Open Issues in Local Star Formation, Vol. 299 (Astrophysics and Space Science Library)

OPEN ISSUES IN LOCAL STAR FORMATION

ASTROPHYSICS AND SPACE SCIENCE LIBRARY VOLUME 299

EDITORIAL BOARD Chairman W.B. BURTON, National Radio Astronomy Observatory, Charlottesville, Virginia, U.S.A. ([email protected]); University of Leiden, The Netherlands ([email protected]) Executive Committee J. M. E. KUIJPERS, Faculty of Science, Nijmegen, The Netherlands E. P. J. VAN DEN HEUVEL, Astronomical Institute, University of Amsterdam, The Netherlands H. VAN DER LAAN, Astronomical Institute, University of Utrecht, The Netherlands MEMBERS I. APPENZELLER, Landessternwarte Heidelberg-Königstuhl, Germany J. N. BAHCALL, The Institute for Advanced Study, Princeton, U.S.A. F. BERTOLA, Universitá di Padova, Italy J. P. CASSINELLI, University of Wisconsin, Madison, U.S.A. C. J. CESARSKY, Centre d'Etudes de Saclay, Gif-sur-Yvette Cedex, France O. ENGVOLD, Institute of Theoretical Astrophysics, University of Oslo, Norway R. McCRAY, University of Colorado, JILA, Boulder, U.S.A. P. G. MURDIN, Institute of Astronomy, Cambridge, U.K. F. PACINI, Istituto Astronomia Arcetri, Firenze, Italy V. RADHAKRISHNAN, Raman Research Institute, Bangalore, India K. SATO, School of Science, The University of Tokyo, Japan F. H. SHU, University of California, Berkeley, U.S.A. B. V. SOMOV, Astronomical Institute, Moscow State University, Russia R. A. SUNYAEV, Space Research Institute, Moscow, Russia Y. TANAKA, Institute of Space & Astronautical Science, Kanagawa, Japan S. TREMAINE, CITA, Princeton University, U.S.A. N. O. WEISS, University of Cambridge, U.K.

OPEN ISSUES IN LOCAL STAR FORMATION Edited by JACQUES LÉPINE University of São Paulo, Brazil and

JANE GREGORIO-HETEM University of São Paulo, Brazil

KLUWER ACADEMIC PUBLISHERS NEW YORK, BOSTON, DORDRECHT, LONDON, MOSCOW

CD-ROM available only in print edition eBook ISBN: 1-4020-2600-5 Print ISBN: 1-4020-1755-3

©2004 Springer Science + Business Media, Inc. Print ©2003 Kluwer Academic Publishers Dordrecht All rights reserved No part of this eBook may be reproduced or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, without written consent from the Publisher Created in the United States of America

Visit Springer's eBookstore at: and the Springer Global Website Online at:

http://www.ebooks.kluweronline.com http://www.springeronline.com

Contents

Contents of the CD-ROM

1

Foreword

7

List of Participants

9

Part I

Stellar Groups and Associations

Major Unsolved Problems in Star Formation (invited talk) Hans Zinnecker

17

The stellar mass function of Galactic clusters and its evolution Guido De Marchi

25

Encounters and Close Fly-Bys of Galactic Clusters and Associations in the past 50 Myr 33 V.V. Makarov Solar-Type Post-T Tauri Stars in the Nearest OB Subgroups Eric E. Mamajek

39

NGC 2362: The shape of the pre-main-sequence from A-stars to brown dwarfs. 47 A. Moitinho, C.J. Lada, N. Hu´elamo, J. Alves Young stars and their circumstellar disks in the σ Orionis cluster J.M. Oliveira, R.D. Jeffries, J.Th. van Loon, M.J. Kenyon

55

The Oph-Sco-Lup-Cen-Cru-Mus-Cha star-formation region (invited talk) Jacques R. D. L´epine, Mar´ılia J. Sartori

63

The gas-to-dust ratio and metallurgy of nearby dark clouds probed by X-ray 73 absorption measurements (invited talk) Thierry Montmerle, M˜y H` a Vuong SACY - a Search for Associations Containing Young stars 83 Carlos A. O. Torres, Germano R. Quast, Ramiro de la Reza, Licio da Silva, Claudio H. F. Melo, Michael Sterzik Age determination of the Ursa Major Association Brigitte K¨ onig

v

91

vi Part II

OPEN ISSUES IN LOCAL STAR FORMATION Young Stellar Objects

Accretion Powered Emission in Young Stellar Objects (invited talk) Nuria Calvet, James Muzerolle

99

The pre-main sequence spectroscopic binary AK Sco 107 S. H. P. Alencar, L. P. R. Vaz, C. H. F. Melo, C. Dullemond, J. Andersen, C. Batalha, R. D. Mathieu Survey of Young Stellar Objects associated with Molecular Clouds 115 Z. Abraham, A. Roman-Lopes, J. R. D. L´epine, T. Dominici, A. Caproni The Stellar population of Embedded Galactic Massive Star Clusters A. Damineli, R. D. Blum, P. S. Conti, E. Figuerˆedo, C. L. Barbosa

121

Spectroscopic Analysis of 131 Herbig Ae/Be Candidate Stars 127 S. L. A. Vieira W. J. B. Corradi S. H.P. Alencar L. T. S. Mendez , C. A. Torres, G. Quast, M. M. Guimar˜ aes, L. da Silva Classification of the Pico dos Dias Survey Herbig Ae/Be stars Mar´ılia J. Sartori, Jane Gregorio-Hetem, Annibal Hetem Jr.

133

Chemical composition study of an accretion episode in the Herbig candidate star PDS080 141 M. M. Guimar˜ aes, S. L.A. Vieira, S. H.P. Alencar, W. J.B. Corradi Magnetically channeled accretion in T Tauri stars (invited talk) J. Bouvier, S.H.P. Alencar, C. Dougados

147

Properties of Young Stellar Objects from High Resolution Near Infrared Spectroscopy 159 G.W. Doppmann, D.T. Jaffe, R.J. White Probing the circumstellar structure of pre-main sequence stars Jorick S. Vink, Janet E. Drew, Tim J. Harries, Ren´e D. Oudmaijer

169

Accretion signatures in the X-ray spectrum of TW Hya Beate Stelzer, J¨ urgen H. M. M. Schmitt

177

Post-T Tauri Stars Rotation in Associations Ramiro de la Reza, Giovanni Pinz´ on

185

Star Formation in Canis Majoris Jane Gregorio-Hetem, Thierry Montmerle, Edson R. Marciotto

193

X-ray emission from NGC 2362 199 Nuria Hu´elamo, Beate Stelzer, Andr´e Moitinho, Jo˜ ao F. Alves, Charles Lada Accretion and Ejection (invited talk) Lee Hartmann

205

vii

Contents

The origin of jets in young stars 213 Catherine Dougados, Jonathan Ferreira, Nicolas Pesenti, Sylvie Cabrit, Paulo Garcia, Darren O’Brien Non coeval young multiple systems? Gaspard Duchˆene, Andrea Ghez, Caer McCabe Part III

223

Brown Dwarfs

Brown dwarfs in young open clusters Estelle Moraux, J´erˆ ome Bouvier

235

Brown Dwarf Companions Michael F. Sterzik, Richard H. Durisen

245

Observational Clues to Brown Dwarf Origins 251 Ray Jayawardhana, Subhanjoy Mohanty, Gibor Basri, David R. Ardila, Beate Stelzer, Karl E. Haisch, Jr. Surface Gravity & Mass in Young Brown Dwarfs and Planemos 259 Subhanjoy Mohanty , Ray Jayawardhana , Gibor Basri , France Allard , Peter Hauschildt, David Ardila Part IV

Disks, Outflows and Jets

FU Orionis Eruptions and the Formation of Close Binaries (invited talk) 269 Bo Reipurth SEDs of Flared Dust Disks Michaela Kraus

279

On the alignment of T Tauri stars with the local magnetic field Gaspard Duchˆene, Franc¸ois M´enard

287

Temporal evolution of main sequence dusty disks Manoj Puravankara, H. C. Bhatt

295

Numerical Simulations of Young Stellar Jets: From One Tenth to Parsec Scales 303 E. M. de Gouveia Dal Pino, A. C. Raga, A. H. Cerqueira, E. Masciadri Alfven Waves in Disks, Outflows and Jets Reuven Opher

311

viii Part V

OPEN ISSUES IN LOCAL STAR FORMATION Early Stages of Star Formation

Submillimeter Studies of Protostellar Cores (invited talk) Philippe Andr´e Detection of a collimated jet towards the high-mass protostar IRAS 16547−4247 Kate J. Brooks, Guido Garay, Diego Mardones

319

331

Disks and Halos in Pre-Main-Sequence Stars ˇ Ivezi´c, M. Elitzur A. S. Miroshnichenko, D. Vinkovi´c, Z.

339

Radiative Transfer in Prestellar Cores: a Monte Carlo approach D. Stamatellos, A. P. Whitworth

347

Part VI

The ISM Conditions for Star Formation

The Initial Conditions for Star Formation (invited talk) Diego Mardones

359

Chemical signatures of new star formation toward young stellar clusters Jonathan Williams, Sandrine Bottinelli, Carlos Rom´ an-Z´ un ˜iga

367

Collapse of Molecular Clouds Leading to Star Formation A. Allen, H. W. Chuang, M. Choi, Z. Y. Li

375

Star formation in globules 383 Rolf Chini, Marcus Albrecht , Luis Barrera, Katrin K¨ ampgen, Markus Nielbock SIMBA survey toward high-mass star forming regions in the southern hemisphere 389 Santiago Fa´ undez, Leonardo Bronfman, Guido Garay , Rolf Chini , Lars-¨ ake Nyman Conference summary and conclusions

397

Author Index

401

CONTENTS OF THE CD-ROM POSTER CONTRIBUTIONS Session I: Stellar Groups and Associations Photometric and Spectroscopic Study of Young Galactic Open Clusters: Search for Pre-Main Sequence Stars Balog Z., Vinko J., Kenyon S.J. The low-mass, dispersed young stellar population in Orion OB1 Brice˜ no C., Calvet N., Hartmann L., Vivas A.K. New Members of the Galactic OB Association Bochum7 Corti M., Bosch G., Niemela V. New catalogue of optically visible open clusters and candidates Dias W.S., Alessi B.S., L´ epine J., Moitinho A. Li Abundance of Post-T Tauri Stars from the Li I Lines at 6104A and 6708A Drake N.A., de la Reza R., Quast G.R., da Silva L., Torres C.A.O. NTT infrared imaging of star cluster candidates towards the central parts of the Galaxy Dutra C.M., Ortolani S., Bica E., Barbuy B., Zoccali M., Momany Y. The Stellar Content of the Obscured Galactic Giant HII Region G333.1-0.4 Figuerˆedo E., Damineli A., Blum R.D., Conti P.S. η Muscae: a multiple system with a PMS component Hensberge H., Nitschelm C., Bouzid M.Y., Clausen J.V., David M., Freyhammer L.M., Helt B., Olsen E.H., Sterken C., Vaz L.P.R. A deep survey of the M17 cluster Hoffmeister V.H., Chini R. Radial velocities of hot stars in the SCO-CEN Association Jilinski E., Cunha K., de la Reza R. Str¨ omgren-Hβ photometry of OB Clusters and Associations: the Norma Field Kaltcheva N. A Gallery of Neglected Nearby Star Forming Regions Kun M.

1

2

OPEN ISSUES IN LOCAL STAR FORMATION

A photographic study of photometric variability in the M42 region Lima G.H.R.A., Vaz L.P.R., Reipurth B. TWA as a Rapidly Expanding Mini-Association and Vega as Its Possible Member Makarov V.V., Gaume R.A. A Calibration of EK Cephei Marques J.P., Fernandes J., Monteiro J.P.F.G. Investigating the accretion-rotation link in the low-mass members of the Orion Nebular Cluster Melo C., Bouvier J., Delfosse X., Pasquini L. The binary population in the Sco-Cen Complex: the present state of knowledge and preparation of future research Nitschelm C. The dusty side of NGC 3603: TIMMI2 data reveal a striking diversity of mid IR sources Nuernberger D., Stanke T. Lithium depletion and age of NGC 2547 Oliveira J.M., Jeffries R.D., Devey C.R., Navascu´es D.B., Naylor T., Stauffer J.R., Totten E.J. A Method to Search for Associations of Young Stars Quast G.R., Torres C.A.O., Melo C.H.F., Sterzik M., de la Reza R., Silva L. X-Ray Emission During the Early Stages of Star Formation: New Results from Chandra and XMM-Newton Skinner S., Gagn´e M., Gudel M., Zhekov S., Barbosa C. Optical Astrometry of X-ray sources and possible PMS stars candidates Teixeira R., Ducourant C., Sartori M.J., Benevides-Soares P., Mui˜ nos J.L., P´ eri´e J.P., Guibert J., Mallamaci C.

Session II: Young Stellar Objects The Metallicity of post-T Tauri Stars: Preliminary Results Almeida R., Oliveira I., de la Reza R., Silva L., Torres C.A., Quast G. Infrared Survey for the Characterization of Stellar Population of: Spiral Arms and of a Bar at the Galactic Center Amˆ ores E.B., L´ epine J.R.D. Mid-Infrared Imaging of Massive YSOs in NGC 3576 Barbosa C., Damineli A., Blum R.D., Conti P.S.

Contents of the CD-ROM

3

Observation of flourescence process in T-Tauri stars, HAeBe stars and symbiotic stars Camperi J., Pereira C.B. Atmospheric Parameters and Metallicity for a sample of HAEBE stars Castilho B.V., Sartori M.J., Daflon S. The possible connection of PDS Herbig Ae/Be stars and the main star forming regions Corradi W.J.B., Guimar˜ aes, M., Vieira, S., Torres, C.A. The Evolutionary Stage of 3 Southern Galactic Unclassified B[e] Stars: HD45667, HD50138 and HD87643. Are They Herbig Ae/Be Stars? Fernandes M.B., Ara´ ujo F.X. A Near-Infrared Multiplicity Survey of Class I/Flat-Spectrum Systems in Six Nearby Molecular Clouds Haisch Jr. K.E., Greene T.P., Barsony M. An ISOCAM mid-IR survey of TMC-2: Searching for sub stellar objects Hony S., Prusti T., Ott S. Brazilian participation in the COROT project Janot-Pacheco E. On the structure of magnetic field of T Tau Lamzin S.A., Smirnov D.A., Fabrika S.N. Two-stream accretion model for CTTS Lamzin S.A., Kravtsova A.S. High-resolution spectroscopy of Herbig Ae/Be stars Miroshnichenko A., Bjorkman K.S., Klochkova V.G., Panchuk V.E., Gray R.O. Anusual phenomena in the circunstellar envelope of an unique magnetic Ae star HD 190073 Pogodin M.A., Franco G.A.P., Lopes D.F. Chemical Abundances of Weak T Tauri Stars Rojas G.A., Gregorio-Hetem J. Discovery of a Young Massive Stellar Cluster associated with IRAS source 16177-5018 Roman-Lopes A., Abraham, Z., L´ epine, J. Spectral Variability of Southern T Tauri Stars: GQ Lupi Seperuelo E., Batalha C., Lopes D. Testing Pre-Main Sequence Evolution Theory: Discovery and Analysis of a Young, Low-Mass Eclipsing Binary Stassun K.G., Vaz L.P.R., Mathieu R.D., Stroud N.S.

4

OPEN ISSUES IN LOCAL STAR FORMATION

Improvements in the Wilson-Devinney Codes for Eclipsing Binaries Vieira L.A., Vaz L.P.R. Young A7e Herbig star HD 144432: typical and unusual characteristics of its spectral behaviour Vieira S., Pogodin M.A., Guimar˜ aes, M.

Session III: Disks, Outflows and Jets Line Emission from the Accretion Shock in Classical T Tauri Stars Ardila D.R., Johns-Krull C. Proper Motions and New HH Knots in the L1551 region Armond A.C., Reipurth B., Vaz L.P.R., Bally J. Precession in pulsed jets from young stellar objects: a glimpse in the MHD case Cerqueira A.H., Dal Pino E.M.G. Radio-Continuum Emission from stellar flows in low mass stars Gonz´ alez R.F., Cant´ o J. T Tauri Disk Model: Confidence Levels for Parameter Estimation Hetem Jr. A., Gregorio-Hetem J. Rotational evolution of T Tauri Stars using two polytropes Pinz´ on G., de la Reza R. Heating of magnetic tubes of young low mass stars driven by damped Alfv´en waves Vasconcelos M.J., Jatenco-Pereira V., Opher R.

Session IV: The ISM Conditions for Star Formation and Early Stages of Star Formation Magnetic Field Geometry in the Region of the Star Forming Cloud Lupus 1 Alves F.O., Franco G.A.P. The Minimum Mass for Opacity-Limited Fragmentation Boyd D., Whitworth A.P. Protostellar Objects in M17 - shadows of a star birth Chini R., Hoffmeister V.H., Schilp S. The role of dust-cyclotron damping of Alfv´en waves in star formation regions Falceta-Gon¸calves D., de Juli M.C., Jatenco-Pereira V.

Contents of the CD-ROM The structure of the Cometary Globule CG 12, NGC 5367 Haikala L.K., Reipurth B., Olberg M. C18O mapping of Chameleon I cloud with SEST Haikala L.K., Harju J., Mattila K., Toriseva M. SIMBA observations of the R Corona Australis molecular cloud Kampgen K., Chini R., Reipurth B. On the Formation of Massive Stellar Clusters Tenorio-Tagle G., Palous J., Silich S., Medina-Tanco G.A., Mu˜ noz-Tu˜ no ´n C. Rotating Low-Mass Stellar Models with Angular Momentum Redistribution Mendes L.T.S., Vaz L.P.R., D’Antona F., Mazzitelli I. The Kinematics of Massive Star Forming Regions in Early Stages of Evolution Pineda J., Garay G., Mardones D. Copper and Zinc in Bulge-like Stars Pomp´ eia L. Polarization of Herbig Ae/Be candidates and their environment Rodrigues C., Sartori M., Gregorio-Hetem J., Magalha˜es A.M., Batalha C.

SUPPLEMENTS Further considerations of the formation of the β Pictoris moving group Oral Contribution from Vladimir Ortega SACY - a Search for Associations Containing Young stars Media player files (Torres et al.) Star formation in Canis Majoris Appendix I (Gregorio-Hetem et al.) Collapse of molecular clouds leading to star formation Colored figures (Allen et al.) Radiative transfer in prestellar cores Colored figures (Stamatellos & Whitworth)

5

This page intentionally left blank

Foreword The international Colloquium “Open Issues in Local Star Formation and Early Stellar Evolution” was held in Ouro Preto (Brazil) from April 05 to 10, 2003. The Colloquium took place in the “Parque Metal´ urgico”, an old iron industry that has been transformed into a nice modern conference center. Ouro Preto is a 18th century colonial city that has been declared a Cultural Heritage of Mankind in 1980. It is situated in the hills of the State of Minas Gerais at about 100 km from Belo Horizonte. The meeting was attended by 115 participants from 15 countries. The participants were in general very happy with the high level of the presentations and with the friendly ambiance of the discussions. The talks and poster sessions were focused on the physics of young stellar objects, which are being observed with increasing angular resolution provided by the new generation of telescopes, and on the processes that triggered large scale star-formation in the solar neighborhood or in the Galaxy. The motivation for organizing the Colloquium at Ouro Preto in 2003 was the growing interest of the Brazilian community in this field. This interest was triggered in part by the Pico dos Dias Survey (PDS). The PDS discovered a large number of new T Tauri and Herbig Ae/Be stars, as well as isolated groups of T Tauri stars. Several groups of researchers devoted efforts to understand the mechanisms of star-formation, and confronted the available models with the observed space motions. Other members of the community have directed their research towards individual objects, studying their spectral energy distribution, the evolution of circumstellar disks, and the mechanism of outflows. The Pico dos Dias Observatory and many of the researchers devoted to star- formation are in Minas Gerais. This was a good reason for looking for a convenient city in that State, far from the sea, contrary to the previous international astronomy meetings in Brazil. Another motivation for the organization of the Colloquium was the need to prepare the star-formation community to the use of the SOAR (4m) telescope which will be soon operating, with 30% of Brazilian time. The meeting was sponsored by the federal research agency CNPq, and the research agencies of the States of S˜ao Paulo, Rio de Janeiro, and Minas Gerais States (FAPESP, PAPERJ, FAPEMIG). We acknowledge the support from the involved universities: Universidade de S˜ ao Paulo (IAG/USP) and Universidade Federal de Minas Gerais (UFMG), and research institutes: Laborat´ orio Nacional de Astrof´ısica (LNA/MCT) and Observat´ orio Nacional (ON/MCT). All the members of the Scientific Organizing Committee had an important role in suggesting names of invited speakers and giving advice on

7

8

OPEN ISSUES IN LOCAL STAR FORMATION

how to organize the scientific program. Unfortunately, Bruce Elmegreen, Charles Lada and Celso Batalha could not come. Besides the members of the Local Organizing Committee, several people contributed strongly to the success of the Colloquium. We should mention Annibal Hetem Jr., who kindly took care of the projection systems and prepared the CDROM for these Proceedings. We also thank Sonia P. de Carvalho, who gave assistance to secretaries of the meeting, and organized the city tour in the touristic Tuesday afternoon and the musical event at the “Teatro Municipal”. A special thank goes to the UFMG students, who received the participants at the Belo Horizonte Airports and gave logistic support during the whole meeting. In particular, we thank Natalia Landin, David Carvalho and Geraldo Oliveira for the help with the microphones along the oral presentations. A very special thank goes the secretaries of the meeting: Marina Freitas, Marcia Pina Albe, and Maria Jos´e F. Ferrer.

Scientific Organizing Committee C. Batalha (Brazil) N. Calvet (USA) R. de la Reza (Brazil) B. Elmegreen (USA) G. Franco (Brazil) J. Gregorio-Hetem (Brazil) C. Lada (USA) J. L´epine (Brazil) B. Reipurth (USA) C. A. Torres (Brazil) H. Zinnecker (Germany)

Local Organizing Committee S. Alencar (Universidade Federal de Minas Gerais) W. Corradi (Universidade Federal de Minas Gerais) K. Cunha (Observat´ orio Nacional) J. Gregorio-Hetem (Universidade de S˜ ao Paulo) M. J. Sartori (Laborat´ orio Nacional de Astrof´ısica) S. L. Vieira (Universidade Federal de Minas Gerais) L. P. Vaz (Universidade Federal de Minas Gerais)

List of Participants ABRAHAM Zulema Universidade de S˜ ao Paulo, Brazil AGUIAR Bruno Universidade Federal de Minas Gerais, Brazil ALENCAR Silvia Universidade Federal de Minas Gerais, Brazil ALLEN Anthony Ac. Sinica Inst. of Astron. and Astroph., Taiwan ALMEIDA Roberta Observat´ orio Nacional, Brazil ALVES Felipe Universidade Federal de Minas Gerais, Brazil ALVES Jo˜ ao European Southern Observatory, Germany ˆ AMORES Eduardo Universidade de S˜ ao Paulo, Brazil ´ Philipe Service d’Astrophysique, CEA, France ANDRE ARDILA David Johns Hopkins University, USA ARMOND Ana Cristina Universidade Federal de Minas Gerais, Brazil BALOG Zoltan University of Szeged, Hungary BARBOSA Cassio Universidade de S˜ ao Paulo, Brazil BARBUY Beatriz Universidade de S˜ ao Paulo, Brazil BOUVIER J´ erˆ ome Observatoire de Grenoble, France BOYD Douglas Cardiff University, UK BROOKS Kate Universidad de Chile, Chile CALVET Nuria Smithsonian Astrophysical Observatory, USA CAMPERI Javier Observat´ orio Nacional, Brazil CARVALHO David Universidade Federal de Minas Gerais, Brazil CASALI Mark Royal Observatory Edinburgh, UK CASTILHO Bruno Laborat´ orio Nacional de Astrof´ısica, Brazil CERQUEIRA Adriano Universidade Estadual de Santa Cruz, Brazil CHAVARRIA Luis Universidad de Chile, Chile CHINI Rolf Astronomisches Institut Bochum, Germany CORRADI Wagner Universidade Federal de Minas Gerais, Brazil CORTI Mariela Universidad Nacional de La Plata, Argentina DA SILVA Licio Observat´ orio Nacional, Brazil DAMINELI Augusto Universidade de S˜ ao Paulo, Brazil DE GOUVEIA DAL PINO Elisabete Universidade de S˜ ao Paulo, Brazil DE LA REZA Ramiro Observat´ orio Nacional, Brazil DE MARCHI Guido European Space Agency, USA DIAS Wilton Universidade de S˜ ao Paulo, Brazil DOPPMANN Gregory NASA Ames Research Center, USA DOUGADOS Catherine Observatoire de Grenoble, France DRAKE Natalia Observat´ orio Nacional, Brazil DUCHENE Gaspard UCLA, USA FALCETA-GONC ¸ ALVES Diego Universidade de S˜ ao Paulo, Brazil FAUNDEZ Santiago Universidad de Chile, Chile FERNANDES Marcelo Observat´ orio Nacional , Brazil ˆ FIGUEREDO Elysandra Universidade de S˜ ao Paulo, Brazil FRANCO Gabriel Universidade Federal de Minas Gerais, Brazil GARAY Guido Universidad de Chile, Chile GONZALEZ Ricardo Universidade de S˜ ao Paulo, Brazil GREGORIO-HETEM Jane Universidade de S˜ ao Paulo, Brazil ˜ GUIMARAES Marcelo Universidade Federal de Minas Gerais, Brazil HAIKALA Lauri European Southern Observatory, Chile HAISCH Karl University of Michigan, USA

9

10

OPEN ISSUES IN LOCAL STAR FORMATION

HARTMANN Lee Smithsonian Astrophysical Observatory, USA HENSBERGE Herman Royal Observatory of Belgium, Belgium HETEM JR. Annibal UNIP - Universidade Paulista, Brazil HOFFMEISTER Vera Astronomisches Institut Bochum, Germany HONY Sacha European Space Agency, The Netherlands ´ HUELAMO Nuria European Southern Observatory, Chile JANOT-PACHECO Eduardo Universidade de S˜ ao Paulo, Brazil JATENCO-PEREIRA Vera Universidade de S˜ ao Paulo, Brazil JAYAWARDHANA Ray University of Michigan, USA JILINSKI Evgueni Laborat´ orio Nacional de Computa¸ca ˜o Cient´ıfica, Brazil KAEMPGEN Katrin Astronomisches Institut Bochum, Germany KALTCHEVA Nadejda University of Wisconsin, USA ¨ KONIG Brigitte Max-Planck Institut fuer Extraterrestrische Physik, Germany KRAUS Michaela University Utrecht, The Netherlands KUN Maria Konkoly Observatory, Hungary LAMZIN Sergei Sternberg Astronomical Institute, Russia LANDIN Natalia Universidade Federal de Minas Gerais, Brazil ´ LEPINE Jacques Universidade de S˜ ao Paulo, Brazil LIMA Gustavo Universidade Federal de Minas Gerais, Brazil LYRA Vladimir Observat´ orio do Valongo, Brazil MAKAROV Valeri Universities Space Research Assoc., USA MAMAJEK Eric University of Arizona, USA MANOJ Puravankara Indian Institute of Astrophysics, India MARDONES Diego Universidad de Chile, Chile MARQUES Jo˜ ao Centro de Astrof´ısica da U. Porto, Portugal MEDINA TANCO Gustavo Universidade de S˜ ao Paulo, Brazil MELO Claudio European Southern Observatory, Chile MENDES Luiz Universidade Federal de Minas Gerais, Brazil MIROSHNICHENKO Anatoly University of Toledo, USA MOHANTY Subhanjov Harvard-Smithsonian Center for Astrophysics, USA MOITINHO Andr´ e Observat´ orio Astron´ omico de Lisboa, Portugal MONIN Jean-Louis Observatoire de Grenoble, France MONTMERLE Thierry CEA/Saclay & Observatoire de Grenoble, France MORAUX Estelle Observatoire de Grenoble, France NIEVA Maria Fernanda Observat´ orio Nacional, Brazil NICKELER Dieter Institute for Astrophysics and Space Science, Germany NITSCHELM Christian University of Antwerp, Belgium NUERNBERGER Dieter European Southern Observatory, Chile OLIVEIRA Isabel Observat´ orio Nacional, Brazil OLIVEIRA Cristina Universidade Federal de Minas Gerais, Brazil OLIVEIRA Geraldo Universidade Federal de Minas Gerais, Brazil OLIVEIRA Joana Keele University, UK OPHER Reuven Universidade de S˜ ao Paulo, Brazil ORTEGA Vladimir Observat´ orio Nacional, Brazil PINEDA Jaime Universidad de Chile, Chile ´ Giovanni Observat´ PINZON orio Nacional, Brazil ´ POMPEIA Luciana Universidade de S˜ ao Paulo, Brazil QUAST Germano Laborat´ orio Nacional de Astrof´ısica, Brazil REIPURTH Bo Institute for Astronomy, USA ROJAS Gustavo Universidade de S˜ ao Paulo, Brazil

List of Participants

11

ROMAN-LOPES Alexandre Universidade de S˜ ao Paulo, Brazil SAMPSON Leda Observat´ orio Nacional, Brazil SARTORI Mar´ılia Laborat´ orio Nacional de Astrof´ısica, Brazil SEPERUELO Eduardo Observat´ orio Nacional, Brazil SKINNER Stephen University of Colorado, USA SMITH Verne University of Texas at El Paso, USA STAMATELLOS Dimitris Cardiff University, UK STELZER Beate INAF - Osservat. Astron. di Palermo, Italy STERZIK Michael European Southern Observatory, Chile TORRES Carlos Alberto Laborat´ orio Nacional de Astrof´ısica, Brazil VASCONCELOS Maria Jaqueline Universidade Estadual de Santa Cruz, Brazil VAZ Luiz Paulo Universidade Federal de Minas Gerais, Brazil VIEIRA Leandro Universidade Federal de Minas Gerais, Brazil VIEIRA Sergio Universidade Federal de Minas Gerais, Brazil VINK Jorick Imperial College London, UK WILLIAMS Jonathan Institute for Astronomy, USA ZINNECKER Hans Astroph. Inst. of Potsdam, Germany

This page intentionally left blank

I

STELLAR GROUPS AND ASSOCIATIONS

This page intentionally left blank

MAJOR UNSOLVED PROBLEMS IN STAR FORMATION Hans Zinnecker Astrophysikalisches Institut Potsdam Germany [email protected]

Abstract

In this talk I address a list of 12 topics in the field of star formation that I believe are the most important open issues. I discuss a number of them in some detail: the universality of the IMF, the origin of massive stars, the origin of Brown Dwarfs, and the origin of protostellar jets.

Introduction This is the first time I speak on a Sunday. While Sundays are often reserved for religious activities, there can be sportive events too. Today, on this particular Sunday, we have the Grand Prix Formula I Race in Sao Paulo, and so I speak with a red Ferrari hat that I got in the airport. I wish the conference would have started on Monday, so I could have watched the race live at the race course this afternoon. As it is, I could not sneak out: this is the first time I have been asked to give the introductory talk of a conference, and I would not want to miss this honour and opportunity, especially given the exciting topic and the wonderful location. This is the second time I’ve come to Ouro Preto. So, unlike most participants who are here for the first time, I knew what I was in for. Thank you, SOC and LOC, for choosing this special colonial setting for your conference – and thank you for inviting me.

1.

Overview

In my talk I will address the following 12 open issues in the field of star formation, including the basic motivations and ideas behind them (I can only discuss the first half of them here): 17 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 17-24. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

18

OPEN ISSUES IN LOCAL STAR FORMATION

1 universality of the stellar IMF 2 origin of massive stars 3 origin of brown dwarfs 4 origin of the Sun and the solar system 5 origin of binary and multiple systems 6 origin of protostellar jets 7 pre-MS evolutionary tracks 8 solution of the angular momentum problem 9 role of magnetic fields in star formation 10 the role of supersonic turbulence 11 role of metallicity and dust-to-gas ratio in star and planet formation 12 origin of globular clusters

2. 2.1

Individual topics Universality of the stellar IMF

The stellar Initial Mass Function (IMF) describes the distribution of stellar masses at birth. It was introduced nearly 50 years ago Salpeter (1955). Its functional form is a power law dN/dlog ≈ M −1.35 , originally defined between 0.4 and 10 solar masses. We now know that at the low mass end the slope is much flatter and may even become positive beyond the limiting mass for hydrogen burning (0.08 M
). An up-to-date review of the observed shape and theoretical explanations is given by Kroupa (2002) and Larson (2003), respectively.

Major Unsolved Problems in Star Formation (H. Zinnecker)

19

While the shape of the sub-solar IMF may vary from place to place, the power-law behaviour and its index for M ≥ 1M
appear to be remarkably universal, although the errors of the index are sometimes as large as 30 – 50 %. As the value for the power-law index plays such a huge role in cosmic population synthesis and galaxy evolution, the issue is whether the value is indeed universal, and whether apparent variations are in fact due stellar mass segregation in young clusters where the IMF is often determined (more massive stars preferentially near the center, Portegies Zwart 2000) or due to inadequate stellar sampling (too small clusters with too few stars, Kroupa 2001). Other pressing questions include those of whether the stellar IMF is deficient in low-mass stars, as a consequence of a reduced heavy element abundance and cooling of star forming clouds (less metals in the early universe) and in response to an increased level of the star formation rate (higher in starburst systems). Various claims have been made in the literature in the past (e.g. Rieke et al. 1980), but even in 2003 the jury is still out as more solid evidence needs to be collected (cf. Larson 1998).

2.2

The origin of massive stars

Unlike low-mass stars, high-mass stars (M ≥ 10M
) exert a tremendous dynamical and chemical influence on their immediate and wider environment, and possibly on the next generation of stars to form in their vicinity. Massive stars are the progenitors of type II supernovae that keep enriching the universe with heavy elements over cosmic time. The pressure of massive stars (supernovae, winds, HII region) can trigger additional massive star formation, either in a systematic sequence (Elmegreen & Lada 1977) or in a stochastic self-propagating way (Gerola & Seiden 1978). Massive stars are the energy source of the interstellar medium and appear to regulate the star formation rate in galaxies (negative feedback). For all those reasons, a clear picture of the origin of massive stars is vital. Do massive stars always form in dense clusters or can they also form in isolation? The conditions leading to massive star formation must be understood to be able to realistically model galaxy formation and evolution. Currently two scenarios of massive star formation are fiercely debated: the accretion scenario (Maeder & Behrend 2002, Yorke & Sonnhalter 2002) and the coalescence scenario (Bonnell et al. 1998, Stahler, Palla, & Ho 2000). One of the issues is to explain the high incidence of tight binaries (many short-period SB2 among O-stars) in young clusters ( Garc´ia & Mermilliod 2001) which may point to tidal capture in near-

20

OPEN ISSUES IN LOCAL STAR FORMATION

miss collisions and hence to the coalescence scenario (Zinnecker & Bate 2002). On the other hand, there is evidence (e.g. Lamers et al. 2000) that massive stars often thought to form only in dense young star clusters can also form quasi in isolation, a fact that the coalescence model cannot easily explain. Similarly, the formation of massive stars in loose OB associations may be a problem for the coalescence model. It has also been argued that the presence of collimated molecular outflows from massive protostars (Beuther et al. 2002) is an indication that the accretion scenario is more likely to be correct (the accretion-ejection paradigm). Overall the situation is still unclear and perhaps both scenarios may have their merit depending on the precise circumstances (i.e. the density of the protostellar environment).

2.3

Origin of Brown Dwarfs

It is only in the last decade that the existence of brown dwarfs (masses of the order 1/10 to 1/100 M
) has been firmly established. We now know that they exist in substantial number (about 1 brown dwarf for each star), but they are not numerous enough to account for a substantial fraction in the galactic mass budget (e.g. Chabrier 2002). Brown dwarfs can also show up as companions to low-mass stars, preferentially in wide pairs. Surprisingly, they do not occur as close companions; this observational fact, a by-product deduced from radial velocity searches for giant planets, has become known as “the brown dwarf desert”. Systems of two brown dwarfs (i.e. binary brown dwarfs) also exist, at the 10 % level among the free-floating brown dwarfs, but their orbital separations are significantly smaller (a ≤ 10AU ) than those of solar-type double stars (Bouy et al. 2003, Close et al. 2003). The open issue is: How do these findings constrain the formation mechanism of brown dwarfs? There are two main competing theories at present: 1) Brown dwarfs form by cloud fragmentation like ordinary stars (Padoan & Nordlund 2002), and 2) brown dwarfs form by ejection from mini-clusters; the premature ejection deprives the embryonic object from further gas accretion, so it cannot aquire enough mass to become a star (Reipurth & Clarke 2001). Recently, studies of young brown dwarfs have shown that they carry a thermal infrared excess, indicative of a disk. While this seems to favour the fragmentation origin, an ejection origin is not ruled out, as evidenced by 3D numerical simulations of cluster evolution (Bate et al. 2003). A third concept for the origin of brown dwarfs has been proposed and partly worked out: brown dwarfs may form from photo-eroded cloud cores exposed to the UV illumination of nearby massive stars, for exam-

Major Unsolved Problems in Star Formation (H. Zinnecker)

21

ple in a young cluster (Whitworth & Zinnecker 2003, Richling & Yorke 2003). This theory may, however, not be general enough to become the leading contender to explain the formation of brown dwarfs, yet it may be a contributor in specific physical situations. Let us end with an observational issue which may have new implications for brown dwarf origins. This concerns the frequency of brown dwarfs in the field vs. that in various young clusters and other star forming regions (e.g. Orion, Rho Oph; Taurus, Cha). If, as it seems to me, the young cluster/ young association IMF is deficient (by a factor of 2 or so) in brown dwarfs compared to the brown dwarf field IMF ( Reid et al. 1999, Briceno et al. 2002), then extra brown dwarf formation processes hitherto unrevealed must operate in other formation sites (e.g. Adams & Myers 2001), so that the superposition of all progenitor sites adds up to the brown dwarf field population.

2.4

Origin of the Sun and the solar system

The major issue here is whether the solar system originated in isolation (say in a Bok globule) or in a dense cluster like the Trapezium cluster. One hint in favour of the latter birth-place is the fact that the Sun’s spin axis is not exactly orthogonal to the ecliptic plane of the planets, but off by some 7 degrees. This may suggest a gravitational interaction of the original circumstellar disk, most likely in a young cluster. The istopic anomalies of the certain meteorites also seem to point in the direction of a birthplace where massive stars have exploded and polluted the protosolar cloud core (e.g. in an OB association).

2.5

Origin of binary and multiple systems

This subsection can be very short, as the problem has been discussed in a recent IAU-Symposium (No. 200). One of the open issues here is the problem of the formation of close spectroscopic binaries with orbital periods of 1 year and under. Fission, i.e. the splitting of a contracting rotating protostar, does not seem to work, but fragmentation of a circumstellar disk is a possibility (Bonnell & Bate 1994). However, another more unexpected possibility is the Kozai mechanism which involves a triple system. Here the third outlying member orbits the inner binary in a non-coplanar orbit which leads the inner binary to evolve into an orbit with an eccentricity approaching unity, which in turn leads to tidal dissipation, thus circularising and shrinking the inner orbit. In this way a close binary may be formed (Sterzik, personal comm.). Another issue is whether the relatively frequent formation of triple and quadruple systems can be understood in the new star formation scenario of turbu-

22

OPEN ISSUES IN LOCAL STAR FORMATION

lent fragmentation (Klessen, personal comm.). Finally, the multiplicity of stellar systems appears to increase with increasing primary mass (cf. Sterzik, this volume), another poorly understood observational result – perhaps suggestive of the importance of N-body gravitational effects in small N sub-clusters.

2.6

Origin of protostellar jets

There are two major ingredients for any protostellar jet model: 1 Jets originate from protostellar sources with accretion disks. However, the disk-jet coupling, the Holy Grail of jet formation theory, is not yet completely understood. 2 Magnetic fields play an essential role for jet launching, acceleration and collimation. However, the role of magnetic jets for the final stellar mass and disk-star angular momentum evolution remains unclear. While we have some basic knowledge about how the acceleration and collimation mechanisms work, the intrinsic model of jet launching is still uncertain. The overall picture of the jet structure is that of nested (axisymmetric) collimating surfaces enclosing the mass and magnetic flux (Camenzind 1990). Acceleration and collimation can be understood within the MHD framework by decomposing the Lorentz force  FL ∼  × B into a component along the flux surfaces, φ FL,|| ∼ ⊥ × B (accelerating) and perpendicular to these surfaces  FL,⊥ ∼ || × B (collimating). Also the launching mechanism which lifts the accreting matter out of the disk into the outflow has been shown to be magnetic (Ferreira 1997). A vertical decrease of the total electric current leads to toroidal Lorentz force which (depending on the strength of the magnetic field) accelerates the matter in toroidal direction to super-Keplerian speed leading to outward directed centrifugal forces (beads on a string a la Blandford and Payne).

Major Unsolved Problems in Star Formation (H. Zinnecker)

23

However, due to the complexity of the governing equations, the different time-scales for the disk and the jet, there is no complete, selfconsistent model calculation of MHD jet formation up to now. Whether the jet is launched as a disk wind or a X-wind may be a secondary question (Fendt, personal comm., but see Shu et al. 1994). Performing numerical simulations of the jet formation out of the accretion disk is one of the most essential numerical problems in the future (e.g. Fendt & Cemeljic 2002 or Casse & Keppens 2002). Other major issues include the effect of the jet on the accretion process, both on the final stellar mass and on the angular momentum transport in the accretion disk. While jets may not be powerful enough to substantially reverse the infall, they may indeed carry away disk angular momentum efficiently. Thus, it might well be that protostars with jets will rotate more slowly than those without a jet. Why are magnetic jets and winds efficient mechanisms to remove disk angular momentum? This is due to the length of the lever arm which in the magnetic case is given by the Alfv´en radius rA located typically at about 10 times the radius of the foot point r0 of the magnetic field line:


rA r0

2

≡ λ  100

If all disk angular momentum is carried away by the disk wind, the lever arm ratio is also related to the mass flow ratio,


rA r0

2



M˙ acc M˙ wind

However, the question of how the stellar angular momentum evolution is affected by the magnetic field, is far from clear. The magnetic coupling between the disk and the star can either remove or add angular momentum to the star depending on the foot point of these field lines in the disk (within or outside the corotation radius).

Acknowledgments I thank Jane Gregorio-Hetem and Jacques L´epine as well as C. Fendt and U. Hanschur for their help in finishing this paper. I also thank the LOC for their financial support. Finally I thank Annibal and his friends for providing a special Brazilean touch (music!) to the meeting.

References Adams F.C. & Myers P.C. 2001, ApJ 553, 744 Bate, M. R.; Bonnell, I. A.; Bromm, V. 2003, MNRAS 339, 577

24

OPEN ISSUES IN LOCAL STAR FORMATION

Bonnell, I. A., Bate, M. R. & Zinnecker, H. 1998, MNRAS 298, 93 Bonnell, I.A. & Bate, M.R. 1994, MNRAS, 271, 999 Bouy, H. & Brandner, W. et al. 2003, Multiplicity of Nearby Free-floating Late M and L Dwarfs: HST-WFPC2 Observations of Candidates and Bona Fide Binary Brown Dwarfs. Proceedings of IAU Symposium 211, San Francisco: Astronomical Society of the Pacific, p. 245 Beuther, H., Schilke, P. 2002, A&A, 383, 892 Briceno, C., Luhman, K. et al. 2002, ApJ, 580, 317 Camenzind, M. 1990, Magnetized Disk-Winds and the Origin of Bipolar Outflows. Reviews of Modern Astronomy 3, Heidelberg: Springer-Verlag, p. 234 Casse, F. & Keppens, R. 2002, ApJ, 581, 988 Chabrier, G. 2002, ApJ, 567, 204 Close, L. M. & Siegler, N. et al. 2003, ApJ 587, 407 Elmegreen, B. G. & Lada, C. J. 1977, ApJ 214, 725 Fendt, Ch. & Cemeljic, M. 2002, A&A 395, 1045 Ferreira, J. 1997, A&A 319, 340 Garc´ia, B. & Mermilliod, J. C. 2001, A&A 368, 122 Gerola, H. & Seiden, P. E. 1978, ApJ 223, 129 Kroupa, P. 2002, Science 295, 82. Kroupa, P. 2001, The Local Stellar Initial Mass Function. ASP Conference Series, Vol. 228, San Francisco: Astronomical Society of the Pacific, p. 187 Lamers, H. J. G. L. M.; Nugis, T. et al. 2000, he Dependence of Mass Loss on the Stellar Parameters. ASP Conference Series, Vol. 204, San Francisco, Astronomical Society of the Pacific, p. 395 Larson, R. B. 2003, The Stellar Initial Mass Function and Beyond (Invited Review). ASP Conference Series, Vol. 287, San Francisco: Astronomical Society of the Pacific, p. 65 – 80 Larson, R. B. 1998, MNRAS 301, 569 Maeder, A. & Behrend, R. 2002, Ap&SS 281, 75 Padoan, P. & Nordlund, A. 2002, ApJ 576, 870 Portegies Zwart, S. F. 2000, Is the Galactic Center Populated with Young Star Clusters? ASP Conference Series, Vol. 228, San Francisco: Astronomical Society of the Pacific, p. 131 Reid, I. N.; Kirkpatrick, J. D. et al. 1999, ApJ 521, 613 Reipurth, B. & Clarke, C. 2001, AJ 122, 432 Richling, S. & Yorke, H. W., in preparation Rieke, G. H.; Lebofsky, M. J.; Thompson, R. I. et al. 1980, ApJ 238, 24 Salpeter, E. E. 1955, ApJ 121, 161. Shu, F.; Najita, J.; Ostriker, E. et al. 1994, ApJ 429, 781 Stahler, S. W.; Palla, F. & Ho, P. T. P. 2000, The Formation of Massive Stars. Protostars and Planets IV, Tucson: University of Arizona Press, p. 327 Whitworth & Zinnecker, H. 2003, in preparation Yorke, H. W. & Sonnhalter, C. 2002, ApJ 569, 846 Zinnecker, H. & Bate, M. R. 2002, Multiplicity of Massive Stars – a Clue to their Origin? ASP Conference Proceedings, Vol. 267, San Francisco, Astronomical Society of the Pacific, p. 209

THE STELLAR MASS FUNCTION OF GALACTIC CLUSTERS AND ITS EVOLUTION

Guido De Marchi European Space Agency, Space Telescope Division, USA [email protected]

Abstract

The stellar mass function of a large sample of Galactic clusters (young and old) can be well reproduced with a tapered power law distribution function with an exponential truncation of the form dN/dM ∝ m−α [1 − exp(−m/mp )β ]. The average value of the power-law index α is very close to Salpeter (∼ 2.3), whereas the peak mass mp is in the range 0.1 − 0.6 M
and does not seem to vary in a systematic way with the present cluster parameters such as metal abundance and central concentration. A remarkable correlation with age, however, is seen in that older disc clusters have higher mp , although this trend does not extend to globular clusters, whose value of mp is lower than that of old open clusters. This trend most likely results from the onset of mass segregation following early dynamical interactions in the loose cluster cores. Differences between globular and younger clusters may depend on the initial environment of star formation.

Introduction To understand whether and how the stellar mass function (MF) varies with the environment it is useful to parametrise it and to look for correlations between its shape and the physical conditions of the clouds in which different stellar populations are born. Observations conducted in environments as diverse as young star forming regions (e.g. Luhman et al. 2000) and old globular clusters (Paresce & De Marchi 2000) suggest that, near the Hydrogen burning limit, stars do not form as copiously and efficiently as they do at higher masses. Whilst it is now clear that 25 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 25-32. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

26

OPEN ISSUES IN LOCAL STAR FORMATION

a simple power-law does not provide an adequate representation of the MF over the entire stellar mass range (Scalo 1998), the proposed MF parametrisation by means of multiple power-law segments (Kroupa 2001) is also not completely satisfactory, since Nature seems to always vary smoothly. Furthermore, the freedom allowed by a segmented power-law MF could make it difficult to identify and characterise possible trends with the physical properties of the environment. It is likely that some of the controversy as to the universality of the initial MF (IMF; see e.g. Eisenhauer 2001, Gilmore 2001) would be recomposed if an analytical functional form were found which is consistent with the MF of many stellar populations. It has been recently suggested by Elmegreen (1999) that star formation, and thus the IMF, is the result of two concurrent processes: one, random sampling from a hierarchical fractal cloud, is responsible of the power-law part of the IMF; the other, the inability of gas to form stars below the thermal Jeans mass at typical temperatures and pressures, produces the flattening at low masses. In this scenario, the thermal Jeans mass is the only relevant scale in the problem and the MF takes on the form of a tapered power-law (TPL), which can be written analytically as: f (m) =

  dN β ∝ m−α 1 − e(−m/mp ) dm

(1)

where mp is the peak or scale mass, α the index of the power-law portion for high masses and β the tapering exponent which causes the MF to flatten at low masses. In the following I show that a function of this type reproduces remarkably well the MF of various Galactic clusters of different ages (both in the disc and halo). This allows one to use the values of the parameters, and particularly the scale mass mp , to study the effect of the environment on the star formation process.

1.

The data sample

I report here on the preliminary results of an ongoing investigation (De Marchi, Paresce & Portegies Zwart 2003) which uses a physically homogeneous data sample comprising of observations of a large number of similar stellar populations, namely Galactic stellar clusters, spanning ages from ∼ 1 Myr to several Gyr. The sample includes twelve globular clusters (GC; see Paresce & De Marchi 2000 for details on these objects) and eight disc clusters, as listed in Table 1.

27

The stellar mass function of Galactic clusters (De Marchi) Table 1.

Properties of the young (top) and globular (bottom) clusters in the sample

Name

Reference

Age

Orion Neb. Cl. (in) Orion Neb. Cl. (out) Trapezium Rho Oph IC 348 Pleiades Pleiades (inner) M 35 Hyades Praesepe Praesepe (outer) Praesepe (inner)

HC00 HC00 L00 LR99 L98 H99 A01 B01 GRM99 H95 A02 A02

0.5 Myr 1.0 Myr 0.5 Myr 0.5 Myr 1.0 Myr 100 Myr 100 Myr 160 Myr 625 Myr 750 Myr 750 Myr 750 Myr

NGC104 NGC5139 NGC5272 NGC6121 NGC6254 NGC6341 NGC6397 NGC6656 NGC6752 NGC6809 NGC7078 NGC7099

PDM00 PDM00 PDM00 PDM00 PDM00 PDM00 PDM00 PDM00 PDM00 PDM00 PDM00 PDM00

12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr 12 Gyr

r/rh

α

β

mp

< 0.1 < 1.0 < 0.1 core < 0.5 < 3.3 < 1.2 < 1.1 < 0.9 < 1.9 > 1.1 < 1.1

2.4 2.3 1.8 2.0 2.0 2.5 2.5 2.6 2.4 1.9 2.5 2.3

2.5 2.5 1.8 2.0 2.2 2.7 2.6 2.9 2.4 2.0 2.9 2.4

0.2 0.1 0.2 0.2 0.2 0.4 0.5 0.6 0.6 0.1 0.2 0.4

1.6 0.9 3.6 1.3 1.3 4.5 1.8 0.8 1.5 0.9 4.6 4.6

2.3 2.3 2.3 2.2 2.3 2.3 2.3 2.3 2.3 2.3 2.3 2.3

2.4 2.7 2.9 2.9 2.4 2.3 2.7 2.7 2.9 2.5 2.7 2.3

0.35 0.35 0.35 0.37 0.35 0.33 0.35 0.35 0.37 0.35 0.28 0.29

The ratio r/rh indicates the region of the cluster covered by the observations in units of the half-mass radius.

Paresce & De Marchi (2000) have studied the GC discussed here, using HST observations, and have concluded that their stellar global MF below 1 M
can be reproduced rather well by a log-normal distribution. They find a characteristic mass of mc = 0.33±0.03 M
and a width σ = 0.34± 0.04. If the MF of GC held to this log-normal shape also above the mainsequence turn-off mass (0.8 M
), where it cannot be presenlty observed, it would depart significantly from that typically found elsewhere, since in the range 1 − 10 M
the Salpeter IMF is practically ubiquitous (see e.g. Scalo 1998; Kroupa 2001). However, the TPL distribution of Equation 1 reproduces equally well the MF of all these GC in the observed mass range, whilst taking on a power-law shape above the turn-off. In fact, since the value of the index α has an almost negligible effect on the shape of the MF around mp , its value cannot be constrained in GC. One can, therefore, assume that the latter be the canonical Salpeter value (α =

28

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 1. Mass functions (solid lines) of the young clusters listed in Table 1, fitted with TPL distributions (dashed lines).

2.3) and use the observed MF to determine mp and β for all clusters. The values of mp and β that best fit the observations span a very narrow range around their mean values (mp = 0.35 ± 0.04, β = 2.6 ± 0.03), as shown in Table 1 (bottom), fully consistent with that found by Paresce & De Marchi (2000) for the log-normal distribution, thus confirming the intrinsic similarity of these MF. A sample of young clusters as homogeneous as that of the GC shown here does, unfortunately, not yet exist. The best that one can do is to select from the literature those objects whose MF has been recently measured on high quality data with reliable error estimates (see references in Table 1). Rather than using the data themselves, we have taken as MF for each cluster the power-laws that best fit the data, either as given by each author or measured by ourselves on their figures. Due to the nature of these MF, usually two and sometimes three power-law segments are necessary. These MF are shown in Figure 1 as thick solid lines. Like in the case of GC, one can fit each observed MF using a TPL

The stellar mass function of Galactic clusters (De Marchi)

29

distribution, marked as dashed lines. The parameters α, mp and β that best fit the data are listed for each cluster in Table 1 (top). As in the case of GC, no correction for stellar multiplicity has been applied.

2.

Discussion

Figure 1 reveals that the shape of the MF can change considerably from cluster to cluster, with the peak varying largely in mass (although α and β span a narrow range around their average values). We have found no correlation between the shape of the MF of GC stars and cluster properties such as metallicity, mass, space motion parameters and dynamical state (Paresce & De Marchi 2000). Unfortunately, the dynamical properties of the young clusters are largely unknown, so these objects cannot currently be subjected to a similarly detailed investigation. More accurately known, however, are their ages which increase from top to bottom in Figure 1 (see Table 1). Even a casual inspection reveals immediately a strong trend: the peak of the MF shifts to higher masses with increasing age. The most likely origin of this trend is the combined effect of mass segregation and the limited area of the cluster covered by the observations. In the absence of tidal interactions with the Galaxy, one expects the global MF of a cluster to vary slowly with time due to evaporation (Spitzer 1987). For massive GC this process can take several tens or hundreds Gyr (Gnedin & Ostriker 1997), but in young open clusters mass segregation and the resulting evaporation proceed more rapidly (Raboud & Mermilliod 1998), yet not as fast as to justify the spread in the shape of the MF observed in Figure 1. In fact, Portegies Zwart et al. (2001) have shown that the global MF of a 600 Myr old cluster with a mass of 1600 M
differs only marginally from the IMF, even when the enhanced erosion induced by the Galactic potential is included in the calculations. However, most of the MF plotted in Figure 1 do not reflect the global MF of the clusters, since they are based on data collected in the central regions, with the unavoidable consequence that the artefacts of mass segregation are impressed on them (Terlevich 1987). Only for the Pleiades and Praesepe are there measurements of the MF which extend well beyond the half-mass radius (Hambly et al. 1999; Adams et al. 2001; Hambly et al. 1995; Adams et al. 2002), although they are not in agreement with one another. For the rest, as Table 1 shows, information on the low-mass range of interest here comes from the inner cluster regions, where Portegies Zwart et al. (2001) find the local MF to change rather rapidly with time.

30

Figure 2. 1600 M
.

OPEN ISSUES IN LOCAL STAR FORMATION

Temporal evolution of the MF within the half-mass radius of a cluster of

Although the details of a quantitative analysis cannot be given here (see Portegies Zwart et al. 2001; De Marchi et al. 2003), Figure 2 shows the temporal evolution of the MF of the stars contained within the halfmass radius (rh ) of a 1600 M
model cluster. The IMF is assumed to be that of Scalo (1986) with a peak at ∼ 0.4 M
which drifts to higher masses as time progresses, in the same way as we observe in Figure 1. Since the average stellar mass increases towards the cluster centre due to mass segregation, the location of the MF peak depends steeply on the fraction of cluster area sampled by the data: the wider is the latter, the lower is the peak mass. Thus, although not specific to any one of the clusters in our sample, the simulation shown in Figure 2 proves rather convincingly that mass segregation, combined with limited sampling of the cluster population, can explain much of the variation noticed in Figure 1. An interesting outcome of this simulation is that the flattening of the MF below mp can only be reproduced by evolving in time an IMF which is already peaked at a lower mass. In other words, if the IMF were to be a pure power-law function all the way through to the sub-stellar mass limit, dynamical evolution would not be sufficient to explain the knee observed in the MF. Furthermore, the models suggest that, regardless of

The stellar mass function of Galactic clusters (De Marchi)

31

the adopted IMF, changes in the shape of the MF are most pronounced in the 0.5 − 1.0 M
mass range, in agreement with what is seen in Figure 1.

3.

Conclusions

The considerations exposed in the previous section suggest that the young clusters of Figure 1 share a rather similar IMF, since it is reasonable to attribute the discrepancies amongst their shapes to the effects of dynamical evolution. Such an IMF would have to be very similar to the MF of the youngest clusters (∼ 1 Myr old) and have a peak mass mp  0.2 M
or, more likely, mp  0.15 M
when account is taken of binaries (see Kroupa 2001). As discussed by Paresce & De Marchi (2000), the similarity amongst the MF of GC suggests that they as well could all share the same IMF. The logical question to pose is whether the IMF is the same for both old and young systems. More precisely, one could ask whether dynamical evolution in GC might have proceeded in such a way that the peak of the MF has moved from an initial value of mp  0.15 M
to the presently observed ∼ 0.35 M
. Insight on any intrinsic characteristic mass differences between young and globular clusters should be searched in the properties of the stellar MF of the Galactic disc, bulge and halo. Since stars do not form in isolation, but rather in clusters and associations, it seems very reasonable that the stars making up the halo, bulge and disc were originally part of clusters that were disrupted, through evaporation and/or dynamical friction in the Galactic tidal field (Surdin 1995; Gnedin & Ostriker 1997; Murali & Weinberg 1997; Dehnen 1998). Evidence of on-going GC disruption is found for instance in NGC 6712 (De Marchi et al. 1999). Presently available measurements of the MF of the Galactic disc (Kroupa 2001; Reid & Gizis 1997; Zheng et al. 2001) and bulge (Holtzman et al. 1998; Zoccali et al. 2000) can be acceptably reproduced with a Salpeter TPL distribution peaked at ∼ 0.25 M
and ∼ 0.4 M
, respectively, and with a tapering exponent β which takes on the value of ∼ 2 in both cases. These values are, thus, consistent with the hypothesis that the disc and bulge were built up over time by contributions from, respectively, young and old clusters with a MF of the same type but with an intrinsically different peak mass. It seems, therefore, unlikely that GC originally had a much smaller mp than today. In any case, regardless as to whether the IMF has a peak at ∼ 0.15 M
or ∼ 0.35 M
, it appears that it has a characteristic scale mass, thus lending support to the theoretical predictions of Adams & Fatuzzo (1996), Larson (1998), Elmegreen (1999) and Bonnell et al. (2001) who suggest that the environments and/or processes through which the for-

32

OPEN ISSUES IN LOCAL STAR FORMATION

mation of high and low mass stars takes place are physically different. The difference between the peak mass of globular and younger clusters could then reflect the initial physical conditions of the star forming environment, thus offering the exciting new possibility of learning more about the star formation process from the properties of the IMF itself.

References Adams, F., Fatuzzo, M. 1996, ApJ, 464, 256 Adams, J., Stauffer, J., Monet, D., et al. 2001, AJ, 121, 2053 (A01) Adams, J., Stauffer, J., Skrutskie, M. et al. 2002, AJ, 124, 1570 (A02) Barrado y Navascu´es, D., Stauffer, J., Bouvier, J., Mart´in, E. 2001, ApJ, 546, 1006 (B01) Bonnell, I., Clarke, C., Bate, M., Pringle, J. 2001, MNRAS, 324, 573 Dehnen, W. 1998, AJ, 115, 2384 De Marchi, G., Paresce, F., Leibundgut, B., Pulone, L. 1999, A&A, 343, L9 De Marchi, G., Paresce, F., Portegies Zwart, S. 2003, in preparation Eisenhauer, F. 2001, in Starburst Galaxies: Near and Far, Ed. L. Tacconi & D. Lutz (Heidelberg: Springer), 24 Elmegreen, B. 1999, ApJ, 515, 323 Gilmore, G. 2001, in Starburst Galaxies: Near and Far, Ed. L. Tacconi & D. Lutz (Heidelberg: Springer), 34 Gizis, J., Reid, I., Monet, D. 1999, AJ, 118, 997 (GRM99) Gnedin, O., Ostriker, J. 1997, ApJ, 474, 223 Hambly, N., Hodgkin, S., Cossburn, M., Jameson, R. 1999, MNRAS, 303, 835 (H99) Hambly, N., Steele. I., Hawkins, M., Jameson, R. 1995, MNRAS, 273, 505 (H95) Hillenbrand, L., Carpenter, J. 2000, ApJ, 540, 236 (HC00) Holtzman, J., Watson, A., Baum, W., Grillmair, C., et al. 1998, AJ, 115, 1946 Kroupa, P. 2001, MNRAS, 322, 231 Larson, R. 1998, MNRAS, 301, 569 Luhman, K., Rieke, G. 1999, ApJ, 525, 440 (LR99) Luhman, K., Rieke, G., Lada, C., Lada, E. 1998, ApJ, 508, 347 (L98) Luhman, K., Rieke, G., Young, E., Cotera, A. et al. 2000, ApJ, 540, 1016 (L00) Murali, C., Weinberg, M. 1997, MNRAS, 288, 749 Paresce. F, De Marchi, G. 2000, ApJ, 534, 870 (PDM00) Portegies Zwart, S., McMillan, S., Hut, P., Makino, J. 2001, MNRAS, 321, 199 Raboud, D., Mermilliod, J. 1998, A&A, 333, 897 Reid, I., Gizis, J. 1997, AJ, 113, 2246 Scalo, J. 1986, Fundamentals of Cosmic Physics, 11, 1 Scalo, J. 1998, in ASP Conf. Ser. 142, Ed. G. Gilmore & D. Howell (San Francisco: ASP) 201 Spitzer, L. 1987, Dynamical evolution of globular clusters, (Princeton Univ. Press) Surdin, V. 1995, Ast. Lett., 21, 4 Terlevich, E. 1987, MNRAS, 224, 193 Zheng, Z., Flynn, C., Gould, A., et al. 2001, ApJ, 555, 393 Zoccali, M., Cassisi, S., Frogel, J., et al. 2000, ApJ, 530, 418

ENCOUNTERS AND CLOSE FLY-BYS OF GALACTIC CLUSTERS AND ASSOCIATIONS IN THE PAST 50 MYR V.V. Makarov U.S. Naval Observatory, Universities Space Research Association, USA [email protected]

Abstract

Collisions and close fly-bys of molecular clouds and associations may be an important mechanism of star formation in the solar neighborhood. A giant molecular cloud may be directly impacted by the gas in another cloud upon a cloud-cloud collision (L´epine & Duvert 1994) or it may be shocked by bubbles blown up by stellar winds and supernova explosions in a chance fly-by OB association (de Geus 1991). Clues of such events during the last 50 Myr are sought for ∼ 80 nearby open clusters, OB associations and star forming regions using the epicycle approximation of the Galactic orbital motion. Several possible events are identified, and their times, relative velocities and impact distances are estimated. In some cases, the events are more recent than the assumed age of one of the parties involved, e.g., the possible encounter of the Coma cluster with the Taurus-Auriga star forming region 10 Myr ago. In other cases, the estimated time of the event matches fairly well the isochrone ages of both parties, e.g., the encounter of the LCC and IC 2602, 36 Myr ago. Relative velocities found thus far are between 5 and 20 km s−1 .

Introduction Most of the new stars in the local part of the Galaxy are formed in gravitationally unbound OB and T associations, and a smaller fraction in gravitationally bound clusters. This probably implies that the mechanisms triggering star formation in molecular clouds do not necessarily require very massive clouds to work on, and do not always produce large numbers of stars in limited volumes of space. The character of those triggering mechanisms is a matter of dispute. It is widely recognized that expanding bubbles around young associations and star-forming regions (SFR), blown by supernova explosions and collective high-velocity winds 33 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 33-38. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

34

OPEN ISSUES IN LOCAL STAR FORMATION

from the most massive stars, may play an important role in the propagation of star formation from one location to another (de Geus 1991). An alternative scenario of star formation was suggested by L´epine and Duvert (1994). High-velocity impacts of molecular clouds can also cause the gas to compress and cool down, generating the appropriate conditions for formation of massive stars. Numerical simulations showed that head-on collisions at 10 km s−1 could produce gas fragments as dense as 104 cm−3 and significantly cooler than the surrounding areas (Marinho and L´epine 2000). There are other conceivable scenarios of interaction between gas and molecular clouds and foreign individual stars or clusters that can bring about the first new stars. A passage of a massive wind-generating star thorough a sufficiently dense interstellar gas or a cloud can generate wake channels and swarms of dense globules (Tenorio-Tagle and MedinaTanco 1998). Such globules may become the seeds of new stars. Even in the absence of wind-generating massive stars and bubbles, upon a collision with an open cluster, a diffuse, low-mass cloud undergoes collapse, followed by an outward shock wave parallel to the impact direction ( Theuns 1992). Evidence has been presented that the interstellar gas inside and around the Pleiades is a separate cloud undergoing a chance encounter with the cluster at  18 km s−1 (White and Bally 1993). All these scenarios imply a collision or a close fly-by of a molecular cloud and another cloud, OB association or cluster. Some of them imply a close interaction of the cloud with massive stars or a cluster, while the shock-wave formation scenario can work at larger distances between the parties. In all of the scenarios, the counterparts will continue to travel on their individual paths after the star-triggering event and will therefore move away from each other, with the exception of a head-on collision of two clouds, which will probably result in a single massive starforming region. The emerging new clusters and associations will have the same age in cloud-cloud encounters, somewhat different age in cloudassociation encounters and quite different age in cloud-cluster collisions. This paper presents the first results of a search for such events in the past by tracking nearby clusters, associations and star-forming regions backward in time.

1.

Method

The epicycle approximation of the Galactic motion (e.g., Asiain et al. 1999; Makarov and Olling 2003) is used to compute the complete 6D paths in the phase space of about 60 nearby clusters, associations and star-forming regions. The vertical motion is assumed to be harmonic

35

Encounters of clusters in the past (Makarov)

oscillation, which is an accurate approximation for heliocentric vertical velocities less than 20 km s−1 . Comparison with integrated orbits has also shown that the planar epicyclic orbit may become increasingly imprecise after some 50 Myr. We assume a vertical period Pz = 2π/ν of 80 Myr, and numerical values for the Oort constants of A=0.0148km s−1 pc−1 and B=−0.0124 km s−1 pc−1 . The basic equations are u0 v0 − 2Ax0 sin κt + (1 − cos κt) κ 2B u0 v0 y = y0 − (1 − cos κt) + (Aκt − (A − B) sin κt) 2B κB 2A(A − B) −x0 (κt − sin κt) κB w0 z = z0 cos νt + sin νt ν

x = x0 +

(1)

where (x0 , y0 , z0 ) denotes the position, and (u0 , v0 , w0 ) the velocity of the test particle with regard to the Sun at t = 0 (now). The horizontal epicycle frequency is related to the Oort’s constants A and B as  κ = −4B(A − B); the vertical frequency ν specifies the common period of test particles oscillations around the Galactic plane. Position of the particle in the past can be directly computed by reversing t in the above equations. Instantaneous heliocentric velocities are obtained by differentiating the equations with respect to time. 120

separation pc

100 80

60

40

20

0 0

-5

-10 Myr

-15

Figure 1 Distance between the R CrA star-forming region and the Upper Centaurus Lupus OB association in the past. The most probable distance is shown with a thick line, the median over 501 Monte-Carlo simulations with a dot-dashed line, and the 25th percentile distance with a thin line.

Due to the uncertainties of the input data, the results obtained should be expressed in statistical terms. Monte-Carlo simulations for each object are conducted much along the lines described in Makarov et al. (2003). The resulting distributions of parameters of interest (coordinates and velocities, respective distances, times of closest approach) are

36

OPEN ISSUES IN LOCAL STAR FORMATION

used to estimate various statistics (e.g., the median and most probable values and sigma intervals). The respective distance (separation), a positively defined quantity, is the most heavily biased parameter. This fact is reflected in Fig. 1, where the most probable distance (thick line) deviates significantly and goes much lower around the minimum than the median or the 25th percentile distances. The time of encounter, defined as the time of the minimum most probable separation or the time of the minimum median distance, will not be the same. Typically, the time of the minimum median separation tends to be later than the time of the minimum expected separation (cf. Table 2). Table 1.

Association and Cluster Parameters

Name

(, b) deg

LCC OB IC 2602 NGC 6475 IC 2391 Coma RXJ0405.7+2248 HIP 19176 HIP 21852 UCL OB R CrA η Cha Collinder 135 Collinder 140

(298.0, +7.0) (289.6, −4.9) (355.8, −4.5) (270.3, −6.9) (221.3, 84.0) (170.9, −21.5) (173.0, −23.1) (173.4, −12.6) (327.0, 13.0) (359.9, −17.9) (292.5, −21.7) (248.8, −11.2) (245.2, −7.9)

(µ∗ , µb ) mas yr−1 (−32.1, −13.1) ± 0.1 (−2.5, 1.5) ± 0.2 (−3.0, −4.7) ± 0.3 (−33.1, −6.0) ± 0.2 (6.6, −12.9) ± 0.2 (14.4, −6.3) ± 1.5 (15.6, −4.3) ± 1.5 (15.4, −14.7) ± 1.5 (−30.1, −9.1) ± 2.0 (−23.1, −15.2) ± 2.0 (−39.5, −10.7) ± 0.3 (−10.3, −6.8) ± 0.5 (−7.4, −5.5) ± 0.5

RV km s−1 +12.0 ± 1.5 +16.2 ± 0.3 −14.7 ± 0.2 +14.1 ± 0.2 −0.1 ± 0.2 +16.4 ± 1.5 +15.5 ± 0.5 +14.1 ± 0.5 +4.9 ± 0.5 −2.6 ± 1.3 +16.1 ± 0.5 +16.4 ± 3.0 +22.4 ± 3.0

Distance pc 118 ± 15 140 ± 8 280 ± 25 146 ± 5 87 ± 2 150 ± 50 156 ± 56 115 ± 20 140 ± 10 130 ± 20 97 ± 5 300 ± 15 375 ± 15

Col. (2): Galactic coordinates; Col. (3): mean proper motion in Galactic coordinates; Col. (4): mean radial velocity; Col. (5): mean distance from the Sun.

2.

Data

For a number of open clusters within ∼ 200 pc, the Hipparcos-based positions, proper motions and mean parallaxes were adopted from Robichon et al. (1999). Similar astrometric data and radial velocities for nearby OB associations were found in de Zeeuw et al. (1999). Hipparcos proper motions and parallaxes become too imprecise for stellar ensembles beyond 200 pc, and the required information has to be culled from the literature sources of nonuniform quality. Hoogerwerf et al. (2001) listed relevant data for a number of clusters and associations that were not well represented in Hipparcos, which were adopted in this paper too. Estimates of ages were adopted from the WEBDA site

37

Encounters of clusters in the past (Makarov)

(http://obswww.unige.ch/webda/) and from the literature. The input data used for the calculation of possible encounter events are given in Table 1 for some part of the general sample. The largest uncertainties in the data concern the often poorly known radial velocities and, for more distant objects, distances. The very sparse Taurus-Auriga SFR is represented by three high-fidelity members with relatively accurate observational data, RXJ0405.7+2248, HIP 19176 and HIP 21852. The data for the R CrA star-forming region were drawn from Neuh¨ auser et al. (2000). Table 2.

Possible Encounters and Fly-bys

Name

(1)

(2)

(3)

(4)

(5)

(6) 5.5

(7)

(8)

(9)

35

10 32

LCC OB IC 2602

54

10

0.002

0.06

49+17 −11

IC 2391 NGC 6475

305

15

0.13

0.42

29+32 −17

8.3

35+10 −7

36

46 300

Collinder 135 Collinder 140

80

4

0.07

0.46

26+35 −12

7.6

10+10 −5

14

26 35

UCL OB R CrA SFR

108

20

0.03

0.22

43+27 −21

9.3

11+3 −3

12

15 10

LCC OB η Cha

55

12

0.12

0.61

21+22 −10

6.4

8+2 −2

9

10 10

Taurus-Auriga Coma Ber

∼ 180

∼ 15

0.07

0.35

20–40

17.5

10+2 −2

11

10? 450

Col. (1): separation now in pc; Col. (2): minimum expected separation in pc; Col. (3): probability that the minimum separation is less than 10 pc; Col. (4): probability that the minimum separation is less than 25 pc; Col. (5): median minimum separation in pc; Col. (6): velocity of fly-by in km s−1 ; Col. (7): median time of fly-by in Myr; Col. (8): expected time of fly-by in Myr; Col. (9): age in Myr.

3.

Results and Discussion

Possible and probable encounters and fly-by events are summarized in Table 2. We have identified, at various levels of confidence, several possible events responsible for the formation of clusters and OB associations. For example, the encounters of IC 2602 with the LCC progenitor cloud, UCL with the R CrA cloud 12 Myr ago and the Taurus-Auriga SFR with the Coma Ber cluster 11 Myr ago were identified. The encounter of LCC and the η Chamaeleontis cluster was discovered by Mamajek et al. (2000). The encounter between the Collinder 135 and Collinder

38

OPEN ISSUES IN LOCAL STAR FORMATION

140 clusters appears to be more recent than their isochrone ages, but this result is statistically uncertain due to the large formal errors in the input parameters, especially in the radial velocities (Table 1). The velocities of fly-bys and encounters range between 5 and 20 km s−1 . Most of the identified events were less energetic, with velocities below 10 km s−1 . This may imply that high-velocity impacts of molecular clouds moving through the disk system of clouds are rare occurrences compared to low-velocity approaches of clouds with nearly co-planar motions. Indeed, the local system of molecular clouds seem to be involved in a systemic motion similar to that of the Gould Belt, with perhaps an internal dispersion less than 10 km s−1 . Due to the roughly cyclic character of the motion in the Galactic plane, a certain pattern of clouds may converge together time and again, in every epicyclic period (∼ 160 Myr), experiencing collisions between its parts and, subsequently, reforming iself due to the emerging bubbles.

References Asiain, R., Figueras, F., and Torra, J. 1999, A&A, 350, 434 de Geus, E.J. 1991, in The Formation and Evolution of Star Clusters, (K. Janes), San-Francisco: ASP, 40 de Zeeuw, P. T., et al. 1999, AJ, 117, 354 Hoogerwerf, R., de Bruijne, J. H. J., and de Zeeuw, P. T. 2001, A&A, 365, 49 L´epine, J.R.D., and Duvert, G. 1994, A&A, 286, 60 Makarov, V.V., and Olling, R.P. 2003, ApJ, submitted Makarov, V.V., Gaume, R.A., and Andrievsky, S.M. 2003, These Proceedings Mamajek, E. E., Lawson, W. A., and Feigelson, E. D. 2000, ApJ, 544, 356 Marinho, E.P., and L´epine, J.R.D. 2000, A&AS, 142, 165 Neuh¨ auser, R., et al. 2000, A&AS, 146, 323 Robichon, N., Arenou, F., Mermilliod, J.-C., and Turon, C. 1999, A&A, 345, 471 Tenorio-Tagle, G., and Medina-Tanco, G.A. 1998 ApJ, 503, L171 Theuns, T. 1992, A&A, 259, 493 White, R.E., and Bally, J. 1993, ApJ, 409, 234

SOLAR-TYPE POST-T TAURI STARS IN THE NEAREST OB SUBGROUPS

Eric E. Mamajek Steward Observatory, The University of Arizona, 933 N. Cherry Ave., Tucson AZ 85721 [email protected]

Abstract

I discuss results from the recent spectroscopic survey for solar-type preMS stars in the Lower Centaurus-Crux (LCC) and Upper CentaurusLupus (UCL) OB subgroups by Mamajek, Meyer, & Liebert (2002, AJ, 124, 1670). LCC and UCL are subgroups of the Sco-Cen OB association, and the two nearest OB subgroups to the Sun. In the entire survey of 110 pre-main sequence stars, there exists only one Classical T Tauri star (PDS 66), implying that only ∼1% of ∼1 M
stars are still accreting at age 13±7 (1σ) Myr. Accounting for observational errors, the HRD placement of the pre-MS stars is consistent with the bulk of star-formation taking place within 5-10 Myr. In this contribution, I estimate conservative upper limits to the intrinsic velocity dispersions of the post-T Tauri stars in the LCC and UCL subgroups ( 0.6, excluding well known RS CVn, W UMa, giants, etc from SIMBAD. We restricted our sample to stars later than G0 because we use the Li I 6707˚ A equivalent width as an age indicator. We obtained high resolution spectra for the selected candidates with the FEROS ´echelle spectrograph (Kaufer et al. 1999) (resolution of 50000; spectral coverage of 5000 ˚ A) of the 1.52 m ESO telescope at La Silla or with the coud´e spectrograph (resolution of 9000; spectral coverage of 450 ˚ A, centered at 6500 ˚ A) of the 1.60 m telescope of the Observat´ orio do Pico dos Dias. For some stars we obtained radial velocities with CORALIE at the Swiss Euler Telescope at ESO (Queloz et al. 2000). We derived spectral classifications, radial velocities and equivalent widths of Li I 6707 ˚ A lines. In particular, the Li I line is important since it can provide a crude age estimate (Jeffries 1995) for late type stars. If the Li I line equivalent width is larger than the highest values for stars stars belonging to the Local Association (Neuh¨auser 1997), the star is flagged as young. This is shown in Figure 1. In Figure 2 we plot, in a polar projection, the complete sample observed.

1.1

Statistics

There are 9574 RASS bright sources in the Southern Hemisphere, 2071 of them having counterparts with (B-V) > 0.6 in TYCHO-2. We

85

SACY (Torres et al.)

