. 1 by solid lines, while the upper bound on m ~ s pfrom Eq. ( l l ) , for r n b ( M ~= ) 2.684 [3.092] GeV, is depicted by a dashed [dotted] line. We observe the following:
64
3000
2500
2000
c3 v u) 1600 u)
Q
E 1000 500
-
0 'LOO
200
300
400
600
600
700
Figure 2. The mass parameters m~ and Msusy as functions of mLSp for various values of A,, which are indicated on the curves. We take p > 0, A0 = 0, m b ( M ~=) 2.888 GeV and cu,(Mz) = 0.1185.
3000
2500
0
'LOO
200
300
m,B
400
600
600
700
(GeV)
Figure 3. The mass parameters m A and Msusy versus m ~ for~ p p> 0, A0 = 0, A, = 1, a,(Mz) = 0.1185 and with m b ( M z ) = 2.684 GeV (dashed lines), 3.092 GeV (dotted lines) or 2.888 GeV (solid lines).
65
0
The lower bounds on mLSp are not so sensitive to the variations of mb(MZ)* The lower bound on mLSp from Eq. (15) overshadows all others. The upper bound on mLSp from Eq. (11) is very sensitive to the variations of mb(MZ). In particular, one notices the extreme sensitivity of the almost vertical part of the corresponding line, where the LSP annihilation via an A-boson exchange in the s-channel is36 by far the dominant process, since mA, which is smaller than 2 m ~ s pis , always very close to it as seen from Fig. 2. This sensitivity can be understood from Fig. 3, where V Z A is depicted versus mLSp for various mb(MZ)’s. We see that, as mb(MZ) decreases, mA increases and approaches 2 m ~ s p .The A-pole annihilation is then enhanced and RLsph2 is drastically reduced causing an increase of the upper bound on mLsp. For A?2 < 0.25, bino-stau c~annihilations~~ take over leading to a very pronounced reduction of R ~ s h2, p thereby enhancing the upper limit on mLSp.
For p > 0, a,(Mz) = 0.1185 and mb(MZ) = 2.888 GeV, we find the following allowed ranges of parameters: 176 GeV 5 mLSp
5 615 GeV,
0 5 A, 585tanP559, 0.145~50.17.
5 1.8,
7. The Inflationary Scenario One of the most promising inflationary scenarios is hybrid inflation37, which uses two real scalars: one which provides the vacuum energy for inflation and a second which is the slowly varying field during inflation. This scheme is naturally i n ~ o r p o r a t e din~ ~ SUSY GUTS, but in its standard realization has the following problem39: if the GUT gauge symmetry breaking predicts monopoles (and this is the case of Gps), they are copiously produced at the end of inflation leading to a cosmological catastrophe4’. One way to remedy this is to generate a shifted inflationary trajectory, so that GPS is already broken during inflation. This could be achievedl8 in our SUSY GUT model even before the introduction of the extra Higgs superfields, but only by utilizing non-renormalizable terms. However, the introduction of 4 and 4 very naturally gives rise17 to a shifted inflationary path with the use of renormalizable interactions only.
66
7.1. The Shifted Inflationary Path The superpotential terms which are relevant for inflation are given by
W = r;S(H"I? - M 2 ) - pSq52 + m&
+ X$H"H",
(17)
where M , m MGUT N 2.86 x 10l6 GeV, and K , p and X are dimensionless coupling constants with M , m, K , X > 0 by field redefinitions. For simplicity, we take p > 0. (The parameters are normalized so that they correspond to the couplings between the SM singlet components of the superfields.) The scalar potential obtained from W is given by N
+ 12pS4 - mr$I2 + Im4 + X H c 8 c \ 2 + x6l2 ( 1 ~ ~ +1 21 1 ~ 1 ~ ) + D - terms. (18)
V = I K ( H ' P - M 2 ) - ,@b2l2
+
(Ks
Vanishing of the D-terms yields H" * = ei'HC (H", H c lie in their right handed neutrino directions). We restrict ourselves to the direction with 2f) = 0 which contains the shifted inflationary path and the SUSY vacua (see below). Performing appropriate R and gauge transformations, we bring S , H" and H c to the positive real axis. From the potential in Eq. (18), we find that the SUSY vacuum lies at
HCHC
-2M2
II
(z)2
7
1 = - (I2<
4
2
J1-), = -.J"P (2) M '
(19)
with 5' = 0 and 6 = 0, where E = pX2M2/tcm2 < 1/4. The potential possesses a shifted flat direction (besides the trivial one) at
-+) H'H" M2 with S
2
=
2K2(&+1)+$ 2(K2 X2)
+
,
p / z ,(3
K
= --s, X
(20)
> 0 and a constant potential energy density VOgiven by
which can be used as inflationary path. VO# 0 breaks SUSY on this path, while the constant non-zero values of H", H" break the GUT gauge symmetry too. The SUSY breaking implies the existence of one-loop radiative correction^^^ which lift the classical flatness of this path, yielding the necessary inclination for driving the inflaton towards the SUSY vacuum. The one-loop radiative corrections to V along the shifted inflationary trajectory are calculated by using the Coleman-Weinberg formula42:
67
where the sum extends over all helicity states i, Fi and Mi2 are the fermion number and mass squared of the ith state, and A is a renormalization mass scale. In order to use this formula for creating a logarithmic slope which drives the canonically normalized real inflaton field (T = d m S / X towards the minimum, one has first to derive the mass spectrum of the model on the shifted inflationary path. This is a quite complicated task and we will skip it here.
7.2. Inflationary Observables The slow roll parameters are given by (see e.g. Ref. 43)
where the primes denote derivation with respect to the real normalized inflaton field (T and mp 21 2.44 x 10l8 GeV is the reduced Planck scale. The conditions for inflation to take place are E 5 1 and lql 5 1. Calculating the number of e-foldings NQ that our present horizon scale suffered during inflation, we obtain the following relation (see e.g. Ref. 43):
where uf [(TQ] is the value of o at the end of inflation [when our present horizon scale crossed outside the inflationary horizon] and T, N lo9 GeV is the reheat temperature taken to saturate the gravitino c o n ~ t r a i n t ~ ~ . The quadrupole anisotropy of the cosmic microwave background radiation can be calculated as follows (see e.g. Ref. 43):
which is its central value from the cosmic Fixing ( ~ T / T )E Q6.6 x background explorer (COBE)45 (assuming that the spectral index n = l ) , we can determine one of the free parameters (say ,6) in terms of the others (m, K and A). For instance, we find ,6 = 0.1, for m = 4.35 x 1015 GeV and IC = X = 3 x In this case, the instability point of the shifted path lies at (T, 2 3.55 x 10l6 GeV, af 21 1.7 x 1017 GeV and OQ N 1.6 x 10l8 GeV43. Also, M 21 2.66 x 10l6 GeV, NQ 21 57.7 and n N 0.98. Note that the slow roll conditions are violated and, thus, inflation ends well before reaching the instability point at ( T ~ We . see that the COBE constraint can be easily satisfied with natural values of the parameters. Moreover, superheavy SM
68
non-singlets with masses << MGUT,which could disturb the unification of the MSSM gauge coupling constants, are not encountered.
7.3. Supergravity Corrections Here, inflation takes place at quite high u ’ s . So, supergravity (SUGRA) corrections are important and could easily invalidate inflation, in contrast to the standard hybrid inflation case, where they can be kept46 under control. This catastrophe can be avoided by invoking a specific Kahler potential (used in no-scale SUGRA models) and a gauge singlet 2 with a similar Kahler potential, as is suggested in Ref. 47. Assuming a superheavy VEV for 2 via D-terms, one can achieve an exact cancellation of the inflaton mass corrections on the shifted path. So, inflation remains intact, but gets considerably corrected via the kinetic terms of (T. We find that, for the (T’S under consideration, the SUGRA corrections have only a small influence on (TQ if we use the same input values for the free increases parameters as in the global SUSY case. On the contrary, (ST/T)Q considerably. However, we can easily readjust the parameters so that the is now N 6.6 x COBE constraint is again met. For instance, (ST/T)Q obtained with m = 3.8 x 1015 GeV keeping K = X = 3 x lop2, p = 0.1 as in global SUSY. In this case, ( T ~N 2.7 x 10l6 GeV, ( ~ N f 1.8 x 1017 GeV and (TQ N 1.6 x lo1* GeV. Also, M 2: 2.6 x 1OI6 GeV, NQ 2: 57.5 and n 2: 0.99.
8. Conclusions We have reviewed the construction of a SUSY GUT model based on the PS gauge group which naturally yields a YQUC, allowing an acceptable b-quark mass within the CMSSM with p > 0. We found that there exists a wide and natural range of parameters consistent with the data on the CDM abundance in the universe, b + sy, the muon anomalous magnetic moment and the Higgs boson masses. Moreover, the model gives rise to a new version of the shifted hybrid inflationary scenario, which avoids overproduction of monopoles at the end of inflation by using only renormalizable interactions.
Acknowledgments We would like to thank M.E. Gbmez, R. Jeannerot and S. Khalil for fruitful and pleasant collaborations from which this work is culled. This work was supported by European Union under the RTN contracts HPRN-CT-200000148 and HPRN-CT-2000-00152.
69
References 1. G. L. Kane, C. Kolda, L. Roszkowski and J. D. Wells, Phys. Rev. D49, 6173 (1994), hepph/9312272. 2. G. Lazarides and C. Panagiotakopoulos, Phys. Lett. B337, 90 (1994), hepph/9403316; S. Khalil, G. Lazarides and C. Pallis, ibid. 508, 327 (2001), hep-ph/0005021. 3. L. Hall, R. Rattazzi and U. Sarid, Phys. Rev. D 50, 7048 (1994), hepph/9306309; M. Carena, M. Olechowski, S. Pokorski and C. E. M. Wagner, Nucl. Phys. B426, 269 (1994), hepph/9402253. 4. D. Pierce, J . Bagger, K. Matchev and R. Zhang, Nucl. Phys. B491,3 (1997), hepph/9606211. 5. M. Carena, D. Garcia, U. Nierste and C. E. M. Wagner, Nucl. Phys. B577, 88 (2000), hep-ph/9912516. 6. S. F. King and M. Oliveira, Phys. Rev. D63,015010 (200l), hepph/0008183. 7. S. Abel et al. (SUGRA Working Group Collaboration), hep-ph/0003154. 8. M. E. G6mez, G. Lazarides and C. Pallis, Nucl. Phys. B638, 165 (2002), hepph/0203131. 9. H. Baer, J. Ferrandis, K. Melnikov and X. Tata, Phys. Rev. D66, 074007 (2002), hepph/0207126. 10. T. Blaiek, R. DermEek and S. Raby, Phys. Rev. Lett. 88, 111804 (2002), hepph/0107097; Phys. Rev. D65, 115004 (2002), hepph/0201081. 11. D. Auto et. al., J . High Energy Phys. 06, 023 (2003), hep-ph/0302155. 12. U. Chattopadhyay, A. Corsetti and P. Nath, Phys. Rev. D66, 035003 (2002), hepph/0201001; C. Pallis, Nucl. Phys. B678, 398 (2004), hep-ph/0304047. 13. G. Lazarides, Q. Shafi and C. Wetterich, Nucl. Phys. B181, 287 (1981); G. Lazarides and Q. Shafi, ibid. B350, 179 (1991). 14. M. E. G6mez, G. Lazarides and C. Pallis, Phys. Rev. D67, 097701 (2003), hepph/0301064; C. Pallis and M.E. G6mez, hep-ph/0303098. 15. D. Spergel et al., Astrophys. J . Suppl. 148, 175 (2003), astro-ph/0302209. 16. R. Barate et al. (ALEPH Collaboration), Phys. Lett. B429, 169 (1998); K. Abe et al. (BELLE Collaboration), ibid. 511, 151 (2001), hep-ex/0103042; S. Chen et al. (CLEO Collaboration), Phys. Rev. Lett. 87, 251807 (2001), hepex/0108032. 17. R. Jeannerot, S. Khalil and G. Lazarides, J . High Energy Phys. 07, 069 (2002), hep-ph/0207244. 18. R. Jeannerot, S. Khalil, G. Lazarides and Q. Shafi, J. High Energy Phys. 10, 012 (2000), hepph/0002151. 19. G. Lazarides, hepph/0011130; R. Jeannerot, S. Khalil and G. Lazarides, hepph/0106035. 20. I. Antoniadis and G. K. Leontaris, Phys. Lett. B216, 333 (1989). 21. G. Lazarides and Q. Shafi, Phys. Rev. D58,071702 (1998), hepph/9803397. 22. M. E. G6mez, G. Lazarides and C. Pallis, Phys. Rev. D61, 123512 (2000), hepph/9907261; Phys. Lett. B487, 313 (2000), hepph/0004028. 23. M. E. G6mez and C. Pallis, hepph/0303094 (in the SUSYO2 Proceedings). 24. J . Ellis, K. A. Olive, Y. Santoso and V.C. Spanos, Phys. Lett. B565, 176
70
25.
26. 27. 28. 29. 30.
31. 32. 33. 34.
35. 36. 37. 38. 39.
40. 41. 42. 43. 44. 45. 46. 47.
(2003), hepph/0303043; A. B. Lahanas and D. V. Nanopoulos, ibid. 568,55 (2003), hep-ph/0303130; H. Baer and C. Balbs, JCAP 05, 006 (2003), hepph/0303114; U. Chattopadhyay, A. Corsetti and P. Nath, Phys. Rev. D68, 035005 (2003), hepph/0303201. H. Goldberg, Phys. Rev. Lett. 50, 1419 (1983); J. R. Ellis, J. S. Hagelin, D. V. Nanopoulos, K. A. Olive and M. Srednicki, Nucl. Phys. B238, 453 (1984). G. BBlanger, F. Boudjema, A. Pukhov and A. Semenov, Comput. Phys. Commun. 149, 103 (2002), hep-ph/0112278. J. Ellis, T. Falk, K. A. Olive and M. Srednicki, Astropart. Phys. E13, 181 (2000); ibid. 15, 413 (2001), hepph/9905481. A. Djouadi, J. Kalinowski and M. Spira, Comput. Phys. Commun. 108, 56 (1998), hepph/9704448. A. L. Kagan and M. Neubert, Eur. Phys. J. C7, 5 (1999), hepph/9805303; P. Gambino and M. Misiak, Nucl. Phys. B611,338 (2001), hepph/0104034. M. Ciuchini, G. Degrassi, P. Gambino and G. Giudice, Nucl. Phys. B527, 21 (1998), hepph/9710335; G. Degrassi, P. Gambino and G. F. Giudice, J . High Energy Phys. 12, 009 (2000), hepph/0009337. S. Martin and J. Wells, Phys. Rev. D64, 035003 (2001), hep-ph/0103067. M. Davier, hepex/0312065 (to appear in the SIGHAD03 Proceedings). G. W. Bennett et al. (Muon 9-2 Collaboration), Phys. Rev. Lett. E89, 101804 (2002); ibid. 89, 129903 (2002), hepex/0208001. ALEPH, DELPHI, L3 and OPAL Collaborations, The LEP Higgs working group for Higgs boson searches, hepex/0107029; LHWG-NOTE/2002-01, [http://lephiggs.web.cern.ch/LEPHIGGS/papers/July2002SM/index.html] . S. Heinemeyer, W. Hollik and G. Weiglein, hepph/0002213. A. B. Lahanas, D. V. Nanopoulos and V. C. Spanos, Phys. Rev. D62,023515 (ZOOO), hepph/9909497. A. D. Linde, Phys. Rev. D49, 748 (1994), astro-ph/9307002. E. J. Copeland, A. R. Liddle, D. H. Lyth, E. D. Stewart and D. Wands, Phys. Rev. D49, 6410 (1994), astro-ph/9401011. G. Lazarides and C. Panagiotakopoulos, Phys. Rev. D52, 559 (1995), hepph/9506325; R. Jeannerot, S. Khalil and G. Lazarides, Phys. Lett. B506, 344 (2001), hepph/0103229. T. W. B. Kibble, J. Phys. A9, 1387 (1976). G. Dvali, R. Schaefer and Q. Shafi, Phys. Rev. D73, 1886 (1994), hepph/9406319. S. Coleman and E. Weinberg, Phys. Rev. D7, 1888 (1973). G. Lazarides, Lect. Notes Phys. 592, 351 (2002), hep-ph/0111328; hepph/0204294. M. Yu. Khlopov and A. D. Linde, Phys. Lett. B138, 265 (1984); J. Ellis, J. E. Kim and D. V. Nanopoulos, ibid. 145, 181 (1984). C. L. Bennett et al., Astrophys. J . L1, 464 (1996), astro-ph/9601067. G. Lazarides, R. K. Schaefer and Q. Shafi, Phys. Rev. D56, 1324 (1997), hepph/9608256. C. Panagiotakopoulos, Phys. Lett. B459, 473 (1999), hep-ph/9904284.
SUPERSYMMETRIC GRAND UNIFICATION: THE QUEST FOR THE THEORY
A. M E L F O ( ~ )G. , SENJANOVIC(~) ('1 Centro d e Fisica Fundamental, Universidad de Los Andes, Mkrida, Venezuela (2) International
Centre for Theoretical Physics, 34100 Trieste, Italy
With the advent of neutrino masses, it has become more and more acknowledged that SO(10) is a more suitable theory than SU(5): it leads naturally t o small neutrino masses via the see-saw mechanism, it has a simpler and more predictive Yukawa sector. There is however a rather strong disagreement on what the minimal consistent SO(10) theory is, i.e. what the Higgs sector is. The issue is particularly sensitive in the context of low-energy supersymmetry.
1. Introduction Supersymmetric Grand Unification has been one of the main extensions of the Standard Model (SM) for now more than two decades. Today, however, it is in search of a universally accepted minimal, consistent model. With the growing evidence for neutrino masses,' it is becoming more and more clear that the SU(5) theory is not good enough: it contains too many parameters in the Yukawa sector. The situation is much more appealing in the SO( 10) scenario, which is custom fit to explain small neutrino masses in a simple and fairly predictive manner. The main dispute lies in the breaking of SO(10) down to the Minimal Supersymmetric Standard Model (MSSM), in the delicate question of the choice of the Higgs superfields. Roughly speaking, there are two schools of thought: one that sticks to the small representations, which guarantees asymptotic freedom above MGUT,but must make use of higher dimensional operators, suppressed by M p l ; one that argues in favor of the renormalizable theory only, even at the price of becoming strong between MGUT and the Planck scale. Each program has its pros and cons. The first one in a sense goes beyond grand unification by appealing to the string picture in order to provide additional horizontal symmetries needed to simplify the theory plagued by many cou71
72
plings. The second one is based on pure grand unification, with the hope that the Planck scale physics plays a negligible role. It is the second one that we discuss at length in this talk.
2. Why grand unification and why supersymmetry ?
No excuse needs to be offered for the natural wish to unify the strong and electro-weak interactions. This appealing idea has two important generic features: proton decay and the existence of magnetic monopoles. They are by themselves sufficient reason to pursue the unification scenario. There are three important reasons to incorporate low-energy supersymmetry in this program: i) the hierarchy problem of the Higgs mass, ii) the gauge coupling unification, and iii) the Higgs mechanism in the form of radiative symmetry breaking. Let us briefly discuss them. 0
0
0
Supersymmetry per se says nothing about the smallness of the Higgs mass (the hierarchy problem), it just keeps the perturbative effect small, the way that chiral symmetries protect small Yukawa couplings. The old feelings that this might not be such a big deal, since the cosmological constant does not get protected in a similar way, are becoming more widespread today. Gauge coupling unification of the MSSM is a rather remarkable phenomenon, but its meaning is not completely clear. Namely, if one believes in a desert between MWand M G ~ T then , this becomes a crucial ingredient. The desert is a property of the minimal gauge group, SU(5), which is not a good theory of neutrino masses. In SO(lO), on the other hand, supersymmetry is not essential at all; the theory works even better without supersymmetry since then it predicts intermediate scales in the range lo1' - 1014GeV, ideal for neutrino masses via the see-saw mechanism13and for lept~genesis.~ It is worth stressing though that supersymmetric grand unification was anticipated already in 1981, and it gave a rationale for a heavy top quark with a mass around 200GeV (needed to increase the p parameter and help change sin2Bw from its then accepted value of 0.20 to the current 0.23). Radiative symmetry breaking and the Higgs mechanism. The tachyonic property of the Higgs mass term has bothered people for a long time. It is of course purely a question of taste, for either sign of M$ is equally probable. Since the charged sfermion mass terms are definitely not tachyonic, in supersymmetry one could
73
ask why is the Higgs scalar so special. The answer is rather simple: if the nature if the scalar mass terms is determined by some large scale, and if all m2 > 0, it turns out that the Higgs doublet coupled to the top quark rather naturally becomes tachyonic at low scales due to the large top Yukawa coupling. This was kind of prophetic more than twenty years ago, and it could be a rationale for such a heavy top quark. Admittedly a little fine-tuning is still needed between the so-called p term and the stop mass, but wether this is a small or a large problem is still disputable. Suppose we accept low energy supersymmetry as natural in grand unification. We then face the task of identifying the minimal consistent supersymmetric grand unified theory and then hope that it will be confirmed by experiment. Due to the miraculous gauge coupling unification in the MSSM, we are then tempted to accept the idea of the desert. Since the desert is a natural property of SU(5), it is not surprising that SU(5) was considered for a long time the main candidate for a supersymmetric GUT. Why not stick to this idea? The point is that SU(5), at least in its simplest form, favors massless neutrinos. Higher dimensional operators can only provide small neutrino masses, not large enough to explain Amzdot I I 10-'eV2 and especially Am; I I10-3eV2. While we can try to remedy this in one way or another, the fact is that the realistic theory needs too many parameters to be predictive. Let us try to justify this claim by analyzing SU(5) in some detail.
3. Supersymmetric SU(5) The minimal Higgs sector needed to break the symmetry completely down U ( l ) e mx s U ( 3 ) , consists of the adjoint 2 4 and ~ two fundamentals 5 8 and 5 ~ The . Higgs superpotential is quite simple
td
+ X ( 2 4 ~ +) @~ H
W H =m(24p)
5H
+a 5 H 2
4 S ~H
(1)
and so is the Yukawa one WY = Y d 1oF 5 F 5 H
+ y.u 1oF 1oF 5 H
(2)
since the charged fermions belong to 5~ and 1 0 ~ . The above theory is usually called the minimal supersymmetric SU(5) theory. It apparently has a small number of parameters: 3 real
Yd
(after diagonalization)
74
0
0 0
2 x 6 = 12 real yu (yu is a symmetric matrix in generation space) 2 real p , m (after rotations) 2 x 2 = 4 real A, a
In total, 21 real parameters. The trouble is that the theory fails badly.5 Neutrinos are massless, and thus Ve = 1; furthermore, the relations me = md at MCUT fail, except for the third generation. The most conservative approach would be to blame the failure in the absence of higher dimensional operators, this way no change in the structure of the theory is needed. Neutrino masses are then given by the Weinberg-type operator
W," = y v
5 5 5 5
H
MPl
(3)
giving 0
6 x 2 = 12 new real parameters in yv.
Similarly, we must add higher dimensional operators to (2),
(we have omitted SU(5) indices, and represented contraction with indices in 2 4 with ~ a bracket). This means 0
9x 2x2
+ 6 x 2 x 2 = 60 real parameters
and then most predictability is gone. What remains? First, mb = mT is still there, still a success. Gauge couplings unify as we know, but the GUT scale is not predicted precisely as often claimed in the literature. The point is that, for the sake of consistency, one should add higher dimensional operators to (l),so that one expects AWH = CI-
Tr24; MP1
(33-24:)' +cz
Mp1
(5)
and if the coupling X in (1) were to be small, these terms would become important. But X is a Yukawa type coupling, i.e. it is self-renormalizable, so it can be naturally small. This point is worth discussing further. At the renormalizable tree level, one gets the same masses for the color octet and the s U ( 2 ) ~triplet in 2 4 ~ :m8 = m3. This is almost always assumed when the running from MGUT to MW is studied. Now, if X is small, the ci terms in ( 5 ) can dominate; if so, one gets m3 = 4mg. This
75
fact alone suffices to increase MGUT by an order of magnitude above the usually quoted value MSUT N 10l'GeV (calculated with ci = 0). Similarly, the masses of the colored triplets T and T in 5H and S H would get increased by a factor of about 30, and the d = 5-induced proton life-time by about lo3. More precisely, one obtains
where the superscript
denotes the tree-level value m3 = ma. In this case
ma
N
M&JT MP1
so that
With MSUT N 101'GeV, this means
MGUT N lOM&,, , mT
N
32mT. 0
(10)
It should be stressed that X small is natural technically, as much as a small electron Yukawa coupling. Taking X N 0(1)and ruling out the theory would be equivalent to finding that the SM does not work with all the Yukawa couplings being of order one, and insisting on this as if you didn't know fermion masses. Taking into account non-renormalizable interactions can thus save the theory. It is important to recall that without them, the minimal SU(5) does not make sense anyway, predicting as it does m, = 0 and md = me;once this is corrected the theory is still valid. Of course, if one prefers the renormalizable theory, one needs new states such as 45H (in order to correct md = me),or 1 5 to~ give neutrino masses, or three (at least two) singlet right-handed neutrinos. This introduces even more uncertainties in the computations of MGUT and ~ p In . short, the minimal realistic supersymmetric SU(5) theory is not yet ruled out. It is indispensable to improve the experimental limit on ~p by two-three orders of magnitude. Grand unification needs desperately a new generation of proton decay experiments.
76
In the supersymmetric version of SU(5), there is yet another drawback.
As much as the MSSM, it allows for the d = 4 proton decay through terms like
AW = mX' 1 0 S ~F S F
(11)
which contains both
+
A' (vcDcDc QLD") (12) This is a disaster (unless A' 5 10-l'). A way out is assumed through the imposition of R-parity, or equivalently matter parity M : F -+ -F, H -+ H , where F stands for the fermionic (matter) superfields and H for the Higgs ones. Grand unification ought to do better than this, and SO(10) does it as we shall see. In any case, SU(5) does a poor job in the neutrino sector and in the charged fermion sector it is either incomplete or it has too many parameters. One would have to include extra horizontal symmetries, and this route is in some sense beyond grand unification and often needs strings attached. If we stick to the pure grand unification, we better move on to SO(10). 4. Towards unification: Pati-Salam symmetry Quark-Lepton unification can be considered a first step towards the complete SO( 10) unification of a family of fermions in a single representation. Many interesting features of SO(10) GUTS, such as a renormalizable seesaw and R-parity conservation, are already present in partial unification based on the Pati-Salam group GPS = SU(4), x S u ( 2 ) ~ x S u ( 2 ) ~so , it is instructive to review the situation there. Later, when we turn to SO(lO), decomposition of representation under the Pati-Salam subgroup will prove to be the most useful. To simplify the discussion, imagine a two-step breaking of the PS symmetry down to the MSSM
- -
s u ( 2 ) L x s u ( 2 ) R x su(4)c
su(2)L x s u ( 2 ) R x U(1)B-L x S u ( 3 ) ~ SU(2)L x W(1)y x SU(3),.
(13)
The first steps breaks GPS down to its maximal subgroup, the LR (LeftRight) group,8 and it is simply achieved through the vev of and adjoint representation (the numbers in parenthesis indicate the GPS representations)
A = (15,1,1).
(14)
77
In turn, the breaking of the L R group can be achieved by having s U ( 2 ) ~ triplets fields, with B - L = 2, acquiring a vev. Triplets will couple to ferniions and give a mass to right-handed neutrino, providing the see-saw mechanism at the renormalizable level. Right-handed doublets could also do the job, but then non-renormalizable operators have to be invoked, which means effective operators resulting from a new theory at a higher scale, but this theory we will discuss explicitly in the next section. There is a more profound reason for preferring the triplets. They have an even B - L number, and thus preserve matter parity as we defined above. This in turn means R-parity is not broken at a high scale. But then it can be easily shown that it cannot be broken afterwards, at the low energy supersymmetry breaking or electroweak scale. More precisely, a spontaneous breakdown of R-parity through the sneutrino VEV (the only candidate) would result in the existence of a pseudo-Majoron with its mass inversely proportional to the right-handed neutrino mass. This is ruled out by the Z decay ~ i d t h . ~ ? This l O fact is completely analogous to the impossibility of breaking R-parity spontaneously in the MSSM, where the Majoron is strictly massless. In terms of PS representations, the LR triplets are contained in the fields
C(3,1, lo), Z(3,1, fO), Cc(1,3, CO), zc(l,3,lO).
(15)
The matter supermultiplets are
$J(2,1,4), &(I, 2,a)
(16)
and the minimal light Higgs multiplet is
@,
231).
(17)
The most general superpotential for the fields (15) is
+
W = mTrA2 + M T r ( C Z + CcCc) TT(CAC- C,AC,)
(18)
where we assume the following transformation properties under Parity C4C,,
Z-C,,
A + -A.
(19)
We choose A to be a parity-odd field in order to avoid flat directions connecting left- and right-breaking minima. It is straightforward to show that the SM singlets in A , C, and C, take vevs in the required directions to achieve the (in principle two-step) symmetry breaking
=Mc
=Mr,
=Mr,
(20)
78
with
As discussed in detail in,l1?l2the SU(2)~-breakingvev lies in a flat direction that connect them with charge-breaking vacua. It can be eliminated if the soft breaking terms break also su(2)R. If not, one would have to appeal to operators coming from a more complete theory as studied in the next section. The interesting point here is that the breaking in the minimal model leaves a number of fields potentially light.13 There is a larger, accidental S U ( 3 ) symmetry broken down to SU(2) by the right-handed triplet fields, hence five Nambu-Goldstone bosons. But the gauge symmetry s u ( 2 ) R x U ( ~ ) B - Lis broken down to U(l)y, so that three of them are eaten, leaving us with states S$+, 8$+ that acquire a mass only at the scale of supersymmetry breaking. These states are common in supersymmetric theories that include the Left-Right group, and have been subject of experimental search.14 In a similar way, a color octet in A has a mass of order M:/M,, and could in principle be light. The unification constraints give the interesting possibility 103GeV 5 M R
< 107GeV
10"GeV
5 M, 5 10I4GeV
opening up the possibility of the LHC discovering them at the TeV scale. For larger M R , which would be necessary if one wants to fit neutrino masses without additional fine-tuning, these particles become less accessible to experiment. However, the large number of fields in this theory implies the loss of perturbativity at a scale around lOM,, and non-renormalizable effects suppressed by this new fundamental scale can be shown to guarantee that they have comparable masses.12 Namely, if these effects are included, the only consistent possibility is the single-step breaking
MR N M,
I IlOl0GeV
(22)
Surely the most interesting feature of a low scale of PS symmetry breaking is the possibility of having U ( ~ ) B - Lmonopoles, with mass r n =~ lOM,. If produced in a phase transition via the Kibble mechanism, the requirement that their density be less than the critical density then implies M, 1012GeV. We see that the single-step breaking at M , M R lOl0GeV (in a theory including non-renormalizable terms) offers the interesting possibility of potentially detectable intermediate mass monopoles, as long as one manages to get rid of the false vacuum problem of supersymmetric theories.
- -
<
79
One final note about PS symmetry and neutrino masses. In LR theories the see-saw mechanism is in general non-canonical, or type 11. That is, there is a direct left-handed neutrino mass from the induced vev of the left-handed triplet fields in C (which we shall call A ) 1571671791
N<
A
>2L
M&/MR.
Namely, in non- supersymmetric theories the symmetry allows for a coupling in the potential
A V = XA+2A"+ M2A2
(23)
resulting in
In supersymmetry such terms are of course not present, but one could have interactions with, for example, a heavy field S transforming as (1,1,3) under GPS
W = 42S+ AA'S
+MS2.
(25)
Integrating out S would then give a contribution
1
AW = -A+2Ac M producing the required small vev. Or one could have a couple of heavy fields X = (2,2, TO) and
8 = (2,2, lo), which through terms like
W = $AX
+ +AcX + MXX
(27)
would give the same effect. These representations in fact are the ones appearing in the minimal SO(10) theory of the next section. The absence of the S,X, 8 fields in the minimal PS theory guarantees a type I see-saw at the supersymmetric level. Breaking of supersymmetry can generate a nonvanishing but negligible vev for A:"
which contributes by a tiny factor (m3p/Mc)25 to the usual see-saw mass term m, 2~ m&/m,,. In short, the minimal PS model has a clean, type I see-saw. In spite of providing only a partial unification, PS theory has interesting features, namely potentially light states and the possibility of intermediate monopoles, that could be a way of differentiating it from other theories at
80
a high scale. We are however interested in grand unification, so let us move on. 5. SO(10) grand unified theory
If not for anything else, but for the fact that matter parity is a finite gauge rotation, SO(10) would be a better candidate for a supersymmetric GUT. But, as is well known, it also unifies a family of fermions, has charge conjugation as a gauge transformation, has right-handed neutrinos an through the see-saw mechanism leads naturally t o small neutrino masses. And, most important, at the renormalizable level, if one is willing to accept large representations, it has fewer parameters than SU(5). We will elaborate on this point as we go along. The issue here, and the main source of dispute among the experts in the field, is the choice of the Higgs sector. Before deciding on this, a comment on d = 4 proton decay in the MSSM is in order. The basic problem is the impossibility of distinguishing the leptonic and the Higgs doublets, both being superfields. This persists in SU(5) where you have both 5~ and 5~ superfields. In SO(10) fermions belong t o 1 6 and ~ the "light" Higgs to 1 0 ~ This . difference should be taken seriously, and all efforts should be made t o maintain it. Not all researchers agree on this. Certainly not the people who pursue minimality by choosing mall representations, like the set ( 45H, 1 6 ~i ,6 ~ ) , in order t o break SO(10) down to S U ( 3 ) , x s U ( 2 ) x~ U(1)y. This way, through ( 1 6 ~ = ) ( 1 6 ~ #) 0, matter parity will be broken at MGUT;hence the catastrophic d = 4 proton decay. One is then forced t o postulate extra discrete symmetries in order to save the theory. In any case, more flavor symmetries are needed, since both the symmetry breaking and the fermion masses need higher-dimensional, Planck suppressed operators whose number is rather large (at least thirteen complex couplings in W H ,the Higgs superpotential). The Yukawa superpotential takes the form
wy= ylo 1 6 r~ l o H +- 1 [c1(i6Fri6F) ( 1 6 ~ r i 6 +~ c) 1 ( 1 6 ~ I ' 1 6 ~( i)6 H r T G ) MP1 ~ 3 1 6 ~ r 45H ~ 116 0~+~ . (29) At MGUT,one arrives at the prediction
81
which works very well as we know. Also, the see-saw takes the so-called type I form. From (29), mvR N
c4- M ~ U T 21
MPl
1012 -
GeV
(31)
which fits nicely the light neutrino masses. The type I1 contribution, obtained when 1 6 gets ~ a small vev N M w ,
is too small to explain either atmospheric or solar v data (maybe relevant for small mass splits in the case of degenerate neutrinos). Once flavor symmetries are added, one can do the texture exercise and look for the most appealing model. But this program goes beyond the scope of grand unification. We ought to try to construct the minimal realistic supersymmetric GUT without invoking any new physics.
5.1. The pure renormalixable supersymmetric S O ( l 0 ) Such a theory is easily b ~ i l t ~ ’with - ~ ~ large representations in the Higgs sector
+ --
2108, 1 2 6 ~ 1 2 6 ~1oH , with this content the theory is not asymptotically free any more above kfGuT,23 and the s o ( 1 0 ) gauge couplings becomes strong at the scale Ap 5 1 0 M ~ u The ~ . Higgs superpotential is surprisingly simple
+ m126126H126H + rnlo(10H)~+ ~ ( 2 1 0 ~ ) ~
W H = m210(210H)~
+
+
+ ~ 1 2 6 ~ 1 2 6 H 2 1 0~~ ~ 1 0 ~ 1 2 6 ~ Z2 i1! 01 ~0 ~ ~ 2 1. 0(33) ~ With ( 2 1 0 ~ # ) 0 and ( 1 2 6 ~ = ) (126~) # 0, SO(10) gets broken down to the MSSM, and then ( 1 0 ~ completes ) the job in the usual manner. The Yukawa sector is even more simple W Y = 16F(?/lO1OH
+ ?/126mH)16F
with only 3 real (say) ylo couplings after diagonalization, and 6 x 2 symmetric ?/I26 couplings, 15 in total. From the a and B terms on gets
+
WH = ... a(27 2, 1)lO (2, 27 15)126 f (1,1,15)210 +5(2,2,1)10 ( 2 , 2 , 1 5 ) m (1,1,15)210 f
+
(34) =
12
(35)
82
Now, the success of gauge coupling unification in the MSSM favors a single step breaking of SO(lO), so that ((1,1,15)210) N MGUT. In other words, the light Higgs is a mixture24 of (at least) (2,2,1)10 and ( 2 , 2 , 1 5 ) m ; equivalently ( ( 2 , 2 , 1 5 ) m ) = ((272,l)lO).
(36)
Since ( 2 , 2 , 1 5 ) m is an adjoint of SU(4),, being traceless it give me = -3md, unlike ((2,2,1)10), which implies me = md. In other words, the ( 1 0 ~ must ) be responsible for the mb 2: m, relation at MGUT, and the (126~) for the m, N 3m, relation a t MGUT. In this theory, the GeorgiJarlskog program becomes automatic. ~ Of course, we don’t know anymore why mb N m,, or why 1 0 dominates; admittedly a loss. But not all is lost. Since ( 1 0 ~ = ) ((2,2,1)) is a Pati-Salam singlet, the difference between down quark and charged lepton mass matrices must come purely from (126H)
Md
- Me
0: Y126
.
(37)
Suppose the see-saw mechanism is dominated by the so-called type 11: this is equivalent t o neutrino masses being due t o the triplet (3,1,m)126, with
In other words
or
MvmMd-Me.
(40)
Let us now look a t the 2nd and 3rd generations first. In the basis of diagonal M e , and for the small mixing Cde
obviously, large atmospheric mixing can only be obtained for mb N m,.25 Of course, there was no reason whatsoever to assume type 11 see- saw. Actually, we should reverse the argument: the experimental fact of mb N m, at MGUT,and large Oatrn seem t o favor the type type I1 see-saw. It can be shown, in the same approximation of 2-3 generations, that type I cannot dominate: it gives a small Oatm.26This gives hope t o disentangle
83
the nature of the see-saw in this theory. As a check, it can be shown that the two types of see-saw are really inequivalent.26 The three generation numerical studies supported a type I1 see- saw. with the interesting prediction of a large 613 and a hierarchical neutrino mass spectrum.27 Somewhat better fits are obtained with a small contribution of 1 2 0 H 28 or higher dimensional operators.2g Type 1can apparently be saved with CP phases, see Ref. 30 (for earlier work on type I see Ref. 31.). 5.2. Unification constraints
It is certainly appealing to have an intermediate see-saw mass scale M R , between 1 O I 2 - 1015GeV or so. In the non-renormalizable case, with 1 6 ~ and Z H ,this is precisely what happens: MR N cM&T/Mpl 21 c(1013 1014)GeV.In the renormalizable case, with 1 2 6 ~and 1268,one needs to perform a renormalization group study using unification constraints. While this is in principle possible, in practice it is hard due to the large number of fields. The stage has recently been set, for all the particle masses were c ~ m p u t e d ,and ~ ~ the > ~ preliminary ~ studies show that the situation may be under control.34 It is interesting that the existence of intermediate mass scales lowers the GUT scale32>35(as was found before in models with 5 4 H and 4 5 H ) 3 6 , allowing for a possibly observable d = 6 proton decay. Notice that a complete study is basically impossible. In order to perform the running, you need to know particle masses precisely. Now, suppose you stick to the principle of minimal fine-tuning. As an example, you fine-tune the mass of the W and 2 in the SM, then you know that the Higgs mass and the fermion masses are at the same scale
-
mH
=
-mw, .\/A
m f = -mw Yf
9
,
9
where X is a 44 coupling, and yf an appropriate fermionic Yukawa coupling. Of course, you know the fermion masses in the SM model, and you know mH zmw. In an analogous manner, at some large scale mG a group G is broken and there are usually a number of states that lie at mG, with masses
mi = aimG
,
(43)
where ai is an approximate dimensionless coupling. Most renormalization group studies typically argue that ai N 0 ( 1 ) is natural, and rely on that heavily. In the SM, you could then take mH N mw, m f N mw; while reasonable for the Higgs, it is nonsense for the fermions (except for the top
84
quark). In supersymmetry all the couplings are of Yukawa type, i.e. selfrenormalizable, and thus taking ai = 0(1)may be as wrong as taking all yf 21 O(1). While a possibly reasonable approach when trying to get a qualitative idea of a theory, it is clearly unacceptable when a high-precision study of M G ~ is T called for. 5.3. Proton decay
As you know, d = 6 proton decay gives r p ( d = 6) cx M&,T, while (d = 5 ) gives r p ( d = 5 ) cx M&,T. In view of the discussion above, the highprecision determination of rp appears almost impossible in SO(10) (and even in SU(5)). Preliminary studies 37 indicate fast d = 5 decay as expected. We are ignoring the higher dimensional operators of order MGUT/MP~ N If they are present with the coefficients of order one, we can forget almost everything we said about the predictions, especially in the Yukawa sector. However, we actually know that the presence of 1/Mpl operators is not automatic (at least not with the coefficients of order 1). Operators of the type (in symbolic notation)
are allowed by SO(10) and they give
These are the well-known d = 5 proton decay operators, and for c N 0(1) they give rpN 1023yr. Agreement with experiment requires c5
(46)
Could this be a signal that 1/Mp1 operators are small in general? Alternatively, you need to understand why just this one is to be so small. It is appealing to assume that this may be generic; if so, neglecting 1/Mpl contributions in the study of fermion masses and mixings is fully justified. 5.4. Leptogenesis
The see-saw mechanism provides a natural framework for baryogenesis through leptogenesis, obtained by the out-of-equilibrium decay of heavy right-handed neutrino^.^ This works nicely for large M R , in a sense too nicely. Already type I see-saw works by itself, but the presence of the type I1 term makes things more complicated. One cannot be a priori sure
85
whether the decay of right-handed neutrinos or the heavy Higgs triplets is responsible for the asymmetry, although the hierarchy of Yukawa couplings points towards UR decay. In the type I1 see-saw, the most natural scenario is the VR decay, but with the triplets running in the This and related issues are now under i n v e ~ t i g a t i o n . ~ ~
6. Summary and Outlook We have argued in favor of SO(10) as the minimal consistent supersymmetric grand unified theory. It includes all the interesting features of Left-Right and Pati-Salam symmetries, it is the ideal setting for a see-saw mechanism, and has the MSSM with automatic R-parity as the low energy limit. It can give connections with low energy phenomenology, such as the one relating b - r unification with neutrino mixings, besides being able to provide realistic charged fermion spectrum. As a gauge symmetry accommodating all fermions of one generation in a single representation, there is little doubt on the convenience of SO(10). The question of the Higgs sector is the one unsolved: one can choose between two different approaches. One can insist on perturbativity all the way to the Planck scale and choose small representations, using then 1/Mpl operators to generate the physically acceptable superpotential; it is then necessary to use textures to simplify the theory. In this sense, this programme appeals to physics beyond grand unification. The other approach is to stick to the pure SO(10) theory, at the expense of using very large representations. The couplings then become strong at AF = OMG GUT, but the theory has the advantage of requiring only a small number of couplings, and is a complete theory of matter and non-gravitational interactions. The important question is rather if these versions of the theory can be tested in the near future. Work is in progress by several groups on the possibility of establishing testable constraints on neutrino masses and mixings, proton decay, and the implementation of the leptogenesis scenario. In the pure SO(10) approach, with less parameters, proving the theory wrong might be just a question of time.
7. Acknowledgements We wish to acknowledge many discussions and enjoyable collaboration in the subjects of this talk with Charan Aulakh, Borut Bajc, F'rancesco Vissani, Andrija Raiin, Pave1 Fileviez-PQrez and Thomas Hambye. And we thank the organizers of BW2003, in particular Goran Djordjevic, for a
86
stimulating conference and a good time in VrnjaEka Banja. A. to say hvala, bre.
M. wishes
References 1. For some recent reviews see e.g. B. Bajc, F. Nesti, G. Senjanovic and F. Vissani, “Perspectives in neutrino physics,” Proceedings of 17th Rencontres de Physique de la Vallee d’Aoste, La Thuile, 9-15 Mar 2003, M. Greco ed., page 103-143; S. M. Bilenky, C. Giunti, J. A. Grifols and E. Masso, Phys. Rept. 379 (2003) 69; V. Barger, D. Marfatia and K . Whisnant, J. Mod. Phys. E 12 (2003) 569. A. Y. Smirnov, Int. J. Mod. Phys. A19 (2004) 1180. 2. S. Dimopoulos, S. Raby, F. Wilczek, Phys. Rev. D24 (1981) 1681. L.E. IbAiiez, G.G. Ross, Phys. Lett. B105 (1981) 439. M.B. Einhorn, D.R. Jones, Nucl. Phys. B196 (1982) 475. W. Marciano, G. SenjanoviC, Phys. Rev. D25 (1982) 3092. 3. P. Minkowski, Phys. Lett. B67 (1977) 421. T.Yanagida, proceedings of the Workshop on Unified Theories and Baryon Number in the Universe, Tsukuba, 1979, eds. A. Sawada, A. Sugamoto, KEK Report No. 79-18, Tsukuba. S. Glashow, in Quarks and Leptons, CargZse 1979, eds. M. L6vy. et al., (Plenum, 1980, New York). M. Gell-Mann, P. Ramond, R. Slansky, proceedings of the Supergmvity Stony Brook Workshop, New York, 1979, eds. P. Van Niewenhuizen, D. Freeman (North-Holland, Amsterdam). R. Mohapatra, G. SenjanoviC, Phys. Rev. Lett. 44 (1980) 912. 4. M. Fukugita and T . Yanagida, Phys. Lett. B174 (1986) 45. 5. M. S. Chanowitz, J. R. Ellis and M. K. Gaillard, Nucl. Phys. B128, 506 (1977). A. J. Buras, J. R. Ellis, M. K. Gaillard and D. V. Nanopoulos, Nucl. Phys. B135 (1978) 66. 6. R. N. Mohapatra, Phys. Rev. D34 (1986) 3457. A. Font, L. E. IbGiez and F. Quevedo, Phys. Lett. B228 (1989) 79. S. P. Martin, Phys. Rev. D46 (1992) 2769. 7. J. C. Pati and A. Salam, Phys. Rev. D10 (1974) 275. 8. R. N. Mohapatra and J. C. Pati, Phys. Rev. D11 (1975) 2558. G. SenjanoviC and R. N. Mohapatra, Phys. Rev. D12 (1975) 1502. G. Senjanovit, Nucl. Phys. B153 (1979) 334. 9. C.S. Aulakh, K. Benakli, G. SenjanoviC, Phys. Rev. Lett. 79 (1997) 2188. 10. C. S. Aulakh, A. Melfo, A. R&in and G. SenjanoviC, Phys. Lett. B459 (1999) 557. 11. R. Kuchimanchi and R. N. Mohapatra, Phys. Rev. D48 (1993) 4352 [arXiv:hepph/9306290]. 12. A. Melfo and G. Senjanovic, Phys. Rev. D68 (2003) 035013 [arXiv:hepph/0302216]. 13. C. S. Aulakh, B. Bajc, A. Melfo, A. Rasin and G. Senjanovic, Phys. Lett. B460 (1999) 325 [arXiv:hep-ph/9904352]. 14. D. Acosta et al. (CDF Collaboration], arXiv:hepex/0406073. P. Achard et al. [L3 Collaboration], Phys. Lett. B576 (2003) 18 [arXiv:hep-ex/0309076]. Phys. Lett. B577 (2003) 93 [arXiv:hep-ex/0308052].
