Universitat de - Departament d'Astronomia i Meteorologia ...
Universitat de - Departament d'Astronomia i Meteorologia ... Universitat de - Departament d'Astronomia i Meteorologia ...
UNIVERSITAT DE BARCELONA Departament d’Astronomia i Meteorologia Discovery and study of the microquasar LS 5039 and a search for new microquasars Memòria presentada per Marc Ribó Gomis per optar al grau de Doctor en Ciències Físiques Barcelona, setembre de 2002
- Page 3: Programa de Doctorat d’Astronomia
- Page 7 and 8: Agraïments Agraïments / Agradecim
- Page 9 and 10: al grup, ja ha realitzat aportacion
- Page 11 and 12: March 2000. Thanks to the people th
- Page 13 and 14: Àlex Gómez, Manel, Mercè, Pau, C
- Page 15 and 16: Contents Resum de la tesi: Descobri
- Page 17 and 18: CONTENTS iii 2.8 A proposed scenari
- Page 19 and 20: CONTENTS v 6.3 Results and discussi
- Page 21 and 22: Resum de la tesi: Descobriment i es
- Page 23 and 24: Resum de la tesi ix Be de la seqü
- Page 25 and 26: Resum de la tesi xi instant de temp
- Page 27 and 28: Resum de la tesi xiii D’altra ban
- Page 29 and 30: Resum de la tesi xv concordaven ent
- Page 31 and 32: Resum de la tesi xvii del plasma i
- Page 33 and 34: Resum de la tesi xix ualment es tro
- Page 35 and 36: Resum de la tesi xxi podríem detec
- Page 37 and 38: Resum de la tesi xxiii Al final d
- Page 39 and 40: Resum de la tesi xxv DECLINATION (J
- Page 41 and 42: Resum de la tesi xxvii 1RXS J001442
- Page 43 and 44: Resum de la tesi xxix Apartat 2: 1.
- Page 45 and 46: Bibliografia Casares, J., Ribó, M.
- Page 47 and 48: Chapter 1 Introduction and backgrou
- Page 49 and 50: 1.1. X-ray binaries 3 According to
- Page 51 and 52: 1.1. X-ray binaries 5 Galactic lati
UNIVERSITAT DE BARCELONA<br />
<strong>Departament</strong> d’Astronomia i <strong>Meteorologia</strong><br />
Discovery and study of the<br />
microquasar LS 5039 and a<br />
search for new microquasars<br />
Memòria presentada per<br />
Marc Ribó Gomis<br />
per optar al grau <strong>de</strong><br />
Doctor en Ciències Físiques<br />
Barcelona, setembre <strong>de</strong> 2002
Programa <strong>de</strong> Doctorat d’Astronomia i <strong>Meteorologia</strong><br />
Bienni 1996–1998<br />
Memòria presentada per Marc Ribó Gomis per optar al grau <strong>de</strong><br />
Doctor en Ciències Físiques<br />
Director <strong>de</strong> la tesi<br />
Dr. Josep M. Pare<strong>de</strong>s Poy
Per a la meva neboda Blanca,<br />
que va començar l’aventura <strong>de</strong> viure<br />
tot just al final d’aquesta etapa <strong>de</strong> la meva vida
Agraïments<br />
Agraïments / Agra<strong>de</strong>cimientos / Acknowledgements<br />
Després d’aquesta llarga aventura que és la realització d’una tesi doctoral, hi ha<br />
molta gent a qui vull donar les gràcies. Abans que res, però, <strong>de</strong>mano disculpes per<br />
si em <strong>de</strong>ixo algú que mereixeria ser-hi i no hi és, doncs el cansament acumulat en<br />
aquests moments fa difícil mantenir-se prou <strong>de</strong>spert.<br />
En primer lloc vull donar les gràcies al Josep Maria Pare<strong>de</strong>s, el meu director <strong>de</strong><br />
tesi, no només pel que m’ha ensenyat científicament, sinó també personalment. Vull<br />
agrair epecialment la seva disponibilitat, <strong>de</strong>dicació i seguiment <strong>de</strong> la meva feina,<br />
que ha permès realitzar una gran quantitat <strong>de</strong> treball sense perdre els papers en<br />
cap moment. Sense les seves i<strong>de</strong>es i suggerències aquesta tesi no hagués estat mai<br />
possible.<br />
També vull agrair molt especialment el suport <strong>de</strong>l Josep Martí, que m’ha ensenyat<br />
a analitzar da<strong>de</strong>s ràdio interferomètriques i òptiques. Vull agrair la gran quantitat<br />
<strong>de</strong> preguntes que ha contestat amablement a través <strong>de</strong>l correu electrònic o el telèfon<br />
al llarg d’aquests anys, fos allà on fos, i que sempre han estat molt útils. Gran part<br />
<strong>de</strong> la feina que aquí presento no hagués estat possible sense ell.<br />
I would like to thank very much the support I had from Maria Massi when I was<br />
in Bonn in 1999. I learned to reduce and analyze VLBI data with her, and part<br />
of this thesis has been possible thanks to her useful comments and suggestions. I<br />
would also like to thank her for the very interesting and sometimes truly discussions<br />
that we have mantained during these years. It’s great to see your enthusiasm!<br />
Gràcies a la Marta Peracaula, per l’anglès, per com aclareix i explica <strong>de</strong> manera<br />
planera problemes que <strong>de</strong> vega<strong>de</strong>s semblen irresolubles, especialment pel que fa a
feina no presentada en aquesta tesi. Gràcies per donar-me, encara que potser no en<br />
siguis conscient, paciència i tranquilitat en <strong>de</strong>terminats moments d’aquests darrers<br />
anys. I gràcies també per ensenyar-me moltes coses <strong>de</strong> software i administració <strong>de</strong><br />
màquines.<br />
Muchas gracias a Eduardo Ros, por todo lo que ha hecho por mi tanto a nivel<br />
científico como personal, y especialmente durante mi estancia en Bonn en 1999 y<br />
durante la semana que pasamos analizando datos en Dwingeloo en el año 2000.<br />
Gracias por enseñarme tantas cosas <strong>de</strong> VLBI, por tus sabios consejos y por tu<br />
pulcritud en el trabajo que hemos realizado juntos. Gracias también a Ana, tu<br />
mujer, por el apoyo recibido durante mi estancia en Bonn.<br />
Moltes, moltes gràcies a l’Octavi Fors, per ensenyar-me tantes coses innovadores<br />
<strong>de</strong> software astronòmic, tantes eines, i sobretot astrometria òptica. Estic molt con-<br />
tent d’haver pogut treballar amb tu. A banda d’això has estat un company <strong>de</strong><br />
<strong>de</strong>spatx d’aquells que no voldries perdre mai.<br />
Moltes gràcies al Xavier Otazu. En primer lloc perquè ell va ser el primer que<br />
em va donar la i<strong>de</strong>a <strong>de</strong> realitzar una tesi doctoral. En segon lloc, per ser tan crític<br />
i escèptic en molts <strong>de</strong>ls temes <strong>de</strong> que hem parlat durant aquests anys, i que han fet<br />
que m’hagués <strong>de</strong> replantejar moltes coses. D’altra banda, tenir un article amb tu, i<br />
sobre wavelets, és alguna cosa més que un simple article, <strong>de</strong>sprés <strong>de</strong> tants anys <strong>de</strong><br />
pensar en possibles articles que mai van tirar endavant per falta <strong>de</strong> temps.<br />
Gràcies també al Joan García, pels seus consells tan útils quan m’he hagut<br />
d’enfrontar amb càlculs d’òrbites. També li vull agrair les diverses vega<strong>de</strong>s que ha<br />
revisat l’anglès d’alguns <strong>de</strong>ls meus escrits, i la manera tan acurada com ho ha fet.<br />
Gràcies al Robert Estalella, per ensenyar-me rigurositat i pulcritud a l’hora d’a-<br />
cabar un article, pels consells d’anglès, d’AIPS i pel seu saber fer.<br />
Moltes gràcies a l’Ignasi Ribas, per la gran quantitat <strong>de</strong> preguntes <strong>de</strong> software,<br />
anglès i especialment d’astrofísica estel·lar, que m’ha contestat en aquests darrers<br />
anys. Algú com tu li dóna un valor afegit a un <strong>de</strong>partament.<br />
Gràcies a l’Eduard Masana, pel latex, pel Java, pel Netscape, pels scripts, pel<br />
linux i, com no, pel masanix. Gràcies per la gran quantitat <strong>de</strong> problemes solucionats.<br />
Moltes gràcies al Valentí Bosch, que encara que fa poc temps que s’ha incorporat
al grup, ja ha realitzat aportacions importants que m’han ajudat en alguns aspectes<br />
d’aquesta tesi.<br />
Gràcies al Pau Reig, per contestar amb <strong>de</strong>tall moltes preguntes sobre anàlisi<br />
temporal i espectral <strong>de</strong> da<strong>de</strong>s <strong>de</strong> satèl·lits <strong>de</strong> raigs X, i per estar a l’altra banda <strong>de</strong>l<br />
correu electrònic o <strong>de</strong>l telèfon quan ha calgut.<br />
Gracias a Ignacio Negueruela, por estar varias veces al otro lado <strong>de</strong>l teléfono o<br />
<strong>de</strong>l correo electrónico cuando me han surgido dudas al respecto <strong>de</strong> binarias <strong>de</strong> rayos<br />
X masivas, y por compartir información al respecto <strong>de</strong> LS 5039.<br />
Moltes gràcies al Jorge Casares. Ha estat un plaer, i espero que ho segueixi sent,<br />
treballar amb tu. Recordo amb alegria molts moments passats a París durant la<br />
nostra setmana <strong>de</strong> feina intensiva en el projecte MINE durant la primavera <strong>de</strong> 2002,<br />
així com les interessants explicacions didàctiques sobre espectroscopia òptica tant a<br />
Granada com a Canàries.<br />
Gràcies al Blai Sanahuja, per les potser escasses però interessants i <strong>de</strong> vega<strong>de</strong>s<br />
llargues converses sobre política científica, investigació i “dinàmica <strong>de</strong> grups”, que<br />
han fet que pogués assimilar alguns <strong>de</strong>ls mals tràngols passats durant aquests anys.<br />
I sí, suposo que encara segueixo sent un d’aquest “joves impetuosos”.<br />
També gràcies a la Cesca Figueras, per atendre’m quan he tingut dubtes sobre<br />
estrelles joves i fotometria estel·lar, i en especial per la seva acurada <strong>de</strong>dicació pel<br />
que fa a la disponibilitat <strong>de</strong> revistes electròniques.<br />
Encara que ja no sigui al <strong>de</strong>partament, vull donar les gràcies al Jaume Soley, el<br />
tècnic informàtic que hem tingut durant molts anys. Gràcies no només per solucionar<br />
els problemes, sinó també per les explicacions, per la seva capacitat <strong>de</strong> resposta i<br />
pel seu entusiasme quan hi havia problemes.<br />
També vull donar les gràcies al JR, que sempre té a punt tota la paperassa<br />
necessària, i amb l’ajuda <strong>de</strong>l qual moltes coses semblen senzilles quan realment no<br />
ho són. Gràcies especialment per la cura amb que ha dut la paperassa necessària<br />
per dipositar aquesta tesi.<br />
També, com no, vull donar les gràcies a l’ Óscar Morata i l’Inma Sepúlveda que,<br />
a part <strong>de</strong> proporcionar-me els estils per escriure aquesta tesi, sempre han estat a<br />
prop, encara que fos un cop l’any, per tenir converses i entendre moltes coses, per
<strong>de</strong>sfogar-nos conjuntament, per la seva comprensió i recolzament en molts aspectes<br />
durant aquests anys, i per algunes nits <strong>de</strong> festa en el món exterior.<br />
Gràcies també al Josep Miquel Girart, per moltes i diferents qüestions que m’ha<br />
solucionat, i per algunes converses interessants que hem tingut en aquests darrers<br />
dos anys.<br />
I encara que ja fa molt temps que va marxar, també vull donar les gràcies a la<br />
Maite, per mantenir la seva frescura i alegria durant els durs primers dos anys <strong>de</strong><br />
tesi, i per enviar notícies <strong>de</strong> tant en tant.<br />
Vull donar les gràcies a tots els incondicionals <strong>de</strong> l’hora <strong>de</strong> dinar durant aquests<br />
anys, per les converses i <strong>de</strong>bats, i per compartir amb mi algunes <strong>de</strong> les penes i<br />
alegries que hem passat durant aquest temps: Maite, David, Teresa, Xavi, Octavi,<br />
Marta, Ricard, Eduard, Albert Domingo, Josep Miquel, Pep, Andreu Raig, Ada, i<br />
en algunes ocasions, especialment en aquest darrer i dur estiu, també l’ Àngels.<br />
Gràcies també a tota la gent <strong>de</strong>l <strong>Departament</strong> d’Astronomia i <strong>Meteorologia</strong> que,<br />
d’alguna manera o altra, ha fet més fàcil la meva feina durant aquests anys.<br />
Although he will never read this, I would like to thank Jan van Paradijs, for<br />
pushing the field of X-ray binaries, for the incredible work he did when compiling<br />
the catalog of X-ray binaries, for his friendly character, and for offering me to join<br />
his group on gamma-ray bursts in 1997.<br />
Thanks to Titus Galama, for welcoming me the several times I have been in<br />
Amsterdam, and for his useful and polite comments. Thanks also for contacting<br />
and offering me to work on GRBs.<br />
Gracias también a Mariano Mén<strong>de</strong>z, por acogerme tan bien durante mis visitas<br />
a Amsterdam, por los proyectos <strong>de</strong> colaboración que al final no cristalizaron, y por<br />
su entusiasmo y simpatía.<br />
Thanks to many people of the VLBI group at Bonn, specially to Giuseppe Cimò,<br />
Mauro Sorgente, Andrea Tarchi, Andrei Lobanov, Yoshiaki Hagiwara, Alan Roy,<br />
Richard Porcas, Alok Patnaik, Walter Alef and also Anton Zensus, for allowing me<br />
to join the group during 5 months in 1999.<br />
Thanks to Simon Garrington for his help during my visit to Jodrell Bank in
March 2000. Thanks to the people that were at JIVE helping me in November<br />
2000, specially to Denise Gabuzda, Cormac Reynolds, Mike Garrett, Ian Avruch<br />
(Max), Bob Campbell, Leonid Gurvits and Huib van Langevel<strong>de</strong>.<br />
Thanks to Ginny McSwain, for sharing interesting information on LS 5039 while<br />
I was finishing my thesis.<br />
Muchas gracias a Félix Mirabel, por pensar en mi para un postdoc y por ofre-<br />
cerme la posibilidad <strong>de</strong> trabajar en un proyecto tan interesante como MINE. Siempre<br />
con i<strong>de</strong>as nuevas, siempre con la cabeza en marcha, consiguiendo así ilusionar con sus<br />
interesantes propuestas. Gracias también por como nos acogió en su casa durante<br />
la dura semana <strong>de</strong>l proyecto MINE en París.<br />
Un abrazo y muchas a gracias a los colaboradores argentinos Gustavo Romero<br />
y Paula Benaglia. Ha sido muy enriquecedor y un placer trabajar con vosotros, y<br />
espero que siga siéndolo.<br />
Como, no, un fuerte abrazo a Jorge Combi, por sus continuos ánimos en esta<br />
etapa final <strong>de</strong> mi tesis, y por los buenos momentos pasados durante esa dura semana<br />
en París. Gracias por las risas y las charlas.<br />
Gracias a los “VLBIeros” con los que he coincidido en varios congresos durante<br />
estos años, que tan bien me han acogido y con los que he compartido tantos mo-<br />
mentos simpáticos y divertidos, que le hacían a uno olvidar que se encontraba en<br />
cualquier lugar <strong>de</strong>l mundo, y en especial a Eduardo Ros, José Luís Gómez, Iván<br />
Agudo, Miguel<br />
Ángel Pérez Torres, José Carlos Guirado, Lucas Lara, María José<br />
Rioja, Antxón Alberdi y Jon Marcai<strong>de</strong>.<br />
Thanks to the “VLBIers” Zsolt Paragi and Michele Pestalozzi, for making me<br />
laugh and enjoy some of the conferences during these years.<br />
Since this thesis is based on 7 published papers, I would like to thank all the<br />
people who have ma<strong>de</strong> useful comments or suggestions while writing them, in partic-<br />
ular: O. Fors, X. Otazu, I. Ribas, D. Fernán<strong>de</strong>z, E. Masana, F. Figueras, J. Colomé.<br />
I would also like to thank all those who have read draft versions and ma<strong>de</strong> useful<br />
suggestions: M. Peracaula, J. García-Sánchez, A. Lobanov, R. Porcas, W. Alef, and<br />
J. Bloom. Finally, I would also like to thank the referees for their comments: R.<br />
Fen<strong>de</strong>r, H. Falcke, L. F. Rodríguez and F. Colin.
From the scientifical point of view, I would finally like to thank Rob Fen<strong>de</strong>r, for<br />
the interesting and stimulating discussions and informal conversations mantained in<br />
conferences during the early years of my PhD, that gave me lots of interesting i<strong>de</strong>as.<br />
Thanks also for pushing the field of radio jet X-ray binaries (a.k.a. microquasars).<br />
Gràcies als incondicionals d’ASTER en les sorti<strong>de</strong>s observacionals <strong>de</strong> fa uns anys,<br />
per les bones estones passa<strong>de</strong>s i per tot el que m’heu ensenyat: Jordi A., Montserrat<br />
G., Jordi B., David, Montse P. i Xavi.<br />
Gràcies a la Rosa Maria Ros, per iniciar-me en l’astronomia en les seves classes<br />
d’EATP <strong>de</strong> l’institut.<br />
També vull donar les gràcies a totes aquelles persones que algun cop m’han<br />
preguntat sobre astronomia i han escoltat amb interès, perquè fan necessària i útil<br />
la nostra tasca d’investigadors.<br />
Agraeixo les dues beques que m’ha concedit la CIRIT (Generalitat <strong>de</strong> Catalunya,<br />
referències 1998 BEAI 200293 i 1999 FI 00199).<br />
Des <strong>de</strong>l punt <strong>de</strong> vista personal, moltíssimes gràcies a tota la colla <strong>de</strong>l Balmes,<br />
pels sopars, caps <strong>de</strong> setmana, paelles, festes, tertúlies i un llarg i molt llarg etcètera,<br />
i en especial pel suport i els ànims <strong>de</strong>ls darrers mesos: Toni, Bego, Núria, Iolanda,<br />
Mar, Leti,<br />
Àlex, Joan, Jordi, Raquel, Raül, Laura, Jaume, etc... D’entre ells, vull<br />
agrair especialment el suport <strong>de</strong>l Toni i la Bego en els moments en que calia ser allà.<br />
També vull recordar aquí que el Marc Llimós, que mai llegirà aquests agraïments,<br />
va ensenyar-me que calia lluitar per aconseguir aquelles coses que valien la pena.<br />
Gracias también a Simone y Roman. Lo siento, pero todavía no os puedo escribir<br />
en alemán! Muchas gracias por vuestros continuos ánimos (hace dos minutos que<br />
acabo <strong>de</strong> recibir un mail <strong>de</strong> Roman animándome en la etapa final).<br />
Com no, moltes gràcies al Jordi Alborch, perquè va fer amenes i/o suportables<br />
certes etapes <strong>de</strong> les meves esta<strong>de</strong>s a Bonn i al JIVE. Per les festes <strong>de</strong> Brusel·les, Sant<br />
Feliu, Begur, Barcelona i Madrid, entre d’altres. Per l’alegria i per l’entusiasme, per<br />
la capacitat <strong>de</strong> reacció i diversió, que han conseguit evadir-me en molts casos en què<br />
ho necessitava.<br />
Gràcies també a la gent <strong>de</strong> la “quinta Alborch”, com diuen alguns: Àlex Sánchez,
Àlex Gómez, Manel, Mercè, Pau, Cuca, Mario i Céline. Per les sorti<strong>de</strong>s <strong>de</strong> cap <strong>de</strong><br />
setmana o <strong>de</strong> festa i per les vega<strong>de</strong>s que m’heu fet oblidar els problemes <strong>de</strong> la feina.<br />
Gràcies al Francesc, al Pablo i en especial al Pep, els nois <strong>de</strong> l’Espai (s’Espai), per<br />
l’espontaneïtat, improvització, performances i altres diversions que han aconseguit<br />
fer més fàcils i agradables certs moments durs <strong>de</strong> la meva vida, especialment durant<br />
els primers anys <strong>de</strong> tesi.<br />
I aquí arriba l’agraïment als companys <strong>de</strong> carrera <strong>de</strong>l sopar mensual, per les di-<br />
vagacions, extravagàncies, “puñaladas traperas” que han fet que ens <strong>de</strong>scargoléssim<br />
<strong>de</strong> riure, tertúlies interessants, etc: David Nofre, David Pino, Toni Hernán<strong>de</strong>z, Toni<br />
Magariños (Pato) i Oriol Saladrigas. Moltes gràcies en especial al David Nofre per<br />
les llargues converses que hem mantingut durant aquests anys.<br />
També vull agrair a l’Imma Bastida els ànims que m’ha donat els darrers mesos.<br />
Gracias a Antonio Vallecillos, por los continuos ánimos y por compartir conmigo<br />
mi <strong>de</strong>sahogo <strong>de</strong>portivo (y también psíquico) semanal.<br />
A tota la gent <strong>de</strong>l nucli dur <strong>de</strong> D-Recerca, per la feina feta, la il·lusió, les ganes,<br />
i per intentar canviar allò que creiem que no és just: Teresa, Cesca, Elisenda, Ana,<br />
Xavi <strong>de</strong> Pedro, Eduard Serra, Hèctor, Sílvia, Miquel, María i un llarg etcètera,<br />
d’entre els quals vull afegir l’Italo, per les interessants converses sobre política en el<br />
darrer any.<br />
A toda la gente <strong>de</strong>l núcleo duro <strong>de</strong> Precarios, por <strong>de</strong>sahogarnos juntos, por<br />
llorar esta política científica que tenemos y por intentar hacer algo por cambiarla, y<br />
también, como no, por lo bien que nos lo hemos pasado juntos en varias ocasiones:<br />
Pastora, Teresa, Ana, Roberto, César, Cristina, Toni, Marta, Xavi <strong>de</strong> Pedro, Pablo,<br />
y un largo etcétera.<br />
Gracias a Jacinta, por estar ahí cada semana durante los últimos 25 años, por las<br />
charlas <strong>de</strong> los lunes por la mañana y por los fantásticos platos tantas veces cocinados.<br />
Vull agrair als meus pares, Ramon i Elena, l’educació que m’han donat, el suport<br />
econòmic durant tots aquests anys <strong>de</strong> sou precari, els ànims i el suport moral en els<br />
moments més durs. No us podré agrair mai prou l’oportunitat que m’heu donat <strong>de</strong><br />
realitzar una tesi doctoral. Gràcies.
Gràcies a la meva germana Nàdia, pels anys que hem viscut junts i per ser<br />
una excel·lent companya <strong>de</strong> pis durant els primers anys <strong>de</strong> tesi. De vega<strong>de</strong>s encara<br />
trobo a faltar aquelles xerra<strong>de</strong>s <strong>de</strong> diumenge explicant-nos la setmana. En fi, va ser<br />
important tenir-te al meu costat durant la difícil situació personal/professional <strong>de</strong>ls<br />
primers dos anys <strong>de</strong> tesi sense beca. Gràcies també al Pau pels ànims i el suport.<br />
I finalment, l’agraïment més important. Gràcies Noemí. Gràcies per cuidar-<br />
me, per animar-me, per escoltar-me en els moments en que més ho necessito, per<br />
recolzar-me durant aquest llarg i dur estiu, per ser com ets i per estar sempre al<br />
meu costat.
Contents<br />
Resum <strong>de</strong> la tesi:<br />
Descobriment i estudi <strong>de</strong>l microquàsar LS 5039 i<br />
una cerca <strong>de</strong> nous microquàsars vii<br />
1 Introduction and background 1<br />
1.1 X-ray binaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1<br />
1.1.1 High mass X-ray binaries . . . . . . . . . . . . . . . . . . . . . 3<br />
1.1.2 Low mass X-ray binaries . . . . . . . . . . . . . . . . . . . . . 6<br />
1.1.3 Radio emitting X-ray binaries . . . . . . . . . . . . . . . . . . 8<br />
1.2 Microquasars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10<br />
1.2.1 Projection and special relativity effects . . . . . . . . . . . . . 12<br />
1.2.2 Quasars and microquasars . . . . . . . . . . . . . . . . . . . . 16<br />
1.2.3 Known microquasars . . . . . . . . . . . . . . . . . . . . . . . 17<br />
1.2.4 Accretion disk and jet formation . . . . . . . . . . . . . . . . . 21<br />
1.2.5 Black hole states and radio emission . . . . . . . . . . . . . . 22<br />
1.3 Motivation of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . 24<br />
i
ii CONTENTS<br />
I Discovery and study of the microquasar LS 5039 31<br />
2 Multiwavelength approach to LS 5039 33<br />
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33<br />
2.2 Chronology of findings on LS 5039 . . . . . . . . . . . . . . . . . . . 34<br />
2.3 An interesting target in the search for new microquasars . . . . . . . 35<br />
2.4 The radio counterpart: NVSS J182614−145054 . . . . . . . . . . . . 36<br />
2.4.1 A radio counterpart in the NVSS . . . . . . . . . . . . . . . . 36<br />
2.4.2 VLA observations, discovery of a REXB . . . . . . . . . . . . 37<br />
2.4.3 Long-term GBI monitoring, a persistent radio source . . . . . 42<br />
2.4.4 VLBA observations, discovery of a microquasar . . . . . . . . 46<br />
2.4.5 EVN and MERLIN observations,<br />
confirmation of persistent relativistic radio jets . . . . . . . . . 50<br />
2.5 The optical/IR star: LS 5039 . . . . . . . . . . . . . . . . . . . . . . 55<br />
2.5.1 Photometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55<br />
2.5.2 Spectral type . . . . . . . . . . . . . . . . . . . . . . . . . . . 58<br />
2.5.3 Distance and its uncertainty . . . . . . . . . . . . . . . . . . . 59<br />
2.5.4 Radial velocity curve . . . . . . . . . . . . . . . . . . . . . . . 60<br />
2.6 The X-ray counterpart: RX J1826.2−1450 . . . . . . . . . . . . . . . 65<br />
2.6.1 The ASM/RXTE data . . . . . . . . . . . . . . . . . . . . . . 65<br />
2.6.2 PCA/RXTE observations . . . . . . . . . . . . . . . . . . . . 67<br />
2.6.3 BeppoSAX observations . . . . . . . . . . . . . . . . . . . . . 73<br />
2.7 The γ-ray counterpart: 3EG J1824−1514 . . . . . . . . . . . . . . . . 74
CONTENTS iii<br />
2.8 A proposed scenario to explain the multiwavelength behavior . . . . . 79<br />
2.8.1 Spectral energy distribution . . . . . . . . . . . . . . . . . . . 79<br />
2.8.2 A mo<strong>de</strong>l based on the γ-ray/radio emission . . . . . . . . . . . 80<br />
2.8.3 Energetic consi<strong>de</strong>rations . . . . . . . . . . . . . . . . . . . . . 84<br />
2.8.4 Future prospects . . . . . . . . . . . . . . . . . . . . . . . . . 85<br />
2.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86<br />
3 LS 5039 as a runaway microquasar 93<br />
3.1 Positions and proper motions . . . . . . . . . . . . . . . . . . . . . . 94<br />
3.1.1 Optical positions . . . . . . . . . . . . . . . . . . . . . . . . . 94<br />
3.1.2 Radio positions . . . . . . . . . . . . . . . . . . . . . . . . . . 96<br />
3.1.3 Proper motions . . . . . . . . . . . . . . . . . . . . . . . . . . 96<br />
3.2 Ejection from the galactic plane . . . . . . . . . . . . . . . . . . . . . 98<br />
3.3 The past trajectory of LS 5039 . . . . . . . . . . . . . . . . . . . . . 99<br />
3.4 SNR G016.8−01.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100<br />
3.4.1 A lower limit of the distance . . . . . . . . . . . . . . . . . . . 103<br />
3.4.2 Particle <strong>de</strong>nsity estimates . . . . . . . . . . . . . . . . . . . . 104<br />
3.5 The H i surroundings of LS 5039 . . . . . . . . . . . . . . . . . . . . . 106<br />
3.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108<br />
3.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
iv CONTENTS<br />
II A search for new microquasars 119<br />
4 The cross-i<strong>de</strong>ntification method 121<br />
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121<br />
4.2 Cross-i<strong>de</strong>ntification between RBSC and NVSS catalogs . . . . . . . . 121<br />
4.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128<br />
5 Radio and optical observations 131<br />
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131<br />
5.2 Radio observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131<br />
5.2.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132<br />
5.3 Optical observations . . . . . . . . . . . . . . . . . . . . . . . . . . . 135<br />
5.3.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138<br />
5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139<br />
5.4.1 Discussion on individual sources . . . . . . . . . . . . . . . . . 142<br />
5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145<br />
6 EVN and MERLIN observations 149<br />
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149<br />
6.2 Observations and data reduction . . . . . . . . . . . . . . . . . . . . . 149<br />
6.2.1 MERLIN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150<br />
6.2.2 EVN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151<br />
6.2.3 Combining EVN and MERLIN . . . . . . . . . . . . . . . . . 153<br />
6.2.4 Flux <strong>de</strong>nsity measurements at the 100 m antenna in Effelsberg 153
CONTENTS v<br />
6.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 154<br />
6.3.1 1RXS J001442.2+580201 and its two-si<strong>de</strong>d jet . . . . . . . . . 155<br />
6.3.2 1RXS J013106.4+612035 and its one-si<strong>de</strong>d jet . . . . . . . . . 158<br />
6.3.3 1RXS J042201.0+485610, a non-<strong>de</strong>tected source . . . . . . . . 159<br />
6.3.4 1RXS J062148.1+174736, a compact source . . . . . . . . . . 159<br />
6.3.5 1RXS J072259.5−073131 and its bent one-si<strong>de</strong>d jet . . . . . . 160<br />
6.3.6 1RXS J072418.3−071508, a quasar with a bent one-si<strong>de</strong>d jet . 162<br />
6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163<br />
7 Summary 167<br />
III General conclusions 169
vi CONTENTS
Resum <strong>de</strong> la tesi:<br />
Descobriment i estudi <strong>de</strong>l<br />
microquàsar LS 5039 i<br />
una cerca <strong>de</strong> nous microquàsars<br />
1. Introducció i antece<strong>de</strong>nts sobre el tema<br />
Els microquàsars són estrelles binàries <strong>de</strong> raigs X que mostren jets relativistes en<br />
longituds d’ona ràdio. En aquesta secció farem una breu introduccio a les estrelles<br />
binàries <strong>de</strong> raigs X, per <strong>de</strong>sprés passar a comentar algunes <strong>de</strong> les propietats <strong>de</strong>ls<br />
microquàsars, i concloure amb l’explicació <strong>de</strong> quina va ser la motivació d’aquesta<br />
tesi.<br />
1.1 Estrelles binàries <strong>de</strong> raigs X<br />
Les estrelles binàries <strong>de</strong> raigs X són sistemes binaris que contenen un objecte com-<br />
pacte, tant pot ser un forat negre com una estrella <strong>de</strong> neutrons, que acreta matèria<br />
<strong>de</strong> l’estrella companya.<br />
Per tal d’explicar l’emissió en raigs X <strong>de</strong>tectada en aquests sistemes po<strong>de</strong>m fer<br />
el següent raonament. La matèria que és acretada per l’objecte compacte és accel-<br />
erada a velocitats relativistes en transformar la seva energia potencial gravitatòria,<br />
<strong>de</strong>guda a l’intens camp gravitatori <strong>de</strong> l’objecte compacte, en energia cinètica. D’al-<br />
tra banda, el ritme d’acreció <strong>de</strong> matèria té un valor màxim teòric, anomenat límit<br />
vii
viii Resum <strong>de</strong> la tesi<br />
d’Eddington, que té lloc quan la pressió <strong>de</strong> radiació contraresta la força d’atracció<br />
gravitatòria. Com que la matèria que és acretada té moment angular, habitualment<br />
es forma un disc d’acreció al voltant <strong>de</strong> l’objecte compacte. La matèria <strong>de</strong>l disc<br />
perd moment angular per dissipació viscosa, la qual cosa produeix un escalfament<br />
<strong>de</strong>l disc, i cau vers l’objecte compacte en una trajectòria espiral. La temperatura<br />
<strong>de</strong> cos negre <strong>de</strong> la darrera òrbita estable, en el cas d’un forat negre acretant matèria<br />
en el límit d’Eddington, ve donada per una <strong>de</strong>pendència amb la massa <strong>de</strong> l’objecte<br />
compacte. Si el forat negre té unes poques masses solars la temperatura obtinguda<br />
és <strong>de</strong> l’ordre <strong>de</strong> 10 7 K. A aquesta temperatura l’energia és radiada en el domini <strong>de</strong><br />
raigs X <strong>de</strong> l’espectre electromagnètic. D’altra banda, po<strong>de</strong>m calcular la lluminositat<br />
d’acreció com l’energia cinètica <strong>de</strong> les partícules en arribar al radi gravitacional o<br />
<strong>de</strong> Schwarzschild. Així, po<strong>de</strong>m explicar les lluminositats típiques <strong>de</strong> les estrelles<br />
binàries <strong>de</strong> raigs X, que són <strong>de</strong> l’ordre <strong>de</strong> 10 37 erg s −1 , amb ritmes d’acreció <strong>de</strong><br />
massa <strong>de</strong> l’ordre <strong>de</strong> 10 −9 M⊙ any −1 . Aquesta matèria prové <strong>de</strong> l’estrella companya<br />
a l’objecte compacte. Si com a objecte compacte tenim una estrella <strong>de</strong> neutrons en<br />
comptes d’un forat negre, els ritmes d’acreció <strong>de</strong> massa necessaris són lleugerament<br />
superiors. En qualsevol cas, po<strong>de</strong>m explicar la lluminositat en raigs X <strong>de</strong> les binàries<br />
<strong>de</strong> raigs X amb ritmes d’acreció <strong>de</strong> massa superiors a 10 −10 M⊙ yr −1 .<br />
Si l’objecte compacte és una estrella <strong>de</strong> neutrons amb un camp magnètic intens,<br />
<strong>de</strong> l’ordre <strong>de</strong> 10 12 G, el disc d’acreció quedarà truncat a alguns milers <strong>de</strong> quilòmetres<br />
<strong>de</strong> l’objecte compacte, ja que la matèria seguirà les línies <strong>de</strong> camp magnètic, i<br />
acabarà impactant en els pols magnètics, produint altre cop emissió en raigs X. Si<br />
l’eix magnètic està inclinat respecte l’eix <strong>de</strong> rotació, veurem pulsacions <strong>de</strong> raigs X<br />
si l’emissió en forma <strong>de</strong> feix <strong>de</strong>ls pols magnètics escombra la nostra visual. Aquest<br />
tipus particular <strong>de</strong> binàries <strong>de</strong> raigs X s’anomenen púlsars <strong>de</strong> raigs X.<br />
Per contra, si l’objecte compacte és una estrella <strong>de</strong> neutrons amb un camp<br />
magnètic feble, aproximadament inferior a 10 10 G, el disc d’acreció pot arribar a<br />
impactar l’objecte compacte. En aquest cas els raigs X seran produïts en la part<br />
interna <strong>de</strong>l disc d’acreció i en la capa límit entre el disc d’acreció i l’objecte compacte.<br />
El nombre <strong>de</strong> binàries <strong>de</strong> raigs X conegu<strong>de</strong>s actualment és <strong>de</strong> 280 aproximada-<br />
ment. Depenent <strong>de</strong> la massa <strong>de</strong> l’estrella companya a l’objecte compacte, se solen<br />
dividir en binàries <strong>de</strong> raigs X <strong>de</strong> massa alta i <strong>de</strong> massa baixa.<br />
En les <strong>de</strong> massa alta, <strong>de</strong> les quals se’n coneixen unes 131, la companya és una<br />
estrella jove <strong>de</strong> tipus espectral O o B, i típicament se’n troben <strong>de</strong> dos tipus: estrelles
Resum <strong>de</strong> la tesi ix<br />
Be <strong>de</strong> la seqüència principal o gegants, i estrelles O o B supergegants. Les primeres<br />
tenen masses d’entre 8 i 20 M⊙, són objectes amb una elevada rotació, i tenen un disc<br />
<strong>de</strong> <strong>de</strong>creció equatorial, que proporciona la matèria a ser acretada per part <strong>de</strong> l’objecte<br />
compacte. Típicament aquest és una estrella <strong>de</strong> neutrons amb un camp magnètic<br />
intens, i aquests sistemes solen ser púlsars <strong>de</strong> raigs X. En el cas <strong>de</strong> les supergegants,<br />
<strong>de</strong> més <strong>de</strong> 15 M⊙, aquestes tenen un vent estel·lar intens, que proporciona la matèria<br />
a ser acretada. En alguns casos també s’omple l’anomenat lòbul <strong>de</strong> Roche, amb la<br />
qual cosa la transferència <strong>de</strong> matèria és encara més eficient. Com que en ambdós<br />
casos es tracta d’estrelles joves, aquests sistemes se situen principalment en el disc<br />
galàctic, ja que tracen la població estel.lar <strong>de</strong> tipus I. És important <strong>de</strong>stacar que el<br />
57% <strong>de</strong> les binàries <strong>de</strong> raigs X <strong>de</strong> massa alta a la Galàxia són púlsars <strong>de</strong> raigs X.<br />
En les <strong>de</strong> massa baixa, <strong>de</strong> les quals se’n coneixen unes 149, l’estrella companya<br />
té un tipus espectral posterior a B, i típicament tenen masses inferiors a 2 M⊙. En<br />
aquest cas la transferència té lloc per <strong>de</strong>sbordament <strong>de</strong>l lòbul <strong>de</strong> Roche. La majoria<br />
són fonts transitòries <strong>de</strong> raigs X. Algunes d’elles es classifiquen com a fonts Z (6<br />
sistemes) i fonts Atoló (18 sistemes), <strong>de</strong>penent <strong>de</strong>l patró traçat en diagrames color-<br />
color <strong>de</strong> raigs X. Es creu que les primeres contenen estrelles <strong>de</strong> neutrons amb camps<br />
magnètics <strong>de</strong> l’ordre <strong>de</strong> 10 10 G i ritmes d’acreció propers al límit d’Eddington,<br />
mentre que les segones tenen camps magnètics febles inferiors a 10 8 G i ritmes<br />
d’acreció d’entre una dècima i una centèssima <strong>de</strong>l límit d’Eddington. Les binàries<br />
<strong>de</strong> raigs X <strong>de</strong> massa baixa se situen en el bulb i en l’halo <strong>de</strong> la Galàxia, ja que són<br />
objectes vells <strong>de</strong> Població II. De fet, 13 sistemes es troben en cúmuls globulars.<br />
D’altra banda, aproximadament 43 binàries <strong>de</strong> raigs X també mostren emissió<br />
en la banda ràdio <strong>de</strong> l’espectre electromagnètic, amb <strong>de</strong>nsitats <strong>de</strong> flux superiors<br />
a 0.1–1 mJy. A més, aquesta emissió és produïda en escales angulars petites, <strong>de</strong><br />
manera que no pot tenir un origen tèrmic. En aquest contexte, el mecanisme més<br />
eficient per produir l’emissió ràdio observada és el mecanisme d’emissió sincrotró,<br />
segons el qual electrons altament relativistes interaccionen amb camps magnètics<br />
i produeixen emissió ràdio, un percentatge <strong>de</strong> la qual està linealment polaritzada.<br />
Com en alguns sistemes, com en SS 433, l’emissió ràdio es va observar formant una<br />
estructura allargada o <strong>de</strong> tipus jet, com en els quàsars i nuclis actius <strong>de</strong> galàxies,<br />
es va proposar que fluxos d’electrons relativistes eren ejectats perpendicularment<br />
al disc d’acreció, i eren responsables d’emissió ràdio d’origen sincrotró en presència<br />
<strong>de</strong> camps magnètics. S’han proposat diversos mo<strong>de</strong>ls per explicar la formació <strong>de</strong>ls<br />
jets, tot i que no hi ha encara un mo<strong>de</strong>l àmpliament acceptat. El que si sembla
x Resum <strong>de</strong> la tesi<br />
imprescindible per la formació <strong>de</strong>ls jets és la presència d’un disc d’acreció prou<br />
proper a l’objecte compacte, ja que en els púlsars <strong>de</strong> raigs X no s’ha <strong>de</strong>tectat mai<br />
emissió ràdio d’origen sincrotró.<br />
El percentatge <strong>de</strong> sistemes <strong>de</strong> massa alta que emeten en ràdio (8 objectes) és<br />
aproximadament un 6%, mentre que en els sistemes <strong>de</strong> massa baixa (35 objectes)<br />
és d’un 23%. Si només consi<strong>de</strong>rem els objectes <strong>de</strong> la Galàxia (excloent els sistemes<br />
situats en els núvols <strong>de</strong> Magallanes) i excloem els púlsars <strong>de</strong> raigs X, aquests per-<br />
centatges es<strong>de</strong>venen 22% i 24%, respectivament. Per tant, veiem que d’entre les<br />
binàries <strong>de</strong> raigs X que po<strong>de</strong>n emetre en ràdio (excloent púlsars <strong>de</strong> raigs X) i on<br />
aquesta pot ser <strong>de</strong>tectada (objectes galàctics), ha estat així en un 20–25% <strong>de</strong>ls casos<br />
aproximadament, sense que hi influeixi la massa <strong>de</strong> l’estrella companya.<br />
1.2 Microquàsars<br />
Els microquàsars són estrelles binàries <strong>de</strong> raigs X amb emissió ràdio en forma <strong>de</strong><br />
jets relativistes. El nom va sorgir no només a partir <strong>de</strong> les similituds morfològiques<br />
entre aquestes fonts i els quàsars, sino també <strong>de</strong>gut a les similituds físiques, ja<br />
que quan l’objecte compacte és un forat negre, paràmetres com la temperatura<br />
<strong>de</strong> cos negre <strong>de</strong>l disc d’acreció, les mi<strong>de</strong>s característiques <strong>de</strong>ls jets o les escales<br />
temporals són escalables a partir <strong>de</strong> la massa <strong>de</strong> l’objecte central. En aquest sentit,<br />
els discs d’acreció <strong>de</strong>ls quàsars emeten en l’òptic i en l’ultraviolat, mentre que en<br />
els microquàsars ho fan en ragis X. D’altra banda, fenòmens que tenen lloc en<br />
escales <strong>de</strong> temps d’anys en quàsars, po<strong>de</strong>n ser estudiats en només alguns minuts en<br />
microquàsars.<br />
Actualment es coneixen uns 16 microquàsars, d’entre les 43 binàries <strong>de</strong> raigs X<br />
amb emissió ràdio.<br />
És interessant remarcar que alguns autors han proposat que<br />
totes les binàries <strong>de</strong> raigs X amb emissió ràdio són microquàsars, i que serien vistes<br />
com a tals si es disposés d’observacions amb prou sensibilitat i/o resolució angular.<br />
L’emissió ràdio <strong>de</strong>tectada en els jets <strong>de</strong>ls microquàsars forma habitualment núvols<br />
discrets <strong>de</strong> plasma, que es propaguen a través <strong>de</strong> l’espai amb velocitats relativistes.<br />
Així doncs, cal tenir en compte efectes <strong>de</strong> la relativitat especial a l’hora d’analitzar<br />
les observacions. Si consi<strong>de</strong>rem un parell <strong>de</strong> núvols <strong>de</strong> plasma ejectats perpendic-<br />
ularment al disc d’acreció i a banda i banda <strong>de</strong> l’objecte compacte en el mateix
Resum <strong>de</strong> la tesi xi<br />
instant <strong>de</strong> temps, i consi<strong>de</strong>rem que són intrínsecament iguals, les propietats obser-<br />
va<strong>de</strong>s d’un i altre <strong>de</strong>pendran <strong>de</strong> l’angle que formi la nostra visual amb la direcció<br />
<strong>de</strong>ls jets. Així, es pot <strong>de</strong>mostrar que el moviment propi mesurat en el pla <strong>de</strong>l cel<br />
per la component que s’acosti a nosaltres serà més elevat que no pas en el cas <strong>de</strong><br />
la component que s’allunyi <strong>de</strong> nosaltres. Un efecte interessant és l’anomenat movi-<br />
ment superlumínic, que té lloc quan la velocitat <strong>de</strong>l núvol és propera a la velocitat<br />
<strong>de</strong> la llum, i la velocitat mesurada és aparentment més gran que la <strong>de</strong> la llum. Es<br />
tracta d’un efecte <strong>de</strong> projecció, fàcilment explicable amb les equacions pertinents.<br />
Es pot veure que po<strong>de</strong>m mesurar un moviment superlumínic en una component<br />
que s’apropi amb una velocitat intrínseca <strong>de</strong> 0.71 vega<strong>de</strong>s la velocitat <strong>de</strong> la llum<br />
i movent-se en una direcció que formi un angle <strong>de</strong> 45 ◦ respecte la nostra visual.<br />
Per angles més grans o més petits, calen velocitats intrínseques més eleva<strong>de</strong>s per<br />
<strong>de</strong>tectar moviments superlumínics.<br />
D’altra banda, també es pot <strong>de</strong>mostrar que la <strong>de</strong>nsitat <strong>de</strong> flux observada en el<br />
núvol que s’acosta serà també més elevada que no pas la <strong>de</strong>l núvol que s’allunya.<br />
En els casos extrems, quan l’angle que formi el jet amb la nostra visual sigui molt<br />
petit, només <strong>de</strong>tectarem l’emissió <strong>de</strong> la component que s’apropa, com succeeix en el<br />
cas <strong>de</strong>ls quàsars, en què el seu flux pot arribar a augmentar com el factor <strong>de</strong> Lorentz<br />
elevat al cub.<br />
Així doncs, mesurant els moviments propis i/o les <strong>de</strong>nsitats <strong>de</strong> flux observa<strong>de</strong>s<br />
en les components que s’acosten i s’allunyen respecte l’observador en el cas d’un<br />
microquàsar, es pot obtenir una cota mínima <strong>de</strong> la velocitat β (velocitat expressada<br />
en unitats <strong>de</strong> la velocitat <strong>de</strong> la llum, β = v/c), i una cota màxima <strong>de</strong> l’angle <strong>de</strong>l jet<br />
respecte la visual, θ.<br />
Sis <strong>de</strong>ls 16 microquàsars coneguts són binàries <strong>de</strong> raigs X <strong>de</strong> massa alta, i els<br />
<strong>de</strong>u restants ho són <strong>de</strong> massa baixa. La majoria <strong>de</strong> sistemes <strong>de</strong> massa alta són<br />
persistents pel que fa a l’emissió ràdio, mentre que la majoria <strong>de</strong>ls <strong>de</strong> massa baixa<br />
son fonts transitòries. En 4 microquàsars s’han mesurat moviments superlumínics,<br />
i existeixen indicis d’aquest fet en un cinquè cas. Pel que fa als objectes compactes,<br />
és segur que en 6 d’ells es tracta d’un forat negre, mentre que se sospita <strong>de</strong> la seva<br />
presència en uns altres 6 casos. En dos microquàsars l’objecte compacte és una<br />
estrella <strong>de</strong> neutrons, i se sospita el mateix en una altres dos casos.<br />
Com ja hem comentat, les escales <strong>de</strong> temps en els microquàsars són molt més<br />
petites que en el cas <strong>de</strong>ls quàsars. Així, ha estat possible <strong>de</strong>tectar ràpi<strong>de</strong>s transicions
xii Resum <strong>de</strong> la tesi<br />
<strong>de</strong>ls sistemes realitzant observacions multi-longitud d’ona simultànies. Un exemple<br />
són les observacions en escales <strong>de</strong> minuts <strong>de</strong> GRS 1915+105. En aquest sistema<br />
s’ha pogut inferir que una disminució sobtada <strong>de</strong>l flux <strong>de</strong> raigs X corresponia a<br />
la <strong>de</strong>saparició <strong>de</strong> la part interna <strong>de</strong>l disc d’acreció, a continuació <strong>de</strong> la qual es<br />
produïa l’ejecció <strong>de</strong> plasma en forma <strong>de</strong> jet a velocitat relativista, ja que primerament<br />
s’observava una erupció en l’infraroig que venia seguida d’una erupció en ràdio.<br />
D’altra banda, observacions <strong>de</strong> gran abast temporal <strong>de</strong>l sistema GX 339−4 han<br />
permès estudiar la correlació entre el flux <strong>de</strong> raigs X tous (provinents <strong>de</strong>l disc d’acre-<br />
ció), el flux <strong>de</strong> raigs X durs (provinents d’una corona per sobre i per sota <strong>de</strong>l disc)<br />
i la <strong>de</strong>nsitat <strong>de</strong> flux en el domini ràdio (provinent <strong>de</strong>l jet). S’ha pogut establir una<br />
correlació un a un entre els raigs X durs i l’emissió en ràdio, i s’ha proposat que els<br />
raigs X durs <strong>de</strong> la corona podrien ser <strong>de</strong> fet la base <strong>de</strong>l jet relativista. Els raigs X<br />
tous semblen traçar els canvis d’estat <strong>de</strong>ls forats negres, <strong>de</strong> manera que quan el flux<br />
augmenta molt (el dics arriba molt a prop <strong>de</strong> l’objecte compacte), tant l’emissió <strong>de</strong><br />
raigs X durs com l’emissió ràdio són suprimi<strong>de</strong>s, probablement <strong>de</strong>gut a la <strong>de</strong>saparició<br />
<strong>de</strong>l jet en el sistema.<br />
1.3 Motivació <strong>de</strong> la tesi<br />
Tal com hem vist, actualment es coneixen al voltant <strong>de</strong> 16 microquàsars, tot i<br />
que aquest nombre podria augmentar fins a 43 si tal i com han proposat diversos<br />
autors totes les binàries <strong>de</strong> raigs X amb emissió ràdio contenen jets relativistes. De<br />
tota manera, la situació encara no és clara. A més, hi ha molts tipus diferents <strong>de</strong><br />
binàries <strong>de</strong> raigs X amb emissió ràdio: d’alta i baixa massa, totes les fonts <strong>de</strong> tipus Z<br />
i només 5 fonts <strong>de</strong> tipus Atoló en el cas <strong>de</strong> baixa massa, amb forats negres i estrelles<br />
<strong>de</strong> neutrons com a objectes compactes, etc. Així doncs, quan comencem a dividir<br />
el grup <strong>de</strong> binàries <strong>de</strong> raigs X que emeten en ràdio segons aquests diferents tipus,<br />
no po<strong>de</strong>m realitzar estudis estadísticament significatius. Aquesta situació empitjora<br />
notablement si només consi<strong>de</strong>rem els 16 microquàsars.<br />
A més, s’han proposat algunes correlacions, com per exemple que la velocitat<br />
<strong>de</strong>ls jets relativistes <strong>de</strong>pèn <strong>de</strong> la massa <strong>de</strong> l’objecte compacte. De tota manera, hi<br />
ha pocs casos on la velocitat <strong>de</strong>l jet i la natura <strong>de</strong> l’objecte compacte siguin coneguts.<br />
Així doncs, és impossible realitzar afirmacions d’aquest tipus a partir <strong>de</strong>l nombre<br />
limitat d’objectes coneguts.