600 500

EW(Li) mA

400 300 200 100 0 0,5

1,0

1,5

B-V Figure 1. Distribution of Li line equivalent width for dwarf stars observed. The lines represent the upper and lower limits for Pleiades Association stars.

observed 1096 sources. We also used published information for 99 others, most of them without interest for our search and the young ones taken mainly from Covino et al. (1997). This defines the area in Figure 2, within which the SACY is complete to about 90% . Unfortunately the southern area without observations is in the core of the Sco-Cen Association. We classified 201 stars as giants and 966 as dwarfs, 421 of them being younger than the Pleiades, 174 having the Pleiades age and 371 older than it.

2.

The Young Associations

Kinematical motions (UVW) have been found for the stars measured by HIPPARCOS. According to our age classification, we find some striking concentrations in the young star UVW space that are not in the same loci of the Pleiades age ones. Older stars lack these concentrations. This can be easily demonstrated with the mpeg-movies on the CD anex. There are four main concentrations in UVW space for young stars that probably define distinct associations. In fact, two could be identified as previously known associations: the Tuc-HorA – that we call GAYA (Great Austral Young Association) for its huge size – and the β Pic association. Kinematically and spatially near GAYA there is an-

86

OPEN ISSUES IN LOCAL STAR FORMATION

α =6

α =18

α =0

-60

-30

Young stars Others

α =12

Project SACY Figure 2. Polar representation of the observed stars where the transversal line is the galatic equator.

other group of stars, but even more widespread and distant. The other association (AnA) is less characterized. Most of our young stars have no HIPPARCOS parallaxes, and we applied the following kinematical analysis to find possible associations: Each point in UVW space is taken as a convergence point and we calculate for it the parallaxes of all stars such as to minimize the moduli of the space velocity vectors relative to this point (but, of course, preserving the parallaxes of HIPPARCOS stars). Then we calculate the density of stars in the velocity space around each point of a grid in UVW. Around some points there are density concentrations much larger than the background fluctuation, revealing possible associations. All the main concentrations are also constricted in space, albeit some of them cover large areas.

87

SACY (Torres et al.) Table 1.

Space motions and parallaxes of the Young Associations

Name

U km/s

σ km/s

V km/s

σ km/s

W km/s

σ km/s

π mas

σ mas

N∗

GAYA1 GAYA2 TWA ChaA LCC US YSSA β PicA OctA ArgusA AnA

-9.1 -11.0 -12.1 -7.7 -8.6 -4.7 -4.0 -9.5 -10.4 -21.5 -7.1

1.1 1.0 0.7 0.6 0.8 1.2 1.3 1.2 0.6 0.8 0.8

-20.9 -22.5 -17.1 -19.7 -21.4 -19.3 -13.4 -16.4 -1.5 -13.1 -28.0

1.0 1.0 0.8 0.8 1.1 0.7 1.0 1.1 0.6 1.1 1.1

-1.2 -4.6 -5.5 -8.5 -5.6 -5.2 -8.0 -9.4 -8.0 -5.1 -12.4

0.9 1.1 0.9 0.9 1.1 1.3 1.2 0.9 0.8 1.4 1.2

22.0 11.9 22.4 10.5 8.7 7.5 8.6 34.1 8.9 9.8 19.0

2.2 4.2 5.6 1.3 1.4 1.4 1.9 27.2 1.5 2.7 10.5

16 41 5+3 15 40 43 21+5 16+1 6 14 11

Table 2. Name

GAYA1 GAYA2 TWA ChaA LCC US YSSA β PicA OctA ArgusA AnA

Positions relative to the Sun and ages X pc

σ pc

Y pc

σ pc

Z pc

σ pc

ρmax pc

age Myr

N∗

12 7 9 47 57 132 117 33 65 30 4

14 27 6 7 12 27 22 28 30 30 32

-25 -78 -41 -82 -100 -28 -10 -8 -73 -95 -35

8 33 10 12 20 18 10 16 7 34 28

-33 -31 20 -6 16 22 -2 -16 -55 -10 -35

5 25 5 14 17 18 30 7 6 25 17

26 84 18 34 75 72 50 50 46 83 69

30 20 8 10 15 8 8 15 30 30? 50

16 41 8 15 40 43 26 17 6 14 11

In Tables 1 & 2 we present the properties of the young associations detected in this way. In Table 1 we give the mean kinematical values and the mean parallax. For some known associations we use bonaf ide members not observed in the SACY to help in their definition. Their numbers are indicated in the last column in Table 1. In Table 2 we give the mean XYZ, the age, and the distance (ρmax ) of the most distant member with respect to the calculated center of the association, giving an idea of its size. The method gives no unique solution for some stars, but in almost all cases we can infer a “best membership”.

88

2.1

OPEN ISSUES IN LOCAL STAR FORMATION

Comments about the associations

2.1.1 GAYA1. We are calling GAYA (Torres et al. 2001) two nearby concentrations on the UVW space, separated mainly in W velocities. Both seem adjacent in the real space. GAYA1 is somewhat older and is one of the more well defined of the associations in SACY and the previous HorA and TucA are within it. Some of the proposed members of TucA are outside of the velocity definition (mainly the eastern ones). Of their 16 proposed members 8 have parallaxes in HIPPARCOS. The spread in distance is small and this does not seem an artifact either of the SACY or our analysis as its derived center is at only 45 pc. We tested 14 of its proposed members for spectroscopic binarity without any positive detection. GAYA1 seems deficient in binaries. 2.1.2 GAYA2. GAYA2 is much less well defined, although it shows a clear concentration using HIPPARCOS stars, but reinforced by members of Lower Cru-Cen (LCC). Actually, the UVW are very near the LCC ones and it is adjacent in space. Nevertheless GAYA2 seems older than LCC and closer to us. 2.1.3 TWA. The TW Hya association is not very well defined in SACY since only two members have trigonometric parallaxes. Torres et al. (2003) present a list of proposed members, but many of them lack information for a complete kinematical analysis. Anyway, we try to use all possible data. The convergence method has problems as TWA is near in velocity and space of  ChaA and LCC. We applied it limiting the possible spatial volume but including any star position in Torres et al. list. Nevertheless our solution excludes many stars in their list. We proposed as bonafide kinematical members: TWA 1, 2, 3, 4, 7, 8, 12 plus a new member, CD-39 7538. 2.1.4  ChaA. This association is defined by Mamajek, Lawson & Feigelson (2000). We propose new members, enlarging it. There is ambiguity for about half of the proposed members between  ChaA and LCC, but the solutions show a consistent separation in UVW space. The distance found by us indicates it is in front of the Cha complex. 2.1.5 LCC. This association has UVW near those of  ChaA and GAYA2 and the age seems between both. The LCC found in SACY is very similar to that found by Sartori et al. (2003) for early-type stars. 2.1.6 US. The Upper Sco (US) has UVW near those of LCC and YSSA. US can easily be separated from LCC in real space, but many

SACY (Torres et al.)

89

stars may be assigned both to US and YSSA. Since we have almost no observations in Upper Cen-Lupus we can not say if they would be separated in SACY.

2.1.7 YSSA. This is a group of young stars, spread from ρ Oph to R Cra, with very similar properties, that we are now calling the Young Sco-Sgr Association. The western border engulfs the stars mentioned in Quast et al. (2001) and Neuh¨ auser et al. (2000). The split in space distribution can be explained by the incompleteness in the RASS coverage. Anyway, the convergence process gives some superposition with US association. The distance is near the assumed one for the R CrA cloud. 2.1.8 β PicA. As described by Zuckerman et al.(2001) this association is very close to the Sun. We propose new members, some of then as far as 80 pc, but the distribution in space seems very consistent. Among the new proposed members is V4046 Sgr, a notorious object, classified before as an isolated SB classical TTS (de la Reza et al., 1986; Quast, 1998; Quast et al. 2000). WW PsA and TX PsA could be members (Song et al. 2002), but their parallaxes should be 49.5 mas, closer than HIPPARCOS ones, about 2σ of the HIPPARCOS errors. 2.1.9 OctA. This is a very homogeneous small group of almost aligned stars (all young G stars) near the South Celestial Pole. Since this region belongs to a completely surveyed area of the SACY, new members have to be found by other means. 2.1.10 ArgusA. Although not very well defined, it has a special position in UVW. Since many stars are in Car, Vel and Pup we tentatively propose to call it as Argus A. 2.1.11 AnA. Like ArgusA, the main reason to claim for this possible association are the very special UVW. The majority of the proposed members have parallaxes and, therefore, this is one of the concentrations in the HIPPARCOS sample. GAYA1, GAYA2, LCC, US and YSSA form a decreasing sequence in age, going from west to east, and they seem to form a kind of continuum in UVW space. All the associations but the last three in the tables can represent local aspects of a global star forming process. More details of these associations can be seen in the poster of Quast et al. in the enclosed CD.

90

OPEN ISSUES IN LOCAL STAR FORMATION

Acknowledgments This work was partially supported by a CNPq - Brazil grant to C. A. O. Torres (pr. 200356/02-0).

References Covino, E., Alcal´ a, J.M., Allain, S., Bouvier, J., Terranegra, L., & Krautter, J. 1997, A&A, 328, 187 de la Reza, J. R., da Silva, L., Jilinski, E., Torres, C. A. O., & Quast, G. R. 2001, in ASP Conf. Ser. Vol. 244, Young stars near earth: progress and prospects, ed. R. Jayawardhana & T. P. Greene (San Francisco: ASP), 37 de la Reza, R., Quast, G. R., Torres, C. A. O., Mayor & M. Vieira, G.V. 1986, in Symp.NASA-ESA. New Insights in Astrophysics, ESA S-263, 107 de la Reza, R., Torres, C. A. O., Quast, G. R., Castilho, B.V., & Vieira, G.L. 1989, ApJL, 343, L61 Gregorio-Hetem, J., L´epine, J. R. D., Quast, G. R., Torres, C. A. O., & de la Reza, R. 1992, AJ, 103, 549 Jeffries, R. D. 1995, MNRAS, 273, 559 Kaufer, A., Stahl, S., Tubbesing, S., Norregaard, P., Avila, G., Francois, P., Pasquini, L., & Pizzella, A. 1999, Messenger, 95, 8 Mamajek, E. E., Lawson, W. A., & Feigelson, E. D. 2000, ApJ, 544, 356 Neuh¨ auser, R. 1997, Science, 276, 1363 Neuh¨ auser, R., Walter, F. M., Covino, E., Alcal´a, J. M., Wolk, S. J., Frink, S., Guillout, P., Sterzik, M. F., & Comer´ on, F. 2000, A&AS, 146, 323 Quast, G.R. 1998, thesis ON-Rio de Janeiro Quast, G. R., Torres, C. A. O., de la Reza, J. R., da Silva, L., & Drake, N. 2001, in ASP Conf. Ser. Vol. 244, Young stars near earth: progress and prospects, ed. R. Jayawardhana & T. P. Greene (San Francisco: ASP), 49 Quast, G.R.; Torres, C. A. O.; de la Reza, R., da Silva, L., & Mayor, M. 2000, IAU Symposium No. 200 “Birth and Evolution of Binary Stars”, Potsdam, Germany, 28 Queloz, D., Mayor, M., Naef, D., Santos, N., Udry, S., Burnet, M., & Confino, B. 2000, in VLT Opening Symposium ¿From Extrasolar Planets to Brown Dwarfs, ESO Astrophys. Symp., Springer Verlag, Heidelberg, 548 Sartori, M. J., L´epine, J. R. D., & Dias, W. S. 2003, A&A, 404, 913 Song, I., Bessel, M. S., & Zuckerman, B. 2002 ApJL, 581, L434 Torres, C. A. O., Quast, G., de la Reza, R., Gregorio-Hetem, J., & L´epine, J. R. D. 1995, AJ, 109, 2146 Torres, C. A. O. 1998, Publica¸c˜ ao Especial do Observat´ orio Nacional, 10/99 Torres, C. A. O., da Silva, L., Quast, G., de la Reza, R., & Jilinski, E. 2000, AJ, 120, 1410 Torres, C. A. O., Quast, G. R., de la Reza, J. R., da Silva, L., Melo, & C. H. F. 2001, in ASP Conf. Ser. Vol. 244, Young stars near earth: progress and prospects, ed. R. Jayawardhana & T. P. Greene (San Francisco: ASP), 43 Torres, G., Guenther, E. W., Marschall, L. A., Neuh¨auser, R., Latham, D. W., & Stefanik, R. P. 2003, AJ 125, 825 Zuckerman, B., Sing, I., Bessell, M. S., & Webb, R. A 2001, ApJL, 562, L87 Zuckerman, B., & Webb, R. A. 2000, ApJ, 535, 959

AGE DETERMINATION OF THE URSA MAJOR ASSOCIATION The companion of

χ1

Orionis

Brigitte K¨onig Max-Planck-Intitut f¨ ur extraterrestrische Physik, Germany [email protected]

Abstract

The center of the Ursa Major association is located at ∼ 25 pc from the sun making it one of the closest associations which host pre-mainsequence stars. The age derived for the association ranges between 200 to 500 Myr where the canonical age is 300 Myr derived by Soderblom & Mayor (1993). More recent age determinations are from K¨ onig et al. (2002) who derive an age of 100 ± 30 Myr based on Baraffe et al. (1998) pre-main-sequence models and from King et al. (2003) based on post-main-sequence models for the A-stars similar to Soderblom & Mayor (1993). In the further we will discuss the age of the Ursa Major association derived from pre-main-sequence models of different authors for the low-mass companion to χ1 Orionis A.

Introduction The Ursa Major association (UMa) is a young association with its center located at about 25 pc. It contains high-mass evolved stars such as the big dipper stars, a white dwarf as a companion to Sirius, F-, Gand late K-stars which have already arrived on the main sequence and pre-main sequence M-stars. Thorough studies to derive the age of the association have been carried out by Soderblom & Mayor (1993) who have used the evolved A-stars of the association to place them into a Hertzsprung-Russel diagram and compare to post-main-sequence models of Vandenberg (1985). The age they derive is the often quoted canonical age of the UMa of 300 Myr. Recently, King et al. (2003) have revised this work and derive a much older age of about 500 Myr with the same method but comparing the position of the A-stars in the H-R diagram to the tracks and isochrones of Vandenberg (1985). It is known that Sirius host a white dwarf companion which is well studied. Holberg et 91 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 91-96. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

92

OPEN ISSUES IN LOCAL STAR FORMATION

al. (1998) gives an age of 160 Myr and a lifetime of the progenitor of 60-70 Myr. More recent calculations from Salaris et al. (2000) revised these values to 111 Myrs plus the lifetime of its progenitor of 46 Myrs. In the UMa we are therefore in the unique situation to be able to compare three different methods to derive ages for stars.

The dynamical mass of χ1 Ori A & B

1.

The binary χ1 Ori A & B is a well-studied object. Han & Gatewood (2002) have published precise radial velocity and astrometric measurements of the orbit. All orbital elements were up to then known except the apparent separation, or the semi-major axis. Table 1. The orbital parameters of the radial velocity and astrometric measurements published by Han & Gatewood (2002) (left), and the measured values from the direct imaging of K¨ onig et al. (2002) (right). Han & Gatewood (2002) Parameter unit α1 Ω i ω T0 P e K1 f (m) q = MB /MA

[mas] [deg] [deg] [deg] [JD] [day]

π

[mas]

[km/s] [M
]

value

89.662 ± 0.880 126.360 ± 0.583 95.937 ± 0.790 111.527 ± 0.230 2, 451, 468.2 ± 3.083 5156.291 ± 2.508 0.452 ± 0.002 1.876 ± 0.003 2.50 × 10−3 0.15 ± 0.005

115.69 ± 0.74

K¨ onig et al. (2002) parameter unit value app. sep. phys. sep. pos. angl. Tobs mH MH

[”] [AU] [◦ ] [MJD] [mag] [mag]

0.4976 ± 0.0036 4.33 ± 0.08 123.22 ± 0.12 52334.33952 7.70 ± 0.15 8.01 ± 0.15

πHIP

[mas]

115.43 ± 1.08

The direct detection of the secondary succeeded in February 2002 using the Keck II telescope equipped with the NIRC 2 camera and the adaptive optics system. This enabled us to measure the separation and calculate the semi-major axis und thus the dynamical masses of χ1 Ori A & B: MA = 1.01 ± 0.13 M
, and MB = 0.15 ± 0.02 M
using the Hipparcos parallax or MA = 1.02 ± 0.13 M
, and MB = 0.15 ± 0.02 M
using the parallax form Han & Gatewood (2002), respectively.

2.

Ages derived from pre-main-sequence models χ1

Ori A is a G-type which has already arrived on the main sequence. But we know that the star is a member of the UMa so that we can assume χ1 Ori has the same age as the association. We place the companion B

The age of the Ursa Major Association (K¨ onig)

93

in the H-R diagram and compare its position to isochrones of pre-mainsequence models.

2.1

Baraffe et al. (1998) models

Baraffe et al. (1998) (BCAH1998) have calculated tracks and isochrones in the mass range of the primary and the secondary. We place both stars into the H-R diagram and overlay the tracks and isochrones.

Figure 1. Left: (a) The primary together with tracks and isochrones from BCAH1998. These tracks were especially calculated to fit the sun ([Fe/H]=0.00, Y=2.82, lmix = 1.9HP ). The primary is fit quite well especially if one would apply corrections for the iron abundance of the star of [Fe/H]= 0.07 ± 0.07. Grey shaded is the position of the star within its errors. Right: (b) Same as for Fig. 1a but with the tracks and isochrones calculated for the model with [Fe/H]=0.00, Y=2.75, lmix = HP . The tracks do not reproduce the temperature of the primary, especially its position in the H-R diagram.

Figure 2. Same as for Fig. 1b but at the position of the secondary. The upper grey shaded area is the position of the secondary at a mass of 0.15 ± 0.02 M
and an absolute H-band magnitude of 8.01 ± 0.15 mag. The lower grey shaded area marks the 200 to 300 Myr age range of the UMa.

94

OPEN ISSUES IN LOCAL STAR FORMATION

From the position of the secondary in Fig. 3 we derive an age of the secondary of 100 ± 30 Myr.

2.2

D’Antona & Mazzitelli (1998) models

DAM1998 provide tracks and isochrones only for the mass range of the secondary not including both masses in on set of model calculations. We therefore cannot check the consistency of the tracks with the position of the primary as we did for the Baraffe et al. (1998)Baraffe, Chabrier, Allard, & Hauschildt model. The models of DAM1998 are calculated for a grey atmosphere. We have to convert the bolometric luminosity they published to absolute H-band magnitudes we measure. The bolometric corrections are taken from BCH1998, Kenyon et al. (1994) (KGMH1994), and Siess et al. (2000) (SDF2000). The resulting isochrones in the mass - H-band magnitude plot are shown below. For the DAM1998 models we

Figure 3. D’Antona & Mazzitelli (1998) isochrones for the initial Deuterium mass fraction of 2 · 10−5 . The bolometric corrections are taken from BCAH1998 and from KGMH1994. The shaded area mark the mass of the secondary with error range once calculated as the dynamical mass and one from the modeling of the primary.

Figure 4. D’Antona & Mazzitelli (1998) isochrones for the initial Deuterium mass fraction of 2 · 10−5 . The corrections are taken from SFD1997. Shaded area as in Fig. 3.

also derive an age of 100 ± 30 Myr if we use the bolometric corrections from Baraffe et al. (1998)Baraffe, Chabrier, Allard, & Hauschildt, or from KGMH1994. The bolometric corrections from Siess et al. (1997) ( SFD1997) make the star to come to lie at a very old age which does not seem reasonable.

The age of the Ursa Major Association (K¨ onig)

2.3

95

Siess et al. (2000) models

Siess et al. (2000)Siess, Dufour, & Forestini have also calculated grey models for the mass range of the secondary once without and once with overshooting. The bolometric luminosity again has to be converted to Hband magnitudes. We use again KGMH1994, and SFD1997. BCAH1998 bolometric corrections at a mass of 0.15 M
are similar to KGMH1994 bolometric corrections.

Figure 5. SDF2000 isochrones of the model without convective overshooting and with bolometric corrections from KGMH1994.

Figure 6. SDF2000 isochrones of the model without convective overshooting and with bolometric corrections from SFD1997.

The SDF2000 models using bolometric corrections from KGMH1994 give a slightly older age for χ1 Ori B of 150 ± 30 Myr while the SDF2000 models with bolometric corrections from SFD1997 again make the star come to lie outside the range of the model calculations.

Figure 7. SDF2000 isochrones of the model with convective overshooting and with bolometric corrections from KGMH1994.

Figure 8. SDF2000 isochrones of the model with convective overshooting and with bolometric corrections from SFD1997.

96

OPEN ISSUES IN LOCAL STAR FORMATION

The SDF2000 model with overshooting and bolometric corrections from KGMH1994 result in a somewhat younger age of the secondary of 110 ± 30 Myr.

3.

Conclusions

The age of the UMa ranges from 160 Myr derived from cooling models for the white dwarf companion of Sirius and 500 Myrs derived from postmain-sequence modeling of cluster members. We compare these results to the ages we derive from pre-main-sequence tracks and isochrones from BCAH1998 (100 ± 30 Myr), DAM1998 (100 ± 30 Myr), and SDF2000 (150 ± 30 Myr, 110 ± 30 Myr). The bolometric corrections and/or the model assumption of the atmosphere influence the age for the UMa derived by pre-main-sequence models. Compared to the post-main-sequence models as well as the white dwarf cooling models we are somewhat inconclusive about the true age of the association. Yet, the spectra of the F-, G-, and late K-type stars would support an age of 200 Myr rather than 500 Myr regarding the activity indicators such as Hα, Ca H&K, Mg-Ib line core filling-in and the Lithium absorption at 6708 ˚ A.

Acknowledgments The author wishes to thank D. Charbonneau, R. Jayawardhana, and R. Neuh¨ auser for the possibility to observe at the Keck II telescope.

References Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H., 1998, A&A, 337, 403 D’Antona, F. & Mazzitelli, I., 1998, ASP Conf. Ser. 134: Brown Dwarfs and Extrasolar Planets, 442 Han, I. & Gatewood, G., 2002, PASP, 114, 224 Holberg, J. B., Barstow, M. A., Bruhweiler, F. C., Cruise, A. M., & Penny, A. J., 1998, ApJ, 497, 935 Kenyon, S. J., Gomez, M., Marzke, R. O., Hartmann, L., 1994, AJ, 108, 251 King, J. R., Villarreal, A. R., Soderblom, D. R., Gulliver, A. F., & Adelman, S. J., 2003, AJ, 125, 1980 K¨ onig, B., Fuhrmann, K., Neuh¨ auser, R., Charbonneau, D., & Jayawardhana, R., 2002, A&A, 394, L43 Salaris, M., Garc´ıa-Berro, E., Hernanz, M., Isern, J., & Saumon, D., 2000, ApJ, 544, 1036 Siess, L., Dufour, E., Forestini, M., 2000, A&A, 358, 593 Siess, L., Forestini, M., Dougados, C., 1997, A&A, 324, 556 Soderblom, D. R. & Mayor, M., 1993, AJ, 105, 226 Vandenberg, D. A., 1985, ApJS, 58, 711

II

YOUNG STELLAR OBJECTS

This page intentionally left blank

ACCRETION POWERED EMISSION IN YOUNG STELLAR OBJECTS Nuria Calvet Smithsonian Astrophysical Observatory [email protected]

James Muzerolle Steward Observatory, University of Arizona

Abstract

We discuss evidence supporting the magnetospheric accretion mode of mass transfer onto the star for a much larger range of masses that previously known, ∼ 0.03M
− ∼ 5M
, and show corresponding mass accretion rate determination over this mass range. We find that the mass accretion rate M˙ increases with stellar mass M∗ , roughly as M˙ ∝ M∗2 . The physical principles behind this correlation are unknown at present; they must involve mechanisms for transporting angular momentum outwards in the disk which must be active for accretion to take place. The recently found correlation between X-ray luminosity and stellar mass may play a role, by influencing the ionization degree in the disk, necessary for the magnetorotational instability, the most accepted mechanism driving accretion, to operate.

Introduction Magnetospheric accretion is now the accepted model for mass transfer from the disk onto the star in the low mass pre-main sequence T Tauri stars (TTS) (Hartmann 1998, and references therein). In this model, the stellar magnetic field disrupts the disk at a few stellar radii, and matter falls onto the star along magnetic field lines, merging with the photospheric material through an accretion shock on the stellar surface. Many lines of evidence support this model. On the one hand, stellar magnetic field strength of the order of a few KG (Johns-Krull et al. 1999; Guenther et al. 1999) and average mass accretion rates of ∼ 10−8 M
yr−1 (Gullbring et al. 1998; Hartmann et al. 1998) in TTS are consistent with disk truncation at a few stellar radii. On the other, 99 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 99-106. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

100

OPEN ISSUES IN LOCAL STAR FORMATION

magnetospheric accretion can naturally explain many of the emission properties that characterize TTS. For instance, the broad emission lines present in the spectra of TTS can be very well understood as forming in the magnetospheric accretion flow (Figure 1a, Muzerolle et al. 2001 and references therein), while the excess flux which veils the optical spectra of TTS and dominates the emission in the ultraviolet (UV) can be explained by accretion shock emission (Figure 1b, Calvet & Gullbring 1998).

Figure 1a. Hα line profile of BP Tau, solid line, compared to the prediction of the magnetospheric infall model, dashed line (from Muzerolle et al. 2003)

Figure 1b. UV-optical SED of BP Tau, heavy solid line, compared to the prediction of the accretion shock model, light solid line (from Gullbring et al. 2001). The contributions from the photosphere and the shock are indicated, with the emission from the heated photosphere (dashed) and the preshock region shown as well.

By means of magnetospheric accretion, potential energy of infalling matter accounts for the excess above purely stellar/photospheric energy in TTS. The accretion luminosity, Lacc = GM∗ M˙ /R∗ , can be derived from the luminosity of the excess flux identified as shock emission. In average it corresponds to 10% of the stellar energy, but in some cases can be equal or even larger. The mass accretion rate, M˙ , can be derived from Lacc , with knowledge of the stellar mass and radius, M∗ and R∗ , from the position in the HR digram. So far, the validity of magnetospheric accretion has been checked for stars in the mass range 0.1M
to 1 M
. We have now evidence that it applies to a much larger range of masses from ∼ 0.01M
to ∼ 5M
, covering most of the stellar mass spectra. We discuss here this evidence and its implications.

101

Accretion in YSO (Calvet & Muzerolle)

1.

Magnetospheric accretion in the substellar limit

The broad line profiles which are the signposts of magnetospheric accretion have been detected in a number of stars near and below the sub-stellar limit (Muzerolle et al. 2000, 2003a; Jayawardhana, Mohanty, & Basri 2002; White & Basri 2003). Quantitative estimates of the accretion properties can be made by modeling the accreting region with magnetospheric models that successfully describe the higher mass TTS. Such models have been constructed by Muzerolle et al. (200x, 2003a) for a sample of 13 very low mass objects (VLMS), which were found to have broad line profiles which could be explained in terms of magnetospheric accretion; Figure 2 shows a representative line profile, compared to theoretical predictions. This fitting procedure yielded mass accretion rates for the VLMS, which turned out to be very low, ˙ ≤ 10−9 M
yr−1 , consistent with the lack of optical 10−12 M
yr−1 ≤ M veiling.

Figure 2. Hα line profile of the Very Low Mass object IC348-336, solid line, compared to the magnetospheric model prediction (Muzerolle et al. 2003a).

Figure 3. Hα line profile of the Herbig Ae/Be star UX Ori compared to the magnetospheric model prediction (Muzerolle et al. 2003b).

The importance of Hα as an indicator of very low accretion is highlighted by the study of the VLMS. Accretion shock emission is too low to be detected, and even at short wavelengths as in the UV, it would be difficult to separate it from stellar chromospheric emission. The Hα line, on the other hand, is a high opacity line and thus even low density gas in the magnetospheric flow gives significant, broad emission, down to M˙ ∼ 10−12 M
yr−1 (Muzerolle et al. 2003a).

102

2.

OPEN ISSUES IN LOCAL STAR FORMATION

Magnetospheric accretion in the 1 - 5 M
range

The best known pre-main sequence stars in the mass range 1 - 5 M
range are the Herbig Ae/Be stars. The pre-main sequence nature has been now well established for the Ae stars in this group (HAe, Natta et al. 2001), which show broad line profiles consistent with magnetospheric accretion, as shown in Figure 3 (Muzerolle et al. 2003b). Two effects may the typical magnetospheric line profile in HAe different in appearance to that of the lower mass TTS: (1) red-shifted absorption components at velocities of a few hundred kms−1 are often seen in TTS; in contrast, the absorption component in HAe is broad and central to slightly red-shifted. In TTS, this absorption arises when the line of sight intersects material on the accretion flow infalling at free-fall velocities, which is located just above the shock region (Muzerolle et al. 2001), emitting in the optical at a color temperature of ∼ 8000K (CG98). In the HAe case, the whole photosphere is emitting at ∼ 8000K, so these photons can be absorbed by the entire magnetospheric flow, covering a much wider velocity range, from nearly zero near the disk, to several hundred kms−1 near the star. (2) The HAe are rotating much faster than the TTS, so effects of rotation in the line profile are more noticeable. The higher rotation rates also restrict the size of the magnetospheric region to at most ∼ 3R∗ , since material cannot accrete from outside the corotation radius. However, even if magnetospheric accretion is taking place in HAe, it is difficult to measure their mass accretion rate. Profile modeling is consistent with low mass accretion rates, M˙ ≤ 10−7 M
yr−1 , but the degeneracy between mass accretion rate and temperature (empirically set, see Muzerolle et al. 2001 for a discussion), yields a large uncertainty in the M˙ determination 1 . On the other hand, shock emission, with similar color temperature as the underlying photosphere, is difficult to separate from it. This hampers attempts to measure the luminosity of the excess flux due to the shock. However, there is a seldom recognized set of stars, the intermediate mass T Tauri stars (IMTTS), which in the HR diagram are located in the same radiative tracks as the HAe, but in an earlier evolutionary time. Thus, these stars cover the same range as the HAe, but have

1 Note that this uncertainty is much smaller in the VLMS because at the high temperatures ˙ of these stars, H is essentially all ionized, see discussion in Muzerolle required at the low M et al. (2003a).

Accretion in YSO (Calvet & Muzerolle)

103

cooler photospheres (early K to early G), making it easier to extract the shock emission, specially in the near UV. With this in mind, HST STIS spectra for a sample of IMTTS were obtained and analyzed with shock models previously applied to low mass TTS (Calvet et al. 2003). The excess flux in the near UV can be well understood in terms of terms of shock emission, as shown by a representative spectrum in Figure 4. By these procedure, mass accretion ˙ < 10−6 M
yr−1 have been obtained rates in the range 10−8 M
yr−1 ≤ M for these stars.

Figure 4. Ultraviolet fluxes of the Intermediate Mass T Tauri star EZ Ori, solid light lines, and simultaneous optical photometry, filled dots, together with prediction of the accretion shock model, heavy solid lines, and the photosphere, dotted lines. The inset shows a detail of the NUV region (Calvet et al. 2003).

3.

Correlation between mass accretion rate and stellar mass

After determining the mass accretion rate for an unprecedented twoorder of magnitude range of masses, from ∼ 0.03M
to ∼ 5M
, essentially covering the mass range where all stars are found, we find a striking correlation of mass accretion rate with mass, shown in Figure 5. Measurements from several sources, in addition to our own, are shown in this figure, with credits in the Figure caption. Note that all the objects shown in Figure 5 have similar ages, a few million years, so that age effects are not expected to play a role in the observed dependence. The correlation can be fitted with M˙ ∝ M∗2 .

104

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 5. Mass accretion rate vs. stellar mass for pre-main sequence stars. Squares from Muzerolle et al. (2003a); crosses from White & Ghez (2001); filled circles from Gullbring et al. (1998); open circles from White & Basri (2003); triangles from Calvet et al. (2003).

The large spread, specially at low masses, could be accounted for uncertainties in the mass accretion rate determination. An intrinsic spread, due to variability (cf. Alencar and Batalha 2002) is also expected. It could be asked if the observed correlation is rather an upper envelope, in the sense that low M˙ are difficult to detect. This does not seem to be the case at least for the the best studied objects, the TTS, for which the limit for detection based on broad Hα is 10−11 M
yr−1 , way below the lower envelope of the correlation in Figure 5 (Muzerolle et al. 2003a). Similar correlations between mass accretion rate and stellar mass have been found before, for instance, in Bondi accretion (Bondi 1952). However, the physics behind the correlation shown in Figure 5 and other types of accretion involving only gravity is profoundly different. Although gravity from the central star plays a role, mechanisms for transporting angular momentum outwards in the disk must be active for accretion to take place. These mechanisms are presently poorly known. The most promising among them is the magnetorotational instability (MRI, Balbus & Hawley 1992). Under the MRI, different parts of the disk rotating at different velocities are tied up by magnetic fields, effectively transferring angular momentum outwards. Still, the MRI can only

Accretion in YSO (Calvet & Muzerolle)

105

operate if a minimum ionization degree is present, resulting in active accretion only in the upper layers of the inner denser part of the disk, where the required degree of ionization can be maintained by cosmic rays and/or stellar X-rays (Gammie 1996; Glassgold et al. 1997). Nonetheless, the origin of the observed M˙ vs. M∗ correlation may lay in the operation of the MRI. A correlation between X-ray luminosity and stellar mass has been observed (Flaccomio et al. 2003; Feigelson et al. 2003). An increased X-ray luminosity may determine a higher ionization degree which may led to a higher accretion rate as stellar mass increases. However, this a complete open question at the present. Calculations for the transfer of X-rays and resultant ionization degree, including trapping of charge particles by dust particles, in realistic pre-main sequence disk models need to be carried out. This new result represents an important challenge to theoretical studies aiming to understanding the fundamental principles of disk accretion in YSO.

Acknowledgments The work described here refers to the latest results of ongoing research into the nature of the accretion processes in YSO, carried out with collaborators James Muzerolle, Lee Hartmann, Erik Gullbring, C´esar Brice˜ no, Paola D’Alessio, Lynne Hillenbrand, Jes´ us Hern´andez, and Jose Luis Saucedo. This work has been partially supported by Origins of Solar Systems Program under NASA grant NAG5-9670 and by grant GO08317.01-97A from the Space Telescope Science Institute.

References Alencar, S. H. P. & Batalha, C. 2002, ApJ, 571, 378 Balbus, S. A. & Hawley, J. F. 1992, ApJ, 400, 610 Bondi, H. 1952, MNRAS, 112, 195 Calvet, N. & Gullbring, E. 1998, ApJ, 509, 802 Calvet, N. et al. 2003, in preparation. Feigelson, E. D., Gaffney, J. A., Garmire, G., Hillenbrand, L. A., & Townsley, L. 2003, ApJ, 584, 911 Flaccomio, E., Micela, G., & Sciortino, S. 2003, AA, 402, 277 Gammie, C. F. 1996, ApJ, 457, 355 Glassgold, A. E., Najita, J., & Igea, J. 1997, ApJ, 480, 344 Gullbring, E., Hartmann, L., Briceno, C., & Calvet, N. 1998, ApJ, 492, 323 Gullbring et al 2001 Gullbring, E., Calvet, N., Muzerolle, J., & Hartmann, L. 2000, ApJ, 544, 927 Guenther, E. W., Lehmann, H., Emerson, J. P., & Staude, J. 1999, AA, 341, 768 Hartmann, L. 1998, Accretion processes in star formation / Lee Hartmann. Cambridge, UK ; New York : Cambridge University Press, 1998. (Cambridge astrophysics series ; 32) Hartmann, L., Calvet, N., Gullbring, E., & D’Alessio, P. 1998, ApJ, 495, 385

106

OPEN ISSUES IN LOCAL STAR FORMATION

Jayawardhana, R., Mohanty, S., & Basri, G. 2002, ApJL, 578, L141 Johns-Krull, C. M., Valenti, J. A., & Koresko, C. 1999, ApJ, 516, 900 Muzerolle, J., Calvet, N., & Hartmann, L. 2001, ApJ, 550, 944 Muzerolle, J., Brice˜ no, C., Calvet, N., Hartmann, L., Hillenbrand, L., & Gullbring, E. 2000, ApJL, 545, L141 Muzerolle, J., Hillenbrand, L., Calvet, N., Brice˜ no, C., & Hartmann, L. 2003a, ApJ, (in press) Muzerolle, J., D’Alessio, P., Calvet, N., & Hartmann, L. 2003b, ApJ, (submitted) Natta, A., Prusti, T., Neri, R., Wooden, D., Grinin, V. P., & Mannings, V. 2001, AA, 371, 186 White, R. J. & Ghez, A. M. 2001, ApJ, 556, 265 White, R. J. & Basri, G. 2003, ApJ, 582, 1109

Nuria Calvet

THE PRE-MAIN SEQUENCE SPECTROSCOPIC BINARY AK SCO S. H. P. Alencar, L. P. R. Vaz Departamento de F´ısica, ICEx-UFMG, Brazil

C. H. F. Melo European Southern Observatory, Chile

C. P. Dullemond Max Planck Institut f¨ ur Astrophysik, Germany

J. Andersen Niels Bohr Institute for Astronomy, Physics, and Geophysics, Denmark

C. Batalha Observat´ orio Nacional, Brazil

R. D. Mathieu University of Wisconsin-Madison, U.S.A.

Abstract

We present an analysis of 32 high-resolution echelle spectra of the premain sequence spectroscopic binary AK Sco obtained during 1998 and 2000, as well as a total of 71 photoelectric radial-velocity observations from the period 1986-1994. These data allow considerable improvement of the period and other orbital parameters of AK Sco. Our analysis also includes eight series of photometric observations in the uvby and Geneva seven-color systems. No eclipses or other periodic variations are seen in the photometry, but the well-determined HIPPARCOS parallax allows us to constrain the orbital inclination of the system and obtain physical parameters for the two stars. 107

J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 107-114. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

108

OPEN ISSUES IN LOCAL STAR FORMATION Disk models have been fitted to the spectral energy distribution of AK Sco from 350 nm to 1100 µm. The stellar parameters permit a consistent solution with an inner rim temperature of 1250 K, instead of the usual 1500 K corresponding to the dust evaporation temperature. Dynamical effects due to tidal interaction of the binary system are supposed to be responsible for pushing the disk inner radius outwards. The spectrum of AK Sco exhibits emission and absorption lines that show substantial variety and variability in shape. The system displays variations at the binary orbital period in the Hα equivalent width, although with appreciable scatter. We find no evidence of enhanced accretion near the periastron passage in AK Sco as expected theoretically and observed previously in DQ Tau, a similarly young binary system with a mass ratio near unity and an eccentric orbit.

Introduction Classical T Tauri stars are young stars of near-solar mass that exhibit a wide variety of permitted and sometimes also forbidden emission lines, together with excess continuum emission that ranges from millimeter to ultraviolet wavelengths. Their spectral energy distribution is consistent with the presence of a circumstellar disk that appears to play a major role in the regulation of the mass flows in the star–disk system. In a binary system, disks can surround the individual components (circumstellar disks) as well as the binary system as a whole (circumbinary disk). From a theoretical point of view, it is believed that dynamical interactions due to tidal and resonant forces will open gaps between the circumbinary and circumstellar disk(s), preventing the flow of material between them (Artymowicz & Lubow 1994). Accretion is known to occur in binary systems at the same level as in single stars, regardless of the separation between the components ( Mathieu 1994). In spectroscopic binaries with periods of a few days, the components are too close to allow for the existence of circumstellar disks around the individual stars, suggesting that matter is probably being accreted directly from a circumbinary disk. Artymowicz & Lubow (1996) have shown that under special circumstances, a flow of matter can occur in the form of time-dependent gas streams. According to their calculations, in the case of binary systems with mass ratios near one and eccentric orbits (e ∼ 0.5), the mass transfer rate would reach a maximum at orbital phases 0.8-1.15, i.e. near periastron. Mathieu et al. (1997) and Basri et al. (1997) have shown that in the binary system DQ Tau, both spectroscopic and photometric variations indicate that the system is very likely experiencing episodes of pulsed accretion near the periastron passage.

The PMS spectroscopic binary AK Sco (Alencar et al.)

109

In order to test this scenario in another well-observed eccentric binary system, we have undertaken a reanalysis of the Pre-Main Sequence (PMS) spectroscopic binary AK Sco (HD 152404; F5IVe; Vmax =8.9), which became known as a PMS star when Herbig & Rao (1972) discovered strong Hα emission and Li I absorption lines in its spectrum. Later, Andersen et al. (1989) showed that AK Sco is actually a double-lined spectroscopic binary of short period (13.6 days) and large eccentricity (e = 0.47). No evidence for eclipses could be found in the photometry then available in the literature. We have obtained considerable amounts of new spectroscopic and photometric data for AK Sco and discuss in the following some of our results. A more detailed analysis can be found in a companion paper (Alencar et al. (2003)).

1.

Spectroscopic observations

We obtained CORAVEL radial velocities, usually for both stars, for a total of 72 epochs during the period 1986 - 1994. In order to study the line profile variations of AK Sco in more detail, we also obtained high-resolution spectra of the system in 1998 and 2000 with the Swiss 1.2m Euler telescope and CORALIE echelle spectrograph, and in 2000 with the ESO 1.52m telescope and FEROS echelle spectrograph, both at ESO, La Silla, Chile. Projected rotational velocities (v sin i) were computed using the calibrations given by Melo et al. (2001) for FEROS and by Santos et al.(2002) for CORALIE, giving mean rotational velocities of v sin iA =18±1 kms−1 and v sin iB =19±1 kms−1 . We adopt an average v sin i for both components of 18.5±1.0 kms−1 .

2.

Photometry

Figure 1 shows the y magnitude differences between AK Sco and the comparison star SAO 208122 obtained from 1987 to 1994. Despite the good coverage of the critical phases where eclipses were expected to occur, no trace of any eclipses or any other periodic variation was found, showing that the orbital inclination cannot be very high. The irregular light variations in AK Sco are similar to those of other Herbig Ae/Be stars (see e.g. Bibo & Th´e 1991). The variations are quite large, with brightness changes of up to one magnitude between consecutive nights and globally exceeding 1.5 magnitudes.

3.

Orbital and stellar parameters

From the radial-velocity data we recompute the spectroscopic orbital elements for both components, using the SBOP program (Etzel 1985).

110

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 1. y magnitude differences between AK Sco and SAO 208122 obtained from 1987 to 1994, folded with the orbital period in phases relative to the periastron passage. The dashed lines show phases where eclipses were expected to occur.

We obtained P (days) = 13.609453 ± 0.000026, e = 0.4712 ± 0.0020, MB /MA = 0.987 ± 0.005, a sin i (AU) = 0.14318 ± 0.00005, and the final orbital solution is shown in Fig. 2. There is good agreement with the results by Andersen et al. (1989), but our errors are substantially smaller.

Figure 2. Radial velocity curve of AK Sco. The filled circles show the CORAVEL measurements and the open circles are the FEROS/CORALIE results. The standard deviation of the fit is shown at lower right.

The PMS spectroscopic binary AK Sco (Alencar et al.)

111

Figure 3. Hα equivalent width folded at the orbital period. Different symbols refer to different orbital cycles. Negative values indicate that the line was mostly in absorption.

4.

Accretion related lines

The Balmer line profiles are highly variable in AK Sco. We searched for periodic variations in these lines using our spectra from April and May 2000, the only dataset that can be used to study variations with the orbital period (13.61 days), as we have 22 spectra taken over a period of 40 nights. The equivalent widths of the Hα line profiles display variations at the orbital period (with some scatter, see Fig. 3). We find no evidence of the accretion outbursts near the periastron passage predicted for eccentric binaries with mass ratio close to one. Instead, we observe Hα equivalent width variations that are smooth with an amplitude of more than 3 times the mean value, and which look rather like the mass accretion rate variations predicted by Artymowicz & Lubow (1996) for low eccentricity systems with unequal components (see their Fig.2 – top panel), which is rather puzzling.

5.

Constraining the system parameters

The masses and radii of the stars in AK Sco are key parameters in any theoretical modeling of the system. Unfortunately, as no eclipses are observed, the orbital inclination and hence the absolute masses and radii cannot be determined directly. However, indirect evidence allows us to

112

OPEN ISSUES IN LOCAL STAR FORMATION

constrain the parameters within narrow ranges. Briefly, the absence of eclipses defines an upper limit to the inclination, hence a lower limit to the masses, while the distance from the Hipparcos parallax and the known luminosity and effective temperatures of the stars constrain the radii, as do the observed rotational velocities. Exploiting these facts, we constrain the orbital inclination of the system to the range 65◦ < i < 70◦ , leading to the following physical parameters for the two stars: MA = 1.36 ± 0.07 M
and RA = 1.59 ± 0.35 R
;MB = 1.34 ± 0.07 M
and RB = 1.58 ± 0.35 R
.

6.

Disk models

We have derived the spectral energy distribution (SED) of AK Sco from optical to millimeter wavelengths, using our own photometry and data from the literature. Despite its relatively late spectral type of F5, AK Sco shows a near infrared bump in the SED, which is typical of Herbig Ae systems (Natta et al. 2001). We decided to fit the SED of AK Sco with the disk model for Herbig Ae/Be stars by Dullemond, Dominik & Natta (2001, henceforth DDN01). This is a modified version of the Chiang & Goldreich (1997) flaring disk model, with the central disk regions removed and a puffed-up inner rim included. In the original model of DDN01 the inner radius is fixed by assuming it to have the dust evaporation temperature of 1500 K. But dynamical effects could push this inner radius farther outwards, yielding a cooler rim and therefore a near infrared bump that is slightly shifted towards longer wavelengths. This is precisely what appears to be the case in AK Sco: instead of the usual inner-rim temperature of 1500 K, a temperature of 1250 K (at Rin = 0.4 AU) is required to fit the near-infrared bump. Since AK Sco has a projected semi-major orbital axis of a sin i = 0.143 AU, thus occupying an appreciable fraction of the central gap, the tidal interaction of the binary with the circumbinary disk could plausibly cause that gap to be wider than expected. The best model has Rout =12 AU, Mdisk =0.05M
, and constant surface density as a function of radius. We assume a gas-to-dust ratio of 100. The SED computed for an inclination of i=70◦ , consistent with the results of Sect. 5, is shown as a solid line in Fig. 4, where the stellar spectrum is approximated by a blackbody curve. Jensen & Mathieu (1997), hereafter JM97, obtained as good a fit to the SED of AK Sco as we, but within a different physical scenario. They used a geometrically thin disk with a power-law surface-density distribution and a hole to simulate disk clearing by the binary. In order

The PMS spectroscopic binary AK Sco (Alencar et al.)

113

to fit the AK Sco SED they had to include ∼10−9 M
, of optically thin material inside the hole. Based on the SED alone, it is difficult to distinguish between the the models of DDN01 and JM97. This is not totally unexpected since the physical state of disks, especially in binaries, is not yet well understood. However, even the small optically thick disk should be resolvable with the large submillimeter array ALMA now under construction in Chile, so it may become possible to settle the issue in the relatively near future.

Figure 4. Best model fit to the AK Sco SED. The de-reddened photometric data are shown as squares and the fit as a solid line.

Acknowledgments This research is based on data collected at the ESO 1.52m telescope at La Silla, partly operated by the Observat´ orio Nacional-CNPq as a result of the Brazilian-ESO agreement, and at the 0.5m and 1.54m Danish telescopes at ESO, La Silla, Chile. We thank the observers at the Geneva 70 cm photometric telescope at ESO La Silla, and G. Burki, N. Cramer and B. Nicolet for the careful reduction of these data. S.H.P.A. acknowledges support from FAPESP (grant number 00/06244-9) and CAPES (PRODOC program), and J.A. acknowledges support from the Danish Natural Science Research Council and the Carlsberg Foundation.

References Alencar, S.H.P., Melo, C.H.F., Dullemond, C.P., Andersen, J., Batalha, C., Vaz, L.P.R. & Mathieu, R.D. 2003, submitted to A&A Andersen, J., Lindgren, H., Hazen, M.L. & Mayor, M. 1989, A&A, 219, 142

114

OPEN ISSUES IN LOCAL STAR FORMATION

Artymowicz, P. & Lubow, S.H. 1994, ApJ, 421, 651 Artymowicz, P. & Lubow, S.H. 1996, ApJ, 467, L77 Basri, G., Johns-Krull, C.M., Mathieu, R.D. 1997, AJ, 114, 781 Bibo, E.A. & Th´e, P.S. 1991, A&AS, 89, 319 Chiang, E.I. & Goldreich, P. 1997, ApJ, 490, 368 Dullemond, C.P., Dominik, C. & Natta, A. 2001, ApJ, 560, 957 Etzel, P.B., 1985, “SBOP”, Program Manual, UCLA Herbig, G.H. & Rao, N.K. 1972, ApJ, 174, 401 Jensen, E.L.N. & Mathieu, R.D. 1997, AJ, 114, 301 Mathieu, R.D. 1994, ARA&A, 32, 465 Mathieu, R.D., Stassun, K., Basri, G. et al. 1997, AJ, 113, 1841 Melo, C., Pasquini, L., De Medeiros, J.R. 2001, A&A 375, 85 Natta, A., Prusti, T., Neri, R. et al. 2001, A&A, 371, 186 Santos, N., Mayor, M., Naef, D., Pepe, F., Queloz, D., Udry, S., Burnet, M., Clausen, J. V., Helt, B. E., Olsen, E. H., Pritchard, J. D. 2002, A&A, 392 215

Silvia Alencar

SURVEY OF YOUNG STELLAR OBJECTS ASSOCIATED WITH MOLECULAR CLOUDS Z. Abraham, A. Roman-Lopes, J. R. D. L´epine, T. Dominici, A. Caproni University of S˜ ao Paulo, Brazil [email protected]

Abstract

We present the preliminary results of a survey in the near-infrared of young and massive stellar objects embedded in dense molecular clouds. The survey was conducted at the LNA, Brazil, in the J, H, and nbK bands, in the direction of IRAS sources with colors characteristic of compact HII regions that also present strong CS(2-1) line emission. The coordinates of the IRAS sources were improved to 1.9 resolution with the help of the MSX catalog, allowing us to associate them univocally to the detected stars in our sample. Several IR nebulae were detected, indicating also that young stellar clusters are being formed in the core of the molecular clouds. From the total number of 127 observed regions we found that 105 presented stars in the H band inside the IRAS error ellipse, and were associated to a MSX point source. We obtained the integrated bolometric luminosities and main-sequence spectral types of the detected stars, assuming that the bolometric luminosity coincided with the integrated IR luminosity, and using the fact that for massive stars the evolution towards the main sequence proceeds at constant luminosity. The Initial Mass Function of the sample follows a Salpeter’s law with index 1.6 between spectral types B0 and O3.

Introduction The main objective of our survey is the study of young massive stars and associations. This is a difficult task because massive stars evolve very fast and they can leave the main sequence still embedded in the molecular cloud from which they were formed. The association of IRAS sources with ultra compact HII regions (Wood & Churchwell 1989) gave a starting point for the search of young massive stars and the latter 115 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 115-120. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

116

OPEN ISSUES IN LOCAL STAR FORMATION

development of near IR imaging allowed the study of several individual regions, generally associated to known molecular clouds (ex. Persi et al. 1994,1997; Tapia et al. 1996; Massi et al. 2000). Maser emission is also known to occur at the sites of recently formed stars (Garay & Lizano 1999) and it was suggested that type II methanol masers at 6.7 GHz occur in circumstellar disks (Norris et al. 1998, Minier, Booth & Conway 2000). Walsh et al. (1997) presented a survey of methanol masers in the direction of UC HII regions, selected from their IRAS colors. For the detected maser sources they calculated the IR luminosities, using the IRAS integrated flux densities and the kinematic distances derived from the maser line velocities. Finally they calculated the spectral types of the embedded stars assuming that the IR luminosity represents the bolometric stellar luminosity reprocessed by the surrounding gas and dust. The number of stars associated with methanol masers for each spectral type is presented in the histogram shown in Figure 1. 50 45 Methanol Masers

40 35

Number

30 25 20 15 10 5 0 B3

B2

B1

B0

O9

O8

O7

O6

O5

O4

O3

Spectral Type

Figure 1. Spectral type distribution of stars associated to 6.7 GHz methanol maser (Walsh et al. 1997)

We can see that the maximum in the distribution corresponds to spectral type O6 , that is, to very massive stars. An important question arises from this work: are methanol masers related mostly to very massive stars, or their distribution reflects the masses of the stars that are currently being formed in our Galaxy? To answer this question we started a survey in the near IR (J, H and nbK bands) in the direction of dense

Survey of YSO associated with Molecular Clouds (Abraham et al.)

117

molecular clouds with IRAS colors characteristic of UC HII regions. The objective is to obtain the spectral types and distribution of the detected stars both from their near IR colors and from their IRAS integrated fluxes, using the velocities of the molecular lines, independently of the existence of associated maser emission.

1.

Observations and Results

We observed a complete sample of sources which have IRAS colors of ultra compact HII regions (Wood & Churchwell 1989), with CS(2-1) antenna temperatures TA > 2 K and galactic coordinates 24◦ < l < 224◦ , extracted from the survey of Bronfmann, Nyman & May (1996). We used the Near Infrared Camera (CamIV), equipped with a Hawaii 1024×1024 pixel HgCdTe array, mounted on the 0.6 m Boller & Chivens telescope at the Laborat´orio Nacional de Astrof´ısica (LNA), in Brazil and obtaine 8 × 8 frames centered in the IRAS source, in the J, H and nbK bands. Data was reduced using the IRAS1 subroutines and positions were obtained, with a precision of about 1 , using the R frames from the Digitized Palomar Sky Survey.2 The next step was to identify the stars associated with the IRAS sources. In many cases, more than one star was found inside the IRAS error ellipse, for that reason we used the Midcourse Space Experiment - MSX source catalog3 , which has an image resolution of 19 arcsec and a global astrometric accuracy of 1.9 arcsec (Price et al. 2001). From the 127 regions observed, in 105 of them a star in the H band (where our sensitivity is greater) could be associated to the IRAS and MSX sources. Photometry for these stars was obtained using the point spread function algorithm in the DAOPHOT package. Although 105 stars detected in the H band coincided in position with IRAS and MSX sources, not all of them were detected at the J band (where the absorption is larger) or in the nbK band (where the sensitivity of our IR camera is lower). For that reason, for many stars it was not possible to obtain the amount of reddening from the (J − H) × (H − K) diagram, a quantity necessary to correct the observed magnitude for absorption. Besides, in most cases in which the three colors were available we verified, as expected, that the stars presented a color excess at 2 µm, indicating that they are embedded in a dusty envelope. In this case, the correction to the observed magnitude cannot be obtained either.

1 IRAF

is distributed by the National Optical Astronomy Observatories, which is operated by the Assotiation of Universities for Research in Astronomy, Inc. under contract to the National Science Foundation 2 http://archive.stsci.edu/cgi-bin/dss-form 3 http://www.ipac.caltech.edu/ipac/msx/msx.html

118

OPEN ISSUES IN LOCAL STAR FORMATION 30

Molecular Clouds

25

Number

20

15

10

5

0 B4

B3

B2

B1

B0

O9

O8

O7

O6

O5

O4

O3

Spectral Type

Figure 2. Spectral type distribution of stars associated to dense molecular clouds (our work)

Therefore, we used a different approach to calculate the stellar spectral types and masses. We integrated the total IR luminosity (1 - 100 µm) and assumed that this IR luminosity corresponds to the stellar bolometric luminosity, reprocessed by the dusty envelope. The distances used to transform flux density in luminosity are the kinematic distances obtained from the CS line velocities. When two distances were allowed, the nearest was used. From the bolometric luminosities we determined the main sequence spectral types and stellar masses. Notice that even if the stars were still in the pre-main sequence, the calculated mass would be correct, since massive stars (above ∼ 9 M
) evolve at constant luminosity from the Hayashi track towards the main sequence (Iben 1965). In Figure 2 we show our results, in number of detected stars as a function of the spectral type, which presents a strong maximum in the distribution for spectral type B0.

Discussion From Figure 2 we can see that the distribution of stars as a function of the spectral index is very different from that of the methanol masers, shown in Figure 1. From our detected sources, 63 are common to the sample of Walsh et al. (1997) and 44 of them detected as methanol masers. In Figure 3 we present the distribution of stars as function of

Survey of YSO associated with Molecular Clouds (Abraham et al.)

119

30 Molecular Clouds

Methanol Masers

25

Not detected

Number

20

15

10

5

0 B4

B3

B2

B1

B0

O9

O8

O7

O6

O5

O4

Spectral Type

Figure 3. Comparison between the distribution of stars as function of the spectral index in the present work and methanol masers and no maser detection from Walsh et al. (1997) (black, grey and white histograms, respectively)

the spectral type for our sample and for the methanol masers and no detections in the maser survey (histograms in black, grey and white, respectively). It is evident that there is a large difference in both distributions, suggesting that methanol maser emission is associated mainly to very massive stars. In Figure 4 we present the logN (M ) versus logM distribution, where M is the stellar mass. We see that the distribution follows a power law N (M ) ∝ M −α between spectral types B0 and O3, with index α = 1.6. This is an unexpected and not obvious result, since we found a mass distribution similar to Salpeter’s law, but not in the distribution of stars formed in a given dust cloud, but in the individual stars that are being formed at the present time in different places in the Galaxy and in very well defined state of evolution, namely, still surrounded by a dusty envelope.

Acknowledgments This work was partially supported by the Brazilian agencies CNPq, FAPESP and FINEP.

120

OPEN ISSUES IN LOCAL STAR FORMATION

1.6

B0

N α M-1.6

log N

1.2

0.8

O3 0.4

0.0 0.5

0.8

1.0

1.3

1.5

1.8

2.0

log M/M
Figure 4.

Number N of detected stars as function of mass M in our sample

References Bronfman, L. Nyman, L. A., May, J. 1996, A&ASS, 115, 81 Garay, G., Lizano, S. 1999, PASP, 111, 1049 Iben, I. Jr. 1965, ApJ, 141, 993 Massi, E., Lorenzetti, D., Giannini, T., Vitali, E. 2000, A&A, 353, 598 Minier, V., Booth, R. S., Conway, J. E. 2000, A&A, 362, 1093 Norris, et al. 1998, ApJ, 508, 275 Persi, P., Roth, M., Tapia, M., Ferrari-Toniolo, M., Marenzi, A. R. 1994, A&A, 282, 474 Persi, P., Felli, M., Lagage, P. O., Roth, M., Testi, L. 1997, A&A, 327, 299 Price, S. D., Egan, M. P., Carey, S. J., Mizuno, D. R., Kuchar, T. A. 2001, AJ, 121, 2819 Tapia, M., Persi, P., Roth, M. 1996, A&A, 316, 102 Walsh, A. J., Hyland, A. R., Robinson, B., Burton, M. G. 1997, MNRAS, 291, 261 Wood, D. O. S., Churchwell, E. 1989, ApJ, 340, 285

THE STELLAR POPULATION OF EMBEDDED GALACTIC MASSIVE STAR CLUSTERS A. Damineli IAG-USP, Brazil [email protected]

R. D. Blum Cerro Tololo Interamerican Observatory, Chile [email protected]

P. S. Conti JILA, University of Colorado, USA [email protected]

E. Figuerˆedo IAG-USP, Brazil [email protected]

C. L. Barbosa IAG-USP, Brazil [email protected]

Abstract

We present results for an on-going project designed to assess the stellar content of obscured Galactic Giant HII regions (GHIIR) in the nearinfrared. Based on deep JHK imaging and K-band spectroscopy, we derive spectroscopic parallaxes, reddening and excess emission. This enables the determination of distances and identification of massive YSOs. Our preliminary results indicate distances shorter than derived from radio techniques, implying in lower star formation rates in our Galaxy than presently believed. 121

J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 121-126. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

122

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction The study of GHIIR (> 1050 photons.s−1 in the Lyman continuum) is important to address the question of the morphological type of our Galaxy and to progress in understanding the formation of massive (M >10M
) stars. The accretion mechanism for those stars is little known, as compared to low mass stars. The main problem is that massive stars remain hidden in their natal cocoon in the earliest 10-15% of their lives, in the optical and even in the near-infrared spectral region. This is due to the fact that massive stellar clusters are embedded in molecular clouds and high mass stars have dense environment, in addition to the fact that they are seen under large interstellar reddening. The present knowledge of massive star formation in our Galaxy is based mostly on radio data. Distances are derived from radial velocities measured on recombination lines and by adopting a Galactic rotation model. The number of ionizing photons per second (NLyc) is derived from the radio flux, enabling to evaluate the number of equivalent O7V stars. The star formation rate through the Milky Way can be estimated by assuming an appropriate Initial Mass Function (IMF). The variation of the IMF as function of the Galactocentric distances is of special interest, since star formation may depend on metallicity and gas density. Since GHIIRs trace very well the spiral arms, the same data are used also to draw the Galactic spiral arms pattern. However, there are problems with such a method: a) the rotation model degenerates in the inner Galaxy, giving only upper and lower distances, and b) the motion of the gas probably deviates from circular motion due to collisions between molecular clouds. Those difficulties led us to propose an imaging survey of GHIIRs in the near-infrared (NIR). JHK images toward strong radio sources (Smith et al. 1978) are obtained in 4-m class telescopes, aiming to select candidates for spectroscopy follow-up. K-band spectra are taken in 4-m and 8-m class telescopes to look for spectral types. Identification of photospheric features are made in the scheme of Hanson, Conti and Rieke (1996), enabling spectral type classification and distance determination. Absence of photospheric features and or identification of CO band indicate that the source still enshrouded in the natal cocoons. Detailed results for several GHIIR: W31, W42, W43, G333.1-0.4 and NGC3576 may be found in Blum, Damineli and Conti (1999); Blum, Conti and Damineli (2000); Blum, R. D., Damineli, A., Conti, P. S. (2001); Figuerˆedo, E., Blum, R. D., Damineli, A., Conti, P. S.( 2002),

Massive Stars (Damineli et al.)

123

Figuerˆedo, E., Blum, R. D., Damineli, A., Conti, P. S. (2003) and Barbosa et al. (2003). In all those cases we identified clearly the ionizing clusters. In the case of the former four objects we were able to classify the spectrum of brightest cluster members and derive the spectroscopic parallaxes.

1.

Accretion process

It is interesting that for NGC3576 none of the eight brightest stellar objects did show photospheric features, indicating a very early evolutionary stage. This could be already expected when looking to the K x H − K color-magnitude diagram (CMD): it shows a number of excess emission objects above the interstellar reddening vector. We observed several of these objects in the mid-infrared (MIR) with Gemini South telescope and confirmed that they are surrounded by warm dust. The almost diffraction limited images (∼ 0.3 arcsec) resolved the previously known object NGC3576/IRS1 into four sources inside 1.5 arcsec radius (Barbosa et al. 2003). The MIR luminosities of these sources indicate they are earlier than B5. This compares quite well with spectral types derived from NIR imaging by Figuerˆedo et al. (2002). Source #48, for example is a B1V, after correcting for the excess emission in the NIR and is earlier than B3V from the MIR luminosity. This is a very interesting object, showing a double-peaked line profile in the spectral line Brγ, that can be interpreted as signature of a torus. This object displays the CO band-head in emission, indicating a dense region excited by UV radiation. Our present understanding of the situation is that those features represent the leftover of an equatorial accreting region. A few similar objects were found in other clusters, with emission line profiles indicative of a disk and/or a torus (W31/#1, G333.1-0.4/#4, M17/B275, M17/268, and M16/331). This is important, since the accretion process of high mass stars is yet completely unknown.

2.

Distances and location of the spiral arms

Although we have determined distances for only a few GHIIRs, they are systematically smaller than that derived from radio techniques by a large factor, as can be seen in table 1. In Figure 1 we plot arrows departing from radio positions given by Georgelin & Georgelin (1976) ending to the positions determined by our spectroscopic parallaxes. If these preliminary results are confirmed by measurements we are making for additional GHIIRs, the implication would be that the star formation rate (SFR) in the Milky Way is much smaller than derived by radio techniques. The location of the spiral arms also would be different from

124

OPEN ISSUES IN LOCAL STAR FORMATION

presently believed. As far as we can see, there are no systematic effects in our method. The main source of errors is the uncertainty in M v, of the order of 0.6 magnitudes. Moreover, the errors of our method will decrease on statistical basis, as more and more spectra for each cluster is accumulated.

3.

Universality of the IMF

The slope of the K−band luminosity function is almost the same for all the clusters we have studied: α = 0.35 ± 0.05 We didn’t translate the luminosity function into an IMF, since we don’t have enough data to correct for stellar multiplicity and excess emission. If we admit that such corrections are the same for all clusters this implies that the IMF slope is almost the same for all clusters. This result favors a universal IMF. This is in agreement with results obtained by other authors, when comparing young clusters in the Magellanic Clouds and the Milky Way.

Figure 1. The spiral structure of our Galaxy as presented by Georgelin & Georgelin (1976). Arrows represent the shift between positions derived from radio (Smith 1978) and our measurements.

Massive Stars (Damineli et al.) Table 1.

125

Distances to Giant HII Regions

ID W31 W42 W43 G333.1-0.4

Radio (kpc) >4.1 3.7 7.0 2.8-11.3

SP Parallax (kpc) 3.4± 0.3 2.2 ± ? 5.0 ± 0.7 4.5 ± 0.6

Acknowledgments E. Figuerˆedo, C. Barbosa and A. Damineli thank FAPESP and PRONEX/G for support.

References Barbosa, C. L., Damineli, A., Blum, R. D., Conti, P. S 2003 - submitted to AJ. Blum, R. D., Damineli, A., Conti, P. S. 1999, AJ, 117, 1392 Blum, R. D., Conti, P. S., Damineli, A. 2000, AJ, 119, 1860 Blum, R. D., Damineli, A., Conti, P. S. 2001 AJ, 121, 3149 Figuerˆedo, E., Blum, R. D., Damineli, A., Conti, P. S. 2002, AJ, 124, 2739 Figuerˆedo, et al. 2003 - in preparation Georgelin, Y.M. and Georgelin, Y.P. (1976), A&A, 49, 57 Hanson, M., Conti, P.S., Rieke, M.J., 1996 ApJS, 107, 281 Smith, L. F. et al. 1978, ApJ, 327, 128.

A. Damineli, M. Freitas, V. Jatenco-Pereira, D. Falceta-Gon¸calves

126

OPEN ISSUES IN LOCAL STAR FORMATION

G. Quast, E. Mamajek

B. K¨onig, A.C. Armond, K. Kaempgen

SPECTROSCOPIC ANALYSIS OF 131 HERBIG AE/BE CANDIDATE STARS S. L. A. Vieira W. J. B. Corradi S. H.P. Alencar L. T. S. Mendez Universidade Federal de Minas Gerais slvieira@fisica.ufmg.br

C. A. Torres, G. Quast Laborat´ orio Nacional de Astrof´ısica

M. M. Guimar˜ aes Universidade Federal de Minas Gerais

L. da Silva Observat´ orio Nacional

Abstract

We present a premliminar analisys of 131 Herbig Ae/Be candidate stars based on low and/or medium resolution spectroscopy. The objects present a great variety of Hα line profiles and were separated according to them.The presence of [O i] and [S ii] forbidden lines is used together with the Hα line profiles to discuss the circumstellar environment of the Herbig Ae/Be candidates. The great variety of line profiles in the Balmer series and the distribution of [O i] and [S ii] forbidden lines favor the hypothesis that the circumstellar environment of these stars are observed on different evolutionary stages.

127 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 127-132. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

128

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction Determining whether a star belongs to the Herbig Ae/Be (HAeBe) group or not has been a topic of great discussion, with various criteria being used in the literature for the classification. In general, to be undoubtedly considered a HAeBe star, a candidate should present the following characteristics: to be of spectral type A or earlier, with emission lines; to be located in an obscured region;

to have a fairly bright nebulosity in its immediate vicinity; to present an anomalous extinction law;

to show infrared excess; to be photometrically variable; to display line profiles of MgII (2800 ˚ A) in emission. The first three criteria were proposed by Herbig (1960) in order to define pre-main-sequence (PMS) stars of intermediate mass. The last four are an extension proposed by Th´e et al. (1994) to encompass the large set of new candidates. However, very few stars satisfy all of them. One possible explanation is that the HAeBe group spans a large range of evolutionary stages (Malfait et al. 1998). The aim of this work is to present a statistical analysis of a sample of 131 HAeBe stars, for which we have low and medium resolution spectra, covering the spectral range of 4300 ˚ A up to 6800 ˚ A.

1.

Observations

Low and medium resolution spectra of all the objects were used to carry out a statistical analysis of the features supposedly associated with the HAeBe class. Examples of such spectra are shown in Figure 1, covering the total observed wavelength range at each resolution (low: 4790 ˚ A– 6880 ˚ A and medium: 6290 ˚ A– 6745 ˚ A, except for three stars where the spectra start at 6500 ˚ A).

2.

Hα Line Profiles

Reipurth et al. (1996) proposed a classification system for Hα line profiles in PMS objects. In this system they are separated in four groups, exemplified in Figure 2:

Analysis of Herbig Ae/Be Candidate Stars (Vieira et al.)

129

Figure 1. Medium resolution spectrum of PDS 225 (left panel) and low resolution spectrum of PDS 353 (right panel). Both spectra show the Hα line in emission.

Type I profiles are symmetric without or with only very shallow absorption features. Type II profiles are double peaked, with the secondary peak having more than half the strength of the primary. Type III profiles are double peaked, with the secondary peak having less than half the strength of the primary. Type IV profiles have P Cyg line characteristics. The objects in our sample present Hα line profiles according to this system as follows: Type I: 29 stars (24%);

Type II: 53 stars (43%); Type III: 23 stars (18.5%); Type IV: 18 stars (14.5%); This distribution of line profiles agrees with what was found by Reipurth et al. (1996) with a totally different and independent sample, type II being the most common and type IV the least common line profile among

130

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. Examples of observed line profiles with the classification proposed by Reipurth et al. (1996). (a) Type I; (b) Type II; (c) Type III; and (d) Type IV.

HAeBe stars. This fact suggests that most of the time a HAeBe star will present type II Hα line profile. Due to the high variability of the Hα line profiles in a period of months, weeks or even days, it is important to notice that the line profile is not correlated to the stellar mass or the stellar evolutionary status, the proposed classification being a snapshot of the current star properties. The objects that changed their Hα line profile from one type to another were classified using the most recent available spectrum. Since they are only 8 stars, our choice does not affect significantly our statistics.

131

Analysis of Herbig Ae/Be Candidate Stars (Vieira et al.)

Another group of 7 objects has a very poor signal to noise ratio, and it is impossible to assign any of the discussed types to them.

3.

[O i](6300 ˚ A and 6364 ˚ A) and [S ii](6716/6731 ˚ A) forbidden lines

Using the low, medium and high resolution spectra obtained for all the stars, [O i] and/or [S ii] lines were detected in 59 out of 128 objects (3 stars have spectra beginning at 6500 ˚ A and we consequently cannot say anything about the presence of [O i]). This amount corresponds to 46% of our sample and it can be regarded as a lower limit, since we do not have high resolution spectra of all the stars and could therefore be missing some low intensity forbidden lines in the low and medium resolution observations. In some cases, however, forbidden lines detected in low and medium resolution spectra were not present in the high resolution observations (taken at a different epoch), suggesting that those lines also present variable intensities. The distribution of the detected forbidden lines among different spectral types and Hα line profile types can be seen in Table 1. We have an almost equal number of A (40%) and B (45%) stars in our sample, therefore the results in Table 1 point to a significantly higher occurrence of forbidden lines among B stars. The forbidden line distribution among different Hα profile types is nearly identical to the Hα profile distribution in our sampleindicating that forbidden lines are evenly distributed among each Hα line profile type. Table 1. Distribution of the forbidden lines among different spectral types (left) and Hα line profile types (right).

B 54%

4.

Sp. Type A F 30% 8%

O 8%

I 29%

Hα profile II III 41% 18%

IV 12%

Conclusions

Although HAeBe objects present a great variety of Hα line profiles, our study shows that type II line profile seems to be the most common. This is the same result of Reipurth et al. (1996), with a different sample, suggesting that most of the time a HAeBe star will present type II Hα line profile. The distribution of forbidden lines among different Hα line profiles is the same as the distribution of objects among Hα profile types. This fact

132

OPEN ISSUES IN LOCAL STAR FORMATION

shows that there is no correlation between the presence of forbidden lines and the Hα line profile type, contrary to what was found by Corcoran & Ray (1997); Corcoran & Ray (1998) and B¨ ohm & Catala (1994). The forbidden lines also tend to occur more frequently among B than A stars supporting the results of Corcoran & Ray (1997,1998)

Acknowledgments S. L. A. Vieira, W.J.B. Corradi and L.T.S. Mendes acknowledge CNPq (grant number 471537/2001-02) and FAPEMIG for partial financial support for doing this work. S.H.P.A. acknowledges support from FAPESP (grant number 00/06244-9) and CAPES (PRODOC program). M.M.G. acknowledges support from CAPES. This research has made use of of the SIMBAD database, operated at CDS, Strasbourg, France.

References B¨ ohm T., Catala C. 1994, A&A, 290, 167 Corcoran M., Ray T.P. 1997, A&A, 321, 189 Corcoran M., Ray T.P. 1998, A&A, 331, 147 Herbig G.H. 1960, ApJS, 4, 337 Malfait K., Bogaert E., Waelkens C. 1998, A&A, 331, 211 Reipurth B., Pedrosa A., Lago M.T.V.T. 1996, A&AS, 120, 229 Th´e P.S., de Winter D., P´erez M.R. 1994, A&AS, 104, 315

S´ergio Vieira

CLASSIFICATION OF THE PICO DOS DIAS SURVEY HERBIG AE/BE STARS Mar´ılia J. Sartori Laborat´ orio Nacional de Astrof´ısica/MCT, Brazil [email protected]

Jane Gregorio-Hetem Universidade de S˜ ao Paulo, IAG/USP, Brazil [email protected]

Annibal Hetem Jr. ICET/Universidade Paulista, Brazil [email protected]

Abstract

In this work we analyzed the circumstellar matter distribution of a large sample of Herbig Ae/Be (HAeBe) stars identified by the Pico dos Dias Survey. We adopted a simple model to fit the spectral energy distribution (SED) of 99 PDS stars (80 candidates and 19 well-known HAeBe stars). From this model, we estimated Sc , the contribution of the circumstellar components (dust disk and/or envelope) to the total emitted flux. We classified the sample stars in 3 groups based on the shape of their SEDs. We analyzed this classification and Sc in relation to other properties of the these stars.

Introduction The Pico dos Dias Survey (PDS), a search for T Tauri (TT) stars based on IRAS colors (Gregorio-Hetem et al. 1992; Torres et al. 1995; Torres 1999), has revealed several new TT stars, as well as other very interesting objects: 108 stars were classified as Herbig Ae/Be (HAeBe) candidate stars. All these stars have shown strong Hα emission in at least one spectrum, they have spectral type earlier than F5 and luminosity classes III–V. The criteria used by the PDS group to classify these stars are described in Vieira et al. (2003; see also Vieira et al., these 133 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 133-140. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

134

OPEN ISSUES IN LOCAL STAR FORMATION

proceedings). In this sample there are only 20 previously known HAeBe stars (Th´e et al. 1994; Malfait et al. 1998). All efforts focused on confirming the nature of the HAeBe candidates have shown that this class of stars does not have exclusive observational characteristics. Many of the spectral features, for example, are also present in more evolved objects (e.g. Fernandes et al., these proceedings). The use of IRAS colors to identify young candidates has been also used to select evolved stars candidates, since stars in different evolutionary phases show far-IR excess due to circumstellar dust. Therefore, this useful tool must be applied with caution to avoid incorrect classifications. The analysis of other characteristics is essential to confirm the nature of the candidates. Thus, to confirm the young nature of the PDS HAeBe candidates, Vieira et al. (2003) analyzed their positions in the HR diagram. Other studies (e.g. V band polarization, see Rodrigues et al., these proceedings) have been performed to complement the original PDS data and to improve the classification of these HAeBe candidates. Our main goal in this work is to contribute with the classification of this sample of HAeBe candidates. We investigated the circumstellar matter distribution of these stars through an analysis of their spectral energy distributions (SEDs).

1.

Spectral Energy Distribution

In the literature of the past 15 years there are several models developed to explain the SEDs of the HAeBe stars (see the reviews by P´erez & Grady 1997, Waters & Waelkens 1998, and Natta et al. 2000). For example, although spherical envelopes models and disk models with high accretion rates have successfully fitted the SEDs of some HAeBe stars, several observations have pointed to the presence of disks, however with low accretion rates. Therefore, both disk and envelope may contribute to the observed SEDs. Recently, passive irradiated circumstellar disk models have been used to reproduce the SEDs of HAe stars (e.g. Dullemond et al. 2001; Dominik et al. 2003). As the HAeBe stars show many different SED shapes, it seems that there is not a single model that could explain the SEDs of all HAeBe stars. Gregorio-Hetem & Hetem Jr. (2002) applied to a selected sample of TT stars a simple model considering a system composed of a central star surrounded by a geometrically thin and optically thick, passive disk and a spherical dust envelope (for more details, see Hetem Jr. and Gregorio-Hetem, these proceedings). As the IRAS colors of the PDS HAeBe stars are similar to those of the TT stars, we decided to adopt this model in order to analyze the circumstellar matter distribution of

Classification of the PDS HAeBe stars (Sartori et al.)

135

our sample. In this case, we are interested on the use of the infrared SED to derive the contribution of the circumstellar components (dust disk and/or envelope) to the total emitted flux. By considering that the flat disk simple model could not be adequate to reproduce circumstellar structure of the HAeBes, we did not intend to derive constraints on the geometry of the circumstellar matter. From the whole sample of stars classified as HAeBe by the PDS, we excluded from our study 3 pairs of binaries; 2 stars that are actually proto-planetary nebulae; and 1 object with no available optical photometry. For the construction of the SEDs of the 99 PDS stars we used orio do photometric data: U BV (RI)c from PDS (obtained at Observat´ Pico dos Dias/LNA, Brazil), near- and mid-IR from the literature (for ∼ 68% of the sample) and IRAS fluxes. Besides the photometric fluxes, another important input parameter is the effective temperature of each star. We took the Teff from the spectral type–Teff relation of de Jager & Nieuwenhuijzen (1987). We adopted the spectral type classification made by Torres (1999) and when it was not available, from the literature. In the adopted model, the disk is basically responsible for the emission in the near- and mid-IR, and the envelope is responsible for the emission in the far-IR, as we can see in the example shown in Fig. 1. From all the outputs of the SED fits, we are especially interested on the contribution of the circumstellar components (dust disk and/or envelope) to the total emitted flux (Sc ) and also on an evaluation of the quality of the fits – the goodness-of-fit (gof). For approximately 30 % of the sample we obtained gof > 0.4, which is far from a good fit, confirming that this model is not appropriate for all the stars of this sample, but for a significant part of it the model reproduces rather well the observed SEDs. PDS 076 (HD 142666)

-2

λFλ [W.m ]

1E-11

star disk envelope

1E-12

1E-13

1E-14 1

10

λ [µm]

100

Figure 1 Example of the SED of HD 142666 (PDS 076), a known HAeBe star. In this diagram we indicate the individual contributions of the star (dashed line), of the disk (dotted line), and of the envelope (dash-dotted line) to the total emitted flux.