87
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.
M. Magg and C. Wetterich, Phys. Lett. B94 (1980) 61. R. N. Mohapatra and G. Senjanovic, Phys. Rev. D23 (1981) 165. G. Lazarides, Q. Shafi and C. Wetterich, Nucl. Phys. B181 (1981) 287. C. Wetterich, Nucl. Phys. B187 (1981) 343. C.S. Aulakh, R.N. Mohapatra, Phys. Rev. D28 (1983) 217. T. E. Clark, T. K. Kuo and N. Nakagawa, Phys. Lett. B115 (1982) 26. D. Chang, R. N. Mohapatra and M. K. Parida, Phys. Rev. D30 (1984) 1052. X. G. He and S. Meljanac, Phys. Rev. D41 (1990) 1620. D. G. Lee, Phys. Rev. D49 (1994) 1417. D.G. Lee and R. N. Mohapatra, Phys. Rev. D51 (1995) 1353. C. S. Aulakh, B. Bajc, A. Melfo, G. SenjanoviC and F. Vissani, Phys. Lett. B588, 196 (2004). C. S. Aulakh, arXiv:hepph/0210337. K. S. Babu and R. N. Mohapatra, Phys. Rev. Lett. 70,2845 (1993). B. Bajc, G. SenjanoviC and F. Vissani, Phys. Rev. Lett. 90 (2003) 051802. B. Bajc, G. SenjanoviC and F. Vissani, arXiv:hep-ph/0402140. H. S. Goh, R. N. Mohapatra and S. P. Ng, Phys. Lett. B570 (2003) 215. H. S. Goh, R. N. Mohapatra and S. P. Ng, Phys. Rev. D68 (2003) 115008. S. Bertolini, M. F'rigerio and M. Malinsky, arXiv:hepph/0406117. W. M. Yang and Z. G. Wang, arXiv:hep-ph/0406221. B. Dutta, Y. Mimura and R. N. Mohapatra, arXiv:hep-ph/0406262. B. Dutta, Y. Mimura and R. N. Mohapatra, Phys. Rev. D69 (2004) 115014. K. Matsuda, Y . Koide and T. Fukuyama, Phys. Rev. D64 (2001) 053015. T. Fukuyama and N. Okada, JHEP 0211 (2002) 011. L. Lavoura, Phys. Rev. D48 (1993) 5440. B. Brahmachari and R. N. Mohapatra, Phys. Rev. D58 (1998) 015001. K. Y. Oda, E. Takasugi, M. Tanaka and M. Yoshimura, Phys. Rev. D59 (1999) 055001. B. Bajc, A. Melfo, G. SenjanoviC and F. Vissani, Phys. Rev. D70 (2004) 035007. T. Fukuyama, A. Ilakovac, T. Kikuchi, S. Meljanac and N. Okada, arXiv:hep ph/0401213. C. S. Aulakh and A. Girdhar, arXiv:hep-ph/0204097. C. S. Aulakh and A. Girdhar, arXiv:hep-ph/0405074. H. S. Goh, R. N. Mohapatra and S. Nasri, arXiv:hepph/0408139. C. S. Aulakh, B. Bajc, A. Melfo, A. %in and G. SenjanoviC, Nucl. Phys. B597 (2001) 89. H. S. Goh, R. N. Mohapatra, S. Nasri and S. P. Ng, Phys. Lett. B587 (2004) 105. T.Fukuyama, A. Ilakovac, T. Kikuchi, S. Meljanac and N. Okada, JHEP 0409 (2004) 052. T. Hambye and G. Senjanovib, Phys. Lett. B582 (2004) 73. S. Antusch and S. F. King, Phys. Lett. B597 (2004) 199. P.J. O'Donnell and U. Sarkar, Phys. Rev. D49 (1994) 2118. E.Ma and U. Sarkar, Phys. Rev. Lett. 80 (1998) 5716. G. Lazarides and Q. Shafi, Phys. Rev. D58 (1998) 071702. P. Fileviez-Pkrez, T. Hambye and G. Senjanovit, work in preparation.
SPIN FOAM MODELS OF QUANTUM GRAVITY
A. MIKOVIC Departamento de Matemdtica e CiEncias de Computaczo Universidade Lusdfona de Humanidades e Tecnologias Av. do Camp0 Grande, 376, 1749-024 Lisbon, Portugal E-mail: amikovicQu1usofona.pt We give a short review of the spin foam models of quantum gravity, with an emphasis on the Barret-Crane model. After explaining the shortcomings of the Barret-Crane model, we briefly discuss two new approaches, one based on the 3d spin foam state sum invariants for the embedded spin networks, and the other based on representing the string scattering amplitudes as 2d spin foam state sum invariants.
1. Introduction The spin foam models originate from the Ponzano-Regge model of 3d Euclidian quantum gravity.I The idea there was to use the simplical complex, i.e. the spacetime triangulation, whose triangles had integer lengths, which were proportional to the spins of the S U ( 2 ) group. Then the 3d gravity path integral (PI) was defined as a sum over the spins of the products of the 6 j symbols which were associated to the tetrahedrons of the simplical complex. A length cut-off was introduced in order to regularize the path integral. Since 3d gravity is a topological theory, the corresponding path integral would be a topological invariant of the 3d manifold. However, the Ponzano-Regge path integral was not a topological invariant , because the topological invariance required the quantum S U ( 2 ) group at a root of unity, which was discovered by Turaeev and Viro.’ Still, the idea of the Ponzano-Reggemodel was useful, because the model can be understood as a path integral for the SU(2) BF t h e ~ r yThis . ~ then inspired Ooguri to consider a 4d version of the model, as a PI for the 4d S U ( 2 ) BF t h e ~ r yIn . ~this case the areas of the triangles are integer valued, i.e. proportional to the spins. The PI was formally topologically invariant, but divergent. A well-defined topological invariant was obtained by Crane, 88
89
Yetter and K a ~ f f m a n ,who ~ replaced the S U ( 2 ) group by the quantum S U ( 2 ) group at a root of unity, so that the S U ( 2 ) spins become bounded by a maximal spin. However, the corresponding invariant was not new, contrary to the 3d case, since it gave a signature of the 4-manifold. At the same time, Baez proposed the idea of the spin foams,6 as a way of understanding the results of loop quantum g r a ~ i t y from , ~ a spacetime perspective, i.e. a spin foam is a time evolution surface of a spin network, so that the spins are naturally associated to the faces of the spin foam. Given that the Einstein-Hilbert (EH) action of GR can be understood as a constrained BF theory, this then prompted Crane and Barrett to look for a constrained version of the CYK topological spin foam state 2. The Barret-Crane Model
The EH action can be written as the SO(3,l) BF theory action
where F a b = d @ a b + @ z A @ c a is the curvature two-form for the spin connection and the two-form B-field is constrained by
0,
Bab
= E,bcdec A ed ,
(2)
where the are the tetrad one-forms. The BF theory path integral can be written as
where 1 and f are the edges and the faces of the dual two-complex F for the simplical complex T ( M ) ,while A are the triangles of T . The variables A l and BA are defined as A and B respectively, while Ff = Jf F . By performing the B integrations one obtains
sa
which can be defined as
90
where gf =
n,,,, 91. By using the well-known identity a(g) = EdimAXA(9) 7
(6)
A
where A’s are the irreducible representations (irreps) of the group and x’s are the characters, one obtains
z= Af+i
ndimAfnA,(Af,Ll), f
(7)
V
where A, is the vertex amplitide associated to the 4-simplex dual to the vertex u. This amplitude is given by the evaluation of the corresponding 4-simplex spin network, known as the 15j symbol. The sum (7) is called a spin foam state sum, because it is a sum of the amplitudes for the colored two complex F , i.e. a spin foam. One can now conjecture that exists a quantization procedure such that the quantities BA become the 4d rotations algebra operators J A , since the 4d rotation group irreps are labelling the triangles A, or the dual faces f. Then one can show that the constraint (2) becomes a constraint on the representations labelling the triangles A, given by8lg pbcdj J ab cd
-0*
(8)
In the Euclidian case the irreps are given by the pairs of the SU(2) spins ( j , j ’ ) , so that the constraint (8) implies j = j’. In the Minkowski case, requiring the hermiticity of the B operators implies that one needs the unitary irreps of the Lorentz group. These are infinite-dimensional irreps and they are given by the pairs ( j , p ) ,where j is the S U ( 2 ) spin and p is a continuous label. The constraint (8) implies that A = ( 0 , p ) or A = ( j ,0). One can argue that the spacelike triangles should be labelled by the ( 0 , p ) irreps, while the time-like triangles should be labelled by the ( j , O ) irreps. Since a spacetime triangulation can be built from the spacelike triangles, Barrett and Crane have proposed the following spin foam state sum (integral) for the quantum general relativityg
U
where A, is an amplitude for the corresponding 4-simplex spin network, given by
91
This is as an integral over the fifth power of the hyperboloid H = S0(3,1)/S0(3) of a propagator K p ( z y) , on that space. The propagator is given by
The expression (9) is not finite for all triangulations, but after a slight modification, consisting of including a non-trivial edge amplitude &I, . . . ,p 4 ) , the partition function becomes finite for all non-degenerate triangulations. lo This was a remarkable result, because it gave a perturbatively finitea quantum theory of gravity, which was not based on string theory. The main difficulties with the BC type models are: 1) It is difficult to see what is the semi-classical limit, i.e. what is the corresponding effective action, and is it given by the EH action plus the O(Zp) corrections, where Zp is the Planck length. 2) Coupling of matter: since matter couples to the gravitational field through the tetrads, one would need a formulation where a basic field is a tetrad and got the B 2-form. In the case of the YM field, the coupling can be expressed in terms of the B field,’l so that one can formulate a BC type m o d e l ~ . ~However, ~J~ for the fermions this is not possible, and a tetrade based formulation is necessary. In Ref. 11 an algebraic approach was proposed in order to avoid this problem, and the idea was to use a result from the loopquantum gravity, according to which the fermions appear as free ends of the spin networks. Hence including open spin networks gives a new type of spin foams,14and this opens a possibility of including matter in the spin foam formalism. However, what is the precise form of the matter spin foam amplitudes remains an open question.
3. New Directions Given the difficulties of the BC model, we have proposed two new directions how to use the spin foam state sum formalism in order to arrive at a desirable quantum theory of gravity. In Ref. 15 it was proposed to use the 3d spin foam state sum invariants in order to define the relevant quantities in the loop quantum gravity formalism. The idea is to use the representation of a quantum gravity state a Zdepends ~ ~on a triangulation, in accordance with the fact that 4d gravity is nontopological, and hence one should also sum over the triangulations in order t o obtain a well-defined quantity. How to do this it is not clear a t present, so that one can obtain only the perturbative results.
92
I*)
in the spin network basis
I@) =
c Ir)(rl@).
(12)
Y
The expansion coefficients are then invariants of the embedded spin networks in the spatial manifold C, and can be formally expressed as
where A is a 3d complex S U ( 2 ) connection, W,[A] is the spin network wave-functional (generalization of the Wilson loop functional) and f [A] is a holomorphic wave-functional satisfying the quantum gravity constraints in the Ashtekar representation. In the case of non-zero cosmological constant A, a non-trivial solution is known, i.e. the Kodama wavefunction
@[A]= e x2 JE T T ( A A ~ A + ; A A A A A ,)
(14)
while in the X = 0 case a class of formal solutions is given by
i.e. a flat-connection w a v e f u n c t i ~ n In . ~ ~the X = 0 case one can show that the corresponding spin network invariant is given by a 3d spin foam state sum for the quantum SU(2) at a root of unity.lS In the X # 0 case, it is conjectured that the corresponding spin network invariant is given in the Euclidian gravity case by the Witten-ReshetikhinTuraeev invariant for q = e 2 n i / ( k + 2where ), k E N and X = k/l:, while in the Minkowski case, the invariant is given by an analytical continuation of the Euclidian one, as k -+ik.16 In Ref. 17 it was proposed to use the 2d spin foam state sums in order to define a string theory as a quantum theory of gravity. The main idea is to use the string theory formal expression for the scattering amplitude of n gravitons (or any other massless string modes), given as
A(m, ...,P,) = where
s,
d a
.
do,
( m ). . . Vpn( 0 ; 2 ) ) ,
(Q1
(16)
93
and argue that (16) should represent a 2d BF theory ivarisnt for the 0, spin network embedded in the string world-sheet manifold C. The B F theory group is given by the group of isometries of the spacetime background metric, and in Ref. 17 a simple possibility for the amplitude was considered
where the isometry group was taken to be SU(2). Then the labels p l , . . . , p , become the SU(2) spins, and there is a maximal spin, because the PI (18) becomes a state sum for the quantum S U ( 2 ) a t a root of unity. Because the B F theory is a topological theory, one can expect that the amplitude (18) will correspond to a topological string theory.
Acknowledgments This work is supported by the FCT grants POCTI/FNU/49543/2002 and POCTI/MAT/45306/2002.
References 1. G. Ponzano and T. Regge, Spectroscopy and Group Theortical Methods in Physics, ed F. Block et al, North-Holland, Amsterdam (1968). 2. V. G. Turaeev and 0. Y. Viro, Topology 31,865 (1992). 3. D. V. Boulatov, Mod. Phys. Lett. A7, 1629 (1992). 4. H. Ooguri, Mod. Phys. Lett. A7,2799 (1992). 5. L. Crane, D. N. Yetter and D. H. Kauffman, Knot Theor. Ramif. 6, 177 (1997). 6. J. Baez, Lect. Notes Phys. 543, 25 (2000). 7. C. Rovelli, Living Rev. Rel. 1, 1 (1995). 8. J. W. Barrett and L. Crane, J. Math. Phys. 39,3296 (1998) 9. J. W. Barrett and L. Crane, Class. Quant. Grav. 17,3101 (2000). 10. L. Crane, A. Perez and C. Rovelli, Phys. Rev. Lett. 87,181301 (2001). 11. A. MikoviC, Class. Quant. Grav. 19,2335 (2003). 12. A. Mikovik, Class. Quant. Grav. 20, 239 (2003). 13. D. Oriti and H. Pfeiffer, Phys. Rev. D66, 124010 (2002). 14. A. MikoviC, Int. J. Mod. Phys. A18S2,83 (2003). 15. A. MikoviC, Class. Quant. Grav. 20, 3483 (2003). 16. L. Smolin, Quantum Gravity with a Positive Cosmological Constant, hepth/0209079. 17. A. MikoviC, String Theory and Quantum Spin Networks, hep-th/0307141.
RIEMANN-CARTAN SPACE-TIME IN STRINGY GEOMETRY
B. SAZDOVIC Institute of Physics, P. 0.Box 57, 12001 Belgrade, Serbia and Montenegro E-mail: [email protected]. yu We consider the classical equations of motion for the string propagating in the target space. It is known that, in the case of Riemann space time, the world-sheet is a minimal surface. It is specified by the requirement that all mean extrinsic curvatures, corresponding to the normal vectors, are zero. The presence of the antisymmetric field in the string action, leads to the space-time torsion and the target space becomes of the Riemann-Cartan type. We define the mean torsion, and its orthogonal projection as the dual mean extrinsic curvature. In this language, the world-sheet is C-dual surface, which means that mean extrinsic curvature is equal to dual mean extrinsic curvature. The string feels the group manifold as RiemannCartan type space-time. The metric tensor is the well known Killing group metric and topological Wess-Zumino term is origin of the parallelizing torsion. We apply above consideration to this particular example.
1. Introduction
We will investigate bosonic string propagation in space-time M D(G,,, B,, , a), defined by x , depended background gravitational field G,,, antisymmetric tensor field B,, = -B,, and dilaton field a. By X,(T,(T) ( p , v = O , l , ...,D - l), we denote position of the string and by E“ = {to= T,E l = o} the coordinates of the world-sheet. The bosonic string action1
(1) has been investigated in the literature.2 Here, gup is intrinsic world-sheet metric and R(2)is corresponding two dimensional scalar curvature. As a preparation, let us first consider the point particle case. Then, instead of world-sheet we have world-line parameterized by one parameter, 94
95
tff
-+ r , so that &xi' -+ x p . Consequently, only the first term, with the background field G,,, survives in the action (1) and we obtain
S = K,
[drG,,(x)xpx".
(2)
J
The equation of motion is
2 + rrgkPxu = D,Y
=0,
(3)
where
is Christoffel connection. So, the test particle in external gravitational field follows geodesic line. It feels the target space as Riemann space-time of general relativity, VD= (M D ,r, G ) . We will try to understand: how test string feels the external spacetime M D , which we shall call the stringy ~pace-time.~ For better understanding, we will first introduce general space-time geometry and investigate string equations of motion.
2. Affine Space-time The metric tensor G,, and the affine connection "I?&, generally are independent variables. They define the afine space-time A D = (G,,, 'I?&). The metric tensor defines the scalar product of the vectors V . U =
VPG,,U". The affine connection, defines the rule for vector parallel transport V p ( x ) -+ "yr = V p + "6Vp where "6VC" = -"I?raVPdxo.
(5)
The covariant derivative is a difference of two vectors at the same point x dx: one living in this point and the other parallely transported from the point x
+
"DV'" = V'(X + dx) -
"7;"= dVI"-
O W / "
3 "D,V'dd'
,
(6)
where "D,VP = a,V, + Ol?;,VP. The rule for the parallel transport of the covector U,(z), "SUP = 'I'$,UpdxU, can be obtained from the invariance of the vector product under parallel transport, "6(VpU,) = 0. In the Riemann space-time connection is symmetric, r$ = I?$p. The metric postulate, DG,, = 0, is satisfied, which means that lengths do not change during the parallel transport.
96
In the affine space-time, the antisymmetric part of the connection defines the torsion
and a covariant derivative of the metric tensor defines the nonmetricity
A
Figure 1. Geometrical meaning of the torsion.
Let us explain their geometrical meaning. Interpretation of the torsion is illustrated in the Fig. 1. Here, iY(5;) are the positions of the tangent vectors tY(t;) after parallel transport along the geodesics l z ( l , )-, respectively, where we used AB = d l l = d i l = CD2 and AC = dCz = dlz = BD1. In fact, the torsion measures the non-closure of the curved rectangle ABCD
z'(D2) - z'(D1) = O T p p , , t:tz d l l d l z .
(9)
To understand the meaning of nonmetricity, we can compare the square of the vector Vp at the point x, V z ( x )= G,,(z)VP(z)V"(z),with its parallel transport to the point z + d z , ' y ? ( z + d z ) = G,,(z+dz)"y;" "l$ Because '. the scalar product is invariant under parallel transport, we have "SV2= "Vi?(z+ d x ) - V 2 ( z )= "DG,,V"V"
3
- ~ z P ~ Q ~ ~(10) ~ V ~ V
So, in affine space-time, after parallel transport the metric tensor is not equal to the local one. The nonmetricity measures the breaking of the metric postulate.
97
Following Ref. 4,we can decompose the connection Christoffel connection, contortion and nonmetricity
in terms of the
With the help of the Schouten braces, {ppu} = upp + pup - ppu, the Christoffel connection can be expressed as rp,pa = id+Gpoj and the contortion OKpPois defined in terms of the torsion OKppo= ~ o T { o p p ) . The first and third terms in Eq. (11) are symmetric in p , indices. ~ In the second term we can separate symmetric and antisymmetric parts, O K p p o = oKp(p,) ioTppo which produces
+
3. Induced and Extrinsic Geometry The geometry of world-sheet embedded in curved ~pace-time,~ has been generalized for the space-times with nontrivial torsion and n~nmetricity.~ We will shortly repeat the main results. 3.1. Induced Metric Tensor and Induced Connection The geometry of the world-sheet is defined by world-sheet metric tensor Gap and world-sheet connection O r & . We are interested in the case when this geometry of the world-sheet is induced from the space-time. The world-sheet induced metric tensor is defined by the requirement that world-sheet intervals measured by the induced and space-time metrics have the same lengths, so that
Gap = G p y d a ~ p d p ~ u .
(13)
The infinitesimal parallel transport of the world-sheet tangent covector V p , along world-sheet line from the point (" to the point (" d(", (Fig. 2), produces the covector = VF "6V;, where
'7:
+
+
Generally, it does not belong to the world-sheet. The best we can do is to relate the projections of the vectors VF and O q L . The induced connection, O r & , defines the rule for the parallel transport of the world-sheet covector v, = VFd,xp, along the same world-sheet
98
\
\
M D
Figure 2.
The parallel transport of the world-sheet tangent covector.
+
line, to the projection "v& = dffxP(t d ( ) " v ; . relations o
x = ,qff
21,
+ O6vff,
So, from the standard
Obv, = oI'&vpdtY,
(15)
we find the expression for the induced connection ffP = G
~
O r r
~
~
~
~
(16) ~
where OD, V P = & X " ~ D ~ VisP space-time covariant derivative along world-sheet direction.
3.2. Second Fundamental Form
Vk,
The parallel transport of the covector orthogonal to the world-sheet = "6V;, where "bVk = O r "PP V*dxp " = (Fig. 3), produces V>&xpd
Vk +
aax"(t
+ d t ) " ? ~=~
1 - 0
I -
io
vllff - v biapdt',
(17)
defines second fundamental form (SFF) O b i f f p = n ~ G P v o D p d a x=u -daxV0Dp(GP,nr),
(18)
where nr (i = 2,3, ..., D - 1) are local unit vectors, normal to the worldsheet. In the spaces with nonmetricity there are two sets of SFFs as well as two sets of the induced connections and induced covariant derivatives. We
G
99
\ MD
/ Figure 3. The parallel transport of the world-sheet orthogonal covector.
will use the variables defined above, because these two sets of variables are equal up to terms with nonmetricity. The induced connection and the SFF are projections of the space-time covariant derivative of the world-sheet basis vectors dffxP,to the worldsheet and its normal, respectively. Consequently, we have the generalization of the Gauss-Weingarten equation
oDpdaxP = ol?&dyxp + "bhP nf .
(19)
In analogy with the general rule for connection decomposition we can decompose the induced connection and SFF Ory,ap
= ry,olp
1
+ OKyap + z o Q { y a p }
1
=hap
+ 'Kim' + z1o Q { i a p } .
(20) Here, the indices for induced variables can be obtained by the rules Tia = n f d a x v T p v , for T = { O T a p r , OQapy, 'Tia', OQiap}, and we also have OKyap = :oT{yap}, biap = nfGpvDpdax", and O K i f f p = ;oT{iap}. 3.3. Mean Extrinsic Curvature The curvature of the normal section ti,as a space-time curve, is orthogonal projection of the covariant derivative along the curve, "ki = OD,tPGPvn:, or explicitly
"ki = Obipa tatP= Obi(pff) tat'.
(21)
100
It depends on the direction of the world-sheet tangent vector t". The i-th mean extrinsic curvature (MEC) is mean value of the ith normal section curvature, with fixed normal n:. It is well defined in Euclidean spaces. In Minkowski spaces the curvature (21) is divergent in the light-like directions. For the time-like region we define regularized mean curvature "lcfl(a)da,
("lc;,(a)= "66cosh2(a) -
"61sinh2(a)) ,
(22) where principal curvatures X i are eigenvalues of the quadratic forms "b& with respect to the metric Gap. For the space-like normal section we have "lctl(a)= "66 sinh2(a)- "6f cosh2(a) and similar expression as Eq. (22). We define the MEC as a mean value of the time-like mean curvature with sign and space-like mean curvature with sign " - "
+,
1 1 lim -["H;,(A) - " H i , ( h ) ]= - ( O K ; 2 2
A+w
+OK;)
= "Ha.
(23)
The condition for nontrivial solutions of the eigenvalue problem, det ( "b& " K ~ G= ~0 produces ~ ) " ~ 6 =, ~"Hi f J("Hi)2 - OKi. Here
detabha -
io i
is just the MEC of the expression (23) and OKi = - tc0 fiC1(no summation over i) is Gauss curvature. For "6; = "6;= "6' the line curvature does not depend on the tangent vector direction, "lcf,(a)= Ai = O H , , , and " k ~ , ( a= ) -"tea = "Hi,. Consequently, the MEC has the same interpretation + ( O H ; , - O H : , ) = "62 = "Hi. The surface defined by the equation "Ha = 0 is the minimal ~ u r f a c e . ~
4. C-duality The above considerations are torsion independent, because the antisymmetric part of the SFF disappears from Eqs. (21) and (24). In Ref. 3 we investigated the case when the torsion is present and the SFF is not symmetric in a, /3 indices
Let us first generalize the eigenvalue problem. We introduce the dual eigenvalue problem, such that linear transformation of the vector v a
101
with operator "b& is proportional to the two dimensional dual vector *u, = G G E f f p up ObiP
UP = * K Z
*u,.
(26)
d(*Hi)2 +
The dual eigenvalues, *fig,l = *Hi f O K i , are solutions of the condition det(*bi - * k i E ~ ) = 0. ~ In Panalogy with the previous case, we will call them dual principal curvatures, and the variable
det O b i p
the dual mean extrinsic curvature (DMEC). Here OKi = * figi * rc1i (no summation over i) is the same Gauss curvature as before.
-
What is the geometrical meaning of the DMEC? In the case when ty and are world-sheet tangent vectors, (Fig. l), we can rewrite Eq. ( 9 ) in the form
ti
Here, dP12 = a d e t ( % ) d l l d & is area of the parallelogram, spanned by the vectors d t y = tyd-tl and dtg = t g d t z . The variable O T P does not depend on the directions ty and t;, and on the lengths del and d&. So, we will call it the mean torsion. Its normal projection, the extrinsic mean torsion
is exactly the same variable as DMEC, defined in (27). If we introduce a dual SFF
we can reformulate the dual eigenvalue problem ( 2 6 ) as an ordinary one (*bhp - * K ~ G ~ ~=)0.V " Let us introduce transformation which interchanges the roles of the symmetric and the antisymmetric parts of the SFF and maps SFF to dual SFF, "bLp 4 *b"&. It relates MEC and DMEC, allowing the exchange of the mean curvature and mean torsion. We called it C-duality (Curvature duality) .3 The conditions for self-dual and self-antidual configurations
102
define C-dual and antidual surfaces. In the torsion free case, they turn to the standard minimal surface O H i = 0.
5. String Equations of Motion For u2 # 0 ( a ,
= a,@), the equations of motion for the action (1)are6 [PI = vFokx, + *r;puaf2pdFxu = 0 , (32)
[iF]E
+ ~2 ( D r , ~ , ) & d ‘ d=~0~. ~
(34)
We used a tangent basis where the light-cone components of the vector V, have the form V* = .&(Vo h+Vl). The world-sheet covariant derivative denoted by 8,. The expression
+
*r$,, = r;,
up f PTpuB& + --,a,,
(35)
a2
which appears in the [PIequation, transforms as a connection under spacetime general coordinate transformations. The first term r;, is the Christoffel connection, D, is the corresponding covariant derivative and Bpvp
= D,B,p
+ D,Bp, + DpB,,
(36)
is field strength of the antisymmetric tensor. The projection operator in the second term is defined as PTPu= G,, Let us apply the general considerations to the string case. Instead of mark the O , the mark * indicates the presence of dilaton field @, and lower indices f indicate the presence of the antisymmetric field B,, in the corresponding two forms of connections (35). The stringy torsion and the stringy nonmetricity take the forms
y.
*T*& = f2PTP,B“PU
1 *Qfppo
= 2D+(apao).
(37)
The form of Eq. (36) suggests that B,, is a torsion p ~ t e n t i a l .In ~ the presence of the dilaton field @, the metric G,, is not compatible with the stringy connection *r&,and during stringy parallel transport, the length deformation depends on the vector field a,.
103
6. Classification of the Space-times With the help of the field equations (32)-(34), we can make the space-time clas~ification,~ depending on the background fields. We will see that the probe string feels torsion and nonmetricity, which is not the case for the probe particle. 6.1. Riemann and Riemann- Cartan Space-time In the absence of the dilaton field a, the action is conformaly invariant. The field F and the corresponding equation [iF]are absent. Both intrinsic world-sheet metric tensor and intrinsic connection are equal to the induced ones from the space-time, up to the conformal factor. There are two forms of connection I'$, = I?$, fBE, and corresponding two forms of covariant derivative D+. The torsion T*Z, = f2Bg,, is proportional to the field strength of the antisymmetric tensor. The orthogonal projection of the [J,] equation obtains the form of selfduality condition
and world-sheet is C-dual (antidual) surface. The string with background fields G,, and B,, feels the target space as Riemann-Cartan space-time.
VD Figure 4. Classification of space-time. A D is a s n e space-time, S D is stringy spacetime, S D is stringy torsion free space-time, UD is stringy Riemann-Cartan space-time and VD is R i e m a n n space-time.
104
For B,, = 0, the last equation turns to the standard condition of minimal surface
H a,
1 -G"Pbb, = 0 . (39) 2 The string with background field G,, feels the target space as R i e m a n n space-time of general relativity and does not see torsion and nonmetricity.
6 . 2 . Stringy Space-time In the presence of the dilaton field CP, the theory loses conformal invariance. We found the relations between intrinsic and induced metric tensors and connections 1 ap - S X:, - $Aa GapGy6X6 - -*QIYap1, w;p = *ry gap = .;Gap,
+
2
(40) where A, = 4 , In and a; = Gapa,ap is a length of the world-sheet projection of the vector field a,. The string propagating in the presence of all three background fields G,,, B,, and Q,, feels both space-time torsion and nonmetricity. The corresponding target space we call stringy space-time. The normal projection of the [J,] equation takes the form
6,
so that the world-sheet is stringy C-dual surface. For B,, = 0, Eq. (41) turns to the stringy minimal surface condition
1 * H . - -Gap *bipa = 0 ,
"2 because all stringy MECs vanish.
7. String Propagation on the Group Manifold
We are going to apply above results to the case where the group manifold takes the role of the curved space-time. Let us put dilaton field CP to zero, Q, = 0. Instead of the string coordinate x, we introduce group manifold coordinates q a . We also substitute spacetime metric G,, with the Cartan one Yap = E&,"E*pb&b
where
,
Ef," are qa depending vielbeins on the group manifold.
(43)
105
Christoffel connection takes the form
and does not depend on the f indices. Instead of the connections (35) we obtain
rZP7= rZ7fT;,
= ya6E$6dpE!&3ab,
(45)
where Tap7
= aa7p-y
1 + 8 y T a p + ap7-p = F5fabcEzaEipE$, .
(46)
The fields r a p and ~~p~ have the roles of antisymmetric field B,, and its field strength BPup. Note that Wess-Zumino-Novikov-Wittenaction (Eq. (3.29) of Ref. 8)
exactly coincides with the action (1),if we substitute space-time with group manifold and put 6, = 0. The second term in Eq. (47) is topological WessZumino term.g The principal differences between space-time and group manifold cases are group conditions on the vielbeins ap~ga aa~g= p fabCEqaEip>
(48)
and on the antisymmetric tensor ~~p (46). The group manifold is space with nontrivial torsion
TfQp7= f2~;,,
(49)
and vanishing nonmetricity, Q& = 0. The Riemann curvature is different from zero, but generalized Riemann-Cartan curvature vanishes, so that the torsion parallelizes the m a n i f ~ l d . ~ 8. The Geometry Induced from the Group Manifold Following the general case, we can define two-dimensional induced metric and induced connection from the group manifold. We will change the notation. Let (P (to= T , E l = a) be the coordinates of two-dimensional world-sheet and a, = the corresponding derivatives. We will use the local space-time basis, relating with the coordinate one by
&
106
the vielbein Zz = {apqa,nF}. Here, dpq" = {Ga,q'a} is the local worldsheet basis and n: (i = 2,3, ..., D - 1) are local unit vectors, normal to the world-sheet . It is useful to introduce the quantities
(50)
u*i = t,Epanq.
u*p = t,E&dpq",
Here, t , are generators of the gauge group G , with normalization (t,, t b ) &b = -atr{t,tb} and ( x , Y )is the Cartan inner product. Starting with the definition of the induced metric (13), we obtain ~
p
=u TapapqQauqP = (u*p,v*u)
>
=
G p i = r a p a p q a niP -- ( u f p , U k i ) .
(51) In analogy with (16), the induced connection is defined by the expression
qfU = G"'~uqaTapD*oa,qP
=
(.$,
apUh7)
,
(52)
+
where D*,VP = d,q"D*,V - a,q"(daVP rtraV.) is the group manifold covariant derivative along world-sheet direction. The second fundamental form, with the help of Eq. (18), takes the form &ipu
= nqy,pa uq7D*r~p9 = ( U h i , a p v * u ) .
(53)
Consequently, we are ready to calculate mean extrinsic curvature and dual mean extrinsic curvature
(54)
*H*.a -=
1 &PU apu
=
1 &PU 1 T(u+i,-a u u * p ) = m ( u & i ,d-u*+-a+u*-).
2 Gb*'
d=
(55) The equations of motion in the form of self-duality conditions turn into the expression
1 H+i f *H*i = - ( ~ * i ,
G
~ T U * * )= 0 .
(56)
Because the inner product with all u*i vanishes, it follows that d,u** = 0. In terms of the group elements g, we have u + = ~ gd,g-l, u - = ~ g-'d,g, and the equations of motion obtain the standard forms LL(g8,g-l)
=0 ,
a+(g-ld-g) = 0 .
(57)
107
9. Conclusions We investigated the space-time geometry felt by the bosonic string propagating in nontrivial background. Particularly, we were interested in the case when group manifold takes the role of the space time and the action turns to the Wess-Zumino-Novikov-Witten one. We introduced affine space-time and defined torsion and nonmetricity. We also obtained expressions for the stringy torsion and stringy nonmetricity (37) originating from the antisymmetric tensor field B,, and dilaton fields @, respectively. We considered induced and extrinsic geometry, when the surface is embedded into space-time with torsion and nonmetricity. We introduced Cduality which maps MEC to DMEC and defined C-dual surface by the conditions O Hi = f* Hi. We made space-time classification depending on the background fields. The [ PI equation defines world-sheet as a stringy C-dual surface. In two particular cases -the vanishing torsion and the vanishing nonmetricitythe field equations turn to the equations of stringy minimal world-sheet *Hi = 0 and C-dual world-sheet Hi = f*HTirrespectively. In the case of Riemann space-time, when both torsion and nonmetricity vanish, they turn to the equations of minimal world-sheet Hi = 0. Using the above results, we investigated bosonic string propagation on group manifold. We offered the geometrical interpretation of the group manifold felt by the string. We found that the antisymmetric tensor field, which corresponds to the Wess-Zumino topological term, is the origin of the group manifold torsion. This field also parallelizes the manifold. The string feels the group manifold as a Riemann-Cartan space-time. In this language, the standard equations of motion (57) have the form of C-duality Hi f *Hi= 0. The difference between left and right chirality equations is a consequence of a nontrivial torsion. Acknowledgements This work is supported in part by the Serbian Ministry of Science, Technology and Development under contract No. 1486. References 1. M. B. Green, J. H. Schwarz and E. Witten, Superstring Theory, Cambridge University Press (1987); J. Polchinski, String Theory, Cambridge University Press (1998).
108
2. E. S. F’radkin and A. A. Tseytlin, Phys.Lett. B158, 316 (1985); Nucl.Phys. B261, 1 (1985); C. G. Callan, D. Friedan, E. J. Martinec and M. J. Perry, Nucl.Phys. B262, 593 (1985); T. Banks, D. Nemeschansky and A. Sen, Nucl.Phys. B277, 67 (1986); A. A. Tseytlin, Int. J . Mod. Phys. A4, 1257 (1989). 3. B. SazdoviC, hep-th/0304086. 4. M. BlagojeviC, Gravitation and gauge symmetries, IoP Publishing, Bristol (2002); F. W. Hehl, J. D. McCrea, E. W. Mielke and Y. Ne’eman, Phys. Rept. 258,1 (1995). 5. B. M. Barbashov and V. V. Nesterenko, Introduction to relativistic string theory, World Scientific, Singapore (1990); D. Sorokin, Phys. Rept. 329, 1 (2000). 6. B. SazdoviC, hep-th/0304085. 7. T. L. Curtright and C. K. Zachos, Phys. Rev. Lett. 53,1799 (1984); S. Mukhi, Phys.Lett. B162,345 (1985). 8. B. SazdoviC, Phys. Rev. D62,045011 (2000). 9. J. Wess and B. Zumino, Phys. Lett. B37,95 (1971); E.Witten, Comm. Math. Phys. 92,455 (1984). I
CAN BLACK HOLE RELAX UNITARILY?
S . N. SOLODUKHIN School of Engineering and Science, International University Bremen, P.O. Box 750561, Bremen 88759, Germany E-mail: [email protected] We review the way the BTZ black hole relaxes back to thermal equilibrium after a small perturbation and how it is seen in the boundary (finite volume) CFT. The unitarity requires the relaxation t o be quasi-periodic. It is preserved in the CFT but is not obvious in the case of the semiclassical black hole the relaxation of which is driven by complex quasi-normal modes. We discuss two ways of modifying the semiclassical black hole geometry to maintain unitarity: the (fractal) brick wall and the worm-hole modification. In the latter case the entropy comes out correctly as well.
1. Introduction Any thermodynamical system initially in equilibrium at finite temperature and then perturbed tends to return to the equilibrium if the perturbation is not too big and does not last too long. The important parameter which characterizes this process is the relaxation time T.' In fact, the process of relaxation back to the equilibrium is a particular and most easily tractable example of more general phenomenon of the thermalization, when system initially far from being thermal gets thermalized to a state characterized by certain temperature. The way how the thermalization goes for different systems is an important and still poorly understood problem. Gravitational physics gives yet another example of system which behaves thermally. This system is the black hole. The black hole formation can be viewed as another example of the thermalization: non-thermal collapsing body and the flat space-time geometry in the beginning transform to (thermal) black hole state in the end of the gravitational collapse. On the other hand, the already formed and stayed in equilibrium black hole can be perturbed by exciting a pulse of matter field in the exterior of black hole. The subsequent relaxation is very well studied in the literature and is known to be characterized by the 109
110
so-called quasi-normal modes. These modes are eigen values of the radial Schrodinger type equation subject to certain boundary conditions. These are dissipative boundary conditions saying that the perturbation should leave the region through all possible boundaries. In general, there are two such boundaries: black hole horizon and spatial infinity. Formulated this way, the boundary value problem is not self-adjoint so that the quasi-normal modes are typically complex w = W R - i w with ~ a negative imaginary part. In most cases there is a discrete set of such modes parameterized by integer number n. The imaginary value of the lowest ( n = 1) quasi-normal mode sets the relaxation time T = l / w ~ . In fact, what we have said in the beginning about the relaxation of any system back to thermal equilibrium should be made more precise: the system should be in infinite volume. In finite volume and if the evolution of the system is unitary, any perturbation once created never leaves the system so that the complete returning to the initial unperturbed state is not possible. Thus information about the perturbation never disappears completely and always can be restored. The characteristic time during which the perturbation (as well as the whole state of the system) is guaranteed to come back is set by the Poincark recurrence time. All this means that the characteristic frequencies which run the perturbation of unitary system in finite volume should be real and discrete. Depending on these frequencies the evolution of the perturbation is quasi-periodic or even chaotic. But it can never be dissipative. Thus, strictly speaking for the thermalization we need infinite volume. Of course, nothing is infinite in the real world. The system still may be considered as thermal during the interval of time which is considerably less than the Poincark recurrence time. What this implies for the black holes? More specifically, for asymptotically Ads black holes the state of thermal equilibrium of which is well defined and can last infinitely long? Such a black hole can be viewed as system put in the box with the size set by the Ads radius. So one would have to expect this black hole to behave as any other system in the finite volume and in particular to show the Poincark recurrences (for the discussion of this in de Sitter space, see Ref. 2). This however does not happen in the semiclassical black hole: the complex quasi-normal modes are always there. The presence of these modes is related to the very existence of the horizon. Once there is black hole horizon there will always be complex frequencies which govern the time evolution of the perturbation. This problem is a manifestation of the long-time debated issue of whether the black hole evolution is actually unitary (see Refs. 2, 3).
111
A refreshed look at the whole issue is offered by the AdS/CFT correspondence (see review in Ref. 4 ) . According to this correspondence the gravitational physics in the bulk of asymptotically Ads space has a dual description in terms of a Conformal Field Theory (CFT) living on the boundary. Thus, the black hole in the bulk corresponds to a thermal CFT. The relaxation of the black hole than has a dual description as relaxation of the CFT after a perturbation driven by certain conformal operator has been applied to the system. The quasi-normal modes thus set the time scale for the relaxation in the boundary CFT.5 The effect of the finite size however is rather delicate issue. It has been studied in Ref. 6 and is reviewed in section 3 of this note. The similar conclusions have been made in Ref. 7. The recent reviews on the issue of black hole relaxation and unitarity are Ref. 8 and Ref. 9. 2. Relaxation in Black Hole: Quasi-normal Modes
We consider (2+1)-dimensional BTZ black hole with metric given by ds2 = - sinh' y dt2
+ dy2 + cosh2y d+2 ,
(1)
where for simplicity we consider non-rotating black hole and set the size of the horizon r+ = 1 and Ads radius I = 1. The coordinate is periodic with period L so that the boundary has topology of cylinder and L sets the finite size for the boundary system. A bulk perturbation @(m,s) of mass m and spin s should satisfy the quasi-normal boundary condition, i.e. it should be in-going at the horizon and have vanishing flux at the infinity. The latter condition comes from the fact that in the asymptotically Ads space-times the effective radial potential is growing at infinity so that there can be no propagating modes as well as no leakage of the energy through the boundary. The relevant radial equation takes the form of the hypergeometric equation, which is exactly solvable. The quasi-normal modes in general fall into two setsloill
+
27r
w = -m-
L
where the left- and right-temperatures TL = TR = 1/2n and ( h ,h) have the meaning of the conformal weights of the dual operator O(h,h),corresponding to the bulk perturbation @ ( m , s ) , with h 71 = A(m), h - h = s and A(m) is determined in terms of the mass m.
+
112
For comparison, in the case of global anti-de Sitter space the horizon and respectively the quasi-normal modes are absent. But, instead, one can define the normalizable modes which form a discrete set of real frequenciesI2 w = 2nm/L
+ 4n(n + h)/L,
n EN,
(3)
where the size of the boundary is also set to be L as in the black hole case. 3. Relaxation in CFTz
The thermal state of the black hole in the bulk corresponds to the thermal state on the CFT side. In fact, the boundary CFT factorizes on left- and right-moving sectors with temperature TL and TR respectively. The bulk perturbation corresponds to perturbing the thermal field theory state with operator C7(h,~).The further evolution of the system is described by the so-called Linear Response Theory (see Ref. 1). According to this theory one has to look at the time evolution of the perturbation itself. More precisely, the relevant information is contained in the retarded correlation function of the perturbation at the moments t and t = 0 (when the perturbation has been first applied). Since the perturbation is considered to be small, the main evolution is still governed by the unperturbed Hamiltonian over the thermal state so that the correlation function is the thermal function at temperature T. Thus, the analysis boils down to the study of the thermal 2-point function of certain conformal operators. Such a function should be double periodic: with period 1/T in the direction of the Euclidean time and with period L in the direction of the compact coordinate 4. This can be first calculated as a 2-point function on the Euclidean torus and then analytically continued to the real time.
3.I. Universality In general the correlation function on torus can be rather complicated since its form is not fixed by the conformal symmetry. The conformal symmetry however may help to deduce the universal form of the 2-point function in two special cases: when size L of the system is infinite (temperature T is kept finite) and when inverse temperature is infinite (the size L is finite). The universal form of the (real time) 2-point function in the first case is
113
which for large t decays exponentially as e-2rrT(h+h)t. The information about the perturbation is thus lost after characteristic time set by the inverse temperature. It is clear that this happens because in infinite volume the information may dissipate to infinity. In the second case correlator
has the oscillatory behavior. Notice that the oscillatory behavior in the second case should have been expected since the system lives on the circle. The perturbation once created at the moment t = 0 at the point q5 = 0 travels around the circle with the speed of light and comes back to the same point at t = L. Thus, the information about the perturbation is never lost. The correlation function (5) as a function of time represents a series of singular picks concentrated at t = &4 n L , n E N. In fact, this behavior should be typical for any system with unitary evolution in finite volume. It is interesting to see what happens in the intermediate regime when both L and 1/T are kept finite. In this case the behavior of the correlation functions is not universal, may depend on the (se1f)interaction in the system and is known only in some cases. We consider two instructive examples: the free fermion field and the strongly coupled CFT which is dual to the gravity on Ad&.
+
3.2. Intermediate Regime: Free Fermions The two point function of free fermions on the torus is known explicitly (e.g. Ref. 13). The real time correlation function is
+
were w = i(t 4) and u characterizes the boundary conditions for +(w). For finite temperature boundary conditions we have u = 3,4. Using the properties of &functions, it is then easy to see that (6) is invariant under shifts w + w 1/T and w -+ w iL. It is then obvious that the resulting real time correlator (6) is a periodic function o f t with period L. Zeros of the theta function Bl(wT[iLT)are known13 to lie at w = m/T+inL, where m, n are arbitrary relative integers. Therefore, for real time t , the correlation function (6) is a sequence of singular peaks located at (t 4) = nL. Using the standard repre~entation'~of the &functions, we also find that in the
+
+
+
114
limit LT -+ rn the correlation function (6) approaches the left-moving part of (3) with h = 1 / 2 that exponentially decays with time,
TT
(1(I(w)1(I(0))3(4) = 4 sinh TT(t + 4) [l f 2e-TLT cosh 27rT(t
+ 4 ) + . .I
. (7)
In the opposite limit, when LT -+ 0, it approaches the oscillating function (5). A natural question is how the asymptotic behavior (7), when size of the system is taken to infinity, can be consistent with the periodicity, t -+ t L , of the correlation function (6) at any finite L? In order to answer this question we have to observe that there are two different time scales in the game. The first time scale is set by the inverse temperature 7 1 = 1/T and is kept finite while the second time scale is associated with the size of the system 1-2 = 1/L. When L is taken to infinity we have that 72 >> 7 1 . Now, when the time t is of the order of 7 1 but much less than 1-2 the asymptotic expansion (7) takes place. The corrections to the leading term are multiplied by the factor e-TLT and are small. The 2-point function thus is exponentially decaying in this regime. It seems that the system has almost lost information about the initial perturbation (at t = 0). But it is not true: as time goes on and approaches the second time scale t 72 the corrections to the leading term in (7) become important and the system starts to collect its memory about the initial perturbation. The information is completely recovered as t = 72 and the time-periodicity is restored. This example is instructive. In particular, it illustrates our point that there can be thermalization in the finite volume for relatively small intervals of time, i.e. when t << 1-2.