Resum <strong>de</strong> la tesi xiii<br />
D’altra banda, hi ha una pregunta important a contestar: <strong>de</strong> quin tipus és la<br />
matèria que forma els jets? Fins ara només s’han <strong>de</strong>tectat línies espectrals atòmiques<br />
en el microquàsar SS 433, que indiquen que en aquest cas els jets son bariònics. De<br />
totes maneres, alguns <strong>de</strong>l mo<strong>de</strong>ls proposats per explicar els processos d’ejecció al<br />
voltant d’objectes compactes només preveuen jets leptònics, formats per electrons i<br />
positrons. Així doncs, és fonamental <strong>de</strong>svetllar el tipus <strong>de</strong> matèria <strong>de</strong>ls jets per tal<br />
<strong>de</strong> po<strong>de</strong>r construir mo<strong>de</strong>ls adients.<br />
Un altre aspecte interessant és la possible emissió <strong>de</strong> raigs gamma d’alta energia<br />
en els microquàsars. Abans d’aquest treball, només el sistema LS I +61 303 podia ser<br />
associat amb una d’aquestes fonts. Actualment, però, el candidat més prometedor és<br />
el microquàsar LS 5039, com es veurà posteriorment. En qualsevol cas, un grup <strong>de</strong><br />
dos objectes és realment pobre, i val la pena realitzar un esforç per tal d’ampliar-lo.<br />
Així doncs, van ser bàsicament dues les raons que van motivar l’inici d’aquest<br />
projecte a llarg termini enfocat en la cerca <strong>de</strong> nous microquàsars a la Galàxia:<br />
• Cada nou microquàsar ha permès nous i interessants <strong>de</strong>scobriments fenomenològics,<br />
que han <strong>de</strong> ser tinguts en compte pels mo<strong>de</strong>ls proposats per explicar els<br />
fenòmens d’acreció/ejecció que tenen lloc al voltant d’objectes compactes.<br />
• Augmentar el grup reduït <strong>de</strong> microquàsars coneguts i permetre estudis es-<br />
tadístics en el futur.<br />
En l’apartat 2 presentem el <strong>de</strong>scobriment i subsegüent estudi <strong>de</strong>l microquàsar<br />
LS 5039. Aquesta va ser la primera font <strong>de</strong>scoberta dins <strong>de</strong>l mèto<strong>de</strong> més global<br />
presentat en l’apartat 3, on expliquem com s’han utilitzat els millors catàlegs <strong>de</strong><br />
raigs X i <strong>de</strong> ràdio per tal <strong>de</strong> buscar noves candidates a binàries <strong>de</strong> rais X amb emis-<br />
sió ràdio/microquàsars. Finalment, llistem les nostres conclusions <strong>de</strong>sprés d’aquest<br />
treball en l’apartat 4.<br />
2. Descobriment i estudi <strong>de</strong>l microquàsar LS 5039<br />
Els catàlegs disponibles en diferents rangs <strong>de</strong> l’espectre electromagnètic (bàsicament<br />
raigs X, òptic i ràdio) proporcionen una eina fonamental en la cerca <strong>de</strong> fonts amb
xiv Resum <strong>de</strong> la tesi<br />
un comportament multi-longitud d’ona conegut. Així doncs, com a primer pas en<br />
la nostra cerca <strong>de</strong> nous microquàsars, vam inspeccionar en longituds d’ona ràdio<br />
aquells objectes que havien estat classificats com a candidats a binàries <strong>de</strong> raigs X.<br />
En aquest context, Motch et al. (1997) van dur a terme una i<strong>de</strong>ntificació creua-<br />
da entre el catàleg <strong>de</strong> raigs X ROSAT All Sky Survey (RASS) i les estrelles <strong>de</strong><br />
tipus espectrals O i B presents en la base <strong>de</strong> da<strong>de</strong>s SIMBAD. Quasi bé 20 nous ob-<br />
jectes candidats a estrelles binàries <strong>de</strong> raigs X <strong>de</strong> massa alta van ser proposats fruit<br />
d’aquest treball. En particular, l’estrella LS 5039, situada a una distància <strong>de</strong> 3 kpc<br />
aproximadament, mostrava un espectre <strong>de</strong> raigs X dur, compatible amb un sistema<br />
binari on un forat negre o una estrella <strong>de</strong> neutrons acretés matèria proporcionada<br />
pel vent <strong>de</strong> l’estrella LS 5039, <strong>de</strong> tipus espectral O7V((f)).<br />
Per tal <strong>de</strong> trobar contraparti<strong>de</strong>s ràdio a les estrelles proposa<strong>de</strong>s per Motch et al.<br />
(1997) com a noves binàries <strong>de</strong> raigs X, vam inspeccionar les seves posicions en el<br />
catàleg radio NRAO VLA Sky Survey (NVSS). Aquest catàleg cobreix el cel per<br />
sobre d’una <strong>de</strong>clinació <strong>de</strong> δ = −40 ◦ (82% <strong>de</strong> l’esfera celest) a una freqüència <strong>de</strong><br />
1.4 GHz (20 cm <strong>de</strong> longitud d’ona) utilitzant el Very Large Array (VLA) en les<br />
seves configuracions més compactes, D i DC. Conté mes <strong>de</strong> 1.8 × 10 6 fonts amb<br />
una <strong>de</strong>nsitat <strong>de</strong> flux superior a 2.5 mJy, el seu límit <strong>de</strong> completesa. L’error en la<br />
posició <strong>de</strong> les fonts és inferior a un segon d’arc per les fonts amb una <strong>de</strong>nsitat <strong>de</strong><br />
flux superior a 15 mJy i <strong>de</strong> set segons d’arc per les fonts més febles <strong>de</strong>tecta<strong>de</strong>s.<br />
Com a resultat <strong>de</strong> la nostra cerca, vam <strong>de</strong>scobrir una font ràdio <strong>de</strong> 24 mJy<br />
coinci<strong>de</strong>nt amb la posició <strong>de</strong> LS 5039, i catalogada com a NVSS J182614−145054<br />
en el NVSS. A continuació passem a <strong>de</strong>tallar l’estudi multi-longitud d’ona realitzat<br />
d’aquest objecte.<br />
2.1 Estudi multi-longitud d’ona <strong>de</strong> LS 5039<br />
Per tal <strong>de</strong> confirmar l’associació entre la font <strong>de</strong> raigs X, l’objecte òptic i la font<br />
ràdio, vam realitzar observacions amb l’interferòmetre VLA en la seva configuració<br />
A (la més estesa i que proporciona una millor resolució angular). Les observacions<br />
van ser realitza<strong>de</strong>s en 4 èpoques diferents entre febrer i maig <strong>de</strong> 1998 i a 4 freqüències<br />
diferents, per tal d’obtenir informació sobre la variabilitat i l’espectre ràdio, respec-<br />
tivament. Com a resultat, vam po<strong>de</strong>r confirmar que les posicions ràdio i òptiques
Resum <strong>de</strong> la tesi xv<br />
concordaven entre elles, amb una diferència inferior a 0.1 segons d’arc. La probabil-<br />
itat d’una associació errònia per coincidència aleatòria era inferior a 10 −7 . D’altra<br />
banda, l’espectre obtingut, que patia lleugeres variacions al llarg <strong>de</strong>l temps, podia<br />
ajustar-se amb un ín<strong>de</strong>x espectral <strong>de</strong> l’ordre <strong>de</strong> α −0.5 (on la <strong>de</strong>nsitat <strong>de</strong> flux<br />
ve donada per Sν ∝ ν +α ). L’elevada temperatura <strong>de</strong> brillantor, així com l’espectre<br />
observat, suggerien emissió ràdio d’origen sincrotró. L’ús d’equacions d’equipartició<br />
proporcionava una estimació <strong>de</strong> l’energia màxima <strong>de</strong>ls electrons, E ≤ 4 × 10 41 erg,<br />
i <strong>de</strong>l camp magnètic mínim, B ≥ 0.01 G. Així doncs, vam concloure (Martí et al.<br />
1998) que es tractava d’una nova binària <strong>de</strong> raigs X <strong>de</strong> massa alta amb emissió ràdio<br />
d’origen sincrotró.<br />
A continuació vam <strong>de</strong>manar que aquest objecte fos observat regularment dins<br />
la campanya GBI-NASA Monitoring Program (GBI), que proporcionava mesures<br />
<strong>de</strong> flux ràdio a dues freqüències (2.25 i 8.3 GHz). Es van obtenir da<strong>de</strong>s durant un<br />
primer perío<strong>de</strong> d’onze mesos (setembre 1998–agost 1999) i un segon perío<strong>de</strong> d’un<br />
sol mes (setembre 2000). L’anàlisi d’aquestes da<strong>de</strong>s (preliminar a Ribó et al. 1999,<br />
i complet aquí) mostra que la font és persistent, mo<strong>de</strong>radament variable i amb un<br />
ín<strong>de</strong>x espectral α −0.5, com s’havia vist anteriorment. D’altra banda, una anàlisi<br />
temporal <strong>de</strong> les da<strong>de</strong>s mostra que no hi apareix cap senyal periòdica, un fet esperable<br />
tenint en compte la relació senyal-soroll relativament baixa <strong>de</strong> les da<strong>de</strong>s.<br />
El pas següent per veure si es tractava d’un microquàsar va consistir en realitzar<br />
observacions d’interferometria <strong>de</strong> molt llarga base (Very Long Baseline Interferom-<br />
etry, VLBI) amb l’interferòmetre Very Long Baseline Array (VLBA). Les observa-<br />
cions es van realitzar el 8 <strong>de</strong> maig <strong>de</strong> 1999, durant 4 hores i a una freqüència <strong>de</strong><br />
5 GHz, en el mo<strong>de</strong> d’observació v6cm-256-8-2-L, que proporciona un ample <strong>de</strong> ban-<br />
da <strong>de</strong> 64 MHz. La imatge resultant d’aquestes observacions es mostra en la Fig. 1,<br />
on es pot veure emissió en forma <strong>de</strong> jet bipolar sorgint a banda i banda d’una font<br />
central intensa. Com es pot observar, el jet que apunta cap al sud-est (inferior es-<br />
querra) és més brillant i es troba a una distància major respecte el centre que no pas<br />
el jet nord-oest (superior dret). Aquesta asimetria pot ser interpretada segons els<br />
efectes relativistes mencionats anteriorment, que impliquen una velocitat intrínseca<br />
mínima <strong>de</strong> 0.15c (on c és la velocitat <strong>de</strong> la llum) pel plasma en els jets, i un angle<br />
màxim <strong>de</strong> posició <strong>de</strong>ls jets respecte la visual <strong>de</strong> 81 ◦ . Assumint equipartició, s’obté<br />
una estimació <strong>de</strong> l’energia <strong>de</strong>ls electrons, E ∼ 5 × 10 39 erg, i el camp magnètic,<br />
B ∼ 0.2 G. Aquests resultats, confirmaven que LS 5039 era un nou microquàsar a<br />
la Galàxia (Pare<strong>de</strong>s et al. 2000).
xvi Resum <strong>de</strong> la tesi<br />
MilliARC SEC<br />
4<br />
2<br />
0<br />
-2<br />
-4<br />
-6<br />
VLBA<br />
6 4 2 0 -2 -4 -6<br />
MilliARC SEC<br />
Figura 1: Imatge autocalibrada <strong>de</strong> LS 5039 obtinguda amb el VLBA a 5 GHz.<br />
A partir d’aleshores diversos grups van començar a estudiar aquest objecte,<br />
bàsicament en longituds d’ona òptiques, i es van obtenir resultats com el perío<strong>de</strong><br />
orbital o una classificació espectral més acurada, tal i com s’explica més endavant.<br />
Com que havíem resolt els jets <strong>de</strong> LS 5039 en un moment d’emissió persistent<br />
en ràdio, sense que hi hagués hagut cap erupció prèvia en raigs X o en ràdio, era<br />
d’esperar que els jets fossin sempre presents en aquest microquàsar. Per tal <strong>de</strong><br />
confirmar-ho, vam realitzar observacions amb la xarxa europea <strong>de</strong> VLBI (European<br />
VLBI Network, EVN) i la xarxa <strong>de</strong> sis antenes angleses connecta<strong>de</strong>s (Multi-Element<br />
Radio-Linked Interferometer Network, MERLIN). Les observacions es van realitzar<br />
l’1 <strong>de</strong> març <strong>de</strong> 2000, utilitzant el mateix mo<strong>de</strong> d’observació que l’emprat amb el<br />
VLBA. Els resultats confirmen la presència persistent <strong>de</strong> jets relativistes a LS 5039,<br />
que s’estenen <strong>de</strong>s <strong>de</strong> ∼ 10 a ∼ 1000 unitats astronòmiques (Pare<strong>de</strong>s et al. 2002a).<br />
A més també s’observen les asimetries <strong>de</strong>gu<strong>de</strong>s als efectes relativistes, la qual cosa<br />
permet confirmar els resultats anteriors pel que fa als limits <strong>de</strong> la velocitat mínima
Resum <strong>de</strong> la tesi xvii<br />
<strong>de</strong>l plasma i l’angle màxim <strong>de</strong>ls jets respecte la visual.<br />
Poc <strong>de</strong>sprés <strong>de</strong>l <strong>de</strong>scobriment <strong>de</strong> l’emissió ràdio <strong>de</strong> LS 5039, vam realitzar ob-<br />
servacions fotomètriques en l’òptic amb el telescopi <strong>de</strong> 1.5 m <strong>de</strong> l’Observatorio As-<br />
tronómico Nacional (OAN), el juny <strong>de</strong> 1998, que van permetre <strong>de</strong>tectar variacions<br />
<strong>de</strong> l’ordre <strong>de</strong> 0.05 magnituds en intervals d’un dia, amb un valor mitjà <strong>de</strong> V = 11.36<br />
(Martí et al. 1998).<br />
Posteriorment, Clark et al. (2001) van realitzar observacions fotomètriques i<br />
espectroscòpiques, que van <strong>de</strong>mostrar que LS 5039 patia variacions fotomètriques<br />
molt mo<strong>de</strong>ra<strong>de</strong>s en l’òptic (0.1 magnituds) i més apreciables en l’infraroig prop-<br />
er (0.5 magnituds) en escales temporals d’anys. Gràcies a les observacions espec-<br />
troscòpiques, aquests mateixos autors van concloure que el tipus espectral <strong>de</strong> l’estrel-<br />
la era O6.5V((f)). Amb aquesta nova classificació i utilitzant noves calibracions per<br />
estrelles joves, nosaltres hem <strong>de</strong>duït una distància a LS 5039 <strong>de</strong> 2.9 ± 0.3 kpc (Ribó<br />
et al. 2002a).<br />
D’altra banda, McSwain et al. (2001) van obtenir la corba <strong>de</strong> velocitat radial<br />
<strong>de</strong>l sistema binari, i van <strong>de</strong>duir els següents paràmetres orbitals: un perío<strong>de</strong> <strong>de</strong> 4.12<br />
dies, una velocitat radial <strong>de</strong>l sistema <strong>de</strong> 4.6 km s −1 , una amplitud <strong>de</strong> variació <strong>de</strong><br />
la velocitat radial <strong>de</strong> 15 km s −1 , una excentricitat <strong>de</strong> 0.41, una funció <strong>de</strong> massa <strong>de</strong><br />
0.0010 M⊙ i una velocitat rotacional projectada <strong>de</strong> 131 km s −1 .<br />
Per tal <strong>de</strong> confirmar aquests resultats, hem realitzat noves observacions espec-<br />
troscòpiques amb el telescopi <strong>de</strong> 2.5 m Isaac Newton Telescope (INT), <strong>de</strong>l 23 al 31<br />
<strong>de</strong> juliol <strong>de</strong> 2002. Aquestes observacions recents confirmen els valors <strong>de</strong>l perío<strong>de</strong><br />
orbital i la velocitat rotacional projectada (Casares et al. 2002).<br />
Pel que fa a l’emissió en raigs X <strong>de</strong> LS 5039, hem dut a terme una anàlisi <strong>de</strong>ls<br />
sis anys i mig <strong>de</strong> da<strong>de</strong>s <strong>de</strong>l seguiment realitzat amb el All Sky Monitoring a bord <strong>de</strong>l<br />
satèl·lit Rossi X-ray Timing Explorer (ASM/RXTE) <strong>de</strong> la contrapartida en raigs X,<br />
anomenada RX J1826.2−1450. Els resultats, igual que en les anàlisis <strong>de</strong> dos anys<br />
i mig <strong>de</strong> da<strong>de</strong>s presenta<strong>de</strong>s a Ribó et al. (1999), indiquen que no hi ha variabilitat<br />
periòdica en aquestes da<strong>de</strong>s. En particular no trobem el perío<strong>de</strong> orbital obtingut<br />
espectroscòpicament. Tenint en compte l’excentricitat <strong>de</strong> l’òrbita, serien d’esperar<br />
variacions <strong>de</strong>l ritme d’acreció <strong>de</strong> massa per part <strong>de</strong> l’objecte compacte amb el perío<strong>de</strong><br />
orbital, i per tant també <strong>de</strong> la lluminositat en raigs X d’aquest sistema binari. De<br />
tota manera, la relació senyal-soroll <strong>de</strong> les da<strong>de</strong>s segurament no és prou bona com
xviii Resum <strong>de</strong> la tesi<br />
per po<strong>de</strong>r <strong>de</strong>tectar aquestes variacions, que també s’haurien d’observar en el domini<br />
ràdio, i no ha estat així en les observacions GBI, tot i que com ja s’ha dit també<br />
tenen una relació senyal-soroll força baixa.<br />
Per tal d’estudiar l’emissió en raigs X amb cura, vam realitzar observacions<br />
amb l’instrument Proportional Counter Array <strong>de</strong>l satèl·lit RXTE (PCA/RXTE), el<br />
febrer <strong>de</strong> 1998, que van permetre mesurar un espectre <strong>de</strong> raigs X significativament<br />
dur fins a 30 keV, on <strong>de</strong>stacava la presència d’una línia <strong>de</strong> ferro a una energia<br />
<strong>de</strong> 6.6 keV. D’altra banda, no s’observaven pulsacions (com era d’esperar ja que el<br />
sistema presentava emissió ràdio) ni variacions quasi-periòdiques en escales <strong>de</strong> temps<br />
<strong>de</strong> 0.02 a 2000 segons. La lluminositat obtinguda a una distància <strong>de</strong> 2.9 kpc és <strong>de</strong><br />
5 × 10 34 erg s −1 , relativament baixa per una binària <strong>de</strong> raigs X (Ribó et al. 1999).<br />
Posteriorment vam realitzar noves observacions amb el satèl·lit <strong>de</strong> raigs X Bep-<br />
poSAX l’octubre <strong>de</strong> 2000. Aquestes observacions van indicar una <strong>de</strong>vallada <strong>de</strong> la llu-<br />
minositat en quasi bé un ordre <strong>de</strong> magnitud respecte les observacions PCA/RXTE.<br />
Això pot ser <strong>de</strong>gut a la variació <strong>de</strong>l vent estel·lar <strong>de</strong> LS 5039, i per tant <strong>de</strong>l seu ritme<br />
<strong>de</strong> pèrdua <strong>de</strong> massa, i en conseqüència <strong>de</strong>l ritme d’acreció <strong>de</strong> matèria per part <strong>de</strong> l’ob-<br />
jecte compacte. A més, aquestes observacions van cobrir el rang <strong>de</strong> fases orbitals <strong>de</strong>l<br />
sistema on s’esperaria trobar un eclipsi <strong>de</strong> raigs X, sense que aquest hagi estat <strong>de</strong>tec-<br />
tat. Aquest fet, dóna suport a la suggerència que el sistema binari té una inclinació<br />
força baixa, al voltant <strong>de</strong> 30 ◦ , segons semblen indicar observacions d’espectroscopia<br />
òptica recents (McSwain & Gies 2002). Finalment, hem pogut obtenir un valor per<br />
la <strong>de</strong>nsitat columnar d’hidrogen a l’objecte, que és <strong>de</strong> NH = 1.0 ± 0.3 × 10 22 cm −2<br />
(Reig et al. 2002).<br />
Poc <strong>de</strong>sprés <strong>de</strong>l <strong>de</strong>scobriment <strong>de</strong> la natura microquàsar <strong>de</strong> LS 5039, vam in-<br />
speccionar els catàlegs <strong>de</strong> fonts d’altes energies per tal <strong>de</strong> buscar-hi una possible<br />
contrapartida. Com a resultat, vam <strong>de</strong>scobrir la presència d’una font <strong>de</strong> raigs gam-<br />
ma d’alta energia no i<strong>de</strong>ntificada en una posició concordant amb LS 5039. A més, el<br />
microquàsar és l’única font <strong>de</strong> raigs X brillant, i l’única font <strong>de</strong> raigs X que presenta<br />
emissió ràdio, en el camp <strong>de</strong> la font gamma, anomenada 3EG J1824−1514 segons<br />
les coor<strong>de</strong>na<strong>de</strong>s en el tercer catàleg <strong>de</strong>l <strong>de</strong>tector EGRET. D’altra banda, tant l’e-<br />
missió ràdio <strong>de</strong> LS 5039 com l’emissió gamma <strong>de</strong> 3EG J1824−1514 són persistents<br />
i mo<strong>de</strong>radament variables al llarg <strong>de</strong>l temps. A més, una anàlisi <strong>de</strong> la variabilitat<br />
gamma sembla excloure un púlsar o un remanent <strong>de</strong> supernova com a origen <strong>de</strong> la<br />
font gamma. Finalment, l’espectre en raigs gamma no concorda amb els que habit-
Resum <strong>de</strong> la tesi xix<br />
ualment es troben en els púlsars <strong>de</strong> raigs gamma, úniques fonts EGRET d’origen<br />
galàctic i<strong>de</strong>ntifica<strong>de</strong>s fins el moment, excloent una erupció solar. Tots aquests fets<br />
suggereixen una associació entre els dos objectes, com vam proposar a Pare<strong>de</strong>s et al.<br />
(2000). A més <strong>de</strong> la possibilitat <strong>de</strong> <strong>de</strong>tectar els microquàsars en raigs gamma d’alta<br />
energia, és segurament més interessant el fet que hi ha prop <strong>de</strong> 170 fonts EGRET no<br />
i<strong>de</strong>ntifica<strong>de</strong>s, 72 d’elles a menys <strong>de</strong> 10 ◦ <strong>de</strong>l pla galàctic. Així doncs, hem proposat<br />
que algunes d’aquestes fonts podrien ser microquàsars persistents i “silenciosos”,<br />
com LS 5039, que encara han <strong>de</strong> ser <strong>de</strong>scoberts.<br />
Per tal d’explicar el comportament multi-longitud d’ona <strong>de</strong> LS 5039 <strong>de</strong>s <strong>de</strong>l<br />
domini ràdio fins als raigs gamma, hem proposat un mo<strong>de</strong>l que explica l’emissió<br />
gamma com a resultat <strong>de</strong> l’efecte Compton invers que pateixen els fotons ultraviolats<br />
<strong>de</strong> l’estrella per part <strong>de</strong>ls electrons relativistes <strong>de</strong>l jet. L’elevat camp <strong>de</strong> radiació <strong>de</strong><br />
fotons ultraviolats fa <strong>de</strong> fet inevitable la producció <strong>de</strong> raigs gamma <strong>de</strong>gut a l’efecte<br />
Compton invers si tenim en compte la presència d’electrons altament relativistes<br />
(amb factors <strong>de</strong> Lorentz <strong>de</strong> l’ordre <strong>de</strong> 10 3 –10 4 ), que posteriorment, i a mida que<br />
s’allunyen <strong>de</strong>l camp <strong>de</strong> radiació, donaran lloc a emissió ràdio d’origen sincrotró. A<br />
més, el mo<strong>de</strong>l suggereix que el jet és força col·limat fins a les escales <strong>de</strong>tecta<strong>de</strong>s <strong>de</strong><br />
∼ 1000 unitats astronòmiques (Pare<strong>de</strong>s et al. 2002a).<br />
Finalment, hem realitzat algunes consi<strong>de</strong>racions energètiques i hem suggerit ob-<br />
servacions que es podrien dur a terme en el futur, per tal <strong>de</strong> po<strong>de</strong>r millorar el<br />
mo<strong>de</strong>l proposat i entendre millor els fenòmens d’acreció/ejecció en el cas d’aquest<br />
microquàsar.<br />
2.2 LS 5039 com un microquàsar en fuga<br />
La formació d’un objecte compacte en una estrella binària <strong>de</strong> raigs X, requereix<br />
necessàriament que hagi tingut lloc una explosió <strong>de</strong> supernova que no hagi trencat<br />
el sistema binari.<br />
És d’esperar que aquest es<strong>de</strong>veniment explosiu canviï consi<strong>de</strong>r-<br />
ablement les propietats cinemàtiques <strong>de</strong>l sistema.<br />
Per tal <strong>de</strong> comprovar quina és la situació en el cas <strong>de</strong>l microquàsar LS 5039 hem<br />
recopilat set posicions òptiques, que cobreixen quasi bé cent anys <strong>de</strong> temps, i dues<br />
posicions ràdio, que només cobreixen dos anys <strong>de</strong> temps. Hem realitzat ajustos <strong>de</strong><br />
moviments propis tant en les posicions òptiques com en les ràdio, i hem obtingut
xx Resum <strong>de</strong> la tesi<br />
valors força similars en ambdós casos. Així, po<strong>de</strong>m dir que basats només en da<strong>de</strong>s<br />
astromètriques som també capaços <strong>de</strong> confirmar que ambdues emissions, en l’òptic<br />
i en ràdio, s’originen en el mateix objecte.<br />
Utilitzant totes les da<strong>de</strong>s conjuntament, hem obtingut els següents moviments<br />
propis: µα cos δ = (4.7 ± 1.1) × 10 −3 ′′ any −1 , µδ = (−10.6 ± 1.0) × 10 −3 ′′ any −1 .<br />
Aquests, juntament amb la nova estimació <strong>de</strong> la distància <strong>de</strong> 2.9 ± 0.3 kpc i la<br />
velocitat radial <strong>de</strong>l sistema <strong>de</strong> Vr = 4.6 ± 0.5 km s −1 (McSwain et al. 2001), han<br />
permès calcular una velocitat espacial <strong>de</strong> (U = 51, V = −71, W = −118) km s −1<br />
respecte el seu entorn local (Regional Standard of Rest). Aquest resultat implica<br />
que LS 5039 és un microquàsar en fuga (runaway), amb una velocitat sistèmica <strong>de</strong><br />
vsis 150 km s −1 , escapant <strong>de</strong>l seu entorn local amb una elevada component <strong>de</strong> la<br />
velocitat perpendicular al pla galàctic. Això és probablement <strong>de</strong>gut a l’explosió <strong>de</strong><br />
supernova que va crear l’objecte compacte en aquest sistema binari.<br />
Hem calculat la trajectòria <strong>de</strong> LS 5039 en el passat, i trobat dues associacions OB<br />
properes al seu camí traçat en el pla <strong>de</strong>l cel. De totes maneres, aquestes associacions<br />
OB són massa properes a nosaltres per estar relaciona<strong>de</strong>s amb el microquàsar. D’al-<br />
tra banda, també hem trobat tres remanents <strong>de</strong> supernova a prop <strong>de</strong>l camí traçat<br />
per LS 5039. Després <strong>de</strong> <strong>de</strong>scartar dues d’elles basats en les distàncies que ens en<br />
separen, hem centrat la nostra atenció en el remanent SNR G016.8−01.1. L’estudi<br />
d’aquesta font no ha pogut ni confirmar ni refusar una possible associació amb el<br />
microquàsar, bàsicament per les incerteses en l’estimació <strong>de</strong> la <strong>de</strong>nsitat <strong>de</strong> flux ràdio<br />
<strong>de</strong>l remanent. Aquest fet segurament justifica noves observacions d’alta sensibilitat<br />
per tal <strong>de</strong> buscar remanents dèbils en aquesta regió <strong>de</strong>l cel.<br />
D’altra banda, també hem trobat una cavitat semi-oberta <strong>de</strong> H i a prop <strong>de</strong> la<br />
posició <strong>de</strong> LS 5039. Encara que l’estrella O((f)) en aquest microquàsar sembla ser<br />
el creador d’aquesta bombolla, no po<strong>de</strong>m <strong>de</strong>scartar una contribució <strong>de</strong> l’estrella<br />
progenitora <strong>de</strong> l’objecte compacte.<br />
Basats en càlculs teòrics, hem estat capaços d’explicar tant la velocitat espa-<br />
cial elevada com l’excentricitat observa<strong>de</strong>s en un escenari d’explosió <strong>de</strong> supernova<br />
simètrica, amb una important pèrdua <strong>de</strong> massa <strong>de</strong> ∆ M ∼ 17 M⊙.<br />
Finalment, la velocitat espacial elevada i el possible temps <strong>de</strong> vida <strong>de</strong>l micro-<br />
quàsar indiquen que aquest podria assolir una latitud galàctica <strong>de</strong> b = −12 ◦ . Per<br />
tant, si l’associació entre LS 5039 i la font EGRET 3EG J1824−1514 és correcta,
Resum <strong>de</strong> la tesi xxi<br />
podríem <strong>de</strong>tectar microquàsars <strong>de</strong> raigs gamma fins a latituds galàctiques <strong>de</strong> l’ordre<br />
<strong>de</strong> |b| 10 ◦ . En particular, els microquàsars en fuga podrien estar relacionats amb<br />
algunes <strong>de</strong> les fonts febles, variables i relativament toves encara no i<strong>de</strong>ntifica<strong>de</strong>s <strong>de</strong>l<br />
tercer catàleg EGRET, situa<strong>de</strong>s per sobre i per sota <strong>de</strong>l pla galàctic. Tots aquests<br />
resultats van ser publicats a Ribó et al. (2002a).<br />
3. Una cerca <strong>de</strong> nous microquàsars<br />
Com ja hem comentat anteriorment, el reduït nombre <strong>de</strong> microquàsars coneguts ens<br />
va donar un bon motiu per començar un projecte a llarg termini per tal <strong>de</strong> cercar<br />
nous microquàsars.<br />
3.1 Mèto<strong>de</strong> d’i<strong>de</strong>ntificació creuada<br />
Per tal <strong>de</strong> <strong>de</strong>scobrir nous microquàsars d’una manera sistemàtica, hem realitzat una<br />
i<strong>de</strong>ntificació creuada entre els millors catàlegs <strong>de</strong> raigs X i ràdio disponibles.<br />
Pel que fa als raigs X, hem utilitzat el catàleg <strong>de</strong> fonts brillants <strong>de</strong>l satèl·lit<br />
ROSAT (ROSAT All Sky Brigth Source Catalog, RBSC), que conté en la seva versió<br />
final un total <strong>de</strong> 18.806 fonts <strong>de</strong>tecta<strong>de</strong>s en el rang d’energia entre 0.1 i 2.4 keV, i<br />
que és un extracte <strong>de</strong>l ROSAT All Sky Survey (RASS). En el RBSC hi ha informació<br />
en quatre ban<strong>de</strong>s d’energia expressa<strong>de</strong>s en keV: A=0.1–0.4, B=0.5–2.0, C=0.5–0.9,<br />
D=0.9–2.0. A partir d’aquestes ban<strong>de</strong>s, es calculen dos ín<strong>de</strong>xs <strong>de</strong> duresa: HR1 =<br />
(B − A)/(B + A), HR2 = (D − C)/(D + C). Els errors a 1σ en la posició <strong>de</strong> les<br />
fonts són típicament <strong>de</strong> 10–20 ′′ . Pel que fa al domini ràdio, hem emprat el catàleg<br />
NVSS, que ja ha estat <strong>de</strong>scrit en l’apartat 2.<br />
Com que el nostre objectiu era maximitzar la probabilitat <strong>de</strong> retenir només les<br />
estrelles binàries <strong>de</strong> raigs X amb emissió ràdio, hem adoptat els següents criteris <strong>de</strong><br />
selecció:<br />
1. Hem seleccionat les fonts amb latituds galàctiques |b| < 5 ◦ <strong>de</strong>l catàleg RBSC.<br />
Com que el NVSS té un limit <strong>de</strong> δ > −40 ◦ , només les fonts per sobre d’aquesta<br />
<strong>de</strong>clinació han estat consi<strong>de</strong>ra<strong>de</strong>s. Així doncs, hem cobert aproximadament el<br />
75% <strong>de</strong> l’àrea |b| < 5 ◦ (l 347–259 ◦ i α 17.2–8.6 h ).
xxii Resum <strong>de</strong> la tesi<br />
2. D’entre les fonts RBSC restants, les que tenien indicadors <strong>de</strong> fonts properes que<br />
contaminaven les mesures o indicadors <strong>de</strong> problemes amb les <strong>de</strong>terminacions<br />
<strong>de</strong> les posicions, han estat excloses.<br />
3. A partir d’estudis estadístics <strong>de</strong>l RBSC, Motch et al. (1998) van concloure<br />
que les binàries <strong>de</strong> raigs X eren fàcilment i<strong>de</strong>ntificables pel seu espectre dur<br />
en raigs X. Per tant, en un intent d’evitar nuclis actius <strong>de</strong> galàxies i variables<br />
cataclísmiques, només hem seleccionat aquelles fonts que tenien un ín<strong>de</strong>x <strong>de</strong><br />
duresa HR1 + σ(HR1) major que 0.9. Encara que aquest criteri podria evitar<br />
la <strong>de</strong>tecció <strong>de</strong> microquàsars en erupció, cal notar que aquesta situació dura<br />
un temps molt curt en la vida d’aquests objectes, i que la probabilitat <strong>de</strong><br />
<strong>de</strong>scartar-ne algun per aquest motiu és molt baixa. Un total <strong>de</strong> 241 fonts <strong>de</strong>l<br />
RBSC romanen en la nostra mostra en aquesta fase <strong>de</strong>l procés.<br />
4. Les posicions en raigs X i en ràdio han <strong>de</strong> concordar dins <strong>de</strong>ls errors. Així,<br />
hem seleccionat les fonts NVSS dins <strong>de</strong> caixes d’error 2σ (probabilitat <strong>de</strong>l<br />
95%) <strong>de</strong> les posicions <strong>de</strong> les fonts RBSC. També hem utilitzat la restricció<br />
d’una diferència màxima <strong>de</strong> 40 ′′ entre les posicions RBSC i NVSS, ja que per<br />
distàncies majors, no s’espera cap i<strong>de</strong>ntificació creïble.<br />
5. No s’ha permès continuar en el procés <strong>de</strong> selecció cap font ràdio resolta, ja que<br />
qualsevol binària <strong>de</strong> raigs X amb emissió ràdio hauria d’aparèixer compacta<br />
vista amb la resolució <strong>de</strong>l NVSS 1 .<br />
Les fonts així selecciona<strong>de</strong>s, un total <strong>de</strong> 35, han estat filtra<strong>de</strong>s amb informació<br />
complementària en l’òptic, utilitzant els següents criteris:<br />
1. Hem inspeccionat les bases <strong>de</strong> da<strong>de</strong>s SIMBAD i NASA/IPAC Extragalactic<br />
Database (NED) per cada font, i hem exclòs aquelles llista<strong>de</strong>s com a objectes<br />
extragalàctics.<br />
2. També hem inspeccionat les imatges <strong>de</strong>l Digitized Sky Survey, DSS1 i DSS2-<br />
red (la primera versió i les plaques en filtre R <strong>de</strong> la segona versió), i hem<br />
buscat possibles contraparti<strong>de</strong>s òptiques compatibles amb les posicions <strong>de</strong> les<br />
fonts NVSS. Els candidats pels quals hem trobat un objecte òptic extès, és a<br />
dir, amb aparença <strong>de</strong> galàxia, també han estat exclosos <strong>de</strong> la mostra.<br />
1 Les fonts resoltes únicament en un <strong>de</strong>ls dos eixos i amb una mida angular inferior a la resolució<br />
<strong>de</strong> l’altre eix no han estat excloses.
Resum <strong>de</strong> la tesi xxiii<br />
Al final d’aquest procés, hem obtingut un total <strong>de</strong> 17 fonts emissores en raigs<br />
X i en ràdio. Entre elles, hem trobat la ràdio binària <strong>de</strong> raigs X LS I +61 303 i<br />
els microquàsars àmpliament coneguts LS 5039, SS 433 i Cyg X-3.<br />
És interessant<br />
comentar que totes aquestes fonts són binàries <strong>de</strong> raigs X <strong>de</strong> massa alta, i que cap<br />
<strong>de</strong>ls microquàsars coneguts amb companyes <strong>de</strong> massa baixa no ha estat recuperat<br />
per aquest mèto<strong>de</strong>. Això és fàcilment explicable pel fet que la majoria <strong>de</strong> ràdio<br />
binàries X <strong>de</strong> massa baixa són fonts transitòries, i que les persistents són massa<br />
dèbils com per ser presents en el NVSS. A més, hem recuperat tots els microquàsars<br />
amb companyes massives presents en el rang <strong>de</strong> latituds galàctiques explorat, és a<br />
dir, LS 5039, SS 433 and Cyg X-3, excepte Cyg X-1. La no <strong>de</strong>tecció <strong>de</strong> Cyg X-1 pel<br />
nostre mèto<strong>de</strong> és <strong>de</strong>guda a la baixa <strong>de</strong>nsitat <strong>de</strong> flux d’aquesta font en l’època en que<br />
es va realitzar el NVSS. Tot això fa que tinguem confiança en el mèto<strong>de</strong> utilitzat.<br />
Un cop excloses les fonts ja conegu<strong>de</strong>s, hem separat les 13 fonts restants en dos<br />
grups. En el primer grup, format per 8 fonts prioritàries, les diferències <strong>de</strong> posició<br />
entre el RBSC i el NVSS són inferiors a la incertesa 1σ <strong>de</strong> la posició RBSC. En el<br />
segon grup, format per 5 fonts, aquestes diferències estan en el rang 1–2σ.<br />
3.2 Observacions òptiques i ràdio<br />
Després <strong>de</strong> realitzar la i<strong>de</strong>ntificació creuada <strong>de</strong> catàlegs presentada en l’apartat an-<br />
terior, el següent pas lògic era obtenir posicions ràdio acura<strong>de</strong>s per tal <strong>de</strong> buscar<br />
possibles contraparti<strong>de</strong>s òptiques. Amb aquesta finalitat, vam realitzar observacions<br />
ràdio amb el VLA en configuració A en tres èpoques diferents, separa<strong>de</strong>s una set-<br />
mana entre elles, el juliol <strong>de</strong> 1999, i a dues longituds d’ona (3.6 i 6 cm) per tal<br />
d’obtenir informació espectral. Les fonts a observar eren les 8 prioritàries, tot i que<br />
<strong>de</strong>gut a condicions <strong>de</strong> visibilitat, només vam po<strong>de</strong>r observar les 6 primeres, és a<br />
dir, 1RXS J001442.2+580201, 1RXS J013106.4+612035, 1RXS J042201.0+485610,<br />
1RXS J062148.1+174736, 1RXS J072259.5−073131 i 1RXS J072418.3−071508. Per<br />
totes elles vam obtenir posicions ràdio acura<strong>de</strong>s fins a 0.01 ′′ . D’altra banda, totes<br />
les fonts apareixien puntuals excepte 1RXS J072259.5−073131, que mostrava indicis<br />
d’un jet cap a l’est a escales <strong>de</strong> segon d’arc.<br />
En el domini òptic, vam realitzar observacions fotomètriques <strong>de</strong>l 24 al 30 <strong>de</strong><br />
novembre <strong>de</strong> 1998 amb el telescopi <strong>de</strong> 1.5 m <strong>de</strong> l’OAN, i observacions astromètriques<br />
l’onze <strong>de</strong> <strong>de</strong>sembre <strong>de</strong> 1999 amb el telescopi <strong>de</strong> 2.2 m <strong>de</strong>l Centro Astronómico His-
xxiv Resum <strong>de</strong> la tesi<br />
pano Alemán (CAHA). Com a resultat d’aquestes observacions, vam obtenir posi-<br />
cions acura<strong>de</strong>s, amb errors a 1σ <strong>de</strong> 0.3 ′′ , i fotometria en els filtres V , R i I <strong>de</strong><br />
Johnson, per cadascun <strong>de</strong>ls objectes.<br />
En la Fig. 2 mostrem les imatges obtingu<strong>de</strong>s amb el telescopi <strong>de</strong> 2.2 m <strong>de</strong>l<br />
CAHA amb el filtre I <strong>de</strong> Johnson i les imatges ràdio resultants <strong>de</strong> concatenar totes<br />
les observacions a 3.6 cm <strong>de</strong> longitud d’ona per cadascun <strong>de</strong>ls objectes. Com es<br />
pot observar, vam po<strong>de</strong>r trobar contraparti<strong>de</strong>s òptiques <strong>de</strong> totes les fonts ràdio. De<br />
fet, una anàlisi estadística mostra que la probabilitat d’associació per coincidència<br />
aleatòria es inferior al 0.3% en tots els casos.<br />
Després d’un estudi exhaustiu <strong>de</strong> les da<strong>de</strong>s obtingu<strong>de</strong>s, concloem que els dos<br />
primers objectes, és a dir, 1RXS J001442.2+580201 i 1RXS J013106.4+612035, són<br />
bons candidats a binàries <strong>de</strong> raigs X amb emissió ràdio. D’altra banda, l’objecte<br />
1RXS J042201.0+485610 no és tan bon candidat, ja que presenta un ín<strong>de</strong>x espec-<br />
tral <strong>de</strong> fins α = +1.6 a freqüències radio eleva<strong>de</strong>s, fet insòlit en una ràdio binària<br />
X, i a més a més, la seva contrapartida òptica és lleugerament estesa (15% més<br />
que estrelles <strong>de</strong> la mateixa magnitud en la imatge CAHA), la qual cosa suggereix<br />
un possible origen extragalàctic. Un estudi acurat <strong>de</strong> la contrapartida òptica <strong>de</strong><br />
1RXS J062148.1+174736, mostra que és un 30% més estesa que estrelles <strong>de</strong> magni-<br />
tud similar en la mateixa imatge, apuntant cap a un origen extragalàctic per aquest<br />
objecte. La situació no és gens clara en el cas <strong>de</strong> 1RXS J072259.5−073131, perquè<br />
les da<strong>de</strong>s òptiques són compatibles amb una natura <strong>de</strong> ràdio binària X, mentre que<br />
el jet ràdio cap a un sol costat a escales <strong>de</strong> segon d’arc (que no es mostra en la<br />
Fig. 2) és més comú en objectes extragalàctics. Finalment, la darrera font observa-<br />
da, 1RXS J072418.3−071508, és un quàsar i<strong>de</strong>ntificat amb posterioritat a la nostra<br />
i<strong>de</strong>ntificació creuada <strong>de</strong> catàlegs. Així doncs, <strong>de</strong>sprés d’aquest estudi, dos objectes<br />
són bons candidats a ràdio binàries <strong>de</strong> raigs X. Tots aquests resultats, juntament<br />
amb els <strong>de</strong> l’apartat 3.1, han estat publicats a Pare<strong>de</strong>s et al. (2002b).<br />
El següent pas lògic era l’obtenció d’espectres òptiques d’aquestes fonts, per tal<br />
<strong>de</strong> <strong>de</strong>svetllar la seva natura galàctica o extragalàctica. Es van dur a terme diversos<br />
intents durant els anys 2000 i 2001, però les condicions meteorològiques durant les<br />
observacions van impedir l’obtenció <strong>de</strong> resultats. Un nou intent està programat<br />
per la tardor <strong>de</strong> 2002. D’altra banda, observacions VLBI d’aquestes sis fonts seran<br />
publica<strong>de</strong>s properament (Ribó et al. 2002b), i es presenten en el següent apartat.
Resum <strong>de</strong> la tesi xxv<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
1RXS J001442.2+580201 I band<br />
58 03 00<br />
02 45<br />
30<br />
15<br />
00<br />
01 45<br />
30<br />
15<br />
58 02 02.0<br />
00 14 48 46 44 42 40 38 36<br />
RIGHT ASCENSION (J2000)<br />
01.5<br />
01.0<br />
00.5<br />
VLA 3.6cm<br />
00 14 42.25 42.20 42.15 42.10 42.05 42.00<br />
RIGHT ASCENSION (J2000)<br />
17 48 30<br />
15<br />
00<br />
47 45<br />
30<br />
15<br />
00<br />
46 45<br />
17 47 36.0<br />
35.5<br />
35.0<br />
34.5<br />
1RXS J062148.1+174736 I band<br />
06 21 51 50 49 48 47 46 45 44<br />
RIGHT ASCENSION (J2000)<br />
VLA 3.6cm<br />
06 21 47.82 47.80 47.78 47.76 47.74 47.72 47.70 47.68<br />
RIGHT ASCENSION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
1RXS J013106.4+612035 I band<br />
61 21 30<br />
15<br />
00<br />
20 45<br />
30<br />
15<br />
00<br />
19 45<br />
61 20 34.0<br />
01 31 14 12 10 08 06 04 02 00<br />
RIGHT ASCENSION (J2000)<br />
33.5<br />
33.0<br />
32.5<br />
-07 30 45<br />
VLA 3.6cm<br />
01 31 07.35 07.30 07.25 07.20 07.15 07.10<br />
RIGHT ASCENSION (J2000)<br />
31 00<br />
15<br />
30<br />
45<br />
32 00<br />
15<br />
30<br />
-07 31 34.0<br />
1RXS J072259.5-073131 I band<br />
07 23 03 02 01 00 22 59 58 57 56<br />
RIGHT ASCENSION (J2000)<br />
34.5<br />
35.0<br />
35.5<br />
VLA 3.6cm<br />
07 22 59.74 59.72 59.70 59.68 59.66 59.64 59.62<br />
RIGHT ASCENSION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
48 57 00<br />
56 45<br />
30<br />
15<br />
00<br />
55 45<br />
30<br />
15<br />
1RXS J042201.0+485610 I band<br />
04 22 06 04 02 00 21 58 56<br />
RIGHT ASCENSION (J2000)<br />
48 56 04.5<br />
04.0<br />
03.5<br />
03.0<br />
-07 14 30<br />
45<br />
15 00<br />
15<br />
30<br />
45<br />
16 00<br />
15<br />
VLA 3.6cm<br />
04 22 00.60 00.55 00.50 00.45<br />
RIGHT ASCENSION (J2000)<br />
1RXS J072418.3-071508 I band<br />
07 24 21 20 19 18 17 16 15 14<br />
RIGHT ASCENSION (J2000)<br />
-07 15 19.5<br />
20.0<br />
20.5<br />
21.0<br />
VLA 3.6cm<br />
07 24 17.36 17.34 17.32 17.30 17.28 17.26 17.24 17.22<br />
RIGHT ASCENSION (J2000)<br />
Figura 2: Imatges òptiques en filtre I i imatges ràdio a 3.6 cm, agrupa<strong>de</strong>s per parelles,<br />
<strong>de</strong> les sis fonts estudia<strong>de</strong>s. Els cercles en les imatges òptiques, que no són caixes d’error,<br />
indiquen les contraparti<strong>de</strong>s òptiques <strong>de</strong> les fonts ràdio. Les creus en les imatges ràdio,<br />
indiquen els errors 1σ <strong>de</strong> les posicions <strong>de</strong> les contraparti<strong>de</strong>s òptiques.
xxvi Resum <strong>de</strong> la tesi<br />
3.3 Observacions amb EVN i MERLIN<br />
Per tal <strong>de</strong> <strong>de</strong>tectar possibles estructures en forma <strong>de</strong> jets ràdio a escales angulars<br />
<strong>de</strong> mil·lèssimes <strong>de</strong> segon d’arc, vam dur a terme observacions VLBI <strong>de</strong> les sis fonts<br />
estudia<strong>de</strong>s anteriorment. Les observacions es van dur a terme amb la EVN i MER-<br />
LIN, <strong>de</strong>l 29 <strong>de</strong> febrer <strong>de</strong> 2000 a l’1 <strong>de</strong> març <strong>de</strong> 2000 (23:30–23:00 <strong>de</strong> temps universal)<br />
a una freqüència <strong>de</strong> 5 GHz. També es van dur a terme mesures <strong>de</strong> la <strong>de</strong>nsitat <strong>de</strong><br />
flux <strong>de</strong> les fonts amb l’antena <strong>de</strong> 100 metres <strong>de</strong> diàmetre que el Max Planck Institut<br />
für Radioastronomie té a Effelsberg, Alemanya.<br />
Les observacions es van realitzar utilitzant la tècnica <strong>de</strong> referència <strong>de</strong> fase, im-<br />
prescindible per po<strong>de</strong>r <strong>de</strong>tectar les fonts més febles. Es van <strong>de</strong>tectar 5 <strong>de</strong> les 6<br />
fonts estudia<strong>de</strong>s (totes excepte 1RXS J042201.0+485610) i es van obtenir imatges<br />
amb MERLIN, amb la EVN i combinant els dos dispositius d’antenes. Aquestes<br />
imatges es mostren en la Fig. 3, on en les imatges EVN+MERLIN hi ha indicats els<br />
coeficients <strong>de</strong> “disminució” (tapering) utilitzats en aquests casos.<br />
Tal i com es pot veure, la primera font, 1RXS J001442.2+580201, mostra un jet<br />
bipolar que emana d’una font central en la imatge EVN, i on hi ha indicats els noms<br />
<strong>de</strong> les components nord i sud, numera<strong>de</strong>s segons l’ordre en que van ser ejecta<strong>de</strong>s <strong>de</strong>s<br />
<strong>de</strong> l’objecte compacte. Una anàlisi <strong>de</strong>tallada en el marc <strong>de</strong> la relativitat especial<br />
mostra que el jet te una velocitat superior a 0.2c, i forma un angle inferior a 78 ◦<br />
respecte la visual.<br />
La segona font, 1RXS J013106.4+612035, apareix bàsicament compacta, tot i<br />
que hi ha indicis d’un jet relativista en un únic sentit, amb velocitat superior a 0.3c,<br />
i un angle inferior a 72 ◦ respecte la visual, tal i com es pot <strong>de</strong>duir <strong>de</strong> la imatge EVN.<br />
La tercera font, 1RXS J042201.0+485610, no va ser <strong>de</strong>tectada, per problemes<br />
amb la referència <strong>de</strong> fase i <strong>de</strong>gut a que és una font molt dèbil.<br />
La quarta font, 1RXS J062148.1+174736, apareix compacta en totes les imatges<br />
(els allargaments visibles són <strong>de</strong>guts a la convolució amb el feix sintetitzat, que es<br />
mostra en la part inferior esquerra <strong>de</strong> cada imatge).<br />
La cinquena font, 1RXS J072259.5−073131, presenta un jet en un únic sentit que<br />
implica una velocitat superior a 0.3c i un angle respecte la visual inferior a 73 ◦ . És<br />
remarcable el torcement <strong>de</strong>l jet a mida que s’allunya <strong>de</strong> la font central, una propietat
Resum <strong>de</strong> la tesi xxvii<br />
1RXS J001442.2+580201<br />
1RXS J013106.4+612035<br />
1RXS J062148.1+174736<br />
1RXS J072259.5-073131<br />
1RXS J072418.3-071508<br />
Figura 3: Imatges a 5 GHz <strong>de</strong> les cinc fonts <strong>de</strong>tecta<strong>de</strong>s, obtingu<strong>de</strong>s amb MERLIN,<br />
MERLIN+EVN i amb la EVN. Les unitats <strong>de</strong>ls eixos són mil·lèssimes <strong>de</strong> segon d’arc.<br />
N1<br />
N2<br />
S2<br />
S1
xxviii Resum <strong>de</strong> la tesi<br />
habitual en objectes extragalàctics com els blàzars.<br />
Finalment, el quàsar 1RXS J072418.3−071508, mostra també un jet tort i en un<br />
únic sentit, que implica una velocitat superior a 0.5c i un angle respecte la visual<br />
inferior a 60 ◦ .<br />
3.4 Sumari<br />
Després d’una anàlisi <strong>de</strong>tallada <strong>de</strong>ls resultats obtinguts, concloem que les dues<br />
primeres fonts <strong>de</strong> la mostra, 1RXS J001442.2+580201 i 1RXS J013106.4+612035,<br />
són bons candidats a microquàsar. La tercera font, 1RXS J042201.0+485610, no ho<br />
és <strong>de</strong>gut al seu espectre ràdio invertit a altes freqüències, que suggereix un origen<br />
tèrmic per aquesta font, que a més a més apareix lleugerament estesa en l’òptic i<br />
no es <strong>de</strong>tecta en observacions VLBI. La quarta font, 1RXS J062148.1+174736, és<br />
probablement una galàxia, ja que la contrapartida òptica és estesa. La cinquena<br />
font, 1RXS J072259.5−073131, és probablement extragalàctica <strong>de</strong> tipus blàzar, ja<br />
que el torcement <strong>de</strong>l seu jet ràdio és una propietat comuna en aquests objectes. Fi-<br />
nalment, la sisena font, 1RXS J072418.3−071508, és un quàsar ja i<strong>de</strong>ntificat. Tots<br />
aquests resultats, juntament amb els <strong>de</strong> l’apartat 3.3, han estat publicats a Ribó<br />
et al. (2002b).<br />
Cap d’aquestes sis fonts apareix en el tercer catàleg EGRET, <strong>de</strong> fonts <strong>de</strong> raigs<br />
gamma d’alta energia. D’altra banda, recor<strong>de</strong>m que un nou intent per tal d’obtenir<br />
espectres òptics, que revelaran la natura galàctica o extragalàctica <strong>de</strong>ls objectes, es<br />
durà a terme la tardor <strong>de</strong> 2002.<br />
Finalment les dues fonts prioritàries <strong>de</strong> la i<strong>de</strong>ntificació creuada que no han estat<br />
posteriorment observa<strong>de</strong>s, així com les fonts no prioritàries, s’estudiaran possible-<br />
ment en el futur.<br />
4. Conclusions generals<br />
Aquí resumim breument les conclusions a les quals hem arribat <strong>de</strong>sprés <strong>de</strong> realitzar<br />
aquest treball.
Resum <strong>de</strong> la tesi xxix<br />
Apartat 2:<br />
1. Després <strong>de</strong> realitzar una acurada inspecció <strong>de</strong> mo<strong>de</strong>rnes bases <strong>de</strong> da<strong>de</strong>s i dur<br />
a terme observacions ràdio interferomètriques, hem <strong>de</strong>scobert el microquàsar<br />
LS 5039. Per tant, concloem que estudis similars po<strong>de</strong>n revelar una població<br />
anteriorment inadvertida <strong>de</strong> microquàsars “silenciosos”, que fins ara no han<br />
presentat cap explosió que ens fes saber <strong>de</strong> la seva existència, com LS 5039.<br />
Si això és correcte, el fenomen <strong>de</strong>ls microquàsars podria ser molt més habitual<br />
<strong>de</strong>l que fins ara s’havia consi<strong>de</strong>rat.<br />
2. Hem realitzat un estudi en profunditat <strong>de</strong>l comportament multi-longitud d’ona<br />
<strong>de</strong> LS 5039, cobrint <strong>de</strong>s <strong>de</strong> longituds d’ona ràdio fins a raigs gamma d’alta<br />
energia. Per tal d’arribar a aquest fi, hem dut a terme observacions amb<br />
l’objectiu d’estudiar el flux, la variabilitat d’aquest i l’espectre en longituds<br />
d’ona ràdio, òptiques i en raigs X. En el domini ràdio hem obtingut imatges<br />
interferomètriques en el continuu i mapes en línies espectrals realitzats amb<br />
antena única. Totes aquestes da<strong>de</strong>s, juntament amb observacions en raigs<br />
gamma presents en la literatura, ens han permès proposar un escenari i un<br />
mo<strong>de</strong>l per tal d’explicar la fenomenologia observada en LS 5039. Els estudis<br />
multi-longitud d’ona són claus per entendre els processos d’acreció/ejecció <strong>de</strong><br />
matèria al voltant d’objectes compactes.<br />
3. Hem proposat una associació entre LS 5039 i una <strong>de</strong> les ∼ 170 fonts EGRET<br />
no i<strong>de</strong>ntifica<strong>de</strong>s, associació que sembla plausible <strong>de</strong>gut a l’emissió persistent<br />
i mo<strong>de</strong>radament variable <strong>de</strong> la font tant en ràdio com en raigs gamma. Si<br />
es confirma, aquesta seria la primera associació entre un microquàsar i una<br />
font EGRET, i suggerim que altres fonts <strong>de</strong> raigs gamma d’alta energia no<br />
i<strong>de</strong>ntifica<strong>de</strong>s podrien ser microquàsars “silenciosos” encara no <strong>de</strong>scoberts.<br />
4. Hem <strong>de</strong>scobert que LS 5039 és un microquàsar en fuga, amb una velocitat<br />
sistèmica <strong>de</strong> vsis 150 km s −1 , escapant <strong>de</strong>l seu entorn local amb una elevada<br />
component <strong>de</strong> la velocitat perpendicular al pla galàctic. Hem estat capaços<br />
d’explicar tant la velocitat espacial elevada com l’excentricitat observa<strong>de</strong>s en<br />
un escenari d’explosió <strong>de</strong> supernova simètrica, amb una important pèrdua <strong>de</strong><br />
massa <strong>de</strong> ∆ M ∼ 17 M⊙.<br />
5. Finalment, la velocitat espacial elevada i el possible temps <strong>de</strong> vida <strong>de</strong>l micro-<br />
quàsar indiquen que aquest podria assolir una latitud galàctica <strong>de</strong> b = −12 ◦ .
xxx Resum <strong>de</strong> la tesi<br />
Per tant, si l’associació entre LS 5039 i la font EGRET 3EG J1824−1514<br />
és correcta, podríem <strong>de</strong>tectar microquàsars <strong>de</strong> raigs gamma fins a latituds<br />
galàctiques <strong>de</strong> l’ordre <strong>de</strong> |b| 10 ◦ . Així, els microquàsars en fuga podrien es-<br />
tar relacionats amb algunes <strong>de</strong> les fonts EGRET febles, variables i relativament<br />
toves encara no i<strong>de</strong>ntifica<strong>de</strong>s, situa<strong>de</strong>s per sobre i per sota <strong>de</strong>l pla galàctic.<br />
Apartat 3:<br />
1. Hem <strong>de</strong>senvolupat un mèto<strong>de</strong> d’i<strong>de</strong>ntificació creuada per buscar noves ràdio<br />
binàries X, que son microquàsars en potència. Els resultats obtinguts donen<br />
confiança en el mèto<strong>de</strong>, ja que la llista <strong>de</strong> fonts selecciona<strong>de</strong>s inlcoïa tots els<br />
microquàsars amb companyes massives en la regió |b| < 5 ◦ excepte un. La<br />
mostra final, un cop excloses les fonts conegu<strong>de</strong>s, conté 13 noves ràdio fonts<br />
X, <strong>de</strong> les quals 8 han estat classifica<strong>de</strong>s com a objectes prioritaris.<br />
2. Hem estudiat 6 d’aquests 8 objectes, i hem trobat contraparti<strong>de</strong>s òptiques per<br />
tots ells. Hem obtingut posicions tant en ràdio com en òptic, perfectament<br />
compatibles entre elles. També hem obtingut espectres ràdio i magnituds<br />
òptiques <strong>de</strong> les fonts.<br />
3. Hem presentat observacions realita<strong>de</strong>s amb la EVN i MERLIN d’aquestes sis<br />
fonts. Cinc d’elles han estat <strong>de</strong>tecta<strong>de</strong>s i hem obtingut imatges, que mostren<br />
diferents morfologies: una font té un jet bipolar, tres fonts tenen un jet en un<br />
únic sentit i una font és compacta.<br />
4. Després d’una anàlisi <strong>de</strong>tallada <strong>de</strong> les observacions, concloem que les fonts<br />
1RXS J001442.2+580201 i 1RXS J013106.4+612035 són bons candidats a mi-<br />
croquàsar. 1RXS J042201.0+485610 no ho és <strong>de</strong>gut al seu espectre ràdio in-<br />
vertit a altes freqüències, que suggereix un origen tèrmic per aquesta font, que<br />
a més a més apareix lleugerament estesa en l’òptic. 1RXS J062148.1+174736<br />
és probablement una galàxia, ja que la contrapartida òptica és estesa. La font<br />
1RXS J072259.5−073131 és probablement extragalàctica <strong>de</strong> tipus blàzar, ja<br />
que el torcement <strong>de</strong>l seu jet ràdio és una propietat comuna en aquests ob-<br />
jectes. Finalment, la font 1RXS J072418.3−071508 és un quàsar ja i<strong>de</strong>ntificat.<br />
5. Treball en curs. Observacions espectroscòpiques en l’òptic <strong>de</strong> les cinc primeres<br />
fonts la tardor <strong>de</strong> 2002, per tal <strong>de</strong> <strong>de</strong>svetllar la seva natura galàctica o extra-<br />
galàctica. D’altra banda, la resta <strong>de</strong> fonts <strong>de</strong> la mostra s’estudiaran possible-<br />
ment en el futur.
