<%BANNER%>

X-Ray and Infrared Spectral and Timing Observations of Galactic Interacting Binary Stars and Associated Relativistic Jets

Permanent Link: http://ufdc.ufl.edu/UFE0022086/00001

Material Information

Title: X-Ray and Infrared Spectral and Timing Observations of Galactic Interacting Binary Stars and Associated Relativistic Jets
Physical Description: 1 online resource (173 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: infrared, jet, massive, microquasar, relativistic, spectroscopy, star, survey, xray
Astronomy -- Dissertations, Academic -- UF
Genre: Astronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: I present results of a campaign to find and identify counterparts to X-ray sources in the Galactic Center (GC), searching for accretion disk signatures in the form of Br-gamma line emission via infrared (IR) spectroscopy. IR and X-ray light penetrate the obstacles that veil optical light, and give us unique opportunities to find and study obscured GC sources that can help us understand the mechanisms of massive star evolution and the extreme physical processes that occur in the vicinity of compact objects. I have surveyed ten candidate counterparts and found one with remarkable Br-gamma emission. Follow-up observations of the source, CXO J174536.1-285638 (hereafter, Edd-1), revealed a 189 +/- 6 day periodicity. I compare Edd-1 to X-ray sources with similar features and find it consistent with both Colliding Wind Binary (CWB) and X-ray binary (XRB) systems. CWBs contain two massive stars with strong winds that collide and produce X-rays. XRBs produce X-rays when matter is accreted from a companion onto a compact object. I compile several known CWBs and XRBs into an Atlas to facilitate the comparison of other unknown sources like Edd-1. A subset of XRBs, called microquasars, produce jet ejections similar to extragalactic quasars. The short dynamical timescales associated with microquasar jets allow us to study their properties in depth. I use high time resolution X-ray data from RXTE to study jet ejection events in the archetype microquasar GRS 1915+105. Examining a series of ejection events, I find strong correlations between the X-ray spectral phenomenology and the strength of the ejection, observed as an IR flare. Jet outflows are modeled with accretion flows because jets are a mechanism to transport angular momentum from the system. The Accretion Ejection Instability (AEI) model employs Magnetohydrodynamics and General Relativity to make predictions about outflows in XRBs. Using my observations of GRS 1915+105, I find a connection between the X-ray spectral behavior and the X-ray timing behavior independently (and uniquely) predicted by the AEI model.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Eikenberry, Stephen S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022086:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022086/00001

Material Information

Title: X-Ray and Infrared Spectral and Timing Observations of Galactic Interacting Binary Stars and Associated Relativistic Jets
Physical Description: 1 online resource (173 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: infrared, jet, massive, microquasar, relativistic, spectroscopy, star, survey, xray
Astronomy -- Dissertations, Academic -- UF
Genre: Astronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: I present results of a campaign to find and identify counterparts to X-ray sources in the Galactic Center (GC), searching for accretion disk signatures in the form of Br-gamma line emission via infrared (IR) spectroscopy. IR and X-ray light penetrate the obstacles that veil optical light, and give us unique opportunities to find and study obscured GC sources that can help us understand the mechanisms of massive star evolution and the extreme physical processes that occur in the vicinity of compact objects. I have surveyed ten candidate counterparts and found one with remarkable Br-gamma emission. Follow-up observations of the source, CXO J174536.1-285638 (hereafter, Edd-1), revealed a 189 +/- 6 day periodicity. I compare Edd-1 to X-ray sources with similar features and find it consistent with both Colliding Wind Binary (CWB) and X-ray binary (XRB) systems. CWBs contain two massive stars with strong winds that collide and produce X-rays. XRBs produce X-rays when matter is accreted from a companion onto a compact object. I compile several known CWBs and XRBs into an Atlas to facilitate the comparison of other unknown sources like Edd-1. A subset of XRBs, called microquasars, produce jet ejections similar to extragalactic quasars. The short dynamical timescales associated with microquasar jets allow us to study their properties in depth. I use high time resolution X-ray data from RXTE to study jet ejection events in the archetype microquasar GRS 1915+105. Examining a series of ejection events, I find strong correlations between the X-ray spectral phenomenology and the strength of the ejection, observed as an IR flare. Jet outflows are modeled with accretion flows because jets are a mechanism to transport angular momentum from the system. The Accretion Ejection Instability (AEI) model employs Magnetohydrodynamics and General Relativity to make predictions about outflows in XRBs. Using my observations of GRS 1915+105, I find a connection between the X-ray spectral behavior and the X-ray timing behavior independently (and uniquely) predicted by the AEI model.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Eikenberry, Stephen S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022086:00001