136

2.

OPEN ISSUES IN LOCAL STAR FORMATION

Classification

Hillenbrand et al. (1992) proposed a classification of HAeBe stars in 3 groups based on the slope of the IR continuum. We made here a similar classification because although all stars of our sample have similar IRAS colors, their SEDs show different shapes. Our classification is based on the spectral index measured in the visible to mid-IR range – from the V band (0.55 µm) to the IRAS 12 µm band, given by β 1 = 0.75 log(F12 /FV ) − 1. Group 1: β 1 > 0 ⇒ the far-IR emission is greater than the optical emission ⇒ Sc approximately > 70 % (see Fig. 5) ⇒ 44.4 % of the sample (only 3 of them are known HAeBe stars). In this group there are two sub-groups: the stars whose SEDs increase in a “continuos” way (Fig. 2a) and the stars with decreasing spectrum in the near-IR range and increasing spectrum from the mid- to far-IR range (Fig. 2b).

Group 2: −1 ≤ β 1 ≤ 0 ⇒ the far-IR emission is smaller than the optical emission ⇒ Sc > 10 % and approximately < 70 % (see Fig. 5) ⇒ 49.5 % of the sample (only 15 of them are known HAeBe stars). In this group there are also two sub-groups: the stars whose SEDs decrease in a “continuos” way (Fig. 3a) and the stars with decreasing spectrum in the near-IR range and increasing spectrum from the mid- to far-IR range (Fig. 3b). Group 3: β 1 < −1 ⇒ the far-IR emission is much smaller than the optical emission (Fig. 4) ⇒ Sc ≤ 10 % (see Fig. 5) ⇒ 6.1 % of the sample (only 1 of them is a known HAeBe star). We denominated the groups following the idea of an evolutionary scenario for the HAeBe stars, proposed by Malfait et al. (1998), similar to what is seen in TT stars. In this scenario, our group 1 corresponds to embedded objects, like R CrA, and group 3 represents the stars with more evolved dust disks, like Vega. The objects of group 2 are in between these both extreme categories, showing SEDs similar to several known HAeBe stars. The shapes of the SEDs of our group 1 are similar to the SEDs of the group II of Hillenbrand et al. (1992), while the SEDs of our group 2 are similar to their group I.

3.

Discussion

We analyzed our classification and the contribution of the circumstellar components to the total emitted flux, Sc , given by the model, by

137

Classification of the PDS HAeBe stars (Sartori et al.) PDS 027 (DW CMa) 1E-11

1E-12

-2

λFλ [W.m ]

-2

λFλ [W.m ]

1E-11

1E-13

PDS 322N (IRAS 10501-5556)

1E-12

1E-13

1E-14

1E-14 1

10

1

100

10

100

λ [µm]

λ [µm]

Figure 2. Examples of HAeBe candidates classified as group 1: (a) PDS 027; (b) PDS 322N. PDS 126 (IRAS 06111-0624)

PDS 031S (HD 72106)

1E-11

1E-12

-2

λFλ [W.m ]

-2

λFλ [W.m ]

1E-11

1E-13

1E-12

1E-13

1E-14

1E-14 1

10

100

1

10

100

λ [µm]

λ [µm]

Figure 3. Examples of HAeBe candidates classified as group 2: (a) PDS 126N; (b) PDS 031S. The IRAS 100 µm flux is an upper limit value.

PDS 545 (HD 174571)

-2

λFλ [W.m ]

1E-11

1E-12

Figure 4 PDS 545: an example of HAeBe candidate classified as group 3. The IRAS 100 µm flux is an upper limit value.

1E-13

1E-14 1

10

100

λ [µm]

comparing them with the spectral index measured in the near- to mid-IR range, αIR , and with the spectral types. We estimated αIR for each star using the synthetic SED curve. This spectral index was first proposed by Lada (1987) to classify TT stars in 3 evolutionary classes. Their divisions, adopting the revised values by Andr´e & Montmerle (1994), are indicated by horizontal dashed lines in Fig. 5. The distribution of αIR as a function of Sc , for the 3 groups (Fig. 5), shows a good correlation.

138

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 5 Distribution of the spectral index, αIR , as a function of Sc for the 3 groups. The divisions of the TT classes are indicated by dashed horizontal lines and the approximate limits of Sc for the 3 groups are indicated by vertical dotted lines.

1.0

Class I

group 1 group 2 group 3

0.5

0.0

αIR

-0.5

-1.0

Class II -1.5

Class III

-2.0 0

10

20

30

40

50

60

70

80

90

100

Sc [%]

The distribution of Sc as a function of the Teff of the stars, for the 3 groups (Fig. 6), shows that, although there is no particular concentration of any group at any spectral type, the distribution of Sc is not the same for the Herbig Ae (spectral types later than B6) and for the Herbig Be (spectral types earlier than B5) stars. Most HBe stars show Sc ≥ 50%, while the HAe stars have Sc distributed along all values. This result is in agreement with other differences observed between the higher mass and the lower mass stars of this class of young stars, for example, the faster evolution toward the main sequence of the HBe stars compared to the HAe stars.

100 F5 A6 A0

B7

B5

B3

B1

B0

90 80 70

Sc [%]

60 50 40 30

group 1 group 2 group 3

20 10 0 10000

15000

20000

T eff [K]

25000

30000

35000

Figure 6 Distribution of Sc as a function of the Teff of the stars, for the 3 groups. Some spectral types are indicated at the top of the diagram. The vertical dashed line indicates the adopted limit between the HAe and the HBe stars and the horizontal dotted lines indicate the approximate limits of Sc for the 3 groups.

Classification of the PDS HAeBe stars (Sartori et al.)

139

Acknowledgments We thank PDS group for useful discussions and permission to access their list prior to publication. MJS acknowledges the brazilian agencies FAPESP and CNPq for the postdoc fellowships (No. 00/06954-6 and No. 300758/01-4), and FAPEMIG for the travel grant to attend the meeting (CEX-5/03). This work has been partially supported by FAPESP grant (No. 2001/09018-2). This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France.

References Andr´e, P., & Montmerle, T. 1994, ApJ, 420, 837 de Jager, C., & Nieuwenhuijzen, H. 1987, A&A, 177, 217 Dominik, C., Dullemond, C. P., Waters, L. B. F. M., & Walch, S. 2003, A&A, 398, 607 Dullemond, C. P., Dominik, C., & Natta, A. 2001, ApJ, 560, 957 Gregorio-Hetem, J., L´epine, J. R. D., Quast, G. R., Torres, C. A. O., & de la Reza, R. 1992, AJ, 103, 549 Gregorio-Hetem, J., & Hetem Jr., A. 2002, MNRAS, 336, 197 Hillenbrand, L. A., Strom, S. E., Vrba, F. J., & Keene, J. 1992, ApJ, 397, 613 Lada, C. J. 1987, in IAU Symp. 115, Star Forming Regions, ed. M. Peimbert & J. Jugaku (Dordrecht: D. Reidel Publishing Co.), p. 1 Malfait, K., Bogaert, E., & Waelkens, C. 1998, A&A, 331, 211 Natta, A., Grinin, V., & Mannings, V. 2000, in Protostars & Planets IV (Tucson: University of Arizona Press), p. 559 P´erez, M. R., & Grady, C. A. 1997, Space Science Reviews, 82, 407 Th´e, P. S., de Winter, D., & P´erez, M. R. 1994, A&AS, 104, 315 Torres, C. A. O., Quast, G. R., de la Reza, R., Gregorio-Hetem, J., & L´epine, J. R. D. 1995, AJ, 109, 2146 Torres, C. A. O. 1999, Publ. Especial 10, Observat´orio Nacional, Brazil Vieira, S. L. A., Corradi, W. J. B., Alencar, S. H. P., Mendes, L. T. S., Torres, C. A. O., Quast, G., Guimar˜aes, M. M., & da Silva, L. 2003, accepted by AJ Waters, L. B. F. M., & Waelkens, C. 1998, Ann. Rev. Astron. Astrophys., 36, 233

140

OPEN ISSUES IN LOCAL STAR FORMATION

M. Sartori, B. V. Castilho

B. Reipurth, J. L´epine, R. Chini

CHEMICAL COMPOSITION STUDY OF AN ACCRETION EPISODE IN THE HERBIG CANDIDATE STAR PDS080 M. M. Guimar˜ aes, S. L.A. Vieira, S. H.P. Alencar, W. J.B. Corradi Universidade Federal de Minas Gerais, Brazil mmg@fisica.ufmg.br

Abstract

We present an investigation of an accretion episode in the Herbig candidate PDS 080, based in spectra obtained with the 1.52m ESO telescope and FEROS echelle spectrograph at La Silla (Chile) during three nights in May 2002. Redshifted absorption patterns, due to infall of circumstellar matter, were observed in the Balmer lines Hα, Hβ, Hγ, Hδ and other metallic lines such as Na i D and Ca ii. In a preliminary chemical study of the infalling structure we analysed these line profiles to determine whether the redshifted components are due to a gaseous structure or the evaporation of a solid body.

Introduction Recent studies using coronagraphic imaging techniques and the Hubble Space Telescope (Grady et al. 2000, 2001) have shown circumstellar disks around Herbig Ae type stars. Since these stars are pre-main sequence objects, the study of the interaction between the star and the circumstellar medium may provide some understanding of this stage of the stellar evolution. Such interaction may occur as a gain or loss of matter resulting in a change of the star’s angular momentum. The physical process of accretion and/or ejection of matter can be studied using absorption patterns observed in line profiles of hydrogen lines like, Hα, Hβ, Hγ, Hδ and other metallic lines such as Na i D and Ca ii. An accretion episode may be identified by a redshifted absorption component (RAC) and a chemical analysis of these RACs characterize 141 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 141-146. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

142

OPEN ISSUES IN LOCAL STAR FORMATION

the episode as produced by gaseous structures from the inner parts of the disk or by evaporation of a comet-like body in a star-grazing orbit. Using an approach similar to the one applied by Natta et al. (2000) for UX Ori, we have investigated an accretion episode in the Herbig candidate PDS080. In a preliminary analysis of the infalling structure’s chemical composition we aim to determine the nature of the RACs.

1.

PDS080

The Herbig candidate star PDS080 is an A8 iii/ iv star also known as V718 Sco. Located at 130 ± 20 pc (Perryman et al. 1997) it is probably associated to the ρ Oph cloud (110 - 160 pc, e.g. Franco 1990). Herbig Ae/Be stars show large brightness variations that are irregular and similar to the minima of eclipsing binaries (Algol-type minima) in the case of A type stars. Based on such behavior some authors (e.g. Koch et al. 1963) classified PDS080 as an eclipsing binary of Algol type, but according to Friedemann et al. (1996) and van der Veen et al. (1994) its binarity still remains uncertain. Such photometric variability allied to the presence of RACs was observed in UX Ori by Grinin et al. (2001) and also in HD100546 by Vieira et al. (1999), who explained it in terms of obscuration of the star by opaque clouds intersecting the line of sight during the infall of matter. In the case of PDS080 we do not have simultaneous photometric and spectroscopic data that could be used to improve the understanding of the accretion episode in these sense. The observations of PDS080 were made with the 1.52m ESO telescope together with the FEROS echelle spectrograph at La Silla, Chile, from May 7 to 9, 2002 and they were automatically reduced in a standard way using routines of the MIDAS package.

2.

The identification of the RACs

To identify a RAC it is necessary to separate photospheric contributions from circumstellar ones. The synthetic photospheric spectra for PDS080 were made with the SME (Spectroscopy Made Easy) code by Piskunov et al. (1995), using the table of atomic lines from the VALD (Viena Astronomic Lines Database) and stellar parameters such as Teff , v sin(i) and log(g), were determined based on the absorption lines that presented only the photospheric contributions (“pure” photospheric lines). Since the photospheric line profile should be constant, the pure photospheric lines were obtained subtracting the spectrum of each night from those of the other nights and stable lines were pre-selected. Some lines

Chemical composition of PDS080 (Guimar˜ aes et al.)

143

that are purely photospheric in T Tauri stars can present a circumstellar contribution in Herbig Ae/Be stars, as is the case of Mg ii 4481˚ A and the O i triplet (7772, 7774 and 7775 ˚ A) found in UX Ori by Grinin et al. (2001). Such lines were discarded from the present analysis. Two lines of Fe i (4226 and 4957 ˚ A) were used to fit the temperature of PDS080 and they can be seen in the panels (a) and (b) of Figure 1. Both lines are more sensitive to temperature than surface gravity. As an example panels (c) and (d), in Figure 1, show the dependence of the Fe i 4957˚ A line on surface gravity and temperature, respectively. They were constructed keeping one parameter constant and varying the other in the range of interest. For panel (c) we kept the temperature in 7300K and varied log(g) from 3.0 to 4.4; and for panel (d) log(g) was kept constant at 3.8 and the temperature varied from 6500K to 7600K. The wings of Hδ and Hγ and the wings of the Ca ii IR triplet lines (8498 and 8662 ˚ A) were used to fit the surface gravity. Panels (e) and (f), in Figure 1, show the dependence on gravity of Hγ and Ca ii 8498˚ A respectively. They were also constructed keeping the temperature at 7300K and varying log(g) from 3.0 to 4.4. Unfortunately the Ca ii IR 8452˚ A line could not be used because it is located where the orders of the echelle spectrum are merged.

Figure 1. Panels (a) and (b) show the SME fit (dashed line) to the observed Fe i lines (4226 and 4957˚ A) (solid lines). Panels (c) and (d) show respectively the dependence on surface gravity and temperature of the Fe i 4957˚ A line and panels (e) and (f) show how the Hγ and the Ca ii 8498˚ A line vary with log(g).

The derived stellar parameters using the above mentioned procedure are: Teff = 7300K, log(g) = 3.8 and v sin(i) = 115 Km/s.

144

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. Photospheric synthetic spectrum (dashed line) and observed spectrum (solid line) of Hγ. The numbers above each spectrum represent the date of observation.

Figure 3. Velocity of the circumstellar components. The straight solid line represents the zero velocity in the star rest frame, using the radial velocity calculated by the SME code.

The synthetic photospheric spectrum, built with these parameters, was subtracted from the observed one to obtain the circumstellar contribution. As an example, the synthetic photospheric spectrum and the observed spectrum of Hγ are shown in Figure 2. The circumstellar components for the same line are shown in Figure 3 where RACs can be seen.

3.

Preliminary chemical analysis of the accretion episode

To quantify the absorption produced by the circumstellar environment, the depth of the circumstellar absorption component (τ ) is defined following Natta et al. (2000) as: τ =1−

Fobs Fsyn

(1)

where Fobs and Fsyn represent respectively the observed and the synthetic photospheric intensity at peak wavelength. If τ = 1 the circumstellar line is saturated and if τ = 0 there is no circumstellar contribution. If the infalling material is produced by evaporation of solid bodies in star-grazing orbits then the RACs must appear in metal lines and

Chemical composition of PDS080 (Guimar˜ aes et al.)

145

simultaneously be absent in Balmer lines since such material should be strongly depleted in hydrogen. The presence of RACs in Balmer lines implies that the infalling material is rich in hydrogen but does not exclude the possibility of RACs in metal lines such as Ca ii and Na iD, since these elements are present in circumstellar structures of solar like chemical composition. As can be seen in Figure 4, RACs of different lines were observed with similar velocities and presented a similar behavior, which means that the accretion event is common to all lines. The fact that the infalling material decelerates with time, contrary to the predictions of the models by Beust et al. (1998), can be explained in terms of the variation of the projected velocity of the material falling toward the star through magnetic field lines.

Figure 4. Velocity of the circumstellar component as a function of the date of observation.

Figure 5. Depth of the circumstellar absorption component as a function of the date of observation.

Using Hβ, Hγ, Hδ and Ca ii IR triplet lines, we have found RACs with values of τ bigger in Balmer lines than in Ca ii lines, as can be seen in Figure 5. The depth of the circumstellar absorption component of all lines decrease from the first night to the second which means that the concentration of material in the line of sight has also decreased. Based on the presence of RACs in Balmer lines we suggest that the gas falling onto PDS080 was not produced by evaporation of a cometlike body, being a gaseous structure from the disk/envelope the most probable cause of the RACs in the spectra of this star. Such gaseous

146

OPEN ISSUES IN LOCAL STAR FORMATION

structures can be possibly part of the inner disk driven toward the star by magnetic fields in an analogous fashion to the T Tauri stars.

Acknowledgments The authors acknowledge CAPES (PROF and PRODOC programs), CNPq (471537/2001-2) and FAPEMIG for the financial support.

References Beust, H., Lagrange, A.-M., Crawford, I.A., et al. , 1998, A&A, 338, 1015 Franco, G.A.P., 1990, A&A, 227, 499 Friedemann, C., G¨ urtler, J., L¨ owe, M., 1996, A&ASS, 117, 205 Grady, C.A., Devine, D., Woodgate, B., Kimble, R., et al. , 2000, ApJ, 544, 895 Grady, C.A., Polomski, E.F., Henning, Th., et al. , 2001, AJ, 122, 3396 Grinin, V.P., Kozlova, O.V., Natta, A., et al. , 2001, A&A, 379, 482 Koch, R.H., Sobieski, S., Wood, F.B., 1963, Pub. Univ. of Pennsylvania, Astr. Ser. 9 Natta, A., Grinin, V.P., Tambovtseva, L.V., 2000, ApJ, 542, 421 Perryman, M.A.C., Lindegren, L., Kovalevsky, J., Hog, E., Bastian, U., et al. , 1997, A&A, 323, 49 Piskunov, N.E., Kupka, F., Ryabehikova, T.A., Weiss, W.W., Jeffery, C.S., 1995, A&AS, 112, 525 van der Veen, W.E.C.J., Waters, L.B.F.M., Trams, N.R., Matthews, H.E., 1994, A&A, 285, 551 Vieira, S.L.A., Pogodin, M.A., Franco, G.A.P., 1999, A&A, 345, 559

Marcelo Guimar˜ aes

MAGNETICALLY CHANNELED ACCRETION IN T TAURI STARS A dynamical process J. Bouvier Observatoire de Grenoble, France

S.H.P. Alencar Departamento de F´ısica, ICEx-UFMG, Brazil

C. Dougados Observatoire de Grenoble, France

Abstract

We review observational evidence and open issues related to the process of magnetospheric accretion in T Tauri stars. Emphasis is put on recent numerical simulations and observational results which suggest that the interaction between the stellar magnetosphere and the inner accretion disk is a highly time dependent process on timescales ranging from hours to months.

Introduction T Tauri stars are low-mass stars with an age of a few million years at most, still contracting down their Hayashi tracks towards the main sequence. Many of them, the so-called classical T Tauri stars, show signs of accretion from a circumstellar disk (see, e.g., M´enard & Bertout 1999 for a review). Understanding the accretion process in young solar type stars is one of the major challenges in the study of T Tauri stars. Indeed, accretion has a significant and long lasting impact on the evolution of low mass stars by providing both mass and angular momentum while the evolution and ultimate fate of circumstellar accretion disks have become an increasingly important issue since the discovery of extrasolar planets and planetary systems with unexpected properties. Deriving the properties of young stellar systems, of their associated disks and 147 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 147-158. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

148

OPEN ISSUES IN LOCAL STAR FORMATION

outflows is therefore an important step towards the establishment of plausible scenarios for star and planet formation. Strong surface magnetic fields have long been suspected to exist in TTS based on their powerful X-ray and centrimetric radio emissions (Montmerle et al. 1983, Andr´e 1987). Surface fields of order of 1-3 kG have recently been derived from Zeeman broadening measurements of CTTS photospheric lines (Johns Krull et al. 1999, 2001; Guenther et al. 1999) and from the detection of electron cyclotron maser emission (Smith et al. 2003). These strong stellar magnetic fields are believed to significantly alter the accretion flow in the circumstellar disk close to the central star. Based on models originally developped for magnetized compact objects in cataclysmic binaries (Ghosh & Lamb 1979) and assuming that T Tauri magnetospheres are predominantly dipolar on the large scale, Camenzind (1990) and K¨onigl (1991) showed that the inner accretion disk is expected to be truncated by the magnetosphere at a distance of a few stellar radii above the stellar surface for typical mass accretion rates of 10−9 to 10−7 M
yr−1 in the disk (Basri & Bertout 1989; Hartigan et al. 1995; Gullbring et al. 1998). Disk material is then channeled from the disk inner edge onto the star along the magnetic field lines, thus giving rise to magnetospheric accretion columns. As the free falling material in the funnel flow eventually hits the stellar surface, accretion shocks develop near the magnetic poles. The basic concept of magnetospheric accretion in T Tauri star is illustrated in Figure 1. The successes and limits of current magnetospheric accretion models (MAMs) in accounting for the observed properties of classical T Tauri systems (CTTS) are reviewed in Section 1, while in Section 2 we discuss recent results which suggest that the interaction between the star’s magnetosphere and the inner disk is a highly dynamical and time dependent process.

1.

Magnetospheric accretion : the static view

The magnetospheric accretion scenario makes clear predictions regarding the existence of an inner (magnetospheric) cavity with a radial extent of a few stellar radii, magnetic accretion columns filled with free falling plasma, and localized accretion shocks at the surface of CTTS. Observational support for these various predictions is reviewed in this section.

1.1

Magnetospheric cavity and IR excess

Disk truncation radii of typically 3-8R , as predicted by MAMs, are required to reproduce the near-IR spectral energy distribution of some

Magnetically channeled accretion in T Tauri stars (Bouvier et al.)

149

Figure 1. A sketch illustrating the basic concept of magnetospheric accretion in T Tauri stars (from Camenzind 1990).

CTTS (Bertout et al. 1988) and to account for the properties of CTTS infrared excess on a statistical basis (Meyer et al. 1997). However, recent near-infrared monitoring studies of accreting T Tauri stars have revealed large variability of some systems occurring on short timescales at these wavelengths (Eiroa et al. 2002) which are not easily accounted for by static MAMs and probably points to time variable accretion in the inner disk. Moreover, the near-IR veiling measured in classical T Tauri systems is often larger than predicted by standard disk models (Folha & Emerson 1999, Johns-Krull & Valenti 2001) which suggests that the inner disk structure is significantly affected by its interaction with the stellar magnetosphere and departs from flat disk models. Observational evidence for an inflated inner disk edge has been reported by Bouvier et al. (1999) and is thought to result from the interaction between the disk and an inclined stellar magnetosphere (Lai 1999, Terquem & Papaloizou 2000).

1.2

Accretion columns and line profiles

The common occurrence of inverse P Cygni profiles in the higher Balmer lines of CTTS spectra, with redshifted absorption components

150

OPEN ISSUES IN LOCAL STAR FORMATION

reaching velocities of several hundred km s−1 , points to gas being accreted onto the star at free fall velocity from a distance of several stellar radii (Edwards et al. 1994). The relative success of MAMs in reproducing the shape and strength of optical and near IR emission line profiles of some CTTS provides additional support to the formation of at least part of the line emission in accretion columns (Hartmann et al. 1994; Muzerolle et al. 2001; for a more critical view see Ardila et al. 2002b, Alencar & Basri 2000, Folha & Emerson 2001). Statistical correlations between line fluxes and mass accretion rates predicted by these models have been reported in samples of TTS for a variety of emission lines in the UV, optical, and near IR range (e.g. Beristain et al. 2001, Johns-Krull et al. 2000, Alencar & Basri 2000, Folha & Emerson 2001, Muzerolle et al. 2001). Correlations have also been reported between mass loss rate and mass accretion rate (e.g., Ardila et al. 2002) as expected from MAMs which predict that part of the accretion flow is diverted into the wind by the magnetic field (e.g. Shu et al. 1994, Ferreira 1997). In a few CTTS, synoptic studies have revealed periodic variations of the line profiles intensity and shape over a rotation timescale. The rotational modulation of line flux and radial velocity has been interpreted as resulting from the changing visibility and projected geometry of an inclined magnetosphere interacting with the disk (Johns & Basri 1995, Petrov et al. 1996, 2001, Oliveira et al. 2000, Batalha et al. 2002). Finally, two Doppler imaging studies of CTTS (Unruh et al. 1998, JohnsKrull & Hatzes 1997) have reported the existence of localized spots at the surface of CTTS whose maximum visibility was found to coincide with the developement of high velocity redshifted absorptions in emission line profiles, as expected from MAMs when the line of sight to the accretion shock intersects the accretion column. All these observational results thus tend to support the existence of (tilted) magnetospheric accretion columns in CTTS. To be fair, however, one must stress that the modulation of inflow and outflow signatures in line profiles has been reported so far for only a few (well studied) systems. Moreover, even in these cases, multiple periods are sometimes observed and their relationship to the stellar rotation period is not always clear (e.g. Alencar & Batalha 2002, Oliveira et al. 2000). In addition, some of the basic predictions of MAMs, e.g. correlated time variations of line flux (from the accretion columns) and continuum excess flux (from the accretion shocks) are not always observed (Ardila & Basri 2000, Batalha et al. 2002). And some of the statistical correlations expected from MAMs for CTTS samples are not always fulfilled (Johns-Krull & Gafford 2002).

Magnetically channeled accretion in T Tauri stars (Bouvier et al.)

151

Hence, while several key properties of CTTS emission line profiles are consistent with MAMs predictions and a few systems indeed appear to behave as qualitatively expected from these models, additional processes are likely present which blur the clear spectral signatures of steady state magnetospheric accretion and associated mass loss. The occasional failure to observe some of the basic predictions of static MAMs may result at least in part from the intrinsically unstable nature of the magnetospheric accretion process as we discuss in the Section 2.

1.3

Accretion shocks, UV lines and veiling

Further observational support for MAMs comes from the detection of the rotational modulation of CTTS luminosity by bright surface spots. Modelling the periodic light variations observed on a timescale of the stellar rotation period suggests hot spots covering of order of one percent of the stellar surface. The localized hot spots are thus identified with the accretion shocks expected to develop near the magnetic poles at the base of magnetic accretion columns (Bouvier & Bertout 1989; Vrba et al. 1993). Continuum flux excesses from the UV to the optical range are expected to be associated with the accretion shock and, at least in some cases, correlated variations between the system’s brightness and veiling or excess flux supports this expectation (e.g. Chelli et al. 1999). Accretion shock models predict the formation of high excitation lines in the shock region (Calvet & Gullbring 1998, Lamzin 1995) and empirical evidence is consistent with the formation of narrow line profiles in the postshock gas (e.g. HeI, Beristain et al. 2001) while broader profiles would arise from the preshock region (e.g. Gomez de Castro & Lamzin 1999). However, the width, strength and kinematics of high excitation UV lines have so far eluded model predictions being broader (∼200 km/s, Ardila et al. 2002a), stronger (Lamzin & Kravtsova, this volume) and exhibiting lower centroid velocities (Ardila & Johns-Krull, this volume) than expected from models. Solving the discrepancies between MAMs predictions and observed line profiles may require the inclusion of a turbulent component in the accretion flow (Ardila et al. 2002b, Alencar & Basri 2000). Turbulence may additionally provide a heating source for the plasma in the funnel flow, up to values deduced from the observations (≤104 K) which non turbulent models hardly reach (e.g. Martin 1996). Thus, turbulence could conceivably generate magnetic waves whose damping in the accretion column would contribute to the heating of the plasma (see Vasconcelos et al., this volume). Clearly, improved shock models are required to yield

152

OPEN ISSUES IN LOCAL STAR FORMATION

a self-consistent description of the physical conditions in the magnetic accretion region of CTTS.

1.4

Magnetic braking and angular momentum regulation

T Tauri stars have surprisingly low rotation rates considering that they accrete high angular momentum material from their disk. In the magnetospheric accretion framework, the lever arm provided by extended magnetic field lines threading the disk outside the corotation radius1 allows the star to transfer angular momentum outwards with a net braking effect (Collier Cameron & Campbell 1993; Armitage & Clarke 1996; Shu et al. 1994). This possibly accounts for the slower rotation rates of accreting T Tauri stars compared to non accreting ones, as originally reported by Bouvier et al. (1993) and Edwards et al. (1993). Recent and more complete studies of TTS rotation rates have however failed to reveal the expected correlation between low rotation and disk accretion, at least for very low mass T Tauri stars (see Stassun et al. 1999; Herbst et al. 2000). Nevertheless, evidence has been accumulating that low mass pre-main sequence stars evolve down their Hayashi tracks without accelerating (Rebull et al. 2002), thus confirming the apparent paradox that the central star loses large amounts of angular momentum during the accretion phase.

2.

Magnetospheric accretion : a dynamical process

Many of the expected signatures of magnetically mediated accretion have thus been observed in CTTS. Evidence, however, comes mostly from snapshot studies and it is only recently that the stability of the phenomenon has started to be addressed. Time dependent modelling of the interaction between the inner disk and the stellar magnetosphere requires heavy numerical simulations and observational clues can only be obtained through long term monitoring studies combining high resolution spectroscopy and multi band photometry. In this section, we briefly review model predictions regarding the temporal evolution of the process and recent observational results which seem to indicate that it is highly time dependent indeed.

1 The corotation radius is the radius in the disk where the keplerian angular velocity of the disk material equals the angular velocity of the star. It is of order of a few stellar radii in TTS.

Magnetically channeled accretion in T Tauri stars (Bouvier et al.)

2.1

153

Time dependent models

Most MAMs assume that the stellar magnetosphere truncates the disk close to the corotation radius. Field lines threading the disk at this radius thus corotate with the central object. However, due to the finite radial distance over which the stellar magnetosphere interacts with the inner disk, the footpoints of most field lines rotate differentially, one being anchored into the star, the other into the keplerian disk. Recent numerical simulations indicate that magnetic field lines are thus substantially distorted by differential rotation on a timescale of only a few keplerian periods at the inner disk. One class of models predict that differential rotation leads to the expansion of the field lines, their opening and eventually their reconnection which restores the initial (assumed dipolar) magnetospheric configuration (e.g. Aly & Kuijpers 1990; Goodson et al. 1997; Goodson & Winglee 1999). Magnetospheric inflation cycles are thus expected to develop and to be accompanied by violent episodic outflows as field lines open and reconnect as well as time dependent accretion rate onto the star (Hayashi et al. 1996; Romanova et al. 2002). The most recent 3D MHD simulations of disk accretion onto an inclined stellar magnetosphere are presented in Romanova et al. (2003) and illustrate well the extreme complexity of the process. Other models, however, suggest that the field lines respond to differential rotation by drifting radially outwards in the disk, leading to magnetic flux expulsion (Bardou & Heyvaerts 1996). The response of the magnetic configuration to differential rotation mainly depends upon the magnitude of magnetic diffusivity in the disk, a parameter of the models which is poorly constrained from basic principles.

2.2

Preliminary observational evidence

Due mostly to the lack of intense monitoring of CTTS on proper timescales, the observational evidence for a time dependent interaction between the inner disk and the stellar magnetosphere is at present limited. Episodic high velocity outbursts, possibly connected with magnetospheric reconnection events, have been reported for a few systems based on the slowly varying velocity of blueshifted absorption components in emission line profiles on a timescale of hours to days (Alencar et al. 2001; Ardila et al. 2002a). Possible evidence for magnetic field lines being twisted by differential rotation has been reported for SU Aur by Oliveira et al. (2000). These authors measured a time delay of a few hours between the appearance of high velocity redshifted absorption components in line profiles formed at different altitudes in the accretion column. They interpreted this result as the crossing of an azimuthally

154

OPEN ISSUES IN LOCAL STAR FORMATION

twisted accretion column on the line of sight. Another possible evidence for magnetic field lines being twisted by differential rotation and leading to quasi-periodic reconnection processes has been reported for the embedded protostellar source YLW 15 based on the observations of quasi-periodic X-ray flaring (Montmerle et al. 2000).

2.3

AA Tau : the prototype of time variable magnetospheric accretion ?

Large scale synoptic studies of a handful of CTTS have been performed in the last years and provided new insight into the magnetospheric accretion process and its temporal evolution. Two of these have targetted the CTTS AA Tau which proved to be ideally suited to probe the inner few 0.01 AU of the disk-magnetosphere interaction region. Due to its high inclination (i 75o , see Bouvier et al. 1999), the line of sight to the star intersects the region where the stellar magnetosphere threads the disk. The peculiar orientation of this otherwise typical CTTS maximizes the variability induced by the modulation of the magnetospheric structure and thus provides the strongest constraints on the inner disk and the magnetospheric cavity. A first monitoring campaign, whose results are reported in Bouvier et al. (1999), led to the discovery of recurrent eclipses of the central object with a period of 8.2 days. The eclipses were attributed to a non axisymmetric warp of the inner disk edge which periodically obscures the the central star as it orbits it at keplerian speed. Such an inner disk warp is expected to develop as the disk encounters an inclined magnetosphere (Terquem & Papaloizou 2000; Lai 1999; Romanova et al. 2003). A second campaign which combined photometry and high resolution spectroscopy and whose results are reported in Bouvier et al. (2003) was performed in 1999. The spectroscopic variability provided evidence for accretion columns and associated hot spots with signatures (redshifted absorptions, continuum excesses) modulated on the rotation timescale of the system. In addition, a time delay of about 1 day was reported between the flux variations of lines forming at a different altitude in the accretion column, from Hα near the disk inner edge to HeI close to the stellar surface. The measured time delay is consistent with accreted gas blobs propagating downwards along the accretion column at free fall velocity from a distance of about 8 R , the radius at which the stellar magnetosphere disrupts the inner disk. These line flux variations indicate non steady accretion along the magnetic funnel flows onto the star on short timescales.

Magnetically channeled accretion in T Tauri stars (Bouvier et al.)

155

On longer timescales, of order of one month, line and continuum excess flux variations were also observed, indicative of a smoothly varying mass accretion rate onto the star. Simultaneously, a tight correlation was observed between the radial velocity of the blueshifted and redshifted absorption components in the Hα emission line profile (see Fig.2). Since the former is a wind diagnostics while the latter forms in the accretion flow, this correlation provides additional evidence for a physical connection between (time dependent) inflow and outflow in CTTS. Bouvier et al. (2003) argued that the associated flux and radial velocity variations can be consistently interpreted in the framework of magnetospheric inflation cycles, as predicted by recent numerical simulations of the star-disk interaction.

Figure 2. The correlation between the radial velocity of the blueshifted and redshifted absorption components in the Hα emission line profile of AA Tau (from Bouvier et al. 2003).

Figure 3. A sketch of the magnetospheric inflation scenario. The arrow on the right side indicates the line of sight to the AA Tau system (from Bouvier et al. 2003).

This is schematically illustrated in Figure 3. As magnetic field lines expand due to differential rotation between the star and the inner disk, the radial velocity of the accretion (resp. wind) flow decreases (resp. increases) due to projection effects on the line of sight. At the same time, the loading of disk material onto inflated field lines becomes increasingly difficult owing to the large angle field lines make relative to the disk plane. This results in a reduced accretion rate onto the star, as deduced from the depressed line and continuum fluxes observed at this phase. The last synoptic campaign on AA Tau thus yields the first evidence for global instabilities developping on a timescale of a month in the large scale structure of the magnetosphere, a result which provides support to

156

OPEN ISSUES IN LOCAL STAR FORMATION

the predictions of time dependent models of magnetospheric accretion. Whether these magnetospheric instabilities are cyclic, being driven by differential rotation as predicted by numerical models, will require additional monitoring lasting for several months.

3.

Conclusion

Observational evidence for magnetospheric accretion being instrumental in classical T Tauri stars has accumulated in recent years. Several key properties of these young stars are naturally accounted for by assuming that the stellar magnetic field governs the accretion flow close to the star. The strong variability of CTTS on all timescales, from hours to months (and possibly years, Bertout 2000), further suggests that the magnetically mediated interaction between the accretion disk and the central object is a highly dynamical and time dependent process. The implications of the non steadiness of magnetospheric accretion in CTTS are plentiful and remain to be fully explored. They range from the evolution of their angular momentum (Agapitou & Papaloizou 2000), the origin of inflow/ouflow short term variability (Woitas et al. 2002, LopezMartin et al. 2003), the modelling of the near infrared excess of CTTS and of its variations both of which will be affected by a non standard and time dependent inner disk structure (Carpenter et al. 2001, Eiroa et al. 2002, Johns-Krull & Valenti 2003), the origin of CTTS variability which is expected to be a complex combination of modulation by hot and cold spots and variable circumstellar extinction (e.g. DeWarf et al. 2003), and possibly the halting of planet migration close to the star (Lin et al. 1996).

Acknowledgments We thank the organisers for a very fruitful and enjoyable meeting, as well as for financial support.

References Agapitou, V. & Papaloizou, J. C. B. 2000, MNRAS, 317, 273 Alencar, S.H.P., & Basri, G. 2000, AJ, 119, 1881 Alencar, S. H. P. & Batalha, C. 2002, ApJ, 571, 378 Alencar, S. H. P., Johns-Krull, C. M., & Basri, G. 2001, AJ, 122, 3335 Aly, J. J. & Kuijpers, J. 1990, A&A, 227, 473 Andr´e, Ph. 1987, in: Protostars and Molecular Clouds, eds. T. Montmerle & C. Bertout, p.143 Ardila, D. R. & Basri, G. 2000, ApJ, 539, 834

Magnetically channeled accretion in T Tauri stars (Bouvier et al.)

157

Ardila, D. R., Basri, G., Walter, F. M., Valenti, J. A., & Johns-Krull, C. M. 2002a, ApJ, 566, 1100 Ardila, D. R., Basri, G., Walter, F. M., Valenti, J. A., & Johns-Krull, C. M. 2002b, ApJ, 567, 1013 Armitage, P. J. & Clarke, C. J. 1996, MNRAS, 280, 458 Bardou, A. & Heyvaerts, J. 1996, A&A, 307, 1009 Basri, G. & Bertout, C. 1989, ApJ, 341, 340 Batalha, C., Batalha, N. M., Alencar, S. H. P., Lopes, D. F., & Duarte, E. S. 2002, ApJ, 580, 343 Beristain, G., Edwards, S., & Kwan, J. 2001, ApJ, 551, 1037 Bertout, C. 2000, A&A, 363, 984 Bertout, C., Basri, G., & Bouvier, J. 1988, ApJ, 330, 350 Bouvier, J. & Bertout, C. 1989, A&A, 211, 99 Bouvier, J., Cabrit, S., Fernandez, M., Martin, E. L., & Matthews, J. M. 1993, A&A, 272, 176 Bouvier, J., Chelli, A., Allain, S., et al. 1999, A&A, 349, 619 (B99) Bouvier, J., Grankin, K., Alencar, S., et al. 2003, A&A, in press Calvet, N. & Gullbring, E. 1998, ApJ, 509, 802 Camenzind, M. 1990, Reviews of Modern Astronomy, 3, 234 Carpenter, J. M., Hillenbrand, L. A., & Skrutskie, M. F. 2001, AJ, 121, 3160 Chelli, A., Carrasco, L., M´ ujica, R., Recillas, E., & Bouvier, J. 1999, A&A, 345, L9 Collier Cameron, A. & Campbell, C. G. 1993, A&A, 274, 309 DeWarf, L.E., Sepinsky, J.F., Guinan, E.F., Ribas, I., Nadalin, I. 2003, ApJ, in press Edwards, S. et al. 1993, AJ, 106, 372 Edwards, S., Hartigan, P., Ghandour, L., & Andrulis, C. 1994, AJ, 108, 1056 Eiroa, C. et al. 2002, A&A, 384, 1038 Ferreira, J. 1997, A&A, 319, 340 Ghosh, P. & Lamb, F. K. 1979, ApJ, 234, 296 Gomez de Castro, A. I. & Lamzin, S. A. 1999, MNRAS, 304, L41 Goodson, A. P. & Winglee, R. M. 1999, ApJ, 524, 159 Goodson, A. P., Winglee, R. M., & Boehm, K. 1997, ApJ, 489, 199 Guenther, E. W., Lehmann, H., Emerson, J. P., & Staude, J. 1999, A&A, 341, 768 Gullbring, E., Hartmann, L., Briceno, C., & Calvet, N. 1998, ApJ, 492, 323 Folha, D. F. M. & Emerson, J. P. 1999, A&A, 352, 517 Folha, D. F. M. & Emerson, J. P. 2001, A&A, 365, 90 Hartigan, P., Edwards, S., & Ghandour, L. 1995, ApJ, 452, 736 Hartmann, L., Hewett, R., & Calvet, N. 1994, ApJ, 426, 669 Hayashi, M. R., Shibata, K., & Matsumoto, R. 1996, ApJl, 468, L37 Herbst, W., Rhode, K. L., Hillenbrand, L. A., & Curran, G. 2000, AJ, 119, 261 Johns, C. M. & Basri, G. 1995a, ApJ, 449, 341 Johns-Krull, C. M. & Gafford, A. D. 2002, ApJ, 573, 685 Johns-Krull, C. M. & Hatzes, A. P. 1997, ApJ, 487, 896 Johns-Krull, C. M. & Valenti, J. A. 2001, ApJ, 561, 1060 Johns-Krull, C. M. & Valenti, J. A. 2003, ApJ, in press Johns-Krull, C. M. & Valenti, J. A. 2000, ASP Conf. Ser. 198: Stellar Clusters and Associations: Convection, Rotation, and Dynamos, 371

158

OPEN ISSUES IN LOCAL STAR FORMATION

Johns-Krull, C. M., Valenti, J. A., & Koresko, C. 1999, ApJ, 516, 900 Johns-Krull, C. M., Valenti, J. A., Saar, S. H., & Hatzes, A. P. 2001, ASP Conf. Ser. 223: 11th Cambridge Workshop on Cool Stars, Stellar Systems and the Sun, 11, 521 K¨ onigl , A. 1991, ApJl, 370, L39 Lai, D. 1999, ApJ, 524, 1030 Lamzin, S. A. 1995, A&A, 295, L20 Lin, D. N. C., Bodenheimer, P., & Richardson, D. C. 1996, Nature, 380, 606 Lopez-Martin, L., Cabrit, S., Dougados, C. 2003, A&A, in press Martin, S. C. 1996, ApJ, 470, 537 M´enard, F. & Bertout, C. 1999, NATO ASIC Proc. 540: The Origin of Stars and Planetary Systems, 341 Meyer, M. R., Calvet, N., & Hillenbrand, L. A. 1997, AJ, 114, 288 Montmerle, T., Koch-Miramond, L., Falgarone, E., & Grindlay, J. E. 1983, ApJ, 269, 182 Montmerle, T., Grosso, N., Tsuboi, Y., & Koyama, K. 2000, ApJ, 532, 1097 Muzerolle, J., Hartmann, L., & Calvet, N. 2001, ApJ, 550, 944 Oliveira, J. M., Foing, B. H., van Loon, J. T., & Unruh, Y. C. 2000, A&A, 362, 615 Petrov, P. P., Gullbring, E., Ilyin, I., Gahm, G. F., Tuominen, I., Hackman, T., & Loden, K. 1996, A&A, 314, 821 Petrov, P. P., Gahm, G. F., Gameiro, J. F., Duemmler, R., Ilyin, I. V., Laakkonen, T., Lago, M. T. V. T., & Tuominen, I. 2001, A&A, 369, 993 Rebull, L. M., Wolff, S. C., Strom, S. E., & Makidon, R. B. 2002, AJ, 124, 546 Romanova, M. M., Ustyugova, G. V., Koldoba, A. V., & Lovelace, R. V. E. 2002, ApJ, 578, 420 Romanova, M. M., Ustyugova, G. V., Koldoba, A. V., Vick J.W., ,& Lovelace, R. V. E. 2003, ApJ, in press Shu, F., Najita, J., Ostriker, E., Wilkin, F., Ruden, S., & Lizano, S. 1994, ApJ, 429, 781 Stassun, K. G., Mathieu, R. D., Mazeh, T., & Vrba, F. J. 1999, AJ, 117, 2941 Smith, K., Pestalozzi, M., G¨ udel, M., Conway, J., Benz, A.O. 2003, AA, in press (astro-ph/0305543) Terquem, C. & Papaloizou, J. C. B. 2000, A&A, 360, 1031 Unruh, Y. C., Collier Cameron, A., & Guenther, E. 1998, MNRAS, 295, 781 Vrba, F. J., Chugainov, P. F., Weaver, W. B., & Stauffer, J. S. 1993, AJ, 106, 1608 Woitas, J., Ray, T. P., Bacciotti, F., Davis, C. J., & Eisl¨offel, J. 2002, ApJ, 580, 336

PROPERTIES OF YOUNG STELLAR OBJECTS FROM HIGH RESOLUTION NEAR INFRARED SPECTROSCOPY G.W. Doppmann NASA Ames Research Center, USA [email protected]

D.T. Jaffe Department of Astronomy, University of Texas, USA [email protected]

R.J. White California Institute of Technology, USA [email protected]

Abstract

We analyze high resolution (R=50,000) spectra at 2.2 µm of 10 Class II young stellar objects in the ρ Ophiuchi dark cloud. We measure effective temperatures, continuum veiling, and v sin i rotation from the shapes and strengths of atomic photospheric lines by comparing to spectral synthesis models at 2.2 µm. We measure surface gravities in 2 stars from the integrated line flux ratio of the 12 CO line region at 2.3 µm and the Na I line region at 2.2 µm. Although the majority (8/10) of the Class II stars have similar effective temperatures (3530 K ± 100 K), they exhibit a large spread in bolometric luminosities (factor ∼8), as derived from near–IR photometry. In the two stars where we have surface gravity measurements from spectroscopy, the photometrically derived luminosities are systematically higher than the spectroscopic luminosities. The spread in the photometrically derived luminosities in our other sources suggests either a large spread in stellar ages, or nonphotospheric emission in the J–band since anomalous and significant veiling at J has been observed in other T Tauri stars. Our spectroscopic luminosities result in older ages on the H–R diagram than is suggested by photometry at J or K. Most of our sources show a larger amount of continuum excess (FKex ) than stellar flux at 2.2 µm (FK∗ ), substantially higher in many cases (rK ≡ FKex /FK∗ = 0.3–4.5). The derived 159

J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 159-168. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

160

OPEN ISSUES IN LOCAL STAR FORMATION v sin i rotation is substantial, but systematically less than the rotation measured in class I.5 (flat) and Class I sources from other studies in Ophiuchus.

Introduction High resolution spectroscopy can be a valuable tool for extracting stellar properties of highly extincted young stellar objects (YSOs). Spectroscopy offers a more direct method than photometry for measuring properties of obscured young stars. Absorption line equivalent widths and ratios are not dependent on extinction. The intrinsic shapes of photospheric absorption lines, which can be fully resolved in cool stars (Teff < 5500 K) by high spectral resolution (R ≡ λ/∆λ ≥ 10,000–20,000) observations, contain stellar kinematic and temperature information. Observations of photospheric lines in obscured stars at high spectral resolution permit us to measure (1) precise radial velocities from line shifts, (2) v sin i rotational velocities from line widths, (3) effective temperatures and surface gravities from line shapes and line ratios, and (4) the continuum “veiling” by hot dust from the line depths. Observers have used many techniques to investigate the properties of stars in very young embedded clusters. In well–studied regions like the ρ Ophiuchi cloud core, previous investigators have used photometric surveys in the near–IR (NIR) (Wilking & Lada 1983, Greene & Young 1992, Barsony et al. 1997), in the mid–IR (Bontemps et al. 2001), and low–to–moderate resolution (R ≡ λ/∆λ = 500–2000) NIR spectroscopy (Casali & Matthews 1992, Greene & Meyer 1995, Greene & Lada 1996, Kenyon et al. 1998, Luhman & Rieke 1999) to estimate temperatures, luminosities, and the amount of excess (non–photospheric emission) in the infrared for the sources. In embedded clusters, however, even NIR photometry and low resolution spectroscopy suffer under disadvantages not inflicted upon these techniques when applied to main sequence stars or to less heavily extincted young stars. Problems include extremely high extinction (e.g. the central part of the ρ Ophiuchi molecular cloud where Av = 40 ± 10.9 magnitudes, Luhman & Rieke 1999), and excess emission in the near– and mid–IR (from warm dust in the circumstellar disks; Greene et al. 1994, Strom et al. 1995).

1.

The Spectral Fitting Technique

We have developed a spectral fitting technique that provides measurements of physical parameters in YSOs (Doppmann & Jaffe, 2003). We use information about the line shapes and relative strengths to derive effective temperature, v sin i rotational broadening, and continuum

YSO properties from IR Spectroscopy (Doppmann et al.)

161

veiling from fits to spectral synthesis models and are able to measure surface gravities from the equivalent width ratios between the Na lines at 2.2 µm and 12 CO lines at 2.3 µm. Since we compare our spectra to photospheric models with different effective temperatures, we measure Teff directly without introducing additional uncertainty due to conversion from spectral type. Our best fit stellar parameters are determined by the position of the minimum in the error space of residuals to the model fits.

1.1

Effective Temperature, Rotation, and Veiling

Using a wavelength region in the K–band at 2.200–2.210 µm (see Fig. 1), we develop a grid of synthetic spectra to which to compare our data and determine stellar properties. The synthetic spectra are generated from NextGen stellar atmosphere models (Hauschildt et al. 1999) using the spectral synthesis program, MOOG (Sneden 1973). The grid of synthesis models covers a range in cool effective temperatures (Teff = 3000–5000 K) and YSO surface gravities (log g = 3.5–5.0). We assume solar metallicity in generating the spectra. We add rotational broadening (v sin i = 4–40 km s−1 in steps of 1 km s−1 ) and 2.2 µm continuum veiling (rK = 0–4.0 in steps of 0.1) to these model spectra and compare each modified spectrum to the observed spectrum at a given radial velocity. We define the 2.2 µm veiling as the ratio of continuum excess at 2.2 µm (the wavelength of the Na interval) to the continuum arising from the stellar photosphere (rK ≡ F2.2µm excess /F2.2µm stellar ). A range of radial velocities (± 30 km s−1 ) is tested by shifting the model in increments of 0.8 km s−1 . We assume the infrared excess is uniform in intensity across our wavelength interval at 2.2 µm. The best model spectrum is selected by finding the minimum RMS of the residuals to the fit from selected wavelength sub–intervals that are centered on the Na lines. We start with an assumed gravity of log g = 3.5, which corresponds to ∼1–2 Myr old objects in stellar evolutionary models (Baraffe et al. 1998, Siess et al. 2000, Palla & Stahler 2000), consistent with age estimates of the central embedded cluster in Ophiuchus (Wilking et al. 1989, Greene & Meyer 1995, Bontemps et al. 2001). Next, we compare the observed spectrum to models over the entire range of our synthesis grid in effective temperature, v sin i rotation, and continuum veiling. We then find the best fit values of these parameters by examining the RMS deviation of the models from the observed spectrum over the regions where lines are

162

OPEN ISSUES IN LOCAL STAR FORMATION

present, and then interpolating between these models to find the values with the minimum RMS.

1.5

1

0.5

0 2.204

2.206

2.208

2.21

Wavelength (microns)

Figure 1. An illustration of the spectral fitting routine with GY 314 (a Class II YSO). a) The algorithm finds a best model fit with Teff =3500K, log g=3.5, rK (veiling)=1.0, and v sin i rotation=20 km s−1 . b) The residuals to the fit are shown where the minimum RMS of the region containing spectral lines selects the best fit. c) Spectral line depths diminish as veiling increases, while all other parameters are held fixed. (rK = 0.5, 1.0, 1.5 ) d) Increased rotational broadening causes spectral lines to become wider and shallower. (v sin i = 10, 20, 30 km s−1 )

1.2

Spectroscopic Determination of Log G

For late–type stars, the RMS difference in the 2.2 µm region between data and families of models in the temperature–gravity plane running in the direction from lower temperature and lower gravity to higher temperature and higher gravity can be very similar. If we were to adopt log g values that are larger or smaller by 0.5 dex than the value adopted here, the derived temperatures will change by ± ∼200 K and yield similar quality fits (Fig. 6 in Doppmann & Jaffe 2003). To resolve this ambiguity, we can use spectra from the (2–0) 12 CO line region (2.2925–2.2960 µm). The strength of the 12 CO bandhead is a sensitive luminosity diagnostic in late type stars (Baldwin et al. 1973, Kleinmann & Hall 1986, Ali et al. 1995, Ram´ırez et al. 1997). CO is only mildly sensitive to changes in effective temperature. The ratio of photospheric line strengths between the Na and CO spectral regions depends strongly on gravity at a given temperature but is only weakly dependent on temperature at a fixed gravity. We can there-

YSO properties from IR Spectroscopy (Doppmann et al.)

163

fore employ this ratio as an indicator of surface gravity in the two stars for which we have measurements of both spectral regions, GY 314 and IRS 2. We first find the best temperature at an assumed log g of 3.5 using our spectral fitting routine. For these Teff values, we then use the 12 CO/Na line strength ratio (Fig. 2) to estimate the gravity. We iterate between determinations of Teff and log g until the values converge (Doppmann & Jaffe 2003). This method gives log g values of 4.3 for IRS 2 and 3.5 for GY 314.

Figure 2. Isogravity contours as a function of the ratio of 12 CO/Na line flux and effective temperature. Equivalent widths were computed from 2.2020–2.2120 µm in the Na interval and 2.2925–2.2960 µm in the 12 CO interval. We calculate the surface gravity of 2 Ophiuchus sources from their effective temperatures and ratio of photospheric line fluxes measured in two regions at high spectral resolution, scaled to the slope of the observed spectral energy distribution at low resolution. T eff errors (horizontal error bars) are dominated by systematics in fitting observed spectra to the synthesis models (± 140 K, i.e. one spectral subclass). Errors in the measured line flux ratio (vertical error bars) are dominated by the determination of the continuum level, a function of S/N. Surface gravities derived using this technique are independent of uncertainties in the line–of–sight extinction and continuum veiling that are problematic in photometric measurements.

2.

Results

We apply our fitting technique to a modest sample of Class II YSOs in the Ophiuchus dark cloud. The derived properties are listed in Table 1.

164

OPEN ISSUES IN LOCAL STAR FORMATION

Table 1.

Measured stellar parameters of YSOs in ρ Ophiuchi sample

Source

Taeff (K)

Spb

v sin i (km s−1 )

rK

vLSR (km s−1 )

AcK

Ldbol (L/L
)

log g e (cm s−2 )

GY 319 GY 314 GY 23 GY 292 EL 24 SR 4 GSS 28 GY 308 GY 20A IRS 2

3410 3435 3490 3500 3520 3585 3610 3715 4460 4640

M3 M3 M2 M2 M2 M1.5 M1.5 M1 K5 K3

26 22 20 28 21 14 17 17 39 12

0.9 1.1 1.2 4.0 4.5 3.7 1.2 1.6 1.6 0.3

3.8 4.0 0.1 2.8 3.8 5.7 4.2 4.1 6.4 4.2

0.14 0.83 1.36 1.24 1.16 0.35 0.34 1.15 0.83 0.67

1.20 0.69 3.03 0.66 1.69 0.63 0.63 0.95 5.19 1.91

3.5 3.5f 3.5 3.5 3.5 3.5 3.5 3.5 3.5 4.3f

a Errors

b Based

are ±140K (Doppmann & Jaffe 2003) on Teff and on adopting the spectral type–Teff relation of de Jager & Nieuwenhuijzen

(1987) c Determined

using J and H photometry of Barsony et al. (1997), our measured effective temperatures, the extinction law of Martin & Whittet (1990), and the spectral type–color relation from table A5 of Kenyon & Hartmann (1995) d Determined from the observed K–band flux of Barsony et al. (1997), corrected for extinction, and the presence of continuum veiling that we derive at K e Assumed log g = 3.5 unless otherwise noted f Log g values measured using spectroscopy

2.1

Luminosities Determined from Photometry

We derive photospheric luminosities for our sources listed in Table 1 in two ways, both based on the NIR photometry of Barsony et al. (1997). First, we use the published J–band apparent magnitudes and assume there is no excess present. We then correct for the extinction at J by using the observed J–H colors, adopting the extinction law of Martin & Whittet (1990), and applying our measured stellar temperatures to the color–temperature relation in Table A5 of Kenyon & Hartmann (1995). Using a distance modulus of 5.81 for Ophiuchus (de Zeeuw et al. 1999), we obtain the absolute magnitude at J. We convert to bolometric luminosity using the derived effective temperature from our best model fit to the data and the bolometric correction factor from Table A5 of Kenyon & Hartmann (1995). Our second way of deriving luminosities from the photometry follows the same procedure as above except that we use the published K–band apparent magnitudes and remove the flux that is due to the continuum excess in the K–band that we have measured from our best model fits to the spectra. The K–band flux from the stellar photosphere is

YSO properties from IR Spectroscopy (Doppmann et al.)

165

FK∗ =FKtot /(1+rK ), where FKtot is the flux corresponding to the apparent magnitude of the source. We then compare the luminosities derived in these two ways by plotting these sources against their measured effective temperatures on an H–R diagram together with the model tracks of Baraffe et al. (1998) (Fig. 3). The dashed lines in Figure 3 connect the stellar luminosities derived from the dereddened J–band magnitudes to the luminosities derived from the dereddened K–band magnitudes from which we have subtracted the non–photospheric excess using the values of rK derived from our high resolution spectroscopy. The luminosities based on the J–band photometry are systematically higher (by an average of 0.35 in the log) than the luminosities derived from the corrected K–band magnitudes.

Figure 3. Differences in derived photometric luminosities of Class II PMS objects in Ophiuchus superposed on theoretical evolutionary tracks Baraffe et al. (1998). The open squares show the luminosities determined from J–band photometry with an assumed excess at J of zero. The filled squares show the same sources (connected by dashed lines) with luminosities calculated from K photometry, and accounting for our measured K–band excesses. Additionally, the solid circles show luminosities determined from spectroscopy for two sources (labeled). The solid vertical bar in the upper right corresponds to a downward shift in the derived luminosity when the J–band veiling increases from zero to rJ = 0.57, the average veiling value at J seen by Folha & Emerson (1999) in their sample of T Tauri Stars.

166

OPEN ISSUES IN LOCAL STAR FORMATION

2.2

Luminosities Determined from Spectroscopy

For IRS 2 and GY 314, where we were able to determine log g directly from spectra, we can determine the luminosity independent of the K– band photometry if the stellar mass is known. We use the PMS model tracks of Baraffe et al. (1998) to fit our observations of log g and Teff to select a unique mass track. Using the relationship between gravity and mass, we derive a luminosity: L = 4πσG

4 M Teff g

(1)

The conversion from gravity to luminosity is dependent on a particular evolutionary model, and we use the Baraffe models to derive luminosities from our spectroscopically measured surface gravities for IRS 2 and GY 314. These luminosities along with our measured effective temperatures for both sources are plotted on the H–R diagram against the Baraffe stellar evolutionary model tracks. (Fig. 3).

3.

Conclusions

We have carried out a high resolution NIR spectroscopic survey of YSOs in the ρ Ophiuchi dark cloud. We measure effective temperatures, continuum veiling, and v sin i rotation in ten Class II (T Tauri) sources. For two YSOs where we have high resolution spectra at 2.2 µm and 2.3 µm, we measure surface gravities using a new spectroscopic technique we have developed and compare our results to luminosities derived from photometry. From our measurements, we draw the following conclusions. 1. Most of the brightest Class II sources in the core of the ρ Oph cloud have cool derived effective temperatures that fall within a narrow range (< Teff > = 3530 K ± 100). The typical Class II source in our sample has a continuum excess larger than the stellar flux at 2.2 µm (< rK > = 2.3). With high resolution spectroscopy we are able to split the degeneracy between veiling and temperature which otherwise causes a systematic offset to higher temperatures and less veiling, evident when we compare our results to low resolution fitting techniques. 2. Using high resolution spectroscopy of spectral intervals including the Na lines at 2.2 µm and the (2–0) 12 CO bandhead at 2.3 µm, we derive luminosities for IRS 2 and GY 314 that are smaller than their photometrically derived luminosities. The ages inferred from the spectroscopic measurements are therefore older than the ages derived from

YSO properties from IR Spectroscopy (Doppmann et al.)

167

photometry. Furthermore, these sources do not appear co–eval based on the luminosities derived by either photometry or spectroscopy. Note, however, that the luminosities may be underestimated if the stars have strong magnetic fields. Future derivations of stellar parameters from high resolution spetra will need to take the effects of magnetic fields into account. 3. The v sin i rotations of YSOs in ρ Ophiuchi Class II sources (this study), along with Class I.5 and I sources (other high resolution studies) indicate that Class II YSOs have already spun down from earlier evolutionary stages (Class I.5 & I).

Acknowledgments We thank Chris Sneden for his advice and support in adapting MOOG for use with YSOs. We are grateful to Tom Greene and Charles Lada for providing us with CSHELL data on YSOs. We thank Ken Hinkle for help with the PHOENIX observations. Discussions with Isabelle Baraffe and Carlos Allende–Prieto were very helpful. We thank Chris Johns– Krull for insightful comments and for generously providing us with some spectral synthesis models that included the effects of magnetic fields. Support by the National Research Council is gratefully acknowledged.

References Ali, B., Carr, J.S., DePoy, D.L., Frogel, J.A., & Sellgren, K. 1995, AJ, 110, 2415 Baldwin, J.R., Frogel, J.A., & Persson, S.E. 1973, ApJ, 184, 427 Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 1998, A&A, 337, 403 Barsony, M., Kenyon, S.J., Lada, E.A., & Teuben P.J. 1997, ApJS, 112, 109. Bontemps, S. et al. 2001, A&A, 372, 173 Casali, M. M. & Matthews, H. E. 1992, MNRAS, 258, 399 de Jager, C. & Nieuwenhuijzen, H. 1987, A&A, 177, 217 de Zeeuw, P. T., Hoogerwerf, R., de Bruijne, J. H. J., Brown, A. G. A., & Blaauw, A. 1999, AJ, 117, 354 Doppmann, G. W. & Jaffe, D. T. 2003, AJ, submitted Folha, D. F. M. & Emerson, J. P. 1999, A&A, 352, 517 Greene, T.P., & Young, E.T. 1992, ApJ, 395, 516 Greene, T. P., Wilking, B. A., Andr´e, P., Young, E. T., & Lada, C. J. 1994, ApJ, 434, 614 Greene, T. P. & Meyer, M. R. 1995, ApJ, 450, 233 Greene, T. P. & Lada, C. J. 1996, AJ, 112, 2184 Hauschildt P.H., Allard, F. & Baron, E. 1999, ApJ, 512, 377 Kenyon, S. J. & Hartmann, L. 1995, ApJS, 101, 117 Kenyon, S. J., Brown, D. I., Tout, C. A., & Berlind, P. 1998, ApJ, 115, 2491 Kleinmann, S. G., & Hall, D. N. B. 1986, ApJS, 62, 501 Luhman. K. L. & Rieke, G. H. 1999, ApJ, 525, 440

168

OPEN ISSUES IN LOCAL STAR FORMATION

Martin, P. G. & Whittet, D. C. B. 1990, ApJ, 357, 113 Palla, F. & Stahler, S. W. 2000, ApJ, 540, 255 Ram´ırez, S.V., DePoy, D.L., Frogel, J.A., Sellgren, K., & Blum, R.D. 1997, AJ, 113, 1411 Siess, L., Dufour, E., & Forestini, M. 2000, A&A, 358, 593 Sneden, C. 1973 PhD Thesis, University of Texas at Austin Strom, K. M., Kepner, J., & Strom, S. E. 1995, ApJ, 438, 813 Wilking, B.A., & Lada, C.J. 1983, ApJ, 274, 698 Wilking, B. A., Lada, C. J., & Young, E. T. 1989, ApJ, 340, 823

Greg Doppmann

PROBING THE CIRCUMSTELLAR STRUCTURE OF PRE-MAIN SEQUENCE STARS Jorick S. Vink, Janet E. Drew Imperial College London, Blackett Laboratory, UK [email protected]

Tim J. Harries University of Exeter, UK

Ren´e D. Oudmaijer Department of Astronomy and Physics, UK

Abstract

We present Hα spectropolarimetry of a large sample of pre-main sequence (PMS) stars of low and intermediate mass, and argue that the technique is a powerful tool in studying the circumstellar geometry around these objects. For the intermediate mass (2 – 15 M
) Herbig Ae/Be stars we find that 16 out of 23 show a line effect, which immediately implies that flattening is common among these objects. Furthermore, we find a significant difference in Hα spectropolarimetry behaviour between the Herbig Be and Ae groups. For the Herbig Be stars, the concept of an electron scattering disc is shown to be a useful concept to explain the depolarizations seen in this spectral range. At lower masses, more complex Hα polarimetry behaviour starts to appear. The concept of a compact source of Hα emission that is formed close to the stellar surface, for instance by hot spots due to magnetospheric accretion, is postulated as a working hypothesis to qualitatively explain the Hα spectropolarimetry behaviour around Herbig Ae and lower mass (M < 2 M
) T Tauri stars. The striking resemblance in spectropolarimetric behaviour between the T Tauri star RY Tau and the Herbig Ae stars suggests a common origin of the polarized line photons, and hints that low and higher mass pre-main sequence stars may have more in common than had hitherto been suspected.

169 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 169-176. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

170

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction One of the most intriguing open issues in star formation concerns the formation of massive stars (e.g. Zinnecker; these proceedings). Although there is a well-established paradigm for the formation of low mass T Tauri stars, namely via magnetospheric accretion, it is as yet unclear whether such a scenario would also apply to the more massive stars. To be able to answer the question whether Nature allows the scaling-up of the formation mechanism of the Sun to the most massive stars, it first needs to be established whether the conditions known to prevail in the lower mass pre-main sequence (PMS) T Tauri stars, such as the presence of circumstellar discs and stellar magnetic fields, persist up to the intermediate mass (2-15 M
) PMS Herbig Ae/Be stars. The question as to whether Herbig Ae/Be stars in general are embedded in circumstellar discs, is still under debate. Although there are clear indications for flattening from millimeter imaging on larger spatial scales (a few hundred AU) for at least some objects (Mannings & Sargent 1997), other studies, probing smaller spatial scales, yield results that seem contradictory. For instance, the IR interferometry of MillanGabet et al. (2001) probes scales of only a few AU, and in this regime the geometry is found to be rather more spherical. Nonetheless, to be able to study the circumstellar geometries around PMS stars at the closest spatial scales, one needs to resort to the tool of spectropolarimetry, as this is the only technique that may probe the geometry on scales of stellar radii (equivalent to ∼ 0.05 AU) compared to the > 1 AU scales probed by other methods.

1.

The Tool of Linear Spectropolarimetry

In principle, the detection of linear polarization of ∼ 2 %, would teach us directly that a specific source is non-spherically symmetric on the sky. However, such a level of polarization may also be due to polarization by dust grains in the interstellar medium operating between the source and the observer. Unfortunately, properly correcting for this interstellar contribution has been proven to be a difficult task (e.g. McLean & Clarke 1979). This is one of the prime reasons as to why spectrallyresolved polarization changes across emission lines are so valuable, as the interstellar polarization affects the continuum and the line in exactly the same way: any observed change in the polarization across the line has to be intrinsic to the source. Although spectropolarimetry has widely been applied to more evolved early-type stars, such as classical Be stars (e.g Poeckert 1975), the technique has only recently been applied to pre-main sequence stars (Oud-

171

The circumstellar structure of PMS stars (Vink et al.)

maijer & Drew 1999, Vink et al. 2002, 2003). For classical Be stars, the dominant effect is known to be due to unpolarized line emission in the presence of intrinsic continuum polarization (e.g. Clarke & McLean 1974). This ‘depolarization’ effect across emission lines occurs because the line photons are formed over a larger volume (in the circumstellar disc) than the continuum photons and are therefore scattered to a lesser extent off free electrons in the disc than are the continuum photons. Consequently, a drop in the polarization percentage is seen (see Fig 1a for an example).

1a)

1b)

Figure 1: Triplots of the observed polarization spectra at Hα of the Herbig Be star BD+40 4124 (LHS) and the Herbig Ae star XY Per (RHS). On both plots, the Stokes I spectrum is shown in the lowest panel, the %Pol is indicated in the middle panel, while the position angle, θ, is plotted in the upper panel. The data have been rebinned to constant errors of 0.05 % for BD+40 4124 and 0.12 % for XY Per, as calculated from photon statistics. The data are taken from Vink et al. (2002).

In certain circumstances however, it is feasible that the converse occurs: a proportion of the line photons originate from a compact source and are scattered and polarized themselves (McLean 1979; Wood et al. 1993). Observationally, such effects have only recently been detected in intermediate and low mass Herbig Ae and T Tauri stars (Vink et al. 2002, 2003) using medium/high resolution (R  8000) spectropolarimetry. Here, the Hα line is believed to be polarized by scattering in a rotating non-spherically symmetric medium, most likely an accretion

172

OPEN ISSUES IN LOCAL STAR FORMATION

disc. Examples of both types of line effect, i.e. depolarization versus line polarization, are presented for respectively Herbig Be and Ae stars in Sects. 2 and 3 below.

2.

The Herbig Be Stars

A polarization spectrum for the Herbig Be star BD+40 4124 is shown in Fig. 1(a), and the observed behaviour across the Hα line profile in both polarization percentage (%Pol) and position angle (PA) is considered to be consistent with depolarization. The reason the PA shows a change across the line as well is attributed to the vector addition of the interstellar polarization contribution. Note that the smooth and broad depolarization effect is represented in the QU diagram of Fig. 2(a) by a more or less linear excursion of the line points out from the dense knot representing the continuum at (Q, U ) = (−0.3,1.25). The angle between this knot and the linear line excursion is directly related to the direction of the flattening of the presumed electron scattering disc around BD+40 4124.

2a)

2b)

Figure 2: QU representations of the observed polarization spectra of the same data as in Fig. 1. The arrow denotes the sense of increasing wavelength. The more or less linear excursion of the Hα line data for the Herbig Be star (LHS) is consistent with depolarization. The Herbig Ae data (RHS) is represented by a loop in the QU diagram. Note that the plot axis directions +Q, +U , −Q, −U correspond to sky PAs of respectively 0o , 45o , 90o , and 135o (i.e. ∆U/∆Q = tan 2θ).

In the event that all Herbig Be stars are embedded in electron scattering discs, one would not expect a 100% detection rate of Hα depolarisations, as at least some of the sources would have their discs too “pole-on” with respect to the observer to yield a %Pol drop large enough

The circumstellar structure of PMS stars (Vink et al.)

173

to be detectable. To estimate the expected fraction of depolarization detections, we turn to a comparison with classical Be stars (Poeckert & Marlborough 1976), objects for which the presence of a circumstellar disc is well-established. Applying the same detection threshold to the Poeckert & Marlborough sample as in ours, we expect a detection rate of about 54%. Returning now to our Herbig Be star sample, we find that 7 out of 12 (i.e. 58%) show a detectable depolarisation (Oudmaijer & Drew 1999, Vink et al. 2002). We conclude that, given both the statistics, as well as the smooth and broad depolarization behaviour in the Herbig Be Hα data, that all early Herbig Be stars are likely embedded in electron scattering discs.

3.

The Herbig Ae Stars

When observing the later spectral type Herbig Ae stars, one may expect to see a sharp decrease in the frequency of line effect detections, as the circumstellar ionization as well as the amount of free electrons that can scatter and polarize, are expected to drop among later type stars. Another reason to expect a decrease in the frequency of line effects going from Herbig Be to the later Ae stars is that there appears to be a general absence of Hα polarization changes in the even later type PMS T Tauri stars (Bastien 1982; but see Sect. 5). However, this turns out not to be the case at all. The number of line effects in Herbig Ae stars is found to be particularly high: 9 out of 11 Herbig Ae stars show a significant line effect (Vink et al. 2002). XY Per is included here as an example, as represented in Figs. 1(b) and 2(b): not only is there a line effect, it is noticeably different from the depolarization behaviour seen in the Herbig Be stars. First, the drop in the %Pol is not as broad as it is for the Herbig Be stars. Second, the behaviour in PA is not smooth. Instead, a line-center flip of the PA is clearly noticeable in the upper panel of the triplot in Fig. 1(b). This PA rotation translates into a “loop” in the equivalent QU diagram of Fig. 2(b). The interpretation of these QU loops in Herbig Ae stars is still a matter of ongoing investigation, but in the next section (Sect. 4), we show that photons arising from a compact source of hot spots (a natural consequence of the magnetospheric accretion model) on the stellar surface that are subsequently scattered in a rotating circumstellar disc may explain the observed Hα spectropolarimetry data of Herbig Ae (and possibly T Tauri stars; see Sect. 5).

174

4.

OPEN ISSUES IN LOCAL STAR FORMATION

Polarimetric Line Profiles from a Spotty Star

We employ a 3D Monte Carlo scattering model torus (Harries 2000) to simulate a compact source of photons arising from two diametrically opposed accretion hot spots onto a circumstellar disc. The disc has an inner hole of 3 stellar radii, and is assumed to be flat. As the Hα emission is compact in this model, we may expect intrinsic line polarization to occur. Figures 3(a) and 3(b) represent the two extreme cases for an almost edge-on (LHS) and an almost pole-on (RHS) disc respectively. As expected for an edge-on disc, there is a significant level of continuum polarization of a few percent. Due to the asymmetry in the velocity field, one hot spot – positioned at an angle of 65o from the equator – illuminates the redshifted part of the disc that is rotating away from us, while the other (diametrically opposed) hot spot illuminates the blueshifted part of the disc moving towards us. Both spots illuminate the disc at similar angles, so that no significant PA changes occur across the line.

3a)

3b)

Figure 3: Monte Carlo models for an edge-on (LHS) and a pole-on (RHS) disc. The photon source is asymmetric applying two diametrically opposed hot spots.

For the more pole-on case, the first difference is the lower level of continuum %Pol, as the scattering is now more spherically symmetric (positive and negative Q and U cancel, resulting into a drop in the net polarization). As far as the line polarization is concerned, the asymmetry in the velocity field is now diminished, and this results in an effective

175

The circumstellar structure of PMS stars (Vink et al.)

merging of the double-peaked %Pol profile into one single peak. Finally, because the illumination of the pole-on disc no longer occurs under similar angles, there is a rotation in the PA in the upper panel of Fig. 3(b), which is indeed the same type of behaviour that is commonly observed in Hα spectropolarimetry of Herbig Ae stars (see Fig. 1b).

5.

The T Tauri Star RY Tau

Partly because of a general absence of polarization changes across Hα using narrow band filters in T Tauri stars (e.g. Bastien 1982), the commonly accepted view of the origin of T Tauri polarization is that it is due to scattering off extended dusty envelopes. However, in Figure 4(a) we show the first spectropolarimetric measurements of an object classified as a T Tauri star, and we notice similar complexity across Hα for RY Tau as seen in Herbig Ae stars. Most notable is the rotation in the PA, which translates into a loop when plotted in QU space (Fig. 4b). This change in the polarization percentage and the PA across Hα suggests that line photons are scattered in a rotating disc. We are able to derive the value of the PA from the slope of the loop in the (Q, U ) diagram and find it to be 146 ± 3o . This is close to perpendicular with respect to the PA of the disc of 48 ± 5o as deduced from submillimeter imaging by Koerner & Sargent (1995). These findings are consistent, as the PA of the imaged millimeter disc is expected to lie at 90o to the scattering PA deduced from the polarization data.

4a)

4b)

Figure 4: Triplot (LHS) and QU diagram (RHS) of the observed polarization spectrum of the classical T Tauri star RY Tau. The data are binned to a constant

176

OPEN ISSUES IN LOCAL STAR FORMATION

error of 0.09 %. The RY Tau data are presented in Vink et al. (2003). Note the flip in PA and the corresponding loop in QU space. We finally note that if these medium/high resolution (R ∼ 8000) data were averaged – as in the narrow Hα filter observations of Bastien (1982)– they would most likely have produced a null result.

6.

Summary

Hα spectropolarimetry has been shown to be a powerful tool in studying the circumstellar geometry around low and intermediate mass PMS stars. For the Herbig Ae/Be stars we found that 16 out of 23 show a line effect, which immediately implies that flattening is common among intermediate mass PMS stars. Furthermore, we noticed a clear difference in Hα spectropolarimetry behaviour between the Herbig Be and Ae groups. For the Herbig Be stars, the concept of an electron scattering disc has been shown to be able to explain the depolarizations. At lower masses, more complex Hα polarimetry behaviour starts to appear. The concept of a compact source of near-stellar Hα emission, for instance associated with magnetospheric accretion, has been proposed to qualitatively explain the linear polarization changes across Hα seen in Herbig Ae and RY Tau. The striking resemblance in spectropolarimetric behaviour between these PMS stars may suggest a common origin for the polarized line photons, and hint that low and higher mass pre-main sequence stars may have rather more in common than had hitherto been suspected.

References Bastien P., 1982, A&AS 48, 153 Clarke D., McLean I.S., 1974, MNRAS 167, 27 Harries T.J., 2000, MNRAS 315, 722 Koerner D.W., Sargent A.I., 1995, AJ 109, 2138 Mannings V., Sargent A.I., 1997, ApJ 490, 792 McLean I.S., Clarke D., 1979, MNRAS 186, 245 McLean I.S., 1979, MNRAS 186, 265 Millan-Gabet R., Schloerb F.P., Traub W.A., 2001, ApJ 546, 358 Oudmaijer R.D., Drew J.E., 1999, MNRAS 305, 166 Poeckert R., 1975, ApJ 152, 181 Poeckert R., Marlborough J.M., 1976, ApJ 206, 182 Vink J.S., Drew J.E., Harries T.J., Oudmaijer R.D., 2002, MNRAS 337, 356 Vink J.S., Drew J.E., Harries T.J., Oudmaijer R.D., Unruh Y.C., 2003, A&A, in press Wood K., Brown J.C., Fox G.K., 1993, A&A 271, 492 Zinnecker H., these proceedings

ACCRETION SIGNATURES IN THE X-RAY SPECTRUM OF TW HYA Beate Stelzer Osservatorio Astronomico di Palermo, Italy [email protected]

J¨ urgen H. M. M. Schmitt Hamburger Sternwarte, Germany [email protected]

Abstract

We present the first XMM-Newton high-resolution spectrum of an accreting pre-main sequence star. The target, TW Hya, is the most nearby classical T Tauri star, and presents a unique testcase for the origin of X-ray emission in this class of objects. We find that its X-ray properties are rather untypical for scaled-up solar-like magnetic activity. In particular, the high density, low temperature, and peculiar elemental abundances seem to point at a different emission mechanism, possibly related to an accretion shock at the hot spot where the magnetically funnelled accreted disk material hits the star.