Bibliografia<br />
Casares, J., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, en preparació<br />
Clark, J. S., Reig, P., Goodwin, S. P., et al. 2001, A&A, 376, 476<br />
Martí, J., Pare<strong>de</strong>s, J. M., & Ribó, M. 1998, A&A, 338, L71<br />
McSwain, M. V., & Gies, D. R. 2002, ApJ, 568, L27<br />
McSwain, M. V., Gies, D. R., Riddle, R. L., Wang, Z., & Wingert, D. W. 2001,<br />
ApJ, 558, L43<br />
Motch, C., Haberl, F., Dennerl, K., Pakull, M., & Janot-Pacheco, E. 1997, A&A,<br />
323, 853<br />
Motch, C., Guillout, P., Haberl, F., et al. 1998, A&AS, 132, 341<br />
Pare<strong>de</strong>s, J. M., Martí, J., Ribó, M., & Massi, M. 2000, Science, 288, 2340<br />
Pare<strong>de</strong>s, J. M., Ribó, M., Ros, E., Martí, J., & Massi, M. 2002a, A&A, 393, L99<br />
Pare<strong>de</strong>s, J. M., Ribó, M., & Martí, J. 2002b, A&A, 394, 193<br />
Reig, P., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, en preparació<br />
Ribó, M., Reig, P., Martí, J., & Pare<strong>de</strong>s, J. M. 1999, A&A, 347, 518<br />
Ribó, M., Pare<strong>de</strong>s, J. M., Romero, G. E., et al. 2002a, A&A, 384, 954<br />
Ribó, M., Ros, E., Pare<strong>de</strong>s, J. M., Massi, M., & Martí, J. 2002b, A&A, 394, 983<br />
xxxi
xxxii BIBLIOGRAFIA
Chapter 1<br />
Introduction and background<br />
Microquasars are radio emitting X-ray binaries displaying relativistic radio jets. In<br />
this chapter we will first explain briefly the accretion mechanisms in X-ray binaries,<br />
as well as perform an approach to the catalogued systems, in Sect. 1.1. We will then<br />
focus on the basic properties of microquasars in Sect. 1.2. We will finish by stating<br />
what was the motivation of this work, the search for new microquasars, in Sect. 1.3.<br />
1.1 X-ray binaries<br />
Since X-rays cannot propagate through the atmosphere and cannot reach the Earth<br />
surface, astronomers had to wait until the 1960s to discover the first extrasolar X-<br />
ray source, namely Sco X-1, thanks to observations with <strong>de</strong>tectors onboard rockets.<br />
However, the real take off of X-ray astronomy can be consi<strong>de</strong>red to be in 1970, when<br />
the UHURU (freedom in Swahili) satellite was launched. It provi<strong>de</strong>d the first all-sky<br />
X-ray map down to a flux of 1 mCrab, which contained 339 sources. It also allowed<br />
the <strong>de</strong>tection of X-ray pulsations and X-ray eclipses in some of these sources. The<br />
UHURU observations allowed to find the optical counterparts of such objects, that<br />
were called X-ray binaries:<br />
• An X-ray binary is a binary system containing a compact object, either a<br />
Neutron Star (NS) or a stellar-mass Black Hole (BH), accreting matter from<br />
the companion star.<br />
1
2 Chapter 1. Introduction and background<br />
Several scenarios have been proposed to explain the X-ray emission of X-ray bina-<br />
ries, <strong>de</strong>pending on the nature of the compact object, its magnetic field in the case<br />
of neutron stars and on the geometry of the accretion flow. In any case, the ac-<br />
creted matter is accelerated to relativistic speeds, transforming its potential energy<br />
provi<strong>de</strong>d by the intense gravitational field of the compact object into kinetic energy.<br />
Assuming that this kinetic energy is finally radiated, we can compute the accretion<br />
luminosity as:<br />
Laccr 1<br />
2 ˙<br />
MaccrV 2 =<br />
GMX ˙<br />
Maccr<br />
RX<br />
, (1.1)<br />
where ˙<br />
Maccr is the accretion rate, V is the free fall speed <strong>de</strong>fined as V = 2GMX/RX,<br />
G is the gravitational constant and MX and RX are the mass and radius of the com-<br />
pact object, respectively. On the other hand, there is a maximum theoretical value<br />
for the accretion rate, when the radiation pressure balances gravity, called the Ed-<br />
dington limit and expressed as:<br />
˙<br />
MEdd = 4πmpcRX<br />
σT<br />
, (1.2)<br />
where mp is the proton mass, c is the speed of light and σT is the Thomson cross<br />
section. The corresponding luminosity can be expressed as:<br />
LEdd = 4πGMXmpc<br />
σT<br />
. (1.3)<br />
In an X-ray binary, the accreted matter carries angular momentum, and on its<br />
way to the compact object it usually forms an accretion disk around it. The matter<br />
in the disk looses angular momentum due to viscous dissipation, which produces a<br />
heating of the disk, and falls towards the compact object in a spiral trajectory. The<br />
black body temperature of the last stable orbit in the case of a BH accreting at the<br />
Eddington limit is given by:<br />
<br />
T<br />
2 × 10<br />
K<br />
7<br />
−1/4 MX<br />
M⊙<br />
. (1.4)<br />
For a compact object of a few solar masses the obtained temperature is ∼ 10 7 K.<br />
At this temperature the energy will be mainly radiated in the X-ray domain of the<br />
electromagnetic spectrum.<br />
Using Eq. 1.1 we can see that for an accreting BH, taking as final radius the<br />
Schwarzschild radius <strong>de</strong>fined as RS = 2GMX/c 2 , the emitted accretion luminosity<br />
is:<br />
Laccr 1<br />
2 ˙<br />
Maccrc 2 . (1.5)
1.1. X-ray binaries 3<br />
According to this result, a significant fraction of the rest mass of the accreted matter<br />
can be transformed into radiation. If we consi<strong>de</strong>r this fraction to be 1/10 instead<br />
of the quoted 1/2 value in Eq. 1.5, and use a typical luminosity of 10 37 erg s −1 for<br />
X-ray binaries, we obtain ˙<br />
Maccr 2 × 10 −9 M⊙ yr −1 . This mass would be given by<br />
the companion or donor star. If instead of an accreting BH we consi<strong>de</strong>r a 2M⊙ NS<br />
with a 10 km radius we obtain a slightly larger value of ˙<br />
Maccr 6 × 10 −9 M⊙ yr −1 .<br />
In any case, it is clear that a donor star mass loss rate greater than 10 −10 M⊙ yr −1 ,<br />
available and sustainable over long periods of time, can easily explain the observed<br />
X-ray luminosities in X-ray binaries.<br />
If the compact object is a NS with a strong magnetic field, B ∼ 10 12 G, the<br />
accretion disk will be disrupted at several thousand kilometers from the compact<br />
object, and matter will follow the magnetic field lines until impacting onto the<br />
magnetic poles, producing again X-ray emission. If there is a misalignment between<br />
the rotation and magnetic axes, X-ray pulsations will be observed if the beamed<br />
emission from the magnetic poles rotates through the line of sight. These particular<br />
X-ray binaries are called X-ray pulsars.<br />
On the contrary, if the compact object is a weak magnetic field NS, B < 10 10 G,<br />
the accretion disk may reach it or come close to it. X-rays will be then produced in<br />
the inner accretion disk and the boundary layer between the disk and the compact<br />
object.<br />
The number of known X-ray binaries has increased consi<strong>de</strong>rably in the last years.<br />
There were 63 classified systems in 1983, and 188 in 1995. The current number of<br />
catalogued X-rays binaries is ∼ 280 (Liu et al. 2000, 2001). According to the mass<br />
of the donor, X-ray binaries are divi<strong>de</strong>d in High Mass X-ray Binaries (HMXBs) and<br />
Low Mass X-ray Binaries (LMXBs).<br />
1.1.1 High mass X-ray binaries<br />
In HMXBs the donor star is an early type star, with an O or B spectral type.<br />
HMXBs are conventionally divi<strong>de</strong>d into two subgroups: systems containing a B star<br />
with emission lines, namely Be stars, and systems containing a supergiant (SG) O<br />
or B star. Here we list some basic typical properties:
4 Chapter 1. Introduction and background<br />
• Be stars: B or late O stars with luminosity classes III–V with emission lines,<br />
indicative of an equatorial <strong>de</strong>cretion disk. Masses in the range ∼ 8–20 M⊙.<br />
Orbital periods in the range ∼ 10–250 days, with eccentricities 0.3. Do not<br />
fill their Roche lobe, and accretion onto the compact object is produced via<br />
mass transfer through the <strong>de</strong>cretion disk. Most of these systems are transient<br />
X-ray sources during periastron passage. The compact object is typically a<br />
strong magnetic field neutron star, and most of these systems are X-ray pulsars.<br />
The lifetime of Be/X-ray binaries is around ∼ 10 7 years.<br />
• OB SG stars: O or B stars with luminosity classes I–III. Masses above 15 M⊙.<br />
Orbital periods typically below 10 days, with eccentricities 0. The mass<br />
transfer is due to a strong stellar wind and/or Roche lobe overflow, with mass<br />
loss rates up to 6 × 10 −6 M⊙ yr −1 . The X-ray emission is persistent, and<br />
large variability is usual. The lifetime of SG/X-ray binaries is between ∼ 10 5 –<br />
10 6 years.<br />
The typical X-ray to optical luminosity ratio is in the range LX/Lopt 10 −3 –10,<br />
although highly absorbed systems in the optical have higher values.<br />
The most recent catalog of HMXBs was compiled by Liu et al. (2000), and<br />
contains 130 sources. We have ad<strong>de</strong>d to this catalog a source previously classified<br />
as a LMXB. We show in Fig. 1.1 the distribution of catalogued HMXBs in galactic<br />
coordinates. As can be seen, these systems are located close to the galactic plane (in<br />
or near spiral arms) because are young stellar systems of Population I. The presence<br />
of HMXBs in the Small and Large Magellanic Clouds (SMC and LMC) is also clearly<br />
seen.<br />
In Table 1.1 we list the number of different types of HMXBs within the SMC,<br />
the LMC, the Galaxy and total values. In the Be and SG types we have consi<strong>de</strong>red<br />
all the confirmed cases and also the suspected ones based on X-ray properties or<br />
orbital solutions. NF stands for those sources that do not fit in the Be or SG types.<br />
Un<strong>de</strong>r the question mark we list the number of sources without a clear i<strong>de</strong>ntification<br />
or with a suspected white dwarf (WD) compact object. All of them amount to 131<br />
systems, of which some are X-ray Pulsars (XPs) and/or transient sources, as quoted<br />
also in the table. Finally, the number of Radio Emitting X-ray Binaries (REXBs)<br />
has also been inclu<strong>de</strong>d (see Sect. 1.1.3).
1.1. X-ray binaries 5<br />
Galactic latitu<strong>de</strong> [<strong>de</strong>g]<br />
90<br />
60<br />
30<br />
0<br />
−30<br />
−60<br />
High Mass X−ray Binaries<br />
−90<br />
180 120 60 0 300 240 180<br />
Galactic longitu<strong>de</strong> [<strong>de</strong>g]<br />
Figure 1.1: Distribution of catalogued HMXBs in galactic coordinates. SMC and LMC<br />
stand for Small and Large Magellanic Clouds, respectively. The dashed line represents<br />
the galactic plane.<br />
SMC<br />
LMC<br />
Table 1.1: Types and location of catalogued HMXBs.<br />
Location Be SG NF ? Total XPs Transients REXBs<br />
SMC 23 1 0 1 25 16 17 0<br />
LMC 12 2 3 3 20 6 8 0<br />
Galaxy 42 15 3 26 86 49 36 8<br />
Total 77 18 6 30 131 71 61 8
6 Chapter 1. Introduction and background<br />
The galactic latitu<strong>de</strong> distribution of the 86 galactic HMXBs has a mean of −0.5 ◦ ,<br />
and a standard <strong>de</strong>viation of 4.6 ◦ . In fact, a total of 80 sources are located within<br />
|b| < 10 ◦ , 78 of them in the |b| < 5 ◦ area. Hence, around 93% and 91% of the galactic<br />
HMXBs are located within the previously quoted regions, respectively. Galactic<br />
sources with |b| > 10 ◦ are nearby systems (d 1 kpc), which have in fact normal<br />
distances to the galactic plane when compared to the other ones.<br />
An inspection of Table 1.1 reveals at first sight that among the 60 i<strong>de</strong>ntified<br />
HMXBs in the Galaxy the most common type is the Be/X-ray binary (∼ 70%),<br />
followed by the SG/X-ray binary type (∼ 25%). Another remarkable fact is the<br />
large number of XPs: ∼ 54% in total and ∼ 57% in the Galaxy.<br />
1.1.2 Low mass X-ray binaries<br />
In LMXBs the donor has a spectral type later than B, and has a mass ≤ 2 M⊙.<br />
Although typically is a non-<strong>de</strong>generated star, there are some examples where the<br />
donor is a WD. The orbital periods are in the range 0.2–400 hours, although the<br />
typical value is < 24 hours. The orbits are circular, and mass transfer is due to<br />
Roche lobe overflow. Most of LMXBs are transients, probably as a result of an<br />
instability in the accretion disk or a mass ejection episo<strong>de</strong> from the companion. The<br />
lifetime of LMXBs is estimated to be ∼ 10 8 –10 9 years. The typical ratio between<br />
X-ray to optical luminosity is in the range LX/Lopt 100–1000, and the optical<br />
light is dominated by X-ray heating of the accretion disk and the companion star.<br />
A consi<strong>de</strong>rable fraction of LMXBs display X-ray bursts, which are due to ther-<br />
monuclear flashes on the surface of accreting NSs. These sources are usually called<br />
bursters. On the other hand, some LMXBs are classified as ‘Z’ and ‘Atoll’ sources,<br />
according to the pattern traced out in the X-ray color-color diagram. The first ones<br />
are thought to be weak magnetic field neutron stars of the or<strong>de</strong>r of 10 10 G with<br />
accretion rates around 0.5–1.0 ˙<br />
MEdd, while the second ones are believed to have<br />
even weaker magnetic fields of 10 8 G and lower accretion rates of 0.01–0.1 ˙<br />
MEdd.<br />
The most recent catalog of LMXBs was compiled by Liu et al. (2001), and<br />
contains 150 sources. We have removed from this catalog a source recently classified<br />
as a HMXB. We show in Fig. 1.2 the distribution of catalogued LMXBs in galactic<br />
coordinates. As can be seen, and contrary to the distribution of HMXBs shown in
1.1. X-ray binaries 7<br />
Galactic latitu<strong>de</strong> [<strong>de</strong>g]<br />
90<br />
60<br />
30<br />
0<br />
−30<br />
−60<br />
Low Mass X−ray Binaries<br />
−90<br />
180 120 60 0 300 240 180<br />
Galactic longitu<strong>de</strong> [<strong>de</strong>g]<br />
Figure 1.2: Distribution of catalogued LMXBs in galactic coordinates. LMC stands for<br />
Large Magellanic Cloud. The dashed line represents the galactic plane.<br />
Fig. 1.1, these systems do not trace the galactic disk, and are located in the galactic<br />
bulge and in globular clusters at relatively high galactic latitu<strong>de</strong>s, because are old<br />
stellar systems of Population II. Only two LMXBs are catalogued in the LMC, and<br />
none in the SMC.<br />
In Table 1.2 we list the number of different types of LMXBs within the SMC,<br />
the LMC, the Galaxy and total values. There are only 5 XPs, i.e., ∼ 3% of the<br />
149 catalogued LMXBs, compared to the ∼ 54% in HMXBs. This fact tells us that<br />
most of LMXBs are low magnetic field NSs or BHs. Two of these XPs are also X-ray<br />
bursters. There are 63 X-ray bursters, all of them containing NS compact objects.<br />
A total of 6 Z sources have been found in the Galaxy, while 1 is suspected in the<br />
LMC (omitted here). There are 18 Atoll sources, of which 16 are X-ray bursters,<br />
LMC
8 Chapter 1. Introduction and background<br />
Table 1.2: Types and location of catalogued LMXBs.<br />
Location Total XPs Bursters Z sources Atoll sources Transients REXBs<br />
SMC 0 0 0 0 0 0 0<br />
LMC 2 0 0 0 0 0 0<br />
Galaxy 147 5 63 6 18 75 35<br />
Total 149 5 63 6 18 75 35<br />
including 5 X-ray transients. Approximately half of LMXBs are transients. REXBs<br />
will be discussed in Sect. 1.1.3.<br />
The galactic latitu<strong>de</strong> distribution of the 147 galactic LMXBs has a mean of +0.4 ◦ ,<br />
and a standard <strong>de</strong>viation of 12.2 ◦ , to be compared with the 4.6 ◦ value obtained<br />
for HMXBs. It is clear the higher scatter in galactic latitu<strong>de</strong>, compatible with<br />
Population II sources. In fact, 13 LMXBs are located in globular clusters.<br />
1.1.3 Radio emitting X-ray binaries<br />
The first X-ray binary known to display radio emission was Sco X-1 in the late 1960’s.<br />
Since then, many X-ray binaries have been <strong>de</strong>tected at radio wavelengths with flux<br />
<strong>de</strong>nsities ≥ 0.1–1 mJy. Moreover, the flux <strong>de</strong>nsities <strong>de</strong>tected at cm radio wavelengths<br />
are produced in small angular scales, and cannot be produced by thermal emission<br />
mechanisms.<br />
In this context, the most efficient known mechanism for production of intense<br />
radio emission from astronomical sources is the synchrotron emission mechanism, in<br />
which highly relativistic electrons interacting with magnetic fields produce intense<br />
radio emission which tends to be linearly polarized. The observed radio emission can<br />
be explained by assuming a spatial distribution of non-thermal relativistic electrons,<br />
usually with a power-law energy distribution, interacting with magnetic fields.<br />
Since some REXBs, like SS 433, were found to display elongated or jet-like<br />
features, like in AGNs and quasars, it was proposed that flows of relativistic elec-<br />
trons were ejected perpendicular to the accretion disk, and were responsible for<br />
synchrotron radio emission in the presence of a magnetic field. Several mo<strong>de</strong>ls have<br />
been proposed for the formation and collimation of the jets, including the presence
1.1. X-ray binaries 9<br />
of an accretion disk close to the compact object, a magnetic field in the accretion<br />
disk, a high spin for the compact object, etc. However there is no clear agreement<br />
on what mechanism is exactly at work.<br />
Among high mass REXBs we find very different types of donors, such as Be,<br />
B[e], SG and Wolf Rayet stars, as well as systems that do not fit in the standard<br />
categories. Among low mass REXBs we find all six Z sources, five Atoll sources, two<br />
persistent sources near the Galactic center and a large number of transient sources.<br />
As quoted in Tables 1.1 and 1.2, there are around 8 radio emitting HMXBs and<br />
35 radio emitting LMXBs. The corresponding abundances are ∼ 6% and ∼ 23%,<br />
respectively. There is a significant difference between those values. However, the<br />
sources located in the Large and Small Magellanic Clouds, at ∼ 50 kpc and ∼ 58 kpc,<br />
respectively, are around 5 times farther than the most distant galactic REXBs,<br />
and the flux <strong>de</strong>nsity of those sources in quiescence is expected to be ∼ 25 times<br />
lower, preventing any radio <strong>de</strong>tection up to now. If we exclu<strong>de</strong> these sources, then<br />
the abundances change into ∼ 9% and ∼ 24%, respectively. Moreover, since the<br />
strong magnetic field X-ray pulsars disrupt the accretion disk at several thousand<br />
kilometers from the neutron star, there is no inner accretion disk to launch a jet<br />
in these systems, and no synchrotron radio emission has ever been <strong>de</strong>tected in any<br />
of those sources. Hence, consi<strong>de</strong>ring only the galactic HMXBs and LMXBs and<br />
excluding XPs, the previous numbers change into ∼ 22% and ∼ 24%, respectively,<br />
which are in<strong>de</strong>ed quite similar. As pointed out by several authors and as we have<br />
seen, although the division of X-ray binaries in HMXBs and LMXBs is useful for<br />
the study of binary evolution, it is probably not important for the study of the<br />
radio emission in these systems, where the only important aspect seems to be the<br />
presence of an inner accretion disk capable of producing radio jets. However, the 8<br />
radio emitting HMXBs inclu<strong>de</strong> 6 persistent sources and 2 transient sources, while<br />
among the 35 radio emitting LMXBs we find 11 persistent sources and 24 transient<br />
sources. It is clear that the difference between the persistent and transient behavior<br />
highly <strong>de</strong>pends on the mass of the donor.<br />
In any case we can say that, excluding X-ray pulsars, 20 to 25% of the catalogued<br />
galactic X-ray binaries have been <strong>de</strong>tected at radio wavelengths, regardless of the<br />
nature of the donor. The corresponding ratio of <strong>de</strong>tected/observed sources is proba-<br />
bly much higher. However, it is difficult to give reliable numbers, since observational<br />
constrains arise when consi<strong>de</strong>ring transient sources observed in the past (large X-ray
10 Chapter 1. Introduction and background<br />
error boxes, single dish and/or poor sensitivity radio observations, etc.), and we do<br />
not know how many non-<strong>de</strong>tections have not been published.<br />
1.2 Microquasars<br />
A microquasar is a radio emitting X-ray binary displaying relativistic radio jets. The<br />
name was given not only because of the observed morphological similarities between<br />
these sources and the distant quasars but also because of physical similarities, since<br />
when the compact object is a black hole, some parameters scale with the mass of<br />
the central object. A schematic illustration comparing some parameters in quasars<br />
and microquasars is shown in Fig. 1.3.<br />
In this context, we can see that the use of Eq. 1.4 leads to temperatures of<br />
∼ 10 7 K in the case of microquasars containing stellar-mass black holes and ∼ 10 5 K<br />
in the case of quasars containing supermassive black holes (10 7 –10 9 M⊙). This<br />
explains why in microquasars the accretion luminosity is radiated in X-rays, while<br />
it is done in the optical/UV domain in the case of quasars.<br />
On the other hand, the characteristic jet sizes seem to be proportional to the<br />
mass of the black hole, since radio jets in microquasars have typical sizes of the<br />
or<strong>de</strong>r of light years, while radio jets in quasars reach distances up to several million<br />
light years in giant radio galaxies.<br />
Last, but not least, the timescales appear to be directly scaled with the mass<br />
of the black hole following τ RS/c = 2GMX/c 3 ∝ M. Therefore, phenomena<br />
that take place in timescales of years in quasars can be studied in minutes in micro-<br />
quasars.<br />
In this sense, one can say that microquasars mimic, on smaller scales, many of<br />
the phenomena seen in AGNs and quasars, but allow a better and faster progress to<br />
un<strong>de</strong>rstand the accretion/ejection processes that take place near compact objects.<br />
The number of currently known microquasars is ∼ 16, among the 43 catalogued<br />
REXBs. It is interesting to note that some authors have proposed that all REXBs<br />
are microquasars, and would be <strong>de</strong>tected as such provi<strong>de</strong>d that there is enough<br />
sensitivity and/or resolution in the radio observations.
1.2. Microquasars 11<br />
Figure 1.3: The quasar-microquasar analogy (from Mirabel & Rodríguez 1998).
12 Chapter 1. Introduction and background<br />
1.2.1 Projection and special relativity effects<br />
The <strong>de</strong>tected radio emission in the jets of microquasars is usually in the form of<br />
blobs, i.e., plasma clouds (or plasmons) emitting synchrotron radiation, which prop-<br />
agate through space at relativistic bulk velocities. Therefore, projection and special<br />
relativity effects have to be taken into account when analyzing the observations.<br />
υ<br />
¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤ ¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
υ t cosθ<br />
¢<br />
¡ ¢ £ £<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤ ¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦<br />
υ t<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤ ¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦ ¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦¥¦<br />
¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤¥¤<br />
θ<br />
towards the observer<br />
υ t sinθ<br />
Figure 1.4: Scheme of an ejection of plasma clouds in opposite senses.<br />
Projection effects: jet parameters and superluminal motion<br />
Let us consi<strong>de</strong>r a pair of relativistic plasma clouds ejected simultaneously from a<br />
central source and moving at intrinsic velocity v in opposite senses, as shown in<br />
Fig. 1.4, in a direction forming an angle θ with respect to the line of sight. If<br />
we obtain several images at different epochs, we will measure the following proper<br />
motions for the approaching and receding components, respectively:<br />
υ<br />
β sin θ c<br />
µa =<br />
, (1.6)<br />
(1 − β cos θ) D<br />
β sin θ c<br />
µr =<br />
, (1.7)<br />
(1 + β cos θ) D
1.2. Microquasars 13<br />
where β is the velocity of the clouds in units of the speed of light (β = v/c) and<br />
D is the distance between the observer and the source. Hence, the approaching<br />
component will have an apparent proper motion, projected in the plane of the sky,<br />
higher than the receding one (e.g., see Fig. 1.5). This pair of equations can be<br />
directly transformed into:<br />
β cos θ = µa − µr<br />
µa + µr<br />
, (1.8)<br />
c tan θ (µa − µr)<br />
D = .<br />
2 µaµr<br />
(1.9)<br />
It is interesting to note that if the proper motions are measured, we can directly<br />
compute the product β cos θ. Moreover, since both variables, β and cos θ, take values<br />
between 0 and 1, it is clear that knowing β cos θ allows us to obtain a lower limit for<br />
the velocity and an upper limit for the angle. We can also compute an upper limit<br />
for the distance to the source given by:<br />
D ≤<br />
c<br />
√ µaµr<br />
. (1.10)<br />
On the other hand, if we have a single image and we do not know the epoch<br />
of the ejection, we can neither measure nor infer the proper motions. However, if<br />
we <strong>de</strong>tect the central source, we can cancel the time variable by using the relative<br />
distances to it, da and dr, and transform Eq. 1.8 into:<br />
β cos θ = µa − µr<br />
µa + µr<br />
= da − dr<br />
da + dr<br />
. (1.11)<br />
As in Eq. 1.8, we can obtain a lower limit for the velocity and an upper limit for the<br />
angle, but in this case no upper limit for the distance to the source can be obtained.<br />
Finally, using Eq. 1.6 we can directly solve for the apparent velocity of the<br />
approaching component (va = µaD):<br />
va =<br />
β sin θ<br />
(1 − β cos θ)<br />
c . (1.12)<br />
We can see that if β cos θ approaches 1, the <strong>de</strong>tected apparent velocity will be higher<br />
than the speed of light. In fact, this will always happen if the condition<br />
β ><br />
1<br />
sin θ + cos θ<br />
(1.13)<br />
is satisfied. The minimum β to <strong>de</strong>tect the so-called superluminal motions is ∼ 0.71<br />
for an angle of 45 ◦ . Higher values of β are required for different angles, according<br />
to Eq. 1.13.
14 Chapter 1. Introduction and background<br />
E<br />
N<br />
1"<br />
MAR 27<br />
APR 03<br />
APR 09<br />
APR 16<br />
APR 23<br />
APR 30<br />
Figure 1.5: VLA multiepoch observations at 3.6 cm wavelength of the superluminal<br />
ejection of GRS 1915+105 in 1994. The approaching component, to the left, appeared<br />
to move at 1.25c and be brighter than the receding one, to the right, which moved with<br />
an apparent velocity of 0.65c (from Mirabel & Rodríguez 1994).
1.2. Microquasars 15<br />
Special relativity effects: Doppler shift and boosting<br />
Taking into account that the plasma clouds move at relativistic speeds, there will<br />
be a shift in the observed to emitted wavelengths. The ratios of the observed (νa<br />
and νr) to emitted frequency (ν0) are given by the Doppler factors:<br />
δa = νa<br />
ν0<br />
δr = νr<br />
ν0<br />
=<br />
=<br />
1<br />
Γ(1 − β cos θ)<br />
1<br />
Γ(1 + β cos θ)<br />
, (1.14)<br />
, (1.15)<br />
where Γ is the bulk Lorentz factor given by Γ = (1 − β 2 ) −1/2 . δr is always smaller<br />
than 1, while δa is smaller than 1 for large values of θ and rises above 1 for small<br />
values of θ, i.e., when the jet is pointing to us. Hence, an eventually observed<br />
receding spectral line will always be red-shifted, while the blue or red-shift of the<br />
approaching one will <strong>de</strong>pend on the ejection angle. If for a particular event we were<br />
able to measure a frequency shift in a spectral line produced in the plasmons, and<br />
also compute β cos θ from Eq. 1.8, we could then find β, and hence θ and D by<br />
using Eq. 1.8 and Eq. 1.9, respectively. This would allow us to solve for the whole<br />
parameters of an ejection.<br />
On the other hand, the measured flux <strong>de</strong>nsities of the approaching and receding<br />
clouds will also suffer from special relativity effects. The ratios of the observed (Sa<br />
and Sr) to emitted flux <strong>de</strong>nsity (S0) from a twin pair of optically thin, isotropically<br />
emitting jets are given by:<br />
Sa<br />
S0<br />
Sr<br />
S0<br />
= δ k−α<br />
a , (1.16)<br />
= δ k−α<br />
r , (1.17)<br />
where k equals 2 for a continuous jet and 3 for discrete con<strong>de</strong>nsations, and α is the<br />
spectral in<strong>de</strong>x of the radio emission (Sν ∝ ν +α ). Typical values for α in the optically<br />
thin regime are between −0.5 and −1. Hence, the exponent in the equations above<br />
will typically be in the range 2.5–4. As a consequence, the <strong>de</strong>tected flux <strong>de</strong>nsity of<br />
the receding plasma cloud will be diminished (<strong>de</strong>boosting), while for the approaching<br />
one will <strong>de</strong>pend on the angle, as the Doppler factor does, i.e., we will have <strong>de</strong>boosting<br />
for large angles and consi<strong>de</strong>rable boosting for small angles. Using these equations,<br />
the ratio of approaching to receding flux <strong>de</strong>nsities measured at equal distance from<br />
the core is:<br />
Sa<br />
Sr<br />
=<br />
k−α 1 + β cos θ<br />
1 − β cos θ<br />
. (1.18)
16 Chapter 1. Introduction and background<br />
This ratio is always greater than 1, implying that we will <strong>de</strong>tect the approaching<br />
component brighter than the receding one (e.g., see Fig. 1.5). Solving for β cos θ in<br />
Eq. 1.18 we obtain:<br />
β cos θ = (Sa/Sr) 1/(k−α) − 1<br />
(Sa/Sr) 1/(k−α) + 1<br />
. (1.19)<br />
As we have seen, at any given time da > dr. Therefore, if we only have a single<br />
image of an ejection event, the receding component will be closer to the core than<br />
the approaching one. Since the flux <strong>de</strong>nsity <strong>de</strong>creases with increasing distance from<br />
the core (due to adiabatic losses, and assuming no interaction with the interstellar<br />
medium), the measured ratio Sa/Sr will be lower than the one that should be used<br />
in the equation above, which is only valid at equal distances from the core. As a<br />
consequence, in such a case Eq. 1.19 only allows us to obtain a lower limit for β cos θ.<br />
Finally, it may also happen that we have a single image where only the approach-<br />
ing component is <strong>de</strong>tected, either because the relativistic <strong>de</strong>boosting of the receding<br />
one is very strong or because of lack of sensitivity. In any case, we can use the fact<br />
that we do not <strong>de</strong>tect the counter-jet, and replace Sr with the 3σ level of the image<br />
in Eq. 1.19. This will only provi<strong>de</strong> a lower limit to β cos θ, expressed as<br />
β cos θ > (Sa/3σ) 1/(k−α) − 1<br />
(Sa/3σ) 1/(k−α) + 1<br />
1.2.2 Quasars and microquasars<br />
. (1.20)<br />
After the introduction on accretion and the former explanation about special rel-<br />
ativity effects, we can comment on some specific differences between quasars and<br />
microquasars.<br />
In the distant quasars radio emission was <strong>de</strong>tected in the early times of radioas-<br />
tronomy, because the jets in the <strong>de</strong>tected quasars have small angles with respect to<br />
the line of sight, and the flux <strong>de</strong>nsities of the approaching components are signifi-<br />
cantly Doppler boosted. Hence, these objects appear in the sky as very bright radio<br />
sources, showing one-si<strong>de</strong>d jets in high resolution radio interferometric observations,<br />
which usually reveal superluminal motions. The probability of having a quasar with<br />
a jet pointing close to our line of sight is small, but provi<strong>de</strong>d enough number of<br />
quasars, these objects are easy to <strong>de</strong>tect, because their emission is persistent. As-<br />
tronomers searched for the optical counterparts, and found objects that looked like
1.2. Microquasars 17<br />
stars but without stellar-like spectra. Therefore, these objects were called ’quasi<br />
stellar radio sources’ (quasars). In fact, the <strong>de</strong>tected optical emission has its origin<br />
in an un<strong>de</strong>rlying galaxy, which can be clearly imaged for the closer sources. This<br />
galaxy supplies matter for the accretion process taking place in the central super-<br />
massive black hole of 10 7 –10 9 M⊙, responsible for launching the jet that we <strong>de</strong>tect<br />
as a quasar. As commented above, the inner part of the accretion disk radiates in<br />
the optical/UV domains, and quasars do not show strong X-ray emission. Typical<br />
luminosities are around 10 47 erg s −1 , implying accretion rates of ∼ 10 M⊙ yr −1 .<br />
If a quasar has a jet pointing almost directly to us (within a few <strong>de</strong>grees) it is<br />
called blazar (after the object BL Lac). Blazars show extreme variability, bending<br />
of the jets (due to projection effects) and strongly polarized radio emission.<br />
If a quasar has a jet pointing away from our line of sight the radio emission will<br />
be <strong>de</strong>boosted. Hence, if it is a distant object it will not be <strong>de</strong>tected. However, if<br />
it is relatively close (∼ 100 Mpc <strong>de</strong>pending on the resolution and sensitivity of the<br />
observations), it may be that both the approaching and receding jet components are<br />
<strong>de</strong>tected.<br />
Since microquasars are nearby sources there is no need of extreme Doppler boost-<br />
ing in or<strong>de</strong>r to <strong>de</strong>tect them at radio wavelengths. Hence, it is not necessary to have<br />
ejections with angles close to the line of sight, and most of times both the approach-<br />
ing and receding components are <strong>de</strong>tected. The <strong>de</strong>tected radio emission is most of<br />
times transient, and at optical wavelengths are not surprising objects. Hence, as-<br />
tronomers had to wait until the launch of X-ray satellites, capable of <strong>de</strong>tecting the<br />
persistent/transient strong and usually hard X-ray emission coming from accretion<br />
disks in microquasars, to discover the galactic relatives of the distant quasars.<br />
1.2.3 Known microquasars<br />
The first microquasar to be discovered was SS 433, although it was consi<strong>de</strong>red as a<br />
very strange object for several years. In early 1990’s, two microquasars were discov-<br />
ered near the galactic center, and in 1994 it was discovered the first superluminal<br />
source in the Galaxy, namely GRS 1915+105. The current census of microquasars<br />
contains around 16 sources, <strong>de</strong>pending on the specific properties one imposes to<br />
consi<strong>de</strong>r a source as a microquasar. In Table 1.3 we list some of the interesting prop-
18 Chapter 1. Introduction and background<br />
erties of the sources we consi<strong>de</strong>r as known microquasars in the Galaxy, including<br />
the position, spectral type of the donor and nature of the compact object, distance<br />
to the system, orbital period, optical magnitu<strong>de</strong> in V or other filters, mass of the<br />
compact object, X-ray and radio luminosities (typical for the persistent sources and<br />
peak values for the transient sources), the activity at radio wavelengths (persistent<br />
or transient), apparent and intrinsic velocity, inclination angle of the <strong>de</strong>tected ejec-<br />
tion(s) and jet size. The objects are grouped in HMXBs and LMXBs and or<strong>de</strong>red in<br />
increasing right ascension. In some objects only hints of relativistic jets have been<br />
seen or are clearly suspected.<br />
Several public databases, such as SIMBAD, GBI/NASA monitoring program,<br />
ASM onboard RXTE have been consulted to compile the information present in<br />
Table 1.3, as well as the catalogs by Liu et al. (2000, 2001) and lots of references<br />
therein. This compilation is not complete nor exhaustive, and is only provi<strong>de</strong>d to<br />
give an i<strong>de</strong>a about the observed properties in microquasars.<br />
The table inclu<strong>de</strong>s well known high mass REXBs: LS I +61 303, that displays<br />
radio outbursts every 26.5 days; SS 433, with their precessing radio jets moving<br />
at 0.26c; Cygnus X-1, the first binary where evi<strong>de</strong>nces of the existence of a BH<br />
were found, and recently classified as a microquasar; and Cygnus X-3, famous for<br />
their strong radio outbursts reaching flux <strong>de</strong>nsities of several Jy. The other massive<br />
systems are V4641 Sgr, a candidate to be a microblazar source (in an analogy with<br />
the extragalactic blazars) with extreme superluminal motion, and LS 5039, which<br />
will be discussed in Part I. We point out that the only other massive REXBs are<br />
γ Cas, which is a very peculiar system in many aspects, and CI Cam, where radio<br />
emission seems to be produced in a fairly isotropic nebula.<br />
Regarding low mass REXBs in Table 1.3 we find: XTE J1118+480, which con-<br />
tains a black hole and has a Galactic halo orbit; Circinus X-1, which experiences<br />
outbursts every 16.6 days; XTE J1550−564, a system that contains a black hole<br />
and where hints of superluminal motion have been found; Scorpius X-1, an histor-<br />
ical source recently ad<strong>de</strong>d to the group; GRO J1655−40, the second superluminal<br />
source to be discovered in the Galaxy; GX 339−4, where the jet has not clearly been<br />
imaged but strong evi<strong>de</strong>nces of its existence have been found; 1E 1740.7−2942, the<br />
great annihilator near the Galactic center; XTE J1748−288, a superluminal source<br />
where <strong>de</strong>celeration and brightening of jet components due to the interaction with<br />
the interstellar medium have been observed; GRS 1758−258, a persistent jet source
1.2. Microquasars 19<br />
Table 1.3: Microquasars in the Galaxy.<br />
§¨©<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
©<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
©<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
§<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
§<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
§<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
§<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
§
20 Chapter 1. Introduction and background<br />
near the Galactic center; and GRS 1915+105, the first superluminal source to be<br />
discovered in the Galaxy, and recently found to have the most massive black hole<br />
as well as the longest orbital period in the microquasar group.<br />
We can see that up to now, four sources have shown superluminal motions,<br />
namely GRS 1915+105, GRO J1655−40, XTE J1748−288 and V4641 Sgr, while<br />
such phenomenon is also suspected in XTE J1550−564.<br />
We show in Fig. 1.6 the distribution in galactic coordinates of the microquasars<br />
listed in Table 1.3, together with their respective names.<br />
Galactic latitu<strong>de</strong> [<strong>de</strong>g]<br />
90<br />
60<br />
30<br />
0<br />
−30<br />
−60<br />
XTE J1118+480<br />
LS I +61 303<br />
Microquasars in the Galaxy<br />
Cyg X−1<br />
Cyg X−3<br />
GRS 1758−258<br />
GRS 1915+105<br />
SS 433<br />
LS 5039<br />
Sco X−1<br />
V4641 Sgr<br />
XTE J1748−288<br />
GRO J1655−40<br />
Cir X−1<br />
GX 339−4<br />
1E1740.7−2942<br />
XTE J1550−564<br />
LMXB<br />
LMXB (rel. jets ?)<br />
HMXB<br />
HMXB (rel. jets ?)<br />
−90<br />
180 120 60 0 300 240 180<br />
Galactic longitu<strong>de</strong> [<strong>de</strong>g]<br />
Figure 1.6: Distribution of known microquasars in galactic coordinates. Filled circles<br />
represent those sources where relativistic jets have been imaged, while open circles are<br />
used for those where hints of relativistic jets have been seen or are clearly suspected.
1.2. Microquasars 21<br />
1.2.4 Accretion disk and jet formation<br />
It is currently accepted that the formation of relativistic radio jets requires the<br />
presence of a compact object (a potential well) and an accretion disk surrounding<br />
it. The mechanism of jet formation and collimation is not yet well un<strong>de</strong>rstood,<br />
but we can gain knowledge and test proposed mo<strong>de</strong>ls after performing observations<br />
like the ones of GRS 1915+105 shown in Fig. 1.7. We can see that after a large-<br />
amplitu<strong>de</strong> quasi-periodic oscillation, the X-ray flux drops substantially, a behavior<br />
interpreted as the emptying of the inner accretion disk. At the same time, the X-<br />
ray spectrum becomes significantly hard, until when an X-ray spike is seen. Then<br />
the X-ray spectrum softens again and we can see the starting of a flare at infrared<br />
wavelengths, which is interpreted in terms of synchrotron emission by an expanding<br />
plasma cloud, as shown above the lightcurves, ejected just when the X-ray spike<br />
took place. As the plasma cloud expands adiabatically, the maximum of the emitted<br />
energy shifts to longer wavelengths, and a subsequent flare is <strong>de</strong>tected in the radio.<br />
Luminosity [arbitrary units]<br />
1.0<br />
0.5<br />
0.0<br />
X−RAYS<br />
Disk<br />
emptying<br />
280.000 <br />
<br />
<br />
INFRARED<br />
Ejection<br />
RADIO<br />
km/s<br />
7.9 8.0 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9<br />
UT Time [hours]<br />
Figure 1.7: Simultaneous multiwavelength behavior at X-ray, radio and infrared wave-<br />
lengths of GRS 1915+105 during 1997 September 9 (from Mirabel et al. 1998).
22 Chapter 1. Introduction and background<br />
1.2.5 Black hole states and radio emission<br />
Black holes mainly display two different kinds of behavior at X-ray energies. On one<br />
hand, when the total X-ray luminosity is found to be low, they display a spectrum<br />
significantly hard up to a few hundreds of keV. This is the so-called low/hard state.<br />
On the other hand, when they exhibit a high X-ray luminosity, their spectrum is<br />
substantially softer. This is the so-called high/soft state. The lower energy photons,<br />
responsible for the enhancement of X-ray luminosity and steepening of the spectrum<br />
in the high/soft state, are produced by a multicolor blackbody in the inner accretion<br />
disk, while the hard X-ray tail in the low/hard state is thought to be the result of a<br />
(inverse) Comptonizing corona above the disk. Apart from these two states, there<br />
is also the off state, when X-ray emission is almost or totally suppressed, and the<br />
hybrids very high and intermediate states, where both spectral characteristics (soft<br />
disk photons and hard tail) coexist.<br />
The accepted i<strong>de</strong>a was that the accretion rate increased from the off state,<br />
through the low/hard state, the intermediate state up to the high/soft state (and<br />
eventually the very high state). However, this picture is not so clear now, because<br />
the same characteristics of a given spectral state (X-ray spectral and timing proper-<br />
ties) have been observed at very different X-ray luminosities (a traditional indicator<br />
of accretion rate).<br />
An important insight into the possible scenario related to the black hole states<br />
has been provi<strong>de</strong>d by radio observations. In this sense, long-term multiwavelength<br />
campaigns such as the one shown in Fig. 1.8 are very useful. There we can see two<br />
years of monitoring of GX 339−4 at radio, hard X-rays and soft X-rays. In the<br />
upper part of the figure it is marked the X-ray spectral state of the source along<br />
time. Starting in the low/hard state, a transition is observed into the high/soft<br />
state, which lasted around one year, followed by a transition back to the low/hard<br />
state and ending with an off state. The radio flux is clearly correlated with the<br />
hard X-rays in all spectral states. The correlation between the radio flux and the<br />
soft X-rays is more complicated, but can be shown to behave as follows: during<br />
the low/hard state radio emission is correlated with the soft X-ray photons, while<br />
during the high/soft state radio emission is suppressed. Faint radio emission is<br />
<strong>de</strong>tected during the off state, compatible with it being a lower luminosity low/hard<br />
state. It is interesting to note that the spectral in<strong>de</strong>x of the radio emission is flat<br />
all the time, indicative of a compact and partially self-absorbed jet, except when
1.2. Microquasars 23<br />
Low-hard state High-soft state<br />
L-H state Off state<br />
Figure 1.8: Two-year monitoring of the black hole candidate GX 339−4 at radio wave-<br />
lengths (ATCA and MOST), hard X-rays (BATSE, 20–100 keV) and soft X-rays (RXTE<br />
ASM, 1.5–12 keV) (from Corbel et al. 2000).<br />
transitions between spectral states take place. Then, the radio emission is optically<br />
thin, probably as a result of major ejections of relativistic plasma clouds.<br />
After these observational results, it was proposed that compact radio jets may<br />
always be present during the low/hard states (as observed in other sources), while<br />
are suppressed in the high/soft states. The reason is not clear yet, but a coupling<br />
between the radio jet and the hard X-ray emitting corona is firmly established. A<br />
suggested possibility is that the corona is in fact the base of the radio jet.<br />
On the other hand, the synchrotron radiation mechanism, which is at work in<br />
the jet, has been seen up to infrared and possibly optical wavelengths. Moreover,<br />
it has been suggested that it could also generate photons up to X-ray energies in<br />
its optically thin part. If this is correct, then the jets would be extremely powerful,<br />
and certainly any mo<strong>de</strong>l explaining accretion should account for their presence.
24 Chapter 1. Introduction and background<br />
1.3 Motivation of the thesis<br />
As we have seen, the current number of microquasars is ∼ 16. This number could<br />
increase dramatically if, as it has been proposed by Fen<strong>de</strong>r & Hendry (2000), radio<br />
emission from X-ray binaries always arises from relativistic jets. This would imply<br />
that all REXBs, ∼ 40 systems, are microquasars. However, the situation is not<br />
so clear yet and REXBs appear as a rather heterogeneous group including both<br />
high-mass (e.g., Cyg X-3, LS I +61 303) and low-mass companions (e.g., Aql X-1,<br />
GX 339−4). Moreover, in the latter group, radio emission has been <strong>de</strong>tected in all<br />
six Z sources and only in a few Atoll sources. In addition, the <strong>de</strong>tection of radio<br />
emission does not seem to be limited to a particular type of compact object, since<br />
there are REXBs harboring black holes (e.g., GRO J1655−40, Cyg X-1) and neutron<br />
stars (e.g., Aql X-1, Sco X-1). Moreover, when one begins to further divi<strong>de</strong> the group<br />
of REXBs according to their optical or compact companion or whether radio jets are<br />
present or not, the number of systems is no longer enough to perform a meaningful<br />
statistical study. The situation is even worse when we focus on microquasars.<br />
There is an important question to be answered: what is the matter content of<br />
the jets? Up to now, atomic emission lines have only been observed in the jets<br />
of SS 433, indicating that the jets are baryonic. However, some of the proposed<br />
mo<strong>de</strong>ls can only account for leptonic (e + e − ) jets. Therefore, it is crucial to unveil<br />
the matter content if we want to correctly mo<strong>de</strong>l the accretion/ejection processes<br />
that take place near compact objects.<br />
Another interesting issue is the possibility of microquasars being sources of high<br />
energy γ-rays. Before this work, the 26.5 d periodic REXB LS I +61 303 was<br />
the only X-ray binary that could be associated with a high-energy γ-ray source<br />
(E > 100 MeV) of the third EGRET catalog (Hartman et al. 1999), namely<br />
3EG J0241+6103 (Kniffen et al. 1997). The search for new REXBs could allow<br />
new findings. In<strong>de</strong>ed, the current most promising candidate is LS 5039 (see Part I),<br />
which appears to be associated with the source 3EG J1824−1514 (Pare<strong>de</strong>s et al.<br />
2000). Although these associations will only be confirmed, or discar<strong>de</strong>d, by future<br />
missions such as GLAST, clearly a sample of two sources is a very poor one, and any<br />
attempt to discover new REXBs/microquasars, and look for possible high energy<br />
γ-ray emission, is warranted.<br />
Hence, two reasons gave us a strong motivation to start a long-term project
1.3. Motivation of the thesis 25<br />
focused on the search for new microquasars in the Galaxy:<br />
1. Every new microquasar has allowed new interesting phenomenological findings,<br />
which have to be taken into account in the proposed mo<strong>de</strong>ls to explain the<br />
accretion/ejection processes taking place near compact objects.<br />
2. To enhance the reduced sample and allow meaningful statistical studies in the<br />
future.<br />
In Part I of this thesis we present the discovery and subsequent study of the<br />
microquasar LS 5039. This was the first source to be discovered within the most<br />
global approach presented in Part II. There we explain how we have used the best<br />
available X-ray and radio catalogs to search for new REXB/microquasar candidates.<br />
We state our general conclusions in Part III.
26 Chapter 1. Introduction and background
Bibliography<br />
Blandford, R. D., & Königl, A. 1979, ApJ, 232, 34<br />
Blandford, R. D., & Payne, D. G. 1982, MNRAS, 199, 883<br />
Blandford, R. D., & Znajek, R. L. 1977, MNRAS, 179, 433<br />
Corbel, S., Fen<strong>de</strong>r, R. P., Tzioumis, A. K., et al. 2000, A&A, 359, 251<br />
Drake, S. A., A Brief History of High-Energy Astronomy,<br />
http://heasarc.gsfc.nasa.gov/docs/heasarc/headates/heahistory.html<br />
Fen<strong>de</strong>r, R. P. 2001, MNRAS, 322, 31<br />
Fen<strong>de</strong>r, R. P. 2002, in Proc. of Relativistic Flows in Astrophysics, ed. A. W. Guth-<br />
mann, M. Georganopoulos, A. Marcowith & K. Manolakou, Lecture Notes in<br />
Physics, 589, 101<br />
Fen<strong>de</strong>r, R. P., & Hendry, M. A. 2000, MNRAS, 317, 1<br />
Fen<strong>de</strong>r, R. P., & Kuulkers, E. 2001, MNRAS, 324, 923<br />
Fen<strong>de</strong>r, R. P., Southwell, K., & Tzioumis, A. K. 1998, MNRAS, 298, 692<br />
Fomalont, E. B., Geldzahler, B. J., & Bradshaw, C. F. 2001, ApJ, 558, 283<br />
Hartman, R. C., Bertsch, D. L., Bloom, S. D., et al. 1999, ApJS, 123, 79<br />
Hjellming, R. M., & Han, X. 1995, Radio properties of X-ray binaries, in X-ray Bi-<br />
naries, ed. W. H. G. Lewin, J. van Paradijs, & E. P. J. van <strong>de</strong>n Heuvel (Cambridge<br />
Univ. Press, Cambridge), 308<br />
Koi<strong>de</strong>, S., Shibata, K., Kudoh, T., & Meier, D. L. 2002, Science, 295, 1688<br />
27
28 BIBLIOGRAPHY<br />
Lewin, W. H. G., van Paradijs, J., & Taam, R. E. 1995, X-ray bursts, in X-ray Bina-<br />
ries, ed. W. H. G. Lewin, J. van Paradijs, & E. P. J. van <strong>de</strong>n Heuvel (Cambridge<br />
Univ. Press, Cambridge), 175<br />
Liu, Q. Z., van Paradijs, J., & van <strong>de</strong>n Heuvel, E. P. J. 2000, A&AS, 147, 25<br />
Liu, Q. Z., van Paradijs, J., & van <strong>de</strong>n Heuvel, E. P. J. 2001, A&A, 368, 1021<br />
Livio, M. 1999, Phys. Rep., 311, 225<br />
Markoff, S., Falcke, H., & Fen<strong>de</strong>r, R. 2001, A&A, 372, L25<br />
Martí, J., Pare<strong>de</strong>s, J. M., & Peracaula, M. 2001, A&A, 375, 476<br />
Meier, D. L., Koi<strong>de</strong>, S., & Uchida, Y. 2001, Science, 291, 84<br />
Mirabel, I. F., & Rodríguez, L. F. 1994, Nature, 371, 46<br />
Mirabel, I. F., & Rodríguez, L. F. 1998, Nature, 392, 673<br />
Mirabel, I. F., & Rodríguez, L. F. 1999, ARA&A, 37, 409<br />
Mirabel, I. F., Dhawan, V., Chaty, S., et al. 1998, A&A, 330, L9<br />
Mirabel, I. F., Rodríguez, L. F., Cordier, B., Paul, J., & Lebrun, F. 1992, Nature,<br />
358, 215<br />
Negueruela, I., The X-ray binaries page,<br />
http://www.astro.soton.ac.uk/∼bexmgr/xbp.html<br />
Pare<strong>de</strong>s, J. M., Martí, J., Ribó, M., & Massi, M. 2000, Science, 288, 2340<br />
Rees, M. J. 1966, Nature, 211, 468<br />
Rees, M. J. 1984, ARA&A, 22, 471<br />
Shakura, N. I., & Sunyaev, R. A. 1973, A&A, 24, 337<br />
Stirling, A. M., Spencer, R. E., <strong>de</strong> la Force, C. J., et al. 2001, MNRAS, 327, 1273<br />
Tanaka, Y., & Lewin, W. H. G. 1995, Black-hole binaries, in X-ray Binaries, ed.<br />
W. H. G. Lewin, J. van Paradijs, & E. P. J. van <strong>de</strong>n Heuvel (Cambridge Univ.<br />
Press, Cambridge), 126
BIBLIOGRAPHY 29<br />
van Paradijs, J., & McClintock, J. E. 1995, Optical and ultraviolet observations of<br />
X-ray binaries, in X-ray Binaries, ed. W. H. G. Lewin, J. van Paradijs, & E. P. J.<br />
van <strong>de</strong>n Heuvel (Cambridge Univ. Press, Cambridge), 58<br />
Verbunt, F., & van <strong>de</strong>n Heuvel, E. P. J. 1995, Formation and evolution of neutron<br />
stars and black holes in binaries, in X-ray Binaries, ed. W. H. G. Lewin, J. van<br />
Paradijs, & E. P. J. van <strong>de</strong>n Heuvel (Cambridge Univ. Press, Cambridge), 457<br />
White, N. E., X-ray binaries,<br />
http://lheawww.gsfc.nasa.gov/users/white/xrb/xrb.html<br />
White, N. E., Nagase, F., & Parmar, A. N. 1995, The properties of X-ray binaries,<br />
in X-ray Binaries, ed. W. H. G. Lewin, J. van Paradijs, & E. P. J. van <strong>de</strong>n Heuvel<br />
(Cambridge Univ. Press, Cambridge), 1
30 BIBLIOGRAPHY
Part I<br />
Discovery and study of the<br />
microquasar LS 5039<br />
31
Chapter 2<br />
Multiwavelength approach to<br />
LS 5039<br />
2.1 Introduction<br />
After our discovery of the microquasar nature of LS 5039 in year 2000, several<br />
groups have studied this source at different wavelengths. As a consequence, some<br />
parameters of the system have been unveiled since then. Therefore, although we<br />
could have reproduced all our findings 1,2,3,4,5,6,7 in a chronological way, we have<br />
preferred to provi<strong>de</strong> here a multiwavelength approach to the source using the new<br />
available information to reanalyze some of the data.<br />
Nevertheless, we think it is worth to start giving a chronology of findings on<br />
LS 5039, which is briefly done in Sect. 2.2, to provi<strong>de</strong> the rea<strong>de</strong>r with a general view.<br />
We then perform the multiwavelength approach in Sects. 2.4–2.7, that contains our<br />
discovery of the LS 5039 microquasar nature, the research done by us and comments<br />
on what has been done by other groups. We propose a scenario to explain the<br />
observational data in Sect. 2.8 and we finally state our conclusions in Sect. 2.9.<br />
1 Published in Martí, J., Pare<strong>de</strong>s, J. M. & Ribó, M. 1998, A&A, 338, L71.<br />
2 Published in Ribó, M., Reig, P., Martí, J., & Pare<strong>de</strong>s, J. M. 1999, A&A, 347, 518.<br />
3 Published in Pare<strong>de</strong>s, J. M., Martí, J., Ribó, M., & Massi, M. 2000, Science, 288, 2340.<br />
4 Published in Ribó, M., Pare<strong>de</strong>s, J. M., Romero, G. E., et al. 2002, A&A, 384, 954.<br />
5 Published in Pare<strong>de</strong>s, J. M., Ribó, M., Ros, E., Martí, J., & Massi, M. 2002, A&A, 393, L99.<br />
6 Casares, J., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, in preparation.<br />
7 Reig, P., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, in preparation.<br />
33
34 Chapter 2. Multiwavelength approach to LS 5039<br />
2.2 Chronology of findings on LS 5039<br />
The V = 11.2 star LS 5039, located at an estimated distance of ∼ 3 kpc and close<br />
to the galactic plane (l = 16.88 ◦ , b = −1.29 ◦ ), was initially proposed by Motch et al.<br />
(1997) to be the optical counterpart of the X-ray source RX J1826.2−1450, a likely<br />
High Mass X-ray Binary (HMXB).<br />
Using data from the NRAO VLA Sky Survey and follow-up VLA observations,<br />
Martí et al. (1998) discovered that the source was also a non-thermal radio emitter<br />
with mo<strong>de</strong>rate variability.<br />
X-ray observations of RX J1826.2−1450 carried out by Ribó et al. (1999) showed<br />
that the X-ray spectrum was significantly hard (Γ = 1.95) and that emission was<br />
<strong>de</strong>tected up to 30 keV, with a strong Gaussian iron line at 6.6 keV, and neither<br />
pulsed nor periodic emission was found on timescales of 0.02–2000 s and 2–200 days,<br />
respectively. These authors also found that LS 5039 appeared to be a persistent non-<br />
thermal radio emitter based on the GBI-NASA Monitoring Program observations.<br />
Pare<strong>de</strong>s et al. (2000) discovered that the system displays relativistic radio jets<br />
(β > 0.15) at milliarcsecond scales thanks to VLBA observations, revealing the<br />
microquasar nature of LS 5039, and proposed an association with the high-energy<br />
γ-ray source 3EG J1824−1514.<br />
While in the past the mass donor had been classified as an O7V((f)) star, optical<br />
and near-IR spectroscopic observations by Clark et al. (2001) show that it is in fact<br />
an O6.5V((f)) star.<br />
McSwain et al. (2001) obtained the radial velocity curve of the system, with<br />
their fitted parameters being a short period of P = 4.117 days, a high eccentricity<br />
of e = 0.41, a radial velocity of the system of V0 = 4.6 km s −1 and a mass function<br />
of f(m) = 0.00103 M⊙.<br />
Ribó et al. (2002) used positions at optical and radio wavelengths obtained at<br />
several epochs to compute in<strong>de</strong>pen<strong>de</strong>nt optical and radio proper motions, which<br />
were perfectly compatible, therefore confirming that both, the optical and the radio<br />
emission, originate in the same object. These authors also found that LS 5039 is a<br />
runaway X-ray binary with a space velocity ∼ 150 km s −1 , probably due to the mass<br />
loss experienced during the SN event that created the compact object in this binary
2.3. An interesting target in the search for new microquasars 35<br />
system. They also used all the available optical photometry to obtain an improved<br />
distance to the source of 2.9 ± 0.3 kpc.<br />
McSwain & Gies (2002), based on wind accretion mo<strong>de</strong>ls, have recently sug-<br />
gested a neutron star (1–3 M⊙) as the compact object in LS 5039, and proposed an<br />
inclination of i 30 ◦ for this binary system.<br />
Pare<strong>de</strong>s et al. (2002), thanks to EVN and MERLIN observations, have recently<br />
confirmed the persistent nature of the relativistic radio jets in LS 5039, and proposed<br />
a scenario based on the γ-ray/radio emission, in terms of the inverse Compton<br />
mechanism for the γ-ray emission and synchrotron mechanism for the radio emission.<br />
Finally, Casares et al. (2002) have just confirmed the ∼ 4 d orbital period, while<br />
Reig et al. (2002) have <strong>de</strong>termined that NH = 1.0 ± 0.3 × 10 22 cm −2 .<br />
2.3 An interesting target in the search for new<br />
microquasars<br />
X-ray, optical and radio catalogs provi<strong>de</strong> a fundamental tool for the search of new<br />
sources with known multiwavelength behavior. In a first step, the most obvious<br />
objects to inspect at radio wavelengths were those new X-ray binary candidates<br />
already proposed from previous researches.<br />
In this context, a systematic cross-i<strong>de</strong>ntification of the ROSAT All Sky Survey<br />
(RASS) (Voges et al. 1996, 1999) with OB star catalogs in the SIMBAD database<br />
was carried out by Motch et al. (1997), hereafter M97. Nearly two tens of OB/X-<br />
ray accreting binary candidates resulted from this work with different <strong>de</strong>grees of<br />
reliability. In particular, LS 5039 was proposed by M97 to be an X-ray binary<br />
system of the massive type with a high <strong>de</strong>gree of confi<strong>de</strong>nce (see Fig. 2.1). The<br />
unabsorbed X-ray luminosity in the 0.1–2.4 keV band, at an estimated distance<br />
of 3.1 kpc, amounted to LX 8.1 × 10 33 erg s −1 . This fact, together with the<br />
hardness of its X-ray spectrum were compatible with a neutron star, or a black<br />
hole, accreting directly from the companion’s wind.. The optical counterpart of the<br />
system appeared as a very luminous (V 11.2) star of the main sequence, with an<br />
early O7V spectral type.