+
-
3.3. Strongly Coupled CFT2 Dual to Ad& As an example of a strongly coupled theory we consider the supersymmetric conformal field theory dual to string theory on Ad&. This theory describes the low energy excitations of a large number of D1- and D S - b r a n e ~It . ~can be interpreted as a gas of strings that wind around a circle of length L with target space T 4 . The individual strings can be simply- or multiply wound such that the total winding number is k = I,where c >> 1 is the central charge. The parameter k plays the role of N in the usual terminology of large N CFT. According to the prescription (see Ref. 4),each AdS space which asymptotically approaches the given two-dimensional manifold should contribute to the calculation, and one thus has to sum over all such spaces. In the case of interest, the two-manifold is a torus ( ~ , 4 )where , 1/T and L are
115
the respective periods. There exist two obvious Ads spaces which approach the torus asymptotically. The first is the BTZ black hole in Ads3 and the second is the so-called thermal Ads space, corresponding to anti-de Sitter space filled with thermal radiation. Both spaces can be represented (see Ref. 14) as a quotient of three dimensional hyperbolic space H 3 , with line element 12
ds2 = -(dzdZ
Y2
+dy2),
y
> 0.
In both cases, the boundary of the three-dimensional space is a rectangular torus with periods L and 1/T. We see that the two configurations (thermal AdS and the BTZ black hole) are T-dual to each other, and are obtained by the interchange of the coordinates T H q5 and L t) 1/T on the torus. In fact there is a whole SL(2, Z) family of spaces which are quotients of the hyperbolic space. In order to find correlation function of the dual conformal operators, one has to solve the respective bulk field equations subject to Dirichlet boundary condition, substitute the solution into the action and differentiate the action twice with respect to the boundary value of the field. The boundary field thus plays the role of the source for the dual operator O(,,h). This way one can obtain the boundary CFT correlation function for each member of the family of asymptotically Ads spaces. The total correlation function is then given by the sum over all SL(2,Z) family with appropriate weight. However, for our purposes it is sufficient to consider the contribution of only two contribution^'^
( O ( t ,~ ) o (0)) o ,= e-"TZ(O
o')BTZ
+ e-'Aa
(O O ' ) A d S
,
(9)
where SET,= - h L T / 2 and S A d S = -kn/2LT are Euclidean actions of the BTZ black hole and thermal AdS3, respectively.16 On the Euclidean torus ( )BTZ and ( ) A d s are T-dual to each other. Their exact form can be computed ex~licit1y.l~ For our purposes it is sufficient to note that the (realtime) 2-point function coming from the BTZ part is exponentially decaying, ( )BTZ e-2?rhTt even though it is a correlation function in a system of finite size L. On the other hand, the part coming from the thermal Ads is oscillating with period L, as it should be for a system at finite size. Thus, the total 2-point function (9) has two contributions: one is exponentially decaying and another is oscillating. So that (9) is not a quasi-periodic function of time t. This conclusion does not seem to change if we include sum over SL(2, Z) in Eq. (9). There will always be contribution of the BTZ black hole that is exponentially decaying. This can be formulated also in N
116
terms of the poles in the momentum representation of 2-point function (see Refs. 11, 18). The poles of ( )BTZ are exactly the complex quasi-normal modes (2) while that of ( )Ads are the real normalizable modes (3). Depending on the value of LT, one of the two terms in Eq. (9) dominates.16 For high temperature (LT is large) the BTZ is dominating, while at low temperature (LT is small) the thermal AdS is dominant. The transition between the two regimes occurs at 1/T = L. In terms of the gravitational physics, this corresponds t o the Hawking-Page phase t r a n ~ i t i 0 n . l ~ This is a sharp transition for large k, which is the case when the supergravity description is valid. The Hawking-Page transition is thus a transition between oscillatory relaxation at low temperature and exponential decay at high temperature.
3.4. The Puzzle and Resolution Thus, the AdS/CFT correspondence predicts that the CFT dual to gravity on Ad& is rather peculiar. Even though, it is in finite volume, the relaxation in this theory is combination of oscillating and exponentially decaying functions. This immediately raises a puzzle: how this behavior is consistent with the general requirement for a unitary theory in finite volume to have only quasi-periodic relaxation? A resolution of this puzzle was suggested in Ref. 6. It was suggested that additionally to the size L there exists another scale in the game. This scale appears due to the fact that in the dual CFT at high temperature the typical configuration consists of multiply wound strings which effectively propagate in a much bigger volume, Leff kL. The gravity/CFT duality however is valid in the limit of infinite k in which this second scale becomes infinite. So that the exponential relaxation corresponds to infinite effective size Leff that is in complete agreement with the general arguments. At finite k the scale Leff would be finite and the correlation function is expected to be quasi-periodic with two periods: 1/L and l/Leff. The transition of this quasi-periodic function to combination of exponentially decaying and oscillating functions when Leff is infinite then should be similar to what we have observed in the case of free fermions when L was taken to infinity. N
4. Black Hole Unitarity: Finite k
That relaxation of black hole is characterized by a set of complex frequencies (quasi-normal modes) is a mathematically precise formulation of the lack of unitarity in the semiclassical description of black holes. The unitarity prob-
117
lem was suggested to be resolved within the AdS/CFT correspondence. l5 Indeed, the theory on the boundary is unitary and there should be a way of reformulating the processes happening in the bulk of black hole space-time on the intrinsically unitary language of the boundary CFT. The analysis of the relaxation is helpful in understanding how this reformulation should work. Before making comments on that let us note that the loss of information in semiclassical black hole is indeed visible on the CFT side. It is encoded in that exponentially decaying contribution to the 2-point correlation function. For the CFT itself this however is not a problem. As we discussed above, the finite size unitarity is restored at finite value of k. This however goes beyond the limits where the gravity/CFT duality is formulated. Assuming that the duality can be extended to finite k an important question arises: What would be the gravity counter-part of the duality at finite k ? Obviously, it can not be a semiclassical black hole. The black hole horizon should be somehow removed so that the complex quasinormal modes (at infinite k) would be replaced by real (normal) modes when k is finite. Below we consider two possibilities of how it may happen.
4.1. Fractal Brick Wall
It was suggested in Ref. 7 that the quantum modification of the black hole geometry, needed for the restoring the PoincarC recurrences, can be modeled by the brick wall. Here we elaborate on this interesting idea. The brick wall is introduced by placing a boundary at small distance E from the horizon and cutting off a part of the space-time lying inside the boundary. The effect of the boundary on the quantum fields is implemented by imposing there the Dirichlet boundary condition. Originally, the brick wall was introduced by 't Hooft2' for regularizing the entropy of the thermal atmosphere out-side black hole horizon. With this regularization the quantum entropy S, correctly reproduces the proportionality of the black hole entropy to the horizon area A rf-'. Assuming that E is taken to be of the order of the Planck length, so that Newton's constant is G c d P 2 , one can argue that the black hole entropy is correctly reproduced in this approach. Later on it was, however, realized that the brick wall divergence is actually a UV divergence. One can introduce a set of the Pauli-Villars fields with masses set by parameter p, which plays the role of the UV regulator. Taking into account the contribution of the regulator fields in the entropy of the quantum atmosphere the brick wall can be removed.21 The entropy then is proportional to certain power of the UV regulator, S,
-
-
-
-
118
In our story of black hole relaxation the brick wall indeed gives the wanted effect: once the brick wall has been introduced the quasi-normal modes disappear completely and are replaced by a set of the real (normal) modes. This happens because the effective infinite size region near horizon is now removed and the whole space is the finite size region between the brick wall and the boundary at spatial infinity. In such a system we expect periodicity with the period set by the brick wall parameter E as tbw l/Tln(l/c). This periodicity shows up in the boundary CFT correlation functions rather naturally. Indeed, these correlation functions are constructed from the bulk Green’s function which describes propagation of the perturbation between two points on the boundary through the bulk. In the present case the perturbation from a point 4 on the boundary goes along null-geodesic through the bulk, reflects at the brick wall and returns to the same point 4 on the boundary. The time which the perturbation travels gives the periodicity for the boundary theory and it equals tbw. Matching tbw and l/Leff gives the relation between brick wall regulator E and parameter k of the large N boundary CFT. This probably should be enough for the explaining and reproducing the second time scale of the boundary CFT from the gravity side. The time tbw is however much smaller than the Poincar6 recurrence time which is A expected t o be of the order, tp e: . So how t o get this time scale in the model with the brick wall? We notice that the brick wall should not be ideally spherical. The possible complexity of the shape is not restricted. It may even be fractal. In order to serve as a regulator for the quantum entropy calculation brick wall should just stay at mean distance E from the horizon, but its shape can be arbitrary. For the recurrence time the shape is however crucial. In the absence of the spherical symmetry the perturbation emitted from the point 4 on the boundary (which is still a circle) at spatial infinity goes along null-geodesic through the bulk, reflects from the brick wall, goes back and arrives at completely different point 4’ on the boundary at spatial infinity. Only after a number of back and forth goings between two boundaries the perturbation can manage to arrive on the boundary at the same point where it was initially emitted. This number can be very large and it sets the periodicity for the boundary theory. The emerging geometric picture is standard set up for the system having classical chaos. Indeed, generic deviations from the spherical symmetry of one of the boundaries leads to chaotic behavior of the geodesics. This means that the 2-point functions on the boundary would generically have chaotic time evolution. The optical volume V between two boundaries N
N
119
seems to be the right quantity to measure the size of the phase space of the chaotic geodesics. Since S, V the recurrence time t p ev gives the right estimate for the Poincark time. In this picture the information sent to black hole eventually comes back. The characteristic time during which it should happen is set by the Poincark recurrence time t p . The classical chaos of the geodesics manifests in the (normal) frequencies. The latter are the eigenvalues of the Laplace-type operator considered on the classical geometry. As we know from the relation between classical and quantum chaos, the chaos of the geodesics in the classical system manifests in that the eigen values of the quantum problem are randomly distributed. Thus, the normal frequencies will be random numbers. This again means that the 2-point function on the boundary (we expect that the normal modes are still poles in the momentum representation of the correlation function) is chaotic function of time. The irregularity of the shape of the brick wall may actually be physically meaningful. It can model the fluctuating quantum horizon. It may also be a way of representing the so-called stretched horizon (see Ref. 2 2 ) . N
N
4.2. Worm-hole Modification: BTZk
The horizon can be removed in a smooth way by modifying the black hole geometry and making it looks like a worm-hole. As an example we present here a modification of the BTZ metric ( l ) , 1 ds2 = -(sinh2 y -) dt2 d y 2 cosh2 y d 4 2 , k2 which we call BTZk. The horizon which used t o stay at y < 0 disappears in metric (10) if k is finite. The whole geometry now is that of worm-hole with the second asymptotic region a t y = -m. The two asymptotic regions separated by horizon in classical BTZ metric can now talk to each other leaking the information through the narrow throat. The metric (10) is still asymptotically AdS although it is no more a constant curvature space-time. The Ricci scalar 2 R=[(k2 1) 3k4 sinh4 y 5k2 sinh2 y] (11) (k2 sinh2 y 1)2
+
+
+
+
+ +
+
+
approaches value -6 at infinite y and -2(k2 1) at y = 0 where the horizon used to stay. The normal frequencies in the space-time with metric (10) are real and are determined by the normalizability and the Dirichlet boundary condition at both spatial infinities. Since the space-time (10) is asymptotically AdS one can use the rules of the AdS/CFT duality and
120
calculate the boundary correlation function. Technically it is more difficult than in the standard BTZ case since (10) is not maximally symmetric space. But the result should be a periodic in time function with the period set by parameter k. It would be interesting to do this calculation and see if this correlation function makes sense from the point of view of the expected behavior of the boundary CFT a t finite k. One can calculate the entropy of the thermal atmosphere in the metric (10). It is now finite with no need for introducing the brick wall. The entropy then behaves as S, kA that is the right answer for the Bekenstein-Hawking entropy of BTZ black hole. Thus, the modification (10) gives us the right entropy and solves the unitarity problem. N
Acknowledgments
I would like t o thank D. Birmingham and I. Sachs for enjoyable collaboration and many useful discussions. I also thank G. Arutyunov, J. Barbon, A. Morozov and N. Kaloper for important discussions. References 1. A. L. Fetter and J. D. Walecka, Quantum Theory of Many-Particle Systems, McGraw-Hill Book Company (1971). 2. L. Susskind, hep-th/0204027. 3. S. W. Hawking, Phys. Rev. D14, 2460 (1976). 4. 0. Aharony, S. S. Gubser, J. Maldacena, H. Ooguri and Y. Oz, Phys. Rev. 323,183 (2000). 5. G. T. Horowitz and V. E. Hubeny, Phys. Rev. D62, 024027 (2000). 6. D. Birmingham, I. Sachs and S . N. Solodukhin, Phys. Rev. D67, 104026 (2003). 7. J. L. F. Barbon and E. Rabinovici, JHEP 0311,047 (2003). 8. I. Sachs, Fortsch. Phys. 52,667 (2004). 9. J. L. F. Barbon and E. Rabinovici, Fortsch. Phys. 52,642 (2004). 10. V. Cardoso and J. P. S. Lemos, Phys. Rev. D63, 124015 (2001); D.Birmingham, Phys. Rev. D64,064024 (2001). 11. D. Birmingham, I. Sachs and S. N. Solodukhin, Phys. Rev. Lett. 8 8 , 151301 (2002). 12. V. Balasubramanian, P. Kraus and A. E. Lawrence, Phys. Rev. D59, 046003 (1999). 13. P.Di Francesco, P. Mathieu and D. Senechal, Conformal Field Theory, New York, USA: Springer (1997). 14. S. Carlip and C. Teitelboim, Phys. Rev. D51,622 (1995).
15. J. M. Maldacena, hepth/0106112. 16. J. M. Maldacena and A. Strominger, JHEP 9812,005 (1998).
121
17. L. Chekhov, hep-th/9811146; E. Keski-Vakkuri, Phys. Rev. D59, 104001 (1999). 18. U. H. Danielsson, E. Keski-Vakkuri and M. Kruczenski, Nucl. Phys. B563, 279 (1999). 19. S. W. Hawking and D. N. Page, Commun. Math. Phys. 87,577 (1983). 20. G. 't Hooft, Nucl. Phys. B256,727 (1985). 21. J. G. Demers, R. Lafrance and R. C. Myers, Phys. Rev. D52, 2245 (1995). 22. N. Iizuka, D. Kabat, G. Lifschytz and D. A. Lowe, Phys. Rev. D68, 084021 (2003).
DEFORMED COORDINATE SPACES DERIVATIVES
J. WESS Universitat Munchen, Fakultat fur Physik Theresienstr. 37, 0-80333 Munchen, Germany and Max- Planck- Institut f u r Physik Fohringer Ring 6, 0-80805 Munchen, Germany E-mail: [email protected] The concept of derivatives on quantum spaces is worked out in detail. Special example of the @-deformedcoordinate space is analysed. It is shown that it is possible t o construct a deformed Lorentz symmetry for this space. Fields are defined in such a way that they transform with respect to the deformed symmetry.
1. Introduction
This lecture is based on joint work with Marija Dimitrijevi6, Laxisa Jonke, Frank Meyer, Lutz Moller, Efrossini Tsouchnika and Michael Wohlgenannt.' The aim of this lecture is to clarify the concept of derivatives on quantum spaces.2 These derivatives are an essential input for the construction of deformed field equations such as the deformed Klein-Gordon or Dirac equation^.^ These deformed field equations are in turn the starting point for field theories on quantum spaces. For a given coordinate space there are in general many ways t o define derivatives.* We shall try to develop a general concept of such derivatives into which all the different sets of derivatives fit and that allows us by adding additional requirements - usually based on symmetries - to reduce the number of possible derivatives. 2. Deformed Coordinate Spaces Let me first remind you of the concept of deformed coordinate spaces (DCS) which we will use as quantum spaces. DCS are defined in terms of coordi122
123
nates F , p = 1. . . n and relations. Examples of such relations are 1. Canonical relations5
[P, 27 = i P V ,
(1)
for constant 0 it leads to the so called &deformed coordinate space (0-DCS). 2. Lie-type relations6 where the coordinates form a Lie algebra = iCf”P,
[F, 27
(2)
Cr” are the structure constants. Among these is the r;-deformed quantum space (K-DCS).~ 3. Quantum group relations: 2Pg” = -RP”,$PP“ 1
(3)
9
where the R-matrix defines a quantum group.8 These are the q-deformed spaces (q-DCS). The DCS is the algebra d n , this is the factor space of the algebra freely generated by the elements ri-p divided by the ideal generated by the relation^.^ We not only consider polynomials in d, but formal power series as well. In short all polynomials of the coordinates 2” that can be transformed into each other by using the relations are linearly dependent. For the examples listed it can be shown that the dimensions of the vector spaces of polynomials with given degree are the same as for commuting coordinates. This is the socalled Poincark-Birkhoff-Witt property.1°
3. Derivatives Derivatives are maps of the DCS
5:dn+d*.
(4)
They are usually defined by maps on the coordinates, and therefore on the free algebra defined by them. To define a map on the factor space DCS, derivatives have to be consistent with the relations defining the DCS. They also should lead to a Leibniz rule. A very general ansatz for the action of a derivative on the coordinates is:
[&, 2 7
= 5(;
+
c
A;p’-.pJ
j
5 .. . P1
*
(5)
124
The coefficients AY’””3 are complex numbers. They have to be chosen such that Eq. ( 5 ) is consistent with the relations. Having found such coefficients a Leibniz rule can be derived because ( f i j )can be computed from Eq. ( 5 ) , f and ij are elements of A,. Maps can also be defined on the set of derivatives:
E : 6-+Sl,
6;
= E/(6)&.
(6)
The matrix E depends on the derivatives 8 only, not on the coordinates. Because the derivatives 8 are maps on DCS the new derivatives will be as well. If E is invertible and if the matrix E starts with the Kronecker symbol as derivative-independent term we obtain from Eq. (6) again derivatives in the sense of Eq. (5). All derivatives satisfying the consistency condition that have been found up to now are related by such transformations. We shall discuss the 8-DCS here, this is the simplest case. The relations (1) are consistent with
[a,, ?”I
= 6;
.
(7)
A short calculation shows: ( p y- y p - ip”)= ( p p- p
af
p
-~ 2. 9P”
)af .
(8)
This is sufficient t o prove consistency. Eq. (7) leads to the Leibniz rule by applying Eq. (7) to the product of the two functions f i j
4 ( f i j ) = (&,f)ij + f ( 8 P i j )
*
(9)
A short calculation shows that
[6&]
=0
(10)
is compatible with the relations (7). We can assume that the derivatives commute and define an algebra that way. The Leibniz rule (9) can be algebraically formulated as a comultiplication:
A i P = 6,, @ 1 + 1 8
aP.
It is compatible with the Lie algebra (10): [A6P,A&] = 0 , (12) and it is coassociative. Thus, Eqs. (10) and (11) define a bialgebra, the q-deformed bialgebra of translations in the 8-DCS.
125
Other sets of derivatives can be obtained from bp by a transformation (6). In general, such derivatives will not have defining relations that are linear in 8 such as Eq. (7). They will also have more complicated comultiplication rules. Thus the definition (7) singles out a specific type of derivatives. Moreover they will transform linearly under a &deformed orthogonal or Lorentz group. We shall now show that.
4. Deformed Symmetry Algebra
A deformed orthogonal group or a deformed Lorentz group will be a deformation of the transformations
where wpu are the parameters of the infinitesimal orthogonal or Lorentz transformations. The corresponding Lie algebra satisfies:
[L, 1s: = L x , t , (w x w’),” = - (w,Owb - w; “w):
.
(14)
The map (13) can be obtained from a differential operator (angular moment um)
6,
= -xuwupap,
This concept can be lifted to the 8-DCS.
This result was first obtained in Ref. 11. For 8 = 0 Eq. (16) agrees with the undeformed equation (13). In Eq. (16) coordinates transform into derivatives. The additional terms are needed to make the deformed Lorentz transformation compatible with the relation (1). The map &, is really a map on A?. This can be shown in a short calculation, applying Eq. (16) to the relations (1). We find
,j(i~y -y p -i p u )
=
( p j y - p p - iew)$,
+ ( p p - p p - ipw) wpv + ( y p - p p v
- i8w) wp” .
(17)
Analagous to Eq. (15), the transformation (16) can be generated by a differential operator
126
This allows us to calculate the transformations of the derivatives:
and the algebraic relations of
iW:
[iw,
&]
=iwxw,
That iwis a map on d, follows from the fact that 6 and 2 are. The comultiplication can be calculated by applying Eq. (18) to the product of two functions f i j . We find: i A&, = iw 8 1 1 8 iw - - ( o U p w , p - e’pwup)6p 8 (21) 2 This result has recently been obtained by M. Chaichian et al. in Ref. 12. This coproduct is coassociative because iufij& is associative: iwfijwiL = (iwfij)& + fij(iwwi-L)-2i ( ~ p ~ -, pe’pwup)(8pfij)8p&
ap.
+
=
(iuf)ij& + f(&wijk)- ~ ( B y ’ ” W u p- O u p w u p ) ( ~ p f ) 8 p ( i j &(22) ).
The Lorentz algebra by itself does not form a bialgebra. Derivatives appear in the comultiplication rule (21). We can, however, interpret Eq. (21) as a comultiplication rule for the Poincar6 algebra (translation included). Then Eqs. (lo), (19) and (20) define an algebra, the &deformed Poincar6 algebra with the comultiplication (11) and (21). We have obtained the 8deformed Poincark bialgebra. The algebra relations are the same as for the undeformed Poincar6 algebra, the comultiplication is deformed. &deformed Poincar6 bialgebra:
[&, *
$4 = 0 , [iw, Bp] = wptL2p, A
[6,,6:]
u
= iwwxw‘, (w x w ) ; v = - ( u p :
‘
(23)
-wpuw:),
A& = b p 8 1 + 1 8 G p , i
~i~= iw8 1 + 1 8 iw+ -2 (ep’w,p
-v
~
~
& .8
p 8)
~
(24)
That the algebraic relations and the comultiplication rules are compatible can be verified directly. 5 . Fields
On our way to a field theory we have to define fields. They are elements of with certain transformation properties. For a scalar field we define:
4
A
=
n
-Cpapq5
A
,
.
A
n
and 6 ~ q 5= -JWq5.
(25)
127
The translation is parametrized by the constant vector
A
= dpbT4 = -<'3p$p$
&'2p$
(26)
and A , . . .
= a,&$ *
A
= -iw&$ - [SP,&]$
+w / @ .
= -iw(3p$)
(27)
This is the transformation law of a vector field:
i p v p = -pGpvp- jwvp +w p q p .
(28)
For a tensor or spinor field we define the transformation law as follows: L
A
S ~ T A= - t p a p F A - i,,,TA
where M,'," isfies
+W / M , " A " P ~ ,
(29)
is a representation of the undeformed Lorentz agebra. It sat-
[MP', M"'
1=ll~Xj-p"
+l
-l
l g " ~ ~X l l ~ " ~ g X
l g X ~ ~ " ,
(30)
where qp' is the metric depending on the algebra, Kronecker symbol for SU(n) or Minkowski metric for SU(1,n - 1). It is easy to see that the transformations (29) represent the algebra (23). For the bialgebra we have to specify the comultiplication. For the translations comultiplication is straightforward: &(FA
8F B ) = -
~ P A ( ~ ~8) F(BF) ~
= (&FA) 8 F B )
+FA 8 (&FBI.
(31)
For the Lorentz transformations we have t o use the comultiplication (24). We obtain:
8, ( F A@?B)
2i ( 6 p u W / - ~ p u ~ u p )a p T A 8 8 p F B . A , .
= (8LFA)@?B + F A @ ( 8 L F B ) +
(32) The compatibility of the algebraic relations with the comultiplication can again be verified. We have established a tensor calculus on tensor and spinor fields. After these considerations it is clear that the Klein-Gordon equation and the Dirac equation are covariant
128
1. Klein-Gordon equation:
The sign of m2 depends on the metric qp” 2. Dirac equation:
4
transforms like a spinor and the y’s are t h e usual y matrices. where Invariant Lagrangian with interaction terms can be constructed with the above tensor calculus for tensor and spinor fields.
References 1. M. Dimitrijevit, L. Jonke, L. Moller, E. Tsouchnika, J. Wess and M. Wohlgenannt, Eur. Phys. J. C31, 129 (2003), hep-th/0307149; M. Dimitrijevit, F. Meyer, L. Moller and J. Wess, Eur. Phys. J. C36, 117 (2004), hepth/0310116; M. Dimitrijevit, L. Moller and E. Tsouchnika, Derivatives, forms and vector fields o n the K-deformed Euclidean space, submitted to J . Phys. A, hep-th/0404224. 2. J. Wess and B. Zumino, Nucl. Phys. Proc. Suppl. B18, 3002 (1991); S. L. Woronowic, Commun. Math. Phys. 122, 125 (1989). 3. A. Nowicki, E. Sorace and M. Tarlini, Phys. Lett. B302, 419 (1993), h e p th/9212065; J. Lukierski, H. Ruegg and W. Ruhl, Phys. Lett. B313, 357 (1993). 4. Y. I. Manin, Commun. Math. Phys. 123,163 (1989). 5. C.S. Chu and P.M. Ho, Nucl. Phys. B550, 151 (1999), hep-th/9812219; V. Schomerus, JHEP 9906,030(1999), hep-th/9903205; J. Madore, S. Schraml, P. Schupp and J. Wess, Eur. Phys. J . C16, 161 (2000), hep-th/0001203. 6. B. JurEo, S. Schraml, P. Schupp and J. Wess, Eur. Phys. J . C17,521 (2000), hepth/0006246. 7. J . Lukierski, A. Nowicki, H. Ruegg and V.N. Tolstoy, Phys. Lett. B264, 331 (1991); J. Lukierski, A. Nowicki and H. Ruegg, Phys. Lett. B293,344 (1992). 8. A. Klimyk and K. Schmiidgen, Quantum Groups and Their Representations, Springer (1997). 9. J. Wess, Gauge theory beyond gauge theory, prepared for Workshop on the Quantum Structure of Spacetime and the Geometric Nature of Fundamental Interactions (1st Workshop of RTN Network and 34th International Symposium Ahrenshoop on the Theory of Elementary Particle), Berlin, Germany, 4-10 Oct 2000, Fortsch.Phys. 49,377 (2001). 10. E. Abe, Hopf Algebras, Cambridge University Press (1980). 11. F. Koch and E. Tsouchnika, Construction of 0-Poincare‘ algebras and their invariants on Mo, to be published. 12. M. Chaichian, P.P. Kulish, K. Nishijima and A. Tureanu On a LorentzInvariant Interpretation of Noncommutative Space- Time and Its Implications on Noncommutative QFT, hep-th/0408069.
11. SHORT LECTURES
This page intentionally left blank
DEFORMED COHERENT STATE SOLUTION TO MULTIPARTICLE STOCHASTIC PROCESSES
B. ANEVA INRNE, Bulgarian Academy of Sciences, 1784 Sofia, Bulgaria E-mail: boyka. anevaocern. ch Deformed algebraic generalized coherent states are considered as the q-analogues of the conventional undeformed harmonic-oscillator algebra squeezed states. It is shown that the boundary vectors in the matrix-product states approach t o multiparticle diffusion processes are deformed coherent or squeezed states of a deformed harmonic-oscillator algebra. A deformed squeezed and coherent states solution to the n-species stochastic diffusion boundary problem is proposed and studied.
1. Introduction By origin coherent states are quantum states, but at the same time they are parametrized by points in the phase space of a classical system.lY2 This makes them very suitable for the study of systems where one encounters a relationship between classical and quantum descriptions. From this point of view, interacting many-particle systems with stochastic dynamics provide an appropriate playground to enhance the utility of generalized coherent states. A stochastic process is described in terms of a master equation for the probability distribution P(si,t ) of a stochastic variable si = 0,1,2, ...,n - 1 at a site i = 1,2, ...,L of a linear chain. A configuration on the lattice at a time t is determined by the set of occupation numbers s1,sp, ..., SL and a transition to another configuration s' during an infinitesimal time step dt is given by the probability r(s,s')dt. In the bulk dynamics is restricted to changes of configuration at two adjacent sites only, and the two-site rates F = rff,i, j , k,1 = 0,1,2, ..., n - 1 are independent from the position in the bulk. At the boundaries, i.e. sites 1 and L , additional processes can take place with single-site rates LL and RL, i, k = 0,1, ..., n - 1. In the matrix-product states approach to stochastic dynamics3i4 the stationary probability distribution of a system with nearest-neighbour in131
132
teraction in the bulk and single-site boundary terms can be expressed as a product of (or a trace over) matrices that form a representation of a quadratic algebra determined by the dynamics of the process. For diffusion processes that will be considered in this paper, r;: = g i k and the n-species diffusion quadratic algebra has the form g i k D i D k -9kiDkDi = X k D i
-XiDk,
(1)
where g i k and s k i are positive (or zero) probability rates, xi are c-numbers and i, k = 0,1, ...,n - 1. (No summation over repeated indices in Eq. (I).) The algebra has a Fock representation in an auxiliary Hilbert space where the n generators D act as operators. For open sytems (with boundary processes) the stationary probability distribution is related to a matrix element in the auxiliary vector space P(S1,..*> SL)
with respect to the vectors) . 1 ditions
(Wl(LFDk 4-
Xi)
= (wlDslDs2...DsL ).I
,
(2)
and (wl, determined by the boundary con= 0,
(@Dk
- xi)/.)
=0,
(3)
where the x-numbers sum up to zero, because of the form of the boundary rate matrices n-1
n- 1
L; = -
E L"
3'
j=O
Ri
n- 1
Rj,
=j=O
E X i
=o.
(4)
i=O
These relations simply mean that one associates with an occupation number si = k at position i a matrix Dsi = Dk (i = 1 , 2 , ..., L; k = 0,1, ...,n - 1) if a site i is occupied by a k-type particle. The number of all possible configurations of an n-species stochastic system on a chain of L sites is nL and this is the dimension in the configuration space of the stationary probability distribution as a state vector. Each component of this vector, i.e. the (unnormalized) steady-state weight of a given configuration, is (a trace or) an expectation value in the auxiliary space given by Eq. (2). The quadratic algebra reduces the number of independent components to only monomials symmetrized upon using the relations (1). In the known examples of exactly soluble 2- and 3-species models, through the matrix product ansatz, the solution of the quadratic algebra is provided by a deformed bosonic oscillator algebra, if both g i k and g k i differ from zero, or by infinite-dimensional matrices, if g i k = 0. It can be shown that in the general n c a ~ e ,if~all ? ~parameters x i are equal to zero on the
133
right hand side of Eq. (l),the homogeneous quadratic algebra defines a multiparameter quantized non-commutative space realized equivalently as a q-deformed Heisenberg algebra'>' of n oscillators. Alternatively a solution of the non-homogeneous algebra with 2-terms on the right hand side of Eq. ( 1 ) can be found trough a relation to a lower-dimensional quantum space realized equivalently as a lower-dimensional deformed Heisenberg algebra. Proposition 1.1. The boundary vectors with respect to which one determines the stationary probability distribution of the n-species diffusion process are generalized, coherent or squeezed states of the deformed Heisenberg algebra underlying the algebraic solution of the corresponding quadratic algebra. We first review the known basic properties of the deformed oscillator coherent states and then define a deformed squeezed state of a pair of deformed oscillators by analogy with the conventional squeezed states as the eigenstate of the deformed boson operators linear combination and study their squeezing properties. Such a q-generalization of the conventional undeformed squeezed states is not known. As a physical application of the deformed coherent and the considered squeezed states we obtain the boundary problem solution of the general n-species stochastic diffusion process.
2. Coherent States of a q-Deformed Heisenberg Algebra We consider an associative algebra with defining relations
aa+ - qa+a = 1,
qNa+ = qa+qN,
N
- -1
4 a-9
(5)
aqN,
where 0 < q < 1 is a real parameter and a+a = = [ N ] . A Fock representation is obtained in a Hilbert space spanned by the orthonormal basis %lo) = In), n = 0, 1,2, ... and (nln') = dnn,: N
alO) = 0 ,
+
aln) = [n]'/21n - 1) ,
+
a+ln) = [n 1]1/21n 1). ( 6 )
The Hilbert space consists of all elements = Cr=o fnln) with complex fn and finite norm with respect to the scalar product = C,"==, The q-deformed oscillator algebra has a Bargmann-Fock representation on the Hilbert space of entire analytic functions. Generalized or q-deformed coherent s t a t e ~ are ~ ~ defined '~ as the eigenstates of the deformed annihilation operator a and are labelled by a con-
If)
(flf)
lfnI2.
134
tinuous (in general complex) variable z:
&.
These vectors belong to the Hilbert space for 1 . ~ 1<~ [m] = The scalar product of two coherent states for different values of the parameter z is non-vanishing
)'.I(
=
c !I.[ O0
(52')"
-
,
-- egZ ,
0
and they can be properly normalized with the help of the q-exponent on the RHS of Eq. (8):
The q-deformed coherent states reduce to the conventional coherent states of a one-dimensional Heisenberg algebra in the limit q -+ 1-. These generalized coherent states carry the basic characteristics of the conventional ones, namely continuity and completeness (resolution of unity I = J Iz)(zI exp,(-lz12)diz). Hence one can expand any state in the coherent states If) = Jdilz) exp,(-1z12)(21 f) and thus
(zla+lf)
= .f (> .
7
(.I.lf)
= Qf
d
(.I
7
( W l f )= z-&f (2)
7
(10)
which is the Bargmann-Fock representation of the deformed oscillators and number operator. 3. Squeezed States of a Deformed Oscillator Algebra
Following the analogy with the (conventional) squeezed oscillator states,11v12generated by the action of a squeezed operator through a linear transformation, we are lead by this idea t o keep the linear structure of the deformed squeezing operator and assume an analogous definition. Proposition 3.1. Let a7a+ and qN generate a deformed Heisenberg algebra with the equivalent f o r m of defining relations [a7
.+I
= qN7
qN a = q - l a q N 7
qNa+ = q a + q N .
(11)
Then there is a two-parameter-dependent linear map to a pair of "quasioscillators" with a "quasiparticle" number operator N
A = pa + ua+ ,
A+=jia++Da.
(12)
135
These operators generate a deformed Heisenberg algebra with relations
[A,A+]= $/ ,
$/A'
qNA = q - ' A $ / ,
= qA'$/,
(13)
provided 4."/ = (lp12 - Iv12)qN.In the limit q + 1- the relation between the parameters of the conventional squeezed state is rec0vered.l' In the deformed "quasi"-oscillator algebra Fock representation space with a vacuum lo), one can define a normalizable coherent state I<), as the eigenvector of the annihilation operator A:
Proposition 3.2. A squeezed state of the deformed creation and annihilation operators is a normalized solution of the eigenvalue equation: (pa
+ .+)Cl ,
p, 4
s
= CIC, P, 4
s
=4
0 s .
(15)
This proposition is motivated by the analogy with the non-deformed case and by the fact that such normalized eigenstate vectors of the written above linear combination of q-deformed oscillators appear in the solution of the boundary problem of a many-particle non-equilibrium system. To show the effects of squeezing we consider the Hermitian quadrature operators x = h ( a a+) and p = &(a - a+), where the boson operators obey the relations of the form (5) and consequently the operators x and p satisfy the deformed canonical commutation relation [ z , p ] = i q N . The variances (62)' = ((x - (x))') and similary (Sp)' = ( ( p - (p))') in any state obey a generalized Heisenberg-Robertson inequality of the form (Sx)2(6p)22 i l ( [ x , p ] ) l ' . One can calculate now the variances in the deformed coherent states 12) and finds that the deformed uncertainties are equal:
+
1
( 6 x 1= ~ ( ~ p ) ' = 5 exp, ( ( q - 1)1z1')
.
(16)
The deformed coherent states are thus states of equal uncertainties only. Minimum-uncertainty states are labeled by the value z for which the qexponent in Eq. (16) has a minimum. In the limit q -+ 1- for 0 < )z)' < 00 the equality of the undeformed uncertainties in the Glauber coherent states is recovered. We proceed further with the discussion of the algebraic states I<, p,v,q ) which reveal stronger squeezing properties, generalizing thus the undeformed case. To calculate the uncertainties with respect to these states
136
we first write the inverse of the linear map in Eq. (12)
a=
P 1P12 -
-
U
ITA- IPI2- 1Ul2
A',
a+=
-U
Id2- l V l 2
A+
1PI2 - 1Ul2
A+, (17) \
I
where lp12 - 1uI2 # 0, being the Jacobian of the linear transformation (12). Exploring the eigenvalue properties of the normalized coherent eigenstates of A,
(CIAIC)S = 4-7
(CIA+lC)S =
c
7
(CI4NlOs
= (CIq[%
IC)S
= ep-l)'['z
, (18)
we calculate the corresponding mean values. This yields a non-equality of the q-deformed uncertainties which read explicitly
At this step we recall the known definition of a squeezed state13 requiring for one of the variances to be smaller than the equal uncertainties common minimal value determined by the equality (16) as the minimum in the variable z of the q-exponential function. F'rom the analyses of the function 'z1 exp, ( ( q - 1)1<12)it follows that the minimum of this function is the finite
+
limit (l+(l-q))T for I() is a squeezed state if
+
&.Hence, according to the expression (19),
which is satisfied provided the parameters (in general complex) p , v of the linear transformation (12) are chosen in such a way that O<
IP - Y l 2 (1PI2- 14")"
and thus the criterion
(W2 holds. The ratio at the very right hand side of Eq. (22) is the basic hypergeometric series 1a.0((1- q)ICI2;q , (q - 1)). Alternatively, from Eq. (20)
137
is satisfied if
which gives
For 0 < q < 1 the values p = f u are not admissible. The inequality (23) (or (26)) together with the condition (22) (or (25)) for the parameters p,Y, q define the eigenstates I<) of the linear combination pa va+ of the deformed boson operators as generalized squeezed states. In the limit q -+ 1- the corresponding expressions for the x , p uncertainties with respect to the conventional harmonic oscillator squeezed states12 are recovered. This analogy with the squeezing properties of the quadratures of the boson creation and annihilation operators justifies, in our opinion, the proposed definition of a q-deformed squeezed state in Eq. (15) as a q-generalization of the conventional squeezed states.
c,
+
4. Deformed Squeezed and Coherent State Solution to the n-Species Diffusion Process We consider a diffusion process with n species on a chain of L sites with nearest-neighbour interaction with exclusion, i.e. a site can be either empty or occupied by a particle of a given type. On successive sites the species i and k exchange places with probability g i k d t , where i, k = 0,1,2, ...,n - 1. With i < k, g i k are the probability rates of hopping to the left, and S k i to the right. The process is totally asymmetric if all jumps occur in one direction only, and partially asymmetric if there is a different non-zero probability of both left and right hopping. In most studied examples of open systems one considers phase transitions inducing boundary processes when a particle of type k, k = 1 , 2 , ..., n - 1 is added with a rate LO, and/or removed with a rate LE at the left end of the chain, and it is removed with a rate RE and/or added with a rate Ri at the right end of the chain. The advantage of the matrix-product state approach is that important physical quantities such as multiparticle correlation functions, currents, density profiles, phase diagrams can be obtained from the representations of the quadratic algebra (1). The problem to be solved is to find matrix representations of the quadratic algebra consistent with the boundary conditions (3), namely that the combinations (LF& xi) and ( R f D k - X i )
+
138
have common vectors with eigenvalue zero, where the only nonvanishing boundary rates are L,k,Ljh, RE, Rjh, k = 1 , 2 , ..., n - 1. The algebra for the n-species open asymmetric exclusion process of a diffusion system coupled at both boundaries to external reservoirs of particles of k e d density has the form
Dn-lDO
- qDoDn-l =
20 -
Xn- 1 Do
Qn-l,O
where k , 1 = 1 , 2 , ..., n - 2, xo 4'-
QO,n-1
Sn-l,o
,
+ xn-l
q k l = -gkl ,
7
Qn-l,O
= 0 and
q k = - gk0 =%.
glk
gOk
gk,n-l
(28)
The equalities in the last formula, together with the relations g k = QOk = g k , n - l i g O k - gkO = g k , n - l - gn-1,k = g o p - 1 - gn-l,O, yield a mapping to the commutation relations of a q-deformed Heisenberg algebra of n - 1 oscillators a h , a:, k = 0 , 1 , 2 , ...,n - 2. A solution is obtained by a shift of the oscillators ao, a$
and by the identification of the rest of the generators with the remaining n - 2 creation operators a t
Dk,
k = 1 , 2 ,...)n - 2
For the phase transition inducing boundary processes, the system (3) for the boundary vectors reduces to the equations:
for k = 1 , 2 , ...n - 2, and the pair of equations
139
Making use of the explicit solution for Dn-l and DO as shifted deformed oscillators (with 20 = -21 = l),we rewrite Eq. (32) as
The latter equations, in accordance with Eq. (15), determine the boundary vectors as squeezed coherent states of the deformed boson operators ao, a$ corresponding to the eigenvalues
The explicit form of these vectors is readily written, namely (wl = a
-3.w
2.w
w
and ).I = eP- Cn=o~~nllIn)* We therefore conclude: the left and right boundary vectors are squeezed coherent states of the shifted deformed annihilation and creation operators D,-1 and Do,associated with the non-zero boundary parameters xn-l and 2 0 , and with eigenvalues depending on the right and left boundary rates: W n
I.( Cn=om
e q
n .
where the eigenvalues u and w are given by Eqs. (39-40). From the inverse linear maps, with R;-lL;-l - L;j.-'REPl # 0 , we obtain
with the help of which the mean values of the generators
DO, Dn-l and the
140
rest Dk for k = 1,2, ..., n - 2 are readily found
With these expressions at hand, it is easy to calculate the expectation value of any monomial of the form (wID,,D s2...Ds~1v) (where DSi = Dj for i = 1,2, ...,L, j = 0,1,2, ..., n - l),which enters the stationary probability distribution, the current, the correlation functions. One first makes use of the algebra to bring all generators Dk for k = 1 , 2 , ...,n - 2 to the very right or to the very left, which results in an expression of the expectation value as a power in DOand Dn-l. Then one writes the arbitrary power of Do, Dn-l as a normally ordered product of A and A+ to obtain, upon using the eigenvalue properties of the latter, an expression for the relevant physical quantity in terms of the probability-rate-dependent boundary eigenvalues v and w. Our results generalize the known solved examples. We note that if the boundary processes are such that there are only incoming particles of ( n- 1)th-type at the left boundary and only outgoing ( n - 1)th-type particles at the right boundary, i.e. Lt-l = R:-l = 0 in (75), then the eigenstate equations define the boundary vectors Iv) and (wl as q-deformed coherent states. Using the eigenvalue properties of the latter one can likewise obtain the physical quantities of interest for the system. The value q # 0 corresponds to a partially asymmetric while q = 0 to a totally asymmetric diffusion in the bulk of the n - 1-type particle. The deformed oscillator coherent states defined for 0 < q < 1 and for q = 0 provide a unified description of both the partially and the totally asymmetric hopping of a given type of particle.
Acknowledgements Most of this work was completed during a stay at the LMU, Munich and the author is very grateful to Julius Wess for the opportunity t o join his
141
theory group at the Physics Department there. The author is happy t,o take part in the organization and to participate the BW2003. References 1. A. M. Perelomov, Generalized Coherent states and their applications, Springer, NY, (1986). 2. W. M. Zhang, D. H. Feng and R. Gilmore, Rev. Mod. Phys. 6 2 , 867 (1990). 3. B. Derrida, M. R. Evans, V. Hakim and V. Pasquier, J. Phys. A 2 6 , 1493 (1991). 4. K. Krebs and S. Sandow, J. Phys. A 3 0 , 3163 (1997). 5. A. Isaev, P. Pyatov and V. Rittenberg, cond-matt/0103603. 6. B.Aneva, J. Phys.A35, 859 (2002). 7. J. Wess and B. Zumino, Nucl. Phys.(Proc. Suppl.) B 1 8 , 302 (1991). 8. D. Fairlie and C. Zachos, Phys. Lett. B256, 43 (1991). 9. M. Aric and D. D. Coon, J. Math. Phys. 17, 524 (1976). 10. B. Jurco, Lett. Math. Phys.21, 51 (1991). 11. D. Stoler, Phys. Rev D 1 , 3217 (1970); D 4 , 2308 (1971). 12. H. P. Yuen, Phys. RevA13, 2226 (1976). 13. K. Wodkiewicz and J. Eberly, J . Opt. SOC.Am.B2, 458 (1985).
NON-COMMUTATIVE GUTS, STANDARD MODEL AND C , P, T PROPERTIES FROM SEIBERG-WITTEN MAP
P. ASCHIERI Dipartimento d i Scienze e Tecnologie Avanzate, Universitci del Piemonte Orientale and INFN, Piazza Ambrosoli 5, I-15100, Alessandm‘a, Italy and Max-Planck-Institut fur Physik, Fohringer Ring 6, 0-80805 Munchen and Sektion Physik, Universitat Miinchen, Theresienstmpe 37, 0-80333 Munchen, Germany E-mail: [email protected] Noncommutative (NC) generalizations of Yang-Mills (YM) theories using SeibergWitten (SW) map are in general not unique. We study these ambiguities and see that SO(10) GUT, at first order in the noncommutativity parameter 8, is unique and therefore is a truly unified theory, while SU(5) is not. We then present the noncommutative Standard Model compatible with SO(10) GUT. We next study the reality, hermiticity and C, P, T properties of the Seiberg-Witten map and of these noncommutative actions at all orders in 8. This allows t o compare the Standard Model (SM) discussed in Ref. 1 with the present GUT inspired one.