This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20110318_AAAAMA INGEST_TIME 2011-03-18T21:40:21Z PACKAGE UFE0022086_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 94638 DFID F20110318_AACGPI ORIGIN DEPOSITOR PATH mikles_v_Page_004.jp2 GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
0d82783369cbb4b8fabb94097bcf0663
SHA-1
7c2a33ea59a43aaabf29a6cc368928df34cbe767
34397 F20110318_AACGOU mikles_v_Page_160.QC.jpg
67fbd31aa5ebb9fb77e2f07997dd9be2
85dd1a1b814d54c0c6368089f7732df7614fa352
59707 F20110318_AACFMG mikles_v_Page_114.pro
6e4dd4418df8b4280ced1ff2d1f917c4
bd7cf5ebbbce88c4d9c249412d71848cdec2564b
7977 F20110318_AACFLS mikles_v_Page_067thm.jpg
262ed6b77d907f5bab9d3d32afb81d24
76853c8afde311a166523064384abfcdbacf147f
1051980 F20110318_AACGPJ mikles_v_Page_005.jp2
2b5d582f2c23b7a35dea27e777daa5d9
77926df598a9725aed4dbf30597c9177b567d762
12968 F20110318_AACGOV mikles_v_Page_164.QC.jpg
b4ceef682c444f9e170b95a6a6e1fd7f
7d72037628c71657071b78ddb1937f2a19c57dc5
1440 F20110318_AACFMH mikles_v_Page_136.txt
9395e5841843873bd51c729cff50ee12
f62c000fce3ed16ce2efeff0deb5acd7aa4185a4
59054 F20110318_AACFLT mikles_v_Page_015.pro
0d9f7caeeedfec0c154159ac508898b6
3208c6919f0cbb849111b79ff60b7b72435e043b
1051961 F20110318_AACGPK mikles_v_Page_007.jp2
d663b66e7ba075d66448281edebfc518
68898a86de4f1a1b2ac1f5b6f305e6f3c4e0d5b5
104646 F20110318_AACGOW mikles_v_Page_165.jpg
7cf4aae2f7a12dbaec92bb5b0157f6ee
126e55afcaf84a8c64d601a5405fa30712e9f404
1053954 F20110318_AACFMI mikles_v_Page_165.tif
577a3ad922c55b1aa602e210eff66f42
bc04f958fe002acbd899fe9a3651a328c7d49e21
8381 F20110318_AACFLU mikles_v_Page_021thm.jpg
47019a690118f6dbbecc427687ccc84d
49f5e701fe9f58ebb76ddaf7fc5a822f2606129d
1051981 F20110318_AACGPL mikles_v_Page_008.jp2
90def3a686cd81e61e69f54cd5f8b9da
f54de4b8842b179ccb410f6ec81d1a3dc5ede1cf
124275 F20110318_AACGOX mikles_v_Page_166.jpg
577784229b8cae0d964d9a94eeea3791
ed58a8ff0e526f6c41b3e1d88683163ff326c722
795 F20110318_AACFMJ mikles_v_Page_053.txt
3a1d8b5a9d7dc95beb46198f230b158e
a218daed768c5b689587c3dac3b4222d6e573f5d
17993 F20110318_AACFLV mikles_v_Page_158.QC.jpg
3ea513680ecc80ddd2004a6b18f8f3e6
0a69f6d4583cb0ee44f091c51dc4b57509cef0cb
680893 F20110318_AACGQA mikles_v_Page_034.jp2
4ca77999bc13d4bab6e602d16426808d
70195e8a200f02a60a89cf400d1cc931908ab577
1051985 F20110318_AACGPM mikles_v_Page_009.jp2
7a66ef8ac7d8f68325682ba02dc361c8
ade3cac0b54e9c66f732153f1bd93909cff364ee
101174 F20110318_AACGOY mikles_v_Page_167.jpg
e444bbe5f1c2b03c2f97f412ad53b21f
7abc09519db794c63f8f80cf1ed1e80adfa3de3b
741 F20110318_AACFLW mikles_v_Page_078.txt
ae1c2806f6cf6368b15ca7bf54273750
172a7e86d9c529808b9ba3729d54227603c0d5c9
6489 F20110318_AACFMK mikles_v_Page_058thm.jpg
626f11b8defaa51dda8f5e4f4c0ab3bc
ca68a1710440716658ffc4767437fc15f831f084
1051922 F20110318_AACGQB mikles_v_Page_036.jp2
c7bbd3d4e0231a00c92067476486d49a
7c77c513505f58c699e194f1d9a1b645af921eba
1051982 F20110318_AACGPN mikles_v_Page_010.jp2
e7dd021344d8e51b934caa8dbe042e0a
12a802435133c49bfd9a55b582c64225a46e1f99
1851 F20110318_AACFML mikles_v_Page_097.txt
ae1497676679417540b91711eb275f2d
9b700df755101d20b0323c75fc456b059a46c21c
667405 F20110318_AACGQC mikles_v_Page_037.jp2
f314aba6722a4ca30547b3227c38c119
a2f33bc3f983e34532ad7e1d83d6e72fe9d71a62
37553 F20110318_AACGPO mikles_v_Page_012.jp2
cd9e510358204b514815b006a4dc50bf
e5f9dac367557d7785ecb1e3b271657baef3fb55
31376 F20110318_AACGOZ mikles_v_Page_167.QC.jpg
f108f9ad16a912aff7ab02eb1285c31b
82013dd5c7097bafdab1798f973ec91e96665b09
109626 F20110318_AACFLX mikles_v_Page_124.jpg
1d19ba037e360b04f4c10202ccd05a1a
d1622b48510bba5679de9a9659a6f550024e8f1e
35385 F20110318_AACFNA mikles_v_Page_120.QC.jpg
0be5c27a03245b0b4ec23bf8748751bb
68ab80721071a31e51b679c06ec89e187a09a44b
61451 F20110318_AACFMM mikles_v_Page_034.jpg
1c5767fcfe1219be36bfad0681d667d5
e93f398b1edaf28cdfd426b870c3084dc28fdc19
F20110318_AACGQD mikles_v_Page_039.jp2
e617823d74f09e087fff6a453a49a423
bc974bd311eb4038e7b69fa555d1dcc883b8445c
1051984 F20110318_AACGPP mikles_v_Page_014.jp2
6abc37eb914f15a106c127208d3ca5d3
788c9662a5f6ff2d2ee8de3fbecd915fd0107f9e
37976 F20110318_AACFLY mikles_v_Page_143.QC.jpg
7a4626b010fb492812aec7980feafe89
d201dee066ef853e50e9ad3379f2a697e33b7043
4777 F20110318_AACFNB mikles_v_Page_107thm.jpg
d22499998a678b09efc89a90f5e6b689
cdc75ff11c711045c689dd7df2000620d931fa16
70230 F20110318_AACFMN mikles_v_Page_025.jp2
5c3d553d0ad41737f9fb2b9df16f2252
69f214be2dc73fad466ef761b719d561f084b7c3
1051974 F20110318_AACGQE mikles_v_Page_040.jp2
84d419ea2268f93497d23c93a04e0e23
3b5fe64b5427a1f280ca3102f56995bedc092de2
F20110318_AACGPQ mikles_v_Page_017.jp2
2ceb1b9920ccec06bccc4fcb55658135
fc0c95ec2722e00c0aa20c95b94e4e27bc71e237
2359 F20110318_AACFLZ mikles_v_Page_009.txt
119a61ba1d7b5d536c68c09a0ce15db2
01278f8d8859fb83bffca641474acc329fa33988
35567 F20110318_AACFNC mikles_v_Page_166.QC.jpg
15cace876e6c87ba863ab0f5824f6bd4
1a04f176de5db353fb3ba39c42e8bef321012ca7
2408 F20110318_AACFMO mikles_v_Page_122.txt
9853cb29fb8aa6bcd5eb316f27f10031
9894149b96c2d85410c5bafc12cdaa6a16d299ae
1051958 F20110318_AACGQF mikles_v_Page_044.jp2
c87c538c7622b86a3b0eb6e9724b8c8e
63cf66227f6e1b345cb8fd5ebe44e90c369cc2c5
1051986 F20110318_AACGPR mikles_v_Page_018.jp2
30afa83b1dd8519495b47c8c8eebbf27
d1ed79c9b8093c44628f0d5331b600536efca807
116674 F20110318_AACFND mikles_v_Page_031.jpg
24df7f54509e932d8135958dd1f6f61c
d14d2f2ce2362561e5c45331b4fb9940f928e52f
32743 F20110318_AACFMP mikles_v_Page_170.QC.jpg
26db808344e6058b7f83fcca01f18aa1
f41cf5a0ba3e2b0472399cf6414b616d19ba62fa
1051908 F20110318_AACGQG mikles_v_Page_061.jp2
98c2284169a1b0f9e1c82e20936b2650
ebd1c20a339bc56ec3e4bee0e04a3e97c6f4f2b4
1051983 F20110318_AACGPS mikles_v_Page_021.jp2
951f129343a515245052891a30f95ccb
2307cf48080d8c5cb68ea9ad0673d825b9eb42cb
110603 F20110318_AACFNE mikles_v_Page_013.jpg
a018adc4f17e024e17b65f79f8d1f6e7
d075cc6841892b4540dc3ebd7b5c58612c2000e3
35935 F20110318_AACFMQ mikles_v_Page_162.QC.jpg
cfb4e49400becbeffd87c591c15dc268
581fe1248e8d7db078f122ae893a8e5e5b0b358f
1051971 F20110318_AACGQH mikles_v_Page_063.jp2
2a6aa5fe377885ca9ecd82f4543236ab
d21de31f5d80707bca8d1f48c53b57f64b187591
1051936 F20110318_AACGPT mikles_v_Page_022.jp2
54347ebd99e7900815d6b5d008b38df6
c8043ca9ff16fb58ee608eb99b9ee25a0428b558
F20110318_AACFNF mikles_v_Page_139.jp2
8bc2924cfb7263cb8e3e5325be4dcde5
8a68b8ea9abe1e3aadee8b2256557d7bff03f922
8410 F20110318_AACFMR mikles_v_Page_015thm.jpg
3d702113f32301b61ab6b2370b9cc389
0b15ba23b671651777132f2482bd9d21c920831b
1033934 F20110318_AACGQI mikles_v_Page_065.jp2
652fc2f49adcec4868d916ba115a4f33
74d64866e7da6e8709dc0712ecdf680257ac108c
631472 F20110318_AACGPU mikles_v_Page_026.jp2
e841153edccba97f1cc5f9f2061d36d8
1137a13cb53db1665f365ed5879ec39b5d2fc1db
12562 F20110318_AACFNG mikles_v_Page_163.QC.jpg
bb37c64d38627131c9682031c207329d
3e82a469fe29c2c417455bbbc6660b2df7ed789d
85503 F20110318_AACFMS mikles_v_Page_023.jp2
05eb925aa7f14174ff80d3ee7bb3ae78
04ecab5b4fa472ea996062e13326d56a546367cb
1051955 F20110318_AACGQJ mikles_v_Page_068.jp2
312c9cbad2128f539da74cf503989098
34f80eebd979bdab7f237a8d5f295b6a3bbbd58a
628542 F20110318_AACGPV mikles_v_Page_027.jp2
ea686b2f88857279f99236e9c5f71aa9
3c43e6ca9265296f1ccb1e30df56047a5c2b98df
56170 F20110318_AACFNH mikles_v_Page_148.pro
c8f090df28a83bbb73ef55eb993248d4
648ad0299ad71777ee824f85c1184b5caaec0fde
105520 F20110318_AACFMT mikles_v_Page_170.jp2
a84281427724d5e61ca3e6f33ffb666c
506a6ebf5d8ea0fda7aed0e391141c1a868922d9
1006951 F20110318_AACGQK mikles_v_Page_069.jp2
4cd9624c6ee5335ec83963a3a8d064b1
9c1e1d98f89ec694bbe9c6671fa4061ebf04cef2
1051977 F20110318_AACGPW mikles_v_Page_028.jp2
bf2cfd884619000d8d0f941633b1b144
9dd356d61fdc07a760b58d5dbbcfea4392d7cdf2
1051972 F20110318_AACFNI mikles_v_Page_042.jp2
46a307ce6fd5f99adf07f0dfe1a9ef45
d46cd869c4037330f91046b304dc378d3d22a9e7
2409 F20110318_AACFMU mikles_v_Page_113.txt
7d12d031dc93c58cc6f6c98612d90d06
caea02bb02f8d8bdeb922709966b1b07d4afb941
F20110318_AACGQL mikles_v_Page_070.jp2
f376d08159b204fa5cfd05fbc2f23c3a
3589bd29ade1dc17a6cfb11c37902cddf3469480
F20110318_AACGPX mikles_v_Page_029.jp2
0f611e0b70db265592eb56b7d7b440b2
5b7a743545ef39746c5d0ed1fb6dd58e98b6244a
27235 F20110318_AACFNJ mikles_v_Page_097.QC.jpg
e576c2fcba6bdfb66d9049485bc88e0f
f673c51822ae5e4c619c5bd5cf7271ccc385d24e
1984 F20110318_AACFMV mikles_v_Page_170.txt
0ba5c5e37e84b834ecb6782e42d48002
3578c1bd23ac5a78c513bed4162b7a9635abf8f0
371732 F20110318_AACGRA mikles_v_Page_107.jp2
6ed6a4f2b8b8d449a3fad1d0aa275ca2
50934bf94f9d84e5edea67ae7c7bdad564b8b358
616143 F20110318_AACGQM mikles_v_Page_074.jp2
85dfef5aade28b833697a540bcf645a0
ce015de3b33f3d069d549c3d41ccc522e268db4f
F20110318_AACGPY mikles_v_Page_032.jp2
6dfa9ad58912fea0ed868361746540ef
1285f1d0705bc3995c8b52267bdc3e6a2b3c0d1f
90937 F20110318_AACFNK mikles_v_Page_156.jpg
0c5bba70b6dc3ab9b52e14a24f362fb3
bf4041d7c916f1d185c3bab20a8dcabd53dbf797
25271604 F20110318_AACFMW mikles_v_Page_153.tif
c3d07dfdfd73dad75d4d2eea839b220f
047550b52f62fe252571ccce8f63dd0c07294852
639719 F20110318_AACGRB mikles_v_Page_108.jp2
41768ba5020f25a4c805d920aabb04df
7fa5d3b88ed741ff2372411deed51e9c8d73b5f9
25266 F20110318_AACGQN mikles_v_Page_076.jp2
9394b292c0c6b2fb22684f9ae766e26a
43ff4486a6b01c68d1eed51c8b8d281fd043b019
F20110318_AACGPZ mikles_v_Page_033.jp2
1771b33fbd2b9cfee668beb49ba54cb1
29fc7808b86f8446583ef4dd146ac4527415e74b
700097 F20110318_AACFNL mikles_v_Page_158.jp2
6822e69793c077b264d48f0e0fdfb45d
1bae4594729b0257a186b220c6a6ad2d2320ac77
105934 F20110318_AACFMX mikles_v_Page_093.jpg
6ac57ee688ac8fe7532d8f2a991eaf6c
ef439b7e664610abef9df261169f4becfaddbfe1
1051884 F20110318_AACGRC mikles_v_Page_109.jp2
bedb28b5ac825630b2b7dfb486a0efe8
8c9b07a77c53c933ae45f34c20a760feca001093
F20110318_AACGQO mikles_v_Page_086.jp2
aa3f019fe9c0f51a064c283ee615905d
4a0e66621d8022b1b4e2d55792dff0f06f94e2b3
9534 F20110318_AACFOA mikles_v_Page_027.pro
c5c67977a61db6ed436989f75e0bccd3
a5255822ca76f98ebce263535dd9879d1e83f382
35748 F20110318_AACFNM mikles_v_Page_088.QC.jpg
4baec41858d34b38e3d06a555ef12604
3849cd2b43ae6137c0f84781ee11d42baf892f59
1051976 F20110318_AACGRD mikles_v_Page_110.jp2
21d057d5b66ff3dd2a3a617f87598d21
98472967d3386bcfcfa76dd5ee263b3e49b71cfb
1051899 F20110318_AACGQP mikles_v_Page_087.jp2
8705e10830b36129e5f60acce00b2f5f
f84050f577aa9c13c07ea9a205127cf77360466b
20234 F20110318_AACFOB mikles_v_Page_003.jpg
19c3382a061dea282ebce3a524de4f10
2115aa1717ed94f29784624f5c1b41ebd08892c3
6006 F20110318_AACFNN mikles_v_Page_052thm.jpg
2d34a6872ffe62c092483c1800d3d8fd
31fe870fa685719ac5c05b03e088cac75a5e5105
662 F20110318_AACFMY mikles_v_Page_158.txt
ba6f44c1446d22ed184b1545be59f074
4bd457f7872417555df9b4f4f0f846fdd7744a7a
1051967 F20110318_AACGRE mikles_v_Page_113.jp2
501a3f09eebe22979ffa4bac06d96761
73a29e63b7f5c7c74295f746d37938da340e5be4
117775 F20110318_AACGQQ mikles_v_Page_088.jp2
4e35652068f040e2da943ba62d9b29a7
da1dd4f0fda4da21b795a937f2ba98f6864fc497
2260 F20110318_AACFOC mikles_v_Page_120.txt
444c219d97c2b096ee34b901aeb877ee
1ebc238eb185ab956930151c5040243c18902e5d
8114 F20110318_AACFNO mikles_v_Page_099thm.jpg
cb0f4d80a1e6cc32e8b26014a5062087
94b6af13e19736d87c8afc27d12e25eca0f8048c
F20110318_AACFMZ mikles_v_Page_143.jp2
8606a3fbf36ae1746db2ce224132d7d1
324d8464fee8ab324079a6d8a5e2579c1ccb5349
1051950 F20110318_AACGRF mikles_v_Page_115.jp2
5ba9d5fa6ed276d8a39c67a3cc08e8cb
99ad30657faadffafddd4e6948d6ea817acd223e
1051963 F20110318_AACGQR mikles_v_Page_089.jp2
a8abbae5cd7b4b99042c79b53663441f
1e3a94f9f32d699f70804ee15263725503b83567
8444 F20110318_AACFOD mikles_v_Page_117thm.jpg
603285b41b8f17e122048c10265df617
7cdca4ea1cd06ed8e4381b743dabafb31958a319
266 F20110318_AACFNP mikles_v_Page_049.txt
6c9e12dd04ffb930a1030fc2f82110a4
1f624e151dbe15b202346ee47316a316d4ae2b13
1051979 F20110318_AACGRG mikles_v_Page_116.jp2
bfe42d58db07a72a16542309618b3fc9
a303e25dacc5a9afe7b5b8af98421b39ce1f5a69
110848 F20110318_AACGQS mikles_v_Page_090.jp2
b46cb4a340c157305fd2147520563f10
1315154f4aaee4fbb087fe90646129c9ca8f2129
56783 F20110318_AACFOE mikles_v_Page_039.pro
6663c48ca4847764ed18088b11b60c36
68ab10601c342397dfbfdf3b86eca6d0bd022e18
15153 F20110318_AACFNQ mikles_v_Page_049.QC.jpg
a79678e6cec2f12870d3bba78f754245
f21f38ced906fa485546ed78959b97e3f4a7aedd
F20110318_AACGRH mikles_v_Page_122.jp2
5252a66eb7c9ae677ec12178b33b7f1b
66dfc2e9e327e28701284971924b50fa1c387b93
976139 F20110318_AACGQT mikles_v_Page_091.jp2
01352db39c744526e2167c0296574f4c
533ca25bd4e9bbfb1d7a2ebc4f228dbe57266755
638720 F20110318_AACFOF mikles_v_Page_024.jp2
2d1df18c07fd91be6f6ec911bd961c15
eff831c3520ce84f849c7b13926fb07ef7f89abd
F20110318_AACFNR mikles_v_Page_128.tif
fd097405854283241c69d2bde0dcf4f0
48e2f76c7a51ba3c71e50079fcbb9fe10b24c48a
F20110318_AACGRI mikles_v_Page_123.jp2
7c62f25c535bd4318fa31d6e1edfabe3
a6a84a7704be2daedd6aba227d9f8e2e373caecd
437418 F20110318_AACGQU mikles_v_Page_095.jp2
6090fd8899e7b779811f42372a39e0ac
3f730876ed8602514067a263a52f31958c2bd725
109393 F20110318_AACFOG mikles_v_Page_139.jpg
5ab2869c6f02dd7354bf1fac4e165505
49a26902ed2c78e34997b1ed443dff00db48c67b
114011 F20110318_AACFNS mikles_v_Page_117.jpg
f7a745df79619dcf5ba7501ec4cc1b7c
6b2aedc84b4cfa7a8937d7d8f09782c9a96c1048
955983 F20110318_AACGRJ mikles_v_Page_128.jp2
733fbf105be2ab2348d2ef767e902472
02339e97567b721a38e4accc137606ec6d8d5f2e
75775 F20110318_AACGQV mikles_v_Page_096.jp2
22efc9c4b227209352fd7474860eee16
a29580d61606afd22c25a79829fb71964d6e6bba
22718 F20110318_AACFOH mikles_v_Page_072.pro
bb0cb40300ca012efac0303d839e03ca
baec0213eec8e042d6d006e75ba26ee98c050d15
5009 F20110318_AACFNT mikles_v_Page_049.pro
b123966b1739b3e46245128c3d94540e
67b4226604bd824151bfeaec823a00a7835d77ed
637483 F20110318_AACGRK mikles_v_Page_129.jp2
33192352b92e7f11c64f61a843327324
d9e536a50a11e188ce4628e6ae0f6a465b5ff54e
100656 F20110318_AACGQW mikles_v_Page_097.jp2
6ad0f065bfa6834ded1dd67fcddeb6ac
ca51a31767ab0922b8f68d4c232deae6c9d0dd10
30012 F20110318_AACFOI mikles_v_Page_057.pro
eeed68758fdc734012c82e4d1b477b7b
44c800c69af65398e36533ecec43b101e924c5a7
29322 F20110318_AACFNU mikles_v_Page_008.QC.jpg
aadeb53eec0d6e85636251fe3c926260
af5c17aa915a605a6b26b51932256b01301abc6c
57363 F20110318_AACGRL mikles_v_Page_131.jp2
17a19779d23932eb6e49447d4bb8cb94
720898b649285fa05310e25467acb72da7f1559a
1051970 F20110318_AACGQX mikles_v_Page_100.jp2
b15f6df0a9797e7d7f942f92b630a8d7
71eed0f8412035105ff0138e53caf7cf40cf97a0
107844 F20110318_AACFOJ mikles_v_Page_067.jpg
e1ed8e662ff4c74528b7f0a6206d49ed
f92dddcd912536cc0e1b24287e1cc7f7bdb4622a
3430 F20110318_AACFNV mikles_v_Page_056thm.jpg
a98c641d11a0daaed43885d4ee8c29dd
e33c17aa02fee45720a308f282d2067aedee37fe
42636 F20110318_AACGSA mikles_v_Page_164.jp2
376119d53a6bc1efa757a91b6dad2d79
9825af8a99e74a3688046006b37ea9a003c3fe18
411676 F20110318_AACGRM mikles_v_Page_132.jp2
b5df537005497bea6976c5da4e70b1a5
c725843bee1eaf88cdc80c61baf1f9afe54bba99
1040441 F20110318_AACGQY mikles_v_Page_104.jp2
55011b6b17bd35e0bcdb209eaabfb1f0
1bd76bfc0723b655e98cc7f009651b0e66aeb35e
8666 F20110318_AACFOK mikles_v_Page_107.pro
8ac543150bd8b5db16cb44955fd9ddd0
0a05fb66899ec730f844f7d7d85a4b7cc994a212
3030 F20110318_AACFNW mikles_v_Page_078thm.jpg
d83d0cca07386cc47c6a08cdbf9186df
e656d02567a5c88e07b169cc88a9e7750808118b
126156 F20110318_AACGSB mikles_v_Page_166.jp2
568d4c807bf4a5b9d461d0d09396f2ea
2fe60cbcb4acf25769e00f3f9e8ab9dcac969c82
69833 F20110318_AACGRN mikles_v_Page_134.jp2
4e4dd875a56f445d216bf714ff8cfe5d
155073015667ff3a300a3d852a3be3af47f631ed
343159 F20110318_AACGQZ mikles_v_Page_105.jp2
09d73fa8ca1c486abc37d0914d07239f
316df9ccd9fafab9c3d1975b11d6871e1f54ae8a
43242 F20110318_AACFOL mikles_v_Page_051.jpg
3bfc6812a93aee7ede96ce92c718df81
3b82c174153acc169ea856811b94732e715ad1a3
467354 F20110318_AACFNX mikles_v_Page_053.jp2
b6e091526b7938ef336076f6ea71e284
3097d7e0b4a891278b09491b1f49e0fcac11ff9e
104064 F20110318_AACGSC mikles_v_Page_167.jp2
79af25fa678b73666b1761551321dc78
e52132158fb589f185bbfa669a4cb54eb848c0a1
55317 F20110318_AACGRO mikles_v_Page_137.jp2
814f5b05607d4e95356029912c5cc2de
6d72ecd101e028f108fa6eb864b8d3e76d7d1b5d
1051975 F20110318_AACFOM mikles_v_Page_112.jp2
4d8f66cbd07c9edd93592723ccecc3ab
3ce35f8102ce49d3f0f3cb730a5acd706789f263
120338 F20110318_AACFNY mikles_v_Page_032.jpg
8a531950701e13aaabac6b241fd51270
ec16aca8304dc3ceb8f8fcd1f89adc6e967ca0b4
31274 F20110318_AACFPA mikles_v_Page_010.QC.jpg
0a3a58e095fd54a674ef8b202b163f9c
8e5857f3b48493cd591fe6a84bb6dcca1a15ac91
110769 F20110318_AACGSD mikles_v_Page_173.jp2
a4a4732a3924d340d5b7593c8aca8761
7ff5afa3bd2d12152cf42568e5b220aeb5495cb0
F20110318_AACGRP mikles_v_Page_140.jp2
61a2ed83f4e017910ed9fbe1194ad563
faecb13c54137963da9e60118f12ea7a5beca63f
1295 F20110318_AACFON mikles_v_Page_007.txt
68c9fc9db912d52484c8c33947e091a1
c4409f700e68cb1127bbf5bec1e0c74267ac0887
115070 F20110318_AACFPB mikles_v_Page_014.jpg
603c9f94b65d4e0ac5314346eae4913a
3efa4c52b794cffcf56e4a6d4ec181e5854404f7
1975 F20110318_AACGSE mikles_v_Page_001thm.jpg
48a93f1b465d524fb1e4e99c2d708970
a5c1554e9f0612bf891aaed0f223498da8bb0b7c
F20110318_AACGRQ mikles_v_Page_141.jp2
264c22831756f5b02d4417ddf53928ec
fcc5aaadda14af09143168ee8903b4a01259ea4f
F20110318_AACFOO mikles_v_Page_106.tif
1e53a2ce39bcc14ccb08feadd7141512
f24bbc591136f38fd64ef4618da25f9685e65405
8347 F20110318_AACFNZ mikles_v_Page_061thm.jpg
8da7c2fac7508267194676ec02efaafb
36f31ecb9d736e7009f68ae835fd3a1a895b663f
453840 F20110318_AACFPC mikles_v_Page_082.jp2
71121c628ee66f9897c5d5bed848df1f
a2a39f91ae0b625d8bee18a2b6750bd9b550ab4c
6684 F20110318_AACGSF mikles_v_Page_005thm.jpg
caf4fa87b9aedb6f2cfa8a9fa31fb356
f239d57814b32095775c81936ef1741ccee2f7be
1051925 F20110318_AACGRR mikles_v_Page_142.jp2
9f3389379e59d8a932394cc48adfd787
7d120de1022b6ffe9aa39add6ac2b2adc8dc96c6
2214 F20110318_AACFOP mikles_v_Page_148.txt
587aaab814740a3c5d3b4193c5410194
cbc8912d04310cfec89b6a4f869784cbcff562b7
9197 F20110318_AACFPD mikles_v_Page_155.QC.jpg
dd42a4e6ef1601ff68c564aa2ce3316d
a6b5def7d007c67dc306eeb1867a36265520932f
7423 F20110318_AACGSG mikles_v_Page_006thm.jpg
3bad04ca36f57a908aeb29c20de6372c
6215bcc5975687a997bd7f7771c3575d530ccbdd
1051953 F20110318_AACGRS mikles_v_Page_144.jp2
2ab53770100fe24824b65d72b47ff078
c8a70563a42b526edfeaca78046451e6b6fd581a
39828 F20110318_AACFOQ mikles_v_Page_023.pro
123e93076374b9333b52b5d5805f5acb
67eb10b70a30442779f28bd873310cd005dffc3f
958155 F20110318_AACFPE mikles_v_Page_081.jp2
e633312d16a373618628fdc8ad0a5826
e35912441a997ac1ee6bdeddcb71110a5d693c10
7267 F20110318_AACGSH mikles_v_Page_011thm.jpg
c0e6868cb037b7132f4fd3af8cc69922
b4f622df0f2fdbe3c99d3e25f69e144572884292
1051969 F20110318_AACGRT mikles_v_Page_145.jp2
1efd7ef7a03f594fb6afa5793314b644
2c752e583e00716c9b595865de3821798a449c1b
403 F20110318_AACFOR mikles_v_Page_051.txt
858192bdca6d834b438345a28c130d2b
27057ba42500349d7ecddd4b0c1bf8b8baf4968d
72842 F20110318_AACFPF mikles_v_Page_126.jp2
d02c9dee53a6b1a3020a8ecc80819750
a2ce81fb8bf11bb58dbd0b0032dba8c7e258f832
2569 F20110318_AACGSI mikles_v_Page_012thm.jpg
df9b1d3e4b203068d59bd64b95edd15c
8deb976b8dd3aa56f00adde35e81116a9d3578b6
1051949 F20110318_AACGRU mikles_v_Page_146.jp2
604505da8c12c7c34ba064dd9a396d94
d4baf46b6cce7c69624b575b40ad3b73de5c206f
2185 F20110318_AACFOS mikles_v_Page_093.txt
7f38d528da319ab819067617475892c6
cc75335ef7777e6aedb14b0e8c2c57e49bc17e0b
7403 F20110318_AACFPG mikles_v_Page_066thm.jpg
e7bf82925bdf49e7fdf84ce31a930fac
66a1850b4600588115c562ca736cdb64e14d91b3
8195 F20110318_AACGSJ mikles_v_Page_031thm.jpg
201b3145733521bc93ab856b8e6aa418
8055e319c4755b0510c9a28b8f1572973a27a57c
F20110318_AACGRV mikles_v_Page_147.jp2
cef96fceed8c0855756306a1fb7c5588
e2022681d01d425f84c37f4854a74a2f7176b967
36694 F20110318_AACFOT mikles_v_Page_141.QC.jpg
918de7e90fa11b1b0fd6c48e78303f83
e2ada6d59eebf0f24849209afdc6491bcfe603d0
107 F20110318_AACFPH mikles_v_Page_172.txt
cd063b430b65695ccda4e089476e520b
739749042c320cea49a676625b7cc7687915c9f3
8580 F20110318_AACGSK mikles_v_Page_032thm.jpg
39ff703d1e3c7ce7a87b0641d87fe656
098b9b74e925bfc726dda53fc9d6be3bfc7cd6d9
F20110318_AACGRW mikles_v_Page_148.jp2
93357f14ff2db5939ed1375ac9b00f4b
32f09ffa3180032246807a1d2df966cae2215827
315945 F20110318_AACFOU mikles_v_Page_106.jp2
a367978653fe7122801b94a1ed67682f
2439cd57c1395fde2a3eff0187b42579530d930b
8524 F20110318_AACFPI mikles_v_Page_051.pro
64d6dcf879d10b4580a8d1e68855abfa
66f3ea9e54903d2c8f0400b88f4f4485eea4d3bc
8382 F20110318_AACGSL mikles_v_Page_033thm.jpg
55603b9023e9ffd435f289e22fa4ab00
9c27fe9b96c26535531defef25550558df90bffb
36034 F20110318_AACGRX mikles_v_Page_151.jp2
aa95861ca27f17ecf6e749ff75df4690
d5896b4584936c10d25d421152918f6cea057cb7
2107 F20110318_AACFOV mikles_v_Page_165.txt
c90c15ed3863569ece026dff9ab72f45
9d4a6b606a6ead987bf7c3762e5049b6123502f2
32287 F20110318_AACFPJ mikles_v_Page_169.QC.jpg
6c462899c5215835a9af90fec970b57f
3aa1404aa87de850a8aab59b5983bc9091020e79
7374 F20110318_AACGTA mikles_v_Page_071thm.jpg
8a9144c75caf0e437592466062d7a14d
761d530d9cdb68355b61716c9f26729553d9fb65
8935 F20110318_AACGSM mikles_v_Page_035thm.jpg
ed062862d36a2e7b95823065461368c0
d38b6a9f13424e0dd2fb183b8a83e1fc2fd84716
37982 F20110318_AACGRY mikles_v_Page_152.jp2
2ae5a652117bc5c4a91c244e4479e3f8
d69d011ada3338c84c1cf012209d60e530e97802
988663 F20110318_AACFOW mikles_v_Page_156.jp2
3e275336a47b9b302347242214640c54
4e936950c2ecb7f42db88cfdd1cacec4bd2c0213
8360 F20110318_AACFPK mikles_v_Page_141thm.jpg
27886778778b2546d456ceec258b9dc6
8733f6a42f7239ac5080a16ef067080ee43d1228
4642 F20110318_AACGTB mikles_v_Page_072thm.jpg
95ba8ffc9bdce696917a75731c441a3d
6942771829af198760aa849a5878b50b88c3fb32
8406 F20110318_AACGSN mikles_v_Page_040thm.jpg
3323fced070a6bc7e28bb2145e5cef4f
f9bde8368a348ebfa635df7d5e61f49ac71a66e1
42786 F20110318_AACGRZ mikles_v_Page_163.jp2
b4bd7ae43dd5d2bf995ea2520cbfc447
b7a80285073598dc1a67d1c735e69660be55d8b0
849 F20110318_AACFOX mikles_v_Page_025.txt
e481d900851acc949c83e858c8b79bc0
d938d6a25b7257c4fd46fbd82b4499dcbdfaeac5
37386 F20110318_AACFPL mikles_v_Page_041.QC.jpg
b7da6c8c69a7e370a5e842d283dbafe8
3d2b24244d699f2f6d52f99333f8f1fba10209f2
5253 F20110318_AACGTC mikles_v_Page_075thm.jpg
7e386c169288a73aea1b6a43aaf75e43
28eebc9f8d6fae755ec1b7a77fb0a5753965c6c6
8469 F20110318_AACGSO mikles_v_Page_042thm.jpg
e94b90ce889937b976d70fdba2b775ad
95b4ac43681134ab50ab19f667b1b50fdc0d43d7
4739 F20110318_AACFOY mikles_v_Page_027thm.jpg
cebb2bc99c2d50cd25296f412a9f6445
88680c0a8bc527a72ca976167fab8ce676ad29fa
2254 F20110318_AACFQA mikles_v_Page_116.txt
1bf50948b89d9dfdc40147668b34b58d
e805ef7200722ea587b463ee1059399da358924c
35668 F20110318_AACFPM mikles_v_Page_022.QC.jpg
7f6e25b74c01cac7c174a266df99ef2f
6722d5b3ef86e6c6dd7e1474517796dca9d405a3
5000 F20110318_AACGTD mikles_v_Page_077thm.jpg
92cfbba40ba24975b9bfa80272f61795
298a475d6ca8e38c607889035a4e7368820f726e
8289 F20110318_AACGSP mikles_v_Page_043thm.jpg
b74f195aa4e3692c77ea487ec0812d81
242d6bdf9f195f91789e11162a8519a2e754682a
113728 F20110318_AACFOZ mikles_v_Page_060.jpg
9e46bf71f6fe6ebc8a02ebb513bfd318
844f84d8fb2cb948413dbd10b8f0e2b9c631c7ae
13358 F20110318_AACFQB mikles_v_Page_159.QC.jpg
49c4eb253f7c0cf02b8d6fce06ebef4a
f97437c29baf14d075bbc9d0ce2e8ada9796ba76
2431 F20110318_AACFPN mikles_v_Page_155thm.jpg
801b39d077b64c2ad6be885de6a4e6b0
3fcce61a7a669167460fd59e46a4762c545148c0
5919 F20110318_AACGTE mikles_v_Page_079thm.jpg
d3e2fd1a1d44a9864e472b190d0070e8
5a119f2cc1f4a092ca0106156b8e30922222456a
8664 F20110318_AACGSQ mikles_v_Page_044thm.jpg
c6a5f3dcceeda7e2b78b8f1ddf8e4fe5
3a047665131842187d4aed4c1c255c77a7e21696
88831 F20110318_AACFQC mikles_v_Page_004.jpg
1f986b6dcf4b018426c9d45178829c29
0658f7896f162585a094eb7a5bf27ff9a83887a7
8767 F20110318_AACFPO mikles_v_Page_073thm.jpg
a8f4a1cf59e3edd9d48558d558af4eff
9c9cba3d4274c81a8187022230b4d2ed32369219
4222 F20110318_AACGTF mikles_v_Page_083thm.jpg
ebafc096df420e529adcd990daddfc24
41572120308a236cefe5c45422a3656bde0836d4
8543 F20110318_AACGSR mikles_v_Page_045thm.jpg
11a096406412c653189fb05a6ce7449e
a1fe10af1b5f698864f8ae0acf0e8d2918953ef4
10897 F20110318_AACFQD mikles_v_Page_012.QC.jpg
e0d8f97b29dd51c368073f55479672ad
0a4885ce4ad096796a319845ee4cf25b43dacea0
94616 F20110318_AACFPP mikles_v_Page_125.jp2
892cf54ee1e862acbbbb8ae4932fe5e9
bfd8bbfe345dbe9b34d9d7eb4d62e64564204d83
7953 F20110318_AACGTG mikles_v_Page_085thm.jpg
39101faa6c051ade8ca65736eca7ce96
c20b990e0a69ccc2c075e1555b594c61f136e4a6
4591 F20110318_AACGSS mikles_v_Page_048thm.jpg
ec3c9f1ea6795c48d7cb4b878aaae89e
13cd89fd0807a8b3936de563f09767431bf05dc5
560033 F20110318_AACFQE mikles_v_Page_133.jp2
8971f8d63e381533e93a7378c85e2ca4
7829da3e2d585cef4e2acbf9c593070eb597f6a3
57915 F20110318_AACFPQ mikles_v_Page_033.pro
81aede19f5f82a46d4030cd5e1d91aef
d9d03772bd356f6dea0d0c716ab96dce0f8f216f
8366 F20110318_AACGTH mikles_v_Page_086thm.jpg
0d184965bb642b8acbe9ea0e1a1eab6c
44ff94d6e64b03005d11328050a0a6192841879e
4766 F20110318_AACGST mikles_v_Page_049thm.jpg
44323ce6b1e72a93c1fae6f889029652
c642029782bae320310fd024bcbf5d87c1469ade
1398 F20110318_AACFQF mikles_v_Page_096.txt
3bf9eb38aacb736b79ba7d6a7bd7ccef
46b0cf56016bad531858f5b941debc6d6bee22a2
66622 F20110318_AACFPR mikles_v_Page_127.jp2
494a4894725110a337a51c337571ef2d
53a49e3438b9ad493b9917d3b4bf1ab3066648ec
7518 F20110318_AACGTI mikles_v_Page_087thm.jpg
fb2808fac999d036bd461972c662a83b
a656a93120c0f96b37ce73ad31b87dd2308b5649
4554 F20110318_AACGSU mikles_v_Page_050thm.jpg
23a78c9ff6a626ed0b849984443265c0
9c2d10c0955cc403f6066c533e74c357d02cdcd7
26431 F20110318_AACFQG mikles_v_Page_109.QC.jpg
4e098037f59419111e5e650db867f083
a5f84886a2138820bd968c3732e101bdbfe91d78
8091 F20110318_AACFPS mikles_v_Page_013thm.jpg
92e4b55475e06afe134790faaf186f18
ec7eeb4883be85c234f62c3b7f16b64712f12798
8031 F20110318_AACGTJ mikles_v_Page_088thm.jpg
6fdd124614cfdd34d7a193abb8db7c0e
f84f6c33d76b361853aa2ea0d4195ba26340ee74
4483 F20110318_AACGSV mikles_v_Page_051thm.jpg
91c01902664b0b7d2b373657e9b38093
d6a8c7ff3ec6cec0cabeae1c748e43575d0b3577
F20110318_AACFQH mikles_v_Page_111.tif
a4005e66e45359d307ba979a13519be1
80ba09dabc6a3b49c910f6d7879927dddab443b6
56534 F20110318_AACFPT mikles_v_Page_088.pro
cab5594244f6e6c9e56cf5a48865c7d8
fca0b02160122f97abc32bf0aac14e2134a8378f
7375 F20110318_AACGTK mikles_v_Page_090thm.jpg
03d087bdfbed3f1ac51617d73694cfa7
fbe3b95fb1724e69000265076966b28cf1ba00dd
8483 F20110318_AACGSW mikles_v_Page_060thm.jpg
c28040eed973ac215b01fed7a9ce0836
bfc27748b38ecf3144d77feb7c1b813dd7a10be0
17658 F20110318_AACFQI mikles_v_Page_077.QC.jpg
3f19310fbadccde32dfec75d9ed2b5af
086041db29e9f3019737bc426eb22dd4914395e8
55428 F20110318_AACFPU mikles_v_Page_160.pro
daa827f73c3af1bf226c0fead30e7ae9
2e88bdc034eb1c09191f73516e6a675d3fd43120
3619 F20110318_AACGTL mikles_v_Page_095thm.jpg
00614f8dc8771129c337e50553d6191b
5d00af217b7804672a21cc657f04abf0ccc67c40
8322 F20110318_AACGSX mikles_v_Page_063thm.jpg
69183014918fca7243243738ad7189df
a180c3a48b49b015fc93157ca7bb02b58140ffd3
60311 F20110318_AACFQJ mikles_v_Page_029.pro
1d3406141ca1195e75cc44f971f9ec52
2892179dddc9a0d733b5d577c276214403405f82
2184 F20110318_AACFPV mikles_v_Page_092.txt
47c446246fdeb32693ef71893e5187c4
81da37f3b8a95ddccfe346c7dc2b374c2bf1bf16
4053 F20110318_AACGUA mikles_v_Page_135thm.jpg
371276aa56a2505f723721e11a9c0028
c42d3c524220c003f34f02413bd435ac195281c2
6989 F20110318_AACGTM mikles_v_Page_097thm.jpg
83af5d780d7981e84c52049ea132df78
9c5008542827857394ef1f20ea6a9a2ec78fda39
8446 F20110318_AACGSY mikles_v_Page_064thm.jpg
f0a527eca4da9820fa0860ace56ff727
89cac02145b44022bc0bbc784e1e0fe854f1942d
18925 F20110318_AACFQK mikles_v_Page_075.QC.jpg
fb13d91b38d33849247085f960a47535
0afbb34fb8f925d24386dc803e7909f73c301f25
2013 F20110318_AACFPW mikles_v_Page_167.txt
9ba46ac797faf0e7a31a7b82c396afe4
271e332a2a08f243fdf3e8cffb441497d0ed0858
5165 F20110318_AACGUB mikles_v_Page_136thm.jpg
b926ace9804292e258d55d706ba89150
9868cf6dd15ed2430d14325e84b16d42d9cdc7db
2606 F20110318_AACGTN mikles_v_Page_098thm.jpg
8fcc1aa1c599d1a7f27c58f53290892b
4adb16d9d758e124ca71c4636cc80ca3eab20d21
7012 F20110318_AACGSZ mikles_v_Page_065thm.jpg
0f060120a4319d41e503f77f8ea09a9b
1888759424546ff1d65b0089a2431fc2a59c1331
62700 F20110318_AACFQL mikles_v_Page_133.jpg
2468446f87a5b773b73aa979c99b8569
d4a187a3b86a22efebcad3e899fa97a7a10a941f
80943 F20110318_AACFPX mikles_v_Page_136.jpg
bdd06ba7c2f6fdbf3ce0443427e52e8b
932447bc6fcbec9f0ab8db7e371facc40b3e976c
4128 F20110318_AACGUC mikles_v_Page_137thm.jpg
0234f6630cdddc71b76dc9fc0a11a3b3
ba86a536613b45c4b883e365a26929e9dc19a4b8
8584 F20110318_AACGTO mikles_v_Page_103thm.jpg
a1d227e335909ff2298b948cf4b462a7
02eebeed2a83ed9e75fbf283a36ca851a2fe4c7a
19984 F20110318_AACFRA mikles_v_Page_053.pro
ace5c1e908db92a99149a6e1bdd2173b
7ed4ef8a9af3d33b01f10d45fa560759d9550360
57271 F20110318_AACFQM mikles_v_Page_063.pro
f56c8880a7dbd2d9280cd5399df43a4c
d3d9e9feeec1790ea281f3a838b9535296bedf68
29968 F20110318_AACFPY mikles_v_Page_065.QC.jpg
469c57a0f0d0a0376ce3d0595ccac30b
141de8d08296b9945c47bc9f5786fb80d22de44c
2991 F20110318_AACGUD mikles_v_Page_138thm.jpg
84b88fec8064ebadae46d2ef02cd1c1b
3d9a317896ebcbf6c3d780d76dacea03163460fe
6362 F20110318_AACGTP mikles_v_Page_108thm.jpg
f265a9adb29514a519214eeee928d021
44da282f9869bff376c28380370f5e54be8f8126
101714 F20110318_AACFRB mikles_v_Page_010.jpg
1160cfff0392506d91a7012007a0e50b
17b155ca856950f7581ec2d5281f89a51f0762cd
9373 F20110318_AACFQN mikles_v_Page_098.QC.jpg
5e4aaaa044f35ca6db235a4f563808d8
03555b5fdf4de4fc713a62a5ba97a5c1a7c36acf
10581 F20110318_AACFPZ mikles_v_Page_074.pro
990feaaa111426888a6f3bfbf2cb6a7d
223148d1eca1feb2912f16940448f2ce98a4d5aa
8280 F20110318_AACGUE mikles_v_Page_140thm.jpg
7a0a5cfea8c3c096f55bb51dce5ee56b
2a08b3f1aba6be1ea6dcdb96497d5835b0662082
6862 F20110318_AACGTQ mikles_v_Page_109thm.jpg
024c7f04f67e32170e87c01bf3fb66b1
c9dc683b12797ee734fa913f1d1e77e08b635457
10163 F20110318_AACFRC mikles_v_Page_108.pro
5da2d619e3addefd5310fb1a087d6077
18abec1c686294edd76f90b071e7ea1eaecdb0fc
F20110318_AACFQO mikles_v_Page_119.tif
242a2d1efd4eeb698d3aa26fb53d3b29
d11b329b864d053deb4189e17e2624451121a4c6
8441 F20110318_AACGUF mikles_v_Page_145thm.jpg
d3fe468b3b9d01d7451485cdb49ffe68
8e86210da1565a9ceaee82335cba1c8e589b0f8a
8343 F20110318_AACGTR mikles_v_Page_114thm.jpg
56c329e0c36d90335812b1cb9596af7c
9c2a6b4cc7307802e7e7a94fbf7490816d25cd3a
F20110318_AACFRD mikles_v_Page_020.jp2
d77a0c16e31f4c4f0cda91254887a142
0ef4144f59e922fddae299828ea31b6ac4c776ae
F20110318_AACFQP mikles_v_Page_108.tif
2d5240887f1d04e80d503c0d5ec3485e
0207727ebe42f97ed74f2afd6fb52af5ebb46b20
8184 F20110318_AACGUG mikles_v_Page_148thm.jpg
2eb1a8a4341ef2d4728707b669e921ff
a30d03f7801f51b3cb1d19ce97977cab6cbad320
8199 F20110318_AACGTS mikles_v_Page_115thm.jpg
8806971fc8feb40b1422bdc7b71a7d54
646cfde113c9563b770cee5161c1da8edbe8081e
1818 F20110318_AACFRE mikles_v_Page_125.txt
8308b9f3add62c747d07beadb8ad8a56
6f3e07fe45b613b75ce377f70e608d8315b9d911
7939 F20110318_AACFQQ mikles_v_Page_119thm.jpg
93a96b700ace8cce42b38e42deeaa9dc
4d2b3edfb65ce26dd6160e97820c95d60f241131
2461 F20110318_AACGUH mikles_v_Page_151thm.jpg
00327306ca71303a0eca40df86680671
93074c694b2667d144558f65bf762d584db7664e
F20110318_AACGTT mikles_v_Page_120thm.jpg
d000d19782e22f6c894dd6fdc0e33a92
7e0409edbe49315d1c81a89a66cbbc3ef506cef4
1051887 F20110318_AACFRF mikles_v_Page_041.jp2
d7399fbc3e804c2ccf42ece438b67363
b8fd2653cb2e9bf286dd9e02b04aebd40afabdaf
7348 F20110318_AACFQR mikles_v_Page_009thm.jpg
648ff91640ca15af3a84aac2d246bd7b
f2f6847cf41be91fde31e1d1d8e8f4b8afa268ca
8115 F20110318_AACGUI mikles_v_Page_161thm.jpg
795d2af1f8a0a1cfb7d0c19f4f5e3c82
469f48159de4858b4010ca00d7161a85bec6cf7e
8572 F20110318_AACGTU mikles_v_Page_123thm.jpg
9e24ff0f6f3adcb2cbb46840287d27de
5420580d05ba0e5574a579c3be154353cd2c9b8f
27903 F20110318_AACFRG mikles_v_Page_149.pro
e4f6a20248c4926bf6ab2a510f595afa
54afa6b0ad40a1c07b0a7575d1dade2db6f275df
101777 F20110318_AACFQS mikles_v_Page_038.jpg
4ccc6d21838ced3f9c242b5986cb5787
44369e2122af35cb566cd402bd12bba304f638e1
8116 F20110318_AACGUJ mikles_v_Page_162thm.jpg
ecfc0c6b8e5bbd300bc1423950cc94f3
fbd33e0fd72687e2a562f8cd2f3f4cff1b19cb89
6831 F20110318_AACGTV mikles_v_Page_125thm.jpg
d7f4a5d051152207c964979cf705e6db
9a6ac41abe72fd243e3838a40c19fbdb25870675
962640 F20110318_AACFQT mikles_v_Page_072.jp2
1f7c281753c769ffc8ae0e4467977647
eaf3aec72537a5925fe97d6f44cc312c6661736b
60050 F20110318_AACFRH mikles_v_Page_031.pro
d39f1a7fbbdfd95f4c5e8b23563c08ab
bd3ed498141114ba3f98e3a27ecabbc87ee4b907
8208 F20110318_AACGUK mikles_v_Page_166thm.jpg
d92c45c83a1593df420036764b765177
3265f4c7ba24da57ea3f34ef493e6ec23cbf4041
7382 F20110318_AACGTW mikles_v_Page_128thm.jpg
7d7882421a3ec9ad44f90ab6446234f6
31a8064a94f295213058c113c78501c30612c908
54789 F20110318_AACFQU mikles_v_Page_026.jpg
6bad6675bd18d3ccb871a372910c2e36
4047a5c2aebcbaa7edf149a07d61b00504b7c8cd
65247 F20110318_AACFRI mikles_v_Page_131.jpg
4ce5a6f5f013dcac82f4279563c43d67
2ab004fa5e683e5f055b6048ca296771be6ce4a5
7505 F20110318_AACGUL mikles_v_Page_167thm.jpg
1edf726d0362d59d7d2683f089a489fa
081d0a059e9c1fb48cc01ecad4a79bd927b8da9b
5986 F20110318_AACGTX mikles_v_Page_129thm.jpg
b26b12fb26be4a7ced57c6b37f270646
22e33573af795cf6e2f67fce3f531d2f4b2fcff7
54147 F20110318_AACFQV mikles_v_Page_067.pro
4cfae00e7a872cf6dde01bfb9bd022b8
757e5547e9691d9db7cbe177e6af67fb831f80e0
1531 F20110318_AACFRJ mikles_v_Page_081.txt
ef62935c2eae15113cbf3be8cec0a7a3
704348a8b2ee4b92af0786485aebd493aa8973e7
7618 F20110318_AACGUM mikles_v_Page_168thm.jpg
d02d65f36e6312ea62680dc45171ff34
fe2eea00d16636a5e016cbccf8f8978aba7d1625
5965 F20110318_AACGTY mikles_v_Page_133thm.jpg
e32e36258a17f51c4c94d91f9201bc62
e3f199aebcc6bce7f14f4232450aa7fb354f0f3b
35979 F20110318_AACFQW mikles_v_Page_103.QC.jpg
1d2be62cdd58f37d2ec9c528d0c0eb95
5b4244704d1254a8c8dbc9bdeb99ff5344d7a2da
40009 F20110318_AACFRK mikles_v_Page_095.pro
9ac99c8bc23cde6ec5ceee7177da9a60
1fb9abff3acb4657b6db4f5a2f2cdf1bbe0773cc
799 F20110318_AACGUN mikles_v_Page_172thm.jpg
08d09500e70c248da2c24512aa7cecae
f9c5b0b4ab57f3d52e0ba663aa085e4ba2aad91e
6235 F20110318_AACGTZ mikles_v_Page_134thm.jpg
4679af27fe3755dd12b150d6fa0e809f
0c1725112dfcba4cbb29afd681e7bea2ca9af204
435 F20110318_AACFQX mikles_v_Page_151.txt
2e8766eb477057fe1a9ed90dd2824b8b
dcadd8ae36c4ea66d96ad550c7d8a5958920c302
7563 F20110318_AACFRL mikles_v_Page_019thm.jpg
fe3b9645c31f31340b6f5fe4ed9f34a8
a8a5e23a3ca4e220ace5eed938708cabbf4650b3
202687 F20110318_AACGUO UFE0022086_00001.mets FULL
52cba6916c03ab616fd68e84d8ab979e
b7f162bb129a5b1b4aa6efec5bb8444ebbb551b6
BROKEN_LINK
mikles_v_Page_002.tif
mikles_v_Page_003.tif
mikles_v_Page_004.tif
mikles_v_Page_005.tif
mikles_v_Page_006.tif
mikles_v_Page_007.tif
mikles_v_Page_008.tif
mikles_v_Page_009.tif
mikles_v_Page_010.tif
mikles_v_Page_011.tif
mikles_v_Page_012.tif
mikles_v_Page_013.tif
mikles_v_Page_014.tif
mikles_v_Page_015.tif
mikles_v_Page_016.tif
mikles_v_Page_019.tif
mikles_v_Page_020.tif
mikles_v_Page_021.tif
mikles_v_Page_022.tif
mikles_v_Page_023.tif
mikles_v_Page_024.tif
mikles_v_Page_025.tif
mikles_v_Page_026.tif
mikles_v_Page_027.tif
mikles_v_Page_028.tif
mikles_v_Page_029.tif
mikles_v_Page_030.tif
mikles_v_Page_031.tif
mikles_v_Page_032.tif
mikles_v_Page_033.tif
mikles_v_Page_034.tif
mikles_v_Page_035.tif
mikles_v_Page_036.tif
mikles_v_Page_037.tif
mikles_v_Page_038.tif
mikles_v_Page_039.tif
mikles_v_Page_040.tif
mikles_v_Page_041.tif
mikles_v_Page_042.tif
mikles_v_Page_043.tif
mikles_v_Page_044.tif
mikles_v_Page_045.tif
mikles_v_Page_046.tif
mikles_v_Page_047.tif
mikles_v_Page_048.tif
mikles_v_Page_049.tif
mikles_v_Page_050.tif
mikles_v_Page_051.tif
mikles_v_Page_052.tif
mikles_v_Page_053.tif
mikles_v_Page_054.tif
mikles_v_Page_055.tif
mikles_v_Page_056.tif
mikles_v_Page_057.tif
mikles_v_Page_058.tif
mikles_v_Page_059.tif
mikles_v_Page_060.tif
mikles_v_Page_061.tif
mikles_v_Page_062.tif
mikles_v_Page_063.tif
mikles_v_Page_064.tif
mikles_v_Page_065.tif
mikles_v_Page_066.tif
mikles_v_Page_067.tif
mikles_v_Page_068.tif
mikles_v_Page_069.tif
mikles_v_Page_070.tif
mikles_v_Page_071.tif
mikles_v_Page_072.tif
mikles_v_Page_073.tif
mikles_v_Page_074.tif
mikles_v_Page_075.tif
mikles_v_Page_076.tif
mikles_v_Page_077.tif
mikles_v_Page_078.tif
mikles_v_Page_079.tif
mikles_v_Page_080.tif
mikles_v_Page_081.tif
mikles_v_Page_082.tif
mikles_v_Page_083.tif
mikles_v_Page_084.tif
mikles_v_Page_085.tif
mikles_v_Page_086.tif
mikles_v_Page_087.tif
mikles_v_Page_088.tif
mikles_v_Page_089.tif
mikles_v_Page_090.tif
mikles_v_Page_091.tif
mikles_v_Page_092.tif
mikles_v_Page_093.tif
mikles_v_Page_094.tif
mikles_v_Page_095.tif
mikles_v_Page_096.tif
mikles_v_Page_097.tif
mikles_v_Page_098.tif
mikles_v_Page_099.tif
mikles_v_Page_100.tif
mikles_v_Page_101.tif
mikles_v_Page_102.tif
mikles_v_Page_103.tif
mikles_v_Page_104.tif
mikles_v_Page_105.tif
3548 F20110318_AACFQY mikles_v_Page_082thm.jpg
31806df6f8fcf269cb425af475237361
0cb281901fbb199a9b73ace4f2b1b4fce799e7cc
F20110318_AACFSA mikles_v_Page_144.tif
2596fd089653541fee8eb7dfaf1e2b53
a9c99c8790ebc5095877342d0bca7c43331c8263
637 F20110318_AACFRM mikles_v_Page_012.txt
74dc4e168fbb55760949c3eac524165b
d4e86c3bbe54fc501471bc222569e52368e6dfab
409 F20110318_AACFQZ mikles_v_Page_150.txt
de7590e752ba1b15e9af4b826473fe99
34d757eb9a815056e4be22af87d805b42f3f4b36
7181 F20110318_AACFSB mikles_v_Page_156thm.jpg
13802c0ab2e381c3f578e0af992f73e1
181cf44d9323f9adec040dcf564601e95a15071b
982857 F20110318_AACFRN mikles_v_Page_066.jp2
138e2c0a8f274cf6146a840040db5b9e
435f56e6e2f08c14d454bcd31490c45415582c32
F20110318_AACFSC mikles_v_Page_046.jp2
e1f533cd5bb71f658554d79f07b0bb2e
3d0141a611a6883b3df596ea10907a5e6e43636e
2282 F20110318_AACFRO mikles_v_Page_062.txt
6dd0725c20ef219434392bd6d89b7885
1366197ac5d12d0785b53fcdcce91d324f338cee
51624 F20110318_AACFSD mikles_v_Page_077.jp2
bf2df47669c38c81a437eb8b794a9e20
ae1f7de867abac207c3d4b535ccf1e798aa3a1ed
F20110318_AACFRP mikles_v_Page_015.jp2
f4809cd102c78d5d85d0fb5f1e21230b
2fb4a8bbf5940ac7eda14790f0ac642815583eed
2150 F20110318_AACFSE mikles_v_Page_173.txt
36bc75306c0101e649b07d1eb379eed1
aa68a51b86ce84964132fdd7806f0175c577a6da
F20110318_AACFRQ mikles_v_Page_152.tif
390711e6444ce631c3fab97ff80aab99
8881a60d357f1c50669bdf6ed4f6650099623be9
13689 F20110318_AACFSF mikles_v_Page_127.pro
2db8bb488a8f2940f0a4533060f5527d
1c961106fed12582f8763fa4e170da8e70e0b466
482 F20110318_AACFRR mikles_v_Page_107.txt
d4d649f69352fa1a84e43171908f438e
ae27588b1b7feae4ff213ee023314274600ff8b2
F20110318_AACFSG mikles_v_Page_030.jp2
4f515f4d8c6fbbada8d38376a3eb1be1
97a0142a2f48c4ae73ea4ed3373713d44d4087fd
56564 F20110318_AACFRS mikles_v_Page_146.pro
eb959db10c686d4338e4786743aaa6cf
ffda2487238b830beadf63afe7a495f55b4d7201
34740 F20110318_AACFSH mikles_v_Page_152.jpg
3d5c63499cbbe49ebe37dc5fe4dca3f4
131000c63d56258767a1c853ac9a4e5ab482e6c1
117522 F20110318_AACFRT mikles_v_Page_079.jp2
9ccff4ceda4a5f2adec44f1e4aa9e330
ed5e69564d85809b78e8123c834db8d30a78251a
31689 F20110318_AACFSI mikles_v_Page_171.QC.jpg
e2ac012c29a278ba7397b6789ed8fbf2
581b4c5c42e22b72470675942fe54ebec33a20b0
45877 F20110318_AACFRU mikles_v_Page_004.pro
0817bedd71bd1837488b88cbf559a8e8
edf9c2d501bf5b3cb5ec632853215c8b7d425d86
59856 F20110318_AACFSJ mikles_v_Page_041.pro
de9065bc3de6bc0ff328923f98fc61a2
555d8043de4778bd4be8395964760c7404ea692a
31873 F20110318_AACFRV mikles_v_Page_165.QC.jpg
4f7098968ed4ca3173fe2643bb7360df
069ba5867ffd9d0f266784be8450d8f0372e9e1c
2232 F20110318_AACFSK mikles_v_Page_059.txt
3350ca40fb4374c6c0c26f785e8d6e63
ad579b8ff8ae3d4f02a19546a3e9721735a24019
105967 F20110318_AACFRW mikles_v_Page_092.jpg
affb7a16f20b61406252e917b30d1f24
ef2d3bc5669bd7a9b8cfab0049e1bdea52e32330
F20110318_AACFSL mikles_v_Page_141.tif
5f0fd35dca795b3e04e8bb40fd9cc095
ba4e4eefaf41d6b5defd974e10de4d9c7a3feeb5
F20110318_AACFRX mikles_v_Page_075.txt
69dc1a91d3983863e98e1471188b22c5
459ca5215a4f7ce06122b7274b74b576d21cc81a
58154 F20110318_AACFTA mikles_v_Page_162.pro
e8c5ca06f6626af696aaae93d2ac22db
a8ea1dc9c94c7498a6f8d9fffa938b360959522a
35797 F20110318_AACFSM mikles_v_Page_161.QC.jpg
9cbb14a0f79d96da097d7d22c8020e5b
c9501a32aa411a17dfe5fc54ec8a4d865206f6d7
117963 F20110318_AACFRY mikles_v_Page_068.jpg
b570fab8849c7e3e63ce89b1f727ddfc
6b9b9479219104703ab354e78a2c6618c68f27ee
28362 F20110318_AACFTB mikles_v_Page_128.QC.jpg
04a329463c05cdf4fc1227348982dc88
6d3de64687b8bba2c9af910990c1af089c20ca11
23053 F20110318_AACFSN mikles_v_Page_130.QC.jpg
4d0e0318af39d9136117fbb28cb0476e
d48d9ef88932fb64786026911b3f1c0ca5825ca4
18255 F20110318_AACFRZ mikles_v_Page_111.QC.jpg
d4617b498465ceecfd82264d4cc92251
07614d117cd44485cf6a20e4fc3a2262da5d2c51
730030 F20110318_AACFTC mikles_v_Page_054.jp2
7da34314ebc4085f8a1785e41a89c074
fa2224bdcbbb97a5d286af98fc8b245deee1f2fe
1195 F20110318_AACFSO mikles_v_Page_129.txt
9120a410486f0b7891bdda5a20195e53
a60b3a7c55511c3eb84d9f9a6753820c264e80a6
F20110318_AACFTD mikles_v_Page_122thm.jpg
5718d81742c4e2b2a123f4a4cba6d7e4
390d07283e284e6387e8d2f27699eb913a1d13c0
38187 F20110318_AACFSP mikles_v_Page_144.QC.jpg
9617dc7ad961f10b7c72c41d46967393
0f85fc3bb17ff53b344ebeb84b76bfcf116544ec
2191 F20110318_AACFTE mikles_v_Page_124.txt
f883f31067e6a2ac9301e2d51ef68b84
881ca59f4df819bbf4e90fdfc6377fddea1b8a7f
64431 F20110318_AACFSQ mikles_v_Page_111.jpg
52711bad2c641bd52e49fc5e74afa593
9f599e38fd951188e9ac87475b4f645667083a96
54626 F20110318_AACFTF mikles_v_Page_093.pro
fccba59be37752ac12052c358cd1fd4c
a15dec01815840af25f6a4452b9fb039c27a6ef8
16200 F20110318_AACFSR mikles_v_Page_027.QC.jpg
8ddf2825bb248c1d55a557013d37e4de
9a144111bf3e182b84197dc5be776930d37dcfa6
49509 F20110318_AACFTG mikles_v_Page_025.jpg
6bc0575bc0da951a90e4c8642849afc8
6ed93208d64eb09f4e27398627c4006ac651f20d
7959 F20110318_AACFSS mikles_v_Page_139thm.jpg
2f6357bb1bc3a4d367a3da30cc94e3ff
e2555c62f5e3073f8c00110b8ee6cca7ad4e3d8e
41320 F20110318_AACFTH mikles_v_Page_050.jpg
c2627acde4359ac3836db01d8e59ea62
5fb30796f25b3788b3f8cd4920fcc9a5afdd0a1d
87314 F20110318_AACFST mikles_v_Page_073.jpg
6dbee8a55086cad6d30fb347eb163d16
57bcdd61a1be45e0af1137ca0af1d7f7b1598d0e
2331 F20110318_AACFTI mikles_v_Page_146.txt
3b81b242f0ce91e233912a49b696d574
0de292f25e6a71e2235c5d3f48653a30f57d6fa3
1051964 F20110318_AACFSU mikles_v_Page_093.jp2
af45176375cfb915a779dffdb7ed6eb9
d4557d1b7bc722ac647bbc15406ec272bfd8df41
36268 F20110318_AACFTJ mikles_v_Page_115.QC.jpg
d00e453206570556445b8cf6dc4ca7ea
eb1030c6e65ac664bcb045746ac830e40c5c2927
93872 F20110318_AACFSV mikles_v_Page_104.jpg
8523dc5a696b2de9362959ebb4a90c27
6ff0776d720276f6e281284e6e43e93b98efc95f
2154 F20110318_AACFTK mikles_v_Page_047.txt
81705bd6a07a18ac3b3d68e2a27e0f57
9c3e0ade231661b82132b27d7d56e5f5cb503c50
3512 F20110318_AACFSW mikles_v_Page_159thm.jpg
8634a191ad783329c45f881ceca902b4
7d8554edcb852b3e3ea34fd56bed883eb4372bb4
1051978 F20110318_AACFTL mikles_v_Page_060.jp2
1673bb754ba354e9316719d4ddc7ae40
cc1f0a5bfe1d487b604a979d3f30cca20ac39eb3
15855 F20110318_AACFSX mikles_v_Page_025.QC.jpg
c9e060a2f66b1ffd049c8092936f37b0
1d95340eea938d959b85c17139dbc6c6afec0892
2352 F20110318_AACFTM mikles_v_Page_044.txt
f59d3395e58d471184679e9e5c738eb8
27dbcd1e16b2fea099975e15228bdf8d6b2dca20
F20110318_AACFSY mikles_v_Page_132.tif
c92b0bf40b27cde176af7b0b87d5f76b
ff7ee2f9ee64ea51d449aaecb14f771ac22955f7
49965 F20110318_AACFUA mikles_v_Page_083.jpg
540359943939e93f04552b1ac12693ea
738e19285cf342a19acf282aecba0fa78bc9101f
117264 F20110318_AACFTN mikles_v_Page_030.jpg
6639f35947d6a8c5ba2369262d6c0c90
3ee3e6096cc8083a4eb22fb9147995e3b77e868d
530 F20110318_AACFSZ mikles_v_Page_002thm.jpg
59b28edca4d47406079769924cac53ea
7b77c5050d3b008dbfa53b88875d6f90a96e5fb2
F20110318_AACFUB mikles_v_Page_155.tif
2ba52642ede808c9a8307a09b0c9607d
635ba630d3cf6802dffa2451dbfc2f62beb7cba5
F20110318_AACFTO mikles_v_Page_107.tif
3d48f2fbd4970ce9f0c78df71f5122cf
fe6cfbb2967b0e725a38277b40c2df20a941bad6
51806 F20110318_AACFUC mikles_v_Page_169.pro
8be6c1fc2b7eddaeaf8b6da56156b65a
1c923b124303b4865f1dc4929b737fb33a744d96
20484 F20110318_AACFTP mikles_v_Page_083.pro
cac4d480dfb83f4f7378306fc94e9d1d
4e955b0487fb6dcc9449f1c194987c05d7cff968
1930 F20110318_AACFUD mikles_v_Page_091.txt
60a6a06df763b0645b135dbbf85968d4
b88750e8a305c8bbaa61de32eb98877bd83cce5c
28008 F20110318_AACFTQ mikles_v_Page_091.QC.jpg
c2e36b49e884d2a7277c099e5535a9fa
3ae42afe146036a81cd30bab925291b25be6cbbf
113522 F20110318_AACFUE mikles_v_Page_161.jpg
632a1f49e7d8d49c9c4a7cdd68c6edee
102a69df806b916827e8021dd9c3deb1b3c92a36
36051 F20110318_AACFTR mikles_v_Page_070.QC.jpg
2d8f1f4797d4ef89bc83cfbbd0a77ddb
9659160ed8285b03309783ce2260cee6e53ad950
60907 F20110318_AACGAA mikles_v_Page_145.pro
3dc7d5416a8e289e5e7d41c34b5da845
3f35c373e146a6e0d62051d53b056acecaa987e9
F20110318_AACFUF mikles_v_Page_156.tif
53b6553b21e374c8b7c71b3377c35942
d24e98a4c0ba8eafe71294178f7a6337644cf121
106571 F20110318_AACFTS mikles_v_Page_119.jpg
a939c9f5e0845bbf686fdbcad23635a5
0ff945616a16f1f6c8d3e8bf2bfc837e719da537
29186 F20110318_AACGAB mikles_v_Page_034.pro
1bfd800f0b0898078fbfd629b3faaecb
a60231c0bf6581ca588f1093aa8802ac90dcada8
319136 F20110318_AACFUG mikles_v_Page_055.jp2
57ae2a163ef6519417598587f50110e2
e60e795347cac8bd6927cfcbd80f8f801c2bd75c
104656 F20110318_AACFTT mikles_v_Page_169.jp2
6ad61030b16e2366e856ec5de717e53e
8010c3b0a77c07cf0c161dd786dd75d31afc0c20
14006 F20110318_AACGAC mikles_v_Page_051.QC.jpg
ba19e3e6e1e65d66293102c2a7272aa8
668ed260aff489eb4f16784b8a2e277219df21db
59123 F20110318_AACFUH mikles_v_Page_161.pro
6e48f4f77abb801d233f32db3a163b63
fdea2a671a20549d357d812f7b9e30d9a800fa85
37462 F20110318_AACFTU mikles_v_Page_101.QC.jpg
34f188cf20e53619102c78e8ee0d2b66
144cd438b386a6e628c1d4d046f11edcbd34240b
5029 F20110318_AACGAD mikles_v_Page_037thm.jpg
1b72f9a8390e9f313a3a019d757161a4
14dd0fa92c6287a2a35c6fa93ee109871194c598
F20110318_AACFUI mikles_v_Page_084.jp2
c6143d68af7a367e1cdc586980a5e070
206bf7eaefb9f5851ab372aee44f5c265a9911e5
40237 F20110318_AACFTV mikles_v_Page_105.jpg