36 Chapter 2. Multiwavelength approach to LS 5039<br />
DECLINATION (J2000)<br />
-14 49 00<br />
30<br />
50 00<br />
30<br />
51 00<br />
30<br />
52 00<br />
30<br />
LS 5039 V band<br />
LS 5039<br />
ROSAT<br />
53 00<br />
18 26 20 15 10<br />
RIGHT ASCENSION (J2000)<br />
05<br />
Figure 2.1: CCD image of LS 5039 obtained on 1998 June 7 with the 1.52 m OAN tele-<br />
scope in the V Johnson filter. The circle with 35 ′′ radius represents the 90% confi<strong>de</strong>nce<br />
position from the ROSAT PSPC pointed observations quoted in M97.<br />
2.4 The radio counterpart: NVSS J182614−145054<br />
2.4.1 A radio counterpart in the NVSS<br />
The NRAO VLA Sky Survey (NVSS) (Condon et al. 1998) provi<strong>de</strong>s a valuable tool<br />
to search for new REXBs. This catalog covers the sky north of δ = −40 ◦ (82% of the<br />
celestial sphere) at a frequency of 1.4 GHz (20 cm wavelength) using the Very Large<br />
Array (VLA) configurations D and DnC. It contains over 1.8 × 10 6 sources stronger<br />
than its 2.5 mJy completeness limit. The rms positional uncertainties are less than<br />
1 ′′ for sources stronger than 15 mJy and 7 ′′ for the faintest <strong>de</strong>tectable sources.<br />
For all the sources listed in M97 we examined the corresponding NVSS maps<br />
at the 20 cm wavelength in a search for possible radio counterparts. We found one<br />
interesting object in the M97 list that <strong>de</strong>served special attention, namely LS 5039<br />
(l = 16.88 ◦ , b = −1.29 ◦ ).
2.4. The radio counterpart: NVSS J182614−145054 37<br />
The existence of a radio counterpart to LS 5039 was first suspected after inspec-<br />
tion of the corresponding NVSS image. The 24 mJy source NVSS J182614−145054<br />
lied outsi<strong>de</strong>, but very close to, the 90% confi<strong>de</strong>nce error circle of RX J1826.2−1450,<br />
with 22 ′′ radius. This was the RASS source i<strong>de</strong>ntified with LS 5039 by M97. These<br />
authors also quoted another pointed ROSAT observation that yiel<strong>de</strong>d a 90% con-<br />
fi<strong>de</strong>nce radius of 35 ′′ , again consistent with NVSS J182614−145054. In addition,<br />
the NVSS coordinates were found to agree within 2 ′′ with the optical position of<br />
LS 5039, as <strong>de</strong>rived from the USNO-A1.0 catalog (Monet 1996, Monet et al. 1999).<br />
All these coinci<strong>de</strong>nces together stimulated our interest about this source and<br />
lead us to carry out higher resolution radio observations. This was consi<strong>de</strong>red to<br />
be the next logical step in or<strong>de</strong>r to better measure both its position and spectral<br />
properties. Therefore, we conducted a multiepoch and multifrequency radio study<br />
of LS 5039 with the VLA.<br />
2.4.2 VLA observations, discovery of a REXB<br />
Observations and results<br />
We observed 8 LS 5039 on several epochs at the wavelengths of 20, 6, 3.6 and 2.0 cm<br />
with the Very Large Array (VLA) interferometer of the NRAO 9 . The VLA data were<br />
processed following standard procedures within the aips package of NRAO. 3C 286<br />
was used as the amplitu<strong>de</strong> calibrator, while the phase calibrators observed were<br />
1834−126 at 20 cm, 1820−254 at 6 and 3.6 cm, and 1911−201 at 2 cm, respectively.<br />
The results of the interferometric runs on LS 5039 are summarized in Table 2.1,<br />
where the flux <strong>de</strong>nsity at several wavelengths is listed for the different observing<br />
dates. Some 20 cm values collected from the literature have been also inclu<strong>de</strong>d. The<br />
VLA data obtained in A configuration were further concatenated at 3.6 and 2.0 cm<br />
in or<strong>de</strong>r to obtain very sensitive sub-arcsec resolution maps. These are displayed in<br />
Fig. 2.2, where LS 5039 appears always as an unresolved point source (≤ 0.1 ′′ ). For<br />
the epochs when nearly simultaneous multifrequency observations are available, we<br />
have plotted in Fig. 2.3 the observed radio spectrum of LS 5039.<br />
8 Published in Martí, J., Pare<strong>de</strong>s, J. M. & Ribó, M. 1998, A&A, 338, L71.<br />
9 The National Radio Astronomy Observatory is a facility of the USA National Science Foun-<br />
dation operated un<strong>de</strong>r cooperative agreement by Associated Universities, Inc.
38 Chapter 2. Multiwavelength approach to LS 5039<br />
DECLINATION (J2000)<br />
Table 2.1: VLA observations of LS 5039.<br />
Date Julian Day VLA S20 cm S6 cm S3.6 cm S2.0 cm<br />
[JD−2400000] Conf. [mJy] [mJy] mJy] [mJy]<br />
1989 May 02 (a) 47649 B 36 ± 1 — — —<br />
1996 Jun 19 (b) 50254 D 23.7 ± 0.9 — — —<br />
1998 Feb 11 50856.2 D 28 ± 2 23.4 ± 0.1 18.1 ± 0.1 12.3 ± 0.3<br />
1998 Mar 10 50883.1 A 37 ± 1 24.0 ± 0.1 19.1 ± 0.1 14.1 ± 0.2<br />
1998 Apr 09 50913.0 A 40 ± 1 23.6 ± 0.1 16.9 ± 0.1 12.0 ± 0.2<br />
1998 May 12 50946.0 A 44 ± 1 25.7 ± 0.1 20.2 ± 0.1 15.1 ± 0.2<br />
(a) Helfand et al. (1992). (b) NVSS value from Condon et al. (1998).<br />
-14 50 52.0<br />
52.5<br />
53.0<br />
53.5<br />
54.0<br />
54.5<br />
55.0<br />
55.5<br />
56.0<br />
LS 5039 3.6 cm<br />
USNO<br />
18 26 15.15 15.10 15.05 15.00<br />
RIGHT ASCENSION (J2000)<br />
DECLINATION (J2000)<br />
-14 50 52.0<br />
52.5<br />
53.0<br />
53.5<br />
54.0<br />
54.5<br />
55.0<br />
55.5<br />
56.0<br />
LS 5039 2.0 cm<br />
USNO<br />
18 26 15.15 15.10 15.05 15.00<br />
RIGHT ASCENSION (J2000)<br />
Figure 2.2: Left. Self-calibrated map of LS 5039 at 3.6 cm after concatenation of<br />
all VLA A configuration data. Natural weight of the visibilities was used. The thick<br />
cross indicates the optical position of LS 5039 as listed in the USNO-A1.0 catalog.<br />
Contours are −3, 3, 5, 7, 10, 20, 30, 50, 100, 200, 400, 600, 800, 1000 and 1200 times<br />
0.015 mJy beam −1 , the rms noise. The synthesized beam, shown at the bottom left<br />
corner, is 0.40 ′′ × 0.25 ′′ , with position angle in 9.1 ◦ . Right. The same at 2.0 cm.<br />
Contours are −3, 3, 5, 10, 20, 30, 50, 80, 100, 120 and 140 times 0.082 mJy beam −1 ,<br />
the rms noise. The synthesized beam is 0.23 ′′ × 0.14 ′′ , with position angle in 13.4 ◦ .
2.4. The radio counterpart: NVSS J182614−145054 39<br />
Flux <strong>de</strong>nsity [mJy]<br />
50<br />
40<br />
30<br />
20<br />
15<br />
10<br />
May 12<br />
Apr 09<br />
Mar 10<br />
Feb 11<br />
1 2 5 8 10 15 20<br />
Frequency [GHz]<br />
Figure 2.3: The radio spectrum of LS 5039 on different epochs during the 1998 obser-<br />
vations with the VLA. Error bars not shown are smaller than the symbol size.<br />
Astrometric i<strong>de</strong>ntification<br />
The J2000.0 ICRS coordinates of the LS 5039 radio position <strong>de</strong>rived from the VLA<br />
A configuration maps in Fig. 2.2 were found to be: α = 18 h 26 m 15.056 s ± 0.001 s ,<br />
δ = −14 ◦ 50 ′ 54.24 ′′ ± 0.01 ′′ . These coordinates agree within 0.1 ′′ with those of<br />
LS 5039 in the optical USNO-A1.0 catalog, whose typical astrometric error is about<br />
0.25 ′′ . Therefore, we conclu<strong>de</strong>d that the i<strong>de</strong>ntification of the LS 5039 radio counter-<br />
part on the basis of astrometrical coinci<strong>de</strong>nce at the sub-arcsec level was virtually<br />
certain. From source count analysis (e.g., Led<strong>de</strong>n et al. 1980), the a priori probabil-<br />
ity of having a background extragalactic source with S20cm 30 mJy, or brighter,<br />
within 0.2 ′′ of the LS 5039 optical position is as low as 10 −7 . To our knowledge,<br />
this discovery increased to four the number of confirmed massive X-ray binaries with<br />
associated radio emission at that epoch, the other three being SS 433, Cygnus X-1<br />
and LS I +61 303 (Cygnus X-3 was a rather strange system with a Wolf rayet com-<br />
panion, γ Cas a very peculiar system in many aspects, and the transients V4641 Sgr<br />
and CI Cam were not yet discovered).
40 Chapter 2. Multiwavelength approach to LS 5039<br />
Radio behavior<br />
The radio spectra of LS 5039 shown in Fig. 2.3 are very suggestive of non-thermal<br />
radio emission. A typical one can be well represented by<br />
<br />
Sν<br />
ν<br />
= (52 ± 1)<br />
mJy<br />
GHz<br />
−0.46±0.01<br />
. (2.1)<br />
Given the unresolved angular size (θ ≤ 0.1 ′′ ), the brightness temperature, expressed<br />
as <br />
TB<br />
= 1.76 × 10<br />
K<br />
3<br />
<br />
ν<br />
<br />
−2<br />
GHz<br />
θ<br />
arcsec<br />
−2 <br />
Sν<br />
mJy<br />
, (2.2)<br />
can be estimated, for example at the 20 cm wavelength, as TB ≥ 4 × 10 6 K. Such<br />
a high value, together with the significantly negative spectral in<strong>de</strong>x α (where Sν ∝<br />
ν +α ), safely rules out a thermal emission mechanism. Non-thermal synchrotron<br />
radiation remains therefore as the most plausible interpretation for the LS 5039 radio<br />
emission. We can also obtain a lower limit to the source angular size by preventing<br />
catastrophic inverse Compton losses to occur (TB ≤ 10 12 K). This condition yields<br />
θ ≥ 0.2 milliarcsecond (mas), implying that further progress in studying the LS 5039<br />
structure could be feasible by the use of VLBI techniques (Sect. 2.4.4). We ignored<br />
at that epoch if the radio emission was originated in collimated milliarcsec radio jets<br />
or, alternatively, some other mechanism was at work.<br />
Assuming equipartition arguments for synchrotron radio sources (Pacholczyk<br />
1970), the total energy content and magnetic field in LS 5039 can be expressed as:<br />
Etotal = c13(1 + k) 4/7 φ 3/7 R 9/7 L 4/7<br />
rad<br />
, (2.3)<br />
H = 4.5 2/7 (1 + k) 2/7 c 2/7<br />
12 φ −2/7 R −6/7 L 2/7<br />
rad , (2.4)<br />
where Lrad is the integrated radio luminosity, R the linear size of the emitting<br />
region, φ the fraction of its volume covered by the magnetic field, k the electron<br />
to proton energy ratio, and c12 and c13 some special functions of the synchrotron<br />
theory tabulated in Pacholczyk (1970).<br />
For a distance of 2.9 kpc (Ribó et al. 2002), and assuming that the spectrum<br />
extends from 0.1 to 100 GHz, the total radio luminosity of the system can be eval-<br />
uated as Lrad ∼ 1 × 10 31 erg s −1 . The radio to X-ray luminosity ratio is thus<br />
Lradio/LX ∼ 10 −3 . Adopting φ = 1 and k = 1, the range of angular sizes found<br />
above (0.2 mas ≤ θ ≤ 0.1 ′′ ) translates into the following allowed source parameters:<br />
1.4 × 10 38 ≤ Etotal ≤ 4.3 × 10 41 erg and 0.01 ≤ B ≤ 2.0 G.
2.4. The radio counterpart: NVSS J182614−145054 41<br />
The radio emission level seems also to be persistent on timescales of, at least, few<br />
years. This is <strong>de</strong>duced from the <strong>de</strong>tections listed in Table 2.1. On shorter timescales,<br />
the 20 cm flux <strong>de</strong>nsity varies from 27 to 44 mJy between our first and last observing<br />
epochs, that are separated by four months. We also find some indications of radio<br />
variability for the remaining shorter wavelengths. The amplitu<strong>de</strong>s of variations at<br />
6, 3.6 and 2.0 cm are however not larger than ∼ 10%. The overall spectral evolution<br />
in Fig. 2.3 can be conceivably un<strong>de</strong>rstood in terms of a progressive <strong>de</strong>crease in<br />
the synchrotron self-absorption opacity. Unfortunately, the time sampling was not<br />
frequent enough to <strong>de</strong>termine what sort of radio events could configure the LS 5039<br />
light curve during our VLA monitoring. In any case, both the non-thermal spectral<br />
indices and the radio variability observed from month to month are usual features<br />
in REXBs (see, e.g., Hjellming & Han 1995).<br />
From the radio point of view, the LS 5039 behavior appeared so far consistent<br />
with the general properties of REXBs, thus supporting the M97 i<strong>de</strong>ntification. In<br />
particular, the radio emission of LS 5039 was somehow reminiscent of LS I +61 303.<br />
This is another massive REXB whose individual Lradio/LX ratio is remarkably similar<br />
to that of LS 5039 (Taylor et al. 1996; Pare<strong>de</strong>s et al. 1997). The energy and magnetic<br />
field estimates for LS I +61 303 (Massi et al. 1993) would also fit comfortably well<br />
within the limits found above for LS 5039. The analogy between both sources was<br />
however not complete at this moment, i.e., there was no evi<strong>de</strong>nce for LS 5039 radio<br />
outbursts in timescales of days or weeks.<br />
Conclusions<br />
After these observations (Martí et al. 1998) the star LS 5039 was proposed to be<br />
a new high mass REXB. The radio counterpart was consistent, within astrometric<br />
errors, with both the optical and X-ray position of the system. Its appearance at<br />
radio wavelengths was that of a compact source, even at the highest VLA resolution.<br />
The high brightness temperature and observed radio spectrum were suggestive of<br />
non-thermal synchrotron emission. LS 5039 also exhibited some evi<strong>de</strong>nces of radio<br />
variability during our VLA runs carried out on a monthly basis. The variations<br />
were specially important at the 20 cm wavelength, where the source flux <strong>de</strong>nsity<br />
increased by almost a factor of two. This kind of behavior, although mo<strong>de</strong>rate,<br />
strongly supported the i<strong>de</strong>ntification of the radio counterpart.
42 Chapter 2. Multiwavelength approach to LS 5039<br />
2.4.3 Long-term GBI monitoring, a persistent radio source<br />
In or<strong>de</strong>r to follow the radio behavior of LS 5039, it was inclu<strong>de</strong>d at our request<br />
in the list of radio sources routinely monitored at the Green Bank Interferometer<br />
(GBI) 10 . The source was also inclu<strong>de</strong>d in the RXTE All Sky Monitor (ASM) service<br />
at our request, to follow its behavior in the X-ray domain. The aims of the requested<br />
observations were basically to be able to <strong>de</strong>tect possible flaring events, X-ray/radio<br />
correlations and to allow the search for the orbital period of the binary. The RXTE<br />
ASM data will be discussed in Sect. 2.6.<br />
GBI observations<br />
The GBI observations consist of daily flux <strong>de</strong>nsity measurements with the GBI<br />
within the GBI-NASA Monitoring Program. LS 5039 was observed at the frequen-<br />
cies of 2.25 and 8.3 GHz (13.1 and 3.6 cm wavelength, respectively) during two<br />
periods. The first one spans 340 days, from 1998 September 17 to 1999 August 23<br />
(MJD 51073.9–51413.9). The corresponding data set, hereafter Set1, contains 284<br />
entries, with 282 successful flux <strong>de</strong>nsity measurements at both frequencies. We show<br />
the flux <strong>de</strong>nsity lightcurves of the Set1 data in the top panel of Fig. 2.4, while the<br />
spectral in<strong>de</strong>x α (where Sν ∝ ν +α ) computed between the two frequencies is shown<br />
in the bottom panel. The second period of observations spans 24 days, from 2000<br />
September 12 to October 6 (MJD 51799.0–51822.9). The corresponding data set,<br />
hereafter Set2, contains 24 successful flux <strong>de</strong>nsity measurements at both frequen-<br />
cies. The flux <strong>de</strong>nsity lightcurves and spectral in<strong>de</strong>x for the Set2 data are shown in<br />
Fig. 2.5. Since there is a ∼ 1 year gap between both data sets, we have preferred to<br />
analyze them separately.<br />
Set1 data<br />
An analysis of the first ∼ 4 months of GBI observations was reported in Ribó et al.<br />
(1999), while Clark et al. (2001) already analyzed this data set.<br />
The mean GBI flux <strong>de</strong>nsities and standard <strong>de</strong>viations are S2.25 GHz = 31.3 (±6.2)<br />
10 The Green Bank Interferometer is a facility of the USA National Science Foundation operated<br />
by NRAO in support of the NASA High Energy Astrophysics programs.
2.4. The radio counterpart: NVSS J182614−145054 43<br />
Flux <strong>de</strong>nsity [mJy]<br />
Spectral in<strong>de</strong>x<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
0<br />
−1<br />
−2<br />
GBI−NASA Monitoring Program (Set1 data)<br />
2.25 GHz<br />
8.3 GHz<br />
−3<br />
51050 51100 51150 51200 51250 51300 51350 51400 51450<br />
Modified Julian Date [JD−2400000.5]<br />
Figure 2.4: Top: GBI radio light curves of LS 5039 Set1 data at the frequencies of 2.25<br />
and 8.3 GHz. Representative ± 1σ error bars have been plotted for the first data point.<br />
Bottom: The corresponding spectral in<strong>de</strong>x. Error bars are also ± 1σ.<br />
mJy and S8.3 GHz = 14.8 (±5.7) mJy. It must be noted that the 1σ errors of the<br />
individual data points are 4 and 6 mJy, respectively. Hence, the source is <strong>de</strong>tected<br />
at the ∼ 8σ level at 2.25 GHz, and only at the ∼ 2.5σ level at 8.3 GHz. However,<br />
the GBI is noise dominated below 10 mJy. Therefore, as pointed out by Clark<br />
et al. (2001), the mean and standard <strong>de</strong>viation at 8.3 GHz should be estimated<br />
with a Gaussian distribution truncated at 10 mJy. The result of this analysis is<br />
S8.3 GHz = 13.7 (±6.7) mJy, indicating that LS 5039 is <strong>de</strong>tected, on average, at the<br />
∼ 2.0σ level at this frequency, preventing any reasonable variability analysis.<br />
A careful inspection of the 2.25 GHz data in Fig. 2.4 reveals that the typical<br />
day-to-day variability does not exceed ∼ 30 %. Ribó et al. (1999) noted that there<br />
could be some exceptions to this behavior, for example around MJD 51075, 51086
44 Chapter 2. Multiwavelength approach to LS 5039<br />
and 51176, when the flux <strong>de</strong>nsity of LS 5039 seemed to have varied by more than a<br />
factor of ∼ 2 on less than one day. However, inspection of the GBI operational notes<br />
reveals that technical problems such as cryogenics warming, or weather problems<br />
such as snow on the dishes, were responsible of this behavior. The apparent bump<br />
between MJD 51180 and MJD 51200 is also a result of cryogenics warming. Hence,<br />
it seems that most of the features in the lightcurve are due to technical problems,<br />
because of the relatively low flux <strong>de</strong>nsity of the source.<br />
Nevertheless, since Pooley et al. (1999) could <strong>de</strong>tect the 5.6 d orbital period<br />
of Cygnus X-1 in similar radio lightcurves, we have performed a timing analysis of<br />
the LS 5039 lightcurves at both frequencies. Given the span of the observations,<br />
the search was restricted between 2 and 100 d. The methods employed were the<br />
Phase Dispersion Minimization (PDM) (Stellingwerf 1978) and the CLEAN algo-<br />
rithm (Roberts et al. 1987). Unfortunately, no convincing period was <strong>de</strong>tected in<br />
this process. In particular, we do not <strong>de</strong>tect the 4.117 d orbital period recently<br />
found by McSwain et al. (2001) from radial velocity measurements. Folding of the<br />
data with such a period does not reveal any significant difference on radio emission<br />
along the orbit. Since the orbit is quite eccentric (e=0.41), and this is a wind fed<br />
system (where the primary does not fill its Roche lobe), we expect the accretion rate<br />
onto the compact object to change significantly between periastron and apastron.<br />
Hence, the non <strong>de</strong>tection of the orbital period in the GBI data can be attributed to<br />
its relatively high noise due to several facts (intrinsic measurement noise, technical<br />
problems, weather conditions, etc.).<br />
Although the 8.3 GHz data is dominated by noise, we can extract some infor-<br />
mation about the spectral in<strong>de</strong>x. The weighted mean of individual spectral indices<br />
is found to be α = −0.4 ± 0.3. On the other hand, if we use the mean flux <strong>de</strong>nsities<br />
at both frequencies we obtain a spectral in<strong>de</strong>x of α = −0.6 ± 0.4. These value<br />
are in good agreement with the results presented in Sect. 2.4.2 (Martí et al. 1998)<br />
obtained a few months before, thus suggesting that the non-thermal radio spectrum<br />
is a persistent property of the source.<br />
Set2 data<br />
This data set is not useful for timing analysis or other purposes due to its limited<br />
time span and total number of points. However, we have explicitly inclu<strong>de</strong>d the
2.4. The radio counterpart: NVSS J182614−145054 45<br />
Flux <strong>de</strong>nsity [mJy]<br />
Spectral in<strong>de</strong>x<br />
100<br />
80<br />
60<br />
40<br />
20<br />
0<br />
0<br />
−1<br />
−2<br />
GBI−NASA Monitoring Program (Set2 data)<br />
2.25 GHz<br />
8.3 GHz<br />
−3<br />
51795 51800 51805 51810 51815 51820 51825<br />
Modified Julian Date [JD−2400000.5]<br />
Figure 2.5: Same as Fig. 2.4 for the Set2 data. The flare is probably false (see text).<br />
light curve in Fig. 2.5 to comment on the intriguing flare seen at 8.3 GHz around<br />
MJD 51803. Although there could be an increase in flux <strong>de</strong>nsity at 2.25 GHz, thus<br />
supporting a possible outburst, the GBI operational notes reveal that spikes were<br />
occasionally occurring at that time, and it is stated that single high scans should be<br />
ignored. Hence, we conclu<strong>de</strong> that the flare shown in Fig. 2.5 is probably false.<br />
Conclusions<br />
LS 5039 is a persistent radio source. In the GBI monitoring it is clearly <strong>de</strong>tected<br />
at 2.25 GHz, with a mean flux <strong>de</strong>nsity of ∼ 31 mJy, and marginally <strong>de</strong>tected at<br />
8.3 GHz, with a mean flux <strong>de</strong>nsity of ∼ 14 mJy. The spectral in<strong>de</strong>x of the source is<br />
α ∼ −0.5, confirming earlier results. In addition to the negative spectral in<strong>de</strong>x, the<br />
brightness temperature estimates for LS 5039 clearly yield to non-thermal values,
46 Chapter 2. Multiwavelength approach to LS 5039<br />
hence supporting the mechanism of synchrotron radiation as the responsible for the<br />
<strong>de</strong>tected radio emission.<br />
Excluding false data due to technical or weather problems, LS 5039 does not<br />
show day-to-day variability larger than ∼ 30%, and no strong radio outbursts have<br />
been <strong>de</strong>tected. No orbital period is <strong>de</strong>tected in the GBI data in the range between<br />
2 and 100 days. In general terms, the observed radio behavior confirms the earlier<br />
suggestions by Martí et al. (1998) concerning the persistent and mo<strong>de</strong>rately variable<br />
nature of the radio emission.<br />
2.4.4 VLBA observations, discovery of a microquasar<br />
As we have seen in Sect. 2.4.2, LS 5039 appeared unresolved to the VLA in A<br />
configuration, i.e., with a size ≤ 0.1 ′′ . On the other hand, a lower limit to the source<br />
size of ≥ 0.2 mas was <strong>de</strong>rived to prevent catastrophic inverse Compton losses to<br />
occur. Therefore, we <strong>de</strong>ci<strong>de</strong>d to conduct observations using the Very Long Baseline<br />
Interferometry (VLBI) technique, aimed to reveal the mas structure of the source.<br />
Observations and results<br />
We observed 11 LS 5039 at 5 GHz frequency (6 cm wavelength) on 1999 May 8<br />
(MJD 51306) with the 10×25-m antennas of the NRAO 12 Very Long Baseline Array<br />
(VLBA) and the 27×25-m antennas of the VLA in its phased array mo<strong>de</strong> in its D<br />
configuration. In this mo<strong>de</strong>, the VLA has the same sensitivity as a dish of ∼ 115 m<br />
diameter, and provi<strong>de</strong>s sensitive baselines with the VLBA antennas. The VLA<br />
phasing process was successful, and it worked without problems as another VLBI<br />
station. On the other hand, in theory the VLA could also record data on its own.<br />
This would allow to perform a very sensitive VLA map and to follow the radio flux<br />
<strong>de</strong>nsity during the VLBA run. However, there was an air conditioning failure in the<br />
VLA correlator, and it could not record data.<br />
The source 3C 345 was used as a fringe-fin<strong>de</strong>r, whereas J1733−1304 was the<br />
phasing source for the VLA. We could not use the VLBI phase-reference technique<br />
11 Published in Pare<strong>de</strong>s, J. M., Martí, J., Ribó, M., & Massi, M. 2000, Science, 288, 2340.<br />
12 The National Radio Astronomy Observatory is a facility of the USA National Science Foun-<br />
dation operated un<strong>de</strong>r cooperative agreement by Associated Universities, Inc.
2.4. The radio counterpart: NVSS J182614−145054 47<br />
MilliARC SEC<br />
4<br />
2<br />
0<br />
-2<br />
-4<br />
-6<br />
VLBA<br />
6 4 2 0 -2 -4 -6<br />
MilliARC SEC<br />
Figure 2.6: VLBA self-calibrated image of LS 5039 at 5 GHz. Contours are 6, 8, 10,<br />
12, 14, 16, 18, 20, 25, 30, 40, and 50 times 0.085 mJy beam −1 , the rms noise. The<br />
synthesized beam, shown at the bottom right corner, is 3.4 mas×1.2 mas, with position<br />
angle in 0 ◦ .<br />
because there was no obvious nearby calibrator to be used. Data were recor<strong>de</strong>d<br />
in VLBA mo<strong>de</strong> v6cm-256-8-2-L, i.e., at 6 cm wavelength, at 256 Mbps, using 8<br />
baseband channels, with 2-bit sampling at left hand circular polarization, yielding<br />
a full bandwidth of 64 MHz. The data were correlated at the VLBA correlator in<br />
Socorro, New Mexico, using an integration time of 4 s. The LS 5039 coordinates<br />
used for correlation were those obtained after the VLA A configuration observations,<br />
i.e., α = 18 h 26 m 15.056 s , δ = −14 ◦ 50 ′ 54.24 ′′ , expressed in J2000.0 ICRS. The<br />
output data were then calibrated using standard procedures in unconnected radio<br />
interferometry within the aips and difmap software packages.<br />
The resulting pattern of the observed visibility amplitu<strong>de</strong>s, <strong>de</strong>caying as a function<br />
of baseline length, indicated that LS 5039 had structure at milliarcsecond scales. The
48 Chapter 2. Multiwavelength approach to LS 5039<br />
Flux <strong>de</strong>nsity at 2.25 GHz [mJy]<br />
80<br />
60<br />
40<br />
20<br />
VLBA run 1999 May 8<br />
0<br />
51290 51295 51300 51305 51310 51315 51320<br />
Modified Julian Date [JD−2400000.5]<br />
Figure 2.7: GBI radio monitoring of LS 5039, at 2.25 GHz, during the weeks before and<br />
after the date of our VLBA observations, indicated by the vertical bar.<br />
final synthesis map (see Fig. 2.6) shows a two-si<strong>de</strong>d jet emerging from a central core.<br />
A <strong>de</strong>convolved angular size of about 2 mas is estimated for the core. The jets extend<br />
over 6 mas on the plane of the sky, oriented along a position angle of ∼ 125 ◦ with<br />
respect to the North, and they account for 20% of the total 16 mJy flux <strong>de</strong>nsity. At<br />
a distance of ∼ 3 kpc the total projected length of the jets is ∼ 20 AU.<br />
Discussion<br />
To obtain some or<strong>de</strong>r of magnitu<strong>de</strong> estimates, we will assume that the overall size<br />
of the radio source is approximately 6 mas×2 mas. This implies a high brightness<br />
temperature of ∼ 9.4 × 10 7 K, indicative of synchrotron radiation. As has been seen<br />
in Sects. 2.4.2 and 2.4.3, the LS 5039 radio spectrum as a function of frequency<br />
often displays a negative spectral in<strong>de</strong>x which is in agreement with a non-thermal<br />
optically thin emission mechanism (Martí et al. 1998, Ribó et al. 1999).<br />
The <strong>de</strong>tection of jets occurred at a time when the source was at its typical<br />
persistent level of radio emission, and only mo<strong>de</strong>rately variable, as inferred from<br />
concurrent radio monitoring by the Green Bank Interferometer (GBI) (see Fig. 2.7).<br />
The absence of any precursor outburst for the radio jets strongly suggests that they<br />
are always present and continuously emanating from the core.
2.4. The radio counterpart: NVSS J182614−145054 49<br />
The data can be mo<strong>de</strong>l fitted with three components: a central core of 12.8 mJy,<br />
a southeast component located at an angular distance of 3.8 ± 0.2 mas from the<br />
core and with a flux <strong>de</strong>nsity of 2.1 ± 0.1 mJy, and a northwest component located<br />
at 2.8 ± 0.2 mas from the core having 1.0 ± 0.1 mJy. It seems reasonable to assume<br />
that the different distance from the components to the core and the different flux<br />
<strong>de</strong>nsities reflect relativistic effects (Rees 1966, Blandford & Königl 1979, see also<br />
Sect. 1.2.1). Hence, using the distances to the core and<br />
β cos θ = µa − µr<br />
µa + µr<br />
= da − dr<br />
da + dr<br />
(2.5)<br />
we obtain β cos θ = 0.15 ± 0.04. On the other hand, based on the flux <strong>de</strong>nsity<br />
asymmetry we can use the equation<br />
β cos θ > (Sa/Sr) 1/(k−α) − 1<br />
(Sa/Sr) 1/(k−α) + 1<br />
(2.6)<br />
to obtain a lower limit on β cos θ. Assuming that the jet flow is continuous, k = 2,<br />
and using α = −0.5±0.4 we obtain β cos θ 0.15±0.03. Both results are in excellent<br />
agreement. If we consi<strong>de</strong>r that the jet is composed by discrete con<strong>de</strong>nsations, k = 3,<br />
then β cos θ 0.11±0.02. The value <strong>de</strong>duced by the use of the distance asymmetry,<br />
β cos θ = 0.15 ± 0.04, allows us to find a lower limit for the jet velocity of β ≥<br />
0.15 ± 0.04 and an upper limit for the ejection angle of θ ≤ 81 ◦ ± 2 ◦ .<br />
Additional information on the source energetics can be obtained by assuming<br />
energy equipartition between the relativistic electrons and the magnetic field (Pa-<br />
cholczyk 1970). We are forced to use the overall source parameters observed, because<br />
not enough information is yet available for appropriate calculations in the rest frame<br />
of the ejecta. The corresponding results are nevertheless expected to be within an<br />
or<strong>de</strong>r of magnitu<strong>de</strong> for a mildly relativistic system. Un<strong>de</strong>r these assumptions, the<br />
observed radio properties of LS 5039 imply a total energy content in relativistic<br />
electrons of Ee ∼ 5 × 10 39 erg, with an equipartition magnetic field of ∼ 0.2 G.<br />
Conclusions<br />
The VLBA observations showed that LS 5039 exhibited a two-si<strong>de</strong>d jet morphology.<br />
The brightness and distance to the core asymmetry of the <strong>de</strong>tected jet components<br />
implies, assuming that they are intrinsically equal, that the jets have a bulk motion<br />
higher than 0.1c, thus revealing the microquasar nature of LS 5039.
50 Chapter 2. Multiwavelength approach to LS 5039<br />
2.4.5 EVN and MERLIN observations,<br />
confirmation of persistent relativistic radio jets<br />
The persistent radio emission of LS 5039 suggests that the jet is always present.<br />
With the aim to confirm this hypothesis, and also to <strong>de</strong>tect the jet at larger angular<br />
scales than those imaged with the previous Very Long Baseline Array (VLBA) mea-<br />
surements, we observed this microquasar with the European VLBI Network (EVN)<br />
and the Multi-Element Radio-Linked Interferometer Network (MERLIN).<br />
Observations and results<br />
We observed 13 LS 5039 simultaneously with the EVN 14 and MERLIN 15 on 2000<br />
March 1 (3:20–7:10 UT, MJD 51604.2) at 5 GHz. Single dish flux <strong>de</strong>nsity measure-<br />
ments were carried out with the MPIfR 100 m antenna in Effelsberg. We scheduled<br />
the observation introducing the phase-reference calibrator J1823−1228, the fringe-<br />
fin<strong>de</strong>r DA 193 and the flux <strong>de</strong>nsity calibrator 3C 286. Unfortunately, the phase-<br />
reference calibrator was not <strong>de</strong>tected and could not be used during the correlation.<br />
The EVN observations were performed with the array <strong>de</strong>scribed in Table 2.2.<br />
Data were recor<strong>de</strong>d in MkIV mo<strong>de</strong> with 2-bit sampling at 256 Mbps at left hand<br />
circular polarization, yielding a full bandwidth of 64 MHz (v6cm-256-8-2-L mo<strong>de</strong>).<br />
The data were correlated at the Joint Institute for VLBI in Europe (JIVE). An inte-<br />
gration time of 4 s was used. Interferometer fringes for LS 5039 were <strong>de</strong>tected in all<br />
baselines. A later fringe fitting of the residual <strong>de</strong>lays and fringe rates was performed<br />
within aips. A first a priori visibility amplitu<strong>de</strong> calibration was performed using<br />
antenna gains and system temperatures measured at each antenna. We averaged<br />
in frequency the data and exported them to be imaged and self-calibrated into the<br />
difference mapping software difmap. The final imaging was carried out on those<br />
data after editing and averaging of the visibilities in 32 s blocks.<br />
MERLIN recor<strong>de</strong>d data with 2-bit sampling at dual polarization and a total<br />
13Published in Pare<strong>de</strong>s, J. M., Ribó, M., Ros, E., Martí, J., & Massi, M. 2002, A&A, 393, L99.<br />
14The European VLBI Network is a joint facility of European, Chinese and other radio astronomy<br />
institutes fun<strong>de</strong>d by their national research councils.<br />
15MERLIN is operated as a National Facility by the University of Manchester at Jodrell Bank<br />
Observatory on behalf of the UK Particle Physics & Astronomy Research Council.
2.4. The radio counterpart: NVSS J182614−145054 51<br />
Table 2.2: EVN array used in the LS 5039 observations.<br />
Antenna Co<strong>de</strong> Location Diameter DPFU (a) Tsys (b)<br />
[m] [K Jy −1 ] [K]<br />
Effelsberg EB Germany 100 1.47 27<br />
Jodrell Bank JB U.K. 25 0.11 40<br />
Cambridge CM U.K. 32 0.21 38<br />
Westerbork WB Netherlands 14×25 1.0 67<br />
Medicina MC Italy 32 0.15 53<br />
Noto NT Italy 32 0.16 47<br />
Toruń TR Poland 32 0.15 31<br />
(a) Degrees per flux unit. (b) Best value during the experiment.<br />
32 MHz bandwidth. We analyzed the left hand circular polarization data excluding<br />
1 MHz channel at both edges of the band, yielding a final bandwidth of 14 MHz.<br />
The correlator integration time was of 4 s. Standard reduction and imaging analysis<br />
were carried out using aips and difmap.<br />
To complement the amplitu<strong>de</strong> calibration, we interleaved cross-scans (in azimuth<br />
and elevation) with the 100 m Effelsberg antenna to measure the radio source flux<br />
<strong>de</strong>nsities (A. Kraus, private communication). We fitted a Gaussian function to the<br />
flux-<strong>de</strong>nsity response for every cross-scan, and we averaged the different Gaussians.<br />
We linked the flux <strong>de</strong>nsity scale by observing primary calibrators such as 3C 286,<br />
3C 48, or NGC 7027 (see e.g. Kraus 1997 or Peng et al. 2000).<br />
Unfortunately, the flux <strong>de</strong>nsity monitoring with the Effelsberg antenna was not<br />
reliable due to contamination effects by the galactic plane diffuse emission and/or by<br />
the close SNR G016.8−01.1. However, inspection of the shortest MERLIN baselines<br />
reveals a constant flux <strong>de</strong>nsity during the full run. That allowed us to image the<br />
source directly without splitting the data in time blocks. We present the obtained<br />
EVN and MERLIN images in Figs. 2.8 and 2.9, respectively. The MERLIN image is<br />
presented convolved with a circular beam, equivalent in solid angle to the interfero-<br />
metric synthesized beam of 142×46 mas at position angle (P.A.) 1.15 ◦ . Our images<br />
clearly show that LS 5039 has a two-si<strong>de</strong>d jet emanating from a central core. In<br />
both images the southeast (SE) jet is brighter and larger than the northwest (NW)<br />
one, as can be seen in Table 2.3, where we have quoted the flux <strong>de</strong>nsities of the core<br />
and the jets and the lengths and P.A. of the jets. Their total lengths are ∼ 60 mas<br />
in the EVN image and ∼ 300 mas in the MERLIN one.
52 Chapter 2. Multiwavelength approach to LS 5039<br />
Figure 2.8: EVN self-calibrated image of LS 5039 at 5 GHz obtained on 2000 March 1.<br />
Axes units are in mas. The synthesized beam, plotted in the lower left corner, has a size of<br />
7.60 mas×6.96 mas in P.A. of −14 ◦ . The first contour corresponds to 0.3 mJy beam −1 ,<br />
while consecutive ones scale with 3 1/2 .<br />
Table 2.3: Flux <strong>de</strong>nsities at 5 GHz, length and P.A. of the LS 5039 structure features<br />
as <strong>de</strong>tected by EVN and MERLIN.<br />
EVN MERLIN<br />
S 5 GHz Length P.A. S 5 GHz Length P.A.<br />
[mJy] [mas] [ ◦ ] [mJy] [mas] [ ◦ ]<br />
Core 29.3 — — 31.6 — —<br />
NW jet 2.6 24 −42 4.0 128 −29<br />
SE jet 3.3 34 140 4.2 174 150<br />
Note: the errors in flux <strong>de</strong>nsity, lenght and P.A. are 0.1 mJy beam −1 , 2 mas and 4 ◦<br />
for the EVN features and 0.4 mJy beam −1 , 12 mas and 5 ◦ for the MERLIN ones.
2.4. The radio counterpart: NVSS J182614−145054 53<br />
Figure 2.9: MERLIN self-calibrated image of LS 5039 at 5 GHz obtained on 2000<br />
March 1. Axes units are in mas. The convolving circular beam, plotted in the lower left<br />
corner, has a diameter of 81 mas. The first contour corresponds to a flux <strong>de</strong>nsity of<br />
1 mJy beam −1 , while consecutive ones scale with 3 1/2 .<br />
Discussion<br />
These results confirm the existence of a two-si<strong>de</strong>d radio jet in LS 5039 reported in<br />
previous VLBA 5 GHz observations (Pare<strong>de</strong>s et al. 2000, see also Sect. 2.4.4). This<br />
source does not present any strong outburst or, at least, none has been <strong>de</strong>tected<br />
during the eleven month monitoring carried out by the Green Bank Interferome-<br />
ter between 1998 September 17 and 1999 August 23 (Clark et al. 2001, see also<br />
Sect. 2.4.3). On the other hand, inspection of the RXTE All Sky Monitor data<br />
(Levine et al. 1996) reveals that the X-ray flux was at the typical low level (Ribó<br />
et al. 1999) at the epoch of the EVN and MERLIN observations and several weeks<br />
before. All this suggests that the jets are persistent, as the VLBI images obtained<br />
up to now seem to indicate.<br />
In the observations reported here, the jets extend further away than the ∼ 6 mas
54 Chapter 2. Multiwavelength approach to LS 5039<br />
of the VLBA observations at the same frequency. In all the images the jets have<br />
similar position angles, ∼ 125 ◦ in the VLBA image, ∼ 140 ◦ in the EVN image, and<br />
∼ 150 ◦ in the MERLIN image. These results suggest a bending of the jets with<br />
increasing distance from the core and/or precession.<br />
The brightness and length asymmetry of the jet components may involve special<br />
relativity effects (Rees 1966, Blandford & Königl 1979, see also Sect. 1.2.1). Hence,<br />
assuming that this is the reason for the <strong>de</strong>tected length asymmetry of the jets, we<br />
can estimate some parameters by using the following equation:<br />
β cos θ = µa − µr<br />
µa + µr<br />
= da − dr<br />
da + dr<br />
. (2.7)<br />
Although we do not know the epoch of ejection of the terminal plasma, we can<br />
i<strong>de</strong>ntify da and dr with the lengths of the approaching and receding jet, respectively.<br />
Using Eq. 2.7 and the EVN values in Table 2.3 we obtain β cos θ = 0.17 ± 0.05,<br />
and hence β > 0.17 ± 0.05 and θ < 80 ◦ ± 3 ◦ . For the MERLIN values we obtain<br />
β > 0.15 ± 0.06 and θ < 81 ◦ ± 3 ◦ . These values are similar to those previously<br />
<strong>de</strong>rived from the VLBA image, of β > 0.15 ± 0.04 and θ < 81 ◦ ± 2 ◦ . We have not<br />
consi<strong>de</strong>red an eventual larger size for the NW jet with a brightness below the image<br />
noise level. McSwain & Gies (2002) have recently proposed an inclination of i 30 ◦<br />
for LS 5039. If we assume that the jet is perpendicular to the accretion disk, and<br />
that the disk lies in the orbital plane of the binary system, then θ = i = 30 ◦ , and<br />
using the values from the EVN image, we obtain β = 0.20 ± 0.06, which indicates<br />
a mildly relativistic jet. In this case, the apparent velocity and proper motion of<br />
the approaching components would be 0.12c and 7.2 mas day −1 . For the receding<br />
components we would have 0.085c and 5.1 mas day −1 .<br />
The total length of the EVN and MERLIN jets is ∼ 60 and ∼ 300 mas, respec-<br />
tively. Consi<strong>de</strong>ring that the source is located at 2.9 kpc (Ribó et al. 2002), these<br />
angular lengths translate into linear lengths in the plane of the sky of ∼ 175 and<br />
∼ 870 AU, respectively. Assuming that θ = 30 ◦ we obtain intrinsic total lengths<br />
of ∼ 350 and ∼1740 AU, respectively, and lengths of the SE jet arm of ∼ 200 and<br />
∼1000 AU, respectively. Moreover, the jet width is smaller than one synthesized<br />
beam even in the EVN image. This implies a jet half opening angle ≤ 6 ◦ .<br />
Further discussion on the LS 5039 jets is reported in Sect. 2.8.
2.5. The optical/IR star: LS 5039 55<br />
Conclusions<br />
The EVN and MERLIN observations have confirmed the existence of relativistic<br />
radio jets in LS 5039, as previously found in VLBA observations (Pare<strong>de</strong>s et al.<br />
2000, see also Sect. 2.4.4). Moreover, the recurrent <strong>de</strong>tection of the jets strongly<br />
suggests that in LS 5039 relativistic radio jets are always present.<br />
2.5 The optical/IR star: LS 5039<br />
The star LS 5039 was originally classified as the star number 5039 in the catalog<br />
Luminous stars in the southern Milky Way, by Stephenson & Sanduleak (1971). It<br />
is an early type star, O6.5V((f)), with a visual magnitu<strong>de</strong> ∼ 11.2. In this section<br />
we report photometric and spectroscopic results that have allowed to obtain the<br />
spectral type, the distance and the orbital parameters of the binary system.<br />
2.5.1 Photometry<br />
Optical photometry of LS 5039 has been carried out by several groups. Here we<br />
start by reporting photometry obtained by us, and then provi<strong>de</strong> a compilation of<br />
optical and near infrared photometry obtained by other groups.<br />
Optical observations with the 1.5 m OAN telescope<br />
Soon after the VLA runs presented in Sect. 2.4.2, we carried out optical CCD ob-<br />
servations 16 at Calar Alto observatory (Spain). Our main goal was to search for<br />
photometric variability that could evi<strong>de</strong>nce the active nature of LS 5039 as an X-<br />
ray binary. We used the 1.52 m telescope of the Spanish Observatorio Astronómico<br />
Nacional (OAN) from 1998 June 1 to 8. The Ritchey-Chrétien focus was available<br />
together with a Tektronics TK1024AB chip. This provi<strong>de</strong>d a scale factor of 0.4 ′′ per<br />
pixel and a 6.9 ′ × 6.9 ′ field of view. Images were acquired through the V RI Johnson<br />
filters and they were reduced using standard procedures based on the iraf image<br />
processing system. In Fig. 2.10 we show a 30 s exposure of the LS 5039 field in<br />
16 Published in Martí, J., Pare<strong>de</strong>s, J. M. & Ribó, M. 1998, A&A, 338, L71.
56 Chapter 2. Multiwavelength approach to LS 5039<br />
DECLINATION (J2000)<br />
-14 49 00<br />
30<br />
50 00<br />
30<br />
51 00<br />
30<br />
52 00<br />
30<br />
LS 5039 V band<br />
LS 5039<br />
ROSAT<br />
53 00<br />
18 26 20 15 10<br />
RIGHT ASCENSION (J2000)<br />
05<br />
Figure 2.10: CCD image of LS 5039 obtained on 1998 June 7 with the 1.52 m OAN<br />
telescope in the V Johnson filter. The circle with 35 ′′ radius represents the 90% confi-<br />
<strong>de</strong>nce position from the ROSAT PSPC pointed observations quoted in M97. The two<br />
comparison stars used for differential photometry are labeled as C1 and C2.<br />
the V band. Differential photometry was performed against two nearby comparison<br />
stars (C1 and C2 in Fig. 2.10).<br />
Based on several observations of Landolt (1992) standards, the adopted magni-<br />
tu<strong>de</strong>s of the comparison stars are V = 12.53, R = 12.01, I = 11.55 and V = 10.10,<br />
R = 9.36, I = 8.68 for C1 and C2, respectively. This absolute photometry is accu-<br />
rate to ±0.03 mag. Relative variations in the final photometric results of Table 2.4<br />
can be nevertheless traced at the ±0.01 mag level.<br />
From our OAN observations we propose that there may be day-to-day variations<br />
with amplitu<strong>de</strong>s of a few hundredths of magnitu<strong>de</strong>. This appears specially evi<strong>de</strong>nt<br />
in the latest two days of Table 2.4. Here, the LS 5039 brightness <strong>de</strong>creased by about<br />
0.04 mag in V and 0.05 mag in R and I from one day to the next. In doing so,<br />
all the differences between the two comparison stars remained constant within less<br />
than 0.01 mag, thus supporting the reality of this variation.<br />
C1<br />
C2
2.5. The optical/IR star: LS 5039 57<br />
Table 2.4: Optical photometry of LS 5039.<br />
Date Julian Day V R I<br />
[JD−2400000]<br />
1998 Jun 2 50966.5 11.35 — —<br />
1998 Jun 3 50967.6 11.33 — —<br />
1998 Jun 7 50971.6 11.35 10.64 9.90<br />
1998 Jun 8 50972.5 11.39 10.69 9.95<br />
Compilation of optical and near infrared observations<br />
Optical and/or near infrared photometry of LS 5039 has been published in the past<br />
by Drilling (1975, 1991), Lahulla & Hilton (1992), Kilkenny et al. (1993), Spencer<br />
Jones et al. (1993), Martí et al. (1998, see Table 2.4) and Clark et al. (2001).<br />
We have compiled all simultaneous broad band (at least 3 filters) measurements in<br />
Table 2.5, where the initials and numbers refer to the cited papers. As can be seen,<br />
all values are consistent with a long-term variability smaller than ∼ 0.1 mag in all<br />
filters except in V , where it varies ∼ 0.2 mag. This suggests that no strong optical<br />
variations appear to be present on timescales of years.<br />
Nevertheless, Clark et al. (2001) also report broad band JHK photometry in the<br />
period 1995–2000 that indicates variability up to ∼0.4 mag in H band and ∼0.5 mag<br />
in K band. These authors also searched for medium-term variability in the B and V<br />
bands during 21 nights spanning 43 days. They only could find aperiodic variability<br />
at the hundredth of magnitu<strong>de</strong> level, on single nights and from night to night, that<br />
can be attributed to variations in mass loss rate or to non-radial pulsations.<br />
Table 2.5: Compilation of optical and near infrared photometry of LS 5039.<br />
Observation U B V R I J H K L<br />
D91 12.02 ± 0.02 12.18 ± 0.01 11.23 ± 0.01 — — — — — —<br />
LH92 11.97 ± 0.02 12.15 ± 0.02 11.20 ± 0.02 — — — — — —<br />
K93, SJ93 — — 11.24 ± 0.01 10.59 ± 0.01 9.88 ± 0.01 9.02 ± 0.01 8.79 ± 0.02 8.57 ± 0.03<br />
1996 Oct, C01 — 12.18 ± 0.02 11.33 ± 0.02 10.65 ± 0.02 9.87 ± 0.02 9.05 ± 0.02 8.75 ± 0.02 8.60 ± 0.02 8.69 ± 0.05<br />
1998 Jun 7 M98 — — 11.35 ± 0.03 10.64 ± 0.03 9.90 ± 0.03 — — — —<br />
1998 Jun 8, M98 — — 11.39 ± 0.03 10.69 ± 0.03 9.95 ± 0.03 — — — —<br />
2000 Sep, C01 — 12.17 ± 0.03 11.32 ± 0.01 10.61 ± 0.01 9.91 ± 0.01 — — — —
58 Chapter 2. Multiwavelength approach to LS 5039<br />
Apart from optical photometry in the Johnson filters, Strömgren and Hβ pho-<br />
tometry have also been conducted. Kilkenny & Whittet (1993) reported the fol-<br />
lowing Strömgren photometry for LS 5039: V = y = 11.223 ± 0.014, (b − y) =<br />
0.766 ± 0.008, (v − b) = 0.624 ± 0.010, (u − b) = 1.204 ± 0.014, m1 = −0.142 ± 0.014<br />
and c1 = −0.044 ± 0.022. They also computed the following <strong>de</strong>-re<strong>de</strong>nned val-<br />
ues: V = y = 11.223 ± 0.014, (b − y) = 0.743 ± 0.008, (v − b) = 0.597 ± 0.010,<br />
(u − b) = 1.202 ± 0.014, m1 = −0.146 ± 0.014 and c1 = −0.008 ± 0.022. Finally,<br />
Kilkenny (1993) reports the following Hβ photometry: β = 2.576 ± 0.010.<br />
2.5.2 Spectral type<br />
Kilkenny (1993) was the first one to suggest a spectral type for LS 5039. He classified<br />
it as an O7IV star based on Strömgren and Hβ photometry. Later on, Motch et al.<br />
(1997) took a low resolution λλ 3800–4700 (all λ are expressed in ˚A) spectrum and<br />
a medium resolution red spectrum. Among O stars, those showing N iii in emission<br />
at λλ 4630–34 and He ii strongly in emission at λ 4686 are classified as Of stars.<br />
When the N iii is in emission and the He ii is weakly in absorption or emission they<br />
are classified as O(f). Finally, when the N iii is in emission and the He ii is strongly<br />
in absorption they are classified as O((f)). Motch et al. (1997) reported an O7V<br />
star with well marked He ii λ 4684 absorption and some evi<strong>de</strong>nce for weak N iii<br />
λλ 4634–4642 emission, and therefore suggested an O7V((f)) classification. No Hα<br />
emission was present in their medium resolution red spectrum.<br />
Clark et al. (2001) obtained higher resolution optical and near infrared spectroscopy<br />
of LS 5039. All Balmer lines were in absorption, preventing hence a classification<br />
as an Oe/Be star. The red-end spectrum (λλ 5650–7500) contained interstellar<br />
lines and helium lines in absorption. The blue-end spectrum (λλ 4050–4950)<br />
showed hydrogen Balmer series and He ii lines. The ratio between He ii λ 4541 and<br />
He i λ 4471 correspon<strong>de</strong>d to a spectral type O6.5, while the presence of He ii 4686<br />
strongly in absorption indicated that LS 5039 is still in the main sequence. Finally,<br />
the presence of very weak emission from N iii λλλ 4634–40–42 allowed them to add<br />
the ((f)) classification. Therefore, they classified LS 5039 as an O6.5V((f)) star. Furthermore,<br />
their K band (2.05–2.20 µm) and H band (1.49–1.72 µm) spectroscopy<br />
was perfectly consistent with such classification, and they stated that there was no<br />
evi<strong>de</strong>nce from any additional emission feature that could be attributed to emission<br />
from a jet or accretion disk.
2.5. The optical/IR star: LS 5039 59<br />
2.5.3 Distance and its uncertainty<br />
Motch et al. (1997) proposed a distance of 3.1 kpc to LS 5039 based on the star<br />
color excess, and stated that this estimate could have a large uncertainty. This value<br />
was computed assuming an O7V((f)) spectral type. However, as we have just seen,<br />
recent observations by Clark et al. (2001) show that the optical companion has a<br />
spectral type of O6.5V((f)). On the other hand, old calibrations for the intrinsic<br />
color in<strong>de</strong>x and absolute magnitu<strong>de</strong> (Johnson 1966, Deutschman et al. 1976) were<br />
used by Motch et al. (1997). Hence, we have performed a new estimate of the<br />
distance taking into account the new spectral type and more recent calibrations.<br />
The optical photometry available up to now and containing at least B and V<br />
magnitu<strong>de</strong>s (see Table 2.5) comes from Drilling (1991), Lahulla & Hilton (1992)<br />
and Clark et al. (2001), and is listed in Table 2.6. Using an intrinsic color in<strong>de</strong>x<br />
of (B − V )0 = −0.30 ± 0.02 for an O6.5V star (Schaerer et al. 1996, Lejeune &<br />
Schaerer 2001), we can compute the color excess E(B − V ) for all the observations.<br />
Finally, using the relationship AV = [3.30 + 0.28 (B − V )0 + 0.04 EB−V ] E(B − V )<br />
(Schmidt-Kaler 1982) and MV = −4.99 ± 0.3 for an O6.5V star (Vacca et al. 1996)<br />
we obtain the distance estimates listed in Table 2.6. The weighted mean of these<br />
values is 2.9 ± 0.3 kpc, and is the distance to LS 5039 used throughout this work.<br />
Finally, assuming that the Sun is at 8.5 kpc from the galactic center, we obtain a<br />
galactocentric distance of 5.8 ± 0.3 kpc for LS 5039.<br />
Table 2.6: Distance estimates to LS 5039 <strong>de</strong>rived from optical photometry.<br />
V B − V E(B − V ) AV d [kpc]<br />
11.23 ± 0.01 (a) 0.95 ± 0.01 (a) 1.25 ± 0.02 4.1 ± 0.1 2.68 ± 0.06<br />
11.20 ± 0.02 (b) 0.95 ± 0.02 (b) 1.25 ± 0.03 4.1 ± 0.1 2.64 ± 0.06<br />
11.33 ± 0.02 (c) 0.85 ± 0.02 (c) 1.15 ± 0.03 3.8 ± 0.1 3.26 ± 0.07<br />
11.32 ± 0.01 (d) 0.85 ± 0.02 (d) 1.15 ± 0.03 3.8 ± 0.1 3.25 ± 0.07<br />
(a) Drilling (1975, 1991).<br />
(b) Lahulla & Hilton (1992).<br />
(c) 1996 October observations by Clark et al. (2001).<br />
(d) 2000 September observations by Clark et al. (2001).