1. Introduction There are different examples of noncommutative theories. We here concentrate on the case where noncommutativity is described by a constant parameter 6””. The commutation relations among the coordinates read [x” x”]= x” * x” - x” * x” = i6””, where the star product between functions f , g is given by f * g = f e:op” g . We do not claim that spacetime has exactly this noncommutativity, rather we are interested in investigating a mathematically sound gauge theory based on this easiest noncommutative structure. General aspects of this noncommutative theory will then probably be in common with more refined choices of 9. In particular, the c
+
142
143
choice P” = constant breakes the Lorentz group in a spontaneous way; in a bigger theory we would like to consider Bp”” (or the related B field) dynamical and not frozen to a constant value, thus recovering Lorentz covariance. One can also consider gauge theories with 0 nondynamical but frozen to a particular nonconstant value, linear in the coordinates, such that one has a (kappa) deformed Poincark symmetry, see Ref. 2. Using Seiberg-Witten map13 that relates commutative gauge fields to noncommutative ones in such a way that commutative gauge transformations are mapped in NC gauge transformations, one can construct NC gauge theories with arbirary gauge g r o u p ~ . ~These t ~ theories are invariant under both commutative and noncommutative gauge transformations. Along these lines noncommutative generalizations of the standard model and GUT theories have been studied.ly6 The SW map and the product allow us to expand these noncommutative actions order by order in 0 and to express them in terms of ordinary commutative fields so that one can then study the physics properties of these &expanded commutative actions, see for example Ref. 7. It turns out that given a commutative YM theory, SW map and commutative/noncommutative gauge invariance are in general not enough in order to single out a unique noncommutative generalization of the original YM theory. One can follow different criteria in order to select a specific noncommutative generalization. We here focus on a classical analysis, in particular imposing the constraint that the noncommutative generalization of the Standard Model should be compatible with noncommutative GUT theories. Another issue would be to single out a noncommutative SM or GUT that is well behaved at the quantum level. We refer to the problems relative to renormalization, see for example Ref. 8. On the other hand chiral gauge anomalies are absent in these model^.^ In this talk, following Ref. 6 , we present a general study of the ambiguities that appear when constructing NCYM theories. We then see that at first order in 0 there is no ambiguity in SO(10) NCYM theory. In particular no triple gauge bosons coupling of the kind BFFF is present. We further study the noncommutative SM compatible with SO(10). We next study the reality, hermiticity, charge conjugation, parity and time reversal properties of the SW map and of &expanded NCYM theories. This constraints the possible freedom in the choice of a “good” SW map. In Ref. 10 the C, P,T properties of NCQED were studied assuming the usual C, P and T transformations also for noncommutative fields. We here show that the usual C, P,T transformation on commutative spinors and
*
144
nonabelian gauge potentials imply, via SW map, the same C, P,T transformations for the noncommutative spinors and gauge potentials. We also see that CPT is always a symmetry of noncommutative actions. In Ref. 11 CPT is studied more axiomatically. The reality property of the SW map is then used to analyze the difference between the SM in Ref. 1 and the GUT inspired SM proposed here. It is a basic one, and can be studied also in a QED model. While in Ref. 1, and in general in the literature, left and right handed components of a noncommutative spinor field are built with the same SW map, we here use and advocate a different choice: if noncommutative left handed fermions are built with the SW map then their right handed companions should be built with the -0 SW map; this implies that both noncommutative ~ ! J L and q!JcL = -202 ~!JR* are built with the +6’ SW map. In other words, with this choice, noncommutativity does not distinguish between a left handed fermion and a left handed antifermion, but does distinguish between fermions with different chirality. This appears to be the only choice compatible with GUT theories.
2. Seiberg-Witten Map and NC Particle Models Consider an ordinary “commutative” YM action with gauge group G, and one fermion multiplet, J d42 $ Tr(F,,F,”) +Ti@!$. This action is gauge invariant under 6% = ipa(A)!$ where pa is the representation of G determined by !$. Following Ref. 5 the noncommutative generalization of this action is given by
where the noncommutative field strength @ is defined by d v z , - iff[&, A^,].The covariant derivative is given by
F,,
h
=
a,A,
-
The action (1) is invariant under the noncommutative gauge transformations
6G = i p a ( X )* 5 , 62, = a,A + iq[A, A,]. h
A
h
-The fields A, and X are functions of the commutative fields A, !$
and the noncommutativity parameter
(3) !$,A
via the SW map.3 At first order in
145
0 we have
&[A, e] = At
+ ~1 e P v { AdVP, A ~+} ~ B P " { F P ~ , +A ,o(e2), }
(4)
1
QA,A, el = A + , e y a , n ,
A , ) + o(e2), (5) @[*,A,O]= @ +1T ~ ~ " ~ ~ ( A , ) ~ ~ @ + ~ ~ ~ Y [ P Y IO(e2) ( A ~. () 6, )~ ~ ( A ,
+
A
8 In terms of the commutative fields the action (1) is also invariant under the ordinary gauge transformation SAP = d,A i [ A , A , ] ,6Q = i p @ ( A ) @ .In Eq. (1) the information on the gauge group G is through the dependence of the noncommutative fields on the commutative ones. The commutative gauge potential A and gauge parameter A are valued in the G Lie algebra, A = AaTa,A = AaTa;and from Eqs, (4)and ( 5 ) it follows that and are valued in the universal enveloping algebra of the G Lie algebra. However, due to the SW map, the degrees of freedom of are the same as that of A. Similarly to A^, also P is valued in the universal enveloping algebra of G. Now expression (1) is ambiguous because in Tr(FPV* p P v ) we have not specified the representation p(Ta). We can render explicit the ambiguity in (1) by writing
+
A^
A^
1
-Tr(&, g2
* W )=
c
cpTr(p(Fpv)* p(E'"V)) ,
(7)
P
where the sum is extended over all unitary irreducible and inequivalent representations p of G. The real coefficients cp parametrize the ambiguity in Eq. (7). They are constrained by requiring that in the commutative limit, 0 4 0, (7) becomes the correctly normalized commutative gauge kinetic term. The ambiguity (7) in the action (1) can also be studied by expanding (7) in terms of the commutative fields *,A , F . At first order in 0 we have A
Sgauge
.1
dimG
4g2
a=l
=--]d4x
FZ,FaPv
where 1 1 -DpabC Tr(p(Ta){p(Tb), p ( T C ) }= ) A(p)"r(ta{tb,tc}) TA(p)dabc.(9) 2 Here ta denotes the fundamental representation, and we are using that the completely symmetric DpabCtensor in the representation p is proportional
146
to the dabc one defined by the fundamental representation. In particular for all simple Lie groups, except S U ( N ) with N 2 3 , we have DZbC= 0 for any representation p. Thus from Eq. (8) we see that at first order in 8 the ambiguity (7) is present just for S U ( N ) Lie groups. Among the possible representations that one can choose in Eq. (7) there are two natural ones. The fermion representation and the adjoint representation. The adjoint representation is particularly appealing if we just have a pure gauge action, then, since only the structure constants appear in the commutative gauge kinetic term CaFivFaP”, a possible choice is indeed to consider only the adjoint representation. This is a minimal choice in the sense that in this case only structure constants enter (7). (It can be shown6 that in this case the gauge action is even in 0). If we also have matter fields then from Eq. (2) we see that we must consider the particle representation pq given by the multiplet Q (and inherited by G). In Eq. (7) one could then make the minimal choice of selecting just the pa representation. Along the lines of the above NCYM theories framework we now examine the SO(lO), the S U ( 5 ) and the Standard Model noncommutative gauge theories.
2.1. Noncommutative SO( 10) We consider only one fermion generation: the 16-dimensional spinor representation of SO(10) usually denoted 16+ (no relevant new effects appear considering all three families). We write the left handed multiplet as
VT
where i is the S U ( 3 ) color index and = -ia2 v ~ is * the charge conjugate of the neutrino particle V R (not present in the Standard Model). The gauge and fermion sector of noncommutative SO(10) is then simply obtained by replacing with 5; in Eq. (1). Notice that no linear term in 8, i.e. no cubic term in F can appear. This is so because SO(10) is anomaly free: D a b c = 0 for all p. In other words, at first order in 8, noncommutative SO(10) gauge theory is unique.
G
2.2. Noncommutative SU (5) The fermionic sector of S U ( 5 ) has the $Pr,multiplet that transforms in the 3 of S U ( 5 ) and the X L multiplet that transforms according to the 10 of S U ( 5 ) . In this case we expect that the adjoint, the 3 and the 10 representations enter in (7). In principle one can consider the coefficients
147
q # c10, i.e. while the (qCL, X L ) fermion rep. is 5 @ 10, in (7) the weights cp of the 5 and the 10 can possibly be not the same. It turns out that cPDzbc# 0 in (8). We see that, already at first only if q # c10 then order in the noncommutativity parameter 0, noncommutative S U ( 5 ) gauge
cp
theory is not uniquely determined by the gauge coupling constant g, but also by the value of C , cpDzbc. Thus S U ( 5 ) is not a truly unified theory in a noncommutative setting. It is tempting t o set q = c1O so that exactly the fermion representation 10 enters (8). We then have C , C,,D,”~“ = 0, (however this relation is not protected by symmetries).
s@
2.3. (GUT Inspired) Noncommutative Standard Model One proceeds similarly for the SM gauge group. The full ambiguity of the gauge kinetic term is given in Ref. 6. About the fermion kinetic term, the fermion vector Q L is constructed from Q L = (ui, di , -up , d p , v,e- , e f ) L . The covariant derivative is as in Eq. (2), with Q + Q L and with A, = A $ T A , where { T A } = {Y,Tf,TA}are the generators of U(1) 8 S U ( 2 ) @ S U ( 3 ) . The fermion kinetic term is then as in Eq. (1). This Standard Model is built using only left handed fermions and antifermions. We call it GUT inspired because its noncommutative structure can be embedded in SO(10) GUT. Indeed Q L and differ just by the extra neutrino v% = -ia2 v ~ ;* moreover under an infinitesimal gauge transformation all fermions in Q L transform with on the left. This GUT inspired Standard Model differs from the one considered in Ref. 1; indeed here we started from the chiral vector Q L , while there the vector 9’= ( u i , d i ,u k d k , v ~ , e LeR) , is considered. In the commutative caSe JV@W= JE@QL, but in the h
Qi
x,
h
noncommutative case (see later) this is no more true:
J QI
- A
#J -
-
*
@ QL ; if we change 0 into -0 in the right handed sector of J @ * @ @ I , -
- A
A
then the two expressions coincide. Finally it is a natural choice t o consider in the SM gauge kinetic term only the adjoint rep. and the fermion rep., we then have that a t first order in 0 there are no modifications t o the SM gauge kinetic term. This is so because the fermion rep. is anomaly free: DAA’A” = 0, and because for Pferrnlon U(1) the adjoint rep. is trivial. 2.4. Higgs Sectors
While the noncommutative Higgs kinetic and potential terms are given by
(E,@ * E,? + p2J-t * 4-A & + * &*$t * 3,
(11)
148
a noncommutative version of the SM and GUT Yukawa terms is not straigh-
forward and requires the introduction of the hybrid Seiberg-Witten maps on fermions. A typical noncommutative Yukawa term then reads
--H
Jt * L L * e;l + -31-
where L, =
(z)
herrn. wnj. ,
. Under an infinitesimal U(1) 8 SU(2) @I SU(3) gauge
4
-31
-
4
-
3
1
transformation A, L, transforms as 6 L, = ++(A) * L, - ZL, *pe;,(A). We see that in the hybrid SW map appears both on the left and on the right of the fermions, moreover the representation of is inherited from -31 the Higgs and fermions that sandwich L , . The Yukawa term (12) is thus invariant under noncommutative gauge transformations. Of course in the 9 -+ 0 limit we recover the usual gauge transformation for the leptons. An explicit formula for the hybrid SW map at first order in 9 is in Refs. 1, 6. The Yukawa terms (12) differ from those studied in Ref. 1. There the
A
A i
hybrid SW map is considered on 4, in particular there q!~ is not invariant under SU(3) gauge transformations (and this implies that in Ref. 1 gluons couple directly to the Higgs field). The Higgs sector in the SO(10) and S U(5 ) models can be constructed with similar techniques.6
3. Hermiticity and Reality Properties of SW Map From Eq. (4) we see that if A is hermitian then A^ is also hermitian. Actually, to all orders in 9, and can be chosen hermitian if A and A are hermitian. Otherwise stated, SW map can be chosen to be compatible with hermitian conjugation. Compatibility of SW map with complex conjugation reads, -* a*=*, (13)
A^
where = SW[@,p,p(A),O]denotes the SW map of @ constructed with the representation p,p of the potential A , and the SW map of the complex conjugate spinor @* is defined by h
@*
(14)
SW[@*,p,p* ( A ) ,-6'1,
where p,p. is the representation conjugate to ~ , p Notice . ~ that in Eq. (14) the noncommutativity parameter 0 appears with opposite sign with respect aGiven the group element g =e i A = eiAaT" we h a v E ( g ) E are real, we have pq* (A) = -pp(A) , p q - (A)= -pq(A) .
a
and, since Aa, A"
149
-
-
to the 8 in the SW map of 9.Similarly to (13) we have pa* ( A ) = -p*(A)
:nd p * * ( R ) = - p ~ r ( R ) where p w ( A ) = A^[pq8(A),-8]and p*.(h) = A [ ~ Q * ( A ) , ~ ~ * ( RThe ) , - proof ~ ] . of Eq. (13) relies on showing that the SW differential equations3 (obtained by requiring that gauge equivalence classes of the gauge theory with noncommutativity 8 68, correspond to gauge equivalence classes of the gauge theory with noncommutativity 8 ) are themselves compatible with complex conjugation.6 One proceeds similarly for the case of hermitian conjugation.
+
3.1. Noncommutativity and Chirality We can now discuss a further ambiguity of noncommutative gauge theories, and resolve it by requiring compatibility with grand unified theories. For simplicity we consider noncommutative QED. Let ?I, be a 4-component Dirac spinor, and decompose it into its Weil spinors $L and $R. Their charge conjugate spinors are $Lc = $% = -iu2$; and $Rc = $% = iaz$i. Consider the noncommutative left-handed spinor $L = SW[$L,p+L ( A ) ,81, we then have the k8 choice A
'& = sw[$R,P+'R ( A ) *'9] 1
(15)
1
for the right handed one. In the literature, the choice +8 is usually considered so that for the 4-component Dirac spinor $ we can write = SW[$, A , 81, d$ = ix * We here advocate the opposite choice
6
6.
h
(-8) in Eq. (15). Indeed from Eq. (13) we have that therefore
GLC = -202
-*
$R
h
so that with the -8 choice in Eq. (15), both left handed fermions $ L , h
-
and
-
qLC
are associated with 8 , while the right handed ones $ R , GRC are associated with -8. In GUT theories we have multiplets of definite chirality and therefore this is the natural choice to consider in this setting. These observations allow us to compare QED+ with QED-, the two different QED theories obtained with the two different f 0 choices (15). This difference immediately extends to the fermion kinetic terms of nonabelian gauge theories and allows us to compare the NCSM discussed in Ref. 1 with the present GUT compatible one. We have (up to gauge kinetic terms)
150
where the GUT inspired QED- is obtained using the left handed spinor h
(:c)
so that
$f = sW[$f,p+f(A),8].Now -t
A
from
h
n *
$Lc
=
-Pt
- '
2 ~ 2 1 4 and ~
- P e p
from u matrix algebra we have S $Lc *i@qLC = J $R i@ $R , where we have emphasized that we are using the -8 convention in the SW map by writing w p instead of - . We conclude that in order to obtain QEDfrom QED+ we just need to change 8 into -8 in the right handed fermion sector of QED+. *OP
4. C , P, T Properties of SW Map and of NCYM Actions
Using compatibility of SW map with complex conjugation and the tensorial properties of SW map (i.e. that SW map preserves the space-time index) one can study the properties of SW map with respect to the C, P and T operations. In particular these same expressions as in the commutative case holds:
h*
-CP
G c p = i a 2 @ L , QR
-*
-CP
=-iazQ~, A,
-A
-
=(-Ao,Ai),
(18)
where the action of the P and C operators on spinors is given by
-T
while the time inversion is given by QL = SW[QT,p q L ~ ( A TB T) ,, d T , -i] . In these expressions we have written explicitly the dependence on the partial derivatives, and the imaginary unit i in the last slot marks that the coefficients in the SW map are in general complex coefficients. The -i in the last expression means that we are considering the complex conjugates of the coefficients in the SW map, this is so because T is antilinear and multiplicative. Relations (17) and (18) hold provided that Owv transforms under C, P, T as a U(1) field strenght FPy.If we choose +8 in Eq. (15) then parity and charge conjugation sepatately assume the same expression as in the commutative case. Now we discuss the transformations properties of NCYM actions under C, P and T . With the +8 choice (15) we have that NCYM actions are invariant under C, P and T iff in the commutative limit they are invariant. On the other hand, with the -8 choice NCYM actions are invariant under CP and T iff in the commutative limit they are invariant. For the fermion kinetic term these statements are a straighforward consequence of
s QL-t
*
151
@ g= JQL-t
-
@QL. Since ? transforms like F under CP and T , and in the +8 case also under C and P separately, the C, P,T properties of the gauge kinetic term Ti-(@*@) = Tr(@F)easily follow. Inspection of the fermion gauge bosons interaction term leads also to the same conclusion. We have studied the C, P and T symmetry properties of NCYM actions where 6' transforms under C, P and T as a field strength. Viceversa, if we keep 8 fixed under C, P and T transformations, we in general have that NCYM theories break C, P and T symmetries. Finally, U(1) field strength is invariant under the combined CPT transformation, and therefore 8 does not change. This implies that CPT is always a symmetry of NCYM actions.
s
Acknowledgments
It is a pleasure to thank the organizers for the nice and stimulating atmosphere at the conference and its efficient organization. References 1. X. Calmet, B. Jurco, P. Schupp, J. Wess and M. Wohlgenannt, Eur. Phys. J. C23,363 (2002), hep-ph/Olllll5. 2. M. Dimitrijevic, F. Meyer, L. Moller and J. Wess, Gauge theon'es on the kappa-Minkowski space time, hep- th/03 10116. 3. N. Seiberg and E. Witten, JHEP 9909,032 (1999), hep-th/9908142. 4. J. Madore, S. Schraml, P. Schupp and J. Wess, Eur. Phys. J . C16, 161 (2000), h e p th/0001203; B. Jurco, S. Schraml, P. Schupp and J. Wess, Eur. Phys. J. C17,521 (2000), hepth/0006246. 5. B. Jurco, L. Moller, S. Schraml, P. Schupp and J. Wess, Eur. Phys. J. C21, 383 (2001), hepth/0104153. 6. P. Aschieri, B. Jurco, P. Schupp and J. Wess, Nucl. Phys. B651,45 (2003), hepth/0205214. 7. P. Schupp, J. Trampetic, J. Wess and G. Raffelt, The photon neutrino interaction in non-commutative gauge field theory and astrophysical bounds, hepph/02 12292. 8. M. Buric and V. Radovanovic, JHEP 0402, 040 (2004), hep-th/0401103. 9. F. Brandt, C. P. Martin and F. R. Ruiz, JHEP 0307, 068 (2003), hepth/0307292. 10. M. M. Sheikh-Jabbari, Phys. Rev. Lett. 84, 5265 (2000), hep-th/0001167. 11. M. Chaichian, K. Nishijima and A. Tureanu, Phys. Lett. B568, 146 (2003), hep-th/0209008 ; L. Alvarez-Gaume and M. A. Vazquez-Mozo, Nucl. Phys. B668,293 (2003), hepth/0305093; N. Mahajan, Phys. Lett. B569,85 (2003), hep-th/0305105.
SEESAW, SUSY AND SO(10)
BORUT BAJC JoEef Stefan Institute Jamova 39 1001 Ljubljana, Slovenia E-mail: borut. [email protected]
The seesaw mechanism can explain why the neutrino masses are so tiny with respect to the charged fermion masses. In the canonical version it is the presence of the right-handed neutrino that is responsible for it. In renormalizable grand unified theories with left-right gauge symmetry it is possible to show quite generically that there is another type of seesaw contribution, mediated by a heavy weak triplet. In this talk I will show that such non-canonical seesaw mechanism can very nicely connect b - T Yukawa unification with the large atmospheric neutrino mixing angle in the context of a SO(10) grand unified theory. Also, a fit to the available low energy masses and mixings points towards the domination of this non-canonical contribution with respect t o the canonical one. Finally I will explicitly present the minimal supersymmetric SO(10) model which has all the above nice features.
1. Introduction
The charged and neutral fermion sectors in the Standard Model (SM) have quite different structures: 1) though very different among themselves, the charged fermions have much larger masses than neutrinos, 2) the mixing angles in the quark sector tend to be much smaller than the corresponding ones in the leptonic (neutrino) sector. While the first point can be easily accounted with the famous seesaw mechanism,l the second issue is more controversial and its solution still debated. The first nontrivial framework that could be predictive in both charged fermion and neutrino sectors is SO(10) grandunification. This is because it automatically incorporates the righthanded neutrino, thus providing for a theory of the seesaw mechanism. Most important, the model connects the neutrino and charged fermion mass matrices, providing some nontrivial relations between them at the grandunification scale. Then, with the assumption of the desert, motivated by gauge coupling unification, one can extrapolate the values of these masses and mixings down to low energies 152
153
and here compare them with experimental data. I will describe in some details the origin and form of the neutrino mass matrix, and connect it to charged fermion masses in the context of a SO(10) grand unified theory. 2. Seesaw
The left-handed neutrino u (EUL) is a component of a weak SU(2)L doublet with B - L = -1. If left-right symmetry is assumed, than a right-handed neutrino uc (s Cv,') must exist as part of a doublet of S u ( 2 ) ~gauge symmetry, with B - L = +l. Since uc is a standard model singlet, its mass is given by the scale of S u ( 2 ) ~breaking, which can be achieved in a renorrnalizable version by the vev of a S u ( 2 ) ~ triplet (with B - L = -2), MvR O( (AR). This, Majorana mass, must then be added to the Dirac mass MUDoc (a) with a a su(2)Lxsu(2)R bidoublet:
+
Lm = - y c T M v R ~ ' U ~ ~ M , + , Uh.c. . After integrating out the heavy uc one gets the famous seesawl formula for the light neutrino mass:
Mu = -MTDM;: Mv, ,
(2) which clearly connects the smallness of the neutrino mass with the largeness of the S u ( 2 ) ~breaking scale. There is however another contribution to the seesaw,2 several times neglected, but as we will see, of great potential importance. It comes essentially from the following arguments: as we said, the large right-handed Majorana mass comes from the vev of a S u ( 2 ) ~ triplet via the term u C T A ~ v C , so due to left-right symmetry an analog term uTALu must exist, with AL a triplet under S u ( 2 ) ~ .With these two interaction terms plus the usual Dirac term uCT@uone can easily show that the one-loop box diagram contribution to A R @ ~ AisLUV divergent, so that such a term must be present already at tree level. Thus, the potential for the triplets looks like
+
V = - M 2 (A; +A:) ARGAL, (3) with M a large mass (not much less than MGUT).The last term represents a tadpole for the lefthanded triplet:
154
since (AR) M M and (@) M M w . So the term
"T(wu gives another contribution to the seesaw. All together we thus have
+
MN = -MTD M&l MvD My,, where the first term represents the type I or canonical seesaw formula mediated by the S u ( 2 ) singlet ~ u C ,while the second term is the S u ( 2 ) triplet ~ contribution t o the type I1 or non-canonical see-saw formula. Of course, in general, the matrices in generation space M,,, M,, and MVD are arbitrary, so the above formula is not very useful. It is thus important t o connect the above matrices to the charged fermion sector, which is experimentally better known. For this one needs a framework. We will choose the most economical one, i.e. a SO(10) grand unified theory.
3. SO(10) Although GUTS are not theories of flavours, they still put some constraints on the possible Yukawa interactions, since different SM fields live in the same representations. In SO(10) all the light fermions of each generation plus the right-handed neutrino are grouped together in the 16-dimensional spinorial representation, while in the most economical version the two light Higgs doublets live in a fundamental 10-dimensional complex Higgs representation. There is thus only one Yukawa matrix in generation space (3 x 3). An O(3) rotation of the 16's in generation space can diagonalize it, giving one good prediction (yb = yT = yt) and several bad ones (equality of charged lepton, up and down quark Yukawas of the second and first generations, and no mixing at all). A nontrivial quark mixing could be achieved adding a new 10-dimensional Higgs, but that would not help in improving the relations among Yukawas of the first two generations. To see it, one can remember that the 10-dimensional Higgs gets a vev in the ( 2 , 2 , 1 ) direction of the Pati-Salam S U ( ~ ) L X S U ( ~ ) R X Ssubgroup. U ( ~ ) ~ This means that such a vev cannot break SU(4)c and thus the equality between leptons and quarks. Thus, a nontrivial multiplet in SU(4)c (and bidoublet in the left-right sector) is needed to get such a splitting. An ideal possibility is given by the Higgs in the representation 126 (5 index antisymmetric and anti-self-dual) of SO(10): its vev in the SU(3)c colour singlet state of
155
the (2,2,15) direction is a possible candidate. The Yukawa terms in the Lagrangian can thus be written as3
These terms can be decomposed in the usual Pati-Salam basis as
+ (1,2,z), 126~ = (1,3,lO) + ( 3 , 1 , m )+ (2,2,15)+ (1,1,6), 16 = (2,1,4)
(9)
(10)
where the fields in boldface have SU(3)cxU(l),, singlets and can thus generate a nonzero vev. The two colour singlet bidoublets are actually needed to develop a vev in order t o fit the light fermion masses, as we have just seen. On top of that, the S u ( 2 ) triplet ~ (1,3,10) above is ideally suited to give a large Majorana mass to the heavy right-handed neutrino v c from (2,1,z): from here we can see the double role that the 126-dimensional Higgs can play. Finally, the S u ( 2 ) ~triplet in (3,1,m), if nonzero as conjectured in the previous section, can contribute to the type I1 see-saw formula. Denoting uydd and u2";dG the vevs from the bidoublets in 1 0 and ~ 1268 and WL,R the vevs from the L or R triplets, one gets from (7) the following expressions for the fermion masses:
+ uY26y126 M D = W t ~ ) y l O + v&vjy126 MU = u;bylO
Mv,
= ~yoY10- 3vy2,y126
ME = V f o y l O
7
(11)
7
(12) (13)
7
- 3v&?6y126
(14)
1
MVR= ((1,31 10)m)Y126 = WRY126 MVL= ((37 1,m)m)y126 = VLy126 . 7
(15)
( 16)
Notice that the factor -3 from the 126~ contribution in the lepton sector with respect to the quark one comes because the SU(3)c colour singlet direction in 15 of SU(4)c is proportional to B - L c( diug(l,l, 1,-3). Also, the above relations are valid at the GUT scale only, so that the known experimental values of the masses and mixings must first be run up to that energy using the renormalization group equations. Expressions (11) and (12) are needed to evaluate the matrices Y10 and Y126 in terms of the better known Mu and MO. Their expressions are then
156
used in (14) as well as in the various neutrino matrices (13), (15) and (16) to be used in (6). Defining
one gets two matrix equations:
4. Type I Versus Type I1 Seesaw What we want to find out is which type of seesaw is compatible with data. To make the analysis simpler, let us study the situation of the 2nd and 3rd generations only, as well as no CP violation (real parameter^).^ Equation (19) has 6 known masses ( n ~ mt,c, ~ , ~m b,, + ) , one known mixing angle (0, = O c b ) , but 3 unknown parameters, x,y and the angle between the orthogonal matrices that diagonalize ME and M D (call it 6,). Since the matrices are all symmetric (due t o its SO(10) origin), we have 3 equations, just enough. After determining these parameters one can attack equation (19). We are interested mainly in the leptonic mixing angle and this is determined as a function of one single parameter, a. In the limit a -+ 00 one remains with a pure type I seesaw, while in the opposite case a -+ 0 and the seesaw is of type 11. To get a feeling, let us first consider the idealized situation of small second generation masses (0 M m2 << ms), but still finite quark mixing 0, (although << 1). It is easy to show that the leptonic atmospheric mixing angle is given by
tan201 = with
sin 20, 2sin2 8, - A '
157
1 mb - mr [-5a (1 - 4a) €1, E = 1-9a m b In the limit a -+ 00 one gets A = (5 4 ~ ) / 9 ,while the opposite case a + 0 gives A = E . Since the parameter E is experimentally small (approximate b - T unification), the small a regime is the one that predicts a large atmospheric angle: type I1 seesaw is thus favoured in these type of models. Now we restore a finite mass of the second generation, but keep in mind that they are small, i.e. that
+
A=---
+
There are two solutions: lStsolution (small OD): writing in a schematic way (coefficients of order 1 are not explicitly written) the neutrino mass matrix is 6 6/E hdN = -a! (6,€ l,€)
+
(: :)
where the first (second) matrix is the type I (11) seesaw contribution. The atmospheric mixing angle is tan281 x
If we further assume that
E
+
6 (1 a / € ) 6(1 + a ) E a/€. M O(6), we get
81 x O(1)
+ +
* a 5 O(62) .
(25)
The dominant type I seesaw ( a ---t m) is clearly excluded, since it would predict a small mixing angle 81 M O(6). Notice that without b--7 unification (small E ) 81 would be small. Znd solution (large OD): here the relevant matrices are 62
M N o - f f ( 6 6, 1
+
(; i)
and the atmospheric mixing angle tan281 M
1+Cd l+a
158
so that now a large angle is easily obtained:
We see that both type I1 or the mixed seesaw are compatible with data, but in no case can the type I ( a -+ m) seesaw dominate and predict a large atmospheric mixing angle. In this case the b - r unification parameter E needs not be small. It is possible to understand why a large atmospheric mixing angle emerges in type I1 ~ e e s a w :the ~ point is that the triplet contribution to the Majorana neutrino mass is proportional t o Y l 2 6 (16), but from (12) and (14) this Yukawa is proportional to the difference M D - M E , so that in pure type I1 seesaw
Assuming small mixings, one immediately obtains approximately
so that an unexpected connection between large atmospheric neutrino angle and b - r unification emerges from type I1 seesaw and the assumption of small large
Batm
u b-r
unification.
(31)
Notice however, that in the large OD case b - r unification is not really needed to get a large atmospheric mixing angle. 5 . SUSY
Up to now we did not specify which is the grandunification theory we were considering: the only requirements were that the Yukawas came from the interactions of matter 16-dimensional representations with two complex Higgses - 1 0 and ~ 126~, and that they got some nonzero vevs in the relevant directions. In order to show that this is possible and to find these vevs as functions of parameters in the Lagrangian, we need to write down an explicit model. I will here shortly describe the minimal supersymmetric
159
SO(10) model, which has all the nice above features and is still consistent with data. First of all, why supersymmetry? There are different reasons to believe that supersymmetry is present in some way or another. The first is of course the (in)famous hierarchy problem, i.e. why are scales in nature so different (for example MW and MGUT),since it is known that quantum corrections tend to destabilize them. One of the few possibilities is to consider supersymmetric extensions, in which such quantum corrections are miracolously stable. The second reason is the origin of the spontaneous breaking of electroweak symmetry: although all scalars could have a positive mass square at a very high scale (Mplanck ?), the coupling of the SM Higgs to the large top Yukawa makes the Higgs mass tachyonic, thus triggering its vev.6 Finally, the presence of the supersymmetric partners in the TeV region modifies the beta functions of the SM gauge couplings exactly in the right way to get (under the assumption of desert) unification at the scale of approximately 10l6 GeV.7 As we saw, in order to fit the fermion masses we need two Higgs representations, 1 0 and ~ 126~. In supersymmetry, the presence of 126 is particularly welcome, since its vev does not break R-parity. In fact (1,3, lo), which vev gives a large mass to the right-handed neutrino, has B - L = -2. R-parity is given by
R = (-1)3(B-L)+2S
(32)
and since the spin S of any vev is 0, the vc mass is R-parity even and so does not break it at the large scale.8 One can show that this is true all the way to the electroweak scale, i.e. R-parity is exact.g This means, among other things that the lightest supersymmetric partner is stable and is thus a good candidate for dark matter. Now, in supersymmetry we need more than just 1 0 and ~ 126~, since the right-handed neutrino mass must be very large, on the order of the GUT scale or so. This would strongly break supersymmetry by the D-terms, so another Higgs - 1 2 6 (5 ~ index antisymmetric and self-dual) - must be used to cancel it. SO(10) symmetry allow the matter 16 to be coupled in the superpotential only to 10,126 and 120 (3 index antisymmetric) dimensional representations, so this new 1 2 6 cannot ~ change (7). Finally, we need to get two (practically) massless Higgs doublets, while heavy all colour triplets. This can be achieved by first introducing the Higgs representation 2 1 0 ~(4 index antisymmetric) and then by one fine-tuning of the parameters in the superpotential.
160
Such a renormalizable theory, with three matter 16, and with the Higgs and 2 1 0 ~(see Ref. 10) has the correct sector made of 1 0 ~ 1, 2 6 ~1268 , symmetry breaking pattern'' and has been shown to be the minimal grand unified theory,12 i.e. the most predictive one. 6. Further Developments The easy example with two generations and real parameters is instructive, and it can be generalized for a more realistic three generations set-up. It has to be said however, that one should not take the results in this case too seriously, since various uncontrolled corrections (like for example soft susy breaking terms) can in principle take place for the tiny first generation masses. The three generation generali~ation'~ roughly confirms the consistency of the type I1 seesaw with data, and predicts a quite large Ue3 x 0.15. Type I see-saw domination seems however still possible, especially after CP violating phases are considered.l4 An interesting and necessary check of the model is to calculate the proton decat rates of d = 5 operators. A preliminary though incomplete study15 shows that these decays could easily be too fast and thus rule out the minimal model. The question is of course of great importance, but a detailed analyses requires the knowledge of the SO(10) Clebsch-Gordan coefficients, which are now available.16 Another issue is the gauge coupling unification: recently it has been shown, that at least in some of the relevant regions of parameter space the threshold corrections are under control in spite of the large representations involved.17
Acknowledgments
It is a pleasure to thank the organizers of the conference for the stimulating environment and especially Goran Djordjevib for his excellent work. I thank my friends Charan Aulakh, Alejandra Melfo, Goran Senjanovib, and Francesco Vissani for collaboration. This work is supported by the Ministry of Science, Sport and Education of the Republic of Slovenia.
References 1. T. Yanagida, Proceedings of the Workshop o n Unified Theories and Baryon Number in the Universe, Tsukuba, Japan (1979) (edited by A. Sawada and
161
2.
3. 4. 5. 6.
7.
8.
9.
10. 11.
12. 13. 14.
15. 16.
17.
A. Sugamoto, KEK Report No. 79-18, Tsukuba); S. L. Glashow, Quarks and Leptons, Cargkse (1979), eds. M. LBvy, et al., (Plenum 1980 New York); M. Gell-Mann, P. Ramond and R. Slansky, Proceedings of the Supergravity Stony Brook Workshop, New York, (1979), eds. P. Van Niewenhuizen and D. Freeman (North-Holland, Amsterdam); R. N. Mohapatra and G. Senjanovit, Phys. Rev. Lett. 44, 912 (1980). G. Lazarides, Q. Shafi and C. Wetterich, Nucl. Phys. B181, 287 (1981); R. N. Mohapatra and G. Senjanovit, Phys. Rev. D23, 165 (1981). K. S. Babu and R. N. Mohapatra, Phys. Rev. Lett. 70, 2845 (1993). B. Bajc, G. Senjanovit and F. Vissani, hep-ph/0402140. B. Bajc, G. Senjanovik and F. Vissani, Phys. Rev. Lett. 90, 051802 (2003); B. Bajc, G. Senjanovit and F. Vissani, hep-ph/0110310. K. Inoue, A. Kakuto, H. Komatsu and S. Takeshita, Prog. Theor. Phys. 68, 927 (1982), [Erratum-ibid. 70 (1983) 3301; L. Alvarez-Gaume, J. Polchinski and M. B. Wise, Nucl. Phys. B221, 495 (1983). S. Dimopoulos, S. Raby and F. Wilczek, Phys. Rev. D24, 1618 (1981); L. E. Ibanez and G. G. RQSS, Phys. Lett. B105, 439 (1981); M. B. Einhorn and D. R. T. Jones, Nucl. Phys. B196,475 (1982); W. J. Marciano and G. Senjanovit, Phys. Rev. D25, 3092 (1982). D. G. Lee and R. N. Mohapatra, Phys. Rev. D51, 1353 (1995); C. S. Aulakh, B. Bajc, A. Melfo, A. RGin and G. Senjanovit, Nucl. Phys. B597, 89 (2001). C. S. Aulakh, K. Benakli and G. Senjanovic, Phys. Rev. Lett. 79,2188 (1997); C. S. Aulakh, A. Melfo and G. Senjanovik, Phys. Rev. D57, 4174 (1998); C. S. Aulakh, A. Melfo, A. R S i n and G. Senjanovik, Phys. Lett. B459, 557 (1999). C. S. Aulakh and R. N. Mohapatra, Phys. Rev. D28,217 (1983); T. E. Clark, T. K. Kuo and N. Nakagawa, Phys. Lett. B115, 26 (1982). X. G. He and S. Meljanac, Phys. Rev. D40, 2098 (1989); D. G. Lee, Phys. Rev. D49, 1417 (1994). C. S. Aulakh, B. Bajc, A. Melfo, G. Senjanovit and F. Vissani, Phys. Lett. B588, 196 (2004). H. S. Goh, R. N. Mohapatra and S. P. Ng, Phys. Lett. 33570, 215 (2003), and Phys. Rev. D68, 115008 (2003). K. Matsuda, Y. Koide, T. Fukuyama and H. Nishiura, Phys. Rev. D65, 033008 (2002) [Erratum-ibid. D 65 (2002) 0799041; T. Fukuyama and N. Okada, JHEP 0211, 011 (2002); B. Dutta, Y. Mimura and R. N. Mohapatra, hepph/0402113. H. S. Goh, R. N. Mohapatra, S. Nasri and S. P. Ng, Phys. Lett. B587, 105 (2004). C. S. Aulakh and A. Girdhar, hep-ph/0204097; T. Fukuyama, A. Ilakovac, T. Kikuchi, S. Meljanac and N. Okada, hep-ph/0401213; B. Bajc, A. Melfo, G. Senjanovik and F. Vissani, hepph/0402122. T. F'ukuyama, A. Ilakovac, T. Kikuchi, S. Meljanac and N. Okada, hep-ph/0405300. C. S. Aulakh and A. Girdhar, hep-ph/0405074.
ON THE DYNAMICS OF BMN OPERATORS OF FINITE SIZE AND THE MODEL OF STRING BITS
S. BELLUCCI AND C. SOCHICHIU* INFN - Laboratori Nazionali d i Frascati, Via E. Fermi 40, 00044 Frascati, Italy E-mail: [email protected]
We consider the discretization effects of a string bit model simulating the nearBMN operators in the super-Yang-Mills model. The fermionic sector of this model is altered by the so called species doubling. We analyze the possibilities to cure this disease and propose an alternative formulation of the fermionic sector free from the above drawbacks. Also we propose a formulation of string bits with exact supersymmetry, which produces however an even number of continuous strings in the limit J -+ m.
1. Introduction AdS/CFT correspondence is originally formulated1y2 as a duality relation between string theory on Ads5 x S5 space and conformal theory of N = 4 Super Yang-Mills (SYM) on four dimensional Minkowski space, which is the boundary of Ads (see Ref. 3 for a review). Being a true duality, AdS/CFT correspondence relates the weakly coupled SYM regime to the strongly coupled string theory and vice versa. This makes it a strong predictive tool, however, a difficult one to test, since there are no means to probe either SYM or string theory at strong coupling. Beyond this, the solution of string theory on Ads background is not known (although some progress towards this was recently achieved, see Ref. 4 and references therein). A few years ago, Berenstein-Malda~ena-Nastase~-~ proposed to consider the particular limit of Ads5 x S5 geometry known as pp-wave back*On leave from: Bogoliubov Lab. Theor. Phys., JINR, 141980 Dubna, Moscow Reg., RUSSIA and Institutul de Fizics Aplicatg AS, str. Academiei, nr. 5, Chiginh, MD2028 MOLDOVA. 162
163
ground, on which the string theory is s o l ~ a b l e This . ~ ~ ~situation of the overlap of the applicability of both string theory and SYM was further extended t o spinning string backgrounds10-12 (see also the contribution of G. Arutyunov to the present proceedings). In the absence of a more adequate tool to describe string theory near the BMN limit it was proposed t o use the string bit mode1.13-15 This model consists of the set of point-like interacting particles - string bits which in the limit of the number of particles going t o infinity is expected t o produce the continuous pp-wave string. Later on, however, the exact description in terms of spin chains was found in the planar limit of SYM.16>17 Nevertheless, string bits remained an efficient tool t o take care of nonplanarity which correspond, via AdS/CFT, to string production. Although the string bit model was quite useful in describing the dynamics of the bosonic sector in the near BMN limit, it suffers from internal inconsistencies due t o the fermionic spectrum doubling. In the theory of discrete fermions there is a well-known no-go theorem due to Nielsen-Ninomiya" which limits the possibilities of having discrete fermions. Therefore, one should make sure that these limits do not interfere with the AdS/CFT correspondence. For this purpose, one needs to prove the existence of a discrete supersymmetric model which in the continuum model reduces to the pp-wave string. This is the aim of the present note. It is based mainly on the Refs. 19, 20 where this problem was approached (for a staggered fermion approach see also Ref. 21).
2. Fermionic Doubling After fixing the permutation symmetry, the one-string sector of the string bit model is given by the following Hamiltonian:
with commutation relations
where i = 1,. . . , 8 are vector and, respectively, a = 1,. . . ,8 spinor indices of SO(8) appearing in the light-cone quantization of the pp-wave string. II
164
is the matrix in SO(8) spinor space given in terms of 16 x 16 dimensional y-matrices in chiral representation by
n2=1.
r I = y1-2 y y3-4 y,
(3)
The Hamiltonian (l), together with the shift operator
where
afn
= (l/a)(fn+i - f n ) are generated by the “supercharges” {&a, Qb}
= % b ( H -k P ) ,
where
Q =a
{ Q a , Qb} = 2 d a b ( H - P ) ,
C [piyien - ~ L ~ T I B+, arcnyien1 ,
(5)
(64
n
In spite of the fact that the supercharges generate according to Eq. (5) both the Hamiltonian and the shift operator, the supersymmetry fails because the supercharges do not commute
{&,&I# 0.
(7)
The fermionic doubling is better seen if we pass to the Fourier mode description
where fn stands for zn, p,, 8, and fin while f k represents their Fourier modes. In terms of Fourier modes the fermionic part of the Hamiltonian has the form
where the doubling is visible due to additional zeroes of the sin function in Eq. (9). Due to the fact that modes around these additional zeroes have low energy, they survive and contribute in the continuum limit, though they correspond to fast oscillating fields. In contrast, in the bosonic part one has a factor with the cosine function which becomes large in the continuum limit and one has no doubling there. Therefore, contrary to what can be expected, supersymmetry is not restored for a + 0.
165
In Ref. 21 it was proposed to use a staggered fermion approach in order to cure the fermion doubling. It consists in placing only a half number of fermions, e.g. in odd points ( n = 2k 1) one has only 8, while in even ( n = 2 k ) there are only 8,. Due to doubling the number of fermions in the continuum limit is the correct one. The supersymmetry is also restored in the continuum limit. However, at any finite lattice size it is still broken.
+
3. Almost supersymmetry
In fact, it appears possible to make the theory supersymmetric with the exception of a single mode by properly choosing the lattice derivative function. If this mode is placed in the high energy region of the model one may expect that it does not contribute in the continuum limit. On the lattice one can have various definitions of the derivative of a field. The only criterion is that the lattice derivative should give the usual derivative in the continuum limit. In general the lattice derivative could be written as the operator m
an,
where = condition that
8,
are the elements of the derivative matrix describes a lattice derivative reads
a.
The
Earn= 0, ESmm= 1, m
m
and
In terms of the Fourier mode expansion this means that the Fourier transshould have the limit form
&
&la+)
-+ ipk = 27rik.
Substituting the forward lattice derivatives d in Eq. (6) by a generic derivative 8 and requiring that we get the same bosonic part of the Hamiltonian, we find the condition of restoration of supersymmetry to be
S = i(a*a)f.
(13)
The condition (13) is trivially satisfied in the continuum limit, since (in Fourier mode) dk = -ipk is an anti-Hermitian operator. In contrast,
166
on the lattice 6' has both Hermitian and anti-Hermitian p a r h a Therefore, extracting the square root in Eq. (13) is a nontrivial procedure also because of its ambiguity. In terms of Fourier modes equation (13) is reduced to the extraction of the square root for each mode
where f k = f l is the sign, which can be chosen separately for each mode. The natural choice f k = sgn k yields a jump in the derivative function at k = f J / 2 . Also, at this point, the solution (14) fails to satisfy equation (13). The jump of the derivative function is related t o the nonlocality of the fermionic derivative
an,
2
= -(-l)n-m a
n(n-m) cos( 5 )sin( 7)
cos(
5 ) - cos(q).
Although Eq. (15) looks nonlocal and ugly, one can checkb that its square is a local operator given by
which is the next-to-neighbor second derivative. Let us see that the ambiguity of the momentum function does not affect physical observables like the energy. 4. Fermion Doubling/Nondoubling in the BMN
Correspondence Earlier we argued that having a well-defined supersymmetric discrete model is indispensable for confirming the self-consistence of the BMN limit in the fermionic sector. Now we want to define precisely the quantities to be compared. In the mode expansion of the pp-wave string described by Met~aev'?~ one has the energy of each mode given by the square root
aIn general, the Hermitian part of the derivative is the one responsible for fermionic doubling. bThe simplest way to do this is to use the Fourier transform.
167
where X is the 't Hooft coupling (for simplicity, other parameters here are put to unity). On the other hand, the Yang-Mills contribution at k loops yields a quantity of order X 2 k which corresponds to the k-th term in the expansion of the square root ( 1 7 ) . The planar Yang-Mills contribution up to three loops nowadays is well known and is described by integrable spin chains. 16,1722 N
In the string bit model the argument of the square root is replaced by 19,20
zn
N
41 + X 2 d k 2 .
(18)
Let us consider, for definiteness, the contribution of order X2 which corresponds to the Yang-Mills one-loop approximation. Then, the energy of the string bit mode in this approximation is given by 1-loop wk
N
A282 = P ( a * a ) k ,
(19)
where 8 is the usual bosonic forward lattice derivative afn = (l/a)(f,+l f n ) , while a * is the backward one a*f, = ( l / u ) ( f n- fn-l). The one loop energy (19) corresponds to the following term in the fermionic Hamiltonian: n
n
n
a2
where we integrated by parts after using Eq. (16), in order to express in terms of next neighbor derivatives a and a * . Let us note that the fields entering in Eq. (20) are inverse Fourier transforms of Hamiltonian eigenmodes rather than the original fermionic fields On and &. The transformation which relates them is a singular one, what makes it possible for the Hamiltonian to be ill-defined in terms of On and &, although in terms of there is no such problem. In particular, this means that the eigenvalues of the fermionic Hamiltonian are well-defined and compatible with supersymmetry in spite of the discontinuity in Eq. (14).
$A*)
$i*)
5. Doubling and Supersymmetry
In the last section we have shown that one can consistently define the supersymmetric model with the right spectrum, provided one does not insist
168
on the worldsheet ferrnionic structure. Let us show now that one can have more. Namely, let us construct an exactly supersymmetric string bit model. As we have established, supersymmetry is present when the fermionic discrete derivative and the bosonic one d satisfy the relation (13). The problem arises when we try to solve equation (13) for the fermionic 6 with a given bosonic d, corresponding to the next neighbor lattice derivative. In this section we want to show that the problem disappears if one relaxes the latter condition and allows also for an arbitrary bosonic derivative with unambiguous square root (13). Since such a derivative should vanish at the end of the Brillouin zone (k = f J / 2 , in the case of an even number of bits), the bosonic sector should be also doubled! In other words, since we are unable to avoid the fermionic doubling without the alteration of the model let us allow the doubling also for the bosons, then one can expect to have supersymmetry. The simplest choice for a “fermionizable” bosonic derivative function dk can be obtained by choosing it to be the “doubled” fermionic one
a
1 arcneu, = - sin (27rka) . 2a
(21)
This corresponds to the symmetric derivative of the bosonic fields, anewfn =
(1/2a)(fn+1 - fn-1).