Introduction The origin of X-ray emission from late-type pre-main sequence stars, i.e. T Tauri stars (TTS), is commonly interpreted in terms of scaledup solar-type magnetic activity (see e.g. Feigelson & Montmerle 1999). While this is certainly a good approximation for stars approaching the main-sequence, the situation is less clear for those stars that prevail in the phase of disk accretion, the so-called classical TTS. In stars that undergo accretion an alternative mechanism of X-ray production is possible which is directly linked to the accretion process: In this scenario the emission site is above a hot spot where a magnetically funnelled accretion flow hits the surface of the star (Lamzin 1999). A distinction between a corona or an accretion funnel as the site of Xray production requires a precise assessment of the physical conditions in the emitting region. An adequate characterization of the emission region 177 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 177-184. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

178

OPEN ISSUES IN LOCAL STAR FORMATION

has only recently become feasible with the availability of high-resolution grating spectrometers onboard the XMM-Newton and Chandra X-ray observatories. The high signal-to-noise ratio required for a meaningful analysis of high-resolution X-ray spectra favors the study of the nearest star forming regions. One of the best-studied association of young stars in the solar neighborhood is the TW Hya association (d ∼ 57 pc; Wichmann et al. 1998). Its most prominent member, TW Hya, is the only classical TTS in this group. Because of its proximity and X-ray brightness giving rise to a high S/N of the X-ray spectrum, TW Hya represents the only classical TTS for which a high-resolution X-ray spectrum was obtained so far. TW Hya was observed with Chandra (see Kastner et al. 2002), and about one year later with XMM-Newton. Here we present the first results from the 30 ksec long XMM-Newton observation.

1.

Observations and Data Reduction

TW Hya was observed with XMM-Newton on July 9, 2001 with the Reflection Grating Spectrometer (RGS) as prime instrument. The European Photon Imaging Camera (EPIC) was operated in full-frame mode with the medium filter inserted. This instrument provides a mediumresolution spectrum (E/∆E ∼ 10 at 1 keV) contemporaneous to the high-resolution spectrum of the RGS (λ/∆λ ∼ 800 at 1 keV). XMMNewton and its X-ray instruments are described by Jansen et al. (2001), Str¨ uder et al. (2001), Turner et al. (2001), and den Herder et al. (2001). We analysed the data using the standard XMM-Newton Science Analysis System (SAS), version 5.3.0. The first order RGS spectrum of TW Hya is shown in Fig. 1. Lines from H-like and He-like carbon, nitrogen, oxygen, and neon are recognized, while iron lines are weak or absent, especially those near 15 and 17 ˚ A.

2.

X-ray properties

Density. The high spectral resolution of the RGS allows us to use flux ratios between individual emission lines for plasma diagnostics. In this respect the triplets of He-like ions are particularly important, because ratios between the intensity of resonance r, intercombination i, and forbidden f line are sensitive to electron temperature and density (Blumenthal et al. 1972). The critical parameter for estimating the density is R = fi . In Fig. 2 we zoom in on the spectral region around the neon triplet. A remarkable feature of this triplet is the weakness of the forbidden line. In the spectra of all other late-type stars observed so far the forbidden

X-ray spectrum of TW Hya (Stelzer & Schmitt)

179

Figure 1. First order XMM-Newton RGS count rate spectrum of TW Hya; exposure time is 29 ksec and binsize is 0.02 ˚ A. Emission lines typical for stellar coronae are indicated by labels and dashed lines. Straight horizontal lines represent gaps due to CCD chain failure or individual chip separation.

180

OPEN ISSUES IN LOCAL STAR FORMATION

line of neon is stronger than its intercombination line. In hotter stars or cool stars that form a close binary with a hot star the situation may be different because the strong UV radiation field of a hot star can induce transitions between the upper level of the f and the upper level of the i line. The net effect is a decrease of the R ratio. However, the UV radiation of TW Hya is rather low at the critical wavelength (Costa et al. 2000), such that this effect is negligible. In general the analysis of the Ne IX triplet is complicated by blending with iron lines (e.g. Ness et al. 2002a). However, TW Hya shows a curious lack of even the strongest Fe L-shell transitions in the region between 15 − 17 ˚ A. Therefore, no significant contribution from iron lines is expected for the Ne IX triplet. Comparison of the measured R ratio to calculations for collisional ionization equilibrium provides the density. We use the models by Porquet et al. (2001), and summarize the results for the Ne IX and for the O VII triplet in Table 1.

Figure 2. He-like triplet of Ne IX measured with the RGS together with best fit Lorentzians. The continuum close to each triplet is approximated by a straight horizontal line.

In Fig. 3 we confront the G and R ratios of TW Hya to a number of late-type main-sequence stars, and compare the temperatures and densities derived from the model calculations by Porquet et al. (2001). For this comparison we selected only those stars for which the UV field is negligible from the literature (Ness et al. 2002b; Stelzer et al. 2002). All main-sequence stars are grouped near the low-density limit, while TW Hya stands out with very high density.

181

X-ray spectrum of TW Hya (Stelzer & Schmitt)

Table 1. Line fluxes for the three components of He-like triplets of oxygen and neon, deduced R ratio, and electron density derived with the calculations by Porquet et al. (2001). Ion

r

17.4 ± 2.0 25.0 ± 1.1

O VII Ne IX

0.6

i ph photon flux [10−5 s cm 2] 10.2 ± 1.6 17.1 ± 2.8

f

R

ne [cm−3 ]

0.7 ± 0.5 8.1 ± 0.8

0.07 ± 0.04 0.40 ± 0.06

> 2 × 1012 1013

4.0

3.0

0.8

G

2.0

1.0

1.5

12.7 12.0

1.0 0.0

1.2

4

11.0 10.0

3

O VII

2 R

1

0

Figure 3. G versus R ratio of O VII and model grid of electron temperature (dotted lines) and density (dashed lines) for a plasma in collisional ionization equilibrium and negligible UV radiation field. Date: filled circle - TW HYa, diamonds - YY Gem (RGS and Chandra LETGS measurements), squares AD Leo, x-point Eri, triangle - α CenA, arrow for upper limit of α CenB.

Temperature. Despite its large overall X-ray luminosity (Lx = A) the RGS spectrum of TW Hya does 1.4 × 1030 erg/s in 0.45 − 2.25 ˚ not contain any high temperature lines. This is confirmed with the medium-resolution EPIC spectrum. A global fit to this spectrum with a model for optically thin emission from a hot plasma (MEKAL; Mewe et al. 1985, Mewe et al. 1995) yields a dominating temperature component of ∼ 3 MK, and a hotter component (T = 7.15 MK) with much lower emission measure. In contrast to the typical EPIC spectrum of other late-type stars that provide similar S/N, no third high-temperature component is required. The X-ray emission of TW Hya is unusually soft. The temperature of the main component is in agreement with the pronounced peak in the differential emission measure distribution presented by Kastner et al. (2002) on basis of the Chandra HETGS spectrum of TW Hya.

182

OPEN ISSUES IN LOCAL STAR FORMATION

Elemental Abundances. The strongest lines in the spectrum of TW Hya (besides the triplets from He-like ions) are the Lyα lines from H-like ions. These lines are clearly detected for neon (at 12.4 ˚ A), oxygen (19.0 ˚ A), nitrogen (24.8 ˚ A), and carbon (33.7 ˚ A). Intensity ratios between lines of different elements that form at similar temperature can be used to obtain information on the relative elemental abundances. In Fig. 4 we examine various energy flux ratios to the theoretical expectation for a solar abundance model from the CHIANTI database (Young et al. 2003). Fig. 4 (top left) shows the flux ratio between the Lyα line of C VI and the He-like r line of O VII. Assuming a plasma temperature of 3 MK, as found from the EPIC spectrum, the abundances of oxygen and carbon are within a factor two of each other. On the other hand we find that the observed line ratio between the r line of Ne IX and the Lyα line of O VIII can not be explained by any single temperature or combination of temperatures under the assumption of solar abundances for both elements (Fig. 4 top right). We conclude that neon must be overabundant by a factor of 10 relative to oxygen. Finally we investigated the flux ratio between the Lyα lines of N VII and C VI in the same way (Fig. 4 bottom left). The result points at a slight overabundance of nitrogen with respect to carbon (and oxygen), albeit less pronounced than the overabundance of neon.

3.

Conclusions

The three properties distinguishing the X-ray spectrum of TW Hya from that of ’normal’ late-type stars are (1) the high density of the Xray emitting region, (2) the absence of a hot component with large emission measure, and (3) the peculiar elemental composition, in particular the low abundance of iron. These characteristics suggest either strong evolutionary effects on coronal structures or an entirely different origin for the X-ray emission of TW Hya. X-ray emission from a shock at the bottom of an accretion column provides a plausible mechanism for the latter possibility. The observed temperature of 3 MK is a typical shock temperature. Interpreting the measured density as the pre-shock particle density, and making use of the strong shock formula, we estimate the thickness of the post-shock layer (Lpost ≈ 2500 km), the emitting area (Ashock ≈ 4×1018 cm2 ), and finally the mass accretion rate. The derived value of ∼ 10−11 M
/yr is quite small, but consistent within one order of magnitude of the measurement derived from UV data (see Muzerolle et al. 2000).

X-ray spectrum of TW Hya (Stelzer & Schmitt)

183

Figure 4. Temperature dependence of various line ratios measured in the RGS spectrum of TW Hya. Solid lines are model calculations from CHIANTI (Young et al. 2003) for solar abundances, shaded regions denote the 1 σ range observed with the RGS. top left C VI Lyα / O VII He r (in linear scale), top right - Ne IX He-like r / O VIII Lyα , and bottom left - N VII Lyα / C VI Lyα .

Acknowledgments The authors wish to thank B. Aschenbach and R. Neuh¨ auser for their contribution to the realisation of this observation within the Guaranteed Time program at the Max-Planck Institut f¨ ur extraterrestrische Physik in Garching.

References Blumenthal G. R., Drake G. W. F. & Tucker W. H., 1972, ApJ 172, 205 Costa V. M., Lago M. T. V. T., Norci L., Meurs E. J. A., 2000, A&A 354, 621 den Herder, J. W., Brinkman, A. C., Kahn, S. M., et al. 2001, A&A, 365, L7 Feigelson E. D. & Montmerle Th., 1999, ARA&A 37, 363 Jansen, F., Lumb, D., Altieri, B., et al. 2001, A&A, 365, L1 Kastner J. H., Huenemoerder D. P., Schulz N. S., Canizares C. R., 2002, ApJ 567, 434 Lamzin S. A., 1999, Astr. Lett. 25, 7 Mewe R., Kaastra J., S., Schrijver C. J., van den Oord G. H. J. & Alkemade F. J. M., 1995, A&A 296, 477 Mewe R., Gronenschild E. H. B. M. & van den Oord G. H. J., 1985, A&AS 62, 197 Muzerolle J., Calvet N., Brice˜ no C., et al., 2002, ApJ 535, L47 Ness J.-U., Schmitt J. H. M. M., Burwitz V., Mewe R., Raassen A. J. J., 2002a, A&A 394, 991

184

OPEN ISSUES IN LOCAL STAR FORMATION

Ness J.-U., Schmitt J. H. M. M., Burwitz V., Mewe R. & Predehl P., 2002b, A&A 387, 1032 Porquet D., Mewe R., Dubau J., Raassen A. J. J. & Kaastra J. S., 2001, A&A 376, 1113 Stelzer B., Burwitz V., Audard M., et al., 2002, A&A 392, 585 Str¨ uder, L., Briel, U. G., Dennerl, K., et al. 2001, A&A 365, L18 Turner M. J. L., Abbey A., Arnaud M., et al. 2001, A&A 365, L27 Wichmann R., Bastian U., Krautter J., Jankovics I. & Rucinski S. M., 1998, MNRAS 301, L39 Young P. R., Del Zanna G., Landi E., Dere K. P., Mason H. E. & Landini M., 2003, ApJSS 144, 135

S. Mohanty, B. Stelzer, J.-L. Monin

POST-T TAURI STARS ROTATION IN ASSOCIATIONS Ramiro de la Reza, Giovanni Pinz´ on Observat´ orio Nacional-MCT, Brazil [email protected], [email protected]

Abstract Nearby associations or moving groups of Post-T Tauri stars with ages between 10 and 30 Myr are excellent objects for the study of the initial spin up phase during the PMS evolution. An empirical approach is made here to infer their rotation properties and relations to X-ray emission. Three associations with distances less than 100 pc are considered. The TW HYa association (TWA) with an age of 10 Myr, the Beta Pictoris Moving Group (BPMG) with an age of 12 Myr and a combination of Horologium and Tucana associations (30 Myr). Two high and low rotation modes are considered for each association corresponding to high mass 1-2 solar mass and low mass, less than 1 solar mass respectively. Whereas the low mode practically doesn’t change during this interval of ages, the high mode changes drastically. Following this, we propose that the high mode could be used as a fine tuning gyrochronometer for associations. These mean velocity enabled us to examine the behavior of the main indicators of X-ray activity, Lx and Lx / Lbol , with rotation and age.

Introduction Studies of the rotational evolution of pre-main sequence (PMS) stars from the initial stages of T Tauri stars (TTS) up the arrival to the main sequence (MS) are mainly devoted to angular momentum transfer processes, X-ray emissions and on possible actions of dynamo mechanisms. Due to the recent use of wide field CCD image applied to dense PMS clusters an espectacular increase of photometric rotation measured periods with very high accuracy appeared in the literature. Also the use of ROSAT and specially CHANDRA X-ray satellites enabled the measurement of hundreds of PMS in the Orion Nebula Cluster (ONC) (Feigelson 185 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 185-192. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

186

OPEN ISSUES IN LOCAL STAR FORMATION

et al. 2003, Flaccomio et al. 2003a). These observations, refers to classical T Tauri stars (CTTS) having important accreting disks and to the weak T Tauri stars (WTTS) with much less important accretion disks. CTTS have in general longer rotation periods; this is considered to be the result of a magnetic braking acting between the disk and the central star of the CTTS system. WTTS in fact, present shorter periods. Concerning the more evolved Post T Tauri stars (PTTS) with ages ≥ 5 Myr up to near 40 Myr, measured periods are scarce. is, due not only because these PTTS have been only recently detected, but also to the inherent difficulties of measuring stars, due to their proximity (≤ 100 pc) distributed in vast areas of the sky. Photometric periods have been obtained nevertheless, for 34 PTTS in Lupus (Wichmann et al. 1998) with a mixture of ages between near 1 Myr and 37 Myr. Also for 9 PTTS belonging to visual binaries Lindroos systems where the primary is a late B star (Hu´elamo, 2002). Also here, a large mixture of ages is present. Even if PTTS can be generally defined as low mass stars with ages between 10 to 100 Myr (Jensen, 2001) it is not clear which are the real limits in both extremes. The purpose of this work is to introduce, as far as we know, for the first time studies of rotation of PTTS belonging to coeval nearby associations. Then, contrary to the PTTS mentioned before, these stars belong to groups having definite ages. We believe that the use of discret ages systems will introduce a fine-tunning and a more clear picture of the evolution of PTTS in the interval of 10 up to 30 Myr. For this, we will use here only published data of the following associations: TWA with an age of 10 Myr (see for instance Torres et al. 2003), BPMG (Zuckerman-Song, 2001a) with an age of 12 Myr and a combination of the Horologium and Tucana association (Tuc-HorA) (Torres et al. 2000) (Zuckerman et al. 2000, 2001b) considered to belong to a larger association called The Great Austral Young Association (GAYA) (Torres et al. 2001). The ages of HorA and TucA were independently found to be near 30 Myr. Unhappily there are not yet measurements of periods of individual members of the PTTS associations but only projected vsini rotation velocities. We present here a method to derive representative equatorial velocities from the vsini data in order to estimate mean periods.

1. 1.1

The Stellar Rotation Determination of mean equatorial velocities

A distribution of projected vsini rotational velocities values φ(vsini) = φ(y) is related by an integral equation (Chandrasekhar & Munch, 1950)

PTTS in Associations (de la Reza & Pinz´ on)

187

TWA

6

V0 =18 Km s-1 , 193 Km s-1 Age~9 Myr

5

f

4 3 2

1

25

50

75

4

V0 =15 Km s-1 , 150 Km s-1 Age~12 Myr

3 f

100 125 150 175 200 vsin i BPMG

2 1

25

50

75

100 125 150 175 200 vsin i

Hor-TucA V0 =20 Km s-1 , 140 Km s-1 Age~30 Myr

8

f

6

4

2

25

Figure 1.

50

75

100 125 150 175 200 vsin i

vsini distributions for the three associations studied in this work. The mean equatorial velocity v0(eq) of low modes corresponding to the fitting procedure explain in the text are indicated at upper right of each panel. For the high modes the mean velocity are also indicated (see text). Two dashed lines are ploted corresponding to a = 5 Km/s and a = 10 Km/s (lower curve).

188

OPEN ISSUES IN LOCAL STAR FORMATION

to the true rotational velocities considered here to be the equatorial velocities v0(eq) having a probability distribution f (v). Assuming that the rotation axis are ramdomly distributed in space this relation is 

φ(y) =

f (v)dv/v(v 2 − y 2 )

(1)

It is not an easy task to derive f (v) from φ(y) by a numerical inversion procedure (Gaig´e, 1993) because φ(y) has to be differenciated. For this it is necessary that the observed distribution φ(vsini) (normally in histograms) be transformed into a continuous function. This is possible when we dispose of a large collection of good quality vsini data. Queloz et al. (1998) realized this inversion for the Pleiades young cluster (age 70-125 Myrs) using vsini data for 235 members stars. Differently from young clusters where in general more than one hundred stars can be measured, the PTTS associations contain much less members. That is the case of TWA and BPMG each of which possess near 20 stellar systems with a large proportion of visual binaries. Adding all probable and possible members of HorA and TucA (where much less binaries are present) we obtain a total of near 40 objects. Due to the low number of PTTS in associations and also considering the uncertainties of vsini values of the order of 2 to 5 Km/s for low and high rotators respectively in the best of the cases, a different strategy must be followed. We adopt here the suggestion of Chandrasekhar & Munch (1950) with consist, in case of few data, of instead invert the integral equation, to assume a parametric form of f (v). We can then compute the integral in (1) and obtain the corresponding φ(vsini) distribution which will be compared to the observational distribution of vsini for each association. (Brown, 1950) tested different functions from a simple rectangular function to a more elaborated parabolic function. This last one, similar to the gaussian form used by Chandrasekhar & Munch produces indistinguishable results from those produced by the rectangular function. We then use this rectangular function as proposed by Brown by adopting f (v)dv, the probability of ocurrence of v between v and v + dv, to be equal to dv/2a for v > v0 − a and v < v0 − a and equal to zero otherwise. The rectangle centered at the probable v0 velocity has a height of 1/2a and a total width of 2a. To compare with observations a cumulative distribution function of the projected velocities must to be defined as  Φ(y) = φ(y)dy. This integration is realized over finite intervals of y corresponding to the grouping of the observations. The fraction of stars with (vsini)l < y < (vsini)l+1 will be given by Φ(vsini)l+1 − Φ(vsini)l . By multiplying this difference by the total number of stars in the as-

PTTS in Associations (de la Reza & Pinz´ on)

189

sociation we obtain the total number of stars in that interval of vsini values. The resulting cumulative function Φ(y) is a monotonically increasing function from zero to one as y increases. To calculate Φ(y) we used the relation (6) of (Brown, 1950). The derivative of Φ(y) using a hundred succesive intervals of y will finally be compared with the observed φ(vsini) distributions.

1.2

Main Results

In Fig 1 we present three histograms corresponding to the frequency distribution of vsini velocities of TWA, BPMG and Hor-TucA. The main source of data of TWA is that of (Torres et al. 2003); we gave preference to their vsini values than those appearing in the literature excepting for TWA stars not appearing in table 3 of this reference. That is the case for example of the only high rotator of TWA, one A star (HR 4796A) of this association where vsini = 152 Km/s. The data of BMPG is that of Zuckerman et al (2001a). Data for HorA and TucA are respectively are taken from Torres et al. (2000) and Zuckerman et al. (2001b). In order to reproduce the data in case of a bi-modal behavior we have to accept a priori the existence of two laws of rotation in an association. These two modes correspond to two different stellar mass intervals. The low mode to stars with masses ≤ 1 M
and the high mode to stars with masses between 1 and 2 M
. Due to the nature of our solution, the assumed f (v) is not necessarily unique. We tested several of them and found that they are not so different. Considering the values of the width a = 5 Km/s, we consider that they are representative mean distributions. The histograms of Fig 1 are disposed in an age sequence with ages of 10, 12 and 30 Myr respectively. Only BPMG appears to have a quite consolidated age value. In fact, two completely different approaches produced the same age. On one hand the stellar evolutionary age obtained with a HR diagram gave a value of 12 Myr (Zuckerman et al. 2001b). On the other hand, a dynamical age of 11.5 is obtained by the first focussing of time reversal 3D orbits under a Galactic potential (Ortega et al. 2002). See also Ortega et al. in these proceedings. A low rotation mean mode is obtained by adjusting the distribution of vsini velocities up to a maximum value vsini = 40 Km/s for all the associations. The best ajustments are obtained with the following v0(eq) values of 18±5, 15±5 and 20±5 Km/s. for TWA, BPMG and Hor-TucA respectively. The high rotation mode distribution is so different that our fiting procedure can only be applied approximately only for Hor-TucA, where a bi-modal pattern with two peaks appears to be emerging. For

190

OPEN ISSUES IN LOCAL STAR FORMATION

BPMG, where no second peak is present, we have taken the mean of the vsini values larger than 50 Km/s and multiplied by 4/π (ChandrasekharMunch, 1950) to obtain the mean equatorial velocity. In TWA only one single high rotating star represent the high mode. For this particular situation we estimate (exeptionally not using a mean) by multiplying its vsini value by the same factor 4/π. The mean high modes are then represented by the following equatorial velocity: 193, 150 and 140 ± 5 for TWA, BPMG and Hor-TucA respectively. Fig 1 clearly represents the different behavior of the two rotation modes with age. Whereas the low mode practically don’t change between these ages, this is not the case of the high mode. It is specially remarkable the difference found between TWA and BPMG for only 2 Myr age difference. This large difference of pattern leave us to suspect that TWA could be younger than 10 Myr. Because these PTTS are in general no more related in principle to a gaseous disk which could produce a magnetic braking, they are free to spin up their rotations due to stellar contraction. Because, the low mode represent stars with stellar masses ≤ 1 M
and the high mode stars with higher masses between 1 and 2 M
, the behavior shown in Fig 1 is then quite natural. The low mode don’t change significantly because stellar contraction is slow for low mass stars whereas for the higher mass the contraction is faster (see for example, Bouvier et al. 1997). If in the future, we will dispose of measured periods for individual stars we can use some high mass stars as gyrochronometers (G stars for exemple) to follow the initial spin up masses for different associations. A new tool can then be used to estimate the age of an association by examining their high mass rotations content. One of the main purposes of studies of X-ray radiation in PMS stars concern the search for correlations with stellar rotation in order to shed some light on what kind of dynamo is into action (see Feigelson et al. 2003 for a recent discussion on different dynamos mechanisms). Flaccomio et al. 2003a have shown that is the central star and not the disk the main source of X-ray production. This conclusion is the result of hundred of X-ray measurements in the ONC (age of 1 to 3 Myr) made using the CHANDRA observatory. In fact, Flaccomio et al. (2003a) found that ONC stars with low accretion have an order of magnitude higher luminosities than the high accreting ones. As concerning rotation, despite such a large volume of data, not a strong clear correlation of X-ray luminosities with rotation have been found in ONC. Whereas Feigelson et al. (2003) found a slight increase of X-ray

PTTS in Associations (de la Reza & Pinz´ on)

191

luminosities with periods from 0.4 to 20 days1 , Flaccomio et al. (2003a) found no correlation with rotation (even if is the same ONC which have been observed it is not completely clear if the sample of these two groups of authors is similar2 ). Our purpose here is to investigate if the saturation is retained for the next 30 Myr. Our approach considering more evolved PMS stars than TTS have the interest to investigate the role of central stars in an eventual desaturation tendancy. This is because, with some very few exceptions, PTTS have no more TT type accreting disks and produce less X-ray absorption. Our obtained mean v0(eq) values enabled us to calculate the mean periods of low and high rotation modes for each association. For this (also to estimate the stellar mass values) we have calculate a mean stellar radius for each mode using the stellar PMS models of (Siess et al. 2000). With those periods we explore the behavior of the two main X-ray activity indicators Lx and Lx / Lbol , where Lx and Lbol are respectively the X-ray and bolometric luminosities. The stellar distances used to calculate these luminosities were taken from the respective literature of these associations mentioned above. We have not found any correlation of Lx or Lx / Lbol , with the periods in the two modes of rotation difficulting this way any interpretation of the importance of rotation in an eventual dynamo mechanism in PTTS. This behavior is similar as that found for TTS in ONC. Recently Flaccomio et al. (2003b) studied the age variation of Lx and Lx / Lbol for PMS stars from ages of 1 Myr up to ages of young MS belonging to the Pleiades and NGC 2516 with ages of 100 Myr and 140 Myr respectively. The present study permit to fill partially a gap in the Flaccomio et al. (2003b) study for ages between 7 and 100 Myr. Our results show that for PTTS with masses from 1.0 to 2.0 M
, Lx / Lbol continue to be below the canonical saturation value of Lx / Lbol = 10−3 . For PTTS with masses less than one solar mass there is a tendancy to desaturation only at the age of 30 Myr. Lx values continue to be constant for the low mass range up to 30 Myr, whereas Lx diminish for the high mass range at 30 Myr converging to the Lx values of young MS corresponding to the Pleiades.

References Brown A. 1950, ApJ 111, 366 Chandrasekhar S., Munch G. 1950, ApJ 111, 142 Feigelson E., Gaffney J., Garmire G., Hillenbrand A. 2003, ApJ 584,911 Flaccomio E., Micela G., Sciortino S. 2003a, A&A 402, 277

1 this

is due to the existence of low rotating large X-ray emitters in ONC communication with T. Montmerle

2 private

192

OPEN ISSUES IN LOCAL STAR FORMATION

Flaccomio E., Micela G., Sciortino S. 2003b, A&A 397, 611 Gaig´e Y. 1993, A&A 269, 267 Hu´elamo N. 2002, PhD disertation, Universidad Aut´ onoma de Madrid, Departamento de F´ısica Te´ orica Jensen E. 2001, Young Stars Near Earth: Progress and Prospects, ASP Conference Series Vol. 244. Edited by Ray Jayawardhana and Thomas Greene. San Francisco: Astronomical Society of the Pacific Ortega V., de la Reza R., Jilinski E., Bazzanella B. 2002, ApJ 575L, 750 Queloz D., Allain S., Mermilliod J., Bouvier J., Mayor M. 1998, AA 335,183 Siess L., Dufour E., Forestini M. 2000, A&A 358, 593 Torres C., da Silva L., Quast G., de la Reza R., Jilinski E. 2000, AJ 120, Issue 3 Torres G., Guenther E., Marschall L., Neuhauser R., Latham D., Stefanik R. 2003, AJ 125,825 Wichmann R., Bouvier J., Allain S., Krautter J. 1998, A&A 30, 521 Zuckerman B., Song I., Bessell, M., Webb R. 2001a, ApJ 562, L87 Zuckerman B., Song I., Webb R. 2001b, ApJ 559, 388 Zuckerman B., Webb R. 2000, ApJ 535, 959

Ramiro de la Reza

STAR FORMATION IN CANIS MAJORIS The nature of the X-ray sources Jane Gregorio-Hetem Universidade de S˜ ao Paulo; IAG/USP,Brazil [email protected]

Thierry Montmerle CEA Saclay & Observatoire de Grenoble, France

Edson R. Marciotto Universidade de S˜ ao Paulo; IAG/USP, Brazil

Abstract

The X-ray emission detected by ROSAT is used to search for unidentified young stars in the Canis Majoris (CMa R1) cloud, which distance is ∼ 1kpc. We studied the properties of 61 X-ray sources detected in the PSPC image. Most of them have optical and infrared counterparts, which probably are low- and intermediate-mass young stars, as indicated by the strong correlation of the X-ray emission and other stellar characteristics. CMa R1 shows a ring emission nebula that coincides with an expanding H I region, suggesting that a supernova remnant (SNR) could be responsible for inducing the star formation in this region. The ROSAT image shows an extended feature that could be due either to an unresolved stellar cluster (the PSF is very degraded towards the edge of the PSPC field) or truly diffuse emission, which would correspond to the optical diffuse feature considered a possible old SNR. The nature of the extended features in CMa R1 remains an open question that will be investigated using recently approved observations with XMM-Newton.

Introduction This work is the follow-up of a study based on ROSAT data of three giant molecular clouds. Previously, we used the X-ray emission to search for unidentified young stars in the molecular clouds Monoceros and Rosette (Gregorio-Hetem et al. 1998). The ROSAT images provided 193 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 193-198. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

194

OPEN ISSUES IN LOCAL STAR FORMATION

information about the presence of low- and intermediate-mass young stars. The nature of the sources was verified on the basis of the strong correlation revealed by the X-ray emission and the optical and infrared data, which is found to be similar to the correlation found by Feigelson et al. (1993) and Casanova et al. (1995) for the Chamaeleon I and ρ Ophiuchi clouds, respectively. The moderately distant cloud CMa R1 (d ∼ 1 kpc) is an interesting star forming region, which is related to CO molecular cores; emission, reflection and dark nebulae; and several Hα emission line stars (Shevchenko et al. 1999, Tjin et al. 2001, Pereira et al. 2001). CMa R1 also contains OB stars associated with an optical ring emission nebula coincident with an expanding HI shell. The whole structure could be a supernova remnant inducing the star formation in this region, or alternatively a fossil HII region (Herbst & Assousa 1977). The ROSAT PSPC 2o image was obtained with an exposure of 20 ksec and pointed towards α = 07h 01m , δ = -11o 24’, the J2000 coordinates of the center of the CMa R1 field. Due to a strong background emission of instrumental and cosmic origin in the spectral range 0.1-0.4 keV (see Gregorio-Hetem et al. 1998), the sources were analyzed in the 0.4-2.4 keV range. The PSPC image of the CMa cloud is displayed in the Appendix of this paper (see the CDROM of these Proceedings: Figure A1), as well as the source list derived from the ROSAT image analysis (Table A1). Sixty-one X-ray sources were detected, 48 of them having S/N > 3.5, with X-ray luminosities in the range of 6 1030 erg/s. to 8 1032 erg/s. Following Gregorio-Hetem et al. (1998) the approximate X- ray luminosities ˜ X ) were derived by using the correspondence between count-rate and (L X-ray flux given by 1 cnt/ks  9x10−15 erg s−1 cm−2 . The assumed visual extinction, temperature and distance are respectively AV =1 mag, kTX =1 keV, and d ∼ 1100 pc for all the sources. According the J-H vs. H-K diagram obtained for most of these sources from the 2MASS catalog, which is shown in Figure 1, it is found that they generally do not suffer a high extinction, so that essentially no correction is needed to determine the X-ray luminosities.

1.

X-ray emission vs. optical and IR data

Optical counterparts were identified in a POSS(R) plate digitized by MAMA1 and R magnitudes could be estimated. The distribution of the sources in a diagram of MR (absolute magnitude) as a function of X-ray

1 Machine

Automatique ` a Mesurer pour l’Astronomie, Observatoire de Paris

Star Formation in Canis Majoris (Gregorio-Hetem et al.)

Figure 1. The distribution of the near-IR counterparts of the X-ray sources detected in CMa R1 (open squares) in the diagram J-H vs. HK. The full lines represent the locus of the main sequence and red giant branch. Dashed lines are used to show the direction of the interstellar reddening vector. By comparing with the data of the Cha I cloud members (Lawson et al. 1996), represented by “x”, it can be noted that the extinction effect is smaller for the CMa R1 sources.

195

Figure 2. The distribution of the Xray sources detected in CMa R1 (open squares) in the diagram of log(LX ) as a function of de-reddened absolute magnitude (MJo ). The Cha I X-ray data, represented by “x”, were obtained from Feigelson et al. (1993). The full line indicates the fit obtained for part of the Cha I sources, which have well determined LX and Jo. The dashed lines are used to indicate the 2σ deviation found by the linear fit.

˜ X [erg/s]) = 31.1(±0.3) − luminosity shows a correlation given by log(L 0.27 × MR , which is similar to that found for T Tauri stars, which are the main X-ray emitters associated with the Chamaeleon I cloud (Cha I) at d 150 pc. The 2MASS catalog was also searched for near-IR sources inside the position error circle of the ROSAT sources. In several cases we found multiple candidates related to one X-ray source that could correspond to possible embedded clusters. Taking into account that the extinction effect at 1 µm is almost the same at 1keV, the X-ray emission was compared to the J band magnitude in order to verify if these data have the same correlation shown by sources from other young stellar populations. ˜ X ) versus The distribution of the CMa R1 sources in the diagram of log(L MJo (de-reddened absolute magnitude, see Casanova et al. 1995) is presented in Figure 2. The Cha I sources were also plotted for comparison. It can be noted that the CMa R1 sources follow a distribution similar

196

OPEN ISSUES IN LOCAL STAR FORMATION

to that of the Cha I sources, for which the data were fitted by a linear ˜ X [erg/s]) = 31.0(±0.4) − 0.33 × MJo . regression represented by log (L Using the same method as Carpenter et al. (1997), we present in Figure 3 the MJ vs J-H diagram to compare our candidates with the locus of low-mass young stars. This locus is indicated by the 1 Myr isochrone and the 0.1 and 2.5 M
tracks (D’Antona & Mazzitelli, 1994). The distribution of the infrared counterparts of the CMa R1 X-ray sources suggests a sample composed mainly of low-mass and intermediate-mass PMS objects. Figure 3 Diagram of absolute J magnitude as a function of the J-H color (both corrected for extinction). The CMa R1 data are represented by squares, while the Cha I data are represented by “x”. The Zero Age Main Sequence (ZAMS) is indicated by the full line. Dashed lines are used to show the region of low-mass young stars defined by the 0.1 and 2.5 M
evolutionary tracks and the 1 Myr isochrone.

2.

The stellar content of CMa R1

We found at least one optical and/or infrared counterpart in 70% of the 61 ROSAT sources. We were able to verify in the MJ vs. J-H diagram the position of 43 near-IR counterparts and our conclusion is that 57% of them have colours similar to low-mass young stars, 37% are similar to intermediate-mass young objects, and 8% are massive stars. The stellar content of CMa R1 could be better studied mainly in the central region of the ROSAT field, where the point-like sources are better resolved. In the upper right part of the diagram in Figure 2 it can be noted some X-rays sources for which the LX is higher than the expected, as compared with the J magnitude of the near-IR counterpart. In these cases, other 2MASS candidates were found to be present within the error circle at the position of the corresponding ROSAT source, which could be considered as possible multiple counterparts not resolved in

Star Formation in Canis Majoris (Gregorio-Hetem et al.)

197

the ROSAT image. It can be concluded that many faint X-ray sources are missing in our study and more sensitive observations are needed to reveal the entire young population of CMa R1.

3.

The problem of the extended features

The eastern part of the CMa R1 cloud shows a feature extended in X-rays, whose nature remains unclear: it may be an unresolved stellar cluster, since the PSF of the PSPC is very degraded towards the edge of the field; or a truly diffuse emission, since it is in the same position of the optical diffuse feature, thought to be an old supernova remnant or a fossil HII region (Herbst & Assousa 1977). The fact that HII regions are not in general X-ray emitters (but see below) tends to support a priori the SNR interpretation. The ROSAT observation of Monoceros and Rosette regions, previously studied by us, indicated the presence of dozens of low- and intermediatemass young stars associated with the clouds. The census of the stellar population of these clouds has been much improved recently, thanks to the subarcsec spatial resolution of the new generation of X-ray telescopes, Chandra and XMM-Newton. Kohno, Koyama & Hamaguchi (2002) obtained Chandra observations covering the 3.2’ x 3.2’ central region of Mon R2 where 154 sources were detected, 85% of them having near-IR counterparts with characteristics of ZAMS or PMS stars. Townsley et al. (2003) identified more than 300 X-ray sources belonging to the Rosette cloud using Chandra data, which also revealed the existence of faint extended X-ray emission: perhaps as much as 80% of the extended emission may be attributed to unresolved faint, low-mass stars, but the remainder is very likely related to stellar wind shocks from early O stars, as is the case in M17 (see discussion in Townsley et al. 2003). So unresolved emission of a large number of faint low-mass stars is a real possibility, in addition to a SNR or a fossil HII region, to explain the extended emission seen by ROSAT. Based on the recent results for these distant clouds we are expecting an improvement in our knowledge of the X-ray cluster discovered by ROSAT in the CMaR1 cloud. Such an advancement could be achieved through spatial and spectral studies conducted with better statistics and spectral resolution provided by the new generation of X-ray observatories. Recently approved observations of the CMaR1 region with XMM-Newton (PI: J. Gregorio-Hetem) will help solve the mystery of its extended X-ray emission.

198

OPEN ISSUES IN LOCAL STAR FORMATION

Acknowledgments This work was partly funded by FAPESP (grant number 2001/090182). E.M acknowledges support from CNPq.

References Carpenter, J.M., Meyer, M.A., Dougados, C., Strom, S.E., Hillenbrand, L.A. 1997, ApJ, 114, 198 Casanova, S., Montmerle, T., Feigelson, E.D., Andr´e, P. 1995, ApJ, 439,752 D’Antona, F., Mazzitelli, I. 1994 ApJS, 90, 467 Feigelson, E.D., Casanova, S., Montmerle, T., Guibert, J. 1993, ApJ, 416, 623 Gregorio-Hetem, J., Montmerle, T., Feigelson, E.D., Casanova, S. 1998, A&A, 331, 193 Herbst, W. & Assousa, G.E. 1977, ApJ, 217, 473 Kohno, M., Koyama, K., & Hamaguchi, K. 2002, ApJ, 567, 423 (Erratum in ApJ, 580, 626; 2002) Lawson, W.A., Feigelson, E.D., Huenemoerder, D.P., 1996, MNRAS 280, 1071 Pereira, C.B., Schiavon, R.P., de Ara´ ujo, F.X., Landaberry, S.J.C., 2001, AJ, 121, 1071 Shevchenko, V.S., et al. 1999, MNRAS, 310, 210 Tjin A Djie, H. R. E., et al. 2001, MNRAS, 325, 1441 Townsley, L., Feigelson, E. D., Montmerle, T., Broos, P.S., Chu, Y.-H., Garmire G. P. 2003, ApJ, in press

T. Montmerle, J. Gregorio-Hetem, H. Zinnecker

X-RAY EMISSION FROM NGC 2362 Nuria Hu´elamo European Southern Observatory, Chile [email protected]

Beate Stelzer Osservatorio Astronomico di Palermo, Italy [email protected]

Andr´e Moitinho Observatorio Astronomico de Lisboa, Portugal [email protected]

Jo˜ao F. Alves European Southern Observatory, Germany [email protected]

Charles Lada Harvard-Smithsonian Center for Astrophysics, USA [email protected]

Abstract

We present an X-ray study of the young and massive star cluster NGC 2362. The aim of this work is to analyze the activity level of the pre-main sequence (PMS) late-type population of the cluster and compare it with that from late-type stars with similar ages but located in intermediate and low-mass star forming regions. In this way we could understand if the environment plays an important role in the magnetic activity of PMS late-type stars. This is still an open issue in early stellar evolution.

199 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 199-204. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

200

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction Several works have been devoted to the study of the activity level of PMS late-type stars in different star forming regions (SFR), (e.g. Orion, Taurus, Chamaeleon, NGC 2264). The result of these works is somehow controversial: although they agree on the fact that PMS late-type stars are very active, it remains unclear to date whether the presence of circumstellar disks influences their X-ray emission level (e.g. Stelzer & Neuh¨ auser 2001, Preibisch & Zinnecker 2002, Feigelson et al. 2002, Flaccomio et al. 2003). However, before studying if accretion processes affect the activity level of PMS late-type stars, we must first understand if non-accreting PMS stars with similar ages but born in different star formation environments display the same level of X-ray emission. In other words, it is important to understand if the stellar activity in the PMS phase is unique for a given age or if it is affected by the environment in which the stars are born. In order to shed light on this important problem it is crucial to observe diskless PMS late-type stars with similar ages but born in different environments. An ideal test case for this picture would be to compare, for the same age, a nearby stellar cluster containing high-mass stars with well-known low-mass SFRs of the same age. NGC 2362 is a young cluster containing high-mass stars. It can be considered a template of early stellar evolution given (i) its youth (5 Myr), (ii) its low (AV =0.3 mag) and non variable extinction (Moitinho et al. 2001, MAHL01 hereafter), (iii) the negligible spread in the age of its members, suggesting a single and quick episode of massive star formation (MAHL01), and (iv) the absence of circumstellar disks, which presumably have already been dissipated (Alves et al. 2002). These unique characteristics make NGC 2362 an ideal laboratory to study the X-ray emission of PMS late-type stars at a few million years in a highmass SFR.

1.

ROSAT observations

NGC 2362 was observed twice by ROSAT. The observations were performed with the two main detectors available: the Position Sensitive Proportional Counter (PSPC), which was used either in survey or in pointed mode, and the High Resolution Imager (HRI). The X-ray telescope and the instrumentation onboard the ROSAT satellite have been described in detail by Tr¨ umper (1983), Pfeffermann et al. (1998) and David et el. (1996).

201

X-ray emission from NGC 2362 (Hu´elamo et al.) V−band image

27’

ROSAT PSPC pointing

ROSAT HRI pointing

120’

38’

Figure 1. Optical and X-ray data of NGC 2362. The optical data corresponds to the study presented by MAHL01. The ROSAT data (middle and right panels) shows the PSPC and HRI data analyzed in this work. In each of the X-ray frames we have overplotted the area covered by our optical survey.

The data reduction, source detection and identification, have been carried out using the source detection routines provided by the Extended Scientific Analysis System (EXSAS; Zimmermann et al. 1997). These routines are based on a Maximum Likelihood (ML) technique (Cruddace et al. 1988). The X-ray luminosities have been derived assuming a Raymond-Smith thermal spectrum with a temperature of k Tx = 1 keV (where k is Boltzmann’s constant) and a visual extinction of AV = 0.3 mag.

2.

Data Analysis

A preliminary analysis of the ROSAT X-ray emission of NGC 2362 was presented by Bergh¨ofer & Schmitt (1998). We have re-analyzed the ROSAT X-ray data of NGC 2362 and cross-correlated it with the optical data presented by MAHL01. We have also compared the X-ray properties of NGC 2362 with those from two approximately similar age low-mass SFRs, Chamaeleon I (e.g. Lawson et al. 1997) and Taurus (e.g. Stelzer & Neuh¨auser 2001).

2.1

Optical and X-ray data

The optical and X-ray images of NGC 2362 (both PSPC and HRI) are displayed in Figure 1. It is noticeable the large number of X-ray sources all over the ROSAT field of view (see PSPC frame), suggesting a cluster radius (r) of r ≥ 9 . In spite of the long exposure time of the ROSAT observations (>80ks), the sensitivity is not enough to detect the fainter X-ray emitters of the cluster due to the large distance to the source: 1480 pc (MAHL01).

202

OPEN ISSUES IN LOCAL STAR FORMATION

5

5

10

10

15

15

20

20

0

1

2

3

0

1

2

3

Figure 2. Color-magnitude diagram with stars in the field of NGC 2362. Note the clear separation between the field and the (PMS) cluster members. The left panel shows all the sources within a circle with radius 9 and centered on Tau CMa, the O9-type star in the center of NGC 2362. The right panel shows the same figure but selecting a radius of 5 . The large filled circles represent the X-ray detections from ROSAT data. As seen, most of the X-ray emission arises from the late-type PMS population of the cluster. Within a radius of 5 , most of the X-ray sources show one optical counterpart. In the outer parts of the cluster (radius > 5 ) we still find strong X-ray emitters, but with multiple optical counterparts.

We have cross-correlated optical data from MAHL01 with the positions of the X-ray detections. We have only used ROSAT HRI data because it provides better spatial resolution than the PSPC. The result of the cross-correlation is displayed in Figure 2. As seen, most of the X-ray emitters are identified with the late-type PMS population of the cluster. This identification becomes more uncertain at radius larger than

X-ray emission from NGC 2362 (Hu´elamo et al.)

203

5 , due to the degradation of the point-spread function of the X-ray data at large off-axis distances. The ROSAT HRI observation is only sensitive to the strongest X-ray emitters of the cluster, with Lx (erg/s) > 29.9.

2.2

X-ray luminosity function of NGC 2362

We have derived the X-ray luminosity function (XLF) of NGC 2362, including both detections and upper limits. The X-ray emission of PMS late-type stars is mass-dependent (e.g. Flaccomio et al. 2003, Stelzer et al. 2001), so we have chosen a small mass range (0.8–1.2 M
) to build the XLF. Given that our X-ray observations are not sensitive enough, the XLF of NGC 2362 only shows detections in the range of lg Lx (erg/s) = 30 − 31. The X-ray sources in that range are the most active stars in the cluster. We have compared the XLF of NGC 2362 with that from two ∼5 Myr intermediate- and low-mass SFRs: Taurus and Chamaeleon I. We have included stars without indications of accretion processes, i.e. Weak-lined T Tauri stars (WTTSs). The result is displayed in Figure 3. The highluminosity tail of the XLFs, which is populated by the strongest X-ray emitters, is similar for the three SFRs. This preliminary result indicates a weak or negligible influence of the stellar environment on the X-ray properties of PMS late-type stars, at least for the most active stars.

3.

Conclusions

The late-type population of NGC 2362 displays a high level of magnetic activity according to the ROSAT data. The high X-ray luminosity tail of its XLF is similar to that found in ∼5 Myr old low-mass SFRs: Taurus and Chamaeleon I. This result indicates a weak influence of the stellar environment on the X-ray properties of very active PMS late-type stars. This preliminary result will be re-examined by upcoming Chandra observations of NGC 2362. Given that Chandra will allow us to derive a more complete XLF of the cluster, we could study the shape of the XLF for less active and less massive late-type stars.

Acknowledgments The ROSAT project is supported by the Max-Planck-Gesellschaft and Germany’s federal government (BMBF/DLR).

204

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 3. X-ray luminosity function (XLF) of ∼ 1 M stars in NGC 2362 (filled circles). We have overplotted the XLF’s of two star forming regions with approximately similar ages but different environmental conditions: Taurus and Chamaeleon I. Note that we have only selected diskless stars (WTTS) with masses between 0.8-1.2 M
for this comparison. As seen, the three stellar groups display a very similar XLF in the overlapping region between lg Lx (erg/s) = 30 − 31. This result suggests that the environment does not influence the X-ray properties of very active PMS late-type stars.

References Alves J.F, Lada C.J., Lada E., et al., 2003, A&A submitted Bergh¨ ofer, T.W. & Schmitt J.H.M.M., 1998, Cool Stars, Stellar Systems and the Sun, ASP conferences series, Vol. 154, eds. R. A. Donahue and J. A. Bookbinder Cruddace, R.G., Hasinger,G.R., Schmitt, J.H.M.M., 1988, The application of a maximum likelihood analysis to detection of sources in the ROSAT data. In: Murtagh, F. and Heck, A. (eds), ESO Conf. and Workshop Proc. 28, Astronomy from large databases, Garching, p.177 David L.P., Harnden F.R., Kearns K.E. et al. 1996, The ROSAT High Resolution Imager calibration report, SAO Technical Report, p.6. Feigelson E., Broos P., Gaffney J.A. et al., 2002, ApJ 574, 258 Flaccomio E., Micela G. and Sciortino S., 2003, A&A 397, 611 Lawson W.A., Feigelson E.D. and Huenemoerder D.P., 1997, MNRAS 280, 1071 Moitinho, A., Alves, J., Hu´elamo, N. and Lada, C.J., 2001, ApJ 563, L73 Pfefferman E., Briel U., Hippman H. et al. 1988, Proc. SPIE 733, 519 Preibisch T. & Zinnecker H. 2002, AJ 122, 866 Stelzer B. & Neuh¨ auser R., 2001, A&A 377, 538 Tr¨ umper, J., 1983, Adv. Space Res. Vol.2, N 4, 241 Zimmermann, H.U. et al., 1997, EXSAS User’s Guide, MPE, Garching

ACCRETION AND EJECTION Lee Hartmann Smithsonian Astrophysical Observatory, USA [email protected]

Abstract

1.

The processes by which stars stars accrete their mass are not fully understood. Accretion in early stellar evolution appears to be highly timevariable. Since accretion and mass ejection are closely connected – the energy for winds and jets comes from accretion – this means that mass loss is also highly time-variable. Understanding the origin of this variable accretion is an important outstanding issue in star formation that has not received sufficient attention.

The luminosity problem The internal energy of a completely convective star is E∗ = −(3/7)GM∗2 /R∗ ,

(1)

where M∗ and R∗ are the stellar mass and radius, respectively. Since the original protostellar cloud has a much larger radius than that of the final star, its original energy is effectively zero. Thus the negative energy of equation (1) must be the energy that is released during star formation; in particular, most if not essentially all, of this energy must be radiated away for the star to form. Putting in typical values for a T Tauri star, M∗ = 0.5M
, R∗ = 2R
, the expected energy release for forming a typical low-mass star is roughly E∗ = 17 L
× 105 yr.

(2)

I write the energy in this form to illustrate that the observable, the accretion luminosity, expected as a result of star formation depends upon the timescale for accretion. The standard theory for collapse of quiescent molecular cloud cores into low-mass stars predicts infall timescales of roughly 105 yr (Adams, Lada, & Shu 1987); this prediction is consistent with the statistics of protostellar sources, for example, in the Taurus molecular cloud complex 205 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 205-212. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

206

OPEN ISSUES IN LOCAL STAR FORMATION

(Kenyon, Calvet, & Hartmann 1993). Similarly, the prediction of the theory for the accretion rate agrees with infall rates estimated from dusty envelope models (Kenyon et al. 1993); at the estimated rates of ∼ 4×10−6 M
yr−1 , enough stellar mass can be built up during the infall timescale.

Figure 1. Luminosity distributions of young stellar objects in Taurus. After Kenyon et al. (1993).

The problem is that observed protostellar luminosities in Taurus are not compatible with equation (1). Figure 1 shows the observed distribution of luminosities of Class I objects (protostars) along with Class II and III systems (T Tauri stars with and without disks, respectively). One observes that the Class I objects have essentially the same luminosity distributions as the slowly-accreting or non-accreting stars, with only one object (L1551 IRS5) having a luminosity comparable to the predicted value. One might try to avoid the accretion luminosity problem by invoking protostars of much larger radii or several times smaller masses. However, such a resolution seems unlikely. For one thing, making the stellar radius much larger makes the stellar luminosity much larger; already the Class I luminosities are comparable to the T Tauri stellar luminosities (Figure 1). Making the mass smaller does not help if one wishes to form stars of similar masses to those observed for T Tauri stars. The “birthline” theory of Stahler (1988) predicts that low-mass protostars do not have

Accretion and Ejection (Hartmann)

207

large radii; furthermore, this theory predicts that protostellar properties should scale roughly as R∗ /M∗ (birthline) ≈ 6.4R
/M
.

(3)

With this constraint, the accretion luminosity becomes ˙ −5 L
, Lacc (birthline) ≈ 50 M

(4)

˙ −5 is the mass accretion rate in units of 10−5 M
yr−1 . The where M resulting predicted accretion luminosity for the inferred typical Taurus ˙ −5 = 0.4 is ∼ 20L
. Recent estimates of Class I stellar infall rate of M properties by Kenyon et al. (1998) and Greene & Lada (2002) indicate reasonable consistency with birthline R∗ /M∗ estimates. Finally, more direct evidence for low accretion rates can be found from the work of Muzerolle et al. (1998), who applied an empirical correlation of Brγ emission line luminosity and accretion luminosity in T Tauri stars, and found low Class I accretion rates. An alternative might be to suppose that most of the stellar mass is accreted during the Class 0 protostellar phase, not the Class I stage. However, the length of the Class 0 phase seems to be an order of magnitude shorter than the Class I lifetime (Andre, Ward-Thompson, & Barsony 2000), so that the mass accretion rates need to be an order of magnitude larger to accrete the same stellar mass. Typical Class 0 luminosities are of order 10L
in nearby regions (Andre et al. 2000), i.e. only about an order of magnitude larger than the typical Taurus Class I luminosity, so this does not solve the luminosity problem. Moreover, even if accretion rates are higher in Class 0 sources, this would still not eliminate the evidence for higher infall rates than accretion rates in Class I sources.

2.

Infall vs. disk accretion

The direct implication of the foregoing is that infall from protostellar envelopes to disks generally exceeds the accretion rate from the disk onto the central (proto)star. Such a discrepancy is easy to understand, at least in principle; infall to the disk occurs roughly at free-fall velocities, and does not require angular momentum transfer, while the accretion through the disk onto the star demands substantial angular momentum transfer. Material cannot pile up in the disk forever, of course; the mass has to be accreted eventually. If the piled-up mass is released in a short burst, as suggested by Kenyon et al. (1991), then the luminosity problem can be avoided because most of the time, the system is usually observed in the low (accretion rate) state.

208

Log Mass Accretion Rate (solar masses/yr)

OPEN ISSUES IN LOCAL STAR FORMATION

T Tauri star

Protostar

FU Ori outburst

10-4

EXor outburst?

10-5 10-6

T Tauri accretion

10-7 10-8

Disk accretion

Infalling Envelope

105

106

107

Age (yr)

Figure 2. Schematic of the suggested evolution of infall rates to disks and disk accretion rates onto central stars.

Figure 2 indicates this scenario of accretion. The rapid accretion states are identified with FU Orionis outbursts (Hartmann & Kenyon 1996, and references therein). FU Ori outbursts involve rapid increases in the accretion rate up to 10−4 M
yr−1 , in which the luminosity of the accretion disk is far greater than that of the central star. FU Ori objects are either surrounded by nebulosity and/or are highly extincted, and frequently show substantial far-infrared excess emisson, consistent with the proposal that they represent outbursts of accretion during the protostellar phase (Kenyon et al. 1991). The proposal that low-mass stars accrete most of their mass during short-lived, FU Ori-like outbursts of accretion seems to me to be the most likely resolution of the luminosity problem. However, this solution is not without its difficulties. Currently, we know of about a dozen probable FU Ori objects (including objects with the right spectroscopic characteristics, acknowledging that we likely have missed outbursts of heavily-extincted systems). If the accretion rates of these systems are ∼ 10−4 M
yr−1 , then we can account for a total star formation rate of roughly 10−3 M
yr−1 . Now, within 1 Kpc of the Sun, which encompasses the known FU Ori objects, the total star formation rate is estimated to be of order 10−2 M
yr−1 , i.e. roughly an order of magnitude larger than can be accounted for by known Fuors. Putting it another way,

Accretion and Ejection (Hartmann)

209

even though event statistics are such that FU Ori outbursts must be recurrent in at least some systems, we can account for only ∼ 10% of the mass of low-mass stars forming in such outbursts. So where might the missing rapid accretors/FU Ori-like systems be? I think that the problem is probably that the missing objects are too heavily extincted to be observable at short enough wavelengths. Assuming free-fall, spherical infall to a radius rin , the extinction at 2.2µm is ˙ −5 M−1/2 (rin /AU)−1/2 , AK ∼ 30M (5) 1/2 where M1/2 is the central mass in units of 0.5M
. Although this equation is for spherical collapse, it is roughly appropriate for rotating collapse if rin is identified with the “centrifugal radius”, i.e. the scale at which the infalling matter’s angular momentum becomes important. (e.g., Hartmann 1998, p. x). It is clear that during the initial phases of collapse, for ˙ −5 > which infall rates are likely to be high, of order M ∼ 1, the protostellar object will be effectively invisible at 2.2 µm. At longer wavelengths, dust emission from the envelope is likely to dominate the spectrum. Hence it will be very difficult to identify the nature of the central protostellar object from near-infrared spectroscopy. FU Ori-like, rapidly-accreting systems might be misidentified as higher-mass, more luminous stars, in the absence of any spectral information on the protostellar photosphere. One might still be able to detect the central object in favorable inclinations, i.e. viewing along a hole in the envelope driven by the object’s outflow. V1057 Cyg and V1515 Cyg probably are specific examples of this fortuitous situation (Goodrich 1987). In this regard, it is encouraging that Reipurth & Aspin (1997) and Sandell & Aspin (1998) have found a few more FUor-like candidates in heavily extincted systems from near-infrared spectroscopy.

3.

Why outbursts?

The magnetorotational instability (MRI) is the favored mechanism to produce the required anomalous viscosity needed for disk accretion (Balbus & Hawley 1998, and references therein). Application of the MRI is uncertain for protostellar disks, however, because such disks are too cold to produce enough ions thermally to couple the magnetic field effectively to the disk gas. This led Gammie (1996) to suppose that protostellar disks might generally accrete only through surface layers which can be ionized by cosmic rays (or by X-rays; Glassgold et al. 1997). The limited column densities that can be externally ionized result in a bottleneck for accretion, ensuring that a disk being fed by infall at typical rates will simply pile up mass. This situation cannot go on

210

OPEN ISSUES IN LOCAL STAR FORMATION

indefinitely, and Gammie (1996) suggested that eventually gravitational instabilities would arise to cause accretion in bursts. Armitage, Livio, & Pringle (2001) calculated simple one-dimensional disk models with parameterized gravitational torques in terms of an effective viscosity, and found that indeed most of the mass gets accreted in outbursts. The Armitage et al. models do not exhibit the extreme variations in mass accretion rates seen in FU Ori variables. The likely mechanism for this is thermal instability in the inner disk. Bell & Lin (1994) showed that the same type of limit cycle invoked for cataclysmic variables can be used to explain FU Ori outbursts. The attractiveness of this mechanism is that it naturally predicts maximum temperatures during outburst that are similar to those observed, as well as explaining the rapid rise times (see Bell & Lin 1994 for details). Charles Gammie and I have computed modified versions of the layered disk models, similar to those of Armitage et al. , except that they do undergo thermal instabilities in inner regions, providing a much better comparison with observations. More work needs to be done in this area, as the time-dependence of FU Ori outbursts may hold important clues to the physics of protostellar accretion disks.

4.

Accretion-powered outflows

Figure 3 demonstrates the clear relationship between mass loss and mass accretion. More than this, it shows that outflows are powered by accretion. The energy loss in the wind, neglecting the work to escape 2 , where M ˙ w is the mass loss rate ˙ w v∞ from the potential well, is 0.5M in the wind and v∞ is the terminal velocity of the wind. The accretion ˙ 2 , where vK is the Keplerian velocity energy release can be written as Mv K in the inner disk. Thus, if v∞ is of order a few times vK , as seems to be ˙ implies that a very large fraction of the ˙ w ∼ 10−1 M the case typically, M accretion energy is carried away by the outflow. In the case of FU Ori systems, it is clear that the only source of energy large enough to drive the wind is the disk accretion. If accretion is highly time-variable, it follows that outflows must also be highly variable. Moreover, the mass and momentum injected into the surroundings must be dominated by phases of rapid accretion. This suggests that to understand outflows we need to understand mass loss from FU Ori systems. Most studies have concentrated on outflows from T Tauri stars and/or low-luminosity systems, simply because these objects are much more common and thus there are many closer systems to analyze. However, this understandable bias should not be an excuse to ignore the very real differences between systems.

Accretion and Ejection (Hartmann)

211

Figure 3. Relation between mass accretion rate and mass loss rate for both T Tauri stars and FU Ori objects. From Calvet (1998).

Certain popular models of outflows, such as the X-wind theory (Shu et al. 1994; Shang et al. 2002) incorporate truncation of the disk by a stellar magnetosphere as an important, if not essential, component. In the specific case of X-wind models, the outflow is supposed to arise entirely from the truncated edge of the disk. However, there is no evidence for magnetospheric truncation of FU Ori disks (Kenyon et al. 1989), which is probably the result of the high accretion ram pressure crushing the magnetosphere back against the star. Indeed, detailed analysis of the line profiles of FU Ori itself indicates that the wind must come out of a finite area of the disk, not simply a narrow annulus (Calvet et al. 1993; Hartmann & Calvet 1995). This means that the magnetospheric truncation cannot be an essential feature of accretion disk mass ejection, especially during rapid accretion which should dominate the overall mass and energy injection. If mass ejection is powered by accretion, and accretion itself is produced by a turbulent magnetic viscosity (the MRI), then one would surely expect that the mass loss will be time dependent at differing locations of the disk, due to both changes in the magnetic field topology and local energy fluxes which might assist acceleration. Indeed, MHD turbulence naturally generated in the disk may provide energy fluxes which heat the wind (Miller & Stone 2000; Shang et al. 2002). The re-

212

OPEN ISSUES IN LOCAL STAR FORMATION

sult would be short-timescale fluctuations of mass loss, both in velocity and in mass flux, which are seen quite clearly in the Na I D line profiles of the FU Ori objects (e.g., Reipurth et al. 2002). Although it is difficult to make a direct quantitative statement from the observations so far, the data clearly do not support simple steady flow models for mass ejection from accretion disks. The strong variability on short timescales (days) is at least qualitatively consistent with what one might expect from a variable magnetic geometry resulting from MRI turbulence.

Acknowledgments I wish to thank the organizers for the invitation to this conference. This work was supported in part by NASA grant NAG5-9670.

References Adams, F. C., Lada, C. J., & Shu, F. H. 1987, ApJ, 312, 788 Andre, P., Ward-Thompson, D., & Barsony, M. 2000, Protostars and Planets IV (eds. A.P. Boss & S.S. Russell (Tucson: University of Arizona Press), 59 Armitage, P. J., Livio, M., & Pringle, J. E. 2001, MNRAS, 324, 705 (ALP01) Bell, K.R., & Lin, D.N.C. 1994, ApJ, 427, 987 Balbus, S. A. & Hawley, J. F. 1998, Reviews of Modern Physics, 70, 1 Calvet, N. 1998, in Accretion Processes in Astrophysical Systems, eds. S.S. Holt & T.R. Kallman, AIP Conf. Proc. 431, 495 Calvet, N., Hartmann, L., & Kenyon, S. J. 1993, ApJ, 402, 623 Gammie, C.F. 1996, ApJ, 457, 355 (G96) Glassgold, A. E., Najita, J., & Igea, J. 1997, ApJ, 480, 344 Goodrich, R. W. 1987, Pub. Astr. Soc. Pac., 99, 116 Greene, T. P. & Lada, C. J. 2002, AJ, 124, 2185 Hartmann, L. 1998, Accretion Processes in Star Formation. Cambridge University Press Hartmann, L. & Calvet, N. 1995, AJ, 109, 1846 Hartmann, L. & Kenyon, S. J. 1996, ARA&A, 34, 207 Kenyon, S. J., Brown, D. I., Tout, C. A., & Berlind, P. 1998, AJ, 115, 2491 Kenyon, S. J., Calvet, N., & Hartmann, L. 1993, ApJ, 414, 676 Kenyon, S. J. & Hartmann, L. W. 1991, ApJ, 383, 664 Kenyon, S. J., Hartmann, L., Imhoff, C. L., & Cassatella, A. 1989, ApJ, 344, 925 Miller, K. A. & Stone, J. M. 2000, ApJ, 534, 398 Muzerolle, J., Hartmann, L., & Calvet, N. 1998, AJ, 116, 2965 Reipurth, B. & Aspin, C. 1997, AJ, 114, 2700 Reipurth, B., Hartmann, L., Kenyon, S. J., Smette, A., & Bouchet, P. 2002, AJ, 124, 2194 Sandell, G. & Aspin, C. 1998, A&Ap, 333, 1016 Shang, H., Glassgold, A. E., Shu, F. H., & Lizano, S. 2002, ApJ, 564, 853 Shu, F., Najita, J., Ostriker, E., Wilkin, F., Ruden, S., & Lizano, S. 1994, ApJ, 429, 781 Stahler, S.W. 1988, ApJ, 332, 804

THE ORIGIN OF JETS IN YOUNG STARS MHD disk wind models confronted to observations Catherine Dougados, Jonathan Ferreira, Nicolas Pesenti Laboratoire d’Astrophysique de Grenoble, France [email protected]

Sylvie Cabrit Observatoire de Paris, France

Paulo Garcia, Darren O’Brien University of Porto, Portugal

Abstract

We discuss in this contribution constraints on the origin of mass-loss from young stars brought by recent observations at high angular resolution (0.1 = 14 AU) of the inner regions of winds from T Tauri stars. Jet widths and collimation scales, the large extent of the velocity profile as well as the detection of rotation signatures agree with predictions from magneto-centrifugal disk wind ejection models. However dynamically cold disk wind solutions predict too large terminal velocities and too low jet densities and ionisation fractions, suggesting that thermal gradients (originating in an accretion heated disk corona for example) may play an important role in accelerating the flow.

Introduction The physical mechanism by which mass is ejected from young stars and collimated into jets remains a fundamental open issue in star formation theory. The strong correlation between ejection and accretion found in PMS stars, with a mass flux ratio as high as 0.1 (Cabrit et al. 1990; Hartigan et al. 1995) has favored accretion-driven magnetohydrodynamic (MHD) wind models. However, it is not yet established 213 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 213-222. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

214

OPEN ISSUES IN LOCAL STAR FORMATION

˚ + continuum narrow-band image of the DG Tau Figure 1. Deconvolved [O i] 6300 A microjet obtained with the adaptive optics system PUEO at CFHT. Achieved spatial resolution is 0.1 = 14 AU at the Taurus distance. Adapted from Dougados et al. (2002).

whether the jet originates from the stellar surface (Sauty & Tsinganos 1994) the magnetosphere/disk interface (Shu et al. 1995), or a wide range in disk radii (“disk winds”: Ferreira 1997, hereafter F97); and whether it is launched mostly by magneto-centrifugal forces or by a strong thermal pressure gradient in an accretion-heated corona. Distinguishing between these various scenarii is crucial not only for the jet phenomenon in itself, but also for models of exoplanet formation, as they have distinct implications on the internal structure, angular momentum transfer, and heating/irradiation processes in the inner regions of protoplanetary disks. The small scale jets detected in the vicinity of optically revealed T Tauri stars (see Fig. 1) offer a unique opportunity to test current ejection theories: they give access to the innermost regions of the wind (≤ 100 AU) where models predict that most of the collimation and acceleration processes occur; in addition, the stellar and accretion disk properties are well characterized in these systems. We summarize in this contribution constraints brought by recent high-angular resolution studies of a few prototypical T Tauri jets, conducted either from the ground with adaptive optics techniques or from space with HST. Most studies have concentrated so far on two of the highest accretion sources in Taurus: DG Tau and RW Aur. Observations are compared to the detailed predictions for the class of disk wind models developed by F97.

1.

MHD Disk winds: models and predictions

Two classes of stationary magnetic wind models currently reproduce the observed correlation between ejection and accretion rates: the selfsimilar disk winds and the X-wind solution. These two models represent

The origin of jets in young stars (Dougados et al.)

215

two extreme situations in the range of disk radii involved in the ejection process. In the X-wind solution of F. Shu & collaborators, mass-loss originates from a single point in the disk (the inner radius), which leads to an intrinsically non self-similar behaviour. We concentrate here on the extended disk wind solutions for which detailed observational predictions are currently available. The disk wind solutions computed by F97 extend the original treatment of Blandford & Payne (1982). Both assume that (1) matter is ejected along a large scale bipolar magnetic field threading the accretion disk, (2) jet enthalpy is negligible for accelerating the flow (the wind is ”cold”), and (3) the structure is steady-state, axisymmetric, and self-similar with disk radius. The solutions of F97 describe selfconsistently the accretion flow in the underlying resistive keplerian disk, especially the wind mass-loading at the slow-point (see Ferreira & Pelletier 1995). This global description imposes additional constraints on the self-similar structure, which is then specified by only three dimensionless parameters: (1) ξ ≡ dlogM˙ acc /dlogr, the ejection efficiency parameter, (2)  = h/r0 , where h is the disk scale height at the cylindrical launching radius r0 , and (3) αm = νm /VA h, where νm is the required turbulent magnetic diffusivity and VA is the Alfv´en velocity on the disk midplane. With  = 0.1 (as estimated in HH 30 by Burrows et al. 1996) and αm = 1 (high magnetic torque), solutions that extend far from the Alfv´en surface are then found for ξ between 0.005 and 0.012 (F97). The parameter ξ is related to the total ejection/accretion ratio by M˙ ej /M˙ acc = ξ × ln(Re /Ri), where Re and Ri are the outer and inner cylindrical radii of the MHD disk-wind structure. The thermal and ionization structure for this class of magnetic winds has been computed a posteriori, considering two main heating mechanisms: ambipolar diffusion (Garcia et al. 2001a), and extra mechanical energy deposition due to the dissipation of shocks and/or Alfv´en waves and parametrized as α × ρV 3 /R (O’Brien et al 2003). Computation is performed along streamlines anchored in the disk at radii ranging from R0 = 0.07 AU, the typical disk corotation radius for a T Tauri star, to 1 AU, where molecules should start to form. In the following, a central stellar mass of 0.5 M
and accretion rates M˙ acc = 10−5 − 10−6 M
yr−1 , typical of the sources investigated here, will be used. The computed emissivity grids in forbidden lines are projected onto the plane of the sky, to construct emission maps and long slit spectra. Synthetic maps are then convolved by a two-dimensional gaussian beam, representative of the instrumental spatial and spectral resolutions.

216

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. Jet widths and collimation scales: jet FWHM (symbols and dotted lines) derived from CFHT/PUEO and HST/STIS observations (see text for more details) are compared with predictions from cold disk wind models (solid lines) with varying ξ values: model A (ξ=0.01), B (ξ=0.007), C (ξ=0.005). Models are convolved with a 14 AU (FWHM) gaussian beam. Adapted from Garcia et al. (2001b).

2.

Jet widths and collimation scales

Shown in Fig 2 are transverse FWHM of the optical line emission for the 5 microjets for which we currently have constraints on scales ≤ 200 AU from the central driving source (HL Tau, HH 30: Ray et al 1996; DG Tau, CW Tau, RW Aur: Dougados et al. 2000, Woitas et al. 2002). At the highest spatial resolution achievable today (0.1 =14 AU), the inner collimation region is still unresolved. Jet widths are resolved beyond 30 AU with typical FWHM of 20-40 AU at zproj = 100 AU then slowly increase with opening angles ≤ 5◦ . The self-similar disk wind solutions investigated here are characterized by a wide opening of the field lines beyond the Alfv´en surface (located at z/r  1). The maximum expansion factor R∞ /R0 increases from 50 to 2000 when ξ decreases from 0.01 to 0.005 (F97). The strong variation of the collimation properties of the model with the ejection efficiency parameter is illustrated in Fig 2 where we plot predicted jet widths for 3 cold disk wind solutions with varying ξ values. Both observed collimation scales and jet widths clearly favor solutions with moderate to high ejection efficiency: ξ ≥ 0.007, i.e. M˙ ej /M˙ acc ≥ 2 × 10−2 , in agreement with observations (Hartigan et al. 1995).

3.

Kinematics

The kinematical properties of the jet put additional strong constraints on the launching mechanism. Forbidden line emission in T Tauri stars is

The origin of jets in young stars (Dougados et al.)

217

Figure 3. Position-velocity diagrams along the DG Tau jet: Left Observations in the [Fe ii] 1.257 µm line by Pyo et al. (2003) of the central 200 AU of the microjet, Right Cold disk wind model predictions (i=45◦ , ξ=0.01, M˙ acc = 10−6 M
yr−1 ) on the same spatial scale convolved with a 22 AU × 30 km s−1 beam. Adapted from Pesenti et al. (2003b).

characterized by a large velocity distribution close to the central source, with emission ranging typically from 0 to –400 km s−1 with respect to the systemic velocity (Fig. 3). This mostly blueshifted emission is interpreted by the fact that the receding part of the flow is obscured by the circumstellar accretion disk surrounding the central star (with typical outer radius of 100 AU). Two velocity peaks are identified in integrated profiles: a low velocity component (LVC) at  –10 km s−1 , spatially unresolved and centered on the continuum position, a high velocity component with velocities ranging from  –100 to –400 km s−1 , spatially extended along the jet over a few arcseconds. Strong velocity gradients generally occur along the jet within the central 2-3  , accompanied by a decrease of line width. Projected terminal velocities range from –100 km s−1 to – 200 km s−1 . In addition, the degree of collimation tends to increase with flow velocity (Bacciotti et al. 2000). The disk wind models reproduce the large velocity distribution close to the central souce (the range in poloidal velocities naturally arises from the range in initial disk launching radii R0 ) and the decrease of line width with distance, resulting from the gradual collimation of streamlines (Fig. 3). Due to the rapid cylindrical density stratification (see Cabrit et al. 1999), inner streamlines quickly dominate the emission and produce a high-velocity, collimated jet component (the so-called jet optical illusion, already noticed for the X-wind model by Shang et al. 1998). However, two major discrepancies occur between model and observations. Firstly, the disk wind model predicts too little emission at intermediate velocities ( 100 km s−1 ). A more efficient heating mecha-

218

OPEN ISSUES IN LOCAL STAR FORMATION

nism seems required in the outer field lines at the wind base. Secondly, predicted terminal poloidal velocities appear to exceed observed ones by a factor 1.5 to 3 (Dougados et al. 2003). Under the assumption that all available energy is finally converted into jet kinetic power, the asymptotic value of the poloidal velocity, independent of the ejection model, can be written as (F97): √ Vpol = Vkep (R0 ) × 2λ − 3, (1)

where λ  (RA /R0 )2 is the magnetic lever arm (RA is the cylindrical radius at the Alfv´en surface) and Vkep (R0 ) is the keplerian velocity at the disk launching cylindrical radius R0 . Decreasing the terminal poloidal velocities could be achieved by either increasing R0 or decreasing the magnetic lever arm λ. Bacciotti et al. (2002) recently reported the detection of rotation signatures in the DG Tau microjet from STIS/HST observations. Velocity shifts of 6 to 15 km s−1 are detected at a few tens of AU from the jet axis and between 20 and 90 AU above the disk plane, in the component with poloidal velocity  –80 km s−1 . Using eq. (4.1) combined with the conservation of angular momentum above the Alfv´en surface, Bacciotti et al. (2002) constrain λ  10 and R0  2 AU for this flow component. Indeed, the cold F97 disk wind solutions investigated here (with λ  50) predict on similar spatial scales velocity shifts that exceed the observed ones by a factor 2 to 3 (Pesenti et al. 2003a). Therefore, in DG Tau, the intermediate velocity component (at Vpol = 80 km s−1 ) almost certainly originates in the outer field lines of an extended disk wind component with moderate magnetic lever arm.

4.

Excitation conditions and jet densities

Constraints on the excitation conditions (ne , Te , xe = ne /nH ), independently of the heating process, can be derived from the analysis of ratios involving the prominent optical forbidden emission lines, using the inversion method developped by Bacciotti & Eisl¨offel (1999). The main assumptions of this method are: optically thin emission, homogeneous emitting region (no strong gradients present), and ionisation equilibrium with H satisfied. This method has been successfully applied to the study of the inner regions of the DG Tau and RW Aur microjets (Lavalley et al. 2000; Bacciotti 2002; Dougados et al. 2002, 2003). In both cases, strong excitation gradients are observed in the central regions: ne decrease from > 2 × 104 cm−3 to a few 103 cm−3 and xe increase from a few 10−2 to a few 10−1 over the central 200 AU. Higher electronic densities and ionisation fractions are observed at larger flow velocity (Lavalley-Fouquet et al. 2000; Bacciotti 2002).

The origin of jets in young stars (Dougados et al.)

219

This density and ionisation stratification with flow velocity is indeed expected in disk wind models where the faster streamlines originate from the inner, denser regions of the disk. However, cold disk winds heated by ambipolar diffusion alone fail to account for the observed ionisation fractions. Line ratios in the inner regions of the DG Tau and RW Aur jets appear best reproduced by planar shock models with moderate shock velocities (Lavalley-Fouquet et al. 2000; Dougados et al. 2002). Internal shocks due to time variability in the ejection process therefore seem to play a dominant role in the excitation process. Derived total jet densities range from a few 105 to > 106 cm−3 at projected distances ≤ 100 AU. These values exceed cold disk wind model predictions for a typical streamline launched at R0 =0.15 AU by one order of magnitude (proper comparison awaits taking into account projection and beam convolution effects). Cold disk wind solutions also fail to reproduce integrated line fluxes in a large sample of cTTS, even with very high ionisation levels inconsistent with observations (O’Brien et al. 2003). Higher density wind models appear required to solve this discrepancy.

5.

Concluding remarks

High spatial resolution studies of the inner regions of the wind in T Tauri stars, compared with detailed predictions from stationary MHD wind models, brought fundamental new constraints to ejection models. Observed jet widths, as well as derived upper limits for both the collimation and acceleration scales (≤ 14 AU), are consistent with predictions from cold disk wind solutions with moderate to high ejection efficiency. The onion-like structure of the flow (with faster, denser material concentrated along the axis), the distribution of the velocity profile close to the central source as well as the detection of rotation in the DG Tau microjet also strongly support magneto-centrifugal disk ejection scenarii. However, cold disk wind solutions, with high magnetic lever arm (λ  50), predict too large poloidal and azimuthal velocities (by a factor 1.5 to 3) and too low integrated fluxes on scales > 30 AU. An additional heating source at the base of the wind (e.g. an accretionheated disk corona) would help enhance the mass-loading efficiency on the field lines and reduce the magnetic lever arm, allowing both to decrease terminal wind velocities and to increase jet densities. These warm disk wind solutions, taking into account the effects of thermal gradients, have been investigated by Casse & Ferreira (2000). Computation of their thermal structure, allowing detailed comparison with observations, is currently under way. The existence of a hot gas component on scales

220

OPEN ISSUES IN LOCAL STAR FORMATION

< 1 AU is suggested by the detection, at the base of the DG Tau jet, of the near-infrared HeI 1.08 µm line (Takami et al. 2002), which requires excitation temperatures ≥ 20 000 K. Blueshifted UV emission lines are also detected in the close vicinity of RY Tau, implying temperatures 4.7 omez de Castro & Verdugo ≤ Log10 (Te) ≤ 5.0 at distances  38 R (G´ 2001). Obviously, the detailed analysis of a larger sample of microjets and comparison with predictions from other classes of wind models, including stellar winds (Sauty et al. 2002) and X-winds (Shang et al 2002), are critically required before deriving general conclusions on the origin of mass-loss in young stars.