60 Chapter 2. Multiwavelength approach to LS 5039<br />
2.5.4 Radial velocity curve<br />
Optical spectroscopy aimed to reveal the radial velocity curve of LS 5039 was carried<br />
out by McSwain et al. (2001). They performed the observations during three runs<br />
in 1998 August, 1999 June and 2000 October using the 0.9 m coudé feed telescope at<br />
Kitt Peak National Observatory, that provi<strong>de</strong>d a resolution R = λ/δλ = 9500. The<br />
obtained radial velocity data appeared to have minima every ∼ 4 days. Therefore,<br />
the data set was then analyzed in the search for an orbital period, and the following<br />
orbital elements were <strong>de</strong>rived after a least-squares fit: P = 4.117 ± 0.011 d, T =<br />
JD 2451822.12±0.09, K = 14.7±0.9 km s −1 , V0 = 4.6±0.5 km s −1 , e = 0.41±0.05,<br />
w = 217 ± 9 ◦ , rms = 3.3 km s −1 , f(m) = 0.00103 ± 0.00020 M⊙, and a1 sin i =<br />
1.09 ± 0.07 km s −1 . They also obtained a projected rotational velocity of V sin i =<br />
131 ± 6 km s −1 . The epoch T corresponds to the time of periastron, and the epoch<br />
of inferior conjunction of the optical star (which would correspond to the time of an<br />
X-ray eclipse if one occurs) is 0.25 days later, corresponding to a phase of 0.06. We<br />
show in Fig. 2.11 the radial velocity curve obtained by McSwain et al (2001).<br />
RADIAL VELOCITY (km s -1<br />
)<br />
30<br />
20<br />
10<br />
0<br />
-10<br />
-20<br />
-30<br />
0.0 0.5 1.0<br />
ORBITAL PHASE<br />
Figure 2.11: Radial velocity measurements and orbital solution plotted against orbital<br />
phase. Phase zero corresponds to periastron (from McSwain et al. 2001).
2.5. The optical/IR star: LS 5039 61<br />
Since no Hα emission was seen, the star could not fill its Roche lobe even at<br />
periastron, and consi<strong>de</strong>ring a radius of R = 10 R⊙ for an O7V star, they directly<br />
inferred a minimum mass of 15 M⊙ for the optical star, compatible with the expected<br />
mass of 36 M⊙. On the other hand, in or<strong>de</strong>r to avoid breakup speed of the star, the<br />
lowest inclination acceptable would be i 9 ◦ , and this would limit the maximum<br />
mass for the compact object to be MX 8 M⊙.<br />
We note that then it was unknown the improvement on the spectral type classifi-<br />
cation as an O6.5V((f)) star. However, the <strong>de</strong>rived values do not change significantly.<br />
In fact, consi<strong>de</strong>ring a mass of 40 M⊙ for an O6.5V((f)) star, as we have discussed<br />
in Sect. 2.5.3, we obtain a maximum mass for the compact object of 9 M⊙.<br />
Spectroscopic observations with the INT<br />
With the aim to confirm the results obtained by McSwain et al (2001), we have<br />
conducted new medium resolution spectroscopic observations of LS 5039 during 9<br />
consecutive nights. We have observed 17 LS 5039 using the Intermediate Dispersion<br />
Spectrograph (IDS) attached to the 2.5 m Isaac Newton Telescope (INT) 18 on the<br />
nights of 2002 July 23–31. A total of 96 spectra were obtained with the combination<br />
of the 235 mm camera and the R900V grating, which provi<strong>de</strong>d a useful wavelength<br />
coverage (free from vignetting) of λλ 3900–5500. The seeing was variable (1 ′′ –2 ′′ )<br />
during our run and we used a 1.2 ′′ slit which resulted in a resolution of 83 km s −1<br />
(FWHM). In addition, we also obtained five spectra of LS 5039 with the holografic<br />
grating H2400B for the purpose of measuring the rotational broa<strong>de</strong>ning (V sin i)<br />
of the companion’s absorption lines. The spectral resolution of these spectra was<br />
30 km s −1 and we varied the central wavelength in or<strong>de</strong>r to fully cover the range<br />
λλ 3900–5100, where most of the prominent He i and He ii lines lie. The spectral<br />
type standards HR 8622 (O9V spectral type) and HD 168075 (O7V((f)) spectral<br />
type) were also observed with the same instrumental configurations. A complete log<br />
of the observations is provi<strong>de</strong>d in Table 2.7.<br />
The images were <strong>de</strong>-biased and flat-fiel<strong>de</strong>d, and the spectra subsequently ex-<br />
tracted using conventional optimal extraction techniques in or<strong>de</strong>r to optimize the<br />
signal-to-noise ratio of the output (Horne 1986). CuAr and CuNe comparison lamp<br />
17 Casares, J., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, in preparation.<br />
18 The INT is operated on the island of La Palma by the Royal Greenwich Observatory in the<br />
Spanish Observatorio <strong>de</strong>l Roque <strong>de</strong> Los Muchachos of the Instituto <strong>de</strong> Astrofísica <strong>de</strong> Canarias.
62 Chapter 2. Multiwavelength approach to LS 5039<br />
Table 2.7: Log of the spectroscopic observations carried out with the INT.<br />
Date Object Exp. Time Wave. Range Dispersion<br />
(2002 July) [s] [λλ] [˚A pix −1 ]<br />
23 LS 5039 1×600 4045–4910 0.85<br />
24 ,, 1×600 ,, ,,<br />
25 ,, 1×750 ,, ,,<br />
26 ,, 1×600 ,, ,,<br />
27 ,, 5×300 3900–5500 0.63<br />
” ,, 7×600 ,, ,,<br />
” HR 8622 2×60 ,, ,,<br />
28 LS 5039 3×300 ,, ,,<br />
” ,, 6×600 ,, ,,<br />
29 LS 5039 1×300 ,, ,,<br />
” ,, 12×600 ,, ,,<br />
” ,, 11×900 ,, ,,<br />
30 ,, 1×300 ,, ,,<br />
” ,, 17×600 ,, ,,<br />
” ,, 4×900 ,, ,,<br />
” HD 168075 1×600 ,, ,,<br />
31 LS 5039 1×300 ,, ,,<br />
” ,, 23×600 ,, ,,<br />
” ,, 3×600 3900–4550 0.23<br />
” HD 168075 1×200 ,, ,,<br />
” LS 5039 2×600 4550–5100 0.22<br />
” HD 168075 1×200 ,, ,,<br />
images were obtained every 15–30 minutes, and the λ-pixel scale was <strong>de</strong>rived through<br />
4/6th-or<strong>de</strong>r polynomial fits to 53/86 lines (<strong>de</strong>pending on the set-up), resulting in an<br />
rms scatter < 0.04 ˚A. The calibration curves were interpolated linearly in time.<br />
All the spectra were rectified, by fitting a low-or<strong>de</strong>r spline to the continuum, and<br />
rebinned into a uniform velocity scale of 83 km s−1 . The continuum was subsequently<br />
subtracted and every individual spectrum of the target was then cross-correlated<br />
with the template HD 168075, after masking the interstellar absorption lines and<br />
bands at λ 3934 (CaII K), λ 4430, λ 4501, λ 4726, λ 4762, λ 4885, λ 5449 and<br />
λλ 5487–5550. The resulting velocities are presented in Fig. 2.12, where it is clearly<br />
seen a night-to-night variability that suggests a periodicity of 4 d, in agreement<br />
with the results of McSwain et al. (2001).
2.5. The optical/IR star: LS 5039 63<br />
Radial velocity [km/s]<br />
60<br />
40<br />
20<br />
0<br />
−20<br />
−40<br />
52479.0 52481.0 52483.0 52485.0 52487.0<br />
Modified Julian Date [JD−2400000.5]<br />
Figure 2.12: Radial velocity measurements obtained with the INT on 2002 July 23–31.<br />
Although our data only span 9 nights, we have performed a timing analysis<br />
in or<strong>de</strong>r to search for possible periods, in particular the proposed ∼ 4 d orbital<br />
period. The methods employed were the Phase Dispersion Minimization (PDM)<br />
(Stellingwerf 1978) and the CLEAN algorithm (Roberts et al. 1987). The two<br />
obtained periodograms show a clear period at 4.0 ± 0.2 d, and the corresponding<br />
1-day alias at 1.33 ± 0.3 d with a similar significance level.<br />
We note that an O6.5V star would not fit in the Roche lobe for a period below<br />
∼ 2 d. A possibility for the binary to have a ∼ 1.3 d orbital period would be if<br />
LS 5039 were a subdwarf. However, with such an orbital period, we would expect<br />
to find a completely circularized orbit (i.e., e = 0), which is not the case if we fold<br />
our measurements with the ∼ 1.3 d period. Therefore, the 4.0 ± 0.2 d period seems<br />
to be the correct one, thus confirming the earlier results by McSwain et al. (2001).
64 Chapter 2. Multiwavelength approach to LS 5039<br />
Radial velocity [km/s]<br />
30<br />
20<br />
10<br />
0<br />
−10<br />
−20<br />
−30<br />
0.0 0.5 1.0<br />
Orbital phase<br />
Figure 2.13: Radial velocity measurements obtained with the INT on 2002 July 23–31,<br />
fol<strong>de</strong>d using the ephemeris of McSwain et al. (2001).<br />
We show in Fig. 2.13 our obtained data with the INT fol<strong>de</strong>d with the ephemeris<br />
of McSwain et al. (2001), and plotted using the same X, Y intervals as in Fig. 2.11.<br />
While with those ephemeris the minimum takes place at phase 0.96 (see Fig. 2.11),<br />
here it occurs at phase ∼0.1–0.2. This indicates that the period used is not accurate<br />
enough, which is compatible with the error quoted in the McSwain et al. (2001)<br />
ephemeris (possible improvement of ephemeris is discussed in Casares et al. 2002).<br />
In or<strong>de</strong>r to obtain the rotational broa<strong>de</strong>ning, we have followed the technique<br />
applied to A0620-00 by Marsh et al. (1994) and <strong>de</strong>scribed in their paper. Essentially,<br />
we subtract different broa<strong>de</strong>ned versions of a O7V((f)) template (HD 168075) from<br />
the Doppler corrected sum of our LS 5039 spectra and perform a χ 2 test on the<br />
residuals. The O7V spectrum was broa<strong>de</strong>ned by convolution with the rotational<br />
profile of Gray (1976) which assumes a linearized limb darkening coefficient ɛ. We<br />
have taken ɛ=0.20 which is appropriate for the stellar parameters of our star (Teff <br />
35000 K, log g = 4) in the B-band. We performed the analysis in<strong>de</strong>pen<strong>de</strong>ntly for<br />
the two groups of spectra at different resolutions, 83 and 30 km s −1 , and we find
2.6. The X-ray counterpart: RX J1826.2−1450 65<br />
that the minimum χ 2 is achieved for the spectrum broa<strong>de</strong>ned by σbroad = 105 ± 1<br />
and 126±12 km s −1 , respectively. Since the template has a rotational broa<strong>de</strong>ning of<br />
σHD 168075 = 79 ± 3 km s −1 (Penny 1996), we need to sum this quadratically to σbroad<br />
in or<strong>de</strong>r to get the true rotational broa<strong>de</strong>ning of LS 5039. Therefore, we obtain<br />
131 ± 3 and 149 ± 12 km s −1 for the two resolutions. These numbers are in excellent<br />
agreement with the value V sin i = 131 ± 6 km s −1 obtained McSwain et al. (2001).<br />
2.6 The X-ray counterpart: RX J1826.2−1450<br />
The X-ray binary name that appears in the Liu et al. (2000) catalog of HMXBs,<br />
which always comes from X-ray satellites, is RX J1826.2−1450, after the ROSAT All<br />
Sky Suvey (RASS). However, in the <strong>de</strong>finitive version of the ROSAT Bright Source<br />
Catalog (RBSC) it is listed as 1RXS J182615.1−145034. In this section we present<br />
the long-term X-ray lightcurve obtained by ASM/RXTE, pointed observations by<br />
PCA/RXTE and pointed observations by BeppoSAX.<br />
2.6.1 The ASM/RXTE data<br />
As we have mentioned in Sect. 2.4.3, LS 5039 was inclu<strong>de</strong>d at our request in the<br />
catalog of the All Sky Monitor (ASM), onboard the Rossi X-ray Timing Explorer<br />
(RXTE), in or<strong>de</strong>r to follow its behavior in the X-ray domain. An analysis of the<br />
first ∼ 33 months of ASM/RXTE data was presented in Ribó et al. (1999). Here<br />
we report ASM/RXTE 19 observations spanning almost six and a half years (MJD<br />
50136.78–52501.41), from 1996 February to 2002 August. The dataset contains more<br />
than 1800 daily flux measurements in the energy range 1.5–12 keV. Each data point<br />
represents the one-day average of the fitted source fluxes from a number (typically 5–<br />
10) of individual ASM dwells, of ∼90 s each (see Levine et al. 1996 for more <strong>de</strong>tails).<br />
The one-day average light curve is shown in Fig. 2.14. The big gap between Modified<br />
Julian Date ∼ 50400 and ∼ 50500 corresponds to the passage of the Sun close to<br />
the source during the first year of observations. This gap repeats the following years<br />
(near MJD 50800, 51150, 51900 and 52250). That is the reason why we have a<br />
dataset of only 1800 one-day average points spanning nearly ∼ 2400 days.<br />
19 Quick-look results provi<strong>de</strong>d by the ASM/RXTE team.
66 Chapter 2. Multiwavelength approach to LS 5039<br />
Most of time the source is at the threshold of ASM <strong>de</strong>tectability. Nevertheless, we<br />
have searched for possible periodicities in the range from 2 to 1000 d. The methods<br />
employed were the Phase Dispersion Minimization (PDM) (Stellingwerf 1978) and<br />
the CLEAN algorithm (Roberts et al. 1987). Our approach here is essentially the<br />
same as in Pare<strong>de</strong>s et al. (1997) when analyzing the periodic behavior in the X-ray<br />
lightcurve of LS I +61 303.<br />
ASM units [count/s]<br />
6<br />
4<br />
2<br />
0<br />
6<br />
4<br />
2<br />
0<br />
ASM units [count/s] 50100 50150 50200 50250 50300 50350 50400 50450 50500 50550 50600<br />
6<br />
4<br />
2<br />
0<br />
ASM units [count/s] 50600 50650 50700 50750 50800 50850 50900 50950 51000 51050 51100<br />
6<br />
4<br />
2<br />
0<br />
ASM units [count/s] 51100 51150 51200 51250 51300 51350 51400 51450 51500 51550 51600<br />
6<br />
4<br />
2<br />
0<br />
52100 52150 52200 52250 52300 52350 52400 52450 52500 52550 52600<br />
Modified Julian Date [JD−2400000.5]<br />
ASM units [count/s] 51600 51650 51700 51750 51800 51850 51900 51950 52000 52050 52100<br />
Figure 2.14: ASM/RXTE one-day average lightcurve of RX J1826.2−1450 / LS 5039<br />
in the 1.5–12 keV energy band.<br />
After applying both the PDM and CLEAN methods to the ASM data, a period<br />
of ∼ 52 d is found. This periodicity corresponds to the <strong>de</strong>tection of some kind of<br />
active events that appear rather evi<strong>de</strong>nt at first glance in Fig. 2.14. Nevertheless,
2.6. The X-ray counterpart: RX J1826.2−1450 67<br />
a careful inspection of the data reveals a suspicious <strong>de</strong>tail. All those active events<br />
take place when the data by dwell coverage is rather poor (less than 5 dwells per<br />
day), thus reducing the statistical significance of the corresponding one-day average.<br />
For some instrumental reason, the ASM coverage becomes poorer than normal every<br />
∼ 52 d or so and, in the case of a weak X-ray source like RX J1826.2−1450, this can<br />
affect somehow the period analysis. Therefore, the ∼ 52 d period is very likely to<br />
be an instrumental artifact. In<strong>de</strong>ed, after removing all daily points resulting from<br />
less than 5 dwells (∼ 26% of total), the timing analysis reveals no significant period<br />
in the range from 2 to 1000 d when comparing the results from both periodograms.<br />
We point out that we have not attempted to fold the ∼ 6.5 years of ASM/RXTE<br />
data using the ephemeris by McSwain et al. (2001), since the orbital period is not<br />
accurate enough (as we have seen in Sect. 2.5.4), and any small <strong>de</strong>viation from the<br />
true period would completely modify the obtained fol<strong>de</strong>d lightcurve due to the span<br />
of the data.<br />
2.6.2 PCA/RXTE observations<br />
We conducted 20 <strong>de</strong>tailed X-ray observations with the Proportional Counter Array<br />
(PCA) onboard RXTE 21 on 1998 February 8 and 16 (MJD 50852 and MJD 50860).<br />
The total on-source integration time was 20 ks. The PCA is sensitive to X-rays in<br />
the energy range 2–60 keV and comprises five i<strong>de</strong>ntical co-aligned gas-filled propor-<br />
tional counter units (PCUs), providing a total collecting area of ∼ 6500 cm 2 , an<br />
energy resolution of
68 Chapter 2. Multiwavelength approach to LS 5039<br />
Table 2.8: Log of the PCA/RXTE observation of RX J1826.2−1450.<br />
Date Start time Stop time MJD Flux (a) Luminosity (a,b)<br />
[TT] (c) [TT] (c) [erg s −1 cm −2 ] [erg s −1 ]<br />
08/02/98 01:05:33 04:18:14 50852.1 5.6 × 10 −11 5.6 × 10 34<br />
08/02/98 20:07:34 00:57:14 50852.9 4.6 × 10 −11 4.6 × 10 34<br />
16/02/98 18:42:25 20:35:14 50860.8 4.6 × 10 −11 4.6 × 10 34<br />
(a) In the energy range 3–30 keV. (b) For an assumed distance of 2.9 kpc.<br />
(c) Terrestrial Time = International Atomic Time + 32.184 s.<br />
30<br />
25<br />
20<br />
15<br />
0<br />
25<br />
5000 10000 15000<br />
20<br />
15<br />
10<br />
25<br />
20<br />
15<br />
10<br />
70000 75000 80000<br />
755000 760000 765000<br />
Figure 2.15: 3–30 keV lightcurve of RX J1826.2−1450 / LS 5039 covering the entire<br />
PCA/RXTE observation. Time 0 is JD 2450852.547 and the bin size is 180 s.
2.6. The X-ray counterpart: RX J1826.2−1450 69<br />
Figure 2.16: Characteristic power spectrum of RX J1826.2−1450 / LS 5039. The dashed<br />
line represents the 95% confi<strong>de</strong>nce <strong>de</strong>tection limit.<br />
Timing<br />
The PCA observations were used to study the time variability on various timescales.<br />
Continuous stretches of clean data were selected from the light curve of the entire<br />
observation. To reduce the variance of the noise powers, these intervals were divi<strong>de</strong>d<br />
up into segments of 8192 bins each, with a bin size of 10 ms. Then the power <strong>de</strong>nsity<br />
spectra for each segment were calculated and the results averaged together. Fig. 2.16<br />
shows the characteristic power spectrum in the frequency range 0.01–50 Hz. The<br />
dashed line represents the 95% confi<strong>de</strong>nce <strong>de</strong>tection limit (van <strong>de</strong>r Klis 1989). As it<br />
can be seen no power exceeds this value; the distribution of powers is flat at a level<br />
of 2, consistent with Poissonian counting statistics. Following van <strong>de</strong>r Klis (1989),<br />
we can set a 95% upper limit of 60% on the rms of a pulsed source signal in the<br />
range 0.01–50 Hz. This relatively high limit is a consequence of the faintness of the<br />
source. On longer timescales, longer intervals (∼ 3200 s) were consi<strong>de</strong>red but no<br />
evi<strong>de</strong>nce for pulsations was found either.<br />
Likewise, we fol<strong>de</strong>d the lightcurve onto a set of trial periods with the FTOOLS<br />
software package (a technique very similar to the PDM) and looked for a peak in<br />
the χ 2 versus period diagram. None of the peaks found were statistically significant<br />
enough. Thus, we conclu<strong>de</strong> that no coherent periodicities were <strong>de</strong>tected in the range<br />
∼ 0.02 to ∼ 2000 s.
70 Chapter 2. Multiwavelength approach to LS 5039<br />
The mean X-ray intensity in the energy range 3–30 keV shows a slight <strong>de</strong>creasing<br />
trend with 19.7 ± 0.2 count s −1 at the beginning of the observation (upper panel of<br />
Fig. 2.15) compared to 16.6 ± 0.1 count s −1 and 16.2 ± 0.2 count s −1 for the middle<br />
and bottom panels of Fig. 2.15, respectively. The fractional rms of the 3–30 keV<br />
lightcurve corresponding to the entire observation is 9%.<br />
The fact that no X-ray pulsations have been found in RX J1826.2−1450 is con-<br />
sistent with the proposed i<strong>de</strong>a that radio emission and X-ray pulsations from X-ray<br />
binaries seem to be statistically anti-correlated (Fen<strong>de</strong>r et al. 1997, Fen<strong>de</strong>r & Hendry<br />
2000), i.e., no X-ray pulsar has ever shown significant radio emission.<br />
We have not computed the orbital phase of each one of the PCA/RXTE obser-<br />
vations according to the ephemeris by McSwain et al. (2001), since the ∼ 1000 d<br />
difference between their T value and our observations is too high to obtain reli-<br />
able results due to the lack of accuracy for the orbital period (as we have seen in<br />
Sect. 2.5.4).<br />
Spectrum<br />
Since the lightcurve of the entire observation does not show sharp features, i.e.,<br />
there is no significant spectral change throughout the observation, we obtained one<br />
average PCA energy spectrum from the complete observation, that can be seen in<br />
Fig. 2.17.<br />
Acceptable fits of the X-ray continuum were obtained with an unabsorbed power-<br />
law mo<strong>de</strong>l, giving a reduced χ 2 =1.14 for 56 <strong>de</strong>grees of freedom (dof). A multicolor<br />
disk mo<strong>de</strong>l, as expected from an optically thick accretion disk (Mitsuda et al. 1984)<br />
plus a power-law gave a reduced χ2 ν =1.11 for 53 dof. Bremsstrahlung and two blackbody<br />
component mo<strong>de</strong>ls did not fit the data. Although the addition of a blackbody<br />
component to the power-law formally produces an acceptable fit, the value of the<br />
blackbody normalization was very low, with the error bar close to zero. In fact, an<br />
F-test shows that the inclusion of a blackbody component is not significant. The<br />
most salient feature that appears in the spectrum of RX J1826.2−1450 is a strong<br />
iron line at ∼ 6.6 keV (Fig. 2.17). A Gaussian fit to this feature gives a line centered<br />
at 6.62±0.04 keV, with an equivalent width (EW ) of 0.75±0.06 keV and a F W HM<br />
of 0.9 ± 0.2 keV. The best-fit results are given in Table 2.9.
2.6. The X-ray counterpart: RX J1826.2−1450 71<br />
Figure 2.17: PCA spectrum of RX J1826.2−1450 / LS 5039. The continuous line<br />
represents the best-fit power-law mo<strong>de</strong>l plus a Gaussian component for the iron line.<br />
When this component is omitted the line shows up clearly in the residuals.<br />
Table 2.9: Spectral fit results for the power-law mo<strong>de</strong>l plus a Gaussian component for<br />
the iron line in RX J1826.2−1450 / LS 5039. Uncertainties are given at 90% confi<strong>de</strong>nce<br />
for one parameter of interest. The spectrum was fitted in the energy range 3–30 keV.<br />
Parameters Fitted values<br />
NH [×10 21 cm −2 ] 2 +1<br />
−2<br />
Γ 1.95 ± 0.02<br />
Eline(Fe) [keV] 6.62 ± 0.04<br />
EWline(Fe) [keV] 0.75 ± 0.06<br />
F W HMline(Fe) [keV] 0.9 ± 0.2<br />
χ 2 r (dof) 1.14 (56)
72 Chapter 2. Multiwavelength approach to LS 5039<br />
In HMXBs the presence of an emission line around 6.4 keV is commonly seen,<br />
with a variety of strengths. It is ascribed to fluorescent reprocessing by neutral or<br />
mildly ionized iron in relatively cold circumstellar material (Fe Kα line). A higher<br />
energy implies the presence of highly ionized iron (He-like ∼ FeXXV) or a different<br />
origin than Kα emission. One possibility would be radiative recombinations of H-<br />
like iron followed by casca<strong>de</strong> processes in a relatively hot corona. Unfortunately, the<br />
PCA energy resolution prevents from distinguishing between these possible mo<strong>de</strong>ls.<br />
On the other hand, the high EWline(Fe) value indicates that a large amount of<br />
circumstellar matter is present in the system. Since LS 5039 lies at low galactic<br />
latitu<strong>de</strong>, we cannot rule out that part of the emission feature comes from the Galactic<br />
ridge (Yamauchi & Koyama 1993). Nevertheless, this contribution cannot account<br />
for the total flux of the line. In addition, Wang et al. (2002) have recently shown<br />
that the ∼ 6.7 keV emission near the Galactic Center region arises from discrete<br />
sources rather than from the diffuse continuum, giving support to the i<strong>de</strong>a that the<br />
∼ 6.7 keV line comes from LS 5039.<br />
A hydrogen column <strong>de</strong>nsity of ∼ 2 +1<br />
−2 × 10 21 cm −2 is found from the fit. This<br />
value is, however, not very well constrained. In fact, it is consistent with zero. The<br />
difficulty in constraining the hydrogen column <strong>de</strong>nsity from our X-ray data can be<br />
attributed to the fact that the interstellar gas mainly absorbs X-ray photons with<br />
energies lower than 2–3 keV, i.e., outsi<strong>de</strong> of the energy range consi<strong>de</strong>red here. We<br />
note that from the λ 4430 and λ 6284 interstellar bands Motch et al. (1997) found<br />
E(B − V ) = 0.8 ± 0.2. Using the relation NH = 5.3 × 10 21 cm −2 E(B − V ) (Pre<strong>de</strong>hl<br />
& Schmitt 1995), we obtain that NH ∼ (4 ± 1) × 10 21 cm −2 , which is consistent,<br />
within the errors, with the X-ray observation value. However, if we consi<strong>de</strong>r the<br />
obtained photometric values for E(B − V ) quoted in Table 2.6 ( 1.2 ± 0.1), we<br />
obtain NH (6.4 ± 0.5) × 10 21 cm −2 , which seems clearly inconsistent with the<br />
results after the PCA/RXTE observations. In any case, the lack of sensitivity of<br />
the <strong>de</strong>tector below 3 keV may explain why we have <strong>de</strong>tected such low values of NH.<br />
Possible nature of the compact object based on the X-ray data<br />
The non <strong>de</strong>tection of X-ray pulses does not necessarily mean that RX J1826.2−1450<br />
/ LS 5039 contains a black hole, since it could contain a low magnetic field neutron<br />
star. Both types of objects have been found in REXBs.
2.6. The X-ray counterpart: RX J1826.2−1450 73<br />
On the other hand, the multicolor disk mo<strong>de</strong>l, although formally fitting the data,<br />
does not help either. First, given the low luminosity (LX < 10 35 erg s −1 in the energy<br />
range 2–10 keV) the system would be in the so-called low/hard state. We would<br />
not expect then to <strong>de</strong>tect a strong soft component. Second, the fit provi<strong>de</strong>s an<br />
unrealistic value of the disk internal radius of Rin cos 1/2 (θ) ∼ 0.3 km.<br />
Unfortunately, the source is too faint at energies above 30 keV to be <strong>de</strong>tected<br />
with the HEXTE/RXTE instrument. Thus, we cannot confirm from the present<br />
data whether the hard tail that usually characterizes the energy spectrum of black<br />
holes at high energies is in<strong>de</strong>ed present.<br />
2.6.3 BeppoSAX observations<br />
With the aim to confirm and improve the obtained results after the PCA/RXTE<br />
observations, we observed 22 RX J1826.2−1450 / LS 5039 with the BeppoSAX X-<br />
ray satellite during ∼ 80 ksec on 2000 October 8 (MJD 51825.29–51826.21). Un-<br />
fortunately, while the flux <strong>de</strong>tected during the PCA/RXTE observations was of<br />
∼ 5 × 10 −11 erg s −1 cm −2 , it was only of ∼ 5 × 10 −12 erg s −1 cm −2 during the<br />
BeppoSAX observations, i.e., one or<strong>de</strong>r of magnitu<strong>de</strong> lower. Hence, we could not<br />
confirm the <strong>de</strong>tection of the ∼ 6.7 keV iron line due to lack of signal-to-noise ra-<br />
tio. However, we could improve the value of the hydrogen column <strong>de</strong>nsity, NH =<br />
1.0 ± 0.3 × 10 22 cm −2 , and obtain a new value of the photon in<strong>de</strong>x, Γ = 1.8 ± 0.2 (in<br />
the energy range 3–10 keV).<br />
Using the relation by Pre<strong>de</strong>hl & Schmitt (1995) the NH value implies E(B−V ) =<br />
1.9±0.6, nearly compatible with the E(B−V ) 1.2±0.1 values <strong>de</strong>rived from optical<br />
photometry (see Table 2.6) and much higher than the E(B − V ) 0.4 +0.2<br />
−0.4 value<br />
<strong>de</strong>rived from the PCA/RXTE observations, which was poorly constrained. On the<br />
other hand, the value of Γ is compatible with the one found in the PCA/RXTE<br />
observations, and typical of black holes in the low/hard state, although we note<br />
that McSwain & Gies (2002) have recently suggested a neutron star nature for the<br />
compact object in LS 5039.<br />
The BeppoSAX observations took place in JD 2451825.79–2451826.71, hence<br />
very close to the T =JD 2451822.12 value used in the McSwain et al. (2001)<br />
22 Reig, P., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, in preparation.
74 Chapter 2. Multiwavelength approach to LS 5039<br />
ephemeris. Therefore, the lack of accuracy in the orbital period is not important in<br />
this case in or<strong>de</strong>r to compute the orbital phase, which ranges from 0.89 to 0.11. As<br />
discussed in Sect. 2.5.4, the time of an X-ray eclipse if one occurs is near phase 0.06<br />
according to these ephemeris, and its maximum duration, for a star of 40 M⊙ and<br />
an inclination angle of 90 ◦ , is expected to be of ∼ 0.2 d, or ∼ 0.05 phase interval<br />
(McSwain, private communication). Hence, we observed the source just around the<br />
time of a possible X-ray eclipse, but there is no signature of such an event in the<br />
BeppoSAX data.<br />
The non-<strong>de</strong>tection of an X-ray eclipse is consistent with the proposed inclination<br />
of the system of ∼ 30 ◦ (McSwain & Gies 2002). We also note that McSwain &<br />
Gies (2002), based on the broad residual emission in the Hα profile, obtained a<br />
higher mass loss rate of the optical star in 1998 than in 2000, which may explain<br />
the difference in the <strong>de</strong>tected flux between the PCA/RXTE observations (carried<br />
out in year 1998) and the BeppoSAX observations (carried out in year 2000).<br />
Therefore, we conclu<strong>de</strong> that the source was intrinsically faint during these obser-<br />
vations, and that there seem to be no X-ray eclipses in RX J1826.2−1450 / LS 5039,<br />
in good agreement with the proposed inclination of this binary system.<br />
2.7 The γ-ray counterpart: 3EG J1824−1514<br />
The highly energetic processes that take place near compact objects in X-ray binaries<br />
are observable from radio to hard X-rays (Mirabel & Rodríguez 1999) and possibly<br />
beyond. Therefore, soon after the discovery of the microquasar nature of LS 5039, we<br />
inspected the most recent high-energy catalogs in or<strong>de</strong>r to find possible counterparts<br />
to LS 5039. While there was no counterpart in the COMPTEL (0.75–30 MeV)<br />
source catalog (Schönfel<strong>de</strong>r et al. 2000), there was a very interesting uni<strong>de</strong>ntified<br />
high-energy γ-ray source in the Third EGRET Catalog (E > 100 MeV) (Hartman<br />
et al. 1999), namely 3EG J1824−1514.<br />
In Fig. 2.18 we show the location map of 3EG J1824−1514, together with the<br />
X-ray sources from both the ROSAT All Sky Bright Source Catalog (RBSC) (Voges<br />
et al. 1999) and Faint Source Catalog (RFSC) (Voges et al. 2000), as well as the<br />
radio sources brighter than 20 mJy (for clarity) from the NRAO VLA Sky Survey<br />
(NVSS) (Condon et al. 1998). As can be seen, there is only a RBSC source, namely
2.7. The γ-ray counterpart: 3EG J1824−1514 75<br />
Galactic Latitu<strong>de</strong><br />
0.0<br />
-1.0<br />
-2.0<br />
ROSAT Bright<br />
ROSAT Faint<br />
NVSS (>20 mJy)<br />
17.0 16.0<br />
Galactic Longitu<strong>de</strong><br />
15.0<br />
Figure 2.18: Location map of 3EG J1824−1514. The contours represent, from insi<strong>de</strong><br />
to outsi<strong>de</strong>, the 50%, 68%, 95%, and 99% statistical probability that a γ-ray source lies<br />
within the given contour. The open big circles are ROSAT bright sources, the small<br />
open circles are the ROSAT faint sources and the filled circles are radio sources from the<br />
NVSS survey. The only source with X-ray and radio emission (filled circle insi<strong>de</strong> an open<br />
big circle, l = 16.88 ◦ and b = −1.29 ◦ ), and well insi<strong>de</strong> the 95% contour, is LS 5039.
76 Chapter 2. Multiwavelength approach to LS 5039<br />
RX J1826.2−1450 / LS 5039, within the 3EG J1824−1514 contours, and well insi<strong>de</strong><br />
the 95% confi<strong>de</strong>nce contour. Moreover, this is the only X-ray source, bright or faint,<br />
that also displays radio emission (addition of NVSS sources below 20 mJy does not<br />
change this discussion). Such a good position agreement between an EGRET source<br />
and a radio jet X-ray binary strongly suggests that both objects may be the same.<br />
Thus, this microquasar system is likely associated with an EGRET source.<br />
As can be seen in Fig. 2.19, the γ-ray emission observed from 3EG J1824−1514<br />
reveals a rather persistent flux of > 100 MeV photons for the last <strong>de</strong>ca<strong>de</strong>. This<br />
persistent behavior is also seen at radio wavelengths in LS 5039, thus giving support<br />
to the proposed association.<br />
s −1<br />
]<br />
photon cm −2<br />
γ−ray flux [10 −8<br />
80<br />
60<br />
40<br />
20<br />
EGRET (>100 MeV)<br />
VLA (1.45 GHz)<br />
GBI (2.25 GHz)<br />
0<br />
1988 1990 1992 1994 1996 1998 2000<br />
0<br />
2002<br />
Year<br />
Figure 2.19: Radio and γ-ray lightcurves of LS 5039 and 3EG J1824−1514. The radio<br />
flux <strong>de</strong>nsities are the VLA ones at 20 cm wavelength (1.45 GHz) quoted in Table 2.1<br />
and the GBI ones presented in Sect. 2.4.3. The error bars for the GBI data (±4 mJy)<br />
are not shown for clarity, whereas those of the VLA are usually smaller than the symbol<br />
size. The γ-ray fluxes correspond to the four viewing periods of EGRET.<br />
80<br />
60<br />
40<br />
20<br />
Radio flux <strong>de</strong>nsity [mJy]
2.7. The γ-ray counterpart: 3EG J1824−1514 77<br />
Figure 2.20: EGRET spectrum of 3EG J1824−1514 (from Hartman et al. 1999).<br />
The average γ-ray flux for all EGRET viewing periods in Fig. 2.19 is Φγ =<br />
(35.2±6.5)×10 −8 photon cm −2 s −1 , with photon spectral in<strong>de</strong>x p = 2.18±0.18, where<br />
Φγ ∝ E −p<br />
γ<br />
(see Fig. 2.20 where α = p). The corresponding integrated luminosity<br />
at a distance of 2.9 kpc (Ribó et al. 2002) amounts to Lγ(> 100 MeV) ∼ 3.6 ×<br />
10 35 erg s −1 , compared to an X-ray luminosity of LX(3–30 keV) ∼ 5 × 10 34 erg s −1<br />
(Ribó et al. (1999) with a distance of 2.9 kpc, see also Table 2.8) .<br />
We note that 3EG J1824−1514 is persistent but variable. In fact, its in<strong>de</strong>x of<br />
γ-ray variability as <strong>de</strong>fined in Torres et al. (2001) is I = 3. This value probably<br />
rules out a pulsar or a SNR nature for the EGRET source. Moreover, the EGRET<br />
photon in<strong>de</strong>x p = 2.18 ± 0.18 of LS 5039 is steeper than the p < 2 values usually<br />
found for pulsars (Merck et al. 1996). This gives further support to the association.
78 Chapter 2. Multiwavelength approach to LS 5039<br />
The importance of the association between LS 5039 and 3EG J1824−1514 is not<br />
only because of the discovery of high-energy γ-ray emission in a microquasar, but<br />
mainly because of the large number of still uni<strong>de</strong>ntified EGRET sources. The Third<br />
EGRET catalog contains a total of 271 sources, of which 103 are i<strong>de</strong>ntified objects:<br />
66 high-confi<strong>de</strong>nce AGN, 27 low-confi<strong>de</strong>nce AGN, 7 pulsars, 1 <strong>de</strong>tection from the<br />
Large Magellanic Cloud, 1 solar flare and the radiogalaxy Centaurus A. Among the<br />
remaining 168 uni<strong>de</strong>ntified sources (62%), 72 of them have |b| < 10 ◦ . Therefore, this<br />
association opens the possibility that some of the high-energy uni<strong>de</strong>ntified γ-ray<br />
sources at low galactic latitu<strong>de</strong>s could be silent microquasars to be discovered in the<br />
near future.<br />
+180<br />
Active Galactic Nuclei<br />
Uni<strong>de</strong>ntified EGRET Sources<br />
3EG J1824-1514 / LS 5039<br />
Third EGRET Catalog<br />
E > 100 MeV<br />
+90<br />
-90<br />
Pulsars<br />
LMC<br />
Solar FLare<br />
Figure 2.21: The Third EGRET source Catalog shown in galactic coordinates. The<br />
size of the symbol represents the highest intensity seen for this source by EGRET. The<br />
white cross insi<strong>de</strong> a black circle (l = 16.88 ◦ and b = −1.29 ◦ ) indicates the γ-ray source<br />
3EG J1824−1514, to be associated with LS 5039. Adapted from Hartman et al. (1999).<br />
-180
2.8. A proposed scenario to explain the multiwavelength behavior 79<br />
2.8 A proposed scenario to explain the multiwave-<br />
length behavior<br />
2.8.1 Spectral energy distribution<br />
In Fig. 2.22 we show the observed spectral energy distribution of LS 5039 ranging<br />
from radio wavelengths to γ-rays. We have quoted the domains within the electro-<br />
magnetic spectrum and the different names of LS 5039 according to each wavelength<br />
range.<br />
log (ν F ν [erg s −1 cm −2 ])<br />
−7<br />
−9<br />
−11<br />
−13<br />
−15<br />
−17<br />
NVSS J182614−145054<br />
LS 5039<br />
RADIO NIR/OPTIC<br />
RX J1826.2−1450<br />
3EG J1824−1514<br />
X−RAY GAMMA−RAY<br />
10 15 20 25<br />
log (ν [Hz])<br />
Figure 2.22: Observed spectral energy distribution of LS 5039. The radio data corre-<br />
spond to the fit in Eq. 2.1, the near infrared (NIR) and optical data are those quoted in<br />
Table 2.5, the X-ray data comes from the fit to the PCA/RXTE data shown in Fig. 2.17,<br />
and the γ-ray data comes from the fit to the spectrum shown in Fig. 2.20.
80 Chapter 2. Multiwavelength approach to LS 5039<br />
As we have already discussed, the radio emission originates as synchrotron ra-<br />
diation in a jet-like flow in the presence of a magnetic field. The near infrared and<br />
optical light arises from the high mass companion star. The X-rays probably arise<br />
in the accretion disk around the compact object. Finally, the γ-rays are probably<br />
produced by inverse Compton scattering of the ultraviolet photons of the companion<br />
star by the relativistic electrons present in the jet flow, as we discuss below.<br />
2.8.2 A mo<strong>de</strong>l based on the γ-ray/radio emission<br />
The movement of matter in the system is proposed to be as follows. The early type<br />
O6.5V((f)) star LS 5039 has a strong wind that accounts for a mass loss rate of<br />
∼ 10 −6 M⊙ yr −1 , according to McSwain & Gies (2002). While leaving the star,<br />
part of this matter will be captured by the intense gravitational field of the nearby<br />
compact object (probably a neutron star). The matter will then fall into the compact<br />
object and form a disk around it, which by viscous heating will emit in X-rays. Part<br />
of the X-ray emission can also be produced in a corona above the disk. A fraction<br />
of the matter forming the disk will be ejected perpendicular to it and in opposite<br />
senses, probably as a result of the Blandford & Payne (1982) mechanism, as seems to<br />
be the case in the quasar M87 (Junor et al. 1999). While flowing away into opposite<br />
jets, the relativistic electrons (matter) will be unavoidably exposed to a huge output<br />
of ultraviolet (UV) photons from the hot optical star. As a consequence, the UV<br />
photons will experience inverse Compton scattering, and γ-rays will be naturally<br />
produced, as <strong>de</strong>tected by EGRET. Later on, the relativistic electrons will produce<br />
synchrotron radio emission. This scenario is schematically reproduced in Fig. 2.23.<br />
It is interesting to use the presence of 5 GHz synchrotron emitting particles far<br />
away from the compact object, ∼ 1000 AU as seen in Sect. 2.4.5, in or<strong>de</strong>r to constrain<br />
the jet physical parameters. The starting point of the following calculations will be<br />
the scenario for the proposed EGRET emission of LS 5039 due to IC scattering. We<br />
recall that the energy shift in the IC process may be as high as Eγ ∼ 4γ 2 e EUV, where<br />
the energies of the γ-ray and the stellar UV photon are related through the squared<br />
Lorentz factor of the relativistic electron.<br />
We will adopt here a very simple mo<strong>de</strong>l of an expanding jet. Cylindrical coor-<br />
dinates, z and r measured parallel and perpendicular to the jet axis, are the best<br />
choice in our case. The jet is assumed to form at a distance z0 from the compact
Figure 2.23: Scenario where the multiwavelength emission originates in LS 5039.<br />
Synchrotron<br />
Radiation<br />
Inverse Compton<br />
Scattering<br />
X-ray<br />
L 3-30 keV ~ 5×10 34 erg/s<br />
v jet ≥ 0.15c<br />
e -<br />
e -<br />
e -<br />
e -<br />
e -<br />
e -<br />
γ-ray, E > 100 MeV, L γ ~ 4×10 35 erg/s<br />
γ e ~ 10 3<br />
Radio, L 0.1-100 GHz ~ 1×10 31 erg/s<br />
UV, E ~ 10 eV<br />
O6.5V((f))<br />
L opt ~ 1×10 39 erg/s<br />
2.8. A proposed scenario to explain the multiwavelength behavior 81
82 Chapter 2. Multiwavelength approach to LS 5039<br />
object and flows with a velocity v = β c. The lateral expansion of the jet is parame-<br />
terized as r = r0(z/z0) ɛ , where r0 is the initial jet radius. A freely expanding conical<br />
jet would correspond to ɛ = 1, while for slowed lateral expansion we would have<br />
ɛ < 1 (see e.g. Hjellming & Johnston, 1988). The expansion velocity perpendicular<br />
to the jet axis is thus v⊥ = dr/dt = ɛvr/z. Concerning the magnetic field, we will<br />
only consi<strong>de</strong>r the component perpendicular to the jet axis, because it has the slowest<br />
<strong>de</strong>cay if conservation of the magnetic flux is assumed. The magnetic field along the<br />
jet will be thus parameterized as B = B0(r/r0) −1 = B0(z/z0) −ɛ .<br />
A relativistic electron injected at the base of the jet will <strong>de</strong>crease its energy<br />
mainly through adiabatic expansion, IC and synchrotron losses according to:<br />
dE<br />
dt<br />
v⊥E<br />
= −2<br />
3 r − αICUradE 2 − αSB 2 E 2 . (2.8)<br />
The factor 2/3 in the adiabatic expansion term comes from the lateral expansion of<br />
the jet. The constants αIC = 3.97 × 10 −2 and αS = 2.37 × 10 −3 (c.g.s. units) are the<br />
coefficients of the terms accounting for IC and synchrotron losses, respectively. The<br />
radiation energy <strong>de</strong>nsity Urad is assumed to be dominated by stellar UV photons from<br />
the O6.5V star, whose luminosity above 10 eV amounts to L∗ 5 × 10 38 erg s −1<br />
and Urad = L∗/[4πc(a 2 + z 2 )]. With the recently <strong>de</strong>termined orbital parameters<br />
(McSwain et al. 2001, McSwain & Gies 2002), the likely semimajor axis of the<br />
orbit is a = 2.6 × 10 12 cm. Adopting these parameters, the radiation energy <strong>de</strong>nsity<br />
close to the compact object is so high that electrons with energies of 10 −3 –10 −2 erg<br />
(γe ∼ 10 3 –10 4 ) will naturally produce γ-ray photons with energies of 100–1000 MeV,<br />
as <strong>de</strong>tected by EGRET. Moreover, integration of Eq. 2.8 taking only into account<br />
the IC losses gives:<br />
E(t) =<br />
E0<br />
1 + αICL∗E0<br />
4πcav tan−1<br />
, (2.9)<br />
vt<br />
a<br />
which reveals that the energy of the electrons <strong>de</strong>cays very fast to values of ∼ 5 ×<br />
10 −4 erg, regardless of their initial energy (the expression above is asymptotic).<br />
If we consi<strong>de</strong>r the IC contribution in Eq. 2.8 to be comparable to adiabatic losses<br />
at the base of the jet (later on, the adiabatic expansion term will soon become<br />
dominant due to its slower <strong>de</strong>cay as z −1 ), at the base of the jet we can request the<br />
following:<br />
αIC<br />
L∗<br />
4πc(a 2 + z 2 0) E2 0<br />
2ɛ vE0<br />
<br />
3 z0<br />
. (2.10)
2.8. A proposed scenario to explain the multiwavelength behavior 83<br />
We can use the values quoted above, together with v = 0.2c and an energy of<br />
E0 = 5 × 10 −4 erg for the electrons injected at the base of the jet, to solve for z0 in<br />
this quadratic equation. The obtained results are 0.35, 0.85 and 1.30 AU for ɛ = 1,<br />
1/2 and 1/3, respectively. The synchrotron term can be shown to be negligible a<br />
posteriori and has not been consi<strong>de</strong>red here.<br />
Hereafter, we will proceed taking into account only the adiabatic expansion term,<br />
because it is dominant compared to the IC one, and rewrite Eq. 2.8 as:<br />
dE<br />
dt<br />
−2ɛ<br />
3<br />
vE<br />
z<br />
. (2.11)<br />
This can be easily integrated to give E E0(z/z0) −2ɛ/3 . As the energy <strong>de</strong>cays, the<br />
synchrotron emission of relativistic electrons will shift towards lower frequencies. In<br />
particular, the synchrotron frequency expected at a distance z is given by:<br />
νc = 6.27 × 10 18 BE 2 6.27 × 10 18 B0E 2 0 (z/z0) −7ɛ/3 . (2.12)<br />
The electrons still producing 6 cm (νc = 5 × 10 9 Hz) synchrotron emission at about<br />
z = 1000 AU (the size of the MERLIN jets) are probably those originally injected<br />
with energies of E0 5 × 10 −4 erg. Then, the magnetic field B0 at the base of the<br />
jet can be roughly estimated as:<br />
B0 8.0 × 10 −10 E −2<br />
0<br />
z<br />
z0<br />
7ɛ/3<br />
. (2.13)<br />
This provi<strong>de</strong>s values of 3.6 × 10 5 , 12 and 0.6 G for ɛ = 1, 1/2 and 1/3, respectively.<br />
If magnetic flux is conserved, the corresponding field at the end of the MERLIN<br />
jet would be B1000 AU = 128, 0.4 and 0.06 G. For comparison, the only in<strong>de</strong>pen<strong>de</strong>nt<br />
estimates of the magnetic field in the LS 5039 jets come from simple equipartition<br />
arguments. The results are merely indicative of the average magnetic field in the<br />
solid angle of the sky covered by the jets. Using the total flux <strong>de</strong>nsity and size of the<br />
MERLIN jets given in Table 2.3 we <strong>de</strong>rive a field value of > 8×10 −3 G, although this<br />
is only a lower limit because the jet width is unresolved. The VLBA image presented<br />
in Sect. 2.4.4, allowed us to estimate an approximate magnetic field of ∼ 0.2 G for<br />
radio jets reaching up to a few mas. It is clear that a field of such intensity is more<br />
consistent with the ɛ < 1 results than with the too high magnetic fields implied<br />
by a simple conical jet (ɛ = 1). Therefore, a slowly expanding jet, and hence well<br />
collimated, is favored to power the LS 5039 radio emission up to ∼ 1000 AU by<br />
the electrons that previously contributed to the γ-ray emission, without in situ<br />
acceleration being required. This is also in agreement with our upper limit of ≤ 6 ◦<br />
for the jet half opening angle <strong>de</strong>rived from the EVN map.
84 Chapter 2. Multiwavelength approach to LS 5039<br />
2.8.3 Energetic consi<strong>de</strong>rations<br />
The central engine in LS 5039 must be supplying at least ˙ Ee ∼ Lγ ∼ 3.6×10 35 erg s −1<br />
in the form of relativistic electrons to power the <strong>de</strong>tected γ-rays (see Sect. 2.7). The<br />
electron energy distribution is expected to be a power law kE 2α−1 dE, which can be<br />
estimated as kE −2 dE according to the observed spectral in<strong>de</strong>x α −0.5. Assuming<br />
electron energies in the range mec 2 ≤ E ≤ γmaxmec 2 we have:<br />
<br />
˙Ee = k<br />
E 2α−1 <br />
EdE k<br />
E −2 <br />
EdE = k<br />
E −1 dE = k ln γmax . (2.14)<br />
Therefore, if the proton mass is mp 1800 me = 1800E/c 2 for every relativistic<br />
electron, the proton mass flow into the jets can be written as:<br />
<br />
<br />
Mjet<br />
˙ = mpk<br />
E 2α−1 dE 1800E<br />
c2 k<br />
E −2 dE 1800<br />
k . (2.15)<br />
c2 From Eq. 2.14 we can obtain the value of the constant, k = ˙ Ee/ ln γmax, and rewrite<br />
Eq. 2.15 as:<br />
˙<br />
Mjet 1800 ˙ Ee<br />
c 2 ln γmax<br />
. (2.16)<br />
Using then ˙ Ee ∼ Lγ ∼ 3.6 × 10 35 erg s −1 and γmax ∼ 10 4 , we obtain:<br />
˙<br />
Mjet ∼ 7.8 × 10 16 g s −1 ∼ 1.2 × 10 −9 M⊙ yr −1 , (2.17)<br />
which is weakly <strong>de</strong>pen<strong>de</strong>nt on the maximum energy cutoff assumed (γmax). Con-<br />
si<strong>de</strong>ring a mass loss from the companion star of 10 −6 M⊙ yr −1 , as previously seen,<br />
this tells us that approximately a fraction of 1/1000 of the mass that leaved the<br />
companion is finally ejected forming relativistic jets perpendicular to the accretion<br />
disk. The equivalent kinetic luminosity is given by Lk = (Γ − 1) ˙<br />
Mjetc 2 , where Γ is<br />
the Lorentz factor of the bulk motion of the jet. Hence, the kinetic luminosity highly<br />
<strong>de</strong>pends on the velocity of the flow, and for values of β in the range 0.15–0.4, and<br />
hence Γ ∼ 1.01–1.1, we obtain Lk ∼ 10 36 –10 37 erg s −1 . This kinetic power is about<br />
four to five or<strong>de</strong>rs of magnitu<strong>de</strong> lower than that estimated for the strong ejections<br />
of the superluminal microquasar GRS 1915+105 (Mirabel & Rodríguez 1994), but<br />
interestingly much higher than the ∼ 5 × 10 34 erg s −1 X-ray luminosity of LS 5039.
2.8. A proposed scenario to explain the multiwavelength behavior 85<br />
2.8.4 Future prospects<br />
We carried out two runs of global VLBI observations of LS 5039 on 2000 June 2 and<br />
7, spanning a total of 14 hours each. Preliminary reduction of these data indicates<br />
the presence of the jets in a similar position angle as previously found with the<br />
VLBA, EVN and MERLIN. These data will be analyzed soon, and we expect to<br />
measure proper motions of individual components within the jets, that will help to<br />
constrain the jet parameters and improve our proposed mo<strong>de</strong>l.<br />
On the other hand, the high eccentricity of LS 5039 provi<strong>de</strong>s both, a variable<br />
accretion rate along the orbit implying a variable rate of electrons injected into the<br />
jet, and a variable radiation energy <strong>de</strong>nsity close to the compact object. Hence, we<br />
predict variability in the radio and γ-ray luminosities correlated with the orbital<br />
period, which may hopefully be <strong>de</strong>tected by INTEGRAL and/or GLAST missions<br />
at high energies, and with already scheduled VLA observations at low energies. On<br />
the other hand, if the star exhibits intrinsic variations in the UV photon flux, a<br />
variability in γ-rays would be seen correlated to it.<br />
Since LS 5039 seems to have persistent radio jets, it would be interesting to study<br />
the possibility of <strong>de</strong>tecting Doppler-shifted lines. As the jet flow has an estimated<br />
velocity of ∼ 0.2c, the amount of Doppler-shift is expected to be comparable to that<br />
found in SS 433 (Kotani et al. 1996; Marshall et al. 2002). Hence, spectroscopic<br />
X-ray observations with XMM or Chandra could be very useful to constrain the<br />
physics of the relativistic flow and to study the matter content within the jet.<br />
Finally, as pointed out by Levinson & Waxman (2001) and by Distefano et al.<br />
(2002), if the jets are hadronic we would expect the formation of TeV neutrinos,<br />
that could be <strong>de</strong>tected in the future with <strong>de</strong>tectors of km 2 -scale effective area.<br />
Overall, LS 5039 is a very good target to be studied with new instruments in<br />
or<strong>de</strong>r to gain knowledge on accretion/ejection processes and jet physics.
86 Chapter 2. Multiwavelength approach to LS 5039<br />
2.9 Conclusions<br />
1. Most of the known microquasars have been discovered only after un<strong>de</strong>rgoing<br />
a noticeable outburst, that has triggered the <strong>de</strong>tection by a battery of satel-<br />
lites and ground-based observatories. In contrast, we have discovered a new<br />
microquasar, namely LS 5039, after careful examination of mo<strong>de</strong>rn archive<br />
databases and follow-up interferometric radio observations with different res-<br />
olution (with the VLA, VLBA, EVN and MERLIN).<br />
2. We have found that LS 5039 displays persistent radio emission in the form of<br />
relativistic jets, with a high brightness temperature of ∼ 10 8 K and an optically<br />
thin spectral in<strong>de</strong>x of α −0.5. These results are indicative of non-thermal<br />
synchrotron radiation mechanism.<br />
3. We have obtained the following jet parameters: β 0.15, θ 80 ◦ , a size up to<br />
∼ 1000 AU with a position angle of 125–150 ◦ and a half opening angle smaller<br />
than 6 ◦ . The asymmetry in flux <strong>de</strong>nsity and distance to the core between<br />
the approaching and receding jet seems to be persistent along time. Small<br />
differences in the position angle of the jets, observed at different epochs and<br />
with different resolutions, suggest bending and/or precession of the jets.<br />
4. The radio emission is mo<strong>de</strong>rately variable but no signatures of variability cor-<br />
related with the orbital period, expected because of the LS 5039 eccentric<br />
orbit, have been found. This is probably because of the lack of sensitivity of<br />
the long-term radio observations available up to now (GBI).<br />
5. We have shown that LS 5039 displays slight variability at optical wavelengths,<br />
and mo<strong>de</strong>rate variability at near infrared wavelengths. Based on the new<br />
spectral type classification, O6.5V((f)), and recent calibrations of early type<br />
stars, we have estimated a distance of 2.9 ± 0.3 kpc to LS 5039.<br />
6. We have presented new radial velocity measurements (obtained with the INT)<br />
that confirm the 4.1 d orbital period of this binary system.<br />
7. We have analyzed six and a half years of X-ray monitoring (ASM/RXTE) of<br />
LS 5039, and found no evi<strong>de</strong>nces of periodic variability. In particular, we have<br />
not found the orbital period of the system, expected because of variable accre-<br />
tion rate along an eccentric orbit, probably because the source is marginally<br />
<strong>de</strong>tected in the analyzed dataset.