Since the derivative (21) is purely imaginary, it solves also Eq. (13) for the fermionic derivative: = Pew. Then, the supersymmetric Hamiltonian is a combination of the bosonic part with derivative (21) and the naive fermionic part
a
which generates the supersymmetry algebra (5) together with the supercharges (6) and the shift operator (4), where we put 0 = = Pew. Let us note that no even bit n in Eq. (22) interacts with an odd one in both fermionic and bosonic sectors. Therefore, the model is doubled in the bosonic sector as well as in the fermionic one. This is confirmed by the inspection of the bosonic Hamiltonian which now reads
a
169
As it can be seen from Eq. (23), sin2(2rka) has two zeroes: one at the origin k = 0 and one at the edge of the Brillouin zone k = fJ. Each of these zeroes contributes in the continuum limit a superstring. Therefore one ends up with two superstrings. One can consider a derivative with any even number 2s of zeroes of the derivative function &. In this case one will have 2s superstrings in the continuum limit. Let us note that, since the supersymmetry is present at any stage of the discrete model, the continuum theory is also supersymmetric.
6. Discussion In this note we reviewed the fermionic spectrum of the string bit model. In the naive formulation the model suffers from inconsistencies due to the fermion spectrum doubling and supersymmetry breaking. Supersymmetry is not restored automatically in the continuum limit. One can “optimize” the fermionic sector for the given form of the bosonic Hamiltonian, in order for supersymmetry breaking to be minimal. In fact, one can have supersymmetry well-defined on all modes but one placed on the edge of the Brillouin zone. We have shown that this discrepancy does not affect, however, the physical spectrum of the model, since it depends on the square of the fermionic derivative which is a well-defined function. In the case one allows doubling also for bosonic part, one can construct a fully supersymmetric discrete theory. The continuum limit of this model is expected to contain a pair or, in general, any even number of superstrings. It is interesting to find out if one can use a correlated doubling removing procedure for both the bosonic and the fermionic sector, like the Wilson generalization or some other procedure, in order to get a single copy of string in the continuum limit and preserve supersymmetry. We hope to return to this elsewhere. Finally, our analysis leads us to the conclusion that the obstructions on the existence of discrete supersymmetric models are not enough to make the BMN limit in SYM theory problematic. Beyond this, there is no topological nor any other obstruction to this, since the theory is not chiral. In this sense, this case is similar to SU(2) chiral fermions on the lattice, where one can have a discrete theory preserving all perturbative ~ y m r n e t r i e s . ~ ~
Acknowledgements This work is supported by Russian grant for support of leading scientific schools 2052.2003.1, NATO Collaborative Linkage Grant PST.CLG.
170
97938, INTAS-00-00254 grant, RF Presidential grants MD-252.2003.02, NS-1252.2003.2, INTAS grant 03-51-6346, RFBR-DFG grant 436 RYS 113/669/0-2, RFBR grant 03-02-16193 and the European Community’s Human Potential Programme under contract HPRN-CT-2000-00131 Quantum Spacetime. We thank G. Arutyunov, N. Beisert and J. Plefka for useful discussions and the organizers of the Conference for warm hospitality and creative atmosphere. References 1. J. M. Maldacena, Adv. Theor. Math. Phys. 2,231 (1998), hep-th/9711200. 2. S. S. Gubser, I. R. Klebanov, and A. M. Polyakov, Phys. Lett. B428, 105 (1998), hepth/9802109. 3. 0. Aharony, S. S. Gubser, J. M. Maldacena, H. Ooguri, and Y . 0 2 , Phys. Rept. 323,183 (ZOOO), hep-th/9905111. 4. I. Bena, J. Polchinski, and R. Roiban, Phys. Rev. D69, 046002 (2004), hepth/0305116. 5. D. Berenstein, J. M. Maldacena, and H. Nastase, JHEP 04, 013 (2002), hepth/0202021. 6. D. Berenstein and H. Nastase, On lightcone string field theory from super Yang-Mills and holography, hep-th/0205048. 7. D. Berenstein, E. Gava, J. M. Maldacena, K. S. Narain, and H. Nastase, Open strings on plane waves and their Yang-Mills duals, hepth/0203249. 8. R. R. Metsaev, Nucl. Phys. B625, 70 (2002), hep-th/0112044. 9. R. R. Metsaev and A. A. Tseytlin, Phys. Rev. D65, 126004 (2002), hepth/0202109. 10. S. Frolov and A. A. Tseytlin, Nucl. Phys. B668, 77 (2003), hep-th/0304255. 11. G. Arutyunov, S. Frolov, J. RUSSO,and A. A. Tseytlin, Nucl. Phys. B671,3 (2003), hep-th/0307191. 12. A. A. Tseytlin, Spinning strings and AdS/CFT duality, hep-th/0311139. 13. H. Verlinde, Bits, matrices and 1/N, hep-th/0206059. 14. D. Vaman and H. Verlinde, Bit strings from N = 4 gauge theory, h e p th/0209215. 15. J.-G. Zhou, pp-wave string interactions from string bit model, h e p th10208232. 16. J. A. Minahan and K. Zarembo, JHEP 03,013 (2003), hep-th/0212208. 17. N. Beisert and M. Staudacher, The n/ = 4 SYM integrable super spin chain, hepth/0307042. 18. H. B. Nielsen and M. Ninomiya, Phys. Lett. B105, 219 (1981). 19. S. Bellucci and C. Sochichiu, Phys. Lett. B571, 92 (2003), hep-th10307253. 20. S. Bellucci and C. Sochichiu, Phys. Lett. B564, 115 (2003), hep-th/0302104. 21. U. Danielsson, F. Kristiansson, M. Lubcke, and K. Zarembo, String bits without doubling, hepth/0306147. 22. N. Beisert, The su(2--3) dynamic spin chain, hep-th/0310252. 23. C . Sochichiu, Phys. Lett. B422, 227 (1998), hep-lat/9710009.
DIVERGENCIES IN &EXPANDED NONCOMMUTATIVE S U ( 2 ) YANG-MILLS THEORY
M. BURIC AND
v. RADOVANOVIC
Faculty of Physics, University of Belgrade P.O. Box 368, 11001 Belgrade, Serbia and Montenegro E-mail: [email protected], [email protected] We analyze the divergent part of the one-loop effective action for the noncommutative SU(2) gauge theory coupled to the fermions in the fundamental representation. We show that the divergencies in the 2-point and the 3-point functions in the 8linear order can be renormalized, while the divergence in the 4-point fermionic function cannot.
1. Introduction Although discussed for quite some time, the question of renormalizability of field theories on noncommutative (NC) R4 has not been settled in a satisfactory way yet. Noncommutativity of the coordinates, i.e., a relation of the type
[P, 27 = i
(1) puts the lower bound on the coordinate measurements, so one would expect that it also implies a natural ultraviolet cut-off and acts as a regulator. However, this idea has not been successfully implemented in models. Usually one represents the noncommutative field &(?) by a function &(z) on R4 and encodes the noncommutativity in the multiplication rule (*-product). For example, for constant P”, the field multiplication is given by the Moyal-Weyl product: i OPY
8
e y
=+?iqz)x(u)/v+”
. (2) This product is nonlocal, and so is the field theory defined by it. In this paper we study the renormalizability of NC S U ( 2 ) in the so called &expanded approach. As shown by Seiberg and Witten,4 noncommutative and commutative gauge theories are in some aspects equivalent. This equivalence is realized by a mapping relating the representation
4 ( x ) * x ( z ) = e5
171
172
of noncommutative gauge symmetry to the fields carrying the representation of its commutative counterpart. The Seiberg-Witten (SW) map is given as a series in powers of 0,". Classical action also can be expanded in 6: in the zero-th order it reduces to the action of ordinary gauge theory; additional terms can be treated as couplings. Renormalizability of the &expanded NC gauge theories has been addressed in Refs. 5-8 for the gauge group U (1) with and without fermionic matter and for its supersymmetric extension. Here we discuss the SU(2) theory coupled to fermions in the fundamental representation. Although the lagrangian of the SU(2) gauge theory technically differs from the abelian U(l), the conclusion concerning renormalizability is the same one. If fermions are massive, theory is not renormalizable. For massless fermions the theory is 'almost renormalizable', meaning that there is only one divergent term in the effective action which cannot be absorbed by the SW field redefinition scheme. The method which we use to calculate the divergent contributions is the background field method. As it was thoroughly explained in Refs. 8, 9, we skip the technical points here, referring also to the standard literature.'' The calculations for SU(2) are quite involved so we discuss only the 0-linear order; we find the divergent parts of the 2-point, 3-point and fermionic 4point functions. Previously, the results for U(1) in the 6-linear order were given in Refs. 5, 6; for the 2-point functions they were extended in Ref. 8 to the 02-order.
2. The Model The general construction of gauge theories on noncommutative space and their relation to the SW map were introduced in Refs. 1-3. The classical action for NC SU(2) Yang-Mills theory is given by
S=
J
's
d4x I$ * (iypD, - rn)G - 4
where the noncommutative field strength
F,,
= a,A,
-
a,A,
-
F,,
d4x Tr (F,,
*Fpy),
is defined with
i(A, *A, - A, *A,),
and the covariant derivative is
If we denote the commutative fields by
(3)
(4)
173
Dpd = q L $
- iApd
(7)
1
(matter, $(z) is in the fundamental representation of SU(2): T” = %), the noncommutative fields are related to the commutative ones by the SW map A p ( x ) = A p ( x )-
41 8’”” { A p ( x )&Ap(z) , + J ’ v p ( x ) } +...
(8)
1
1 i $(x) = $(z) - - ~ p u A p ( x ) d u $ ( z )4e””Ap(X)Av(x)$(x) . . . 2
+
+
to the first order in 0 . Inserting the expansion (8) into Eq. (3), we get the classical action in &linear order:3
To simplify the notation we introduced the symbol A:;; defined as A:,”,’ = 6:6;6;
- 6;6!6;
+ (cyclic c@y)
= - E ~ ~ ~ ~ E , ~(12) ~ x .
The bosonic &linear term S 1 , A vanishes (unlike in the U ( 1 ) case) because it is proportional to the symmetric coefficients dab‘ = Tr ({T”,Tb} Tc),and for S U ( 2 ) these coefficients are zero for all irreducible representations. In the functional integration we will treat ( A p l$) as a multiplet, so we want both gauge and matter fields to be real (or to be complex). Therefore, we write the Dirac spinor $J in terms of the Majorana spinors +1,2. For the charge-conjugated spinor QC = CqT Majorana spinors are given by $ 1 , ~ = ~(.ICI*$J~); vice versa: .1c, = $1 +id2 . To express the action in terms of Majorana spinors we have to use explicitly the form of Pauli matrices (i.e., the representation of the group generators). CTZ is antisymmetric and D ~ , C Tare Q symmetric matrices and therefore A: couples to Majorana spinors
174
A;. The action reads:
differently from A:,
This is the initial point for the quantization.
3. One-loop Effectiwe Action Background field method is one of the standard methods t o obtain divergent and finite quantum contributions to the classical action.1° In the first step, one expands fields around their classical configuration, i.e. splits the fields into the background (classical) part and the quantum correction: A,
--+
A,
+ A,
,
$1,2
+
A,’ + *1,2.
(15)
Quantum fields are denoted here by A,, Q1,2. The functional integration over the quantum fields in the generating functional is then performed; the effective action, r, is the Legendre transformation of the generating functional. In the saddle-point approximation, the integration gives: I’[Ap,$l,$zl = S[A,,$i,$2]
+ i S d e t logS(2)[A,,$1,$zl.
(16)
Sdet denotes the functional superdeterminant and S(’) is the second functional derivative of the classical action. For polynomial interactions, the second derivative can be obtained from the quadratic part of the action; it is an expression of the type:
175
where we wrote B instead of S(2)as, in fact, we have to include the gauge fixing term, too. We use the background field analogue of the Feynman't Hooft gauge,"
S,,
=
-12 /d4z(D,A,a)2,
with the gauge k i n g parameter equal to 1. D, is the background covariant derivative, D,A"" = d,A"" PbcA:Auc . To calculate the one-loop correction
+
i I"') = 1 log Sdet B = - STr logB 2 2 perturbatively, one expands logB. In our notation B can be written as a 3 x 3 block matrix
B11 Bl2 B13
B31 B32 B33 The submatrices B12, B13, B21 and B31 are Grassmann-odd, while the rest are Grassmann-even; the supertrace is defined by STrB = TrBll TrB22 - TrB33. B depends on the classical fields. One should keep in mind that A; (for a = 1,2,3) is a triplet, while $1 and $2 are dublets of the SU(2) group. In the absence of the interaction, B is the inverse propagator; in general case, one can separate the kinetic part: $gapS"bO 0 0 i0) O ip) + M .
B=(
We are to expand logB around identity I = diag(g,,bab, 1,l).To achieve this, we multiply B by C:ll 2 0
c = (o-i? 0 0
0 0
-ip
),
and then for the one-loop correction we obtain i I?(') = - STr log(BC) + STr log C-l 2 2 i i = - STr log(1 O-lMC) STr log C-' STr log 0 . (19) 2 2 The second and the third terms, being independent on the fields, can be included in the infinite renormalization. Note that now the propagator for
+
+
+1
176
all fields is 0 - l . The operator O-lMC defines the rules in the perturbation expansion. To get the structure of the expansion more clearly, we decompose MC into the sum
MC = NO
+ + + + T2 +T3 N1
N2
Tl
(20)
with respect to the number of fields - indices denote their number in a given term. NO,N1 and N2 originate from the commutative theory, while Ti, T2 and T3 are the noncommutative interactions linear in 6'. One can read No.. .T3 from the action (13-14), after the separation of the part quadratic in the quantum fields. The details of the structure of (20) are given in Ref. 9; we shall omit them here. Let us just mention that all operators have the same form Qii Q12
Qi3
Q31 -Q23
Q22
which is a consequence of the fact that the action was originally in Dirac spinors. In the case of the massless fermions all matrices simplify; No = 0.
4. Divergencies Introducing the decomposition (20) for the one-loop correction (19) we get i 00 (-)n+l r(l)= -2i s n log (I + O - ~ M C )= -2 n n=l
C
XSTr
~
(o-%J+ o-W1+ 0-1N2 + 0-1T1 + 0-9, + O-1T3y
Our notation allows to easily extract the contributions from different terms in this expansion. Parts of the effective action which give the 2-point functions have two classical fields. This means that the sum of indices in the monomials which we are interested in is equal to 2. Since there is an operator with the index 0, in principle, infinitely many terms contribute to the n-point functions. However, as we are calculating the divergencies only, we will need just finite number of terms: O-'No behaves as p-l in the momentum space, and the integrals become convergent for ( O - l N ~ ) kof a high enough degree. For the 2-point function in the 6'-linear order, potentially divergent terms are: STr(0-1 N1 O-lT1), STr( (O-lNoO-lNi 0-N1 - NO)0-2'1 ) , STr (0-NO0-T2) and STr ((0-No) 0-N1
+
'
177
'
+O- N0O-l NICI-~NO + I T 1 N1( 0-1 N o ) ~ O - ' T ~Adding ). their contributions with the appropriate coefficients after the dimensional regularization we obtain the divergent part
We shall need the covariant form of this formula; in m = 0 case it reads
The result ( 2 2 ) is nice: comparing it with the &linear correction for the 2-point functions in noncommutative electrodynamics18 we see that in the SU(2) case also, only fermionic propagator gets a correction; the correction is up to a factor the same as the one for U(1). This has further consequences. In NC QED we argued that the massive terms obstruct renormalization, as only for the case m = 0 one can redefine fields in such a way that the divergent terms disappear. The analysis can be repeated for SU(2) without change. Hence, we come to the conclusion: the NC gauge theories with massive fermions are not renormalizable. For this reason in the calculations of 3-point and 4-point functions we focus to the massless case. One might add that the calculations become so cumbersome that otherwise they would hardly be doable. The divergent parts of the 3-point functions are in the terms STr ((O-1N1)20-1T1) and STr (0-1N10-1T2). After a tedious but straightforward calculation one obtains the result which we present in its covariant form:
+ 2i47,FvaDV + i$7,paFva)$ 1 + -58 Gmpp4757P(D,Fa99 - fp"ap475YsYa(DaFa9$ 16 1 + 8 E,"ap(24757PFPaDp7b + 4rs7"DpFP")i)) *
(24)
4-fermionic vertex occured to be very important for the discussion of renormalizability. The corresponding 4-point function is relatively easy to find: to this end one can put A; = 0 in N 1 . . .T3. The divergent part
178
comes from STr ((O-1N1)20-1T2) and STr ((O-1N1)30-1Z’1); one gets
The equations (23), (24) and (25) are the principal results of this paper. 5 . Discussion
As it is well known,12 SW map is not unique. The redefinition of fields which it allows changes the action by terms of the form: d4x ( D , F p ” ) A P ) ,
AS$) =
/
+ iidn)ip$),
d4x (sip!D@)
(26)
(27)
written for the case of massless fermions. Here A?) and !D(n) are gauge covariant expressions of the n-th order in 0. The important thing in Eqs. (26 and 27) is that, except for the fields Fp” and lc,, a t least one derivative must be present. The obtained divergencies in Eqs. (23), (24) and (25) are such that they cannot be subtracted by the usual counterterms. However, if they were of the types (26-27), we could include them in the field redefinitions; thus the theory would be renormalizable in a generalized sense. Analyzing the divergencies, we see that the situation with the NC SU(2) is pretty much the same as with the electrodynamics. We already noted that the propagator correction (23) breaks the renormalizability, unless m = 0. The 3-point functions in the massless case present no problem, too. Gluon 3and 4-vertices get no quantum correction in the &linear order - there is no classical gluon vertex in Eq. (14) in that order, either. In this respect the behavior is again similar t o U(1), where 0-linear 3-photon vertices did exist: quantum one-loop corrections were precisely of the same form as the corresponding classical vertices.6 Further, the fermion-gluon 3-vertex (24) is already written in the form (27) with
(
+ ir;2uppFvp+ i ~ 3 ~ ~ , , , , , ~+564upVD2)lc,. F~~ (28)
i P E r (= l ) Op” 61Fpy
We observe that the fermions in the renormalized theory would be redefined via the gluon fields, i.e., noncommutativity would be mixed with (or partly immersed into) the gauge interactions.
179
However, there is a divergent term which spoils the renormalizability: the 4-point function (25). It predicts the current-current interaction, and there is no simple way to circumvent this coupling induced by noncommutativity. In fact, from the dimensional analysis we can see that in the massless case the propagators in NC gauge theories are renormalizable to all orders. Namely, for gauge bosons the n-th order corrections are of the form e . . . e A a ... a ~ , * k n
while for fermions they are e...eqya w wa $ . n
k
The power counting gives the number of derivatives k: for the gluon propagator k = 2 + 2n; for the fermions, k = 1 2n. This shows that one can, in all orders, transform (29-30) into the desired forms (26-27). The use of the background field method guarantees the gauge covariance. On the other hand, the vertices are potentially problematic. From the power counting we see that in the linear order the ‘wrong’ vertex could be
+
e(13;r$)2 ,
(31)
while in the quadratic order we could have, e.g.,
e 2 ~, 4 e2(&q2F.
(32)
These terms contain no derivatives and therefore break the SW generalized renormalization scheme. The term (31) is present for both U(1) and SU(2) theories coupled to fermions. An interesting fact, however, is that in both theories it has the same form (a different coefficient), namely e ~ ” q L ” p u 4 Y 5 Y ‘ ~4 Y P $
1
(33)
whereas the other combinations allowed by covariance, as, e.g. e ~ ’ ~ p v p u ~ y g y$oqyp+$ + or e ~ j y ~ y ,$$ y,,$, never show up. This opens the possibility that this divergence might cancel in a gauge theory based on the product of gauge groups. However, we are not in favor of a theory which needs too much fine tuning. Thus we are inclined to interpret our results (and the previous one^^^^^^) as an indication that NC gauge theories coupled to fermions are not renormalizable. But before a definite conclusion, one should certainly check whether the specific €J2-corrections,as 4A vertex (32), vanish. The
180
presence of the €J2F4divergence would prove non-renormalizability, possibly even for the pure gauge theories. It would also be interesting t o understand if there is some systematics in the behavior of various divergent terms.
Acknowledgment This work is supported by the project 1468 of the Serbian Ministry of Science, Technology and Development.
References 1. J. Madore, A n Introduction to Noncommutative Differential Geometry and its Physical Applications, Camb. Univ. Press, Cambridge (1999). 2. J. Madore, S. Schraml, P. Schupp and J. Wess, EUT.Phys. Jour. C16, 161 (2000). 3. B. Jureo, L. Moller, S. Schraml, P. Schupp and J. Wess, Eur. Phys. J. C21, 383 (2001). 4. N. Seiberg and E. Witten, JHEP 9909, 032 (1999). 5. A. A. Bichl, J. M. Grimstrup, H. Grosse, L. Popp, M. Schweeda and R. Wulkenhar, JHEP 0106, 013 (2001). 6. R. Wulkenhaar, JHEP 0203, 024 (2002). 7. A. A. Bichl, M. Ertl, A. Gerhold, J. M. Grimstrup, H. Grosse, L. Popp, V. Putz, M. Schweda and R. Wulkenhaar, Non-commutative U ( 1) Super- YangMills Theory: Perturbative Self-Energy Corrections, hep-th/0203141. 8. M. Burit and V. Radovanovit, JHEP 0210, 074 (2002). 9. M. Burit and V. RadovanoviC, JHEP 0402, 040 (2004). 10. M. E. Peskin and D. V. Schroeder, A n Introduction to Quantum Field Theory, Addison-Wesley, Reading, MA (1995). 11. A. G. Barvinsky and G. A. Vilkovisky, Phys. Rev. 119, 1 (1985). 12. T. Asakawa and I. Kishimoto, Comments on Gauge Equivalence in Noncommutative Theory, hep-th/9909139.
HETEROTIC STRING COMPACTIFICATIONS WITH FLUXES
G. L. CARDOSO, G. CURIO, G. DALL'AGATA AND D. LUST Humboldt Uniwersitiit z u Berlin, Institut fur Physik, Newtonstrasse 15, 12489 Berlin, Germany E-mail: gcardoso, curio, dallagat, [email protected] We review some of our work on compactifications of heterotic string theory to four dimensions in the presence of H-flux.
1. Introduction String compactifications in the presence of fluxes have been revived recently as an appealing way to address the moduli problem. Turning on fluxes in the ten-dimensional string theories produces, at the level of the effective four-dimensional action, a potential. Upon minimization of this potential one finds new vacua with (generically) less moduli. Moreover, one expects to find Minkowski vacua only when the deformation of the internal manifold balances the presence of the fluxes. Supersymmetric compactifications of the heterotic string in the presence of three-form H-flux requires the six-dimensional compactification manifold to be n~n-Kahler.'-~ More precisely, one finds that a non-vanishing H-flux is associated to a non-vanishing exterior derivative of the complex structure J,5
1 H = --e-'+ 2
* d(e84J).
(1)
This condition, which follows from the analysis of the supersymmetry rules, can also be understood from the perspective of rewriting the tendimensional effective action of heterotic string theory as a sum of squares.6 This rewriting results in the following term (among others),
181
182
This makes it clear that in order to obtain a vanishing four-dimensional effective potential V = 0, the condition (1) has to be imposed. Therefore, one obtains supersymmetric Minkowski vacua by fixing the (2,l) + (1,2) Hodge components of the three-form flux, H(’y1) and H(l>’),in terms of the internal geometry. The other Hodge components, i.e. H(3t0)and H(073), must vanish in order to have unbroken supersymmetry. In the following, let us review the rewriting leading to Eq. (2).
2. Fluxes and Torsion
The bosonic part of the Lagrangean up to second order in a’ is given by7
This action is written in the string frame and its fermionic completion makes it supersymmetric using the three-form Bianchi identity given by
dH = a’ (tr R+ A R+ - t r F A F ) ,
(4)
where the curvature RS is the generalized Riemann curvature built from the generalized connection V+ (i.e. from wf = w H ) . In the search for a BPS rewriting of Eq. ( 3 ) , 6 we will assume that the ten-dimensional space is given by the warped product of four-dimensional Minkowski spacetime with a six-dimensional internal space admitting an SU(3) structure. In order to consistently obtain that setting to zero the BPS-like squares implies a solution to the equations of motion, we also impose that the only degrees of freedom for the various fields are given by expectation values on the internal space and are functions only of the internal coordinates. To simplify the discussion we limit ourselves to the case with dilaton and warp factor identified, i.e. q5 = A, but the generalization of the following results is straightforward. After various manipulations, the action (3) can
183
be written as6
-
14 J d6y &eg+
Nmnpgmqgnrgps N~~~
In this expression the traces are taken with respect to the fiber indices a, b, . . ., whereas the Hodge type refers to the base indices m, n,. . . of the curvatures. The other geometrical objects appearing in the above expression are the Lee-form 3 0 JAdJ = - J"" qrnJnpldxP , (6) 2 the Nijenhuis tensor
NmnP= JmqdiqJnlP- Jnq6jq J,,',
(7)
and the generalized curvature R, which is constructed using the Bismut connection built from the standard Levi-Civita connection and a totally antisymmetric torsion T B proportional to the complex structure,
The action ( 5 ) will now be used to find the conditions determining the background geometry by demanding the vanishing of ( 5 ) . Setting the squares to zero yields 0
0
the vanishing of the Nijenhuis tensor the vanishing of some components of the generalized Riemann curvature constructed from the V+ connection, the vanishing of
1 d4+-0=O, 8
(9)
184
the vanishing of
1 H+-*e-'+d(e'+J) 2 0
=0,
(10)
the vanishing of
The vanishing of the Nijenhuis tensor states that the internal manifold is complex. The conditions on the R+ curvature can be translated into the requirement of SU(3) holonomy for the V- connection. The proof requires the identity b R :
cd
= R z ab - (dH)abcd
(11)
7
which relates the R+ and R- curvatures with the base and fiber indices swapped. Using this identity and the fact that d H gives higher order terms in a', the conditions on the base indices of R+ become conditions on the R- fiber indices, to lowest order in a', R-
(2,o) = R- (0,2)
= JabR-
ab
=0
'
(12)
These conditions precisely state that the generalized curvature R- is in the adjoint representation of SU(3) c SO(6) and therefore its holonomy group is contained in SU(3). The conditions in the gauge sector are that the gauge field strength is of type (1,l)and J traceless. On a complex manifold, the condition (10) yields
=0 , H(3,0)+(0>3)
(13)
where we also used (9). On the solution, R+ A R+ and F A F are of type (2,2). Therefore, the Bianchi identity (4) implies that (dH)(311)f(193)= 0, which is indeed satisfied. We conclude our discussion with remarks about the possible superpotential describing such vacua in the effective four-dimensional theory. It has been argued that a candidate superpotential describing the N = 1 vacua of the heterotic theory in the presence of fluxes is given by6,'
where R denotes a holomorphic (3,O)-form on Ms.This conjecture is motivated by the hope of getting the conditions (13) out of W = DW = 0.
185
For instance, assuming that H is constant under variation of the almost complex structure moduli and that &s2 = kiR xi,where xi are a set of ( 2 , l ) forms and ki are constants over the internal manifold, one obtains t h a t t h e (1,2) part of H $ d J is vanishing. On a complex manifold, this yields the locking of the t h r e e f o r m flux onto the geometry, namely
+
+
H(ly2)+ 8 J = 0 , and since H is a real form, it follows that
References 1. A. Strominger, Nucl. Phys. B274, 253 (1986). 2. G. L. Cardoso, G. Curio, G. Dall’Agata, D. Lust, P. Manousselis and
G. Zoupanos, Nucl. Phys. B652, 5 (2003). 3. K. Becker, M. Becker, K. Dasgupta and P. S. Green, JHEP 04, 007 (2003). 4. J. P. Gauntlett, D. Martelli and D. Waldram, Phys. Rev. D69,086002 (2004). 5. J. P. Gauntlett, N. w. Kim, D. Martelli and D. Waldram, JHEP 0111, 018 (2001). 6. G. L. Cardoso, G. Curio, G. Dall’Agata and D. Lust, JHEP 0310,004 (2003). 7. E.A. Bergshoeff and M. de Roo, Nucl. Phys. B328, 439 (1989). 8. K. Becker, M. Becker, K. Dasgupta and S. Prokushkin, Nucl. Phys. B666, 144 (2003).
SYMMETRIES AND SUPERSYMMETRIES OF THE DIRAC-TYPE OPERATORS ON EUCLIDEAN TAUB-NUT SPACE
I. I. COTAESCU West University of Tamigoara, V. P6rvan Ave. 4, RO-1900 Timigoara, Romania E-mail: [email protected]
M. VISINESCU Department of Theoretical Physics, National Institute for Physics and Nuclear Engineering, P.O.Box M. G.-6, Magurele, Bucharest, Romania E-mail: [email protected]
The role of the Killing-Yano tensors in the construction of the Dirac-type operators is pointed out. The general results are applied to the case of the four-dimensional Euclidean Taub-Newman-Unti-Tamburino space. Three new Dirac-type operators, equivalent to the standard Dirac operator, are constructed from the covariantly constant Killing-Yano tensors of this space. The Dirac operators are related among themselves through continuous or discrete transformations, In this space there is also a non covariantly constant Killing-Yano tensor connected with hidden symmetries. The Runge-Lenz operator for the Dirac equation in this background is written down pointing out its algebraic properties.
1. Introduction The (skew-symmetric) Killing-Yano (K-Y) tensors, that were first intrcduced by Yanol from purely mathematical reasons, are profoundly connected to the supersymmetric classical and quantum mechanics on curved manifolds where such tensors do exist.2 The K-Y tensors play an important role in theories with spin and especially in the Dirac theory on curved spacetimes where they produce first order differential operators, called Dirac-type operators, which anticommute with the standard Dirac one, D,.3 Another virtue of the K-Y tensors is that they enter as square roots in the struc186
187
ture of several second rank Stackel-Killing tensors that generate conserved quantities in classical mechanics or conserved operators which commute with D,. The construction of Carter and McLenaghan depended upon the remarkable fact that the (symmetric) Stackel-Killing tensor K,, involved in the constant of motion quadratic in the four-momentum p p
has a certain square root in terms of K-Y tensors f,,: Kpv =
The K-Y tensor here is a 2-form
fpu
fpxf.,A . .
(2)
= -f,,, which satisfies the equation
These attributes of the K-Y tensors lead to an efficient mechanism of supersymmetry, especially when the Stackel-Killing tensor K,, in Eq. (1) is proportional with the metric tensor gPv and the corresponding K-Y tensors in Eq. (2) are covariantly constant. Then each tensor of this type, fi, gives rise to a Dirac-type operator, Di, representing a supercharge of the superalgebra {Di, Dj} 0: D2&. The general results are applied to the case of the four-dimensional Euclidean Taub-Newman-Unti-Tamburino (Taub-NUT) space. The Euclidean Taub-NUT metric is involved in many modern studies in physics. This metric might give rise to the gravitational analog of the Yang-Mills i n ~ t a n t o nThe . ~ Kaluza-Klein monopole of Gross and Perry5 and of Sorkin‘ was obtained by embedding the Taub-NUT gravitational instanton into fivedimensional Kaluza-Klein theory. On the other hand, in the long-distant limit, the relative motion of two monopoles is approximately described by the geodesics of this space.? The Euclidean Taub-NUT space which is a hyper-Kahler manifold possessing three covariantly constant K-Y tensors. Using these covariantly constant Killing-Yano tensors it is possible to construct new Dirac-type operators3 which anticommute with the standard Dirac ~ p e r a t o r . * ~ ~ The Taub-NUT space also possesses a Killing-Yano tensor which is not covariantly constant. The corresponding non-standard operator, constructed with the general rule3 anticommutes with the standard Dirac operator but is not equivalent to it.lo This non-standard Dirac operator is connected with the hidden symmetries of the space allowing the construction of a conserved vector operator analogous to the Runge-Lenz vector of the Kepler problem. We shall discuss the behavior of this operator under
188
discrete transformations pointing out that the hidden symmetries are in some sense decoupled from the discrete symmetries studied here. 2. Dirac Equation on a Curved Background In what follows we shall consider the Dirac operator on a curved background which has the form (4)
D, = y p V p .
In this expression the Dirac matrices y,,are defined in local coordinates by the anticommutation relations {y,,y’} = 2gp’I and V, denotes the canonical covariant derivative for spinors. Carter and McLenaghan showed that in the theory of Dirac fermions for any isometry with Killing vector R, there is an appropriate ~ p e r a t o r : ~
which commutes with the standard Dirac operator (4). Moreover, each Killing-Yano tensor f p u produces a non-standard Dirac operator of the form A
1
D, = - i Y Y f ( f , ” V u - p ” r V p u ; p )
(6)
1
which anticommutes with the standard Dirac operator D,.
3. Dirac Operators of the Taub-NUT Space Let us consider the Taub-NUT space and the chart with Cartesian coordinates z@(p, v = 1,2,3,4) having the line element
+
+
where 2 denotes the three-vector 2 = ( T , 0, cp), (dZ)2 = ( d ~ l ) ( ~ d ~ ~ ) ~ ( d ~ ’ and ) ~ A’ is the gauge field of a monopole divA’=O,
-
2 B =rotA=p-.
r3
(8)
The real number p is the parameter of the theory which enters in the form of the functions P+r f ( r ) = g-l(T) = V-’(r) = -. (9) r
189
In the Taub-NUT geometry there are four Killing vector^.'^^^^ On the other hand, in the Taub-NUT geometry there are known to exist four KillingYano tensors of valence 2. The first three are covariantly constant
D,
fr' = 0 ,
i, j , k = 1 , 2 , 3 .
( 10)
These first three Killing-Yano tensors of the Taub-NUT space16 are rather special since they are covariantly constant. The f i define three anticommuting complex structures of the Taub-NUT manifold, their components realizing the quaternion algebra fifj+ fjfi
=-2&,
fifj
- f j f i = - 2 & . .k f k .
(11)
The existence of these Killing-Yano tensors is linked to the hyper-Kahler geometry of the manifold and shows directly the relation between the geometry and the N = 4 supersymmetric extension of the theory.2*18 For the theory of the Dirac operators in Cartesian charts of the TaubNUT space, it is convenient to consider the local frames given by tetrad fields .a(.) and 2"(x) while the four Dirac matrices y" satisfy {y", yP} = 2S@, where all of them are self-adjoint. In addition, we consider the matrix y5 = yiyiyiyd which is denoted by yo in Kaluza-Klein theory explicitly involving the time." The standard Dirac operator of the theory without explicit mass term is defined D, = y"V&, where the spin covariant derivatives with local indices, V", depend on the momentum operators, Pi = -i(ai - A&) and P4 = -4,and spin connection,12such that the Hamiltonian 0 p e r a t o r l ~ 7 ~ ~
can be expressed in terms of Pauli operators,
involving the Pauli matrices, ~ i .These operators give the (scalar) KleinGordon operator of the Taub-NUT space: A = -V,gp""V, = We specify that here the star superscript is a mere notation that does not represent the Hermitian conjugation because we are using a non-unitary representation of the algebra of Dirac operators. Of course, this is equivalent to the unitary representation where all of these operators are self-adjoint.12
190
Moreover we can give a physical interpretation of these Killing-Yano tensors defining the spin-like operators,
that have similar properties to those of the Pauli matrices. In the pseudoclassical description of a Dirac particle27l8the covariantly constant KillingYano tensors correspond to components of the spin which are separately conserved. Here, since the Pauli matrices commute with the Klein-Gordon operator, the spin-like operators (14) commute with H 2 . Remarkable the existence of the Killing-Yano tensors allows one to construct Dirac-type operators3
which anticommute with D , and y5 and commute with H.l0 Another Dirac operator can be defined using the fourth Killing-Yano tensor, but this will be discussed separately in Sec. 6. 4. Equivalent Representations
In Ref. 12 we have shown that in the massless case the operators Qi (i = 1,2,3) and the new supercharge QO = iD, = iy5H form the basis of an N = 4 superalgebra obeying the anticommutation relations {QA, Q B ) = ~
~ A B H A,B ~ ,... , =0,L2,3,
(16)
linked t o the hyper-Kahler geometric structure of the Taub-NUT space. In addition, we associate to each Dirac operator Q A its own Hamiltonian operator QA = - i y ' Q ~ , obtaining thus another set of supercharges
Qo =H ,
Qi = i [ H , C i ] ,
(17)
which obey the same anticommutation relations as Eq. (16). Thus we find that there are two similar superalgebras of operators with precise physical meaning. Obviously, since all of these operators must be self-adjoint we have to work only with unitary representations of these superalgebras, up to an equivalence. The concrete form of these supercharges depends on the representation of the Dirac matrices which can be changed at any time with the help of a non singular operator T , such that all of the 4 x 4 matrix operators of the Dirac theory transform as X -+ X' = T X T - l . In this way one obtains
191
an equivalent representation which preserves the commutation and the anticommutation relations. We used such transformations for pointing out that the convenient representations are equivalent to a unitary one.12 We note that some properties of the transformations changing representations in theories with two Dirac operators and their possible new applications are discussed in a paper by K1i~hevich.l~ For example, simple and convenient transformations can be chosen of the form:
U(d
E SU(2). where O(p,f) = e - i P 6 ( f i E U(2) = U(1) 8 SU(2) with This is because among these transformations one could find those linking equivalent Dirac operators. It is interesting to observe that the SU(2) transformations are generated just by the above defined spin-like operators as
<
+
If we take now = 2 9 5 with 151 = 1 and cp E [O,.rr], we find that +
-
..
~ (= e-iE.d/2 0 = 1 2 cos cp - is.:sin
cp ,
(20)
and after a little calculation we can write the concrete action of Eq. (19) as
Qb = U ( f ) Q o U + ( d = QO cos cp Q: = U ( d Q i U ' ( d
+
niQi sin cp
+
= Qi coscp - (niQo
,
&ijknjQk)
(21) sincp.
(22)
Hereby we see that the supercharges are mixed among themselves in linear combinations involving only real coefficients. In addition, we observe that these transformations correspond to an irreducible representation since the supercharges transform like the real components of a Pauli spinor. In other words, the usual SU(2) transformations $Q 4 = U+(()$Q of the spinor-operator
$b
'Q=
QO - iQ3
(Q2-iQl)
give just the transformations (21) and (22).
'
192
5. Discrete Transformations Let us focus now only on the transformations which transform the supercharges QA among themselves without to affect their form. From (21) and (22) we see that there exists particular transformations, Qk
, k = 1,2,3 ,
= UkQou:
(24)
where the matrix u k = diag(-ick112) is given by -ick E s u ( 2 ) . In addition, we consider the parity operator P = P-l = -y5 which changes the sign of supercharges, PQAP=-QA,
A=0,1,2,3.
(25)
Then it is not hard to verify that: the identity I = 1 4 , P and the sets of matrices u k and P u k (k = 1,2,3) form a discrete group of order eight the multiplication table of which is determined by the following rules
P2= I ,
Puk = UkP,
u12 = u22
U1U2 = U3 ,
= u32 = P ,
(26)
U2U1 = PU3 , ... etc.
We denote this group by GQ since it is a realization of the quaternion group Q which is isomorphic with the dicyclic group (2, 2, 2).20721In the usual representation of the y-matricesl8 its operators are defined by proper unitary matrices (which satisfy G-' = G+ and detG = 1,V G E GQ) constructed using the elements f l 2 , fiel, f i e 2 , fie3 of the natural realization of Q as a discrete subgroup of SU (2). 6. Hidden Symmetries and the Fifth Dirac Operator
In the Taub-NUT space, in addition to the above discussed covariantly constant Killing-Yano tensors, there exists a fourth Killing-Yano tensor , fY = 2p(dx+cosOdq) A d r + 2 r ( 2 r $ p ) ( l + L)sinedOAdp, P
having a non-vanishing covariant derivative r fy,.o;p = 2 ( 1 + -)r sin 0 . P
(27)
(28)
In Taub-NUT space there is a conserved vector analogous to the RungeLenz vector of the Kepler-type p r ~ b l e m l ~ ? ~ ~ (29)
193
where the conserved energy is E = ~ g , , x ~ x " . The components Kip" involved with the Runge-Lenz vector (29) are Stackel-Killing tensors satisfying the equations Ki(,v;~)= 0, Kip" = Ki",, and they can be expressed as symmetrized products of the Killing-Yano tensors f i , f y and Killing vector~.~~>~~ As in the case of the Dirac operators (15), one can use f Y for defining the fifth Dirac operator
called here the non-standard or hidden Dirac operator t o emphasize the connection with the hidden symmetry of the Taub-NUT problem. It is denoted by Q,' to point out its relation to the standard Dirac operator since it can be put in the form
Q,'
= iT [Qo
P
(:
g,v-l)] 0
9
where B, = 3 .Z / r . We showed that QF commutes with QO = H and anticommutes with QO and y5.lo This operator is important because it allowed us to derive the explicit form of the Runge-Lenz operator, 2, of the Dirac field in Taub-NUT background establishing its properties." We recall that the components of the conserved total angular momentum, f, and the operators Ri = F-lKi with F2= P42 - H 2 are just the generators of the dynamical algebra of the Dirac theory in Taub-NUT b a ~ k g r o u n d . ' ~ ? ~ ~ Starting with Q,' we can construct a new orbit, Ry, of GQ defining
(for k = 1 , 2 , 3 ) and observing that P Q Z P = -QZ , A = 0 , 1 , 2 , 3 . From the explicit form (32) we deduce that, in contrast with the operators of the orbits RQ and those of the orbit RY have more involved algebraic properties. We can convince that calculating
a~, +
H2(Q,')2 = H4
(33)
and it is worth comparing it with Eq. (16). The Dirac-type operators Q A are characterized by the fact that their quanta1 anticommutator close on the square of the Hamiltonian of the theory. No such expectation applies to the non-standard, hidden Dirac operators QZ which close on a combination of different conserved operators. Also from Eq. (33) it results that ( Q f ; ) 2#
194
(QZ)2 if A # B (because ? does not commute with u k ) . Moreover, one can show that the commutators [QX, Q';] have complicated forms which can not be expressed in terms of operators Q,'. Therefore, neither the commutator nor the anticommutator of the pairs of operators of this orbit do not lead t o significant algebraic results as the anticommutation relations (16) of the operators Q A , ( A = 0 , 1 , 2 , 3 ) . The hidden symmetries of the Taub-NUT geometry are encapsulated in the non-trivial Stackel-Killing tensors Kip",(i = 1,2,3). For the Dirac theory the construction of the Runge-Lenz operator can be done using products among the Dirac-type operators Q y and Qi. Let us define the operator:8
Ni = m { Q Y , The components of the operator lowing commutation relations
Q i } - JiP4.
fl commutes with H
[Ni, p4] = 0 , [Ni, Jj] = i & i j k N k , [Ni, Qo] = 0 [Ni, Q j ] = i E i j k Q k p 4 9
[Ni, N j ] = i S i j k J k F 2 +
(34) and satisfy the fol-
>
(35)
2 -&ijkQiH,
2
where F 2 = P42 - H 2 . In order to put the last commutator in a form close to that from the scalar case,16917 we can redefine the components of the Runge-Lenz operator, g , as follows:
Ki = Ni
+ -21H - l ( F
- P4)Qi,
having the desired commutation relation:8
7. Concluding Remarks In the study of the Dirac equation in curved spaces, it has been proved that the Killing-Yano tensors play an essential role in the construction of new Dirac-type operators. The Dirac-type operators constructed with the aid of covariantly constant Killing-Yano tensors are equivalent with the standard Dirac operator. The non-covariantly constant Killing-Yano tensors generates non-standard Dirac operators which are not equivalent to the standard Dirac operator and they are associated with the hidden symmetries of the space.
195
The Taub-NUT space has a special geometry where the covariantly constant Killing-Yano tensors exist by virtue of the metric being self-dual and the Dirac-type operators generated by them are equivalent with the standard one. The fourth Killing-Yano tensor fY which is not covariantly constant exists by virtue of the metric being of type D.The corresponding non-standard or hidden Dirac operator does not close on H as it can be seen from Eq. (33) and is not equivalent to the Dirac-type operators. As it was mentioned, it is associated with the hidden symmetries of the space allowing the construction of the conserved vector-operator analogous to the Runge-Lenz vector of the Kepler problem. Acknowledgments
M.V. thanks Professor Goran Djordjevic for his kind invitation and pleasant hospitality. This work is partially supported by a grant of the Romanian Academy. References 1. K. Yano, Ann. Math. 55,328 (1952). 2. G. W. Gibbons, R. H. Rietdijk and J. W. van Holten, Nucl. Phys. B404,42 (1993). 3. B. Carter and R. G. McLenaghan, Phys. Rev. D19,1093 (1979). 4. S. W . Hawking, Phys. Lett. A60,81 (1977). 5. D. J. Gross and M. J. Perry, Nucl. Phys. B226,29 (1983). 6. R. Sorkin, Phys. Rev. Lett. 51, 87 (1983). 7. M. F. Atiyah and N. J. Hitchin, The geometry and dynamics of magnetic monopoles, Princeton, Princeton University Press (1987). 8. I. I. Cotbscu and M. Visinescu, Gen. Rel. Grav. 35,389 (2003). 9. I. I. Cotbscu and M. Visinescu, Class. Quant. Grav. 21, 11 (2004). 10. I. I. Cotbscu and M. Visinescu, Phys. Lett. B502,229 (2001). 11. I. I. Cot&scu and M. Visinescu, Mod. Phys. Lett., A15, 145 (2000). 12. I. I. Cotbscu and M. Visinescu, Int. J. Mod. Phys. A16,1743 (2001). 13. I. I. Cotbscu and M. Visinescu, Class. Quant. Grav. 18,3383 (2001). 14. I. I. Cotbscu and M. Visinescu, J. Math. Phys. 43, 2978 (2002). 15. M. Visinescu, Int. J. Mod. Phys. A17,1049 (2002). 16. G. R. Gibbons and P. J. Ruback, Comrnun. Math. Phys. 115,267 (1988). 17. G. W. Gibbons and N. S. Manton, Nucl. Phys. B274,183 (1986). 18. J. W. van Holten, Phys. Lett. B342,47 (1995). 19. V. V. Klishevich, Class, Quant. Grav. 17,305 (2000). 20. A. 0. Barut and R. b c z k a , Theory of Group Representations and Applications, Warszawa, PWN (1977). 21. H. S. M. Coxeter and W. 0. J. Moser, Generators and Relations for Discrete Groups, Berlin, Springer-Verlag (1965).
196
22. D. Vaman and M. Visinescu, Phys. Rev. D54,1398 (1996). 23. D. Vaman and M. Visinescu, Fortschr. Phys. 47, 493 (1999).