217cbd543b87bbead2005dfad67c8ccb
847ef50dcd2e6e9d9606987ff475c1b20b52e442
2367 F20110318_AACGAE mikles_v_Page_029.txt
e10a5ddd98d1151ddfef6124109891e7
fe260198548127e8b9847b4973a0f53932b45e1e
35076 F20110318_AACFUJ mikles_v_Page_052.jp2
47b5dd632b887590603cf97dcfb7ff30
751d83a63817c532fc2c0e92f1bcd71576223744
F20110318_AACFTW mikles_v_Page_173.tif
facca6c33ebf1bb687396ce5745b0b45
0129201551d1bd27751c37225615b983756d16a1
57044 F20110318_AACGAF mikles_v_Page_062.pro
834fe3912244a115807f96d2daec7507
27b934f30cf995edaa78c7f99879e1777bd7ddd7
5789 F20110318_AACFUK mikles_v_Page_127thm.jpg
b0e615b5944a3f386a597834cac8815b
27c0495e680eca2495ba9f3c07886555f1247a1a
52150 F20110318_AACFTX mikles_v_Page_019.pro
b29714a1e65f2cd901da2580ca56c6d7
6cfef8ccbcc5659be0cd94f9f1f213ea09e3622b
8055 F20110318_AACGAG mikles_v_Page_169thm.jpg
ff8ebdea31f4ab8f0527c6e9a24d2abf
b83ac23a27c1e666de7b7ef9e6ab741119af772b
33908 F20110318_AACFVA mikles_v_Page_092.QC.jpg
763e85ffe7874c9c3d203061a9fef7d4
e1a93b8fa0303fef04484536995238c1dafed928
57373 F20110318_AACFUL mikles_v_Page_102.pro
80e98372959d082fd523156393f6b754
5f92d657805bf946f31d52a43b5caa0c83c06fdd
117090 F20110318_AACFTY mikles_v_Page_043.jpg
1cd53bae51099b6d9b406353ac1fe929
5e383842fc90b215d44c7a7b17ec939d39814e5d
34561 F20110318_AACGAH mikles_v_Page_146.QC.jpg
b42faf8ff427f9ed01b11fd4e6a97f38
2beac310893fe7690ed2d567a0a7e75a712010c6
3036 F20110318_AACFUM mikles_v_Page_164thm.jpg
fb112910c549a9a388f2814705454f40
e1a114076d77ff3ba90948e2338dd4889797b7f3
38799 F20110318_AACFTZ mikles_v_Page_046.QC.jpg
4f7fe5e994fba5f16be85af0485642ce
480e3a86ad2c6f281ea5b4436952a8ce618c9c93
34235 F20110318_AACGAI mikles_v_Page_126.pro
28eaaa98092b490bfcd24861ecc13aa4
82b4555e9f3300d74deb7bdfeacb7ce93322e490
111545 F20110318_AACFVB mikles_v_Page_088.jpg
9af92ba1b39a5bf7ee8fd5f087605ff3
ff92d51eab002ae20188a3f0be01549030f9e685
1051968 F20110318_AACFUN mikles_v_Page_121.jp2
440370d22417d4bd4985896c29ccadf1
d3df0a83a8f63975fc437363370016931847243b
50660 F20110318_AACGAJ mikles_v_Page_167.pro
1f12c219779eeecb66c4fc758da67428
b5a69be1bcd01d7e9efe89a99e12b6eef2af1e2f
2586 F20110318_AACFVC mikles_v_Page_154thm.jpg
5174a5419580d85e6bf7f803139e8f5b
f83b7d59abc37d17a15cb77e9e72bc0958eaa8da
68446 F20110318_AACFUO mikles_v_Page_127.jpg
e3fb008637c4d6a7509bda255ddf9d62
a1e2d264955c992271f1c1851cd0c51c81bfd816
2269 F20110318_AACGAK mikles_v_Page_060.txt
427983fb1b5935240915d25b8716f4da
c15770137178503d606c663709b59197bde040a3
12487 F20110318_AACFVD mikles_v_Page_106.QC.jpg
9a019fd6d5a41fde8865c22dcd42c4ba
57802edc6cbd74a1ea868cc37dffac8fded6ba12
106277 F20110318_AACFUP mikles_v_Page_087.jpg
1cb92911cdb9d2a9e806b7d581a37524
6002e23415630619b04cce11d06a58cf35bd2152
108235 F20110318_AACFVE mikles_v_Page_011.jp2
6fcebc6a4297ed007f1455d004c58ef2
25ef56fd63e9dd26b5e212304de52f9290bc5eda
F20110318_AACFUQ mikles_v_Page_143.tif
2148e94a2a886b6ba7956f1c2e739c20
c5128e05e47d3d7e31f27dea4136840b931fb63f
9156 F20110318_AACGAL mikles_v_Page_036.pro
0c18f7b1f00e31a98f48da1b79213772
d8c64c463389f584201d29aa7ee7c4ffa3b9e7dc
1100 F20110318_AACFVF mikles_v_Page_159.txt
a7724cba6b883fb1118520527f81a871
37833659ea2033a82489a2d6a08f725123f68f64
13009 F20110318_AACFUR mikles_v_Page_082.QC.jpg
04f0982021934e9c089a502d4418fc3e
f90b8563643fec5c579c5e2284ab291ff33836ea
14047 F20110318_AACGBA mikles_v_Page_055.pro
4af369572b4316f75cc743b25da5f2ac
cd617a00a06f6b889e177809e5134711c308a33f
31952 F20110318_AACGAM mikles_v_Page_036.QC.jpg
c0797f5814398c84afbee2be06264188
342bd4a4eada16592fd3d1d01320aa3e5087c1f2
13435 F20110318_AACFVG mikles_v_Page_095.QC.jpg
7c13a572d476821a903f24b1659fbd3f
4e6aa7e0fc72a0124d772296e408fab53f3871e3
13255 F20110318_AACFUS mikles_v_Page_133.pro
641a6cd3017449c61c77e715190cf7e0
f38334c9c2a7ae66ff59531994bc311d4fe77377
4924 F20110318_AACGBB mikles_v_Page_025thm.jpg
b5b213cda02b013989990e98ce502800
fff4207aabe97470e2a2dbf17aa262493bb568a0
105883 F20110318_AACGAN mikles_v_Page_019.jpg
7922aa8a940d7fbcd81c9fb9f566d415
a84ffeaed48cb76276ebfde49959c466706cafe4
24459 F20110318_AACFVH mikles_v_Page_081.QC.jpg
c0effa2c2005d3e505765be403cd1c96
d18a650389993e6e8f803a1bc52bb070a86bc61c
8423998 F20110318_AACFUT mikles_v_Page_109.tif
003fe6e02b5aa5264b9a003c7155cb37
4f0d2ff685cb279d7ef7e151f74ffa2543589613
18070 F20110318_AACGBC mikles_v_Page_149.QC.jpg
707f998faa7d011622a77ad65ce040e3
c34d0b4dc699373d15056a99d339e80fc6efbf90
10244 F20110318_AACGAO mikles_v_Page_153.QC.jpg
718a71e0180f490203c13d1a6ddbec5f
b5b8b64d7adf5236c0b752ace962b873bbd05eac
116617 F20110318_AACFVI mikles_v_Page_102.jpg
8aefdb8eb7f5ce9eb76f4f0fa0fe7c42
3e095f5472cf0c38b65ae04d3fee308dfe68a717
653 F20110318_AACFUU mikles_v_Page_035.txt
2380b930e9348a821c012c0c5f191822
64d3f86d9ebe94e6b50458d7617ffd470568df56
13523 F20110318_AACGBD mikles_v_Page_024.pro
b1b15206a8e9ae5bfa99412edab6196f
c8f03fe4adc01b981e0955743e1bd87634ea6dc1
111211 F20110318_AACGAP mikles_v_Page_162.jpg
f9f538984eac3b3b78d95f419f57d67a
8958e00eeb858519b8efc82c02669b2ef7845adb
6793 F20110318_AACFVJ mikles_v_Page_010thm.jpg
2f57fe07c59ddcf0b11de4790ea62cd8
0ad1d6f0aea3a0af78dffcd8c6f21d306f2a0413
14807 F20110318_AACFUV mikles_v_Page_105.QC.jpg
9f71fbbb1058598f89ad583192d8cd0e
f917cd66fc049bb712d74e883acb76e314317239
2598 F20110318_AACGBE mikles_v_Page_005.txt
852d4f8005962f5bd9e1c0995641bd9b
a28af598df5c8435b38e3be255163b4185ecba56
8174 F20110318_AACGAQ mikles_v_Page_102thm.jpg
cf915276ef952ea32e979a24af1be047
26db3b2084d65e2cae34b4b7257536ba99e3c251
30689 F20110318_AACFVK mikles_v_Page_168.QC.jpg
7a7e80a93ce47709b88d37969c8c5672
0711076905633a1bf29c682a435958f4e9e6b849
8502 F20110318_AACFUW mikles_v_Page_118thm.jpg
fab31ffebc041f9f958c1a263462b26a
cb9b996de9396abd234de3bd9159bc5189e4c211
7930 F20110318_AACGBF mikles_v_Page_160thm.jpg
b6dffbf250eb33e07eea042d9a7d9b82
6dbfd9562ac868018cc9404df0091c46797d11a8
23822 F20110318_AACFWA mikles_v_Page_137.pro
76468253a3a72bb21fc65da906094408
856c113732ff8051ff4533bde4c31281409acf85
2776 F20110318_AACGAR mikles_v_Page_153thm.jpg
ebaf0cecb5bab468e09ef7a4d95af896
4371734c6d474efc7248dc2e393b63ed45a6bd26
8151 F20110318_AACFVL mikles_v_Page_039thm.jpg
5c3a49e2108c375f72037ec8a045dcd3
fe78b95ee792e827b50f2b572bf0f80b1299acf1
11157 F20110318_AACFUX mikles_v_Page_153.pro
b075fc6290eb2b75754504931a0e9015
3b39e4f14335b1d33b867ecec4971d52dff585c2
36328 F20110318_AACGBG mikles_v_Page_020.QC.jpg
e8c5763dd1b0a24e1fbf0e67f1333288
37bfce861fc88f684a15ec178f1f2910422cf09e
38128 F20110318_AACFWB mikles_v_Page_044.QC.jpg
4facbfc48f7e262078c97ddc354cd8fe
7048dc5e0add5fa274b94d832d5f9c50f3506066
648 F20110318_AACGAS mikles_v_Page_152.txt
45ac2a787b86e0c92cd49b8376e54bf2
a3987ea1d42506ce64d4c876fbc2c6e13d2343ec
23856 F20110318_AACFVM mikles_v_Page_004.QC.jpg
a78172e72d846ea7f0205924c195705d
bee3224b87b6e720ba29cf0b6401a7a1d8af99d5
2322 F20110318_AACFUY mikles_v_Page_028.txt
4a2238000d0255df427ed79cbabc1d30
a08338f9a5229c87e6e5764c68775af718060a38
7875 F20110318_AACGBH mikles_v_Page_059thm.jpg
aad3d82d6ef771ab4ea01c575127e4a1
28afec46262eb1238a5397ed7859fc256e3370f5
7961 F20110318_AACGAT mikles_v_Page_036thm.jpg
b5e0cb814d7352b27b35b51557a1b112
42bd06f191f4b8cd4bf4a88dd5aff18af6c81607
7898 F20110318_AACFVN mikles_v_Page_142thm.jpg
f7036cf11c690127c57eddea9faeaa91
4301ddfcd8e0e0738ec119869f25ffdb40a60956
18477 F20110318_AACFUZ mikles_v_Page_080.QC.jpg
b63a1c9ab102eadfbc3ec6605857311c
816a02dda701822f7cba7a064d95aadfc4a92257
11794 F20110318_AACGBI mikles_v_Page_138.QC.jpg
b9b5b390345ffa7b1c5d7b0b600a1ef4
5a137a35533da687383483a1084569d1d59220a8
8123 F20110318_AACFWC mikles_v_Page_093thm.jpg
0ae803021aababca6e5549b6959f62c9
f28c237acbed5f4026adb0eae4209f05eb6a763c
2215 F20110318_AACGAU mikles_v_Page_058.txt
8970db62f073e87fa84f425df1048a0a
0188d75dc2f6deeeda9d420a5a0f0af222c871f2
35759 F20110318_AACFVO mikles_v_Page_048.jp2
3fdf132a7ac1c56c63d5a9ac0b6753dc
b1866ab749d5c11e2e3cee7f4927d9fec9d7d9df
35716 F20110318_AACGBJ mikles_v_Page_155.jp2
d45649d39e7a8514791eb9ee8c36b4cf
e09398f8e48cd350e1a5180445bd7bcc7a59ba12
5361 F20110318_AACFWD mikles_v_Page_024thm.jpg
c19aecd32dbf82e74b366f3bd5103799
b41e6a1c869691d490f6cbc34f7a4c21c80532c8
43905 F20110318_AACGAV mikles_v_Page_125.pro
9ee5eefadd61cd42cf039109b151c229
63d7a6491ce241f4b667be2eeefc831fa0784d83
60359 F20110318_AACFVP mikles_v_Page_061.pro
7a182d2e97ed623bf6397275b381f179
f229ae1aff0cc866632551b8ff85a9b9c6df7bda
2313 F20110318_AACGBK mikles_v_Page_162.txt
56d7343330cd150c3973225d2206c39b
6a9de7e97e293d48b4cb2103a549b09a75b76e47
40486 F20110318_AACGAW mikles_v_Page_049.jpg
978bb7a87a86a7bfe4a7aad9b8086319
d8a3c261d6963c2333b3028dd8a73749dbc7a015
8326 F20110318_AACFVQ mikles_v_Page_106.pro
738b069795d8403cee48bf722c6cfbd9
7e3892be6f397a729d4c99d3f66c538414caa80d
57315 F20110318_AACGBL mikles_v_Page_040.pro
84439fc2dbf27bbbb6a22ae763e45132
9afb0f618039eb9da0b2f71df6f77ffb8d5360b2
38464 F20110318_AACFWE mikles_v_Page_154.jp2
28c5c5c1924ee893f1f11161cb844353
319440e996c56a42e54a1bc83e1108f993bb1d63
126560 F20110318_AACGAX mikles_v_Page_031.jp2
3fcb5d40421fd7d5398cc3dbe98be590
9bc56c74c364a8f15c357cd887a243a7af0a8b2b
9612 F20110318_AACFVR mikles_v_Page_052.pro
f83cf29f2bf122ece48acac8e24813b6
f368234ca1cc89dca19ad9b57ada8b6a14c46cc8
4667 F20110318_AACGCA mikles_v_Page_080thm.jpg
d832f69000f4eb87cd4a17ca3cf73dbb
1e5bf4025e66fd4b69c6fdab66057132c893aadd
2366 F20110318_AACFWF mikles_v_Page_114.txt
bf42bdc3e5c7b1fc3687a204907cec2a
aa6a6d695495b37d87cd653b6c5aad3d4467913d
43374 F20110318_AACGAY mikles_v_Page_138.jp2
958f016a4beb1663b1746ab59d9381ad
a9b5098aa2becfa842e50eccc266fe9667399e18
46373 F20110318_AACFVS mikles_v_Page_097.pro
e48cc9e841f0eddfe890c1596a44f901
2a188671a276215a669131e84306e3897ce5ab98
7529 F20110318_AACGCB mikles_v_Page_170thm.jpg
e53c4f7b43117b363d09b644b124547a
8c9b5bdaf2bc45a7cdfcc70a98fbca94b8e7145f
F20110318_AACGBM mikles_v_Page_119.jp2
21807d3c599322d5f5b3ad3c763bd1a6
aaaa74e8d0d1980285f10dd4c5028722c605eef1
45816 F20110318_AACFWG mikles_v_Page_104.pro
b9bd1dfd608e775e98749afd28a62aaf
f499370bfd2504ee2354483b7f20d502cc757030
412 F20110318_AACGAZ mikles_v_Page_003.txt
8ebc517276968e6844818316f945b96e
2eddf64cefe229dfa422ed26692244a6d0d38154
F20110318_AACFVT mikles_v_Page_013.jp2
c8ceca1b9c92645f1597344f99552227
1d476e0179a0aa93422d449e30472cf21b0c5a9b
F20110318_AACGCC mikles_v_Page_112.tif
1cc68d1115a54087d91eba32f35c29e6
4f81da10bec512a3bcf7b8a2e839f03d11b8fccb
36699 F20110318_AACGBN mikles_v_Page_145.QC.jpg
58a3b2f32922bbbbd8adaa4181146b7d
78bdd743621b9f317c04a51199d80a6654279708
668064 F20110318_AACFWH mikles_v_Page_080.jp2
49ff9848da422869481eacf577befdb9
18de18a04f6a9c4f892c9595eafe66674135e7d7
48415 F20110318_AACFVU mikles_v_Page_038.pro
d4ce7855aae00fea4f116a5d9e426690
de91e9bf2aeee831b5a1fd5ca0fe8f51f77852f1
F20110318_AACGCD mikles_v_Page_001.tif
b4bfc4e06e4994aae3bf6084efd032ed
0b0fa02952828ee242099e779fc3ef903234d13e
F20110318_AACGBO mikles_v_Page_043.txt
baa61affbac4d3e289f89c9228ad5d8a
d2649bef4f7e03178cdb589b635107bb6c2f171c
61567 F20110318_AACFWI mikles_v_Page_113.pro
7c50e7e31cffd3b48a4c1720c5ef3173
d0e10678e4d3054db05af2a42016c8a4d91b6df8
2074 F20110318_AACFVV mikles_v_Page_087.txt
8e4e023191b80778e29b62d7a1a42b65
04595934cbb2c66f1a4e38a168d96f4e0f5788cc
60368 F20110318_AACGCE mikles_v_Page_021.pro
01bf59d2e04f74aea57e4d411748c704
cf980a186c25b481bff419661f3b9c9e1b02d84c
1102 F20110318_AACGBP mikles_v_Page_137.txt
4c4d3a0b4e16854877f27ad7d68a588e
8d4de418a8a2966e7615b56af3538e89d52c8711
60071 F20110318_AACFWJ mikles_v_Page_005.pro
649e4106add9f6c72a2a6970da1eed42
5127dee72434785fc4817eb7abfb4bae8b12b7d8
1158 F20110318_AACFVW mikles_v_Page_034.txt
ae30676f542144e1c4469c90a1ef0e96
7bb0235d5996f8154ef0a4dc0ee7256e656938ef
8618 F20110318_AACGCF mikles_v_Page_144thm.jpg
796a9125250d73c8bbff35a4f5abe0aa
fe23d33f61428858f810ad1d405cce655cf29789
27461 F20110318_AACGBQ mikles_v_Page_080.pro
28fe202c3ced0d77417d679cccedfca8
042abceb015890be3decec895f9aab9cc1638015
58106 F20110318_AACFWK mikles_v_Page_149.jpg
775b958872ac65346f49d70d2bc3d365
fcefa7dae1ef0d02b71c8594970ee32a75dcd3c5
8190 F20110318_AACFVX mikles_v_Page_124thm.jpg
2e2a226ad83232719e73cbc478e3ac89
b74fdcc8f280f2204b219204d735e14470d4dd96
110383 F20110318_AACGCG mikles_v_Page_058.jpg
9233146bb1eb1445537fd8f7d40d15ce
598feee49e5bbfd3a6dfdca407cfcecf1e9ecde3
8537 F20110318_AACFXA mikles_v_Page_062thm.jpg
a0e6183ec6f6000bacee725d9378402f
8b7342543e9fcba2d4c1967b53c2c75ae18a2abe
1837 F20110318_AACGBR mikles_v_Page_095.txt
b3a646a600e6b001aa82a11a2432df9d
d10afedde91d9bcd689f3773bd3b760a22cef349
F20110318_AACFWL mikles_v_Page_172.tif
4e6e3b528e9770b28845bc3882a93481
713ed5805eca758f74e89fb583ffcd71fe37b273
118130 F20110318_AACFVY mikles_v_Page_140.jpg
ffe6f006ac967ae3971b5f90db46bd0d
ac90ceb397c021a3523d650f9d5be9636ef26411
F20110318_AACGCH mikles_v_Page_114.jp2
5540565e4ebf640799a15805475a4c40
a715a213a4808fc7b617bd8421088910724b2500
30775 F20110318_AACFXB mikles_v_Page_050.jp2
1ba4f81f37488d3b67eeccc282efbd48
85367de5f5cbee1e2f628772ca5d4b51ffd90de6
517634 F20110318_AACGBS mikles_v_Page_083.jp2
6ec32d8c8249a4c330d5f5bb6c0020c4
8b03ee648381dfed7e39033d3f77f8866d6af457
35832 F20110318_AACFWM mikles_v_Page_063.QC.jpg
f218f19bd94cb309b7a9fbb128b4e141
5e1e4ca3a789d71f730afc9a4a0714f76b57c3a7
767481 F20110318_AACFVZ mikles_v_Page_035.jp2
ef626dbc8d5fccace89ce88914c0d78c
4931044328a50941859988de65a109ce58d5968a
2276 F20110318_AACGCI mikles_v_Page_017.txt
5bb221996976b2d619e1d08b09dc18fc
ef5be7c01c8e34d100eb4667735d201acf37d670
88494 F20110318_AACFXC mikles_v_Page_091.jpg
844f6f3e78cfa2506e7d1c2afbf5ea69
2dff583d53bb8739e0edb852d492987d119b6a91
578303 F20110318_AACGBT mikles_v_Page_157.jp2
c631635cfd6c7a5edd610b1733153669
01efbefdfd6111f5ffab2877ff64802c89efc3ea
99902 F20110318_AACFWN mikles_v_Page_089.jpg
d37d7870780dcc95cc7ce2a2a30a7606
92729f034a90c426783a6684af867de3edf3de65
8364 F20110318_AACGCJ mikles_v_Page_017thm.jpg
66fe5bb6cf0c801b10c82b600f0cd1a4
eda9245cdc05388d0fe8a70391a7528f2b905451
672 F20110318_AACGBU mikles_v_Page_133.txt
529950671b4a00022b064f5eff86a4e9
3f03e0fad27f7ec97b3b0e5d97bd2a7cbd422b03
119594 F20110318_AACFWO mikles_v_Page_041.jpg
1047ea7c42a9b89493b6a49ed6115715
5cc0836a7b2c9b9b0b6fb9f9259c2cda4b34811d
18874 F20110318_AACGCK mikles_v_Page_037.QC.jpg
1033065c96e71fb5f3c7285dac0e2268
a4d7c6e5d8936d432c3090f23ae342398de3fe67
F20110318_AACFXD mikles_v_Page_135.tif
bc4c14a6493af8f31e4731b87680a864
b2f56663a8b2967e8d0576bc3f73150d776e3ca5
35074 F20110318_AACGBV mikles_v_Page_018.QC.jpg
5b90d204c1105f0e82cdc44961282a54
54b5dadbcb7489406638acf5d20862396d0374db
116576 F20110318_AACFWP mikles_v_Page_064.jpg
bfff9633d84b85c5a5f205d89371e549
213ccc08b61bb181a6874b385fa071fd5fd2d01c
18786 F20110318_AACGCL mikles_v_Page_138.pro
3841e4f52cb7d4f82fadc6e2cf095413
2457db5b350bfaacffdc037063f1f8c349b8970f
20252 F20110318_AACFXE mikles_v_Page_052.QC.jpg
e4d83060b87334d0f077999a9659695f
ab0a720f2739697a4667720399895a65b8401fc9
1051929 F20110318_AACGBW mikles_v_Page_120.jp2
425d39ab5e2e01c8da039e5cd012e354
020393bf35da3fe0e4a78bc1d3a61c797f094372
F20110318_AACFWQ mikles_v_Page_168.tif
223d8ae1ba8ee2607b57a6bb7df583b5
cd2b7c1f46a034c7d6eaaadd5f5ae1bc16f70ffc
1064 F20110318_AACGDA mikles_v_Page_083.txt
e68b864d78a582edda77da4e00e492c0
37a8cf112617680732093110cab778a0f825ad65
29092 F20110318_AACGCM mikles_v_Page_104.QC.jpg
ed3ed608cf7097d64d02a7a55b7b2dfc
f0e40d79f409be7f4166f01d6334a9b101c9fe6f
5239 F20110318_AACFXF mikles_v_Page_026thm.jpg
c9946c63efaf6c2d02c36e7291ca0c19
5ba180cd8a4c2bc6a5cbc936da5b8ed6cebd146a
8610 F20110318_AACGBX mikles_v_Page_029thm.jpg
e55f97b842f92ac42d497bfff5ef9cbc
18b0ff887f43dc8355bf99b5da31770d554c0dd6
66929 F20110318_AACFWR mikles_v_Page_057.jp2
ab79194811c93d03fc041367e72919d0
df885062316caa204186ade89caca635ee44a54f
20161 F20110318_AACGDB mikles_v_Page_133.QC.jpg
6543021e25a5908951e9f87b77955309
f8771c8a15fcbcb8032cd7f07dd91080e2f0e437
2330 F20110318_AACFXG mikles_v_Page_161.txt
4b3b04cf94ef9b3ce880a33c0dba5464
f4dac187c5e118fb1b800c9fe67dcae617a35c5b
652204 F20110318_AACGBY mikles_v_Page_111.jp2
2254d6b1fb1ee4521dde441c9aee9f3e
9af292f01c3f7209756279bf6302269ba784946d
5100 F20110318_AACFWS mikles_v_Page_158thm.jpg
79cfb34f2e28d74e1dfa626f5cc3738d
be287ad01f8e27b0f87046649b73278a4036d035
6927 F20110318_AACGCN mikles_v_Page_050.pro
8e76ed6637faa63192cc5106680a47ef
2b49d24b54cac0aa51a5a2bf8385c7e2203e179f
43049 F20110318_AACFXH mikles_v_Page_107.jpg
737aaee9ebc86459541b36ab6beb7ed0
9cab8edd2d5bd28e12dee5f0d02caf00d463fd73
1051907 F20110318_AACGBZ mikles_v_Page_085.jp2
842a218176d0d11db450cd81b6b309cb
5257bd9fb71bfa22c3b1998b0ef3b5d0102e9f28
34245 F20110318_AACFWT mikles_v_Page_012.jpg
d75db9911978f851c3933ccda4590425
31329ae2da7ece5f1a2eef4b7fc74b2592a240aa
8176 F20110318_AACGDC mikles_v_Page_116thm.jpg
1ce93d76003145c670ade2c976b662b8
325febbedf8719981646809a1ecf18707d43b62d
59047 F20110318_AACGCO mikles_v_Page_140.pro
6d9048e970fc1068888c6eee884b7ed6
2b6ebff6509bc9473353ce9b21555736554ccd74
53777 F20110318_AACFXI mikles_v_Page_059.pro
2fd91536b59ab00b910c6bed1358ed30
4f570922b1d0eec23f30314bf503988baf0f44dc
37597 F20110318_AACFWU mikles_v_Page_122.QC.jpg
c744f9de042a0d4b046329960fbbd808
3234944c8e986f40ce3ef798e95b69646a329174
2320 F20110318_AACGDD mikles_v_Page_086.txt
90204fc56dfb61046251a12bbc71ada2
59fc7d9771ae5cd471145e4d906fe696d2902b97
93673 F20110318_AACGCP mikles_v_Page_128.jpg
dea2b1203ebc13dcf56b009d2f644dc9
e5017d270bf337c398fcfa812d56a649d7c92a66
F20110318_AACFXJ mikles_v_Page_102.jp2
6062d6434a3cf90f8b2a1f7d37e65382
28938f966c0415ef24fb445c2cf96fe43bf1e40f
1985 F20110318_AACFWV mikles_v_Page_065.txt
845d7e4ac09e3a4257a6bd672e0e2315
5047822c4711013a1da29742421f48a514c016a3
77744 F20110318_AACGDE mikles_v_Page_136.jp2
121570fbf6bdc0f81855922e4c4dacf1
7b14a9405369d76b3647543b999d0437eceaec9d
121375 F20110318_AACGCQ mikles_v_Page_029.jpg
0fec47e4a9f9a4d3781a6417c622061e
9d65d2b90b0af9de48ff3f3d26327ea30665b6fb
40861 F20110318_AACFXK mikles_v_Page_094.pro
4dfd7c148cb29eb6c140af611356c1aa
b80f3dd48fdf361218d0ba6922d165b3120d011b
8229 F20110318_AACFWW mikles_v_Page_030thm.jpg
e0c6d361388eaa8075880a26cd354f33
76f130af7d57524667a2c25140be745ac2da818a
103005 F20110318_AACGDF mikles_v_Page_173.jpg
2c6a2d3934c18919824cec352163db55
32d7383de147e686287b1650d1a4456682c7224b
18860 F20110318_AACFYA mikles_v_Page_025.pro
3d8a278d5de330419f00e0ce6d7c1fda
b65d7aace2caa0972d131d6ec356d44bf435023e
8984 F20110318_AACGCR mikles_v_Page_172.jp2
9dc2fee13ec4bbe1047d6ef32e948d4d
4cc5d2369b3f0efb75215b9ce0dbefea439dd5f9
99588 F20110318_AACFXL mikles_v_Page_168.jp2
4190d0b3cef6da4a95ef4d02bdaf6bcf
c0832b08bbfd9c338bbd94e98e21eb13c8622174
F20110318_AACFWX mikles_v_Page_148.tif
2b3064919d718c14517aab56a127bb7f
2a0c1a74bd89c2a452659b200cb593008981ce06
107261 F20110318_AACGDG mikles_v_Page_165.jp2
82dd98cace56db86ebdfd4b98ee3fa20
25e8d194b33c4f54ec12923b3abcb8f1e9acec18
2545 F20110318_AACFYB mikles_v_Page_055thm.jpg
c5fdfac1f8e6b542277a28baa7fe83b8
7de8826dc5b39c13a92fadf0f1327c267f67bb71
5435 F20110318_AACGCS mikles_v_Page_004thm.jpg
f8135bba958e1d21487c84d29f0491e5
63cb04222708e8bf811c52c35234f014c5dbfc7d
26153 F20110318_AACFXM mikles_v_Page_156.QC.jpg
16a1bcc264c5b5095da3acdf6994a9b0
39d203ab7465eb4dff01c7e255c501126af0739e
35559 F20110318_AACFWY mikles_v_Page_099.QC.jpg
44fd28fdacb7c2f1380fb722b3f3b681
ddbc97954906b84e1130d924fa53c4158c8b090f
15865 F20110318_AACGDH mikles_v_Page_012.pro
e78e023fd418651caa084cca5b50b1c8
9ccce00b5654e5d5f305d5e2fc505097ba852683
84941 F20110318_AACFYC mikles_v_Page_125.jpg
a18b5d5f1a78269183dc53e49ce44645
28ee65f2664cc10cf8a11b559d9edfa36e2fcb95
364 F20110318_AACGCT mikles_v_Page_076.txt
f13a7a4c53b8ffd734aa0e009db3bde2
614e6fd8b99bc66b2d6fbed50f8ab3804e61384c
78211 F20110318_AACFXN mikles_v_Page_035.jpg
83d37d851781808dc3fe20e69509309d
b7eb66f6d02f352981d6d820f09f0c035c303486
110528 F20110318_AACFWZ mikles_v_Page_147.jpg
58432e9a857375c097b50faff280c8ff
0d3b4288969402c61e30b2accf1d40c05a27590e
2348 F20110318_AACGDI mikles_v_Page_041.txt
66bbc8567e60742034d2908a9c876ef5
6e2165652962946db579f446c0e93b27a958b2a2
F20110318_AACFYD mikles_v_Page_110.tif
821c18530ce1beae7615066c256c8c12
488d985c2e02a72fc0c82a8b3eb314c04c1b45d1
427 F20110318_AACGCU mikles_v_Page_106.txt
5b903e14a8563ed8436175a2f46a6e79
ae4643facf4f25ae19e50e2bec21f53e2dd0e0af
7569 F20110318_AACFXO mikles_v_Page_047thm.jpg
4161a9560478490113fc63594a0dd675
872b765595958fa1976fcb67592be4f733090eac
2339 F20110318_AACGDJ mikles_v_Page_121.txt
15d041c09fd6ca26b381e157736675bd
a6b28af23d372a57a6d5a476aced42a5cac9a784
42015 F20110318_AACGCV mikles_v_Page_078.jpg
3e33e25779f42ad5e5dd146eef50b5aa
655cc7916d3ce133f88e21d7c4e6fda7559f0f7a
19264 F20110318_AACFXP mikles_v_Page_034.QC.jpg
5b0d1e64042db3d4326e4a5e7ca66fda
4dfdff4512be06e6af6f7758cdda1898ab6d2109
2302 F20110318_AACGDK mikles_v_Page_103.txt
8b74ebfc167de5420884709413b2de5c
ff1e0679ee93457000778ab634144c2543f72ea9
4819 F20110318_AACFYE mikles_v_Page_126thm.jpg
06e6616a5d874c26a719b2e7837f7dca
87b2c6b32a580ab360792335f55f5a13b4d70a4f
24327 F20110318_AACGCW mikles_v_Page_023.QC.jpg
e7b74563ba9e98db365d6d4e7be9083b
d093e79701df764af8c5999c4992f03b9582061f
1911 F20110318_AACFXQ mikles_v_Page_168.txt
4dc638aba2a3526b7e18c1fdea6a3df7
868fa18e1be26d81769b4f833717549ea2131540
112098 F20110318_AACGDL mikles_v_Page_018.jpg
61853891a59b833efc5cfa02c4597a51
8143bd9f93130e788cfcd991e68f655f5dfdce70
34879 F20110318_AACFYF mikles_v_Page_153.jpg
c87072b2c49dec90447798634aeb97e7
5bf020a383971d68c24ff91c0dd597b6de666265
F20110318_AACGCX mikles_v_Page_018.tif
d3fd63123d51f04e651928a90e48b72f
8ea7039f16439ef74feecf68f47590fc3889f950
8339 F20110318_AACFXR mikles_v_Page_151.pro
da7c64b9889518932eb7c724ba1dc2d7
0374a912a6e4f062328f6acffaf62f9a619b53ec
57389 F20110318_AACGDM mikles_v_Page_016.pro
67f59b37edc385b31ffa0f9c27c8e757
ed4fae42c0211d590fdca1175d03bf56988f4988
8435 F20110318_AACFYG mikles_v_Page_143thm.jpg
40df4a6831ffc35eef443ef75cecb7af
27e90d45d4a7ff7ff9be5f4caad8f486b06ad211
1864 F20110318_AACGCY mikles_v_Page_104.txt
471a07ff80bfc2140b9fbb919f190214
3f152127ae6828fd7f271a7bf8eefe4c40604c76
4244 F20110318_AACFXS mikles_v_Page_111thm.jpg
6f8c8176a112ddde1ac54c2efeb75b32
5d4c81c560c4cad4cb7d773497a16a2c549fa8b4
20329 F20110318_AACGDN mikles_v_Page_129.QC.jpg
a29dfe2411ef422cc573517a16ecc11e
727b75889e19808aeac829aebd3f882a1c1ea87f
792 F20110318_AACFYH mikles_v_Page_138.txt
4e0158856463a13b672a8dc9fcea33a3
431052c89b7456986553f46a2be83e257052898a
7355 F20110318_AACGCZ mikles_v_Page_110thm.jpg
8e848c81c05fa0b56e8b62b42340002b
189297f2aa636a9543a978c4233e4373b46498b0
2869 F20110318_AACFXT mikles_v_Page_163thm.jpg
142598d82a2b7d530e095c32a97864b3
c47ea00c821495927cb0d2781df5fd485ad7c9f2
F20110318_AACGEC mikles_v_Page_113.tif
f78e51669952342776576f9efec513c4
12ea72ac4e2de971ccb1670cfb4bb8e3d0887cc3
121329 F20110318_AACFYI mikles_v_Page_113.jpg
953812344fa8dcd14d03e1defbd5d34a
5646d4285a542b9f8a293f29851fcd8056d831bc
2422 F20110318_AACFXU mikles_v_Page_123.txt
502d1be9062f2e3bfa993be1d4e06100
287ca50e25c66b79bb2a963d5a7886afdf05ef55
F20110318_AACGED mikles_v_Page_114.tif
9c75d60c7e08f4e93d749cb7a697079b
8162464a858aa376a5932fbabce44c0576a7d5a2
120671 F20110318_AACGDO mikles_v_Page_162.jp2
3038756a9bc82b8475f14f6b6e9ebeff
397711327157cf0cd84afde4ae720a08c7fc03e6
15752 F20110318_AACFYJ mikles_v_Page_073.pro
eae5da7eaf420ddfd5da9ef729eae48f
16b4ffd0ff6248959134c7a9ea0447631ad2313f
7740 F20110318_AACFXV mikles_v_Page_165thm.jpg
bc759adcb41f95db3dfc32ff0f77dc2f
120cbafd0740a3c9393cfb7eba82278bdb75ee4e
F20110318_AACGEE mikles_v_Page_115.tif
16134bee16a10b1b4b53c0c1fe7fcd4b
a2dcb0339b06585fc7d1d74721ba56200e102d75
22523 F20110318_AACGDP mikles_v_Page_159.pro
c005851e0570b65349bb6c27b17fecf0
8d3e2f387c65b8efa04407c5800e67c6818a2200
F20110318_AACFYK mikles_v_Page_120.tif
cc1cbfa2b0ab53e52bddd8124ddc7cfe
2881a5d854f58f7a2c81cc8e3f86dea13b9b0691
119274 F20110318_AACFXW mikles_v_Page_045.jpg
5bc292eeb163dae74f8fda94d605b436
f3130fbc71d7ffbd542503b07f45f210f5c594f9
F20110318_AACGEF mikles_v_Page_117.tif
407256fa29aa9b62f52080344eaad847
e988b0665895bd11e024b6224802279369f0f4c8
F20110318_AACGDQ mikles_v_Page_160.tif
767fd68b52b30b7842b22a1e93618245
3318cc96b71fa17a379c952cb127a1b9775856f8
109056 F20110318_AACFYL mikles_v_Page_059.jpg
065f4d4c36bc83755742074ab0fd1af2
778d70687d1992f22919d39750c7129fd9872af0
853 F20110318_AACFXX mikles_v_Page_164.txt
eb00bc953e815caa0e7646ec69cc8582
910f10c3f8540f7b1983844de0fec6778a8140d4
F20110318_AACGEG mikles_v_Page_118.tif
94b7b403f7fbe04a0323d756521341ae
1b41e6df2fbdd250b420cb88cccab54035e008d6
19569 F20110318_AACFZA mikles_v_Page_131.pro
c53a8d76f284d2d10d2723b3e8c64447
ade3ea003d3c92404a016a2acce03263896eb064
107607 F20110318_AACGDR mikles_v_Page_047.jpg
3903ee85d85966c1f63b4ad7bd72322b
9773ab18c1a4c62a2fe0c52155663b5522ca4ef0
90780 F20110318_AACFYM mikles_v_Page_008.jpg
32622a832b2d1c015eaf4f1811a03afe
7a16fbfdc17bfdc94b48c708aec4570401e6b314
380881 F20110318_AACFXY mikles_v_Page_153.jp2
827b7a292a5794a5f3f07e51286c9c12
3ef5d51bf6decc0c7de5017d0fdd16232cc4bc50
F20110318_AACGEH mikles_v_Page_121.tif
3e28019f9be8fc425b79e8ecc31adee7
63f491489566e1d6d35d502c1ac1b9297d99c014
2795 F20110318_AACFZB mikles_v_Page_172.QC.jpg
0977925fd4ec032ca1824612eae1bc59
2afe7e67fdda330e6475d9899737ebfe19eab6c2
681 F20110318_AACGDS mikles_v_Page_077.txt
6236c5be8c4ccbd7d213bb4ea4591497
17709c6a061b4cc5a20119d6521ff9406e381ec9
7360 F20110318_AACFYN mikles_v_Page_069thm.jpg
37cf9b3e76eecc3f1673c6d49ab5b8b3
e3258d0c4363d63b1a1e0676e55469ab9a7c81d4
F20110318_AACFXZ mikles_v_Page_139.tif
f26c5cc65a1a6ee09295143582707630
aa78fe97c104bf39d907516804e009264d02c8a4
F20110318_AACGEI mikles_v_Page_122.tif
66075258a7ca0602e8ed2f369f223730
60a15c84aeb5afe4496dadde1f75efc8dc7adf70
F20110318_AACFZC mikles_v_Page_133.tif
cf1a76d818d12f14a4a0dddda8b0b851
a12e601ae2e596a00f66a745abc94b9ef2269500
57558 F20110318_AACGDT mikles_v_Page_118.pro
91cf755d33cad521aebe19ea7091e5ff
56875fdf0d723b1191784aec6c6d289f9b0bf4aa
F20110318_AACFYO mikles_v_Page_147.tif
e4d819a9b37abc4341e2aa46bbd86df2
83aaadfa86f81996bebe3707543b919d8203e54e
F20110318_AACGEJ mikles_v_Page_123.tif
6718ac35c04b0d7d43e439e68d8da3a5
91b8ee60188d4b6f2522f0ee4676b315056148da
4938 F20110318_AACFZD mikles_v_Page_054thm.jpg
73cf17a3f48b616bf4d4c77547984c8d
6383feb1549ea2d8449edffb2672997b18166249
8673 F20110318_AACGDU mikles_v_Page_041thm.jpg
74ad241e9b7980ca82bcbafbab956a12
d8a50ffc1e1eb7d0454776382cec7a2c06f4b893
49467 F20110318_AACFYP mikles_v_Page_010.pro
66adc20eabb261761b52c96c26570b95
e75ddf76003d7c860ff3684d452cc3b2dba5400e
F20110318_AACGEK mikles_v_Page_125.tif
e0272441b6d8f29016102816c089eed6
15a74dba4992678a6ec6b27c58c40d35406ebdeb
F20110318_AACFZE mikles_v_Page_130.tif
b1f9401c42b02677716866837bb630fa
298a89809ae5d05862a2f94e65bd42d62e01a3cd
6249 F20110318_AACGDV mikles_v_Page_131thm.jpg
55eb8afc9418ac93d743226ef1098073
00b507bd6678ddf1b7cc24a1606b4ba08d251d4d
2207 F20110318_AACFYQ mikles_v_Page_085.txt
476e517fb1f7f490f0a0844a3eda2519
a23ebacdf9b45ba01021ced49e778e3ff346ccaf
F20110318_AACGEL mikles_v_Page_127.tif
667f995d240a76ea5ad3010b1938dd7b
6a2653b4ed7ecf05788038620fbade87888882b0
8278 F20110318_AACGDW mikles_v_Page_014thm.jpg
02ee5a22bfe2ce6f678b14db1c128ae7
480e379bb2f97a2016faa109e4979d3409eeea07
F20110318_AACFYR mikles_v_Page_160.jp2
722e216186addbe89b8c8ee5c9721527
afa15525fe13b0b85faf95c47451db9ad1fe78f6
F20110318_AACGFA mikles_v_Page_169.tif
e715c91194dbd71ce99ce67ee6a8b1e8
49198dd9da0e3ebdc56fcc3f03ff1cec99279b20
F20110318_AACGEM mikles_v_Page_129.tif
5ce7c13b520b527fcfd48ba49859c9bc
69709a1bd5e9a56353778988e1c99970c9f104c7
51793 F20110318_AACFZF mikles_v_Page_087.pro
5de160d121fdaed8f4fe0b6f240c7bcf
e3477afad26bc7c803f78d90ac4ebec0e9c9b146
17063 F20110318_AACGDX mikles_v_Page_026.QC.jpg
6e72f0097a2bd0542cd84ac272eed31f
c99cb8a90fa5cc1d0b16dadae97e168e61c8be48
1051966 F20110318_AACFYS mikles_v_Page_045.jp2
dc2804f667aa63e1354991dc8178fa8c
d35e41529861613ddf332265a90208f8809ef845
F20110318_AACGFB mikles_v_Page_170.tif
146291c9fa67f1fcaa9d573a14505557
937429b9acb7cc4557c9f0f9a79fb30f452e4c8c
F20110318_AACGEN mikles_v_Page_131.tif
f9ac5d27aa645384329096a0ab02c33f
33535ab7d6028f1186601dee7e29836f841d5bca
1278 F20110318_AACFZG mikles_v_Page_128.txt
682bfc59e0cb1aad0325d1421ddee1c8
932f7652f266f8511e9b0bc333d666a5c018b519
8407 F20110318_AACGDY mikles_v_Page_084thm.jpg
57193fdec84be8a7ecb07998ac3c3985
7234b90a19d032d7d8250f9dd3363a559a64c137
8689 F20110318_AACFYT mikles_v_Page_046thm.jpg
51176d79d7ab9d934d30deb514116a28
c9d241e4966e8ac52a6f52a57e912ea548a94c72
108 F20110318_AACGFC mikles_v_Page_002.txt
068797ef8c09264c39c4db156df04671
bd3e154f246234aa2ad02a9a9b2ea213ea2ab9e3
F20110318_AACGEO mikles_v_Page_137.tif
53480da6081ecfb8e45ae1cf69774a13
ce5469de29f32a9e59aebc7f6f2fb3270fde70e7
F20110318_AACFZH mikles_v_Page_166.tif
051bab67217d5b7a10f47becf08b4bef
b87bae119679ac3a57265db7810c03551aa42fd7
278672 F20110318_AACGDZ UFE0022086_00001.xml
9679873da084c3caed27cb620848ad4f
0ef100875fd5f42420aadf8ba95c6b221be14224
mikles_v_Page_002.tif
mikles_v_Page_003.tif
mikles_v_Page_004.tif
mikles_v_Page_005.tif
mikles_v_Page_006.tif
mikles_v_Page_007.tif
mikles_v_Page_008.tif
mikles_v_Page_009.tif
mikles_v_Page_010.tif
mikles_v_Page_011.tif
mikles_v_Page_012.tif
mikles_v_Page_013.tif
mikles_v_Page_014.tif
mikles_v_Page_015.tif
mikles_v_Page_016.tif
mikles_v_Page_019.tif
mikles_v_Page_020.tif
mikles_v_Page_021.tif
mikles_v_Page_022.tif
mikles_v_Page_023.tif
mikles_v_Page_024.tif
mikles_v_Page_025.tif
mikles_v_Page_026.tif
mikles_v_Page_027.tif
mikles_v_Page_028.tif
mikles_v_Page_029.tif
mikles_v_Page_030.tif
mikles_v_Page_031.tif
mikles_v_Page_032.tif
mikles_v_Page_033.tif
mikles_v_Page_034.tif
mikles_v_Page_035.tif
mikles_v_Page_036.tif
mikles_v_Page_037.tif
mikles_v_Page_038.tif
mikles_v_Page_039.tif
mikles_v_Page_040.tif
mikles_v_Page_041.tif
mikles_v_Page_042.tif
mikles_v_Page_043.tif
mikles_v_Page_044.tif
mikles_v_Page_045.tif
mikles_v_Page_046.tif
mikles_v_Page_047.tif
mikles_v_Page_048.tif
mikles_v_Page_049.tif
mikles_v_Page_050.tif
mikles_v_Page_051.tif
mikles_v_Page_052.tif
mikles_v_Page_053.tif
mikles_v_Page_054.tif
mikles_v_Page_055.tif
mikles_v_Page_056.tif
mikles_v_Page_057.tif
mikles_v_Page_058.tif
mikles_v_Page_059.tif
mikles_v_Page_060.tif
mikles_v_Page_061.tif
mikles_v_Page_062.tif
mikles_v_Page_063.tif
mikles_v_Page_064.tif
mikles_v_Page_065.tif
mikles_v_Page_066.tif
mikles_v_Page_067.tif
mikles_v_Page_068.tif
mikles_v_Page_069.tif
mikles_v_Page_070.tif
mikles_v_Page_071.tif
mikles_v_Page_072.tif
mikles_v_Page_073.tif
mikles_v_Page_074.tif
mikles_v_Page_075.tif
mikles_v_Page_076.tif
mikles_v_Page_077.tif
mikles_v_Page_078.tif
mikles_v_Page_079.tif
mikles_v_Page_080.tif
mikles_v_Page_081.tif
mikles_v_Page_082.tif
mikles_v_Page_083.tif
mikles_v_Page_084.tif
mikles_v_Page_085.tif
mikles_v_Page_086.tif
mikles_v_Page_087.tif
mikles_v_Page_088.tif
mikles_v_Page_089.tif
mikles_v_Page_090.tif
mikles_v_Page_091.tif
mikles_v_Page_092.tif
mikles_v_Page_093.tif
mikles_v_Page_094.tif
mikles_v_Page_095.tif
mikles_v_Page_096.tif
mikles_v_Page_097.tif
mikles_v_Page_098.tif
mikles_v_Page_099.tif
mikles_v_Page_100.tif
mikles_v_Page_101.tif
mikles_v_Page_102.tif
mikles_v_Page_103.tif
mikles_v_Page_104.tif
mikles_v_Page_105.tif
F20110318_AACFYU mikles_v_Page_092.jp2
a55a800ddfead7ebd3004c8e992b815f
7bf03016686a44c4bef8fc3d73729273bc53bb11
3383 F20110318_AACGFD mikles_v_Page_006.txt
e130cd79dd788fab6d6a6e27897f38a3
b73933e1865ceaf1b7a50db1ca590b6b52af07b3
55030 F20110318_AACFZI mikles_v_Page_075.jpg
2c43f9aa6b020d3406b32c43c124b374
e26d0bcb046a1fe90d0859f13dc6fbfdb5b7d0bc
77613 F20110318_AACFYV mikles_v_Page_023.jpg
17f5b0d83118e7e5257ee76c2846670d
3770bedefa9a6459d0181588559fe24bb39a86a7
2009 F20110318_AACGFE mikles_v_Page_010.txt
90f335945f256e398b99d787a7e138f7
e12d479ad3564a9d81426c20d238e929f794542b
F20110318_AACGEP mikles_v_Page_138.tif
8dd0daedb307fb6f2c567e231ae9bfc2
af05903aace874004db075eb19e20f66e32b0ac8
6848 F20110318_AACFZJ mikles_v_Page_081thm.jpg
ac887ed2a8c6db8a354cdbcaa3da8f2e
3f2df8634541a1b170456d8fbdc1abdc4584ebdb
32753 F20110318_AACFYW mikles_v_Page_071.QC.jpg
74467a9d403612fcd9a064101ceab225
80f86a09cd032f7b184c53800fbbb80c08ab0b56
2122 F20110318_AACGFF mikles_v_Page_011.txt
82f14fbd455935ddb4f79a17c284eac9
ee37b0e5f2ecb73b90c3a86c69cea64793c869a0
F20110318_AACGEQ mikles_v_Page_140.tif
5415fe38cd8f418f4e7b540075a8850b
6f84745501323518a76168c4f691a310c62537ea
1051962 F20110318_AACFZK mikles_v_Page_067.jp2
ceda1c3547b85fea4c68d6054388fb94
867b2b8eafe99329cf1d9d663b73ef0d112cc5c6
640 F20110318_AACFYX mikles_v_Page_055.txt
243ba273e75e7bbfe98d0d5239725d7b
e5f54f46bbaeb8c5a865aebd1ba1cfa92c45a697
F20110318_AACGFG mikles_v_Page_014.txt
523072f8cdc4591a5e7e26f8946349b2
28387412cb3ad6727abb197777092cdbc994617b
F20110318_AACGER mikles_v_Page_142.tif
6518b0644a3bf6bd2807213353bd609d
afad176db0772b00c722b7d8c60b3ea557e451e5
118488 F20110318_AACFZL mikles_v_Page_141.jpg
184ec3d781e5296a820363161632ea2e
e62fc1abfbdcbd32571a66f54069b352cf3effde
35478 F20110318_AACFYY mikles_v_Page_060.QC.jpg
67c267b83a0c3b2e2e1c0c1d15d6d69e
e3c56570b499d0abcbcf546bec14bb50476982d0
2255 F20110318_AACGFH mikles_v_Page_016.txt
888bd1623e1c66607348187386c7626f
fcd2772f8fde5697f1e8d4a653f0cbac611d9099
F20110318_AACGES mikles_v_Page_149.tif
7e7d7089286de98ea64179b32c994edf
87047d728dda9dde5d560945c7d3b6247ed19b54
113952 F20110318_AACFZM mikles_v_Page_115.jpg
a3602acf8a2a34e37969555be5ce83b1
bdefed736d34e0cd1537e4e7248f126ecee23787
378 F20110318_AACFYZ mikles_v_Page_036.txt
c8c2684e3ad68fe2ae1583eaaac41442
8dcad60472a70efdc88c83bc99024253fa36340d
2073 F20110318_AACGFI mikles_v_Page_019.txt
83dfb67cc1b9ce967763a2c4ef1465a1
b101c50101b36aaae1425f68a58ba6b71e726b2c
F20110318_AACGET mikles_v_Page_150.tif
f4b286768ad68275947d38084383f3ad
0581da9090287798df3b8c4703144217d53cc073
63409 F20110318_AACFZN mikles_v_Page_037.jpg
93c744e43d5b2113503830f7ebcf7114
c52c61c429090c53043904f4d282c98b21723d64
2370 F20110318_AACGFJ mikles_v_Page_021.txt
af9c7846ea5e26f90b8a384d4687adc0
4af8b0cf5db9ed7cb282dba9dfd4adf59ed314df
F20110318_AACGEU mikles_v_Page_151.tif
07fcac613f1e0cd5ac23dddb0df59ffb
797bf44ea230f268a0c4bd279be0ea685aa818dc
8439 F20110318_AACFZO mikles_v_Page_121thm.jpg
ad52649eeeb9b3118c9362a1dc3e5189
9a80bd0f7916fa79e339748bbc4ebff5c6a38385
445 F20110318_AACGFK mikles_v_Page_026.txt
09dfdccd68cea05d0de9978673eb093c
afe565328ea89106b5f4f9860d29f5b723cc5f0a
F20110318_AACGEV mikles_v_Page_154.tif
8c56116fd505672ecc9e74d989783cec
6dc39a94093d6f154a542c8ddab2ecaa9460d0db
37392 F20110318_AACFZP mikles_v_Page_028.QC.jpg
e68f508685c870070d832c93e1b117e9
49178bddbe8ffd39f968cb57519fc07d685ac4a7
2316 F20110318_AACGFL mikles_v_Page_030.txt
26a192b6ae97f0d5f0ac9af33e8c1fde
fbe17069444d023810386ec005186ec649883952
F20110318_AACGEW mikles_v_Page_157.tif
9541b5f30e84b71726869f9b73e17265
b6e3298473e02a6a9f39b39a3f93668848f079c3
2270 F20110318_AACFZQ mikles_v_Page_099.txt
455acdbece937ef6192f5d6accfbfb0e
9564b8bba5e29c62367666ba799322e5af15df3e
1175 F20110318_AACGFM mikles_v_Page_037.txt
131b1e455eae9d8a4cfade00b2dd258b
c3fcdff5527d31dbf712665bfdcf980041f9f7be
F20110318_AACGEX mikles_v_Page_158.tif
2672b78952d2e9b6d87b04b5cbcc9380
5cef05d8430814e970cc5a50ca5f83debc3a983f
31178 F20110318_AACFZR mikles_v_Page_073.QC.jpg
cd6248e3ad99927f5483125326ba9296
0aed6ef8f755bd32c761b66ce0cdbae8ae9b6ce9
1101 F20110318_AACGGA mikles_v_Page_080.txt
e785063fc470dc8504faaed6da621e10
d6c3d6cc518af5604523a6d1c3e464cc07e23fca
F20110318_AACGFN mikles_v_Page_039.txt
bc087f609c552a547169356d6f9c713e
bbff957f1823bae7252f62188a69866dc59e0e95
F20110318_AACGEY mikles_v_Page_159.tif
67cbb6f9dde3ba5bb9e79525540c4519
2662165339abe40b89ae8074f553932308893357
6377 F20110318_AACFZS mikles_v_Page_008thm.jpg
530e0829ea81587f19a68c478bec3069
8a62b4258c4f13e9c58761eee2b70e76c89f47be
2410 F20110318_AACGGB mikles_v_Page_084.txt
2c7ff002a86ec6549f103c85dbcf6dd3
5825593542e3947b049cb2b495a2d3818ffa3c4b
2259 F20110318_AACGFO mikles_v_Page_040.txt
372c653bc261c435f0424e0e71b64389
d1c30d2703efa5e1c47078af4ddf9763439ccff2
F20110318_AACGEZ mikles_v_Page_167.tif
fd3269c2980a36c03265107ab98721ea
c860aff3324126010137c46c8000b544d97e7e68
F20110318_AACFZT mikles_v_Page_163.tif
1f56815b2db7be145d69141dd2d3e9a9
f2dca79046f56b042ca412d8e85ad97e9792f650
2248 F20110318_AACGGC mikles_v_Page_088.txt
2434633e1fed6170b03629d7862055b5
d61118c1f12c8c3ef6beffd701d54e464042699e
2387 F20110318_AACGFP mikles_v_Page_046.txt
18cb4111e246b928018d83891f059ff3
0ade0aec6f727f3acbe19e18eac71a9dc7067f78
F20110318_AACFZU mikles_v_Page_006.jp2
210db76e1855bcc6dccf8145cc01aae6
0ac7ba0a4325b9b697ca267b6c9835623420c0f8
1967 F20110318_AACGGD mikles_v_Page_089.txt
775e5fbdaad3beb467a3bf96375b6e88
7ef79b1d26f0e64dad9459a299ada082dce113b4
406709 F20110318_AACFZV mikles_v_Page_150.jp2
a354d391e8d491937e4dc1e237e65db5
6d8e0e83ee1717d928ecd22796846d1ef80bf73d
2091 F20110318_AACGGE mikles_v_Page_090.txt
da537f7ff0a2e9b90636371d270e0332
d99a4edf5b65e244042cc95e83014e7779134b69
527 F20110318_AACGFQ mikles_v_Page_048.txt
ad4d239288691074dfb6c0945c10c54e
11bc9541846e36992a436ef824e00e72f3da6895
4896 F20110318_AACFZW mikles_v_Page_057thm.jpg
6dd718839cd9a6d3bf1c895ca0b7578f
fc791cbb21a5d75eb8bf61270739a0d182f68bb4
1668 F20110318_AACGGF mikles_v_Page_094.txt
7238d5e406cec02b6210be3307010f08
3c18c530ab42dbb6010ebb3472ab3f83cdc8a9dd
524 F20110318_AACGFR mikles_v_Page_052.txt
ee683bccbf162d4e02f1fbfe6ae471f0
fcef3c3348885ec5c41bbee23ff0eb69f0b5c424
F20110318_AACFZX mikles_v_Page_033.txt
6e4dc11c7e14cbdf97f4965df8850fdd
a1c21297ab29b2e18ee0b73016ce575d6674a692
F20110318_AACGGG mikles_v_Page_098.txt
349e13eed187f8e11e321f4cd5876440
33e3f3f0b1f14a21de92a1c13d6cb1c1a2e15bba
1218 F20110318_AACGFS mikles_v_Page_054.txt
813a4cb4409d962e7d7321fb39951477
f87c67cc0c5f5fffa815d0a0dea84115a56d94b7
57622 F20110318_AACFZY mikles_v_Page_022.pro
338b64b1278b837b652badbc30d0714a
8b721a41aaea6787d44e401c7f495020f5450465
2230 F20110318_AACGGH mikles_v_Page_100.txt
13942d0cbc5377fdcdea97879d8c9e1b
2bd90f78c42abcaa054e507ef79d43f1bb59d402
2301 F20110318_AACGFT mikles_v_Page_064.txt
af0b41aa0c38b85c44ee1155f6e797c0
2d0cabb45b4325f915ae3eb9c85a1b526f66cbdd
78080 F20110318_AACFZZ mikles_v_Page_006.pro
2cf9876ed2698f7d75e4e829175b9afd
9ab0a0cb73cfab304f42b8f5f0261cef7712e403
420 F20110318_AACGGI mikles_v_Page_105.txt
4e26758527b9b57bfd5d0a982eb07b3d
474b1d403b751cd4daee3ee702194e035f0db7a9
2274 F20110318_AACGFU mikles_v_Page_066.txt
005489aa5430f2acaae2ec4b4ab63f38
92b40b3fbcd7178528e365c739034e1e3e8ed476
518 F20110318_AACGGJ mikles_v_Page_109.txt
33a729642f6f422da2cf38ae3349272b
85c8fdf96e79e99992d5f63c5f649869bad6b91b
2281 F20110318_AACGFV mikles_v_Page_068.txt
6f8c3c29c752a37cfeb936e723322c0c
9eb2cb37b6f6009d43c2d6ba4986e1dfa406be5b
1011 F20110318_AACGGK mikles_v_Page_110.txt
23d9bbbd4d9d1e3d4375ddadffc2287d
362bb3b2b010549e3a0f6f46a640278086872197
2342 F20110318_AACGFW mikles_v_Page_070.txt
5e20f6c9dbc88af2104ef6e1a579a7ac
677adb0d915b8ed46b080e6ea6af49941e45f08a
1256 F20110318_AACGGL mikles_v_Page_111.txt
ea6f143952e1aba1e96b47efbd63efc7
9503d9dba321fbc6375c5b3f185323446042ff3e
1388 F20110318_AACGFX mikles_v_Page_072.txt
bae0ca5e19ccc544cee0313783ddfc60
b9d9c9519436659602587e931f48ff972f2beec9
691 F20110318_AACGHA mikles_v_Page_153.txt
a07c4818aa5074fe9dfa205404bdf38c
0009addfdf6931f9e08063945249997dc9011f32
F20110318_AACGGM mikles_v_Page_112.txt
18d4b908b4c15944b92515fe6a02621a
423d8b9091bc3838ec7567512aa2d38385ba55e0
666 F20110318_AACGFY mikles_v_Page_073.txt
653dc368579f9c0ef1540765ca7a4b59
9062e60dfd9a5a165ece795243095411b131a42b
1119 F20110318_AACGHB mikles_v_Page_154.txt
0b66dfca56b211e0c964d2c26f9eb57a
d3156b459f55d3404d8ad7d7d90a431b999a3b6c
2307 F20110318_AACGGN mikles_v_Page_117.txt
d48d2e5af4244587db6f5fa10d4a104a
26df2c38d45eff4769a666f5522eead70f5360c1
2411 F20110318_AACGFZ mikles_v_Page_079.txt
522b927ec49a79765e3b008848a22e36
849ebbd1c72c3f504e3a0c6da32a515b1dcd4f9e
3760 F20110318_AACFEA mikles_v_Page_076thm.jpg
25afcd9170ba904a5e282e51a4887de7
aca185884e8a2903553966c27cb8716c573d46ba
497 F20110318_AACGHC mikles_v_Page_155.txt
ec7ca142c6b98a5c40963de18ae73106
23b93b8bfe80d662bee2bee0ae39142384818d49
2264 F20110318_AACGGO mikles_v_Page_118.txt
d282953b27585f188e0e401936ff5e32
cda8c07d21045561264e2d76f94d2d7b241e18bf
18171 F20110318_AACFEB mikles_v_Page_135.QC.jpg
dc0a8248e65e82cbb9f68e8d8eb3bd85
f0972b39dc3e06448cae222ffbf3ad439e6636de
2273 F20110318_AACGHD mikles_v_Page_160.txt
5320b458a807e0bba419f17ff2ea4496
22b925f60c193eb93a84614f050bc9075c982357
2129 F20110318_AACGGP mikles_v_Page_119.txt
9d1e7b9deceaa06a3a0218f3a57248bb
4a741ce7dc4dc13583af7663ce6c5bfe5ce25b52
1520 F20110318_AACFEC mikles_v_Page_156.txt
f372813579c10ce4d6be7e7207c8efe6
b2efddcb66a387b4107e24151aea2139509600f6
2481 F20110318_AACGHE mikles_v_Page_166.txt
96374067ff152a916f2802da70aac833
12619e9ea39e1e7a4bb7a09b16a3d9613f4dfddf
683 F20110318_AACGGQ mikles_v_Page_130.txt
b8d6474ee8598b32672927a9103b378d
0df2f1ad4caf2e025b885d7339bc92e985aab51a
44254 F20110318_AACFED mikles_v_Page_138.jpg
5b326c64361462ebc454342e78b432bd
9aa5d1c2b0d53b7ff53be00db9443cf4d2f4974b
2046 F20110318_AACGHF mikles_v_Page_169.txt
3526187a0f2ce4bd66995e982d666652
cf90157d0a09a2809b5dcc8c891ed9369f945beb
57745 F20110318_AACFEE mikles_v_Page_060.pro
147232753ded51480128d8f3006d3bc8
e68521cdaaccccc96bc849119eaa96fe338b0565
9477 F20110318_AACGHG mikles_v_Page_001.pro
7f05654da4c3d9ccbe4567353f771878
6e098a3987aa396e0ab7a3004c1846e1c2e1f521
954 F20110318_AACGGR mikles_v_Page_131.txt
41964bdab0d6e13961727fa2eebb128b
49230d4cd5ab0a501a9e38e3050d99dcbebba2e1
2299 F20110318_AACFEF mikles_v_Page_022.txt
e76c22dcfe4c496bacad9b5aefe6befb
d1c1552f0e97ee370075d1e4ba64cc8df4a9cfa2
1072 F20110318_AACGHH mikles_v_Page_002.pro
f99318bcc2f776f37323bdd22f896c09
56def0bd5ff03dc99b617190683d494e19269a31
984 F20110318_AACGGS mikles_v_Page_132.txt
99a7792bf7f7011e56b16bbdff6960fb
130f606d9bafcc7cb1d6194351c1b29b7e4bd61f
2001 F20110318_AACFEG mikles_v_Page_071.txt
5c69c0a396c0024ac92fcf91c5063deb
41793c1d02b2e548aecd2cad4c606102d7f771a0
31933 F20110318_AACGHI mikles_v_Page_007.pro
a972a71f469fa839214530b5cd1f30b1
79a9ae370fb0e3ed5d17203eb6eb7395db9fdb00
2222 F20110318_AACGGT mikles_v_Page_139.txt
a282d4813edfbcaa6d60f4e0a16ce172
2a5d98aa7dfffce0d34a20ab1c9656b4f13edcb4
42649 F20110318_AACFEH mikles_v_Page_048.jpg
1a9abbebac4d05edcd423d805c9188d8
03b6d086fc64c0d8cd45f47b4583f3807684adac
35729 F20110318_AACGHJ mikles_v_Page_008.pro
f9e61288429157531ea79cf63e136af3
9c433d9fb572c031cc718eb9fc92eae647ae9086
2356 F20110318_AACGGU mikles_v_Page_141.txt
b4fc314255a9aa44182382a58d020374
6d13c4812e36ab6bf99fb905bd8ff5afb3fe2c72
48289 F20110318_AACFEI mikles_v_Page_168.pro
363d8e01a643f04a65bc7f23d694063c
d8f3a5175b793b684a9822e67e0a63c42d48bc8e
49741 F20110318_AACGHK mikles_v_Page_011.pro
c4282ed862392facf6bfb2e14c2225ba
94d651e6bbf1f5f3426ec564c90f449b6d18691f
2360 F20110318_AACGGV mikles_v_Page_142.txt
86aea6d15635f41b13b4dfd1351e5a77
b12b9a4d7474db7bb720fb07919120f3eca52f59
38071 F20110318_AACFEJ mikles_v_Page_032.QC.jpg
043b2334c54d93b7ba466369538bb7ee
7070166187638a7fa6a1bcd81645c762ff2c8597
56237 F20110318_AACGHL mikles_v_Page_013.pro
9ef965e1aadcfdddbd7266117dca06f2
fbfb81496c0fb60e2d1f3d5ea552a3ce98255f02
4724 F20110318_AACFDU mikles_v_Page_105thm.jpg
e8e4c798615c186d19b21b82bb1cb97c
bf364a12f999ccc65c119a9703e9bb8b813b2185
2430 F20110318_AACGGW mikles_v_Page_143.txt
b10e735a76b86124c824ba061b71fdb3
080239ee52092ea305054211ac50d7811938a37d
27239 F20110318_AACGIA mikles_v_Page_056.pro
3bd3a792e86368a548c8d1528c379aa2
75d7cfe1e652014f4dcb23e9586a66b62892342a
10020 F20110318_AACFEK mikles_v_Page_003.pro
964a555caadb2fcb4f4d682594424c14
0a18a6168a5fc040654d89718e597fdb92ae9e5d
56465 F20110318_AACGHM mikles_v_Page_014.pro
8d59fa9e92855080a6cb37f7934fc23c
f30333b941be41594768f1e4b64aa606abe41f3b
96520 F20110318_AACFDV mikles_v_Page_097.jpg
c94b99854eeb5668b0ff31191a22802f
17639e68fdb15e13ff0d215e882a43075060eaa6
2399 F20110318_AACGGX mikles_v_Page_144.txt
88cfc4a7a4e4d1425392d45f343e8d34
56a56fa5c17109369902ba70065e923591a7e258
52859 F20110318_AACGIB mikles_v_Page_058.pro
fb7cf61b8b2bb0a6e9865a240880cd3f
747803f9d203e521c7f179e6454214d98d23f972
6242 F20110318_AACFEL mikles_v_Page_130thm.jpg
b6bbb4771f853c643f262432ee8fc263
57fbe34525d0157969f5e42fbd0c415d1de0618c
57868 F20110318_AACGHN mikles_v_Page_017.pro
20b128395ccee9f56dc8f34c9ca87be6
d474c9fb035c0603783b40eef938ea8f6d72e386
37192 F20110318_AACFDW mikles_v_Page_006.QC.jpg
f784da71b0d5f4e300aff436c074b19f
c073288a603a305df41a75e02dbffc277374a4ab
2226 F20110318_AACGGY mikles_v_Page_147.txt
a2e323150069bb9249cd4bf4e94bf8aa
7872ed6e99822a4976edf351c92cf5240a0f4d51
58531 F20110318_AACGIC mikles_v_Page_064.pro
a251f644c0a87d6ea0b06b55f8688620
ee667af5cdb8eafb25bb19dc64bfcacf98b89485
47301 F20110318_AACFEM mikles_v_Page_065.pro
726c861ad56103fb46963eee46411dae
c85873c622390aaa3224d29446b12dbd6db2640c
56440 F20110318_AACGHO mikles_v_Page_018.pro
f3d383e1c6a02d0e764b6cfd78f46991
7c1ab8114e8a9a9c206ff7e04fdbbcf04546a2bc
5639 F20110318_AACFDX mikles_v_Page_074thm.jpg
f1ae1a967f9c102c035ad96fb74007f7
99003af89ae03e4210806aa2b535a76b6a6791f8
1108 F20110318_AACGGZ mikles_v_Page_149.txt
9eba9bc1bf29fb94add9f01c23a746d9
c2e9370ea8a87ee8ed9e85e0bba16303e51582ff
2286 F20110318_AACFFA mikles_v_Page_020.txt
fa82a405c2e844f8fd82cf49e58505e0
8657cd20f44f3178006f19751fe61d6f468459fc
46428 F20110318_AACGID mikles_v_Page_066.pro
2b96505836f6480400d6a09ab5f999c2
e5b750b8663096d77e076105fd4638425e2f6b0e
5439 F20110318_AACFEN mikles_v_Page_132thm.jpg
d3ce85fd796eb90428e5cff567f44411
d22a4e8812fe9b6edb871c0c70858533821a5c5a
10756 F20110318_AACGHP mikles_v_Page_026.pro
8d64df5e79bad9dcfc759c04770a2b28
543cdc31050fe8e3e0d8c90abb31d897905ae396
41285 F20110318_AACFDY mikles_v_Page_075.jp2