2.9. Conclusions 87<br />
8. We have carried out pointed X-ray observations (PCA/RXTE and BeppoSAX)<br />
that reveal a spectrum significantly hard up to 30 keV, with a strong iron line<br />
around 6.6 keV. No pulsations or quasi periodic oscillations have been found.<br />
Based on our X-ray data, we have estimated a hydrogen column <strong>de</strong>nsity of<br />
NH = 1.0 ± 0.3 × 10 22 cm −2 . We have found X-ray variability of an or<strong>de</strong>r of<br />
magnitu<strong>de</strong> on timescales of years, that can be related to the variable mass loss<br />
rate of the companion star. There seem to be no X-ray eclipses in LS 5039, in<br />
good agreement with the low proposed inclination (∼ 30 ◦ ).<br />
9. We have proposed an association between LS 5039 and one of the ∼ 170<br />
uni<strong>de</strong>ntified EGRET sources, which seems plausible due to the persistent and<br />
mo<strong>de</strong>rately variable emission of the source at both radio and γ-ray wave-<br />
lengths. If confirmed, this would be the first association between a microquasar<br />
and an EGRET source, and we suggest that other uni<strong>de</strong>ntified high-energy<br />
γ-ray sources could be silent microquasars waiting to be discovered.<br />
10. We have suggested a possible scenario to explain the multiwavelength behavior<br />
of LS 5039, and proposed a mo<strong>de</strong>l based on the γ-ray/radio emission. This<br />
mo<strong>de</strong>l suggests a slowly expanding and well collimated radio jet.<br />
11. Finally, we have ma<strong>de</strong> some energetic consi<strong>de</strong>rations and suggested future<br />
observations that could help to better constrain the proposed mo<strong>de</strong>l.<br />
12. Overall, we think that a careful examination of mo<strong>de</strong>rn archive databases may<br />
reveal a previously unnoticed population of microquasars like LS 5039. If this<br />
is correct, the microquasar phenomenon may not be as rare as it seems.
88 Chapter 2. Multiwavelength approach to LS 5039
Bibliography<br />
Blandford, R. D., & Königl, A. 1979, ApJ, 232, 34<br />
Blandford, R. D., & Payne, D. G. 1982, MNRAS, 199, 883<br />
Casares, J., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, in preparation<br />
Clark, J. S., Reig, P., Goodwin, S. P., et al. 2001, A&A, 376, 476<br />
Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693<br />
Deutschman, W. A., Davis, R. J., & Schild, R. E. 1976, ApJS, 30, 97<br />
Distefano, C., Guetta, D., Waxman, E., & Levinson, A. 2002, ApJ, 575, 378<br />
Drilling, J. S. 1975, AJ, 80, 128<br />
Drilling, J. S. 1991, ApJS, 76, 1033<br />
Fen<strong>de</strong>r, R. P., & Hendry, M. A. 2000, MNRAS, 317, 1<br />
Fen<strong>de</strong>r, R. P., Roche, P., Pooley, G. G., et al. 1997, in Proc. of the 2nd INTEGRAL<br />
Workshop ‘The Transparent Universe’, ESA SP-382, 303<br />
Gray, D. F. 1976, The Observations and Analysis of Stellar Photospheres, Wiley-<br />
Interscience, New York<br />
Hartman, R. C., Bertsch, D. L., Bloom, S. D., et al. 1999, ApJS, 123, 79<br />
Helfand, D. J., Zoonematkermani, S., Becker, R. H., et al. 1980, ApJ, 80, 211<br />
Hjellming, R. M., & Han, X. 1995, Radio properties of X-ray binaries, in X-ray Bi-<br />
naries, ed. W. H. G. Lewin, J. van Paradijs, & E. P. J. van <strong>de</strong>n Heuvel (Cambridge<br />
Univ. Press, Cambridge), 308<br />
89
90 BIBLIOGRAPHY<br />
Hjellming, R. M., & Johnston, K. J. 1988, ApJ, 328, 600<br />
Horne, K. 1986, PASP, 98, 609<br />
Johnson, H. L. 1966, ARA&A, 4, 193<br />
Junor, W., Biretta, J. A., & Livio, M. 1999, Nature, 401, 891<br />
Kilkenny, D. 1993, South African Astronomical Observatory Circular, 15, 53<br />
Kilkenny, D., & Whittet, D. C. B. 1993, South African Astronomical Observatory<br />
Circular, 15, 40<br />
Kilkenny, D., van Wyk, F., Marang, F., Spencer Jones, J. H. & Whittet, D. C. B.<br />
1993, South African Astronomical Observatory Circular, 15, 65<br />
Kotani, T., Kawai, N., Matsuoka, M., & Brinkmann, W. 1996, PASJ, 48, 619<br />
Kraus, A. 1997, PhD Thesis, Friedrich-Wilhelms-Universität Bonn, Germany<br />
Lahulla, J. F., & Hilton, J. 1992, A&AS, 94, 265<br />
Landolt, A.U. 1992, AJ, 104, 340<br />
Led<strong>de</strong>n, J. E., Bro<strong>de</strong>rick, J. J., Condon, J. J., et al. 1980, AJ, 85, 780<br />
Lejeune, T., & Schaerer, D. 2001, A&A, 366, 538<br />
Levine, A. M., Bradt, H., Cui, W., et al. 1996, ApJ, 469, L33<br />
Levinson, A., & Waxman, E. 2001, Phys. Rev. Lett., 87, 171101<br />
Liu, Q. Z., van Paradijs, J., & van <strong>de</strong>n Heuvel, E. P. J. 2000, A&AS, 147, 25<br />
Marsh, T. R., Robinson, E. L., & Wood, J. H. 1994, MNRAS, 266, 137<br />
Marshall, H. L., Canizares, C. R., & Schulz, N. S. 2002, ApJ, 564, 941<br />
Martí, J., Pare<strong>de</strong>s, J. M., & Ribó, M. 1998, A&A, 338, L71<br />
Massi, M., Pare<strong>de</strong>s, J. M., Estalella, R., & Felli, M. 1993, A&A, 269, 249<br />
McSwain, M. V., & Gies, D. R. 2002, ApJ, 568, L27<br />
McSwain, M. V., Gies, D. R., Riddle, R. L., Wang, Z., & Wingert, D. W. 2001,<br />
ApJ, 558, L43
BIBLIOGRAPHY 91<br />
Merck, M., Bertsch, D. L., Dingus, B. L., et al. 1996, A&AS, 120, 465<br />
Mirabel, I. F., & Rodríguez, L. F. 1994, Nature, 371, 46<br />
Mirabel, I. F., & Rodríguez, L. F. 1999, ARA&A, 37, 409<br />
Mitsuda, K., Inoue, H., Koyama, K., et al. 1984, PASJ, 36, 741<br />
Monet, D. 1996, BAAS, 188, 5404<br />
Monet, D. G., Bird, A., Canzian, B., et al. 1999, USNO-A2.0 CD-ROM (U.S. Naval<br />
Observatory, Washington DC)<br />
Motch, C., Haberl, F., Dennerl, K., Pakull, M., & Janot-Pacheco, E. 1997, A&A,<br />
323, 853 (M97)<br />
Pacholczyk, A. G. 1970, Radio Astrophysics, Freeman, San Francisco, CA<br />
Pare<strong>de</strong>s, J. M., Martí, J., Peracaula, M., & Ribó, M. 1997, A&A, 320, L25<br />
Pare<strong>de</strong>s, J. M., Martí, J., Ribó, M., & Massi, M. 2000, Science, 288, 2340<br />
Pare<strong>de</strong>s, J. M., Ribó, M., Ros, E., Martí, J., & Massi, M. 2002, A&A, 393, L99<br />
Peng, B., Kraus, A., Krichbaum, T. P., & Witzel, A. 2000, A&A, 145, 1<br />
Penny, L. R. 1996, ApJ, 463, 737<br />
Pooley, G. G., Fen<strong>de</strong>r, R. P., & Brocksopp, C. 1999, MNRAS, 302, L1<br />
Pre<strong>de</strong>hl, P., & Schmitt, J. H. M. M. 1995, A&A, 293, 889<br />
Rees, M. J. 1966, Nature, 211, 468<br />
Reig, P., Ribó, M., Pare<strong>de</strong>s, J. M., & Martí, J. 2002, in preparation<br />
Ribó, M., Reig, P., Martí, J., & Pare<strong>de</strong>s, J. M. 1999, A&A, 347, 518<br />
Ribó, M., Pare<strong>de</strong>s, J. M., Romero, G. E., et al. 2002, A&A, 384, 954<br />
Roberts, D. H., Lehár, J., & Dreher, J. W. 1987, AJ, 93, 968<br />
Schaerer, D., <strong>de</strong> Koter, A., Schmutz, W., & Mae<strong>de</strong>r, A. 1996, A&A, 312, 475
92 BIBLIOGRAPHY<br />
Schmidt-Kaler, Th. 1982, Landolt-Börnstein: Numerical Data and Functional Re-<br />
lationships in Science and Technology, New Series ”Group 6 Astronomy and As-<br />
trophysics”, Volume 2<br />
Schönfel<strong>de</strong>r, V., Bennett, K., Blom, J. J., et al. 2000, A&AS, 143, 145<br />
Spencer Jones, J. H., Whittet, D. C. B., & He, L. 1993, South African Astronomical<br />
Observatory Circular, 15, 73<br />
Stellingwerf, R. F. 1978, ApJ, 224, 953<br />
Stephenson, C. B., & Sanduleak, N. 1971, Publ. Warner & Swasey Obs., 1, 1<br />
Taylor, A. R., Young, G., Peracaula, M., Kenny, H. T., & Gregory, P. C. 1996,<br />
A&A, 305, 817<br />
Torres, D. F., Romero, G. E., Combi, J. A., et al. 2001, A&A, 370, 468<br />
Vacca, W. D., Garmany, C. D., & Shull, J. M. 1996, ApJ, 460, 914<br />
van <strong>de</strong>r Klis, M. 1989, in Timing Neutron stars, ed. H. Ögelman & E. P. J. van <strong>de</strong>n<br />
Heuvel (Dordrecht: Kluwer), 27<br />
Voges, W., Boller, Th., Dennerl, K., et al. 1996, in Proc. of the Conference Rönt-<br />
genstrahlung from the Universe, MPE Report, 263, 637<br />
Voges, W., Aschenbach, B., Boller, Th., et al. 1999, A&A, 349, 389<br />
Voges, W., Aschenbach, B., Boller, Th., et al. 2000, IAU Circ., 7432, 3<br />
Wang, Q. D., Gotthelf, E. V., Lang, C. C. 2002, Nature, 415, 148<br />
Yamauchi, S., & Koyama, K. 1993, ApJ, 404, 620
Chapter 3<br />
LS 5039 as a runaway microquasar<br />
The formation of the compact object in an X-ray binary necessarily requires a super-<br />
nova explosion that does not disrupt the system. This explosive event is expected<br />
to consi<strong>de</strong>rably change the kinematic properties of the binary star. The change<br />
may be rather extreme (i.e., kick velocities approaching ∼ 10 3 km s −1 ) for highly<br />
asymmetric supernovae, as it has been proposed for the microquasar Circinus X-1<br />
(Tauris et al. 1999). This mechanism may also be responsible for ejecting X-ray<br />
binaries into the halo of the Galaxy. The fast moving X-ray nova XTE J1118+480 is<br />
a possible example (Mirabel et al. 2001), although it could have been ejected from<br />
a globular cluster in the past. On the other hand, the microquasars GRO J1655−40<br />
and Cygnus X-1, also display runaway velocities (Shahbaz et al. 1999; Kaper et al.<br />
1999). With in<strong>de</strong>pen<strong>de</strong>nce of their origin in the galactic plane or in the halo, the<br />
existence of runaway microquasars is a new and important issue that <strong>de</strong>serves an<br />
in-<strong>de</strong>pth study.<br />
In this chapter 1 we focus on the kinematic properties of LS 5039 and its sur-<br />
roundings. In Sect. 3.1 we estimate the proper motions of the system, using optical<br />
and radio data. In Sect. 3.2 we discuss the space velocity of LS 5039, whereas in<br />
Sect. 3.3 we analyze the trajectory of the binary system in the past. In Sect. 3.4 we<br />
study the possible association of this microquasar with a supernova remnant, and<br />
in Sect. 3.5 we analyze their H i surroundings. Finally, we make a global discussion<br />
in Sect. 3.6 and summarize our main conclusions in Sect. 3.7.<br />
1 Published in Ribó, M., Pare<strong>de</strong>s, J. M., Romero, E. G., Benaglia, P., Martí, J., Fors, O., &<br />
García-Sánchez, J. 2002, A&A, 384, 954.<br />
93
94 Chapter 3. LS 5039 as a runaway microquasar<br />
3.1 Positions and proper motions<br />
3.1.1 Optical positions<br />
LS 5039 (V 11.2, B 12.2) is not inclu<strong>de</strong>d in most accurate astrometric catalogs,<br />
like Hipparcos and Tycho-1 (which are complete up to V ∼ 9.0 and V ∼ 10.0,<br />
respectively). Although it appears in Tycho-2, its position has a large uncertainty<br />
when compared to the average error within its magnitu<strong>de</strong> range (Høg et al. 2000).<br />
As a result, the catalogued proper motions lack of the required precision. Moreover,<br />
other astrometric catalogs which were not inclu<strong>de</strong>d in Tycho-2 estimate of the proper<br />
motions, have been released since then. In view of all these facts, we <strong>de</strong>ci<strong>de</strong>d to carry<br />
out a thorough study of all available catalogs. First, we did a search using the query<br />
forms within the VizieR database (Ochsenbein et al. 2000) at the Centre <strong>de</strong> Données<br />
astronomiques <strong>de</strong> Strasbourg (CDS), and afterwards we looked for new information<br />
on several public web sites. The results of this search are listed in Table 3.1, with<br />
all coordinates in ICRS, and commented below.<br />
The ol<strong>de</strong>st astrometric information for LS 5039 is dated around 1905 and comes<br />
from the Astrographic Catalog 2000 (AC 2000, Urban et al. 1998). However, a<br />
re-reduction of this catalog was performed recently, using an improved reference<br />
catalog that allowed a better handling of the systematic errors (S. E. Urban, private<br />
communication). This new catalog will soon be released as AC 2000.2. Since changes<br />
were most significant in the faint stars of the southern AC 2000 zones, the position<br />
of LS 5039 suffered a noticeable shift. Actually, two positions corresponding to<br />
different epochs are given in AC 2000.2.<br />
The following available astrometry, dated around 1950, comes from the scanning<br />
and reduction of the Palomar Observatory Sky Survey plates, carried out by Monet<br />
et al. (1999), which is part of the USNO-A2.0 catalog.<br />
We inspected the Second Cape Photographic Catalog (CPC2, Zacharias et al.<br />
1999), a Southern Hemisphere astrometric catalog, containing observations between<br />
1962 and 1973. Unfortunately, LS 5039 is not listed in it, because the limiting<br />
magnitu<strong>de</strong> of this catalog was V 10.5.<br />
The Twin Astrographic Catalog version 2 (TAC 2.0) provi<strong>de</strong>s positional informa-
3.1. Positions and proper motions 95<br />
Table 3.1: Compilation of optical and radio positions, with associated errors, of LS 5039.<br />
Wavelength Epoch α (ICRS) σα cos δ δ (ICRS) σδ Catalog Ref.<br />
domain [h, m, s] [mas] [ ◦ , ′ , ′′ ] [mas] name<br />
Optical 1905.45 18 26 15.0194 255 −14 50 53.300 247 AC 2000.2 1<br />
1907.43 18 26 15.0177 255 −14 50 53.075 247 AC 2000.2 1<br />
1951.577 18 26 15.034 250 −14 50 53.59 250 USNO-A2.0 2<br />
1979.484 18 26 15.0557 64 −14 50 54.075 47 TAC 2.0 3<br />
1986.653 18 26 15.054 300 −14 50 54.29 300 GSC 1.2 4<br />
1991.75 18 26 15.0427 149 −14 50 54.229 120 Tycho-2 5<br />
2000.289 18 26 15.0563 13 −14 50 54.277 13 UCAC1 6<br />
Radio 1998.24 18 26 15.056 10 −14 50 54.24 10 VLA obs. 7<br />
2000.42 18 26 15.0566 4 −14 50 54.261 6 VLBA obs. 8<br />
1. Urban, private communication. 2. Monet et al. 1999. 3. Zacharias & Zacharias 1999. 4. Morrison<br />
et al. 2001. 5. Høg et al. 2000. 6. Zacharias et al. 2000. 7. Martí et al. 1998. 8. Ribó et al. 2002.<br />
tion around 1980. Apart from the internal error of σα cos δ = 62 mas and σδ = 45 mas<br />
associated to the LS 5039 position, an external error of 15 mas has been taken into<br />
account following an estimate carried out by Zacharias & Zacharias (1999).<br />
The last photographic catalog used is the Gui<strong>de</strong> Star Catalog version 1.2 (GSC 1.2,<br />
Morrison et al. 2001). This is a re-reduction of the GSC (Lasker at al. 1990), after<br />
removing plate-based systematic distortions that are a function of magnitu<strong>de</strong> and<br />
radial distance from the plate center. In addition, by using the Starlink library<br />
slalib, we have transformed the catalogued position from FK5 to ICRS.<br />
As mentioned above, LS 5039 appears as well in the Tycho-2 catalog (Høg et al.<br />
2000). The average internal error in position for a V 11.2 star is around 50 mas,<br />
while the errors in α cos δ and δ for LS 5039 are 93 and 120 mas, respectively. This<br />
fact tells us that the position estimate is not as good as one would expect. On the<br />
other hand, the scatter-based errors from the 4 positions present in the Tycho-2<br />
data are 149 and 60 mas, respectively. Hence, we have consi<strong>de</strong>red the higher error<br />
estimate in each coordinate: σα cos δ = 149 mas and σδ = 120 mas.<br />
A second epoch of the GSC was performed during the nineties. However, the<br />
positional errors in the current version of this catalog, GSC 2.2.01, are inten<strong>de</strong>d<br />
as indicators for operational use only, and cannot be used for scientific studies.
96 Chapter 3. LS 5039 as a runaway microquasar<br />
Hence, we have consi<strong>de</strong>red that this catalog does not provi<strong>de</strong> accurate astrometric<br />
information to be used for estimating proper motions.<br />
Finally, we have used the first release of the US Naval Observatory CCD Astro-<br />
graph Catalog (UCAC1, Zacharias et al. 2000). As can be seen in Table 3.1, this is<br />
by far the best optical astrometric catalog available.<br />
3.1.2 Radio positions<br />
In<strong>de</strong>pen<strong>de</strong>nt astrometric information can be obtained from radio observations of<br />
LS 5039. The first available astrometric information comes from the NVSS by<br />
Condon et al. (1998). However, the position error in this catalog, around 1 ′′ ,<br />
is too large to be of any use for our purposes. Martí et al. (1998) carried out<br />
VLA 2 -A configuration observations which provi<strong>de</strong>d accurate (10 mas) astrometry for<br />
LS 5039. Finally, phase-referencing VLBA 2 observations conducted by Ribó et al.<br />
(2002) provi<strong>de</strong> the last radio position, with the best accuracy among all available<br />
data. Details of the astrometry for the last two cases are listed in Table 3.1. The<br />
relatively large location error of few mas in the VLBA data is due to confusion<br />
produced by the high level of Galactic electron scattering in this region (for both<br />
the reference source and LS 5039).<br />
We must state that no correction for parallax has been applied to the obtained<br />
positions, since at an estimated distance of 2.9 ± 0.3 kpc, this would translate into<br />
less than half a mas, which is always much smaller than the available uncertainties.<br />
3.1.3 Proper motions<br />
It is clear from the astrometric data that both, the optical and radio sources, are<br />
almost in the same position of the sky. However, in or<strong>de</strong>r to show that this is not<br />
a chance coinci<strong>de</strong>nce and that both emissions originate in the same object, we have<br />
computed in<strong>de</strong>pen<strong>de</strong>nt proper motions for the optical and radio data.<br />
2 The VLA and the VLBA are operated by the National Radio Astronomy Observatory (NRAO).<br />
The NRAO is a facility of the National Science Foundation operated un<strong>de</strong>r cooperative agreement<br />
by Associated Universities, Inc.
3.1. Positions and proper motions 97<br />
Table 3.2: Proper motions estimates for LS 5039.<br />
Data set µα cos δ µδ<br />
(mas yr −1 ) (mas yr −1 )<br />
Optical 4.7 ± 1.3 −11.0 ± 0.8<br />
Radio 4.0 ± 4.9 −9.6 ± 5.3<br />
Optical+Radio 4.7 ± 1.1 −10.6 ± 1.0<br />
Although the position uncertainties from the AC 2000.2 and USNO-A2.0 cat-<br />
alogs are relatively large, the epoch span achieved justifies their inclusion in the<br />
estimate of the proper motions. Hence, from the optical data, and taking into<br />
account the astrometric uncertainties, we obtain the following results: µα cos δ =<br />
4.7 ± 1.3 mas yr −1 , µδ = −11.0 ± 0.8 mas yr −1 , where the errors come directly<br />
from the least squares fit. The two accurate radio positions give proper motions of:<br />
µα cos δ = 4.0 ± 4.9 mas yr −1 , µδ = −9.6 ± 5.3 mas yr −1 . Although this last result<br />
has a large error because only two points are available, we can say that both, the<br />
optical and radio sources, have very similar proper motions. Therefore, based only<br />
on astrometric data, we are able to confirm that both, the optical and the radio<br />
emission, originate in the same object. Hence, we will use all the data listed in<br />
Table 3.1 to compute accurate proper motions for LS 5039, which happen to be:<br />
µα cos δ = 4.7 ± 1.1 mas yr −1 , µδ = −10.6 ± 1.0 mas yr −1 . All these results are<br />
summarized in Table 3.2. We can now transform the proper motions into galactic<br />
coordinates and obtain µl = −7.2 ± 1.0 mas yr −1 and µb = −9.1 ± 1.0 mas yr −1 .<br />
It is clear from these results that there is a noticeable motion perpendicular to the<br />
galactic plane and moving away from it.<br />
According to our least squares fits, and <strong>de</strong>fining t = yr − 2000.0, the predicted<br />
ICRS values for α and δ near t = 0 are:<br />
α = 18 h 26 m 15.0565 s +<br />
δ = −14 ◦ 50 ′ 54.260 ′′ −<br />
<br />
4.7 t/ cos δ<br />
, (3.1)<br />
mas<br />
<br />
10.6 t<br />
, (3.2)<br />
mas<br />
being 9 + 1.2 t 2 / cos 2 δ and √ 9 + t 2 the errors in mas in α and δ, respectively.<br />
Comparing the fits with the positions in Table 3.1 and the proper motions in Ta-<br />
ble 3.2 we can say that, approximately, the offsets are <strong>de</strong>termined by the VLBA data
98 Chapter 3. LS 5039 as a runaway microquasar<br />
∆α cos δ (mas)<br />
200<br />
0<br />
−200<br />
−400<br />
−600<br />
−800<br />
1900 1920 1940 1960<br />
Time (yr)<br />
1980 2000 2020<br />
∆δ (mas)<br />
1400<br />
1200<br />
1000<br />
800<br />
600<br />
400<br />
200<br />
0<br />
−200<br />
−400<br />
1900 1920 1940 1960<br />
Time (yr)<br />
1980 2000 2020<br />
Figure 3.1: Offsets in Right Ascension (left) and Declination (right), relatives to the<br />
fitted ones for year 2000.0, versus time for all the positions listed in Table 3.1. Open<br />
circles represent the optical positions, and filled circles the radio ones. The solid lines<br />
represent the least squares fits to the whole data sets.<br />
point because of the small error in position, while the proper motions are obtained<br />
by the optical points due to the huge time span.<br />
In Fig. 3.1 we have plotted the offsets in α cos δ and δ from the fitted values for<br />
year 2000.0 versus time, for the data listed in Table 3.1, together with the respective<br />
fits to the data. Notice that in both plots the UCAC1 and VLBA data are almost<br />
superimposed.<br />
3.2 Ejection from the galactic plane<br />
If we assume a distance of 2.9 ± 0.3 kpc to LS 5039, as seen in Sect. 2.5.3, and a<br />
systemic Vr = 4.6 ± 0.5 km s −1 (McSwain et al. 2001), the proper motions estimates<br />
translate into (U = 40 ± 5, V = −82 ± 16, W = −118 ± 19) km s −1 in the Local<br />
Standard of Rest (LSR) <strong>de</strong>fined by (U⊙ = 9, V⊙ = 12, W⊙ = 7) km s −1 . This gives<br />
a total systemic velocity of vsys = (U 2 + V 2 + W 2 ) 1/2 = 149 ± 18 km s −1 . Using a<br />
value of 215 km s −1 for the galactic rotation at 5.8 kpc from the galactic center (Fich<br />
et al. 1989) we can transform the galactic plane space velocities (U and V ) from the<br />
LSR into the LS 5039 regional standard of rest (RSR). Applying this transformation<br />
we find U = 51 ± 6 and V = −71 ± 16 km s −1 , and vsys = 147 ± 17 km s −1 . All<br />
these results are summarized in Table 3.3.
3.3. The past trajectory of LS 5039 99<br />
Table 3.3: Space velocity estimates, in km s −1 , for LS 5039 related to the LSR and to<br />
its RSR.<br />
Frame U V W vsys<br />
LSR 40 ± 5 −82 ± 16 −118 ± 19 149 ± 18<br />
RSR 51 ± 6 −71 ± 16 −118 ± 19 147 ± 17<br />
Taking into account a cosmic dispersion for early-type stars of (σU, σV , σW ) <br />
(7, 8, 4) km s −1 (Torra et al. 2000) we can conclu<strong>de</strong> that this source is escaping<br />
from its own RSR and has a large velocity component perpendicular to the galactic<br />
plane. Since LS 5039 is a HMXB containing an early-type star, it seems very un-<br />
likely to be a high speed halo object crossing the galactic plane. Hence, the most<br />
plausible explanation for the observed velocity is that the binary system obtained<br />
an acceleration during the supernova event that created the compact object in this<br />
microquasar.<br />
3.3 The past trajectory of LS 5039<br />
An interesting thing to do once a position and space velocity estimates are available,<br />
is to compute the trajectory of LS 5039 in the past, and then to look for OB<br />
associations and SN Remnants (SNRs) in its path, in or<strong>de</strong>r to establish possible<br />
relationships. For this purpose, we computed the galactic orbital motion of this<br />
system un<strong>de</strong>r the gravitational field of the Galaxy. The galactic mass mo<strong>de</strong>l by<br />
Dauphole & Colin (1995) was adopted for the integration, which was performed<br />
using a Runge-Kutta fourth-or<strong>de</strong>r integrator using time steps of 1000 years.<br />
It is helpful to search for OB associations in the past trajectory of runaway X-ray<br />
binaries, as done by Ankay et al. (2001) with HD 153919/4U 1700−37, because if<br />
evi<strong>de</strong>nces are found that the runaway system originated in an OB association, it is<br />
possible to estimate the age of the binary system after the SN explosion. According<br />
to the computed trajectory in the past for LS 5039, there are two OB associations<br />
in or close to its path in the plane of the sky: Sct OB3 and Ser OB2 (Melnik &<br />
Efremov 1997). Unfortunately, the distance to Sct OB3 is ∼ 1.5 kpc (Melnik &<br />
Efremov 1997), and the distance to Ser OB2 is 1.9 ± 0.3 kpc (Forbes 2000). Hence,
100 Chapter 3. LS 5039 as a runaway microquasar<br />
the two OB associations found in the path of LS 5039 are too close to us to be<br />
related to it.<br />
Since we have not been able to fix a limit on the integration time in the past<br />
thanks to a possible relationship with an OB association, we will use the vertical<br />
distribution of early-type stars in the Galaxy for this purpose. O-type stars are<br />
typically located at distances within 45 ± 20 pc from the galactic midplane (Reed<br />
2000), i.e., from Z −65 to Z +65 pc. According to the galactic latitu<strong>de</strong> of<br />
LS 5039 and using a distance of 2.9 ± 0.3 kpc, its actual height is Z = −65 ± 7 pc.<br />
Since it is escaping from the galactic plane, we can compute the trajectory backward<br />
in time up to when it had a height of Z = +65 pc, which was ∼ 1.1 Myr ago. Hence,<br />
the SN explosion probably took place in the last ∼ 1.1 Myr.<br />
Thus, for a possible association of LS 5039 with a SNR, we will focus our attention<br />
on the galactic trajectory of the binary system for the last ∼ 1 Myr. In addition,<br />
the likelihood of <strong>de</strong>tecting a SNR <strong>de</strong>creases with time, so it is not expected that the<br />
radio emission of the SNR can be observed much beyond the time interval consi<strong>de</strong>red<br />
above (e.g., Shklovskii 1968).<br />
A search in a catalog of galactic supernova remnants (Green 2000) reveals the<br />
presence of several SNRs near the path of LS 5039 on the plane of the sky, though<br />
the center of none of them is crossed by its trajectory. Only three of them, namely<br />
SNR G016.7+00.1, SNR G016.8−01.1, and SNR G018.8+00.3, are within 1 ◦ of the<br />
LS 5039 path. The distance to SNR G016.7+00.1 was estimated to be at ∼ 14 kpc<br />
by Reynoso & Mangum (2000), making impossible an association with LS 5039.<br />
Dubner et al. (1999) conclu<strong>de</strong>d that SNR G018.8+00.3 is located at 1.9 ± 0.5 kpc,<br />
and has an age of ∼ 16 000 yr. Although the distance is not very different from<br />
the one to LS 5039, the age of this SNR is by far too short to be associated with<br />
the microquasar. Hence, the only remaining candidate is SNR G016.8−01.1, which<br />
<strong>de</strong>serves a <strong>de</strong>tailed study.<br />
3.4 SNR G016.8−01.1<br />
A wi<strong>de</strong> field radio map of the surrounding of LS 5039 is shown in Fig. 3.2, with<br />
the microquasar (contours) and the nearby SNR G016.8−01.1 (grey scale), together<br />
with the position of the H ii-region RCW 164 (which is the stronger source in the
3.4. SNR G016.8−01.1 101<br />
DECLINATION (J2000)<br />
-14 30<br />
35<br />
40<br />
45<br />
50<br />
55<br />
-15 00<br />
LS 5039<br />
SNR G016.8-01.1<br />
o<br />
b = -0.8<br />
RCW 164<br />
18 26 30 15 00 25 45 30 15 00 24 45 30<br />
RIGHT ASCENSION (J2000)<br />
Figure 3.2: Wi<strong>de</strong> field radio map of LS 5039, its nearby shell-like SNR G016.8−01.1<br />
and the H ii-region RCW 164 (which is the stronger source in the field). The grey scale<br />
emission is taken from the Parkes-MIT-NRAO tropical survey at the 6 cm wavelength<br />
(Tasker et al. 1994). The overlaid contours correspond to the NVSS map of the region<br />
at the 20 cm wavelength (Condon et al. 1998). The arrow marks the proper motion<br />
sense (see text). The dashed line is the computed trajectory for the last 10 5 yr, with<br />
the corresponding error boxes in position at that epoch and 5 × 10 4 yr ago. The dotted<br />
line represents positions with a galactic latitu<strong>de</strong> of −0.8 ◦ for reference purposes.
102 Chapter 3. LS 5039 as a runaway microquasar<br />
field). The arrow in this figure marks the proper motion sense of LS 5039 as it would<br />
be seen from the LSR (i.e., correction for the peculiar velocity of the Sun has been<br />
applied), while the dashed line represents the trajectory in the past up to 10 5 yr,<br />
with the corresponding error boxes in position at that epoch and 5 × 10 4 yr ago.<br />
From this figure we can see that LS 5039 crosses the projection in the plane of the<br />
sky of SNR G016.8−01.1 between ∼ 4 × 10 4 and ∼ 1.3 × 10 5 years ago.<br />
If one of the two stars forming a close binary system experiences a SN explosion,<br />
it may form a SNR and an X-ray binary. Since conservation of the linear momentum<br />
after the SN explosion is required, if the SNR and the X-ray binary are related, we<br />
expect them to be aligned with the proper motion of the latter, and any <strong>de</strong>viation<br />
from this behavior should be explained by the projection in the plane of the sky of<br />
the initial space velocity of the binary system prior to the SN event. This does not<br />
seem to be the case here, because SNR G016.8−01.1 and LS 5039 are not well aligned<br />
with the LS 5039 proper motion. Moreover, if we assume that the center of the shell-<br />
like remnant is located at the minimum of radio emission in Fig. 3.2, approximately<br />
located at α = 18 h 25 m 03 s , δ = −14 ◦ 37 ′ 30 ′′ (or l = 16.94 ◦ , b = −0.93 ◦ ), a<br />
peculiar velocity in the plane of the sky for the original system of 70 km s −1<br />
is obtained. This velocity is much higher than the typical ∼ 10 km s −1 found for<br />
early-type stars. In or<strong>de</strong>r to reduce this velocity we can use in our favor the error in<br />
the proper motion angle, and find that it turns out to be 50 km s −1 , which is still<br />
too high. However, the center of the shell-like remnant could be in another position,<br />
closer to the LS 5039 trajectory, and the <strong>de</strong>rived peculiar velocity for the system<br />
prior to the SN explosion could fit in the typical values of early-type stars. Hence,<br />
from the kinematical point of view, we are not able to firmly discard a possible<br />
association between both objects. Therefore, we have performed an in-<strong>de</strong>pth study<br />
of the remnant.<br />
SNR G016.8−01.1 was discovered by Reich et al. (1986) using a combination<br />
of survey data analysis and new multifrequency observations. The exten<strong>de</strong>d radio<br />
source has a complex structure due to the superposition of the H ii-region RCW 164<br />
(Rodgers et al. 1960), which is coinci<strong>de</strong>nt with the peak of the continuum emission<br />
(in black in Fig. 3.2). The diffuse radiation observed around the peak is strongly<br />
polarized (Rodgers et al. 1960), indicating a synchrotron nature. Polarization <strong>de</strong>-<br />
creases towards the position of the foreground H ii-region as a result of the thermal<br />
contribution. Thermal and non-thermal components of the radiation cannot be<br />
clearly separated, but the total flux from the SNR at 5 GHz seems to be ∼ 1 Jy
3.4. SNR G016.8−01.1 103<br />
Normalized Intensity<br />
1,4<br />
1,2<br />
1,0<br />
0,8<br />
0,6<br />
0,4<br />
0,2<br />
0,0<br />
-0,2<br />
-0,4<br />
-0,6<br />
-100 -80 -60 -40 -20 0 20 40 60 80<br />
Velocity (km/s)<br />
Figure 3.3: H166α line observations of the H ii-region RCW 164. The solid line repre-<br />
sents a Gaussian fit to the data, with its maximum being at v = 16.5 ± 0.8 km s −1 .<br />
(J. A. Combi, private communication). The angular diameter of the remnant is<br />
∼ 30 ′ , and its distance remains unknown.<br />
3.4.1 A lower limit of the distance<br />
In or<strong>de</strong>r to <strong>de</strong>termine a lower limit of the distance to SNR G016.8−01.1, we car-<br />
ried out H166α recombination line (1424.734 MHz) observations of the foreground<br />
H ii-region on 2000 November 17. We used a 30-m radiotelescope at the Instituto<br />
Argentino <strong>de</strong> Radioastronomía (IAR), Villa Elisa. The receiver is a helium-cooled<br />
HEMT amplifier with a 1008-channel autocorrelator at the back end. The HPBW<br />
at a wavelength of 21 cm is 30 ′ and the temperature of the system on the cold sky<br />
during the observations was about 35 K. The H166α line was <strong>de</strong>tected at a veloc-<br />
ity of 16.5 ± 0.8 km s −1 (see Fig. 3.3) after 1.5 hours of integration time, with a<br />
signal-to-noise ratio of ∼ 4. At a location of l ≈ 16.8 ◦ , b ≈ −1.1 ◦ , standard galactic<br />
rotation mo<strong>de</strong>ls (Fich et al. 1989) indicate that the observed velocity corresponds<br />
to a distance d ∼ 1.8 kpc. Consequently, SNR G016.8−01.1 should be farther than<br />
∼ 2 kpc, as suggested by its relatively small angular size.
104 Chapter 3. LS 5039 as a runaway microquasar<br />
3.4.2 Particle <strong>de</strong>nsity estimates<br />
If we adopt for SNR G016.8−01.1 the same distance to LS 5039, we can make<br />
some estimates of the parameters that would characterize the SNR in case of being<br />
associated with the X-ray binary. Using a size of 30 ′ and a distance of 2.9 kpc, the<br />
inferred radius would be R = 12.7 pc and it would be, thus, still in the adiabatic<br />
expansion phase. Using the standard Sedov (1959) solutions, we can express the<br />
particle <strong>de</strong>nsity of the ambient medium as<br />
n = 4.44 × 10 −8 t 2 E51 R −5<br />
1 cm −3 , (3.3)<br />
where t is the age of the remnant in years, E51 is the original energy release of the<br />
SN explosion in units of 10 51 erg and R1 is the remnant radius in units of 10 pc.<br />
Since the binary system is bound after the SN explosion, its total mass has to be<br />
larger than the mass of the SNR (van <strong>de</strong>n Heuvel 1978), which, therefore, should<br />
be moving at least with the same space velocity in the opposite direction. Hence,<br />
the maximum age of the remnant if it were related to LS 5039, would be around<br />
5 × 10 4 yr. If we use this age and assume that E51 = 0.4 (Spitzer 1998), we find<br />
an ambient <strong>de</strong>nsity of n ∼ 13.5 cm −3 . Consi<strong>de</strong>ring the uncertainty in the distance,<br />
we can say that 8 < n < 23 cm −3 . Although such number <strong>de</strong>nsities would not be<br />
very unusual towards this direction of the inner part of the Galaxy, we have ma<strong>de</strong><br />
H i observations in or<strong>de</strong>r to estimate the typical column <strong>de</strong>nsities in the region of<br />
interest.<br />
The H i observations were performed during 2000 November 29 and December<br />
1, with the already mentioned IAR telescope. The H i line was observed in a hybrid<br />
total-power mo<strong>de</strong> with a sampling on a 0.25 ◦ -lattice around the position of SNR<br />
G016.8−01.1, covering a total of 4 ◦ × 4 ◦ . The velocity resolution obtained was<br />
1.05 km s −1 , with a total coverage of ±450 km s −1 . The rms level in brightness<br />
temperature was ∼ 0.1 K. Those channel maps in the velocity range from 10 to<br />
50 km s −1 (corresponding to distances between 1.5 and 4.5 kpc) were inspected<br />
looking for clouds or a local minimum that could be associated to the SNR.<br />
The maps show a strong gradient of brightness temperature towards the galactic<br />
plane, but no clear evi<strong>de</strong>nce for a structure that could be associated with the SNR.<br />
In Fig. 3.4, we show the integrated column <strong>de</strong>nsity map for the velocity interval<br />
between 26–34 km s −1 , where the contour labels are in units of 10 19 cm −2 . With<br />
the column <strong>de</strong>nsities observed in the region where SNR G016.8−01.1 is located
3.4. SNR G016.8−01.1 105<br />
Galactic latitu<strong>de</strong><br />
1.0<br />
0.0<br />
-1.0<br />
-2.0<br />
-3.0<br />
-19.0 | -18.0 | -17.0 | -16.0 | -15.0 |<br />
Galactic longitu<strong>de</strong><br />
Figure 3.4: Neutral hydrogen column <strong>de</strong>nsity distribution towards SNR G016.8−01.1, in-<br />
tegrated over the velocity interval 26–34 km s −1 . Contour levels are in units of 10 19 cm −2 .<br />
The circle marks the position of SNR G016.8−01.1, while the star represents LS 5039.<br />
(∼ 4 × 10 20 cm −2 ), the ambient <strong>de</strong>nsity should be ∼ 5 cm −3 . This is nearly a factor<br />
of 3 lower than the estimated <strong>de</strong>nsity from the expected size of the SNR if it is at<br />
2.9 kpc. Taking into account the uncertainty in the distance, the ambient <strong>de</strong>nsity<br />
<strong>de</strong>termined from the remnant’s size (8 < n < 23 cm −3 ) would be slightly larger than<br />
that estimated from the H i observations. It could be that there is more material<br />
un<strong>de</strong>r the form of other molecular species, or that the original SN energy release<br />
could have been different from what we have assumed, or, finally, that the distance<br />
to the SNR is not 2.9 kpc and it is not related to LS 5039. The case, consequently,<br />
remains unsolved.
106 Chapter 3. LS 5039 as a runaway microquasar<br />
3.5 The H i surroundings of LS 5039<br />
Notwithstanding the absence of structures that can be clearly associated with the<br />
SNR in the H i distribution, the channel maps in the interval 10–46 km s −1 (see<br />
Fig. 3.5) reveal the existence of a large, semi-open cavity which is very similar to<br />
bubbles blown out by early-type stars in regions with steep <strong>de</strong>nsity gradients (e.g.,<br />
Benaglia & Cappa 1999). This phenomenon has also been observed around another<br />
HMXB (HD 153919) by Benaglia & Cappa (1999). The ambient material is thought<br />
to be swept up by the strong wind of the massive early-type star in the binary system<br />
creating a local minimum around the star. Density gradients towards the galactic<br />
plane in the original matter distribution can result in a preferred escape direction for<br />
the wind, yielding open structures. Even in some cases the star appears displaced<br />
from the actual minimum in the H i distribution.<br />
The bubble <strong>de</strong>tected around LS 5039 is centered at a systemic velocity of vsys <br />
33 km s −1 , with an expansion velocity vexp 12 km s −1 . The galactic rotation mo<strong>de</strong>l<br />
by Fich et al. (1989) indicates that it is located at a distance of 3.2 kpc, quite<br />
consistent with the distance to the microquasar. The total mass removed in or<strong>de</strong>r<br />
to create the cavity is ∼ 1.5 × 10 4 M⊙, implying a kinetic energy ∼ 2 × 10 49 erg s −1 ,<br />
which can be consi<strong>de</strong>red as a rough estimate of the energy <strong>de</strong>posited by the wind<br />
of the star into the ISM. Such a value is in accordance with previous estimates for<br />
similar Of star wind blown bubbles (e.g., Benaglia & Cappa 1999). With a radius of<br />
∼ 60 pc, the dynamical age of the bubble is ∼ 3×10 6 yr, which should be consi<strong>de</strong>red<br />
as an upper limit for its actual age since it ignores possible contributions from the<br />
SN explosion and the wind of the progenitor of the compact star.<br />
We have also searched for other contributors to the cavity, among stars earlier<br />
than B2 that were projected over the cavity, and whose distances were in agree-<br />
ment with that of LS 5039. It seems unlikely that the Wolf Rayet star WR 115<br />
(WN6+OB?) could contribute in some way, because its distance is 2 kpc, and is<br />
located over the shell of the cavity (l = 16.98 ◦ , b = −1.03 ◦ ). The other luminous<br />
stars in the field, LS 5005, LS 5017, LS 5022, LS 5047, LS 5048 and HD 169673 can<br />
sum up at most 20% of the energy required to create such a minimum in the H i<br />
distribution. The O((f)) star in LS 5039, then, seems to be the main agent forming<br />
the bubble.<br />
However, the actual space velocity indicates that the system has been close to the
3.5. The H i surroundings of LS 5039 107<br />
Galactic latitu<strong>de</strong><br />
1.0<br />
0.0<br />
-1.0<br />
-2.0<br />
-3.0<br />
1.0<br />
0.0<br />
-1.0<br />
-2.0<br />
-3.0<br />
1.0<br />
0.0<br />
-1.0<br />
-2.0<br />
-3.0<br />
-19.0 -18.0 -17.0 -16.0<br />
v = 12 km/s v = 16 km/s v = 20 km/s<br />
v = 24 km/s v = 28 km/s v = 32 km/s<br />
v = 36 km/s v = 40 km/s v = 44 km/s<br />
-19.0 -18.0 -17.0 -16.0<br />
Galactic longitu<strong>de</strong><br />
-19.0 -18.0 -17.0 -16.0 -15.0<br />
| | | | | | | | | | | | |<br />
Figure 3.5: Neutral hydrogen channel maps for the velocity interval 10–46 km s −1 ,<br />
towards LS 5039. Each map covers 4 km s −1 , the central velocity is given in each label.<br />
A local minimum and an open low <strong>de</strong>nsity cavity can be clearly seen close to the HMXB.<br />
The cavity is open towards the direction opposite to the strongest mass concentrations<br />
near the galactic plane, as expected from a wind-blown structure.
108 Chapter 3. LS 5039 as a runaway microquasar<br />
cavity only during the last ∼ 2 × 10 5 yr, which is much shorter than the upper limit<br />
found for the dynamical age of the bubble (∼ 3 × 10 6 yr). Nevertheless, the cavity<br />
could be originated by LS 5039 if the pre-SN system was close to the actual position,<br />
in such a way that the progenitor of the compact object could have also contributed<br />
to the formation of the bubble. In such a case, the X-ray binary would be only about<br />
∼ 2 × 10 5 yr old, and we would expect to find the radio SNR originated in the SN<br />
event with a current surface brightness of Σ408 ∼ 3×10 −22 W m −2 Hz −1 sr −1 (Caswell<br />
& Lerche 1979). Filtering techniques of the diffuse galactic emission (e.g., Combi<br />
et al. 1998) allow to <strong>de</strong>tect very low brightness SNRs with Σ408 ∼ 10 −22 W m −2<br />
Hz −1 sr −1 , but nothing has been found in this region except for SNR G016.8−01.1,<br />
whose surface brightness at 408 MHz would be Σ408 ∼ 6 × 10 −22 W m −2 Hz −1 sr −1<br />
for a typical spectral in<strong>de</strong>x α = −0.5. This level of flux is expected at this latitu<strong>de</strong><br />
for a younger SNR, say with t ∼ 10 5 yr, but in such a case the particle <strong>de</strong>nsity of<br />
the medium around SNR G016.8−01.1 should be higher than the estimates given in<br />
Sect. 5.2, with values of n ∼ 50 cm −3 , in or<strong>de</strong>r to confine the SNR to its inferred size<br />
at d ∼ 2.9 kpc. These <strong>de</strong>nsities are 1 or<strong>de</strong>r of magnitu<strong>de</strong> higher than those <strong>de</strong>rived<br />
from the H i observations. However, a clear rejection of SNR G016.8−01.1 cannot be<br />
established on this basis alone, due to the flux contamination from the H ii region,<br />
which could be responsible for errors as large as 100 % in the flux <strong>de</strong>nsity estimate<br />
of the source. Consequently, a picture where the H i bubble has been created by<br />
both the stellar winds from the binary plus the SN explosion remains as an open<br />
possibility.<br />
3.6 Discussion<br />
There are two ways to accelerate a binary system by a supernova explosion: ejec-<br />
tion of material from the binary in a symmetric SN or an additional velocity kick,<br />
produced by asymmetries in the SN itself. In Nelemans et al. (1999), the authors<br />
conclu<strong>de</strong> that there is no need of additional velocity kicks in systems like Cygnus X-1<br />
or GRO J1655−40, both containing black hole candidates as the compact objects,<br />
to explain their observed large runaway velocities. On the other hand, from the<br />
study of Be/X-ray binaries containing neutron stars, van <strong>de</strong>n Heuvel et al. (2000)<br />
conclu<strong>de</strong> that, in or<strong>de</strong>r to explain the observed large eccentricities and low space<br />
velocities of these systems, a kick in an asymmetric SN explosion scenario is nee<strong>de</strong>d.<br />
Moreover, the microquasar Circinus X-1, containing a neutron star, seems to have
3.6. Discussion 109<br />
experienced a highly asymmetric SN explosion (Tauris et al. 1999). Unfortunately,<br />
the nature of the compact object in LS 5039 still remains unknown. However, we<br />
can try to study it in the symmetric SN scenario.<br />
Let MX and MO be the masses of the compact object and the optical companion,<br />
respectively. Taking into account the stellar evolution tracks by Schaerer et al.<br />
(1996), we can assume that MO 40 M⊙. Consi<strong>de</strong>ring that it must rotate slower<br />
than the breakup speed, and using the mass function f(m) = 0.00103±0.00020 M⊙<br />
by McSwain et al. (2001), an upper limit of MX < 9 M⊙ is obtained. The minimum<br />
mass for the compact object would be that of a neutron star of the Chandrasekar<br />
mass, MX > 1.4 M⊙. Hence, a total mass of ∼ 41–49 M⊙ remains in the system.<br />
Since the actual eccentricity of the system is e = 0.41 ± 0.05, and has a short<br />
orbital period of P = 4.117 ± 0.011 days, both parameters are expected to <strong>de</strong>-<br />
crease with time due to tidal forces which act to re-circularize the orbit. Hence,<br />
the eccentricity just after the SN explosion was epost−SN > 0.41. Using the equation<br />
∆ M = epost−SN (MX + MO) and MX = 1.4 M⊙ we find that at least 17 M⊙ were<br />
lost in the SN explosion in the symmetric case. If we adopt MX = 9 M⊙ the min-<br />
imum mass loss is 20 M⊙. Moreover, in or<strong>de</strong>r to have the system bound after the<br />
explosion, i.e., epost−SN < 1, the mass loss during the SN event must be less than<br />
the remaining mass in the system, i.e., < 49 M⊙. Hence, the values for the reduced<br />
mass of the system, <strong>de</strong>fined as µ = 1/(1 + epost−SN), range from 0.5 to 0.7. Using<br />
Eq. (5) of Nelemans et al. (1999) with the actual values of P and e we find that<br />
the re-circularized period will be Pre−circ 3 days. Since the period of the initial,<br />
pre-SN, binary system is given by Pi = µ 2 Pre−circ, the orbital period before the SN<br />
explosion was in the range 0.8 to 1.5 days. Using Kepler’s third law and assuming<br />
the latter value to allow the maximum separation possible between the progenitor<br />
stars, initial total masses between 60–100 M⊙ lead to semimajor axis in the range<br />
22–26 R⊙. This means that the two stars did not evolve separately, but probably<br />
through a common envelope phase. Hence, a <strong>de</strong>tailed evolution of the system should<br />
be carried out in or<strong>de</strong>r to try to compute the initial mass of the progenitor star of<br />
the compact object, the maximum age of the system before the SN event, and the<br />
mass loss during the SN explosion itself. This kind of study is beyond the scope of<br />
this work.<br />
Nevertheless, we can try to explain the measured vsys in the context of the<br />
symmetric SN explosion scenario using the following equation from Nelemans et al.
110 Chapter 3. LS 5039 as a runaway microquasar<br />
(1999):<br />
<br />
vsys<br />
km s −1<br />
1<br />
−<br />
∆ M MO Pre−circ<br />
= 213<br />
M⊙ M⊙ day<br />
3 MX + MO<br />
M⊙<br />
− 5<br />
3<br />
. (3.4)<br />
Assuming that MO = 40 M⊙ and Pre−circ = 3 days we can find two minimum values<br />
of vsys <strong>de</strong>pending on the adopted mass for the compact object. For MX = 1.4 M⊙,<br />
and hence ∆ M > 17 M⊙, we find vsys 200 km s −1 . Similarly, for MX = 9 M⊙,<br />
and hence ∆ M > 20 M⊙, we find vsys 180 km s −1 . Our measured recoil velocity<br />
is vsys 150 ± 20 km s −1 , which would be perfectly reached in the SN explosion<br />
scenario, as we have just seen. In fact, the measured value is a little bit lower than<br />
the computed ones, which could indicate that part of the eccentricity of the system<br />
was reached thanks to a kick in an asymmetric SN explosion. However, we must<br />
be cautious with the use of the value for vsys obtained after Eq. 3.4, because some<br />
effects have been ignored, as pointed out by Nelemans et al. (1999), and also the<br />
errors in e and Pre−circ could reduce the computed values. Hence, we can say that<br />
both, the high eccentricity measured by McSwain et al. (2001), and the high space<br />
velocity reported here, are perfectly compatible, within errors, with a symmetric<br />
SN explosion scenario. In other words, the actual eccentricity of the system is<br />
naturally explained with the mass loss necessary to explain the measured space<br />
(recoil) velocity of LS 5039. We note that McSwain & Gies (2002) have recently<br />
arrived to very similar conclusions using our proper motions.<br />
Another common mechanism referred to produce runaway stars is based on dy-<br />
namical ejection from young open clusters. Numerical simulations (Kiseleva et al.<br />
1998) show that it can explain some mo<strong>de</strong>rately high-velocity stars observed in the<br />
Galactic disk, but only about 1% of the stars would be ejected with velocities higher<br />
than 30 km s −1 . Hence, it seems difficult to explain the high observed velocity of<br />
LS 5039 as a result of such a mechanism.<br />
It is interesting to note that, contrary to other HMXB with low orbital peri-<br />
ods, LS 5039 is still in the circularization process. Using the actual value of P , a<br />
typical value of R = 10.3 R⊙ for the radius of the O6.5V star, and different for-<br />
malisms (Claret et al. 1995; Claret & Cunha 1997) we obtain typical circularization<br />
timescales of the or<strong>de</strong>r of (0.5 – 5) × 10 5 yr. These values are compatible with the<br />
system being still in the circularization process, since its formation just in the galac-<br />
tic midplane would take 5 × 10 5 yr to bring it to the actual position according to<br />
the integration of its trajectory.
3.6. Discussion 111<br />
At this point we might ask what is the point of studying runaway microquasars.<br />
In the case of LS 5039, if the SN explosion had taken place recently, we could expect<br />
the system to survive for the following ∼ 5 Myr, because the O6.5 star is still in<br />
the main sequence. Hence, according to the computed trajectory, it would reach<br />
a height of Z −600 pc, which translates into b −12 ◦ . On the other hand,<br />
LS 5039 could be associated with the high energy γ-ray source 3EG J1824−1514,<br />
as suggested by Pare<strong>de</strong>s et al. (2000). Hence, if this association is correct, we could<br />
be able to <strong>de</strong>tect γ-ray microquasars up to values of |b| > 10 ◦ , although practically<br />
all confirmed microquasars lie within |b| < 5 ◦ . In particular, runaway microquasars<br />
could be connected with some of the uni<strong>de</strong>ntified faint, variable, and soft γ-ray<br />
EGRET sources above/below the galactic plane, as suggested by Romero (2001)<br />
and Mirabel et al. (2001).<br />
Hence, it would be interesting to study the proper motions and radial velocities<br />
of as many microquasars as possible. Unfortunately, this kind of information is not<br />
easy to obtain for a wi<strong>de</strong> sample of systems. Proper motions can be obtained either<br />
from optical data or either from radio data. However, most optical counterparts<br />
of these sources are faint objects not present in old astrometric catalogs. Hence, it<br />
is mandatory to acquire new positions. In this context, it is better to do it in the<br />
radio domain because the uncertainties are always smaller than those obtained in<br />
the optical. On the other hand, it is necessary to carry out optical spectroscopy to<br />
obtain the radial velocity curve of these sources and use the correct value for the<br />
radial velocity of the system.<br />
An alternative approach could be the search for signatures of bow shocks in the<br />
microquasar vicinity. However, recent studies by Huthoff & Kaper (2002) of runaway<br />
OB stars, reveal that the success of this method is highly <strong>de</strong>pendant on the <strong>de</strong>nsity<br />
of the ISM around the runaway object. In particular, in only one object, namely<br />
Vela X-1, a bow shock has been <strong>de</strong>tected. Finally, one could search for signatures<br />
of explosive events in the ISM. Sensitive spectral line observations in the radio are<br />
probably the best tool for this purpose, as done in GRO J1655−40 by Combi et al.<br />
(2001).<br />
In any case, the study of runaway microquasars such as LS 5039 is likely to<br />
contribute significantly to different areas of mo<strong>de</strong>rn high-energy Astrophysics with<br />
in<strong>de</strong>pen<strong>de</strong>nce of the observing technique.
112 Chapter 3. LS 5039 as a runaway microquasar<br />
3.7 Conclusions<br />
After an in-<strong>de</strong>pth study of the proper motions and surroundings of LS 5039 our<br />
main conclusions are:<br />
1. Positions at optical and radio wavelengths have been used to compute in<strong>de</strong>-<br />
pen<strong>de</strong>nt optical and radio proper motions, which are perfectly compatible.<br />
Therefore, based only on astrometric data we are able to confirm that both,<br />
the optical and the radio emission, originate in the same object.<br />
2. From the combined optical and radio positions we have computed an accurate<br />
proper motion for LS 5039. This, together with the new estimate of 2.9 ±<br />
0.3 kpc for the distance (Sect. 2.5.3) and the radial velocity of the system<br />
(McSwain et al. 2001), allows us to compute a space velocity of (U = 51,<br />
V = −71, W = −118) km s −1 in its Regional Standard of Rest (RSR). This<br />
result implies that LS 5039 is a runaway microquasar with vsys 150 km s −1 ,<br />
escaping from its own RSR with a large velocity component perpendicular to<br />
the galactic plane. This is probably the result of the SN event that created<br />
the compact object in this binary system.<br />
3. We have computed the past trajectory of LS 5039. Two OB associations have<br />
been found close to its path in the plane of the sky. However, they are too close<br />
to us to be related to the microquasar. On the other hand, we have also found<br />
three SNRs near the path of LS 5039. After discarding two of them based on<br />
distance arguments, we have focused our attention on SNR G016.8−01.1. A<br />
study of this source could not clearly confirm nor reject the association due to<br />
the large uncertainties in the estimated radio flux <strong>de</strong>nsity of the SNR. This fact<br />
perhaps justifies future, high sensitivity searches of low-brightness remnants<br />
in this region.<br />
4. We have found a semi-open H i cavity close to the LS 5039 position. Although<br />
the O((f)) star in this microquasar seems to be the main agent forming the<br />
bubble, a contribution from the progenitor of the compact object cannot be<br />
ruled out.<br />
5. We have been able to explain both, the high space velocity and the high<br />
eccentricity observed, in a symmetric SN explosion scenario with a mass loss<br />
of ∆ M ∼ 17 M⊙.