REAL AND P-ADIC ASPECTS OF QUANTIZATION OF TACHYONS
G.S. DJORDJEVIk AND LJ. NESIC Department of Physics, Faculty of Sciences P.O.box 22.4, 18001 NiS, Serbia and Montenegro E-mail: gorandj, [email protected]. yu A simplified model of tachyon matter in classical and quantum mechanics is constructed. p-Adic path integral quantization of the model is considered. Recent results in using padic analysis, as well as perspectives of an adelic generalization, in the investigation of tachyons are briefly discussed. In particular, the perturbative approach in path integral quantization is proposed.
1. Introduction An increasing number of researchers are testing again an (almost twenty year) old idea that padics and padic string theory' can be useful in attempts to understand ordinary string theory, D-brane solutions and various aspects of tachyon^.^?^ Possible cosmological implications4 are also an interesting matter. pAdic strings have many properties similar to ordinary strings, but p adic ones are much simpler. For example, one can find an exact action for this theory, as well as for practically all variations of padic string theory. In addition, it turns out that padic string could be a very useful model for testing Sen% conjecture on tachyon condensation in open string field theory (see e.g. Ref. 5). Corresponding padic classical solution in string field theory can be explicitly found. An important fact should be also considered, that is, padic string field theory in the p 4 1 limit reduces to the tachyon effective action.6 It might be the case that results for padic strings are applicable t o boundary string field theory. The next reason for further investigation is similarity between padic D-branes (without problems with strong coupling problems) and the results found in vacuum cubic string filed theory. It is interesting that the effective energy-momentum tensor is equivalent to that of nonrotating dust.' 197
198
Generally speaking, after almost twenty years of “discoveringL‘ padic strings, our understanding of physics on padic spaces is still very poor. From many point of view, including pedagogical one, it is very useful to understand (quantum) mechanical analogies of new, unfamiliar objects, including tachyons. Considering “padic tachyons L L we stress results in foundation of padic quantum mechanics, padic quantum cosmology and their connection with standard theory on real numbers (and corresponding spaces: Minnkowsky, Riemman ...) and adelic quantum theory. It has been noted that path integrals are extremely useful in this a p proach. Following S. Kar’s’ idea on the possibility of the examination of zero dimensional theory of the field theory of (real) tachyon matter, we extend this idea to the padic case. In addition, we note possibilities for adelic quantum treating of tachyon matter. After some mathematical background in Section 2, we present padic path integrals in Section 3. Section 4 is devoted to padic strings and tachyons. Simple quantum mechanical analog of padic tachyons is considered in Section 5. The paper is ended by a short conclusion and suggestion for further research.
2. pAdic Numbers and Related Analysis Let us recall that all numerical experimental results belong to the field of rational numbers Q. The completion of this field with respect to the standard norm I (absolute value) leads to the field of real numbers R = Qoo. Completion of Q with respect to the padic norms yields the fields of padic numbers Q p ( p is a prime number). Each non-trivial norm (valuation) on Q, due to the Ostrowski theorem, is equivalent either to a padic norm I I p or to the absolute value function. Any padic number x E Q p can be presented as an expansiong
lo
x = x”(x0 + x1p
+ x2p2 + . * . ),
v E 2,
(1) where xi = 0,1, . . . , p - 1. pAdic norm of any term in (1) is P-(”+~). The padic norm is the nonarchimedean (ultrametric) one. There are a lot of exotic features of padic spaces. For example, any point of a disc &( a ) = {x E Qp : Ix - alp 5 p”} can be treated as its center. It also leads to the total disconnectedness of padic spaces. For the foundation of the path integral approach on padic spaces it is important to stress that no natural ordering on Q p exists. However, one can define a linear order as follows: x < y if lxlp < Iylp, or when 1xIp = Iylp, when there is an index m 2 0 such that the following is satisfied:
199
z o = yo,z1 = yl, ..., x ,-1
= ym-l, z , < ym.l0 Generally speaking, there are two analysis over Q p . One of them is connected with map 4 : Q p -+ Q p , and the second one is related to the map ?c, : Q p + C. In the case of padic valued function, derivatives of 4(z) are defined as in the real case, using padic norm instead of the absolute value. pAdic valued definite integrals are defined for analytic functions M
n=O as follows:
Jd
M
b
4(t)dt =
4n x(bn+l
-ant1). (3) n=O In the case of mapping Q p -+ C , standard derivatives are not possible, and several different types of pseudodifferential operators have been introduced.’~~~ Contrary, there is a well defined integral with the Haar measure. Of special importance is Gauss integral
~ where xp(u) = exp(27ri{u},) is a padic additive character, and { u } denotes the fractional part of u E Q p . X,(CY) is an arithmetic, complex-valued function.’ To explore the existence of a vacuum state in padic quantum theory we need
where 2, is the ring of padic integers (a set of all padics z, lzlp 5 1) and R is the characteristic function of 2,. It should be noted that R is the simplest vacuum state in padic Quantum Theory. There is quite enough similarity between real numbers and padics and the corresponding analysis for the so called adelic approach in mathematics” and physics (see e.g. Ref. 4), that in a sense unifies real and all padic number fields. 3. Path Integral in Ordinary and p-Adic Quantum Mechanics According to Feynman’s idea,13quantum transition from a space-time point (z’,t’) to another (d’”’’) is a superposition of motions along all possible
200
paths connecting these two points. Let us remind the corresponding probaz*i s bility amplitude is (x”,t”(x’,t’) = C , e T [,I. Dynamical evolution of any quantum-mechanical system, described by a wave function $ ( x , t ) ,is given by $(XI’,
t”) = J
K(x”,t”;X I , t’)$(x‘, t’)dx‘,
Q-
(6)
where K(x”,t”;X I , t’) is a kernel of the unitary evolution operator U(t”,t‘). In Feynman’s formulation of quantum mechanics, K(x”,t”;x’, t’) was postulated to be the path integral
where x” = q ( t ” ) and x’ = q(t’). As we know, for a classical action S(X”,t”;x’, t’),which is a polynomial quadratic in XI‘ and X I , the corresponding kernel K (for one-dimensional quantum system) reads
( i 823 > + exp (2ai -S(X”, t“;x‘, t’) K(X”,t”;2 1 , t’) = -hax~~ax/ h It can be rewritten in the form more suitable for generalization (at least from the number theory point of view)
,/-
where & = = lalz2A,(-a). In (9), x,(a) = exp(-27ria) is an additive character of the field of real numbers R. D-dimensional generalization of the transition amplitude is:
where ,A
is defined as
(,) and x = (x,), a = 1 , 2 , . . . , D. By defining AO this A,-function satisfies the same properties as A.,
= 1, one can see that
201
In padic quantum mechanics dynamical differential equation of the Schrodinger type does not exist and padic quantum dynamics is defined by the kernel Kp(x”, t”;x’,t’) of the evolution operator:
$,(x”, t”) = Up(t”,t’)$,(x’, t’) =
s,,
Kp(x”,t”;x’,t’)$,(z’, t’)dx’. (12)
All general properties which hold for the kernel K(x”,t”; t’) in standard quantum mechanics also hold in padic case, where integration is now over Q,. pAdic generalization of (7) for a harmonic oscillator was done in” starting from
( h E Q and q , t E Q,). In (13) d q ( t ) is the Haar measure and padic path integral is the limit of a multiple Haar integral. This approach was extended in Ref. 14. A rather general path integral approach, valid for analytical classical solutions, was developed for quadratic padic quantum systems in Ref. 15 (in one-dimension)
Kp(x”,t”;
t’) = A,
1
s2s
(-2hm)
1 4 xp(--S(x”, 1t”;x‘, t’)) l%mlp h a2S
(14) and Ref. 16 (two-dimensional case). The obtained padic result (14) has the same form as (11) in the real case. The higher-dimensional padic kernel was also c0nsidered.l’ Considering real-ordinary and all padic quantum mechanics on the same foot, adelic formulation is also possib1e.l’ It could be a good starting point to consider quantum mechanical analog of “real“, “padic“ and, possibly, ‘adelic“ tachyons . 4. p-Adic Strings and Tachyons
The padic open string theory can be deduced from ordinary bosonic open string theory on a D-brane by replacing the integral over the real worldsheet by padic integra1.l’ A tachyon was defined as a particle that travels faster than light, and consequently has negative mass2. Surely, it is not a convincing case for the tachyon. Quantum field theory offers a much better framework for considering such a pretty exotic physical model. If we would carry out
202
perturbative quantization of the scalar field by expanding the potential around $J = 0, and ignore higher (cubic, ...) terms in the action, we would find a particle-like state with mass2 = V”(0).In the case of V”(0)< 0 we have again a particle with negative mass2, i.e., a tachyon. The physical interpretation is that the potential V ( 4 ) has a maximum at the origin and hence a small displacement of 4 will make it grows exponentially. It is associated with the instability of the system and a breakdown of the theory. Conventional formulation of string theory uses a first quantized formalism. In this formulation one can get a state-particle with negative muss2, i.e. tachyons. The simplest case appears in 26 dimensional bosonic string theory. This approach is, unfortunately, not suitable for testing tachyon‘s solutions,2 but there are supersting theories defined in (9+1) dimensions that have tachyon free closed string spectrum. In addition, some string theories contain open string excitations with appropriate boundary conditions at the two ends of the string. So, one can ask: is there a stable minimum of the tachyon potential around which it is possible to quantize the theory. In the last few years there many papers devoted to this problem have made some progress, but we will not consider them in this paper. In padic string theory all tree level amplitudes involving tachyons in the external states can be computed. The padic (open) string theory is obtained from ordinary bosonic (open) string theory on a D-brane by replacing the integrallg over the real world-sheet coordinates by padic integral associated with a prime number p. There have been somewhat different approaches, but we will not consider the constructions of all these theories. Let us see the exact effective action for the ptachyon field. It is described by the lagrangian21
This form, obtained by computing Koba-Nielsen amplitudes for a prime p, makes sense for all (integer) values of p . The classical equation of motion derived from (15) is
Besides the trivial constant solutions $J = 0,1, a soliton solution is admitted. The equations separate in the arguments and for any spatial
203
direction x we get
a gaussian lump whose amplitude and spread are ~ o r r e l a t e d . ~
5. Quantum Mechanical Analogue of Tachyon Matter Now, we will concentrate on a relatively new field theory - the field theory of tachyon matter was proposed by Sen a few years ago.2o The derivation of its action is based on a rather involved argument. The obtained form is pretty strange and different from the actions we used to be familiar ones
Let us consider padic analogue of the above action, originally considered as real one, i.e. 700 = -1, ,,v = ,S where p , v = 1,2, ..., n. T ( x ) is a padic scalar tachyon field and V ( T )is the tachyon potential: V ( T )= exp(-crT/2). In the bosonic case, n = 25, cr = 1, and for superstring n = 9, Q = fi.The square root appearing in action (18) (and its multiplication to tachyon potential) makes this theory so unusual. Here we examine a lower (zero-dimensional) mechanical analogue of the field theory of padic tachyon matter (whatever it would physically mean). As usually, the correspondence can be obtained by the correspondence xi + t , T + x , V ( T )+ V ( x ) .The corresponding zero-dimensional action reads
so = -
I
dtV(X)drn,
(19)
where integration has to be performed over padic time. From the above action it is not difficult to get the classical equation of motion
where function f ( x ) denotes
) . ( f
=
---.1 d V
v dx
Partial differentiation of padic valued function is well defined, although in this case it can be replaced by the ordinary one, because V = V ( x ) .
204
Keeping in mind that exponential padic function (the tachyon potential V ( z )= exp(-ax) should be understood as an analytic function with corresponding radius of convergenceg r l/p, we obtain as in the real case N
?+a* 2 = a .
By the replacements motion
j. = ~ y a, y =
mji
6 and
(21) = g we get the equation of
+ pg2 = mg,
(22) which describes motion of a particle of mass m moving in a constant (say gravitational, Newtonian) field with quadratic friction. It is interesting that this equation can be derived from the (padic) action
Surprisingly or not, the zero dimensional analog of the (Sen's) field theory of tachyon matter offers an action integral formulation for the system under gravity in the presence of (quadratic) damping. The solution of the equation of motion (22) reads Y = y o + - lm n(
2P
:),
g-Pv2 g-,v
with initial t = 0 conditions for position y(0) = yo and velocity w(0) = wo. This solution has the same form in the real and padic case, but the radius of convergence is rather different. Faced with the increasing interest in various aspects of tachyon field theory, including its padic aspect, this connection with the field theory of tachyons thorough action integral formulation seems worth mentioning and examining in general. Also, quantization of the theory in path integral language might be very useful and, as we know, very general (for real, p adic and adelic path integrals see, e.g. Ref. 17). However, a kernel of the operator of evolution that corresponds to the action (23) is still unknown, even in the real case. Because of that the square root and exponential for small P should be expanded. If we treat p padicaly small, in respect to padic norm, we obtain
205
We have already calculated the path integral for the particle in constant external field.14 Here we have a slightly changed form (y” = Y(T), = Y(O), h = 1)
It makes it possible to check the existence of the simplest tachyonic vacuum state (invariant in respect t o the evolution operator), of the corresponding quantum mechanical model, i. e.
1
Kp(Yll,7; d ,O)fl(lY’lp)dY’. (27) IY’ll Using (5) and (26) we find that for the existence of the “ground“ (52) state of (quantum) padic tachyons (here some technical details and case p=2 are omitted) the following is necessary 5 1 for 5
52(lY”lp) =
1, or 12y” -
$1
5 1 for P
I &l p >
15
(s) I I5I P
1. Possible physical implication on
constraints for quantities related to the starting tachyon action (18) will be discussed elsewhere. The existence of R state opens the “door“ for further adelic generalization and investigation of higher-excited states. As in the real case8 a quadratic damping effect could enter explicitly into the play treating it as a perturbation over classical solution of the equation for a particle in constant external field without friction. Damping effect would be ascertained and understood through its dependence of ,B term.22 6. Conclusion
In this paper we show that quantum mechanical simplification of the tachyon field theory, besides the real case, is possible in a padic context. Also, an adelic generalization looks possible, i. e. without some obvious principal obstacles. Path integral formulation of zero-dimensional padic tachyons has been done and some ”minimal” conditions for their existence have been found. Of course, how much this approach could be useful for deeper understanding of the whole string theory and of its tachyon sector requires time and further, in-depth research. The fact, that the exact effective tachyon action in the usual string theory is not known, while in
206
padic string theory it is, is quite enough motivation for this and similar and investigation. We would propose a few promising lines for further investigation. The exact formula for quadratic quantum padic systems in two and more dimensions17 could be useful for multidimensional generalization of padic tachyons. It is tempting to extend our approach t o 1 1 dimensional field theories, even nonlinear field theories would be here quite nontrivial problem. Finally, padic string theory could be a very useful guide to difficult question in the usual string theory. It requires deeper understanding of padic string theory itself, especially of closed padic strings (strings on p adic valued worldsheet and target space as well). It is a worthwhile task to explore padic strings in nontrivial backgrounds. It will naturally lead to noncommutative formulation on padic quantum theory and examination of the corresponding Moyal-product, introduced in a context of the Noncommutative Adelic Quantum Mechanics.23 Recently, the Moyal-product has been applied in the calculation of the Noncommutative Solitons in padic string theory.24
+
Acknowledgments The research of both authors was partially supported by the Serbian Ministry of Science and Technology Project No 1643. A part of this work was completed during a stay of G. Djordjevic at LMU-Munich supported by DFG Project "Noncommutative space-time structure-Cooperation with Balkan Countries". Warm hospitality of Prof. J. Wess is kindly acknowledged.
References 1. I.V. Volovich, Theor. Math. Phys. 71 (1987) 337. 2. A. Sen, Tachyon Dynamics in Open String Theory, hepth/0410103. 3. D. Ghoshal and T. Kawano, Towards p-Adic String in Constant B-Field, hepth/0409311. 4. G. S. Djordjevic, B. Dragovich, Lj. D. Nesic and I. V. Volovich, Int. J.Mod.Phys. A17 (2002) 1413. 5. D. Ghoshal and A. Sen, N d P h y s . B584 (2000) 300. 6. A.A. Gerasimov and S.L. Shatashvili, JHEP 0010 (2000) 034. 7. A. Sen, Znt. J . Mod.Phys. A18 (2003) 4869. 8. S. Kar, A simple mechanical analog of the field theory of tachyon matter, [hep-th/0210108],
207
9. V. S. Vladimirov, I. V, Volovich and E. I. Zelenov, p-Adic Analysis and Mathematical Physics, World Scientific, Singapore 1994. 10. E. I. Zelenov, J. Math. Phys. 3 2 (1991) 147. 11. D. Dimitrijevic, G. S. Djordjevic and LJ. Nesic, Fourier Transformation and Pseudodifferential Operator with Rational Part, Proceedings of the Fifth General Conference of the Balkan Physical Union BPU-5, Vrnjacka Banja, Serbia and Montenegro, August 25-29, (2003) 1231. 12. I. M. Gel’fand, M. I. Graev and I. I. Piatetskii-Shapiro, Representation Theory and Automorphic Functions (Saunders, London, 1966). 13. R. P. Feynman, Rev. Mod. Phys. 20 (1948) 367; R. P. Feynman and A. Hibbs, Quantum Mechanics and Path Integrals, McGraw Hill, New York, 1965. 14. G. S. Djordjevik and B. Dragovich, O n p-Adic Functional Integration, in Proc. of the I1 Math. Conf. in PriStina, PriStina (Yugoslavia) (1996) 221. 15. G. Djordjevik, B. Dragovich, Mod. Phys. Lett A12 (1997) 1455. 16. G.S. Djordjevic, B. Dragovich and Lj. Nesic, p-Adic Feynmads Path Itegrals, Int. Conference: FILOMAT 2001,Nis, 26-29 August 2001. FILOMAT 15 (2001) 323. 17. G. S. DJordjevic and LJ. Nesic, Path Integrals f o r Quadratic Lagrangians in Two and More Dimensions, Proceedings of the Fifth General Conference of the Balkan Physical Union BPU-5, Vrnjacka Banja, Serbia and Montenegro, August 25-29, (2003) 1207. 18. G. Djordjevic, B. Dragovich and Lj. Nesic, Infinite Dimensional Analysis, Quantum Probability and Related Topics 6 (2003) 179. 19. P. G. 0. Freund and E. Witten, Phys. Lett. B199 (198’7) 191. 20. A. Sen, JHEP 0204 (2002) 048. 21. L. Brekke, P.G.O. Freund, M. Olson and E. Witten, Nucl. Phys. B302 (1988) 365. 22. G.S. Djordjevic and Lj. NeSic, Quantum p-tachyons, work in progress. 23. G.S. Djordjevic, B. Dragovich and Lj. Nesic, Adelic Quantum Mechanics: Nonarchimedean and Noncommutative Aspects, Proceedings of the NATO ARW ”NONCOMMUTATIVE STRUCTURES IN MATHEMATICS AND PHYSICS”, Kiev, Ukraine, September 2000, Eds. S. Duplij and J. Wess, Kluwer. Publ. (2001) 401-415. 24. D. Ghoshal, JHEP 0409 (2004) 041.
SKEW-SYMMETRIC LAX POLYNOMIAL MATRICES AND INTEGRABLE RIGID BODY SYSTEMS
V. DRAGOVIC AND B. GAJIC Mathematical Institute SANU Kneza Mihaila 35, 11 000 Belgrade, Serbia and Montenegro E-mail: [email protected], [email protected] Skew-symmetric matrix Lax polynomials are considered. Few rigid body systems with such representations are presented. Lagrange bitop is completely integrable four-dimensional rigid body system. Algebro-geometric integration procedure, using Lax representation, for that system is performed. This integration is based on deep facts from the theory of Prym varieties such as the Mumford relation and Mumford-Dalalyan theory. Class of isoholomorphic integrable systems is established and importation class of its perturbations is observed, generalizing classical Hess-Appel’rot system.
1. Introduction
Lax representation is one of the most powerful tools in the modern theory of integrable systems. The Lax equation
,with L(X),A(X) being matrix polynomials in so called spectral parameter X were studied by different authors in the middle of seventies. Approach, based on the Baker-Akhiezer functions, is developed by Dubrovin (see Ref. 1). Different version of that theory, based on Ref. 2 is presented by Adler and van Moerbeke in Ref. 3. Both of the theories were applied to the rigid body motion in Refs. 4,5, 3, 6 and 7. Recently in Refs. 8 and 9 we have found Lax representation for the classical Hess-Appel’rot system of rigid body, and for a new completely integrable four-dimensional rigid body system that we called Lagrange bitop. It appears that in both cases matrices L and A are skew-symmetric. This was not the case in Manakov top.4 In the case of Lagrange bitop matrix 208
209
L(X) satisfies (in proper basis) Ll2 = L2l = L34 = L43 = 0,
(1)
which is excluded in both theories (Dubrovin’s and Adler-van Moerbeke’s). Analysis of the spectral curve and the Baker-Akhiezer function shows that dynamics of the system is related to certain Prym variety ll (which splits according to the Mumford-Dalalyan theory and evolution of divisors of some meromorphic differentials 0;. Then the condition (1) requires that some of these differentials have to be holomorphic during the whole evolution. Compatibility of this requirement with dynamics put a strong constraint on the spectral curve: its theta divisor should contain some torus. The conditions (1)and holomorphicity conditions create a new situation from the point of view of the existing integration techniques. Such systems we call isoholomorphic systems (see Ref. 9). Recently, in Ref. 13 we have constructed families of partially integrable rigid body systems with skew-symmetric Lax matrices. These families represent higher-dimensional generalizations of the Hess-Appel’rot system in any dimension. They are certain perturbations of n-dimensional Lagrange top and Lagrange bitop. 2. Lagrange bitop
We will consider motion of a heavy n-dimensional rigid body fixed at a point. Equations on semidirect product so(n) x sa(n) in moving frame are introduced in Ref. 6
Ail = [ M R I + [F’X],
= [F’RI 7
(2)
+
where the matrix I is diagonal, diag(l1,. . . ,In).Here Mij = (Ii Ij)Rij E so(n) is the kinetic momentum, R E so(n) is the angular velocity, x E so(n) is a given constant matrix (describing a generalized center of the mass), r E so(n). Then Ii Ij are the principal inertia momenta. The Lagrange bitop is system defined by (see Ref. 9):
+
0 13 = 1 4 = b and ’‘=I2=’
x12
.=(-:2
0
with the conditions a
# b,
0
0
0
x3 O4)
-x34
0
# 0 , l x l ~# l 1x341.
x12,x34
(3)
210
Proposition 2.1. T h e equations of m o t i o n ( 2 ) under the conditions (3) have a n L - A pair representation L(X) = [L(X),A(X)] ,(see Refs. 8,9) where
L(X) = X2C+ AM and C = ( u
+ r,
A(X)
= XX
+ R,
(4)
+b ) ~ .
One can observe that both leading terms in the operators L and A (matrices C and x) are skewsymmetric, while in Refs. 1 and 14 C is symmetric and M is skew-symmetric. Before analyzing spectral properties of the matrices L(X),we will change the coordinates in order to diagonalize the matrix C. In this new basis the matrices L(X) have the form L(X) = U-lL(X)V,
The spectral polynomial p(X, p ) = det
(E(X) - p . 1) has the form
where
P(X) = AX4 + BX3 + DX2+EX
+ F,
&(A)
= GX4
+ HA3 + I X 2 + J X + K .
21 1
Their coefficients
A = C t 2 + C i 4 = (C+,C+)+(C-.,C-), B=2C34M34+2C12M12 = 2 ( ( C + , M + ) + ( C - , M - ) ) ,
D
=~ =
3 + 22c12r12 ~ + 2c34r34
, 2+, M ? ~+ M ; ~+ M ? ~+ ~
(M+,M+)+(M-,M-)+2((c+,r+)+(c-,r-)),
+ + 2r14w4 + 2 r 2 3 ~+223r 2 4 ~+224r 3 4 ~ 3 4 = 2 (F+, M+) + F-, M-)), F= + ri3+ r:4 + r;3 + r;4+ ri4= (r+,r+)+ (r-,r-), E
= 2r12~1 2r13w3 2
G = c 1 2 c 3 4 = (c+, c-),
+
H = c34M12 + c12M34 = (c+, M - ) (c-,M + ) , I = c34r12r34c12M12M34 M23M14 - M13M24 = (C+,F-) + (C-,F+) + ( M + , M - ) , J = ~~~r~~ + ~~~r~~ + ~~~r~~~~~r~~ - r 1 3 ~-2r42 4 ~ 1 3
+
+
+
+
W+W + w-,r+), r34r12 + r23r14 - r13r24 = (r+,r-). integrals of motion of the system (a), (3). =
K are
=
M+, M- E R3 which correspond to
Mij E
We used two vectors so(4) according to
(7)
Here M$ are the j-th coordinates of the vector M+. The system (2), (3) is Hamiltonian with the Hamiltonian function
1 2
= -(M13013+M14fl14+M23023+M12012 +M34034) Sx12r12 +x34r34*
The algebra so(4) x so(4) is 12-dimensional. The general orbits of the coadjoint action are 8-dimensional. According to Ref. 6, the Casimir functions are coefficients of A', A, A4 in the polynomials [ d e t z ( A ) ] 1 / 2and - I2T T ( L ( A ) )and ~ they are J , K , E , F . Nontrivial integrals of motion are B , D, H, I , and they are in involution. When 1x121 = 1x341, then 2H = B or 2H = -B and there are only 3 independent integrals in involution. Thus, for 1x121 # 1x341, the system (2), (3) is completely integrable in the Liouville
212
Ratiu introduced in Ref. 6 two families of integrable Euler-Poisson equations: the generalized symmetric case, defined by the conditions
x arbitrary;
I1 = . . . = I,,,
and the generalized Lagrange case is defined by I1 =
IZ = a, I3 = ... = I,,
= b, xij = 0 if ( i 7 j )
6 ((17 2)7 (271)).
The system (2), (3) doesn’t fall in any of those families and together with them it makes the complete list of systems with the L operator of the form
+
L ( A ) = A2C AM
+ r.
More precisely, if x 1 2 # 0 then the Euler-Poisson equations (2) could be written in the form ( 4 ) (with arbitrary C ) if and only if the equations (2) describe the generalized symmetric case, the generalized Lagrange case or the Lagrange bitop, including the case X ~ = Z f ~ 3 4 . ~ ~ ’ Starting from the well-known decomposition so(4) = so(3) CB so(3), introducing
1 1 M2 = -(M+ - M - ) M I = -(M+ + M - ) 2 2 (and similar for R, r,x),where M+, M- are defined with (7), equations (2) become
a1+ rl x xl), = 2(hf2 x a2+r2x x 2 ) ,
A& = 2 p 1 x
r1= 2(r1 x nl), r2= 2(r2 x 02),
(8)
and
As it is shown in Ref. 9 after changing of variables equations (8) can be written in the form: 2i; = Pi(?&),
i = 1,2,
P ~ ( u=) -4u3 - 4uZBi+ 4uCi + Di,
i = 1,2;
where the constants Bi, Ci, Di, i = 1,2 depends of the first integrals and the parameters of the system. For more details see Ref. 9. From the previous relations, we have
213
So, the integration of the system (8) leads to the functions associated with the elliptic curves El, E2, given with:
Ei : y2 = P ~ ( u )i, = 1 , 2 .
(9)
3. Properties of spectral curve The L(X) matrix (4) for the Lagrange bitop (2), (3) is a quadratic polynomial in the spectral parameter X with matrix coefficients. There are two general theories describing the isospectral deformations with matrix coefficients. The first one, based on the Baker-Akhiezer function is developed by Dubr0vin.l The other one, given by Adler and van Moerbeke in Ref. 3 was based on Ref. 2. Both of these theories were applied to the rigid body motion in Refs. 4, 5 and 3, 6, 7, respectively. As it is shown in Ref. 9, non of these two theories can be directly applied in our case. So, we are going to make certain modifications, and then we will integrate the system (2), (3). As usual in the algebro-geometric integration, we consider the spectral curve
I? : det i ( A ) - p . 1) = 0.
(
By using ( 5 ) , ( 6 ) , we have
r : p4+p2(A;, +
+ 4p3p3*+ 4p4p4*)+[A12A34+ 2i(p3*p4- p3~4*>i2 = 0.
There is an involution (T : (A, p ) -+ (A, - p ) on the curve r, which corresponds to the skew symmetry of the matrix L(X). Denote the factor-curve by rl = r/o.
Lemma 3.1. (a) The curve rl is a smooth hyperelliptic curve of the genus g(r1) = 3. The spectral curve is a double covering of r l . The arithmetic genus of is ga(r)= 9. (b) The spectral curve r has f o u r ordinary double points Si, i = 1,. . . , 4 . The genus of its normalization is five. (c) The singular points Si of the curve are fixed points of the involution (T. The involution o exchanges the two branches of I' at Si . In general, whenever matrix L(X)is antisymmetric, the spectral curve is reducible in odd-dimensional case and singular in even-dimensional case.
214
Before starting the study of the analytic properties of the BakerAkhiezer function, let us give the formulae for (nonnormalized) eigenvectors of the matrix L(A). An eigen-vector of L , i.e. such that L(A)f = pf is given by
f1 =
+ p2)(iA34
- p ) - 2p(P3P3*
f 2 = W P 3 - iP4)(iP$
+ P4P4*) + 2 A 1 2 ( @ 3 f l i - P 4 P ; ) ,
+Pi),
+ '$4) [(iAi2- P ) ( i A 3 4 - P ) + 2i(P3Pi P;P4)] , f4 = (iP: + P i ) [(iAi2+ P ) ( i A 3 4 - P ) -t2i(P3Pi - P;P4)] . f 3 = (-P3
-
Corollary 3.1. The eigenvectors f' normalized by the condition
f: + f; + f; + fi = 1, f',
have different values in the points Sl,S,!l E points SiE I?.
which cover the singular
Following ideas of Dubrovin and Krichever (see Refs. 15, 16, 1, and bibliography therein) , we consider the next eigen-problem
where $k are the eigenvectors with the eigenvalue p k . Then $ k ( t , A) form 4 x 4 matrix with components $ i ( t ,A). Denote by cp: the corresponding inverse matrix. Let us introduce gj(t1 (A, P k ) )
=
$i(t,A)
'
$(t, A)
or, in other words g ( t ) = $ k ( t ) 18 p(t)k. Matrix g is of rank 1,and we have d$/at = --A$, dcp/dt = cpA, dgldt = [ g ,A].w e can consider vector-functions $ k ( t , A) = ($i(t,A), ...,$:( t , A ) ) ~ T
as one vector- function $(t,(A, p ) ) = ($'(t, (A, p ) ) , ...,7,b4(t,(A, p ) ) ) on the curve r defined with lCli(t,(A, p k ) ) = $ i ( t , A). The same we have for the matrix 9:. The relations for the divisors of zeroes and poles of the functions $2 i cpi in the afine part of the curve I? are:
(g;), = d j ( t )
+ dZ(t) - D,
-
D$,
(10)
where D, is the ramification divisor over X plane (see Ref. 1) and D , is the divisor of singular points, D: 5 D,. One can easily calculate deg D , = 16, degD, = 8.
215
Lemma 3.2.
(a) The matrix g has a representation
+
+
where a1 = L , a2 = P E L 2 ,a3 = P L L3. (b) For the Lax matrix L and for X i such that &(Xi) = 0 , it holds a3 = 0. (c) The matrix g has no poles at the singular points of the curve F. From now on we will consider all the functions in this section as functions on the normalization ? of the curve F. The matrix elements gj(t,(A, P k ) ) are meromorphic functions on the curve I?. We need their asymptotics in the neighborhoods of the points Pk, which cover the point X = 00. Let be the eigenvector of the matrix z ( X ) normalized in Pk by the condition = 1, and let be the inverse matrix for We will also use another decomposition of matrix elements of g : 9;.= dJ:p$ = $i+$.It is an immediate consequence of the proportionality of the vectors dJk and (and p k and g k ) . Using asymptotics of the functions i g$ in neighborhoods of the points Pk we getg that the divisors of matrix elements of g are
'& 4;
4:.
+:
I .
&
(9;) = 2
4;
+ J j - D, + 2 (PI + P2 + P3 + P4)
-
where the divisors c&, & are of the same degree deg & thet are given with:
+ P2, 2' = d1 + P2,
21 = d1
+ Pi, 2' = d2 + P i ,
2 2 = d2
Pa - Pj,
=
deg
&
=
5 and
+ P4, 2 4 = d4 + P3, J3 = d3 + P4, J4 = d4 + P3. 23
= d3
Let us denote with @ ( tA) , the fundamental solution of
(g+
A ( h ) ) q t , A)
= 0,
normalized with Q ( T ) = 1. Then, if we introduce functions
'$(t,T,(A,
Pk))
@ t ( hA)h"(T,(A, P k ) ) ,
= S
where h" are the eigenvectors of L(A) normalized by the condition h s ( t ,(A, p k ) ) = 1, it follows that
c,
216
Proposition 3.1. Ref. 9)
The functions
Gi satisfy the following properties (see Gi
(a) I n the a f i n e part o f f ' the function has 4 time-dependent zeroes which belong to the divisor d i ( t ) defined by formula (lo), and 8 time-independent poles, e.q.
(Gi(t,T , (A, p k ) ) ) a = d i ( t ) - 9, (b) At the points follows:
4 , the functions
di@,
7, (A,
4 2
degV = 8.
have essential sangularities as
P ) ) = exp [-(t -
G i ( t ,7, (A, P ) )
where Rk are given with
and di are holomorphic in a neighborhood of Pk,
Gi(~,7,(X,p)) = h i ( 7 , ( A , p ) ) , di(t,7,Pk) = S t + v : ( t ) z + 0 ( z 2 ) , with
Observeg that the following relation takes place on the Jacobian Jx(f'):
+
A(dj(t) odi(t))= A(dj(7)
+odj(7))
where A is the Abel map from the curve f' to Jac(f'). Thus, the vectors A ( d i ( t ) ) belong to some translation of the Prym variety Il = Pryrn(f'lr1). More details concerning the Prym varieties one can find in Refs. 10, 17 and 12. The natural question arises to compare two two-dimensional tori Il and El x Ez, where the elliptic curves Ei are defined in (9). 4. Geometry of the Prym variety II and Mumford's relation
Lemma 4.1.
The curves Ei are Jacobians of the curves
C1 :
v2 = P(A) + Q(A), 2
C2 : v2 =
Ci
-- &(A). 2
given by:'
217
Theorem 4.1. (For details see Ref. 9) (a) The Prymian II is isomorphic to the product of the curves Ei:
n = Jac(C1) x Jac(C2). The curve f' is the desingularization of rl x p CZ and C1 x p r l .
(b) (c) The canonical polarization divisor E of II satisfies
E = El x
0 2
+ 0 1 x Ez,
where 0i is the theta - divisor of Ei. Theorem 4.1 is based on Mumford-Dalalyan theory and it gives the connection between the curves E l , E2 and the Prym variety II. The Baker - Akhiezer function Q satisfies usual conditions of normalized 4-point function on the curve of genus g = 5 with the divisor 2, of degree d e g B = 8, see Refs. 16, 14. It was proven by Dubrovin in the case of general position, that 0; = g i j d X , i = 1, ..., 4 are a meromorphic differentials having poles at Pi and Pj, with residues vj and -v! respectively. But here we have that the four differentials O i l Oq, fli, 0: are holomorphic during the whole e v o l ~ t i o n . ~ We can say that the condition (1) implies isoholomorphicity. Let us recall the general formulae for v from14
where U = C X ( ~ ) U is( ~ certain ) linear combination of b periods U ( i )of the differentials of the second kind Og) , which have pole of order two at Pi; X i are nonzero scalars, and
(Here v is an arbitrary odd non-degenerate characteristics.) Thus, from
vz1 = 2112 = v43 = 2134 = 0 ,
(12)
using (11)we get that holomorphicity of some of the differentials 0; implies that the theta divisor of the spectral curve contains some tori (see ref. 9). In a case of spectral curve which is a double unramified covering
T : Fj r l ;
218
with g(r1) = g, g(p) = 2g - 1, as we have here, it is really satisfied that the theta divisor contains a torus, see Ref. 10. More precisely, following," let us denote by II- the set
IT-
= L E Pic2g-2?INmL = Krl, ho(L)is odd},
{
where Krl is the canonical class of the curve l?l and N m : Picf' -+ P i c r l is the norm map, see Refs. 10, 12 for details. For us, it is crucial that ITis a translate of the Prym variety II and that Mumford's relationlo holds
II- c QF. Complete formulae for Baker-Akhiezer function are given in Ref. 9. 5 . Conclusion
New family of rigid body systems with skew-symmetric Lax matrices have been constructed recently in Ref. 13. These systems are higher-dimensional generalizations of Hess-Appel'rot system constructed in any dimension. They are certain perturbations of higher-dimensional Lagrange tops. The four-dimensional Hess-Appel'rot system is perturbation of Lagrange bitop and belongs to the class of isoholomorphic systems. Thus, for the integration of the system we need to apply approach given it this paper.
Acknowledgement The research of both authors was partially supported by the Serbian Ministry of Science and Technology Project No 1643.
References B. A. Dubrovin, Funk. Analiz i ego prilozheniya 11,28 (1977 [in Russian]). P. van Moerbeke and D. Mumford, Acta Math. 143,93 (1979). M. Adler and P. van Moerbeke, Advances in Math. 38,318 (1980). S. V. Manakov, Funkc. Anal. Appl. 10, 93 (1976) [in Russian]. 0. I. Bogoyavlensky, Soviet Acad Izvestya 48,883 (1984) [in Russian]. T. Ratiu, American Journal of Math 104,409 (1982). T. Ratiu and P. van Moerbeke, 32,211 (1982). V. DragoviC, B. GajiC: An L-A pair for the Hess-Apel'rot system and a new integrable case for the Euler-Poisson equations on so(4) x so(4). Roy. SOC.of Edinburgh: Proc A 131,845 (2001). 9. V. DragoviC, B. GajiC, American Journal of Math., (2004), (to appear) 10. D. Mumford, A collection ofpapers dedicated to Lipman Bers (Acad. Press.) New York, 325 (1974). 1. 2. 3. 4. 5. 6. 7. 8.
219
11. S. G. Dalalyan, Uspekhi Math. Naukh 29, 165 (1974) [in Russian]. 12. V. V. Shokurov, Algebraic Geometry 111, 219 (Berlin: Springer-Verlag, 1998). 13. V. DragoviC, B. GajiC, ”Systems of Hess-Appel’rot type”, Preprint SISSA, (2004). 14. B. A. Dubrovin, Uspekhi Math. Nauk. 36,11 (1981) [in Russian]. 15. I. M. Krichever, Uspekhi Math. Naukh 32, 183 (1977). 16. B. A. Dubrovin, I. M. Krichever and S. P. Novikov, Dynamical systems l V , Berlin: Springer-Verlag, 173. 17. J. D. Fay, Lecture Notes in Mathematics vol. 352, Springer-Verlag, (1973).
SUPERSYMMETRIC QUANTUM FIELD THEORIES
D. R. GRIGORE Department of Theoretical Physics, Institute of Atomic Physics, Bucharest-Miigurele, Romcinia E-mail: [email protected] We consider some supersymmetric multiplets in a purely quantum framework. A crucial point is to ensure the positivity of the scalar product in the Hilbert space of the quantum system. For the vector multiplet we obtain some discrepancies with respect to the literature in the expression of the super-propagator and we prove that the model is consistent only for positive mass. The gauge structure is constructed purely deductive and leads to the necessity of introducing scalar ghost superfields, in analogy t o the usual gauge theories. Then we consider a supersymmetric extension of quantum gauge theory based on a vector multiplet containing supersymmetric partners of spin 3/2 for the vector fields. As an application we consider the supersymmetric electroweak theory. The resulting self-couplings of the gauge bosons agree with the standard model up to a divergence.
1. Introduction
The supersymmetric gauge theories are constructed using the so-called vector supersymmetric multiplet.’ The justification for this choice comes from the analysis of the unitary irreducible representations of the N = 1 supersymmetric extension of the Poincark group; there are two irreducible massive representations R l p N [m,01 @ [m,1/21 @ [m,1/21 @ [m,I], and 01 [m,1/2]@[m,11@[m, 11 @ [m,3/21, containing a spin 1 system; here [m,s] is the irreducible representation of mass m and spin s of the Poincark group.) The standard vector multiplet is constructed such that the oneparticle subspace of the Fock space carries the “simplest” representation
-
%/2.
We use entirely the quantum f r a m e ~ o r k ~avoiding -~ the usual approach based on quantizing a classical supersymmetric theory. In this way we do not have the complications associated to the proper mathematical definition of a super-manifold and we do not need a quantization procedure. A rigorous treatment of the perturbative aspects of these models can be pro220
221
vided using the Epstein-Glaser framework. We start this program analyzing the layout of the model, that is the construction of the quantum multiplet, its gauge structure and the expression of the interaction Lagrangian (or,in the language of perturbation theory, the first order chronological product). The main results are the following. The vector model is consistent only for positive mass. The origin of this result comes from the condition of positivity of the scalar product in the Hilbert space. In the standard literature one constructs the Hilbert space by applying polynomials in the free fields of the model on the vacuum. (In the language of the reconstruction theorem from axiomatic field theory this amounts to the construction of the Borchers algebra). However, this is not enough: one needs to provide the expression of the scalar product and prove that it is positively defined. The comparison with the literature dedicated to this subject shows that this point has not been analyzed. The structure of the Hilbert space of the model is never described explicitly. One can infer it only from the Feynman rules, more precisely the expressions of the (super)propagators. The point is that the assumption that the BRST quantization procedures gives automatically a positively defined Hilbert space structure is wrong. An explicit check should be made and this leads to some conditions on the mass of the supersymmetric multiplet. (For another point of view concerning supersymmetric positivity see Ref. 5). We can determine the gauge structure of this model in a deductive way: it coincides essentially with the expression from the literature but, because the mass of the multiplet is positive, we need to introduce some scalar ghost superfields. We are also able to determine the general expressions for the Feynman super-propagators in a purely deductive way; some discrepancy with the standard literature appears. Similar results are available for extended supersymmetries.6 Next we construct a new vector multiplet based on the representation R1. This multiplet exists for all values of the mass, its gauge structure is very similar to the gauge structure of the usual gauge theories and the interaction Lagrangian can be obtained by simply replacing fields by superfields in the Lagrangian of the standard model. In this way we obtain a supersymmetric extension of the standard model. One can conclude that the new vector multiplet proposed for the first time in Ref. 4 and based only on chiral superfields is a more natural object and it remains as a serious candidate for a possible supersymmetric extension of the standard model.
222
2. Quantum Supersymmetric Theory We remind here the definition of a N = 1 supersymmetric theory in a pure quantum context. Suppose that we have a quantum theory of free fields; this means that we have the following construction: (a) H is a Hilbert space of Fock type (associated to some one-particle Hilbert space describing some choice of elementary particles) with the scalar product (., .); (b) R E H is a special vector called the vacuum; (c) U a ,is~a unitary irreducible representation of inSL(2,C) the universal covering group of the proper orthochronous Poincark group such that a E R4 is translation in the Minkowski space and A E SL(2,C); (d) b j , j = 1,.. . ,NB (resp. f ~ ,A = 1,.. . , N F ) are the quantum free fields of integer (resp. halfinteger) spin. We assume that the fields are linearly independent up to equations of motion and that the equations of motion do not connect distinct fields. In practice, one considers only particles of spin s 5 2. For the standard vector model we consider only the case s 5 1. For the new vector model we consider the more unusual case 1 5 s 5 3/2. If one considers higher-spin fields (more precisely s 2 l), as we have done in Refs. 4, 7, it is necessary to extend somewhat this framework: gauge fields must be considered and we will describe them in the indefinite metric approach (Gupta-Bleuler). That is, we assume the existence of the following objects: (a) A gauge charge operator Q verifying Q 2 = 0 ; (b) A non-degenerate sesqui-linear form < . > which becomes positively defined when restricted to a factor Hilbert space K e r ( Q ) / l m ( Q ) Q ; is called gauge charge and we denote by At the adjoint of the operator A with respect to < ., . >; (c) Hphys = K e r ( Q ) / l m ( Q )will be the physical space of our problem; (d) The interaction Lagrangian t ( x ) is some Wick polynomial acting in the total Hilbert space H and verifying the conditions a,
[Q,t(x)l=ia#’(x)
(1)
for some Wick polynomials t”(x). This condition guarantees that the interaction Lagrangian t ( x ) factorises to the physical Hilbert space K e r ( Q ) / l m ( Q )in the adiabatic limit, i.e. after integration over x; the condition (1) is equivalent to the usual condition of (free) current conservation. The condition (1) has far reaching physical consequences: under some reasonable additional assumptions one can prove that the usual expression of the interaction Lagrangian for a Yang-Mills model is unique, up to trivial terms. It is desirable to generalize this scheme to supersymmetric theories. We note here that our framework is different from the usual treatment of quantum gauge invariance based on the construction of a classical
223
field theory with BRST invariance, some quantization procedure and the quantum action principle imposed on the Green functions. It seems that for ordinary gauge theories the two formalisms give identical results. But this is not the case for the supersymmetric gauge theories. This makes our alternative approach far from being superfluous. Now we define the notion of supersymmetry invariance of the system of Bosonic and Fermionic fields considered above. Suppose that in the Hilbert space H we also have the operators Q a , a = 1 , 2 such that: (i) the following relations are verified:
[&a, P p ]
UiIQaUA = AabQb ,
= 0,
(4)
here Pp are the infinitesimal generators of the translation group and Q b (&b)t; (ii) The following commutation relations are true: i [ Q a , b ] =p(a)f >
{&a,
f}= q(a)b,
=
(5)
where b (resp. f ) is the collection of all integer (resp. half-integer) spin fields and p , q are matrix-valued polynomials in the partial derivatives 8, (with constant coefficients). These relations express the tensor properties of the fields with respect to (infinitesimal) supersymmetry transformations. If these conditions are true we say that Qa are super-charges and b, f are forming a supersymmetric multiplet. As emphasized in Ref. 4, the matrix-valued operators p and q are subject to various constraints: ( 1 ) From the compatibility of (5) with Lorentz transformations it follows that these polynomials are Lorentz covariant. (2) Next, we start from the fact that the Hilbert space of the model is generated by vectors of the type 9 = b(zp) f(x,) R E H . It follows that the relations (3) are true if the left hand sides commute with all the fields of the model. Using the (graded) Jacobi identities we obtain:
n
n
[&a, [ Q b , W ] ]
[&a,
[QG,
w ] l + [&a,
[&a,
= -(a
~ 1 =1 -2i
++
b) ,
apw;
7
(6)
here w = b , f and [.,.I is the graded commutator. (3) The equation of motion are supersymmetric invariant. (4) The (anti)commutation relations have the implication that one and the same vector from the Hilbert space H can be expressed in distinct ways. This means that the supercharges are
224
well defined via (3) iflsome new consistency relations are valid following again from graded Jacobi identities; the non-trivial ones are of the form:
[b(z),{ f ( ~ 7 &all ) = -{f(y)
[&a, b(z)l}
.