9c0988854344f21958c91ea366c96e06
649a68768043f1089875ca870516aaacc8fe7970
39035 F20110318_AACFFB mikles_v_Page_163.jpg
0a16993f39266eb8bfc8ed0bd943a5d9
4403bf81b2025a4bf4d0153948be5e438866f958
57906 F20110318_AACGIE mikles_v_Page_068.pro
c4f086496b74564f3a2c4e02dadf5800
d6f423ad5f1ba6f4a35fb6eee66166b3db2f3de1
2257 F20110318_AACFEO mikles_v_Page_063.txt
6a0d8cb07f7d72b2ba20eedcbf64dffb
5d05052f7142a8e7a7325af80be8407c65246e97
57219 F20110318_AACGHQ mikles_v_Page_028.pro
6f08b8361e4f93af56af4d6f5ae81882
22bdba8d825918fda9a6881479c720067ceaf47d
540 F20110318_AACFDZ mikles_v_Page_027.txt
2d6fe5ed30ee16b14befb7a721618a10
ec552ddcfbb6601bafc9550e29bbbc0e8df49810
F20110318_AACFFC mikles_v_Page_164.tif
d9a5439e50b4b108b44f9b58b1f931ab
81c61fb88441d38722e7dd9627e1487ed4a8e2d4
58669 F20110318_AACGIF mikles_v_Page_070.pro
217c1846b4060c5ca07d89b3ef11ed93
c83bf4ecdeb76ef850677bc2b745ec0b177cc24e
8418 F20110318_AACFEP mikles_v_Page_068thm.jpg
6e2b884da4cf0bf90218b69dc6a3ffc3
23f68c68816cd9a456ea888a2856b2baa5febb34
58270 F20110318_AACGHR mikles_v_Page_030.pro
2ebf9d16019181e26b67b2f0cfcf8adf
82bc87c255c878ef9583fcc515c0dabccdd12dd7
F20110318_AACFFD mikles_v_Page_017.tif
0f2f510fb73552397114af90587942d5
8799b17b0e32f592092cf6be305a94e57ef972b9
8219 F20110318_AACGIG mikles_v_Page_075.pro
a73d6e7fcd88d08eecb1289562c0b1d0
4130331b0629a8fea3c0e316ae5d1544be104d4b
738 F20110318_AACFFE mikles_v_Page_163.txt
a9cf7e5de34a548ea9e8afb01c0708d9
778a4b635a4573f0c826b1b1c4c40ae00be0dc67
6632 F20110318_AACGIH mikles_v_Page_076.pro
f034684915007eb007378147c40e5dd3
92711127c7620b920e9be3e04700aaa22c81c15f
32336 F20110318_AACFEQ mikles_v_Page_090.QC.jpg
524d50caf01c2a5a546fc7fa64401d48
3a5d2c15c0e9b5612706bea6b5de4982fb47d715
61566 F20110318_AACGHS mikles_v_Page_032.pro
6f48b029fc2557562e4baabe437b4d45
c4b805a133910899e1325b10698fed14a344cba6
1051957 F20110318_AACFFF mikles_v_Page_117.jp2
10ed358de8520445ccdb81d75d0291ff
4e887fbdc7546dd47c61ec3e37ffa5c95fca433d
17510 F20110318_AACGII mikles_v_Page_078.pro
48a30209286e918f52076bf938740397
1d27789ddc07ef0fc7623de7274637b785e3152b
55553 F20110318_AACFER mikles_v_Page_027.jpg
ceacba1cd95822606eef81a7cfb1dad2
a201500da61f941e69853a5e26ba6f7acf62c199
26991 F20110318_AACGHT mikles_v_Page_037.pro
95f13b110fc198f8d8f9ef6a7b70e26f
047324aa582b9441a9b3279049f9054a59dea18b
F20110318_AACFFG mikles_v_Page_162.tif
0892da0a9a632f71699c9c7ea132c894
e4ce14dde0576279d7453927cd9a13e6efc67c70
54144 F20110318_AACGIJ mikles_v_Page_079.pro
a83c7f17234622075ab808b47fca5d80
86c2703f7f43c9ac5f5a33fd73e40d400371a3e8
58633 F20110318_AACFES mikles_v_Page_157.jpg
18e2c1c9fd03e2e872932ee009c5ed43
90adba5538145a7b5a8f24a260e9f5d1661e486d
59107 F20110318_AACGHU mikles_v_Page_042.pro
57401f9d7cd914ea2988460b0d82b9de
9e19dd71e3b3b5905ced883005c9e2187eefce06
3870 F20110318_AACFFH mikles_v_Page_150thm.jpg
92e892f799cdb15884d6818c87df6397
2e176f6c1f5c8f99f2e63bee19ed4b4fd9a8907c
16578 F20110318_AACGIK mikles_v_Page_082.pro
3fa7482f4e0d948c52bd4df25f8916d9
830724abe59df7a6f28c707994c49472c7942650
2060 F20110318_AACFET mikles_v_Page_038.txt
a162737d39a43b174af3d5093e43bc3d
6a2300bd5bf9f5a046c2a1d6039a39ba73474e60
57374 F20110318_AACGHV mikles_v_Page_043.pro
bf3e29cf07ff8a2595bb4701101de671
04402dd226bffb82f2c5ec896631b7b2019c36c9
2289 F20110318_AACFFI mikles_v_Page_013.txt
4a4a3ce8f841407aadc70be150b51cca
9c75d5192d104feaf0423aeed3a519f9aebb0241
59564 F20110318_AACGIL mikles_v_Page_084.pro
037506cb680d701d9a418fdd07ad5b9d
6cc86db435e8aff0c4082b75ec39cd76a2d14ada
F20110318_AACFEU mikles_v_Page_016thm.jpg
165c420281aaf643abcc167e5340206e
40b4d156e22fc83b58afbe9e3f1b593826e56084
57954 F20110318_AACGHW mikles_v_Page_045.pro
3624f00e958c84f5a50defc47688b48d
4073728d0c89c5fd77efb47ff762b6e57bb06de2
1208 F20110318_AACFFJ mikles_v_Page_002.QC.jpg
b5c2efb2414e9f6527a1862c16ade320
bd99bfa19094ce9ca51b15daafde862cc25da455
56951 F20110318_AACGJA mikles_v_Page_120.pro
21a125303e7009557c2290c876d27cc1
7a823de19d5f775d1bcb30d2365d809caab870e4
58632 F20110318_AACGIM mikles_v_Page_086.pro
25900c011db35a8c88db56e5a096557b
583fc3830fb37a05f46452b182354b50c9b7362a
1883 F20110318_AACFEV mikles_v_Page_004.txt
c4d3fb389a06d09fd9abfd8cfbbe9346
c87c6f426d350fb8a4f31c36671698a47ec60edb
60948 F20110318_AACGHX mikles_v_Page_046.pro
5d331f7af45741739dadad8321dbb18f
c23339e311f9fefde93a31370bcdebb0259b626f
44938 F20110318_AACFFK mikles_v_Page_078.jp2
5c38c946ba0f8b1326241f4af3b3a3f2
ac421e01f1bd588ef61981ae125de448343dc761
59674 F20110318_AACGJB mikles_v_Page_121.pro
15b92fc6da4b9289688952aa98303682
abe572add228791b871568967d2c580baeadeabd
49367 F20110318_AACGIN mikles_v_Page_089.pro
1461d79ae2e4cd88bf7fde1489d986cf
7bf9b7823f6f6fa42a945c15ab1783a508c9be3f
25537 F20110318_AACFEW mikles_v_Page_134.pro
40325f9fca3878cf8547f6af0adecdbd
7d4204f3fa29f54ee9efcfbcc18ae3e172b2adc9
53522 F20110318_AACGHY mikles_v_Page_047.pro
de2f53a133b3dd7cded0cc67fe64743e
269faa31f638c6e05b79949e312ef37c562b078a
119989 F20110318_AACFFL mikles_v_Page_122.jpg
269173ce5432064497db1260df9337d5
8883fa24b8f3dc1c44b97170c8912ad8a2d8f49a
60480 F20110318_AACGJC mikles_v_Page_122.pro
99ef79153f6021678d25a45382aa7560
876096ac1e084ccc99ccf73081ebe2305e277e17
54447 F20110318_AACGIO mikles_v_Page_092.pro
f2b130052aceea13771e70c111a769ce
e19a30912b166abd0742cbec060b8cc039c4efeb
F20110318_AACFEX mikles_v_Page_019.jp2
f430cfe8c53ed8883f04d5cc36c5b07b
00697c63df3b6deec36fa2a6905e71d20d40bdfa
8966 F20110318_AACGHZ mikles_v_Page_048.pro
29348bbdc2f3e45cd96e40c2a4af3202
b9b51a279db97aa38117a0ec78573afb4f9867e8
56309 F20110318_AACFGA mikles_v_Page_100.pro
656d39c4ab428fd4afd328ad62933d9a
6e773075f54de4ebf181d039e1053011fc106807
36389 F20110318_AACFFM mikles_v_Page_030.QC.jpg
63e128b8aad4d1293e5d02868e71bc1b
7b51351a70e3387467bf19248e378dcb97aaec24
61826 F20110318_AACGJD mikles_v_Page_123.pro
a357faf92013badfe3265d70d4c427b4
d67a71144281df30dbe7babdbd64ac11403560a7
34129 F20110318_AACGIP mikles_v_Page_096.pro
02c3675a7a4463a732dbc1b0fcfbc31e
d7854da6e0e52580baa9d84f4d77370f3da5c89b
121026 F20110318_AACFEY mikles_v_Page_143.jpg
48813f7205e2a4787631fd88d955c1d5
2bd96b32a1f59f1363c15dd818f886a6da81ea8b
F20110318_AACFGB mikles_v_Page_074.txt
c8180f1abe50e957e6f7773d83ccb9a8
715c69bd71b28cd8d862d09554723997601ea589
7362 F20110318_AACFFN mikles_v_Page_171thm.jpg
1f2c5703c54fe9a9affc70ec93b351f4
2867b702fe6064541116cdbb485eac7e4fac9ccc
55250 F20110318_AACGJE mikles_v_Page_124.pro
0006d56bb321c14835a1a085e663560c
cb8adb943458f66c09b838c6ad9dec5852142bcd
55314 F20110318_AACGIQ mikles_v_Page_099.pro
80e9ca905543c6e267ab7aab9b2126f7
83cfceb81e3aa780e9a6281b127c21fdec8ac18f
112241 F20110318_AACFEZ mikles_v_Page_062.jpg
53d36ad572e44f0e8f4cb3d267d8e81c
2e174aa4fbe56fd6a3b66f47ffb2440b4fcfff90
37968 F20110318_AACFGC mikles_v_Page_113.QC.jpg
ec4c505323b41420e7b6c2d0e4276bb5
35708a9e4eb84bfad1c79f3aafa4c432548e800d
F20110318_AACFFO mikles_v_Page_136.tif
75f09356d6be6edb1492fc33e9623c1c
34b4b72908080e836d6ae68d023d741cbb739146
20342 F20110318_AACGJF mikles_v_Page_129.pro
252f9b1a63f7920a80ddd5773bde2383
663fdb13769290259d75ad92b391d9cc5f2f6f64
58556 F20110318_AACGIR mikles_v_Page_101.pro
5012ab0e3d8dbbd87c8eb7550fd7eda4
6075b03cd8cf11e5acb3c0ec28eb2fc0eb31f19d
F20110318_AACFGD mikles_v_Page_134.tif
234b98b75a031342c728e3c6dceadfb9
fc35c7e7d64c7fe13ad723a5bc51f524cd690f3f
8124 F20110318_AACFFP mikles_v_Page_146thm.jpg
947a47391e0659fc59775c76269766a9
c63becc2fe3b5f6d3a383c47df3c5ef9635e9b35
35694 F20110318_AACGJG mikles_v_Page_136.pro
24574173b05440418ea1442aa82ca8e7
1ba52d0da07b7d07b12ea126f0463516b46be7e5
57888 F20110318_AACGIS mikles_v_Page_103.pro
1cbecefc9c84827350bf211068f3fcdc
ece3bf9988277afc5e7645b6c5d9c1c4b5ba2055
117261 F20110318_AACFGE mikles_v_Page_145.jpg
ca55a85d9dda3af1e820ed3ddd8fcb43
1fe531df59198ed39465329ca78ccf5ac675679e
34623 F20110318_AACFFQ mikles_v_Page_148.QC.jpg
bb94a63e4b3f2e35af934a5af5b76f99
d29e4575844f77098fb68dd6584983c6aa2a542c
53660 F20110318_AACGJH mikles_v_Page_139.pro
2ece521bb04ab8560b8b1c78a0e4fdf0
762b165ac5c3f2aaacf96eae373c16c4a7c6eb7c
2388 F20110318_AACFGF mikles_v_Page_145.txt
7c10c4801fc266d90fd3470f59c64ca9
2d70031278d7715ebff4b3d1cd9847b2133d478c
59829 F20110318_AACGJI mikles_v_Page_141.pro
50a915838b8b30dd8580d89a9ea7ad2a
9cb940e401c132e53a5aa069d4688cf4e1953c53
9442 F20110318_AACGIT mikles_v_Page_109.pro
f1a8b741b6b049a53fb08135f981921d
54ca1cc5e3b55fefff159fe03dd8c705a1beed84
2328 F20110318_AACFGG mikles_v_Page_140.txt
ea48753a627210ef78ef9d07cd0df7bf
226b07fbb01a8d650d99834a2e5bcaeefc74c0ab
33121 F20110318_AACFFR mikles_v_Page_098.jpg
527cead41a5832a451b0519e4753ff7e
f86afa4a626f66314d2e177e1b08831baa409572
53579 F20110318_AACGJJ mikles_v_Page_142.pro
c001711f013e5ed1b333b925f9478494
cb91d73ef28a1ca34a71ac9b208b658c94669c48
22510 F20110318_AACGIU mikles_v_Page_110.pro
e5172b9f848731377605c92e884c4ea7
7d923ab0143ff39681763173a8abe6ea0fe3ae1a
8330 F20110318_AACFGH mikles_v_Page_022thm.jpg
7f6110f08f15cfcd4f9e4ae86746b224
7af1da56ba25a9d90944d817b71edd24414f7c9d
F20110318_AACFFS mikles_v_Page_161.tif
f0e534ee5a085eabd0306ed5cd596173
3f604262204ee702f2e44c7d36ccc8bfba65f473
62089 F20110318_AACGJK mikles_v_Page_143.pro
f3c9b89e31a8f778698ad8b71e51641d
e7dc60af4c199df29109486d3407862d09943bf7
57396 F20110318_AACGIV mikles_v_Page_112.pro
bc4b013092d0f9b9956ebd5ff3c91806
53342a3108f4a07b87490fc553d567009965851e
457492 F20110318_AACFGI mikles_v_Page_159.jp2
b5a821f2f81442b77393fb1a70f1f9bf
19c52109e07ff378d6544fd6006b70ad61672c71
F20110318_AACFFT mikles_v_Page_070thm.jpg
e68b56dcb88605e3d887bb1056f742cc
2b021c8d29ae2daf60e35d24700c54411c022e09
60712 F20110318_AACGJL mikles_v_Page_144.pro
382145cd8ff3f83eb93c05b5824e51c4
c31ae7d876a591ffb4fcde96e01c03b5dcfb6dad
58362 F20110318_AACGIW mikles_v_Page_115.pro
1810ced89527d72296ec4a72ab14014c
2c6bcb3d75d3bff3ca23606fecd3fbb12cb4a93a
59799 F20110318_AACFGJ mikles_v_Page_044.pro
ce835d001512ed47a12d41bba0fbad68
0a0eb018b4c01a5df604b379f5abea624045c0e0
37669 F20110318_AACFFU mikles_v_Page_051.jp2
42b3c9bd3121a2db1f75ecebe6fda162
26c665f1008ce458fa9030d31f449e1f435837ed
30437 F20110318_AACGKA mikles_v_Page_001.jpg
581069081189f726b14eddeff223b291
4a7679ff9f8d10b5f37b5e24376f082c7c1c543f
55827 F20110318_AACGJM mikles_v_Page_147.pro
b3fdb028d93b68760f95d9aa6779bec1
808910826a74dc1a3344da47373be637e8301ad3
56722 F20110318_AACGIX mikles_v_Page_116.pro
d9edbb4de87ecad22937321bfaa969e8
88161ea6ad2b2b063316e3576b3114c3c34aa512
14928 F20110318_AACFGK mikles_v_Page_130.pro
7b97ea0d588450e84b5aa0892f801c04
813f66047a4d72eda559da452b01b82d9726d376
2576 F20110318_AACFFV mikles_v_Page_152thm.jpg
f1af9f81d8e384023217143b47cf08b1
e9cb77b51acb63a404deefb0f42595ad8046b580
9334 F20110318_AACGKB mikles_v_Page_001.QC.jpg
f069c9e35f60d5faa3788f87afb26c8b
7a4da326b9618736dd69981dca1b92171ed22af2
10278 F20110318_AACGJN mikles_v_Page_150.pro
711c29f677b5037d8336a7031125ca21
a3071abf30fc3684ef0cd641d45c050aa05578e8
58415 F20110318_AACGIY mikles_v_Page_117.pro
0d61d1ba5281688285c8029919346ada
be33530104d9b1e4c765e65a3a0a61ff8ac84aa2
37610 F20110318_AACFGL mikles_v_Page_164.jpg
9e6a7464580111c1ba3329c652e148f5
70dad6ba5e214236ff784f404cc1d128e9a83120
F20110318_AACFFW mikles_v_Page_171.tif
0a8eddce6acd64432350fc5a16a0bbc6
8a2f83745a80c42946bad74f826b7367dfa995d0
4041 F20110318_AACGKC mikles_v_Page_002.jpg
6393b8fd47460e4fadb664205c0fdd42
f8ffb0337a1fa6c9fa98448c9474daa3e45253a9
10076 F20110318_AACGJO mikles_v_Page_152.pro
58fe2a9f1bf7b43790ebff4cef098d6f
2beddcc60c2d24dfc9a666d8defd97714b33b7da
52940 F20110318_AACGIZ mikles_v_Page_119.pro
e3b86926773be0d6c0637f392957e97e
3dfc5e55da447adf27a451b2bd525e05ff3e443e
4223 F20110318_AACFHA mikles_v_Page_106thm.jpg
1607136cc6a17f5443089bde36dce16f
162494b6258c13845136c2c5aa5b18c84882a805
798676 F20110318_AACFGM mikles_v_Page_073.jp2
71df8622a32c10af1915635626fd466d
10d5af0784f1c4be36fa29fcca9d6db2ce161d96
32731 F20110318_AACFFX mikles_v_Page_173.QC.jpg
4b58d3be7f536176060f41a46e15b04b
514e68a2157279afa3bf217bfdc1dbeab0ff56b0
113308 F20110318_AACGKD mikles_v_Page_005.jpg
d8ae7dea8933a4ab312c761ec413300c
ae3f14432e71481383c7c4294f6df260ef5a22f1
16584 F20110318_AACGJP mikles_v_Page_154.pro
2029eba5a1cc2cf6eb96a679d42e77d5
426e6c7118ee51c69b357f94b1191a9096f0f0a4
6429 F20110318_AACFHB mikles_v_Page_094thm.jpg
0dce3580ff6d0aac18eee19d93e04613
9b9881550e41ed1e041d746934e8cd25bf2a6406
56054 F20110318_AACFGN mikles_v_Page_085.pro
a59344dcc8bd3c4f1737d2002d5ca8c9
6330b5853a94b49ab5cdc590c9fa732ee5454959
45564 F20110318_AACFFY mikles_v_Page_056.jp2
a248799f637dfe32adbeee4fea121338
0688e8fd8c7eb3b7d22b2a332aafeb0e57908d6e
33084 F20110318_AACGKE mikles_v_Page_005.QC.jpg
8006d368d2671ba36c6cbc71321324a2
f61f65eecd0457572c8c520d01e8198983034f79
8181 F20110318_AACGJQ mikles_v_Page_155.pro
c1e72060dd9b5786c27ec549f4b110d6
f429f9706fdd24db57572b16e8768e4f4eb87b00
729217 F20110318_AACFHC mikles_v_Page_130.jp2
4ab817c5002c0166964ac5b98c4a997a
6dc3a1d15fac5276ed0360953a0fcc300f016622
8321 F20110318_AACFGO mikles_v_Page_020thm.jpg
5e79e70cedd9227d3d218e6d0bb4a947
1b2211493c713695cf949ced47e8ad349a488af7
102412 F20110318_AACFFZ mikles_v_Page_171.jp2
09c315b1f615abb0fae564e20a52e140
8e1b79d8449a04ed030f0f07858bfd9f0d05ad4e
132952 F20110318_AACGKF mikles_v_Page_006.jpg
359919756418a45d79455018d5e63c27
93e622e83b818a2392da892dccb82725aef0f8bf
36349 F20110318_AACGJR mikles_v_Page_156.pro
02a291064793a9aecccd418c9d38e6a0
c416e345a36b0295303dcf92517d25b223fec9e7
1149 F20110318_AACFHD mikles_v_Page_134.txt
1e81bcfb678a6fe2cc284c277a5fccdb
a9910d0b9e19eac96f750fa234574514b9781580
F20110318_AACFGP mikles_v_Page_116.tif
0dd37ba64482267e281ebe801c9815fc
86f7f1da72043a4fa1a33411608061c4627996dc
74864 F20110318_AACGKG mikles_v_Page_007.jpg
0c66179aab75155d46f3c7eb7319dd6e
d17694251c95a070976be0bcfbe37ca69f0e76c1
19650 F20110318_AACGJS mikles_v_Page_157.pro
54ef4c706c8499e5b8675d3e0b50a4ad
efe448d4dfefd8340ffae118ec6b68b09a182a1e
8624 F20110318_AACFHE mikles_v_Page_113thm.jpg
0c475ced48f290b73e0fa95b462a4f7c
fb717aae04afdfddaef064219e0ae8d2d237457c
67257 F20110318_AACFGQ mikles_v_Page_126.jpg
30f9a7f42749ed56955fa3b579074a06
3f913145363b6d1091dbacb43844829422c3c4d9
21532 F20110318_AACGKH mikles_v_Page_007.QC.jpg
cf82df37ece3751c635fe3f302896041
64465073ede2b60a7afc0317acb8d52bfc3bcad0
15711 F20110318_AACGJT mikles_v_Page_158.pro
b48a61c995899bbb52dec6c21796ad11
f3052853926bedf79f1d15d20784ed98fdd69c4d
37133 F20110318_AACFHF mikles_v_Page_114.QC.jpg
e52032f3f548ebac005633efd30df44a
a702aeb27ea4ebf8c85f80cc3ffc4c164cebca4f
120793 F20110318_AACFGR mikles_v_Page_144.jpg
5de44589e8dd5ff3dd3ee3a637ca5c08
68a4844471ff7a4c26347617ba83f8d6ce51804c
112593 F20110318_AACGKI mikles_v_Page_009.jpg
64864cc1db931d2521e1b3755ae94c34
54fa67771546c6398cdfe76b1fe01ec1917d74b6
611 F20110318_AACFHG mikles_v_Page_127.txt
5bf5d25c1ce103532f04a6bb96223419
1ee6b4991d5a79cedcc1247a67a58785795790f7
103789 F20110318_AACGKJ mikles_v_Page_011.jpg
de7edacdeb1329e0e0ecb361e4bd92be
b8594a28e9dada07e87d1bb7433544676ffaa6f6
18362 F20110318_AACGJU mikles_v_Page_163.pro
97a250d46ebdf65eb4566c7d206389fa
b8007ed2e87e3889594370ac753c78f4ff9f2ee1
52658 F20110318_AACFHH mikles_v_Page_165.pro
e6010f1e275f1c9dc39497dfbb6ef3fc
c1d96075ba04b796175a2e503be446d6f234cce4
2345 F20110318_AACFGS mikles_v_Page_015.txt
48c5c16165f6a72002082ace3facd011
593c88334551ecfbe7b5dd801c8963fd490a8eca
34859 F20110318_AACGKK mikles_v_Page_013.QC.jpg
63a52d161274514b8805feff7383a8eb
c631e778de7b7be704bf05817772d2c8242e6cb7
17625 F20110318_AACGJV mikles_v_Page_164.pro
0a0ab2f1f944f2883727509549bd89c5
df8e7cda4ca524a2a403880bf88fde563ab85095
13247 F20110318_AACFHI mikles_v_Page_076.QC.jpg
ba2bf743daa09641f84fdfe822f623fb
4738aa1cbcd95fae22838415d5532fd4e1b77f9e
548 F20110318_AACFGT mikles_v_Page_024.txt
c27229b984851e023ca8cfc268f87c11
2daafeea0be7b96485a4040e2b97d2ef4fd72594
35798 F20110318_AACGKL mikles_v_Page_014.QC.jpg
8c009b778a4ad95a001a21d3568f4887
8d31e32d65faae1f868829a63e5ce740b384e73c
62848 F20110318_AACGJW mikles_v_Page_166.pro
7aac8ccd767fa07bbc3cc432da5788e7
b47edfe47973db8520f9c7d135d4e84cec96f51e
36891 F20110318_AACFHJ mikles_v_Page_015.QC.jpg
8cd6d6aeb82970c39725048bb0f93c3a
c19c298218cb7190c175544d3992c7cbd8d13116
30771 F20110318_AACFGU mikles_v_Page_011.QC.jpg
5c15463cd813751d2a0bafa6328fea61
6ed4a10b8538ad79ce914aaeac6e2290a84e0446
114748 F20110318_AACGLA mikles_v_Page_039.jpg
2012f07e4e531536bf19b0b9eca49901
7d344376d3d50e0fc5fef05c73a8c597b97ca142
117186 F20110318_AACGKM mikles_v_Page_015.jpg
5acdc6e81c0edc1247d0130da72de121
9e1c25bc4401d5dbb7048af06d9d536f390eb99b
49524 F20110318_AACGJX mikles_v_Page_171.pro
cb9c2ceb180f8ab8c607a5c9de808082
f87a7e9bee52aa814f16d7c7f36df61e36c3070e
F20110318_AACFHK mikles_v_Page_099.jp2
741f622d7c7120cae5be2d98c36c327f
70446c87b4a56df1beb5f8ccc93732410b394d77
F20110318_AACFGV mikles_v_Page_038.jp2
370210a256170666e9b3402db9e75d33
e76d0fb205cf4f21de8e01677de84ddfbedbd97a
35920 F20110318_AACGLB mikles_v_Page_039.QC.jpg
75be1637234e64067334039985180545
145aa3af46d0defc2ec549a31bde00d70d31a855
114031 F20110318_AACGKN mikles_v_Page_016.jpg
fd6d9f6d7665d0b27855339458c935c2
49fc28b23d99a9863af37535d5ee18aa9f88361b
2626 F20110318_AACGJY mikles_v_Page_172.pro
cbb6634fb467b83b69476607c49c24ef
f67e3e6aee5ebe352da0383885f3f958a769bf23
36271 F20110318_AACFHL mikles_v_Page_016.QC.jpg
0265d22f44a4c733246117f33157abe0
2bb3e123e8f12e37669a574bcb837dba80681467
116980 F20110318_AACFGW mikles_v_Page_061.jpg
6b4ea3db7bb310591e3e45412bb3c078
b23591c94c2d0fe9afe10250a926e3ac0bf57da6
114542 F20110318_AACGLC mikles_v_Page_040.jpg
651852ad2179e98710a535a4eb672912
644e192b707ab77620a2d74390565107c3b23c50
114753 F20110318_AACGKO mikles_v_Page_017.jpg
c66b2d76268a72340597a3dad6e2ca42
b33260cabfc793bc141e5b8c98fa9ba023e6c57d
53504 F20110318_AACGJZ mikles_v_Page_173.pro
d0e81ca7017d249abcc4ee42b3ecf1f6
153f09fe147fd678374b9b237888f546a8020fc9
37794 F20110318_AACFHM mikles_v_Page_045.QC.jpg
13c67824bc4249fca6f719465b07a761
1927e620c0e6a77b58d990058ab8a82a93ef11a0
F20110318_AACFGX mikles_v_Page_126.tif
3c00bee916cce2b192cd3b1b3c36e2b7
abab344fe990ccc2ba914dad93db236401e8dd8e
F20110318_AACFIA mikles_v_Page_043.jp2
650d6b0ac8d7d9e4c65db00ad335c95e
2ca20181687fb62d43f2a9494fbc4662bc9d19ae
36207 F20110318_AACGLD mikles_v_Page_040.QC.jpg
4756e26ea6d36e84eb5b682e39d9d8d3
4bd66a2ddbb0142f8ca0a142f15d188aac0092cf
115649 F20110318_AACGKP mikles_v_Page_020.jpg
82584896ce72a335874dbc4b80baf5df
5d7ddb75740e6d67d2db24cc332757aa5101250a
58764 F20110318_AACFGY mikles_v_Page_009.pro
5512dcef33ab8c4e319b3e12cfde8502
a7b4ead6268ef697d42249ff039c20cf6b5cc8f5
1051948 F20110318_AACFIB mikles_v_Page_071.jp2
2916866cdeec4e07736d6e0a037c1974
69464cb8dbd673a6cd3f186ce211ba0697b22e8e
6787 F20110318_AACFHN mikles_v_Page_104thm.jpg
82b02ac720ca47d87d3e254a38c53da1
5931373d77416d30011919d721507234a5b015da
116893 F20110318_AACGLE mikles_v_Page_042.jpg
41031155734f5e92d599d6debeb619ad
481096d679ef4314d8a52092502ab362233f4f1e
119996 F20110318_AACGKQ mikles_v_Page_021.jpg
12b852fc012ff44d41d27de21d8932cd
029751f9af561477c9fde253b3c49610271191e8
2412 F20110318_AACFGZ mikles_v_Page_032.txt
0d31f8ddd6d498f8dbd7058badbfb21a
48ab735bcd5c9d332e6b96ca88419d6da32010a0
31396 F20110318_AACFIC mikles_v_Page_089.QC.jpg
73b1d76f3dc114ba5c4d7a368a2f9601
ffdb2672e3756b528362941207aad4454086c1ab
11796 F20110318_AACFHO mikles_v_Page_077.pro
674a4f43b439238cf9ea1aa2dcd4ae18
a9eedfd9818fab35b8eb5aeb514b6e1958a9b5f1
36786 F20110318_AACGLF mikles_v_Page_042.QC.jpg
1f0aada5bbf6c2119de3bcc8c026f1e0
0e58632b2eb05f9c58cbb089888a9ac3a03ae0e5
37132 F20110318_AACGKR mikles_v_Page_021.QC.jpg
4c9568e684a01c982d524c9ac54dc470
c070fe6562b51892f5a83f0fbc741965f8295371
F20110318_AACFID mikles_v_Page_145.tif
cd1705d2278ad3a00b549a54c88c9544
10bc519f2a7e95c48a3356146ba813796656b189
117388 F20110318_AACFHP mikles_v_Page_028.jpg
0d9a20251284967035c349d589827cba
f734d6d98f16e72f1666326b3cbd45603e5e16fb
120812 F20110318_AACGLG mikles_v_Page_044.jpg
b3233264e795b80b3f4c35e0beefcb72
5080ca1970603b2e7bb9e2e710e1f3d40846a51c
113565 F20110318_AACGKS mikles_v_Page_022.jpg
685c2d1c56d45093bc0a32996b0e201c
3e7ebbf4aee911490e2190ad27f0430f0c32178f
37320 F20110318_AACFIE mikles_v_Page_068.QC.jpg
9f6a8e7daa65c78dd60a8483b14dcdb5
2d5a85a6441b62e23bc3f2f3b760a6ec519fc54b
F20110318_AACFHQ mikles_v_Page_016.jp2
2236b980d500953b1323cd7b34552b9a
af41547b05e238eee0fec054c72380d40b3d3371
33681 F20110318_AACGLH mikles_v_Page_047.QC.jpg
911d33966b32e94e824d653c349c569f
50b6957015dfdc290fa6157af30d814e5e197b01
61982 F20110318_AACGKT mikles_v_Page_024.jpg
26027fb99de82fc2bc4ac0109b248307
b4f71b7d041231fd519474cfc331cb0b7dc963d4
1239 F20110318_AACFIF mikles_v_Page_057.txt
5e5c7238ff12cfd6b970bf7c9fb6458b
9dc09c0199225595a2388891465fbd12aa737dd1
4480 F20110318_AACFHR mikles_v_Page_007thm.jpg
06ee5f3ab348bb5982a8d3dddad3047f
0bf18d77e1f76fc82d0f4c4e2ee6f84ea20e34af
14961 F20110318_AACGLI mikles_v_Page_048.QC.jpg
4515f43d5545e739ddf2901612602a61
6620b8bb064d501eaa4a926ebc129c74bb36172d
19484 F20110318_AACGKU mikles_v_Page_024.QC.jpg
bd3b6d303bb436347b4fc394e9feb25c
4afc9821b3d5b65e0c02e63bb7816a1e5cf2d25d
2256 F20110318_AACFIG mikles_v_Page_102.txt
db618b1d966f11561ac548af274d6cff
5bb738ddebd5a87c4cf433ac9cbea61025f4b810
20456 F20110318_AACFHS mikles_v_Page_096.QC.jpg
f6934711dca6523f2035a9a158b98781
6e8b7f750d11d9345032a132aeabfa910da4de22
14949 F20110318_AACGLJ mikles_v_Page_050.QC.jpg
ff13a159138c5124bc66d0bc8c053008
9a3906d5a4d16b0ae3cb0d08b4dacf30f8d10783
12991 F20110318_AACFIH mikles_v_Page_098.pro
8c2a33b50de19a1b5a399fc443e9ccd9
d1bfd666d5ea0358057e504d80fea83497037fad
18016 F20110318_AACGLK mikles_v_Page_053.QC.jpg
7af1c6d9cd5fc4690d081e2f21c84591
c652623cfa70affaf0b5f261c6b301991b70a73b
37702 F20110318_AACGKV mikles_v_Page_029.QC.jpg
12bb4e72f5451eb5abe4567e671a3c49
353d9a894cc764aec033c0e5bc1eb59dc21305aa
F20110318_AACFII mikles_v_Page_124.tif
cb07ecdbf9dbe0672daec0648275a5f0
2d31449fa0a133c0355917215592cadbe504520a
4082 F20110318_AACFHT mikles_v_Page_149thm.jpg
9648048a062fc5a358ed99d523104791
dd9ae9b55652ed40de56e75d0cfd4fb884d1c460
71223 F20110318_AACGLL mikles_v_Page_054.jpg
e25e4eee92805aaa157fae9324d794ad
f5e31ccffa5775e8247c6eee90610ad1566c9d60
36888 F20110318_AACGKW mikles_v_Page_031.QC.jpg
dc8ab7d1b8055d91d4db744eb6450f20
8b88f0291d180ccb63a291c87026468e353fef80
28287 F20110318_AACFIJ mikles_v_Page_054.pro
5cf9561a58bf08729e7f5a311c61141d
b1db732727d66c73146ac0cb97d915b266d7d826
34862 F20110318_AACFHU mikles_v_Page_100.QC.jpg
a15e5b7f24b904dfda357b1862342361
26509ba81588c42ffca15912c98d86754da8ece4
20901 F20110318_AACGMA mikles_v_Page_072.QC.jpg
84bd1b8892267a74fde49d22758c3621
8bdc8713620e0d769baeefa8d6610c729aa18be9
20170 F20110318_AACGLM mikles_v_Page_054.QC.jpg
d518765d147cb4ee06673f04391e7fc3
f38fc5d39415d8c74aebb9c533c8e76dba9db9c4
115267 F20110318_AACGKX mikles_v_Page_033.jpg
c5bbad3a53b269822d7c15bc3d889097
2fa337e80401c54d8da5c12bf4ef45de15b33fc3
70205 F20110318_AACFIK mikles_v_Page_057.jpg
940479cced5f83fa2c48d86b42bc3c52
c3d811e053f04a0547d700191062b6a4bf3281c0
102924 F20110318_AACFHV mikles_v_Page_071.jpg
feff41f87310f57bf4c93e8496bec1f0
158bd709b077bb6f6c83e0be9d4e3b525508b98e
56384 F20110318_AACGMB mikles_v_Page_074.jpg
9f9fdfdba297d00e14eb2963187ca377
e073a746647d5fc3180f4e33f347f7c0a1239395
32238 F20110318_AACGLN mikles_v_Page_055.jpg
c7c3c28812ecb82e6f2bbdaf26590f19
f27464a393a9f58a7bdf3d48633c7a5347c13e5e
36899 F20110318_AACGKY mikles_v_Page_033.QC.jpg
be934ad3c376871574fa27eae86adb05
07be731cb277de8f7ab59687fe353c68fb17ffb7
61626 F20110318_AACFIL mikles_v_Page_149.jp2
6fcc6ff214058fee8067ab32536e602b
76e1cba072749b7e5d4278d6ff4a984c95a46b1a
7956 F20110318_AACFHW mikles_v_Page_018thm.jpg
3c9ee96391a18ecba7f61ffe02552d5b
5b70a8dc3db77772b8b3623f8d7329db8be919a9
19115 F20110318_AACGMC mikles_v_Page_074.QC.jpg
138c7007fa5231e6f61b2c1cb8cea317
79034d35134c877cdfae5916f059ad9891785947
9133 F20110318_AACGLO mikles_v_Page_055.QC.jpg
4e04cd5cbda0d41747bbe8e19bdf6d7e
2a349f1e5bc918f5129e1291f49ecd2f87a6cabb
31954 F20110318_AACGKZ mikles_v_Page_038.QC.jpg
098402c168c6391b536221d4ab9c6935
753cd066e5733019d208c06c530a58ce08829676
111535 F20110318_AACFJA mikles_v_Page_047.jp2
fc93fdaa6aaa1fa7b2431510b2a7ae55
20a75ad032e0f65383c8b89c3483f35e5e8cfa90
121655 F20110318_AACFIM mikles_v_Page_161.jp2
61ccc901ec3af01897156607cf1e73f8
13e62d4938d1f3b871ad493f903c528135e46bb0
1713 F20110318_AACFHX mikles_v_Page_003thm.jpg
b27179152c001d8757a2664be2d611cc
9b8123173b6ec489b460aac30b5e731af041f07e
34442 F20110318_AACGMD mikles_v_Page_076.jpg
43d756cac63d9d8623b0e490f572063b
e5bd1f4623dfd48746c6f4239672d6c26f3d1f6a
42753 F20110318_AACGLP mikles_v_Page_056.jpg
338e40475f30ff9f0610bd13d7fbfa24
87bf1bfda494553fe837856a5873b9d07d3ef8b8
43652 F20110318_AACFJB mikles_v_Page_095.jpg
21456fd40b90002cb80f7921d687c3a8
413508de6136122a037d9a1f331d2bd5443ba31d
F20110318_AACFIN mikles_v_Page_103.jp2
83409f38b332ee69ca53974b3fbfb202
00c735cecccb060e15fc9d0decb22a672283692d
8530 F20110318_AACFHY mikles_v_Page_112thm.jpg
09a941c89f8974ff52ed4c4aad929562
494dc23ad28a6c653a31c2bcfc711ddb32f2dbdc
53384 F20110318_AACGME mikles_v_Page_077.jpg
0aa7fd90d5e43a624a5901c36c803903
2b0621b461b8f29aec3c480e98cdb2682e413472
12990 F20110318_AACGLQ mikles_v_Page_056.QC.jpg
c1af9a0b9c9bb4202e950b032d502336
5d213593c45ab3f7b8dcd8ba6368b210bddab1d7
34750 F20110318_AACFJC mikles_v_Page_154.jpg
ecb452191dce913d6693290327bc3c52
db157aa59da79a2316a77bf9aba35ebcd287eee1
37627 F20110318_AACFIO mikles_v_Page_081.pro
c4b43764a9c43d7df1a97f72c6e4871d
3bb0236bec31923be82eab2f6aac6e46c542d4e0
F20110318_AACFHZ mikles_v_Page_118.jp2
2126759a239c1ad2aa9dec47499b437f
a8b2291690c3484f3ca3b59e3f130b2fb0336354
105998 F20110318_AACGMF mikles_v_Page_079.jpg
adb1aa532ee649921e7a076659beb2ec
330c945c7a10a8aaaab87b066b01a02a05a7b958
20384 F20110318_AACGLR mikles_v_Page_057.QC.jpg
da93021017e856a37fd0df94bc419cc8
40e640c4a374494d2e1f81380ed674192b20c829
F20110318_AACFJD mikles_v_Page_146.tif
eeb713d5f3dfecf59fde09659ff18197
3e8467394c572be0526a0aa392a30a24ab78f239
7972 F20110318_AACFIP mikles_v_Page_147thm.jpg
54f5755668190d481e0fb26a36353d61
cf2044b960f83925905736f1a08cc034ff345ced
27881 F20110318_AACGMG mikles_v_Page_079.QC.jpg
b73d3ef934c26a0f890796137409a98d
e105bfd7645dbeafbbc28e77747578658f6e2208
34272 F20110318_AACGLS mikles_v_Page_059.QC.jpg
33353ba109382797246b1263fb1200e7
571e208a8628e254b79512fe03792dc447dbb507
5993 F20110318_AACFJE mikles_v_Page_003.QC.jpg
1c34577f02b0fd36397caa48252a8c28
0e1f5d0f8f9ec81fbca6b9b1054f843ddfb58a55
F20110318_AACFIQ mikles_v_Page_115.txt
cdc9d520e5b4b8cfeaa7bc76b68c98f6
a47262666be6112e79cfd6b9440c8ba0ea6e4bb2
64653 F20110318_AACGMH mikles_v_Page_080.jpg
3a2bd76bfefc431b4a79465a004c3553
0970650e2f02cee116ce2a06e6f9955b95d2d3ad
36545 F20110318_AACGLT mikles_v_Page_061.QC.jpg
af7aae6f317f9955cb134c6c959e58d1
843142acda8ad6432f22c4fa78f26d9fe0318d2d
29237 F20110318_AACFJF mikles_v_Page_066.QC.jpg
6b9ff57297e5b62317c3a808164573b6
a3d66c406fb51e4bf4f4bcf0de577d9fe46f0674
7447 F20110318_AACFIR mikles_v_Page_091thm.jpg
7dd3fd560856fde8c3a9f8f79ba331a2
9d519a31a8622ceaa771ac49b5b9099045554149
84438 F20110318_AACGMI mikles_v_Page_081.jpg
92060a98359c0e0d29ee4eaca92e4bc5
75c0d183d3150ea0fae919342f9578302ba779a5
35663 F20110318_AACGLU mikles_v_Page_062.QC.jpg
088fc1356d3513972ea96b5060672ef3
5d85a839fb3f5658a5b6efb634b7f185b60388ee
36365 F20110318_AACFJG mikles_v_Page_017.QC.jpg
9c4d6a3f6ee72cd893835669856df1d1
977da39c8e9893b833bbf42ad8e68c9f1cb96fa5
35305 F20110318_AACFIS mikles_v_Page_085.QC.jpg
8a0b11b7723492b3a43ea78a679a0b41
e6e9b32b41a2aeb8b35e533267d33f20c58c3e00
46035 F20110318_AACGMJ mikles_v_Page_082.jpg
d8b3b4a6f0460cf7d88295940a66b745
e07c62a84644ab96570578c357e001e19efa4594
36937 F20110318_AACGLV mikles_v_Page_064.QC.jpg
6cb3a82d371f2cc275fa538d3d8861b6
23562db4fa9f6e4c060400e0db5115e9f9f8e3d1
52353 F20110318_AACFJH mikles_v_Page_053.jpg
32e6dcaa7bbf8513b1664074025324ec
07fe4aa022500cf9ab71902d29a67b1e910edbd1
2180 F20110318_AACFIT mikles_v_Page_067.txt
a3131222b6e796165cae08428ef3a7d3
bfd4c428f84d20d3b0591268f043e96659d91a66
15168 F20110318_AACGMK mikles_v_Page_083.QC.jpg
5a82948c37ebfaa7e9abf2c93de50c76
6a7f6c046285f69f51bf297c705eb39d856a92ed
8608 F20110318_AACFJI mikles_v_Page_101thm.jpg
1b7b036e94a6b2935a408e4937409f4b
b7ef510eb21cc99fba68b4f96aacbb2343f99895
118991 F20110318_AACGML mikles_v_Page_084.jpg
2039fe571412fd3adda57406cb62c211
0d2e9d1be4d9f1d4023b3bb84787cf365637c6fb
87520 F20110318_AACGLW mikles_v_Page_066.jpg
2450b18f95d736ef3b4a3ac8bbc087bd
9515e3fd368f29fdb0617562dfee52e275a60a61
32109 F20110318_AACFJJ mikles_v_Page_098.jp2
10acf7b8077c075fc54d8e4c9c5130f7
839a5a4842032b680eaac3147b3e68b352a9e429
47910 F20110318_AACFIU mikles_v_Page_159.jpg
7d70e26947e57d714b09d208533147ef
e3ab59d7539ed62a881a078860a578f951645e49
14773 F20110318_AACGNA mikles_v_Page_107.QC.jpg
0e8ca8658703dab3693bf442266b5e2d
e7858c0dd9e02d91d6aa136003cf36aff57a2f6e
38263 F20110318_AACGMM mikles_v_Page_084.QC.jpg
7d39807d68538bde96a37e266a17ab74
202b4a609548e2f64397b2460c1ac0172454a634
34199 F20110318_AACGLX mikles_v_Page_067.QC.jpg
42bc2578fb08332d6ae7eae80a1fbd4e
2877e158d852595b0e610e7836001c0d4e0ddd66
123299 F20110318_AACFJK mikles_v_Page_036.jpg
03a0648db9bd96f19c98a298ebb2b170
930881a8e4b5e15e0b950e005d6cf948c7aaebe0
35833 F20110318_AACFIV mikles_v_Page_106.jpg
046158f972a3f4c3319197eb448456b0
ed8f158c7d1e4753b1e76bd82ae663442bb2a2ed
66799 F20110318_AACGNB mikles_v_Page_108.jpg
65ff47ab73f8761db8218490c84ecfe7
3e243b54224c011898aa30268edd483dc53be577
112032 F20110318_AACGMN mikles_v_Page_085.jpg
90c865981a73d19bb87d2d56e7967ddf
8a4da5a512c84ba8792f372041887c453c576462
88848 F20110318_AACGLY mikles_v_Page_069.jpg
3bae4f7e37a9b605a894369a79a8f860
441276b959ed24d24a8dc7ddaf4e469e0990bd22
5641 F20110318_AACFJL mikles_v_Page_053thm.jpg
90c9fb7d74dd8d24e3974291371dbaf8
39c2c0d95120d4573a340e13332eef883731d7fa
28625 F20110318_AACFIW mikles_v_Page_058.QC.jpg
b6fbbdd09ddcc305b7b303636c871ec7
b88b6a4e0a419064498d53142920098d7ed227d0
22950 F20110318_AACGNC mikles_v_Page_108.QC.jpg
0abe62c758c02abd50a78608980a1939
0491bda9fde73e6ccaeaf9e7a2fa5324383f0be3
117460 F20110318_AACGMO mikles_v_Page_086.jpg
118bbc06c3c18288235042fcc6fa6cc9
f9b4238af5350c62fb725e86a083b3ba3fb36478
114881 F20110318_AACGLZ mikles_v_Page_070.jpg
32d6b1538de1f9fba0c61d29022e76e5
24b3017df9dd004e8774f01d6778707a0118d3d4
491 F20110318_AACFJM mikles_v_Page_001.txt
1a0127b8f79be80b14d07a80ba87ca40
b011b9a0dd8b3e6a272ab800aaa6a460212ac959
57794 F20110318_AACFIX mikles_v_Page_020.pro
513e1218a4b5b90fd71273bf56031fd2
4cf9fb2c2fc32bf422d4f781f1156419b4871fdf
29560 F20110318_AACFKA mikles_v_Page_049.jp2
f2dd7aac4662dcf8fa7b4a58ec5da2e2
7356bcf176d2035da7a9d0e8f85e9deb61a1f25a
95948 F20110318_AACGND mikles_v_Page_109.jpg
30cab5b2873920975534b35264594073
56369e88f26b235305263fd57082ece1bc43899d
36460 F20110318_AACGMP mikles_v_Page_086.QC.jpg
0aee281a4ee825fbf02f597e5f57199e
1bd49455b47aad879185d52023222d4e41283535
7908 F20110318_AACFJN mikles_v_Page_089thm.jpg
aa4d03a8f36c26936591a1bc9e404227
76fd7105de15dc8bd691a432ea19d04a0ff74661
2164 F20110318_AACFIY mikles_v_Page_069.txt
56f4386c4fe91c5ba1b6702f4b8cede9
7c016f7ed736b75f4239d9051c7906150fd66c64
F20110318_AACFKB mikles_v_Page_061.txt
399d26ba484eab3eefde0123268596a0
f3e6a3ad2536cea2ee795701a9d2c6072e94b693
93886 F20110318_AACGNE mikles_v_Page_110.jpg
becb192c416cf99245ffd1e0c17c2bfe
363aa1a8a9e222bdab55f15256482b427169fb1b
32827 F20110318_AACGMQ mikles_v_Page_087.QC.jpg
8f3dadd379500e678326af3b2b054f10
d31e7e61d4ea8028502611f8bf8db72359b67aa0
F20110318_AACFJO mikles_v_Page_101.txt
2851ba86d09815a83aca24d1be453761
90a10d97355ace893274fbb9528d57ab1136078f
2219 F20110318_AACFIZ mikles_v_Page_018.txt
c92c2eb579c7a384d43c595adabd3e70
df477988da31ab461c478e9a2434f531d2d3493f
8351 F20110318_AACFKC mikles_v_Page_028thm.jpg
e061527262b20e1fc550f21aedb5fc5f
01e4b2c3e7fb68573bb232eeb12233658b0b0e32
27531 F20110318_AACGNF mikles_v_Page_110.QC.jpg
5f090bde59cb82db40b8c43d9609986a
80a039c804e69ae40ce2c9e19eab26cc42b20d70
102544 F20110318_AACGMR mikles_v_Page_090.jpg
49df769c4ac553c097d4db4698e3e63e
7268b7c0d4e70b6f9b5119d2694ae692115ad0e3
F20110318_AACFJP mikles_v_Page_062.jp2
7e433d428649f79e958873a933faf1dc
36d7b485f70277767e171b70fb0b2243a000f7a5
12008 F20110318_AACFKD mikles_v_Page_078.QC.jpg
e73050c88b7d9876fc374d333f4e573d
ef12e5fb7676e38967e486818f449aa16cac599d
117903 F20110318_AACGNG mikles_v_Page_112.jpg
307490451e896d0ace36c0830137769d
02ea4b2609cc1a51b3b6d5208cc0efd1744faecc
33906 F20110318_AACGMS mikles_v_Page_093.QC.jpg
b4b19a621e536be3d63d3a6151eb9d74
ea278901ed4fd33bd00aeef63da027c60f33fcbc
111339 F20110318_AACFJQ mikles_v_Page_116.jpg
93f2a0ff10618594b8a313f5135dd0ef
b999753bc457d2ccf5861d7ba505c6a52307ad9e
32819 F20110318_AACFKE mikles_v_Page_019.QC.jpg
4243f6a27d3aee0796547da736a61ed6
21f4c4bb67bf9cb8e2d07cc3b5a2cae3c14664a7
36549 F20110318_AACGNH mikles_v_Page_112.QC.jpg
f873b67778f3f6ec21b3c6cf73bad191
2f4d9ef4e35b98f55e7e270daef01707f75d24df
87148 F20110318_AACGMT mikles_v_Page_094.jpg
778093fff247a60760537aa4db17c746
7ad6f4a58c4727d6991307a0e2acd17fc0f841c5
13658 F20110318_AACFJR mikles_v_Page_132.pro
a48ce48d1eedae2e164b078dab2b4016
0d7c5667ae97dcacf99d52029713bb4999238c0e
10772 F20110318_AACFKF mikles_v_Page_035.pro
354511b2329528c9de2f3827a13f8a02
951bea4be52bbfc492ba826eac3e02ed995f2194
117675 F20110318_AACGNI mikles_v_Page_114.jpg
a20c15247a3c2939198b2a18bd399cd6
f42647fff3ea0b6522adbd28a6a3eb55fe959ae3
27716 F20110318_AACGMU mikles_v_Page_094.QC.jpg
6a66de4ef6ab04e33efb9ef4ca913020
0b614e1a964b0df32d601cbca42f9a9e0bb06370
1852 F20110318_AACFJS mikles_v_Page_008.txt
93781306dcedf16fabbf3fd5bd696431
1a92c5a4bbd4041560657f1f82dc52c4e553f113
1363 F20110318_AACFKG mikles_v_Page_126.txt
32039ebc597b2c01dfb9e33ccbcf0ffb
0e3c7b839b39f8ce28e4adac768c78533f042d8c
35201 F20110318_AACGNJ mikles_v_Page_116.QC.jpg
e9db994413ea0f01e26946bde9de439f
eeb3c63173e214ef540f8e47d3a206fc14631731
74202 F20110318_AACGMV mikles_v_Page_096.jpg
4f1d162af802c240831877bf5e15a299
e96b3e13f60e40bac7664e1973c280315473391d
1051965 F20110318_AACFJT mikles_v_Page_064.jp2
051dd7421746ce790a7c6061f3b0ad09
3254b4348521b1f55923c714f9ffdac9ac44e02e
27930 F20110318_AACFKH mikles_v_Page_111.pro
7be9398a86a12de9ea26a9cb8200e539
2ada9b2121919b659fcbd2700a1e8e41db1183f0
35869 F20110318_AACGNK mikles_v_Page_117.QC.jpg
82c85ce59969cd24c558edc55dcb9c94
f7030153d27a8e9f014ee2a455e1ab6ea008ab5f
113916 F20110318_AACGMW mikles_v_Page_100.jpg
4ad881eec4ffb248926696f7ee9c9751
dec3268a79385af88cd90acfc3368ef9c484a795
112054 F20110318_AACFJU mikles_v_Page_099.jpg
8b0985598790d3a5d317c11702b7a69b
a13790f46711fb42d46398b377acbdbb112876c2
749 F20110318_AACFKI mikles_v_Page_082.txt
39e09b50b3aa68cdd76dc3c5fbd411af
1fc518081d3bf2c2d7826c2f3e8072ef33ef59a7
35772 F20110318_AACGNL mikles_v_Page_118.QC.jpg
41fde5fdbe157c19c198c824ec3e459b
e84086eee6b5c80a2f04ee7030a615aacccb4f55
2349 F20110318_AACFKJ mikles_v_Page_042.txt
4313814227463efbc5f4cf37d902f34e
bed7fa1c9e10b08079a7cfce83404a9956059113
74920 F20110318_AACGOA mikles_v_Page_134.jpg
e196de10de7876d1fdae2a0dae57f7bd
899dca4e460ec0426b347139960a50bfd89cca96
F20110318_AACGNM mikles_v_Page_120.jpg
0d74f4b0e335ec6d99804d7b7c325d21
b2762512ada915a5b4f1c96c88632082e7c2ca9a
117706 F20110318_AACGMX mikles_v_Page_101.jpg
fad51e5694c4ddf5d0dc5e63d95e023b
5cc573d8e420b84749e27ca87e2c4132f0d715b4
5800 F20110318_AACFJV mikles_v_Page_157thm.jpg
0a3a8c39101121359af61e5311075289
c167583a7e1268752b6e60dcd48d22b061fcbe46
F20110318_AACFKK mikles_v_Page_058.jp2
ffa78a6112c2cd9f08a93ffc734ca574
47163901905b61f109bbf303437c1cf69e4d3cf1
22571 F20110318_AACGOB mikles_v_Page_134.QC.jpg
b8169f5cf1c2a03b025bf6ae6eb6d235
d69f12d503dd610430c7a646df16ee575caaea91
117333 F20110318_AACGNN mikles_v_Page_121.jpg
d7bf8734d4349faf719a4364dd686d1b
0bc6acdb59525aa5bab12bb3df84746257266cdc
36367 F20110318_AACGMY mikles_v_Page_102.QC.jpg
795692de5fe28713393e3d42958c1997
ed2e6d8b3ce4320be81abbef3c0045dd6a052835
7451 F20110318_AACFJW mikles_v_Page_038thm.jpg
761bc441969deae5b0302c12e2ae5905
75e7cbd31fc05986b524f1119389e687e86591f4
5623 F20110318_AACFKL mikles_v_Page_023thm.jpg
9282f3dfdc055d0f89299057a0ece033
3b325190a541c4cadd577b7b65eba4f25fe0a73b
66605 F20110318_AACGOC mikles_v_Page_135.jpg
b6d4b5ac1f8fa5320662c7855fae1236
413e67f903bcdc73c490969e9ce3d11042a3de4e
37414 F20110318_AACGNO mikles_v_Page_121.QC.jpg
76ac11a98ef3bd321760fb70fecc5092
c52b44391a035f5aa3ebe79f7e96e744d4792541
114738 F20110318_AACGMZ mikles_v_Page_103.jpg
6357688a5b96d87e78be59b36c23b1cc
87601350ae3172398097e1edc66429d7c4a7efd3
4486 F20110318_AACFJX mikles_v_Page_034thm.jpg
300699014ed2fbab017a2401ddeea885
fd1c529389da4813b74eb07e7afb0086db027401
51459 F20110318_AACFLA mikles_v_Page_052.jpg
30ef2a4348b3e19c745d9cc7ffe391b2
99f933f3dfd4c49be362a9cf2d08a8fc91a5d957
111816 F20110318_AACFKM mikles_v_Page_063.jpg
c291c09d98663c41a6cd5ca3f58261d1
5db5a7e5b33088c7a14913695b6845eb2de4c654
21662 F20110318_AACGOD mikles_v_Page_136.QC.jpg
d2a0cb698789d101b9497de67442f140
fa24103698943e255649c3e1b36a07f59f7657d0
120674 F20110318_AACGNP mikles_v_Page_123.jpg
e5131090a3d2cfd4703c26981833c95b
9448a354982e953c49bc7979f07b385b3f1a7e2c
7579 F20110318_AACFJY mikles_v_Page_173thm.jpg
514960f5b666b377d939778f86172c39
64f1db7b600a345ab093d5a2cc2d3b7cf5b22850
63951 F20110318_AACFLB mikles_v_Page_135.jp2
5e55ca893b58f2deb29cdf00b0459be3
e476d72e9595ac2b09845424f14d8483f2773808
45842 F20110318_AACFKN mikles_v_Page_069.pro
f92a2ffc2a42841bafacf84ceca92307
7e61a370108eadecb0a47e091de3c80c673fa120
60140 F20110318_AACGOE mikles_v_Page_137.jpg
f317837d20637376c78650ecb157413f
3c6b64da6510d05ae30a52caef87bd2b47de5d72
38039 F20110318_AACGNQ mikles_v_Page_123.QC.jpg
79f93092291a8d5ee83d2d4225b3d2ae
8e67fa067df58595d98a3c82a8a803c26b6c5b3e
96765 F20110318_AACFJZ mikles_v_Page_065.jpg
359f61ce02854b00e34810af7ef75306
ef6807cce3454c9327a1f18889ec19ef7dab0245
30734 F20110318_AACFLC mikles_v_Page_035.QC.jpg
9a748d9be8c9c5d7ecc8910d94dc9ab3
94f2ca8e0c4c721adf1a05d834135f005e4c087c
8386 F20110318_AACFKO mikles_v_Page_105.pro
3589ad2039227617df95dc5f179efac3
c350ff1537ccb64ad0e304dde3ad4585e1c5e5ba
34144 F20110318_AACGOF mikles_v_Page_139.QC.jpg
947eb19bff992ea7b32d90adc6a0e8a6
ea8596b63a79334e4b5de8f0f17b29ba298a2d55
35220 F20110318_AACGNR mikles_v_Page_124.QC.jpg
7beadbec0531d03dbd050ffedb2fb6d4
834dbe1bd304e7cc683097bb629223310f2e5453
937149 F20110318_AACFLD mikles_v_Page_094.jp2
c4499e2fa20205dca741ee2fcdd2ec2a
605ae96ebf771efafb59ef7e3dff9e8850cba544
30381 F20110318_AACFKP mikles_v_Page_135.pro
735eb17ac9adc3c85311a3f84a8788a9
8709c8d9aceda12f2fc296f23e243bef06222b3d
37207 F20110318_AACGOG mikles_v_Page_140.QC.jpg
9dd510c6f2cf78e01b7036185f1d2c94
7b2c0e95b317e35d92a107be1d91ec8332ab73ab
27195 F20110318_AACGNS mikles_v_Page_125.QC.jpg
c0c5b774e86bdba09b6f03058deb5319
7c65bccccfa6c3063b9a03c22270cf0672a15008
F20110318_AACFLE mikles_v_Page_124.jp2
bfe4c97819696a670c44a6c32ca10012
ae27f21234c62e5632f12e2b17b2fab63d0f9df4
29678 F20110318_AACFKQ mikles_v_Page_128.pro
16b551e0f1d3769c81231b11c21301e8
de91ae13c8ec507c656309fb8fdd38c7669358c9
107765 F20110318_AACGOH mikles_v_Page_142.jpg
de6ba8a716386b882fee28a6f2313791
484c23a6a2786b58e7f78ef35980ab188e24493c
21410 F20110318_AACGNT mikles_v_Page_126.QC.jpg
fb5d78c3f2c0a086f4502c2727470679
c00cdcc7dca197a76fee4915f6ebdb8f7167f1a6
1584 F20110318_AACFLF mikles_v_Page_023.txt
16e24faeb3061f690f6830aed82d5304
0754cf395c47f6ec22cdc16e211222b79a0f132b
2291 F20110318_AACFKR mikles_v_Page_045.txt
1a3705409d531e64acfe296020377151
7b60bfbf5f744f68bc2474ebb3cf5f8cba89badb
33776 F20110318_AACGOI mikles_v_Page_142.QC.jpg
f9c1e9c1f8b9aaddcc8d46f406307f34
3c30007a7d049a4259f4212b65d56d17261f7ae3
21093 F20110318_AACGNU mikles_v_Page_127.QC.jpg
5195bf517380da6c5c451563ec75d211
a144fe7569e9f2a93a187d6aa5bbe6ae8374de7d
45331 F20110318_AACFLG mikles_v_Page_091.pro
d0ef0b994f96a265dcf4efc6ac14f0c7
8d538ad0c46046d3b0c02eb72e51caa8a409fdac
8023 F20110318_AACFKS mikles_v_Page_100thm.jpg
37e62d09fea75b702a9af26d94ea41ea
e30dbe2e3bf7df73b5d8e6e25444eb9fcff07104
111542 F20110318_AACGOJ mikles_v_Page_146.jpg
1c80be0252101df76a78f12c3e593b74
fccdb283c815399b6d811b926b0182eea0102224
65345 F20110318_AACGNV mikles_v_Page_129.jpg
4804b969d61b01bfc7dfc3c38f09ef0c
1d069019c5ec3adb6bd114572d66bfb880b42811
50210 F20110318_AACFLH mikles_v_Page_170.pro
c9d5ca58e3b2e6e4c8eea8fb6248e349
65ab5c5b35b05f961a1e7046c97c89d07ef50cf5
F20110318_AACFKT mikles_v_Page_101.jp2
092b27016f0ad5c4821eec5ee903db72
c6d59b4ce1b5e596e1146d0f3a839f619c88e840
110122 F20110318_AACGOK mikles_v_Page_148.jpg
a2d1ab03dc0214a1fd1e8d7ae0049183
f2cf29eccd01321f8f09094eba90ae5f324cc945
73156 F20110318_AACGNW mikles_v_Page_130.jpg
1bb2fb1eac2d3e2f598fdf62a9d9d717
8acd8231e93e6b9ce9598e4479257cc63bf33a49
28759 F20110318_AACFLI mikles_v_Page_069.QC.jpg
e6322f2fa32d0c6cdbb4eeddbed9327e
f233c31c930533137f85e4600fbf3565d884761b
908 F20110318_AACFKU mikles_v_Page_157.txt
4de45c977595e63cc24d51f8095caa73
e84f963f4d4c26b06885e702c6e77efd27962780
34899 F20110318_AACGOL mikles_v_Page_150.jpg
6183ba0d9fdd79edc1c61b6a73c7adec
71c000f49853095445a6e326353fb4bb408f4b08
20630 F20110318_AACGNX mikles_v_Page_131.QC.jpg
f3850b581bb2cdf6ffdf6cb5d19ab381
452ae84f81fce541f5973b97dac503ee17d5f206
37337 F20110318_AACFLJ mikles_v_Page_043.QC.jpg
902acd60bb0c9b82e76d80c63e6df788
dbe00765c77a810bde1f33fdffe0fa49ca59b540
114158 F20110318_AACFKV mikles_v_Page_118.jpg
d43e97b6f1f7526981fdc58a4fbc53dc
31fd8eeeeea8b5257f0f77c4290b93bdd980ba37
95377 F20110318_AACGPA mikles_v_Page_168.jpg
a33c8df3af556e7e5e4d993ad6c17611
666cfb3a335e29c76637422ae518790cb5f5e1c0
11225 F20110318_AACGOM mikles_v_Page_150.QC.jpg
f98db99d5809c7f28284cc3b83356d12
958ca204a3cd5658cd7344b1aaf838831c0c734e
461 F20110318_AACFLK mikles_v_Page_108.txt
c98512ccff4ae88b0875f9d452049bd1
56545804759b1c6c432b0786a428f23d5a78f6c8
103656 F20110318_AACGPB mikles_v_Page_169.jpg
d7ffca03c6667f4764e719e13bab3bcd
48cdac191cd0ae56231f9f87d3ce04a12785652f
32269 F20110318_AACGON mikles_v_Page_151.jpg
02c093b18d27d19092011e99dcf684e9
69c98bcb4e272d23c0d01b74458b8e0da2dd5102
50186 F20110318_AACGNY mikles_v_Page_132.jpg
ac44b4246326296f9abe227a45e3b4d6
7de86f2b6e3a760a156613652d595d7cc0450bdf
52912 F20110318_AACFLL mikles_v_Page_090.pro
f5e3ff9a228f2cee2f1104a3040a8789
bc77d422f1b2df7aa5167c1f65f940d559436bd0
33533 F20110318_AACFKW mikles_v_Page_119.QC.jpg
401fe23f6d9225af0603e2dfe2661fc3
2edcc3ddeafcfae4788fd2e3ab62696688f8c5a1
102645 F20110318_AACGPC mikles_v_Page_170.jpg
33583d70766f4b5894d7ed61c72b4185
8616378251629938a0e1bd4f199a4e221c78a647
9382 F20110318_AACGOO mikles_v_Page_151.QC.jpg
c29aa7a8fbd5cb22c93ac749593e4b2d
3e322ac12f7b907f902e9959f9352db683da8ad6
16564 F20110318_AACGNZ mikles_v_Page_132.QC.jpg
4cecee54b0604be677d87682964d3909
199d3de33c7018b85e2101e38b0f439f19478af8
1948 F20110318_AACFMA mikles_v_Page_171.txt
b09d5a5e934035ec750656ab9f990eaf
6051b4f9239f524ea6ef7f4e5283e3261c9b88bd
16803 F20110318_AACFLM mikles_v_Page_137.QC.jpg
1ff46f530962fa449246033e91ae8ce0
30c46309c62c4c87dace4fcb3065c8bfc3a0aeb8
7825 F20110318_AACFKX mikles_v_Page_092thm.jpg
2cc461162c4bc177b9e375e67fe32afa
78ab71945fb4767f69648eaf26ec5634f24f9394
98779 F20110318_AACGPD mikles_v_Page_171.jpg
ecb0aae23372231d275324894b29dad2
a9e1930bb267e8538722a240891161a6dc1a8ae8
9928 F20110318_AACGOP mikles_v_Page_152.QC.jpg
58f20f3dc3bd9e25ba57f6524402bb6b
896122827e309f911d6909e073f9e52081844a9b
F20110318_AACFMB mikles_v_Page_031.txt
fff6ecb214485363aa48c0766cab5837
64b57b372d294fef72dd2cadb4cec9634e6b02b4
124523 F20110318_AACFLN mikles_v_Page_046.jpg
309f08468b765adfb937c03799385061
5fd98ceeb5d2efdaaa18a76329e2fab489f91b0d
18959 F20110318_AACFKY mikles_v_Page_157.QC.jpg
224358d5bba86d7de0fc1bc40149ad7e
3ce66f7619aea087c26b82666a50c178c8b2d1ab
7672 F20110318_AACGPE mikles_v_Page_172.jpg
ec492256cdb4426e216a32d84f8b9cd2
68d678e4372d1f5e373989db65af6453618e22de
9929 F20110318_AACGOQ mikles_v_Page_154.QC.jpg
459eb609c0c1462f0937432d9573b8a2
cf5ac52a7b4c2d4e2ae47d0fb06acca987694797
1251 F20110318_AACFMC mikles_v_Page_135.txt
f759475099f80fbb31c636aa2e234eed
09fcd881a2a11e87cc7cc468d6a7bc0dd96f5b9c
34257 F20110318_AACFLO mikles_v_Page_147.QC.jpg
5f58cd926ecedb1102164aeaf5a2bfe8
cb1a85f099d29edff50e0c99b3b6760d70b395e9
1449 F20110318_AACFKZ mikles_v_Page_056.txt
ae64d5fc74177e301b52dd7e329a99a3
dc91f4d6f3eb9996684846668f2039b20723fec6
28375 F20110318_AACGPF mikles_v_Page_001.jp2
c49cbfad706571b879921c94dc38308c
e415eae0c1eea75008ac713539f1f316a2aa7634
32621 F20110318_AACGOR mikles_v_Page_155.jpg
7ba94f47bfca82bf23f37fce4e315c0f
e1a02d82d70b3728675a8cee62ba2b2c61e38569
81283 F20110318_AACFMD mikles_v_Page_072.jpg
306004d0b634e6f208acd17ed46493b2
af8048921e1a8892b3be182c18958cf69e57fec0
34543 F20110318_AACFLP mikles_v_Page_009.QC.jpg
b7458b5a46879069d70386e42c9b79cc
296daa2af049a9dc787376f3e3957d238d60526b
5386 F20110318_AACGPG mikles_v_Page_002.jp2
0c5668dcfb351296be42d99865407427
267d2ea118c5c60df0c7a94d13ac5d4bcc0afe63
61273 F20110318_AACGOS mikles_v_Page_158.jpg
db6eac70d7beaf45842338fadcd2bb2a
810bd8484c45d2c458a42b8e4abb5e6734ceb944
5551 F20110318_AACFME mikles_v_Page_096thm.jpg
f2f88100ff77ef2d6e5c0def5b95d0e8
31bbd19989002b1db5909bf287e2484bdabf2c8e
F20110318_AACFLQ mikles_v_Page_059.jp2
6e95b433c9480f67f9265617c46c0ecc
d19282e15478182e48f34d36c9653e7016ea0a73
25978 F20110318_AACGPH mikles_v_Page_003.jp2
a243bca7ec95ecb869d96ca2768da2ce
0ae50a5e657cd8d178e23fc1c47d0687b175ad55
110340 F20110318_AACGOT mikles_v_Page_160.jpg
8060f64576122d792a04971c59d9a045
3a0c29a5432cde40613b5e254e7daedbf7d07f3f
F20110318_AACFMF mikles_v_Page_050.txt
b7fe233cceb7654fc06146da7efe99ae
206a5778db1d63b950b60a3f6dcf6adf155fc4b7
49912 F20110318_AACFLR mikles_v_Page_071.pro
e5035f04ad904a382c1b566858b28aee
fe96be33f1288e375412ed6cdf0dcd2c636cc5a0