References Bacciotti, F. 2002, RMxAC 13, 8 Bacciotti, F., & Eisl¨ offel, J. 1999, A&A 342, 717 Bacciotti, F., Mundt, R., Ray, T. P., Eisl¨offel, J., Solf, J., & Camenzind, M. 2000, ApJ 537, L49 Bacciotti, F., Ray, T., Mundt, R., Eisl¨offel, J., & Solf, J. 2002, ApJ 576, 222 Blandford, R. D., & Payne, D. G. 1982, MNRAS 199, 883 Burrows, C. J., et al. 1996, ApJ 473, 437 Cabrit, S., Edwards, S., Strom, S.E., & Strom, K.M. 1990, ApJ 354, 687 Cabrit, S., Ferreira, J., Raga, A. 1999 A&A 343, L61 Casse, F., & Ferreira, J. 2000, A&A 361, 1178 Dougados, C., Cabrit, S., Lavalley, C., & M´enard, F. 2000, A&A 357, L61 Dougados, C., Cabrit, S., & Lavalley-Fouquet, C. 2002, RMxAC 13, 43 Dougados, C., Cabrit, S., Lopez-Martin, L., Garcia, P., O’Brien, D. 2003, proceedings of JENAM2002, Kluwer Academic publishers, in press Ferreira, J. 1997, A&A 319, 340 (F97) Ferreira, J., & Pelletier, G. 1995, A&A 295, 807 Garcia, P. J. V., Ferreira, J., Cabrit, S., & Binette, L. 2001a, A&A 377, 589, Garcia, P. J. V., Cabrit, S., Ferreira, J., & Binette, L. 2001b, A&A 377, 609 G´ omez de Castro, A. I., & Verdugo, E. 2001, ApJ 548, 976 Hartigan, P., Edwards, S., & Ghandour, L. 1995, ApJ 452, 736 Lavalley-Fouquet, C., Cabrit, S., Dougados, C. 2000, A&A 356, L41 O’Brien, D., Garcia, P., 2003 proceedings of JENAM2002, Kluwer Academic publishers, in press Pesenti, N., Dougados, C., Cabrit, S., O’Brien, D., Garcia, P. J. V., & Ferreira, J. 2003a, proceedings of JENAM2002, Kluwer Academic publishers, in press Pesenti, N., Dougados, C., Cabrit, S., O’Brien, D., Garcia, P. J. V., & Ferreira, J. 2003b, A&A, in press Pyo, T. -S., Kobayashi, N., Hayashi, M., Terada, H., et al. 2003 ApJ, in press Ray, T., Mundt, R., Dyson, J., Falle, S., & Raga, A. 1996, ApJ 468, L103 Sauty, C., & Tsinganos, K. 1994, A&A 287, 893 Sauty, C., Trussoni, E., & Tsinganos, K. 2002, A&A 389, 1068 Shang, H., Shu, F., & Glassgold, A. 1998, ApJ 493, L91 Shang, H., Glassgold, A. E., Shu, F .H., & Lizano, S. 2002, ApJ 564, 853 Shu, F. et al. 1995, ApJ 455, L155

The origin of jets in young stars (Dougados et al.)

221

Takami, M., Chrysostomou, A., Bailey, J., Gledhill, T. M., Tamura, M. & Terada, H. 2002, ApJ 568, L53 Woitas, J., Ray, T. P., Bacciotti, F., Davis, C. J., & Eisl¨offel, J. 2002, ApJ 580, 336

Catherine Dougados

222

OPEN ISSUES IN LOCAL STAR FORMATION

Z. Abraham

L. Hartmann, H. Zinnecker, A. Hetem, J. Gregorio-Hetem, P. Andr´e

NON COEVAL YOUNG MULTIPLE SYSTEMS? On the pairing of protostars and T Tauri stars Gaspard Duchˆene, Andrea Ghez, Caer McCabe UCLA – Department of Physics & Astronomy, USA [email protected]

Abstract

We summarize here the observed properties of “infrared companions” to T Tauri stars and argue that their observational properties are identical to those of Class I sources. They may therefore be embedded protostars in a much earlier evolutionary phase than T Tauri stars, in which case these multiple systems are significantly non-coeval as opposed to the majority of young binary systems. They would have formed through a different mechanism than core fragmentation. The only distinction between IRCs and Class I sources is that they lie within a few tens of AU of a T Tauri star, and so they cannot be at the center of a vast optically thick envelope as is believed to be the case for protostars. We discuss whether systems with an IRC are really candidates for noncoeval multiple systems.

Introduction Binary and higher order multiple systems are the most frequent outcome of the star formation process and, as such, they represent a direct probe of this process. In the most widely accepted model to date, giant molecular clouds first give rise to a number of individual clumps, or cores, that subsequently undergo a collapse to finally form a central star. During this collapse, fragmentation may occur, due to rotation, turbulence or ambipolar diffusion of the magnetic field for instance, that leads to the formation of ∼2–5 physically associated objects which eventually evolve into a stable multiple system. One of the basic predictions of this scenario is that binary systems should be tightly coeval, to within a freefall time of the original core ( 106 yrs). Observational campaigns to 223 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 223-232. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

224

OPEN ISSUES IN LOCAL STAR FORMATION

test this prediction on large populations of T Tauri binary systems have been conducted for several years (e.g., Hartigan et al. 1994; White & Ghez 2001) and have concluded that binary systems are remarkably coeval, at least to within < 106 years. Therefore, observations tend to favor core fragmentation as the dominant process of forming binary stars. Among all T Tauri binary systems, there is a small category of remarkable objects named “infrared companions” (IRCs). These are defined as companions to known T Tauri stars that can only be detected in the infrared (IR) and that display “extremely” red IR colors, probably because of a large line-of-sight extinction. The most prominent IRC is the companion to the prototypical object T Tau (Dyck et al. 1982). The status of this class of objects was first discussed by Zinnecker & Wilking (1992) and a good compilation of observations can be found in Koresko et al. (1997), after which a few more IRCs were identified. As we will discuss here, these systems are problematic, as they seem to pair a normal T Tauri star with a much more embedded, and qualitatively much younger, protostar. They would therefore be non-coeval systems, contrasting with most young binary systems. Since the review by Koresko et al., additional high spatial resolution data have been obtained and information about the dynamics of the systems have been gathered by us and others. Here, we summarize those results and discuss how they fit into the binary coevality issue.

1. 1.1

Observational properties of IRCs A naive look at IRCs: an age paradox

In the well-studied Taurus-Auriga star-forming region, 5 IRCs are known among a sample of ∼ 120 stars, representing a mere 4 % of the overall population. The masses and ages of their optically bright primaries are in the ranges ∼0.2–2 M
and ∼0.1–5 Myrs respectively. The projected separations range from ∼ 10 to a few hundred AUs, similar to the range of separations for normal T Tauri binaries. Also, most systems show thermal millimeter emission associated with cold circumstellar material although in general it is not known with which component it is associated. The fact that some of these IRCs have (at times) been detected in the visible raises the possibility that intrinsically different objects are erronneously assembled in the same category. As a first step to determining the nature of IRCs, Koresko et al. (1997) compiled the spectral energy distribution (SED) for each IRC from the visible up to 100 µm. They all showed a similar shape, with a very broad peak centered between 5 and 20 µm depending on the object. The corresponding bolometric temperatures are much cooler than any stellar

Non coeval young multiple systems? (Duchˆene et al.)

225

photosphere, only a few hundred degrees. Assuming that they radiate isotropically, the bolometric luminosities of IRCs range from ∼0.8 to 12 L
. Taking these values at face value and plotting them against other young stellar objects as well as against the protostar evolution models of Myers et al. (1998), one finds that IRCs fall in the same part of the diagram as Class I sources, protostars embedded in a moderately massive, contracting circumstellar envelope. Therefore, if one were to classify IRCs on the basis of their SED only, one would undoubtedly conclude that IRCs are Class I protostars. T Tauri stars in general, and those that have an IRC in particular, have typical ages of 1–5 Myr. On the other hand, it is generally admitted that the small numbers of Class I protostars implies that they are much younger, typically a few 105 yrs. It therefore appears that IRC systems are non-coeval, with multi-Myr age differences. Three general explanations can be proposed to account for these systems: i) some multiple systems are really non-coeval, in which case one must explain how they formed; ii) IRCs only look like protostars but are in fact T Tauri stars disguised as protostars; and iii) the T Tauri stars associated to IRCs are in fact much younger than we think they are and these systems are in fact extremely young, coeval systems.

1.2

A purely geometrical explanation?

Among the three explanations suggested above, the idea that IRCs are in fact normal T Tauri stars in a peculiar geometric configuration is the most widely believed (e.g., Koresko et al. 1997). Such objects would look like protostars if they were heavily extincted by some circumstellar material, as none or very little flux shortward of 1 µm would reach the observer while at long wavelengths one would detect the thermal emission of the dusty material that enshrouds the central star. There are two types of configurations that would lead to the observed SEDs for IRCs. The first one is the case of a star that is embedded in an optically thick dusty envelope so that the only light we receive from the star has been reprocessed. Alternatively, IRCs could be T Tauri stars surrounded by an unresolved optically thick circumstellar disk that is seen at an almost edge-on inclination. Both observations and radiative transfer models of such objects show that their SED is extremely suppressed shortward of ∼ 10 µm, resulting in a predominant mid-IR peak, similar to those of Class I sources (D’Alessio et al. 1999; Wood et al. 2002). In this case, the inferred bolometric luminosity assuming isotropic radiation can be 1–2 orders of magnitudes lower than its actual

226

OPEN ISSUES IN LOCAL STAR FORMATION

value because the visible/near-IR light is predominantly scattered away from the observer’s line of sight. A relatively straightforward observational test can discriminate the two scenarios presented here: if the star is embedded into an optically thick envelope, its near-IR (and visible if observed) spectrum is featureless as it has been reprocessed by warm dust. On the other hand, in the edge-on disk scenario, the received spectrum has only been scattered at the surface of the disk and has retained the intrinsic photospheric features of the central object. This has for example been verified for the edge-on disk source IRAS 04158+2805 (M´enard et al. 2003).

2.

Recent observations of IRCs

Over the last few years, at least two new IRCs have been identified, WL 20 S and V 773 Tau D (Ressler & Barsony 2001, Duchˆene et al. 2003), and several high angular resolution datasets have been obtained, both in imaging and spectroscopic modes, for several systems, allowing a more complete understanding of their properties. First of all, IRCs are extremely variable, by up to several magnitudes even in the mid-IR. This was already known for some of them (e.g., T Tau: Ghez et al. 1991) and has also been observed for V 773 Tau D (Duchˆene et al. 2003). Also strong absorption features of both water ice and silicates have been observed in the IR (Hanner et al. 1998, Beck et al. 2001). These features unambiguously show that IRCs are observed through large column densities of dusty material. The photometric variability is however unlikely to be fully explained by a varying line-of-sight extinction: variations in the emission of the central source has to be present as well (Beck et al. 2001; Leinert et al. 2001). Recent near-IR spectroscopy of several IRCs (Haro 6-10 N, T Tau S, V 773 Tau D) have revealed featureless spectra, with the exception of atomic and molecular hydrogen in emission (Herbst et al. 1995; Beck et al. 2001; Duchˆene et al. 2002, 2003). This excludes the possibility of these IRCs being K- or M-type T Tauri stars extincted by an edge-on circumstellar disk. Note that this result is not inconsistent with IRCs being earlier spectral type (i.e., higher mass) objects seen behind an edge-on disk. This is discussed in more details in the following section. The case of T Tau S is quite revealing since this IRC is located in a triple system (Koresko 2000). Most importantly, Duchˆene et al. (2002) have shown that the tight companion to the IRC, which is located only 10–12 AU away, is a normal, though heavily extincted, T Tauri star with an M0.5 spectral type. This implies that what makes an IRC so special is confined within a few AU of the central object. If it is an opaque

Non coeval young multiple systems? (Duchˆene et al.)

227

Figure 1. Location of the companion to the IRC of T Tau – a.k.a. T Tau Sa – as a function of time since it discovery. Data are from Koresko (2000), K¨ ohler et al. (2000) and Duchˆene et al. (2002). The two most recent points were obtained by us with the adaptive optics system on the Keck II telescope.

circumstellar envelope, it has to be quite dense in order to be optically thick despite such a small radius. Finally, the most exciting new result regarding IRCs concerns their dynamical status. Most of them are located at a few hundred AU and the orbital periods for those systems are on order of a thousand years. However, T Tau S is in a 10 AU-binary and we can expect its orbital period to be about 15–30 yrs. Since its first discovery in 1997, several measurements of the binary separation have been made and clear evidence of orbital motion has been found (see Fig. 1). In November 2000, the relative velocity of the binary in the plane of the sky was on order of 13±4 km/s (based on a quadratic fit). Furthermore, we have also obtained spatially resolved high spectral resolution near-IR spectra of both components of the system. We found a relative velocity of 20±2 km/s. This combines to a three dimension relative velocity of about 24±4 km/s which implies a minimum system mass of MT T auS ≥ (4.2 ± 1.5)( 140Dpc )3 M
, if the system is physically bound. This is much more than the estimated mass of T Tau Sb and T Tau N, suggesting that the IRC is the most massive object in the T Tau multiple system. So far, the measurements do not cover enough of the orbit to allow a complete orbital solution fit but this should be feasible in just a few years. A controversial result concerning this system was recently obtained by Loinard et al. (2003) using archival VLA centimeter data. At these wavelengths, only T Tau N (the well known T Tauri star) and T Tau Sb (the extincted close companion to the IRC) are detected. Still, they concluded that the T Tau Sa–Sb system is unstable and that that T Tau Sb has been ejected in the last few years on a higher (possibly open) orbit. This is reminiscent of the “disrupting triple systems” proposed by Reipurth (2000), which were candidates for forming IRC systems. In this scenario, a very young (a few 105 yrs-old) triple system undergoes an unstable triple encounter and one of them is ejected in one direction while the remaining binary experiences a slow recoil motion in the other direction. The single star would then escape the opaque envelope still

228

OPEN ISSUES IN LOCAL STAR FORMATION

surrounding the system and therefore become optically visible while the other two components would remain heavily obscured (this is the third scenario mentionned in § 2.1). However, this scenario is not very well supported by our IR images, which show a clear slowdown of the motion of T Tau Sb between 2000 and 2002, suggesting that its orbit is bound. Only future observations will tell what type of orbit this star is on. In the case of V 773 Tau, the IRC has not been monitored long enough to allow a proper orbital fit. However, all observations are consistent with the system being hierarchical, as well as possibly coplanar, and therefore dynamically stable. Most other IRCs are located in binary systems and, unless a third component is later discovered, they are necessarily in two-body stable systems. Until further measurements prove otherwise, we conclude that IRCs are in stable systems and that Reipurth’s scenario is not generally the cause for these unusual IRCs.

3.

The high-mass star hypothesis

One of the scenarios presented above to account for the observed properties of IRCs is that they are high- to intermediate-mass stars obscured by a circumstellar disk seen edge-on. This is a likely situation in Glass I, since its spectral type has been estimated to be G5 (Feigelson & Kriss 1989) but is clearly inconsistent with the spectral type of UY Aur B (M2, Duchˆene et al. 1999). In systems with featureless IR spectra, the central star could be an A- or F- type star, preventing any line detection if the star is accreting material (accretion produces hydrogen line emission that fills photospheric features). Also, such objects would have large luminosities,  10 L
, but only a fraction of it would be seen by the observer because of the peculiar geometry. This would explain the observed luminosities of IRCs. Finally, in the case of T Tau S, a relatively large mass is required for the IRC if the system is physically bound. The high-mass star hypothesis is therefore a significant possibility that needs to be studied in more details. One way to test this hypothesis is to obtain high-resolution spectra of IRCs and try to find some photospheric features in them. It is for instance suggestive that, in our radial velocity measurement, we obtained the strongest cross-correlation peak for T Tau Sa with the spectrum of an F8 template star (we used M5 to F8 templates). A larger set of templates is however required to determine accurately the actual intrinsic spectral type of this IRC. Another possible test consists in analyzing the spectrum of the scattered light nebula surrounding T Tau. If the IRC is a 3–5 M
star, it is by far the most luminous object in the system and it should be the dominant source of illumination for the nebula.

Non coeval young multiple systems? (Duchˆene et al.)

4.

229

Are IRCs bona fide protostars?

If IRCs are not high-mass objects, then they have to be surrounded by compact optically thick circumstellar envelopes. It is then natural that they have similar properties as Class I protostars, as they are virtually identical: a central point source surrounded by a lighter but opaque envelope. In this scenario, one wonders why a T Tauri star would be surrounded by an optically thick envelope, since they are usually defined as objects around which the vast majority of the circumstellar material lies in an equatorial disk. It has been proposed that they are in fact deeply embedded because they are experiencing a temporary high accretion rate event similar to FU Ori bursts (Ghez et al. 1991, White & Ghez 2001) so that their opaque envelope would be a transient phenomenon. The number of IRCs would then suggest a relatively short-lived phenomenon (≈ 104 yrs). It remains to be understood how such an opaque envelope suddenly appears around one of the components of the system. This may be the result of star-star dynamics within the systems (Bonnell & Bastien 1992) but the details of this phenomenon still have to be described. There is however one usual property of protostars that IRCs do not share: a vast envelope. Millimeter mapping has shown that protostars have envelopes thousands of AU in radius (e.g., Motte & Andr´e 2001) whereas the fact that IRCs are in multiple systems imply that the outer radius of their envelope is not bigger than ∼2–100 AU depending on the system. From this point of view, IRCs are unlikely to be actual protostars. By analogy, this implies that Class I objects in general, which are defined by their near- to mid-IR SED, should not be considered protostars without further analysis, even though they are not known to have a companion. In fact, as discussed in Andr´e et al. (2000), the presence of a massive and extended envelope is required to consider an object a bona fide protostar. With this criterion, none of the IRCs can be considered a real protostar and the apparent non-coevality of the systems is solved. In summary, this study reminds us, from an unusual perspective, that a Class I-type SED is not the ultimate proof that an object is a protostar and that there might be a few non-coeval multiple systems in starforming regions. If we take the Class I classification of IRCs from their SED at face value, we then consider these systems as not coeval and we must explain how they formed. One possibility is that IRCs formed after a gravitational instability disrupted the circumstellar disk of their companion. This could represent a secondary channel for star formation and, although it only amounts to a few percent of all the stars formed in

230

OPEN ISSUES IN LOCAL STAR FORMATION

an environment such as Taurus, it would be interesting to see if different star-forming regions can form more objects in this way. Alternatively, IRCs might be FU Ori-like objects as old as their T Tauri companion, in which case dynamical interactions within the systems would need to be extremnely strong. Finally, a more original scenario for explaining (some of the) IRCs is that they are high-mass objects extincted by a nearly edge-on disk. If so, this would imply that even the Taurus molecular cloud can form objects as massive ≈ 5 M
. In any case, the unusual properties of IRCs deserve further investigation.

Acknowledgments The authors are grateful to the organizers of the conference for its great atmosphere, highly interesting list of topics and professional organisation. Many thanks also to Hans Zinnecker, Nuria Calvet, Lee Hartmann, Bo Reipurth and Philippe Andr´e for enriching discussions following this presentation. This work has been supported by the National Science Foundation Science and Technology Center for Adaptive Optics, managed by the University of California at Santa Cruz under Cooperative Agreement AST 98-76783.

References Andr´e, P., Ward-Thompson, D. & Barsony, M. 2000, in “Protostar & Planets IV”, eds. Mannings, Boss & Russell, Univ. of Arizona Press, p. 59 Beck, T., Prato, L. & Simon, M. 2001, ApJ, 551, 1031 Bonnell, I. & Bastien, P. 1992, ApJ, 400, 579 D’Alessio, P., Calvet, N., Hartmann, L., Lizano, S. & Cant´o, J. 1999, ApJ, 527, 893 Duchˆene, G., Monin, J.-L., Bouvier, J. & M´enard, F. 1999, A&A, 351, 954 Duchˆene, G., Ghez, A. M. & McCabe, C. 2002, ApJ, 568, 771 Duchˆene, G., Ghez, A. M., McCabe, C. & Weinberger, A. J. 2003, ApJ, 592, 288 Dyck, H. M., Simon, T. & Zuckerman, B. 1982, ApJ, 255, L103 Feigelson, E. D. & Kriss, G. A. 1989, ApJ, 338, 262 Ghez, A. M., Neugebauer, G., Gorham, P. W., Haniff, C. A., Kulkarni, S. R., Matthews, K., Koresko, C. & Beckwith, S. 1991, AJ, 102, 2066 Hanner, M. S., Brooke, T. K. & Tokunaga, A. T. 1998, ApJ, 502, 871 Hartigan, P., Strom, K. M. & Strom, S. E. 1994, ApJ, 427, 961 Herbst, T. M., Koresko, C. D. & Leinert, C. 1995, ApJ, 444, L93 Koresko, C. D., Herbst, T. M. & Leinert, C. 1997, ApJ, 480, 741 Koresko, C. D. 2000, ApJ, 531, L147 Leinert, C., Beck, T. L., Ligori, S., Simon, M., Woitas, J. & Howell, R. R. 2001, A&A 369, 215 Loinard, L., Rodr´ıguez, L. F. & Rodr´ıguez, M. I. 2003, ApJ, 587, L47 M´enard, F., Dougados, C., Magnier, E., Duchˆene, G., Cuillandre, J.-C., Mart´ın, E., Fahlman, G., Forveille, T., Lai, O., Manset, N., Martin, P., Weillet, C. & Magazz`u, A. 2003, ApJ, submitted Motte, F. & Andr´e P. 2001, A&A, 365, 440

Non coeval young multiple systems? (Duchˆene et al.)

231

Myers, P. C., Adams, F. C., Chen, H. & Schaff, E. 1998, ApJ, 492, 703 Reipurth, B. 2000, AJ, 120, 3177 Ressler, M. E. & Barsony, M. 2001, AJ, 121, 1098 White, R. J. & Ghez, A. M. 2001, ApJ, 556, 265 Wood, K., Wolff, M. J., Bjorkman, J. E. & Whitney, B. 2002, ApJ, 564, 887 Zinnecker, H. & Wilking, B. A. 1992, in “Binaries as Tracers of Stellar Formation”, eds. Duquennoy & Mayor, Cambridge Univ. Press, p. 269

Gaspard Duchˆene

232

M. Sterzik, S. Lamzin

OPEN ISSUES IN LOCAL STAR FORMATION

III

BROWN DWARFS

This page intentionally left blank

BROWN DWARFS IN YOUNG OPEN CLUSTERS The Pleiades mass function Estelle Moraux, J´erˆome Bouvier Laboratoire d’Astrophysique, Observatoire de Grenoble, France [email protected]

Abstract

We report new estimate for the lower Pleiades mass function across the stellar/substellar boundary. We find that it is well represented by a single power-law dn/dm ∝ m−0.60±0.11 between 0.03M
and 0.45M
. Over a larger mass domain, however, the mass function is better fitted by a log-normal distribution. We find that the early dynamical evolution of the cluster appears to have had little effect on its present mass distribution at an age of 120 Myr. Comparison between the Pleiades mass function and the Galactic disk mass function suggests that apparent differences may be mostly due to unresolved binaries.

Introduction The initial mass function results directly from physical processes which occur during the star formation. Its determination at low masses is therefore a formidable constraint for stellar and substellar formation models and is one of the main motivations for the quest of brown dwarfs. Young nearby open clusters are ideal locations to search for isolated substellar objects since their youth ensures that brown dwarfs are still bright enough for being easily detected. Moreover, their populations are homogeneous in age and distance and they offer the possibility to estimate complete mass functions from the substellar domain up to massive stars. In this contribution, we present the results obtained from the CFHT wide field survey of the Pleiades. We discuss contamination of the photometric samples and derive the cluster substellar mass function (section 1). We then briefly discuss the potential effects of cluster dynamical 235 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 235-244. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

236

OPEN ISSUES IN LOCAL STAR FORMATION

evolution (section 2) and binarity (section 3) on the shape of the mass function.

1.

The Pleiades mass function

To date, several Pleiades surveys have been performed from which the cluster mass function has been estimated in the upper part of the substellar domain. All the estimates are consistent with a single powerlaw dn/dm ∝ m−α (Zapatero-Osorio et al. 1997: α  1 ± 0.5, Bouvier et al. 1998 : α  0.6 ± 0.15 revised to α  0.51 ± 0.15 by Moraux et al. 2001, Mart´ın et al. 2000 : α  0.5 − 1.0, Hambly et al. 1999 : α ≤ 0.7). However, those results suffer from uncertainties due mainly to small number statistics.

CFHT-PLIZ BD candidates

H 18

J

I G

CFHT-PL (Bouvier et al. 98) 120 Myr isochrone

O

E

NEXTGEN model (BCAH 98) 120 Myr isochrone

K D

DUSTY model (CBAH 00)

20

L

Q

B

R

P

C 22

U

V

T

24

0

0.5

1

1.5

2

2.5

Figure 1. Left panel : Area of the sky covered by the CFHT 2000 Pleiades survey. Each rectangle corresponds to one CFH12K field. The star symbols indicate the 25 brightest stars of the cluster which have been avoided to prevent CCD saturation. Filled dots show the location of detected Pleiades brown dwarf candidates. Right panel : (I, I-z) CMD for all stellar-like objects identified on the long exposure images. The 120 Myr NEXTGEN isochrone from Baraffe et al. (1998) and DUSTY isochrone from Chabrier et al. (2000) shifted to the distance of the Pleiades are shown as a dashed line and a dash-dotted line labelled with mass (in M
unit) respectively. Brown dwarfs detected in the first CFHT survey and recovered here are shown as encircled triangles while other candidates are new. (from Moraux et al. 2003)

A new Pleiades survey was performed with the CFH12K camera in December 2000 covering a wide area (6.4 sq.deg.) and reaching down to 0.025M
. Figure 1 illustrates the covered area on the sky (left) and the resulting (I, I − z) color magnitude diagram (CMD, right). We selected low mass star and substellar candidates on the basis of their location on

Brown dwarfs in young open clusters (Moraux & Bouvier)

237

this CMD which led to the identification of 40 brown dwarf candidates down to 0.03M
, of which 29 are new. One of the major shortcomings of this selection method, however, is the contamination by older and more massive late-type field dwarfs which may lie in the same region of the CMD. Follow-up observations are necessary to confirm the brown dwarfs status of the candidates. The best way to identify true cluster brown dwarfs is to measure the proper motion of the candidates (Moraux et al. 2001). However, even if the Pleiades proper motion is quite large (∼50 mas/yr), we have to wait for at least three years to be able to measure proper motion. A quicker alternative is to obtain near-infrared photometry in the Kband, since the (I, I − K) CMD associated to the (I, I − z) CMD is a powerful membership evaluation tool, and/or low-resolution spectra of the candidates. Figure 2 shows an optical spectrum obtained with FORS2 at the VLT (left) and a near-infrared spectrum obtained with NICS at the TNG (right).

Figure 2. Left panel : Optical spectrum of a brown dwarf candidate obtained with FORS2 at the VLT (solid line). A fit by a synthetic spectra from Allard et al.’s (2001) models is shown as a dashed line. The effective temperature is 2800K, which corresponds to a M6 spectral type. Right panel : Near-infrared spectrum obtained with NICS at the TNG (solid line). The best fit (dashed line) indicates an effective temperature of 2100K corresponding to a L1 spectral type. The discrepancies between the observed spectrum and the synthetic spectrum are mainly due to the fact that the H2 O opacities are not very well constrained (Moraux et al., in prep).

These spectra allow us to check that the candidate spectral type corresponds to very late-type dwarfs (M or L) as expected for young brown dwarfs at ∼100 Myr and to study some gravity features such as the N a doublet (see Figure 3) in order to distinguish between young cluster’s brown dwarfs and field dwarfs (Mart´ın et al. 1999). We thus found a level of contamination of the substellar photometric sample of ∼30%. Taking into account this contamination, the luminosity function was then converted into a mass function (MF) using the (I magnitude, mass) relationship from Baraffe et al. (1998) for stars and Chabrier et al.

238

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 3 Zoom on the gravity sensitive N a doublet of the FORS2 spectrum. The brown dwarf candidate spectrum is shown as a solid line. The spectra of two field dwarfs having the same M6 spectral type are overplotted for comparison. 296A (dashed line) is known to be of similar age to the Pleiades (Thackrah et al. 1997) whereas GJ 866 (dotted line) belongs to the old disk population. The older is a dwarf, the larger is its gravity and the larger is the equivalent width of the N a doublet at a given spectral type. The brown dwarf candidate has a N a equivalent width similar to this of 296A which suggests it is of similar age to the Pleiades and is therefore a genuine brown dwarf member.

(2000) for brown dwarfs. The derived Pleiades MF across the stellar/substellar boundary is shown in Figure 4. It is reasonably well-fitted by a single power-law dn/dm ∝ m−0.60±0.11 over more than one decade in mass, from 0.03 to 0.45M
.

2.

Cluster dynamical evolution

Since the main objective in determining the lower mass function of young open clusters is to constrain the star and brown dwarf formation process(es), a pressing issue is whether this MF is representative of the initial mass function (IMF). In other words, is the cluster population observed at an age of about 100 Myr representative of the initial population of the cluster at the time it formed ? Cluster dynamical processes act to deplete the lowest mass members. Weak gravitational encounters lead to mass segregation and to the evaporation of the lowest mass objects which may affect the shape of the mass function.

Brown dwarfs in young open clusters (Moraux & Bouvier)

239

Figure 4 The Pleiades mass function across the stellar/substellar boundary. Note that all data points are derived from the same survey, using short exposures for the stellar domain and long exposures for the substellar regime. This provides a consistent determination of the slope of the cluster’s mass function in the mass range from The 0.030 to 0.4 M
. data points are fitted by a power law with an index α = 0.60 ± 0.11 (dn/dm ∝ m−α ).

We therefore compare our results with the lower mass function derived for star forming regions such as the Trapezium (Luhman et al. 2000) or σ-Ori (B´ejar et al. 2001) and find that they appear similar. This suggests that dynamical processes are not predominant in the early evolution of open clusters and that the MF derived for the Pleiades cluster at an age of 120 Myr may indeed be representative of the MF at a few Myr. Moreover, current models of dynamical evolution (e.g., de la Fuente Marcos & de la Fuente Marcos 2000) predict that only ∼10% of the low mass stars and brown dwarfs will have escaped a Pleiades-like cluster at an age of 100 Myr, and the predicted loss rate is nearly the same for substellar objects and low mass stars. Hence, the shape of the mass function will not be affected across the stellar/substellar boundary and the secular dynamical evolution of stellar clusters is not expected to significantly deplete the IMF at low masses during the early evolutionary stages. However, in some theoretical models, other dynamical effects intimately linked to the star-formation process may influence significantly the observed MF at Pleiades age. In the Reipurth & Clarke (2001) scenario proposed for the formation of isolated brown dwarfs, the lowest mass fragments of protostellar aggregates are dynamically ejected. If their ejection velocity exceeds the escape velocity, then we expect they rapidly leave the cluster and this may lead to a primordial depletion

240

OPEN ISSUES IN LOCAL STAR FORMATION

of brown dwarfs. Several models of dynamical ejection have been developped along these lines and all suggest ejection velocities of order of a few km/s, though the output of each particular model differs in the details (cf. Sterzik & Durisen 1998, Bate et al. 2002, Delgado-Donate et al. 2003). Figure 5 The cumulative radial distribution of brown dwarfs at an age of 120 Myr computed from numerical simulations of the dynamical evolution of a Pleiades-like cluster. From top to bottom the distributions illustrate the effect of increasing the initial velocity dispersion of cluster BDs. The dashed line represents the Pleiades tidal radius of about 13pc.

Following this scenario, Figure 5 shows the results of a N-Body 2 numerical simulation of the dynamical evolution of a Pleiades-like cluster. This simulation includes 1600 objects whose mass distribution follows a prescribed mass function, with the initial velocity dispersion of brown dwarfs being scaled relative to that of stars, and assuming the whole system is virialized. The various curves in Figure 5 illustrate the computed spatial distribution of brown dwarfs at an age of 120 Myr depending on their initial velocity distribution. It is seen that half or more of the initial cluster brown dwarfs will have left the cluster by an age of 120 Myr if their initial velocity dispersion exceeds that of stars by a factor of 2. Such dynamical processes could therefore modify significantly the lower mass function of young open clusters. Furthermore, this effect should depend upon the specific properties of each cluster, such as initial stellar density and radius, thus presumably leading to different lower MF for different clusters. Yet, no significant difference is found so far between the lower MF of the Pleiades cluster and other young open clusters, such as α-Per. The Pleiades lower MF appears also similar to that of star forming regions as already mentioned. These results suggest that brown dwarfs do not

Brown dwarfs in young open clusters (Moraux & Bouvier)

241

have any peculiar kinematic. We conclude that dynamical processes are probably not predominant in the formation and early evolution of open clusters and the MF derived at ∼100 Myr may indeed be representative of the IMF.

3.

Binarity

Below 0.8M
, the Galactic disk mass function also ought to be representative of the initial mass function since field stars have not had time to evolve off the main sequence. We therefore compare our results to those obtained recently by Chabrier (2001). The log-normal fit to the field mass function in the mass range 0.1-1.0M
is shown as a dashed line in Figure 6. Over the same mass domain, we find that the Pleiades MF is also well fitted by a log-normal approximation (solid line). However, major differences clearly appear at low masses. The log-normal mass function peaks at M  0.1M
for the disk population and at  0.25M
for Pleiades members. Furthermore, Chabrier (2002) estimates that the brown dwarf population in the disk is comparable in number to the stellar one, NBD  N∗ , and that the substellar mass contribution to the disk budget amounts to about ∼ 10%. In the Pleiades, we have instead NBD  N∗ /3 and a brown dwarf mass contribution to the cluster mass of only 1.5%. Part of the observed difference may arise from the effect of binarity, which is not accounted for in the Pleiades mass function. While the Pleiades mass function derived above includes unresolved cluster binaries, the Galactic disk mass function has been derived for the single star population. In order to correct the observed Pleiades mass function for unresolved binaries at low masses, we used a Monte Carlo technique to estimate the difference between single star and system mass functions. Assuming that the properties of field binaries derived by Duquennoy & Mayor (1991) apply to Pleiades systems (Bouvier et al. 1997), we form unresolved binaries by random pairing. The individual objects are drawn from a segmented power law mass function and we adopt a 50% binary fraction. We find no major differences between the star and system mass functions at M > 0.6M
since above this mass the presence of a low mass companion does not significantly affect the determination of the primary’s mass. Below 0.6M
, however, we find ∆α ∼ 0.5 between the power law exponents of the single star and system mass functions. Thus, we estimate the binary corrected Pleiades mass function down to the substellar limit (dotted line in Figure 6). This mass function and the Galactic disk mass function now both peak around 0.13 − 0.1M


242

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 6 The Pleiades MF represented as the number of objects per logarithmic mass units for 0.03 ≤ m ≤ 10M
. The histogram corresponds to the Pleiades star catalog built from the Prosser and Stauffer Open Cluster Database and the large dots are our data points. The log-normal fit to the observed Pleiades mass function is shown as a solid line. The dashed line corresponds to the Galactic disk mass function from Chabrier (2001) and the dotted line to the estimated Pleiades mass function corrected for unresolved binaries. (from Moraux et al. 2003)

and their shapes are roughly identical. This suggests that the difference between the observed Pleiades mass function and Galactic disk mass function in the stellar range is merely the result of unresolved Pleiades binaries. The comparison of the two mass functions in the substellar domain cannot be as detailed, due to the large uncertainties still affecting the derivation of the substellar mass function in the Galactic disk and the brown dwarf binary properties. This similarity confirms that the MF observed at an age of ∼100 Myr is representative of the IMF. Hence, we do not find evidence for massive ejection of the lowest mass objects early in the life of the cluster, which suggests either that dynamical ejection from protostellar groups is not the dominant mode of brown dwarf formation, or that this process yields a velocity dispersion for brown dwarfs which is not different from that of stars.

4.

Conclusion

Our deep wide-field photometric survey of brown dwarfs has yielded an estimate of the Pleiades mass function accross the stellar/substellar boundary. We find that a single power-law dn/dm ∝ m−0.60±0.11 provides a good match in the 0.03-0.45 M
range. Over a larger mass domain, from 0.03 to 10 M
, we find that the cluster mass function is

Brown dwarfs in young open clusters (Moraux & Bouvier)

243

better fitted by a log-normal distribution. Current estimates of the mass function in star forming regions indicates that there is no appreciable differences in the shape of the lower mass function. Furthermore, when unresolved Pleiades binaries are taken into account, the log-normal mass function is not unlike the Galactic disk mass function. This suggests that the dynamical evolution of the cluster has had little effect on its mass content and thus that the currently measured Pleiades MF at an age of 120 Myr is representative of its IMF. It also suggests that the brown dwarf formation process does not lead to the dynamical evaporation of substellar objects at the time the cluster forms.

References Allard, F., Hauschildt, P. H., Alexander, D. R., Tamanai, A., & Schweitzer, A. 2001, ApJ, 556, 357 Baraffe, I., Chabrier, G., Allard, F., Hauschildt, P.H. 1998, A&A, 337, 403 Bate M.R., Bonnell, I.A., & Bromm, V. 2002, MNRAS, 332, L65 B´ejar, V.J.S., Mart´ın, E.L., Zapatero-Osorio, M.R., Rebolo, R., Barrado y Navascu´es D., Bailer-Jones, C.A.L., Mundt, R., Baraffe, I., Chabrier, G., Allard, F. 2001, ApJ, 556, 830 Bouvier, J., Rigaut, F., & Nadeau, D. 1997, A&A, 323, 139 Bouvier, J., Stauffer, J.R., Mart´ın, E.L., Barrado y Navascu´es, D., Wallace, B., B´ejar, V.J.S. 1998, A&A, 336, 490 Chabrier, G., Baraffe, I., Allard, F., Hauschildt, P.H. 2000, ApJ, 542, 464 Chabrier, G. 2001, ApJ, 554, 1274 Chabrier, G. 2002, ApJ, 567, 304 Delgado-Donate, E., Clarke, C., Bate, M.R. 2003, MNRAS, submitted Duquennoy, A., & Mayor, M. 1991, A&A, 248, 485 de la Fuente Marcos, R., de la Fuente Marcos, C. 2000, Ap&SS, 271, 127 Hambly, N.C., Hodgkin, S.T., Cossburn, M.R., Jameson, R.F. 1999, MNRAS, 303, 835 Luhman, K.L., Rieke, G.H., Young, E.T., Cotera, A.S., Chen, H., Rieke, M.J., Sneider, G., Thompson, R.I. 2000, ApJ, 540, 1016 Mart´ın, E.L., Delfosse, X., Basri, G., Goldman, B., Forveille, T., Zapatero-Osorio, M.R. 1999, AJ 118, 2466 Mart´ın, E.L., Brandner, W., Bouvier, J., Luhman, K.L., Stauffer, J.R., Basri, G., Zapatero-Osorio, M.R., Barrado y Navascu´es, D. 2000, ApJ, 543, 299 Moraux, E., Bouvier, J., Stauffer, J.R. 2001, A&A, 367, 211 Moraux, E., Bouvier, J., Stauffer, J.R., Cuillandre, J.-C. 2003, A&A, 400, 891 Reipurth, B., & Clarke, C. 2001, AJ, 122, 432 Sterzik, M.F., & Durisen, R.H. 1998, A&A, 339, 95 Thackrah, A., Jones, H., & Hawkins, M. 1997, MNRAS, 284, 507 Zapatero-Osorio M.R. et al. 1997, ApJ 491, L81

244

Estelle Moraux

T. Montmerle, J. Bouvier

OPEN ISSUES IN LOCAL STAR FORMATION

BROWN DWARF COMPANIONS Predictions from dynamical decay models Michael F. Sterzik European Southern Observatory, Chile [email protected]

Richard H. Durisen Indiana University, U.S.A [email protected]

Abstract

Some popular star formation theories envision binary formation as the end product of cloud core fragmentation followed by dynamical evolution within small-N stellar groups. To explore these scenarios, we simulate the dynamical decay of small-N stellar groups by direct orbit integrations and analyze the multiplicity and contents of the stable decay products. Our simulations include a wide range of component masses, including very-low mass stars and brown dwarfs. We find: (1) Brown dwarfs have a low multiplicity and are in general rare companions to stars. (2) When all the small-N clusters are given the same specific energy, there is a significantly smaller mean separation distribution for brown dwarf binaries than for stellar binaries. (3) Brown dwarfs companions to stars with wide (separations > 100 AU) should be members of hierarchical triples, in the sense that the stellar primary itself is expected to be a binary. (4) Spectroscopic binary brown dwarfs should often be the central binary of a hierarchical triple. Although observations tend to support some of these predictions, puzzles remain.

Introduction Many, if not most, stars are observed to be members of binary or multiple systems. Multiple star properties provide critical constraints on the star formation process, but the detailed mechanisms that define the binary frequency, period, and mass ratios are still poorly understood. Recent observational advances include the determination of 245 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 245-250. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

246

OPEN ISSUES IN LOCAL STAR FORMATION

multiplicity fractions and distributions across a broad range of primary masses. These observations indicate that (1) the multiplicity fraction among very low-mass stars (VLMs) and brown dwarfs (BDs) appears to be about 15%, significantly lower than it is for higher mass stars, (2) the separation distribution for VLMs and BDs is narrower and is peaked towards smaller values than for solar-type stars (4 AU instead of 40 AU, see Close et al. 2003), (3) the so-called “brown dwarf desert” – a strong deficit of BD companions to stars – extends to periods of a few years (see Halbwachs et al. 2000), and (4) the frequency of BD companions to main-sequence stars at wide separations > 1000 AU is apparently “normal” and in accord with expectations from wide stellar companions (see Gizis et al. 2001). The observations mentioned above can be interpreted in the framework of dynamical processes occurring in small-N stellar clusters formed from fragmenting cloud cores. In one popular scenario, BDs are substellar objects because they have been ejected from small, newborn multiples that decay through dynamical interactions (Reipurth & Clarke 2001; Reipurth, this volume). This interpretation qualitatively explains some of these observations. More quantitative predictions are possible when the dynamical evolution of small-N clusters is calculated by detailed numerical integrations. We have recently presented statistically robust predictions of BD properties arising from dynamical interaction calculations (Sterzik & Durisen 2003a). In the following, we will discuss our main results regarding VLMs and BDs as members of binary and multiple systems.

1.

Brown dwarf companion frequencies

Dynamical evolution within a cluster of point-like objects having diverse masses will favor the formation of a relatively hard binary consisting of the two most massive bodies. The less massive objects will most likely become unbound or remain in a hierarchical orbit around the central binary. This basic dynamical behavior is well-known and is sometimes referred to as “dynamical biasing”. If cluster masses are assumed to be drawn randomly from an initial mass function (IMF) extending from the substellar regime to many solar masses, then most VLMs and BDs will become single objects after a short dynamical interplay as unstable members in the primordial multiples. But deviations from strict dynamical biasing are also expected. For instance, the influence of dissipative processes during dynamical evolution (e.g., competitive accretion, disk collisions, and stellar tides) will increase the cross section for binding low mass objects (McDonald & Clarke 1995).

247

Brown Dwarf Companions (Sterzik & Durisen)

We have also found that the way masses are assigned within clusters can significantly alter the final pairing statistics (Durisen, Sterzik & Pickett 2001). If stellar clusters are formed via hierarchical fragmentation of molecular clouds, the mass of each individual core limits the total mass budget for the cluster of stars it will form. This tends to produce a more uniform distribution of stellar masses within clusters relative to a pure random mass assignment from an IMF. In our statistical treatment of small-N clusters, we refer to this as a “two-step” approach to mass assignment, because a cluster mass is chosen (step one) and then the sum of the chosen component masses (step two) is required to equal it. Using this approach to treat large numbers of clusters while also requiring that the overall stellar and substellar IMF be consistent with observations, we find that cluster decays produce the pairing statistics for substellar objects shown in Table 1. Table 1. Companion frequencies (CF) for two mass ranges of sub-stellar objects (T: .01–.05 M
, L: .05–.08 M
) as a function of spectral type of the primary. The mass ranges for the primary stars are indicated. Spectral type Mass range [M
] CF(L) CF(T)

L .05–.08

M.08–.27

M+ .27–.47

K .47–.84

G .84–1.2

F+ >1.2

.02 .06

.04 .05

.03 .04

.03 .03

.01 .02

.01 .01

Our dynamical evolution simulations yield a relatively low percentage of substellar objects as bound members in binary or multiple systems. In fact, in our calculations, only about 8% of L-type BDs and only 2% of all early type stars have BD companions. The observed BD companion frequency of about 15% for VLMs (Close et al. 2003), though somewhat higher than the 9% entry for M- in Table 1, is consistent to within the uncertainties of both the theoretical and empirical numbers. The binary separations for the Close et al. 2003 sample of VLMs and BDs, with typical masses of 0.1M
, are observed be 10 times smaller than for 1M
stars, which agrees well with results from our simulations when we assume that all clusters have similar specific energy. This lends some support to arguments we have made (Sterzik & Durisen 2003b) in favor of this assumption. Although considerable efforts have been undertaken to determine the dependence of BD companion frequencies on their primary mass, it is not easy to compare the predicted and observed fractions quantitatively. The first BD, Gl 229B, was actually found as a visual companion of an M1 star. Initially, because observers expected the frequency of BD

248

OPEN ISSUES IN LOCAL STAR FORMATION

companions to be as high as that for stars, intensive searches by various research groups were conducted, but only a few visual BD companions have been found. The BD desert is characterized by a firm upper limit of 1000 AU seem to be comparatively common (Gizis et al. 2001). Figure 1 illustrates the dependence of the mass ratio on separation for different types of BD companions based on our simulations. With few exceptions, BD binaries

249

Brown Dwarf Companions (Sterzik & Durisen)

(small filled circles) have narrow separations, typically below 10 AU. Their mass ratios are evenly distributed. BDs as secondaries to stars (large open circles) tend to have a similar separation distribution but preferentially populate the lower mass ratio regime. Only a few low mass ratio binary systems have larger separations. BDs as outer components of triple systems are plotted as large, filled circles. In general, they have large separations and small mass ratios. According to this, we expect that stellar primaries of wide, low-q BDs will mostly turn out to be binaries or higher-order systems. Observations are needed to test this prediction. 1.0

mass ratio

0.8

0.6

0.4

0.2

0.0 1

10

100 semi major axis [AU]

1000

10000

Figure 1. Simulated mass ratio versus semi-major axis distribution for various BD secondary populations. Binary BD: small filled circles; BD companions to stars: large open circle; BD companions in triples: large filled circles.

2.3

Very close brown dwarf binaries

Only three spectroscopic VLM or BD pairs are known, namely PPl15, 2M0253 and 2M0952 (Basri & Martin 1999, Reid et al. 2002). The formation of spectroscopic binaries of any mass is still an enigma. Dynamical interactions in small-N clusters cannot alone account for them. It is interesting to note that spectroscopic binaries tend to occur more often in hierarchical systems (Tokovinin & Smekhov 2002; Melo 2003). One way to shrink an inner orbit is through the combined action of the Kozai effect and tidal friction in hierarchical triples (Kiseleva, Eggelton & Mikkola 1998). A perpendicular orbit configuration (which is required

250

OPEN ISSUES IN LOCAL STAR FORMATION

to start the process) occurs frequently in decaying few-body clusters ( Sterzik & Tokovinin 2002). We speculate that an enhanced fraction of very close BD binaries might be found in hierarchical triples. This is another prediction that awaits observational tests. We note, however, that the known spectroscopic BD pairs do not seem to be accompanied by a third, low-mass system member. Their origin remains puzzling.

3.

Summary

We use numerical integrations to simulate the dynamical evolution and decay of small-N stellar clusters in fragmenting cloud cores. The statistical analysis of stable end products containing VLMs and BDs as members of multiple systems yields the following results: (1) Brown dwarfs have a low multiplicity and are in general rare companions to stars. (2) Brown dwarf binaries have smaller mean separations than stellar binaries. (3) Spectroscopic (sub-)stellar binaries cannot be formed by dynamical interactions alone. (4) Wide brown dwarf companions tend to be members of hierarchical triple systems. (5) Triples formed by dynamical interactions have preferred orbit configurations that could lead to significant orbital evolution (shrinkage) of the inner binary. Results (1) and (2) are corroborated by observations. An observational analysis of primaries having wide BD companions will reveal the correctness of prediction (4). The presence of spectroscopic binaries (3) is problematic. Processes related to (5) might be a way to explain them.

References Basri, G., & Martin, E., 1999, AJ, 118, 2460 Close, L.M., Siegler, N., Freed, M., & Biller, B., 2003, ApJ, 587, 407 Durisen, R.H., Sterzik, M.F., & Pickett, B.K 2001, A&A, 371, 952 Gizis, J. E., Kirkpatrick, J. D., Burgasser, A., et al., 2001, ApJ, 551, L163 Halbwachs, J. L., Arenou, F., Mayor, M., Udry, S., Queloz, D., 2000, A&A, 355, 581 Kiseleva, L.G., Eggelton, P.P., & Mikkola, S. 1998, MNRAS, 300, 292 Martin, E. L., Koresko, C. D., Kulkarni, S. R., et al., 2000, ApJ, 529, L37 Melo, C., 2003, A&A, in press McDonald J.M., & Clarke C.J., 1995, MNRAS 275, 671 Reid, I. N., Kirkpatrick, J. D., Liebert, J., Gizis, J. E., Dahn, C.C., & Monet, D.G., 2002, AJ, 124, 519 Reipurth, B., & Clarke, C., 2001, AJ, 122, 432 Sterzik, M.F. & Durisen, R.H. 2003a, A&A, 400, 1031 Sterzik, M.F. & Durisen, R.H. 2003b, In: ‘The Environment and Evolution of Binary and Multiple Stars’, Rev.Mex.AA, in press. Sterzik, M.F. & Tokovinin, A.A 2002, A&A, 384, 1030 Tokovinin, A.A., & Smekhov, M.G., 2002, A&A, 382, 118

OBSERVATIONAL CLUES TO BROWN DWARF ORIGINS Ray Jayawardhana Department of Astronomy, University of Michigan, USA [email protected]

Subhanjoy Mohanty Harvard-Smithsonian Center for Astrophysics, USA

Gibor Basri Department of Astronomy, University of California, USA

David R. Ardila Bloomberg Center, Johns Hopkins University, USA

Beate Stelzer Osservatorio Astronomico di Palermo, Italy

Karl E. Haisch, Jr. Department of Astronomy, University of Michigan, USA

Abstract

Over the past year, we have conducted a multi-faceted program to investigate the origin and early evolution of brown dwarfs. Using highresolution Keck optical spectra of ∼30 objects near and below the substellar boundary in several star-forming regions, we present compelling evidence for a T Tauri-like accretion phase in young brown dwarfs. Our systematic study of infrared L -band (3.8µm) disk excess in ∼50 spectroscopically confirmed young very low mass objects reveal that a significant fraction of brown dwarfs harbor disks at a very young age. Their inner disk lifetimes do not appear to be vastly different from those of disks around T Tauri stars. Taken together, our findings are consis251

J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 251-258. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

252

OPEN ISSUES IN LOCAL STAR FORMATION tent with a common origin for most low-mass stars, brown dwarfs and isolated planetary mass objects.

Introduction Brown dwarfs, which straddle the mass range between stars and planets, appear to be common both in the field and in star-forming regions. Their ubiquity makes the question of their origin an important one, both for our understanding of brown dwarfs themselves as well as for theories on the formation of stars and planets. In the standard framework, a low-mass star forms out of a collapsing cloud fragment, and goes through a “T Tauri phase”, during which it accretes material from a surrounding disk, before arriving on the main sequence. There is ample observational evidence now to support many key aspects of this picture for young solar-mass stars. Whether the same scenario holds for objects at and below the sub-stellar limit is an open question. Studies of young sub-stellar objects could provide valuable clues to address that question. To that end, over the past year, we have undertaken a multi-faceted study of very low mass (VLM) objects in star-forming regions and their immediate circumstellar environment. Here we report on our investigations of accretion signatures and disk excess in young brown dwarfs.

1.

Accretion Signatures

The shape and width of the Hα emission profile is commonly used to discriminate between accretors and non-accretors among T Tauri stars (TTS). Stars exhibiting broad, asymmetric Hα lines with equivalent width larger than 10 ˚ A are generally categorized as classical TTS (CTTS), although this threshold value varies with spectral type. Recently White & Basri (2003) have suggested that a full-width > 270 kms−1 at 10% of the peak emission is a better empirical indicator of accretion, independent of spectral type. However, we find that in VLM accretors, the Hα profile may be somewhat narrower than that in higher mass stars. We propose that low accretion rates combined with small infall velocities at very low masses can conspire to produce this effect, and adopt ∼ 200 kms−1 as a more appropriate, yet conservative, threshold (see Jayawardhana, Mohanty & Basri 2003 for further discussion). We obtained high-resolution Keck optical spectra of ∼30 objects spanning the range of M5–M8 in IC 348, Taurus, Upper Scorpius and ρ Ophiuchus star-forming regions. Putting together now all the young objects near or below the substellar boundary (∼ M5 and later) with published high-resolution optical spectra (Jayawardhana, Mohanty & Basri 2002;

Brown Dwarf Disks and Accretion (Jayawardhana et al.)

253

2003 and references therein), we have 4 objects in ρ Ophiuchus, 10 in IC 348, 14 in Taurus and 11 in Upper Scorpius. Of these, optical spectral signatures of accretion are found in 1 object in ρ Oph (GY 5), 5 in IC 348 (IC 348-165, 205, 355, 415, 382; Fig. 1), 3 in Taurus (CIDA-1, GM Tau, V410 Anon 13) and 1 in Upper Sco (USco 75), adopting an accretion cutoff of ∼ 200 kms−1 in 10% width of the Hα line. The vast majority of ρ Oph VLM objects were inaccessible to our optical spectroscopy because of significant extinction, presumably due to circumstellar as well as interstellar material. Thus, our (small) ρ Oph sample is heavily bi-

Figure 1. Hα line profiles of IC 348 and Taurus targets. Spectra shown have been smoothed by a 3-pixel boxcar; continuum and full width at 10% of the peak levels are marked by dotted lines. Thick black lines indicate accretors with broad Hα as well as CaII and OI emission; grey indicates probable accretors, based on the Hα profileshape and 10% full-width. Insets zoom in on objects with low peak-flux and noisy continuua, to clearly show the Hα detection. For CFHT-3, two spectra are shown, separated by a year; note the similarly strong emission both times. For CFHT-4, note the change in Y-axis scale; the peak flux in this object is much higher than in any other target in our sample.

254

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. Line profiles of OI (8446˚ A) and CaII (8662˚ A) in three IC 348 objects. Arrows indicate line-center. These spectra have been smoothed by a 4-pixel boxcar.

ased against possible accretors, and should not be used to estimate the accreting fraction. Considering the other three clusters, which are much less affected by this bias, we find that ∼50% of the VLM objects show disk accretion at an age ≤ 2 Myr (IC 348), ∼ 20% at age ∼ 3 Myr (Taurus), and ≤10% by ∼ 5 Myr (Upper Sco). While there are uncertainties in the cluster ages, IC 348 is likely to be younger than Taurus and Upper Sco. Thus, we appear to be seeing a decrease in the fraction of accreting young sub-stellar objects with increasing age. Interestingly, three of our ∼ M6 IC 348 targets with broad Hα also harbor broad OI (8446˚ A) and CaII (8662˚ A) emission (Fig. 2), and one shows broad HeI (6678˚ A) emission; these features are usually seen in strongly accreting classical T Tauri stars. Our results constitute the most compelling evidence to date that young brown dwarfs undergo a T Tauri-like accretion phase similar to that in low-mass stars.

2.

Disk Excess

Excesses at infrared wavelengths provide readily a measurable signature of dusty disks around late-type objects. Using the ESO Very Large Telescope, Keck I and the NASA Infrared Telescope Facility, we have carried out a systematic study of infrared L -band (3.8µm) disk

Brown Dwarf Disks and Accretion (Jayawardhana et al.)

255

Figure 3. J − H/K − L color-color diagram for our target sample. Also plotted are the empirical loci of colors for giants (solid) and for main-sequence dwarfs (dashed) from Bessell & Brett (1988) and Leggett et al. (2002) and the reddening vectors (dotted). The filled circles are stars with E(K − L ) > 0.2.

excess in a large sample of spectroscopically confirmed objects near and below the sub-stellar boundary in several nearby star-forming regions (Jayawardhana et al. 2003). Our longer-wavelength observations are much better at detecting disk excess above the photospheric emission and are less susceptible to the effects of disk geometry and extinction corrections than JHK studies (also see Liu, Najita & Tokunaga 2003). We find disk fractions of 40%–60% in IC 348, Chamaeleon I, Taurus and Upper Scorpius regions. Based on ISO observations, Natta & Testi (2001) have already shown that Cha Hα 1, ChaHα 2 and ChaHα 9 harbor mid-infrared spectral energy distributions consistent with the presence of dusty disks. ChaHα 2, which shows a large K − L excess (0.97 mag) in our data is a probable close (∼0.2”) binary with roughly equal-mass companions (Neuh¨auser et al. 2002). It is possible that a few of our targets harbor infrared companions that contribute to the measured excess, but this is unlikely in most cases. The disk fractions we report for IC 348 and Taurus are lower than those found by Liu, Najita & Tokunaga (2003). This is primarily because we use a more conservative criterion of K − L > 0.2 for the presence of optically thick disks whereas Liu et al. consider all objects with K −L > 0 as harboring

256

OPEN ISSUES IN LOCAL STAR FORMATION

disks. In IC 348, our disk fraction is comparable to that derived from Hα accretion signatures in high-resolution optical spectra (Jayawardhana, Mohanty & Basri 2003). However, in Taurus and Upper Sco, which may be slightly older at ∼2-5 Myrs, we find K − L excess in ∼50% of the targets whereas only three out of 14 Taurus VLM objects and one out of 11 Upper Sco sources exhibit accretion-like Hα (Jayawardhana, Mohanty & Basri 2002; 2003). This latter result suggests that dust disks may persist after accretion has ceased or been reduced to a trickle, as also suggested by Haisch, Lada & Lada (2001). In the somewhat older (∼5 Myr) σ Orionis cluster, only about a third of the targets show infrared excess. Neither of the two brown dwarf candidate members of the ∼10-Myr-old TW Hydrae association (Gizis 2002) shows excess. Gizis (2002) reported strong Hα emission (equivalent width ≈ 300 ˚ A) from one of the TW Hydrae objects, the M8 dwarf 2MASSW J1207334-393254, and suggested it could be due to either accretion or chromospheric activity. Given the lack of measurable K − L excess in this object, accretion now appears less likely as the cause of its strong Hα emission. Our findings in the σ Ori and TW Hya associations, albeit for a small sample of objects, could mean that the inner disks are clearing out by the age of these groups. Similar results have been found for T Tauri stars in the TW Hydrae association (Jayawardhana et al. 1999). Our results, and those of Muench et al. (2001), Natta et al. (2002), and Liu, Najita & Tokunaga (2003) show that a large fraction of very young brown dwarfs harbor near- and mid-infrared excesses consistent with dusty disks. While the samples are still relatively small, the timescales for inner disk depletion do not appear to be vastly different between brown dwarfs and T Tauri stars. Far-infrared observations with the Space InfraRed Telescope Facility and/or the Stratospheric Observatory For Infrared Astronomy will be crucial for deriving the sizes of circumsub-stellar disks and providing a more definitive test of the ejection hypothesis for the origin of brown dwarfs.

Acknowledgments We would like to acknowledge the great cultural significance of Mauna Kea for native Hawaiians, and express our gratitude for permission to observe from its summit. We thank Geoff Marcy for useful discussions, constant encouragement and access to Keck (for the L observations) and Kevin Luhman for valuable assistance during the December 2002 Keck run. We are grateful to the staff members of the VLT, Keck and IRTF observatories for their outstanding support. We also thank Fernando

Brown Dwarf Disks and Accretion (Jayawardhana et al.)

257

Comer´on and staff at the ESO User Support Group for their prompt responses to our queries. This work was supported in part by NSF grants AST-0205130 to R.J. and AST-0098468 to G.B.

References Bessell, M. S. & Brett, J. M. 1988, PASP, 100, 1134 Gizis, J.E. 2002, ApJ, 575, 484 Haisch, K.E.,Jr., Lada, E.A., & Lada, C.J. 2001, 553, L153 Jayawardhana, R., et al. 1999, ApJ, 521, L129 Jayawardhana, R., Mohanty, S., & Basri, G. 2002, ApJ, 578, L141 Jayawardhana, R., Mohanty, S., & Basri, G. 2003, ApJ, in press Jayawardhana, R., Ardila, D.R., Stelzer, B., & Haisch, K.E., Jr. 2003, AJ, in press Leggett, S. K., et al. 2002, ApJ, 564, 452 Liu, M.C., Najita, J. & Tokunaga, A.T. 2003, ApJ, 585, 372 Muench, A.A., et al. 2001, ApJ, 558, L51 Natta, A. & Testi, L. 2001, A&A, 376, L22 Natta, A., et al. 2002, A&A, 393, 597 Neuh¨ auser, R., et al. 2002, A&A, 384, 999 White, R. J., & Basri, G. 2003, ApJ, 582, 1109

J. Bouvier, J. L´epine, S. Mohanty, R. Jayawardhana, C. Melo

258

OPEN ISSUES IN LOCAL STAR FORMATION

K. Brooks, S. Mohanty, J. Alves

D. Nickeler, M. Kraus, F. Nieva, V. Lyra, M. Fernandes, M. Corti

SURFACE GRAVITY & MASS IN YOUNG BROWN DWARFS AND PLANEMOS Subhanjoy Mohanty Harvard-Smithsonian Center for Astrophysics, USA

Ray Jayawardhana Department of Astronomy, University of Michigan, USA

Gibor Basri Astronomy Department, University of California at Berkeley, USA

France Allard ´ CRAL, Ecole Normale Sup´erieure, France

Peter Hauschildt Hamburger Sternwarte, Universitaet Hamburg, Germany

David Ardila Bloomberg Center, Johns Hopkins University, USA

Abstract

We analyse high-resolution optical spectra of a sample of low-mass, very young, mid-to late M stellar and substellar objects in Upper Scorpius and Taurus. Precise effective temperatures (± 50K) and surface gravities (± 0.25 dex) are derived from a multi-feature spectral analysis using TiO, NaI and KI, through comparison with the latest synthetic spectra. Our high-resolution analysis is independent of extinction estimates. Combining our Tef f and gravities with distance and photometry information then allows us to derive extinctions, masses, radii and luminosities. We emphasize that our analysis is independent of theoretical evolutionary models. As such, it offers a new way of deriving fundamen259

J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 259-266. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

260

OPEN ISSUES IN LOCAL STAR FORMATION tal parameters for low-mass stellar and substellar objects, and of testing the model evolutionary tracks. We find discrepancies between our results and the theoretical tracks for the coolest, lowest mass targets. Our gravities for these objects are significantly (∼ 0.75 dex) lower than for hotter ones. This is at odds with the predictions of the evolutionary models, and causes improbably low ages to be derived from the theoretical tracks for the coolest objects. Similarly, our radii and luminosities for the coolest, lowest mass objects are much larger than predicted by the evolutionary models. These results lead us to suggest that significant improvements in the evolutionary models may be necessary for very cool, ultra-low mass objects, perhaps in the treatment of convection and/or initial conditions. We also find two of our coolest targets to be planetary mass objects (“planemos”), with mass ∼ 7 MJ . Thus, some late-M PMS objects may be much less massive than current evolutionary models indicate. Equivalently, young planemos may be more luminous than expected. Combined with the detection of planemos by other groups, our results imply that such objects may not be too rare.

Introduction Currently, masses of low-mass stars and substellar objects are largely inferred by comparing observables such as temperature and luminosity to the predictions of theoretical evolutionary tracks. Consequently, constructing accurate evolutionary models is now one of the holy grails of the star-formation discipline, and testing such models is of paramount importance. The simplest test is to derive dynamical masses for the observed components of multiple systems with known orbital parameters, and compare them to the theoretical predictions. In very low mass stars and brown dwarfs, however, there is a paucity of suitable multiple systems known. This is particularly true in the substellar regime: to date, no directly detected object beyond the Solar System has been proven to be substellar by a dynamical measurement of its mass. Dynamical masses have been determined for planemos in circumstellar orbits, but none of these have been directly observed, so the theoretical models cannot be checked through a comparison with their other properties. This situation is especially troubling at very young ages, since even the identification of objects as substellar currently depends, at such ages, on the theoretical tracks (empirical tests of substellarity are largely inapplicable to very young objects). Moreover, the low-mass tracks are most uncertain at precisely such early times (Baraffe et al., 2002), so testing them for young objects is particularly crucial. To address this issue, we have developed a technique for calculating masses for young cluster objects from surface gravity measurements, independent of theoretical evolutionary tracks. We apply it to a sample of M-type bona-fide

Young Brown Dwarfs and Planemos (Mohanty et al.)

261

PMS members of Upper Scorpius and Taurus. Our methods and results are summarized here, and detailed in Mohanty et al. (2003a & b).

1.

Method of Analysis

We first derive rotation velocities (vsini) by cross-correlation with a non-rotating template. Effective temperature (Tef f ) and gravity (log[g]) are then determined by comparing our high-resolution optical spectra to the rotationally broadened synthetic spectra (described below). M-type spectra exhibit a number of very temperature-sensitive (and gravityinsensitive) TiO bandheads. Fitting the observed profile of the tripleheaded band at ∼ 8440 ˚ A allows us to fix Tef f with a precision of ± 50K. Gravity is fixed by simultaneously fitting the profile of the gravitysensitive NaI doublet at ∼ 8200 ˚ A. The gravity-sensitive KI doublet at ∼ 7700 ˚ A provides an independent check on the gravity derived from NaI. The fact that we are able to fit very well all our spectral regions (TiO bandheads, NaI and KI doublets, and surrounding continuua) with synthetic spectra at the same Tef f and gravity, supports the validity of both the model spectra and our derived parameters (Fig.1). Since we consider fairly narrow wavelength regions (∼2600K, log[g]>∼3.0), no dust forms in the models, i.e., the “Cond” and “Dusty” models are indistinguishable. Finally, the models use a mixing-length parameter α=2, in agreement with the results of recent 3-D simulations of convection in M type objects (Ludwig et al., 2002; Ludwig 2003). With Tef f and gravity in hand, we derive extinctions (AV ) by comparing the observed [RC -IC ] colors to those predicted by the synthetic spectra (for the derived Tef f and log[g]). The average distances to Upper Sco and Taurus are known (∼ 140 pc); combining this with the AV , observed IC flux, and predicted IC surface flux from the synthetic spec-

262

OPEN ISSUES IN LOCAL STAR FORMATION

tra, we derive radii. Mass is determined from the radius and gravity, and luminosity by combining radius and Tef f . Our errors are ∼ 25% in radius, a factor of 2 in mass, and ∼ 50% in luminosity.

Figure 1. Best fits (red) to some of our targets (black); all spectra smoothed by 3pixel boxcar. Top to bottom: USco 112, 55, 53 and 130. Left to right, for each object: TiO, NaI doublet (zoom-in), entire NaI order (includes surrounding continuum) and the red lobe of the KI doublet (blue lobe is heavily telluric contaminated, so not fitted). Fit parameters shown in plot. For a given object, good fits are found to all the spectral regions shown at the same Tef f and log[g]. (Telluric lines - dotted absorption lines in the NaI doublet section, and the sharp deep lines in the NaI continuum - are not included in the model spectra, and are excluded from the fitting procedure). In the KI lobe of USco130 (bottom right), the lower S/N (illustrated by the unsmoothed spectrum shown) makes the fit look worse. However, the fits in the less noisy TiO and NaI sections are good, supporting our derived Tef f and log[g].

2. 2.1

Results Effective Temperature & Gravity

In Fig.2, we compare our derived Tef f and log[g] to the predictions of the widely-used evolutionary models of Chabrier et al. 2000 [CBAH00]. For Upper Sco objects with Tef f >∼ 2750K, our derivations are consistent with tracks at 5±2 Myr, the expected age for Upper Sco (Preibisch

Young Brown Dwarfs and Planemos (Mohanty et al.)

263

Figure 2. log[g] vs Tef f . The two lowest gravity USco objects (which we find to be planemos; Figs. 3 & 4) are shown as open diamonds, all other Usco targets as filled circles, and GG Tau Ba & Bb as crosses. Our error bars on T ef f and log[g] are shown on the bottom right (±50K, ±0.25dex). Solid curves are CBAH00 theoretical tracks at various ages (1,2,3,5, & 10 Myr, from top to bottom); dotted curves are the evolutionary paths predicted by CBAH00 for various masses (0.20-0.01 solar masses, left to right). The arrows (bottom left) show the maximum change in derived T ef f (150K), log[g] (0.25dex), and combined position on this plot, expected from cool spots.

et al, 2002). The hotter Taurus object, GG Tau Ba (∼ 2800K), is also compatible (within our errors) with tracks at ∼1 Myr, the expected age of Taurus (White et al, 1999). However, the cooler USco objects (Tef f < 2750K) are completely inconsistent with ∼5 Myr tracks; at the much lower log[g] we derive for them, the tracks imply an age less than even 1 Myr. The cooler Taurus object, GG Tau Bb (∼ 2600K), also appears much younger than the ∼ 1 Myr expected for Taurus. Line-profile variations due to cool surface spots cannot be responsible for the ∼ 0.75 dex gravity variation we find between the hotter and cooler Usco objects: (1) cool spots might appear in all our stars, not just in the cooler ones, and (2) even the maximum estimated error in Tef f and log[g] due to very large spots is insufficient to explain our discrepancy with the tracks (see Fig.2). Line-profile changes due to drastic changes in H2 broadening or in the mixing length are also not expected over the fairly narrow range of Tef f we find. Our derived gravities, and the resultant discrepancy with the tracks in cooler objects, thus appear real.

264

2.2

OPEN ISSUES IN LOCAL STAR FORMATION

Mass, Radius & Luminosity

Figure 3. Mass-Radius plot. Planemo, BD and stellar mass regimes marked. 2 Usco planemos shown as open diamonds, all other USco targets as filled circles, and GG Tau Ba & Bb as crosses. CBAH00 theoretical tracks for 1,2,3,5 & 10 Myr overplotted. Arrows at bottom show the resultant change in derived radius, mass and combined position on mass-radius plot, if any of our objects were equal-mass binaries. Note the larger-than-predicted radii derived for the lowest mass USco and Taurus objects.

In Fig.3, we compare our derived masses and radii to the CBAH00 predictions. We see that (1) at the higher masses, our results agree with the tracks for the expected ages (5±2 Myr for Usco, ∼1 Myr for Taurus), while at the lowest masses, our derived radii are considerably larger than expected (the latter are the same objects that also disagree with the predicted gravity; Fig.2), (2) the two lowest mass Usco objects appear to be planemos (∼6-7 MJ ) - even including our mass errors puts then at the planemo / brown dwarf boundary, and (3) our radii for the younger Taurus objects are slightly larger than for similar-mass but older Usco objects, ie, we find contraction with age, as expected from general considerations; this attests to the general validity of our results. In Fig.4, we plot our derived masses and luminosities against the CBAH00 predictions. Again, there is a discrepancy at the lowest masses: while the luminosities of higher masses are compatible with the tracks for USco and Taurus ages, the lowest masses appear significantly more luminous (primarily because their radii are much larger than predicted;

Young Brown Dwarfs and Planemos (Mohanty et al.)

265

Figure 4. Mass-Luminosity plot. All symbols otherwise same as in Fig.3. Arrows at bottom show the resultant change in derived luminosity, mass and combined position on this plot, if any of our objects were equal-mass binaries. Note the larger-thanpredicted luminosities derived for the lowest mass USco and Taurus objects.

Fig.3). In other words, our derived mass-luminosity relationship is notably shallower than in the evolutionary models, for the coolest, lowest mass objects. We point out that binarity cannot account for this discrepancy. If any of our objects were equal-mass binaries, their components would slide, in the mass-radius and mass-luminosity plots, roughly parallel to the theoretical tracks (see Figs.3 & 4); objects that do not fall on the tracks now would remain discrepant.

3.

Implications & Conclusion

Our derived parameters, for low-mass stellar and substellar objects in Upper Sco and Taurus, are significantly at odds with the predictions of evolutionary models for the coolest, lowest mass objects. The latter appear to have larger radii and luminosities than predicted for Upper Sco and Taurus ages; i.e., they appear much younger than the hotter, more massive objects when compared to the theoretical tracks. However, a real age difference between the hotter and cooler objects appears unviable: for example, the Taurus targets, GG Tau Ba and Bb, belong to the same stellar system and are thus very likely to be coeval.

266

OPEN ISSUES IN LOCAL STAR FORMATION

A possible resolution may lie in the treatment of convection and initial conditions in the widely-used CBAH00 evolutionary models. CBAH00 use a mixing-length parameter of α=1. Recent full 3-D hydrodynamic simulations of convection in PMS M-types indicate that α=2 is more appropriate (Ludwig 2003); simultaneously, Baraffe et el. (2002) have shown that a change in the mixing length at early ages has significant effects on the evolutionary path of very low masses. CBAH00 also assume an initial gravity of 3.0-3.5; Baraffe et al. (2002) have shown that decreasing this value also affects the tracks significantly. Finally, the latest, improved opacities included in our synthetic spectra have not been incorporated into CBAH00. We suggest that a combination of these factors may bring the model tracks into better agreement with our results. Finally, we find two of our objects to be planemos; though the faintest of our targets, they are much brighter than expected for such low masses. If we are correct, then some faint late-M PMS objects may be much less massive than expected; conversely, given their overluminosity compared to predictions, planemos may be easier to detect than thought. Other groups have also recently identified young, isolated planemos; understanding how such isolated, ultra-low masses form presents a challenge.

Acknowledgments The authors wish to thank Drs. I. Baraffe and G. Chabrier for extremely useful discussions, and Dr. R. White for the Taurus spectra.

References Allard,F., Hauschildt,P.H., Alexander,D.R., Tamanai,A., Schweitzer,A., 2001, ApJ, 556, 357 Baraffe,I., Chabrier,G., Allard,F., Hauschildt,P.H., 2002, A&A, 382, 563 Chabrier,G., Baraffe,I., Allard,F., Hauschildt,P.H., 2000, ApJ, 542, 464 [CBAH00] Ludwig,H.-G., Allard,F. & Hauschildt,P., 2002, A&A, 395, 99 Ludwig,H.-G., 2003, “Modeling of Stellar Atmospheres”, IAU Symposium, accepted Mohanty,S., Basri,G., Jayawardhana,R., Allard,F., Hauschildt,P.H., David,A., 2003a, ApJ, submitted Mohanty,S., Jayawardhana,R. & Basri,G., 2003b, ApJ, submitted Preibisch,T., Brown,A.G.A., Bridges,T., Guenther,E., Zinnecker,H., 2002, AJ, 124, 404 White,R.J., Ghez,A.M., Reid,I.N., Schultz,G., 1999, ApJ, 520, 811

IV

DISKS, OUTFLOWS AND JETS

This page intentionally left blank

FU ORIONIS ERUPTIONS AND THE FORMATION OF CLOSE BINARIES Bo Reipurth Institute for Astronomy, University of Hawaii, USA [email protected]

Abstract The discovery of close pairs of stars where both are FU Orionis objects poses important constraints on the triggering mechanism for these dramatic eruptive events among pre-main sequence stars. It is argued here that FUor eruptions are signposts of the last stages of orbital evolution as a close binary is being formed. In this view, FUors represent an evolutionary phase immediately following the formation of Herbig-Haro jets.

Introduction Although pre-main sequence binaries have been known for a long time (e.g. Joy & van Biesbrock 1944; Herbig 1962), it is only within the last decade that major surveys have been carried out, which have demonstrated the ubiquity and importance of binaries in early stellar evolution (e.g. Reipurth & Zinnecker 1993; Ghez, Neugebauer, & Matthews 1993; Leinert et al. 1993). Recently, attention has been focused on the presence of triple and multiple systems among pre-main sequence stars, opening up the rich and complex world of dynamical interactions (e.g. Sterzik & Durisen 1995; Reipurth 2000; Brandeker, Jayawardhana, & Najita 2003). The formation of binaries and multiple stars is a natural outcome of fragmentation during protostellar collapse (e.g. Bodenheimer et al. 2000), but the formation of close binaries is still not fully understood (e.g. Tokovinin 2003), although sophisticated numerical simulations can offer important insights into the problem (Bate, Bonnell, & Bromm 2002a). In this presentation, a possible relation between FU Orionis eruptions and the formation of close binaries is explored.

269 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 269-278. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

270

1.