(7)
(5) If a gauge supercharge Q is present in the model, then it is usually determined by relations of the type (5) involving ghost fields also, so it means that we must impose consistency relations of the same type as above. Moreover, it is desirable to have
{Q,Qa) = 0 , {Q,Q E } =
(8)
0;
this implies that the supersymmetric charges Qa and Qa factorizes to the physical Hilbert space Hphys = K e r ( Q ) / l m ( Q ) . These new consistency relations are of the type ( 6 ) with one of the supercharges replaced by the gauge charge: {&a,
[Q,bl}
= - {Q,[&a, bl>
>
[&a,
{Q,f}] = - [Q,{ & a , f}]
*
(9)
( 6 ) A relation of the type (7) must be also valid for the gauge charge:
[b(z),{ f ( ~Q)1 ) , = -{f(~),[Q,b ( z ) l ) .
(10)
(7) To have Q2 = 0 we must also impose
{ Q , [Q,bl)
=0
[Q,{ Q ,f)] = 0 .
(11)
(8) In the presence of a gauge structure one can relax the condition (3). Indeed, every operator A which (anti)commutes with the gauge charge Q factorizes t o an operator 7r(A) : K e r ( Q ) / l m ( Q ) -+ K e r ( Q ) / l m ( Q ) . In particular, the supercharges anticommute with Q by construction and Pp also commute with Q. So, we can impose Eq. (3) only on the physical Hilbert space7 i.e. we replace Qa -+ 7r(Q,), QE -+ 7 r ( & ) , Pp -+ r(P,>. All these conditions are of pure quantum nature i.e. they can be understood only for a pure quantum model. Some of them do not have a classical analogue so we can interpret the obstacles in constructing supersymmetric quantum models (associated to some classical supersymmetric theories) as some quantum anomalies. It seems to be an essential point to describe supersymmetric theories in superspace. We do this in the following way. We consider the space HG = G 8 H where G is a Grassmann algebra generated by Weyl anticommuting spinors 8, and their complex conjugates & = (8,). and perform a Klein transform such that the Grassmann parameters 0, are anticommuting with all Fermionic fields, the supercharges and the gauge charge.
225
The field operators acting in HG are called superfields. Quite generally one associates to every Wick polynomial w (x) its supersymmetric extension W = S(W) according t o the formula:
W ( x ,8,e) E exp ( i P Q , - i6'Q8)
W(Z)
exp (-ida&,
+ ie"Qa) ,
(12)
and we interpret the exponential as a (finite) Taylor series. Of special interest are the superfields constructed as in Refs. 2, 8 according to the preceding formulz taking w = b, f . It is a remarkable fact that only such type of superfields are really necessary, so in the following, when referring to superfields we mean expressions given by Eq. (12). We will call them super-Bose and respectively super-Fermi fields. For convenience we will denote frequently the ensemble of Minkowski and Grassmann variables by = @,8,s). Moreover, one postulates that the interaction Lagrangian t should be of the form
x
(13) for some supersymmetric Wick polynomial T . This hypothesis makes possible the generalization of the Epstein-Glaser approach to the supersymmetric case as it is showed in Ref. 4. Concerning the gauge invariance of the model there are two possible attitudes. One is to impose only Eq. (1); this "minimal" possibility is certainly consistent from the physical point of view but in general one loses the unicity results concerning the interaction Lagrangian. One can hope to keep this unicity result if one finds out a supersymmetric generalization of Eq. (1). A natural candidate would be the relation:
[Q,T ( X ) ]= ia,T.(X)
+ .. . ,
(14)
where by . . . we mean total divergence expressions in the Grassmann variables. It is clear that Eq. (14) implies Eq. (1) but not conversely. We call Eq. (14) the condition of supersymmetric gauge invariance because it involves only superfields in contrast to Eq. (1). In Ref. 4 we have showed that the stronger condition (14) can be imposed for the 01-vector model and indeed the unicity argument concerning the interaction Lagrangian holds. However for the Ol,z vector model the situation is not so good. If one uses only the LLminimal" gauge invariance condition (1) then one loses the unicity of the interaction Lagrangian. (However some other limitations might come from the condition of gauge invariance in higher orders of perturbation theory.) If one tries to impose the supersymmetric version
226
(14) one finds out that the usual expressions for the interaction Lagrangian suggested by the literature do not fulfill it. Of course, it is in principle possible to find alternative expression for the interaction Lagrangian such that Eq. (14) is true, but this possibility seems to be rather unprobable. So our results concerning the construction of the interaction Lagrangian for the R 1 p vector model must be considered as a criticism of the traditional approaches. The construction of the ghost and anti-ghost multiplets (which are needed in the construction of supersymmetric gauge theories) is done by inverting the spin-statistics assignment: the integer (resp. half-integer) spin fields have Fermi-Dirac (resp. Bose-Einstein) statistics.
3. The Vector Multiplet By definition, the vector multiplet has the following content: the Bosonic fields are some (real) scalars b ( j ) , j = 1,.. . ,s and a real vector field b,; the Fermionic fields are some Majorana fields of spin one-half f(A), A = 1 , . . . ,f . Let us consider that C is one of the scalar fields b(j) (or a linear combination of them). Now we define the following superfield: V E s(C) using formula (12); it is clear that one has the reality condition V t = V. Moreover the generic expression of V must be V(X,
e,B) = C ( X ) + ex(.) + ~z(x)+ e2 4 ( ~+)e-2 4t (x) +(e&7) .,(x) + e2 E ~ ( x+) e2 ex(^) + e2g2 d ( ~ ) ,
(15)
where the fields appearing in the expansion must linear combinations of the basic fields taking into account Lorentz properties and statistics. The action of the supercharges on the components of the multiplet is given in a compact way if we introduce the covariant derivative operators
acting on superfields. Then one imposes
i [ ~ V ( X~, 8,,
e)]= ~ , v 8,( e)~, ,
i[Q,,v
( e,8)] ~ , = D , V ( ~ ,e , e )(17)
and determines the action of the supercharges on the component fields by expanding both sided in the Grassmann variables. The explicit formulae are given in Ref. 7 and are compatible with the Jacoby identities (6); they also coincide with the formulae from the l i t e r a t ~ r e ~ after ~ ' ~some field redefinitions.
227
If C verifies the Klein-Gordon equation (for mass m ) , then the superfield V verifies the Klein-Gordon equation so all the components of the multiplet are verifying Klein-Gordon equation of mass rn. These equations are compatible with the supersymmetry action, i.e., they are left invariant by the supercharges Qa and Q f i . The multiplet (C,4l ZI,, d, A, x a )is irreducible; in particular it follows that the indices j and A take four values, C and d are real scalar fields of mass m, b, is a real vector field of mass m, 4 is a complex scalar field of mass m and xa and A, are Dirac fields of mass m. We now determine the supercommutator of the vector field. Let us consider the causal commutator
[ V ( X l ) V(X2)I , = -iD(X1; X2) 1 .
(18)
The expression D ( X 1 ; X 2 ) is a distribution in the variables xj and a polynomial in the Grassmann variables B j , j = 1 , 2 and verifies the following properties: (a) it is Poincarh covariant; in particular it depends only on the difference 2 1 - 2 2 ; (b) it has causal support; (c) verifies Klein-Gordon equation; (d) verifies the Hermiticity condition: D(X1;X2) = -D(X2; X I ) ; (e) verifies the antisymmetry condition: D(X2;X I ) = -D(X1; X2); and (f) verifies the consistency condition:
+
(0: D2)D(X1;X2) = 0 .
(19)
All properties except (f) are immediate. The property (f) follows from the Jacobi identity [&a,
[V(X1)1V(X2)11+ XI), [V(X2)7&all
+ [V(X2),
[&a,
V(X1)ll = 0
(20) and formula (17). It is easy to prove that the generic solution of the conditions (a)-(f) of the preceding proposition is a linear combination of four expressions. The condition of positivity gives an important result, namely the vector multiplet exists only for m > 0 and also some constraints on the coefficients of the development. If we compare our results with the usual quantization procedure by means of path integrals we can find out the corresponding values of these coefficients and it turns out that the constraints are not verified. One can argue that some of the fields are gauged away in the Wess-Zumino gauge; however, Wess-Zumino gauge can be proved to be in conflict with the supersymmetric transformation rules. Now we mention the (quantum) gauge structure of the vector model. In ordinary quantum gauge theory, one gauges away the unphysical degrees of freedom of a vector field v, using ghost fields. Suppose that the vector field is of positive mass
228
m; then one enlarges the Hilbert space with three ghost fields u,ii, 4 such that: (a) all three are scalar fields; (b) the fields u,ii are Fermionic and 4 is Bosonic. (c) all them have the same mass m as the vector field. (d) the ut = u, ut - = --u;and (e) the Hermiticity properties are; q5t = 4, commutation relations are:
[dW,d(Y)l = -2
{u(x),qY)} = -i Om(a:- Y) 7 (21)
- Y) 7
and the rest of the (anti)commutators are zero. Then one introduces the gauge charge Q according to:
QR=O,
[Q,up] = i a p u ,
Qt=Q,
[Q,41 = i m u ,
{Q,u}= 0 , {Q,ii} = -2 (dpup + m 4). (22) It can be proved that this gauge charge is well defined by these relations i.e. it is compatible with the (anti)commutation relations. Moreover, one has Q2 = 0 so the factor space K e r ( Q ) / l m ( Q )makes sense; it can be proved that this is the physical space of an ensemble of identical particles of spin 1. For details see Refs. 11, 12. One can generalize this structure in the supersymmetric context if one introduces ghost superfields. One obtains essentially the expressions from the literature;" the main difference is due to the fact that we have m > 0. One can compare this gauge structure to the classical gauge structure;' there is no straightforward correspondence, but at least for ordinary Yang-Mills theories, one finds out that the analogy is rather miraculous. It is natural to try the same idea in a supersymmetric context. However, we have found a negative result (the details are in Ref. 7). There are other multiplets associated to the vector multiplet as the linear multiplet and the rotor multiplet. The algebraic structure of these multiplets is also clarified in Ref. 7, but there are obstacles in the construction of a consistent supersymmetric extension of the standard model also in these cases.
4. The New Vector Multiplet We will consider here a new vector multiplet which has the nice property that the corresponding gauge structure is similar to the usual gauge theories. If this model is consistent with the phenomenology it brings new physics. The new vector multiplet is obtained from the Wess-Zumino multiplet if we transform the scalar (resp. the spinor) field in a vector (resp. a Rarita-Schwinger) field wp,Gpa by simply adding the index p ; this multiplet is subject to the following consistency conditions: (a) Hermiticity
229
= Gpii; (b) all fields verify Klein-Gordon (resp. Dirac) equation of mass M . (c) the action of the supercharges can be obtained from the corresponding formulz for the Wess-Zumino multiplet:
[Qalupl = 0 ,
i [Qa,ut1 = 2
+pa
= CJl6avup ;
(23) and (d) (anti)commutation relations can be obtained starting from the commutation relation for the vector field u, and using the consistency relation (10). We call this new multiplet the RS vector multiplet. The associated superfield can be easily constructed using the supersymmetric extension formula. The gauge structure of this multiplet can be obtained if we replace fields by superfields v, + v,, q5 + @, u U, G + 0 where Ul 0 are ghost and anti-ghost multiplets; all these multiplets are of the same positive mass M . One can obtain the action of the supercharge on the field components if one substitutes in the preceding relations the expressions of the superfields in terms of component the fields and the Grassmann variables. It also can be showed that, as for the usual gauge case, the factor space K e r ( Q ) / l m ( Q )describes a system of identical I;21 super-symmetric systems. So, the analogy with the usual gauge case is really remarkable. Moreover, it is quite easy to obtain consistent gauge invariant couplings if we substitute fields by superfields in the known formulz.12 One can prove4 that in this way a consistent supersymmetric extension of the standard model can be obtained. {&a, $pb}
=
EabMup
7
{Qal
$p6}
-
References 1. 2. 3. 4. 5. 6.
7. 8. 9. 10.
11.
S. Ferrara, 0. Piguet, Nucl. Phys. B93,261 (1975). F. Constantinescu, M. Gut, G. Scharf, Ann. Phys. 11,335 (2002). D. R. Grigore, European Phys. Journ. C21,732 (2001). D. R. Grigore, G. Scharf, Annalen der Physik 12,5 (2003). F. Constantinescu, Lett. Math. Phys. 62,111 (2002). D. R. Grigore, G. Scharf, Quantum Extended Supersymmetries, h e p th/0303176, to appear in Annalen der Physik (2003). D. R. Grigore, G. Scharf, Annalen der Physik 12,643 (2003). F. Constantinescu, G. Scharf, Causal Approach to Supersymmetry: Chiral Superfields, hep-th/0106090. P. P. Srivastava, Supersymmetry, Superfields and Supergmuity: an Introduction, IOP Publ. (1986). S. J. Gates Jr., M. T. Grisaru, M. RoEek, W. Siegel, Superspace or One Thousand and One Lessons in Supersymmetry, Cummings (1983), hepth/0108200. D. R. Grigore, Journ. Phys. A33, 8443 (2000).
230
12. G. Scharf, Quantum Gauge Theories: A Tme Ghost Story, Wiley (2001).
PARASTATISTICS ALGEBRAS AND COMBINATORICS
T. POPOV Institute f o r Nuclear Research and Nuclear Energy, Bulgarian Academy of Sciences, bld. Tsarigradsko chauske 72, 1784 Sofia, Bulgaria E-mail: [email protected]. bg We consider the algebras spanned by the creation parafermionic and parabosonic operators which give rise t o generalized parastatistics Fock spaces. The basis of such a generalized Fock space can be labelled by Young tableaux which are combinatorial objects. By means of quantum deformations a nice combinatorial structure of the algebra of the plactic monoid that lies behind the parastatistics is revealed.
1. Introduction Wigner' was the first to remark, back in 1950, that the quantum mechanical equations of motion allow for a commutation relations more general than the canonical commutation relations
A few years latter (in 1953) Green wrote the relations' now known under the name parastatistics commutation relations
Let us define the parafermi algebra p$ and the parabose algebra p B to be the associative algebra (over C ) generated by the creation at and annihilation ai operators with relations (l),(2) and (3) with the upper and lower sign, respectively. These are not all parastatistic relations, one obtains new ones by hermitean conjugation. We shall only consider quantum systems with finite number D degrees of freedom, i, j, k = 1 , . . . ,D. 23 I
232
The parastatistics commutation relations are generalization of the cannonical commutation relations
Indeed, the quadratic relations (4) imply the trilinear parastatistics relations ( l ) , (2) and (3). The representations of the parafermionic and p% parabosonic algebras relevant in physics are labelled by the non-negative integer number p , called the order of parastatistics. A creation operator u! acting on the vacuum state creates one particle state, and the annihilation operator ai annihilates it. One has
ps
uiu; 10) = p6ij 10)
.
(5)
The representation p = 0 is trivial (ui = ul = 0) and corresponds to the vacuum representation. The case p = 1 yields particular representation of (p%) in which the relations (4) hold and one retrieves the ordinary fermionic (bosonic) Fock space. For p 2 2 fermionic generalized Fock spaces are characterized by the property that one can accomodate up to p identical particles in a state. This property is a manifestation of a generalized exclusion principle and the Greens parastatistics2 is the first consistent example of generalized statistics. We denote by (p%+) the associative algebra, subalgebra of ps(p%) generated by the creation operators ui subject of the relations (3). Every generalized parafermionic (parabosonic) Fock space labelled by integer p arises as a quotient of the the creation algebra ( p S + ) , i e . , by imposing further pdependent conditions. Remark The parafermionic parabosonic p% algebra with D degrees of freedom has been identified with the universal enveloping algebra U(so(2D 1)),3 U ( o ~ p ( 1 1 2 D ) )respectively. ~ However, we shall not make use of these identifications in what follows.
ps
ps+
ps+
ps,
+
2. Homogeneous Algebras
A homogeneous algebra of degree N or N-homogeneous algebra is an algebra of the form5
A = A ( E ,R ) = T ( E ) / ( R ) ,
(6)
where E is a finite-dimensional vector space over C, T ( E ) is the tensor algebra of E and ( R )is the two-sided ideal of T ( E )generated by a vector
233
subspace R of E @ N .The homogeneity of ( R ) implies that A is a graded algebra A = e n E ~ A with n A, = Em'" for n < N and
A,=E@'"/
E @ r @ R @ E @for s n>N,
(7)
r+s=n-N
where we have set E@'O= C. Thus A is a graded algebra A = anE~.Rn which is connected ( A= C), generated in degree 1 and such that the A, are finite-dimensional vector spaces. When N = 2 or N = 3 we shall speak of quadratic or cubic algebras respectively. We shall refer to the subspace R as relation space. We set E = @:,Cul N C D . When the parastatistical order p = 1 then the symmetrical S ( E ) = en>oSnE and the exterior A ( E )= @,>o - An E algebras provide the basises of the bosonic and fermionic Fock space respectively. The grading is the number of particles in a state. The creation algebras pS+ and p%+ are cubic algebras. The creation parafermionic algebra pS+ is the cubic algebra6
PS+ = A ( E ,R ) = T ( E ) / ( " x Y, l @ l .I@) Yl t E E , (8) where [x,y]@= x @ y - y @ x and the subspace R is associated to the 7
2 1
relations (3). The creation parabosonic algebra p%+ is the cubic algebra6
P%+
.I@)
= A(E1 R) = T ( E ) / ( [ { x , Y ) @ ,
7
x1 Y l z E E
1
(9)
where the R means that instead of commutators one has t o take supercommutators, ie., the relation space R is the super-counterpart of R. The parabosons are super-parafermions. Note that in this approach the parafermions are even and the parabosons are odd. Choosing a basis in E l which means fixing the isomorphism E N C D ,we identify the linear group G L ( E )with GL(D,C). One has a natural action of the linear group GL(D,C) (denoted GL(D) from now on) on the linear space E N C D given by left matrix multiplication of the column vector ( G z ) ~= Gixj,
2 = X ~ U ;E
E,
G E GL(D) .
The natural action of GL(D) can be continued t o every tensor power
E@'"as G(x~@ ~2 @ . . . x,) = G z @ ~ G x 8.. ~ . Gx,
.
The natural action of the linear group G L ( D ) on E@' preserves the subspaces R and R, ie., they are invariant submodules. It follows that GL(D)-action passes to the quotients pS+ and p%+ respecting the grading.
234
ps+
This implies that G L ( D )acts on and p%+ by automorphisms, i.e., the algebras p$+ and p%+ are modules of the linear group GL(D). 3. The Symmetric and the Linear Group We briefly recall some general facts about the intimately related representation theories of the symmetric and the linear group. Let X = (X1,Xz ,... Xk) with A1 2 Xz 2 .. . 2 X k > 0 be partition of k
n,
C X i = n.
A partition X of n will be identified with a Young diagram
i=l
X with n boxes and k rows, where X i is the length of the i-th row. The number of boxes in X is denoted by 1x1. The filling of a Young diagram with numbers non-decreasing on rows (from left t o right) and strictly increasing on columns (from top t o bottom) is called a Young tableau. A standard Young tableau is a Young tableau strictly increasing in rows, i.e., without repetition in the entries. The group ring C[Sn]is endowed with structure of a right (left) S -, module in a natural way. The regular representation is the representation of S, into C[S,]. The dimension fx of an irreducible representation Sx of the symmetric group S, (with X partition of n) is equal t o the number of the standard tableaux with shape A. It is well known fact from the theory of finite groups that the regular representation contains all irreducible representations and each representation enters with multiplicity equal t o its dimension. Therefore the left S,-module C[Sn]has the decomposition into a direct sum of modules
C[Sn]N
@ (SX)@fX,
fx
= dim Sx ,
(10)
IXI=n
where Sx are the irreducible S,-modules. The Young projectors Yx(T) are idempotent elements of the group algebra C[Sn]of the symmetric group S,. They are labelled by a partition X and a standard tableau T with shape A. The projectors Yx(T)form an orthogonal system
Yx(T)Yx,(T’) = SAA/STT’YX(T).
(11)
When acting from the right on C[Sn]the Young projectors give rise t o the left irreducible representations of the symmetric group S,
S’ = c[s~IY~(T), 1x1 = n .
(12)
235
The Sn-modules SX arising through different Young projectors Yx(T)with one and the same X axe isomorphic. One can endow the tensor degree EBn with a right action of Sn which permutes the slots in E B n . The right &-action commute with the natural left GL(D)-action on E@l"(Schur-Weylduality). Hence the projection E B n Y ~ ( Tis) stable under the natural left GL(D)-action. The physical meaning of the right &-action is a place permutation of the creation operators a+ while the left So-action (which is a subgroup of GL(D)) operates as particle permutations. Schur module is the irreducible representation E X given by
E X21 E@1"Yx(T),
1x1 = 72.
(13)
The Schur modules arising through Young projectors with one and the same X are isomorphic. So the finite-dimensional irreducible representations E X of the linear group GL(D) are parametrized by Young diagrams too. The character & ( E x ) of the representation E X is the Schur polynomial associated to a partition X
and therefore the decomposition of the GL(D)-module EBn
EBn = @ (EX)@fA IXI=n
follows from the well known formula
4. The Algebras p B + and p 5 + as GL(D)-modules
The standard tableaux for X = ( 2 , l ) are two, thus for n = 3 one has @ E(3) = r\3E @ E(2,1)@ E(2,1)@ S 3 E . E@3 E(l,l,l)@ E(2y1)@ E(2~1)
The Jacobi and super-Jacobi identities
b, [Y,.IF1 + [Y,[z, 471 + b, b,YlFl = 0 imply that the GL(D)-modules R and R (corresponding to p$+ and p%+ ) are neither symmetric nor antisymmetric. It follows that R and R are two orthogonal GL(D)-modules associated to the Young diagram ( 2 , l )
RN
N
I?.
(16)
236
Let T be the maximal element in S3, 7 = slszsl = S Z S ~ S Z . One can choose the Young projectors qz,l)(T1)= Y + and Y&)(T2) = Y - to be the eigenvalues of the place permutation action of T , the so called flip (flip(u@bbc) = ( a @ b @ c ) T = c @ b @ u )
f l i p ( y * ) = y * T = &Y* ,
(17)
which determines Y + and Y - uniquely as S('tl)-modules. The corresponding Schur modules isomorphic to E('?l) are
R = EB3Y+,
R = EB3Y-.
(18)
This simple observation will be very useful later.
Theorem 4.1. Each irreducible representations E x of GL(D) appears exactly once in the decomposition of creation parafermionic pg+ and parabosonic p?B+ algebras
pg+ N @EX N p?B+.
x In the classical textbook on paras tat is tic^^ this result has a tedious proof. Chaturvedi' seems to be the first who used combinatorial identities in parastatistics (in this context). Here we give a short proof to the theorem which boils down to an exercise in the Fulton's Proof: Let us endow the space E @ A'E with the bracket defined by [z, y] = z A y if z and y are both in E and [z,y] = 0 otherwise. The so defined bracket is a Lie bracket and E @ A'E is a graded Lie algebra for this bracket if one ascribes the degree 1to the elements of E and the degree 2 to the elements of A'E. By definition pg+ is the universal enveloping algebra of the graded Lie algebra E @ A'E
pg+ = U ( E CB 2 ~ ) . In view of the PoincarbBirkoff-Witt theorem U ( E@ A'E) is isomorphic as graded vector space and as a graded coalgebra to the symmetric algebra
S ( E @ A'E) = S ( E ) @ S(A'E).
(19)
The character of the symmetric algebra S ( E )over the space E (of degree 1
n (l--zi)-l. It follows from (19) that the character D
elements) is ch S ( E ) =
i=l
of pg+ is given by the left hand side of the Schur combinatorial identity
237
while the right hand side is a sum (over all Young diagrams A) of the characters of Schur modules which implies the decomposition of In view of isomorphisms (16) and p%+ must be isomorphic as GL(D)modules which ends the proof. Remark. The direct proof" of the decomposition of p%+ uses the PoincarB-Birkhoff-Witt theorem for super-Lie algebras. One of the definitions of Schur polynomial is
ps+.
ps+
sx(z) =
czT,
where
zT = z1
... zg
,
(21)
T
where ai(T)is the number of times the entry i appears in the Young tableau T and the sum is over all Young tableaux T which are fillings of the diagram X with numbers from the set (1,. . . , D}. Therefore the Young tableaux T with shape X are in 1-1 correspondence with the monomials in the Schur polynomial s~(z),hence with the basis of the representation EX.Due to the homogeneity of the Schur polynomials sx(t,.
. . ,t ) = tlX1sX(l,. . . , 1) = tl'l dim E X ,
one obtains the PoincarQ series of the algebra p$+ (and also of p%+)
5 . The Plactic Algebra
We now introduce another cubic algebra coming from the combinatorics. The set of Young tableaux can be endowed with a structure of associative monoid, the so called plactic monoid" (see also Ref. 9). Every Young tableau T can be encoded with a word written with the entries of T and one can define algebraic operations on these words, i.e., on the set of Young tableaux. The plactic monoid is generated by ordered set of elements { e l , . . . ,e g } (for tableaux with entries in (1,. . . ,D})subjects of the Knuth relations
eizei3ei, = ei,eil ei3 if
il
< i 2 5 i3
= ei3eilei2 if
il
5 i2 < i 3
eilei3ei,
I
(23)
Let us choose the basis ( e i ) of the linear space C D 21 E to be the canonical basis of C D . The set { e l , . . . , e g } is ordered with the natural ordering
238 el < e:! < ..- < e D . To the Knuth relations we associate in an obvious way the subspace Rv c EB3 which generates the ideal (Rv).The plactic algebra p is the cubic algebra6
'$ = A ( C D ,Rv).
(24)
In contrast to R and R, the relation space Rv depends on the basis ( e k ) and on the ordering of ( e k ) . There is no natural action of GL(D) on $ I because GL(D)-action spoils the ordering. We have seen that the Schur module E X has a linear basis labelled by Young tableaux which are fillings of the diagram X with entries out of the set { 1,2,. . . ,dim E } . On the other hand the homogeneous independent elements of the plactic monoid (and so the plactic algebra) are identified with Young tableaux. We conclude that the Poincarh series of the plactic algebra and the creation parafermionic (parabosonic) p5+ ( ~ 2 3 ' ) algebra coincide
P&)
= Ppg+( t )= PPB+( t ).
(25)
This is not just a coincidence, it turns out that the plactic algebra can be obtained by means of the deformation of the algebras pS+ and pB+. 6. Hecke Algebra and the Quantum Linear Group The Hecke algebra '&(q) is the algebra generated from bi, i = 1,.. . ,n - 1 and the unit element with relations
bibi+lbi = bi+lbibi+l bibj = bjbi b: = 1 ( 4 - q-l)bi
+
Z=l, ..., n - 2 , Ji - j l 2 2, i = l , ...,n - 1 .
These are the same relations as for the symmetric group Sn except for the last one which is relaxation of .s: = 1. When q is not root of unity the Hecke algebra NHn(q) is isomorphic to the group algebra C[Sn]
which allows to index the idempotents in 7-lFln(q) called generalized Young projectors Y:(T)I2 in the same way as Yx(T),ie., again by standard tableaux. One has the q-analog of Eq. (11)
Yx"(T)Y,4(T') = bXA'bTT/YXQ(T).
(27)
239
An irreducible ',Jf,(q)-module stems from the counterpart of (12)
'FIX((?) = ',Jfn(q)YXQ(T) 7
1x1 = n .
The right action of 'H,(q) commutes with the left action of the quantum linear group GL,(D) (quantum Schur-Weyl duality). An irreducible GL,(D)module or q-Schur module is defined in the manner of (14)
1x1 = n .
E A ( q )= E@"YZ(T), 7. Deformation of pS+ and p B + and $!3
Let T, be the maximal element in ' , J f 3 ( q ) . The quantum flip, which is the place permutation (ie., the right) action of T,, has two idempotent eigenvalues Yq* q-flip(Y") = Y P f T , = fY,*.
Yqf are generalized Young projectors Y&)(T).
(28)
We now define two
GL,(D)-modules isomorphic to the q-Schur module E(271)(q) R, = E@'3Yq+,
R -E@3yQ-.
(29)
4-
Deformed creation parafermionic p$,f and deformed creation parabosonic p23: algebra are the following cubic algebras"
Ps:
= A(& Rq)
,
p%;
ii,) .
= A(E,
An outcome of lengthy calculations (using a representation of the Hecke algebra given by a R-matrix of Hecke type) are the explicit expressions for the relations of ps: and p23: (see Ref. 10 and also Ref. 6 for D = 2)
t t t t bf2,[ais,ai1lrlq2 + daf,, [ai2,aillrl =0 r.13, .lll,,al2I42-1- q"al,, "Z',lr,az',] = 0
< i2 I i3 il I i2 < i3 il
}
*
(30)
At the classical point q = 1 with the help of the (super)Jacobi identity we recover the relations (3). In the application of the quantum groups in two-dimensional statistical mechanics, the parameter q has merely the sense of temperature, q = e - h . The point q = 0 corresponding to the absolute zero temperature T = 0 is a singular point for the GL,(D)-symmetry. Figuratively speaking at q = 0 the symmetry is frozen. Nevertheless, the relations (30) are regular at q = 0 and the evaluation in this point called the crystal limit yields
240
Comparing the crystal limit of the pTi-relations (the upper sign in (30)) with the Knuth relations (23) we come to the conclusion that
is a crystal limit of the deformed creation that is, the plactic algebra parafermionic algebra p T i . By analogy the crystal limit of pBl is an algebra that might be called the super-plactic algebra
One can speculate that the ordinary parastatistics is the high temperature limit T 4 00, i e . , q = 1 of the more general deformed (or quantum) parastatistics. Then the algebra of the (super)plactic monoid is the zero temperature limit q = 0.
Acknowledgments
I wish to thank to Michel Dubois-Violette for many inspiring discussions. References 1. E. Wigner, Phys. Rev. 77,711 (1950). 2. H.S. Green, Phys. Rev. 90, 270 (1953). 3. C. Ryan, E.C.G. Sudarshan, NucZ. Phys. 47, 207 (1963). 4. A. Ganchev, T. Palev, J . Math. Phys. 21,797 (1980). 5. R. Berger, M. Dubois-Violette, M. Wambst, J . Algebra 261,172 (2003). 6. M. Dubois-Violette, T. Popov, Lett. Math. Phys. 61,159 (2002). 7. Y. Ohnuki, S. Kamefuchi, Quantum field theory and parastatistics, SpringerVerlag (1982). 8. S. Chaturvedi, hepth/9509150. 9. W. filton, Young tableaux, Cambridge University Press (1997). 10. T. Popov, Ph.D. thesis, http://qcd.th.u-psud.fr/preprints_labo/physiquemath/art2OO3/. 11. A. Lascoux, M.P. Schutzenberger, Quaderni de “La ricerca scientzfica” 109, Roma, CNR 129 (1981). 12. D. Gurevich, Algebra i Analiz 2, 119 (1990).
NONCOMMUTATIVE D-BRANES ON GROUP MANIFOLDS
J. PAWELCZYK Institute of Theoretical Physics, Warsaw University, H o i a 69, PL-00-681 Warsaw, Poland E-mail: [email protected]
H. STEINACKER Institut f u r theoretische Physik, LMU Munchen, Theresienstr. 37, 0-80333Munchen, Germany E-mail: Harold.Steinacker@physik. uni-muenchen. de
We propose an algebraic description of (untwisted) D-branes on compact group manifolds G using quantum algebras related to U,(g). It reproduces the known characteristics of D-branes in the WZW models, in particular, their configurations in G, energies, and the set of harmonics.
1. Introduction This report is a brief review of the quantum algebraic description of Dbranes on group manifolds as proposed in Ref. 1. The structure of D-branes in a B field background has attracted much attention recently. The case of flat branes in a constant B background has been studied extensively and leads to quantum spaces with a MoyalWeyl star product. A rather different situation is given by D-branes on compact Lie groups G, which carry a B field which is not closed. It is known from their conformal field theory (CFT) descriptions2 that stable branes are given by certain conjugacy classes in the group manifold. On the other hand, it is expected that these branes are formed as bound states of DO-branes. Attempting to unify these various approaches, we proposed in Ref. 3 a matrix description of D-branes on S U ( 2 ) . This was generalized in Ref. 1,giving a simple and compact description of all (untwisted) D-branes on group manifolds G in terms of quantum algebras related to the quantized universal enveloping algebra U,(g).The main result is that a model based 241
242
on the reflection equation (RE) algebra leads to precisely the same branes as the WZW model. It not only reproduces their configurations in G, i.e. the positions of the corresponding conjugacy classes, but also gives the same (noncommutative) algebra of functions on the branes.
2. CFT and the Classical Description of Untwisted D-branes The CFT description is given in terms of a WZW model, which is specified by a compact group group G and an integer level k.4 We concentrate on the case G = S U ( N ) ,but all constructions work for S O ( N ) and U S p ( N ) as well. The WZW branes can be described by certain boundary states of the Hilbert space of closed strings. We consider here only "symmetrypreserving branes" (untwisted branes), given by the Cardy (boundary) states. They are labeled2v5by a finite set of integral weights
X E I'z = {A
E
P+; X.8 5 k } ,
(1)
(here 8 is the highest root of g), corresponding to integrable irreps of G. Hence untwisted branes are in one-to-one correspondence with X E P l . The energy of the brane X is given by
The CFT also contains the description of branes as quantum manifolds, in terms of boundary primary fields. Their number is finite for any compact WZW model. In the k -+ oa limit, these boundary primaries generate the (noncommutative) algebra of functions on the b r a n e ~ ,see ~ )also ~ Section 4. For finite k , the corresponding algebra, as given in Ref. 6, is not associative. It becomes associative after "twisting", so that it can be considered as algebra of functions of a quantum manifold. Then the primaries become modules of the quantum group U,(g). On a semi-classical level, the D-branes are simply conjugacy classes of the group manifold, of the form
C ( t ) = {gtg-';
g E G} .
(3)
One can assume that t belongs to a maximal torus T of G. These conjugacy classes are invariant under the adjoint action of the vector subgroup Gv
243
GLx GR of the group of motions on G. This reflects the breaking 5~X
~ -+ R
iV.
A lot of information about the spaces C ( t ) can be obtained from the harmonic analysis, i.e. by decomposing scalar fields on C ( t ) into harmonics under the action of the (vector) symmetry G v . One finds
F(C(t))2 @ muzt,,( K t ) v,.
(4)
XEP+
Here X runs over all dominant integral weights P+, V, is the corresponding highest-weight G-module, and r n u l t ~is~the ) dimension of the subspace of V,+ = V,, which is invariant under the stabilizer Kt o f t . As discussed above, there is only a finite set of stable D-branes on G (up to global motions) in the CFT description, one for each integral weight X E P:. They correspond to C ( t x ) for
t,
tn
= q2(HA+HP)
(5)
q=ek+sv.
Here gv is the dual Coxeter number. The location of these branes in G is encoded in s,
= t r(gn) = tr(t") ,
g E C(t), n E
N,
(6)
which are invariant under the adjoint action. For the classes C ( t x ) , they can be easily calculated: s, = trVN ( 42 7 G P + H A ) ) =
c
q274P+A)+
(7)
UCvN
where VN is the defining representation. The s, completely characterize C ( t x ) , and their quantum analogs (11) can be calculated exactly. An equivalent characterization of these conjugacy classes is provided by a characteristic equation of the form P x ( M ) = 0.l 3. Quantum Algebras and Symmetries for Branes We now define the quantum space describing G and its branes in terms of a non-commutative algebra M , which transforms under a quantum symmetry. The quantized algebra M of functions on G is generated by elements Mj with indices i,j in the defining representation of G, subject to some commutation relations and constraints. The relations are given by the socalled reflection equation (RE),' which in a short notation reads R2iMiRi2M2 = M2R21MiRi2.
(8)
244
Here R is the R matrix of U,(g)in the defining representation. For q = 1, this reduces to [M:, Mf]= 0. Because M should be a quantization of G, there must be further constraints. In the case G = S U ( N ) , these are det,(M) = 1, where det, is the so-called quantum determinant (12), and suitable reality conditions imposed on the generators M j . The RE appeared more than 10 years ago in the context of the boundary integrable model^.^ M is covariant under the the transformation
M; -+ ( s - w t ) ; ,
(9)
where si and ti generate algebras BL and GR respectively, which both coincide with the well-known quantum groups Fun,(G), as defined in Ref. 8; i.e. s2slR = RslsZ, t2tlR = Rtlt2 etc. This (co)action is consistent with RE if we impose that (the matrix elements of) s and t commute with M , and in addition satisfy s2tlR = Rtls2. Formally, M is then a right 8' GR - comodule algebra; seeRef. 1 for further details. Notice that Eq. (9) is a quantum analog of the action of the classical isometry group GL x GR on classical group element g. Furthermore, GL B R GR can be mapped to a vector Hopf algebra GV with generators T , by si @I 1 -+ rj and 18 tj -+ T ; . The (co)action of Gv on M is then
M;
-+ (r-lMr);.
.
(10)
The (generic) central elements of the algebra (8) are given by c, = t r q ( M n )E trv,,,
(M"
W)
E
M ,
(11)
where v = T ( q - 2 N p ) is a numerical matrix which satisfies S'(T) = W - ~ T W for the generator T of Gv. These elements c, are independent for n = 1 , 2 , ...,rank(G). One can also show that the c, are invariant under Gv. As we shall see, the c, for n = 1,...rank(G) fix the position of the brane configuration on the group manifold, i.e. they are quantum analogs of the s n , Eq. (7). There is another central term det,(M), the quantum determinant, which is invariant under the full chiral quantum algebra GL BR GR. Hence we can impose the constraint det,(M) = 1.
(12)
Furthermore, there is a realization (algebra homomorphism) of the RE algebra (8) in terms of the algebra U,(g), given by M = (T 8 i d ) ( R z i R i z ) where T is the defining representation. Equation (8) is then a consequence of the Yang-Baxter equation for R.8
245
4. Representations of M and Quantum D-branes We claim that the quantized orbits corresponding to the D-branes of interest here are described by irreps 7rx of M . On any irrep, the Casimirs c, (11) take distinct values, i.e. they become constraints. In view of their explicit form, an irrep of M should be considered as quantization of a conjugacy class C ( t ) , whose position depends on c,. To confirm this interpretation, one can calculate the position of the branes on the group manifold, and study their geometry by performing the harmonic analysis on the branes. The “good” irreps of the algebra M coincide with the highest weight representations Vx of U,(g)for X E P z : they are ~ n i t a r y have , ~ positive quantum-dimension, and are in one-to-one correspondence with the integrable modules of the affine Lie algebra g. As shown in Ref. 1,these irreps 7 r ~ of M for X E P z describe precisely the stable D-branes C ( t x ) , denoted by Dx. It is an algebra of maps from Vx to Vx which transforms under the quantum adjoint action of U,(g).For “small” weights, this algebra coincides with Mat(Vx). There is clearly a one-to-one correspondence between the (untwisted) branes in string theory and these Dx,since both are labeled by X E P z . 4.1. Position of DX
The values of the Casimirs c, on Dx are as follows:1
Here XN is the highest weight of the defining representation VN, and the sum in Eq. (14)goes over all v E VN such that X v lies in P z . The value of q ( X ) agrees precisely with the corresponding value (7) of s1 on C ( t x ) . For n 2 2, the agreement of cn(X) with sn on C ( t x ) is only approximate, becoming exact for large A. The discrepancy can be blamed to operator-ordering ambiguities. M also satisfies a characteristic equation1 similar to the classical one. Therefore the position and “size” of the branes essentially agrees with the results from CFT. Furthermore, the energy of the D-brane is given by the quantum dimension of the representation space
+
vx
246
4.2. The Space of Harmonics on DA
For simplicity, assume that X is not too large. Then
Dx
2
Mut(Vx) = Vx @ V i % CB,N;~+V, ,
(16)
since the tensor product is completely reducible. Here N f x + are the usual fusion rules of 8. This has a simple geometrical meaning if p is small enough: comparing with Eq. (4),one can show' that
D x 2 F(C(ti))1
(17)
. differs slightly from up t o some cutoff in p, where t i = e x p ( 2 ~ i h ) This Eq. (5), by a shift X + X p. The structure of harmonics on Dx is however in complete agreement with the CFT results. Moreover, it is known5 that the structure constants of the corresponding boundary operators are essentially given by the 6 j symbols of V,(g), which in turn are precisely the structure constants of the algebra of functions on Dx. Therefore, our quantum algebraic description not only reproduces the correct set of boundary fields, but also essentially captures their algebra in (B)CFT. Particularly interesting examples of degenerate conjugacy classes are the complex projective spaces @PN-l, which in the simplest case of G = SU(2) become (q-deformed) fuzzy spheres" Si,N.
+
References 1. J. Pawelczyk, H. Steinacker, H. Nucl. Phys. B638, 433 (2002). 2. A. Yu. Alekseev, V. Schomerus, Phys. Rev. D60, 061901 (1999). 3. J. Pawelczyk and H.Steinacker, JHEP 0112,018 (2001). 4. J. Fuchs, AfJine Lie Algebras and Quantum Groups, Cambridge University Press (1992). 5. G. Felder, J. Frohlich, J. Fuchs, C. Schweigert, J. Geom. Phys. 34, 162 (2000).
6. A. Yu. Alekseev, A. Recknagel and V. Schomerus, JHEP 9909, 023 (1999). 7. E. Sklyanin, J. Phys. A21,2375 (1988). 8. L. D. Faddeev, N. Yu. Reshetikhin and L. A. Takhtajan, Algebra Anal. 1, 178 (1989). 9. H. Steinacker, Rev. Math. Phys. 13, No. 8, 1035 (2001). 10. H. Grosse, J. Madore, H. Steinacker, J. Geom. Phys. 38, 308 (2001).
HIGH-ENERGY BOUNDS ON THE SCATTERING AMPLITUDE IN NONCOMMUTATIVE QUANTUM FIELD THEORY
A. TUREANU
P.
Department of Physical Sciences, and Helsinki Institute of Physics, 0. Box 64, 00014 University of Helsinki, Helsinki, Finland E-mail: [email protected]
In the framework of quantum field theory (QFT) on noncommutative (NC) space time with SO(1,l) x SO(2) symmetry, which is the feature arising when one has only space-space noncommutativity (&i = 0), we prove that the Jost-LehmannDyson representation, based on the causality condition usually taken in connection with this symmetry, leads to the mere impossibility of drawing any conclusion on the analyticity of the 2 + 2-scattering amplitude in cos 8,0 being the scattering angle. A physical choice of the causality condition rescues the situation and, as a result, an analog of Lehmann’s ellipse as domain of analyticity in cos 8 is obtained. However, the enlargement of this analyticity domain to Martin’s ellipse and the derivation of the Froissart bound for the total cross-section in NC QFT is possible only in the special case when the incoming momentum is orthogonal to the NC plane. This is the first example of a nonlocal theory in which the cross-sections are subject t o a high-energy bound. For the general configuration of the direction of the incoming particle, although the scattering amplitude is still analytic in the Lehmann ellipse, no bound on the total cross-section has been derived. This is due to the lack of a simple unitarity constraint on the partial-wave amplitudes.
1. Introduction The development of QFT on NC space-time, especially after the seminal work of Seiberg and Witten,l which showed that the NC QFT arises from string theory, has triggered lately the interest also towards the formulation of an axiomatic approach to the subject. Consequently, the analytical properties of scattering amplitude in energy E and forward dispersion relations have been ~ o n s i d e r e d ,Wightman ~>~ functions have been introduced and the CPT theorem has been p r o ~ e n , and ~ ’ ~as well attempts towards a proof of 247
248
the spin-statistics theorem have been made.5a In the axiomatic approach to commutative QFT, one of the fundamental results consisted of the rigorous proof of the Froissart bound on the high-energy behaviour of the scattering amplitude, based on its analyticity properties.lOyll Here we aim at obtaining the analog of this bound when the space-time is noncommutative. Such an achievement, besides being topical in itself, will also prove fruitful in the conceptual understanding of subtle issues, such as causality, in nonlocal theories to which the NC QFTs belong.12 In the following we shall consider NC QFT on a space-time with the commutation relation [ Z p , ZCYI
= iepv
7
(1)
where 8,, is an antisymmetric constant matrix (for a review, see, e.g., Refs. 13, 14). Such NC theories violate Lorentz invariance, while translational invariance still holds. We can always choose the system of coordinates, such that 813 = 823 = 0 and OI2 = -OZ1 = 8. Then, for the particular case of space-space noncommutativity, i.e. 8oi = 0 , the theory is invariant under the subgroup SO(1,l) x SO(2) of the Lorentz group. The requirement that time be commutative (8oi = 0 ) discards the well-known problems with the unitarity15 of the NC theories and with ~ a u s a l i t y . ~ ~ ? ~ ~ As well, the Boi = 0 case allows a proper definition of the S - m a t r i ~ . ~ In the conventional (commutative) QFT, the Froissart bound was first obtained1' using the conjectured Mandelstam representation (double dispersion relation)," which assumes analyticity in the entire E and cos@ complex planes. The Froissart bound,
expresses the upper limit of the total cross-section ctot as a function of the CMS energy E , when E -+ cy). However, such an analyticity or equivalently the double dispersion relation has not been proven, while smaller domains of analyticity in cos 0 were already known.lg One of the main ingredients in rigorously obtaining the Froissart bound is the Jost-Lehmann-Dyson representation20'21 of the Fourier transform of the matrix element of the commutator of currents, which is based on aIn the context of the Lagrangian approach to NC QFT,the CPT and spin-statistics theorems have been proven in general in Ref. 6; for CPT invariance in NC QED,see Refs. 7, 8, and in NC Standard Model Ref. 9.