X-RAY AND INFRARED SPECTRAL AND TIMING OBSERVATIONS OF GALACTIC
INTERACTING BINARY STARS AND ASSOCIATED RELATIVISTIC JETS



















By

VALERIE JEAN MILES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2008































2008
by
Valerie Jean Mikles













To all those who supported me in prayer.


Ode to an Accretion Disk
(written some time after 4am during an observing run)
Particles heated
GI,i; .:l;, feated
Space time pleated
Accrete and accrete!

Shining bright
Then silencing light
A -Aii,,,l,-l,; 's delight
Accrete and accrete!

A constant fire
Until all fuel expire
Or the Universe retire
Accrete and accrete!









ACKNOWLEDGMENTS


They held me up, though I should fall
They said they'd see me through it all
And now I know at beck and call
They'll shield me from my fear.
Ml 1. f,,,n:l it is here.


I would like to thank Almighty God who saw fit to let me finish this dissertation
before taking me home or returning.
I thank my family for their encouragement and love, and for teaching me life lessons
that cannot be learned at a university. Also, I'd like to thank my aunts, uncles,
cousins, and grandparents, whose pr- ir,-i have held me up and supported me in the
darkest of times.
I am particularly grateful to Alison Klesman, my partner in "This-Land," without
whom I would have lost my sanity. I also thank Audrey Ford and the Family C('!n 1!
Dance Team, who carried me in prayer and creative expression.
I thank my advisor, Dr. Steven Eikenberry, who gave me this project and set me up
with colleagues like Reba Bandyopadhyay, Dave Rothstein, Mike Muno, Shannon
Patel, and Jessica Lavine whose experience jump-started my journey into science.
I am grateful to Dave Clark who helped me navigate the red tape of the dissertation
process and kept my spirits high. I thank Michelle Edwards for the significant time
and energy she has put into science conversations.
I also would like to recognize my committee for their enthusiasm and insight:
Vicki Sarajedini, Guido Mueller, Elizabeth Lada, Fred Hamann, and Reba
Bandyopadhyay. Additionally, I thank ('ihi Packham who stepped in as a last
minute alternate for the defense.
This thesis was funded in part by a UF Alumni Fellowship and an NSF grant (AST
05-07547). This funding was essential to my ability to complete this work and
establish appropriate science collaborations.









TABLE OF CONTENTS
page

ACKNOW LEDGMENTS ................................. 4

LIST OF TABLES ....................... ............. 8

LIST OF FIGURES .................................... 9

ABSTRACT ....................... ........... ...... 11

CHAPTER

1 INTRODUCTION ...................... .......... 13

1.1 The Galactic X-ray Population ................... ... 14
1.2 W ind-colliding Systems ................... ........ 14
1.3 Accreting Binary Systems ................... ..... 15
1.3.1 Persistent vs. Transient Sources ......... ........... 16
1.3.2 Multi-wavelength Study of X-ray Binary Sources . .... 17
1.3.3 Modeling Accretion Flows in X-ray Binary Sources . .... 17
1.3.4 Jets in X-ray Binary Sources ................. . .. 20
1.3.5 Variability in X-ray Binary Sources ................. .. 20
1.4 This Work ................... ... ... ......... 22

2 THE SEARCH FOR COMPACT OBJECTS IN THE GALACTIC CENTER. 28

2.1 M otivation ..... ... .. .. .. .. .. ..... .. ... .. .. .. .. 28
2.2 Target Selection .................. ............ .. .. 30
2.3 Observations and Analysis .................. ........ .. 31
2.4 Summ ary .................. ............... .. .. 33

3 THE INFRARED COUNTERPART TO CXOGC J174536.1-285638 (EDD-1) .38

3.1 Introduction .................. ................ .. 38
3.2 Analysis . . . . . . . . ... .. 38
3.2.1 Infrared . . . . . . . ... .. 38
3.2.2 X-ray .................. ................ .. 40
3.3 D discussion . . . .. . . ... .... 42
3.3.1 Is Edd-1 an Isolated Star? ................ .... .. 43
3.3.2 Is Edd-1 A High-Mass X-ray Binary? ................ .. 44
3.3.3 Is Edd-1 A Colliding-Wind Binary? ................. .. 45
3.4 Summ ary .................. .......... .. 47

4 THE X-RAY PERIOD OF CXOGC J174536.1-2.-. :,; (EDD-1) . ... 59

4.1 Introduction .. .. .. ... .. .. .. ... ... .. ... .. .. 59
4.2 Observations and Analysis ............... ........ ..59
4.2.1 Infrared ............... .............. ..59









4.2.2 X-ray Data .................. ............. .. 60
4.3 Discussion ................... . . ...... 62
4.3.1 Infrared Variability .................. ........ .. 62
4.3.2 X-ray Variability .................. . .. 63
4.3.3 The Orbital Period Assumption .............. .. .. 64
4.3.3.1 Wind Obscuration Scenario ................ .. 65
4.3.3.2 The Eclipsing Binary Scenario . . ..... 67
4.3.3.3 Edd-1 as a Wind-Accreting HMXB . . 68
4.4 Summary ................... . . .... 70
4.5 Concluding Remarks About Edd-1 ............ ... .. .. 71

5 AN ATLAS OF KNOWN HIGH-MASS OBJECTS IN THE GALAXY ..... 84

5.1 Background and Motivation ............... ..... .. 84
5.2 Types of sources included/ criteria ................ .. .. 86
5.3 Construction of the database .................. ..... .. 87
5.4 Statistics and Parameters in the Atlas ............ .. .. 88
5.4.1 Object Name and General Information . . ..... 88
5.4.2 Infrared Information .................. ..... .. 88
5.4.3 Reddening .................. ........... .. .. 88
5.4.4 Distance .................. .......... ..89
5.4.5 X-ray Information ........... . . ...... 89
5.4.5.1 Quiescent and Burst X-ray Luminosity . .... 89
5.4.5.2 X-ray Energy Waveband ..... . . ..... 90
5.4.5.3 X-ray Model, thermal temperature, and NH . ... 90
5.4.6 Spectral Information .................. ..... .. 90
5.4.7 Calculated Luminosity .................. ..... .. 91
5.4.8 References .................. ............. .. 91
5.5 The W eb Interface .................. .......... .. 92
5.6 Science Applications ............. . . . .... 93
5.6.1 ('C! i:terizing an Unclassified Source Spectrum . . .. 93
5.6.2 Comparing Classes of Sources ............... . 93
5.6.3 Population Study of a Source Class ... . . 94
5.7 Summary ............... .............. .. 94

6 THE MICROQUASAR GRS 1915+105 ................ .... .. 99

6.1 Introduction .... . . ..... ............... 99
6.2 X-ray Light Curves and Quasiperiodic Oscillations . . 100
6.3 Infrared Flaring Behavior ............... ....... .. 103
6.4 Summary ............... .............. .. 104

7 THE LOW FREQUENCY X-RAY QPO BEHAVIOR OF GRS 1915+105.... 112

7.1 Introduction .. .. .. ... .. .. .. ... .. .. .. .. ... .... 112
7.2 Refining QPO Detection ........... . . .... 114
7.2.1 QPO Frequency Correlation to Spectral Features . ... 115









7.2.2 QPO Time Behavior.. . . .
7.2.3 Differing Ti i.-.- r Spike Morphology in 3-class Light Curves
7.2.4 Associated Infrared Flaring . ............
7.3 D discussion . . . . . . . .
7.3.1 QPO Correlation with Spectral Features .. ........
7.3.2 Infrared Flaring Behavior .. ................
7.3.3 A Cause and Effect Summary .. .............
7.4 Sum m ary . . . . . . . .


. . 118
. . 119
. . 121


8 GRS 1915+105 AND THE ACCRETION-EJECTION INSTABILITY MODEL


Introduction .. .............
X-ray QPO Phenomenology in the 0-, a-,
QPO Models and the Radius Dilemma
The Accretion-ejection Instability Model
Caveats, Conclusions, and Future Work


and 3-classes


9 CONCLUSIONS AND FUTURE WORK ......................

9.1 The Galactic X-ray Population .. ...................
9.2 M icroquasars . . . . . . . . .

APPENDIX

ABBREVIATIONS ..................................

REFERENCES .......................................

BIOGRAPHICAL SKETCH ........................................













LIST OF TABLES


Table


SpeX Observing Log . ............


Table of IR Lines . ..............


Identified Lines . ...............


P Cygni Line Velocity . ...........


Chandra Spectrum of Edd-1 . ........


X-ray and IR Source Comparison . ......


Observing Log: IR Spectra . ........


Observing Log: Chandra . ..


Observing Log: XMM-Newton . .


Mass ratio estimations for the eclipsing scenario


Infrared Line Ratios . ............


Mass ratio estimations for the wind obscuration

of a HMXB .... .................


Summary of scenarios . ...........


Atlas Information . ...............


RXTE Observations . ............


Linear Pearson Correlation Coefficients . .


Partial Correlation Coefficients . .....


Length of Frequency Dip . .........


7-4 Gaussian Fit Parameters .


8-1 Linear Pearson Correlation Coefficients


8-2 QPO Frequency- Radius Fits . ..


scenario


.


.


.


.


.


.


.


.


.


.


the

.


.


.


.


.


.


.


.


.


. .


. .


. .


. .


. .


. .


. .


. .


. .


. .


. .


case

...


. .


. .


. .


. .


. .


. .


. .


. .


. .


page


37


37


54


55


56


57


78


79


80


80


81




82


83


96


111


135


136


137


138


156


159


. . .


: :












Figu

1-1

1-2

1-3

1-4

2-1

2-2

3-1

3-2

3-3

3-4

3-5

3-6


LIST OF FIGURES
re

W ind-colliding binary system ......................

Roche Lobe Overflow in a binary system . ..........

Schematic drawing of a microquasar . ....

Radio jets in the microquasar GRS 1915+105 . ........

K-band spectra of surveyed sources . ..............

Example of spectral image trace with SpeX . .........

K-band spectrum of Edd-1 . ..................

H-band spectrum of Edd-1 . ..................

J-band spectrum of Edd-1 . ...................

P Cygni profiles in the K-band spectrum of Edd-1 . .....

X-ray spectrum of Edd-1 . ...................

Comparison of the X-ray and IR luminosity of Edd-1 to other known
containing massive stars . ....................

K-band spectra of Edd-1 in 2005 and 2006 . ..........

Br7 region of select Edd-1 spectra . ..............

Two representative XMM spectra of Edd-1 separated by 0.7 in phase.

X-ray light-curve and folded light-curve of Edd-1 . ......

Periodogram analysis of Edd-1's X-ray light curve . .....

Theoretical mass ratio as a function of primary mass for binaries with
d4 i period . . . . . . . .

Screen shot of database statistic selection page . .......

Typical 0-class light curve at 1 s resolution . .........

Typical a-class light curve at 1 s resolution . .........

Typical /-class light curve at 1 s resolution . .........


. 49

. 50

. 51

. 52

X-ray systems
. 53

. 72

. 73

. 74

. 75

. 76


189


. ... 77

. . 95

. . 05

. . 06

. . 07


6-4 Eight ms light curve of GRS 1915+105 during a spectrally hard dip and the
power spectrum of select 4 s regions of the light curve .. ..........

6-5 PDS and 1-second light curves for the 0- and a-class .. ............


page

24

25

26

27

35

36

48


1









6-6 PDS and 1-second light curves for the 0-class .................. 110

6-7 A sample light curve showing the X-ray, the infrared, and the QPO behavior
during a jet ejection. .................. .... ......... 110

7-1 Overlay of the 1-second resolution X-ray light curve and fine-binned 4-second
QPO frequency in GR 1915+105. .................. ..... 127

7-2 Scatter plots of QPO frequency with power law flux for a- and p-class light curves. 128

7-3 Scatter plots of QPO frequency with blackbody flux for a- and 3-class light
curves ...................... ........ ... ...... . 129

7-4 Scatter plots of QPO frequency with blackbody temperature for a- and p-class
light curves .................. ................... .. 130

7-5 Time evolution of a Group 1 event in GRS 1915+105. .............. .131

7-6 Time evolution of a Group 2 event in GRS 1915+105. .............. ..132

7-7 Time evolution of a Group 3 event in GRS 1915+105. .............. ..133

7-8 Ti-.-:. r spike of the 3-class light curves in GRS 1915+105. . .... 134

8-1 Cartoon model of jet ejection in an X-ray binary. ............... 150

8-2 QPO frequency vs. total flux for a-, 3-, and 0-classes above 4 Hz. . ... 151

8-3 QPO frequency vs. power law flux for a-, 3-, and 0-classes above 4 Hz. .... 152

8-4 QPO frequency vs. blackbody flux for a-, 3-, and 0-classes above 4 Hz. .... 153

8-5 QPO frequency vs. blackbody temperature for a-, 3-, and 0-classes above 4 Hz. 154

8-6 QPO frequency vs. inner disk radius for a-, 3-, and 0-classes above 4 Hz. . 155

8-7 Observations: Color radius vs. QPO frequency. ................. 156

8-8 Theory: Radius vs. QPO frequency. ................... ..... 157

8-9 Radius fits for the 3-2 event. .................. ..... ..... 158









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

X-RAY AND INFRARED SPECTRAL AND TIMING OBSERVATIONS OF GALACTIC
INTERACTING BINARY STARS AND ASSOCIATED RELATIVISTIC JETS

By

Valerie Jean Mikles

A? ,v 2008

C('! i': Dr. Stephen S. Eikenberry
Major: A-l ii iv

I present results of a campaign to find and identify counterparts to X-ray sources

in the Galactic Center (GC), searching for accretion disk signatures in the form of Br7

line emission via infrared (IR) spectroscopy. IR and X-ray light penetrate the obstacles

that veil optical light, and give us unique opportunities to find and study obscured GC

sources that can help us understand the mechanisms of massive star evolution and the

extreme physical processes that occur in the vicinity of compact objects. I have survn .1

ten candidate counterparts and found one with remarkable Br7 emission. Follow-up

observations of the source, CXO J174536.1-285638 (hereafter, Edd-1), revealed a 189 6

dv: periodicity. I compare Edd-1 to X-ray sources with similar features and find it

consistent with both Colliding Wind Binary (CWB) and X-ray binary (XRB) systems.

CWBs contain two massive stars with strong winds that collide and produce X-rays.

XRBs produce X-rays when matter is accreted from a companion onto a compact object.

I compile several known CWBs and XRBs into an Atlas to facilitate the comparison of

other unknown sources like Edd-1.

A subset of XRBs, called microquasars, produce jet ejections similar to extragalactic

quasars. The short dynamical timescales associated with microquasar jets allow us to

study their properties in depth. I use high time resolution X-ray data from RXTE to

study jet ejection events in the archetype microquasar GRS 1915+105. Examining a series









of ejection events, I find strong correlations between the X-ray spectral phenomenology

and the strength of the ejection, observed as an IR flare.

Jet outflows are modeled with accretion flows because jets are a mechanism to

transport angular momentum from the system. The Accretion Fi. I i. o Instability (AEI)

model employs Magnetohydrodynamics and General Relativity to make predictions about

outflows in XRBs. Using my observations of GRS 1915+105, I find a connection between

the X-ray spectral behavior and the X-ray timing behavior independently (and uniquely)

predicted by the AEI model.









CHAPTER 1
INTRODUCTION

The first rocket-borne X-ray detection instrument was launched in 1948 and was

designed to look only at the Sun (Tousey et al., 1951). The Sun's energy signature was

found to be so weak, it was assumed that no other celestial object would be visible to us

in the X-rays. Despite the implausibility of finding additional celestial sources, a small

X-ray detector was launched in 1962, and Sco X-1 became the first known non-solar

celestial X-ray source (Giacconi et al., 1962). At the time, no one guessed at the wealth

of extreme environments and sources radiating X-rays and filling the sky. Using X-ray

flux variability from identified point sources to constrain the size of an X-ray emitting

region and dynamical constraints from observing counterparts at optical and infrared (IR)

wavelengths, led eventually to the identification of the first compact objects and, more

importantly, confirmation of the existence of black holes. The identification of optical and

IR counterparts to these high energy X-ray sources allows us to target and study systems

exhibiting the extreme physical processes that produce X-rays, in the context of their local

environments.

In the past decade, our knowledge of X-ray emitting sources has advanced significantly

due to the observations of space-based X-ray observatories such as the Rossi X-ray

Timing Explorer (RXTE), the Chandra X-ray Observatory, and XMM-Newton. The

unprecedented angular resolution, energy resolution, and timing resolution of these

instruments let us study heretofore unobservable properties of X-ray sources, including

the shock front produced in a wind-collision region between two massive stars (1.2) and

rapid variability phenomena that occur near the surface of a compact object (1.3.5). In

addition, multi-wavelength observations of these sources from radio to gamma-rays have

resulted in the discovery of relativistic jets in accreting stellar-mass sources (1.3.4). The

incredible influx of X-ray data has led to significant advances in theoretical modeling of

both wind-colliding systems and accreting binary systems.









1.1 The Galactic X-ray Population

Of the hundred billion stars in the Galaxy, less than 0.15'. are massive (> 8Me)

and up to S I' of those massive stars are found in binary or multiple systems (Kroupa,

2001; Lada, 2006). Albeit rare, massive stars greatly affect the Galactic environment

and the chemical evolution of our Universe. The galactic and intergalactic medium is

enriched by the products of the nucleosynthesis that occurs in the cores of massive stars.

These products are distributed to the circumstellar environment by wind-driven mass loss,

and by novae or supernovae events. The study of the stellar X-ray source population in

our Galaxy has increased our knowledge of the evolution of these sources and how they

contribute to the Galaxy's evolution.

1.2 Wind-colliding Systems

Colliding wind binary (CWB) systems are composed of a pair of very massive

(typically > 20M1) stars undergoing strong mass loss. The massive stars comprising

CWB systems are O stars, Wolf-R i-t stars, or some combination thereof. Wolf-R i, t

stars are evolved, massive stars that are characterized by very high mass-loss rates of

O-10 -M/yr, as compared to very massive O-type stars which tend to have mass loss

rates of ~ 10-6M. /yr (Corcoran et al., 2006).

The high resolution and sensitivity of Chandra and XMM-Newton, as well as the

monitoring capabilities of RXTE, have allowed a detailed examination of some of the

brighter (Lx ~ 103erg/s) CWBs and identification of the fainter CWB population down

to Lx ~ 1032erg/s (\!,iii, et al., 2006a). In these sources, X-ray emission primarily

arises in the shock front where the winds from the stars collide (see, e.g. Luo et al., 1990;

Sana et al., 2004; De Becker et al., 2006). The physics of the heating process in the shock

1. -r is poorly understood, but it has been shown that the thermal X-ray luminosity

scales as M2 (Luo et al., 1990). The force balance between the two contributing winds

determines which wind dominates. In Figure 1-1, I show an example of a colliding wind

system containing a Wolf-R ,it star and an O-type companion as derived mathematically









by Usov (1992). The Wolf-R ,it star contributes the dominant wind because the typical

mass loss rate of a Wolf-R ,it is expected to be an order of magnitude higher than that of

an O-type star (Usov, 1992). The shock region thus occurs much closer to the O-type star.

The X-ray spectral energy distribution of a CWB is that of a collisionally ionized

plasma, with electrons accelerated by shock heating (Luo et al., 1990). The luminosity and

absorbing column of the plasma are measured by fitting models to the X-ray spectrum

(Sana et al., 2004). Both the luminosity and the absorption column are known to vary

with time, often as a function of the orbital phase (Luo et al., 1990). Wind parameters

such as the mass loss rates and wind velocities can be characterized by radio intensity

and ultraviolet line properties. Stellar parameters such as mass, temperature, radii, and

rotational velocity can often be estimated via spectral typing in the optical wavelengths

(Corcoran et al., 2006).

The emission from CWBs provides an extremely useful probe of the nature of the

shock and is valuable for investigating the underlying physics of particle acceleration in

conditions too extreme to be reproduced in a laboratory. Additionally, we can infer from

CWBs the nature of the mass loss processes in massive stars, which can then lead to

better understanding of the evolution and end products of massive stars.

1.3 Accreting Binary Systems

An accreting binary system (also called and X-ray Binary or XRB) contains a

compact object, either a Neutron star or a Black Hole, and a less evolved star, such as a

main sequence star, giant, or supergiant. The primary factors affecting stable accretion

of mass onto a compact object are the mode of mass transfer, the ratio of the mass of the

compact object to that of the companion star, and their orbital separation (Lewin and van

der Klis, 2006). There are two common modes of mass accretion: Roche Lobe Overflow

and Wind-fed Accretion. In the example of Roche Lobe Overflow, mass is accreted onto

a compact object when the companion (i.e., the less evolved star) fills its Roche Lobe

and mass then transfers through the inner Lagrangian point as shown in Figure 1-2. The









Roche Lobe is the region surrounding a star in a binary system inside which the star's

material is gravitationally bound to the star. Roche Lobe Overflow occurs in systems

where the mass of the companion is smaller than that of the compact object, meaning

Roche Lobe Overflow is more common in systems where the companion is a low-mass

star (< 1.4M., Lewin and van der Klis, 2006). This is called a low-mass X-ray binary

(LMXB). Due to the low-mass nature of the companion, the optical light of LMXBs is

due mostly to reprocessing of the X-ray flux from the outer accretion flows and LMXBs

are significantly brighter in the X-ray than in the optical wavelengths (Lewin and van der

Klis, 2006).

A compact object can also accrete matter from a companion star that does not fill

its Roche Lobe if the companion star is losing mass in the form of stellar wind. Wind-fed

accretion depends on strong stellar winds (AMr- 10-6 M./yr), such as those present in the

massive O-type stars discussed above. Systems with massive companion stars are called

high-mass X-ray binaries (HMXBs). The optical luminosity in these systems is dominated

by the companion star and the X-ray luminosity depends on the rate of mass transfer

(Frank et al., 2002). I will elaborate upon and calculate the X-ray luminosity resultant

from a wind-fed accretion system in Chapter 4, 4.3.2.2.

1.3.1 Persistent vs. Transient Sources

The X-ray sky is filled with objects that are very broadly divided into two categories:

transient and persistent. A persistent X-ray source maintains its energetic processes and

constantly produces X-rays, though the intensity may vary. Many LMXBs are persistent

in the X-ray while HMXBs are commonly transient (Lewin and van der Klis, 2006). Most

transient sources display outbursts lasting from a few weeks to months, separated by

long periods of inactivity lasting months to decades. During an outburst in a transient

source, the luminosity changes by several orders of magnitude. The length of the outburst

is consistent with viscous timescales of accretion disks. Less than half of known X-ray

sources are transient (Lewin and van der Klis, 2006).