OPEN ISSUES IN LOCAL STAR FORMATION

FUor Eruptions

The FU Orionis eruptive variables are located in regions of star formation and are associated with dark clouds and reflection nebulae (Herbig 1977). The two best studied members of the class, FU Ori and V1057 Cyg, both exhibited increases in optical brightness by 5-6 magnitudes on timescales of about a year, followed by a much slower fading (Wachmann 1939; Herbig 1966, 1977; Kolotilov & Petrov 1985; Ibragimov & Shevchenko 1987). Post-eruption optical spectra show peculiar supergiant F-G type spectra, with blueshifted shell components, and P Cygni profiles in Hα and the Na I resonance lines (Herbig 1977; Bastian & Mundt 1985). The spectral type becomes gradually later with increasing wavelength, showing molecular absorption bands characteristic of late-type stars in the near infrared (Mould et al. 1978). A model in which accretion through the circumstellar disk is greatly increased can explain many observed peculiarities of FU Ori objects, including the broad spectral energy distributions, the variation of spectral type and rotational velocity with wavelength of observation, and many “double-peaked” spectral lines in the optical and the near-infrared (Hartmann & Kenyon 1985,1987a,b; Kenyon, Hartmann, & Hewett 1988). The onset of increased mass transfer through the disk may be triggered by a thermal instability (Bell & Lin 1994), or by the close passage of a companion star (Bonnell & Bastien 1992; Clarke & Syer 1996). While the accretion disk model is successful in explaining many of the observed properties of FUors, some discrepancies remain. It has been suggested by Herbig (1989) and Herbig, Petrov & Duemmler (2003) that a number of observations are better accounted for by a rapidly rotating star near the edge of stability, a concept that has been discussed by Larson (1980). Up to now only a few objects have been generally accepted as members of the FU Ori class, including some for which no eruption has been observed, but which are spectroscopically very similar to the “classical” FUors. Besides FU Ori and V1057 Cyg, these include V1515 Cyg (Herbig 1977), V1735 Cyg = Elias 1-12 (Elias 1978), V346 Nor (Graham & Frogel 1985; Reipurth 1985b), L1551 IRS5 (Mundt et al. 1985; Carr, Harvey & Lester 1987; Stocke et al. 1988), Z CMa (Hartmann et al. 1989; Hessman et al. 1991), and BBW 76 (Reipurth 1985a, 1990; Eisl¨ offel et al. 1990; Reipurth et al. 2002). Several other, more embedded objects have the deep, broadened infrared CO bands in absorption which is characteristic of FUors, including RNO 1B/C and Parsamian 21 (Staude & Neckel 1991,1992; Kenyon et al. 1993), V883 Ori (Strom & Strom 1993; Reipurth & Aspin 1997), Haro 5a IRS, HH 354 IRS and

FU Orionis Eruptions and Close Binary Systems (Reipurth)

271

HH 381 IRS (Reipurth & Aspin 1997), PP 13S (Sandell & Aspin 1998), and AR6A and AR6B (Aspin & Reipurth 2003). Because of the small number of known objects, the frequency of FU Ori eruptions is still poorly known. Herbig (1977,1989) and Hartmann & Kenyon (1985) concluded that FU Ori eruptions must occur many times during the T Tauri phase of a star, suggesting mean times of 104 - 105 yr between successive outbursts. With this frequency and with decay-times of perhaps several hundred years it seems probable that eruptions could have taken place before the times of photographic sky surveys, leaving stars in decaying but still elevated states. As our understanding of the FUor phenomenon increases, we may therefore be able to find and identify such objects. Since the FUor phenomenon is mainly defined by a characteristic outburst behavior (Herbig 1977), we refer in the following to objects as FUor-like if they have spectroscopic FUor characteristics, but no outburst was observed (Aspin & Reipurth 2003). For a detailed review of the FU Ori phenomenon, see Hartmann & Kenyon (1996).

2.

Binary FUors

In a recent infrared study of an area in the NGC 2264 region, Aspin & Reipurth (2003) found two infrared sources, AR 6A and 6B, both of which have spectroscopic characteristics of FUors. Their separation is only 2.8 , corresponding to a projected separation of 2240 AU at the assumed distance of 800 pc. This is well within the binary separation distribution function, and the two are likely to form a physical binary. Only one other pair of FUor-like objects is known, the RNO 1B/1C binary in Lynds 1287 (Kenyon et al. 1993). Given the rarity of FUors, finding two FUor-like objects within a few arcseconds of each other in any star forming region is exceedingly improbable. In other words, whatever has triggered the FUor outburst in AR 6A seems likely to be somehow connected to whatever triggered the FUor outburst in AR 6B. In this connection, it is of interest to note that AR 6A has a faint companion at an angular separation of 0.85 . If this is a physical companion, their projected separation amounts to ∼700 AU. The possibility therefore exists that AR 6A, B and C form a triple system which therefore, in principle, could have interacted leading to the current disturbed states of AR 6A and 6B. However, the separation of AR 6A and AR 6B is sufficiently large that for typical bound or unbound motions it would have taken more than 104 years to reach their current positions from the moment of closest approach. It seems unreasonable

272

OPEN ISSUES IN LOCAL STAR FORMATION

to assume that the present FUor states could have persisted for such a length of time. This apparent conundrum has a solution if we assume that AR 6A and 6B each represent an unresolved binary, i.e. that AR 6A and 6B form a quadruple system. Such systems are most likely born as non-hierarchical multiples which evolve dynamically. These configurations are unstable, and sooner or later, during a close encounter of all the components when energy and angular momentum can be exchanged, the components will be separated into a hierarchical quadruple, with two pairs of much closer binaries. The hierarchical quadruple may be bound or unbound. This transforming event is statistically most likely to occur at such an early stage that all four stars still have substantial circumstellar material surrounding them (Reipurth 2000). The binary star pairs will evolve with significant viscous interactions between their disks, leading to angular momentum transfer and consequent orbital shrinkage (e.g. Artymowicz & Lubow 1996). It appears that we can understand the improbable pairing of two FUor-like objects in both AR 6A/B and RNO1B/C as a consequence of the formation of a hierarchical quadruple system followed by orbital shrinkage until the individual binary components are so close as to trigger disk instabilities leading to FUor eruptions. In an important paper, Bonnell & Bastien (1992) explored the role of binary companions in triggering FUor outbursts. Their model of tidally induced accretion in an eccentric binary system is directly applicable to the present discussion. There is only one major difference between the scenario discussed here and the Bonnell & Bastien model, namely the binary separation. Whereas Bonnell & Bastien considered peri-astron passages of binaries like Z CMa with semi-major axes of thousands of AU, we are here concerned with the formation of close binaries. In another study, Clarke & Syer (1996) considered how a very low mass companion in a close orbit can act as a flood gate, storing material upstream and periodically releasing it. The properties and evolution of small multiple systems are discussed in the following section.

3.

The Evolution of Multiple Systems

The chaotic dynamics which characterize triple systems have no analytical solutions and can only be explored through numerical simulations in a statistical manner. A vast literature exists on the subject, and the subject was recently reviewed in IAU Colloquium No. 191 The Environment and Evolution of Double and Multiple Stars. The motions of the three bodies while they are bound together can basically be divided

FU Orionis Eruptions and Close Binary Systems (Reipurth)

273

into three classes, which interchange at random until the system decays through escape of a member. One class is the interplay, during which the three members perform completely random motions with no periodicity. During interplay two members, often the two most massive, can form a temporary binary, which have frequent two-body encounters, while continuously being perturbed by approaches by the third member, sometimes in the form of a flyby, but occasionally resulting in exchange of binary membership. A second class is the close triple approach, in which all three bodies briefly and more or less simultaneously are brought close together. It is during such events that energy is exchanged between the components. Statistical studies show that close triple approaches are necessary but not sufficient conditions for escape. The third class of motion is the ejection, which will occur following a close triple approach during which a binary is formed that absorbs the potential energy of the third member thereby loosening its ties to the system. At this point the triple transforms into two two-body systems, namely a close binary and an ejected member moving on an approximately elliptic orbit, which eventually brings it back to the close binary. If instead the ejected member moves on a hyperbolic orbit, the ejection becomes an escape. Statistical studies show that it is usually, but not always, the lightest member that is ejected; the escape probability scales approximately as the inverse third power of the mass. Sometimes the decay can occur so early in the collapse phase that the ejected member has not reached the hydrogen burning limit, and it will thus become a free-floating brown dwarf (Reipurth & Clarke 2001; Bate, Bonnell, & Bromm 2002b). Although eventually most non-hierarchical triple systems disintegrate in an escape, this is clearly not always the case, as the presence of stable hierarchical triple systems among main sequence stars demonstrates (e.g. Tokovinin 1997). Finally, it is important to note that numerous studies have shown that the close binaries which form as a result of the escape of a third member usually are highly eccentric, with eccentricities exceeding 0.9 not being unusual (e.g. Valtonen & Mikkola 1991). It is not possible to define the precise lifetime of a triple system, because it can be shown that the decay of a large number of triple systems occurs randomly. However, approximately within a hundred crossing times about 95% of the triple systems have decayed (Sterzik & Durisen 1995,1998; Armitage & Clarke 1997). For parameters which are likely to hold among newly born multiple stars, decay times are of the order of several times 104 yr. In a recent study, Reipurth (2000) postulated that the dynamical decay of triple or multiple systems leads to strong outflow activity. Mas-

274

OPEN ISSUES IN LOCAL STAR FORMATION

sive disk truncation results from close triple approaches, accompanied by large-scale accretion, with a consequent burst of outflow activity, which produces the observed giant HH bow shocks. Some of the material culled from the individual circumstellar disks may settle into a circumbinary disk around the newly bound stellar pair. The small remaining and truncated circumstellar disks are fed from the circumbinary disk through gas streams, and this as well as other dynamical effects cause the binary orbit to shrink (Artymowicz & Lubow 1996). Gas streams together with disk interactions at periastron drive cyclic accretion modulated on an orbital time scale. As the stellar components gradually spiral towards each other, the increasingly frequent mass loss events form chains of HH objects until eventually the binary has a semi-major axis of only 9-12 AU, at which point the closely spaced shocked ejecta appear as a finely collimated jet. Thus, such HH flows can be read as a fossil record of the evolution of orbital motions of a binary, newly formed in a triple disintegration event, as it shrinks from a typical separation of 100 AU or more to 10 AU or less.

4.

FUors as Close Binaries: Open Questions

I here argue that the existence of several visual pairs of FUors is statistically so improbable that one must conclude that the eruptions must be synchronized. This strongly suggests that the pairs at an earlier stage were interacting much more closely. Since the time scale to reach their present separation is so long that it is unlikely that the presently observed disk disturbances originated at that time, it follows that each of the FUors are likely to consist of a close binary, so the whole system is a quadruple one. The original transformation of a non-hierarchical quadruple into two pairs of newly formed binaries in a hierarchical configuration starts a process of orbital decay driven by viscous dissipation due to circumstellar material. In this picture FUor outbursts are the final visible manifestation of the spiraling in of two components about to form a close binary. This view of FUors leads to several corollaries as well as open questions. More details are given in a forthcoming paper. 1. FUors should always be very close binaries with semi-major axes of several AU or less (possibly with a few rare exceptions, see Point 4 below). This may be very difficult to verify observationally, since at the distances of most FUors, only with interferometry can we hope to resolve so close pairs, see, e.g., Malbet et al. (1998). 2. Close visual pairs of FUor-like objects are likely to arise from smallN clusters. But does that imply that all FUors originate in triples

FU Orionis Eruptions and Close Binary Systems (Reipurth)

275

or higher-order multiples? Probably not. The break-up of a smallN cluster will harden the binary or binaries involved and thus facilitate their subsequent orbital evolution towards becoming a close binary (e.g. Bate et al. 2002a). But this does not mean that two stars on their own cannot evolve to become a close binary, since this is mainly driven by the amount of circumstellar material available to transfer angular momentum. Only in those cases where the orbital decay has been triggered by the disintegration of a small multiple system, would one expect to find the FUor associated with another young star or close binary. 3. In a recent paper, Herbig et al. (2003) have speculated that FUor events may not be a property of ordinary T Tauri stars, but of a particular subset of them. In other words, not all young low mass stars may go through FUor phases, in contrast to widely accepted current views. At face value, this appears to be in contradiction to the statistical arguments of Herbig (1977) and Hartmann & Kenyon (1985). However, it should first be recalled that the observational basis for the statistics is highly uncertain, as the authors themselves have emphasized. Furthermore, while fewer stars are involved in FUor outbursts in this new picture, this may be compensated by them going through several and perhaps even many FUor eruptions, depending on the particular orbital characteristics of each binary. This is consistent with the assumption that FUors represent a late stage in the formation of a close binary. We may get an estimate of the frequency of stars formed as close binaries by examining evolved stars, and for high velocity G and K dwarfs Latham et al. (2002) found that 15% are spectroscopic binaries. 4. Could the orbital decay lead to the actual merger of two stars? It is a common characteristic of binaries newly formed in a dissolving multiple system to have extremely eccentric orbits. Furthermore, eccentricity is not a fixed parameter for a newborn binary, but varies as the orbit evolves (e.g. Kozai 1962). And there are no obvious reasons why the eccentricity should not be able to grow large enough to permit periastron passages of less than the sum of the two stellar radii, leading to the merger of the two stars. This would temporarily lead to a bloated star rotating near break-up speed, as advocated by Larson (1980) and Herbig et al. (2003) on theoretical and observational grounds, respectively. However, the energy release, and thus the luminosity, would briefly get much higher than observed, so unless such an event had been missed, it

276

OPEN ISSUES IN LOCAL STAR FORMATION

is unlikely that any of the FUor eruptions observed to date involves a merger.

5. It is well established that FUors have powerful mass loss, and yet none of the classical FUors like FU Ori, V1057 Cyg, or V1515 Cyg are associated with Herbig-Haro (HH) flows. The disk model for FUors indicates that the objects have extremely high accretion rates of up to 10−4 M
yr−1 and so heavy mass loss is expected. But somehow the ability to collimate this mass loss has been lost. In the stellar dynamics jet hypothesis advocated by Reipurth (2000), HH jets are the signposts of a decaying binary system that has contracted to a semimajor axis of the order of 10 AU or less. Further orbital decay leads to the disruption of the inner disk regions which anchor the organized magnetic fields that are believed responsible for the high collimation of HH jets. In this view, FUor eruptions is an evolutionary state that follows the phase of jet formation. There is probably a rather diffuse transition between these two phases, and some FUors could be associated with the last vestiges of HH jets. Z CMa drives a finely collimated jet, but this is actually known to be driven by its infrared companion (Garcia et al. 1999). At least one of the components of L1551 IRS5 is a FUor-like object. Both components drive small jets, suggesting that IRS5 is a quadruple system in the process of achieving its final configuration.

6. The processes that transform a non-hierarchical group of newborn stars into a hierarchical system, whether bound or unbound, are highly stochastic in nature, and can occur at different stages of early evolution. This right away suggests that there are many possible outcomes, and that we are unlikely to find that all FUors will fit into a single well defined mold. Indeed, even the classical FUors first discussed by Herbig (1977) show significant differences in their lightcurves, the light curve of FU Ori being steeply rising, V1057 Cyg having a somewhat slower rise, and V1515 Cyg showing a very gentle rise. All three also have different behaviors in their decay phases. It may well be that the classical FUors represent the most extreme cases of a general phenomenon that can give rise to a wide range of mass loss and time-dependent behavior and that may account for objects as disparate as FU Ori and EX Lupi (Herbig 1977, 1989).

FU Orionis Eruptions and Close Binary Systems (Reipurth)

277

Acknowledgments I am thankful to George Herbig, Colin Aspin, and Cathie Clarke for many illuminating discussions.

References Armitage, P.J., & Clarke, C.J. 1997, MNRAS, 285, 540 Artymowicz, P., Lubow, S.H. 1996, ApJ, 467, L77 Aspin, C.A., & Reipurth, B. 2003, AJ, in press Bastian, U., & Mundt, R. 1985, A&A, 144, 57 Bate, M.R., Bonnell, I.A., Bromm, V., 2002a, MNRAS, 336, 705 Bate, M.R., Bonnell, I.A., Bromm, V., 2002b, MNRAS, 332, L65 Bell, K.R., & Lin, D.N.C. 1994, ApJ, 427, 987 Bodenheimer, P., Burkert, A., Klein, R.I., & Boss, A.P. 2000, in Protostars and Planets IV, ed. V. Mannings, A. Boss, & S. Russell (Tucson: Univ. of Arizona Press), 675 Bonnell, I.A., & Bastien, P. 1992, ApJ, 401, L31 Brandeker, A., Jayawardhana, R., & Najita, J. 2003, AJ, in press Carr, J.S., Harvey, P.M., & Lester, D.F. 1987, ApJ, 321, L71 Clarke, C.J., Syer, D. 1996, MNRAS, 278, L23 Elias, J.H. 1978, ApJ, 223, 859 Eisl¨ offel, J., Hessman, F.V., & Mundt, R. 1990, A&A, 232, 70 Garcia, P.J.V., Thi´ebaut, E., & Bacon, R. 1999, A&A, 346, 892 Ghez, A.M., Neugebauer, G., & Matthews, K. 1993, AJ, 106, 2005 Graham, J.A., & Frogel, J.A. 1985, ApJ, 289, 331 Hartmann, L., & Kenyon, S.J. 1985, ApJ, 299, 462 Hartmann, L., & Kenyon, S.J. 1987a, ApJ, 312, 243 Hartmann, L., & Kenyon, S.J. 1987b, ApJ, 322, 393 Hartmann, L., & Kenyon, S.J. 1996, Ann. Rev. Astron. Astrophys., 34, 207 Hartmann, L., Kenyon, S.J., Hewett, R., Edwards, S., Strom, K.M., Strom, S.E., & Stauffer, J.R. 1989, ApJ, 338, 1001 Herbig, G.H. 1962, Advances Astron, Astrophys. 1, 47 Herbig, G.H. 1966, Vistas in Astronomy, 8, 109 Herbig, G.H. 1977, ApJ, 217, 693 Herbig, G.H. 1989, in ESO Workshop on Low Mass Star Formation and Pre-Main Sequence Objects, ed. B. Reipurth, p. 233 Herbig, G.H., Petrov, P.P., & Duemmler, R. 2003, ApJ, in press Hessman, F.V., Eisl¨ offel, J., Mundt, R., Hartmann, L.W., Herbst, W., & Krautter, J. 1991, ApJ, 370, 384 Ibragimov, M.A., & Shevchenko, V.S. 1987, Astrophysics, 27, 337 Joy, A.H., & van Biesbrock, G. 1944, PASP, 56, 123 Kenyon, S.J., Hartmann, L., & Hewett, R. 1988, ApJ, 325, 231 Kenyon, S.J., Hartmann, L., Gomez, M., Carr, J.S., & Tokunaga, A. 1993, AJ, 105, 1505 Kolotilov, E.A., & Petrov, P.P. 1985, Sov. Astron. Lett., 11, 358

278

OPEN ISSUES IN LOCAL STAR FORMATION

Kozai, Y. 1962, AJ, 67, 391 Larson, R.B. 1980, MNRAS, 190, 321 Latham, D.W., Stefanik, R.P., Torres, G., Davis, R.J., Mazeh, T., Carney, B.W., Laird, J.B., & Morse, J.A. 2002, AJ, 124, 1144 Leinert, C. et al. 1993, A&A, 278, 129 Malbet, F. et al. 1998, ApJ, 507, L149 Mould, J.R., Hall, D.N.B., Ridgway, S.T., Hintzen, P., & Aaronson, M. 1978, ApJ, 222, L123 Mundt, R., Stocke, J., Strom, S.E., Strom, K.M., & Anderson, E.R. 1985, ApJ, 297, L41 Reipurth, B. 1985a, in ESO/IRAM/ONSALA Workshop on (Sub)-Millimeter Astronomy, eds. P. Shaver, K. Kj¨ ar, p. 459 Reipurth, B. 1985b, A&A, 143, 435 Reipurth, B. 1990, in IAU Symp. No. 137 Flare Stars in Star Cluster, Associations and the Solar Vicinity, eds. L.V. Mirzoyan et al., p. 229 Reipurth, B. 2000, AJ, 120, 3177 Reipurth, B., & Aspin, C. 1997, AJ, 114, 2700 Reipurth, B. & Clarke, C.J. 2001, AJ, 122, 432 Reipurth, B., & Zinnecker, H. 1993, A&A, 331, 977 Reipurth, B., Hartmann, L., Kenyon, S.J., Smette, A., Bouchet, P. 2002, AJ, 124, 2194 Sandell, G., & Aspin, C. 1998, A&A, 333, 1016 Staude, H.J., & Neckel, Th. 1991, A&A, 244, L13 Staude, H.J., & Neckel, Th. 1992, ApJ, 400, 556 Sterzik, M.F., Durisen, R.H. 1995, A&A, 304, L9 Sterzik, M.F., Durisen, R.H. 1998, A&A, 339, 95 Stocke, J., Hartigan, P.M., Strom, S.E., Strom, K.M., Anderson, E.R., Hartmann, L.W., & Kenyon, S.J. 1988, ApJS, 68, 229 Strom, K.M., & Strom, S.E. 1993, ApJ, 412, 63 Tokovinin, A.A. 1997, Astronomy Letters, 23, 727 Tokovinin, A.A. 2003, in Galactic Star Formation across the Stellar Mass Spectrum, eds. J. De Buizer & N. van der Bliek, ASP Conf. Ser., p.397 Valtonen, M., & Mikkola, S. 1991, Ann. Rev. Astron. Astrophys., 29, 9 Wachmann, A.A. 1939, Beob. Zirk., 21, 12

SEDS OF FLARED DUST DISKS Radiation Transfer Model Versus 2-Layer Model Michaela Kraus Sterrenkundig Instituut, Universiteit Utrecht, The Netherlands [email protected]

Abstract

A radiation transfer model for the top-layer of a flared dust disk around a low mass star is presented. The disk structure, temperature distribution and spectral energy distribution are compared with the results of the analytical two-layer model of Chiang & Goldreich. We find that both, a proper iterative calculation of the main parameters and proper radiation transfer calculations, are important to not overestimate the disk structure and its emission especially at long wavelengths.

Introduction Many T Tauri stars are observed with a flat spectral energy distribution (SED) in the near infrared (e.g. Beckwith et al. 1990). This is in contradiction with the standard disk which is usually modeled as a flat blackbody. The flat spectrum in the near infrared hints to an additional hot dust component in the circumstellar region of young stars. Several models have been proposed, e.g. a dusty halo (e.g. Miroshnichenko et al. 1999) or a dust shell beyond the outer edge of the disk (e.g. GregorioHetem & Hetem 2002). Both models can fit observed SEDs. However, it is not clear how such a configuration of the circumstellar dust should form. We are dealing here with a different location of the hot dust component, namely the surface of the circumstellar disk. For the surface to become much hotter than the interior, it must be inclined with respect to the infalling stellar radiation, i.e. the disk must be flared. Such a flaring can happen naturally when the disk is in hydrostatic equilibrium in z-direction with a temperature distribution that falls off more slowly than r−1 (Kenyon & Hartmann 1987). Chiang & Goldreich (1997, hereafter CG) developped an analytical two-layer model. Their flared dust disk contains an optically thick 279 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 279-286. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

280

OPEN ISSUES IN LOCAL STAR FORMATION

isothermal mid-layer and a superheated surface layer. CG provide simple analytical expressions to calculate the SEDs which are indeed flat in the near infrared, and their model is used by many people to fit the SEDs of young stellar objects. Our intention was to study the reliability of these analytical two-layer models. We developped a radiation transfer code for the upper disk part to calculate the parameters that determine the disk structure and temperature distribution self-consistently, and we compare our results with the CG model. Note that we compare our results to the original analytical CG model. A comparison between an improved two-layer model with a radiation transfer model has recently been done by Dullemond & Natta (2003).

1.

Description of the dust disk model

Our disk consists of an optically thick (τVmid  1) mid-layer which is isothermal in z-direction. The mid-layer is sandwiched by two toplayers which are marginally optically thick at visual wavelengths but top ≤ 1. still optically thin at IR wavelengths, i.e. τVtop ≥ 1 and τIR Consequently, the infalling stellar light is completely absorbed within the top-layer. The dust particles re-radiate the energy at IR wavelengths, so half the redistributed energy leaves the disk into space, and the other half penetrates the mid-layer and heats it. As a third distinct region we define the disk photosphere which encloses the uppermost part of the top-layer and whose location is defined by the parameter h (see Sect. 1.2). Further, we make the following assumptions: the disk is in hydrostatic equilibrium in z-direction, gas and dust are well mixed throughout the disk, and in the mid-layer gas (g) and dust (d) are in thermal equilibrium at the same temperature Tg = Td = Tmid . The emission and absorption of the gas component of the disk is ignored. We restrict the discussion to passive disks, which means that the only heating source is the star which illuminates the disk surface. In addition, our (gas) disk extends down to the stellar surface with no inner hole. The existence of such a hole would lead to a puffed-up inner rim and selfshadowing effects of the disk (see Dullemond et al. 2001 and Dullemond 2002).

1.1

Stellar illumination of the disk surface

For a razor-thin passive disk around a star with effective temperature T∗ and radius R∗ the monochromatic flux entering the disk perpendicular

281

SEDs of Flared Dust Disks (Kraus)

h

Figure 1 Sketch of a flared disk. The flaring term of the grazing angle is indicated. The inner radius of the dust disk is r0 .

α flare gr

P

r

0

r

to the surface at distance r from the star is 

fν⊥



R∗ R∗ = Bν (T∗ ) arcsin − r r




1−

R∗ r

2

 

(1)

where the star is assumed to emit a Planck spectrum. This flux can be parametrized in terms of the solid angle Ω under which the star is seen from a disk surface element at distance r and the so-called grazing angle αgr , i.e. the mean angle of incidence of the stellar flux fν⊥ = αgr ΩBν (T∗ )

(2)

The grazing angle of a razor-thin disk is therefore 

razor αgr




1 R∗ = arcsin Ω r





R∗ − r




1−

R∗ r

2

 

(3)

which for large distances reduces to the handy formula used by CG 4 R∗ R∗  0.4 (4) 3π r r This relation also holds for a wedge-shaped disk as long as its opening angle is small. In a flared disk the surface is curved and the grazing angle increases with distance from the star (see the the solid lines in Fig. 1). Therefore, the flared disk intercepts more stellar light and is heated more flare efficiently. The grazing angle of the flared disk is by the amount αgr greater than for a flat disk (Fig. 1), and the total grazing angle becomes razor + αflare This flaring term of the grazing angle, αflare , simply αgr = αgr gr gr can be computed from rR

razor αgr −→∗

flare αgr = arctan

dh h d − arctan  r dr r dr




h r

(5)

With this new grazing angle, which is a function of the location h of the photosphere, we can calculate the flux penetrating the flared disk in z-direction.

282

1.2

OPEN ISSUES IN LOCAL STAR FORMATION

The location of the disk photosphere

The photosphere encloses the uppermost part of the top-layer. Its onset is described by the parameter h, which is defined as the height z above the mid-plane where the visual optical depth along the direction of the infalling stellar light, i.e. along the grazing angle, equals 1. This leads to a vertical optical depth of τV⊥ = sin αgr .

(6)

Since the disk is assumed to be in hydrostatic equilibrium, τV⊥ can be calculated ∞ τV⊥



= κV ρ0 (r)

z2

e− 2H 2 dz

(7)

h

where ρ0 is the density in the mid-plane and H is the scale height of the disk given by

kTd H= r3/2 (8) GM∗ µ Here, M∗ and µ are the stellar mass and the mean molecular weight. From equaling Eqs. (6) and (7) h can be determined, but it is a function of temperature and grazing angle.

1.3

The temperature of the mid-layer

The downwards and upwards directed fluxes, F ↓ and F ↑ , are in equilibrium throughout the passive disk. We can calculate these two fluxes explicitely at the boundary between the isothermal mid-layer and the top-layer. We assume that half the incident stellar flux is re-radiated into space and the other half penetrates the mid-layer, leading to F↓ =

1 ⊥ 1 f = 2 2



αgr ΩBν (T∗ ) dν

(9)

The flux leaving the mid-layer in upward direction is given by ↑

F = 2π





Bν (Tmid ) 1 − e−τν /µ µ dµ dν

(10)

where Tmid is the isothermal temperature of the mid-layer, and we set µ = cos θ with θ as the angle measured from the z-axis. The visual optical depth of the mid-layer in vertical direction can be written in the form
−s r τVmid (r) = τVmid (r0 ) (11) r0

283

SEDs of Flared Dust Disks (Kraus)

with the visual optical depth at the inner edge, τVmid (r0 ), and the exponent s as free parameters. The temperature of the mid-layer, Tmid , follows from equaling Eqs. (10) and (9). This temperature is, however, a function of the grazing angle. We have now three important parameters, αgr , h, and Tmid , but none of them can be computed independently, instead they must be calculated iteratively.

1.4

Radiation transfer within the top-layer

We have to specify the visual optical depth in the top-layer in zdirection, τVtop . It should be high enough so that first, the infalling stellar radiation is completely absorbed and second, the dust grains at the bottom of the top–layer reach a temperature close to that of the mid-layer in order to guarantee a smooth transition. On the other hand, τVtop must be small enough to allow the dust emission, which occurs at infrared wavelengths, to escape the top-layer. The radiation transfer equation of a plane-parallel slab, Iν (µ, τν ) = I0 eτν /µ −



Sν (t)e−(t−τν )/µ

dt µ

(12)

with the source function Sν , the incident intensity I0 and µ = cos θ is split into up-streams, Iν+ , that penetrate the top-layer from the midlayer, and down-streams, Iν− , that cross the top-layer starting from the surface. The incident intensity is either the stellar radiation for downwards directed streams (but only for angles θ under which the star can be seen, else it is zero), or the emission of the mid-layer, mid Bν (Tmid )(1 − e−τν /µ ), for the upwards directed streams. With the help of the Feautrier parameters uν =

1 + (I + Iν− ) 2 ν

and

vν =

1 + (I − Iν− ) 2 ν

(13)

the mean intensity Jν becomes 

Jν = Jν (τν ) =

uν (µ, τν ) dµ

(14)

which is needed to calculate the source function and the emission of the grains. The source function itself determines the up- and down-streams of the intensity; the radiation field calculation must therefore be iterated. Finally, the temperature of the grains at each location in the top-layer follows from balancing their absorption from the surrounding radiation field, Jν , and their emission at their equilibrium temperature, Td .

284

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. Drop of the parameter h/H (top, see text), and the height h of the photosphere (bottom).

2.

Figure 3. Temperature of the midlayer (top), and comparison of the top- and mid-layer temperature at the boundary (bottom).

Results

Since we want to compare our results with those of CG we took the same parameters as they did: for the star, T∗ = 4000 K, L∗ = 1.44 L
, R∗ = 2.5 R
, M∗ = 0.5 M
; for the disk, rin  0.07 AU, rout = 270 AU, − 1 .5 ; for the dust grains, spheres of radius a = 0.1 µm τVmid (r)  4 · 105 rAU with mass density ρd = 2 g cm−3 and absorption efficiency Q = 1 for λ ≤ 2πa and Q  2πa/λ for λ ≥ 2πa. The latter implies an absorption coefficient κV  4 × 104 cm2 per gram of dust. Scattering is absent and we use τVtop = 3. The disk is seen face-on. We first compare our height h of the photosphere (Fig. 2) with the CG results. The ratio h/r which determines αgr can be written in terms of the scale height H h h H = (15) r H r

According to CG the ratio h/H drops from 5 at 3 AU to 4 at 100 AU but they use h/H = 4 throughout their calculations which leads to an extremely thick flared disk with h(r = 270 AU) = 270 AU. Our calculations show, however, that h/H drops from about 6 at r0 to ≤ 2.8 at the outer edge, and our disk stays much flatter at large distances. In

SEDs of Flared Dust Disks (Kraus)

285

Figure 4. Temperature distribution within the top-layer at ro (top) and at r = 250 AU (bottom). The star is the CG value, square and triangle are our and the CG mid-layer value.

Figure 5. Comparison between CG and our results for the top-layer (top), mid-layer (middle) and total SED (bottom).

a recent work Chiang et al. (2001) corrected their old results for the disk thickness and their corrected photosphere almost agrees with our results. The temperature in the mid-layer, Tmid (Fig. 3), is not so different in the two models. Our mid-layer is slightly hotter especially in the inner parts (r ≤ 20 AU) but comparable with CG for larger distances. Our temperature calculations in the top-layer result in a temperature gradient between the very hot surface and the bottom of the top-layer. In Fig. 4 we compare our temperature distribution with the CG surface temperature as well as with our and the CG mid-layer temperatures at two different distances. We cannot reproduce the very hot surface temperature of CG, but on the bottom of the top-layer (at τVtop = 3) the temperature approaches everywhere the value of the mid-layer (see bottom panel of Fig. 3). The differences in the disk and temperature structure between the CG model and our more detailed radiation transfer calculations lead of course also to differences in the SED (Fig. 5). The emission from the CG

286

OPEN ISSUES IN LOCAL STAR FORMATION

surface layer which is much hotter than our top-layer is shifted to shorter wavelengths. The emission of the CG mid-layer comes from a narrower wavelength region than in our model and is too low. The total SED of CG therefore deviates from our results especially in the long wavelength region (λ ≥ 30µm).

3.

Conclusions

A radiation ransfer model for the upper disk part of a flared disk is presented and the results are compared with the analytical two-layer model of CG. We show that the major parameters that determine the structure of the disk, αgr , h, Tmid , must be calculated iteratively. In addition we perform detailed radiation transfer calculations in the top-layer to find the temperature structure in the upper disk part. At the bottom of the top-layer the temperature agrees very good with the values of the mid-layer. The very hot surface temperature found by CG could not be reproduced, and our total SED deviates from the one of CG especially in the long wavelength regime. Although the two-layer model gives handy formulae to calculate the SED of a flared dust disk, the results should only be used as a ‘first guess’ for the structure and temperature distribution of the disk and radiation transfer calculations should be performed for a better characterization of the circumstellar dust disk.

Acknowledgments I would like to thank Dr. Endrik Kr¨ ugel for many helpful discussions. This work was supported by the German Deutsche Forschungsgemeinschaft, DFG grant number Kr 2163/2–1.

References Beckwith, S.V.W., Sargent, A.I., Chini, R.S. & G¨ usten, R. 1990, AJ 99, 924 Chiang, E.I., & Goldreich, P. 1997, ApJ 490, 368 Chiang, E.I., Joung, M.K., Creech-Eakman, M.J., Qi, C., Kessler, J.E., Blake, G.A. & van Dishoeck, E.F. 2001, ApJ 547, 1077 Dullemond, C.P. 2002, A&A 395, 853 Dullemond, C.P., & Natta, A. 2003, A&A in press Dullemond, C.P., Dominik, C., & Natta, A. 2001, ApJ 560, 957 Gregorio-Hetem, J. & Hetem, A., Jr. 2002, MNRAS 336, 197 Kenyon, S.J., & Hartmann, L. 1987, ApJ 322, 293 Miroshnichenko, A., Ivezic, Z., Vinkovic, D. & Elitzur, M. 1999, ApJ 520, L 115

ON THE ALIGNMENT OF T TAURI STARS WITH THE LOCAL MAGNETIC FIELD Gaspard Duchˆene Department of Physics & Astronomy, UCLA, USA [email protected]

Fran¸cois M´enard Laboratoire d’Astrophysique, Observatoire de Grenoble, France [email protected]

Abstract

Magnetic field is believed to play an important role in the collapse of a molecular cloud. In particular, due to the properties of magnetic forces, collapse should be easier along magnetic field lines, as supported by the large-scale sheet-like structure of the Taurus giant molecular cloud for instance. Here we investigate whether such a prefered orientation for collapse is present at a much smaller scale, that of individual objects. We use recent high-angular resolution images of T Tauri stars located in the Taurus star-forming region to find the orientation of the symmetry axis of each star+jet+disk system and compare it to that of the local magnetic field. We find that i) the orientations of the symmetry axis of T Tauri stars are not random with respect to the magnetic field, and ii) that young stars that are associated to a jet or an outflow are oriented very differently from those which do not have a detected outflow. We present some implications of this puzzling new result.

Introduction Stars form in molecular clouds as a consequence of the gravitational collapse of dense molecular cores. This process is however sensitive to other phenomena than gravity, such as rotation or magnetic field. For instance, neutral species are only affected by gravity while ions are tightly bound to the magnetic field. Friction between ions and neutrals, known as ambipolar diffusion, modifies the kinematics of the neutrals, leading to protostars that are surrounded by pseudo-disks, even in the absence of rotation (Galli & Shu, 1993). In the presence of straight magnetic 287 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 287-294. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

288

OPEN ISSUES IN LOCAL STAR FORMATION

field lines threading the cloud, Galli & Shu (1993) showed that the major axis of the pseudo-disks are perpendicular to the direction of the original magnetic field, as expected if the collapse occurs preferentially along the field lines. The orientation of young stellar objects (YSOs) can therefore inform us directly on the importance of magnetic field in the collapse process. In the event of a leading role for magnetic forces, we expect circumstellar disks to be oriented so that their symmetry/rotation axis is (roughly) parallel with the local magnetic field in the cloud. On the other hand, if magnetic field does not influence much the collapse, disks might well be oriented at random. In the Taurus-Auriga giant molecular cloud, both the dense gas clouds and the YSOs shows a sheet-like structure fully consistent with a large scale collapse along the magnetic field lines (Goodman et al. 1990; Hartmann 2002). At the individual object scale, Lee & Myers (1999) showed that pre-stellar cores were elongated, with an average aspect ratio of ∼ 2.4. If one attributes this elongation to rotation, it can be shown that their major axis are also preferentially found perpendicular to the magnetic field (Hartmann 2002). These findings support the case for a causal link between the direction of the large scale magnetic field and the orientation of young stars. With the advent of high-angular resolution imaging devices, it is now possible to trace the orientation of already-formed YZOs. First, collimated jets and outflows from Myr-old T Tauri stars are commonly observed and can be used to trace their orientation. Strom et al. (1986) noted, during a deep imaging survey of Taurus sources driving HerbigHaro objects, that jets and outflows have a tendency to align parallel with the local magnetic field. MHD models predict that the ionised atomic jets of young stars are expelled perpendicular to the accretion disk (e.g., Shu et al. 1995; Ferreira & Pelletier 1995), in agreement with striking high resolution observations of objects like HH 30 (Burrows et al. 1996). The preferred jet orientation therefore suggests that the rotation axis of T Tauri circumstellar disks are parallel to the local magnetic field. A similar conclusion was reached by Tamura & Sato (1989) who found linear polarisation vectors of YSOs to be preferentially parallel to the local magnetic field. Although the interpretation of this result is affected by a 90o ambiguity, it has reinforced the general belief that the orientation of young stars is in some way related to the direction of the local magnetic field. Since early surveys for optical jets from T Tauri stars, many new jets and outflows have been identified. Furthermore, it is now also possible to spatially resolve circumstellar disks and accurately estimate their orientation through high angular resolution imaging. It is therefore time

Alignment of TTs with the local magnetic field (Duchˆene & M´enard)

289

to readress this issue and study the link between the orientation of the local magnetic field and the orientation of YSOs in more detail. We will focus here on the well-studied Taurus star-forming regions.

1.

Direction of the magnetic field in Taurus

In the presence of a magnetic field, elongated interstellar grains tend to spin along a preferred direction, with their axis of smallest moment of inertia parallel to the magnetic field (e.g., Lazarian 2002, and references therein). Dichroism is induced as the grains act like a picket-fence to absorb the light of background stars more efficiently along the direction perpendicular to the grains’ rotation axis, i.e., generally perpendicular to the field. A net linear polarisation results in the other direction, i.e., parallel to the projected direction of the magnetic field. This technique has been used by numerous authors to infer the structure of the magnetic field in molecular clouds. For our analysis, we have compiled over 400 linear polarisation measurements of background stars in Taurus (left panel in Fig. 1) from Vrba et al. (1976), Heyer et al. (1987), Moneti et al. (1984), Tamura et al. (1987), Tamura & Sato (1989) and Goodman et al. (1990, 1992). To estimate the direction of the magnetic field at the location of each star in our sample (see below), we searched all the interstellar polarisation data in increasingly larger circles of radii 0.5–2.0o . We selected the smallest zone containing at least 4 different measurements of the magnetic field and retained the median of the measured position angles as the direction of the local magnetic field. As can be seen in Fig. 1, the polarisation vectors form a smooth structure throughout the region and we estimate that the orientation of the local magnetic field can be estimated to within 5o for most objects.

2.

Orientation of T Tauri stars in Taurus

We have first compiled a complete list of T Tauri stars in the TaurusAuriga star forming region from the Herbig & Bell (1988, hereafter HBC) catalogue. We restrict our study to the zone 4h00< α 3) are considered here. The length of all vectors is uniform and not proportional to the polarisation rate. Right: Orientation of T Tauri stars in Taurus as defined by an optical jet or outflow (solid vectors) or by a spatially resolved circumstellar or circumbinary disk (dashed vectors). In both cases, the symmetry axis of the system is shown.

For each objects, we searched the litterature for the presence of a spatially resolved jet and/or disk. Only morphological evidences were used in this process. Collimated jets are usually identified in deep narrowband images obtained at the wavelength of optical forbidden lines (e.g., Mundt et al. 1991; Dougados et al. 2000) or through long slit spectroscopy observations, which locate the redshifted and blueshifted parts of the jet on opposite sides of the star (e.g., Hirth et al. 1997). In most cases, jets are clearly resolved and their position angle is known within 10o or better. Disks around young stars can be identified in two main ways: thermal imaging in the submillimeter and millimeter ranges and scattered light imaging in the optical and near-infrared. Because current instruments have a limited dynamic range, the latter technique is usually more sensitive to edge-on disks (e.g., Burrows et al. 1996) and to circumbinary rings (e.g., Roddier et al. 1996). In the millimeter continuum, disks appear as sources of thermal emission that are spatially resolved when observed with long baseline interferometers (e.g., Dutrey et al. 1996; Kitamura et al. 2002). Furthermore, observations of some disks in CO lines at millimeter wavelength reveal velocity profiles that are consistent with Keplerian rotation (e.g., Simon et al. 2000). When available, we used these resolved CO maps to define the orientation of the disk’s semi-major axis. Disk orientations are known to within 5o or better when a Keplerian velocity gradient is detected and within 5–15o

Alignment of TTs with the local magnetic field (Duchˆene & M´enard)

291

otherwise. More details on the individual sources as well as data tables are presented elsewhere (M´enard & Duchˆene 2003).

3.

Relative orientations of T Tauri stars

Among our targets, we identified 34 objects with a resolved jet and/or disk, including 12 that possess both of them. Among those 12 sources, all but one dubious case (DO Tau, see M´enard & Duchˆene 2003) have perpendicular jets and disks to within 20o or better, as illustrated by the prototypical object HH 30 (Burrows et al. 1996). The symmetry axis of the system is determined by the orientation of the jet if it is present. Otherwise, we assume that the symetry axis lies in a direction perpendicular to the semi-major axis of the disk. As for the magnetic field direction, we only find the projection in the plane of the sky of the symmetry axis of the object. Despite possible projection effects, which we discuss below, it is worthwhile searching for possible correlations, which would trace an underlying physical link between the orientation of YSOs and the magnetid field. The orientation of the symmetry axis for the 34 T Tauri stars for which we could determine it is plotted in the right panel of Fig 1 with different symbols for objects with and without jets. We plot the same axis, i.e. the symmetry axis of the star+jet+disk system, for both categories of objects. Although the direction of the symmetry axes shows significant scatter, two general trends can be identified from Fig 1, especially in the central area of the figure (R.A. ∼ 4h30, Dec. ∼ 25o ). First, T Tauri stars for which a jet or outflow has been spatially resolved (solid vectors) are majoritarily oriented along the same direction as the local magnetic field. Second, T Tauri stars without a jet or an outflow but with a spatially resolved circumstellar disk are roughly oriented perpendicularly to the local magnetic field. While the former conclusion is similar to those reached in past studies, the latter point is revealed here for the first time. We now quantify it in more detail before exploring its implication for star formation studies. For each object with a known orientation in the plane of the sky, we measured the “misalignment angle” between the axis of symmetry of the T Tauri star and that of the local magnetic field. A value of 0o indicate that both are parallel while it is 90o if they are perpendicular. The distributions of these angles for both our subsamples (objects with and without a spatially resolved jet) are shown in Fig. 2 (left panel). The two histograms are statistically different at the ∼ 3 σ level: there is a significant difference in the respective orientations of T Tauri stars with and without jets. Systems with a jet have their symmetry axis

292

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. Left: Distribution of the angle between the orientation of the symmetry axis of T Tauri stars and their local magnetic field. From 0o to 90o , these two directions go from parallel to perpedicular. The dashed histogram represent T Tauri stars with a jet or an outflow while the solid histogram are those for which only a disk could be spatially resolved. Right: Cumulative distributions of angles between the symmetry axis of T Tauri stars and their local magnetic field for objects with a jet/outflow (dashed), with only a disk (solid) and for the whole sample of objects with known orientation (dot-dashed). The dotted line represents the expected distribution if YSOs were randomly oriented with respect to the local magnetic field.

roughly parallel to the local magnetic field (median angle = 25o ) while systems without a jet or an outflow are preferentially perpendicular to it (median angle = 75o ). Interestingly, the distribution of orientations for our complete sample of 34 objects is consistent with random orientation, as illustrated in the right panel of Fig. 2. As pointed out earlier, our sample is unlikely to be 100 % complete and an unknown selection effect may be responsible for this apparent difference. However, the imaging studies used to build our database of stellar orientations are equally sensitive to structures (jets, outflows, disks) in any direction in the plane of the sky. It is difficult to think of a selection effect that would preferentially sample stars oriented in a specific direction. Although there are strong selection effects regarding inclination, in particular against pole-on systems, we do not find any significant difference between the distributions of inclinations of objects with and without a jet. Therefore, we do not believe that the distinction between the two subsamples compared here can be distinguish by specific geometric configurations that lead to undetectable jets for some YSOs.

4.

Open questions

The fact that T Tauri stars do not orient at random with respect to the magnetic field and that the presence or absence of a bipolar jet is statistically linked to the orientation of the object clearly shows that the

Alignment of TTs with the local magnetic field (Duchˆene & M´enard)

293

magnetic field does play an important role in the collapse of individual dense cores and their subsequent evolution. The origin of the differential orientation trend discovered here is not clear yet and only suggestions can be put forward at this point. A first important point to note is that, while jets are most likely launched through a specific magnetic configuration linking the star, its accretion disk and the jet, this involves the stellar magnetic field. On the other hand, our studies of orientations concerns the molecular cloud magnetic field, which may have a different topology and orientation. First, it is unclear whether T Tauri stars retain the same orientation throughout their evolution. It could be that the collapse is strongly driven by the magnetic field, with all systems having their symmetry axis roughly parallel to the local magnetic field, and that some of them later become misaligned through an unknown process. It is worth noting that the orientation of bipolar outflows from more embedded (i.e., younger) Class I sources in Taurus show the same tendency to align parallel to the magnetic field as T Tauri stars that possess a jet (see M´enard & Duchˆene 2003). If all YSOs retain their original orientation, where are the precursors of those T Tauri stars that now only have a disk? Despite this suggestive trend for protostars, we fail to identify any mechanism that could provide enough torque to rotate by 90o a star+disk system. A second possibility is that YSOs are formed with random orientations, as suggested by the distribution of our complete sample, and that subsequent evolution lead to the formation of a powerful jet/outflow only in some conditions. For instance, Ferreira (1997) showed that a quadrupolar magnetic field configuration in the disk lead to a much weaker disk-wind, if any, than a dipolar configuration could achieve. We can then propose that the configuration of the stellar magnetic field will be dipolar if its axis of symmetry is more or less aligned with the local field in the molecular cloud, while only quadrupolar (or high order configurations) can be created if the relative orientation is close to perpendicular. This could be the consequence of a feedback of the cloud magnetic field on the (currently unknown) growth mechanism of the T Tauri star field. The fact that Class I sources do not apper to be randomly oriented is somewhat problematic in this picture though. A third possible explanation to the observed trend is a projection effect, in which T Tauri stars that have a jet are located in a different part of the cloud than those who do not have one. The latter could be seen in projection over the Taurus cloud, while being located well in front, where the magnetic field orientation could be different. In that case however, the field at the periphery of the cloud would need to be twisted by 90o with respect to the field found inside the cloud where

294

OPEN ISSUES IN LOCAL STAR FORMATION

T Tauri stars with jets would presumably be located. It is unclear why this would be the case. Furthermore, one would have to explain why the objects located in the outer parts of the cloud are systematically devoid of jets and outflows. Only tentative interpretation of the observational result presented here can be given for now; none of them is fully satisfying. Conducting similar studies in other star-forming regions, such as ρ Ophiuchus or Orion, and measuring submillimeter polarisation vectors from much younger prestellar cores will help understanding whether this trend is a general one and possibly identify other factors that play a role in this preferred alignment of objects with the magnetic field.

References Burrows, C. J., Stapelfeldt, K. R., Watson, A. M., Krist, et al. 1996, ApJ, 473, 437 Dutrey, A., Guilloteau, S., Duvert, G., Prato, L., Simon, M., Schuster, K. & M´enard, F. 1996, A&A, 309, 493 Ferreira, J. & Pelletier, G. 1995, A&A, 295, 807 Ferreira, J. 1997, A&A, 319, 340 Galli, D. & Shu, F. H. 1993, ApJ, 417, 243. Goodman, A. A., Bastien, P., M´enard, F. & Myers, P. C. 1990, ApJ, 359, 363 Goodman, A. A., Jones, T. J., Lada, E. A. & Myers, P. C. 1992, ApJ, 399, 108 Hartmann, L. 2002, ApJ, 578, 914 Herbig, G. H. & Bell, K. R. 1988, Lick Obs. Bulletin, No. 1111. Heyer, M. H., Vrba, F. J., Snell, R. L., Schloerb, F. P., Strom, S. E., Goldsmith, P. F. & Strom, K. M. 1987, ApJ, 321, 855 Hirth, G. A., Mundt, R. & Solf, J. 1997, A&AS, 126, 437 Kitamura, Y., Momose, M., Yokogawa, S., Kawabe, R., Tamura, M. & Ida, S. 2002, ApJ, 581, 357 Lavalley-Fouquet, C., Cabrit, S. & Dougados, C. 2000, A&A, 356, L41 Lazarian, A. 2002, Elsevier preprint. (astro-ph/0208487). Lee, C. W. & Myers, P. C. 1999, ApJS, 123, 233 M´enard, F., Dougados, C., Magnier, E., Duchˆene, G., Cuillandre, J.-C., Mart´ın, E., Fahlman, G., Forveille, T., Lai, O., Manset, N., Martin, P., Weillet, C. & Magazz` u, A. 2003, ApJ, submitted M´enard, F. & Duchˆene, G. 2003, A&A, submitted. Moneti, A., Pipher, J. L., Helfer, H. L., McMillan, R. S. & Perry, M. L. 1984, ApJ, 282, 508 Mundt, R., Ray, T. P. & Raga, A. C. 1991, A&A, 252, 740 Roddier, C., Roddier, F., Northcott, M. J., Graves, J. E. & Jim, K. 1996, ApJ, 463, 326 Shu, F. H., Najita, J., Ostriker, E. C. & Shang, H. 1995, ApJ, 455, L155 Simon, M., Dutrey, A. & Guilloteau, S. 2000, ApJ, 545, 1034 Strom, K.M., Strom, S. E., Wolff, S. C., Morgan, J. & Wenz, M. 1986, ApJS, 62, 39 Tamura, M., Nagata, T., Sato, S. & Tanaka, M. 1987, MNRAS, 224, 413 Tamura, M. & Sato, S. 1989, AJ, 98, 1368 Vrba, F. J., Strom, S. E. & Strom, K. M. 1976, AJ, 81, 958

TEMPORAL EVOLUTION OF MAIN SEQUENCE DUSTY DISKS Manoj Puravankara Indian Institute of Astrophysics, India [email protected]

H. C. Bhatt Indian Institute of Astrophysics, India [email protected]

Abstract Using the velocity dispersion as a statistical measure of age we constrain the ages of a large sample of Vega-like stars in order to study the temporal evolution of the dust disks around them. Fractional dust luminosity (fd ≡ Ldust /L ) of stars is taken as a measure of ’dustiness’ of the disks. The dustiness (fd ) of the main sequence disks is found to decrease with increasing velocity dispersion. Since velocity dispersion of stars is known to increase with age as ∼ t1/3 , a functional dependence of dustiness (fd ) with age can be derived. fd is found to drop off with stellar age according to the power law fd ∝ (age)−2.38 . Ages of these main sequence dusty systems range from a few times 107 yr to a Gyr. The age distribution of Vega-like stars is found to be consistent with a constant rate formation of 10−14 yr −1 pc−3 in the solar neighborhood.

Introduction Main sequence dusty systems were discovered by IRAS in 1983 (Aumann et al. 1984). Since then several surveys have identified hundreds of such main sequence stars with disks based on the excesses that they exhibited at far infrared wavelenghts. These stars became known as ‘Vega-like’ stars, named after the prototype Vega. The dust disks around Vega-like stars are optically thin and relatively gas-free so that the disk evolution is dominated by the grain dynamics. Because of the short dust survival time scales relative to the stellar ages, the disks are thought to be made of ‘second generation dust’ released from larger parent bodies 295 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 295-302. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

296

OPEN ISSUES IN LOCAL STAR FORMATION

such as asteroids or comets. At least 15% of the main sequence stars are surrounded by such disks (Lagrange et al. 2000). It is believed that planet formation is well under way in these disks and that the dust that we detect is the debris of this formation process. Therefore, these main sequence disks are regarded as the signposts of recent planet formation (Kenyon & Bromely 2002). A study of the temporal evolution of the main sequence dusty systems, then, is of critical importance to the understanding of the formation of planetary systems and disk dispersal timescales. There have only been a few studies on the time evolution of the dusty disks around main sequence stars. From the sub-mm studies of prototype Vega-like stars, Zuckerman & Becklin (1993) found that the amount of dust around main sequence stars declines as rapidly as (time)−2 during the initial 3×108 yr. Similar results have been reported by Holland et al. (1998) from their recent SCUBA observations. Recently, et al. (2001) from a survey of circumstellar disks around pre-main sequence and young main sequence stars in young open clusters of known ages, have found that the fractional dust luminosity fd drops off with stellar age according to the power law fd ∝ (age)−1.76 , a result which is consistent with the theoretically expected relation of fd ∝ (age)−2 for collisionally replenished secondary dust disks. Here we study the kinematics of a large sample of Vega-like stars and use the velocity dispersion as an age indicator to study the evolution of their circumstellar dust disk. It has long been known that there is a strong correlation between the random velocities and ages of stars in the Galactic disk. Velocity dispersion of a group of stars in the solar neighborhood has been found to increase with its age (Wielen 1977; Jahreiss & Wielen 1983). Observationally, velocity dispersion, σ, is found to grow with age at least as fast as t0.3 and more like t0.5 (Wielen 1977; Binney & Tremaine 1987). Dynamical origin of this effect is attributed to the encounters between the disk stars and the massive gas-clouds (Spitzer & Schwarzchild 1951; Spitzer & Schwarzchild 1953) and to transient spiral waves heating up the Galactic disk thereby increasing the velocity dispersion of stars (Barbanis & Woltjer 1967). A combination of these two mechanisms has been found to reproduce the observations reasonably well (Jenkins 1992). Recently, using accurate Hipparcos parallaxes and proper motions Binney et al. 2000 and Denhen & Binney 1998 have shown that for a coeval group of stars, the rms dispersion in transverse velocity, S, which is connected to the principal velocity dispersion by the 2 + σ 2 + σ 2 ], increases with time from 8 kms−1 at relation S 2 = 2/3[σR z φ 1/3 birth as t . We follow this formalism and use the dispersion in transverse velocity to constrain the ages of Vega-like stars and to study the temporal evolution of their dust disks.

Temporal evolution of dusty disks (Manoj & Bhatt)

1.

297

Transverse Velocity and Dustiness of Vega-like stars

We compiled a total of 486 Vega-like stars from several sources in literature (Song 2000; Silvestone 2000; Coulson et al. 1998; Malfait et al. 1998) out of which 342 stars have both Hipparcos and IRAS (PSC/FSC) data. For all the stars in our sample we have proper motions and parallaxes from Hipparcos catalog (ESA 1997). The transverse velocity perpendicular to the line of sight relative to the solar system barycenter is then computed using the relation VT =

Av µ π

where Av = 4.740470 kmyrs−1 , π is the parallax in milli arcseconds and  µ = µ2δ + (µα cos δ )2

with µδ and µα cos δ being the Hipparcos proper motions along declination and right ascension expressed in milli arcseconds. Errors in transverse velocities are estimated from the probable errors in parallaxes and proper motions given in Hipparcos. Transverse velocities of stars thus obtained will have solar motion reflected in them. This will be felt differently by stars in different directions. We, therefore, correct the velocities for solar motion using the values of U = 10.0 ± 0.4kms−1 , V = 5.2 ± 0.6kms−1 , W = 7.2 ± 0.4kms−1 (Binney & Merrifield 1998) for the standard solar motion. To minimize the effect of Galactic differential rotation we consider only stars within 250 pc from the Sun. Further, we include only those stars in our analysis that have fractional error in transverse velocity less than 0.5. We have, then, 270 Vega-like stars with transverse velocities and fractional dust luminosities computed for the final analysis. A good measure of the ‘dustiness’ of the disks around Vega-like stars is the fractional dust luminosity, fd ≡ Ldust /L which represents the optical depth offered by an orbiting dust disk to ultraviolet and visual radiation. We compute fd from IRAS (PSC/FSC) fluxes for the Vegalike stars in our sample using the relation fd = Ldust /L =

10−4 × [6.45e12 + 2.35e25 + 1.43e60 + 0.55e100 ] 10[0.4(4.75−mV −BC)]

(Emerson 1988). In the above equation e12 , e25 , e60 , e100 are the excess flux densities over the photospheric values at IRAS wave bands, mV is the extinction corrected visual magnitude of the star and BC the

298

OPEN ISSUES IN LOCAL STAR FORMATION

bolometric correction. All our sample stars have good quality flux measurements at 12 & 25 µ. Wherever the IRAS (PSC/FSC) flux densities at 60 and 100 µm are upper limits, we used the expected e60 and e100 for the color temperature derived from the ratio e12 /e25 to compute fd . 0

fractional dust luminosity Log (fd)

-1

-2

-3 -4 -5 -6 -7 0

5

10 15 20 25 velocity dispersion S (km/sec)

30

35

Figure 1. Fractional dust luminosity plotted against transverse velocity dispersion for Vega-like stars. The point with downward arrow represents stars with excesses below the significance level and is only an upper limit

2.

f d - velocity dipersion relation

As mentioned earlier the velocity dispersion of a group of stars is an indicator of the average age of the group. Here we look for a correlation between fractional dust luminosity, fd and transverse velocity dispersion, S for stars of similar fd . For this we grouped the stars into bins of a given range in dustiness (fd ). We then computed the velocity dispersion of transverse velocities of stars in each of these bins. The mean value of fd in each bin is then plotted against the velocity dispersion of stars in that bin as shown in Figure 1. Error bars plotted in fd represent the standard deviation from the mean in each bin. The horizontal error bars plotted are the mean errors in the transverse velocities of stars in each bin estimated from the probable errors in the measurements of proper motions and parallaxes as given in Hipparcos. The point with a downward arrow represents stars with barely detectable excesses at IRAS sensitivity. The infrared excesses over the photospheric

299

Temporal evolution of dusty disks (Manoj & Bhatt)

emission determined for stars in this bin are below the significance level and the point can only be taken as an upper limit. It is clear from Figure 1 that there is a systematic decrease in the dustiness (fd ) of the disks with increasing velocity dispersion of stars. As discussed earlier, the velocity dispersion, S of stars, in general, is found to increase with stellar age as S ∝ t1/3 (Binney et al. 2000). Thus the correlation between fractional dust luminosity of Vega-like stars and their velocity dispersion seen in Figure 1 clearly implies a steady decrease in the optical thickness of the dusty disks with stellar age. This is consistent with the earlier findings that the amount of dust in the disks appears to decrease generally with system age. Moreover, the velocity dispersion of Vega-like stars range from ∼ 10kms−1 for fd ∼ 10−2 to ∼ 30kms−1 for fd ∼ 10−6 which translates into a total spread in age of a factor of 30 for Vega-like stars.

3.

Temporal evolution of dusty disks

0

fractional dust luminosity Log (fd)

-1

-2 -3 -4 -5 -6 -7 6.5

7.0

7.5

8.0 Log [Age (yr)]

8.5

9.0

9.5

Figure 2. Fractional dust luminosity of Vega-like stars plotted against their ages obtained from velocity dispersion. The plotted line is a regression fit to the data with slope = −2.38.

In order to quantify the temporal evolution of the dustiness of disks around Vega-like stars, the velocity dispersion has to be related to the stellar age. We follow the formalism of Binney et al. (2000) where the

300

OPEN ISSUES IN LOCAL STAR FORMATION

velocity dispersion as a function of age is given by S = v10 [(τ + τ1 )/(10Gyr + τ1 )]β . In the above equation τ1 determines the random velocity of stars at birth, and v10 and β characterize the efficiency of stellar acceleration. Using values of β = 1/3, v10 = 58kms−1 and τ1 = 0.03 Gyr (Binney et al. 2000) we compute the average stellar ages for each bin. In Figure 2 we plot fd against the stellar ages thus obtained. The vertical error bars in Figure 2 are the same as that in Figure 1. Horizontal error bars represent errors in velocity dispersion translated into errors in age. The steady drop of fractional dust luminosity with increasing stellar age is evident from the figure. A linear regression fit to the data yields a reasonably good fit with a power law of the form fd ∝ (age)−2.38 . This power law index is consistent with the theoretically expected value of −2 for collisionally replenished debris disks and with the evolutionary trend found by Spangler et al. (2001) and Zuckerman & Becklin (1993).

0

-2

fractional dust luminosity Log (fd) -6 -4

-8

frequency Log N

100

10

1

7.5

8.0

9.0 8.5 Log [Age (yr)]

9.5

10.0

Figure 3. Histogram of the age distribution of Vega-like stars. Solid histogram contains stars with fd that are upper limits. The dashed line represents the expected age distribution for a constant rate of formation.

From the above relation between fractional dust luminosity, fd and ages, we also obtain the age distribution of Vega-like stars. In Figure 3 we plot a histogram of age distribution of Vega-like stars in our sample. This is essentially a histogram of fractional dust luminosities of Vega-like

Temporal evolution of dusty disks (Manoj & Bhatt)

301

stars with fd translated into ages using the fd − age relation that we derived above. The solid histogram represents stars that have excesses below the significance level. It can be seen from Figure 3 that the ages of Vega-like stars range from a few times 107 yr up to a Gyr with the age distribution peaking at ∼ 300 to 400 M yr. In Figure 3 dashed line represents the age distribution for a constant rate of formation. Thus it can be seen that the age distribution of Vega-like stars is consistent with a constant rate of formation. The rate of incidence of Vega phenomenon is found to be 10−14 yr−1 pc−3 in the solar neighborhood.

References Aumann, H. H. et al. 1984, ApJ, 278, 23 Barbanis, B. & Woltjer, L., 1967, ApJ, 150, 461 Binney, J., Dehnen, W. & Bertelli, G., 2000, MNRAS, 318, 658 Binney, J. & Merrifield, M. 1998, Princeton University press Binney, J. & Tremaine, S., 1987, Princeton University press Coulson I. M., Walther D. M., & Dent W.R.F.,1998, MNRAS, 296, 934 Denhen, W., Binney J. J., 1998, MNRAS, 298, 387 Emerson, J. P., 1988, in Formation and evolution of low mass stars, eds. Dupree, A. K., & Lago M. T. V. T., p193 ESA, 1997, The Hipparcos and Tycho Catalogues, ESA SP-1200 Holland et al., 1998, Nature, 392, 788 Jahreiss, H. & Wielen, R., 1983, The nearby stars and the stellar luminosity function, IAU Coll. 76, eds. Davis Philip, A. G., Upgren, A. R. (Schenectady: L. Davis Press), p. 277 Jenkins, A., 1992, MNRAS, 257, 620 Kenyon, Scott J. & Bromley, Benjamin C., 2002, ApJL, 577, 35 Lagrange A-M, Backman D. E. & Artymowicz, P., 2000, Protostars and Planets IV, eds. Mannings, V., Boss, A. P., Russell S. S., University of Arizona Press, Tucson,pp 639-672 Malfait,K., Bogaert,E., and Waelkens,C., 1998, A & A, 331, 211 Silverstone, M., 2000, Ph.D. thesis, UCLA Song, I., 2000, Ph.D. thesis, Univ. Georgia Spangler, C., Sargent, A. I., Silverstone, M. D., Becklin, E. E. & Zuckerman, B., 2001, ApJ, 555, 932 Spitzer, L. & Schwarzchild, M., 1951, ApJ, 114, 385 Spitzer, L. & Schwarzchild, M., 1953, ApJ, 118, 106 Wielen, R., 1977, A & A, 60, 263 Zuckerman, B. & Becklin, E. E., 1993, ApJ, 414, 793

302

OPEN ISSUES IN LOCAL STAR FORMATION

Z. Balog, I. Oliviera, G. Oliviera, F. Alves, L. Vieira, G. Lima, P. Manoj

NUMERICAL SIMULATIONS OF YOUNG STELLAR JETS: FROM ONE TENTH TO PARSEC SCALES E. M. de Gouveia Dal Pino Instituto Astronˆ omico e Geof´ısico, USP, Brazil [email protected]

A. C. Raga Instituto de Ciˆencias Nucleares, UNAM, M´exico

A. H. Cerqueira Instituto de F´ısica, DCET Universidade Estadual de Santa Cruz, Brazil

E. Masciadri Max-Planck Institut f¨ ur Astronomy, Germany

Abstract

We here review the structure and evolution of young stellar HerbigHaro (HH) jets with the help of fully 3-D hydrodynamical (HD) and magneto-hydrodynamical (MHD) simulations of radiative cooling jets, giving particular emphasis to the study of giant HH flows, and the effects of magnetic fields on the general jet structure. Also, we will discuss some relevant physical processes that arise from the interaction of these jets with the interestellar environment, like the entrainment of ambient molecular gas in the jet flow, and some implications of the numerical results on observed jet phenomena and the associated central sources.

Introduction In order to get rid of the angular momentum excess, low-mass young stellar objects produce collimated optical outflows (the Herbig-Haro, hereafter HH, jets) that have embedded in them bright emission-line 303 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 303-310. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

304

OPEN ISSUES IN LOCAL STAR FORMATION

knots which are radiating shock fronts propagating with v ∼ few 100 km/s into the ambient medium (see, e.g., Reipurth 2003, these proceedings). Researchers at first considered the most luminous Herbig-Haro objects in the flows to be terminal bow shocks, i.e., to mark the location where the proto-stellar flow impacts its cloud environment. However, these regions are typically displaced only about 0.1 pc from the star. With the recent availability of wide-field CCD arrays, it has become clear that some jets indeed extend to far greater distances (e.g., Devine et al. 1997). For example, HH 111 shows a total extent of ≈ 7.7 pc, HH 34 of ≈ 3 pc and HH 355 a total extent of ≈ 1.55 pc. Giant jets represent a fundamental trace of the historical evolution of the outflow activity over timescales of ∼ 104 yr, i.e., timescales comparable to the accretion time of the outflow sources in their main protostellar phase. The study of such huge jets provides the possibility of retrieving important elements related to the life of the outflow sources. Another important issue in the investigation of HH jets, is the requirement of the presence of magnetic fields to explain their generation and collimation through magneto-centrifugal forces associated either with the accretion disk that surrounds the star (e.g., K¨ onigl & Pudritz 2000), or with the disk-star boundary (in the X-winds; e.g. Shu et al. 1994). Polarization measurements (Ray et al. 1997) have evidenced magnetic field strengths B ∼ 1 G in the outflow of T Tau S at a distance of few tens of AU from the source, which could imply a plasma β = pj,gas /(B 2 /8π)  10−3 for a toroidal field configuration, and ∼ 103 , for a longitudinal field, at distances ∼ 0.1 pc. We here review the structure and evolution of the HH jets by means of fully 3-D hydrodynamical (HD) and magneto-hydrodynamical (MHD) simulations of radiative cooling jets, giving particular emphasis to the aspects above, i.e., the study of giant HH flows, and the effects of magnetic fields in the HH jets structure. Also, we will discuss some relevant physical processes that arise from the interaction of these jets with the interestellar environment, like the entrainment of ambient molecular gas in the jet flow through the knots.

1.

Giant Herbig-Haro Flows

An important characteristic of the giant jets is that they appear to slow down for increasing distances from the outflow source. This effect is seen in the HH 34 (Devine et al. 1997) and in the HH 111 giant jets (Reipurth et al. 1997). In a previous study, de Gouveia Dal Pino (2001) modeled giant jets by performing three-dimensional simulations (with the SPH Lagrangian

HH Jet Simulations (Dalpino et al.)

305

code originally developed by de Gouveia Dal Pino & Benz 1993, 1994) of overdense, radiatively cooling jets modulated with long-period (P ∼ several hundred years) and large amplitude sinusoidal velocity variability at injection (∆v ∼ mean jet flow velocity). Allowing them to travel over a distance well beyond the source, she found that multiple travelling pulses develop and their velocity indeed falls off smoothly and systematically with distance. The deceleration was found to be mainly due to progressive momentum transfer sideways into the surrounding medium by the expelled gas from the travelling pulses. In a more recent study, Masciadri et al. (2002), have carried out 3-D simulations (employing the adaptive grid YGUAZU code developed by Raga et al. 2000) of the full extent of the giant HH34 flow, including not only jet velocity variability, but also a long period precession (of the order of 104 yr), which is suggested by the observed morphology of the source (Devine et al. 1997). They find that the impact of the precession on the deceleration mechanism is considerable. The Hα intensity maps and position-velocity diagrams obtained from this model reproduce the observations of the HH 34 giant jet in a successful way (see Figures 1 and 2). They conclude that the combined effects of jet velocity variability and precession (and the resulting enhanced interaction with the surrounding environment) produce the observed deceleration. What can we learn from the results above about the central source itself? Two points deserve to be addressed: (i) the existence of precession is generally ascribed to tidal forces produced by a companion in a binary or multiple system. Even though in the case of HH 34 the binary source has not yet been resolved, there is some evidence that this source could be a binary, such as the discovery of a second outflow (HH 534) emanating from the source, as well as the abrupt change of direction of the jet axis near knot B (Reipurth et al. 2002). This is in agreement with Reipurth (2003, this conference), who argues that the sources of giant HH jets are binary or multiple systems. However, the results above seem to be less consistent with his proposed thesis that giant jets are fossil records of the evolution of orbital motions in different phases of a disintegrating multiple system, as they are able to reproduce the HH 34 giant flow structure with single precession and variability periods (without needing different ejection properties at different times, reflecting qualitative changes in the outflow source resulting from the different phases of a disintegrating multiple system). Therefore, we conclude that if the HH 34 system does correspond to an outflow history with distinct phases, the evidence for this appears to have been lost probably due to the past complex interaction between the jet and the environment (Raga et al. 2002a, Masciadri et al. 2002);

306

OPEN ISSUES IN LOCAL STAR FORMATION

(ii) The simulations are consistent with the fact that velocity variations applied to the underlying flow at the jet inlet, steepen into internal shocks that produce the observed bright knots. These variations could arise from thermal instabilities in the accretion disk that surrounds the central star, like those believed to be responsible for FU Ori objects (Bell & Lin 1994). Another possibility is that they arise from outbursts produced during violent magnetic reconnection events in the inner edges of the disk (de Gouveia Dal Pino 2003). It is also possible that both processes occur but at different phases and rates of the accretion, and this subject still deserves further study.

Figure 1. Hα map obtained from a 3-D gas dynamics simulation of the giant HH 34 flow with sinusoidal ejection velocity variability (with a mean velocity of 300 km s −1 , a half-amplitude of 100 km s−1 and a period of 1010 yr) and precession of the outflow axis (with a half-angle of 6◦ and a 12000 yr period). The frame corresponds to the time at which the jet extends out to ≈ 3 × 1018 cm from the source. The abscissas and ordinates are labeled in units of 1018 cm. The map is shown with the logarithmic grey-scale given (in erg cm−2 s−1 sr−1 ) by the bar on the left. The similarity with the observed map is remarkable (from Masciadri, de Gouveia Dal Pino, Raga & NoriegaCrespo, 2002).

2.

Molecular Ambient Gas Entrainment

There is a general consensus that the molecular emission in HH jets is mainly due to molecular ambient gas entrained into the atomic flow through the bowshocks (e.g., Masson & Chernin 1993; Raga & Cabrit 1993; Raga et al. 1995; Chernin, Masson, de Gouveia Dal Pino & Benz 1994; Downes & Ray 1999; Lim et al. 2001; see also Raga et al. 2003a for an updated review on this matter). The IR H2 emission of some objects (e.g., HH46, HH47) seems to favor this interpretation, at least qualitatively. As an exception, HH 110 shows evidence of molecular emission along the jet but with an offset to one side of the atomic jet emission. This has been interpreted (Reipurth et al. 1996) and modeled (Raga & Canto 1995; de Gouveia Dal Pino 1999; Hurka et al. 1999) as a result of the collision of an HH atomic jet with a dense cloud. Recent 3-D gas dynamical simulations by Raga et al. (2002b) have shown evidence that

HH Jet Simulations (Dalpino et al.)

307

Figure 2. Radial velocity vs. distance from the HH 34 giant jet source: curve measured by Devine et al. 1997 (bold line), and curves obtained from the numerical models A (dotted line), B (thin line) and Model C (dot-dash-dot line) from Masciadri et al. (2002). The best fit is given by model B which is the same as in Figure 1.

the H2 emission could in fact come from the dense cloud molecular material that has been ablated by the ionic jet (see Figs. 6 and 7 of Raga et al. 2002b). Though more rare, there are objects (e.g., HH111, HH1) that exhibit coincident H2 /atomic knot structure (Reipurth et al. 1999, 2000) that is difficult to explain in terms of entrained molecular gas from the ambient medium. One possible way for generating such coincident atomic and molecular emission structures is to assume that the H2 emission could come from interactions of the atomic jet with ambient molecular clumps (Raga et al. 2003a). In this case, ablated molecular material could fill in the jet beam thus producing a similar molecular and atomic morphology. A potential problem with this interpretation is that in order for the clumps to survive over the lifetime of the jet, they should have masses as large as Mc ∼ 10−4 to 10−1 times the Jupiter mass (Raga et al. 2003a). There is, however, no direct evidence of the existence of such clumps in the surroundings of young stars. Another (perhaps more plausible) possibility, is that the molecular/atomic coincident knot emission arises from a combination of two effects, i.e., the atomic jet could be launched with a small fraction of molecular content and later entrain more molecular material from the surrounding medium through its internal bowshocks, as it seems to be the case in the HH 1 jet (Reipurth et al. 2000). In the past, only hydrodynamical numerical modeling involving either pure atomic or pure molecular jets were performed (see, e.g., V¨olker et al. 1999 for the later case), but presently multidimensional numerical simulations of ionic jets with small contents of molecular material have

308

OPEN ISSUES IN LOCAL STAR FORMATION

become possible and are already in progress (Raga et al. 2003b) to check the possibility above.

3.

Effects of Magnetic Fields

In a search for possible signatures of magnetic fields on the large scales of the HH outflows, several MHD investigations of overdense, radiative cooling jets, have been carried out with the help of multidimensional numerical simulations both in two-dimensions (2-D) (Todo et al. 1993, Frank et al. 1998, 1999, 2000, Lerry & Frank 2000, Gardiner et al. 2000, Stone & Hardee 2000, Gardiner & Frank 2000, O’Sullivan & Ray 2000,), and in three-dimensions (3-D) (de Gouveia Dal Pino & Cerqueira 1996, Cerqueira, de Gouveia Dal Pino & Herant 1997, Cerqueira & de Gouveia Dal Pino 1999, 2001a,b, 2003a, 2003b: these proceedings; see also de Gouveia Dal Pino & Cerqueira 2002 for a review). The main results of these studies can be summarized as follows. The effects of magnetic fields are dependent on both the field-geometry and intensity (which, unfortunately, are still poorly determined from observations). The presence of a helical or a toroidal field tends to affect more the characteristics of the fluid, compared to the purely HD calculation, than a longitudinal field. However, the relative differences which are detected in 2-D simulations involving distinct magnetic field geometries (e.g., Stone & Hardee 2000, O’Sullivan & Ray 2000), seem to decrease in the 3-D calculations (Cerqueira & de Gouveia Dal Pino 2001a, b). In particular, Cerqueira & de Gouveia Dal Pino (2001a, b) have found that features, like the nose cones, that often develop at the jet head in 2-D calculations involving toroidal magnetic fields, are absent in the 3D models a result which is consistent with observations which show no direct evidence for nose cones at the head of protostellar jets. 3-D calculations have revealed that magnetic fields which are initially nearly in equipartition with the gas tend to affect only the detailed structure behind the shocks at the head and internal knots, mainly for the helical and toroidal topologies. In such cases, the Hα emissivity behind the internal knots can increase by a factorof up to four relative to that in the purely hydrodynamical jet (this factor increases to five for β  0.1 jets; Cerqueira & de Gouveia Dal Pino 2001b). The importance of magnetic effects in the HH bow shocks has been recently reinforced by high resolution studies of individual HH objetcs. For instance, detailed shock diagnosis of the HH 7 object (Smith et al. 2003) has revealed that the observed H2 emission is due to a C-shock (where heating is provided by ambiplor diffusion) with a magnetic field strength B ∼ 10−4 G. Further 3-D MHD studies are still required as the

HH Jet Simulations (Dalpino et al.)

309

detailed structure and emission properties of the jets seem to be sensitive to multidimensional effects when magnetic forces are present. Also obvious is the need for further observations and polarization mapping of star formation regions, for a real understanding of their magnetic field structure.

Acknowledgments E.M.G.D.P. wishes to thank the Brazilian Agencies FAPESP (grant 1997/13084-3) and CNPq for partial support. AR acknowledges support from the CONACyT grants 36572-E and 41320, and the DGAPA (UNAM) grant 112602.