249
the causality as well as the spectral conditions (for an overall review, see Ref. 22). Based on this integral representation, one obtains the domain of analyticity of the scattering amplitude in cos0. This domain proves However, the domain to be an ellipse - the so-called Lehmann’s e1lip~e.l~ of analyticity in c o s 0 can be enlarged to the so-called Martin’s ellipse by using the dispersion relations satisfied by the scattering amplitude and the unitarity constraint on the partial-wave amplitudes. Using this larger domain of analyticity, the Froissart bound (2) was rigorously provenll (for a review, see Ref. 23). In NC QFT with 8oi = 0 we shall follow the same path for the derivation of the high-energy bound on the scattering amplitude, starting from the Jost-Lehmann-Dyson representation and adapting the derivation to the new symmetry SO(1,l) x SO(2) and to the nonlocality of the NC theory. In Section 2 we derive the Jost-Lehmann-Dyson representation satisfying the light-wedge (instead of light-cone) causality condition, which has been used so far, being inspired by the above symmetry. In Section 3 we show that no analyticity of the scattering amplitude in cos 0 can be obtained in such a case. However, with a newly introduced causality condition, based on physical arguments, we obtain from the Jost-Lehmann-Dyson representation a domain of analyticity in cos 0 , which coincides with the Lehmann ellipse. In Section 4 we show that the extension of this analyticity domain to Martin’s ellipse is possible in the case of the incoming particle momentum orthogonal t o the NC plane (x1,x2), which eventually enables us to derive rigorously the analog of the Froissart-Martin bound (2) for the total cross-section. The general configuration of incoming particle momentum is also discussed, together with the problems which arise. Section 5 is devoted to conclusion and discussions. 2. Jost-Lehmann-Dyson Representation The Jost-Lehmann-Dyson representation2’t21 is the integral representation for the Fourier transform of the matrix element of the commutator of currents:
where X
X
f ( x ) = ( p ’ l [ j l ( ~ ) , j 2 ( - ~ ) I l p7 )
(2)
satisfying the causality and spectral conditions. The process considered is the 2 -+ 2 scalar particles scattering, k + p -+ k’ + p ’ , and j , and j 2 are the
250
scalar currents corresponding to the incoming and outgoing particles with momenta k and k' (see also Refs. 22, 24). For NC Q F T with SO(1 , l ) x SO(2) symmetry, in Ref. 25 a new causality condition was proposed, involving (instead of the light-cone) the light-wedge corresponding to the coordinates xo and x3, which form a two-dimensional space with the S O ( 1 , l ) symmetry. Accordingly we shall require the vanishing of the commutator of two currents (in general, observables) a t space-like separations in the sense of S O ( 1 , l ) as: X
X
for ~ 2 = x ~ -< ox . ~
[ j l ( 2- ) , j 2 ( - - 2) ] = 0 ,
(3)
The spectral condition compatible with (3) would require now that the physical momenta be in the forward light-wedge:
p2 =pE
-pi
> 0 and
po
>0.
(4)
The spectral condition (4) will impose restrictions on f ( q ) (see Ref. 12 for details), in the sense that f ( q ) = 0 in the region outside the hyperbola Po -
Jm<
qo
< -Po
+b G .
(5)
To derive the Jost-Lehmann-Dyson representation, we further consider the 6-dimensional space-time with the Minkowskian metric (+, -, -, -, -, -) (for details of the derivation, see Ref. 12). Using the standard mathematical procedure,28 one obtains the Jost-Lehmann-Dyson representation in NC QFT, satisfying the light-wedge causality condition
(3): f(q)=
/
d4udK26(qo
- uo)b[(qo- uo)2 - (q3 - u3)2 - K21
x d(q1 - u l N q 2 - u2)4('zL,K 2 ) 7
(6)
where + ( u , K ~ ) = -*. Equivalently, denoting ii = (uo,us),Eq. (6) can be written as:
The function 4(G,41,q 2 , K ~ is ) an arbitrary function, except that the requirement of spectral condition determines a domain in which +(ii, q1, q 2 , K ~ =) 0. This domain is outside the region where the 6 function in (7) vanishes, i.e.
(@- i i ) 2 - K 2 = 0 ,
(8)
25 1
but with 4 in the region given by ( 5 ) , where f(q) = 0. Putting together (8) and (5), we obtain the domain out of which $(ii, 41, 4 2 , K ~ =) 0: a)
1 -(F 2
+@I)
fii are in the forward light-wedge (cf. (4));
(9)
For the purpose of expressing the scattering amplitude, we actually need the Fourier transform fR(4) of the retarded commutator, fR(x) = e(xo)f(x) = ( ~ ' ~ ~ ( ~ 0 ) [ ~ 1 ( ~ ) , ~, 2 ( - ~ ) (10) 1 ~ ~ )
which is obtained in the form:12
Compared to the usual Jost-Lehmann-Dyson representation,
the expression (11) is essentially different in the sense that the arbitrary function $ now depends on 41 and 4 2 . This feature will have further crucial implications in the discussion of analyticity of the scattering amplitude in cos 8. 3. Analyticity of the Scattering Amplitude in cos 0. Lehmann's Ellipse
In the center-of-mass system (CMS) and in a set in which the incoming particles are along the vector = ( O , O , e),b the scattering amplitude in NC QFT depends still on only two variables, the CM energy E and the cosine of the scattering angle, cos 8 (for a discussion about the number of variables in the scattering amplitude for a general type of noncommutativity see Ref. 26).
6
bThe "magnetic" vector p' is defined as pi = iEijk6'jk The terminology stems from the antisymmetric background field B,, (analogous to F,, in $ED), which gives rise to , essentially proportional to B p y (see, e.g., noncommutativity in string theory, with 8 Ref. 1).
252
In terms of the Jost-Lehmann-Dyson representation, the scattering amplitude is written as (cf. Ref. 22 for commutative case):
where +(G, K’, ...) is a function of its SO(1,l)- and SO(2)-invariant variables: - u:, (ko PO)^ - (k3 - ~ 3 ) ~(ki , PI)^ (k2 + ~ 2 ) ~(k:, (kh The function q5 is zero in a certain domain, determined by the causal and spectral conditions, but otherwise arbitrary. For the discussion of analyticity of M ( E ,cos 0 ) in cos 8,it is of crucial importance that all dependence on cos 8 be contained in the denominator of (13). But, since the arbitrary function q5 depends now on (k’- p’)1,2, it also depends on cos 8. This makes impossible the mere consideration of any analyticity property of the scattering amplitude in cos 8. Since the Jost-Lehmann-Dyson representation reflects the effect of the causal and spectral axioms, we notice that the hypotheses (3) and (4) used for the present derivation of JLD representation are too weak, in the sense of their physical implications, since they allow for a much larger physical region, by not at all taking into account the effect of the NC coordinates x1 and 22.
+
+
ui
3.1. Causality in NC QFT In the following, we shall challenge the causality condition f(x) = 0 , for
z2 = x i - x i < 0 ,
(14)
which takes into account only the variables connected with the SO(1,l) symmetry. This causality condition would be suitable in the case when nonlocality in NC variables x1 and 2 2 is infinite, which is not the case on a space with 8. The the commutation relation [xI,x2] = 28, which implies AxlAx2 fact that in the causality condition (14) the coordinates x1 and 2 2 do not enter means that the propagation of a signal in this plane is instantaneous: no matter how far apart two events, are in the noncommutative coordinates, the allowed region for correlation is given by only the condition xg -xi > 0, which involves the propagation of a signal only in the x3-direction, while the time for the propagation along 21- and xz-directions is totally ignored. Admitting that the scale of nonlocality in x1 and 2 2 is 1 i.e. the propagation of interaction in the noncommutative coordinates is instanta-
-
- 4,
253
neous only within this distance 1, we can argue, based on these physical arguments, that the locality condition should indeed be given by:
f(x) = o
, for z2 - (xf+ xi - 1 2 )
3 xi
+ x; - 1’)
- xi - (xf
or, equivalently,
f(x) = 0 , for xi - xi - (x: + xi) < -12, where l 2 is a constant proportional to NC parameter 8. When l2 becomes the usual Iocality condition. Correspondingly, the spectral condition will read as Pi -Pi - (Pf + P i ) 2 0 ,
PO
>0,
(15) -+ 0,
(15)
(16)
since there is no noncommutativity in momentum space. In fact, the consideration of nonlocal theories of the type (15) was initiated by W i g h t m a x ~who , ~ ~ asked the concrete question: whether the vanishing of the commutator of fields (or observables), i.e. f(x) = 0, for xg - x: -xi -xi < -12, would imply its vanishing for xi - x: -xi - xi < 0. It was proven later28-30 (see also Ref. 31) that, indeed, in a quantum field theory which satisfies the translational invariance and the spectral axiom (16), the nonlocal commutativity 2 f(z)= 0 , for xo - xf - xi - xi < -12
implies the local commutativity
f(x) = 0 ,
for xi - xf - xi - xi < 0 .
(17)
This powerful theorem, which does not require Lorentz invariance, can be applied in the noncommutative case, since the hypotheses are fulfilled, with the conclusion that the causality properties of a QFT with space-space noncommutativity are physically identical to those of the corresponding commutative QFT. It is then obvious that the Jost-Lehmann-Dyson representation (12) obtained in the commutative case holds also on the NC space. Consequently, the NC two-particle4two-particle scattering amplitude will have the same form as in the commutative case:
This leads to the analyticity of the NC scattering amplitude in cos 0 in the analog of the Lehmann ellipse, which behaves at high energies E the same
254
way as in the commutative case, i.e. with the semi-major axis as yr, = (COSO),,,
=1
const +E4 .
4. Enlargement of the Domain of Analyticity in cos 0 and Use of Unitarity. Martin's Ellipse Two more ingredients are needed in order to enlarge the domain of analyticity in cos O to the Martin's ellipse and to obtain the F'roissart-Martin bound: the dispersion relations and the unitarity constraint on the partialwave a m p l i t ~ d e s . ~ ~ Imposing the physical nonlocal commutativity condition (15) and reducing it to the local commutativity (17), by using the theorem due to Wightman, Vladimirov and Petrina, leads straightforwardly to the usual forward dispersion relation also in the NC case with a general direction. As for the unitarity constraint on the partial wave amplitudes, the problem has been dealt with in Ref. 26, for a general case of noncommutativity O p v , Qoi # 0. For space-space noncommutativity (Ooi = 0), the scattering amplitude depends, besides the center-of-mass energy, E , on three angular variables. In a system were we take the incoming momentum @in the z-direction, these variables are the polar angles of the outgoing particle momentum, 0 and 4,and the angle a between the vector and the incoming momentum. The partial-wave expansion in this case reads:
p
A ( E ,074,a ) =
C (21' + 1)a~~~rn(E)J'irn(@, 4)8t(cosa)
7
(20)
l,l',m
where J'i, are the spherical harmonics and Pp are the Legendre polynomials. Imposing the unitarity condition directly on (20) or using the general formulas given in Ref. 26, it can be shown that a simple unitarity constraint, which involves single partial-wave amplitudes one at a time, cannot be obtained in general, but only in a setting where the incoming momentum is orthogonal to the NC plane (equivalently it is parallel to the vector In this case, the amplitude depends only on one angle, 0 , and the unitarity constraint is reduced to the well-known one of the commutative cases, i.e.
p).
p,
For this particular setting, @ 11 it is then straightforward, following the prescription developed for commutative QFT, to enlarge the analyticity
255
domain of scattering amplitude to Martin’s ellipse with the semi-major axis at high energies as
and subsequently obtain the NC analog of the Froissart-Martin bound on the total cross-section, in the CMS and for ji 11
8:
E . atot(E)5 c In2 EO
(23)
Thus, the unitarity constraint on the partial-wave amplitudes distinguishes a particular setting (6 11 /?) in which the Lehmann’s ellipse can be enlarged to the Martin’s ellipse and Froissart-Martin bound can be obtained. Nevertheless, this does not exclude the possibility of obtaining a rigorous high-energy bound on the cross-section for p’N and the issue deserves further investigation.
8,
5. Conclusion and Discussions In this paper we have tackled the problem of high energy bounds on the twoparticle+two-particle scattering ampIitude in NC QFT and obtained that, using the causal and spectral conditions (3) and (4) proposed in Ref. 25 for NC theories, it is impossible to draw any conclusion about the analyticity of the scattering amplitude (13) in cos0. However, the physical observation that nonlocality in the noncommuting coordinates is not infinite brought us to imposing a new causality condition (15). We proved that the new causality condition is formally identical t o the one corresponding to the commutative case (17), using the Wightman-Vladimirov-Petrina t h e ~ r e m . ~ ~Thus, - ~ ’ the scattering amplitude in NC QFT is proved to be analytical in cos 0 in the Lehmann ellipse, just as in the commutative case; moreover, dispersion relations can be written on the same basis as in usual QFT. Finally, based on the unitarity constraint on the partial-wave amplitudes in NC QFT, we can conclude that, for theories with space-space noncommutativity (&i = 0), the total cross-section is subject to an upper bound (23) identical to the Froissart-Martin bound in its high-energy behaviour, when the incoming particle momentum p’ is orthogonal to the NC plane. This is the first example of a nonlocal theory, in which cross-sections do have an upper high-energy bound.
256
Acknowledgments T h e financial support of the Academy of Finland under the Project no. 54023 is acknowledged.
References 1. N. Seiberg and E. Witten, JHEP 9909,032 (1999), hep-th/9908142. 2. Y. Liao and K. Sibold, Phys. Lett. B549,352 (2002), hepth/0209221. 3. M. Chaichian, M. N. Mnatsakanova, A. Tureanu and Yu. S. Vernov, Nucl. Phys. B673,476 (2003), hep-th/0306158. 4. L. Alvarez-GaumB and M. A. Vkquez-Mozo, Nucl. Phys. B668,293 (2003), h e p t h/0305093. 5. M. Chaichian, M. N. Mnatsakanova, K. Nishijima, A. Tureanu and Yu. S. Vernov, hep- th/0402212. 6. M. Chaichian, K. Nishijima and A. Tureanu, Phys. Lett. B568, 146 (2003), h e p t h/0209008. 7. M. M. Sheikh-Jabbari, Phys. Rev. Lett. 84,5265 (2000), hep-th/0001167. 8. S. M. Carroll, J. A. Harvey, V. A. Kostelecky, C. D. Lane, T. Okamoto, Phys. Rev. Lett. 87,141601 (2001), hep-th/0105082. 9. P. Aschieri, B. JurEo, P. Schupp and J. Wess, Nucl. Phys. B651,45 (2003), hepth/0205214. 10. M. Froissart, Phys. Rev. 123,1053 (1961). 11. A. Martin,Phys. Rev. 129,1432 (1963); Nuovo Cim. 42,901 (1966). 12. M. Chaichian and A. Tureanu, hep-th/0403032. 13. M. R. Douglas and N. A. Nekrasov, Rev. Mod. Phys. 73,977 (2001), hepth/O 106048. 14. R. J. Szabo, Phys. Rept. 378,207 (2003), hep-th/0109162. 15. J. Gomis and T. Mehen, Nucl. Phys. B591,265 (2000), hep-th/0005129. 16. N. Seiberg, L. Susskind and N. Toumbas, JHEP 0006,044 (2000), h e p th/0005015. 17. L. Alvarez-GaumB and J. L. F. Barbon, Int. J. Mod. Phys. A16,1123 (2001), h e p t h/0006209. 18. S. Mandelstam, Phys. Rev. 112,1344 (1958). 19. H. Lehmann, Nuovo Cimento 10,579 (1958). 20. R. Jost and H. Lehmann, Nuovo Cimento 5, 1598 (1957). 21. F. Dyson, Phys. Rev. 110,1460 (1958). 22. S. S. Schweber, An Introduction to Relativistic Quantum Field Theory, Row, Peterson and Company (1961). 23. A. Martin, Scattering Theory: Unitarity, Analyticity and Crossing, Springer Verlag (1969). 24. N. N. Bogoliubov and D. V. Shirkov, Introduction to the Theory of Quantized Fields, Wiley, New York, 3rd ed. (1980). 25. L. Alvarez-GaumB, J. L. F. Barbon and R. Zwicky, JHEP 0105,057 (2001), hepth/0103069. 26. M. Chaichian, C. Montonen and A. Tureanu, Phys. Lett. B566,263 (2003),
257
hepth/0305243. 27. A. S. Wightman, Matematika 6:4,96 (1962). 28. V. S. Vladimirov, Sou. Math. Dokl. 1, 1039 (1960); Methods of the Theory of Functions of Several Complex Variables, Cambridge, Massachusetts, MIT Press (1966). 29. D. Ya. Petrina, Ukr. Mat. Zh. 13,No. 4, 109 (1961) (in Russian). 30. A. S. Wightman, J. Indian Math. SOC.24, 625 (1960-61). 31. N. N. Bogoliubov, A. A. Logunov and I. T. Todorov, Introduction t o Aziomatic Quantum Field Theory, W. A. Benjamin, Inc., New York (1975).
MANY FACES OF D-BRANES: FROM FLAT SPACE, VIA ADS TO PP-WAVES
M. ZAMAKLAR MPI fur Gravitationsphysik A m Muhlenberg 1 14476 Golm, Germany E-mail: [email protected] We review recent studies of branes in Ads x S and pp-wave spaces using effective action methods based on probe branes and supergravity. We also summarise results on an algebraic study of D-branes in these spaces, using extensions of the superisometry algebras which include brane charges.
1. Introduction
Since their discovery in 1995, D-branes have become the center of intensive research of the string theory community. In this process a lot has been learnt about various manifestations of these objects. It has become clear that, depending on the regime in which one works, D-branes can be described using a variety of techniques. In situations with a small number of branes and weak string coupling, methods of open string conformal field theory are appropriate. Open string CFT techniques have, however, been applied mainly to the study of D-branes in very restricted classes of backgrounds. The main obstacle in the study of branes using these techniques is the lack of knowledge concerning the quantisation of strings in arbitrary backgrounds. The problem can be simplified partially by restricting one’s interest t o low-energy processes. In this case a description of branes using effective actions becomes viable. The effective actions of the closed string CFTs are various supergravity actions. These actions are supersymmetric generalisations of the Einstein-Hilbert action. The open string CFT leads to the Dirac-Born-Infeld action (DBI), which (as the name suggests) is a generalisation of the Dirac action for the relativistic membrane to higher dimensional objects, modified to include the brane’s “inner” degrees of freedom, i.e. gauge fields. The latter are included in a fashion proposed a long time 258
259
ago by Born and Infeld in order to regularise the infinite electromagnetic energy of a classical electron. The total (bulk plus brane) effective action is quite complicated due to the non-trivial coupling between the brane and supergravity fields. Hence in order to use this action, one is often forced to simplify the problem further. Increasing the number of branes, for example, leads to the regime where gravitational back-reaction of the branes cannot be neglected, while the gauge theory on the worldvolume of the branes becomes strongly coupled. In this regime branes can be described as purely gravitational solutions, using only the bulk effective action. On the other side, when the number of branes is very small, the probe brane approach is appropriate: only the worldvolume excitations (i.e. scalars and gauge fields) are dynamical fields, while the background is “frozen”. In the first two parts of this report we will partially survey our recent study of Dbranes in Ads and plane-wave (pp-wave) geometries using the supergravity and Dirac-Born-Infeld effective actions. Finally, a very powerful method for classifying possible brane configurations in arbitrary backgrounds is the so-called algebraic method. The full information about the non-perturbative spectrum of string theory (in a given background) is encoded in the “central” extensions of the appropriate superisometry algebra of the background. Unfortunately, the explicit forms of the algebras are essentially not known beyond flat space. Recently, however, we have made an important step in understanding the construction of central extensions of Ads and pp-wave superisometry algebras. In the last part of this survey we report on these results. 2. The Probe Brane Approach 2.1. h m Flat Space to A d s
The problem of understanding the full brane-background system is simplified dramatically by making a restriction to the effective actions, and then further restricting to the probe brane approach. However, this still leaves one with a generically complicated action which one has to solve in order to find the exact embedding of the brane surface in the target space. In generic backgrounds, and in cases of supersymmetric brane configurations, one often uses kappa-symmetry or calibration methods in order to replace the problem of solving second-order differential equations to the problem of solving first order, BPS-like equations. Both of these methods, however, are in a certain sense ad hoc, since they both require good intuition about the ansatze for the brane embeddings.
260
On the other hand, the situation is simpler in special backgrounds such as Ads or pp-waves. All supersymmetric brane configurations in these spaces are inherited from flat space and can be ‘Lderived”by starting from flat space. Given a brane configuration in flat space one first replaces some of the branes in the configuration by their supergravity solutions and subsequently one focuses on their corresponding near-horizon geometries, while keeping the remaining branes as probes. The actual equations for the embedding of the probe can usually be deduced directly from the flat space equations, using Poincark coordinates for the Ads background. This is due to the fact that in this coordinate system, the relation to the flat space Cartesian coordinates is direct. Moreover, it turns out that the same equations that describe the embedding of the brane in flat space also solve the D B I action in the near horizon geometry in Poincark coordinates. The essential reason why the inheritance property holds is the fact that the brane configurations are supersymmetric. This situation should be contrasted to non-BPS configurations, for example a circular 0 1 string in the space transverse to a 0 3 brane. In flat space one can easily derive the solution that describes shrinking of this string, and one can show that this solution (in Cartesian coordinates), when interpreted in Poincarh coordinates, does not solve the DBI equations of motion of the 0 1 string in the Ads5 x S5space. Instead of listing all possible brane configurations which one can derive using this method, let us illustrate it on a very simple example of two 0 3 branes intersecting over a string, 0301 2 3 ----- D301--4 5----.
(1)
In flat space the embedding equations of the second brane are given by
-
2%. -- cz-const., .
(i=2,6,7,8,9).
(2)
We now we replace the first brane with its near horizon geometry, i.e. with the AdSS x S5 space, which in Poincarh coordinates yields ds2 = R2u2(- dt2 +dx: +dx; +dxz) u2 = ~42
+ ... + X: +
dx:
+ $(dx: +. . .
+ . . . + dx:
= du2 -I-u2dRi.
When all coefficients ci in Eq. (2) are zero, we see that the Eq. (2) define a maximal-curvature Ads2 x S1 submanifold.1*2In the case when some of the ci # 0, the Eq. (2) solve the 0 3 DBI action in the (full) 0 3 brane
261
supergravity background. However, when taking the near horizon limit, one in addition needs to rescale the parameters q to zero in order for the solution to survive this limit. As we focus on the region near the 0 3 brane that becomes the background, we simultaneously have to bring the probe 0 3 brane closer and closer to the horizon. The resulting geometry of the 0 3 brane probe describes a brane which starts at the Ads boundary, extends in the u-direction up to some point and then folds back to the b~undary.~ We have recently extended4 the analysis described above to the cases of supersymmetric brane configuration intersecting under an angle in flat space. Due to supersymmetry the inheritance property goes through, as in the previous situation. The resulting geometries of the AdS branes are, however, different, since they non-trivially mix the Ads and sphere submanifolds, i.e. unlike before, the worldvolume surfaces are not factorisable into a product of Ads and sphere submanifolds. Another interesting type of brane that has appeared in Ref. 4 is one whose worldvolume surfaces axe direct products of Ads and sphere submanifolds, but where mixing is achieved via a mixed worldvolume flux which has one index in the sphere and one index in the Ads part. These branes wrap non-supersymmetric target space cycles which are stabilised only after the mixed worldvolume flux is turned on. To construct them, one starts from the flat space configuration of branes intersecting under an angle and performs T-duality in such a way that branes which will be replaced with the background do not carry any worldvolume flux, while the brane which will become a probe carries flu. Then, as before, one takes the near horizon limit of this configuration.
2.2. From AdS to pp-waves
It was realised a long time.ago by Penrose that an infinitely boosted observer in an arbitrary spacetime, in a neighborhood of its geodesic, sees a very simplified background geometry: the geometry of a gravitational wave. This dramatic simplification has recently been used extensively for a direct check of the gauge-gravity (AdS/CFT) corre~pondence,~ one which avoids the standard strong-weak coupling problems. On the gravity side, the Penrose limit amounts to a suitable rescaling of the coordinates and parameters characterising the (super)gravity solution, in such a way that one focuses on the region close to an arbitrary null geodesic. In the same way in which the background undergoes simplifica-
262
tion, so do different objects present in the initial space. The geometry of the resulting branes can easily be derived by rescaling the embedding equations of the branes in the same way as the target space coordinates. Since the pp-wave space is homogeneous but not isotropic, there are three basic families of D-branes which appear in the limit, depending on the relative orientation of the brane and the wave?’ longitudinal D-branes for which the pp-wave propagates along the worldvolume of the D-brane, transuersal D-branes for which the pp-wave propagates in a direction transverse to the D-brane but the timelike direction is along its worldvolume, and instantonic D-branes for which both the direction in which the pp-wave propagates and the timelike direction are transverse to the D-brane. The first class originates from Ads branes where the geodesic along which the boost was performed belonged to the worldvolume of the brane (before the limit), while for the second case, the brane was co-moving with the observer along the the geodesic (i.e. it was infinitely boosted). The third class of branes can be obtained from the first class by a formal T-duality in the timelike direction of the wave. The pp-wave coordinates split into three groups: the “lightcone coordinates” u and w , and two four-dimensional subspaces with SO(4) x SO(4) isometry group. The split of the transverse coordinates is due to the nonvanishing 5-form flux. In the case of longitudinal branes, the worldvolume coordinates split accordingly into three sets: the “lightcone coordinates” u and v, m coordinates along the first SO(4) subspace and n along the second SO(4). A Dp brane ( m n = p - 1) with such orientation is denoted with (+,-, m, n). The number of preserved supersymmetries depends on the values for (n,m)?
+
0
0
0
+
1/2-BPS D-branes with embedding (+, -, m 2, m ) , for m = 1,.. . ,4, 1/4-BPS D-string with embedding (+, -, 0, 0), non-supersymmetric D-branes with embedding (+,-, m, m ) , for m = 1,2,3.
All these results are valid for the brane placed at the “origin” of the ppwave. If we rigidly move the first or second type of brane outside the origin (without turning worldvolume fluxes), supersymmetry is always reduced t o 1/4. However, the previous three classes do not capture all the branes which can appear in p p - w a v e ~ . In ~ ?the ~ process of Penrose rescaling, not all objects of the initial space will be inherited by the final wave geometry. It
263
is usually said that in order to have a nontrivial Penrose limit of a brane in some background, one needs to take the limit along a geodesic which belongs to the brane. This statement is intuitively understandable: in the Penrose limit an infinitesimal region around the geodesic gets zoomed out. Hence, those parts of the brane which are placed at some nonzero distance from the geodesic get pushed off to infinity. However, this reasoning can be circumvented if the distance between the geodesic and the brane is determined by free parameters of the s ~ l u t i o n In . ~ that case one can take the Penrose limit along a geodesic that does not belong to the brane, as long as the parameter labeling the brane in a family of solutions is appropriately scaled. For example, let us consider the family of solutions corresponding t o two intersecting D-branes and let us take the Penrose geodesic to lie on one of the two branes. Then the Penrose limit of the other brane can be nontrivial if, while taking the Penrose limit of the target space metric, we simultaneously scale the angle between the two branes to zero. It should be emphasized that the final configuration obtained in this way is different from the one which is obtained by first sending the angle to zero and then taking the Penrose limit of the metric. Namely, if we first set the angIe to zero (and hence reduce the problem to taking a limit of orthogonally intersecting branes) the resulting brane will be “flat” (i.e. placed at zi = const. in Brinkman coordinates). However, if we follow the procedure outlined above, the resulting D-brane is a brane with a relativistic pulse propagating on its worldvolume (i.e. with some of xi = const. getting replaced by zi(z+)).The precise form of these worldvolume waves carries information about the position of the brane with respect to the geodesic before the limit was been taken.
3. The Supergravity Approach When the number of branes in some space becomes very large, the probe brane approach is inadequate and a supergravity description takes over. Finding a supergravity solutions for D-branes in AdS spaces is, however, still an open problem; the explicit constructions have been carried out only in a few specific cases. One of the reasons for this is that fully localised supergravity solutions for D-brane intersections in flat space are generically not known. Hence in order to construct the brane solutions in asymptotically Ads and pp-wave spaces, one has to start from scratch. We will present here the construction of the (extremal) D-brane solutions in asymp
264
totically pp-wave spaces. The main difficulty in constructing these solutions consists of identifying a coordinate system where the description of the D-brane is the simplest. This is similar to the problem that one would face if one would only know about Minkowski space in spherical coordinates and would try t o describe flat D-branes in these coordinates. Cartesian coordinates are the natural coordinates to describe infinitely extending D-branes in flat space. So the question that one should first ask is what are the analogues of the Cartesian coordinates for D-branes in pp-wave backgrounds? The answer to this question is more complicated than in flat space, as it depends very much on what kind of D-branes one considers. It was shown8i9 that Brinkman coordinates are the natural coordinates for a description of 1/4-BPS and nonsupersymmetric D-branes, while the natural coordinates for the 1/2BPS D-branes are the Rosen coordinates. For the metric part of the ansatz, one writes a simple standard metric for a superposition of D-branes with waves, ds2 = H(y, y')-i (2du(dv
+ S(z, z', y, y')du)
- dZ2 - d?")
- H(y,y')i(dq2 + d g 2 ) . (4) The metric is given in the string frame, and the D-brane worldvolume co. . ,z", x" = d', . . . ,P ) ,while the directions ordinates are (u, v,zi= d,. transverse to the D-brane are (y" = y', . . . ,Y(~-"), yrA = y'l, . . . ,Y ' ( ~ - ~ ) ) . The function H characterising the D-brane is at this stage allowed to depend on all transverse coordinates y, y'. The ansatz for the RR field strength and the dilaton reads
where '*' in F[S]denotes Hodge duality with respect to the metric (4) and W ( z )is an undetermined function which can depend on all directions transverse to the pp-wave. Also, in the case of the 0 3 brane, one has to add to the form ( 5 ) its Hodge dual. One of the main characteristics and perhaps limitations of this ansatz is that the metric is diagonal in Brinkman coordinates. This property forces
265
one to delocalise the supersymmetric solutions along some directions transverse t o the brane when solving the equations of motion.a The smearing procedure physically means that one is constructing an array of D-branes of the same type with an infinitesimally small spacing. However, as we have seen before, the probe brane results tell us that, unless we turn on additional bulk fluxes (sourced by the worldvolume fluxes of the I/ZBPS D-branes), a periodic array of rigid D-branes in Brinkman coordinates with orientation (+, -,n 2 , n ) is only one quarter supersymmetric. Hence the supersymmetric solutions that we find due to the smearing procedure are only 1/4 BPS. However, these restrictions have to be imposed only on the harmonic function characterising the D-brane, and not on the function characterising the pp-wave. Hence, all our solutions asymptotically tend to the unmodified Hpp-wave. Also, despite the simplicity of the ansatz, the nonsupersymmetric solutions, describing branes with (+,-, m, m) orientation, are fully localised. Pluging the ansatz (4)-(7) into the equations of motion and the Bianchi identities one obtains, depending on the orientation of the branes, solutions with the following characteristics. The presence of the D-brane modifies the function S which characterises the pp-wave, while the function H (which specifies the D-brane) is completely unmodified by the presence of the wave. Therefore, this ansatz does not catch the back-reaction of the pp-wave on the D-brane. For a generic embedding of the D-brane, one expects that the (fully localised) D-brane is modified by the wave. However, as our fully localised, nonsupersymmetric solution demonstrates, this does not have to hold for specific embeddings. By examining the behavior of the radially infalling geodesics, one discovers that if the pure pp-wave was focusing the geodesics, this attractive behaviour is strengthened in the presence of a supersymmetric brane, as one would expect. Surprisingly, however, the non-supersymmetric geometries exhibit repulsion behavior.
+
4. The Algebraic Approach
Rather surprisingly, a modification of the superalgebra of anti-de-Sitter backgrounds which accounts for the presence of D-branes in the string spectrum is still unknown. At an algebraic level, D-branes manifest themselves through non-zero expectation values of bosonic tensorial charges. There exists a widespread, but incorrect, belief that the inclusion of these "This is the same type of restriction that one faces when constructing supergravity solutions for intersecting D-branes, with a simple diagonal ansatz.
266
brane charges into the anti-de-Sitter superalgebras follows the well-known flat-space pattern. In flat space, the inclusion of brane charges leads to a rather minimal modification of the super-Poincar6 algebra: the bosonic tensorial charges appear on the right-hand side of the anti-commutator of supercharges, transform as tensors under the Lorentz boosts and rotations, while they commute with all other generators. The brane charges are therefore often loosely called “central”, and the resulting algebra is referred to as the maximal bosonic “central” extension of the super-Poincar6 algebra. However, despite several attempts to construct a similar modification of anti-de-Sitter superalgebras, a physically satisfactory solution is as of yet unknown. There are two basic physical requirements which have to be satisfied by an anti-de-Sitter algebra which is modified to include brane charges. The algebra has to include at least the brane charges which correspond to all D-branes that are already known to exist, and it also has to admit at least the supergraviton multiplet in its spectrum. Mathematically consistent modifications of anti-de-Sitter superalgebras can be constructed, but all existing proposals fail to satisfy one or both of these physical criteria.1° In Ref. 11 we have identified a simple reason why previous attempts to extend anti-de-Sitter superalgebras with brane charges have failed: such extensions are only physically acceptable when one adds new fermionic brane charges as well. The necessity of including new fermionic brane charges into the modified algebra can be understood from a very simple argument based on Jacobi identities, in combination with the two physical requirements just mentioned.ll Let Consider an anti-de-Sitter superisometry algebra, or a p p wave contraction of it. The bracket of supercharges can, very symbolically, be written in the form
where Q and M are the supercharges and rotation generators respectively (we have grouped together momentum and rotation generators by using a notation in the embedding space). Suppose now that we add a bosonic tensorial brane charge 2 on the right-hand side of this bracket. This extension has to be made consistently with the Jacobi identities. Consider the (Q, Q, 2 ) identity, which takes the symbolic form
267
As the brane charge 2 is a tensor charge, it will transform non-trivially under the rotation generators. This implies that the first term of (9) will not vanish. The Jacobi identity can then only hold if 2 also transforms nontrivially under the action of the supersymmetry generators! (In flat space, only the vanishing bracket [P,Z]appears in the first term of the Jacobi identity (9), because in that case the { Q , Q } anti-commutator closes on the translation generators). The simplest option is to assume that no new fermionic charges should be introduced, and that therefore symbolically
21 = Qa .
(10) Although it is possible to construct an algebra based on (lo), which satisfies all Jacobi identities, it is physically unsatisfactory.1° The essential reason is that brackets like (10) are incompatible with multiplets on which the brane charge is zero (the left-hand side would vanish on all states in the multiplet, while the right-hand side is not zero). In other words, one cannot “turn off” the brane charges. The only other way out is to add new fermionic charges Qb, to the algebra, such that (10) is replaced with [Qa,
[Qa,21 = Q&.
(11) In this case it becomes possible to find representations in which both 2 and the new charge Qb, are realised trivially, as expected for e.g. the supergraviton multiplet, while still allowing for multiplets with non-zero brane charges. This formal argument based on Jacobi identities may come as a surprise, and one would perhaps find it more convincing to see new fermionic brane charges appear in eqcplicit models. In Ref. 11 we have shown that such charges indeed do appear. In order to show this, we have analysed the world-volume superalgebras of the supermatrix model and the supermembrane in a pp-wave limit of the anti-de-Sitter background. These models exhibit, in the absence of brane charges, a world-volume version of the superisometry algebra of the background geometry. When bosonic winding charges are included, the algebra automatically exhibits fermionic winding charges as well. Moreover, configurations on which these charges are non-zero can be found explicitly, or can alternatively be generated from configurations on which the fermionic winding charges are zero. On the basis of these results we have briefly discussed a D-brane extension of the osp* (814) superisometry algebra with bosonic as well as fermionic brane charges, which avoids the problems with purely bosonic modifications as first observed in Ref. 10. A partial construction of this algebra has been carried out in Ref. 12.
268
References 1. A. Bilal and C.-S. Chu, Nucl. Phys. B547,179 (1999), hepth/9810195. 2. J. Gutowski, G. Papadopoulos, and P. K. Townsend, Phys. Rev. D60,106006 (1999), hepth/9905156. 3. A. Karch and E. Katz, JHEP 06, 043 (2002), hep-th/0205236. 4. G. Sarkissian and M. Zamaklar, JHEP 03, 005 (2004), hep-thf0308174. 5. D. Berenstein, J. M. Maldacena, and H. Nastase, JHEP 04, 013 (2002), hepth/0202021. 6. K . Skenderis and M. Taylor, JHEP 06, 025 (2002), hepth/0204054. 7. G. Sarkissian and M. Zamaklar, Symmetry breaking, permutation D-branes o n group manifolds: Boundary states and geometric description, h e p t h/03 12215. 8. P. Bain, P. Meessen, and M. Zamaklar, Class. Quant. G m v . 20, 913 (2003), h e p t h/0205 106. 9. P. Bain, K . Peeters, and M. Zamaklar, Phys. Rev. D67, 066001 (2003), hepth/0208038. 10. P. Meessen, K. Peeters, and M. Zamaklar, O n central eztensions of anti-deSitter alge bras, hep-t h/0302 198. 11. K. Peeters and M. Zamaklar, Phys. Rev. D69, 066009 (2004), hepth/0311110. 12. S. Lee and J.-H. Park, Noncentral extension of the Ads5 x S5 superalgebm: Supermultiplet of brane charges, hep-th/0404051.
ABSTRACTS AND TITLES OF REPORTS NOT INCLUDED IN THE VOLUME
Perturbation of Spectra of Operator Matrices D. Djordjevik Department of Mathematics, Faculty of Sciences, University of Nis, Serbia E-mail: [email protected]. yu
If A , B , C are bounded operators on Banach or Hilbert spaces and
we determined the sets
n
uT(~c),
C
where u7 denotes any of the following part of the spectrum: left (right) spectrum, left (right) essential spectrum, Browder or Weil spectrum.
Boundary Liouville Theory and 2D Quantum Gravity I. Kostov Saclay , France E-mail: [email protected] We study the boundary correlation functions in Liouville theory and
2D quantum gravity. We find that all the fundamental Liouville structure constants obey functional equations similar to the one obtained for the two-point function by Fateev, Zamolodchikov and Zamolodchikov. These take the form of finite-difference equations with respect to the boundary parameters. Then we show how these equations can be derived in a discrete model of 2D quantum gravity and give them a geometrical meaning.
269
270
Manifestation of TeV Gravity A. Nicolaidis Theoretical Physics Department, University of Thessaloniki, Greece E-mail: [email protected] In recent unification models, gravity propagates in 4+d dimensions, while standard model fields are confined to a four dimensional brane. As a consequence, gravity becomes strong not at the Planck scale, but at the TeV scale. We searched for signatures of TeV gravity and extra dimensions in the cosmic rays. We have interpreted the cosmic ray spectrum "knee" (the steepening of the cosmic ray spectrum at energy 10'5.5eV), as due to missing energy from graviton bremsstrahlung. We estimated the graviton production in pp collisions, in the soft graviton approximation. By reproducing the cosmic ray spectrum in the "knee" region, we deduced that the fundamental scale of gravity is M = 8 TeV and the number of extra dimensions is d = 4 (for details see Gen. Rel. Grav. 35, 1117 (2003), hep-ph/0109247). We studied also gravitational scattering in the background of a massive black hole, living in 4+d dimensions. Two regimes appear. For large impact parameters, the deflection angle follows a power-law behavior, reminiscent of the Rutherford-type scattering. For small impact parameters, the deflection angle develops a logarithmic singularity and becomes infinite at a critical b value. This singularity is reflected into a strong enhancement of the backward scattering. We suggest then, as distinctive signature of black hole formation in particle collisions at TeV energies, the observation of the backward scattering events (for details see hep-ph/0307321).
The Rational Topology of Gauge Groups and of Spaces of Connections S. Terzic University of Montenegro, Faculty of Sciences, Cetinjski put bb, 81 000 Podgorica, Serbia Montenegro E-mail: [email protected]. y u In this talk we are going to present general approach which completely solves the problem on computation of the rational homotopy groups and the rational cohomology of the gauge groups and of the space of connections modulo gauge transformations for principal bundles over four-manifolds. The latter are assumed to be equipped with appropriate Sobolev topolo-
271
gies and we also assume that G is a semisimple compact simply connected Lie group and M is compact and simply connected. Note that in some particular cases some of these computations have already been done. Namely, Donaldson computed the cohomology structure of the quotients of spaces of connections for SU(2)-principal bundles over compact simply connected four-manifolds, but the proof essentially uses the fact that the structure group is SU(2). We propose here a general approach which appeals to Sullivan’s minimal model theory. We proceed as follows. First we compute the rational homotopy groups of the gauge group using the result of Singer characterising the weak homotopy type of base point preserving gauge groups, and the result of Milnor giving the homotopy type of a simply connected four-manifold M . Having computed the rational homotopy groups of the quotients of spaces of connections, the nilpotency of the space of connections modulo based point gauge transformations group will make it possible to apply Sullivan’s minimal model theory for the cohomology computation.
String Theory and Matrix Models A . Morozov
Strong Interactions and Stability in Quasi Localized Gravity R. Rattazzi
Strings and Branes in Non-compact Backgrounds V. Schomerus
Two Dimensional Gravity and Liouville Field Theory A . B. Zamolodchikov
This page intentionally left blank
273
Acknowledgment There are many people and institutions we would like to acknowledge at the end of this “scientific adventure”. Many thanks are due to all authors and participants for their cooperation and to the members of the International Advisory Committee, including those who could not attend. The crucial financial support came from UNESCO-ROSTE Venice, and it was a great privilege to work together with Mr. Vladimir Kouzminov, deputy director of this office. Kind assistance of Mrs. Iulia Nechifor from the same office and Ms. Vesna Filipovic-Nikolic from the Permanent Delegation of Serbia and Montenegro in UNESCO-Paris is also acknowledged. The participation of the researchers from developing countries, i.e. their travel expenses, was substantially supported by International Centre for Theoretical Physics (ICTP) - Trieste. We are especially indebted to Prof. Goran Senjanovic, Prof. Seif Randjbar-Daemi and Prof. Faheem Hussain for this support. Optimal conditions for the workshop were provided through joint applications of Institute of Physics, Humboldt University, Berlin and Department of Physics NiS, to Deutscher Akademischer Austauschdienst (DAAD) coordinated by Prof. Dieter Luest, as well as of Department of Physics, LMU-Munich with Physics Department NiS, coordinated by Prof. Julius Wess, to Deutsche Forschungsgemeinschaft (DFG). The support of Serbian Ministry of Science, Technology and Development was greatly appreciated. It was an encouraging sign for the future support of fundamental research. Many thanks are due t o the Assistant Minister, Aleksandar Belic. Last but not least, our gratitude goes out to the International Association of Mathematical Physics (IAMP) and its President, Prof. David Brydges, as well as to JAT Airways company and Faculty of Science, NiS, for their donation in the early stages when the status of the whole conference was questionable. Finally, this Balkan Workshop could not have been realised without great devotion and efforts of the students of Department of Physics, NiS, D. Dimitrijevic, J. Stankovic and G. Stanojevic. For the moral support and kind attendance to the Conference, we are most grateful to Prof. Martin Huber, President of the European Physical Society (EPS), Prof. Ilija Savic, President of the Serbian Physical Society and academician Zvonko Maric, Institute of Physics, Belgrade. For the Organizing Committee Goran DjordjeviC NiS, September 2004
This page intentionally left blank
275
Sponsors We would like to thank the following for support: UNESCO - ROSTE (Regional Bureau for Science in Europe, Venice) ICTP (The Abdus Salam International Centre for Theoretical Physics), Italy DFG Deutsche Forschungsgemeinschaft, Germany Ministry of Science, Technologies and Development, Republic of Serbia DAAD (Deutscher Akademischer Austausch Dienst), Germany IAMP (International Association of Mathematical Physics) JAT (Yugoslav Airlines)
-
This page intentionally left blank
277
Southeastern European Network in Theoretical and Mat hematical Physics Statement of Intention Participants of BALKAN WORKSHOP BW2003 Mathematical, Theoretical and Phenomenological Challenges Beyond Standard Model: Perspectives of Balkans Collaboration (VrnjaEka Banja, Serbia, August 29 September 2, 2003.) from Bulgaria, Greece, Moldova, Romania and Serbia and Montenegro, initiate the mobility-teaching-research NETWORK IN THEORETICAL AND MATHEMATICAL PHYSICS of Southeast part of Europe. Our goal is to establish closer relations between science faculties and research institutes and individual scientists in the region of Southeastern Europe. We are going to improve the teaching in mathematical and theoretical physics on undergraduate and postgraduate levels, as well as, joint scientifical work. The initiators propose establishing a few different working groups of particular interest in selected topics (quantum field theory, gravity, statistical physics, dynamical systems etc.). This initiative is and will be open to all institutions and individual scientists from the region who share the ideas of the Network. We expect that scientists and scientific institutions from the region (of Southeast Europe, in particular: Albania, Bosnia and Herzegovina, Croatia, Former Yugoslav Republic of Macedonia, Turkey) will join this initiative in the near future. Scientists from all over the world are welcome to join the Network activities. The proposed collaboration will be realized in the following ways: through the exchange of professors with maximal duration of one semester (3 months); 0 exchange of students (preferably PhD students, or for preparation of a diploma work/thesis, or Master degree); 0 joint organization of various meetings (workshops, conferences and schools). The main joint event could be the “Summer Institute in Theoretical and Mathematical Physics”. We are expecting to organize it every year in different countries, based on the rotational scheme. The duration of the institute could be 2-4 weeks. The institute should consist of different events (seminars, workshops, schools and so on); 0 short visits: from a couple of days to a couple of weeks, including seminars and research; 0 support of the publications of
278
proceedings, books and other issues and other means for the exchange of scientific information. The participants agree to support the foundation of the Initiative Committee to oversee the establishment and initial development of the Network with the following composition: Boyka Aneva (INRNE and University of Sofia, Bulgaria), Goran DjordjeviC (University of NiS, SCG), Argyris Nicolaidis (University of Thessaloniki, Greece), Corneliu Sochichiu (IAP, Chisinau, Moldova, INFN F'rascati, Italy), Mihai Visinescu (Bucharest, Romania). The membership of the Committee will be extended to assure the participation of all countries which join the Network. One of the prime objectives of this Committee will be the elaboration of the terms of the reference of the Network. The meeting of the Initiative Committee will be held during next 1 2 months. It was agreed that for this initial period, the Faculty of Science and Mathematics of the University of NiS would take responsibility for the coordination the Network activities. A Steering Committee consisting of outstanding and internationally leading researchers from the region and from all over the world is to be established. Based on previous contacts it is expected that financial support will be provided by: 0 local institutions, ministries and other foundations from our countries; 0 European Union, different European foundations, large national foundations in Western Europe; 0 international Institutions, like UNESCO, ICTP, CERN, IAE, JINR, Central European Initiative; private foundations.
Vrnjaeka Banja, Serbia and Montenegro, September 2, 2003.
279
Epilogue This book has almost bean read and is about t o be closed. The last participants of the 2003 conference in VrnjaEka, Banja have left this beautiful resort a long time ago. In the meantime, many new findings and equations have been brought to light, many new papers have been published, and many new relations have been established. However, physicists all around the world, in particular, young ones from the Balkan region (or Southeastern Europe, as others may prefer to call it) are still in need of scientific contact, books, equipment, exchange of ideas and new knowledge. We do hope that Balkan Workshop series and the Southeastern European Network in Mathematical and Theoretical Physics (SEENET-MTP, http://seenet-mtp.pmf.ni.ac.yu)will be proved helpful and valuable in this respect, and it should be our imperative that their scope of impact be broadened. In the course of preparatory activities for the next in the series of the Balkan workshops - BW2005 (http://www.pmf.ni.ac.yu/bw2OO5) we learnt about the Nordic Network Discovery Physics at the LHC (http://www.hep.lu.se/nlhc). The example of the Nordic countries should be kept in mind when tracing further paths of our cooperation. As always, time will be the most crucial judge of all our efforts and endeavors. The people in the Initiative and Steering/Advisory Committee, together with all their friends, supporters, physicists and mathematicians of “good will” will try to offer their own contribution to this long and difficult process of making our beautiful region a respectable part of the world community or a t least, by asking, like billions before us did, what the universe in fact is. Welcome to the SEENET-MTP!