1.3.2 Multi-wavelength Study of X-ray Binary Sources

XRBs are identified by their intense X-ray brightness; however, the properties of the

binary systems, such as their orbital periods and the masses of the companion star and

compact object can be determined only if we observe a counterpart to the X-ray source

in optical, IR, or ultraviolet wavelengths. The optical emission of an XRB is affected

by both thermal emission from the accretion flow and the reprocessing of the X-ray

emission in the outer accretion disk and on the surface of the companion star (Lewin

et al., 1997). If the source is X-ray transient, we can use times of "quiescence" (a time of

low X-ray luminosity, presumably caused by reduced accretion rate) to learn more about

the companion star. In times of quiescence, the optical spectrum is dominated by the

companion star, which allows us to more accurately spectral type the companion and

dynamically measure the mass and binary orbit of the system. Detailed studies of XRBs

at IR, optical, and UV wavelengths are also crucial in order to measure the temperature

profiles, ionization fractions, and abundances of elements in the accretion disk (Lewin and

van der Klis, 2006).

1.3.3 Modeling Accretion Flows in X-ray Binary Sources

Accretion is the process by which matter is transferred from one celestial object

to another. The transferred mass attempts to form a stable circular orbit around the

accreting object, however, the orbiting matter is subject to viscosity which leads to

internal friction. The matter heats up as the directed kinetic energy of the bulk mass

motion is converted into random thermal motion. The orbiting mass that loses energy

spirals inwards while the gas that gains energy spirals out, spreading to form a disk. The

goal of standard accretion disk theory is to understand the physical properties of the disk

and how these properties translate to observables.

The nature of the compact object and the interaction of the accreted material with

the surface of the compact object are also important factors for modeling accretion

flows. Accreted matter will behave differently when accreted onto a Neutron Star, which









has a solid surface, versus a Black Hole, which has an event horizon. The 1i ii, i iy of

the high-energy X-ray and gamma-ray radiation is produced in the region between the

innermost stable orbit of an accretion disk and the surface of the compact object (Lewin

and van der Klis, 2006).

Regardless of the method of mass transfer in an XRB, the motion of a particle in

an accretion disk is determined by the amount of angular momentum it possesses, the

physical processes by which it loses angular momentum, and the radiation processes

by which it cools. In order to obtain a disk solution, we simultaneously solve four

conservation equations: mass, angular momentum, vertical momentum, and energy. If

we assume the accretion disk is geometrically thin, then there is no vertical motion in

the disk and the particles are in hydrostatic equilibrium in that dimension. The standard

model developed by Shakura and Sunyaev (1976) assumes that angular momentum is lost

at high rates due to an unknown process that is proportional to the pressure. This model

also assumes that radiation processes are efficient meaning that accretion flow is relatively

cool. The result is that the accretion disk is geometrically thin (the thickness is much less

than radius). Ultimately, disk structure is controlled by the mass of the system M, the

accretion rate fM, the radius R, the viscosity a, and the pressure ratio / (Shakura and

Sunyaev, 1976). Once defined, most other parameters of the accreting gas (such as the

luminosity of the disk, the density of the region, and the speed and temperature of the

particles) can be calculated.

More sophisticated models have since developed because of the identification

and modeling of the magneto-rotational instability (i\ I), which has affected our

understanding of how turbulence arises and transports angular momentum in ci-1 1,! i--ical

accretion disks (Balbus and Hawley, 1991, 1998). In the MRI model, a seed magnetic field

in the flow starting with infinitesimal strength gets enhanced and tangled. The resulting

turbulence is an efficient mechanism for angular momentum transport.









Further advances in accretion modeling came with the discovery of stable inefficient

accretion flows. In these solutions, the electrons and ions have high temperatures that

differ from each other, the accretion flows are geometrically thick, and most of their

potential energy is advected toward the compact object rather than radiated away

(N ,o ,yan and Yi, 1995). Advection dominated accretion flows (ADAFs) were later shown

to be capable of launching powerful outflows (Blandford and Begelman, 1999).

The study of ADAFs began when independent measurements of the mass and

accretion rate in these under-luminous accreting systems predicted luminosities much

brighter than observed (N ii- i, et al., 1995). The apparent radiative inefficiency of these

objects introduced an energy imbalance in the standard accretion model. Advection was

-,-.:., -1. d as a means to dissipate the excess heat.

Advective cooling assumes a radiatively inefficient environment where energy is stored

in accreting gas as entropy (N ii- i, and Yi, 1995). The system is hot because it is not

i i11 i iiir- and it is geometrically thick because it is hot. In a black hole system, advection

results in cooling because once the gas is accreted across the event horizon, the energy it

carries is lost from the system.

In traditional ADAF models, the gas has a positive Bernoulli parameter and is

gravitationally unbound, meaning that spontaneous winds can develop and produce

outflows. The Bernoulli parameter B is the sum of the kinetic energy, the potential

energy, and the enthalpy of the accreting gas. It measures the likelihood of spontaneous

winds and outflows. The "Advection Dominated Inflow/Outflow Solutions" or ADIOS

was developed by Blandford and Begelman (1999), who created a self-consistent solution

by adding a powerful wind which transports away mass, energy, and angular momentum

in such a way as to keep the Bernoulli number locally negative and thus allow mass to

accrete.









In C'! ipter 8, I discuss the Accretion-Fi', I i on Instability model, which is a more

complicated outflow solution that combines General Relativity, the MRI, and the rapid

X-ray variability observed in XRBs.

1.3.4 Jets in X-ray Binary Sources

Jets are a form of outflow from accreting stellar-mass black holes in the galaxy as

well as accreting supermassive black holes in active galaxies. They have non-thermal,

polarized radio spectra indicating the presence of shock-accelerated relativistic electrons.

These electrons emit synchrotron radiation as they propagate in regions with large-scale

magnetic fields (Lewin and van der Klis, 2006). Radio observations show large flux ratios

between the approaching and receding sides of the jets as expected in relativistic flows

('!il ,'Iel and Rodriguez, 1994). The propagation speeds of jets can be measured on

images by following the kinematics of individual radio knots. XRBs that have relativistic

jets are called microquasars. In Figure 1-3, I show a schematic drawing of a microquasar.

In 1979, Margon et al. (1984) observed the first microquasar, SS 433, a high-energy

source with p-i. --iir- mildly relativistic jets ejecting material at a velocity of 0.26c. In

1992, the WATCH/GRANAT satellite detected the microquasar GRS 1915+105, which

had highly relativistic jets, measured at a velocity of 0.92c. The radio image of the ejecta

is shown in Figure 1-4 (\iil i el and Rodriguez, 1994).

While jet ejections take decades to observe in extragalactic quasars, microquasars

show us a similar event in a matter of hours or d -iv Thus, microquasars are an excellent

laboratory through which we study accretion physics, relativistic jet formation, and

disk-jet interactions. Because disk-jet interactions occur close to the central compact

object, understanding the interaction will help us probe the limits of General Relativity

and answer the question of what happens at the edge of a black hole.

1.3.5 Variability in X-ray Binary Sources

X-ray variability of XRBs allows us to probe the physics of strong gravity and dense

matter. Whereas General Relativity is observationally well-tested in the weak field limit









(GM/R << c2), there are no direct observations of particle motion in strong gravitational

fields, such as those within a few Schwarzschild radii (Rs) of a compact object. Since

characteristic velocities are relativistic (~ 0.5c) near a compact object, the dynamical

timescales can be short (Lewin and van der Klis, 2006). For example, a 1.4 Me neutron

star has a dynamical timescale (r3/GM)1/2 of ~ 0.lms near the surface of the compact

object (15 km), and ~ 2ms at a distance of 100km (Lewin and van der Klis, 2006). A 14

Me black hole, like GRS 1915+105 (Greiner et al., 2001), has a dynamical timescale of

l~ ms at a distance of 3 Rs.

Variability can be used to probe accretion flow dynamics. For a 14 M. black hole,

most of the gravitational energy is released within the inner ~ 100 km of the system,

which is characterized by millisecond dynamical timescales (Lewin and van der Klis,

2006). Since the temperature of the inner accretion disk is high (> 107K), most of the

emission is expected to be in the X-rays. Factors affecting the X-ray luminosity of an

accretion flow include the mass-transfer rate from the companion to the compact object,

the clumpiness of the accreted material, the inward diffusion of matter in the accretion

flow, and the interaction of the accretion flow with the central object.

Perhaps most telling about the inner workings of the accretion disk and the launching

of outflows are X-ray quasi-periodic oscillations (QPOs). QPOs have highly variable

frequencies and may lose coherence after tens or hundreds of seconds (Lewin and van der

Klis, 2006). The observed QPO frequencies (from a few hertz to tens of kilohertz) are

attributed to the characteristic dynamic and hydrodynamic frequencies at the inner part

of the disk. The rapid variability properties of accretion flows provide useful probes into

the physical conditions close to compact objects and the geometry of accretion flows.

Low frequency (2-10 Hz) QPOs are commonly observed in XRBs, but the effort to

tie them to the accretion-flow geometry is complicated because they are much lower than

the dynamical frequencies associated with particles orbiting in the inner accretion disk.

The physical mechanism that generates these low frequency QPOs is not known, though









the strength of the oscillation -i-i- -I that the mechanism is global and organized within

the emitting region (Lewin and van der Klis, 2006). In C'! lpter 8, I discuss the Accretion

Fi. i. oi Instability (AEI) model which combines magnetic instability and Keplerian

motion to explain the observed low frequency QPOs. The QPO behavior and the AEI

model are discussed extensively in C'! Ilpters 6-8.

1.4 This Work

The projects in this dissertation are motivated by a desire to (1) understand the

Galactic X-ray population by identifying infrared counterparts to X-ray sources, and (2)

understand the subset of jet-producing X-ray binary systems by studying the microquasar

GRS 1915+105.

A large number of known X-ray sources are concentrated in the Galactic Center

(GC) and Galactic Plane (GP), though only a fraction of these X-ray sources are known

to be XRBs or CWBs. Multi-wavelength follow-up is imperative to the classification of

these sources. For part of my dissertation, I performed infrared spectroscopic observations

of the GC in order to find counterparts to Chandra X-ray sources in the GC. Heavily

obscured by dust, the GC is virtually unobservable at visible wavelengths. However, IR

observations reveal a dense stellar population traced by X-ray sources, including compact

objects, massive star clusters, and a supermassive black hole. As noted earlier, the GC has

been known to be home to X-ray emitting compact objects since the beginning of such

studies in the 1960s (see, e.g., Giacconi et al., 1972; Gursky, 1972; Forman et al., 1978).

Identifying and studying the population of compact objects in the GC gives important

insights into the massive star formation history of this region. With the sub-arcsecond

resolution of modern IR and X-ray instruments, I am able to penetrate the extinguishing

veil of the interstellar medium and explore this exciting region.

In the IR spectroscopic survey, I seek emission signatures, particularly in the Br7 line

of the IR spectrum, which are indicative of high energy processes. In C!i lpter 2, I detail

the nature of my survey. In C(i lpters 3 and 4, I discuss follow-up observations of a strong









emission-line source identified during this survey as the infrared counterpart to an X-ray

source. Chapter 3 focuses on the initial IR and X-ray spectral analysis of the source, while

('!i lpter 4 uses a detailed timing analysis to place constraints on the nature of the system.

In ('!C lpter 5, I discuss my preliminary work on a multi-wavelength spectral atlas designed

for the rapid characterization of X-ray/IR sources. This Atlas contains multiwavelength

information on XRBs, CWBs, and massive stars.

In addition to exploring the overall Galactic X-ray population in general, I devote

considerable attention to the archetype microquasar GRS 1915+105. Although it suffers

from ~ 30 magnitudes of extinction at visible wavelengths, GRS 1915+105 has shown

great activity in the radio, infrared, and X-ray regimes. Of particular interest in this

study are X-ray light curves showing rapid variability positively linked to jet ejection

events. The second part of my dissertation details my investigation of microquasar

GRS 1915+105, specifically the phenomenological behavior proceeding a jet ejection. In

C'! Ilpter 6, I give a more detailed introduction to GRS 1915+105 and discuss the X-ray

phenomenology during events positively associated with a jet ejection from the source. In

('!C lpter 7, I discuss the X-ray evolution proceeding a jet ejection in the context of infrared

flaring. In C'! Ilpter 8, I compare the observed phenomenology to predictions made by jet

ejection models for microquasar systems.














S2


Figure 1-1.


Example of a wind colliding binary system containing a massive O-type star
and a Wolf-R i- t S1 and S2 are the shock waves and C is the contact surface.
The region of stellar wind collision is hatched. D is the orbital separation
and rwR and roB are the distances of the contact surface to the Wolf-R i-,-t
and O-star respectively. Since the wind from the Wolf-R ti-,t is stronger, the
contact surface is closer to the O-type star. From Usov (1992).


~ILC~~III (IIIII1 1IIII.1IIICIIIl



















Roche Lobe




S--- /
>' *


^ .* K"{.y 0: OX* X4 1
Ax *s ? :

/z ..s i i uiit
-I

4-
V I I I
^-ffs:'?:^ %" %' % % i^
** :;i '' ; ','' S;S-S" ^^i*^.

^v /v ** .. ... A \,, ... \ \ A \* A *^


TL "v "** '*'
.'*f'- / : : ? ; :'j
\il;V;~~ ',,* : 'f **


Figure 1-2.


Op-AN#,


Accretion Disk


The Roche Lobe of a star is he region surrounding a star in a binary system
inside of which the star's material is gravitationally bound to the star. If
the star's envelope fills the Roche Lobe, matter can be transferred to its
companion star. This is a common method of accretion in Low-Mass X-ray
Binaries.


/ 4


-t













MICROQUASAR


X-RAY
RADIATION





ACMCRET1MO
DISK t- l krm)


Figure 1-3. Schematic drawing of a microquasar showing the compact object, accretion
disk, companion star, and relativistic jets. The companion star dominates the
optical and infrared emission, the accretion disk dominates the X-ray emission,
and the endpoints of the relativistic jets dominate radio emission.






















9 Aprll 16 April


~b--~-- O"


-0





Figure 1-4. Jet progression images made from of a series of VLA radio images of the
microquasar GRS 1915+105 taken at wavelength of 3.5 cm, depicting the
rapid ejection of jets over the span of a month (! i'ch 27 to April 30, 1994).
From Mirabel and Rodriguez (1994).


27 March 3 April


23 April 30 April









CHAPTER 2
THE SEARCH FOR COMPACT OBJECTS IN THE GALACTIC CENTER

2.1 Motivation

As early as 1978, the HEAO (Einstein) Observatory resolved the X-ray background

sufficiently well to show that the sky was filled with X-ray point sources (Giacconi

et al., 1979). The Galactic Center (GC) region has been well-known as a home for

X-ray emitting compact objects since the beginning of such studies (Giacconi et al.,

1972). Identifying and studying the population of compact objects in the GC gives

important insights into the massive star formation history of the region as well as

providing individual targets for multi-wavelength follow-up. Despite this, the crowded

stellar fields, high visual extinction, and poor image resolution of previous X-ray missions

have prevented identification of many optical and infrared counterparts to these compact

objects. Steady improvements in detector technology have led to better resolution and

more detailed analyses of X-ray objects. The Chandra X-ray Observatory has taken a

revolutionary leap over its predecessors with angular resolution comparable to optical

telescopes (t 1 arcsec). This allows us to identify X-ray sources and match them to

counterparts at other wavelengths with unprecedented precision. X-ray binaries (XRBs),

which generally appear indistinct from stars in optical and infrared (IR) surve,- shine

brightly in the X-ray, making Chandra an ideal instrument for targeting and studying

XRBs at multiple wavelengths. The advent of the high-resolution capability of Chan-

dra, coupled with sensitive IR imagers and spectrographs capable of penetrating the

extinguishing veil of the interstellar medium, offers a new opportunity to explore the GC

region.

Muno et al. (2003) presented a catalog of 2357 X-ray point sources in a 17'x17'

field around the GC, complete to Lx ~ 1031erg s-1. About I'- of these sources are

detected at < 3a in the hard 2-8 keV energy range and are virtually invisible below

2.5keV (L 'iock et al., 2005). The spectra are cut off at the lower energies due to high

interstellar extinction (L ,i wck et al., 2005; Negueruela and Schurch, 2007). The 2-8 keV









X-ray luminosities of the ini i, ii ily of sources are inconsistent with normal stars, active

binaries, or young stellar objects, none of which are normally persistently luminous enough

to produce the observed fluxes given the high absorption column toward the GC (L i-', ick

et al., 2005). However, sources such as Cat.,., i-i-nic Variables (CVs), pulsars, Colliding

Wind Binaries (CWBs), and XRBs are expected to lie in the luminosity range of the GC

sources (\iiii.) et al., 2003; Mikles et al., 2006).

With an average optical extinction Av ~ 25 mag, searches for counterparts to

the X-ray sources in the GC region are limited to IR wavelengths. Unfortunately, the

infrared sky is significantly more crowded than the X-ray sky and dominated by non-X-ray

emitting sources. The 2MASS survey catalogs over 12,000 sources in the 0.3" x 0.3" region

covered by the Muno et al. (2003) X-ray survey (Cutri et al., 2003). An analysis by Dutra

et al. (2003) quotes an average of 60,000 stars per r = 1 field with 8 mag < Ks < 11 mag.

2MASS is complete to Ks = 14.3mag, a boundary which is ultimately set by the confusion

limit, beyond which it is not possible to resolve individual stars, especially in high-density

regions. However, the vast 1i i i ly of the low-luminosity Chandra sources are expected

to have main sequence counterparts with low (< 1Ml ) masses. Low-mass XRBs and CVs

are not visible in the infrared at GC distances without very deep, high resolution imaging

(Gosling et al., 2007). A maximum of 5. !'. of sources could have counterparts brighter

than Ks = 15mag. The presence of a significant High-Mass X-ray Binary (HMXB)

population would appear between Ks =14 -16 mag (L li ... 1: et al., 2005).

Another caveat in finding IR counterparts is that the extinction due to dust lying

between us and the GC makes the stars appear systematically redder than foreground

stars. Negueruela and Schurch (2007) attempted to use IR photometry to identify possible

counterparts and determined that it is nearly impossible to break the degeneracy between

a red (A or F) star and a highly reddened OB star. In addition, many 2MASS sources in

the GC fields are at the confusion limit of the 2MASS images; the colors of unresolved

blends are essentially meaningless.









Infrared spectroscopic surveys of the GC are being carried out by a number of groups

seeking counterparts to unusual high-energy sources (e.g., Band). i]' i.l!:ay et al., 2005;

Laycock et al., 2005; Mikles et al., 2006). Such surveys for counterparts are further

fueled by the launch of observatories such as INTEGRAL and Son J, Since its launch,

INTEGRAL has found or identified more supergiant XRBs than were previously known

(Walter et al., 2006). This discovery challenges current binary star population synthesis

models, making counterpart discoveries an important matter (Negueruela and Schurch,

2007). Bandyopadhyay et al. (2005) performed a spectroscopic survey and, following

their initial work, analyzed the effectiveness of finding counterparts using their technique.

The success rate is low owing to the number of chance coincidences possible in crowded

regions and large (~ 0.5arcsec) astrometric errors (Bandyopadhyay, 2008). There may

also be physical conditions responsible for the absence of an accretion signature, such

as the system being in a quiescent phase. While some of these physical limitations can

be overcome by deeper or multi-epoch iin -ii.- single-look (or snapshot) spectroscopic

surveys remain the most efficient method for identifying counterparts and eliminating false

matches.

2.2 Target Selection

The target selection was performed by my collaborator Shannon Patel. I describe

his methods here. Using archival Chandra observations of the GC region Muno et al.

(2004b) identify ~400 serendipitous X-ray sources with Lx[D/8kpc]2 ~ 1031 1035erg s-1

within ~10-arcmin of Sgr A*, or ~24 pc of the GC with <1-arcsec positional accuracy.

Patel cross-correlated these Chandra observations with the 2MASS catalog, identifying

possible IR counterparts to the X-ray sources, and in 2004 began a campaign to locate

and identify compact objects in the GC via IR spectroscopy. The X-ray/IR analysis he

employs to identify candidate compact objects is designed to select sources consistent

with reddened accreting compact objects such as cataclysmic variables, relativistic

jet sources, background quasars, and black hole binaries. Taking the astrometrically









matched X-ray/IR catalog, he breaks the sources into categories of "possible matches"

(<1.5-arcsec positional difference between Chandra and nearest 2MASS source) and "close

non-matches" (1.5-3.0-arcsec difference). This produces a list of ~180 Chandra sources

with possible IR counterparts. He then uses the "close non-matches" to estimate the false

positive rate for association in each Chandra observation by taking the number of false

matches, dividing by the area of this annulus in square arcseconds, and multiplying by

the number of Chandra sources and area of the inner 1.5-arcsec-radius 1i1 ii' !!" region. In

this way, he eliminates candidates in crowded regions for which he expects >35' of the

possible matches to be random coincidences. He then compares the X-ray to IR flux ratios

for these sources, and eliminate those with low ratios (log Fx/FIR < 2, reddened) as being

likely due to stellar atmospheric emission rather than compact objects. Finally, he uses

the J K8 colors of the potential 2MASS counterparts to remove all candidates which are

too blue (J K, < 2.0 mag) to be located in the reddened GC. This yields a list of 26

X-ray sources with probable (typically >7.' likelihood depending on the Chandra field)

physically-linked IR counterparts at or beyond the distance of the GC.

IR spectroscopic follow-up is necessary to resolve the nature of each source. Through

spectroscopy, we can identify XRBs though their typical emission line spectra (Br7, Hel

2.058pm, and Hell 2.189pm in the K-band, Brackett series and Hell in H-band, and CO

absorption for low-mass giant donors). Likewise, redshifted lines indicate that the target

is a background AGN or quasar, and emission lines and continuum shape allow us to

determine the appropriate subtype classification for each. Non-accreting X-ray sources also

have distinguishing characteristics, so that these spectroscopic observations should allow

for a robust classification of each candidate.

2.3 Observations and Analysis

I carried out observations at NASA's Infrared Telescope Facility (IRTF), a 3.5m

telescope located in Maunea Kea, Hawaii. Its size makes it optimal for observing our GC

targets (iK ~ 10 12 mag) with sufficient resolution. On 2005 July 01-03 UT, I obtained









near-infrared spectra, simultaneously covering the 1.1 2.4pm bandpass, of nine fields

using SpeX on IRTF (Rayner et al., 2003). This resulted in a total of ten target spectra.

Table 2-1 shows the observing log for these sources. Nodding along the length of the slit,

I obtained six ~120s exposures for a total exposure time of 720s. In the short-wavelength,

cross-dispersed mode, I attained a resolution of R ~ 1200 over the JHK bandpass. The

resolution estimation is confirmed by measurements of OH sky lines. Target observations

were followed by observations of the GOV star HR 6836 at similar airmass for removal of

telluric absorption features. Using the standard SpeX macro cal_sxd_0.5, I obtained flat

fields and wavelength calibration.

I extracted spectra using the standard SpeXTool procedure for AB nodded data,

resulting in a series of -1.:,--~l tracted, wavelength-calibrated spectra (described below;

Vacca et al., 2003; Cushing et al., 2004). Extracting a point source using SpeXTool is

highly automated. To set up the general procedure, one must first set up a sample of

a single AB pair. Inputting the appropriate calibration frame, SpeXTool will load a

subtracted image and then allow the user to select the location of a spatial profile of

the spectrum on the detector. In order to extract spectra for the wavebands separately,

the spatial profiles must be selected separately. Although SpeXTool will attempt to do

this automatically, one is able to adjust the aperture position and manually define an

extraction region if desired. This feature is helpful if observing multiple stars on the slit

during a single observation, as I did for our stars 1 and 2. In Figure 2-2, I show the AB

subtracted image of stars 1 and 2 from our sample. The H and K band are clearly visible

for both sources as the parallel light and dark lines on the image. If the source or band

is not bright (as in the case of the J-band spectra of these stars), then the automated

selection may fail. Manual placement of the aperture allows one to attempt to extract

these spectra. I obtained J, H, and K-band spectra for both sources in this image.

Once the spectra were extracted, I median-combined them using xcombspec. I loaded

a series of source files into the program and viewed them as a group before writing the









sum to a file. By loading them separately, I detected a number of irregularities in the

individual spectra (e.g., hot pixels or accidental blending of sources in a nod frame). Using

xcleanspec, I corrected hot pixels and other irregularities that only occur in one of the

nodded images prior to combining. Both xcleanspec and xcombspec are included in the

SpexTool package.

After combining the source spectra, I used xcleanspec to remove intrinsic Br7

absorption in the standard star. Because I chose a GOV star, this absorption feature

created a false feature at the Br7 region, which is where I primarily hope to find an

emission signature in my targets. To remove the Br7 absorption feature from the summed

standard spectra, I interpolated over the region, careful to only smooth over the range

2.160-2.172pm; any variation in the smoothing technique affects the region strongly. I then

divided the target spectrum by the GOV-star to remove atmospheric absorption bands. To

restore the intrinsic shape of the the target spectrum, I multiplied the resultant spectrum

by a 5900 K blackbody spectrum, corresponding to the temperature of a GOV star.

For the preliminary analysis, I estimated the extinction toward the GC at Av = 26

and correct the spectral shape for the reddening law based on Rieke and Lebofsky (1985).

The resultant spectra are plotted in Figure 2-1. A majority appear to be red stars, of type

K or cooler as is apparent by the CO absorption bands at A > 2.1/m (Wallace and Hinkle,

1997). In Table 2-2, I list the wavelength centers and identification of lines I observe in the

spectra. I also see evidence of Cal triplet absorption and the sodium doublet, common to

these types of sources (L.-.-. tt et al., 1996). Source 3 appears to be a blue star (type F or

hotter; Wallace and Hinkle, 1997), and is relatively featureless in the infrared. Only source

1 has strong Br7 emission. I describe the extended analysis of source 1 (nicknamed Edd-1)

in the next two chapters.

2.4 Summary

IR spectroscopic follow-up is imperative for confirming counterparts to X-ray sources

in the GC and resolving their nature. For the bright sources targeted in the survey, I did









not find a significantly large fraction of massive star or emission line counterparts. It is

possible that the infrared targets are the correct counterpart and that the system is in

quiescence, thus preventing me from identifying an accretion signature. Also, it is possible

that astrometric errors have caused me to misidentify the counterpart and that another

nearby infrared target is the correct counterpart. Given that the technique identifies IR

bright counterparts to low luminosity X-ray sources, I am preferentially seeking HMXBs

as they have Lx/LK << 1. However, Muno et al. (2006a) -ii--.- -1 that a large fraction

of low luminosity Chandra sources (up to 91 .) may be CVs or LMXBs. (LMXBs have

Lx/LK >> 1.) If the sources targeted in the survey are LMXBs, their IR counterparts

may be more than five magnitudes fainter than the target range -beyond the confusion

limit of 2MASS and too faint to be practically observed with IRTF. In that instance,

future surveys like the Flamingos-2 Galactic Center Survey (Eikenberry et al., 2005) would

be necessary to resolve the nature of those sources.











25


I




20

2


3


15 41 -





I7

10







5 I


101





2.1 2.2 2.3 2.4
Wavelength (microns)


Figure 2-1. Spectra obtained from the 2005 July 01-03 observing run. The observing log is
detailed in Table 2-1. The location of Br7 (2.166/pm) is marked with a vertical
line. Other visible lines are listed in Table 2-2.









X Ximgtool


- i X


File View Cursor I Zoomln ZoomOutl Centerl Invert Hp


Ji1, Auto Range, Linear, Zoom]


(600,50, 575.50)


-0.107595


I'
Figure 2-2. The calibrated, subtracted spectral image trace for objects 1 and 2, obtained
simultaneously with SpeX. The K-band spectrum of object 1 is outlined in
green.










Table 2-1. SpeX Observing Log


Target ID


Date


2005-07-01
2005-07-01
2005-07-01
2005-07-02
2005-07-02
2005-07-02
2005-07-02
2005-07-03
2005-07-03
2005-07-03


Time
(UT)

11:17
11:17
12:02
10:20
10:56
11:30
12:01
10:48
11:30
12:15


Exp. Time
(min)


Mag. R.A.


10.33 17h45m36.139s


11.12
10.63
11.64
11.04
11.33
11.85
11.97
11.17


17h45m52.142s
17h45m17.359s
17h46m9.953s
17h45m8.532s
17h45m56.974s
17h45m35.002s
17h45m49.214s
17h46m8.393s


Dec.


-28d56'38.47"

-29d07'43.92"
-29d06'25.03"
-29d03'21.64"
-28d56'51.55"
-28d55'19.71"
-29d07'51.29"
-29d01'37.71"
-29d06'23.81"


Note. -All magnitudes are Ks from 2MASS.


Table 2-2. Ta
Line


Hel
Br7
Hell
Nal doublee)
Cal (triplet)
12CO(2,0)
12CO(3,1)
13CO(2,0)
12CO(4,2)
13CO(3,1)
12CO(5,3)


ible of IR Lines
Wavelength
(p m)

2.114
2.166
2.191
2.207
2.261
2.293
2.322
2.343
2.352
2.375
2.382


Note. -Select K-band
IR lines identified in source
spectra. Wave centers taken
from Bandyopadhyay et al.
(1997).









CHAPTER 3
IDENTIFICATION OF THE INFRARED COUNTERPART TO CXOGC
J174536.1-285638 (EDD-1)1

3.1 Introduction

In this chapter, I present the first results of a campaign to locate and identify

compact objects in the GC. Recent Chandra observations of the GC revealed a new

population of faint X-ray sources with Lx ~ 1031 1033 erg s-1 (\ii!,, et al., 2004b).

Using the criteria discussed in the previous chapter, I identify CXOGC J174536.1-2-.i.

(the Br7 emission source noted in ('! lpter 2; hereafter, Edd-1) as a potential compact

object. In this chapter, I discuss the definitive identification of Edd-1 as an IR counterpart

to one of these new Chandra sources.

3.2 Analysis

3.2.1 Infrared

According to the 2MASS catalog, Edd-1 has IR magnitudes of: J = 15.56 0.08

I ,- H = 12.11 0.06 In i, and Ks = 10.33 0.07 mag. On 2005 July 1 UT, I obtained

1.1 2.4/m spectra of Edd-1 using SpeX on IRTF (Rayner et al., 2003). I discuss the

details of my observation and initial analysis in ('C! pter 2, but I employ a stricter method

here for determining the reddening of the source. To estimate the reddening toward

Edd-1, I assume a GC distance of 8kpc (McNamara et al., 2000). I estimate the infrared

extinction toward Edd-1 in two v- -v-. First, I assume a Cardelli et al. (1989) reddening

law and an intrinsic (H -K)o 0 for a hot star/disk, which gives a value of Av = 29 mag.

As a second method, I estimate the reddening by fitting the K-band spectral continuum to

the Rayleigh-Jeans tail of a T > 104K blackbody. The fit corresponds to Av = 33 mag.

While both values are typical of the GC, I adopt the more conservative value of Av=



1 This work was prepared in collaboration with S. S. Eikenberry, M. P. Muno, R. M.
Bandyopadhyay, and S. Patel. It is published in the A-l i .1lri-- i i1 Journal, 2006, 651: 408.









29 mag. Given this, the dereddened magnitudes are: J = 7.3 mag, H = 6.9 1i i- and

Ks = 6.9 mag.

Figures 3-1, 3-2, and 3-3 show the spectra for the K, H, and J bands respectively,

dereddened by Av= 29 mag. The spectra are dominated by strong hydrogen emission

lines, including Paschen-3, Brackett-7, and Brackett series lines Brl0 Brl4. The Brl3

line is not distinguishable in my spectrum. I observe two neutral Helium lines (A 1.701

and 2.113 pm) and six Hell transitions (A 1.163, 1.736, 1.772, 2.189, 2.038, and 2.348 pm).

The Hell A 1.736 is blended with the Brl0 line. In the K-band I also observe metal lines

from CIII and NIII, consistent with an accretion signature or a colliding wind system.

I fit a Gaussian function to the line profile to determine the line centers and FWHM. I

estimate the spectral resolution of the instrument by measuring the width of OH sky lines

and correct my measured line widths accordingly. I present the line centers, equivalent

widths, and full-width velocity in Table 3-1.

Most of the emission lines are broad with a full-width velocity above 300 km/s.

The Br7 line is strongest and has a full-width velocity of 710 km/s. Given the spectral

resolution here, it is not trivial to deconvolve the Br7 line from neighboring HeI emission

at A2.162 2.166pm. Detailed modeling is required to accurately assess the Br7 line

equivalent width independently of the HeI contribution and such analysis is beyond

the scope of this work. Both Morris et al. (1996) and Hanson et al. (1996) study the

IR spectra of massive stars without quantifying the relative contributions. Morris

et al. (1996) note that the apparent .-i-,ii1! i 1 y of the Br7 line is likely caused by HeI

contribution, but -,v the contribution is relatively weak in the context of LBV and

Ofpe/WN9 stars. Since my determination of spectral classification is based on broad

X-ray and IR spectral features and not on specific line ratios, it is unlikely that the

composite nature of the line affects my results here. Three Hell lines in Edd-1 show

P Cygni profiles at 2.034 pm, 2.189 pm, and 2.348 pm (Fig. 3-4, Table 3-2). I calculate









the differential velocity from the line center to the blue edge and get an average v = 170

km/s. Error due to pixel size and peak location is 70km/s.

3.2.2 X-ray

Muno et al. (2004a) examined the spectrum and variability of the 2-8 keV X-ray

emission from Edd-1 as part of a study of ,2000 X-ray sources detected toward the

Galactic Center. Here, I summarize the properties of the X-ray source derived from that

study, based on 626 ks of Chandra observations taken between 1999 September and 2002

June. The initial X-ray ain i, -; was performed by my collaborator Michael P. Muno and

is described in detail in Muno et al. (2004b). First, the pulse heights of each event were

corrected to account for position-dependent charge-transfer inefficiency (CTI; Townsley

et al., 2002), and the lists were cleaned using standard tools in CIAO version 3.2 to

remove those that did not pass the standard ASCA grade filters, that did not fall within

the good time intervals defined by the Chandra X-ray center, or that occurred during

intervals when background rate flared to > 3a above the mean level. Counts were then

extracted from a contour enclosing 9n'1 of the point spread function around Edd-1, and

binned them as a function of time to create light curves, and as a function of energy

to create spectra. The background was estimated from an annular region surrounding

the source. For each observation, he obtained the instrumental response functions from

Townsley et al. (2002) and computed effective area functions using the CIAO tool mkarf,

and averaged these, weighted by the numbers of counts in each observation.

The X-ray emission in Edd-1 varied in intensity by a factor of -3 in the 2-8keV

range. The three observations in which the source was faint lasted only ~30 ks in total,

which is a small fraction of the total exposure. These did not provide enough signal to

test whether the spectrum varied along with the flux, so only the average spectrum is

examined in detail.

The average X-ray spectrum of Edd-1 is di !,1 'i in Figure 3-5. The most prominent

feature is line emission centered at 6.7 keV from the n=2-1 transition of He-like Fe with









an equivalent width of 2.2keV. In addition, lines are evident from Sulfer at 2.4 keV, Argon

at 3.1 keV, and Calcium at 3.9 keV. Typically, accreting binaries are modeled with a

multi-temperature blackbody plus a power-law model; however, such a model would not

reproduce the prominent line emission observed in Edd-1. The presence of these lines

motivated my collaborator to model the spectrum as a thermal plasma (apec in XSPEC;

Arnaud, 1996) whose emission has been absorbed by interstellar gas and scattered by

interstellar dust. The spectrum is consistent with a kT = 2.00. keV thermal plasma

absorbed by a column density of NH = 3.5+0. x 1022 cm-2. Despite the prominence of the

line emission, the abundances of S, Ar, Ca, and Fe are consistent with the solar values (see

Table 3-3).

However, the absorption column corresponds to a reddening of only Av = 20 mag,

which is much lower than that inferred from the IR spectrum, Av = (29 33). Although it

is common for the absorption column inferred from the X-ray spectrum to be I7.1/, than

would be derived from the IR spectrum (e.g., if the X-rays are produced by a neutron star

embedded in the wind of its companion), there is no known physical situation in which

one would expect the column of material absorbing the X-rays to be smaller. Instead,

the model for the X-ray probably under-estimates the amount of flux produced below 2

keV by Edd-1, which means that the absorption is under-estimated as well. Therefore,

my collaborator has added a second, cooler plasma component to the model of the X-ray

spectrum, and fixed the extinction toward the X-ray source at NH = 5.2 x 1022 cm-2

(from Predehl and Schmitt, 1995, forAv = 29). The spectrum can be adequately modeled

with two plasma components of temperatures kT = 0.7 0.1 keV and kTh = 4.6 0.7

keV, with the cooler component producing ~-:i i. of the observed 2-8 keV flux. However,

when the spectrum is dereddened, the additional soft component produces an enormous

amount of flux between 0.5-2 keV, raising the total inferred luminosity by a factor of

~20 to (1.1 0.3) x 1035 ergs. This would make Edd-1 either one of the most luminous

known colliding wind binaries, or a moderately bright accreting black hole or neutron









star. Although most of this luminosity will never be directly observable given the high

absorption below 2 keV, postulating the existence of a large amount of unseen X-ray flux

is more physically reasonable than supposing that the X-ray flux passes through a smaller

column of gas than does the much larger flux of IR photons. Detailed information about

both models is listed in Table 3-3.

3.3 Discussion

In Figure 3-6, I plot the X-ray and IR luminosity of Edd-1 against known high-mass

stars and massive binary systems, including massive OB-stars, Luminous Blue Variables

(LBVs), HMXBs, and CWBs. I recognize that the X-ray flux (and the IR to a lesser

degree) tends to be variable in binary systems and that my photometric data are not

simultaneous. Thus, the specific points are not as useful as the region subtended by the

general classes. In addition, for these sources, there is often uncertainty in both column

density and distance that can shift the individual points. I identify the distances and

extinction used in placing these sources as well as the individual source names in Table

3-4. I do not plot Low Mass X-ray Binaries (LMXBs) in Figure 3-6 because LMXBs tend

to have Lx/LK >> 1. For Edd-1, Lx/LK ~ 10-4. Even if most of the X-ray emission

from Edd-1 is obscured, and its intrinsic luminosity is 100 times larger than what I have

inferred, the value of Lx/LK is more consistent with an HMXB than a LMXB.

In addition to X-ray and IR color, I observe several interesting spectral features which

may help identify the nature of this source. Because Edd-1 has P Cygni profiles in several

Hell lines, I have searched in the literature for objects with P Cygni profiles in Hell in

the optical and IR. P Cygni profiles tend to appear in the Hel lines of HMXBs and CWBs

(e.g., Cyg X-l: Gies and Bolton 1986; IGR J16318-4848: Filliatre and C I i. 2004; and
rT Carinae: Hillier et al. 2001). P Cygni profiles in Hell lines, such as those observed in

Edd-1, are rare. I have also searched for similar X-ray features such as strong Fe-XXV

emission in Edd-1. Below, I compare Edd-1 to different types of systems containing

massive stars, focusing on similarities to Edd-1's distinguishing spectral features.









3.3.1 Is Edd-1 an Isolated Star?

I show several OB-stars and LBVs in Figure 3-6 with similar color to Edd-1, including

the peculiar Oe-star HD 108. Naz4 et al. (2004) observed HD 108 in the optical and X-ray.

They observe weak Fe emission in the X-ray at 6.6keV. When fit with a two-temperature

plasma, they find kT1 ~ 0.2 keV and kT2 ~1.4keV, cooler than that observed in Edd-1

when using a two-temperature model, but consistent with the low Av model. Although

some have -ii-:.- -i .1 that HD 108 is a binary because of its strong X-ray emission,

long-term observations by Nazd et al. (2004) -,-.:. -1 that HD 108 does not exhibit the

same behavior as classical short- or long-term binaries. The H and Hel lines in HD 108

have been observed to change from strong P Cygni profiles to simple absorptions (Naz6

et al., 2004). The fact that Edd-1 shows P Cygni profiles in Hell, not Hel, -ir-.-.i -I-

the wind in Edd-1 is arising in a hotter region. The K-band spectrum of HD 108 shows

primarily emission, including Br7 and Hel A2.114pm (\! .iii: et al., 1996), consistent with

the IR spectrum of Edd-1. Overall similarities in the X-ray plasma temperatures, infrared

emission features, and the presence of a Helium wind -ii-:.: -1 that Edd-1 and HD 108 may

be similar objects.

The Bp star a Orionis E, also shares some of Edd-1's distinguishing characteristics.

In quiescence, its X-ray thermal temperature is measured between 0.3 and 1.1 keV and

reaches 3.5keV in an outburst, consistent with the Edd-1 models. While Fe-XXV (6.7keV)

is weakly present, the Fe K, (6.4keV) appears in excess during a flare (Sanz-Forcada

et al., 2004). In contrast, Edd-1 shows weak Fe K, emission and strong Fe-XXV emission.

The Helium lines in a Orionis E vary strongly (Groote and Hunger, 1982), but this

variation is interpreted as inhomogeneous chemical abundances at the stellar surface

(Reiners et al., 2000).

LBVs have similar spectra to B[e] stars and are known for -1i 'i- variable Br7 lines

as well as HI and Hel emission. I plot the positions of the LBV P Cygni and the LBV

candidate known as the Pistol star in Figure 3-6. The X-ray luminosities that I cite for









these sources are only upper limits (see Muno et al., 2006b). Bright X-ray emission, such

as that observed in Edd-1, would not be expected from an isolated LBV.

In short, many known isolated LBVs do not show strong X-ray emission. The two

isolated OB stars I have considered are both modeled as having a cooler thermal X-ray

plasma than Edd-1. In addition, the strong Fe-XXV feature in Edd-1 is observed to be

weak in isolated stars. Variable H and He lines are observed in isolated stars and P Cygni

profiles can occur, but are rarely observed in Hell as is seen in Edd-1. Thus while I cannot

rule out this classification based on the spectral and photometric data presented here, I

believe the isolated star scenario is less likely than a binary nature for Edd-1. In the next

chapter, I use X-ray variability data to rule out the isolated star scenario.

3.3.2 Is Edd-1 A High-Mass X-ray Binary?

Another possible classification for Edd-1 is that it is an HMXB. Evidence in favor of

this includes indications of wind activity in the IR spectrum and strong Fe emission. The

HMXBs Vela X-1 and Cen X-3 have shown Fe emission the former in eclipse, the latter

out of eclipse (Schulz et al., 2002a; N ', .- et al., 1992). It is rare, however, to find iron

lines with equivalent widths >1 keV. In Cen X-3, both Fe K, (6.40keV) and the Fe-XXV

triplet are observed (laria et al., 2005). The equivalent widths of these lines are measured

at only a few eV when out of eclipse. In contrast, strong line emission is seen during the

eclipses of Vela X-l, with Fe K, equivalent width as large as 1.3 keV in deep eclipse (e.g.

Choi et al., 1996; Schulz et al., 2002a). Neither of these show dominant Fe-XXV emission.

Both of these HMXBs have a supergiant OB companion. Their optical spectra show

Hydrogen and Helium in both absorption and emission associated with the star (Mouchet

et al., 1980; Dupree et al., 1980).

A more rare HMXB companion, a supergiant-B[e] star (sgB[e]), has been observed

in CI Camelopardalis (CI Cam). While several sgB[e] stars have been observed in the

Magellanic Clouds (Zickgraf et al., 1986), they are rarely observed in the Milky Way. Like

Edd-1, CI Cam has prominent line emission in its IR spectrum, but additionally CI Cam









has forbidden Fe lines. Forbidden lines have not yet been identified in Edd-1. In the

X-ray, CI Cam has a large Fe K, line with an equivalent width of 940 eV (Filliatre and

C'!I ly, 2004), but it does not exhibit strong Fe-XXV emission.

Recently IGR J16318-4848 has been tentatively identified as having a sgB[e]

companion. While the X-ray Fe K, emission from IGR J16318-4848 is negligible, a

Hel wind has been observed with a velocity of 410 40km/s (Filliatre and C'!i I r, 2004).

The Hell lines in Edd-1 indicate a much weaker wind with a velocity of 170 70 km/s.

P Cygni profiles are fairly common among high-mass systems; however, they are more

often observed in H or Hel than Hell (e.g. Vela-X-1, IGR J16318-4848). To date, only one

HMXB has been observed to have a Hell wind: Cyg X-1. Cyg X-1 is a black-hole binary

with an O-star companion. It shows weak Fe K, with an equivalent width of only 16eV.

Gies and Bolton (1986) found that the P Cygni profile of Hell A4686A in Cyg X-1 varied

with orbital phase and the maximum P Cygni velocity was ~ 100 km/s, similar to that of

Edd-1. The varying P Cygni profile in Cyg X-1 is modeled as a focused stellar wind whose

flux related to the mass transfer rate in the system.

The presence of Helium wind and the prominent H and He emission in the IR

spectrum of HMXBs lends support to this scenario for Edd-1. However, although strong

Fe K, lines are seen in HMXBs, this feature is rarely the Fe-XXV line observed in Edd-1.

Finally, while the inferred X-ray luminosity of Edd-1 is more typical of an HMXB than an

isolated star, the IR luminosity is brighter than the classical HMXBs listed in Table 3-4.

So although an HMXB scenario is supported, I cannot make a conclusive classification

on the basis of these data. In the next chapter, I use the X-ray variability of Edd-1 to

constrain physical scenarios in which Edd-1 is consistent with a HMXB.

3.3.3 Is Edd-1 A Colliding-Wind Binary?

Strong wind features may indicate a colliding-wind binary (CWB), which is

an association of two massive stars with strong winds. The CWBs HD 152248 and

HD 150136 have similar X-ray thermal temperature to Edd-1 in the single-temperature









model (Sana et al., 2004; Skinner et al., 2005). For example, 72 Velorum (hereafter 72

Vel), a WC8+07 binary, has a measured temperature kT = 1.5keV (Skinner et al., 2001).

Because the WR star dominates line emission, 72 Vel appears Helium-rich. In WR+O

binaries, weak Brackett series emission may be present, but broad Helium lines dominate

emission (Skinner et al., 2001; Varricatt et al., 2004) It is possible the He-wind observed

in Edd-1 comes from an eclipsed or obscured WR companion. However, because Brackett

series emission dominates Edd-1, I find this scenario less likely. I further consider the

unlikelihood of a WR+O binary system in the next chapter.

Based on the X-ray spectrum, Edd-1 holds the greatest similarity to Eta Carinae. Eta

Carinae is an unusually bright X-ray/IR source whose nature is not definitive, but models

-~ti-'-I a CWB containing an LBV+O star pair. The greatest difference between Edd-1

and Eta Carinae is the infrared luminosity. Eta Carinae is intrinsically much brighter in

the IR by over two orders of magnitude. The similarities between Edd-1 and Eta Carinae

are primarily in the X-ray. Viotti et al. (2004) find that the thermal component of Eta

Carinae has a temperature of 5.5 keV with NH = 4.8 x 1022cm-2, comparable to the

two-temperature model of Edd-1. In addition, Eta Carinae has a sizeable Fe-XXV line

varying between 0.9 and 1.5 keV (Viotti et al., 2004). This is the only source I have found

in the literature with a Fe-XXV line of similar equivalent width to that observed in Edd-1.

Typically assumed to be an LBV in a binary system, the X-ray emission of Eta Carinae is

modeled as arising from the colliding wind. Detection of P Cygni profiles in Eta Carinae's

optical spectrum is consistent with a CWB classification. Steiner and Damineli (2004)

observed variable Hell emission. It is believed to originate in a dense stellar wind, however

the energy supply is still debated (\! irtin et al., 2006). Because wind from the LBV

in Eta Carinae is mostly cool, the Hell emission is believed to be from the companion,

.-.-I_, -I, d to be an O-star. The presence of strong Fe-XXV emission in Eta Carinae, the

similar X-ray thermal temperature to CWBs, and the presence of Helium winds in CWBs

lend strong support to Edd-1 being a CWB.