References Bell, K.R., Lin, D.N.C. 1994, ApJ, 427, 987 Cerqueira, A.H., de Gouveia Dal Pino, E.M., & Herant, M. 1997 ApJ, 489, L185 Cerqueira, A.H., & de Gouveia Dal Pino, E.M. 1999, ApJ, 510, 828 Cerqueira, A.H., & de Gouveia Dal Pino, E.M. 2001a, ApJ, 550, L91 Cerqueira, A.H., & de Gouveia Dal Pino, E.M. 2001b, ApJ, 560, 779 Cerqueira, A.H., & de Gouveia Dal Pino, E.M. 2003a, Ap.Spa.Sci. (in press) Chernin, L.M., Masson, C.R., & de Gouveia Dal Pino, E.M., & Benz, W. 1994, ApJ, 426, 204 de Gouveia Dal Pino, E.M. 1999, ApJ, 526, 862 de Gouveia Dal Pino E.M., 2001, ApJ, 551, 347 de Gouveia Dal Pino E.M, 2003 (in prep.) de Gouveia Dal Pino, E.M., & Benz, W. 1993, ApJ, 410, 686 de Gouveia Dal Pino, E.M., & Benz, W. 1994, ApJ, 435, 261 de Gouveia Dal Pino, E.M., & Cerqueira, A.H. 1996, Astron. Lett., 34, 303 de Gouveia Dal Pino, E.M., & Cerqueira, A.H. 2002, Rev. Mex. A. A. (Conf. Series), 13, 29 Devine D., Bally J., Reipurth B., Heatcote S., 1997, AJ, 114, 2095 Downes, T.P., & Ray, T.P. 1999, A&A, 345, 977 Frank, A., Gardiner, T., Delemarter, G., Lery, T., & Betti, R. 1999, ApJ, 524, 947 Frank, A., Lery, T., Gardiner, T., Jones, T.W., & Ryu, D. 2000, ApJ, 530, 834 Frank, A., Ryu, D., Jones, T.W., & Noriega-Crespo, A. 1998, ApJ, 494, L79 Gardiner, T.A., & Frank, A. 2000, ApJ, 545, L153 Gardiner, T.A., Frank, A., Jones, T.W., & Ryu, D. 2000, ApJ, 530, 834 Hurka, J.D., Schmid-Burgk, J., & Hardee, P.E. 1999, A&A, 343, 558 K¨ onigl, A., & Pudritz, R.E. 2000 in Protostars and Planets IV, Eds. V. Mannings, A. Boss & S. Russel, Tucson: The University of Arizona Press, 2000, 759 Lery, T., & Frank, A. 2000, ApJ, 533, 897 Lim, A.J., Rawlingns, J.M.C., & Williams, D.A. 2001, MNRAS, 326, 1110 Masciadri, E., de Gouveia Dal Pino, E.M., Raga, A.C., & Noriega- Crespo, A. 2002, ApJ, 580 950 Masson, C.R., & Chernin, L.M. 1993, ApJ, 414, 230 O’Sullivan, S., & Ray, T.P. 2000, A&A, 363, 355 Raga, A.C., & Cant´ o, J., 1995, RMxAA, 31, 51

310

OPEN ISSUES IN LOCAL STAR FORMATION

Raga, A.C., & Cabrit, S. 1993, A&A, 278, 267 Raga A. C., de Gouveia dal Pino E. M., Noriega-Crespo A., Mininni P., & Ve´azquez P. F. 2002b, A&A, 392, 267 Raga, A.C., Navarro-Gonz´ alez, R., & Villagr´ an-Muniz, M. 2000. R.MxAA, 36, 67 Raga, A.C., Taylor, S.D., Cabrit, S., & Biro, S. 1995, A&A, 296, 833 Raga A.C., Velazquez, P. F.; Cant´ o, J.; Masciadri, E. 2002a, A&A, 395, 647 Raga, A.C., Velazquez, P.F., de Gouveia Dal Pino, E.M., Noriega- Crespo, A., & Mininni, P. 2003a, RmxAA (Conf. Ser.), 15, 115 Raga, A.C. et al. 2003b (in prep.) Ray, T.P., Muxlow, T.W.B., Axon, D.J., Brown, A., Corcoran, D., Dyson, J., & Mundt, R. 1997, Nature, 385, 415 Reipurth, B., Bally, J., & Rodriguez, L.F. 2000, ApJ, 534, 317 Reipurth, B., Heathcote, S., Yu, K.C., Bally, J., & Devine, D. 1997, AJ, 114, 2708 Reipurth B., Heathcote S., Morse J., Hartigan P., Bally J., 2002, AJ, 123, 362 Reipurth, B., & Raga, A.C., & Heathcote, S. 1996, A&A, 311, 989 Reipurth, B., & Raga, A.C. 1999, in The Origin of Stars ans Planetary Systems, Eds. C.J. Lada & N.D. Kylafis, Kluwer Academic Publishers, 1999, 267 Reipurth, B., Yu, K.C, Rodriguez, L.F., Heathcote, S., & Bally, J. 1999, A&A, 352, L83 Shu, F.H., Najita, J., Ostriker, E., Wilkin, F., Ruden, S.P., & Lizano, S. 1994, ApJ, 429, 781 Smith, M.D., Khanzadyan, T., & Davis, C.J. 2003, MNRAS, 339, 524 Stone, J.M., & Hardee, P.E. 2000, ApJ, 540, 192 Todo, Y., Uchida, Y., Sato, T., & Rosner, R. 1993, ApJ, 403, 164 V¨ olker, R., Smith, M.D., Suttner, G., Yorke, H.W. 1999, A&A 345, 953

Elisabete de Gouveia Dal Pino

ALFVEN WAVES IN DISKS, OUTFLOWS AND JETS Reuven Opher IAG - USP, Brazil [email protected]

Abstract

Observations of young stellar objects, such as classical T Tauri stars (CTTSs), show evidence of accretion disks (ADs), outflows and jets. It is shown that Alfven waves (AWs) are important in all of these phenomena. AWs can be created by turbulence in the AD and magnetosphere of a CTTS. Jets and stellar outflows as well as the heating of the central regions of ADs, may be due to the damping of AWs. Our recent investigations of the heating of the AD and the magnetosphere by AWs as well as the structures produced in outflows from ADs, initiated by turbulent AWs, are discussed. We comment on the possible importance of AWs in resolving the energy and angular momentum problems of CTTSs.

Introduction The current picture of a classical T Tauri star (CTTS) is that of a central protostar, surrounded by a thin accretion disk that was formed as the result of the collapse of matter in a molecular cloud. At the corotation radius, the disk is disrupted due to the star’s dipole magnetic field. The accreting matter then follows the star’s magnetic field lines until it impacts on the stellar surface. This magnetospheric model accounts for the observational signatures seen in CTTSs, such as the excess of optical and ultraviolet continuum flux (veiling) and redshift absorption features in the emission line profiles (inverse P Cygni profiles) (Muzerolle, Hartmann, & Calvet 1998; Hartmann, Hewitt, & Calvet 1994). CTTSs display both outflow and inflow signatures (see, e.g., Edwards, Ray, & Mundt 1993). Magnetic field lines on the order of 1 kG at the surface of the protostar are sufficient to disrupt the disk. The existence of magnetic fields of this magnitude on the star’s surface is inferred from observations (JohnsKrull et al. 1999; Guenther et al. 1999). Coupling of the accreting matter to the star’s magnetic field lines is possible if the temperature 311 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 311-316. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

312

OPEN ISSUES IN LOCAL STAR FORMATION

of the disk at the truncation radius is > 103 K, at which collisional ionization of metal ions comes into effect (Umebayashi & Nakano 1988). A major problem to be resolved concerns the gravitational energy EG of the accreted matter, which is over an order of magnitude greater than ¯ ∗ τ∗ of the CTTS, where L ¯ ∗ is the average the emitted energy E∗ = L optical luminosity and τ∗ is its lifetime. It has been suggested that the occurance of large ejections could possibly get rid of the excess gravitational energy, but no detailed analysis has been made so far. Another problem that must be dealt with is the large amount of angular momentum loss required for the formation of a protostar in a CTTS. The average protostellar mass M∗ in a molecular cloud has an angular momentum 4-5 orders of magnitude greater than a protostar in a CTTS. It has been suggested that the Balbus and Hawley Instability (BHI) (Balbus & Hawley 1991, 1998) in the accretion disk may be able to get rid of the excess angular momentum. Detailed calculations of this possibility have yet to be made, although it has been shown in 3D numerical simulations that angular momentum can, in fact, be transported outward in the disk by the BHI (Hawley & Stone 1998). Energy and angular momentum in CTTSs can be transported over large distances by Alfven waves which are generated by perturbations of the magnetic field, embedded in the plasma of the protostar. In general, AWs are weakly damped, as compared with sonic waves. The wave transports the energy and angular momentum away from the source of the perturbation, generally turbulence, along the magnetic field lines at the Alfven velocity. Classical plasma processes do not allow for accretion to take place in a disk. It is, thus, generally assumed that turbulence exists in the disk, in order to produce the anamolous viscosity required for accretion. This assumption justifies the existence of AWs in the disk. AWs have indeed been observed in a space physics environment, in particular, the solar wind. Near the sun, most of the fluctuations observed are propagating outwards (Bavassano 1990). Tu, Marsch and Thieme (1989) and Tu, Marsch and Rosenbauer (1990) found that, at a distance ∼ 64R
from the sun, the fluctuations are outward-going with periods of ∼ minutes to days. All of the observations indicate that these fluctuations are, in fact, AWs. The waves are observed to have a power law distribution. Jatenco-Pereira, Opher and Yamamoto (1994) showed that the observed AW flux can explain the observed velocity of the solar wind as well as the observed radial temperature distribution in the corona and chromosphere. In the calculations of JatencoPereira et al. (1994), the network heating of Parker (1991) with a flux of 8.7 × 105 ergs cm−2 s−1 , which has a damping length of 0.5 R
, was used. An Alfven flux of 1.3×105 ergs cm−2 s−1 was found to be necessary

Alfven Waves in Disks, Outflows and Jets (Opher)

313

in order to fit the observed data. A coronal hole geometry, in which surface Alfven damping (Lee and Roberts 1986, Hasegawa and Uberoi 1982) at the border was assumed, was studied. The calculations predicted a spectral break at 10−2 Hz for a distance of 64R
, as is observed. This Alfven flux produced a ∼ 600 kms−1 wind, as is also observed. In Section 1, our recent work on AW heating of the accretion disk and magnetosphere of CTTSs is presented. The structures produced in outflows from accretion disks, initiated by turbulent AWs, are treated in Section 2. Concluding remarks and a discussion of the energy and angular momentum problems in CTTS are given in Section 3.

1.

AW heating in the accretion disk and magnetosphere of a CTTS

Accretion in the disk about a CTTS is generally assumed to be due to the BHI. An ionized plasma is required in order for the BHI to occur. However, Gammie (1996) predicted that the accretion disk is unionized in its central region from 0.1 to 5 AU. Based on the evidence by Stone et al. (1996) that the BHI becomes turbulent, Vasconcelos, Jatenco-Pereira and Opher (2000) suggested that the BHI turbulence in the ionized regions produce AWs that propagate into the central region and heat it, so that the entire disk becomes ionized. They found that, assuming nonlinear mode coupling for the damping mechanism of an AW with a frequency 1/10 the ion cyclotron frequency, an AW amplitude of only 0.2% of the Alfven velocity was sufficient to create the necessary heating in the disk. The required temperature profile of the magnetosphere, based on observations, was calculated by Hartmann, Hewitt and Calvet (1994). Martin (1996) studied the heating of the accreting matter as it follows the magnetic field lines in the magnetosphere. He took into account adiabatic compression, photoionization and ambipolar diffusion. He found too low a temperature for the magnetosphere due to these heating mechanisms and concluded that an additional heating source must be present. Vasconcelos, Jatenco-Pereira and Opher (2002) found that for an Alfven frequency of 1/10 the ion cyclotron frequency, an AW amplitude of only 0.3% the Alfven velocity was required to attain the necessary temperature.

2.

Structures produced in the turbulent AW-initiated outflows from accretion disks

In a cold, non-turbulent disk, matter cannot leave the surface since it is bound gravitationally. A magnetic field perpendicular to the AD

314

OPEN ISSUES IN LOCAL STAR FORMATION

is generally assumed in CTTSs. Vitorino, Jatenco-Pereira and Opher (2002, 2003) studied the outflow from the accretion disk in a CTTS along the magnetic field lines, assuming, as did Ouyed and Pudritz (1997), that the corona of the disk is supported by a turbulent AW pressure PA , taken to be 0.39PKR , where PKR is the Keplerian ram pressure, defined as ρVK2 , and ρ and VK are the density in the disk and the Keplerian velocity, respectively. The PA balances the gravitational attractive force of the disk, allowing outflow to take place due to the centrifugal force. It is generally assumed that the twisting of the magnetic field lines due to the rotation of the disk will eventually transform the outflow into a jet. Gas pressure in the corona PG was taken to be 0.01PKR . The initial poloidal magnetic field, perpendicular to the disk, was assumed to be (8πPG )1/2 . It was assumed that the magnitude of the initial toroidal field was equal to that of the initial poloidal field, to which it was perpendicular. The density in the corona was taken to be 0.01ρ. The structure of the outflow after several hundred rotations of the inner Keplerian orbit was studied by Vitorino et al. (2002, 2003), using a 3D ZEUS code. Initially, they perturbed the outflow with a random velocity amplitude proportional to r−a , where r is the radius in the accretion disk and a = −3/2, −1, −1/2, 0, 1/2 and 1. The maximum random velocity perturbation at the inner orbit of the disk was 0.01 the Keplerian velocity. Although the perturbation was random, all the structures that formed had a spacing of 11 times the radius of the inner Keplerian orbit. Periodic perturbations of the outflow were also studied by Vitorino et al. (2003). In this case, the structures that formed had a spacing of T /2 times the radius of the inner Keplerian orbit, where T is the period of the perturbation. They investigated periods 10-80 times that of the inner Keplerian orbit TKi and found that the structures dissipated for small T (∼ 10 − 20TKi ) and tended to fragment for large T (∼ 60 − 80TKi ). It is to be noted that the investigations of Vitorino et al. (2000, 2003) are relevant for the formation of structures very near to the source, as yet unresolved with present telescopes. For example, the spacing of 11 times the radius of the inner Keplerian radius corresponds to a distance of only ∼ 10−6 pc.

3.

Conclusions

In Section 1, it was shown that AWs may be important in ionizing the central unionized region of an AD of a CTTS, making it possible for the BHI to operate and accretion to take place. It was also shown

Alfven Waves in Disks, Outflows and Jets (Opher)

315

in Section 1 that AWs may be important in raising the temperature of the magnetosphere of a CTTS, so that it agrees with observations. In Section 2, turbulent AW-initiated outflow from the disk was seen to create interesting structures. Finally, I would like to comment on the energy and angular momentum problems with respect to CTTSs which were discussed in the introduction. Concerning the problem of the excess angular momentum, it is possible that it may be transported away from the disk by AWs instead of the BHI. As noted, AWs can transport energy and angular momentum over great distances along magnetic field lines. Moreover, turbulent AWs can not only initiate mass outflow, as discussed in Section 2, but can also carry angular momentum out of the disk, depositing it in the molecular cloud. Concerning the energy problem, the major portion of the gravitational energy of the accreted matter is lost when its distance to the protostar decreases from ∼ 3R∗ to ∼ 1R∗ , where R∗ is the radius of the star. In the standard CTTS model, the accreting matter is found in the magnetosphere for radii ≤ 3R∗ . The presence of turbulence is assumed in the standard model for the accretion disk and we may assume that, due to the accreting matter, it is also present in the magnetosphere, where it can create large oscillations. In the region of the magnetosphere which touches the magnetic field lines that thread the disk, these oscillations can generate AWs. The AWs could transport appreciable energy outward from the magnetosphere into the molecular cloud.

Acknowledgments The author would like to thank the Brazilian agencies FAPESP (Proc. No. 00/06770-2) and CNPq (Proc. No. 300414/82-0) as well as the project PRONEX (Proc. No. 41.96.0908.00) for partial support.

References Balbus, S.A. and Hawley, J.F. (1991) ApJ, 376, 214 Balbus, S.A. and Hawley, J.F. (1998) Rev. Mod. Phys., 70, 1 Bavassano, B. (1990) Nuovo Cimento, 13, 79 Edwards, S., Ray, T. and Mundt, R. (1993) in Protostars and Planets III ed. E.H. Levy and J. Lunine Tucson: Univ. Arizona Press, 567 Gammie, C.F. (1996) 457, 355 Guenther, E.W., Lehmann, H., Emerson, J.P. and Staude, J. (1999) A&A, 341, 768 Hartmann, L., Hewett, R. and Calvet, N. 1994, ApJ, 426, 669 Hasegawa, A. & Uberoi, C. 1982, ”The Alfven Waves” (Washington D.C.: Tech. Inf. Center, US DOE) Hawley, J.F. and Stone, J.M. 1998, ApJ, 501, 758 Jatenco-Pereira, V., Opher, R., and Yamamoto, L.C. 1994, ApJ, 432, 409

316

OPEN ISSUES IN LOCAL STAR FORMATION

Johns-Krull, C.M., Valenti, J.A., Hatzes, A.P. and Kanaan, A. 1999, ApJ, 501, L41 Lee, M.A. and Roberts, B. 1986, ApJ, 301, 430 Martin, S.C. 1996, ApJ, 470, 537 Muzerolle, J., Hartmann, L. and Calvet, N. 1998, ApJ, 492, 743 Ouyed, R. and Pudritz, R.E. 1997, ApJ, 482, 712 Parker, E.N. 1991, ApJ, 372, 719 Stone, J.M., Hawley, J.F., Gammie, C.F. and Balbus, S.A. 1996, ApJ, 463, 656 Tu, C.Y., Marsch, E. and Rosenbauer, H. 1990, Geophys. Res. Lett., 17, 283 Tu, C.Y., Marsch, E. and Thieme, K.M. 1989, J. Geophys. Res., 94, 11739 Umebayashi, T. and Nakano, T. 1988, Prog. Theor. Phys. Suppl., 96, 151 Vasconcelos, M.J., Jatenco-Pereira, V. and Opher, R. 2000, ApJ, 534, 967 Vasconcelos, M.J., Jatenco-Pereira, V. and Opher, R. 2002, ApJ, 574, 847 Vitorino, B.F., Jatenco-Pereira, V. and Opher, R. 2002, Astron. Ap., 384, 329 Vitorino, B.F., Jatenco-Pereira, V. and Opher, R. 2003, ApJ, 592, 332

G. Medina, J. Gregorio-Hetem, V. Jatenco-Pereira, R. Opher

V

EARLY STAGES OF STAR FORMATION

This page intentionally left blank

SUBMILLIMETER STUDIES OF PROTOSTELLAR CORES Probing the Initial Conditions for Collapse Philippe Andr´e CEA Saclay, Service d’Astrophysique, France [email protected]

Abstract

1.

Improving our understanding of the initial conditions and earliest stages of protostellar collapse is crucial to gain insight into the origin of stellar masses, multiple systems, and protoplanetary disks. I discuss recent advances made in this area thanks to (sub)millimeter mapping observations with large single-dish telescopes and interferometers.

Collapse Initial Conditions: Theory

The inside-out collapse model of Shu (1977), starting from a singular isothermal sphere (SIS) or toroid (cf. Li & Shu 1996, 1997), is well known and underlies the ‘standard’ picture of isolated, low-mass star formation (e.g. Shu, Adams, & Lizano 1987). Other collapse models exist, however, which adopt different initial conditions. In particular, Whitworth & Summers (1985) have shown that there is a two-parameter continuum of similarity solutions to the problem of isothermal spherical collapse. One of the parameters measures how close to hydrostatic equilibrium the system is initially, while the other parameter reflects how important external compression is in initiating the collapse. In this continuum, the solutions proposed by Shu (1977) and Larson (1969)-Penston (1969) represent two extreme limits. All of the similarity solutions share a universal evolutionary pattern. At early times (t < 0), a compression wave (initiated by, e.g., an external disturbance) propagates inward leaving behind it a ρ(r) ∝ r−2 density profile. At t = 0, the compression wave reaches the center and a point mass forms which subsequently grows by accretion. At later times (t > 0), this wave is reflected into a rarefaction or expansion wave, propagating outward through the infalling gas, and leaving behind it a 319 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 319-330. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

320

OPEN ISSUES IN LOCAL STAR FORMATION

free-fall ρ(r) ∝ r−1.5 density distribution. Several well-known features of the Shu model (such as the expansion wave) are thus in fact common to all solutions. The various solutions can be distinguished by the absolute values of the density and velocity at t ∼ 0. In particular, the Shu (1977) solution has ρ(r) = (as 2 /2π G) r−2 (where as is the isothermal sound speed) and is static (v = 0) at t = 0, while the Larson-Penston (1969) solution is ∼ 4.4 times denser and far from equilibrium (v ≈ −3.3 as ). During the accretion phase (t > 0), the infall envelope is a factor ∼ 7 denser in the Larson-Penston solution. Accordingly, the mass infall rate is also much larger in the Larson-Penston case (∼ 47 as 3 /G) than in the Shu case (∼ as 3 /G). In practice, however, protostellar collapse is unlikely to be strictly selfsimilar, and the above similarity solutions can only be taken as plausible asymptotes. More realistic initial conditions than the SIS are provided by the so-called ‘Bonnor-Ebert’ spheres (e.g. Bonnor 1956), which represent the equilibrium states for self-gravitating isothermal spheres and have a flat density profile in their central region. Such spheres are stable for a center-to-edge density contrast < 14 and unstable for a density contrast > 14 (e.g. Bonnor 1956). Numerical hydrodynamic simulations of cloud collapse starting from such initial conditions (e.g. Foster & Chevalier 1993, Hennebelle et al. 2003) find that the Larson-Penston similarity solution is generally a good approximation near point-mass formation (t = 0) at small radii, but that the Shu solution is more adequate at intermediate t ≥ 0 times, before the expansion wave reaches the edge of the initial, pre-collapse dense core. In general, the mass accretion rate is thus expected to be time-dependent. Observationally, it is by comparing the (density and velocity) structure of prestellar cores such as L1689B (see § 2 and Fig. 1) with the structure of the envelopes surrounding Class 0 protostars such as IRAM 04191 (see § 3) that one may hope to constrain the initial conditions for collapse and to discriminate between the various existing models.

2.

Density Structure of Prestellar Cores

Two main approaches have been used to trace the density structure of cloud cores: (1) mapping the optically thin (sub)millimeter continuum emission from the cold dust contained in the cores, and (2) mapping the same cold core dust in absorption against the background infrared emission (originating from warm cloud dust or remote stars). Mapping the molecular gas component is generally less effective as most molecules tend to freeze out onto dust grains in the dense, cold inner parts of cloud cores (e.g. Walmsley et al. 2001 and Bergin 2003).

Submillimeter Studies of Prestellar Cores and Protostars (Andr´e)

321

Ward-Thompson et al. (1994, 1999) and Andr´e et al. (1996) employed the first approach to probe the structure of prestellar cores (see also Shirley et al. 2000). Under the simplifying assumption of spatially uniform dust temperature and emissivity properties, they concluded that the radial density profiles of isolated prestellar cores were not consistent with the single ρ(r) ∝ r−2 power law of the SIS but were flatter than ρ(r) ∝ r−1 in their inner regions (for r ≤ Rf lat ), and approached ρ(r) ∝ r−2 only beyond a typical radius Rf lat ∼ 2500–5000 AU. (a)

(b)

N ~ r i.e. ρ ∼ r

-1 -2

Figure 1. (a) ISOCAM 6.75 µm absorption image of the prestellar core L1689B (see color scale on the right in MJy/sr). The 1.3 mm continuum emission map obtained by Andr et al. (1996) at the IRAM 30 m telescope is superposed as contours (levels: 10, 30, 50 mJy/13 beam). (b) Column density profile of L1689B (crosses) derived from the absorption map shown in (a) by averaging the intensity over elliptical annuli for a 40◦ sector in the southern part of the core. The dashed curves show the most extreme profiles compatible with the data given the uncertainties affecting the absorption analysis. The solid curve represents the best fit of a Bonnor-Ebert sphere model (embedded in a medium of uniform column density), obtained with the following parameters: ρc /ρout = 40 ± 15 (i.e., well into the unstable regime), Tef f = 50 ± 20 K, Pext /kB = 5 ± 3 × 105 K cm−3 . For comparison, the dotted line shows the N H2 ∝ r¯−1 profile of a SIS at T = 10 K. (Adapted from Bacmann et al. 2000.)

More recently, the use of the absorption approach, both in the mid-IR from space (e.g. Bacmann et al. 2000, Siebenmorgen & Kr¨ ugel 2000) and in the near-IR from the ground (e.g. Alves et al. 2001), made it possible to confirm and extend the (sub)millimeter emission results, essentially independently of any assumption about the dust temperature distribution. The typical column density profile found by these emission and absorption studies of prestellar cores has the following characteristics (see, e.g., Fig. 1b): a) a flat inner region (of radius Rf lat = 5000 ± 1000 AU for L1689B according to the fitting analysis shown in Fig. 1b), b) a region roughly

322

OPEN ISSUES IN LOCAL STAR FORMATION

consistent with N H2 ∝ r¯−1 (corresponding to ρ ∝ r−2 for a spheroidal core), c) a sharp edge where the column density falls off more rapidly than N H2 ∝ r¯−2 with projected radius defining the core outer radius Rout (Rout = 28000 ± 1000 AU for L1689B). Note that, in contrast to prestellar cores, protostellar envelopes are always found to be strongly centrally condensed and do not exhibit any marked inner flattening in their radial (column) density profiles (e.g. Chandler & Richer 2000, Hogerheijde & Sandell 2000, Shirley et al. 2000, Motte & Andr 2001, Jørgensen et al. 2002).

2.1

Comparison with core models

The observational results summarized above set strong constraints on the density structure at the onset of protostellar collapse. The circularlyaveraged column density profiles can often be fitted remarkably well with models of pressure-bounded Bonnor-Ebert spheres, as first demonstrated by Alves et al. (2001) for B68. This is also the case of L1689B as illustrated in Fig. 1b. The quality of such fits shows that BonnorEbert spheroids provide a good, first order model for the structure of isolated prestellar cores. In detail, however, there are several problems with these Bonnor-Ebert models. First, the inferred density contrasts > 20–80 – see Fig. 1b) (from center to edge) are generally larger (i.e., ∼ than the maximum contrast of ∼ 14 for stable Bonnor-Ebert spheres (cf. § 1). Second, the effective core temperature needed in these models (for thermal pressure gradients to balance self-gravity) is often significantly larger than both the average dust temperature measured with ISOPHOT (e.g. Ward-Thompson et al. 2002) and the gas temperature measured in NH3 (Hotzel et al. 2002, Lai et al. 2003). In the case of L1689B, for instance, the effective temperature of the Bonnor-Ebert fit shown in Fig. 1b is Tef f ∼ 50 K, while the dust temperature observed with ISOPHOT is only Td ∼ 11 K (Ward-Thompson et al. 2002). Third, the physical process responsible for bounding the cores at some external pressure is unclear. These arguments suggest that prestellar cores cannot simply be described as isothermal hydrostatic structures and are either already contracting (see § 3 below) or experiencing extra support from static or turbulent magnetic fields (e.g. Curry & McKee 2000). One way to account for large density contrasts and high effective temperatures is indeed to consider models of cores threaded by a static magnetic field and evolving through ambipolar diffusion (e.g. Ciolek & Mouschovias 1994, Basu & Mouschovias 1995). In these models, at any given time prior to protostar formation, the cores are expected to feature a uniform-density central region whose size corresponds to the instan-

Submillimeter Studies of Prestellar Cores and Protostars (Andr´e)

323

taneous Jeans length. This agrees well with the characteristics of the flat inner regions seen in starless cores. Furthermore, the observed sharp edges (e.g. Fig. 1b) are consistent with the model predictions (shortly) after the formation of a magnetically supercritical core. Physically, this is because when a supercritical core forms, it collapses dynamically inward, while the outer, subcritical envelope is still efficiently supported by the magnetic field and remains essentially “held in place”. As a result, a steep density profile develops at the outer boundary of the supercritical core (see Fig. 8 of Basu & Mouschovias 1995). A serious problem, however, with ambipolar diffusion models involving only a static magnetic field is that they require a strong field in the ambient cloud (∼ 30–100 µG, see Bacmann et al. 2000), which seems to largely exceed the values found by Zeeman measurements (e.g. Crutcher 1999, Crutcher & Troland 2000).

2.2

Temperature distribution in starless cores

The models discussed above assume that prestellar cores are isothermal, which is quite a good first approximation (e.g. Larson 1969, Tohline 1982). In actual fact, however, there are good reasons to believe that the central regions of starless cores are somewhat cooler than their outer regions. Indeed, starless cores are heated only from outside by the local interstellar radiation field (ISRF), with no evidence for any central heating source (e.g. Ward-Thompson et al. 2002). In such a situation, recent dust radiative transfer calculations (Evans et al. 2001; Zucconi et al. 2001) predict that there should be a positive temperature gradient from the cores’ centers (with Td as low as ∼ 5–7 K) to their edges (at Td ∼ 15 K). The first calculations published by Evans et al. (2001) and Zucconi et al. (2001) assumed rather a simplistic input radiation field, based on current estimates of the average radiation field in the solar neighborhood (e.g. Black 1994). Bouwman et al. (2003) have more recently carried out similar calculations in which the levels of the diffuse mid-IR and far-IR backgrounds observed toward the cores are used to make more realistic estimates of the effective radiation field directly impinging on their surfaces. Fig. 2 gives an illustration for the prestellar core L1689B discussed earlier (see Fig. 1). The Bouwman et al. (2003) study (see also Stamatellos & Whitworth, this volume) confirms that the dust temperature generally reaches a minimum ≤ 10 K in the centers of prestellar cores and that the central temperature depends primarily on the central optical depth (directly related to the degree of shielding from the external ISRF). However, it appears that the minimum temperature may not be as low as initially

324

OPEN ISSUES IN LOCAL STAR FORMATION 20

Temperature [K]

18 16 14

This paper et al. 2003 Bouwman Evans Evans et.al. et al.(2001) 2001

12 10 8 6 2.5

Core edge

3.0

4.0 3.5 Log R [AU]

4.5

5.0

Figure 2. Comparison of two dust temperature profiles calculated for the prestellar core L1689B: the solid curve shows the model of Bouwman et al. (2003) based on the ISO observations of Bacmann et al. (2000) and Ward-Thompson et al. (2002); the dashed curve corresponds to the model favored by Evans et al. (2001).

reported by Evans et al. (2001) and Zucconi et al. (2001) and that the shape of the temperature profile may differ from that calculated by these authors: It can be seen in Fig. 2 that the major drop in temperature occurs in the outer parts of the core in the Bouwman et al. (2003) model of L1689B (solid curve), while it occurs closer to the center in the Evans et al. (2001) model (dashed curve). Interestingly, all models of L1689B have a relatively high mass-averaged dust temperature, comparable to the temperature of ∼ 11 K estimated by fitting a greybody to the global SED (cf. Fig. 13 of Ward-Thompson et al. 2002). The presence of a temperature gradient changes slightly the density profile expected in hydrostatic equilibrium and modifies the stability properties of prestellar cores compared to strictly isothermal BonnorEbert spheres, allowing stable equilibria with density contrasts up to < ∼ 40 (Galli et al. 2002). The latter effect remains small, however, and is restricted to a narrow range of core masses, so that it is unlikely to account for the large density contrasts observed in real prestellar cores (see Fig. 1b above). The final word on the structure of prestellar cores is likely to come from high-resolution mapping at far-IR and submillimeter wavelengths with future space-borne telescopes such as Herschel to be launched by ESA in 2007 (e.g. Pilbratt et al. 2001). Using Herschel images at 75–500 µm (in combination with ground-based dust continuum mapping at longer submillimeter wavelengths) to construct SED maps for

Submillimeter Studies of Prestellar Cores and Protostars (Andr´e)

325

at least the nearest, spatially resolved prestellar cores, it will be possible to reconstruct their intrinsic temperature and density distributions simultaneously (see Andr´e et al. 2003).

3.

Velocity Structure of Class 0 Objects

Briefly, low-mass prestellar cores are characterized by low levels of internal turbulence (e.g. Caselli et al. 2002), small rotational velocity < 1 km s−1 pc−1 – e.g. Goodman et al. 1993), gradients in general ( ∼ and subsonic, extended infall motions (e.g. Lee, Myers, & Tafalla 2001; Evans 2003). Many Class 0 protostars also show evidence for extended infall motions (e.g. Mardones et al. 1997, Mardones, this volume). Very few quantitative studies of the velocity fields exist, however. Here, we describe in some detail the results of two recent such studies for two contrasted Class 0 protostellar envelopes. The first object, IRAM 04191+1522 (IRAM 04191 for short), is relatively isolated. It was found at 1.3 mm by Andr´e, Motte, & Bacmann (1999) in the southern part of the Taurus cloud (d = 140 pc). IRAM 04191 is associated with a prominent, flattened protostellar envelope, seen in the (sub)millimeter dust continuum and in dense gas tracers such as N2 H+ , C3 H2 , H13 CO+ , and DCO+ (Belloche et al. 2002). All of the maps taken at the IRAM 30m telescope in small optical depth lines show a clear rotational velocity gradient across the envelope of ∼ 9 km s−1 pc−1 (after deprojection), perpendicular to the outflow axis. This gradient is one order of magnitude larger than those typically observed in starless cores (see above). Furthermore, the rotation of the protostellar envelope does not occur in a rigid-body, but differential, fashion: the inner ∼ 3500 AU-radius region rotates significantly faster than the outer parts of the envelope, as indicated by the characteristic “S” shape of the position-velocity diagrams shown by Belloche et al. (2002 – see their Fig. 3). Fast, differential rotation is expected in protostellar envelopes because of conservation of angular momentum during dynamical collapse. In the present case, however, the dramatic drop in rotational velocity observed at r ≥ 3500 AU (Fig. 3a), combined with the flat infall velocity profile (see below), points to losses of angular momentum in the outer envelope (see discussion in § 4). Direct evidence for infall motions over a large portion of the IRAM 04191 envelope is observed in optically thick lines such as CS(2–1), CS(3–2), H2 CO(212 − 111 ), and H2 CO(312 − 211 ). These lines are double-peaked > 40 from source center, which is indicaand skewed to the blue up to ∼ > 5000 AU (cf. Evans 1999, tive of infall motions up to a radius Rinf ∼ 2003). Radiative transfer modeling confirms this view, suggesting a flat,

326

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 3. Rotational velocity (a) and infall velocity (b) inferred in the IRAM 04191 envelope based on radiative transfer modeling of multi-transition CS and C 34 S observations with the IRAM 30m telescope (Belloche et al. 2002). The shaded areas show the estimated domains where the models match the observations reasonably well.

< r < 11000 AU subsonic infall velocity profile (Vinf ∼ 0.1 km s−1 ) for 3000 ∼ ∼ − 0 . 5 < for r ∼ 3000 AU (see and larger infall velocities scaling as Vinf ∝ r Fig. 3b and Belloche et al. 2002 for details). The mass infall rate is estimated to be M˙ inf ∼ 2 − 3 × a3s /G ∼ 3 × 10−6 M
yr−1 (with as ∼ 0.15 − 0.2 km s−1 for T ∼ 6 − 10 K), roughly independent of radius. Another Class 0 object whose kinematics has been quantified in detail is IRAS 4A in the NGC 1333 protocluster (Di Francesco et al. 2001). Using the IRAM Plateau de Bure interferometer, Di Francesco et al. (2001) observed inverse P-Cygni profiles in H2 CO(312 − 211 ) toward IRAS 4A, from which they derived a very large mass infall rate of ∼ 1.1 × 10−4 M
yr−1 at r ∼ 2000 AU. Even if a warmer initial gas temperature (∼ 20 K) than in IRAM 04191 and some initial level of turbulence are accounted for (see Di Francesco et al. 2001), this value of M˙ inf corresponds to more than ∼ 15 times the canonical a3ef f /G value < 0.3 km s−1 is the effective sound speed). This very high (where aef f ∼ infall rate results both from a very dense envelope (a factor ∼ 12 denser than a SIS at 10 K – see Motte & Andr´e 2001) and a large, supersonic infall velocity (∼ 0.68 km s−1 at ∼ 2000 AU – Di Francesco et al. 2001). Evidence for fast rotation, producing a velocity gradient as high as ∼ 40 km s−1 pc−1 , was also reported by Di Francesco et al. (2001).

Submillimeter Studies of Prestellar Cores and Protostars (Andr´e)

4.

327

Conclusions: Comparison to collapse models

In the case of isolated dense cores such as those of Taurus, the SIS model of Shu (1977) describes global features of the collapse reasonably well (e.g. the mass infall rate within a factor ∼ 2) and thus remains a useful, approximate guide. In detail, however, the extended infall velocity profiles observed in prestellar cores (e.g. Lee et al. 2001) and in the very young Class 0 object IRAM 04191 (§ 3 above) are inconsistent with a pure inside-out collapse picture (see also Mardones, this volume). The shape of the density profiles observed in prestellar cores are well fitted by purely thermal Bonnor-Ebert sphere models, but the absolute values of the densities are suggestive of some additional magnetic support (§ 2). The observed infall velocities are also marginally consistent with isothermal collapse models starting from Bonnor-Ebert spheres (e.g. Foster & Chevalier 1993, Hennebelle et al. 2003), as such models tend to produce somewhat faster velocities. This suggests that the collapse of ‘isolated’ cores is essentially spontaneous and somehow moderated by magnetic effects in magnetized, not strictly isothermal versions of Bonnor-Ebert cloudlets. Indeed, the contrast seen in Fig. 3 between the steeply declining rotation velocity profile and the flat infall velocity profile of the IRAM 04191 envelope beyond ∼ 3500 AU is very difficult to account for in the context of non-magnetic collapse models. In the presence of magnetic fields, on the other hand, the outer envelope can be coupled to, and spun down by, the (large moment of inertia of the) ambient cloud (e.g. Basu & Mouschovias 1994). Based on a qualitative comparison with the ambipolar diffusion models of Basu & Mouschovias, Belloche et al. (2002) propose that the rapidly rotating inner envelope of IRAM 04191 corresponds to a magnetically supercritical core decoupling from an environment still supported by magnetic fields and strongly affected by magnetic braking. A magnetic field strength ∼ 60 µG is required at 3500 AU (where nH2 ∼ 1−2×105 cm−3 ), which seems realistic (cf. Crutcher 1999). In this view, the inner ∼ 3500 AU radius envelope of IRAM 04191 would correspond to the effective mass reservoir (∼ 0.5 M
) from which the central star is being built. In protoclusters such as NGC 1333, by contrast, the large overdensity factors measured for Class 0 envelopes compared to hydrostatic isothermal structures (cf. Motte & Andr´e 2001), as well as the fast supersonic infall velocities and very large infall rates observed in some cases (§ 3), are inconsistent with self-initiated forms of collapse and require a strong external influence. This point is supported by the results of recent SPH simulations by Hennebelle et al. (2003). These simulations follow the

328

OPEN ISSUES IN LOCAL STAR FORMATION

SIS

Figure 4. Density profile (solid curve) obtained near point mass formation (t ≈ 0) in SPH numerical simulations of the collapse of an initially stable, isothermal (T = 10 K) Bonnor-Ebert sphere induced by a very rapid increase in external pressure (with Pext /P˙ext ∼ a tenth of the initial sound crossing time) (Hennebelle et al. 2003). Note the large overdensity factor compared to the ρ ∝ r−2 profile of a SIS at 10 K (dotted line).

evolution of a Bonnor-Ebert sphere whose collapse has been induced by an increase in external pressure Pext . Large overdensity factors (compared to a SIS), together with supersonic infall velocities, and large infall > 10 a 3 /G) are found near t = 0 when (and only when) the inrates ( ∼ s crease in Pext is strong and very rapid (e.g. Fig. 4), resulting in a violent compression wave. Such a violent collapse in protoclusters may be conducive to the formation of both massive stars (through higher accretion rates) and multiple systems (when realistic, non-isotropic compressions are considered). Future high-resolution studies with the next generation of (sub)millimeter instruments (e.g., ALMA) will greatly help test this view and shed further light on the physics of collapse in cluster-forming regions.

Acknowledgments I wish to thank my colleagues A. Bacmann, A. Belloche, J. Bouwman, P. Hennebelle, F. Motte, and D. Ward-Thompson for their important contributions to many of the results discussed here.

References Alves, J.F., Lada, C.J., & Lada, E.A. 2001, Nature, 409, 159 Andr´e, P., Bouwman, J., Belloche, A., & Hennebelle, P. 2003, in Chemistry as a Diagnostic of Star Formation, Eds. C.L. Curry & M. Fich, in press (astro-ph/0212492) Andr´e, P., Motte, F., & Bacmann, A. 1999, ApJ, 513, L57 Andr´e, P., Ward-Thompson, D., & Motte, F. 1996, A&A, 314, 625

Submillimeter Studies of Prestellar Cores and Protostars (Andr´e)

329

Bacmann, A., Andr´e, P., Puget, J.-L., Abergel, A., Bontemps, S., & Ward-Thompson, D. 2000, A&A, 361, 555 Basu, S., & Mouschovias, T.C. 1994, ApJ, 432, 720 Basu, S., & Mouschovias, T.C. 1995, ApJ, 453, 271 Belloche, A., Andr, P., Despois, D. & Blinder, S. 2002, A&A, 393, 927 Bergin, E.A. 2003, in Chemistry as a Diagnostic of Star Formation, Eds. C.L. Curry & M. Fich, in press (astro-ph/0211576) Black, J.H. 1994, in The First Symposium on the Infrared Cirrus and Diffuse Interstellar Clouds, ASP Conf. Ser., 58, p. 355 Bonnor, W.B. 1956, MNRAS, 116, 351 Bouwman, J., Andr´e, P., Galli, D. 2003, in preparation Caselli, P., Benson, P.J., Myers, P.C., & Tafalla, M. 2002, ApJ, 572, 238 Chandler, C.J., & Richer, J.S. 2000, ApJ, 530, 851 Ciolek, G.E., & Mouschovias, T.C. 1994, ApJ, 425, 142 Crutcher, R.M. 1999, ApJ, 520, 706 Crutcher, R.M., & Troland, T.H. 2000, ApJ, 537, L139 Curry, C.L., & McKee, C.F. 2000, ApJ, 528, 734 Di Francesco, J., Myers, P.C., Wilner, D.J., Ohashi, N., & Mardones, D. 2001, ApJ, 562, 770 Evans, N.J. II 1999, ARA&A, 37, 311 Evans, N.J. II 2003, in Chemistry as a Diagnostic of Star Formation, Eds. C.L. Curry & M. Fich, in press (astro-ph/0211526) Evans, N.J. II, Rawlings, J.M.C., Shirley, Y.L., & Mundy, L.G. 2001, ApJ, 557, 193 Foster, P.N., & Chevalier, R.A. 1993, ApJ, 416, 303 Galli, D., Walmsley, M., & Goncalves, J. 2002, A&A, 394, 275 Goodman, A.A., Benson, P.J., Fuller, G.A., & Myers, P.C. 1993, ApJ, 406, 528 Hennebelle, P., Whitworth, A.P., Gladwin, P.P., & Andr, P. 2003, MNRAS, 340, 870 Hogerheijde, M.R., & Sandell, G. 2000, ApJ, 534, 880 Hotzel, S., Harju, J., & Juvela, M. 2002, A&A, 395, L5 Jørgensen, J.K., Sch¨ oier, F.L., & van Dischoeck, E.F. 2002, A&A, 389, 908 Lai, S.-P., Velusamy, T., Langer, W.D., & Kuiper, T.B.H. 2003, AJ, in press (astroph/0303642) Larson, R.B., 1969, MNRAS, 145, 271 Lee, C.W., Myers, P.C., & Tafalla, M. 2001, ApJS, 136, 703 Li, Z.-Y., & Shu, F.H. 1996, ApJ, 472, 211 Li, Z.-Y., & Shu, F.H. 1997, ApJ, 475, 237 Mardones, D., Myers, P. C., Tafalla, M., Wilner, D. J., Bachiller, R., Garay, G. 1997, ApJ, 489, 719 Motte, F., & Andr, P. 2001, A&A, 365, 440 Penston, M.V., 1969, MNRAS, 144, 425 Pilbratt, G.L., Cernicharo, J., Heras, A.M., & Prusti, T. (eds.) 2001, The Promise of FIRST, ESA SP-460 Shirley, Y., Evans II, N.J., Rawlings, J.M.C., Gregersen, E.M. 2000, ApJS, 131, 249 Shu, F. 1977, ApJ, 214, 488 Shu, F.H., Adams, F.C., & Lizano, S. 1987, ARA&A 25, 23 Siebenmorgen, R., & Kr¨ ugel, E. 2000, A&A, 364, 625

330

OPEN ISSUES IN LOCAL STAR FORMATION

Tohline, J. E. 1982, Fund. of Cos. Phys., 8, 1 Walmsley, M., Caselli, P., Zucconi, A., & Galli, D. 2001, in The Origin of Stars and Planets: The VLT View, Eds. J. Alves & M. McCaughrean (Springer-Verlag, Berlin), ESO Astrophysics Symposia, p. 29 Ward-Thompson, D., Andr´e, P., & Kirk, J.M. 2002, MNRAS, 329, 257 Ward-Thompson, D., Motte, F., & Andr´e, P. 1999, MNRAS, 305, 143 Ward-Thompson, D., Scott, P.F., Hills, R.E., & Andr´e, P. 1994, MNRAS, 268, 276 Whitworth, A., & Summers, D. 1985, MNRAS, 214, 1 Zucconi, A., Walmsley, C.M., & Galli, D., 2001, A&A, 376, 650

Philippe Andr´e

DETECTION OF A COLLIMATED JET TOWARDS THE HIGH-MASS PROTOSTAR IRAS 16547−4247 Kate J. Brooks Universidad de Chile, Chile and European Southern Observatory, Chile [email protected]

Guido Garay Universidad de Chile, Chile [email protected]

Diego Mardones Universidad de Chile, Chile [email protected]

Abstract We report the discovery towards IRAS 16547–4247 of a chain of H2 2.12 µm emission knots that trace a collimated flow extending over 1.5 pc. IRAS 16547–4247 is a massive young stellar object with a luminosity of 6.2 × 104 L
. It is associated with a thermal radio jet and two nonthermal radio lobes which correspond to the working surfaces of the jet. The geometry of the H2 flow implies that it is driven by the thermal jet. We have also identified an isolated unresolved mid-infrared object associated with the jet and which is likely to be responsible for the excitation of IRAS 16547–4247.

Introduction Whether massive stars (B and O-type) form via an accretion process similar to that for low-mass stars (Osorio et al., 1999; McKee and Tan, 2003) or instead via collisions with lower-mass stars (Bonnell et al., 1998) is currently under debate. Highly collimated radio jets (Rodr´ıguez, 1997; 331 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 331-338. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

332

OPEN ISSUES IN LOCAL STAR FORMATION

Anglada et al., 1999) and Herbig-Haro (HH) flows (Reipurth et al., 2002) are frequently observed towards young low-mass stars. It is now widely accepted that jets are intimately linked to the accretion process and the formation of both bipolar molecular outflows and HH flows (see review by Reipurth and Bally, 2001). The role of collimated jets in massive stars (M > 8 M
) is less certain. In this higher mass regime observations of jets are difficult, primarily because of the greater distances involved and because the evolutionary time scales of such jets are expected to be much shorter. Of all the reported sources to have collimated radio jets, the three with the brightest luminosity are: HH80-HH81 ≈ 1.7 × 104 L
at 1.7 kpc (Mart´ı et al., 1998); IRAS 20126+4104 ≈ 1.3 × 104 L
at 1.7 kpc (Shepherd et al., 2000); and Cepheus A-HW2 ≈ 1 × 104 L
at 725 pc (Gomez et al., 1999). Here we present a series of new data towards IRAS 16547–4247, a young stellar object with a bolometric luminosity of 6.2 × 104 L
, equivalent to that of a single young O8 star. Our findings show that IRAS 16547–4247 is associated with a radio jet and collimated flow, making it the most luminous young stellar object known to be associated with such phenomena.

1.

Observations

IRAS 16547–4247 is one object from a sample of 18 luminous IRAS sources that we are studying in detail. The sources in our sample were chosen from the Galaxy-wide survey of CS(2–1) emission towards 843 IRAS sources with infrared colours typical of compact HII regions ( Bronfman et al., 1996). Each source selected for our sample showed line profiles indicative of either inward or outward motions and in some cases broad wings. Their IRAS luminosities were in the range 2 × 104 − 4 × 105 L
, implying that they all contain at least one embedded massive star (> 8 M
). These sources are thought to be representative of young massive star-forming regions. Our study incorporates radio continuum data at four frequencies (1.4, 2.5, 4.8 and 8.6 GHz) obtained with the Australia Telescope Compact Array (ATCA), 1.2-mm continuum data using the SIMBA 37-channel bolometer array installed at the SEST, as well as a series of SEST molecular-line transitions between 85 and 250 GHz. All of these data were obtained between May 2000 and July 2002. Subsequent near- and mid-infrared data centred on IRAS 16547–4247 were obtained in August 2002. Narrow-band images of the H2 1–0 S(1) 2.12 µm emission line, together with the adjacent continuum at 2.09 µm, were obtained using ISAAC mounted on the ANTU telescope of the Very Large Telescope

A collimated jet towards IRAS 16547−4247 (Brooks et al.) 4.8 GHz

333

8.6 GHz

Figure 1. ATCA maps of the radio continuum emission from IRAS 16547−4247. Beams are shown in the lower left corner of each panel. Contour levels are −1, 1, 2, 3, 5, 7, 9, 12, and 15 times 0.30 mJy beam−1 (1σ = 0.096 mJy beam−1 at 4.8 GHz and 0.070 mJy beam−1 at 8.6 GHz). Also shown are the OH maser position (+) from Caswell (1998) and the H2 O maser position (×) from Forster & Caswell (1989).

(VLT). A mid-infrared image was obtained using TIMMI2 mounted on the 3.6-m telescope in La Silla using the N11.9 µm filter and a pixel scale of 0.3 pixel−1

2.

Radio Continuum Emission

Fig. 1 shows the radio continuum maps at 4.8 and 8.6 GHz towards IRAS 16547–4247. The radio emission arises from three components aligned in a southeast-northwest direction. The outer components are symmetrically located in opposite directions from the central source, with peak positions separated by an angular distance of ∼20 arcsec (0.28 pc at the distance of 2.9 kpc, Bronfman, private communication). Using the radio data at all four frequencies the spectral indices for each of the three components were measured. The spectral index for the central object is 0.49 ±0.12 and is consistent with thermal emission produced by a biconical jet (Reynolds, 1986). The spectral indices of the emission from the outer components are negative (−0.61 ± 0.26 and −0.33 ± 0.04) and are consistent with shock-induced synchrotron emission (Blandford and Eichler, 1987). This non-thermal emission arises at the working surface of the jets, where the collimated winds from the jet interact with the surround medium. The actual thermal radio jet corresponds to the brightest 8.6-GHz emission component in Fig. 1b. There is a fainter source offset to the

334

OPEN ISSUES IN LOCAL STAR FORMATION

southeast and whose spectral index is not known. It is not certain what role (if any) this source plays. It is associated with an H2 O maser spot (Forster and Caswell, 1989). An OH maser is located towards the radio jet (Caswell, 1998). It should be noted that there was no detection of any methanol maser emission towards IRAS 16547–4247 (Walsh et al., 1998).

3.

1.2-mm Dust Emission

The 1.2-mm dust continuum emission detected towards IRAS 16547– 4247 is just resolved having a diameter of 0.4 pc and has a flux density of 16.3 Jy (see Fig. 3). Assuming a dust opacity at 1.2 mm of 1 cm2 g−1 ( Ossenkopf and Henning, 1994) we derive a mass of 1.3×103 M
. Results from the SEST line observations indicate the presence of a molecular gas core with a molecular hydrogen column density N(H2 ) of 6 × 1023 cm−2 , density n(H2 ) of 5 × 105 cm−3 and mass of 9 × 102 M
. These properties are similar to those of other massive star-forming cores (e.g. Garay et al., 2002). The triple radio source is centred on the 1.2-mm continuum emission peak.

4.

H2 2.12 µm emission

A map of the H2 2.12 µm emission towards IRAS 16547−4247 is shown in Fig. 2. There is a complex chain of emission knots extending over 110 arcsec (1.5 pc at the distance of 2.9 kpc). The H2 emission has the morphological characteristics of HH objects arising from the interaction of a collimated flow with the ambient medium. The outermost emission knots are approximately symmetrically offset from the radio jet and are in the shape of bow-shocks, all pointing away from the direction of the radio jet. Their arrangement is consistent with an elongated outflow cavity. Fig. 3 illustrates the comparison between the H2 emission (without any continuum subtraction) and (a) the 1.2-mm dust continuum emission and (b) the 8.6-GHz continuum emission. The chain of H2 emission is oriented in the same direction as the triple radio source and is contained within the molecular core traced by the 1.2-mm continuum emission. One of the H2 emission knots is associated with the southern non-thermal radio component. The striking alignment of the collimated flow delineated by the H2 emission and the central location of the radio jet implies that these phenomena are coupled. It is reasonable to assume that the radio jet is the driving force of the collimated flow.

A collimated jet towards IRAS 16547−4247 (Brooks et al.)

Non−thermal radio lobes

335

Thermal radio source

Figure 2. ISAAC H2 2.12 µm emission image (continuum subtracted) shown in grey scale. Overlaid with contours of the 8.6-GHz continuum emission (see Fig. 1).

336

OPEN ISSUES IN LOCAL STAR FORMATION

(a)

(b)

Non−thermal radio lobe

Thermal radio jet 5"

Non−thermal radio lobe

Figure 3. ISAAC H2 2.12 µm emission image (no continuum subtraction) shown in grey scale. (a) Overlaid with 1.2-mm dust continuum emission. Contour levels are 0.25 (5σ), 0.5, 0.75, 1, 1.5, 2.5, 3.5, 4.5, 5.5, 6.5 Jy beam−1 . (b) Overlaid with contours of the 8.6-GHz continuum emission (see Fig. 1). Also shown are the OH maser position (+) from Caswell (1998) and the H2 O maser position (×) from Forster & Caswell (1989). There is a positional uncertainty of 1 in registering the H2 2.12 µm and 8.6-GHz continuum data.

A collimated jet towards IRAS 16547−4247 (Brooks et al.)

5.

337

Mid-infrared Emission

The 11.9-µm image obtained in this study (with a field of view of 96 × reveals a single unresolved (< 0.6  ) emission source with a flux of 0.28 Jy. Within the 5  positional uncertainty of the mid-infrared data, this source is coincident with the radio jet and is likely to be responsible for the excitation of IRAS 16547–4247. No stellar counterpart to the radio jet was detected in the narrow-band near-infrared images (e.g. see Fig. 3b). 72 )

6.

Summary

IRAS 16547–4247 is the most luminous (6.2×104 L
) embedded young stellar object known to harbor a thermal radio jet. We have detected H2 2.12 µm emission that delineates a collimated flow extending over 1.5 pc. The alignment of the flow and the central location of the radio jet implies that these phenomena are intimately linked. We have also detected an isolated, unresolved 12 µm source towards the radio jet. Our finding of a jet and collimated flow towards such a luminous object supports the accretion scenario for the formation of stars across the entire mass spectrum.

Acknowledgments We thank Vanessa Doublier, Rachel Johnson, Nathan Smith, Michael Sterzik and Leo Bronfman for their help with observations and data reduction. The ISAAC data were obtained through the ESO Director’s Discretionary Time Program. This work has been partly funded by the Chilean Centro de Astrof´ısica FONDAP No 15010003.

References Anglada, G., Villuends, E., Estalella, R., Beltr´an, M. T., Rodr´ıguez, L. F., Torrelles, J. M., and Curiel, S. 1999, AJ, 116, 2953 Blandford, R. and Eichler, D. 1987, Phys. Rep, 154, 1 Bonnell, I. A., Bate, M. R., and Zinnecker, H. 1998, MNRAS, 298, 93 Bronfman, L., Nyman, L. ˚ A., and May, J. 1996, A&AS, 115, 81 Caswell, J. L. 1998, MNRAS, 297, 215 Forster, J. R. and Caswell, J. L. 1989, A&A, 213, 339 Garay, G., Brooks, K. J., Mardones, D., Norris, R., and G., B. M. 2002, ApJ, 579, 678 Gomez, J. F., Sargent, A. I., Torrelles, J. M., Ho, P. T. P., Rodr´ıguez, L. F., Canto, J., and Garay, G. 1999, ApJ, 514, 287 Mart´ı, J., Rodr´ıguez, L. F., and Reipurth, B. 1998, ApJ, 502, 337 McKee, C. F. and Tan, J. C. 2003, ApJ, 585, 850

338

OPEN ISSUES IN LOCAL STAR FORMATION

Osorio, M., Lizano, S., and D’Alessio, P. 1999, ApJ, 525, 808 Ossenkopf, V. and Henning, T. 1994, A&A, 291, 943 Reipurth, B. and Bally, J. 2001, ARevA&A, 2001, 39 Reipurth, B., Heathcote, S., Morse, J., Hartigan, P., and Bally, J. 2002, AJ, 123, 362 Reynolds, S. P. 1986, MNRAS, 304, 713 Rodr´ıguez, L. F. 1997, In Reipurth, B. and Bertout, C., editors, Herbig-Haro Flows and the Birth of Stars; IAU Symposium No. 182, page 83, Netherlands. Kluwer Academic Publishers. Shepherd, D. S., Yu, K. C., Bally, J., and Testi, L. 2000, ApJ, 535, 833 Walsh, A. J., Burton, M. G., Hyland, A. R., and Robinson, G. 1998, MNRAS, 301, 640

Kate Brooks

DISKS AND HALOS IN PRE-MAINSEQUENCE STARS A. S. Miroshnichenko Department of Physics and Astronomy, University of Toledo, USA [email protected]

D. Vinkovi´c Department of Physics and Astronomy, University of Kentucky, USA [email protected]

ˇ Ivezi´c Z. Astrophysical Sciences Department, Peyton Hall, Princeton University, USA [email protected]

M. Elitzur Department of Physics and Astronomy, University of Kentucky, USA [email protected]

Abstract

Two different geometries have been proposed to explain the dust emission from pre-main-sequence stars: flared disks and “classical” geometrically-thin optically-thick disks imbedded in optically thin halos. We show that only imaging observations can differentiate these two morphologies, while flux measurements can never distinguish between them. A model constructed with one geometry implies an equivalent model with the other that produces the identical flux at every wavelength. Even when its optical depth is much smaller than that of the disk, the halo can still have a significant effect on the disk temperature profile.

339 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 339-346. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

340

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction Star formation is expected to produce circumstellar disks during the pre main-sequence phase. Indeed, T Tauri stars (TTS) provide good evidence for the presence of such disks (e.g. Bertout 1989). In contrast, the predominant evidence from visual and IR observations of their intermediate mass (∼ 2–10 M
) counterparts, the Herbig Ae/Be stars (Haebes), is for a different morphology. The 50 and 100 µm imaging of numerous Haebes by Di Francesco et al. (1998) reveal sizes ≥ 104 AU, much larger than the ∼ 100–500 AU typical of accretion disks. Furthermore, the emitting regions are roughly circular and do not show the elongation that disks are expected to produce for all orientations other than faceon. Similar results show near-IR speckle interferometry by Leinert et al. (2001), who find extended (∼ 1000 AU) halos with no more than slight evidence for the elongations expected in a sample of disks in 18 of 31 Haebes. Circular shapes predominate also in the near-IR IOTA interferometry of 10 Haebes by Millan-Gabet, Schloerb & Traub (2001). These findings contrast sharply with those for TTS: using the same observing and reduction techniques and down to the same contrast level, Leinert et al. (1993) found evidence for near-IR halos in only 3 of 71 TTS. At the same time, the CO line and mm-continuum observations of Mannings & Sargent (1997) give conclusive evidence for disks. Miroshnichenko et al (1999, MIVE) note that the seemingly conflicting evidence for extended halos on one hand and disks on the other can be reconciled in a simple composite model: the dust distribution in Haebes is comprised of a compact, optically thick disk embedded in an extended, optically thin halo. This composite morphology works because, as noted earlier by Butner, Natta, & Evans (1994), the disk dominates the mm emission while the halo dominates the shorter wavelengths owing to the different temperature profiles of the two components. This role reversal gives rise to images that are extended at IR but compact at mm wavelengths. No single dust configuration can explain such a decrease. Chiang and Goldreich (1997, CG) suggest that the emission from TTS can be modeled purely in terms of flared disks, and fit a number of SEDs in terms of this morphology. Natta et al (2001) extend this proposal to Haebes. The underlying premise of this approach is that halos can be ignored in the studied sources because even if they do exist, their contribution is negligible in comparison with the flared disk. Here we compare our two-component disk+halo model with the CG flared disk model and suggest an observational test to recognize flared disks.

Disks and Halos in Pre-Main-Sequence Stars (Miroshnichenko et al.)

Figure 1. Geometry of the halo-imbeddeddisk model: a flat geometrically-thin optically-thick disk extends from the stellar surface to radius Rd . An optically thin spherical halo extends from the dust sublimation radius Rs to Rh . The small pillbox at the disk surface serves as a Gaussian surface for flux conservation.

1.

341

Figure 2. Top: Temperature profiles of a disk when heated only by a central star with T = 10,000 K (full line), and when imbedded in a spherical dusty halo with τV = 0.1 or 1, as marked. The halo starts at dust sublimation Ts = 1,500 K and its density profile is ∝ r−2 . Its temperature profile is also shown in each case. Bottom: The fractional contributions of the halo and (attenuated) stellar components to heating of the disk.

Halo-imbedded Disk

In this model the star is surrounded by a disk and a spherical dusty envelope (Fig. 1). The envelope extends from the inner radius Rs to some outer radius Rh = Y Rs . Thanks to scaling (Ivezi´c & Elitzur 1997), instead of these radii we can specify the dust temperature on each boundary. The halo is fully characterized by its dimensionless density profile η and optical depth τV . We assume a geometrically thin flat passive disk, i.e., negligible accretion. Because of its potentially large optical depth, the disk can extend inside the dust-free cavity where its optical depth comes from the gaseous component. The geometrically-thin disk assumption implies that the disk temperature varies only with radius, vertical temperature structure is ignored. This temperature is calculated from radiative flux conservation through the Gaussian surface (the small pillbox in Fig. 1).

342

OPEN ISSUES IN LOCAL STAR FORMATION

We performed detailed model calculations with the code DUSTY (Ivezi´c, Nenkova & Elitzur 1999) which takes into account the energy exchange between the star, halo and disk, including dust scattering, absorption and emission. Because its optical depth is typically τV ≤ 1, the halo is transparent to the disk emission in all the models we consider and we neglect the disk effect on the halo. In all the calculations, the spectral shapes qν of the grain absorption and scattering coefficients are those of standard interstellar mix, the sublimation temperature Ts = 1500 K. The spectral shape of the stellar radiation is taken from the appropriate Kurucz (1994) model atmosphere. Natta (1993) noted that imbedding the disk in a dusty halo can significantly affect its temperature even at small halo optical depths. Our calculations confirm this important point. Even though a halo with τV = 0.1 (Fig. 2) is almost transparent to the stellar radiation, it still causes a large rise in disk temperature. It contributes to the disk heating more than the star inside the dust-free cavity and dominates once the dust is entered. The important property evident in Fig. 2 is that the disk is much cooler than the envelope at all radii at which both exist and can also contain cooler material in spite of being much smaller. The role reversal between halo and disk domination of the SED affects also the image sizes. At IR wavelengths the image is dominated by the halo. A switch from envelope to disk domination provides a simple explanation for the otherwise puzzling decrease in the observed size of MWC 137 between 50 µm and 100 µm (Di Francesco et al. 1998). A similar effect was detected in the dust-shrouded main-sequence star Vega: its 60 µm size is 35 ± 5 (Van der Bliek, Prusti & Waters 1994), yet at 850 µm it is only 24×21 ± 3 (Holland et al. 1998).

2.

Flared Disks

The surface skin of any optically thick object is optically thin. CG note that it could become significant under certain flaring conditions. The stellar radiation penetrates to an optical distance τV = 1 along a direction slanted to the surface by angle α (Fig. 3). The optical depth ˆ is α of the corresponding skin layer along the normal to the surface n at visual and αqν at wavelength ν. The grazing angle of a flat thin disk whose inner radius is determined by dust sublimation is αflat =

α∗ , a

where α∗ =

4 R . 3π Rs

(1)

In Haebes α∗ ∼ 10−2 , and the optical depth of the surface layer cannot exceed this value. Flaring is defined by the radial profile of the disk height H ( ra ) or, equivalently, β = arctan H/ra  H/ra .

Disks and Halos in Pre-Main-Sequence Stars (Miroshnichenko et al.)

Figure 3.

343

Model geometry and notations for a flared disk and its CG surface layer.

Since α = γ −β (Fig. 3) where tan γ = dH/dra , then the grazing angle of a flared disk is α = adβ/da. The CG surface layer serves as an effective optically thin disk atop the underlying optically thick disk core. Since the radiative heating of optically thin dust is independent of geometry, the temperature profile of the CG layer is similar to that of a halo. Furthermore, emission from this surface layer provides additional heating for the underlying optically thick core, just like the halo. Therefore, the CG model can be considered a special case of the general family of models in which the disk is embedded in an optically thin halo, only in this case the “halo” is really the disk surface layer, fully determined from the flaring geometry. And since the CG layer exists in every disk, including flat, one may ask whether it alone suffices to explain all the observations and whether an additional putative halo is even necessary. To address these questions we must find radiative signatures that can discriminate between emission from the CG layer and a spherical halo. For any geometry, the flux from optically thin dust emission is obtained by integration over the volume occupied by this dust. In the case of spherical geometry, the flux can be written as 

Fsph,ν =

4πRs2 qν τV Bν (T ) η y 2 dy, D2

(2)

where T and η are given functions of radius y. For a face-on CG layer 

FCG,ν

2πRs2 = qν D2

Bν (T ) α ada.

(3)

Here T and α are given functions of cylindrical radius a. Since y and a enter only as integration variables, Eq. 2 and Eq. 3 are mathematically identical if  α(y) da η∝ and τV = 21 α . (4) y a

344

OPEN ISSUES IN LOCAL STAR FORMATION

This shows that it is impossible to distinguish a CG surface layer from a spherical halo with flux measurements because in either case we can construct a model with the other geometry and the exact same flux. In addition, the flux from the optically thick core of a CG flared disk and from a flat disk imbedded in the equivalent halo are identical because they both have the same temperature structure and that is the only relevant ingredient in optically thick emission. Therefore, the overall flux from a flared disk and from its equivalent halo-imbedded-disk are identical at all wavelengths. Only imaging can differentiate between these two geometries.

Figure 4. Points on the surface of a flared disk at equal distance from the star lie on a circle centered on the disk axis. The circle retains its shape in pole-on viewing but is deformed into an off-center ellipse in viewing from inclination angle i.

Figure 5. Viewing at inclination angles i = 25◦ , 45◦ and 65◦ of radiation scattered off the surface of a flared disk with 3 radius and constant grazing angle α = 4.9◦ . For each i, the left panel shows the image with brightness contours, the right panel the asymmetry factor A along three azimuthal directions, shown in the top left panel, whose angles are designated from the observer’s direction.

The brightness contours of face-on flared disks are concentric circles centered on the star, same as in spherical halos. Inclined viewing changes the contours substantially. The fundamental reason for image distortion by inclination is that the same projected distance from the star corresponds to widely different locations on the surface of the disk. On that surface, contours of equal distance from the star are circles of radius ra .

Disks and Halos in Pre-Main-Sequence Stars (Miroshnichenko et al.)

345

When viewed face-on from distance D, each contour appears as a concentric circle of radius θa = ra /D. At inclination viewing angle i to the disk axis, the contour is no longer circular. Absent flaring, the contour becomes an ellipse centered on the star with major axis 2θa and minor axis 2θa cos i, aligned with the projection of the disk axis on the plane of the sky. Flaring raises the contour to height H = ra tan β above the equatorial plane, and the star is shifted toward the observer along the minor axis by θa tan β sin i.

2.0.1 Image Asymmetry. Brightness contours not subject to rim occultation are ellipses with eccentricity e = cos i that directly determines the inclination angle irrespective of the flaring profile. The images shown in Fig. 5 possess an additional deviation from circular symmetry, conveniently measured by the ratio of brightness at diametric locations across an axis through the star A(θ, φ) =

I(θ, φ + π) − 1. I(θ, φ)

(5)

This asymmetry parameter vanishes for flat disks at all inclination angles and for pole-on viewing irrespective of the flaring. However, flaring introduces substantial asymmetry even at modest inclination angles (right panels in Fig. 5). Its systematic variation with azimuthal angles cannot be replicated by any other means, easily distinguishing it from deviations from the perfect geometry of idealized models or noise in the data. Each flaring profile produces its own asymmetry. Therefore, measuring A determines the flaring profile once the inclination is determined from the eccentricity of the brightness contours.

3.

Discussion

Our results reveal numerous degeneracies that underscore the severe limitations of attempts to determine the dust morphology from SED analysis without imaging observations. In spite of the attractiveness of the flared disk as a simple, physical model without additional components, imaging observations give irrefutable evidence for the existence of extended halos in many Haebes. Combining space- and ground-based observations of HD 100546 Grady et al (2001) resolve both the disk, extending to 5 , and a halo that extends out to 10 from the star. The origin of these halos has not been studied yet and could involve outflows such as disk winds (e.g., K¨ onigl & Pudritz 2000). Winds have been proposed to explain Haebes spectral line observations (e.g., Bouret, Catala & Simon 1997). Accretion rates of ∼ 10−8 M
yr−1 has also been deduced from UV spectra of both Haebes (Grady et al. 1996) and

346

OPEN ISSUES IN LOCAL STAR FORMATION

TTS (Gullbring et al 1998), and is consistent with halos with τV ∼ 0.1 (MIVE). It is possible that in the halos we get a glimpse of what would eventually evolve into the equivalent of the solar system’s Oort cloud. Whatever their origin, the halos seem prevalent in Haebes and would affect the emission from flared disks. The halo becomes the dominant component of the flux when it contains more dust than the haloequivalent of the disk (whenever τhalo > 41 H(Rd )/Rd , where τhalo is the optical depth across the halo at visual wavelengths). In TTS the halo inner cavity is much smaller than in Haebes, because of the lower T , and the importance of the halo contribution to disk heating is reduced relative to the star. Nevertheless, Kikuchi, Nakamoto & Ogochi (2002) conclude that halos are necessary supplements to flared disks also in flat-spectrum TTS. It appears that in many cases, flared disk in premain-sequence stars cannot be treated as if surrounded by vacuum.

Acknowledgments AM acknowledges financial support of the AAS International Travel Grant Program and LOC.

References Bertout C., 1989, ARA&A, 27, 351 Bouret, J.-C., Catala, C. & Simon, T., 1997, A&A, 328, 606 Butner H.M., Natta A. & Evans N.J. II, 1994, ApJ, 420, 326 Chiang, E.I. & Goldreich, P., 1997, ApJ, 490, 368 (CG) Di Francesco J., et al, 1998, ApJ, 509, 324 Grady C.A., et al., 1996, A&AS, 120, 157 Grady, C.A., et al, 2001, AJ, 122, 3396 Gullbring, E., Hartmann, L., Brice˜ no, C. & Calvet, N., 1998, ApJ, 492, 323 Holland, W.S. et al. 1998, Nature, 392, 788 ˇ Elitzur M., 1997, MNRAS, 287, 799 (IE) Ivezi´c Z., ˇ Nenkova M. & Elitzur M., 1999, User Manual for DUSTY, Internal Report, Ivezi´c Z., University of Kentucky, accessible at http://www.pa.uky.edu/∼moshe/dusty/ Kikuchi, N., Nakamoto, T., & Ogochi, K., 2002, PASJ, 54, 589 K¨ onigl, A. & Pudritz, R. E. 2000, Protostars and Planets IV, 759 Kurucz R.L., 1994, CD-ROM No.19, Smithsonian Astrophys. Observ. Leinert, Ch. et al., 1993, A&A, 278, 129 ´ Leinert, Ch., Haas, M., Abrah´ am, P., & Richichi, A. 2001, A&A, 375, 927 Mannings V. & Sargent A.I., 1997, ApJ, 490, 792 Millan-Gabet, R., Schloerb, F.P. & Traub, W.A., 2001, ApJ, 546, 358 ˇ Vinkovi´c, D. & Elitzur, M., 1999, ApJ, 520, L115 Miroshnichenko, A.S., Ivezi´c, Z., (MIVE) Natta, A. 1993, ApJ, 412, 761 Natta, A. et al., 2001, A&A, 371, 186 van der Bliek, N.S., Prusti, T. & Waters, L.B.F.M. 1994, A&A, 285, 229

RADIATIVE TRANSFER IN PRESTELLAR CORES: A MONTE CARLO APPROACH D. Stamatellos, A. P. Whitworth Department of Physics & Astronomy, Cardiff University, UK [email protected], [email protected]

Abstract

We use our Monte Carlo radiative transfer code to study non-embedded prestellar cores and cores that are embedded at the centre of a molecular cloud. Our study indicates that the temperature inside embedded cores is lower than in isolated non-embedded cores, and generally less than 12 K, even when the cores are surrounded by an ambient cloud of small visual extinction (AV ∼ 5). Our study shows that the best wavelength region to observe embedded cores is between 400 and 500 µm, where the core is quite distinct from the background. We also predict that very sensitive observations (∼ 1 − 3 MJy sr−1 ) at 170-200 µm can be used to estimate how deeply a core is embedded in its parent molecular cloud. Finally, we present preliminary results of asymmetric models of non-embedded cores.

Introduction Prestellar cores are cores that are either on the verge of collapse or already collapsing (e.g. Myers & Benson 1983, Ward-Thompson et al. 2002). They represent the initial stage of star formation and their study is important since theoretical models of star formation are very sensitive to the initial conditions. Prestellar cores have been observed either isolated or embedded in protoclusters. Isolated prestellar cores (e.g. L1544, L43, L63) have > 1.5 × 104 AU and masses 0.5 − 35 M extent ∼

(Ward-Thompson et al. 1999, Andr´e et al. 2000). On the other hand, prestellar cores embedded in protoclusters (e.g. in ρ Oph, NGC2068/2071) are generally smaller, with extent ∼ 2 − 4 × 103 AU and masses ∼ 0.05 − 3 M (Motte et al. 1998, Motte et al. 2001). 347 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 347-356. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

348

OPEN ISSUES IN LOCAL STAR FORMATION

In this paper, we present radiative transfer models of non-embedded and embedded prestellar cores, performed using a 3-D Monte Carlo radiative transfer code we have developed (Phaethon).

1.

The Method: Monte Carlo Radiative Transfer

Our method (Stamatellos & Whitworth 2003) is similar to that developed by Wolf, Henning & Stecklum (1999) and Bjorkman & Wood (2001). We represent the radiation field of a source (star or background radiation) by a large number of monochromatic luminosity packets (Lpackets). These L-packets are injected into the system and interact stochastically with it. If an L-packet is absorbed its energy is added to the local cell and raises the local temperature. To ensure radiative equilibrium the L-packet is re-emitted immediately with a new frequency chosen from the difference between the local cell emissivity before and after the absorption of the packet (Bjorkman & Wood 2001). This method conserves energy exactly, accounts for the diffuse radiation field and its 3-dimensional nature makes it attractive for application in a variety of systems. The code has been thoroughly tested using the thermodynamic equilibrium test (Stamatellos & Whitworth 2003) and also against benchmark (Ivezic et al. 1997) and previous (Bjorkman & Wood 2001) calculations.

2.

Non-Embedded Prestellar Cores

We represent prestellar cores by Bonnor-Ebert (BE) spheres, in which gravity is balanced by gas pressure (Bonnor 1956, Ebert 1955). In many cases this is a good approximation to prestellar cores (e.g. Alves et al. 2001, Ward-Thompson et al. 2002). For the radiation incident on the core, we use the Black (1994) interstellar radiation field (hereafter BISRF), that consists of an optical component due to radiation from giant stars and dwarfs, a component due to thermal emission from dust grains, mid-infrared radiation from non-thermally heated grains and the cosmic background radiation. The opacity of the dust at the low temperatures (5-20K) and high densities (104 −107 cm−3 ) expected in prestellar cores, is quite uncertain. In our study, we use the opacities calculated by Ossenkopf and Henning (1994), for grains that have coagulated and accreted thin ice mantles. We find that the temperature inside non-embedded cores drops from around 17 K, at the edge of the core, to a minimum at the centre, which maybe as low as 7 K, depending on the visual extinction to the centre

Radiative Transfer in Prestellar Cores (Stamatellos & Whitworth)

349

of the core. Our results are similar to those of previous studies (Evans et al. 2001, Zucconi et al. 2001).

3.

Embedded Prestellar Cores

The radiation field incident on cores embedded in molecular clouds is different from that incident on isolated non-embedded cores. The ambient molecular cloud absorbs the UV, optical and NIR part of the radiation and reemits it in the FIR and submm region (Mathis et al. 1983). It also makes the radiation incident on the core anisotropic, because, in general, the ambient cloud is not homogeneous. Another factor that contributes to the anisotropy of the radiation incident on an embedded core is the presence of stars or protostars in the vicinity of the core or the cloud (e.g as in ρ Oph; Liseau et al. 1999).

The Model. In this first approach (Stamatellos & Whitworth 2003), we study a spherical core at the centre of a molecular cloud of uniform density. The radiation incident on the molecular cloud is the BISRF, but the radiation incident on the core is enhanced in the FIR and submm and reduced at shorter wavelengths, as a result of the presence of the molecular cloud around the core. We chose the parameters of our models so as to mimic the embedded prestellar cores and the conditions in the ρ Oph protocluster (see Motte et al. 1998): core sizes 4 − 8 × 103 AU, masses 0.4 − 1.2 M , ambient cloud particle density ntot = 0.96×104 cm−3 and ambient cloud pressure ∼ 106 cm−3 K. In Figs. 1-2, we present our calculations for a supercritical core at gas temperature T =15 K with mass 0.8 M , under external pressure Pext = 106 cm−3 K, surrounded by a spherical ambient cloud with different visual optical depths. Core Temperature Profiles. We find that the presence of even a moderately thick cloud (AV = 5) around the core, results in a less steep temperature profile inside the core than in the case of a core that is directly exposed to the BISRF. When there is no surrounding cloud (Fig. 1, dashed lines), the temperature drops from ∼16 K at the edge of the core to around 6-7 K in the centre (∆T ≈ 9−10 K, depending on the core density), whereas with a τV = 5 ambient cloud (Fig. 1, dotted lines) the temperature drops from around 11 K to 7 K (∆T ≈ 4 K). Our studies show that dust temperatures inside embedded cores are probably lower than 12 K in cores surrounded by even a relatively thin cloud (AV ≈ 5), which seems to be the case for many of the prestellar cores in ρ Ophiuchi. Previous studies (Motte et al. 1998, Johnstone et al. 2000) of cores in this region, assumed isothermal dust at temperatures from 12 to 20 K,

350

OPEN ISSUES IN LOCAL STAR FORMATION

when calculating core masses from mm observations and, thus, they may have underestimated masses by a factor of 2. Recent studies (Andr´e et al. 2003, Andr´e, this volume) find similar temperature profiles, using a different approach, in which they estimate the effective radiation field incident on an embedded core from observations.

Figure 1. (a) Density, (b) dust temperature, and (c) SEDs, for a supercritical core (see text) surrounded by a spherical cloud with visual optical depth 20 (solid lines), 5 (dotted lines) and 0 (dashed lines). The dash-dot line on the SED graph corresponds to the background SED.