3.7. Conclusions 113<br />
6. Finally, the high space velocity and possible lifetime of this microquasar indi-<br />
cate that it could reach a galactic latitu<strong>de</strong> of b = −12 ◦ . Therefore, if the pro-<br />
posed association between LS 5039 and the EGRET source 3EG J1824−1514 is<br />
correct, we could be able to <strong>de</strong>tect γ-ray microquasars up to values of |b| 10 ◦ .<br />
In particular, runaway microquasars could be connected with some of the<br />
uni<strong>de</strong>ntified faint, variable, and soft γ-ray EGRET sources above/below the<br />
galactic plane.
114 Chapter 3. LS 5039 as a runaway microquasar
Bibliography<br />
Ankay, A., Kaper, L., <strong>de</strong> Bruijne, et al. 2001, A&A, 370, 170<br />
Benaglia, P., & Cappa, C. E. 1999, A&A, 346, 979<br />
Caswell, J. L., & Lerche, I. 1979, MNRAS, 187, 201<br />
Claret, A., & Cunha, N. C. S. 1997, A&A, 318, 187<br />
Claret, A., Giménez, A., & Cunha, N. C. S. 1995, A&A, 299, 724<br />
Combi, J. A., Romero, G. E., & Benaglia, P. 1998, A&A, 333, L91<br />
Combi, J. A., Romero, G. E., Benaglia, P., & Mirabel, I. F. 2001, A&A, 370, L5<br />
Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693<br />
Dauphole, B., & Colin, J. 1995, A&A, 300, 117<br />
Dubner, G., Giacani, E., Reynoso, E., et al. 1999, AJ, 118, 930<br />
Fich, M., Blitz, L., & Stark, A. A. 1989, ApJ, 342, 272<br />
Forbes, D. 2000, AJ, 120, 2594<br />
Green, D. A. 2000, A Catalog of Galactic Supernova Remnants, Mullard Radio<br />
Astronomy Observatory, Cambridge, England, UK (available on the World Wi<strong>de</strong><br />
Web at http://www.mrao.cam.ac.uk/surveys/snrs/)<br />
Høg, E., Fabricius, C., Makarov, V. V., et al. 2000, A&A, 357, 367<br />
Huthoff, F., & Kaper, L. 2002, A&A, 383, 999<br />
115
116 BIBLIOGRAPHY<br />
Kaper, L., Camerón, A., & Barziv, O. 1999, in Proc. of the IAU Symp. 193: Wolf-<br />
Rayet phenomena in massive stars and starburst galaxies, ed. K. A. van <strong>de</strong>r Hucht,<br />
G. Koenigsberger, and P. R. J. Eenens, ASP Conference Series, 193, p. 316<br />
Kiseleva, L. G., Colin, J., Dauphole, B., & Eggleton, P. 1998, MNRAS, 301, 759<br />
Lasker, B. M., Sturch, C. R., McLean, B. J., et al. 1990, AJ, 99, 2019<br />
Martí, J., Pare<strong>de</strong>s, J. M., & Ribó, M. 1998, A&A, 338, L71<br />
McSwain, M. V., & Gies, D. R. 2002, ApJ, 568, L27<br />
McSwain, M. V., Gies, D. R., Riddle, R. L., Wang, Z., & Wingert, D. W. 2001,<br />
ApJ, 558, L43<br />
Melnik, A. M., & Efremov, Y. N. 1997, VizieR Online Data Catalog, 902, 110013<br />
Mirabel, I. F., Dhawan, V., Mignani, R. P., Rodrigues, I., & Guglielmetti, F. 2001,<br />
Nature, 413, 139<br />
Monet, D. G., Bird, A., Canzian, B., et al. 1999, USNO-A2.0 CD-ROM (U.S. Naval<br />
Observatory, Washington DC)<br />
Morrison, J. E., Röser, S., McLean, B., Bucciarelli, B., & Lasker, B. 2001, AJ, 121,<br />
1752<br />
Nelemans, G., Tauris, T. M., & van <strong>de</strong>n Heuvel, E. P. J. 1999, A&A, 352, L87<br />
Ochsenbein, F., Bauer, P., & Marcout, J. 2000, A&AS, 143, 23<br />
Pare<strong>de</strong>s, J. M., Martí, J., Ribó, M., & Massi, M. 2000, Science, 288, 2340<br />
Reed, B. C. 2000, AJ, 120, 314<br />
Reich, W., Fuerst, E., Reich, P., Sofue, Y., & Handa, T. 1986, A&A, 155, 185<br />
Reynoso, E. M., & Mangum, J. G. 2000, ApJ, 545, 874<br />
Ribó, M., Pare<strong>de</strong>s, J. M., Martí, J., & Massi, M. 2002, in preparation<br />
Rodgers, A. W., Campbell, C. T., & Whiteoak, J. B. 1960, MNRAS, 121, 103<br />
Romero, G. E. 2001, in Proc. of The Nature of Uni<strong>de</strong>ntified Galactic High-energy<br />
Gamma-ray Sources, ed. A. Carramiñana, O. Reimer, and D. J. Thompson, ASSL<br />
Series of Kluwer Aca<strong>de</strong>mic Publishers 51, p. 65
BIBLIOGRAPHY 117<br />
Schaerer, D., <strong>de</strong> Koter, A., Schmutz, W., & Mae<strong>de</strong>r, A. 1996, A&A, 312, 475<br />
Sedov, L. I. 1959, Similarity and Dimensional Methods in Mechanics, John Wiley<br />
and Sons, New York<br />
Shahbaz, T., van <strong>de</strong>r Hooft, F., Casares, J., Charles, P. A., & van Paradijs, J. 1999,<br />
MNRAS, 306, 89<br />
Shklovskii, I. S. 1968, Supernovae, John Wiley and Sons, New York<br />
Spitzer, L. 1998, Physical Processes in the Interstellar Medium, John Wiley and<br />
Sons, New York<br />
Tasker, N. J., Condon, J. J., Wright, A. E., & Griffith, M. R. 1994, AJ, 107, 2115<br />
Tauris, T. M., Fen<strong>de</strong>r, R. P., van <strong>de</strong>n Heuvel, E. P. J., Johnston, H. M., & Wu, K.<br />
1999, MNRAS, 310, 1165<br />
Torra, J., Fernán<strong>de</strong>z, D., & Figueras, F. 2000, A&A, 359, 82<br />
Urban, S. E., Corbin, T. E., Wycoff, G. L., et al. 1998, AJ, 115, 1212<br />
van <strong>de</strong>n Heuvel, E. P. J. 1978, in Physics and Astrophysics of Neutron Stars and<br />
Black Holes, ed. R. Giacconi and R. Ruffini (Amsterdam: North Holland), 828<br />
van <strong>de</strong>n Heuvel, E. P. J., Portegies Zwart, S. F., Bhattacharya, D., & Kaper, L.<br />
2000, A&A, 364, 563<br />
Zacharias, N., & Zacharias, M. I. 1999, AJ, 118, 2503<br />
Zacharias, N., Zacharias, M. I., & <strong>de</strong> Vegt, C. 1999, AJ, 117, 2895<br />
Zacharias, N., Urban, S. E., Zacharias, M. I., et al. 2000, AJ, 120, 2131
118 BIBLIOGRAPHY
Part II<br />
A search for new microquasars<br />
119
Chapter 4<br />
The cross-i<strong>de</strong>ntification method<br />
4.1 Introduction<br />
As discussed in Sect. 1.3, the reduced population of known microquasars gave us a<br />
strong motivation to start a long-term project focused on the search for new micro-<br />
quasars in the Galaxy. In this context, X-ray, optical and radio catalogs provi<strong>de</strong> a<br />
fundamental tool for the search of new sources with known multiwavelength behav-<br />
ior. Therefore, we have used the best available X-ray and radio catalogs to search<br />
for new REXB candidates on the basis of positional coinci<strong>de</strong>nce, together with very<br />
restrictive selection criteria. As a result we have obtained a sample of 13 radio<br />
emitting X-ray sources, which is expected to contain some galactic X-ray binaries.<br />
4.2 Cross-i<strong>de</strong>ntification between RBSC and NVSS<br />
catalogs<br />
In or<strong>de</strong>r to find new microquasar candidates we have to search for new REXBs. The<br />
first step, hence, has consisted of a positional cross-i<strong>de</strong>ntification 1 of X-ray and radio<br />
sources using the most complete available catalogs at both wavelengths, namely the<br />
ROSAT all sky Bright Source Catalog (RBSC) and the NRAO VLA Sky Survey<br />
(NVSS), which are briefly introduced now.<br />
1 Published in Pare<strong>de</strong>s, J. M., Ribó, M., & Martí, J. 2002, A&A, 394, 193.<br />
121
122 Chapter 4. The cross-i<strong>de</strong>ntification method<br />
In the X-ray domain, the RBSC (Voges et al. 1999) contains in its current<br />
version a total of 18 806 sources in the energy band 0.1–2.4 keV, and is <strong>de</strong>rived<br />
from the ROSAT All Sky Survey (Voges et al. 1996). Four energy bands in the<br />
following ranges of keV are provi<strong>de</strong>d: A=0.1–0.4, B=0.5–2.0, C=0.5–0.9, D=0.9–<br />
2.0. From these bands, two hardness ratios are computed: HR1 = (B −A)/(B +A),<br />
HR2 = (D − C)/(D + C). The 1σ positional uncertainties are typically in the range<br />
10–20 ′′ .<br />
In the radio domain, the NVSS (Condon et al. 1998) catalog covers the sky<br />
north of δ = −40 ◦ (82% of the celestial sphere) at a frequency of 1.4 GHz (20 cm<br />
wavelength) using the VLA configurations D and DnC. It contains over 1.8 × 10 6<br />
sources stronger than its 2.5 mJy completeness limit. The rms positional uncer-<br />
tainties are less than 1 ′′ for sources stronger than 15 mJy and 7 ′′ for the faintest<br />
<strong>de</strong>tectable sources.<br />
Since our aim is to maximize the probability of retaining only galactic REXB<br />
systems by cross-i<strong>de</strong>ntifying the RBSC and the NVSS catalogs, we have adopted the<br />
following selection criteria:<br />
1. Sources with absolute galactic latitu<strong>de</strong> < 5 ◦ have been selected from the RBSC<br />
catalog. Since the NVSS has a limit of δ > −40 ◦ , only the RBSC sources<br />
above this <strong>de</strong>clination have been allowed to continue in the selection process.<br />
Therefore, approximately 75% of the whole |b| < 5 ◦ area is covered (l 347–<br />
259 ◦ and α 17.2–8.6 h ).<br />
2. Among the remaining RBSC sources, those containing screening flags about<br />
nearby sources contaminating measurements or problems with position <strong>de</strong>ter-<br />
minations (Voges et al. 1999) have been rejected.<br />
3. From statistical RBSC i<strong>de</strong>ntification studies, Motch et al. (1998) conclu<strong>de</strong>d<br />
that X-ray binaries are essentially recognizable from their hard X-ray spectra.<br />
Therefore, in an attempt to avoid AGNs and cataclysmic variables, we have<br />
only selected the sources with HR1 + σ(HR1) higher than 0.9. Although<br />
this criterion would exclu<strong>de</strong> microquasars in an outburst, we note that this<br />
situation only lasts a short time of their life, and the probability of discarding<br />
one of them is very low. A total of 241 RBSC sources remain in our sample<br />
at this stage.
4.2. Cross-i<strong>de</strong>ntification between RBSC and NVSS catalogs 123<br />
4. The X-ray and radio positions must agree within errors. For this purpose, we<br />
have selected NVSS sources within the 2σ (95% probability) error boxes of<br />
the RBSC sources. We have also used the constraint of a maximum offset of<br />
40 ′′ between the RBSC and the NVSS positions, since, for higher distances,<br />
no reliable i<strong>de</strong>ntification is expected (Voges et al. 1999).<br />
5. No exten<strong>de</strong>d radio source has been allowed to continue in the selection process,<br />
since any REXB is expected to appear compact at the NVSS resolution 2 .<br />
The sources selected with the RBSC/NVSS cross-i<strong>de</strong>ntification, a total of 35,<br />
were then filtered with complementary optical information using the following cri-<br />
teria:<br />
1. We have inspected the SIMBAD database and the NASA/IPAC Extragalactic<br />
Database (NED) for each source, and if it is listed as an extragalactic object,<br />
we have obviously rejected it from the sample.<br />
2. We have also inspected the Digitized Sky Survey, DSS1 and DSS2-red images 3 ,<br />
and looked for possible optical counterparts in position agreement with the<br />
NVSS sources. If an optical object is present and displays exten<strong>de</strong>d emission,<br />
i.e., with a galaxy-like appearance, it has also been removed from the sample.<br />
At the end of this process the resulting sample contained 16 sources. Among<br />
them, we found the well known REXB LS I +61 303, which displays a one-si<strong>de</strong>d<br />
jet at milliarcsecond scales (Massi et al. 2001). We also recovered the well known<br />
microquasars LS 5039, SS 433 and Cyg X-3. We show in Fig. 4.1 the distribution<br />
of known microquasars in galactic coordinates, together with the area of the sky<br />
covered by our search and the recovered sources clearly marked. It is interesting to<br />
note that all these sources are HMXB, and none of the known microquasars with<br />
low mass companions were found by this technique. This can be easily explained<br />
by the fact that most of LMXB are transients, and the remaining persistent ones<br />
are too faint to be present in the NVSS. Moreover, all but one HMXB persistent<br />
microquasars in our explored range of galactic latitu<strong>de</strong>s, namely LS 5039, SS 433<br />
2 The sources resolved only in one axis and with an angular size smaller than the other axis were<br />
allowed to continue on the process.<br />
3 http://archive.eso.org/dss/dss/
124 Chapter 4. The cross-i<strong>de</strong>ntification method<br />
and Cyg X-3, have been recovered by our cross-i<strong>de</strong>ntification method. The remain-<br />
ing one, namely Cyg X-1, although marginally present in the corresponding NVSS<br />
image, does not appear in the NVSS catalog. This was probably due to the source<br />
low flux <strong>de</strong>nsity and diffuse background around it, which prevented its automatic<br />
<strong>de</strong>tection by the source search software.<br />
Galactic latitu<strong>de</strong> [<strong>de</strong>g]<br />
90<br />
60<br />
30<br />
0<br />
−30<br />
−60<br />
XTE J1118+480<br />
LS I +61 303<br />
Microquasars in the Galaxy<br />
Cyg X−1<br />
GRS 1758−258<br />
GRS 1915+105<br />
SS 433<br />
Cyg X−3<br />
LS 5039<br />
Sco X−1<br />
V4641 Sgr<br />
XTE J1748−288<br />
GRO J1655−40<br />
Cir X−1<br />
GX 339−4<br />
1E1740.7−2942<br />
XTE J1550−564<br />
LMXB<br />
LMXB (rel. jets ?)<br />
HMXB<br />
HMXB (rel. jets ?)<br />
−90<br />
180 120 60 0 300 240 180<br />
Galactic longitu<strong>de</strong> [<strong>de</strong>g]<br />
Figure 4.1: Distribution of known microquasars in galactic coordinates. Filled circles<br />
represent those sources where relativistic jets have been imaged, while open circles are<br />
used for those where hints of relativistic jets have been seen or are clearly suspected.<br />
The solid lines <strong>de</strong>limit the area of the sky covered by our search, where the recovered<br />
sources are marked in boldface.<br />
After removing the previously known sources of our sample, we en<strong>de</strong>d with a<br />
total of 12 new uni<strong>de</strong>ntified REXB candidates. Among them, 7 had offsets between<br />
the X-ray and radio positions within the 1σ RBSC position error, and belong to the<br />
hereafter Group 1 sample. The other 5 sources had offsets between 1–2σ and form
4.2. Cross-i<strong>de</strong>ntification between RBSC and NVSS catalogs 125<br />
the Group 2 sample.<br />
When we started this project in 1998 another source fullfiled all the selection<br />
criteria. Now we have performed the cross-i<strong>de</strong>ntification process again, in or<strong>de</strong>r to<br />
present here an updated version of it, and we have found that this source, namely<br />
1RXS J072418.3−071508, was i<strong>de</strong>ntified by Perlman et al. (1998) as a quasar<br />
(WGA J0724.3−0715 in their paper and listed as PMN J0724−0715 in the NED<br />
database). Although it is clear that it is not any more a REXB candidate, we have<br />
preferred to inclu<strong>de</strong> it in the list and present the observational results obtained up<br />
to now in the following chapters. Since the corresponding offset in position for this<br />
object is less than 1σ, it is inclu<strong>de</strong>d in Group 1.<br />
All the sources belonging to Group 1 (a total of 8), Group 2 (a total of 5)<br />
and the already known REXBs are listed in Table 4.1, where Cols. 1 to 4 contain<br />
the RBSC names (which also provi<strong>de</strong> the positions), the 1σ errors in position, the<br />
count rates, and HR1. In Col. 5 we show the offsets between the RBSC and the<br />
NVSS positions. In Cols. 6 to 8 we list the NVSS names (which provi<strong>de</strong> a limited<br />
position information), the 1σ errors in position and the NVSS flux <strong>de</strong>nsities (at<br />
20 cm wavelength). Finally, galactic coordinates for all sources are listed in Cols. 9<br />
and 10. The horizontal line divi<strong>de</strong>s the two groups of sources and the already known<br />
REXBs.<br />
Here we will focus on the Group 1 sources, while Group 2 sources will be eventu-<br />
ally studied in the future. In or<strong>de</strong>r to have a look at the targets, we show in Fig. 4.2,<br />
for each source of Group 1, the NVSS radio contours overlaid on 6 ′ ×6 ′ optical DSS1<br />
images. To mark the RBSC positions we have plotted as open crosses the 3σ errors<br />
in position, allowing a better display than if we had plotted the 1σ errors.<br />
As can be seen in Table 4.1, the NVSS errors are much smaller than the RBSC<br />
ones. Hence, the radio positions are more accurate than the X-ray ones to look<br />
for optical counterparts. An inspection of the DSS1 images in Fig. 4.2 reveals that<br />
some candidates have an optical counterpart approximately in the middle of the<br />
radio contours, while others do not. This is un<strong>de</strong>rstandable because of the low<br />
galactic latitu<strong>de</strong> of these objects, which implies a high <strong>de</strong>gree of extinction along<br />
the line of sight. On the other hand, crow<strong>de</strong>d fields near the galactic plane also<br />
prevent us to be sure about the counterparts <strong>de</strong>tected. Therefore, the next step of<br />
the process is to obtain accurate radio positions and then search for possible optical<br />
counterparts, which is presented in the next chapter.
126 Chapter 4. The cross-i<strong>de</strong>ntification method<br />
Table 4.1: Selected sources from the RBSC/NVSS cross-i<strong>de</strong>ntification. Cols. 1 to 4<br />
contain the RBSC name (which also provi<strong>de</strong>s the position), the 1σ error in position,<br />
the count rate, and the hardness ratio 1 for each source. In Col. 5 we show the offset<br />
between the RBSC and the NVSS positions, while in Cols. 6 to 8 we list the NVSS name<br />
(constructed with truncated coordinates), the 1σ error in position and the NVSS flux<br />
<strong>de</strong>nsity for each source. Galactic coordinates for all sources are listed in Cols. 9 and<br />
10. The horizontal lines divi<strong>de</strong> Group 1 (top), Group 2 (middle), and the already known<br />
REXB sources (bottom). The source 1RXS J072418.3−071508 is a quasar (see text).<br />
RBSC NVSS Gal. coord.<br />
1RXS name Pos. err. Count rate HR1 Off. NVSS name Pos. err. Flux <strong>de</strong>nsity l b<br />
[ ′′ ] [10 −2 s −1 ] [ ′′ ] [ ′′ ] [mJy] [ ◦ ] [ ◦ ]<br />
J001442.2+580201 9 8.5 ± 1.4 1.00 ± 0.12 3 J001441+580202 3 7.1 ± 0.5 118.07 −4.49<br />
J013106.4+612035 7 25.0 ± 2.4 0.90 ± 0.05 6 J013107+612033 1 19.1 ± 0.7 127.67 −1.16<br />
J042201.0+485610 14 5.1 ± 1.1 1.00 ± 0.24 4 J042200+485607 7 2.3 ± 0.4 154.41 −0.63<br />
J062148.1+174736 8 8.8 ± 1.6 1.00 ± 0.13 4 J062147+174734 1 12.2 ± 0.5 193.78 +1.72<br />
J072259.5−073131 8 17.4 ± 2.1 0.94 ± 0.05 6 J072259−073135 1 84.1 ± 3.1 223.24 +3.52<br />
J072418.3−071508 23 5.2 ± 1.2 1.00 ± 0.13 19 J072417−071519 1 330.4 ± 9.9 223.15 +3.93<br />
(Quasar)<br />
J181119.4−275939 14 14.0 ± 3.1 0.87 ± 0.14 13 J181120−275946 1 27.9 ± 1.0 3.64 −4.43<br />
J185002.8−075833 13 6.1 ± 1.6 1.00 ± 0.24 10 J185002−075842 1 49.9 ± 1.6 25.66 −3.32<br />
J050339.8+451715 11 6.1 ± 1.3 0.85 ± 0.17 16 J050339+451658 1 34.3 ± 1.1 161.81 +2.31<br />
J080451.8−274924 11 7.0 ± 1.8 1.00 ± 0.15 14 J080451−274911 1 821 ± 25 245.80 +2.01<br />
J082404.3−302033 8 17.5 ± 2.4 0.96 ± 0.06 11 J082403−302038 1 85.0 ± 2.6 250.23 +4.11<br />
J190333.1+104355 14 10.8 ± 1.7 1.00 ± 0.04 16 J190333+104409 7 4.7 ± 0.6 43.86 +2.22<br />
J205644.3+494011 16 7.7 ± 1.0 1.00 ± 0.06 17 J205642+494005 1 167.3 ± 5.0 89.32 +2.76<br />
J024033.5+611358 13 5.4 ± 1.1 1.00 ± 0.17 19 J024031+611345 1 42.2 ± 1.3 135.68 +1.09<br />
(LS I +61 303)<br />
J182615.1−145034 11 6.5 ± 1.6 1.00 ± 0.16 21 J182614−145054 1 23.4 ± 0.9 16.88 −1.29<br />
(LS 5039)<br />
J191149.7+045857 8 52.6 ± 3.7 0.96 ± 0.03 3 J191149+045858 1 867 ± 26 39.69 −2.24<br />
(SS 433)<br />
J203226.2+405725 7 284.9 ± 6.5 0.98 ± 0.00 7 J203225+405728 1 87.3 ± 3.2 79.84 +0.69<br />
(Cyg X-3)
4.2. Cross-i<strong>de</strong>ntification between RBSC and NVSS catalogs 127<br />
DECLINATION (J2000)<br />
1RXS J001442.2+580201<br />
58 05<br />
04<br />
03<br />
02<br />
01<br />
00<br />
57 59<br />
00 15 00 14 55 50 45 40 35 30 25 20<br />
RIGHT ASCENSION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
1RXS J042201.0+485610<br />
48 59<br />
58<br />
57<br />
56<br />
55<br />
54<br />
-07 29<br />
30<br />
31<br />
32<br />
33<br />
34<br />
-27 57<br />
04 22 15 10 05 00 21 55 50 45<br />
RIGHT ASCENSION (J2000)<br />
1RXS J072259.5-073131<br />
07 23 10 05 00 22 55 50<br />
RIGHT ASCENSION (J2000)<br />
58<br />
59<br />
-28 00<br />
01<br />
02<br />
1RXS J181119.4-275939<br />
18 11 30 25 20 15 10<br />
RIGHT ASCENSION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
61 23<br />
22<br />
21<br />
20<br />
19<br />
18<br />
1RXS J013106.4+612035<br />
01 31 30 25 20 15 10 05 00 30 55 50 45<br />
RIGHT ASCENSION (J2000)<br />
17 50<br />
49<br />
48<br />
47<br />
46<br />
45<br />
1RXS J062148.1+174736<br />
06 22 00 21 55 50 45 40<br />
RIGHT ASCENSION (J2000)<br />
-07 13<br />
14<br />
15<br />
16<br />
17<br />
18<br />
-07 56<br />
57<br />
58<br />
59<br />
-08 00<br />
01<br />
1RXS J072418.3-071508<br />
07 24 25 20 15 10 05<br />
RIGHT ASCENSION (J2000)<br />
1RXS J185002.8-075833<br />
18 50 15 10 05 00 49 55<br />
RIGHT ASCENSION (J2000)<br />
Figure 4.2: Optical, radio and X-ray composition of Group 1 sources. The NVSS radio<br />
contours are overlaid on the 6 ′ × 6 ′ optical DSS1 images, whereas the open crosses<br />
<strong>de</strong>note the 3σ RBSC astrometric errors.
128 Chapter 4. The cross-i<strong>de</strong>ntification method<br />
4.3 Conclusions<br />
We have presented a cross-i<strong>de</strong>ntification method to search for REXBs, which are<br />
potential microquasar sources. The obtained results give confi<strong>de</strong>nce to the proposed<br />
method, since the output list of objects inclu<strong>de</strong>d all but one known persistent HMXB<br />
microquasars within |b| < 5 ◦ . The uni<strong>de</strong>ntified sources in the list have been divi<strong>de</strong>d<br />
in two groups, <strong>de</strong>pending on the offset between the RBSC and NVSS positions.<br />
The Group 1 source list contains 8 objects with offsets smaller than 1σ, while the<br />
Group 2 source list contains 5 objects with offsets between 1 and 2σ.<br />
Finally, since some microquasars could have been ejected from the galactic plane<br />
(Ribó et al. 2002), we are planning to extend this project up to higher galactic<br />
latitu<strong>de</strong>s (5 ◦ < |b| < 10 ◦ ).
Bibliography<br />
Condon, J. J., Cotton, W. D., Greisen, E. W., et al. 1998, AJ, 115, 1693<br />
Massi, M., Ribó, M., Pare<strong>de</strong>s, J. M., Peracaula, M., & Estalella, R. 2001, A&A,<br />
376, 217<br />
Motch, C., Guillout, P., Haberl, F., et al. 1998, A&AS, 132, 341<br />
Perlman, E. S., Padovani, P., Giommi, P., et al. 1998, AJ, 115, 1253<br />
Ribó, M., Pare<strong>de</strong>s, J. M., Romero, G. E., et al. 2002, A&A, 384, 954<br />
Voges, W., Boller, Th., Dennerl, K., et al. 1996, in Proc. of the Conference Rönt-<br />
genstrahlung from the Universe, MPE Report, 263, 637<br />
Voges, W., Aschenbach, B., Boller, Th., et al. 1999, A&A, 349, 389<br />
129
130 BIBLIOGRAPHY
Chapter 5<br />
Radio and optical observations<br />
5.1 Introduction<br />
After the cross-i<strong>de</strong>ntification of catalogs presented in the previous chapter, the next<br />
logical step of the process 1 is to obtain accurate radio positions, that will be ex-<br />
plained in Sect. 5.2, and then search for possible optical counterparts, that will be<br />
presented in Sect. 5.3. A discussion on the obtained results is carried out in Sect. 5.4,<br />
while we state our conclusions in Sect. 5.5.<br />
5.2 Radio observations<br />
The main goal of the radio observations was to obtain accurate sub-arcsecond po-<br />
sitions, but we also wished to monitor the variability of radio flux and spectrum<br />
of our targets. On the other hand, it was also important to investigate the source<br />
structure beyond the NVSS resolution. For REXBs, most of the source flux <strong>de</strong>n-<br />
sity is expected to be concentrated in a compact core plus possible arcsecond or<br />
sub-arcsecond exten<strong>de</strong>d features. By observing each target with the VLA in A con-<br />
figuration, we were able to verify to what extend our sources were in<strong>de</strong>ed compact,<br />
and to look for possible elongations or jets.<br />
To this end, multi-frequency and multi-epoch observations of the Group 1 sources<br />
1 Published in Pare<strong>de</strong>s, J. M., Ribó, M., & Martí, J. 2002, A&A, 394, 193.<br />
131
132 Chapter 5. Radio and optical observations<br />
were carried out with the NRAO 2 VLA in A configuration. We did not observe<br />
1RXS J181119.4−275939 and 1RXS J185002.8−075833, the last two sources of<br />
Group 1 in Table 4.1, because they had very different right ascensions, compared to<br />
the other sources, and were not visible during the scheduled VLA observing time.<br />
Three VLA A configuration observing sessions were carried out on 1999 July 9,<br />
22 and 30 at 3.6 cm (8.4 GHz) and 6 cm (5 GHz) wavelengths (frequencies). A<br />
typical observation consisted of 14 minutes at 3.6 cm and 4 minutes at 6 cm on each<br />
target, prece<strong>de</strong>d and followed by a 2 minute observation of a phase calibrator. Dur-<br />
ing the last observing session, July 30, we inclu<strong>de</strong>d two additional 20 cm (1.4 GHz)<br />
measurements of 1RXS J072259.5−073131 and 1RXS J072418.3−071508. The am-<br />
plitu<strong>de</strong> calibrator used in all cases was 3C 48, while we used the phase calibrator<br />
J0102+584 for the first two sources, J0359+509 and J0603+177 for the third and<br />
fourth sources, respectively, and finally J0730−116 for the last two sources. The<br />
data were edited and calibrated using the aips software package of NRAO.<br />
5.2.1 Results<br />
All sources were <strong>de</strong>tected at all frequencies, allowing us to obtain accurate positions<br />
and flux <strong>de</strong>nsity measurements. To achieve the first goal, we ma<strong>de</strong> use of the phase-<br />
reference technique to calibrate the data at 3.6 cm wavelength, the ones with the<br />
highest angular resolution. Then we concatenated all these data for each source,<br />
with the task dbcon within aips. Finally, we performed uniform weighted maps<br />
from which positions were obtained after fitting elliptical Gaussians using jmfit<br />
within aips. A realistic estimate of the position error is about 0.01 ′′ .<br />
In or<strong>de</strong>r to obtain good quality maps, once accurate positions were known, we<br />
carried out a similar process, but using phase self-calibration with the previously<br />
obtained positions for each individual snapshot at 3.6 cm wavelength. This was<br />
not possible for 1RXS J042201.0+485610 because the flux <strong>de</strong>nsity was too low and<br />
prevented self-calibration. The resulting uniform weighted maps, together with the<br />
optical images that will be presented in the following section, are shown in Fig. 5.1.<br />
The flux <strong>de</strong>nsities of all sources were measured as follows. First of all we self-<br />
2 The National Radio Astronomy Observatory is a facility of the National Science Foundation<br />
operated un<strong>de</strong>r cooperative agreement by Associated Universities, Inc.
5.2. Radio observations 133<br />
calibrated all snapshots at all wavelengths (again with the exception of the source<br />
1RXS J042201.0+485610). Then we performed natural weighted maps with im-<br />
proved dynamic range, from which flux <strong>de</strong>nsities were measured using again elliptical<br />
Gaussian fits.<br />
The obtained results for all observed sources are summarized in Table 5.1. In<br />
Col. 1 we list the RBSC object names, while in Cols. 2 and 3 we show the radio posi-<br />
tions <strong>de</strong>rived from our phase-referenced uniform weighted maps of all concatenated<br />
3.6 cm wavelength observations. In Col. 4 we list the observing dates and, in Cols. 5<br />
and 6, the corresponding flux <strong>de</strong>nsities at both observing wavelengths, measured on<br />
the self-calibrated natural weighted maps. The flux <strong>de</strong>nsity errors quoted are the<br />
rms noise of the maps in mJy beam −1 . In Col. 7 we show the computed spectral<br />
in<strong>de</strong>x α (<strong>de</strong>fined in such a way that Sν ∝ ν +α , where Sν is the flux <strong>de</strong>nsity at a<br />
given frequency ν) of each source for all epochs. The radio spectra corresponding<br />
to the values listed in Table 5.1 are plotted in Fig. 5.2 for all sources, including the<br />
20 cm wavelength flux <strong>de</strong>nsities measured on July 30 for the last two sources. We<br />
have also plotted, for all sources, the non-simultaneous 20 cm measurements from<br />
the NVSS survey obtained in VLA D configuration.<br />
Inspection of all obtained maps (with phase-reference or self-calibration tech-<br />
niques, with uniform or natural weights, of individual snapshots or after concate-<br />
nating all of them, and at all frequencies), reveals that all sources are point-like<br />
except 1RXS J072259.5−073131, which presents exten<strong>de</strong>d structure towards the<br />
east, marginally present in the 3.6 and 6 cm maps, and visible as a one-si<strong>de</strong>d jet in<br />
the 30 July snapshot observation at 20 cm wavelength (see Fig. 5.3).
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
134 Chapter 5. Radio and optical observations<br />
1RXS J001442.2+580201 I band<br />
58 03 00<br />
02 45<br />
30<br />
15<br />
00<br />
01 45<br />
30<br />
15<br />
58 02 02.0<br />
00 14 48 46 44 42 40 38 36<br />
RIGHT ASCENSION (J2000)<br />
01.5<br />
01.0<br />
00.5<br />
VLA 3.6cm<br />
00 14 42.25 42.20 42.15 42.10 42.05 42.00<br />
RIGHT ASCENSION (J2000)<br />
17 48 30<br />
15<br />
00<br />
47 45<br />
30<br />
15<br />
00<br />
46 45<br />
17 47 36.0<br />
35.5<br />
35.0<br />
34.5<br />
1RXS J062148.1+174736 I band<br />
06 21 51 50 49 48 47 46 45 44<br />
RIGHT ASCENSION (J2000)<br />
VLA 3.6cm<br />
06 21 47.82 47.80 47.78 47.76 47.74 47.72 47.70 47.68<br />
RIGHT ASCENSION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
1RXS J013106.4+612035 I band<br />
61 21 30<br />
15<br />
00<br />
20 45<br />
30<br />
15<br />
00<br />
19 45<br />
61 20 34.0<br />
01 31 14 12 10 08 06 04 02 00<br />
RIGHT ASCENSION (J2000)<br />
33.5<br />
33.0<br />
32.5<br />
-07 30 45<br />
VLA 3.6cm<br />
01 31 07.35 07.30 07.25 07.20 07.15 07.10<br />
RIGHT ASCENSION (J2000)<br />
31 00<br />
15<br />
30<br />
45<br />
32 00<br />
15<br />
30<br />
-07 31 34.0<br />
1RXS J072259.5-073131 I band<br />
07 23 03 02 01 00 22 59 58 57 56<br />
RIGHT ASCENSION (J2000)<br />
34.5<br />
35.0<br />
35.5<br />
VLA 3.6cm<br />
07 22 59.74 59.72 59.70 59.68 59.66 59.64 59.62<br />
RIGHT ASCENSION (J2000)<br />
(See figure caption on next page)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
DECLINATION (J2000)<br />
48 57 00<br />
56 45<br />
30<br />
15<br />
00<br />
55 45<br />
30<br />
15<br />
1RXS J042201.0+485610 I band<br />
04 22 06 04 02 00 21 58 56<br />
RIGHT ASCENSION (J2000)<br />
48 56 04.5<br />
04.0<br />
03.5<br />
03.0<br />
-07 14 30<br />
45<br />
15 00<br />
15<br />
30<br />
45<br />
16 00<br />
15<br />
VLA 3.6cm<br />
04 22 00.60 00.55 00.50 00.45<br />
RIGHT ASCENSION (J2000)<br />
1RXS J072418.3-071508 I band<br />
07 24 21 20 19 18 17 16 15 14<br />
RIGHT ASCENSION (J2000)<br />
-07 15 19.5<br />
20.0<br />
20.5<br />
21.0<br />
21.5<br />
22.0<br />
22.5<br />
23.0<br />
SE<br />
NW<br />
VLA 3.6cm<br />
07 24 17.45 17.40 17.35 17.30 17.25 17.20<br />
RIGHT ASCENSION (J2000)
5.3. Optical observations 135<br />
Figure 5.1: Optical images and radio maps, grouped in pairs, of the six sources listed in<br />
Table 5.1. The 2 ′ × 2 ′ optical images were obtained through the I Johnson filter with<br />
the CAHA 2.2 m telescope on 1999 December 11, and circles have been used to clearly<br />
mark the optical counterparts. The radio maps, computed using uniform weights, are<br />
the result of concatenating, for each source, all VLA 1999 July observations at 3.6 cm<br />
wavelength, while the crosses indicate the optical positions with the 1σ 0.3 ′′ astrometric<br />
errors. The radio maps are 2 ′′ × 2 ′′ wi<strong>de</strong> for all sources except for the last one, where a<br />
4 ′′ × 4 ′′ map has been plotted to allow the inclusion of the close SE optical object. The<br />
synthesized beam is plotted in the bottom left corner of each map. The radio contours<br />
start at a signal-to-noise ratio of 6 for all sources except for the last one, where a value<br />
of 15 has been used, in or<strong>de</strong>r to avoid reproducing cleaning effects at low signal-to-noise<br />
ratios in all cases.<br />
5.3 Optical observations<br />
Having in mind the goal of i<strong>de</strong>ntifying the optical counterparts, CCD observations<br />
were ma<strong>de</strong> at Calar Alto (Almería, Spain) with the 1.5 m telescope of the Observa-<br />
torio Astronómico Nacional (OAN), between 24 and 30 November 1998. We used<br />
the Ritchey-Chrétien focus and a Tektronics TK1024AB chip that provi<strong>de</strong>s a scale<br />
factor of 0.4 ′′ per pixel and a 6.9 ′ × 6.9 ′ field of view. Deep CCD images were ob-<br />
tained for all our candidates, with the exception of 1RXS J181119.4−275939 and<br />
1RXS J185002.8−075833 that were not visible at the epoch of our observations. The<br />
images were taken with the V , R and I Johnson filters, and the exposure times were<br />
typically of 600–1200 seconds. Although the images were useful for photometric<br />
measurements, their quality did not allow us to obtain accurate positions.<br />
Complementary CCD observations were conducted on 1999 December 11 also at<br />
Calar Alto, but using in this case the 2.2 m telescope of the Centro Astronómico<br />
Hispano-Alemán (CAHA). Images were obtained using the Ritchey-Chrétien focus<br />
and CAFOS, leading a scale factor of 0.53 ′′ per pixel and a 8.8 ′ × 8.8 ′ field of view,<br />
through the I Johnson filter and with exposure times between 180 and 600 seconds.<br />
The good quality of these images allowed us to perform accurate astrometric mea-<br />
surements, although no precise photometric information could be obtained since the<br />
images could not be photometrically calibrated.<br />
The observations were reduced using standard procedures within IRAF.
136 Chapter 5. Radio and optical observations<br />
Table 5.1: Radio positions, flux <strong>de</strong>nsities and spectral indices obtained after the<br />
VLA A configuration observations for all sources belonging to Group 1 except<br />
1RXS J181119.4−275939 and 1RXS J185002.8−075833. The coordinates are those<br />
obtained from the uniform weighted maps of all concatenated 3.6 cm wavelength obser-<br />
vations. The three rows of flux <strong>de</strong>nsities and spectral in<strong>de</strong>x for each source correspond<br />
to the three observing runs, i.e., July 9, 22, and 30, respectively. The flux <strong>de</strong>nsities at<br />
3.6 and 6 cm have been obtained from natural weighted self-calibrated maps for each<br />
source, with the exception of 1RXS J042201.0+485610, where self-calibration was not<br />
possible.<br />
1RXS name α (2000) δ (2000) Flux <strong>de</strong>nsity [mJy] Spectral in<strong>de</strong>x<br />
[h, m, s] [ ◦ , ′ , ′′ ] S3.6 cm S6 cm α3.6−6 cm<br />
J001442.2+580201 00 14 42.1262 +58 02 01.219 6.94 ± 0.04 7.90 ± 0.07 −0.24 ± 0.02<br />
± 0.0013 ± 0.010 6.36 ± 0.07 7.05 ± 0.09 −0.19 ± 0.03<br />
5.98 ± 0.06 6.47 ± 0.07 −0.14 ± 0.03<br />
J013106.4+612035 01 31 07.2267 +61 20 33.376 16.95 ± 0.04 17.95 ± 0.08 −0.10 ± 0.01<br />
± 0.0014 ± 0.010 16.62 ± 0.07 16.49 ± 0.09 +0.01 ± 0.01<br />
17.43 ± 0.07 17.95 ± 0.08 −0.05 ± 0.01<br />
J042201.0+485610 04 22 00.5244 +48 56 03.634 0.70 ± 0.03 0.30 ± 0.04 +1.6 ± 0.3<br />
± 0.0010 ± 0.010 0.74 ± 0.05 0.57 ± 0.05 +0.5 ± 0.2<br />
0.71 ± 0.03 0.35 ± 0.04 +1.3 ± 0.2<br />
J062148.1+174736 06 21 47.7522 +17 47 35.078 9.63 ± 0.04 8.37 ± 0.07 +0.26 ± 0.02<br />
± 0.0007 ± 0.010 9.53 ± 0.07 9.87 ± 0.07 −0.06 ± 0.02<br />
11.45 ± 0.05 10.40 ± 0.07 +0.18 ± 0.01<br />
J072259.5−073131 07 22 59.6809 −07 31 34.805 40.90 ± 0.05 47.67 ± 0.08 −0.280 ± 0.004<br />
(a) ± 0.0007 ± 0.010 44.29 ± 0.07 44.5 ± 0.1 −0.009 ± 0.005<br />
45.37 ± 0.08 54.29 ± 0.08 −0.329 ± 0.004<br />
J072418.3−071508 07 24 17.2912 −07 15 20.339 438.5 ± 0.2 423.4 ± 0.3 +0.064 ± 0.002<br />
(b) ± 0.0007 ± 0.010 455.75 ± 0.09 439.3 ± 0.2 +0.067 ± 0.001<br />
428.5 ± 0.3 404.9 ± 0.3 +0.104 ± 0.002<br />
(a) Also observed at 20 cm on 1999 July 30, with a flux <strong>de</strong>nsity of 55.1 ± 0.4 mJy, and showing a<br />
one-si<strong>de</strong>d radio jet.<br />
(b) Also observed at 20 cm on 1999 July 30, with a flux <strong>de</strong>nsity of 248.3 ± 0.9 mJy.
5.3. Optical observations 137<br />
Flux <strong>de</strong>nsity [mJy]<br />
9<br />
8<br />
7<br />
6<br />
20<br />
18<br />
16<br />
3<br />
2<br />
1<br />
0.5<br />
0.3<br />
14<br />
12<br />
10<br />
8<br />
100<br />
80<br />
60<br />
40<br />
500<br />
400<br />
300<br />
200<br />
1RXS J001442.2+580201<br />
1RXS J013106.4+612035<br />
1RXS J042201.0+485610<br />
1RXS J062158.1+174736<br />
1RXS J072259.5−073131<br />
1RXS J072418.3−071508<br />
1 2 5 8 10<br />
Frequency [GHz]<br />
Figure 5.2: Radio spectra of the six sources listed in Table 5.1 at different epochs. The<br />
VLA data from July 9, 22 and 30 are <strong>de</strong>noted by filled circles, open circles and open<br />
triangles, respectively. The non-simultaneous NVSS data, at 20 cm wavelength, are<br />
plotted as filled squares. Both axes are in logarithmic scale. Error bars not visible are<br />
smaller than the symbol’s size.
138 Chapter 5. Radio and optical observations<br />
DECLINATION (J2000)<br />
-07 31 28<br />
30<br />
32<br />
34<br />
36<br />
38<br />
40<br />
42<br />
44<br />
1RXS J072259.5-073131 VLA 20cm<br />
07 23 00.2 00.0 22 59.8 59.6 59.4 59.2<br />
RIGHT ASCENSION (J2000)<br />
Figure 5.3: Natural weighted map of the self-calibrated 20 cm data of<br />
1RXS J072259.5−073131 obtained on the July 30 snapshot observation. Contours cor-<br />
respond to −3, 3, 4, 5, 7, 10, 15, 25, 40, 60, 85, 110 and 130 times 0.4 mJy beam −1 ,<br />
the rms noise.<br />
5.3.1 Results<br />
Promising optical counterparts for all observed sources, i.e., the first six ones in Ta-<br />
ble 4.1, were found in both observing runs. Thanks to the CAHA 2.2 m observations<br />
we were able to obtain accurate positions. For this purpose, we performed a <strong>de</strong>tailed<br />
astrometric reduction of each image using field stars present in the USNO-A2.0 cat-<br />
alog (Monet et al. 1999). First of all we selected point-like objects in the image, not<br />
elliptical or binary objects due to crow<strong>de</strong>d fields, and <strong>de</strong>termined their positions (X,<br />
Y ). We also rejected very faint objects to allow a significant signal-to-noise ratio.<br />
On the other hand, very bright objects were also rejected, because the positions<br />
given in the USNO-A2.0 catalog are from plates obtained around 1950, and nearby<br />
(bright) stars could have experienced a shift in position since then due to relatively<br />
high proper motions. Then we fitted an astrometric solution to our image using<br />
the USNO-A2.0 (α, δ) and image (X, Y ) positions for the common stars in both
5.4. Discussion 139<br />
the catalog and the image, and rejected spurious points above 3σ. We procee<strong>de</strong>d<br />
iteratively until convergence was achieved. This process allowed us to obtain plate<br />
solutions with an rms of ∼ 0.2 ′′ in each coordinate. This relatively high error prob-<br />
ably arises from the non-zero proper motions of the field stars finally used in the<br />
fit, between 36 and 155 <strong>de</strong>pending on each image. We have to quadratically add to<br />
this error the 1σ empirical uncertainty estimate of the USNO-A2.0 catalog relative<br />
to the ICRF, which is also ∼ 0.2 ′′ in each coordinate (see Table 1 of Deutsch 1999).<br />
Therefore, the final 1σ error of our obtained coordinates is estimated to be 0.3 ′′ .<br />
We show the central 2 ′ × 2 ′ of the CAHA 2.2 m images in Fig. 5.1, with the<br />
optical counterparts marked with a circle (which is not any X-ray or radio er-<br />
ror box). All sources appeared basically as point-like objects, except the quasar<br />
1RXS J072418.3−071508, which had a complex structure with two separate opti-<br />
cal components, hereafter NW (northwest) and SE (southeast). The NW position<br />
falls well within the NVSS error box and close to our obtained radio position. On<br />
the contrary, the SE position is clearly out the NVSS error box, and cannot be<br />
consi<strong>de</strong>red as a reliable counterpart to the radio source.<br />
Photometry was obtained thanks to the OAN 1.5 m observations. The absolute<br />
photometry is believed to be accurate only to ±0.1 magnitu<strong>de</strong> in the best cases, be-<br />
cause we could only <strong>de</strong>termine the approximate photometric zero point by observing<br />
a few standard stars from Landolt (1992).<br />
The obtained results are presented in Table 5.2, where Col. 1 gives the RBSC<br />
object name, Cols. 2 and 3 give the J2000.0 ICRS coordinates of the optical po-<br />
sition with the 1σ error of 0.3 ′′ below each coordinate, Col. 4 gives the observing<br />
dates and Cols. 5 to 7 give the Johnson V , R and I magnitu<strong>de</strong>s. Magnitu<strong>de</strong>s for<br />
1RXS J072418.3−071508 correspond to the sum of both the NW and SE compo-<br />
nents.<br />
5.4 Discussion<br />
The observations reported here provi<strong>de</strong> accurate positions in the optical (0.3 ′′ ) and<br />
specially in the radio (0.01 ′′ ) domains. The astrometric agreement for each source<br />
can be seen in the radio maps of Fig. 5.1, and is numerically expressed in Table 5.3,<br />
where we list the offsets in right ascension, in <strong>de</strong>clination and the total offset between
140 Chapter 5. Radio and optical observations<br />
Table 5.2: Optical astrometric and photometric results for all sources listed in Table 5.1.<br />
Accurate positions were obtained from the I Johnson filter CAHA 2.2 m telescope<br />
observations carried out on 1999 December 11. Magnitu<strong>de</strong>s in V , R and I Johnson<br />
filters were obtained in different dates of November 1998 using the OAN 1.5 m telescope.<br />
Note that two positions are given for the source 1RXS J072418.3−071508.<br />
1RXS name α (2000) δ (2000) Day Magnitu<strong>de</strong><br />
[h, m, s] [ ◦ , ′ , ′′ ] Nov. 1998 V R I<br />
J001442.2+580201 00 14 42.111 +58 02 01.32 25 - - > 19.0<br />
± 0.038 ± 0.30 27 - - 20.1 ± 0.4<br />
27 - - 19.7 ± 0.2<br />
J013106.4+612035 01 31 07.267 +61 20 33.60 27 19.5 ± 0.1 18.8 ± 0.1 17.9 ± 0.1<br />
± 0.042 ± 0.30<br />
J042201.0+485610 04 22 00.533 +48 56 03.84 28 20.4 ± 0.2 18.8 ± 0.1 17.4 ± 0.1<br />
± 0.030 ± 0.30 29 - - 17.5 ± 0.1<br />
J062148.1+174736 06 21 47.748 +17 47 35.23 28 19.8 ± 0.1 18.6 ± 0.1 17.5 ± 0.1<br />
± 0.021 ± 0.30 29 - - 17.6 ± 0.1<br />
J072259.5−073131 07 22 59.689 −07 31 34.78 28 18.5 ± 0.1 - -<br />
± 0.020 ± 0.30 29 18.3 ± 0.1 17.7 ± 0.1 16.8 ± 0.1<br />
J072418.3−071508(NW) 07 24 17.299 −07 15 20.46 29 18.8 ± 0.1 18.0 ± 0.1 17.2 ± 0.1<br />
± 0.020 ± 0.30<br />
(SE) 07 24 17.387 −07 15 22.33<br />
± 0.020 ± 0.30<br />
the optical and radio positions. As can be seen, the offsets in each coordinate are<br />
always smaller than the 1σ optical errors in position.<br />
Although the agreements in position are certainly very promising, we have esti-<br />
mated the probability of a random coinci<strong>de</strong>nce. For this purpose, we have obtained<br />
the limiting I magnitu<strong>de</strong> of our CAHA 2.2 m images and counted the number<br />
of objects until this threshold. Then we have estimated the object-<strong>de</strong>nsity of our<br />
images by assigning to each object the area of the 1σ error box in position, i.e.,<br />
(2 × 1σ) 2 = (0.6 ′′ ) 2 . Finally, we have divi<strong>de</strong>d the area occupied by the stars by<br />
the area of the image, and obtained a naive estimate of the probability to find an<br />
optical object, up to a given limiting I magnitu<strong>de</strong>, close to the radio position by<br />
less than 1σ, or 0.3 ′′ , in each coordinate. The limiting I magnitu<strong>de</strong>s and obtained<br />
probabilities of random coinci<strong>de</strong>nce are listed in the last two columns of Table 5.3.
5.4. Discussion 141<br />
Table 5.3: Offsets from the radio to the optical positions in right ascension, in <strong>de</strong>clination<br />
and the total offset. The limiting I magnitu<strong>de</strong> of our CAHA 2.2 m images and the<br />
probability of a random coinci<strong>de</strong>nce are listed in the last two columns.<br />
1RXS name αo − αr [ ′′ ] δo − δr [ ′′ ] o-r [ ′′ ] Limiting I mag. Probability [%]<br />
J001442.2+580201 −0.12 0.10 0.16 22 0.3<br />
J013106.4+612035 0.29 0.22 0.37 21 0.3<br />
J042201.0+485610 0.08 0.21 0.22 20 0.2<br />
J062148.1+174736 −0.06 0.15 0.16 20 0.2<br />
J072259.5−073131 0.12 0.02 0.12 19 0.2<br />
J072418.3−071508 0.12 −0.12 0.17 18 0.2<br />
Since these probabilities are always smaller than 1%, and even much smaller if we<br />
consi<strong>de</strong>r the particular offsets instead of the 1σ optical errors, we conclu<strong>de</strong> that<br />
probably all optical objects are the counterparts of the radio sources, and none of<br />
them is a field star not related to the radio source.<br />
On the other hand, an analysis of observed radio spectra of X-ray binaries can<br />
be found in Fen<strong>de</strong>r (2001). It seems clear that negative spectral indices (−1 ≤ α ≤<br />
−0.2) are <strong>de</strong>tected when observing synchrotron emission from expanding plasmons,<br />
which are the result of discrete ejections after an outburst. On the contrary, flat<br />
or inverted radio spectra (0.0 ≤ α ≤ 0.6) are typical of the low/hard X-ray state<br />
of black hole candidates, and are believed to arise in synchrotron emission from a<br />
partially self-absorbed jet. Hence, a variety of spectral indices can be found in the<br />
objects we are looking for. However, inverted spectra with α > 1 at high radio<br />
frequencies (above 5 GHz) would probably rule out a REXB nature.<br />
REXBs often display a variable flux, although extragalactic sources may vary as<br />
well. Our observations, performed at one/two week interval, allow us to estimate the<br />
<strong>de</strong>gree of variability given by (Smax − Smin)/(Smax + Smin), being Smax and Smin the<br />
maximum and the minimum flux <strong>de</strong>nsity, respectively. As it is clear from Table 5.1<br />
and Fig. 5.2, all sources in the sample show some <strong>de</strong>gree of flux <strong>de</strong>nsity variability at<br />
centimeter wavelengths during the three runs of our monitoring. This variability is<br />
always within 10%, except for 1RXS J042201.0+485610 at 6 cm, where a variability<br />
up to ∼ 30% is <strong>de</strong>tected, although it should be consi<strong>de</strong>red with caution due to the<br />
low emission level.