In summary, Edd-1 shares qualities with a variety of high mass systems including

isolated stars, HMXBs, and CWBs. While an isolated star scenario is supported by the

IR emission spectrum, it is not consistent with the high X-ray luminosity. Edd-1's X-ray

luminosity is fairly common for an HMXB, and although I observe strong Fe K, emission,

I do not observe strong Fe-XXV emission in HMXBs. The CWB Eta Carinae does show

strong Fe-XXV emission, but in general, CWBs are not as X-ray luminous as Edd-1. Each

type of source can show a strong He wind like Edd-1, but the Hell lines in these sources

rarely have P Cygni profiles. The HMXB Cyg X-1 is the only source I have found in

the literature with a Hell wind of similar velocity to Edd-1. Based on the spectral and

photometric observations presented here, the evidence favors either a CWB system similar

to Eta Carinae, or an accreting binary system. Further distinction of these potential

classifications is derived from the variability of the source, presented in the next chapter.

3.4 Summary

Edd-1 is a reddened source with an estimated extinction Av = 29 mag. I have

identified Edd-1 as having prominent emission lines in the X-ray and IR. The Hell lines

show P Cygni profiles consistent with a 170 km/s wind. In addition, Edd-1 has very

strong Fe-XXV emission in the X-ray, the line having an equivalent width of 2.2 keV.

While it is difficult to positively classify Edd-1 based on the X-ray and IR characteristics

presented in this chapter, Edd-1's spectral features indicate the presence of a high mass

star. I have compared Edd-1 to OB stars, LBVs, HMXBs, and CWBs -all of which are

types of systems containing a massive star. The X-ray and IR color of Edd-1 is somewhat

consistent with each of them; however, the prominent spectral features do not match

exactly the characteristics of any of these types. In the next chapter, I present follow-up

infrared and X-ray observations and use the source variability to further refine the possible

nature of Edd-1.











0.5


0.4 I I (J II z


X


> 0.3




0.2-




0.1 ..
1.9 2.0 2.1 2.2
Wavelength (microns)

Figure 3-1. The K-band spectrum of Edd-1 shows strong Bry, Br6,
P Cygni profiles are seen in several of the Helium lines,
wind around a massive star.


2.3 2.4



and Hel emission.
-', ii-; a Helium











0.035


0.030


X

0.025
.6



0.020




0.015


1.5 1.6 1.7 1.8
Wovelength (microns)


Figure 3-2. Brackett series emission dominates the H-band spectrum of Edd-1.












0.8



0.6



0.4



0.2



0.0

1.15 1.20 1.25 1.30 1.35 1.40 1.45
Wavelength (microns)

Figure 3-3. The PaP line is clearly visible in the J-band spectrum of Edd-1. Atmospheric
noise distorts the spectrum at A 1.35 1.42/m and at A < 1.15pm.


































2.030 2.035 2.040 2.045
Wavelength (microns)


2.185 2.190 2.195 2.200
Wavelength (microns)


2.32 2.33 2.34 2.35
Wovelength (microns)


Figure 3-4. P Cygni profiles for the Helium lines. The dotted line shows the approximate
continuum level. The vertical line is placed at the vacuum center wavelength.












S100I
> o-3

U


CO


10-5
4


x 0 ------

-2' t
-4
2 4 6 8
Energy (keV)

Figure 3-5. The average Chandra X-ray spectrum of Edd-1. The source di pl prominent
line emission from the n 2 1 transitions of He-like S at 2.4 keV Ar at 3.1
keV, Ca at 3.9 keV, and Fe at 6.7 keV.











0

A


34
Iog(Lx)


Figure 3-6.


Comparison of the X-ray and IR luminosity of Edd-1 to other known X-ray
systems containing massive stars. When a single source is observed at varying
luminosities, the two points are connected with a line. The lighter symbol
indicates a quiescent state. Because simultaneous X-ray and IR data is not
typically available, I generally have only one K-band data point. Further
information on these sources is available in Table 3-4.


41



40



39


Edd-
Isolat
HMXB
CW9


O
0


38


S
*


A


I
ed Staor










-*



*


37



36



35


30


32


36


38


_______________


I h























Band Line


Table 3-1.
Ac (pm)
(pm)


Identified Lines
EW VFW
(A) (kms-l)


Comments


Hell 7-5
HI 5-3 (Pa3)
unknown
HI 14-4 (Br14)
HI 12-4 (Br12)
HI 11-4 (Br1l)
Hel 4D-3P, 3D-3Po
HI 10-4, Hell 20-8
Hell 19-8
HI 8-4 (Br6)
CIII/ NIII
Hel 4S-3P, 1S-1Po
HI 7-4 (Br7)
NIII


1.163
1.282
1.503
1.589
1.641
1.681
1.701
1.736
1.772
1.945
2.104
2.114
2.166
2.247


-29 4
-19 2
-1.9 0.4
-2.3 0.3
-6.6 0.4
-5.9 0.6
-2.7 + 0.4
bl
-10.0 + 0.5
-36.2 0.6
bl
-13.8 0.5
-36.6 0.3
-1.7 + 0.2


450
310
340
280
470
570
370
bl
2160
640
bl
590
710
310


blend?






blend

Hel blend?
blend
blend? CIII/NIII


Note. -The identified lines transitions, vacuum wavelength, equivalent width,
and full-width velocity. The full-width velocity has been corrected for the intrinsic
line width of the instrument. Vacuum line centers are obtained from Hanson et al.
(1996); Morris et al. (1996); Figer et al. (1997); Wallace et al. (2000); Schultz (2005).
Uncertainty in velocity is up to 70km/s due to uncertainties in measured line width.























Table 3-2. P Cygni Line Velocity

Avac /I. .
Band Line (pm) (km/s)

K Hell 15-8 2.0379 180
K Hell 10-7 2.1891 150
K Hell 13-8 2.3464 180



Note. -The velocity is calculated
based on the difference between
the blue edge and vaccum central
wavelength. Line centers are as
referenced in Table 3-1. Because
of discrepancies in the location
of the Hell 13-8 line, I estimate
the central wavelength using the
equation 0.0911138Z-2n( n).
Uncertainties due to pixel size and
peak location give a AV = 70km/s.




















Table 3-3. Chandra Spectrum of Edd-1
Parameter 1-kT 2-kT

NH (1022 cm-2) 3.5+.3 5.2a
kTs (keV) 0.7'.1
KEM,s (1056 cm-3) ... 32+9
kTh (keV) 2.0+.5 4.6 .7
-0.2 -0.7
KEM,h (1056 cm-3) 5+1 1.4+
Zs/Zs,o 1.105 .90.2
'D-0.4 -0.2
ZAr/ZAr, 110.7 13+0.7
-0.7 -0.4
Zca/Zca, 1.5 +0.8 2.29 9+1.3
'C-0.8 '-0.7
ZFe /ZFe, 1.4 +0.3 1.5 +0.2
'* 0.3 '- 0.2
FFe-xxv (10-7 photon cm-2 s-1) 3 2 2 1
X2/v 125.2/107 105.6/105
Fx (10-13 erg cm-2 S-1) 1.2 1.3
Lx,s [D/8 kpc]2 (1033 erg s-1) .. 110
Lx,h [D/8 kpc]2 (1033 erg s-1) 6 3

aParameter held fixed.


Note. KEM is the emission measure for each plasma
component, f r. i,,,1V [D/ 8 kpc]2. Uncertainties are ,91'.
confidence intervals (A 2 2.76). Fluxes and luminosities
are reported for the 0.5-8.0 keV band; most of the observed
flux is in the 2-8 keV bandpass.









Table 3-4. X-ray and IR Source Comparison


Source


Class


Edd-1
CI Cam
IGR J16283-4838
IGR J16283-4838
XTE J1906+090
XTE J1906+090
GRO J2058+42
GRO J2058+42
X1908+075
Cyg X-1
Cyg X-1
Cen X-3
Vela X-1
HD 152248
HD 152248
HD 150136
72 Vel
72 Vel
rl Car
P Cygni


unknown
sgB[e]+X
Be+NS
Be+NS
Be+P
Be+P
Be+X
Be+X
OBI+NS
09I+BH
09I+BH
OI+NS
BI+NS
08I+O
081+0
03+06V
WC8+07
WC8+07
LBV+O
LBV


log(Lx) a log(Lk)


35.04
33.54 (q)
34 (q)b
35 (b)b
34.84 (q)
36.84 (b)
33.47
33.95
36c
36.80 (q)
37.20 (b)
37.70d
33.32
32.90 (q)
33.04 (b)
33.38
32.89 (q)
33.17 (b)
34.88
<31.0e


38.56
39.23
35.66
35.84
36.06
36.06
37.55
37.55
37.96
37.86
37.86
37.81
37.94
37.87
37.87
37.85
38.10
38.10
40.86
39.02


d
(kpc)
8
5
5
5
4
4
9.0
9.0
7
2.5
2.5
8
1.9
1.757
1.757
1.32
0.278
0.278
2.3
2.1


AK References


3.4
0.3
0.9
1.4
1.8
1.8
1.2
1.2
2.3
0.36
0.36
1.8
0.28
0.15
0.15
0.20
0.99
0.99
2.3
0.2


[1], [2]
[3],[4]
[5]
[5]
[6]
[6]
[7]
[7]
[8]
[9],[10]
[9],[10]
[11],[12]
[13],[14]
[10],[15],[16]
[10],[15],[16]
[10],[17]
[18],[19]
[18],[19]
[22],[23] ,[24]
[20],[21]








Table 3-4. Continued


d
Source Class log(Lx)a log(Lk) (kpc) AK References
Pistol LBV <32.0 39.66 8 3.2 [21],[25]
a Ori E Bp 31.30 35.94 0.40 0.06 [26],[27]
HD 108 06f 33.0 37.60 2.1 0.2 [10],[28],[29]
HD 152408 08Iaf <31.7e 38.29 2.16 0.16 [21],[28]
HD 151804 081 31.9e 38.07 1.66 0.13 [21],[28]
X174516.1 0? 33.3 39.28 8 2.7 [21]
H2 0? 33.1 39.45 8 4.5 [21]


aNote that X-ray luminosities are reported from a variety of sources, most consistent with
a 0.5-8keV range. I further mark objects for which a significantly different energy range is
reported as follows (b-e).
bSource reports bolometric luminosity based on RXTE observations.
cSource reports 1.5-100keV luminosity based on RXTE observations.
dSource reports 2-30keV luminosity based on Ginga observations.
'Source reports 0.1-2keV luminosity based on ROSAT observations.
Identification of sources plotted in Figure 3-6. Here, I specific the source classifications
(when available) as well as the distance and K-band extinction used in calculating
luminosity. The luminosities are in erg/s. The reference numbers are as follows: [1] this
work; [2] Muno et al. (2003); [3] Boirin et al. (2002); [4] Clark et al. (2000); [5] Beckmann
et al. (2005); [6] Gogiiu et al. (2005); [7] Wilson et al. (2005); [8] Morel and Grosdidier
(2005); [9] Schulz et al. (2002b); [10] Mafz-Apelldniz et al. (2004); [11] Coe et al. (1997);
[12] N ';, i-- et al. (1992); [13] Schulz et al. (2002a); [14] Hyland and Mould (1973);[15]
Cassinelli et al. (1981); [16] Sana et al. (2004); [17] Skinner et al. (2005); [18] Schild et al.
(2004); [19] Williams et al. (1990); [20] Turner (1985); [21] Muno et al. (2006b); [22] van
Genderen et al. (1994); [23] Seward et al. (2001); [24] Evans et al. (2003); [25] Figer et al.
(1998); [26] Groote and Hunger (1982); [27] Groote and Schmitt (2004); [28] Leitherer and
Wolf (1984); [29] Nazi et al. (2004)









CHAPTER 4
DISCOVERY AND INTERPRETATION OF AN X-RAY PERIOD IN THE GALACTIC
CENTER SOURCE CXOGC J174536.1-285638 (EDD-1)2

4.1 Introduction

In 2005, I identified a star, hereafter called Edd-1, as the first spectroscopically

confirmed infrared (IR) counterpart to a low luminosity Chandra source (as described in

C! Ilpter 3). Initial analysis of Chandra and XMM-Newton archival data -i-.-,- -1.i the

presence of short time-scale variability which prompted IR followup observations to look

for spectroscopic evidence for a short-period binary. My subsequent analysis, presented

in this chapter, utilizes more X-ray data, and shows Edd-1 to be relatively steady in the

IR and X-ray over short time baselines. I use IR spectroscopy to search for variations

in Edd-1's emission features. In addition, I combine a long-term Chandra monitoring

campaign with archival XMM data to search for periodicity in the X-ray light curve and

find a period of 189 6 di-,v In this chapter, I discuss Edd-l's variability in the IR and

X-ray, and examine the implications of a 189 d period for the nature of the source.

4.2 Observations and Analysis

4.2.1 Infrared

On 2006 Aug 02-04 UT, I obtained J, H, and K band (1.1-2.4 pm) spectra of Edd-1

using SpeX on IRTF (Rayner et al., 2003). Dithering along the 0.5 arcsec slit, I obtained

184 exposures of 120s each over the course of three half-nights, giving us a time baseline

of 3-4 hours per night. In the short-wavelength, cross-dispersed mode, I get a resolution

of R ~ 1200 over the JHK bandpass. Target observations were followed by observations

of the GOV star HR 6836 at similar airmass for removal of telluric absorption features. I

extract spectra using the standard SpexTool procedure for AB nodded data (detailed in

C! Ilpter 2), resulting in a series of -1.:,--l1 tracted, wavelength-calibrated spectra (Vacca



2 This work was prepared in collaboration with S. S. Eikenberry, M. P. Muno, and
R. M. Bandy. .i' .1l:ay. It has been submitted to the A-l i, .1,,1. -i. I Journal.









et al., 2003; Cushing et al., 2004). Using my previous observations taken on 2005 July 1

UT (C'!i ipter 3), I adopt a reddening value of Av = 29 mag and apply this correction to

all data.

I combined the spectra from each night to test variability on multiple time scales.

Figure 4-1 shows the series of twenty-one K-band spectra taken over the course of our

observations, with integration times between 8 and 20 minutes per spectrum. I list the

specific time stamps and exposure times of these spectra in Table 4-1. To search for radial

velocity variations in the emission lines, I track the line centers with two methods: first by

taking a statistical mean of the wavelength around the line center, weighted by flux, and

second by fitting a Gaussian to the line. I find no radial velocity variations, nor do I find

significant flux variations in the lines. I checked for IR variability on 1 year, 3 d i-, 3 hour,

1 hour, and 30 minute baselines and found no evidence of periodic variability or flares in

this sample. The only apparent variation is in the structure of the Br7 line complex (see

Figure 4-2), but this does not often vary more than ~ 5 times the RMS spectral difference

in the vicinity of the A2.164 Helium component. Further I note that this region is affected

by our data reduction process (i.e., the removal of the intrinsic Brackett absorption in the

GOV).

4.2.2 X-ray Data

Muno et al. (2004a) examined the spectrum and variability of the X-ray emission

from Edd-1 as part of a study of ,2000 X-ray sources detected toward the GC. The

analysis is described in detail in Muno et al. (2004b) and summarized in C'i lpter 3. I list

the Chandra data used in our analysis in Table 4-2. During the Chandra observations, the

X-ray emission in Edd-1 varied in intensity by a factor of ~3 in the 2-8 keV range.

In Table 4-3, I list XMM archival data used in our analysis. While the 2001-2002

XMM data encompass fairly short observations (exposure time < 7 h), in 2004 I have four

observations of 40 consecutive hours each. I followed the standard XMM reduction routine

for odf data, first building a cif-file, then using odf ingest and epchain to create









an event list. Once I had an event list, I located Edd-1 and extracted light curves and

spectra from a circle with a 200 pixel radius. The background was calculated from a ring

extending 300-500 pixels from the source center. I set spectral bin size at 200 eV and plot

two representative spectra in Figure 4-3. The flux varies by a factor of five between these

observations. Because of the extremely low count rate, I cannot fit the fainter spectrum

with XSpec models.

I extracted light curves at one-hour intervals over the full 2-8 keV band, as well

as from the "soft" 2-4 keV band, and the Ii ird" 4-8 keV band separately. Using these

one-hour resolution light-curves, I perform a period i i1.,i -;i searching for periodicities

in the range of 0.1-40 hours, but find no significant periods in this range. I find that the

XMM X-ray flux is constant within Poisson errors during a single observation (as long as

40 hours). I do find a 4o variation in consecutive observations separated by five months

(see Fig. 4-3). Thus I calculate a single flux value for each XMM observation epoch

and combine these measurements with the Chandra light curve in Figure 4-4. Using the

combined light curve, I can test for the presence or absence of periodic flux variations on

timescales longer than 40 hours.

Using the method of Home and Baliunas (1986), I perform a periodogram ,:! 1,~-i;

of the combined light curve and find a period of 189 6 d ,v-. In Figure 4-5, I show

the resultant periodogram which tests for periodicity on scales of 1-1500 di The

peak at 189 d is clearly distinct and additional peaks are visible at integer multiples

of the period. In Figure 4-4, I plot the X-ray light curve folded on the 189 d period.

Analytically estimating the significance of a signal in non-uniformly sampled data is

non-trivial. Thus, in order to estimate the confidence of this detection, I perform a Monte

Carlo simulation as follows. I take the existing data set and maintain the same sampling

intervals throughout. For each Monte Carlo realization, I randomly reassign the observed

flux values to the time samples, effectively scrambling the light curve. In 30,000 trials, I

do not achieve a peak power approaching the power of our original periodogram, implying









that the 189-d-,v period is not due to random noise with a confidence level greater than

99.9 7' .

The previous test accounts for white noise variability, but does not consider red

noise, which is a significant source of false peaks in X-ray power spectra of X-ray binaries

(Titarchuk et al., 2007). Red noise is a flux variation in the power spectrum that can

be parameterized with a frequency dependence f-~. A white noise process will generate

a flat power spectrum such that/3 ~ 0; a value of/3 2 describes random walk noise

(Timmer and Koenig, 1995). A 3 ~ 1 dependence has been noted in stellar-mass black

hole candidates and may be strongly related to accretion physics in the system (l\ I. -II:II

et al., 1994; Timmer and Koenig, 1995; Titarchuk et al., 2007). Following the method of

Timmer and Koenig (1995), I test the possibility of red noise creating a false signal which

matches the strength of our periodogram. By simulating a number of red noise dominated

light curves of varying power law slope, /, I find that as 03 increases, more noise gets

shunted near the period frequency, and the significance of our detection decreases. I find

the significance of our period detection remains above 3a for values of 3 < 1.0 and above

2.5a for 3 < 1.5 showing that the significance decreases slowly as red noise is increased.

4.3 Discussion

4.3.1 Infrared Variability

In the initial discovery spectrum, I identified three Hell lines with P Cygni profiles:

A2.0379 pm, 2.1891 pm, and 2.3464 pm (C'! lpter 3). In the analysis of our 2005 data,

I estimated the P Cygni velocity at 170 70 km/s. I repeat our analysis on the 2006

data to search for variations and find the approximate velocity of the wind is 200 70

km/s. (The error is dominated by the spectral resolution.) I find no evidence of changes in

the P Cygni profile or velocity over our three d-,v observations. Also, the 2005 and 2006

spectra have consistent P Cygni profiles and velocities.

In Figure 4-2, I show several close-ups of the Brackett-7 region of Edd-l's spectrum

over the course of our three night IRTF run in 2006. To the left of the A2.164/m marker,









I see minor variances in the Helium contribution to the line. Because this line cannot

be resolved from the larger Br7 contribution, it is difficult to determine the origin of

this change. The RMS spectral difference rarely reaches 5a between any two events

which are separated by ~ 1 hour. The observed differences are primarily in the "i,

of the line (~ 2.164pm or ~ 2.168pm). Higher resolution spectroscopy is required to

determine whether the changes in the Helium contribution are intrinsic to Edd-1 rather

than an artifact of the data analysis. The observed variations do not have any detectable

periodicity. It should also be noted that this region is affected by the data reduction

process, as described in 4.2.

My 2006 IR spectra were obtained about two di,- after the Chandra observations on

Day 2402 in the X-ray light curve (see Fig. 4-4), where the object is transitioning from an

apparent low-flux state to a high-flux state. Because I have no IR data consistent with the

lowest X-ray flux events, it is impossible to know from these data if the apparent Helium

variability at A2.164pm I observe is associated with this X-ray flux transition. Because

the infrared spectra from 2005 and 2006 are taken at nearly the same phase in the X-ray

light-curve, I do not expect to find radial velocity variations. Nor can I observe variations

if the orbital velocity is less than 70km/s.

4.3.2 X-ray Variability

In C'!i pter 3, I argue that Edd-1 contains at least one massive star based on the

presence of P Cygni profiles in the IR spectrum. In comparing Edd-1 to other systems

containing massive stars, I showed that the X-ray to IR luminosity ratio, Lx/LK 10-4,

is consistent with both CWB and HMXB systems. Furthermore, X-ray variability similar

to that seen in Edd-1 is not observed in isolated massive stars (Cohen, 2000), a finding

which favors a binary interpretation.

In the standard models for CWBs, X-ray emission arises from the shock front of

colliding winds in two massive stars (see, e.g. Luo et al., 1990; Sana et al., 2004; De Becker

et al., 2006). Observed variability is often attributed to phase-locked flux modulations









due to the effect of variations in absorption along the line of sight and variations in X-ray

emission as a function of orbital phase. In this situation, the X-ray periodicity reflects

an orbital period. It is possible that stellar rotation or photospheric pulsation may also

produce periodic X-ray modulations. Models of such behavior are often employ, -1 to

explain the 84 d quasi-periodicity in Tr Carinae (Davidson et al., 1998). In these situations,

the modulation of the X-ray flux is correlated to recurrent behavior affecting the wind

emission, but not related to the orbital period.

However, in HMXBs, periodic X-ray flux changes can be the result of either orbital

or superorbital motion. Several HMXBs in the Liu et al. (2006) catalog have orbital

periods of > 100 d. These long-period systems are often X-ray transients with Be star

counterparts (Lewin et al., 1997). Cen X-3 and Cyg X-1 are both high-mass binary

systems showing both orbital and superorbital periods. The superorbital periods of these

systems are 140 d and 142 d respectively, values which are associated with a processing

accretion disk. Their orbital periods are 2.1 d and 5.6 d (Ogilvie and Dubus, 2001).

However, if Edd-1 has a similar short orbital period in addition to the 189 d period, I do

not detect that short period in our current X-ray data (Fig. 4-5). Thus I restrict myself to

exploring the possibility that the 189 d period is orbital rather than superorbital.

4.3.3 The Orbital Period Assumption

For both the CWB and HMXB cases, the X-ray periodicity can trace the orbital

period. CWBs have periods of d ,v- to years while HMXBs have shorter periods ranging

from hours to d ,v- (Vanbeveren et al., 1998; Lewin and van der Klis, 2006). In ('!i Ilter

3, I determine an absolute magnitude MK = -7.6 0.3 for Edd-1 using a distance

of 8 kpc, reddening of AK = 3.4, and a 2MASS magnitude of Ks = 10.33. I can use

Edd-1's exceptional brightness and the X-ray period to place constraints on the nature of

the system. For our purposes, the pliin i y" star (mass, MOB) will refer to the massive

OB-star and the "secoil i. y star (mass, ._.) will refer to the companion whose nature

has yet to be identified.









Using the mass function


f (qsini)3 PU
f (q, 1) orb
(1 + q)2 27rGMOB

where q = M_/MoB, I can generate a parameter space of orbital velocities and mass ratios

for the system. Massive OB stars can range from 20 100Me and still emit strongly

in the infrared (see, e.g. Cox, 2000; Girardi et al., 2002). In Figure 4-6, I plot the mass

ratio as a function of the inferred orbital velocity for a range of primary masses and note

that the orbital velocity is less than our infrared spectral resolution of 70 km/s for cases

of mass ratio q < 0.5. Even for higher mass ratios, a radial velocity variation would

have low signal-to-noise with our current observations. Thus, I require higher resolution

spectroscopy in order to observe radial velocity variations in the infrared associated with

this periodicity.

In the next two sections I discuss the possibility that the modulations in X-ray flux

are caused by (1) obscuration of the X-ray source by stellar wind; and (2) eclipse of the

X-ray source.

4.3.3.1 Wind Obscuration Scenario

Edd-l's X-ray light-curve shows flux variation by a factor of 5 over the course of the

189 d period. Using this information, if I assume that the X-ray emitting source is being

obscured by a windy counterpart, I can calculate the column density of the wind required

to cause such absorption. Because there are insufficient counts in the low-flux state to

fit the X-ray spectrum, I use the model fit from the high-flux state and create a dummy

response with XSPEC to measure the amount of absorption required to decrease the flux

by a factor of five. Given our initial NH = 5.2 x 1022cm-2 (see ('!i Ipter 3), I find the

column density from the obscuring wind must reach NH w 2.5 x 1023CM-2 to cause the

flux variation observed in Edd-1.









To estimate the absorption column caused by a stellar wind with density p(r), I use

the equation:

NH j (r)dl (4 1)

For a spherically symmetric shell, and a star with mass-loss rate M and escape velocity

'Uc

p(r) (4-2)
4vKT2v0
For an edge-on view of the system, dl = dr, thus

/" M dr M
NH d (43)
a 47v00 r2 47TVUROB

Normalizing for typical values of v. = 1000km/s and = 10-6,1[ ir-1 (see, e.g. Mokiem

et al., 2007), this becomes

NH M -1 RoN
= 4.3 x ('(44)
1023cm-2 10-61[ ir-1 1000kms- R

If I am not viewing the system edge-on, I must take into account the angle through which

I am viewing the wind as an effect on the observed absorption column. I can parameterize

this in terms of an impact factor b such that b r cos 0. In this case, dl = bd, and

fM F/2 2M b b R2 b2
NH = --b cos2 OdO = -- arccos R2 (4-5)
47vb 'o 47v b 2 R R2

where cos 00 = b/R. Larger impact values require windier stars to create the same

absorption column, thus the values of M estimated with Equation 4-4 should be

considered a lower limit of the M required to produce the absorption column that effects

the flux change in Edd-1.

I estimate the mass loss for two special cases. In the first case, I postulate that the

infrared light is dominated by a single bright source. In HMXBs, the star is expected to

contribute more heavily to optical and infrared emission than the accretion disk (Lewin

et al., 1997). In certain CWB cases, especially of lower mass ratios, it is possible that

a single source dominates emission (LUpine et al., 2001). Thus for CWB and HMXB









scenarios in which a single star dominates the IR emission, I use Edd-l's infrared

luminosity and estimate stellar characteristics based on the isochrones of Girardi et al.

(2002) and find that a star with MK ~ -7.6 will likely have a radius ROB N~ 80R

valid for a range of masses 20 100Me. Using equation 4-4, I get a mass-loss rate of

M ~ 4 x 10-53 r-1.

In the second case, I consider a system that contains two massive stars each

contributing half of the infrared luminosity which is only consistent for CWBs containing

two stars of similar bolometric luminosity. These stars would have ROB ~ 20R. and

1M ~ 1 x 10-sMoyr-1. Typical massive O-stars are reported to have mass-loss rates

of 10-6 10-s[ ,ir-1 (Mokiem et al., 2007). Thus, while a relatively windy star

is necessary to produce the flux variations that I observe, the mass-loss rate is not

unreasonable. Therefore a wind-obscuration scenario is consistent with both CWB and

HMXB interpretations of Edd-1.

4.3.3.2 The Eclipsing Binary Scenario

By assuming that the low-flux portion of the dip is caused by an eclipse of the X-ray

region, I estimate a transit time of -r 50 d for the putative eclipse, limited by .,ldi ient

observations of the high-flux stage. I convert the transit time to a velocity by estimating

Vrb = 2RoB/r. Combining this with the mass function, I get

(q sin i)3 4PRB -7 (P/189d) r(46)
(1 + q)2 WrGMOBT3 (7r/50d)3 mOB

where moB and roB are in units of solar masses and solar radii respectively. Assuming

sin i = 1, I then solve the cubic equation for different scenarios. In Table 4-4, I list a

series of mass ratios, q, associated with varying fractions rtB/moB. As an example, I

can examine the two cases as I did above. To complete this numerical exercise, I choose

a median primary mass MOB = 40Me (while acknowledging that a wide range of

masses is possible). If two massive stars each contribute half of the IR luminosity, then

RoB ~ 20Re, rO/moB = 200, and the mass ratio is q ~ 0.05. This resulting mass ratio









is inconsistent with my initial assumption of two massive stars contributing equally to the

emission. If a single massive star dominates the infrared emission, then ROB ~ 80Re,

rOB/moB = 12800, and the mass ratio is q ~ 0.2. I find that adjusting the inclination

does not significantly alter this result because eclipsingg scenarios do not exist at low

inclinations (i < 82; Terrell and Wilson, 2005).

Thus if the system is a CWB, it would have to have a relatively low mass ratio with

the infrared emission dominated by a single source. This implies that the wind emission of

one source overwhelms that of its companion (Luo et al., 1990). It is possible for CWBs

to have lower mass ratios if the secondary is a Wolf-R iv t (WR) star. By the time a

massive star reaches the WR stage, it may have a relatively small mass, but still have

enormously powerful winds (Crowther, 2007). For example, 72Velorum is a WR+O star

with a mass ratio q ~ 0.35 (van der Hucht, 2001). In the case of 72Velorum, the WR

star dominates the IR emission, so the source appears Helium rich (Crowther, 2007).

It is possible that the Helium emission I observe in Edd-1 is evidence of an obscured

WR companion. However, because Brackett series emission rather than Helium emission

dominates the IR spectrum, I find this scenario less likely. In Table 4-5, I list line ratios

of Br7 to Hel 2.114pm and Br7 to Hell 2.189pm for known CWBs and XRBs. In known

WR+O binaries, the Hell 2.189pm is notably stronger than Bry. Comparatively, Edd-1

has much stronger Br7 emission, and hence a quite different Br7/ Hell line ratio from

what is observed in WR+O systems. In fact, I note the Bry/HeI and Bry/HeII line ratios

in Edd-1 are more consistent with HMXBs than either 0+0 or O+WR CWBs. Thus if

Edd-1 is a WR+O CWB, it is very unusual. In the eclipsing binary scenario, Edd-1 would

more likely be an HMXB.

4.3.3.3 Edd-1 as a Wind-Accreting HMXB

In C!i lpter 3, I showed that the X-ray luminosity of Edd-1 (1.1 x 1035 erg s-1) is fairly

typical of HMXBs, within the observed range of X-ray luminosities between INTEGRAL

sources identified as HMXBs (~ 1034 erg s-1; Tomsick et al., 2006; Sidoli et al., 2006)








and the canonically bright sources such as Cyg X-1 and Cen X-3 (~ 1037 erg s-1; N I, -.
et al., 1992; Schulz et al., 2002b). I explore the implications of the observed period for the
case where Edd-1 is an accreting binary system with a compact object. Since the IR data
-,i-.-, -1 that Edd-1 contains a high-mass star, I focus on the case of wind-fed accretion.
Taking the standard accretion luminosity as

Lx = Mc2 (4 7)

where c is the efficiency of converting energy into X-ray light and A is the accretion rate,
I can rewrite this in terms of the mass loss rate of the donor star due to wind such that


Lx 5.7 x 037 ( 11_--'- (-) 0 erg s-1 (4-8)
-10 -4 ind) (t10-6M 5 ir-I


I have normalized the mass loss rate of the primary due to wind and the accretion
efficiency of the system with typical values found in Frank et al. (2002). Frank et al.
(2002) estimate the accretion efficiency, M!/Mid, by comparing the mass flux within
an accretion cylinder to the total mass loss of the donor star. The accretion cylinder is
estimated from the gravitational potential of the compact object, giving

M r racc ind (a)
ind 471a2Vwind(a)

where racc ~ 2GV_/vlU4, Vwind ~ (2GMf/R1)1/2, and a is the orbital separation. This
gives us
M (F1 2 ROB 2 (49)
~ n ----- (4-9)

Normalizing to standard values, and using our known values, I get

Lx 3 q3 Mind P
x 1 35c O (410)
035erg s- OB (1 + q) 10-5Meyrr-1 189 d

where roB and moB are normalized to solar radii and solar masses respectively. This
form is useful for exploring the scenarios put forth in the previous sections. Because I am









considering a wind-accreting HMXB, I use my previous standards where a single massive

star dominates the system, for mass between 20 100Me and radius R ~ 80Re.

The wind obscuration scenario gave an estimate of Mwid 4 x 10-slM./yr. I

can then use Equation 4-10 and find that the mass ratio of the system is q ~ 0.01.

This sl--.-, -1 a massive M > 80Me donor for a typical neutron star companion. By

relaxing the estimate of the massive star radius, ROB,, I find that q will increase and more

compact object solutions exist over a wider range of primary masses. The estimate of

ROB = 80Re is derived from the gravitational potential as estimated in the Girardi et al.

(2002) isochrones. In Table 4-6, I list a series of solutions for Equation 4-10.

Because the eclipsing scenario case places firm constraints on the mass ratio of the

system, I use Equation 4-10 to calculate the mass loss rates associated with various

scenarios. I list those values in Table 4-4. For case where the mass ratio is q ~ 0.2, the

associated mass loss rate is low Mind ~ 2 x 10-7'Me/yr, for an efficiency e ~ 0.1. This is

not unreasonable for massive stars (Mokiem et al., 2007). Interestingly, in both the wind

obscuration and the eclisping binary scenario, the X-ray luminosity is consistent with a

low mass ratio for the system.

4.4 Summary

I have searched for evidence of periodic variability in the IR spectra and long-term

X-ray light-curve of the GC X-ray source Edd-1. I find no evidence of IR variability on

short (< 3 d) timescales or between the 2005 and 2006 spectra. I compare the IR line

ratios Bry/HeI and Bry/HeII in Edd-1 to known HMXBs and CWBs and find the relative

emission line strengths more consistent with an HMXB. I have identified an apparent

189 6 d period in the Edd-1 X-ray light curve; I speculate this may be associated

with an orbital period. I find no evidence of periodic X-ray variability at timescales less

than 189 d. Using a Monte Carlo simulation, I test the significance of the 189 d period

detection; despite our fairly sparse time sampling, I find this period is significant with

a confidence level greater than 99.9',7'. I explore several interpretations of the X-ray









modulation and summarize those scenarios in Table 4-7. If the observed period is orbital

in nature, and the X-ray modulation is caused by obscuration of the X-ray source due to

a dense wind, then Edd-1 is consistent with both CWB and HMXB interpretations. The

further constraint of the X-ray luminosity is consistent with a massive (MOB > 80M.)

donor with a neutron star companion. If X-ray modulation is caused by an eclipse, the

mass ratio is low and Edd-1 is more consistent with an HMXB interpretation.

4.5 Concluding Remarks About Edd-1

Edd-1 is a reddened Galactic Center source with prominent line emission at both

X-ray and infrared wavelengths. The observed variability, spectral, and photometric data

collected to date show Edd-1 to be consistent with both CWB and HMXB interpretations.

The X-ray variability -i .-.i- -I that the mass ratio of the system is small. The distinctive

feature in Edd-1 compared to most HMXBs is the strong Fe-XXV line emission at

6.7 keV. Most HMXBs that can rival the intensity of the iron line show the emission in

Fe-Ko (6.4 keV). The higher ionization and unusual strength of the iron line observed in

Edd-1 ensures that the source environment is unusual.

The X-ray variability of Edd-1 is consistent with a CWB scenario if the variability

is interpreted as wind-obscuration of the X-ray source. Further supporting the CWB

scenario is Edd-1's similarity to the CWB Eta Carinae, though admittedly Eta Carinae is

peculiar among its class. However, Edd-1 does not have the Br7/HeI and Br7/HeII line

ratios of known WR+O CWBs, it's X-ray luminosity is brighter than that of modeled

0+0 binaries, and it's IR luminosity is significantly fainter than the LBV+O source, Eta

Carinae. Future X-ray and infrared observations of Edd-1 during the X-ray faint phase of

the source may reveal the stellar components and shed light on this highly unusual source.








































~l~1~2mr,
~s~..~ ,

~~

h
"~-
Yxyiy~iy
iXI~-MxT
i

1S1


-

uJ~
~~u,

w,
II~
,~i~
\vlc~,~in\~--

t
i"~

E~vlh~

""vi~,~
"t ~R

~z~hh.




,IIIIIIIIIIIIIIIIII IIII


2.2

Wavelength


Figure 4-1.


The K-band spectra of Edd-1. I show the original 2005 spectrum at the

bottom and the twenty-minute combinations of the 2006 spectra over the
three nights. These are offset by time of observation, such that the earliest
spectra are lower and later are higher. The relative times of these spectra are
listed in Table 4-1.


6 I.





5

I



4










2i







0
O i


2.0


2.3


2.4


I I I I


I I I I I I I I I I I I I I I II' . . I . I .. I .. I


~u~-~
-Rw,~Ji~~v~
~Y~-~-~JL~
-s,
Y7X~~~L~LL_ ,
Uh~,~ ~V-,
^~xy_____
~

"--~c,
~*"-".
-uw~
"n~- IC~h--
Vy-~/\2

"-~,.~ n~~/v~""-wwh,


u-~ V~v~hhh~





--~s,
--YI,
"I~iY~11~~












1.5


1.0-






L-
0.5








0.0





2.15
2.15


Figure 4-2.


2.1


6 2.17 2
Wavelength (microns)


.18


2.19


The Br7 region of select Edd-1 spectra taken from 2006 Aug 02-04. The region
shows apparent non-periodic variation, mostly around the 2.164/m Helium
contribution. These variations are only occasionally greater than 5-times the
RMS spectral difference. Higher resolution spectroscopy is needed to show
whether this is intrinsic to Edd-1 or an artifact of the data reduction. The
relative times of these spectra are listed in Table 4-1.




































2 4 6 8
Energy (keV)


Figure 4-3.


Two representative XMM spectra separated by 0.7 in phase. While the
strength of the Fe-XXV line is consistent between the two observations, the
continuum level drops significantly. If such variation were caused entirely
by column absorption due to a stellar wind, then NH would increase by
2.5 x 1023cm-2


100


80


60


40


20


0 a
0

























0 500 1000 1500 2000
Days since 2000 Jan 1


OF
0.0


0.5 1.0 1.5
Phase


2500


2.0


Figure 4-4. The X-ray light curve (top) and folded light curve (bottom) of Edd-1. The
light curve is folded on a 189 d period. The squares are XMM data; the
diamonds are Chandra data.






















10



8



6



4



2




0 200 400 600 800 1000 1200 1400
Period (days)


Figure 4-5. A periodogram analysis of the X-ray light curve. The preferred period is 189
6 di,- Subsequent peaks appear at integer multiples of this period.























0.8- j '/ -




o 0.6

0

0.4-




0.2



0.0
0 20 40 60 80 100
Orbitol Velocity (km/s)

Figure 4-6. Using the mass function and the putative period of 189 di,- I calculate the
expected mass ratio, q = [_!MOB, for primary masses MOB = 20 1OOMD.
The primary mass is indicated to the left of each line. The vertical dashed line
represents the limiting IR spectral resolution.


















Table 4-1. Observing Log: IR Spectra


Obs ID Date Time (UT)


2006-08-02
2006-08-02
2006-08-02
2006-08-02
2006-08-02
2006-08-03
2006-08-03
2006-08-03
2006-08-03
2006-08-03
2006-08-03
2006-08-03
2006-08-03
2006-08-04
2006-08-04
2006-08-04
2006-08-04
2006-08-04
2006-08-04
2006-08-04
2006-08-04


6:38
7:05
8:25
8:55
9:46
5:27
5:52
6:37
7:07
7:49
8:20
9:02
9:32
5:51
6:27
6:46
7:23
7:58
8:11
8:47
9:10


Exposure Time (min)

20
20
20
16
8
16
16
20
20
20
20
20
12
20
16
16
20
16
18
20
20


Note. -These observation IDs are associated with
Figures 1 and 2. The d-.- align with d-4.i 2404-2406 on
our X-ray light curves.















Table 4-2. Observing Log: Chandra


Date Time Obs. ID Exp. Time R.A. Declination Roll
(UT) (ks) (J2000) (deg)

2000-10-26 18:15:11 1561a 35.7 266.41344 -29.01281 264.7
2001-07-14 01:51:10 1561b 13.5 266.41344 -29.01281 264.7
2001-07-18 14:25:48 2284 10.6 266.40415 -28.94090 283.8
2002-05-22 22:59:15 2943 34.7 266.41991 -29.00407 75.5
2002-02-19 14:27:32 2951 12.4 266.41867 -29.00335 91.5
2002-03-23 12:25:04 2952 11.9 266.41897 -29.00343 88.2
2002-04-19 10:39:01 2953 11.7 266.41923 -29.00349 85.2
2002-05-07 09:25:07 2954 12.5 266.41938 -29.00374 82.1
2002-05-25 15:16:03 3392 165.8 266.41992 -29.00408 75.5
2002-05-28 05:34:44 3393 157.1 266.41992 -29.00407 75.5
2003-06-19 18:28:55 3549 24.8 266.42092 -29.01052 346.8
2002-05-24 11:50:13 3663 38.0 266.41993 -29.00407 75.5
2002-06-03 01:24:37 3665 89.9 266.41992 -29.00407 75.5
2004-07-05 22:33:11 4683 49.5 266.41605 -29.01238 286.2
2004-07-06 22:29:57 4684 49.5 266.41597 -29.01236 285.4
2004-08-28 12:03:59 5360 5.1 266.41477 -29.01211 271.0
2005-07-24 19:58:27 5950 48.5 266.41519 -29.01222 276.7
2005-07-27 19:08:16 5951 44.6 266.41512 -29.01219 276.0
2005-07-29 19:51:11 5952 43.1 266.41508 -29.01219 275.5
2005-07-30 19:38:31 5953 45.4 266.41506 -29.01218 275.3
2005-08-01 19:54:13 5954 18.1 266.41502 -29.01215 274.9
2005-02-27 06:26:04 6113 4.9 266.41870 -29.00353 90.6
2006-07-17 03:58:28 6363 29.8 266.41541 -29.01228 279.5
2006-04-11 05:33:20 6639 4.5 266.41890 -29.00369 86.2
2006-05-03 22:26:26 6640 5.1 266.41935 -29.00383 82.8
2006-06-01 16:07:52 6641 5.1 266.42018 -29.00440 69.7
2006-07-04 11:01:35 6642 5.1 266.41633 -29.01237 288.4
2006-07-30 14:30:26 6643 5.0 266.41510 -29.01218 275.4
2006-08-22 05:54:34 6644 5.0 266.41484 -29.01202 271.7
2006-09-25 13:50:35 6645 5.1 266.41448 -29.01195 268.3
2006-10-29 03:28:20 6646 5.1 266.41425 -29.01178 264.4











Observing Log: XMM-Newton


Observation ID


0112972101
0111350101
0111350301
0202670501
0202670601
0202670701
0202670801


Date


2001-09-04
2002-02-26
2002-10-03
2004-03-28
2004-03-30
2004-08-31
2004-09-02


Time (h) Exposure Time (h)


01:19:34
03:11:27
06:36:49
14:37:16
14:29:07
02:54:31
02:44:08


Table 4-4. Mass ratio estimations for the eclipsing scenario

rI/monoB [_/MOB MXid/M. yrr1


0.5
0.2
0.09
0.04
0.02


Note. -The mass ratio expected for
a primary of the given mass to radius
ratio in the eclipsing binary scenario.
In Column 1, the ratios are in units of
R /M.. Values of rIB/moB>104 are
more typical of brighter stars (\! Ij -7.6)
and thus consistent with cases where
a single massive star is dominating
Edd-l's infrared emission. Values of
rUB/monB MK~-4 stars such that Edd-l's infrared
emission is composed of the flux from
two bright stars. The estimation of
Mwind is based on Equation 4-10 which
is only valid for the HMXB case.


Table 4-3.












Table 4-5. Infrared Line Ratios


Source


Equivalent Width (A)


Hel
2.114/pm


Edd-1
HMXB
Cir X-1
IGR J16318-4848 (sgB[e])
HD 34921 (BOI)
HD 24534 (09111-Ve)
EX02030+375
V725Tau (09.7IIe)
0+0
HD 93205 (03V)
HD 206267 (06.5V)
HD 152248 (07Ib)
HD 57060 (07Ia)
HD 47129 (08)
HD 37043 (O9III)
HD 47129(07.5I+061)
HD 15558 (05III)
HD 199579 (06V)
O+WR
WR138 (WN5+09)
WR139 (WN5+06)
WR133 (WN4.5+09.5)
WR127 (WN4+09.5)
WR151 (WN4+08)


13.8


5
1
2.7
1.7


Br7
2.166pm


36.6

24.2
45
6
14.5
4
13

2
1.2
4
5
7
1.6
7
1.4
1.4


Ref. Br7/HeI Br7/HeII


Hell
2.189/pm


<2

1.3





<1

1.1
0.4
1.8
1.1
< 0.5
0.2
< 0.5
0.4
0.6

52
66
20
77
81


1 2.65


>18.3

18.62


9
6
5.37
2.35


2.83
1.87

2.56
2.25


> 13

1.82
3
2.22
4.55
> 14
8
> 14
3.5
2.33

0.65
0.42
0.63
0.53
0.44


Note. -Infrared line ratios. I compare the relative strength of Hel and Hell lines
to Br7 in Edd-1 and a selections of HMXBs and CWBs. Note that the Hell 2.189/m
line in Edd-1 has a P Cygni profile. I group 0+0 and O+WR binaries separately
as the former systems are less likely to produce low mass ratios. In known WR+O
systems, the Br7/HeII line ratio is significantly different than that observed in Edd-1.
REFERENCES (1) Ch, .pter 3; (2) Clark and Dolan (1999); (3) Clark et al. (2003); (4)
Figer et al. (1997); (5) Filliatre and C'!i .ir (2004); (6) Hanson et al. (1996).




















Table 4-6. Mass ratio estimations for the wind obscuration scenario in the case
of a HMXB
ROB IR R./M M OB 3[-/M.


20
60
100
20
60
100
20
60
100


0.010
0.011
0.016
0.015
0.022
0.026
0.03
0.06
0.07


Note. -The mass ratio and compact
object mass expected for a primary of the
given mass to radius ratio in the wind
obscuration scenario, valid for the HMXB
case. The estimation of q is based on
Equation 4-10. I use Lx=1.1x1035erg s-1
and assume an efficiency c=0.1, mass loss rate
MiA 4x10-5AM yr-1. The value ROB-80R.
is most consistent with our observed infrared
luminosity (Girardi et al., 2002).






















Table 4-7. Summary of scenarios
WIND OBSCURATION SCENARIO ECLIPSING BINARY SCENARIO


AM O-5Me/y
consistent


M 4 x 10-5M
consistent


M= 4 x 10-5s
q ~ 0.01 (Eq. 4
radius constrain


Two stars contributing equally to the IR luminosity (CWB)
RoB ~ 20Re
r (Eq. 4-4) q 0.05 (Eq. 4-6)
inconsistent with initial assumptions
One star dominating the IR luminosity (CWB)
RoB ~ 80Re
e/yr (Eq. 4-4) q 0.2 (Eq. 4-6)
IR line ratios inconsistent with know
WR+O s,;l-1. m,
One star dominating the IR luminosity (HMXB)
RoB ~ 80Re
Lx 1.1 x 1035erg s-1
./yr (Eq. 4-4) q 0.2 (Eq. 4-6)
-10) M 2 x 10- 'M/yr (Eq. 4-10)
t suggests MOB > 80Me consistent


Note. -See details of more general cases and caveats in Section 4.3.3.3.


'n









CHAPTER 5
AN ATLAS OF KNOWN HIGH-MASS OBJECTS IN THE GALAXY

5.1 Background and Motivation

Galactic X-ray point source radiation from binary sources is produced by high-energy

processes such as wind collision or accretion and thus does not carry information about

the donor stars contributing to the interaction. As shown in previous chapters, X-ray

data alone is often insufficient to classify a source, thus traditionally, we turn to optical

wavelengths. In X-ray Binaries (XRBs), the X-ray waveband shows the accretion region

around the compact object, and the optical waveband shows thermal blackbody emission

from the donor star with contributions from the outer edge of the disk. In Colliding Wind

Binaries (CWBs), the X-rays are generated by the wind-collision region and contain no

information about the donor stars. Optical observations of CWBs tell us the nature of

the massive stars. The combined X-ray and optical information gives us a more complete

picture of the system.

Ongoing Chandra surveys have yielded thousands of new X-ray sources and spawned

a number of campaigns designed to find and classify counterparts. In the GC, the heavy

optical extinction (Av ~ 26 mag) prevents the identification of optical counterparts, thus

we turn to infrared (IR) wavelengths. The near IR (1.1-2.5 pm) detects the same types of

processes dominating the optical waveband (i.e. thermal emission from the star). However,

it can be more difficult to spectrally classify an IR stellar spectrum due to the absence

of key identifiers (e.g., Ha) that aid in spectroscopic identification of massive stars. The

fact that the IR wavebands detect the Rayleigh-Jeans tail of blackbody emission for many

normal stars makes it difficult to establish the temperature class of of the hotter source.