SEDs and Intensity Profiles. At 90 microns the core is seen in absorption against the background (Fig. 2a), and the intensity increases towards the edge of the core. For very centrally-condensed cores the decrease towards the centre is just ∼ 8 − 10 MJy sr−1 , but for less centrally-condensed cores it is even lower. Thus, very sensitive (say ∼ 1 − 3 MJy sr−1 ) observations are needed to detect cores in absorption at 90 µm.

Figure 2. Intensity profiles at (a) 90, (b) 170, and (c) 450 and 850 µm, for the models in Fig. 1. The horizontal solid lines on the profiles correspond to the background intensity at the wavelength marked on the graph.

Radiative Transfer in Prestellar Cores (Stamatellos & Whitworth)

351

At wavelengths near the peak of the core emission (150-250 µm) the intensity increases by a small amount (∼ 5 − 20 MJy sr−1 above the background) towards the edge of the core (Fig. 2b). The higher the increase in the intensity near the core boundary, the less embedded is the core. Thus, very sensitive observations of embedded prestellar cores at 170-200 µm, might allow us to determine the extinction of the cloud surrounding the core, and, thus, to estimate roughly the position of the core inside the cloud. However, more sophisticated modelling is required, with an accurate density profile for the cloud and taking into account the close environment of the core under study. Finally, at submillimeter and millimetre wavelengths (400-1300 µm) the intensity drops toward the edge of the core considerably (Fig. 2c). The core can be easily observed at 400-500 µm, where the contrast with the background is quite large (∼ 50 − 150 MJy sr−1 ). At wavelengths longer than ∼ 600 µm the background radiation becomes important and the core emission is not much larger than the background emission (∼ 20-50 MJy sr−1 larger).

Diagnostics. In Table 1, we list the peak intensities at various wavelengths for cores embedded in molecular clouds with visual optical depths 5 and 20. This table indicates that embedded cores are most easily distinguished from the background radiation around 450 µm. The peak Table 1.

Typical peak∗ intensities for embedded cores λ (µm) 90b 170 450 850 1300

Iλ a (MJy sr−1 ) τcloud = 5 τcloud = 20 5-15 10-15 55-160 20-80 10-40

∼3 ∼3 40-130 15-70 10-25

∗

The term peak refers to the maximum intensity above or below the background (as noted) at a specific wavelength. a Approximate peak intensities for a core embedded in a cloud with visual optical depth 5 and 20. The deeper the core is embedded the less distinct from the background is. The lower value corresponds to a subcritical core and the higher value to a supercritical (i.e. more centrally condensed) core. b At 90 µm the core seen in absorption against the background.

emission from embedded cores could be as low as ∼ 10 MJy sr−1 above the background at 1300 µm, but it’s at least ∼ 40 MJy sr−1 at 450 µm. The wavelength range between 400-500 seems favourable for observing embedded cores but the atmospheric transmission is not good in this

352

OPEN ISSUES IN LOCAL STAR FORMATION

range and space observations are needed. The upcoming Herschel (to be launched in 2007) will be operating in this range.

4.

Asymmetric Models of Non-Embedded Cores

The model. Generally, cores are not spherically symmetric. Here, we assume non-embedded cores with a disk-like asymmetry, i.e. the core is denser on the equatorial plane;

1+A ρ(r, θ) = ρc



r r0

1+

2

r r0

sin(θ)

2 2

,

(1)

where ρc is the the density at the centre of the core, r0 the scale length, and A is a factor that determines how asymmetric is the core. This density profile (Fig. 3, left) resembles the BE sphere density profile (drops as r−2 at large radii and gets flatter near the centre). In Fig. 4, we present our calculations for a core with nc = 106 cm−3 , r0 = 2 × 103 AU, Rcore = 2 × 104 AU, M = 7.3 M and e = 2.5 (e is the ratio of the optical depth at θ = 90, i.e. looking at an edge-on core, to the optical depth at θ = 0).

Figure 3. (a) Density, (b) dust temperature, and (c) SEDs, for an asymmetric core (see text). The core is divided into 15 equal-angle cells (each cell’s angle span is 12 o ). The lower density cell corresponding to θ = 0o (lower curve on the density plot), has higher temperature (upper curve in the temperature plot). The dotted line on the SED graph corresponds to the incident SED.

Temperature Profiles. The dust temperature is θ dependent (Fig. 3, centre). As expected, the ‘equator’ of the core is colder than the ‘poles’. The difference in temperature is around 3-4 K for the core under study (e = 2.5). The difference is expected to be larger for less symmetric cores.

Radiative Transfer in Prestellar Cores (Stamatellos & Whitworth)

353

SEDs and Images. The SED distribution (Fig. 3, right) of the core is the same at any viewing angle, because the core is optically thin to the radiation it emits (FIR and longer wavelengths). However, the isophotal maps are quite different for different viewing angles (Fig. 4). At 200 µm the outer, hotter parts of the core dominate the core emission. The core appears spherical when viewed pole-on and elongated when viewed edge-on. The image at 30o is quite interesting, with two absorption ‘blobs’ near the centre of the core. A quick search through the Kirk (2002) sample of ISO/ISOPHOT observations did not reveal cores with such distinctive features, which is surprising, because one excepts some of the observed cores to be viewed at such angles. Further study is required to see whether this is due to the sensitivity and resolution of observations, due to selection effects, or it is connected with the core structure and its environment (i.e. the ambient cloud). At longer wavelengths, like 850 µm, the core emission is regulated by the column density. Thus, when the core is viewed edge-on the intensity is larger at the centre. For the same reason the core looks elongated when the observer is looking at it at any other direction than pole-on.

5.

Conclusions

We applied a Monte Carlo radiative transfer method to spherical cores embedded inside molecular clouds. We find that the temperature inside these cores is less than 12 K, even for an ambient cloud with moderate visual extinction ∼ 5. We also studied asymmetric non-embedded cores and the preliminary results show that a small disk-like asymmetry in the density distribution will make the core look elongated when viewed at a random angle.

Acknowledgments We acknowledge help from the EC Research Training Network “The Formation and Evolution of Young Stellar Clusters” (HPRN-CT-200000155).

References Andr´e, P., Bouwman, J., Belloche, A., & Hennebelle, P., 2003, astro-ph/0212492, to appear in the proceedings “Chemistry as a Diagnostic of Star Formation” (C.L. Curry & M. Fich eds.) Andr´e, P., Ward-Thompson, D., & Barsony, M. 2000, Protostars and Planets IV, 59 Alves, J., Lada, C. J., & Lada, E. A. 2001, Nature, 409, 159 Black, J. H. 1994, ASP Conf. Ser. 58: The First Symposium on the Infrared Cirrus and Diffuse Interstellar Clouds, 355

354

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 4. Isophotal maps at 200 (left panel) and 850 µm (right panel) at viewing angles 0o , 30o and 90o (top to bottom), for a disk-like asymmetric core (see text). The core appears elongated when viewed at a direction different from θ = 0o . The colour image is available on the CD-ROM.

Bjorkman, J. E. & Wood, K. 2001, ApJ, 554, 615 Bonnor, W. B. 1956, MNRAS, 116, 351 Ebert, R. 1955, Zeitschrift Astrophysics, 37, 217 Evans, N. J., Rawlings, J. M. C., Shirley, Y. L., & Mundy, L. G. 2001, ApJ, 557, 193

Radiative Transfer in Prestellar Cores (Stamatellos & Whitworth)

355

Ivezic, Z., Groenewegen, M. A. T., Men’shchikov, A., & Szczerba, R. 1997, MNRAS, 291, 121 Kirk, J., PhD Thesis, Cardiff, 2002 Johnstone, D., Wilson, C. D., Moriarty-Schieven, G., Joncas, G., Smith, G., Gregersen, E., & Fich, M. 2000, ApJ, 545, 327 Liseau, R. et al. 1999, A&A, 344, 342 Mathis, J. S., Mezger, P. G., & Panagia, N. 1983, A&A, 128, 212 Motte, F., Andre, P., & Neri, R. 1998, A&A, 336, 150 Motte, F., Andr´e, P., Ward-Thompson, D., & Bontemps, S. 2001, A&A, 372, L41 Myers, P. C. & Benson, P. J. 1983, ApJ, 266, 309 Ossenkopf, V. & Henning, T. 1994, A&A, 291, 943 Stamatellos, D. & Whitworth, A. P., 2003, submitted to A&A Ward-Thompson, D., Motte, F., & Andre, P. 1999, MNRAS, 305, 143 Ward-Thompson, D., Andr´e, P., & Kirk, J. M. 2002, MNRAS, 329, 257 Wolf, S., Henning, T., & Stecklum, B. 1999, A&A, 349, 839 Zucconi, A., Walmsley, C. M., & Galli, D. 2001, A&A, 376, 650

D. Stamatellos, D. Boyd

356

OPEN ISSUES IN LOCAL STAR FORMATION

A. Miroshnichenko, N. Drake, J. Vink, E, Jilinski, M. Fernandes

S. Faundez, L. Chavarria, J. Pineda

VI

THE ISM CONDITIONS FOR STAR FORMATION

This page intentionally left blank

THE INITIAL CONDITIONS FOR STAR FORMATION Diego Mardones Universidad de Chile

Abstract

I review some of the observations that have enabled us to probe the kinematics of accretion onto protostars. Special emphasis is given to molecular line profile diagnostics which combined with radiative transfer provide both some of the clearest evidence of accretion onto protostars and also suggest limitations to the current star formation models. A parallel is drawn for the formation of high mass and low mass stars, and evidence is given to suggest they are all formed by accretion processes. Finally, I list some conclusions and current open problems in the field.

Introduction Stars are formed by the gravitational contraction of gas in dense molecular cloud cores. These are found both in isolation (e.g. Bok Globules; Bok & Reilly 1947), and within larger clouds (e.g. Taurus-like regions) or giant molecular clouds (GMCs). The star-forming nature of dense cores was definitely established with the association of pre-mainsequence stars detected by IRAS (Beichman et al. 1986). The bulk of the stars in the Galaxy are formed in a ”cluster environment” within massive cores in GMCs, however, a lot of work has been done in the past in smaller molecular clouds or Bok globules forming exclusively low mass stars. These regions are closer to the Sun, allowing studies with higher effective angular resolution and sensitivity to faint structures than in the more distant massive star forming regions. Furthermore, massive stars are more disruptive of their environment, evolve faster onto the main sequence, and are fewer in the Galaxy, making detailed observations of their formation process generally harder than in the low mass case. On the positive side, massive stars are brighter and can be detected throughout the Galaxy in directions or at wavelengths with low extinction. 359 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 359-366. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

360

OPEN ISSUES IN LOCAL STAR FORMATION

The physical processes important during the star formation process and the observational evidence supporting them are summarized in the excellent reviews by Evans (1999), and Garay & Lizano (1999). In broad terms, star formation can be characterized by the following series of events: (a) a dense cloud core undergoes an initial phase of dynamical contraction; (b) a flattened structure is formed channeling the accretion flow onto a disk and the protostar, and material is ejected in a bipolar outflow; (c) accretion eventually stops and as the disk material dissipates planets are likely to form. This cartoon was originally developed by Shu, Adams, & Lizano (1987) to explain the formation of low mass stars, but may be extended to higher mass stars if they are formed by accretion with the following caveats: (i) massive protostars evolve quickly onto the main sequence, producing a destructive HII region; (ii) the support of massive cores may be fundamentally different from that of low mass cores; (iii) bipolar outflows are ubiquitous around low mass protostars (e.g. Bachiller 1995), but are only recently being discovered around high mass protostars, the ejection mechanism may be different. In this contribution I will quickly review some of the fundamental observations that have shed light on the initial conditions (during the initial dynamical contraction phase) for low mass star formation and then recent discoveries about the formation of high mass stars.

1.

Low mass star formation

A core undergoing gravitational contraction is expected to show generally radial inward gas velocities localized around the accreting protostar(s), and accelerating towards the center. These kinematics are observed together with distracting effects such as bipolar outflows, rotation, turbulence, and spatial variation in chemistry and physical conditions within a core. Moreover, the gravity-induced gas velocities are expected to be generally smaller than those arising from non-thermal motions within dense cores; making their direct observation impossible. However, the observation and modeling of self-absorbed line profiles allows the detection of minute velocity differences between the foreground absorbing and the background emitting gas. The first very convincing arguments in favor of such interpretation were those of Zhou et al. (1993) in the isolated globule B335. Zhou et al. (1993) showed that the Shu (1977) inside-out collapse model provided a better fit to the data than the Larson (1969) model. They also argued against an independent foreground absorbing layer. Since the work of Zhou et al., many new observational and modeling results have made the case for infall much stronger. Gregersen et al.

Initial Conditions (Mardones)

361

(1997) and Mardones et al. (1997) showed that a significant fraction of the Class 0 sources (believed to be the youngest protostars) show evidence of inward motions in their observed molecular line profiles. However, as more observations were published, difficulties with the simplest models became apparent. Tafalla et al. (1998) detected very extended inward motions in the L1544 starless core. The spatial extent and radial velocity required to match the spectra could only be matched by ”old” inside-out collapse models, implying that a protostar should already be present in the core center but none is detected. Lee, Myers, & Tafalla (2001) found similar results towards a large number of starless cores. Thus, the large-scale core contraction motions are likely to be signs of core-formation rather than star-formation. More on starless cores can be found in the contribution by Andre to this volume. Star-forming cores also show widespread and large-scale evidence for inward motions. Typical inward speeds are 0.05 km s−1 at distances of 0.04 pc from the accreting protostars (Mardones et al. 2003). Moreover the model spectra are most sensitive to such large scale motions and are relatively insensitive to the expected inwards acceleration. Thus, even though the detection of contracting cores is now widely recognized, we are left with the uncertainty about the origin of such motions. As in the case of starless cores, we could be detecting remnants of core formation and not the telltale signatures of accreting protostars. In order to find the expected inward acceleration two approaches are possible. The first if to use interferometer techniques to filter out the smooth core kinematics. One of the most spectacular interferometer results to date was the detection of red-shifted absorption against the mm-continuum source in NGC1333-4 by Di Francesco et al. (2001) and Choi et al. (2002). These are the most detailed kinematic studies to date, yielding inward speeds of 0.5–0.7 km s−1 at distances of ∼0.003 pc from a 0.5 M
protostar. However, earlier attempts to observe B335 at high spatial resolution gave negative results, suggesting that strong CS depletion and outflow dominates the spectral line profiles at resolutions of an arc-second (Wilner et al. 2000). The second approach to detect the inwards acceleration close to the protostars relies on line wings in optically thin tracers. Detecting weak line winds is generally harder than detecting self-absorption close to the line centers. Moreover if the tracers are required to be optically thin. Unsuccessful attempts have been made using N2 H+ and N2 D+ transitions with the IRAM 30-m telescope (Bourke, private communication). Finally, submm-continuum observations are sensitive to the dust column density and temperature and provide complementary information to the kinematics learned from the line profiles. Furthermore, optical

362

OPEN ISSUES IN LOCAL STAR FORMATION

and infrared extinction can yield column densities at very high spatial resolution under the right conditions, see the article by Alves in this book for more information.

2.

High mass star formation

Two significantly different scenarios for massive star formation have been proposed. The first assumes the rapid formation of an HII region prevents significant accumulation of mass onto a central protostar, and propose instead that massive stars are formed from the coalescence of lower mass stars in a dense cluster environment (e.g. Bonnel, Bate, & Zinnecker 1998). This model is supported by the fact that massive stars are not found in isolation. However, at typical proto-cluster densities the coalescence time-scales are too long. The alternative scenario assumes massive stars are formed under such high mass accretion rates that the expansion of the (proto)HII region is quenched by mass loading (e.g. Osorio, Lizano & D’Alessio 1999). Thus, the high mass star formation process may be an extension of the low mass case, including the formation of disks and the fast ejection of matter along the poles. In recent years we have carried out an extensive observational effort to determine the physical conditions in high mass star forming regions. We used initially the SEST telescope mm-wavelength receivers for multitransition and multi-molecule observations. We then obtained 1.2-mm continuum data at the SEST and cm-wavelength continuum images at the ATCA. To select the sources we searched for ”interesting” line profiles, meaning those with clear evidence of self-absorption and or high velocity wings. To start the search we looked at every one of the CS 2–1 line profiles observed by Bronfman, Nyman, & May (1996) towards 1300 galactic sources with IRAS colors of UCHII regions. We selected the southern sources (δ < −20◦ ) from the full sample yielding a total of 639 spectra, 90% of which are likely come from massive star forming regions, the rest being nearby low mass protostars. It is interesting to note that the bulk of the line profiles are fairly symmetric, unlike the low mass case. Out of 639 spectra we detect evidence of wings in 95 sources, but only 14 strong cases. We detect apparent self-absorption in 53 sources, 26 of which are indicative of overall expansion motions and 27 of contraction. A third of the self-absorbed line profiles are very clear. We can put these figures in context by comparing the excess of sources with evidence of inward vs outward motions in the sample defined as E = (Nin − Nout )/Ntotal (Mardones et al. 1997).

Initial Conditions (Mardones)

363

Lee, Myers, & Tafalla (2000) detect an infall excess of E ∼ 25% in a sample of 69 starless cores. Gregersen et al. (1997) find E ∼ 20 − 30% in a sample of 23 Class 0 sources observed in HCO+ . Mardones et al. (1997) find E ∼ 40 − 50% in a sample of 24 Class 0 sources and 10% among 23 Class I sources, using CS and H2 CO lines. Williams & Myers (1999) find E ∼ 10% of 19 ”cluster” sources observed in CS. These are all regions of intermediate mass star formation. This work: E ∼ 0.2% of 639 massive sources We must note that since this sample lacks optically thin transitions, some of the self-absorbed line profiles may actually arise from the presence of multiple velocity-components along the line of sight, however, these cases are likely to be few. The bulk of the CS 2–1 spectra from Bronfman et al. (1996) are not dominated by either radial velocities, in which case self-absorption would be more common. This is in part an effect of spatial resolution, and likely an effect of ”complex structures” within the telescope beam, which tends to make lines Gaussian. However, we could still select sources with interesting line profiles as stated above for further studies. Since self-absorbed lines are rare in this sample, they must indicate special conditions, perhaps in only these cases the initial large-scale core kinematics have not yet been erased by the star formation process. Thus, we selected 18 sources for detailed studies with the SEST and ATCA. Here I comment on the two most interesting sources: IRAS 16272–4837 at a distance of 3.4 kpc and has an IRAS luminosity of 2 × 104 L
. From the CS and HCO+ line profiles we derive an average infall speed of vin ∼ 0.5 km s−1 and mass accretion rate of 10−3 M
yr−1 . The SiO and SO spectra show outflow wings at velocities vout ∼ 20 km s−1 (Garay et al. 2002). There is no continuum emission detected down to Sν < 0.2 mJy at 4.8 GHz, making this source an excellent candidate to be the equivalent of a ”massive class 0 protostar.” IRAS 16547-4247 at a distance of 2.9 kpc and has an IRAS luminosity of 6 × 104 L
(O8 star) has a triple radio continuum source. This is the jet arising from the most massive protostar so far (see Brooks in this book for details). This last radio jet provides perhaps the best evidence to date in support for an accretion model of massive star formation.

364

3. 3.1

OPEN ISSUES IN LOCAL STAR FORMATION

Conclusions Low mass star formation

Widespread inward motions have been detected towards a significant fraction of starless cores and Class 0 sources. The spatial scales of these observations are inconsistent with pure inside-out collapse models; the initial conditions in star forming cores can’t be static. Radiative transfer models of the observed molecular line profiles suggest typical inward motions of 0.05 km s−1 are required at distances of 0.04 pc from the protostars. However, the observed line profiles are relatively insensitive to the gas velocities much closer to the protostars, making model comparisons hard to do. Continuum observations of star-forming cores suggest that the inner density profiles are rather flat, consistent with Bonnor-Ebert spheres and also with inside-out collapse predictions.

3.2

High mass star formation

Self-absorbed molecular line profiles are much rarer in massive cores than in low mass cores. These probably arise from more complex kinematics within the radio-telescope beam due to poorer resolution and to multiple sources within the beam. On the other hand, the rare sources with self-absorbed line profiles are likely to be kinematically simpler on large scales, making them promising case studies. The accretion model of star formation is likely applicable at least up to stars of spectral type O8, much higher than previously thought possible. This is possible in cores having very large mass accretion rates (∼ 10−3 M˙
yr−1 ) that can effectively quench the development of compact HII regions which stop the in-flowing gas. Properties of dense cores fill the ∆v vs R space, there is no evidence for ”bimodal” core properties, but instead there is a continuum from the smallest to the largest cores.

4. 4.1

Open Issues Low mass star formation

No molecular cloud radiative transfer modeling has been able to reproduce all observed spectral line profiles: multi-transition, multi-molecule, multi-spatial-scales. In order to fit all observations with the models it is likely we will need to incorporate detailed radius and/or time-dependent chemistry into the model clouds. It is also likely that 2-dimensional or fully 3-dimensional cloud models will be required.

Initial Conditions (Mardones)

365

We can’t decouple the core-formation from the (later) star-formation process. We still have a very poor understanding of the end of the star formation process, thus, no real links to observed IMF are possible yet. In particular, the role of bipolar winds, of companions, and of core boundaries needs to be explored further.

4.2

High mass star formation

What is the maximum stellar mass that can be attained by accretion? Is coalescence still needed to form the highest mass stars? Bipolar jets have been observed to arise from high mass protostars, but their observed frequency is still relatively low. Are the bipolar winds (jets) arising from high mass protostars common? Are they similar in structure and origin to the low mass case? Massive protostars are formed close to the centers of massive dense cores. Are the physical conditions within massive cores similar (scaled) in structure to the lower mass cores? Are the radial profiles in density and temperature similar? The study of the inner structure of massive dense cores will require the advent of sensitive long-baseline radio interferometers like ALMA, expected to be completed in 2012.

Acknowledgments The author wishes to thank Guido Garay, Kate Brooks, and Andr´es Escala for helpful discussions; and the FONDAP Center for Astrophysics 15010003 for financial support.

References Beichman, C.A., Myers, P.C., Emerson, J.P., Harris, S., Mathieu, R., Benson, P.J., & Jennings, R.E., 1986, ApJ, 307, 337 Bok, B.J., & Reilly, E.F., 1947, ApJ, 105, 255 Bonnel, I.A., Bate, M.R., & Zinnecker, H., 1998, MNRAS, 298, 93 Bronfman, L., Nyman, L.A., & May, J., 1996, A&A, 115, 81 Di Francesco, J., Myers, P.C., Wilner, D.J., Ohashi, N., & Mardones, D., 2001, ApJ, 562, 770 Evans, N.J.II 1999, ARA&A, 37, 311 Garay, G., & Lizano, S. 1999, PASP, 111, 1049 Garay, G., Brooks, K.J., Mardones, D., Norris, R.P., & Burton, M.G., 2002, ApJ, 579, 678 Gregersen, E.M., Evans, N.J. II, Zhou, S., & Choi, M. 1997, ApJ, 484, 256 Lee, C.W., Myers, P.C., & Tafalla, M., 2001, ApJ Suppl., 136, 703 Mardones, D., Myers, P.C., Tafalla, M., Wilner, D.J., Bachiller, R., & Garay, G. 1997, ApJ, 489, 719 Mardones, D., Myers, P.C., Tafalla, M., Bachiller, R., Wilner, D.J., & Garay, G. 2003, ApJ, in press

366

OPEN ISSUES IN LOCAL STAR FORMATION

Osorio, M., Lizano, S., & D’Alession, P., 1999, ApJ, 525, 808 Shu, F.H. 1977, ApJ, 214, 488 Tafalla, M., Mardones, D., Myers, P.C., Caselli, P., Bachiller, R., and Benson, P.J., 1998, ApJ, 504, 900 Tafalla, M., Myers, P.C., Caselli, P., Walmsley, C.M., & Comito, C., 2002, ApJ, 569, 815 Wilner, D.J., Myers, P.C., Mardones, D., & Tafalla, M. 2000, ApJ, 544, L69

Diego Mardones

CHEMICAL SIGNATURES OF NEW STAR FORMATION TOWARD YOUNG STELLAR CLUSTERS Jonathan Williams, Sandrine Bottinelli Institute for Astronomy, Hawaii, USA jpw,[email protected]

Carlos Rom´an-Z´ un ˜iga University of Florida, USA [email protected]fl.edu

Abstract

Molecular abundances vary on dynamically short timescales in the dense gaseous envelopes around young stars. Modeling and observations of the chemistry in molecular clouds suggest the use of different species as clocks to follow the evolution of a dense core toward star formation. We present interferometric and singledish observations of CS and N2 H+ in three cluster forming regions, Serpens, Cepheus A, and the Rosette, that show potential new sites of star formation.

Introduction Cold molecular clouds, the birthplaces of stars, emit most strongly at far-infrared and millimeter wavelengths. As the most abundant observable molecule, observations of CO are frequently used to determine global cloud properties such as mass, size, and linewidth (Dame et al. 1987). Clouds are hierarchically structured, however, and stars only form in gravitationally bound, dense clumps (Williams, Blitz, & McKee 2000). Many molecular lines, not just from CO, can be observed in such clumps and, as the sensitivity and mapping speed of millimeter wavelength instrumentation has increased, it has become possible to study the chemistry of star forming regions (e.g., Bergin et al. 1997). In this contribution we present millimeter wavelength observations of N2 H+ and CS toward cluster forming clumps in nearby clouds. Since 367 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 367-374. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

368

OPEN ISSUES IN LOCAL STAR FORMATION

most stars, including all massive stars, form in small groups (Adams & Myers 2001) such observations tell us about the chemical conditions in typical star forming environments. In addition, clustered regions are ideally suited for comparative and statistical studies which can constrain evolutionary scenarios. A practical drawback, however, is that clustered regions are, by nature, crowded and molecular emission associated with different protostars can blend together. Nevertheless, although the study of cluster formation is less advanced than isolated star formation, observations of different molecular species may provide new clues to their origin and early evolution.

1.

Chemical clocks

The abundance of different molecules is observed to vary from one star forming region to another and may signify different evolutionary states (Tafalla et al. 2002). As a clump collapses to higher densities, the shielding of external dissociating radiation increases and the frequency of particle collisions increase. Some molecules deplete onto dust grain surfaces which may open up chemical pathways for other molecules to increase in abundance. The modeling of Bergin & Langer (1997) shows that abundances of several observable molecules can change by an order of magnitude over relatively short timescales. In particular, CS tends to form rapidly but significantly depletes after several 105 yr. On approximately the same timescale, however, the abundance of N2 H+ increases by about an order of magnitude. Since the free-fall timescale of the dense cores at which these tracers are observed, nH2 ∼ 105 cm−3 , is tff = 2 × 105 yr, there should be observable chemical changes as a core dynamically evolves. An example of how chemistry may correlate with stellar content and core dynamical state in the Serpens NW cluster forming region is shown in Figure 1 (adapted from Williams & Myers 1999). This map, as with those in Figures 2 and 3, is a combination of singledish (FCRAO) and interferometer (BIMA) data which achieves relatively high resolution, ∼ 10 , without loss of flux from extended emission. CS emission (upper left panel) shows two cores, one of which contains a Class 0 source, FIRS1, while the other appears starless. Although the CS emission is of similar strength in each core, the N2 H+ is much weaker in the starless 5 core (lower left panel), suggesting a chemically youthful state, < ∼ 10 yr, compared to the FIRS1 core. The correlation between chemistry and star forming content extends also to the dynamical state of the cores. The right panels show the integrated CS and N2 H+ spectra of each core. In each case the CS is self

Chemical signatures in young stellar clusters (Williams et al.)

369

Figure 1. Correlated chemistry, dynamics, and star formation in two cores in Serpens NW. The upper left panel shows integrated CS emission from two cores, one containing the Class 0 protostar FIRS1, indicated by the star symbol, and the other starless. Despite the CS being of similar strength, the starless core is substantially weaker in N2 H+ emission, plotted in the lower left panel. Average spectra toward each core are plotted in the right panels; CS emission is self-absorbed in each case but the spectrum of the starless core (lower panel) is asymmetric indicating collapse.

absorbed but whereas the spectrum is symmetric toward the star forming core (upper right), indicating overall balance between infall and outflow, it is asymmetric towards the starless core (lower right) suggestive of core collapse just prior to star formation. We have since mapped several nearby young stellar clusters in CS and N2 H+ with the goal of seeing how far this result can be generalized and extended. Our preliminary results are reported below.

2.

Preliminary results

Serpens SE A small group of Class 0 sources lies just a few arcminutes to the SE of the cores in Figure 1 (Testi & Sargent 1998). Integrated maps of CS and N2 H+ emission show a very different morphology which may be due to evolutionary differences in the cores. The N2 H+ generally follows the location of the Class 0 sources but there are several cores with strong CS and weak (or undetected) N2 H+ emission. By analogy with Serpens NW we suspect that these are pre-stellar cores. The CS

370

OPEN ISSUES IN LOCAL STAR FORMATION

linewidth is slightly larger, however, and the optical depth is too low for the line to be self-absorbed. It is therefore not clear from these data alone whether these cores are collapsing or not. (Observations of a different, abundant dense gas tracer, such as HCO+ , may be useful.) We note also that there are two linear features with coherent velocity gradients in the CS map that appear to be bipolar outflows.

Figure 2. BIMA+FCRAO map of CS(2-1) and N2 H+ (1-0) in Serpens SE. Stars show the position of 3 mm continuum sources from deeply embedded (Class 0) protostars. Several regions with strong CS and weak N2 H+ emission are outlined. These may be chemically young, pre-stellar cores.

Cepheus A Cepheus is a more distant (725 pc), more luminous (LFIR = 2.4×104 L
) star forming region than Serpens. VLA observations show a chain of bright continuum knots from outflow-ambient gas shocks (Garay et al. 1996). Our BIMA observations reveal two 3 mm continuum sources that appear to be the embedded protostars (or protostellar groups) driving the multiple outflows in this region. Figure 3 shows the centimeter and millimeter continuum sources overlaid on CS and N2 H+ line maps. As with Serpens, the CS is more widely spread than the N2 H+ which clumps into 3 main regions (also seen in NH3 ; Torrelles et al. 1993). In this case however, unlike Serpens above and other low mass star forming regions (e.g. NGC 1333; Di Francesco et al. 2001), the N2 H+ does not peak toward the dust condensations. The different CS/N2 H+ chemistry in this region may be due to the strong outflows or enhanced radiation field. We are currently investigating the excitation and dynamics of the gas through multi-transition observations and radiative transfer modeling.

Chemical signatures in young stellar clusters (Williams et al.)

371

Figure 3. BIMA+FCRAO map of CS(2-1) and N2 H+ (1-0) in Cepheus A. Stars show the position of two BIMA 3 mm continuum sources that are likely embedded protostars. The circles show the location of centimeter VLA sources due to shock interactions from protostellar outflows. To the west, away from the star forming center, lies a prominent, potentially pre-stellar, clump with strong CS but weak N 2 H+ emission.

As with Serpens, we find a prominent core with strong CS but weak N2 H+ emission. Its location away from the main star forming center suggests that the radiation and outflows are not likely to be a major influence on its chemistry. The core is not only strong in CS(2-1) but also in CS(5-4) which suggests a high central density, nH2 ∼ 106 cm−3 .

Rosette The above observations establish the existence of dense, compact cores within cluster forming regions that lack millimeter continuum sources from deeply embedded Class 0 protostars. Their chemical state, high CS abundance relative to N2 H+ , suggests that they are young, 5 < ∼ 10 yr, and perhaps pre-stellar. However, a newly formed star may have a similar effect; by heating up the surrounding core, CS can be liberated from grain mantles, while N2 H+ can be destroyed by its outflow. An alternative possibility, therefore, may be that the cores are substan6 tially older, > ∼ 10 yr, and harbor more evolved Class II/III protostars. Such stars should be observable in the near-infrared. In part to examine this possibility, we have carried out a joint nearinfrared/millimeter mapping study of young stellar clusters in the Rosette cloud. Details of the infrared observations are in Rom´ an-Z´ un ˜iga, Williams, & Lada (2003). Using the IRAM 30 m we mapped several lines, including 13 CO(2-1), CS(2-1), and N2 H+ (1-0) at ∼ 15−20 resolution. We are following up this work with JCMT SCUBA observations to measure the

372

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 4. Gas and dust toward a cluster in the Rosette molecular cloud. The upper left panel shows 13 CO emission outlining the relatively low density gas, the upper right panel shows the total dust column density, and the bottom panels show CS and N2 H+ emission from dense gas. The cluster is breaking out of its molecular envelope which breaks up into three main clumps each with a distinct chemistry. Faint red sources, perhaps newly formed stars, are found in the west most cluster where there is strong dense gas emission.

dust column density directly. Unlike the above interferometric studies of the closer Serpens and Cepheus A regions, these observations provide a large scale view of the chemistry of the envelopes around entire clusters rather than individual star forming cores within a single cluster. Seven clusters were mapped and, again, we found that different species showed a different morphology. Now, however, we found significant chemical variations over much larger (0.5 pc) scales. Dense gas emission was clearly detected toward all seven clusters but in five of the seven, an additional clump was found adjacent to the cluster. Close inspection of deep near-infrared imaging showed that several of these offset clumps contained very red, faint sources (K ∼ 16, J − K > ∼ 3). The high frequency with which we found these clumps and the red sources within them suggest that they are the sites of the next generation of star formation within the cloud. In Figure 4, we see a cluster surrounded on 3 sides by molecular gas. The appearance is of the stars breaking out of their cocoon. There are three condensations in the 13 CO and dust continuum emission but the dense gas is concentrated in the west most

Chemical signatures in young stellar clusters (Williams et al.)

373

clump, well offset from the cluster. A handful of faint red sources are found in this clump and may represent the next stage of star formation in this region. It is tempting to speculate that their formation has been induced by compression of the gas surrounding the main cluster. Detailed examination of the spectral energy distribution of these sources will be necessary to address this issue. We will also use the SCUBA dust continuum map to determine absolute abundances of the different molecules for comparison with chemical models.

3.

Summary

The formation of clusters is an inherently complex process but observations of the chemistry in these regions have the potential to increase our understanding of the timescales and processes involved and even how star formation might propagate across a cloud.

Acknowledgments This work is supported by NSF grant AST-0324328, “The formation of stars in groups”

References Adams, F. C. & Myers, P. C. 2001, ApJ, 553, 744 Bergin, E. A., Ungerechts, H., Goldsmith, P. F., Snell, R. L., Irvine, W. M., & Schloerb, F. P. 1997, ApJ, 482, 267 Bergin, E. A. & Langer, W. D. 1997, ApJ, 486, 316 Dame, T. M., Ungerechts, H., Cohen, R. S., de Geus, E. J., Grenier, I. A., May, J., Murphy, D. C., Nyman, L.-A., & Thaddeus, P. 1987, ApJ, 322, 706 Di Francesco, J., Myers, P. C., Wilner, D. J., Ohashi, N., & Mardones, D. 2001, ApJ, 562, 770 Garay, G., Ramirez, S., Rodriguez, L. F., Curiel, S., & Torrelles, J. M. 1996, ApJ, 459, 193 Rom´ an-Z´ un ˜iga, C. G., Williams, J.P., & Lada, E. A. 2003, RevMexAA (Serie de conferencias), 15, 206 Tafalla, M., Myers, P. C., Caselli, P., Walmsley, C. M., & Comito, C. 2002, ApJ, 569, 815 Testi, L. & Sargent, A. I. 1998, ApJ, 508, L91 Torrelles, J. M., Verdes-Montenegro, L., Ho, P. T. P., Rodriguez, L. F., & Canto, J. 1993, ApJ, 410, 202 Williams, J. P., Blitz, L., & McKee, C. F. 2000, in Protostars and Planets IV (Tucson: University of Arizona Press) eds Mannings, V., Boss, A. P., Russell, S. S., p. 97 Williams, J. P. & Myers, P. C. 1999, ApJ, 518, L37

374

Jonathan Williams

OPEN ISSUES IN LOCAL STAR FORMATION

COLLAPSE OF MOLECULAR CLOUDS LEADING TO STAR FORMATION A. Allen, H. W. Chuang, M. Choi Academia Sinica Institute of Astronomy and Astrophysics

Z. Y. Li University of Virginia

Abstract

Models of molecular cloud cores have been proposed by Frank Shu (Tsing-Hua U.) and Zhi-Yun Li (UVa). These models have been used as initial states for magnetohydrodynamic collapse calculations to study the dynamics of the early stages of star formation. As with the unmagnetized, singular isothermal sphere, constant central accretion rates are found for magnetized toroids. When rotation is present, low-velocity outflows are seen, transporting angular momentum away from the forming protostar into the ambient cloud.

It is generally believed that stars form from the condensation of dense cores of molecular clouds. For some reason, perhaps a trigger or a slow drift of physical parameters beyond a point of unstable force balance, these dense cloud cores collapse and grow to very high densities, eventually forming stars. There are many stages of collapse. There is a rich variety of physics during the collapse processes. Aside from the theoretical considerations, there is a rapidly growing library of observations for comparison and explanation. Molecular cloud cores start in a condition of near force balance. We know this because of their observed velocity distributions and the fact that we can see them (e.g. if they were far from force balance, they wouldn’t last long enough to see). The simplest model for force balance involves two opposing forces. There is always, of course, gravity. But what would oppose gravity in a cloud core? Cloud cores are composed of 375 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 375-382. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

376

OPEN ISSUES IN LOCAL STAR FORMATION

molecular gas, so the internal collisions of the molecules will provide gas pressure. Gravity tries to pull everything together, but pressure tries to push everything apart. Gravity is easy to treat, but what about gas pressure? The cooling time is much shorter than the crossing time for a molecular cloud, e.g. the time it takes to radiate away energy is much less than the time for gas to transverse the cloud, so we can assume the cloud to be isothermal. If we write down the equations for force balance between gravity and pressure, a static 1/r2 density distribution can be easily derived. A cloud can exist in this configuration for a while, but it is an unstable equilibrium; if by some means, the center of the cloud were to condense by eating out a hollow sphere, then the density shell above the hollow would feel no pressure and fall in. Like dominoes, successive gas shells would fall in and a collapse wave would move outward at the isothermal sound speed. The mathematics of this model were fully explored by Frank Shu in 1977 (Shu, 1977). An important result is a constant central accretion rate of a3 /G which also gives a timescale to star formation (M˙ ∗ = 1M
/550, 000yr). The isothermal sphere in pre-collapse and collapsing states is shown in Figures 1ac. This simple model has had great success in star formation, but there is still much room for improvement. There are other forces in space besides gravity and gas pressure. One such force (or more precisely, a fictitious force dependent on reference frame) is due to momentum conservation in a rotating frame–the centrifugal force. This will add support to a rotating molecular cloud in a direction perpendicular to the axis of rotation. A cloud that has some initial rotation, as clouds often exhibit (usually attributed to galactic shear–the galaxy does not show uniform angular velocity as a function of radius, so a cloud occupying a finite range of galactic radii will have some net rotation about the cloud’s center), can also exist in an equilibrium configuration. Such a cloud is an extension of the isothermal sphere to non-zero rotational velocities and its equilibrium structures were explored by Toomre and Hayashi, Narita & Miyama in 1982 (Toomre, 1982; Hayashi et al. ,1982). Like their non-rotating limit, the isothermal sphere, they have the same unstable equilibrium that can lead to runaway collapse. Unlike the isothermal sphere, however, rotating isothermal clouds can not collapse to form a point-mass protostar. Because angular momentum must be conserved and there are no means to transport angular momentum in this simple model involving only three forces, collapse will form a rotationally supported disk. With axisymmetry, the disk may never collapse to a point mass, although some non-axisymmetric ejection instability might perhaps allow a point mass to form by some complicated, chaotic mech-

Collapse of Molecular Clous (Allen)

377

anism. An isothermal toroid initially rotating at 0.25 sound speeds is shown in Figures 1bd in pre-collapse and collapsing states. Another commonly observed feature in molecular clouds is the presence of magnetic fields. Magnetic fields, through their trapping of charged particles along field lines, provide another means of cloud support. Since charged particles can not cross field lines, and since the time for neutral particles to drift through an impediment of charged particles (the ambipolar diffusion timescale) is longer than the dynamical timescales, the field will provide non-isotropic support for the molecular clouds, similar to the support that rotation gives. Such magnetically supported equilibrium structures were explored by Li and Shu (1996). These clouds are also a generalization of the isothermal sphere to include magnetism. Like their unmagnetized cousins, they are subject to a runaway collapse instability. Collapse will proceed quickly along field lines, but slower in perpendicular directions due to the field freezing that requires field lines to be dragged along with the fluid. Collapse toward the midplane will form a pancake like structure, named a pseudodisk by Galli & Shu (1992ab). As the pancake forms, the central protostar feeds from its inner edge, growing in mass linearly with time, as in the case of the unmagnetized, isothermal sphere, but at an accretion rate slightly higher due to the ability of the magnetic field to support extra mass within a given radius. Still threaded by a frozen-in magnetic field, the footpoints of the magnetic field lines are swallowed by the protostar as it dines. Magnetic pressure will push the field lines as far apart as possible. Since they are tied together at the origin, a split monopole will be formed as the field lines fan out isotropically, with inbound(outbound) lines in the lower(upper) half-space. A magnetized, isothermal toroid in pre-collapse and collapsing states is shown in Figures 2ac. Of course, magnetism and rotation can exist in the same cloud. Such clouds can still form structures with force balance (Allen, Li, & Shu, 2003). Such clouds are also subject to runaway collapse. In the presence of magnetic fields, however, angular momentum is not able to stop the growth of a point-mass. A forming pointmass protostar is bound to the split monopole threaded within–if the protostar rotates with finite speed, the field lines must rotate as well. But, the field lines are bound outward to the cloud. Thus, the inner cloud must slow down in angular speed and the outer cloud must speed up. This transport of angular momentum by magnetic braking allows the protostar to form at the origin; as matter passes inward, it passes it’s angular momentum outward by a toroidal field drilling upward. As the toroidal field passes away from the protostar, it will drag a portion of the embedding cloud with it. If the cloud disperses or the fields decouple from the star through ambipolar

378

OPEN ISSUES IN LOCAL STAR FORMATION

diffusion or magnetic reconnection, processes expected in the later stages of star formation, then braking efficiency can drop and enough angular momentum can remain in the cloud to form a planetary system after the central protostar has formed. A magnetized, isothermal toroid initially rotating at 0.25 sound speeds is shown in Figures 2bd in pre-collapse and collapsing states. As physics are added to the collapse models, the results improve and come into greater agreement with observation and expectations about the early stages of star formation. The magnetohydrodynamics simulation results shown in Figures 1 and 2 were obtained using a modified version of zeus2d (Norman & Stone, 1992ab). Details of the modifications and initial conditions, results for different levels of rotation and magnetization, and discussions of the features can be found in Allen, Shu, & Li (2003) and Allen, Li, & Shu (2003) and the references therein. To form a bridge between theory and observation, one must start from the most complete set of data and extract the less complete set. Therefore, this burden must be carried by the theorist, who has access to all variables everywhere in his simulations, to extract what an observer can see only from one side, far away, with a finite field of view. To create expected observations from theory, the density and velocity fields must be integrated along lines of sight, with microturbulent linewidths and appropriate radiative transfer, and convolved with the beam pattern of the target telescope or antennaes. Early examples of such efforts are shown in Figures 3 and 4, where the clouds are assumed to be optically thin and pencil beams are used for the beam patterns. As there is work remaining to make the theory more complete, work remains and is ongoing to finish the bridge from theory to observation.

References Allen A., Shu, F. H., & Li, Z. Y. 2003a, ApJ, submitted May 16, 2003. Allen A., Li, Z. Y., & Shu, F. H. 2003b, ApJ, submitted May 16, 2003. Galli, D., & Shu, F. H. 1993a, ApJ, 417, 220 Galli, D., & Shu, F. H. 1993b, ApJ, 417, 243 Hayashi, C., Narita, S., & Miyama, S. M. 1982, Prog. Theor. Phys., 68, 1949 Li, Z. Y., & Shu, F. H. 1996, ApJ 472, 211 Shu, F. H. 1977, ApJ, 214, 488 Stone, J. M., & Norman, M. L. 1992a, ApJ Suppl, 80, 753 Stone, J. M., & Norman, M. L. 1992b, ApJ Suppl, 80, 791 Toomre, A. 1982, ApJ, 259, 535

Collapse of Molecular Clous (Allen)

379

Figure 1. At top are initial states of exact force balance for (left) the unmagnetized, non-rotating, isothermal sphere and (right) an unmagnetized, isothermal toroid with uniform rotational velocity of 0.25 sound speeds. Colored isodensity contours are highlighted with solid contour lines every order of magnitude in density. At bottom are snapshots of the collapsing sphere and toroid at time t = 2.5e12s. 3 dimensional unit vectors show the direction of fluid flow with the speed indicated by the dotted contour lines in units of the sound speed. The collapse wave can be easily seen in the lower right panel–outside the collapse region, unit vectors point into the page and are invisible.

380

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 2. At top are initial states of exact force balance for (left) the magnetized, non-rotating, isothermal sphere and (right) a magnetized, isothermal toroid with uniform rotational velocity of 0.25 sound speeds. Isodensity contours are highlighted with solid contour lines every order of magnitude in density. Solid Black magnetic field lines are shown with field strength given by the dashed-dotted β contours, where β is the ratio of gas pressure to magnetic pressure. At bottom are snapshots of the collapsing toroids at time t = 2.5e12s. 3 dimensional unit vectors show the direction of fluid flow with the speed indicated by the dotted contour lines in units of the sound speed. In the lower panels, the split monopole is easily seen at the origin, as well as a high-density pseudo-disk in the midplane. At lower right, a slow outflow can be seen near the polar axis and the collapse wave can be easily seen–outside the collapse region, unit vectors point into the page and are invisible.

Collapse of Molecular Clous (Allen)

381

Figure 3. At left are shown moment-0 plots showing the column densities that would be observed during the collapse of the cloud shown in Figure 2bd. Moment-1 plots of density weighted velocity are shown in center. Moment-2 plots of density weighted velocity dispersion are shown at right. Time increases from top to bottom. Linear scales run from −1017 cm to 1017 cm. Please see the CD for the color figure.

382

OPEN ISSUES IN LOCAL STAR FORMATION

Figure 4. At left are shown moment-0 plots showing the column densities that would be observed during the collapse of the cloud shown in Figure 2d. Moment-1 plots of density weighted velocity are shown in center. Moment-2 plots of density weighted velocity dispersion are shown at right. Viewing angle, as measured from the polar axis, increases from top (pole view) to bottom (equatorial view). Linear scales run from −1017 cm to 1017 cm. Please see the CD for color figure.

STAR FORMATION IN GLOBULES Rolf Chini, Marcus Albrecht Astronomisches Institut Ruhr-Universit¨ at, Bochum, Germany [email protected]

Luis Barrera Instituto de Astronom´ıa Universidad Cat´ olica del Norte, Antofagasta, Chile

Katrin K¨ ampgen Astronomisches Institut Ruhr-Universit¨ at, Bochum, Germany

Markus Nielbock European Southern Observatory Santiago, Chile

Abstract

A complete sample of 169 small and opaque southern Bok globules is currently surveyed at 10 and 1200 µm to investigate the formation of individual stars independent of violent environmental effects. According to the association with IRAS sources, the sample has been divided into FIR-loud and FIR-quiet objects. Preliminary results for 36 FIR-loud sources suggest different classes of globules as witnessed by their ratio of bolometric luminosity to gas mass and - as a consequence - by the number and mass of embedded objects. Nine FIR-quiet globules show signs of fragmentation and seem to be at the border of gravitational instability. The present study indicates that about 75% of the FIR-loud and 30% of the FIR-quiet globules do actively form stars.

383 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 383-388. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

384

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction Young clusters are likely to be responsible for producing the majority of stars in the Galaxy. They contain significant numbers of stars covering in general a large range in mass with similar chemical composition in a small volume at the same distance from the sun. Therefore, clusters provide unique opportunities for the study of the early phases of star and planet formation in a statistically meaningful way. However, our current models of star formation predominantly handle the formation of isolated stars – without the influence of nearby massive stars, their stellar winds and their hard radiation field. In contrast, small isolated dark clouds, also known as Bok globules, are the ideal laboratory to study the low–mass star formation. Their simple structure allows them to be modeled in detail, since the number of physical processes occurring within them is likely to be limited. As a result of their optical appearance, in particular the lack of foreground stars, the Bok globules are judged to be relatively nearby (< 500 pc), and so linear scales of a few tenths of a parsec may be readily resolved by the current instrumentation. Their isolation also means that they are relatively unaffected by the problems associated with the study of the larger, more complex molecular clouds, such as large–scale motions which may mask the smaller scale motions of interest, energetic HII regions, and multiple star formation within one cloud which may destroy much of the evidence sought on single star forming events. Earlier studies of Bok globules concentrated on the larger globules (> 10 ), which are relatively massive and may contain multiple sites of star formation. Therefore, in order to examine single low–mass star forming events in an isolated environment, one must turn to the smaller (< 10 ) isolated Bok globules. The largest objects will then have a size of only 0.7 pc at a distance of 500 pc, not much larger than the typical core size found in large dark cloud complexes (∼ 0.3 pc).

1.

The BHR sample of southern globules

Bourke et al. (1995, hereafter BHR) have established a list of 169 small southern molecular clouds of very high opacity and with diameters less than 10 and thus provided a complementary sample to that of Clemens & Barvainis (1988) in the northern sky. 76 of them are associated with IRAS sources while the rest do not have any FIR detections. We have selected the densest and most promising candidates for current star formation and performed a 1200 µm survey with SIMBA. From the FIR-loud sample we have observed 49 sources and detected 76% of them.

Star formation in globules (Chini et al.)

385

From the FIR-quiet sample we did observe so far 28 sources with a detection rate of 30%.

2.

Global properties of FIR-loud globules

In the following we describe some global properties as they follow from the IRAS and our SIMBA measurements.

2.1

Bolometric luminosity

The bolometric luminosity Lbol was calculated by integrating the SED from 12 to 1200 µm; in those cases where an IRAS flux is missing we regard the corresponding luminosity as an upper limit. The distance was assumed to be 500 pc whenever there was no distance determination. In this way, we obtained a luminosity distribution that ranged from 2 to 18 L
with a pronounced maximum around 4 L
. BHR 12, 36, 92 and 23 are exceptions with luminosities of 22, 23, 62 and 1550 L
, respectively.

2.2

Dust temperature

The temperature of the coldest dust carries the bulk of mass in an internally heated volume. Therefore, we have calculated the color temperature between 100 and 1200 µm adopting the mass absorption coefficient κν to be proportional to ν 2 . The derived dust temperatures are 13 K ≤ Td ≤ 24 K with the majority of sources having 15 K ≤ Td ≤ 18 K

2.3

Gas mass

Using the optically thin emission of dust as a tracer for its mass, we convert the observed flux density Sν into a mass by the relation Sν = Mdust × κν × Bν (Tdust )/D2 ,

(1)

where Bν is the Planck function and D the distance to the source. Adopting a gas-to-dust ratio of about 150 we derive a typical gas content of the globules of Mgas ∼ 2 − 4 M
; a few exceptional objects have higher masses between 10 and 125 M
.

2.4

Star formation efficiency

The ratio Lbol /Mgas can be regarded as a measure for the conversion of gas into stars and thus describes the current star formation efficiency. Comparing both quantities with each other leads us to suggest that there are three classes of globules.

386

OPEN ISSUES IN LOCAL STAR FORMATION 4

10

1.3

L~M

BHR23 3

Luminosity [Lo]

10

2

10

BHR92

BHR36 BHR67 1

10

0

10

1

2

10

10

3

10

Gas mass [Mo]

Figure 1. The bolometric luminosity as derived from the SED between 12 and 1200 µm as a function of gas mass as obtained from our SIMBA measurements.

2.4.1 Class A. BHR 23, 36, 67 and 92 form a small group that 1.3 , i.e. their star formation can be characterized by a relation Lbol ∝ Mgas efficiency is highest among all globules. Furthermore, they have in common that they were detected in all four IRAS bands. The number of embedded NIR sources ranges between 10 to 20 and they are typically of Class II - III. The extreme case is BHR 23 which harbors the earliest spectral type (B1V). It should be noted that from a statistical point of view, this sample cannot be considerably enlarged because about 60% of the FIR-loud sample has already been observed; folded with the detection rate only two more objects can be expected here. 2.4.2 Class B. There are 11 globules that form an intermediate 0.8 . In general, they Class which is characterized by a relation Lbol ∝ Mgas are detected at 60 and 100 µm but are missing the shorter IRAS bands. The number of associated NIR sources is below 10 which are typically of Class I - II. 2.4.3 Class C. The remaining 12 globules that have been investigated so far could be detected only at 100 and 1200 µm in most cases. As such, their luminosity is highly uncertain and definitely underestimated. Taking, however, the derived luminosities at their face values, we obtain 0.5 . The number of embedded NIR sources is genera relation Lbol ∝ Mgas ally below 5; their evolutionary stage is between Class 0 and I. The most

387

Star formation in globules (Chini et al.) 2

10

10

0.5

0.5

L~M

BHR71

BHR140

Luminosity [Lo]

Luminosity [Lo]

L~M

2

BHR7 1

10

BHR158

BHR71 1

10

BHR158

BHR110

BHR160

BHR160

BHR107

BHR100

BHR110 BHR107

BHR100

BHR2 BHR3 BHR83

BHR2 BHR3 BHR83

BHR5

BHR5

0

10

BHR140

BHR7

0

0

10

1

10 Gas mass [Mo]

2

10

Figure 2. Same notation as Fig. 1 for the Class B globules; for these objects not all IRAS fluxes are available.

10

0

10

1

10 Gas mass [Mo]

Figure 3. Same notation as Fig. 1 for the Class C globules. In most cases, only 100 µm IRAS fluxes are available; therefore, the values for the luminosity are only upper limits.

prominent example from this group is BHR 72 with an embedded binary protostar.

2.5

MIR appearance

Due to the excellent spatial resolution of TIMMI 2 the 10 µm imaging of embedded IRAS sources resolves their morphology. In the case of BHR 23 the single IRAS source could be resolved into a triple source. Likewise, the increased MIR sensitivity allowed to detect further four previously unknown 10 µm sources, which turns BHR 23 into an aggregate of at least seven MIR sources.

3.

Global properties of FIR-quiet globules

The absence of any further spectral information apart from their emission at 1200 µm makes it difficult to derive physical properties for the FIR-quiet globules. One morphological feature that all of them seem to share is a fragmentation into several emission knots. Adopting a dust temperature of 10 K which is a typical value for externally heated dust clouds the 1200 µm emission may be turned into a gas mass. The range of masses is between 3 and 200 M
depending on the value of κν . With the observed size of the emitting area one obtains hydrogen densities nH ∼ 104 − 106 cm−3 . Using simple arguments to

2

10

388

OPEN ISSUES IN LOCAL STAR FORMATION

estimate the gravitational stability  of the fragments we calculate the Jeans-length by LJ ∼ 1.2 × 1019 T /n. In this way we derive at the conclusion that the radius of the condensations is smaller that (0.1 − 1.0)LJ which means that the fragments are likely to collapse.

4.

Summary

The preliminary results of the present study are still based on low source numbers and therefore, certain values may change during the project. Likewise, the spatial resolution of the IRAS data allow only to discuss global properties while our 10 µm survey will provide information on individual embedded objects. So far, we have reached at the following conclusions: There are three classes of FIR-loud globules with respect to their ratio Lbol /Mgas . α with Class A having We find an empirical relation Lbol ∝ Mgas α ∼ 1.3 (10%), Class B yields α ∼ 0.8 (45%) and Class C α < 0.5 (45%).

The number of embedded YSOs reaches from 10 – 20 in Class A to less than 5 in Class C.

The presence of only a single star within a globule has not yet been observed. The FIR-quiet globules are fragmented into few clumps which seem to be close to gravitational instability. 75% of the FIR-loud and 30% of the FIR-quiet globules do actively form stars.

References Bourke T.L., Hyland A.R., Robinson G., 1995, MNRAS, 276, 1052 Clemens D.P., Barvainis R., 1988, ApJS, 68, 257

SIMBA SURVEY TOWARD HIGH-MASS STAR FORMING REGIONS IN THE SOUTHERN HEMISPHERE Santiago Fa´ undez, Leonardo Bronfman, Guido Garay Departamento de Astronom´ıa, Universidad de Chile, Chile [email protected]

Rolf Chini Astronomisches Institut der Ruhr-Universit¨ at Bochum, Germany

Lars-¨ake Nyman European Southern Observatory, Chile and Onsala Space Observatory, Sweden

Abstract

We summarize the results of a 1.2-mm dust continuum emission survey toward 146 IRAS point-sources, selected from the CS(2-1) survey of Bronfman et al. (1996), thought to be associated with young massive stars. The survey, carried out using the SIMBA Bolometer at the SEST in La Silla, is sensitive to detection of cores of at least 40 solar masses at a typicaldistance of 5 kpc. Regions of ∼ 15 × 10 were mapped with angular resolution of 24” toward each IRAS source, detecting 1.2-mm emission in all observed regions. Half of the maps show the presence of two or more 1.2 mm sources. The 1.2-mm cores associated with the IRAS sources have typically radii of 0.4 pc, masses of 4 × 103 M
, luminosities of 2 × 105 L
and temperatures of 32 K. We have detected, in a handful of regions, 1.2-mm sources with no counterparts in the IRAS and MSX surveys, which we suggest correspond to cold, massive starless cores.

389 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 389-396. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

390

OPEN ISSUES IN LOCAL STAR FORMATION

Introduction High-mass stars (M > 8M
) are important sources of injection of energy into the interstellar medium (ISM) from their birth to their death. The knowledge of the distribution and physical properties of massive stars is thus fundamental to understand the global characteristics of galaxies. Due to their intrinsic scarcity and the short timescales involved in their formation (t ∼ 105 yrs, McKee & Tan 2002), the early evolutionary stages of massive stars have been poorly explored by observations. The formation process of massive mass stars is not yet well established (e.g. Garay & Lizano 1999). There are two main hypothesis to explain the assembling of matter of high-mass stars: by accretion ( Osorio et al. 1999; Yorke & Sonnhalter 2002; McKee & Tan 2002) and by coalescence of low-mass stars (Bonnell et al. (1998)). The recent detection of disks (Shepherd et al. 2001) and jets (Garay et al. 2003) in newly formed massive stars, phenomena that are expected to appear in the early stages of formation in the accretion scenario, seem to favour this hypothesis. However, the scarcity of observed cases makes necessary to increase the observational data concerning the earliest stages of massive star formation. A wealth of molecular line observations of dense cores have been made during the last decade (see Evans 1999 for a review). In particular, CS surveys (Plume et al. 1992; Bronfman et al. 1996) have identified a large sample (about 1000) of high-mass star forming regions. Subsequent studies (e.g. Plume et al. 1997) have shown that the molecular cores associated with massive stars have sizes of ∼ 0.5 pc and masses of about a few thousands solar masses. The properties of dense cores derived from molecular lines are, however, likely to depend on the tracer used. In order to determine the physical properties of cores containing massive stars, independently of molecular line observations, we recently undertook a survey of dust continuum emission toward 146 regions of massive star formation selected from the CS(2-1) survey of Bronfman et al. (1996) of IRAS point-sources with FIR colors of UC HII regions (Wood & Churchwell 1989).

1.

The sample

The observed sources have at least one of the following characteristics. They are either: a) bright in CS (Ta > 2.0 K; 116 sources); b) show the presence of extended wings in the CS(2-1) spectrum (82 objects); and/or c) undetected in the radio continuum survey of Walsh et al. (1998; 21 objects).

SIMBA survey of high-mass star forming regions (Fa´ undez et al.)

1.1

391

Observations

The observations were carried out in three runs, between June 2001 and July 2002, using the SIMBA bolometer array installed at the SEST telescope in La Silla. Observations were made in the fast mapping mode, with azimuthal scans at a velocity of 80  /s, separated in elevation by 8 . The map sizes were 900 × 600 and we made at least two maps of each source. The data reduction was done using the MOPSI software ( Zylka 1998). The beamwidth of the 15m SEST telescope at 250 GHz is 24 . For flux calibrations we observed Uranus and Mars, estimating the uncertainties at the level of 20 %.

2.

Results

We detected 1.2-mm emission toward all the observed regions, and in 46 % of the regions we detected more than one 1.2-mm source. The average number of sources per map is 2.2. In what follows, we first summarize the physical properties of the dust cores associated to the IRAS sources and then discusss the characteristics of massive starless cores found in the neighbourhood of the IRAS sources.

2.1

Massive cores associated with IRAS sources

Sizes. The radius of a core was computed from the geometrical mean of the angular size of the semi-major and semi-minor axis. For distances we used kinematical distances derived from the CS(2-1) velocities reported by Bronfman et al. (1996). In those cases in which the distance ambiguity was unresolved, we adopted the near kinematical distance. We find that the mean radius of the cores is 0.4 ± 0.2 pc. Dust temperatures. The dust temperature was determined from a fit to the spectral energy distribution (SED), constructed using the IRAS fluxes and our 1.2-mm flux. Two modified blackbodies were needed to fit the SED, a cold and a hot component. The dust opacity was assumed to have a power law dependence with frequency, with a typical power law index 2.5 ≤ β ≤ 1.5 (see Evans 1999). The mean temperature of the cold component is 32 ± 5 K. Masses. At 1.2 millimeter the emission is optically thin and hence the dust mass can be computed from the observed flux density Sν . Adopting a single-temperature dust model, the dust mass, Md , can be expressed as Md = Sν D2 /κν Bν (Td ), where κν , Td , Bν (Td ) and D are, respectively, the dust opacity, dust temperature, Planck function and distance. The total mass of a core was computed assuming a dust opacity κ1.2mm of

392

OPEN ISSUES IN LOCAL STAR FORMATION

1.0 cm2 g−1 (Ossenkopf & Henning (1994))and a gas-to-dust mass ratio of 100. We find that the mean mass of the cores is 3.8 × 103 M
.

Luminosities. The bolometric luminosity was computed by integrating the SED. We find that the mean luminosity of the cores is 1.7 × 105 L
, which corresponds to that of a O6.5 star. Densities. Densities were computed assuming that the sources are spherical with the radius derived as above. Assuming a mean mass per particle of µ = 2.29mH , where mH is the massof a hydrogen atom, we find that the mean density is nH2 = 4.3 × 105 cm−3 . Table 1 summarizes the physical parameters of the dust cores associated with high-mass star forming regions. Figure 1 shows the distribution of sizes, temperatures, masses, and luminosities.The values derived from our dust continuum observations are in good agreement with those derived from molecular line observations (Cesaroni et al. (1991); Juvela (1996); Plume et al. (1997)). For instance, using as tracer emission in the CS(5-4) line, Plume et al. (1997) derived that regions of massive star formation associated with H2 O masers have average radius of 0.5 pc and average virial mass of 3.8 × 103 M
.Our results are also in agreement with the results of two recent surveys of dust emission from high-mass star forming regions (Beuther et al. 2002, Mueller et al. 2002). Beuther et al. (2002) observed the 1.2 millimeter emission toward 69 regions and found typical radii of 0.2 pc, masses of 1.5 × 103 M
and dust temperatures of 45 K. Mueller et al. (2002) observed the 350 µm emission toward 51 dense cores associated with water masers and found typical radii of 0.26 pc, masses of 2 × 103 M
and dust temperatures of 29 K. Table 1. Characteristics of dust cores containing young massive stars.a

Size (pc)

Dust temperature (K)

Mass (M
)

Luminosity (L
)

Density cm−3

0.4

32

3.8 × 103

1.7 × 105

4.3 × 105

a Average

2.2

values.

Massive starless cores

The determination of the initial conditions of massive star formation requires the identification of dense, cold and massive pre-stellar objects. The bulk of the radiation from these objects is expected to be emitted

SIMBA survey of high-mass star forming regions (Fa´ undez et al.)

393

Figure 1. Distribution of physical parameters of cores associated with high-mass star forming regions. Upper left: radius distribution. Upper right:temperature distribution. Lower left: mass distribution. Lower right: luminosity distribution.

in the (sub)millimeter range (see fig. 5 of Evans et al. (2002)). In search for massive starless cores we looked in our maps for millimeter sources with no counterparts in the IRAS and MSX surveys, detecting a handful of candidates. The lack of detections in the IRAS bands unables a determination of the dust temperature, but indicates an upper limit of ∼ 20 K. Using this value we derived lower limits for the core masses. The confirmation that these objects are truly massive starless cores requires molecular line observations. Up to date, we have made CS line observations toward three candidates. Wefind that the characteristics of

394

OPEN ISSUES IN LOCAL STAR FORMATION

these cores, summarized in Table 2, are similar to those of cores with embedded massive stars, indicating that they can be genuinely associated with the initial conditions of massive star formation. In particular, their line widths are broad (∼ 4 km s−1 ) which is only characteristic of massive dense cores. Table 2.

Characteristics of massive starless cores.

α(1950)a

Observed δ(1950)a Smm θmm θCS ∆v (Jy) ( ) ( ) (km/s)

13 07 28 -62 27 17 16 17 47 -50 34 15 18 50 47 +01 24 32 a 1.2-mm

3.

2.5 7.5 2.2

37 43 43

37 60 70

3.7 3.7 3.5

Mbdust (M
) 6.7 × 102 1.5 × 103 7.3 × 102

Derived Mvir (M
) 8.6 × 102 1.5 × 103 1.6 × 103

nCS (cm−3 ) 1 × 105 5 × 104 3 × 104

peak position;b Lower limit of mass assuming T = 20 K.

Summary We have made a 1.2-mm dust continuum emission survey toward 146 IRAS sourcesthought to be associated with young massive stars. We find that high-mass stars are typically formed in massive and dense cores with radii of ∼ 0.4 pc, masses of ∼ 3.8 × 103 M
and densities of n ∼ 4 × 105 cm−3 . These parameters are in good agreement with those derived from molecular line observations. The massive dense cores have typically dust temperatures of Td ∼ 32 K, in agreement with that expected from dust cocoons of ∼ 0.4 pc enshrouding UC HII regions (Wolfire & Churchwell 1994).

SIMBA survey of high-mass star forming regions (Fa´ undez et al.)

395

Figure 2. Top: Image of the 1.2-mm SIMBA emission observed towards IRAS 13080-6229 (at the center of the image). Bottom: Image of the MSX emission from roughly the same region. Shown in contours is the 1.2-mm emission. The millimeter source seen toward the northwest has no counterparts in MSX nor in IRAS.

396

OPEN ISSUES IN LOCAL STAR FORMATION

We have found, in our maps, about a dozen of 1.2-mm sources which do not have IRAS nor MSX counterparts. The masses, sizes and densities of these objects,derived from the millimeter observations, are similar to those derived for cores with embedded massive stars. The virial masses, estimated from the CS observations, are in good agreement with the masses estimated from the dust emission. We suggest that these cores will eventually form massive stars, and therefore that their detailed studies will allow to determine the initial conditions for the formation of massive stars.

Acknowledgments The authors wish to thank Dr. M. Nielbok for considerable help in clarifying the use of MOPSI. SF acknowledges support fromFundaci´ on Andes. LB and GG acknowledge support from the Chilean Centro de Astrof´ısica FONDAP 15010003.

References Beuther, H., Schilke, P., Menten, K., Motte, F., Sridhran, T. K. & Wyrowski, F., 2002, ApJ, 566, 945 Bonnell, I., Bate, M. & Zinnecker, H. 1998 MNRAS 298, 93 Bronfman, L., Nymann, L., & May, J. 1996 A & A 115, 81 Cesaroni, R., Walmsley C. M., K¨ompe, C. & Churchwell, E. 1991 A & A 252, 278 Evans, N. J. 1999, ARA&A 37:311 Evans, N. J., Shirley, Y. L., Mueller, K. E. & Knez, C. 2002 Hot Star Workshop III: The Earliest Stages of Massive Star Birth. ASP Conference Proceedings, Vol. 267. Edited by Paul A. Crowther. San Francisco, Astronomical Society of the Pacific, 2002, p.17 Garay, G., & Lizano, S. 1999, PASP, 111, 1049 Garay, G., Brooks, K., Mardones D. & Norris R. 2003 ApJ, 587, 739 Juvela, M. 1996, A&AS 118, 191 McKee, C. & Tan, J., 2002 ApJ 333,333 Mueller, K., Shirley, Y., Evans N. and Jacobson H., 2002 ApJSS, 143, 469 Osorio, M., Lizano S., & D’Alessio P. 1999 ApJ, 525, 808 Ossenkopf, V. & Henning, Th., 1994, A & A, 291, 943 Plume, R., Jaffe, D. T., & Evans, N. J. 1992 ApJSS 78, 505 Plume, R., Jaffe, D. T., Evans, N. J., Mart´ın-Pintado, J. & G´ omez-Gonz´ alez, J. 1997 ApJ 476, 730 Shepherd, D., Claussen, M. & Kurtz, S. 2001 Science, 292, 1513 Walsh, A. J., Burton, M. G., Hyland, A. R. and Robinson, G., 1998, MNRAS, 301, 640 Wolfire, M. G. & Churchwell, E. 1994, ApJ, 427,889 Wood, D. O. S. & Churchwell, E. 1989, ApJ, 340, 265 Yorke, H. & Sonnhalter, C. 2002, ApJ, 569, 849 Zylka, R. 1998, Pocket Cookbook for the MOPSI software

CONFERENCE SUMMARY AND CONCLUSIONS 1.

Open issues

Many unsolved problems were mentioned in the introductory talk by Hans Zinnecker, and pointed out by other participants during the conference. In the concluding session of the conference, Hans was asked to review shortly the open issues. He also summarized the results that he considered as the highlights of the meeting (see Sect. 2). After that, the discussion started and many participants made more comments on the open issues. We next summarize these comments. Hans Zinnecker: There are still many unsolved problems. Are there actually any solved problems? The answer is yes. For instance, observations have shown that OB associations contain not only OB stars but also low-mass stars, i.e., the full IMF is populated (e.g. Preibisch & Zinnecker 1999). Related open issues are the universality of the IMF and the origin of massive stars. Usually the data are not good enough to put universality to a test. We have to plan observations more carefully. Concerning the origin of massive stars: coalescence or accretion - that is the question. Perhaps both, depending on circumstances, e.g. the stellar density. Any theory of the origin of massive stars must address the high incidence of tight massive binaries. Maybe fission works after all! Another question is the origin of Brown Dwarfs. Bo Reipurth showed that the observations are unable to conclude if the BD are the result of fragmentation, or of the ejection from multiple systems. Concerning the PMS evolutionary tracks, there are still some big uncertainties, for example convection theory, or the role of magnetic fields on the stellar surface. The origin of protostellar jets is still an open issue. The binaries are expected to play a role. But the material can also pile up, and this could be the origin of jets. Are the magnetic fields an open issue? People working on turbulence do not like them, but they may be important anyway, certainly for the 397 J. Lépine and J. Gregorio-Hetem (eds.), Open Issues in Local Star Formation, 397-399. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

398

OPEN ISSUES IN LOCAL STAR FORMATION

launch and collimation of jets and outflows and likely for angular momentum transport. A short comment on metallicity: X-ray observations can give us a clue on the metallicity of T Tauri stars. Philipe Andr´e: The origin of the IMF, not universality, is the main issue. Diego Mardones: ALMA will bring a lot of new data. It will become clear that a lot of chemistry is needed to understand what is happening. J´erˆ ome Bouvier: There is a diversity of star formation characteristics in different environments, e.g. Orion versus Taurus associations. Mass functions, angular momentum, binary population all appear to be different in these environments. The origin of these differences must probably be traced back to pre-stellar cores. Nuria Calvet: We need to understand when and how stars form. When → we must investigate the mass reservoir, the infalling envelope, and look for Class 0 and Class I objects. How → this is less known; we must study the accretion and infall processes. We also need to understand how the gas in the disk evolves, and the transition from optically thick to optically thin disks. Ramiro de la Reza: Planets are metal-rich. Post-T Tauri stars are forming planets. Almeida’s poster showed the tendency of the metallicity to increase with age. Important informations about the accretion process and disk evolution can be contained in the variations of metallicity. Thierry Montmerle: The metallicity in dust clouds is related to the magnetic field. We must understand the role of the magnetic field, and the correlation between the small scale and the large scale magnetic fields (this is related to the talk by Gaspard Duchˆene). We need to remember the “big” picture, and not to look only at the small picture. Diego Mardones (again): How does star formation stop, or what stops accretion? Some work on energetics is needed. Bo Reipurth: About the inflow /outflow relation: how is inflow turned into outflow? There are two pictures: i) disk wind models, ii) X-wind models (co-rotation point ejection). Maybe the solution lies in a unification of both (a job for theoreticians).

Conference summary and conclusions

399

Jo˜ ao Alves: Concerning the plea for IMF, we should forget about it for some time and then come back to it in a few years. We should start simple and try to understand chemistry and other stuff. Jonathan Williams: Planet formation is the next issue. Valeri Makarov: The large scale should not be forgotten. For instance, Sco-Crux-XXX-YYY-...How was this structure formed? Reuven Opher: The origin of the first stars (Pop III) is an open issue. Cosmology people do not know about star formation, and stellar formation people do not know about cosmology. A joint effort is needed.

2.

Highlights and other topics.

The following items were the highlights of the conference, according to Hans Zinnecker. The most exciting poster result was that showing an edge-on silhouette disk around a massive protostar in M17 (60 M
disk, 0.1 pc in diameter) by Rolf Chini. About the oral presentations: (i) The role of metallicity was brought up by Thierry Montmerle; (ii) A T Tauri star disk accretion law, with M˙ roughly proportional to the square of the mass of the star, by Nuria Calvet; (iii) The binary star frequency is a function of the primary mass (Michael Sterzik); (iv) The new interpretation for the Gould Belt, by Jacques L´epine; (v) The evidence for a dynamic magnetosphere (inflation) by J´erˆome Bouvier; (vi) Luminosities derived from IR spectroscopy unequal photometric luminosities (Gregory Doppmann); (vii) The use of X-Ray spectroscopy to infer element abundances of YSOs (Beate Stelzer); (viii) The luminosity problem in T Tauri stars; it is too low; FU Ori phases might be a solution (Lee Hartmann); (ix) The evolutionary state of T Tau’s IR companion, coeval or not? (Gaspard Duchˆene); (x) The evolution of multiple systems, with the ejection of a star, can form close binaries, which increases the accretion rate and could explain FU Ori systems (Bo Reipurth); (xi) Observation of non-thermal radial jets, and isolated massive star formation with accretion (Kate Brooks); (xii) New nomenclature “planemos” for planetry mass objects, by Subhanjoy Mohanty. At the end of the session, suggestions were made that the organizers should think about an Ouro Preto II Colloquium on Star Formation.

This page intentionally left blank

Author Index

Abraham, 115 Albrecht, 383 Alencar, 107, 127, 141, 147 Allard, 259 Allen, 375 Almeida, 2 Alves F., 4 Alves J., 47, 199 Amˆ ores, 2 Andersen, 107 Andr´e, 319 Ardila, 4, 251, 259 Armond, 4 Balog, 1 Barbosa, 2, 121 Barbuy, 1 Barrera, 383 Basri, 251, 259 Batalha, 107 Bhatt, 295 Blum, 121 Bottinelli, 367 Bouvier, 147, 235 Boyd, 4 Brice˜ no, 1 Bronfman, 389 Brooks, 331 Cabrit, 213 Calvet, 1, 99 Camperi, 3 Caproni, 115 Castilho, 3 Cerqueira, 4, 303 Chini, 4, 383, 389 Choi, 375 Chuang, 375 Conti, 121 Corradi, 3, 127, 141 Corti, 1 Da Silva, 83, 127 Dal Pino, 303 Damineli, 121 De la Reza, 83, 185 De Marchi, 25

Dias, 1 Dominici, 115 Doppmann, 159 Dougados, 147, 213 Drake, 1 Drew, 169 Duchˆene, 223, 287 Dullemond, 107 Durisen, 245 Dutra, 1 Elitzur, 339 Fa´ undez, 389 Falceta-Gon¸calves, 4 Fernandes, 3 Ferreira, 213 Figuerˆedo, 1, 121 Franco, 3 Garay, 331, 389 Garcia, 213 Ghez, 223 Gonz´ alez, 4 Gregorio-Hetem, 133, 193 Guimar˜ aes, 127, 141 Haikala, 5 Haisch, 3, 251 Harries, 169 Hartmann, 205 Hauschildt, 259 Hensberge, 1 Hetem, 4, 133 Hoffmeister, 1 Hony, 3 Hu´ elamo, 47, 199 Ivezi´ c, 339 Jaffe, 159 Janot-Pacheco, 3 Jayawardhana, 251, 259 Jeffries, 55 Jilinski, 1 K¨ ampgen, 383 K¨ onig, 91 Kaltcheva, 1 Kampgen, 5 Kenyon, 55

401

402 Kraus, 279 Kun, 1 L´ epine, 63, 115 Lada, 47, 199 Lamzin, 3 Li, 375 Lima, 2 M´ enard, 287 Makarov, 2, 33 Mamajek, 39 Manoj, 295 Marciotto, 193 Mardones, 331, 359 Marques, 2 Masciadri, 303 Mathieu, 107 McCabe, 223 Medina-Tanco, 5 Melo, 2, 83, 107 Mendes, 5 Mendez, 127 Miroshnichenko, 3, 339 Mohanty, 251, 259 Moitinho, 47, 199 Montmerle, 73, 193 Moraux, 235 Muzerolle, 99 Nielbock, 383 Nitschelm, 2 Nuernberger, 2 Nyman, 389 Oliveira, 2, 55 Opher, 311

OPEN ISSUES IN LOCAL STAR FORMATION Oudmaijer, 169 Pesenti, 213 Pineda, 5 Pinz´ on, 4, 185 Pogodin, 3 Pomp´eia, 5 Quast, 2, 83, 127 Raga, 303 Reipurth, 269 Rodrigues, 5 Rojas, 3 Rom´ an-Z´ un ˜ iga, 367 Roman-Lopes, 3, 115 Sartori, 2, 63, 133 Schmitt, 177 Seperuelo, 3 Skinner, 2 Stamatellos, 5, 347 Stassun, 3 Stelzer, 177, 199, 251 Sterzik, 83, 245 Teixeira, 2 Tonorio-Tagle, 5 Torres, 83, 127 Van Loon, 55 Vasconcelos, 4 Vaz, 3, 107 Vieira L., 4 Vieira S., 4, 127, 141 Vink, 169 Vinkovi´ c, 339 Vuong, 73 White, 159 Whitworth, 347 Williams, 367