142 Chapter 5. Radio and optical observations<br />
The optical emission seen in microquasars arises from the non-<strong>de</strong>generated star<br />
of the system, i.e., the companion of the compact object. Hence, in microquasars<br />
containing a high mass companion spectral types O or B are found, while in those<br />
containing a low mass companion the spectral type is later than A (White et al.<br />
1995). Therefore, as optical counterpart of a given candidate, we expect to find a<br />
non-<strong>de</strong>generated star with any spectral type, and with a luminosity class ranging<br />
from the main sequence to supergiant. In fact, once photometric magnitu<strong>de</strong>s of an<br />
object through different filters and its distance are known, one can easily <strong>de</strong>duce<br />
its spectral type assuming it is a star. However, we have no information about the<br />
distance to our sources, and no correction for extinction can be applied to our data.<br />
This is particularly important in our case, since our search has been carried out at<br />
low galactic latitu<strong>de</strong>s, where a high <strong>de</strong>gree of extinction is expected, preventing any<br />
spectral type classification from the photometric data alone.<br />
5.4.1 Discussion on individual sources<br />
1RXS J001442.2+580201. The radio counterpart, see Table 5.1, shows a mod-<br />
erate <strong>de</strong>gree of variability but always displaying a negative spectral in<strong>de</strong>x of ∼ −0.2,<br />
suggesting an optically thin or partially self-absorbed synchrotron emitter. As can<br />
be seen in Fig. 5.2, the non-simultaneous NVSS flux <strong>de</strong>nsity measurement at 20 cm<br />
wavelength is compatible with this behavior or even with a flattening of the spec-<br />
trum at longer wavelengths, although this could be due to intrinsic variability. Non-<br />
thermal synchrotron radiation remains therefore as the most plausible interpretation<br />
for the radio emission of this source. The optical counterpart to the radio source<br />
(Fig. 5.1) appears as a point-like source with I ∼ 20, which makes it the weakest<br />
optical object of our sample. No R or V magnitu<strong>de</strong>s could be obtained, although<br />
a typical behavior for low latitu<strong>de</strong> highly absorbed sources is an increase of ∼ 1<br />
magnitu<strong>de</strong> when changing from I to R and from R to V , as can be seen in the other<br />
sources of Table 5.2. All the available information points towards a good REXB<br />
candidate.<br />
1RXS J013106.4+612035. In the radio domain it shows a slightly negative or<br />
close to zero spectral in<strong>de</strong>x, reminiscent of optically thin or partially self-absorbed<br />
jets in REXBs, in which the non-simultaneous NVSS flux measurement fits well, and<br />
it displays a very low <strong>de</strong>gree of variability of 2% at 3.6 cm and 4% at 6 cm. This
5.4. Discussion 143<br />
object has also been observed within the Green Bank 5 GHz radio survey (Gregory<br />
& Condon 1991; Gregory et al. 1996) with a measured flux <strong>de</strong>nsity of 24 mJy. This<br />
value is a little bit higher than the ones reported here, although this discrepancy<br />
could be due to intrinsic variability and/or to the different angular scales sampled<br />
by the interferometric VLA observations. In fact, VLA A configuration observations<br />
carried out by Laurent-Muehleisen et al. (1997) on October 19, 1992, indicate a flux<br />
<strong>de</strong>nsity of 16 mJy at 5 GHz, in good agreement with our data. In the optical domain,<br />
the counterpart to the radio source appears as a point-like object with a nearby field<br />
object at ∼ 3 ′′ to the northeast, visible in Fig. 5.1, which is ∼ 1 magnitu<strong>de</strong> fainter<br />
in the I band. The optical magnitu<strong>de</strong>s in Table 5.2 correspond to the sum of both<br />
objects (unresolved in the OAN 1.5 m images), and are typical of low latitu<strong>de</strong> highly<br />
absorbed sources. Overall it seems a good REXB candidate.<br />
1RXS J042201.0+485610. In the RBSC this source appears catalogued with<br />
an X-ray extension of 17 ′′ , which could suggest an extragalactic origin of the source.<br />
However, X-ray halos produced by the scattering of the original X-ray photons by<br />
interstellar dust grains are seen in some galactic microquasars, like for example in a<br />
Chandra observation of Cyg X-3 (Pre<strong>de</strong>hl et al. 2000). In fact, Cyg X-3 has an X-ray<br />
extension of 63 ′′ within the RBSC, while SS 433 has one of 25 ′′ . Hence, the presence<br />
of an X-ray extension does not rule out a REXB nature for this source. The radio<br />
counterpart shows the lowest radio emission level of all sources in our sample, and<br />
displays an inverted spectrum up to α ∼ +1.6 (see Table 5.1), difficult to account<br />
for self-absorbed synchrotron radiation at such high frequencies. Nevertheless, as<br />
it happens in some AGN, a highly inverted spectrum would be expected if the jet<br />
is stopped in a <strong>de</strong>nse medium on a compact scale. On the other hand, the non-<br />
simultaneous NVSS flux <strong>de</strong>nsity measurement seems to be incompatible with self-<br />
absorbed synchrotron radiation. Although this could be due to intrinsic variability<br />
of the source, this would require variations of an or<strong>de</strong>r of magnitu<strong>de</strong> in flux. This<br />
discrepancy can be explained if we assume a thermal origin for the radio emission and<br />
if we attribute the drop in flux <strong>de</strong>nsity in our VLA A configuration observations due<br />
to over resolution when compared to the NVSS VLA D configuration observations.<br />
The optical counterpart to the radio source has a nearby field object at ∼ 3 ′′ to<br />
the west, marginally visible in Fig. 5.1, which is ∼ 2 magnitu<strong>de</strong>s fainter in the I<br />
band. The optical magnitu<strong>de</strong>s in Table 5.2 correspond to the sum of both objects<br />
(unresolved in the OAN 1.5 m images), although the nearby one is not expected<br />
to contaminate appreciably, and are again typical of low latitu<strong>de</strong> highly absorbed
144 Chapter 5. Radio and optical observations<br />
sources. Although at first sight our target appears as a point-like source in Fig. 5.1,<br />
if we perform Gaussian fits to field objects with similar magnitu<strong>de</strong>s, we find our<br />
target to have a Full Width at Half Maximum (FWHM) ∼ 15% greater, suggesting<br />
that it is in fact an exten<strong>de</strong>d optical object. In summary, the high spectral in<strong>de</strong>x<br />
and the exten<strong>de</strong>d optical emission indicate that this source is not a very promising<br />
REXB candidate.<br />
1RXS J062148.1+174736. In the radio domain, this source displays an inverted<br />
or flat spectrum, which could account for a partially self-absorbed jet. On the<br />
other hand, the mo<strong>de</strong>rate variability observed, around 10%, could explain the higher<br />
non-simultaneous NVSS flux <strong>de</strong>nsity measurement. In the optical domain, this<br />
source appears as a point-like object with magnitu<strong>de</strong>s typical of low latitu<strong>de</strong> highly<br />
absorbed sources. However, a careful inspection of the optical image, reveals that<br />
the FWHM of a Gaussian fit to our target source is ∼ 30% greater than those of<br />
other sources with similar magnitu<strong>de</strong>s. This fact points to an extragalactic nature<br />
for this source. It is worth to mention that this source is the only one in our<br />
sample present in the second incremental release of the Two Micron All Sky Survey<br />
(2MASS) (Cutri et al. 2000). Its position, obtained from images taken in October<br />
1997, is there listed as α = 6 h 21 m 47.749 s ± 0.009 s , δ = +17 ◦ 47 ′ 35.07 ′′ ± 0.13 ′′ ,<br />
in very good agreement with our obtained coordinates. Finally, we would like to<br />
point out that Motch et al. (1998) list 3 stars within the X-ray error box as possible<br />
counterparts to the X-ray source. Our proposed optical counterpart on the basis of<br />
the radio position, would rule out any of these 3 stars as the optical counterpart of<br />
this X-ray source. Overall, it looks like an extragalactic object because of its optical<br />
extension.<br />
1RXS J072259.5−073131. The radio counterpart presents an exten<strong>de</strong>d struc-<br />
ture towards the east, marginally present in the 3.6 and 6 cm maps, and visible as<br />
a one-si<strong>de</strong>d jet in the 30 July 4 minute snapshot observation at 20 cm wavelength<br />
(see Fig. 5.3). Although the one-si<strong>de</strong>d morphology is suggestive of an extragalac-<br />
tic nature for this source, specially if we take into account that we are mapping the<br />
source at arcsecond scales, we cannot rule out a possible galactic nature on the basis<br />
of the <strong>de</strong>tected morphology. The source also shows a mo<strong>de</strong>rate <strong>de</strong>gree of variability,<br />
displaying most of time a negative spectral in<strong>de</strong>x but also close to 0 in the July 22<br />
observation. As can be seen in Fig. 5.2, the non-simultaneous NVSS flux <strong>de</strong>nsity
5.5. Conclusions 145<br />
measurement at 20 cm wavelength could be compatible with a non-thermal optically<br />
thin spectrum. However, simultaneous observations during July 30 show a flattening<br />
of the spectrum at higher wavelengths, which could be due to self-absorption in this<br />
energy range. The difference between the NVSS and our flux <strong>de</strong>nsity measurement<br />
at 20 cm wavelength could be due to intrinsic variability and/or to the different<br />
VLA configurations used. Anyway, non-thermal synchrotron radiation remains as<br />
the most plausible interpretation for the radio emission of this source, specially at<br />
higher frequencies. The optical counterpart appears as a point-like object with the<br />
magnitu<strong>de</strong>s of an absorbed object at low galactic latitu<strong>de</strong>s. These properties are in<br />
good agreement with the expected ones for a microquasar, although the one-si<strong>de</strong>d<br />
radio jet at arcsecond scales is most common in extragalactic objects.<br />
1RXS J072418.3−071508. This source has been recently (March 2002) classi-<br />
fied as a quasar in the SIMBAD database, and it is not any more a REXB can-<br />
didate. It is listed as PMN J0724−0715 in the NED database, and is the source<br />
WGA J0724.3−0715 in Perlman et al. (1998), who reported a faint and quite broad<br />
Hα emission line (rest-frame Wλ = 30.3 ˚A, FWHM=4000 km s −1 ), and classified it<br />
as a flat spectrum radio quasar with z = 0.270. Nevertheless, we have reported here<br />
our observational results for this source, since it was a candidate when we performed<br />
the observations. In the radio it displays a low <strong>de</strong>gree of variability of 3% at 3.6 cm<br />
and 4% at 6 cm, with a flat or even inverted spectrum. In the optical it appears as<br />
a complex double object, with a NW and a SE components, being the first one the<br />
counterpart to the radio source. Magnitu<strong>de</strong>s in V , R and I typical of absorbed objects<br />
at low galactic latitu<strong>de</strong>s were found. On the other hand, it appears catalogued<br />
with an X-ray extension of 40 ′′ in the RBSC.<br />
5.5 Conclusions<br />
After the cross-i<strong>de</strong>ntification of catalogs presented in Chapt. 4, we found 8 interesting<br />
sources to be studied (Group 1 sources). Here we have reported observations<br />
of all these sources except 1RXS J181119.4−275939 and 1RXS J185002.8−075833,<br />
that were not visible during our observations. We have studied the remaining 6<br />
radio sources of Group 1, and found optical counterparts to all of them. We have<br />
obtained accurate positions at both radio and optical wavelengths, perfectly com-
146 Chapter 5. Radio and optical observations<br />
patible between them. We also have obtained radio spectra and optical magnitu<strong>de</strong>s<br />
of the sources.<br />
After a <strong>de</strong>tailed analysis of the obtained data, we conclu<strong>de</strong> that the sources<br />
1RXS J001442.2+580201 and 1RXS J013106.4+612035 are good REXB candidates,<br />
while 1RXS J042201.0+485610 is not so promising due to its highly inverted spec-<br />
trum at high frequencies and its optical exten<strong>de</strong>d emission. A careful study of the<br />
optical counterpart of 1RXS J062148.1+174736 reveals it is exten<strong>de</strong>d, which points<br />
towards an extragalactic nature for this object. The situation is not clear in the case<br />
of 1RXS J072259.5−073131, because the optical data agrees with a microquasar na-<br />
ture, while the <strong>de</strong>tected one-si<strong>de</strong>d radio jet reaching arcsecond scales is most common<br />
in extragalactic sources. The last studied source, 1RXS J072418.3−071508, is an<br />
already i<strong>de</strong>ntified quasar. Hence, two sources in our sample are very promising to<br />
be new microquasars in the Galaxy.<br />
The next logical step was to obtain optical spectroscopy of these sources, to<br />
unveil their galactic or extragalactic nature. Several attempts have been carried out<br />
in years 2000 and 2001, without success due to bad weather conditions, and a new<br />
attempt is in progress for autumn 2002. On the other hand, VLBI observations<br />
of the 6 sources studied here will be published soon (Ribó et al. 2002), and are<br />
reported in the next chapter.
Bibliography<br />
Cutri, R. M., Skrutskie, M. F., Van Dyk, S., et al. 2000, Explanatory Supple-<br />
ment to the 2MASS, Second Incremental Data Release, Caltech (available at<br />
http://www.ipac.caltech.edu/2mass/releases/second/doc/explsup.html)<br />
Deutsch, E. W. 1999, AJ, 118, 1882<br />
Fen<strong>de</strong>r, R. P. 2001, MNRAS, 322, 31<br />
Gregory, P. C., & Condon, J. J. 1991, ApJS, 75, 1011<br />
Gregory, P. C., Scott, W. K., Douglas, K., & Condon, J. J. 1996, ApJS, 103, 427<br />
Landolt, A.U. 1992, AJ, 104, 340<br />
Laurent-Muehleisen, S. A., Kollgaard, R. I., Ryan, P. J., et al. 1997, A&AS, 122,<br />
235<br />
Monet, D. G., Bird, A., Canzian, B., et al. 1999, USNO-A2.0 CD-ROM (U.S. Naval<br />
Observatory, Washington DC)<br />
Motch, C., Guillout, P., Haberl, F., et al. 1998, A&AS, 132, 341<br />
Perlman, E. S., Padovani, P., Giommi, P., et al. 1998, AJ, 115, 1253<br />
Pre<strong>de</strong>hl, P., Burwitz, V., Paerels, F., & Trümper, J. 2000, A&A, 357, L25<br />
Ribó, M., Ros, E., Pare<strong>de</strong>s, J. M., Massi, M., & Martí, J. 2002, A&A, 394, 983<br />
White, N. E., Nagase, F., & Parmar, A. N. 1995, The properties of X-ray binaries,<br />
in X-ray Binaries, ed. W. H. G. Lewin, J. van Paradijs, & E. P. J. van <strong>de</strong>n Heuvel<br />
(Cambridge Univ. Press, Cambridge), 1<br />
147
148 BIBLIOGRAPHY
Chapter 6<br />
EVN and MERLIN observations<br />
6.1 Introduction<br />
A sample of six radio emitting X-ray sources, among the obtained after the cross-<br />
i<strong>de</strong>ntification of catalogs reported in Chapt. 4, has been further studied in Chapt. 5<br />
(see Pare<strong>de</strong>s et al. 2002). Here we present VLBI observations 1 of these six sources,<br />
aimed at revealing possible jet-like features at milliarcsecond scales. We <strong>de</strong>scribe the<br />
observations and the data reduction in Sect. 6.2, present the results and a discussion<br />
in Sect. 6.3, and state our conclusions in Sect. 6.4.<br />
6.2 Observations and data reduction<br />
We observed the six sources studied in Chapt. 5 simultaneously with the Multi-<br />
Element Radio-Linked Interferometer Network (MERLIN) 2 and the European VLBI<br />
Network (EVN) 3 on 2000 February 29th/March 1st (23:30–23:05 UT) at 5 GHz.<br />
Single dish flux <strong>de</strong>nsity measurements were carried out with the MPIfR 100 m<br />
antenna in Effelsberg, Germany.<br />
1Published in Ribó, M., Ros, E., Pare<strong>de</strong>s, J. M., Massi, M., & Martí, J. 2002, A&A, 394, 983.<br />
2MERLIN is operated as a National Facility by the University of Manchester at Jodrell Bank<br />
Observatory on behalf of the UK Particle Physics & Astronomy Research Council.<br />
3The European VLBI Network is a joint facility of European, Chinese and other radio astronomy<br />
institutes fun<strong>de</strong>d by their national research councils.<br />
149
150 Chapter 6. EVN and MERLIN observations<br />
Table 6.1: Observing scheme for target sources and calibrators (EVN and MERLIN).<br />
1XRS name Phase-reference calibrators Cycling scheme (a) Obs. Time<br />
[s] [s]<br />
J001442.2+580201 J0007+5706 (b) 300/150 7200<br />
J013106.4+612035 J0147+5840 2×(300/150)/300/120 8000<br />
4C 61.02 (c)<br />
J042201.0+485610 (c) TXS 0422+496 300/150/300/120/300/150 7500<br />
NRAO 150<br />
J062148.1+174736 J0630+1738 (b) 300/150 6200<br />
J072259.5−073131 1RXS J072418.3−071508 2×(270/180)/240/180 6100<br />
J072418.3−071508 1RXS J072259.5−073131 2×(180/270)/180/240 4700<br />
(a) target/calibrator/2nd calibrator.<br />
(b) Used in the VLBI fringe fitting process to gui<strong>de</strong> the target source fringe <strong>de</strong>tection (see text).<br />
(c) No fringes found to the radio source.<br />
6.2.1 MERLIN<br />
MERLIN is a connected radio interferometer across England, with baselines reaching<br />
up to 217 km length. This array observed with 2-bit sampling at dual polarization<br />
with two blocks of 16 channels, each channel of 1 MHz bandwidth. We analyzed<br />
the left hand circular polarization data excluding one channel at both edges of the<br />
band, yielding a final bandwidth of 14 MHz. The correlator integration time was of<br />
4 s. Some antennas did not observe during short periods due to strong winds.<br />
Since some of the target radio sources were faint, we scheduled the observations<br />
introducing phase-reference calibrators with cycle times of around 7 min (compatible<br />
with the expected coherence times). The observing scheme is <strong>de</strong>scribed in Table 6.1,<br />
where we quote the total on-source time for the target sources in the last column.<br />
We also observed the fringe-fin<strong>de</strong>r DA 193 and the MERLIN flux <strong>de</strong>nsity calibrator<br />
3C 286.<br />
The MERLIN data reduction was carried out at Jodrell Bank Observatory,<br />
using standard procedures within the aips software package. We did not <strong>de</strong>tect<br />
1RXS J042201.0+485610. All other sources were <strong>de</strong>tected, and accurate positions<br />
for the target radio sources were obtained via phase-referencing. These positions,<br />
presented in Table 6.2, were used later as a priori information for the VLBI cor-
6.2. Observations and data reduction 151<br />
Table 6.2: MERLIN positions for the five <strong>de</strong>tected target sources, obtained via phase-<br />
referencing. The uncertainties quoted for the target sources were provi<strong>de</strong>d by the jmfit<br />
task in aips (measuring in the phase-referenced image) and do not inclu<strong>de</strong> systematic<br />
errors. The positions of all phase-reference calibrators except 1RXS J072418.3−071508<br />
have an accuracy better than 1 mas.<br />
Target sources Phase-reference calibrators<br />
1RXS name α (J2000.0) δ (J2000.0) Source name α (J2000.0) δ (J2000.0)<br />
J001442.2+580201 00 h 14 m 42. s 12822 +58 ◦ 02 ′ 01. ′′ 2460 J0007+5706 (a) 00 h 07 m 48. s 47110 +57 ◦ 06 ′ 10. ′′ 4540<br />
±0. s 00013 ±0. ′′ 0022<br />
J013106.4+612035 01 h 31 m 07. s 23210 +61 ◦ 20 ′ 33. ′′ 3752 J0147+5840 (a) 01 h 47 m 46. s 54380 +58 ◦ 40 ′ 44. ′′ 9750<br />
±0. s 00014 ±0. ′′ 0016<br />
J062148.1+174736 06 h 21 m 47. s 75264 +17 ◦ 47 ′ 35. ′′ 0818 J0630+1738 (b) 06 h 30 m 07. s 25870 +17 ◦ 38 ′ 12. ′′ 9300<br />
±0. s 00006 ±0. ′′ 0017<br />
J072259.5−073131 07 h 22 m 59. s 68188 −07 ◦ 31 ′ 34. ′′ 8009 1RXS J072418.3 (c) 07 h 24 m 17. s 2912 −07 ◦ 15 ′ 20. ′′ 339<br />
±0. s 00011 ±0. ′′ 0022 −071508<br />
J072418.3−071508 (d) 07 h 24 m 17. s 2912 −07 ◦ 15 ′ 20. ′′ 339 J0730−116 07 h 30 m 19. s 1125 −11 ◦ 41 ′ 12. ′′ 601<br />
±0. s 0007 ±0. ′′ 010<br />
(a) Position from Patnaik et al. (1992).<br />
(b) Position from Browne et al. (1998).<br />
(c) This position was obtained from VLA observations (listed in the line below).<br />
(d) The entries in this line correspond to VLA observations reported in Chapt. 5 (Table 5.1).<br />
relation. The position given in Table 6.2 for the quasar 1RXS J072418.3−071508<br />
was obtained from the VLA observations presented in Sect. 5.2. The position of<br />
1RXS J072259.5−073131 is <strong>de</strong>duced from the phase-reference offset relative to the<br />
source 1RXS J072418.3−071508 provi<strong>de</strong>d by MERLIN.<br />
To image the sources, we averaged the data in frequency and exported them to<br />
be processed into the difference mapping software difmap (Shepherd et al. 1994),<br />
where we time-averaged the data in 32 s bins after careful editing.<br />
6.2.2 EVN<br />
The EVN observations were performed with the array <strong>de</strong>scribed in Table 6.3, record-<br />
ing in MkIV mo<strong>de</strong> with 2-bit sampling at 256 Mbps with left hand circular polar-<br />
ization. A bandwidth of 64 MHz was used, divi<strong>de</strong>d into 8 intermediate frequency<br />
(IF) bands (v6cm-256-8-2-L mo<strong>de</strong>). Some antennas did not observe during short<br />
periods due to strong winds or snow.
152 Chapter 6. EVN and MERLIN observations<br />
Table 6.3: EVN array used in the observations.<br />
Antenna Co<strong>de</strong> Location Diameter DPFU (a) Tsys (b)<br />
[m] [K Jy −1 ] [K]<br />
Effelsberg EB Germany 100 1.47 27<br />
Jodrell Bank JB U.K. 25 0.11 40<br />
Cambridge CM U.K. 32 0.21 38<br />
Westerbork WB Netherlands 14×25 1.0 67<br />
Medicina MC Italy 32 0.15 53<br />
Noto NT Italy 32 0.16 47<br />
Shanghai SH China 25 0.088 40<br />
Toruń TR Poland 32 0.15 31<br />
Onsala85 ON Swe<strong>de</strong>n 25 0.043 45<br />
(a) Degrees per flux unit. (b) Best value during the experiment.<br />
The data were processed at the EVN MkIV correlator at the Joint Institute for<br />
VLBI in Europe (JIVE), in Dwingeloo, The Netherlands. The correlator integration<br />
time was of 4 s. A first post-processing analysis was also carried out at JIVE. The<br />
data were processed using aips. A first a priori visibility amplitu<strong>de</strong> calibration was<br />
performed using antenna gains and system temperatures measured at each antenna.<br />
The fringe fitting (fring) of the residual <strong>de</strong>lays and fringe rates was performed<br />
for all the radio sources. No fringes were found for 1RXS J042201.0+485610 and<br />
4C 61.02. Fringes for many baselines were missing for 1RXS J001442.2+580201,<br />
1RXS J062148.1+174736 and TXS 0422+496.<br />
To improve the fringe <strong>de</strong>tection on all baselines for 1RXS J001442.2+580201<br />
and 1RXS J062148.1+174736 we used the <strong>de</strong>lay, rate, and phase solutions from<br />
their corresponding phase-reference calibrators (J0007+5706, 1 ◦ 18 ′ separation, and<br />
J0630+1738, 1 ◦ 59 ′ separation, see Table 6.1) and interpolated them to the target<br />
sources using the aips task clcal. We fringe-fitted the target sources again using<br />
narrower search windows and obtained solutions for all baselines. A similar attempt<br />
on 1RXS J042201.0+485610 (with respect to TXS 0422+486 and NRAO 150) was<br />
unfruitful. Effelsberg was used as reference antenna throughout the aips data re-<br />
duction process.<br />
We then averaged the data in frequency and exported them to be imaged and<br />
self-calibrated in difmap. The a priori visibility amplitu<strong>de</strong> calibration was not suffi-
6.2. Observations and data reduction 153<br />
cient to reliably image the weakest radio sources. We improved that by first imaging<br />
in difmap the calibrator sources J0007+5706, J0147+5840, and J0630+1738, with<br />
appropriate amplitu<strong>de</strong> self-calibration. We <strong>de</strong>duced correction factors for each an-<br />
tenna, these being consistent for the three radio sources within 2%. The factors<br />
were: EB:0.98, JB:1.10, CM:0.78, WB:1.18, MC:0.87, NT:1.12, SH:1.03, TR:0.98<br />
and ON:0.97. We corrected the amplitu<strong>de</strong> calibration back in aips using the task<br />
sncor (optype ’mula’) for all the radio sources and exported the data again into<br />
difmap, where the final imaging was performed after editing and averaging of the<br />
visibilities in 32 s blocks.<br />
6.2.3 Combining EVN and MERLIN<br />
The EVN and MERLIN arrays have one common baseline, between JB and CM,<br />
which allows to combine both data sets and map them together. Combining data<br />
sets from both arrays allows us to reach (u, v) resolution ranges from 0.04 Mλ (MK2-<br />
Tabley) to 140 Mλ (NT-SH) at 5 GHz. We processed the data within aips in or<strong>de</strong>r<br />
to combine both arrays. The B1950.0 (u, v) coordinates of the MERLIN data had<br />
to be corrected to the ones of the EVN for the same reference system (J2000.0)<br />
with uvfix. Then, the MERLIN data were self-calibrated with the EVN images<br />
(see below), and the phase solutions were limited to the longest MERLIN baselines.<br />
The MERLIN data were imaged and the peak-of-brightness of both data sets were<br />
checked to be similar. As a next step, the EVN data were averaged in frequency<br />
to correspond to the MERLIN data. The aips hea<strong>de</strong>rs of both data sets were<br />
modified conveniently to match together, and the weighting of both data sets was<br />
also modified to be equal with wtmod. Finally, both data sets were concatenated<br />
(using dbcon) and exported to difmap to be imaged with different data weighting<br />
in (u, v) distance (tapering) after time averaging in 32 s bins.<br />
6.2.4 Flux <strong>de</strong>nsity measurements at the 100 m antenna in<br />
Effelsberg<br />
To complement the amplitu<strong>de</strong> calibration and obtain additional information on the<br />
radio sources, we interleaved cross-scans (in azimuth and elevation) with the 100 m<br />
Effelsberg antenna to measure the radio source flux <strong>de</strong>nsities (A. Kraus, private
154 Chapter 6. EVN and MERLIN observations<br />
Table 6.4: Flux <strong>de</strong>nsities and parameters used to produce the images in Figs. 6.1–6.5.<br />
Single dish Array (a)<br />
1RXS name SEB (taper FWHM) beam size P.A. Stot Speak Smin<br />
[mJy] [Mλ] [mas] × [mas] [ ◦ ] [mJy] [mJy b. −1 ] [mJy b. −1 ]<br />
J001442.2+580201 6.5 ± 0.5 (b) M 57 × 51 37 6.2 5.8 0.1<br />
E+M (10) 7.7 × 7.2 −4 10.1 9.9 0.18<br />
E 1.77 × 0.86 −20 11.5 7.0 0.15<br />
J013106.4+612035 20.1 ± 0.6 M 71 × 39 −63 17.5 17.9 0.5<br />
E+M (15) 7.8 × 6.2 −75 19.2 18.7 0.8<br />
E 1.04 × 0.99 −1 17.6 11.6 0.4<br />
J042201.0+485610 < 5 — — — — — —<br />
J062148.1+174736 7.1 ± 0.7 (b) M 88 × 39 24 5.8 5.6 0.2<br />
E+M (8) 13.1 × 11.1 −24 6.0 6.8 0.3<br />
E 8.8 × 4.1 47 7.0 6.4 0.3<br />
J072259.5−073131 67.9 ± 1.1 M 115 × 66 4 66.0 62.9 0.7<br />
E+M (15) 9.5 × 7.3 −61 52.1 41.9 0.9<br />
E 5.74 × 1.12 11 46.9 36.2 0.9<br />
J072418.3−071508 282.2 ± 4.1 M 120 × 64 8 301.0 285.7 0.9<br />
E+M (10) 11.7 × 9.8 −61 287.0 263.4 2.0<br />
E 5.73 × 1.02 11 248.0 184.8 0.7<br />
(a) M: MERLIN. E+M: EVN+MERLIN, FWHM of the tapering function (weighting of visibilities)<br />
in parenthesis. E: EVN.<br />
(b) Values with low SNR in the Gaussian fits.<br />
communication). We fitted a Gaussian function to the flux-<strong>de</strong>nsity response for<br />
every cross-scan, and we averaged the different Gaussians. We linked the flux <strong>de</strong>nsity<br />
scale by observing primary calibrators such as 3C 286, 3C 48, or NGC 7027 (see e.g.<br />
Kraus 1997; Peng et al. 2000). We list the single dish flux <strong>de</strong>nsity measurements<br />
in the second column of Table 6.4. The flux <strong>de</strong>nsity values for the main VLBI<br />
calibrators were of 7.5 ± 0.1 Jy for 3C 286, and 5.9 ± 0.2 Jy for DA 193.<br />
6.3 Results and discussion<br />
We present all the imaging results in Figs. 6.1–6.5, and the image parameters in<br />
Table 6.4. The total flux <strong>de</strong>nsity values for the different images diverge from each
6.3. Results and discussion 155<br />
Figure 6.1: Images of 1RXS J001442.2+580201 using the different arrays and with the<br />
parameters given in Table 6.4. The axes units are in mas. A (u, v)-tapering with a<br />
FWHM at 10 Mλ has been used to perform the combined EVN+MERLIN image. In the<br />
EVN image, N1, N2, S2 and S1 indicate the components discussed in the text.<br />
other and from the single dish measurements, due to the amplitu<strong>de</strong> self-calibration<br />
process in all cases. Therefore, those values should be consi<strong>de</strong>red with care. The<br />
minimum contours in the images are those listed as Smin in Table 6.4, while con-<br />
secutive higher contours scale with 3 1/2 . Here follows a <strong>de</strong>tailed discussion on each<br />
source.<br />
6.3.1 1RXS J001442.2+580201 and its two-si<strong>de</strong>d jet<br />
As can be seen in our images, shown in Fig. 6.1, this source appears point-like<br />
at MERLIN resolution, partially resolved in the tapered EVN+MERLIN image<br />
and clearly resolved at EVN scales. In this last case, it shows a two-si<strong>de</strong>d jet-like<br />
structure roughly in the north-south direction, with brighter components towards<br />
the south. The trends are visible in the closure phases, giving us confi<strong>de</strong>nce that<br />
the structure observed is not a consequence of si<strong>de</strong>lobes or imaging artifacts. In<br />
the tapered EVN+MERLIN images, the structure extends up to 20–30 mas outsi<strong>de</strong><br />
of the core, and more clearly towards the north. This discrepancy could be due to<br />
calibration problems.<br />
Mo<strong>de</strong>l fitting of the EVN visibilities with circular Gaussians provi<strong>de</strong>s a param-<br />
eterization of the inner structure. Five components reproduce the visibilities. The<br />
central one has 7.8 mJy, with a FWHM of 0.4 mas. Towards the north, one compo-<br />
nent (N2) of 0.6 mJy at 3.6 mas (P.A. −10 ◦ , FWHM 0.7 mas) and another one (N1)<br />
of 0.5 mJy at 8.6 mas (P.A. 2 ◦ , exten<strong>de</strong>d over 3 mas) are nee<strong>de</strong>d. Brighter compo-<br />
N1<br />
N2<br />
S2<br />
S1
156 Chapter 6. EVN and MERLIN observations<br />
Table 6.5: Mo<strong>de</strong>l fitted positions of the components in the 1RXS J001442.2+580201<br />
radio jets and obtained jet parameters.<br />
Comp. Distance P.A. βmin θmax<br />
[mas] [ ◦ ] [ ◦ ]<br />
N1 8.6 ± 0.3 2 0.23 ± 0.02 77 ± 1<br />
N2 3.6 ± 0.2 −10 0.16 ± 0.02 81 ± 1<br />
S2 5.0 ± 0.1 177 0.16 ± 0.02 81 ± 1<br />
S1 13.7 ± 0.3 177 0.23 ± 0.02 77 ± 1<br />
nents are present southwards, one (S2) of 1.1 mJy at 5.0 mas (P.A. 177 ◦ , FWHM of<br />
0.8 mas) and the other one (S1) of 0.4 mJy at 13.7 mas (P.A. 177 ◦ , FWHM below<br />
0.3 mas).<br />
If we assume that components S1 and N1 correspond to a pair of plasma clouds<br />
ejected at the same epoch near the compact object and perpendicularly to the<br />
accretion disk, we can estimate some parameters of the jets by using the following<br />
equation:<br />
β cos θ = µa − µr<br />
µa + µr<br />
= da − dr<br />
da + dr<br />
, (6.1)<br />
β being the velocity of the clouds in units of the speed of light, θ the angle between<br />
the direction of motion of the ejecta and the line of sight and µa and µr the proper<br />
motions of the approaching and receding components, respectively (see Sect. 1.2.1).<br />
Although we do not know the epoch of ejection of the clouds, we can cancel the<br />
time variable by using the relative distances to the core da and dr, as expressed in<br />
Eq. 6.1. Since both variables, β and cos θ, take values between 0 and 1, it is clear<br />
that knowing β cos θ allows us to compute a lower limit for the velocity (βmin) and<br />
an upper limit for the angle (θmax). The same applies for the S2 and N2 components.<br />
In Table 6.5 we list the positions of the components obtained from mo<strong>de</strong>l fitting,<br />
together with the <strong>de</strong>rived values from βmin and θmax for each one of the pairs. The<br />
slightly different results obtained using pair 1 or 2, could be due to the fact that<br />
the position for the S1 component obtained with mo<strong>de</strong>l fitting happens to be at<br />
the lower part of this elongated component, hence increasing β cos θ, or to intrinsic<br />
different velocities for each one of the pairs. Hereafter we will use β > 0.20 ± 0.02<br />
and θ < 78 ± 1 ◦ .<br />
A similar approach to obtain the jet parameters of the source can be performed
6.3. Results and discussion 157<br />
thanks to the brightness asymmetry of the components using the following equation<br />
(see Sect. 1.2.1):<br />
β cos θ = (Sa/Sr) 1/(k−α) − 1<br />
(Sa/Sr) 1/(k−α) + 1<br />
, (6.2)<br />
where Sa and Sr are the flux <strong>de</strong>nsities of the approaching and receding components,<br />
respectively, k equals 2 for a continuous jet and 3 for discrete con<strong>de</strong>nsations, and α<br />
is the spectral in<strong>de</strong>x of the emission (Sν ∝ ν +α ). However, the equation above is<br />
only valid when the components are at the same distance from the core. If this is<br />
not the case (i.e., dr < da), the ratio Sa/Sr will be lower than the one that should<br />
be used in Eq. 6.2 (because the flux <strong>de</strong>nsity <strong>de</strong>creases with increasing distance from<br />
the core). In consequence, Eq. 6.2 only allows us to obtain a lower limit for β cos θ.<br />
In or<strong>de</strong>r to use this approach we will consi<strong>de</strong>r k = 3, because the components<br />
seem to be discrete con<strong>de</strong>nsations, and α = −0.20 ± 0.05, according to the overall<br />
spectral in<strong>de</strong>x reported in Table 5.1, since we do not have spectral in<strong>de</strong>x information<br />
of the components. The use of the flux <strong>de</strong>nsities obtained after mo<strong>de</strong>l fitting gives<br />
β cos θ > 0.09 ± 0.04 for the S2–N2 pair and β cos θ > −0.07 ± 0.08 for the S1–N1<br />
pair. This last value is certainly surprising, although it can be explained by the<br />
fact that the S1 flux <strong>de</strong>nsity obtained after mo<strong>de</strong>l fitting does not account for the<br />
total flux of this plasma cloud. In fact, better estimates of the flux <strong>de</strong>nsities can be<br />
obtained summing together the flux <strong>de</strong>nsities of the clean components obtained<br />
within each one of the four plasma clouds. This yields to β cos θ > 0.13 ± 0.05 for<br />
the S2–N2 pair and β cos θ > 0.17±0.02 for the S1–N1 pair, in good agreement with<br />
the values computed using the distances from the components to the core, shown in<br />
Table 6.5.<br />
If we compare the VLA position reported in Table 5.1 with the MERLIN position<br />
in Table 6.2 we can see that they are different. In fact, taking into account the errors,<br />
the MERLIN−VLA position offsets can be expressed as: ∆α cos δ = 16 ± 10 mas<br />
and ∆δ = 27 ± 10 mas. Assuming that the difference in position is due to intrinsic<br />
proper motions and consi<strong>de</strong>ring the time span between both observations, 224 days,<br />
we obtain: µα cos δ = 26 ± 16 mas yr −1 and µδ = 44 ± 16 mas yr −1 . Although a<br />
proper motion of the source would clearly indicate its microquasar nature, because<br />
no proper motion would be <strong>de</strong>tected in an extragalactic source, we must be cautious<br />
with this result. First of all, the phase-reference sources where different in the<br />
VLA and MERLIN observations, as well as the observing frequencies from which<br />
positions were estimated. On the other hand, none of these results exceeds the 3σ
158 Chapter 6. EVN and MERLIN observations<br />
Figure 6.2: Images of 1RXS J013106.4+612035 using the different arrays and with the<br />
parameters given in Table 6.4. The axes units are in mas. A (u, v)-tapering with a<br />
FWHM at 15 Mλ has been used to perform the combined EVN+MERLIN image.<br />
value, preventing to state that a proper motion has been <strong>de</strong>tected.<br />
Summarizing, our results indicate that this source exhibits relativistic radio jets<br />
with β > 0.20 and, therefore, together with the results reported in Chapt. 5, we<br />
consi<strong>de</strong>r 1RXS J001442.2+580201 as a very promising microquasar candidate.<br />
6.3.2 1RXS J013106.4+612035 and its one-si<strong>de</strong>d jet<br />
We show in Fig. 6.2 the images obtained after our observations. Although the source<br />
appears compact at MERLIN and EVN+MERLIN scales, there is a weak one-si<strong>de</strong>d<br />
radio jet towards the northwest in the EVN image. Mo<strong>de</strong>l fitting with circular<br />
Gaussian components can reproduce the observed visibilities as follows: a central<br />
15.4 mJy component with 0.64 mas FWHM, and a northwest 2.1 mJy component<br />
with a FWHM of 0.74 mas, located at 1.8 mas in P.A. −73 ◦ . As can be seen, this<br />
last component is located at a distance of ∼ 2 times the beam size from the core.<br />
Using the fact that we do not <strong>de</strong>tect a counter-jet, we can use Eq. 6.2 replacing<br />
Sr with the 3σ level value. This, of course, will only provi<strong>de</strong> a lower limit to β cos θ,<br />
expressed as follows:<br />
β cos θ > (Sa/3σ) 1/(k−α) − 1<br />
(Sa/3σ) 1/(k−α) + 1<br />
. (6.3)<br />
Using Sa = 2.1 mJy, 3σ = 0.30 mJy (the 1σ value has been taken as the root mean<br />
square noise in the image), α = −0.05 ± 0.05 (see Table 5.1), and k = 3 to be<br />
consistent with the lowest limit, we obtain β cos θ > 0.31 ± 0.05 (β > 0.31 ± 0.05
6.3. Results and discussion 159<br />
and θ < 72 ± 3). Hence, a lower limit of β ≥ 0.3 is obtained, pointing towards<br />
relativistic radio jets as the origin of the elongated radio emission present in the<br />
EVN image. Although the one-si<strong>de</strong>d jet morphology at mas scales is found mostly<br />
in extragalactic sources, it is also present in some galactic REXBs, like Cyg X-3<br />
(Mioduszewski et al. 2001) or LS I +61 303 (Massi et al. 2001). Hence, we cannot<br />
rule out a possible galactic nature on the basis of the <strong>de</strong>tected morphology.<br />
As done for the previous source, we can compare the VLA position with the<br />
MERLIN one, and find that they differ in ∆α cos δ = 39 ± 10 mas and ∆δ =<br />
−0.8 ± 10 mas. If the offsets are real, this would imply µα cos δ = 64 ± 16 mas yr −1<br />
and µδ = −1 ± 16 mas yr −1 . Hence, it seems possible that we have <strong>de</strong>tected a<br />
proper motion in right ascension at a 4σ level. However, we must be cautious since,<br />
as in the previous case, the phase-reference sources and observing frequencies were<br />
different in each observation.<br />
Overall, these results are indicative of relativistic radio jets, and hence, together<br />
with the results reported in Chapt. 5, allow us to classify this source as a promising<br />
microquasar candidate.<br />
6.3.3 1RXS J042201.0+485610, a non-<strong>de</strong>tected source<br />
This radio source was marginally seen in the cross-scans carried out in Effelsberg. In<br />
fact, this is compatible with the low flux <strong>de</strong>nsity of 2.3±0.4 mJy at 1.4 GHz listed in<br />
the NVSS. The source was not <strong>de</strong>tected with MERLIN or with the EVN, mainly due<br />
to problems with the phase-reference sources, as pointed out in Sect. 6.2.2. In fact,<br />
as discussed in Chapt. 5, VLA A configuration observations showed a flux <strong>de</strong>nsity of<br />
0.4 mJy at 5 GHz and an inverted spectrum with a spectral in<strong>de</strong>x up to α = +1.6,<br />
suggesting thermal radio-emission resolved at higher angular resolutions. However,<br />
we cannot exclu<strong>de</strong> the possibility of having a highly variable source.<br />
6.3.4 1RXS J062148.1+174736, a compact source<br />
As can be seen in our images, shown in Fig. 6.3, the radio source is compact on all<br />
scales. In fact, mo<strong>de</strong>l fitting of the EVN visibilities converges to a point-like radio<br />
source (the FWHM of a circular Gaussian tends to zero). We must note that when
160 Chapter 6. EVN and MERLIN observations<br />
Figure 6.3: Images of 1RXS J062148.1+174736 using the different arrays and with the<br />
parameters given in Table 6.4. The axes units are in mas. A (u, v)-tapering with a<br />
FWHM at 8 Mλ has been used to perform the combined EVN+MERLIN image.<br />
this source was observed it was below the horizon in SH. Therefore, the obtained<br />
beam size for the EVN image is larger than the ones obtained for the other sources,<br />
as can be seen in Table 6.4, hence providing lower angular resolution than in the<br />
other cases. A comparison between the VLA and MERLIN positions reveals that<br />
they are perfectly compatible within the errors, suggesting an extragalactic nature<br />
or a small proper motion if it turns out to be a galactic microquasar. Although the<br />
compactness of the source is not indicative of a galactic or an extragalactic origin,<br />
the exten<strong>de</strong>d optical counterpart, reported in Chapt. 5, suggests an extragalactic<br />
origin for this source. In fact, the <strong>de</strong>tected radio variability from 8 to 11 mJy within<br />
a week reported in Chapt. 5, is compatible with the compactness in an extragalactic<br />
object (see IDV phenomenon, Wagner & Witzel 1995).<br />
6.3.5 1RXS J072259.5−073131 and its bent one-si<strong>de</strong>d jet<br />
The images obtained at different resolutions are plotted in Fig. 6.4. The MERLIN<br />
image presents a compact structure with some elongation eastwards, while the EVN<br />
and combined EVN+MERLIN images show a clear one-si<strong>de</strong>d jet towards the east,<br />
with a slight bent towards the south at larger core separations. The closure phases<br />
clearly show that the source <strong>de</strong>parts from symmetry, with preferred emission to the<br />
east, both in the MERLIN and the EVN data sets. Two distinct components are<br />
present in the EVN image, at 9 and 17 mas from the compact core (P.A. of 89 and<br />
113 ◦ , respectively). Those components can be mo<strong>de</strong>l fitted with elliptical Gaussians,<br />
yielding flux <strong>de</strong>nsities of 3.7 and 2.8 mJy, respectively, for a core of 40.3 mJy.
6.3. Results and discussion 161<br />
Figure 6.4: Images of 1RXS J072259.5−073131 using the different arrays and with the<br />
parameters given in Table 6.4. The axes units are in mas. A (u, v)-tapering with a<br />
FWHM at 15 Mλ has been used to perform the combined EVN+MERLIN image.<br />
Using again the fact that we do not <strong>de</strong>tect a counter-jet, we can use Eq. 6.3 with<br />
Sa = 3.7 mJy (the closest component to the core in the EVN image), 3σ = 0.54 mJy<br />
(as previously done, the 1σ value has been taken as the root mean square noise<br />
in the image), α = −0.25 ± 0.2 (see Table 5.1), and k = 3 to be consistent with<br />
the discrete nature of the components, to obtain β cos θ > 0.29 ± 0.05, and hence<br />
β > 0.29 ± 0.05 and θ < 73 ± 3 ◦ . Therefore, these results point towards relativistic<br />
radio jets as the origin of the elongated radio emission present in the images.<br />
As pointed out for 1RXS J013106.4+612035, the one-si<strong>de</strong>d jet morphology does<br />
not rule out a microquasar nature for this object. However, the bending of the jet<br />
at such small angular scales resembles the ones seen in blazars. In fact, as reported<br />
in Chapt. 5, a one-si<strong>de</strong>d arcsecond scale jet is also present in VLA A configuration<br />
observations at 1.4 GHz, an unusual feature in the already known microquasars.<br />
A comparison between the VLA and MERLIN positions reveals that they agree<br />
within the errors, suggesting an extragalactic nature or a small proper motion if it<br />
turns out to be a galactic microquasar.<br />
Overall, these results are indicative of relativistic radio jets, although this source<br />
shows characteristics more similar to blazars than to microquasars.
162 Chapter 6. EVN and MERLIN observations<br />
Figure 6.5: Images of the already i<strong>de</strong>ntified quasar 1RXS J072418.3−071508 using the<br />
different arrays and with the parameters given in Table 6.4. The axes units are in mas.<br />
A (u, v)-tapering with a FWHM at 10 Mλ has been used to perform the combined<br />
EVN+MERLIN image.<br />
6.3.6 1RXS J072418.3−071508, a quasar with a bent one-<br />
si<strong>de</strong>d jet<br />
This radio source has recently been (March 2002) classified as a quasar in the SIM-<br />
BAD database, and it is not any more a microquasar candidate. It is catalogued in<br />
the Parkes-MIT-NRAO (Griffith et al. 1994) and the Texas Survey (Douglas et al.<br />
1996). It is listed as PMN J0724−0715 in the NED database, and is the source<br />
WGA J0724.3−0715 in Perlman et al. (1998), who reported a faint and quite broad<br />
Hα emission line (rest-frame Wλ = 30.3 ˚A, FWHM=4000 km s −1 ), and classified<br />
it as a Flat Spectrum Radio Quasar (FSRQ) with z = 0.270. Nevertheless, we<br />
have reported here our observational results for this source, since it was a candidate<br />
when we performed the observations. It presents (Fig. 6.5) a one-si<strong>de</strong>d pc-scale jet<br />
oriented towards the northeast, changing from a P.A. of ∼ 50◦ at 5 mas from the<br />
core (EVN image) to 20◦ up to 200 mas, at MERLIN scales.<br />
Mo<strong>de</strong>l fitting of the EVN visibilities with circular Gaussians reveals a compact<br />
core (0.7 mas FWHM) with 245.4 mJy, and two distinct components, one with<br />
11.9 mJy at 6.2 mas (P.A. 46◦ , 1.9 mas FWHM, present in the right panel image of<br />
Fig. 6.5) and another one with the size of the beam, a flux <strong>de</strong>nsity of 1.7 mJy at a<br />
distance of 29.6 mas in P.A. 27◦ (visible in the middle panel image of Fig. 6.5).<br />
Using again the fact that we do not <strong>de</strong>tect a counter-jet, we can use Eq. 6.3<br />
with Sa = 11.9 mJy (the closest component to the core), 3σ = 0.45 mJy, α =
6.4. Conclusions 163<br />
0.07 ± 0.02 (see Table 5.1), and k = 3 to be consistent with the discrete nature of<br />
the components, to obtain β cos θ > 0.51±0.04. Hence, an upper limit of θ < 60±3 ◦<br />
and a lower limit of β > 0.51 ± 0.04 is obtained, pointing towards relativistic radio<br />
jets as the origin of the elongated radio emission present in the images of this already<br />
i<strong>de</strong>ntified quasar.<br />
6.4 Conclusions<br />
We have presented EVN and MERLIN observations of the six sources studied in<br />
Chapt. 5 in a search for microquasar candidates at low galactic latitu<strong>de</strong>s. The<br />
first one, namely 1RXS J001442.2+580201, displays a two-si<strong>de</strong>d radio jet, which<br />
after analysis implies β > 0.20 ± 0.02 and θ < 78 ± 1 ◦ . 1RXS J013106.4+612035,<br />
displays a one-si<strong>de</strong>d radio jet, requiring β > 0.31 ± 0.05 and θ < 72 ± 3 ◦ . The third<br />
one, namely 1RXS J042201.0+485610, was not <strong>de</strong>tected due to its low flux <strong>de</strong>nsity<br />
and/or to phase-referencing problems. 1RXS J062148.1+174736 appeared compact<br />
at all scales. The fifth one, namely 1RXS J072259.5−073131, displays a bent one-<br />
si<strong>de</strong>d radio jet, implying β > 0.29 ± 0.05 and θ < 73 ± 3 ◦ . Finally, the already<br />
known quasar 1RXS J072418.3−071508 shows also a bent one-si<strong>de</strong>d jet, requiring<br />
β > 0.51 ± 0.04 and θ < 60 ± 3 ◦ .<br />
We summarize our obtained results for these sources in the next chapter.
164 Chapter 6. EVN and MERLIN observations
Bibliography<br />
Browne, I. W. A., Wilkinson, P. N., Patnaik, A. R., & Wrobel, J. M. 1998, MNRAS,<br />
293, 257<br />
Douglas, J. N., Bash, F. N, Bozyan, F. A, Torrence, G. W., & Wolfe, C. 1996, AJ,<br />
111, 1945<br />
Griffith, M. R., Wright, A. E., Burke, B. F., & Ekers, R. D. 1994, ApJS, 90, 179<br />
Kraus, A. 1997, PhD Thesis, Friedrich-Wilhelms-Universität Bonn, Germany<br />
Massi, M., Ribó, M., Pare<strong>de</strong>s, J. M., Peracaula, M., & Estalella, R. 2001, A&A,<br />
376, 217<br />
Mioduszewski, A. J., Rupen, M. P., Hjellming, R. M., Pooley, G. G., & Waltman,<br />
E. B. 2001, ApJ, 553, 766<br />
Pare<strong>de</strong>s, J. M., Ribó, M., & Martí, J. 2002, A&A, 394, 193<br />
Patnaik, A. R., Browne, I. W. A., Wilkinson, P. N., & Wrobel, J. M. 1992, MNRAS,<br />
254, 655<br />
Peng, B., Kraus, A., Krichbaum, T. P., & Witzel, A. 2000, A&A, 145, 1<br />
Perlman, E. S., Padovani, P., Giommi, P., et al. 1998, AJ, 115, 1253<br />
Shepherd, M. C., Pearson, T. J., & Taylor, G. B. 1994, BAAS, 26, 987<br />
Wagner, S. J., & Witzel, A. 1995, ARA&A, 33, 163<br />
165
166 BIBLIOGRAPHY
Chapter 7<br />
Summary<br />
After a <strong>de</strong>tailed analysis of our data, we show in Table 7.1 a summary of the results<br />
obtained after the VLA and optical observations presented in Chapt. 5, and the<br />
EVN+MERLIN observations reported in Chapt. 6.<br />
As can be seen, 1RXS J001442.2+580201 and 1RXS J013106.4+612035 are<br />
promising microquasar candidates. 1RXS J042201.0+485610 is probably of thermal<br />
nature due to its highly inverted spectrum at high radio frequencies, and we also note<br />
its exten<strong>de</strong>d appearance at optical wavelengths. 1RXS J062148.1+174736 is proba-<br />
bly extragalactic due to the exten<strong>de</strong>d optical counterpart. 1RXS J072259.5−073131<br />
displays a bent one-si<strong>de</strong>d jet, which is a common property in blazars. Finally, the<br />
source 1RXS J072418.3−071508 is an already i<strong>de</strong>ntified quasar.<br />
We note that none of the 6 sources studied here appears in the Third EGRET<br />
Catalog of high-energy γ-ray sources.<br />
We point out that 1RXS J072259.5−073131 is bright enough at radio wave-<br />
lengths to attempt an H i absorption experiment, that could allow to <strong>de</strong>termine if<br />
this source is galactic or not. In any case, optical spectroscopic observations of the<br />
first five sources are in progress for autumn 2002, to clearly unveil their galactic or<br />
extragalactic nature.<br />
Finally, the two sources belonging to Group 1 that have not been observed yet,<br />
namely 1RXS J181119.4−275939 and 1RXS J185002.8−075833 will be eventually<br />
studied in the future, together with Group 2 sources.<br />
167
¡¢ ¢ <br />
¢<br />
<br />
¡¢<br />
£¤<br />
¡¢<br />
¤ ¢ ¢ <br />
¡¢<br />
<br />
¡¢<br />
¤<br />
¤<br />
¤ £<br />
¡¢<br />
<br />
<br />
<br />
¤<br />
£¦ <br />
<br />
<br />
£¤ ¤ £<br />
¥ ¥<br />
§ ¢<br />
<br />
§ ¢ <br />
<br />
¦ <br />
¦ ¥<br />
¢<br />
¤<br />
¢ <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
§<br />
¢<br />
¨<br />
¤<br />
<br />
<br />
<br />
© <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
© <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¢<br />
¨<br />
¤<br />
<br />
¢<br />
<br />
<br />
<br />
<br />
¦<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¤<br />
<br />
£<br />
¢<br />
<br />
<br />
¥<br />
<br />
<br />
<br />
¢<br />
§<br />
¨<br />
<br />
© <br />
<br />
<br />
<br />
©<br />
<br />
<br />
<br />
<br />
<br />
<br />
¢<br />
¨<br />
<br />
<br />
<br />
¤<br />
<br />
<br />
<br />
<br />
<br />
<br />
£¦<br />
¥ ¢<br />
<br />
<br />
<br />
<br />
£<br />
<br />
<br />
¥ ¢<br />
¦<br />
<br />
<br />
<br />
<br />
<br />
©<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¨<br />
¢<br />
§<br />
¢<br />
¢<br />
¨<br />
¨<br />
¨<br />
¤<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
© <br />
<br />
© <br />
<br />
<br />
<br />
<br />
¥<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¦ <br />
<br />
¤ <br />
<br />
<br />
<br />
<br />
© <br />
<br />
© <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
¢<br />
¢<br />
¨<br />
¨<br />
¤<br />
<br />
<br />
<br />
¥<br />
<br />
<br />
<br />
¢ <br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
<br />
Table 7.1: Summary of the obtained results after the VLA and optical observations<br />
presented in Chapt. 5, and the EVN and MERLIN observations reported in Chapt. 6.<br />
An asterisk indicates a non-expected behavior for microquasars.<br />
168 Chapter 7. Summary
Part III<br />
General conclusions<br />
169
General conclusions 171<br />
Since each chapter contains its own conclusions, here we will briefly list our<br />
general conclusions after this work.<br />
Part I:<br />
1. We have discovered the microquasar LS 5039 after a careful examination of<br />
mo<strong>de</strong>rn archive databases and follow-up interferometric radio observations.<br />
Therefore, we conclu<strong>de</strong> that similar studies may reveal a previously unnoticed<br />
population of silent microquasars like LS 5039. If this is correct, the micro-<br />
quasar phenomenon may not be as rare as it seems.<br />
2. We have performed an in-<strong>de</strong>pth study of the LS 5039 multiwavelength be-<br />
havior, covering from radio wavelengths to high-energy γ-rays. In doing so,<br />
we have carried out observations aimed to study the flux, flux variability and<br />
spectrum at radio, optical and X-ray wavelengths. In the radio domain we<br />
have also obtained interferometric images in the continuum and single dish<br />
spectral line maps. All these data, together with γ-ray observations from the<br />
literature, have allowed us to propose a scenario and a mo<strong>de</strong>l to explain the<br />
observed phenomenology of LS 5039. Multiwavelength observations are clue<br />
to un<strong>de</strong>rstand accretion/ejection processes near compact objects.<br />
3. We have proposed an association between LS 5039 and one of the ∼ 170<br />
uni<strong>de</strong>ntified EGRET sources, which seems plausible due to the persistent and<br />
mo<strong>de</strong>rately variable emission of the source at both radio and γ-ray wave-<br />
lengths. If confirmed, this would be the first association between a microquasar<br />
and an EGRET source, and we suggest that other uni<strong>de</strong>ntified high-energy γ-<br />
ray sources could be silent microquasars waiting to be discovered.<br />
4. We have discovered that LS 5039 is a runaway microquasar with vsys 150<br />
km s −1 , escaping from its own regional standard of rest with a large velocity<br />
component perpendicular to the galactic plane. We are able to explain both,<br />
the high space velocity and the high eccentricity observed, in a symmetric SN<br />
explosion scenario with a mass loss of ∆ M ∼ 17 M⊙.<br />
5. The high space velocity and possible lifetime of this microquasar indicate that<br />
it could reach a galactic latitu<strong>de</strong> of b = −12 ◦ . Therefore, if the proposed asso-<br />
ciation between LS 5039 and the EGRET source 3EG J1824−1514 is correct,<br />
we could be able to <strong>de</strong>tect γ-ray microquasars up to values of |b| 10 ◦ . In
172 General conclusions<br />
particular, runaway microquasars could be connected with some of the uni<strong>de</strong>n-<br />
tified faint, variable, and soft γ-ray EGRET sources above/below the galactic<br />
plane.<br />
Part II:<br />
1. We have presented a cross-i<strong>de</strong>ntification method to search for REXBs, which<br />
are potential microquasar sources. The obtained results give confi<strong>de</strong>nce to<br />
the proposed method, since the output list of objects inclu<strong>de</strong>d all but one<br />
known persistent HMXB microquasars within |b| < 5 ◦ . The obtained list<br />
after removing previously known sources contains 13 new radio emitting X-<br />
ray sources, of which 8 have been classified as high priority objects.<br />
2. We have studied 6 of these 8 sources, and found optical counterparts to all of<br />
them. We have obtained accurate positions at both radio and optical wave-<br />
lengths, perfectly compatible between them. We also have obtained radio<br />
spectra and optical magnitu<strong>de</strong>s of the sources.<br />
3. We have presented EVN and MERLIN observations of these six sources. Five<br />
of the six objects have been <strong>de</strong>tected and imaged, presenting different mor-<br />
phologies: one source has a two-si<strong>de</strong>d jet, three sources have one-si<strong>de</strong>d jets,<br />
and one source is compact.<br />
4. After our observations we conclu<strong>de</strong> that the sources 1RXS J001442.2+580201<br />
and 1RXS J013106.4+612035 are promising microquasar candidates in the<br />
Galaxy. 1RXS J042201.0+485610 is probably of thermal nature due to its<br />
highly inverted spectrum at high radio frequencies, and we note its exten<strong>de</strong>d<br />
appearance at optical wavelengths. 1RXS J062148.1+174736 is probably ex-<br />
tragalactic due to the exten<strong>de</strong>d optical counterpart. 1RXS J072259.5−073131<br />
displays a bent one-si<strong>de</strong>d jet, which is a common property in blazars. Finally,<br />
the source 1RXS J072418.3−071508 is an already i<strong>de</strong>ntified quasar.<br />
5. Work in progress. Optical spectroscopic observations of the first five sources<br />
are in progress for autumn 2002, to clearly unveil their galactic or extragalactic<br />
nature. On the other hand, the remaining sources in the list will be eventually<br />
studied in the future.