However, the IR spectrum is often sufficient for distinguishing hot (B, A, F) stars and cool

(G, K, M) stars (Heras et al., 2002). Additionally, (Heras et al., 2002) show it is possible

to spectrally type low-mass, K and M stars due to absorption bands present in the IR.

The IR/optical spectral typing of the X-ray source counterpart is a vital step toward

distinguishing CWBs and HMXBs. However, the "canonical" X-ray stellar sources defining









the HMXB and CWB classes are not ahv--, extensively observed in the IR, especially if

optical information is available. Thus it becomes difficult to compare new sources, like

Edd-1 (see C'! lpters 3-4), to well-studied CWBs and XRBs. Given the number of exotic

X-ray sources expected in deep GC X-ray surveys (e.g., Chandra: Muno et al. 2006a) and

the complementary influx of IR spectroscopic counterparts expected in IR surveys (e.g.,

the Flamingos-2 Galactic Center Survey (F2GCS): Eikenberry et al. 2005), I concluded

that it would be useful to establish a complete IR/X-ray spectral atlas of archived

observations of known CWBs and HMXBs to better and more quickly characterize new

GC sources observed in these wavebands.

The goal is to create a searchable database of spectroscopic and photometric X-ray,

IR, and optical information on sources that are already classified as either CWBs or

HMXBs. This provides a starting point from which to compare unclassified, new sources

to well-studied systems that may share similar features to see if further clues as to the

nature of new sources can be discerned with less direct observational information about

them. Compiling sources into an Atlas offers considerable scientific benefits as well.

The characteristics of XRBs and CWBs differ within the individual classes, thus many

newly discovered sources are only compared to one or two well-studied sources at a time.

Additionally, much of the compiled information compares the X-ray periodicity of a source

rather than luminosity or line features. This is due, in part, to the known variable nature

of X-ray emitting binaries and the usefulness of variability in classifying a system, and in

part due to practical considerations (e.g., time allocation) involved with taken multi-epoch

spectroscopy of a source. Although this Atlas is not intended to describe the full range of

variability and behaviors of a single source, by collecting information from IR and X-ray

spectra at one or more epochs, I can examine commonalities among a wider cross-section

of sources that are not often considered side-by-side.









5.2 Types of sources included/ criteria

Sources included in the Atlas have both X-ray (between 0.2-l0keV or a subset

of that range) and IR K-band (between 2.0-2.5 pm) spectral information available

in the published literature. In general, papers which present an X-ray spectra also

provide a model-dependent X-ray luminosity value and a reddening value. Given

these minimum basic criteria, an unclassified source like Edd-1 may be considered for

inclusion in the database; however, in the Atlas I target sources that do have CWB and

HMXB classifications. Additional parameters included in the database are described

below and included in the Atlas when such information is available. Sources initially

targeted for inclusion include HMXBs, CWBs, and isolated massive stars (O-supergiants

and LBVs). The Atlas builds upon ongoing Chandra science efforts like the X-Atlas

(http://cxc.harvard.edu/XATLAS/), but is unique in that both X-ray and IR spectra

are compiled for the sources, whereas the X-Atlas includes only X-ray spectra of hot,

luminous stars (Westbrook et al., 2006). My Atlas makes use of existent, published IR

and X-ray spectra, but hopefully will inspire new observations at both wavelengths. The

preliminary interactive online version features (1) a table of common statistics on the

cataloged sources including distance, reddening, X-ray luminosity, and IR luminosity; (2)

the ability to select statistics one wants in an output table; and (3) the ability to browse

all collected information on an individual source in the catalog. The Atlas is hosted at:

http://www.astro.ufl .edu/~mikles/vjm_scrapbook/index.html (user: web-access,

pswd: nadamucho).

The number of isolated massive stars will likely remain small compared to the number

of binary sources because my main focus is massive interacting binary stars. Also, few

isolated massive stars have detectable X-ray emission. However, sources may be added to

ensure that the statistical information on isolated massive stars is representative of the

group. The list of HMXBs and CWBs will be expanded preferentially, so that we ensure

a broad baseline of statistics on these objects. Periodicity may be included in the notes,









but will not necessarily be compiled here, since the goal is to assist in the identification of

individual spectra of newly discovered sources. In addition, Liu et al. (2006) are diligently

maintaining catalogs of periodicity for XRBs.

As the number of compiled objects expands, the object classes currently grouped

together may be further sub-categorized. For example, the CWB class may be divided into

binaries containing two massive O-type stars (0+0) and binaries containing an O-type

star with a Wolf-R v- t companion (O+WR). These sources already have noticeable

spectral distinctions (see previous chapter). Also, HMXBs can be divided into compact

objects with supergiant companions (sg+X), compact objects with Be-type companions

(Be+X), Supergiant Fast X-ray Transients (SFXTs), and microquasars. The compilation

of statistics enabling us to search for common signatures for a given source type (e.g. line

ratios) can open new windows to our understanding of these sources.

5.3 Construction of the database

I used SQL to create a password-protected database and then developed a web

interface to facilitate navigation and updates. I gathered the information about each

selected source from the literature first on a standard form, then input it into the database

together, so that the record would be as complete as possible for each source upon

creation. The database is set up to include IR photometric information, X-ray spectral

fits, reddening information, and references to published spectra (detailed below). A

summary of the compiled data statistics and parameters are listed in Table 5-1. The

general hope is to have a complete source of information for each source, though likely

not uniform across the Atlas, due to the vagaries of what has been published in literature.

While it is useful to browse individual records, the purpose and power of the Atlas is

ultimately to provide general statistical information for classes of objects.









5.4 Statistics and Parameters Included in the Atlas

5.4.1 General Information: Object Name, Alternate Names, RA, Dec,
Spectral Type, Object Class

The Atlas stores a primary name for each source, as well as other catalog identifications

to aid in a name searching feature (not yet developed). Each object is then classified as

a HMXB, CWB, or Isolated Massive Star (abbreviated as OB). Additional object classes

may be added in the future. The "spectral type" field further refines the object class by

identifying the spectral type of the isolated star or the two component stars in the binary

system. For instance, if the compact object in a HMXB system is unknown, the spectral

type is listed as OB+X. If the compact object is known and is a neutron star, the spectral

type is listed as OB+NS. The letters "OB" indicate that the donor is an unclassified

massive star. The exact spectral type can be placed in that field as well.

5.4.2 Infrared: J, H, K, Alternate Magnitudes, References

Often, counterparts to X-ray sources are identified through imaging first, thus the

available IR photometry is stored in the Atlas. In some cases, the IR magnitude are

known to vary. The alternate magnitude field can contain information about the range

of observed magnitudes, or it may simply state magnitudes taken during a different

observation that differ from the initial observation. Since the Atlas is designed for

statistical studies to measure the behavior of an object class, the specific variability

information for each source not available here. The "IR references" state the literature

source for the IR magnitude and the reader would be referred to those resources for more

information on the IR luminosity behavior of the source.

5.4.3 Reddening: NH, Av, AK, E(B V), References

Source reddening is recorded in the literature in a number of different v-i-,

depending in part on the wavelength at which the reddening was calculated. To ensure

the preservation of the literature information, the Atlas has four fields in which to store

reddening information: NH, Av, AK, E(B V). These options encompass most of the









manners in which the reddening toward a source is reported in the X-ray, optical, and IR

wavebands. To be included in the Atlas, it is only necessary for one of these fields to be

filled.

The reddening value is used by the Atlas to calculate an IR luminosity, and for that

reason, there is a separate field in the reddening table called "AK (calculations)", not to

be confused with the observed, value of AK reported in the literature. This field is filled

in manually when an object is added to the Atlas. For the case where Av or E(B V)

is known, I convert optical extinction estimates to AK using a Cardelli et al. (1989)

reddening law assuming Rv 3.1. If an X-ray value of NH is known, I use the estimate of

Predehl and Schmitt (1995) to convert the X-ray values of NH to optical extinction, Av.

5.4.4 Distance

The distance is stored with the reddening data, because the distance is often assumed

in the same reference. The value is stored in kiloparsecs and is used in calculating

IR luminosity. A field for comments gives one opportunity to note alternate distance

estimates from other literature sources.

5.4.5 X-ray: log(Lx,burst), log(Lx,qui), Energy Range, Model, Thermal Temper-
ature, NH, Comments

The X-ray table contains interpreted data from an X-ray spectrum. Rather than

providing raw counts, this table uses flux-calibrated spectra that have been fit with a

model.

5.4.5.1 Quiescent and Burst X-ray Luminosity

X-ray sources tend to vary, and in the case of HMXBs, many have burst states and

quiescent states that can differ by an order of magnitude or more. This table gives one

an opportunity to record measured X-ray luminosities that are from notably different

phases/epochs of the source's light-curve. These two fields, log(Lx,burst), log(Lx,qui), are

not meant to distinguish luminosities calculated from different models.









5.4.5.2 X-ray Energy Waveband

While it would be ideal to standardize the source luminosities to a single energy

range, this is not feasible given the wide variety of X-ray instruments, survey 1- source

types, and literature sources. While I aim for a luminosity in the 2-8keV range, it is still

valuable to gather information at alternate wavebands between 0.2-lOkeV, where similar

processes are likely to dominate the X-ray spectrum, in the hope that the bolometric

luminosity for a well-modeled system would remain statistically relevant. I specify the

energy waveband range used in modeling the source and calculating the source luminosity.

5.4.5.3 X-ray Model, thermal temperature, and NH

X-ray source spectra are rarely interpreted outside the context of models. For

example, XRBs are often fit with a combination of a blackbody and a hard power law

representing the combination of thermal emission from an accretion disk and Comptonized

disk photons. As seen in the case of Edd-1, the choice of model and reddening affected

the estimated X-ray luminosity by two orders of magnitude (C'!i lpter 3). Because the

calculated luminosity is nearly alv--,v- model-dependent, I record the model used in

spectral fitting. I include a field to record the thermal temperature of the source, should

one be presented in the literature

X-rays can be subject to heavy intrinsic extinction that may not extend to a

companion star (e.g, highly absorbed HMXBs). Thus, the X-ray absorption, NH, is

recorded separately from the IR reddening data.

5.4.6 Spectral Information

By compiling spectral information from the literature and identifying sources

for which no IR spectroscopy exists, we can facilitate the creation of target lists for

observations of similar sources. It is the hope that in the future, postscript files of the

single epoch, or snapshot, X-ray and IR spectra will also be directly linked in this Atlas.

For now, the reference papers containing the available spectra are cited.









The information about the lines present in the source spectrum as well as the

observation of a wind in the source and the lines showing P Cygni profiles were designed

directly in response to identifying features similar to Edd-1. Standard literature searches

make it difficult to identify a grouping of sources that, e.g., are all lacking Hydrogen or

show strong Iron emission. In the future, the comments section of this table will allow

the inclusion of more observations of the spectrum, pointing users to sources that share

features with an unknown source.

5.4.7 Calculated Values: LJ, LH,LK

The absolute magnitude of a source is calculated by


MK = rK AK DM


where mK is the apparent magnitude, AK is the extinction in that band, and DM is the

distance modulus calculated as DM = 5 5 log d where d is in parsecs. The IR luminosity

is calculated from the reddening, distance, and IR magnitudes using the equations:


Lj = 39.522 0.4(J 2.44AK) + 2 log dkp, (5-1)


LH = 39.522 0.4(H 1.528AK) + 2 log dkp, (5-2)

and

LK = 39.522 0.4(K AK) + 2 log dkpc. (5-3)

The constants are derived from normalization to solar luminosity (Cox, 2000). This

normalization is reasonable because the accretion disk and winds are a minor component

to the overall infrared luminosity of systems containing massive stars.

5.4.8 References

The standard bibliographic code for all references is kept in a separate table and

is given an ID number by which they are identified for use in the other tables. Before a

reference can be used in the Atlas, the reference must first be added to the reference table.

In this manner, I ensure a centralized, uniform list of references.









5.5 The Web Interface

The web interface gives several options for information output using a combination

of html forms and html generating perl scripts. These scripts are linked to the database

server, providing access to the data. The menu options are (1) browse individual source

records, (2) compare statistics of objects by class, (3) view most commonly requested

statistics, and (4) add an object to the catalog or edit information about an existing

object.

The I .--- option takes the viewer to a list of all the sources in the catalog. By

following the link to a chosen object, one can review all compiled statistics for that object

as they exist in the catalog. This information is most helpful in assessing what is known

about a given source and can inspire literature searches, archival searches, or observing

proposals to fill in the gaps in the Atlas data. I currently have 22 objects in the catalog

and all of these sources are listed when the browsing option is selected. The list can

be narrowed down by selecting the object class. As the number of sources in the Atlas

increases, it will be helpful to have additional general parameters (such as position in

the sky or X-ray luminosity range) by which one can easily narrow down the objects one

is browsing through. In the future, I will add a search feature to aid in the browsing of

individual source records.

The "compare statistics" option takes the viewer to a list of statistical data stored in

the database (see Fig. 5-1). Using the HTML form, the viewer can then select a variety

of parameters to include in a table or plot. For example, by selecting X-ray and K-band

luminosity, one can quickly make a table of the requested statistics that includes each

object in the catalog. It is also possible to restrict the output table to a single class of

objects. The 1-., -- common statistics" catalog bypasses the step of selecting statistics and

produces a table that includes all of the catalog objects and contains basic, X-ray, IR, and

reddening data for the sources.









The II, 1 option is currently a blank form facilitating the collection of data prior

to inputting the source into the database. In the future, I may add an online interactive

form. The hope is to make it possible for astronomers using the database to add their own

data and information to the system.

5.6 Science Applications

5.6.1 Characterizing an Unclassified Source Spectrum

The Atlas was inspired by and designed to put the source Edd-1 into context,

and is based on the information collected in the previous two chapters to achieve said

goal (see specifically, Fig. 3-6 and Table 3-4. The Atlas is thus immediately applicable

to characterizing new source spectra in similar manners. One can easily compare the

X-ray and IR luminosity of a new high-mass candidate stellar source to others of known

classifications, or find sources that share a specific characteristic, such as Helium wind

or prominent Iron emission. The characterization of an unclassified source is a powerful

step toward understanding the nature of the underlying physical processes at work in the

system.

5.6.2 Comparing Classes of Sources

Prominent Iron emission in the X-ray is observed in several HMXBs and CWBs,

however, there is no definitive study comparing the appearance of strong Fe-XXV emission

in either of these two groups of sources. The presence of Fe-Ka (line center at 6.4keV)

in HMXBs is discussed extensively in Lewin and van der Klis (2006),who concluded that

the emission is related to the accretion process near the innermost stable circular orbit.

The Fe-XXV emission in the CWB Tr Carinae is described as arising in the wind collision

region of the source (see C'!i pter 3). However, emission as strong as that observed in

Edd-1 (or TI Car) is not common among either HMXBs or CWBs.

By simply recording the strength of the Iron emission line in each X-ray source, the

Atlas allows us to target objects with similar Iron in order to compare environments of

systems that produce prominent Iron lines and those that produce weak emission or none









at all. This may provide new insight and distinction into the environments of HMXBs and

CWBs and answer puzzles such as why Tr Carinae and Edd-1 both show stronger Fe-XXV

emission than Fe-Ka emission, or some HMXBs show strong Fe-Ka emission and others

do not. Answers to these questions may help us understand the types of environments

that produce X-ray bright massive binary sources.

5.6.3 Population Study of a Source Class

Additionally, the Atlas is easily applied to population studies. For example, in

Figure 3-6, I compared the X-ray and IR luminosity of various HMXBs, CWBs, and

massive stars that are the foundation of this Atlas. With the exception of t7-Carinae, the

CWBs observed to date subtend a relatively compact range of X-ray and IR luminosities

compared to the HMXBs. This leads one to wonder: Is this an authentic representation?

Why would that be? My hope is that the Atlas will lead to many areas of inquiry such as

this.

5.7 Summary

The compilation of multi-wavelength statistics is vital for categorizing and classifying

new XRBs and CWBs, especially in the GC. By establishing a spectral resource of

both archetypal, well-studied systems and those with less available data, we can

better and more quickly characterize new GC sources or other optically obscured

counterparts to X-ray sources. Identifying new high mass binary systems and establishing

a multi-wavelength Atlas will enhance our knowledge of the processes at work in these

sources and inspire scientific inquiry at both the observational and theoretical level.











































COMPARE OBJECT STATS

Select the object classes you wish to review:

* ( All Objects r High-Mass X-ray Binary r Colliding Wind Binary r Isolated Massive Star r Luminous Blue
Variable* C Unknown Class *



Please select the statistics your wish to review


r Object Name r Alternate Object Names
r Object Cais r Spectral Type
" Hmag r Kmag
F Reference 2 (IR) r N_H (cm^-2)
F AK r A.K (for calculations)
r Distance (kpc) F Iog(L_X) burst (erg/cm^2/s)

F X-ray Model r X-ray Thermal Temperature (keV)

r Reference 1 (X-ray) r Reference 2 (X-ray)

r Does Optical Spectrum Exist? r Does Another Spectrum Exist?


r Reference 2: X-ray Spectrum
r Does spectrum have Fel?
Equivalent Width?
r Velocity of wind
F L_H (erg/)

- MR-41 Ji tJ


r Reference: Optical Spectrum
F Does spectrum have FeXXV?
Equivalent Width?
r Reference (wind)
r L_K (erg/s)


r R.A. r Dec.
F General Comments 3 mag
r Alternate Magnitudes r Reference I (IR)
V AV r E(B-V)
r Reference 1 (A_V) r Reference 2 (AV)
F log(LX) qul F X-ray Energy Range
(erg/cm^2/s) (keV)
r X-ray N_H (cm^-2) r Comments (X-ray)

F Does IR Spectrum Exist? Does X-ray Spectrum
Exist?
SReference: IR Spectrum Reference 1: X-ray
Spectrum
F Does spectrum have r Does spectrum have
Hydrogen? Helium?
r Is wind observed? F Elements in wind

r Comments on spectrum r LU_ (erg/s)


Figure 5-1. Screen shot of database statistic selection page.









Table 5-1. Atlas Information


Statistic


Object Name
Alternate Object N ,. -;

R.A.
Dec.
Object Class

Spectral Type

General Comments


J Mag
H Mag
K Mag
Alternate

Reference
Reference


Magnitudes


(IR)
(IR)


NH(cm-2)
AV
E(B-V)


AK
AK (for


calculations)


Reference 1 (Av)
Reference 2 (Av)
Distance (kpc)


Description
Basic Information
The primary identification of the object in the Atlas
Alternate names used for cross referencing at multiple
wavelengths or between catalogs
Right Ascension (hh:mm:ss)
Declination (dd:mm:ss)
Specifies X-ray binary, Colliding Wind Binary, Massive
star, Luminous Blue Variable, or unknown
Classification of the star or binary components if
known
Variability information or other facts known
Infrared Data
Magnitude in J-band
Magnitude in H-band
Magnitude in K- or Ks-band
If the variability range is known or if contradicting
reports exist
Reference for the Infrared data
Additional reference for the Infrared data
Reddening Data
Hydrogen column density
Observed visual extinction
Alternate observed measure of the visual extinction
Observed infrared extinction
Calculated extinction value used in calculating
luminosity
Reference for the reddening data
Reference for the reddening data or the distance
Distance to the source









Table 5-1. Continued


Description


Calcuatled IR/Distance/Reddening


Lj(ergs-1)
LH(ergs-1)
LK(ergs-1)


log(Lx, burst) (ergcm-2 -1)
log(Lx, qui)(ergcm-2 S-1)

X-ray Energy range (keV)

X-ray Model
X-ray Thermal
Temperature (keV)
X-ray NH(cm- 2)
Comments (X-ray)
Reference 1 (X-ray)
Reference 2 (X-ray)

Does the IR Spectrum
Exist?
Reference: IR Spectrum
Does the X-ray Spectrum
Exist?
Reference 1: X-ray
Spectrum
Reference 2: X-ray
Spectrum
Does the Optical Spectrum
Exist?
Reference : Optical
Spectrum
Does the Another Spectrum
Exist?

Does the spectrum have
Hydrogen?
Does the spectrum have
Helium?
Does the spectrum have
Fel? (equivalent width?)


The calculated luminosity value in the J-band
The calculated luminosity value in the H-band
The calculated luminosity value in the K- or Ks-band
X-ray Data
The calculated X-ray luminosity in a high-flux state
The calculated X-ray luminosity in a low-flux state (or
the known luminosity of only one observation exists)
The energy range of the X-ray spectrum that was fit for
the above modeling
The model used to estimate the luminosity
If a thermal blackbody was used, the temperature of
the model
The absorption value used in the model
Comments relevant to X-ray modeling or variability
Reference for the X-ray data
Additional reference for the X-ray data
Spectral Info
(yes or no)

Reference for the IR spectrum should it exist
(yes or no)

Reference for the X-ray spectrum should it exist

Alternated reference for the IR spectrum should it exist

(yes or no)

Reference for the optical spectrum should it exist

(yes or no)


Spectral Line Info
yes or no; relevant to any waveband

yes or no; relevant to any waveband

yes or no; relevant to X-ray waveband


Statistic









Table 5-1. Continued


Description


Does the spectrum have
Fe-XXV? (equivalent
width?)

Is wind observed?
Elements in wind
Velocity of wind
Reference (wind)

Comments on spectra


yes or no; relevant to X-ray waveband



Wind/ P Cygni profile
yes or no; relevant to any waveband
Primarily expect HI or Hel
The P Cygni velocity measured in km/s
Reference for the wind elements and velocity should it
exist
General comments related to the spectra available on
the source


Statistic









CHAPTER 6
THE MICROQUASAR GRS 1915+105

6.1 Introduction

GRS 1915+105 is a microquasar located 11.2 0.8 kpc away on the opposite side of

the Galactic Plane (Fender et al., 1999). A microquasar is an X-ray binary that exhibits

both disk accretion and relativistic jet ejections. GRS 1915+105 appeared in the X-ray

sky in 1992 as an X-ray transient (Castro-Tirado et al., 1992), but unlike any X-ray

transient known before it or since, it has completely failed to disappear. Its longevity

has allowed extended study of the source and jet production events. GRS 1915+105 is

classified as a Low-Mass X-ray Binary (LMXB) with a black hole counterpart. The mass

of the compact object is estimated at 14 Me and the companion is K-M III star (Greiner

et al., 2001). The low-mass companion was identified due to absorption bands present in

the infrared spectrum. LMXBs such as GRS 1915+105 accrete primarily via Roche Lobe

overflow.

Radio imaging of GRS 1915+105 shown in Figure 1-4, shows that jet ejecta

associated with the source move away from the central object at 0.98+ 2jc (\! ihel and

Rodriguez, 1994; Fender et al., 1999). Continued radio, infrared, and X-ray observations

have led to the classification of GRS 1915+105's radio jets into three types: (1) -i Ily"

radio jets, (2) discrete plasma ejection events of 20-40 minute duration in the infrared and

radio, and (3) large superluminal jets akin to the 1994 event that earned GRS 1915+105

the name "microquasar." The steady jet is associated with a radio and X-ray "plateau"

state, meaning there is relatively little variation in the light curves in these regimes. The

steady jets are optically thick, extend for < 200AU from the central source, and have a

velocity of 0.1 0.4c (Yadav, 2006). The discrete jets are associated with radio, infrared,

and X-ray oscillatory behavior. The discrete jets are compact (extending a few hundred

AU), but has a velocity closer to 9,' of the speed of light. They range in strength form

5-200mJy in the infrared and radio (e.g., Eikenberry et al., 1998; Rothstein et al., 2005;









Yadav, 2006). The superluminal radio jets can be up to 1Jy in the radio and extend

several thousand AU (M\ ihel and Rodriguez, 1994; Fender et al., 1999).

Only a handful of X-ray binaries are classified as microquasars displaying jet activity,

though Fender et al. (2005) have -ii--.- -i l that all XRBs will emit superluminall" jets at

some point during their evolution. The presence of a jet can be inferred by observations

in the hard X-ray, IR, or radio regimes. While jet ejections take decades to observe in

extragalactic quasars, microquasars show similar events in a matter of hours or d -,

Thus, microquasars are an excellent laboratory through which we can study accretion

physics, relativistic jet formation, and disk-jet interactions. Because disk-jet interactions

occur close to the central compact object, understanding the interaction will help us probe

the limits of General Relativity and answer the question of what happens at the edge of a

black hole.

6.2 X-ray Light Curves and Quasiperiodic Oscillations

What causes a jet to form? To answer this, I focus on the inner accretion disk,

which dominates the X-ray waveband emission due to the heat of the in-falling material.

The accretion disk of an XRB radiates through both thermal and non-thermal processes.

The thermal radiation reaches high temperatures due to a combination of the scale and

mass of the system. The non-thermal processes are not well constrained, but are generally

associated with synchrotron emission and Comptonization of disk photons. The X-ray

light curve of GRS 1915+105 has many patterns of variability, but only a few have been

positively linked with jet ejection events.

Belloni et al. (2000) define a total of twelve X-ray light curve classes for GRS 1915+105,

distinguishable by general appearance, count rate, and X-ray color (related to hardness).

In Figures 6-1, 6-2, and 6-3, I show the 1-second X-ray light curves of three classes:

0, a, and 3. I focus on these three classes of events because simultaneous X-ray and

infrared analysis reveal the presence of discrete oscillating jets recurring on ~ 30 minute









timescales, manifesting in the infrared as 5 200 mJy flares (Eikenberry et al., 2000;

Rothstein et al., 2005).

The 0-class X-ray light curve is characterized by M-shaped intervals lasting a few

hundred seconds, and spectrally harder dips lasting 100-200 seconds (Fig 6-1). Both the a-

and p-class light curves have extended spectrally hard dips, but the length of the a-class

hard dip is nearly twice the length of the 3-class one. The a-class light curves have a dip

of ~1200 seconds followed by strong X-ray oscillations, marking the end of the hard state

(see Fig. 6-2). The 3-class light curves are more complex. The dip lasts 500 700 seconds

and its ending is marked by a spectrally soft spike followed by a nearly monotonic rise.

After reaching a peak flux, the X-rays begin rapid, large-amplitude oscillations (see Fig.

6-3).

It is well established that the spectrally-hard dips that frequently appear in the X-ray

light curves of GRS 1915+105 are associated with infrared and radio flares (Eikenberry

et al., 1998; Mirabel et al., 1998; Fender and Pooley, 1998; Klein-Wolt et al., 2002). These

spectrally-hard dips are also associated with 2 10 Hz variable quasi-periodic oscillations

(QPOs). QPOs are a form of rapid, periodic variability observed in both black hole and

neutron star binaries. QPOs are not related to periodic variability such as pulsations,

dips, or eclipses (Lewin et al., 1997). It is believed that QPOs probe the accretion flow

dynamics at the innermost part of the accretion disk. I seek to understand how and why

the QPO is related to the jet ejection.

A QPO can not ahv--, be identified within an X-ray light curve with the naked ec-

To resolve the oscillations, exceptional timing resolution is required; such resolution is

available through the Rossi X-,in; Timing Explorer (RXTE) Proportional Counter Array

(PCA). With 8-millisecond resolution light curves, I am able to observe the evolution of

the power spectrum and track the frequency of an oscillation as it changes with time.

The power spectrum of each sampled interval is comprised of a number of variability

components. Broad features are called noise and narrow features are called QPOs. For









my analysis, I divided the light curve into 4-second intervals and took a Fast Fourier

Transform (FFT) of each interval to obtain a series of power spectra. In Figure 6-4, I

show a portion of the 8-millisecond light curve of GRS 1915+105 during a spectrally hard

dip and the FFT of two regions each lasting 4 seconds and separated by 4 seconds. The

QPO is clearly visible in each power spectrum and the frequency of the oscillation shifts

noticeably. Thus, it is clear that estimating the QPO frequency over longer time samplings

(>8 s) can blur the oscillation frequency and make the QPO more difficult to identify and

model. By applying Fourier analysis to the entire length of the hard dip, I can study QPO

behavior without making assumptions about the underlying process.

From several RXTE observations of GRS 1915+105 taken on 14 15 August 1997,

9 September 1997, and 27 28 July 2002 (see Table 6-1), I extract Proportional Counter

Array (PCA) Standard-1 light curves using FTOOLS v5.3 and identify regions showing

a hard X-ray dip. I examine 17 regions here: nine of them are p-class, three are a-class,

and five are 0-class (Belloni et al., 2000). They are numbered sequentially 0-1 through

3-9, a-10 through a-12, and 0-13 through 0-17. For the seventeen regions, I extract

binned mode 8-millisecond light curves in the 2 13 keV range and 4-second resolution

binned and event X-ray spectra in the 2 25 keV range. Using XSPEC v11.3.0, I fit the

spectra with a combination of absorbed multi-temperature disk blackbody and power law

models. This is a common combination of models for describing an X-ray binary system.

The multi-temperature blackbody models the thermal emission from a geometrically

thin accretion disk around a compact object, recognizing that the temperature of the

disk increases at low radii close to the compact object. This model is characterized by

the X-ray flux, and the temperature and radius at the inner part of the accretion disk.

The power law component dominates the hard (>4keV) X-rays, and is attributed to

synchrotron emission or other non-thermal processes. In our analyses, we only use fits

which have a X < 2.









By using a FFT, I calculate the power density spectrum (PDS) from the binned

mode 8-millisecond light curve. I fine-bin the PDS using Fourier interpolation and track

the peak frequency at 4-second resolution (see e.g. Ransom et al., 2002). Fine inning

allows me to calculate a higher-resolution Fourier response by interpolating responses at

non-integer frequencies. In Figures 6-5 and 6-6, I show a gi li--- I. .1 fine-binned PDS

and over plot a one-second resolution light curve. The PDS is gi li--- I. d according to

the relative power of the spectrum such that darker regions indicate power peaks. The

QPO peak frequency appears to evolve smoothly during the progression of the hard dip,

developing a U-shape in the /3-class light curves while tracing the total flux in the a- and

0- classes.

I determine the QPO frequency by fitting a Moffat function to the PDS in the

2 10 Hz frequency range. The Moffat function is a Lorentzian modified with a variable

power law index (Moffat, 1969). The QPO is considered to be detected if it has a quality

factor Q = v/FWHM > 2, where v is the centroid frequency of the Lorentzian and

FWHM is the full-width at half max. This method yields a consistent and repeatable

QPO detection. However, low-frequency noise in the power density spectrum may still

lead to false detections. In the next chapter, I discuss the fine-tuning required to select the

coherent QPO in each light curve class.

6.3 Infrared Flaring Behavior

In all three classes that we study, simultaneous X-ray and infrared observations show

infrared flares rising as the hard X-ray dip ends and the QPO vanishes (Eikenberry et al.,

1998; Mirabel et al., 1998; Rothstein et al., 2005). Eikenberry et al. (1998) observed

a one-to-one correspondence of X-ray dips to infrared flares at 30-minute intervals. I

show a sample X-ray light curve with simultaneous infrared data in Figure 6-7 with the

QPO frequency over-plotted. The X-ray observations were carried out using the RXTE

Proportional Counter Array and simultaneous infrared data was obtained at the Palomar

200-inch telescope. In the figure, the solid line is the X-ray flux. For the first 500 seconds









of this observation, large amplitude X-ray oscillations occur and soft X-ray flux from the

accretion disk dominates the emission. At t ~ 500s, the soft flux diminishes and the

X-rays enter into a spectrally-hard dip. The QPO frequency is shown as plus signs. At

t ~ 1200s, there is a spectrally soft X-ray spike which il i -.-- i an infrared flare (black

circles). At this point, the coherent QPO vanishes. This pattern repeats about every

thirty minutes.

IR flares associated with my a- and /3-class X-ray observations are defined by

Eikenberry et al. (2000) as class C and class B, with the flux of the flares ranging from

30 200 mJy when dereddened by 3.3 magnitudes. Individual class C infrared flares are

smaller (5 10 mJy) and have been observed to be individually associated with the 0-class

(Eikenberry et al., 2000). I discuss the distinct flaring behaviors of these classes more fully

in the next chapter.

6.4 Summary

My goal in studying GRS 1915+105 is to analyze the X-ray and infrared phenomenology

and compare these observations to predictions made by jet ejection models. In this

chapter, I have introduced the X-ray behavior of three light curve classes (a, 3, and 0)

that have been positively linked to jet ejections. I focus on these three classes, because

simultaneous X-ray and infrared information is available, allowing me to study the

behavior of discrete ejection events that range in intensity from 5-200 mJy in the infrared.

In the next chapter, I will discuss the X-ray and infrared phenomenology associated

with ejection events in the a- and /-classes. In C'! plter 8, I expand the phenomenology

analysis and compare my observations in all three classes to predictions made by a

QPO/jet ejection model.


























10


0




Figure 6-1.


1.0 1.2 1.4 1.6 1.8
Time in hours since 1997Sep05 05:02:07 UTC


A typical 0-class light curve at 1 s resolution. The curve has a characteristic
M-shape and becomes spectrally hard at low flux. The observation ID is
derived from the naming scheme in Table 6-1.












































0.2 0.4 0.6 0.8
Time in hours since 2002Jul28 03:42:06 UTC


Figure 6-2.


A typical a-class light curve at 1 s resolution. The a-class curve has notably
longer hard dips and shorter soft flaring times. The observation ID is derived
from the naming scheme in Table 6-1.


10-


5


00
0.0












































5.0 5.2 5.4 5.6
Time in hours since 1997Aug14 04:02:37 UTC


Figure 6-3.


A typical /-class light curve at 1 s resolution. The hard dip ends with a soft
spike that leads into rapid oscillations. The observation ID is derived from the
naming scheme in Table 6-1.


50


40

u
3 30


20


10


0.
4.8





























01 . .
1150 1155
Time in seconds


1160 1165 1170
since 1997Aug14 04:02:37 UTC


4 6 8
Frequency (Hz)


Figure 6-4.


(top) An 8 ms light curve of GRS 1915+105 during a spectrally hard dip.
(bottom) The power spectrum of select 4 s regions of the light curve. The
QPO frequency can change noticably over short time scales.










































200 400 600 200 400 600 200 400 600


0 500 1000
Time


0 500 1000
Time


0 500 1000
Time


Figure 6-5. A gi li--, .'1 fine-binned power density spectrum for the 3- and a-class light
curves. I plot the 1-second resolution light curve over the PDS.


8U
60
Q)
4(
2-
Su


80
60
2)
4
2-


83





Q.
6>
V)
4(
2t






Y
o

(0-S























500
S Time
8o
80
6>
6)

i,
4<

2
-Y


0 500
Time


1000 500
Time


Figure 6-6. A gi i,--l I, .1 fine-binned power density spectrum for the 0-class light curves.
I plot the 1-second resolution light curve over the PDS.


500 1000 1500 2000
Relative Time (seconds)


20

15 u


10
L,
C5
5 n
o


010
2500


Figure 6-7.


A sample light curve showing the X-ray, the infrared, and the QPO behavior
during a jet ejection. The solid line is the X-ray light curve. The dots are the
infrared, and the crosses are the QPO frequency. The QPO evolves while the
X-rays are in a spectrally hard dip. After the spike in the X-ray, the QPO
drops off and an infrared flare begins. This is interpreted as a plasma ejection.


8

c 6

I4


c 6
0 6

I4
L-


80
0-
6>
U
4\

2U
, ^


1000


1000




















Table 6-1. RXTE Observations
Set ID RXTE DATA ID Date Observed Start Time

3-1 20186-03-03-01 1997 Aug 14 04:20:52
3-2 20186-03-03-01 1997 Aug 14 05:50:52
3-3 20186-03-03-01 1997 Aug 14 07:17:12
3-4 20186-03-03-01 1997 Aug 14 09:18:56
3-5 20186-03-03-02 1997 Aug 15 07:33:36
3-6 20402-01-45-03 1997 Sep 09 06:15:32
3-7 20402-01-45-03 1997 Sep 09 08:04:28
3-8 20402-01-45-03 1997 Sep 09 09:21:28
3-9 20402-01-45-03 1997 Sep 09 09:50:36
a-10 50125-01-04-00 2002 Jul 27 07:30:50
a-11 50125-01-04-00 2002 Jul 27 10:24:48
a-12 50125-01-05-00 2002 Jul 28 06:57:37
0-13 30182-01-03-00 1998 Jul 10 05:05:57
0-14 30182-01-03-00 1998 Jul 10 05:46:27
0-15 30182-01-03-00 1998 Jul 10 08:55:27
0-16 30182-01-03-00 1998 Jul 10 05:15:15
0-17 20402-01-45-02 1997 Sep 05 05:02:37



Note. -Observation IDs and dates of the 17 epochs. The Start
Time indicates where spectral fitting began. The IDs are based on
the Belloni et al. (2000) classifications.









CHAPTER 7
DOES LOW FREQUENCY X-RAY QPO BEHAVIOR IN GRS 1915+105 INFLUENCE
SUBSEQUENT X-RAY AND INFRARED EVOLUTION?3

7.1 Introduction

In this chapter, I focus on the X-ray and infrared behavior of the a- and 3-classes

of X-ray flares in GRS 1915+105. The X-ray behavior of these classes is described in

the previous chapter. In both the a- and 3-classes, the 2-25 keV X-ray light curve enters

a spectrally hard dip that lasts for several hundred seconds, and each class shows

an infrared flare rising as the hard X-ray dip ends and soft X-ray oscillations begin

(Eikenberry et al., 1998; Mirabel et al., 1998; Rothstein et al., 2005). While the X-rays are

rapidly ,-, 11 ill i.- the infrared flare peaks and decays without the rapid, large amplitude

variations seen in the X-ray. The infrared flares associated with these classes are defined

by Eikenberry et al. (2000) as class B and class C. When several class C infrared flares

occur in rapid succession, they can appear as a larger flare. Rothstein et al. (2005) show

this to be the case for the a-class light curves in my sample, where the associated infrared

flares range from 10 30 mJy. Rothstein et al. (2005) showed that if each soft X-ray flare

were associated with a 5 10 mJy (class C) sub-flare, then the duration and strength of

the overall infrared flare would be explained. By that analysis, the predicted collection

of infrared sub-flares associated with a 3-class light curve would contribute only a small

fraction of the overall observed 60-200 mJy infrared flux from GRS 1915+105. This small

contribution was observed by Eikenberry et al. (2000) as an i i I1. i1 excess." Although

the period of the rise and decay of the primary class B infrared flare is not coupled to

the period of the X-ray oscillations, the overall duration of the infrared excess is coupled

(Eikenberry et al., 2000; Rothstein et al., 2005). The difference between the dip/flare

cycles associated with a- vs. 3-class curves becomes the presence of a large primary



3 This work was prepared in collaboration with S. S. Eikenberry, and D. M. Rothstein.
It is published in the A-1 i. .l iv-ical Journal 2006, 637:978.









infrared flare which is uniquely associated with 3-class X-ray light curves. The presence

of the primary flare is widely associated with the presence of the X-ray 112 2-- i spike"

(see Fig. 6-7). It is believed that both are linked to the underlying cause of larger plasma

ejections.

Several authors have shown that infrared flares tend to be followed by radio flares

of similar morphology (Fender and Pooley, 1998; Mirabel et al., 1998). The sequence

of a spectrally-hard dip, an infrared flare, and then a radio flare, is generally associated

with a plasma ejection from the source. The ejection is observed as the X-rays transition

from a spectrally-hard to a spectrally-soft state. During the hard dip that precedes an

ejection event, a 2 10 Hz QPO is alv--, observed (Belloni et al., 2000). While the

QPO is present, the disk X-ray emission is greatly reduced in the 2 25 keV range and

the X-ray luminosity is dominated by the harder power law flux. Several groups have

found the QPO peak frequency positively correlated to both power law and thermal disk

components, ii.:.: -lii-; that the QPO arises in the same location as the related emission

or is causally related to it (\iniii, et al., 1999; Feroci et al., 1999). Markwardt et al. (1999)

observed that at frequencies above 4 Hz, the QPO is most strongly correlated to the

thermal disk component, specifically the blackbody disk flux. At lower frequencies, the

QPO shows a broad association with the power law flux. Muno et al. (1999) also show the

QPO is strongly associated with disk temperature during the hard dip. After the tri .-.-r

spike, the power density spectrum becomes smooth (\ i I.:wardt et al., 1999).

In this chapter, I examine the behavior of the QPO during the hard dip and the

relationship of its behavior to subsequent infrared flares. In Section 7.2, I discuss my

observations and ,i" &1 i-;- technique. I calculate and discuss the correlation of the QPO

peak frequency to spectral features in Section 7.2.1 and comment on the time evolution of

the QPO during the dip in Section 7.2.2. In Section 7.2.3, I examine the morphology of

the trigger spike in the 3-class light curve and the indication of a continuum of behaviors

from this class to the a-class, which does not show a tri. -. r spike. In Section 7.2.4, I









compare infrared flaring behavior between the events. I discuss the significance of these

results in Section 7.3 and in Section 7.4 I summarize my conclusions.

7.2 Refining QPO Detection

I use several a- and 3-class RXTE observations of GRS 1915+105 taken on 14

- 15 August 1997, 9 September 1997, and 27 28 July 2002 (see Table 6-1) and use

the reduction procedure described in the previous chapter. I refine my QPO detection

method to account for low frequency noise in the observations. In a-class curves, I require

the QPO frequency to be between 2.6 and 10 Hz to avoid contamination by low frequency

noise. In 3-class curves, I allow detections in the full 2-10 Hz range. However, also visible

in the PDS of the 3-class is a low-frequency noise component, the maximum frequency

of which tends to rise above 2 Hz at the beginning and end of the dip. To avoid spurious

detections from this component, I require that each detected QPO does not vary too

sharply from the QPOs at surrounding times. I make the following "by-cyj assessments,

keeping my refinements as general as possible for the sake of repeatability. For points at

times t > 350 seconds, I take an average of the prior 10 detected frequencies, Vag,. The

next detection is required to be greater than 7 .'. of a,,g. For t < 150 seconds, QPOs

are treated similarly, though Vag, is determined by using points following the reference

point, as opposed to those proceeding it. Although omitting points as non-detections can

leave part of the QPO evolution under-sampled, I believe I have sufficient representation

from the regions to ensure a sound qualitative result and a reasonable quantitative one. In

addition this method yields a consistent and repeatable QPO detection. It is possible that

adjusting the range in this manner causes me to omit real, unresolved oscillations in the

QPO, but visual inspection shows that these points are not part of the primary U-shaped

QPO feature I wish to focus on. This shape is apparent in the PDS shown in Figure 6-5.

In Figure 7-1, I plot the detected QPOs as filled circles against a 1-second resolution light

curve. For illustrative purposes, I show power peaks in the PDS that do not meet the

detection criteria as open circles.









7.2.1 QPO Frequency Correlation to Spectral Features

For each hard dip, I calculate the Linear Pearson Correlation Coefficient, r, between

the QPO frequency and various spectral features. I define events with Irl > 0.70 as highly

correlated. This value is chosen because it means that at least half the variance of the

spectral feature can be accounted for by the variance of the QPO frequency. Values of Irl

between 0.4 and 0.7 I discuss as believable, but generally disregard when more statistically

significant correlations are present.

I list the correlation coefficients for the twelve dips in Table 7-1. In all of the cases,

I observe a highly significant correlation between the QPO frequency and the total flux,

which is generally stronger than those to individual blackbody or power law features.

When considering model-specific spectral features, eleven out of twelve show strong

correlations to power law flux, and only six out of those eleven show strong correlations to

blackbody features. In nine out of those eleven cases, the correlation to the power law flux

is stronger than the correlation to any blackbody feature. This gives the apparent result

that the QPO frequency is more fundamentally tied to the power law component than the

blackbody component.

Figure 7-2 shows a scatter plot of the QPO frequency versus the power law flux for

the twelve cases. The points marked with triangles are detections within the first 100

seconds of the dip and the squares are from the last 100 seconds of the dip. The line is

based on a fourth-order fit to the time evolution of the QPO frequency and the power

law flux and traces the approximate path of the evolution. The open circles are power

peaks between 2 and 10 Hz, where a QPO was not detected (see above). These points are

generally not a part of the observed trends.

In /3-1 through 3-5 (August 1997) the evolution of the QPO-power law flux relation is

generally tighter and the correlations stronger (r > 0.80). The a-10 through a-12 curves

also show a tight correlation, though the QPO frequency varies over a smaller range of

frequencies. Like the a-class curves, I only have detections for 3-8 over a small range of









frequencies, during which I see a tight correlation to the power law flux. In the other three

cases, /3-6, /3-7, and /3-9, I notice a hysteresis effect, observed as a separation between the

entrance (triangles) and exit (squares) points of the dip. In the 3-7 case shown in 7-2, the

hysteresis is extreme and lowers the correlation coefficient drastically despite the initially

linear correlation at the dip entrance. The squares which trace the last 100 seconds of the

dip show the QPO frequency rising sharply while the power law flux is relatively constant,

-,- .. -I ii-; a decoupling of the features.

In /3-1 through 3-7, the correlation between the QPO frequency and the power law

flux is accompanied by a similarly strong correlation to the blackbody flux and blackbody

disk temperature. In Figure 7-3, I see a strong correlation to the blackbody flux is most

prominent above 4 Hz, a point also observed by Markwardt et al. (1999). However,

the correlation does not last over the broad part of the hard dip where the QPO can

change significantly while at relatively constant blackbody flux. In /-8 and /3-9, where the

blackbody flux appears uncorrelated, there are fewer QPO detections above 4 Hz due to

the rise in low frequency noise at the dip exit.

I show the QPO frequency blackbody temperature scatter plots in Figure 7-4.

Clearly the QPO in the a-class curves is not believably correlated to the blackbody

temperature. In /3-1 through 3-5 and /3-8, the correlations to the blackbody temperature

follow a steady trend, but suffer a slightly wider dispersion in their evolution than the

power law flux. In /3-6, 3-7, and /3-9, I see a similar hysteresis to that observed in the

power law flux. To untangle the possible interplay of the blackbody temperature and

power law flux, I apply a partial correlation analysis. The partial correlation coefficients

are listed in Table 7-2. In the table, I calculate coefficients for four scenarios:

1. The QPO frequency -power law flux correlation removing the effect of blackbody

flux.

2. The QPO frequency -power law flux correlation removing the effect of blackbody

temperature.









3. The QPO frequency -blackbody flux correlation removing the effect of power law

flux.

4. The QPO frequency -blackbody temperature correlation removing the effect of

power law flux.

In the first scenario, I see that with the blackbody flux removed, the correlation to

power law flux is still strong. This is expected because the blackbody flux poorly explains

variances at frequencies below 4 Hz (see Fig. 7-3 and Table 7-1). In the second case, I find

the removal of blackbody temperature has a more significant effect. In most cases, the

QPO frequency -power law flux correlation is weak, though believable (r > 0.4). This

means that after removing variations in the QPO frequency and power law flux that can

be explained by variations in blackbody temperature, variations in the QPO frequency

remain that can be at least partially explained by variations in the power law flux. High

significance is seen in a-class cases, which is expected due to their weaker dependence

on blackbody temperature. In the third case, I test the correlation to the blackbody flux

after removing the power law flux. These correlations are believable and occasionally

-1r 'i -, :. iir-::; that a combination of the blackbody and power law sources is required

to explain the QPO frequency. In the final case, the removal of power law flux from

the QPO frequency -blackbody temperature correlation, I see that partial correlation

coefficient drops below significance in half of the cases. This means that once variations in

the QPO frequency that can be explained by variations in the power law flux are removed,

no variation remains that can be explained by blackbody temperature.

In summary, the QPO frequency is often correlated to the power law flux. For

3-class light curves, when this correlation is strongest, I tend to find a correlation to the

blackbody flux and blackbody temperature as well. In cases where there is a slightly

weaker correlation to the power law flux, the correlation to the blackbody features is

less predictable and a hysteresis effect is visible in the power law flux and blackbody

temperature relations. In contrast, a-class QPOs tend to have a strong correlation to the









power law flux and weak or non-existent correlations to the blackbody features. Based on

this evidence, I believe the correlation between the QPO frequency and the power law flux

is more fundamental.

7.2.2 QPO Time Behavior

Noting the hysteresis in the QPO frequency versus power law flux distribution in

several of the 3-class light curves (see Fig. 7-2), it is likely that the spectral evolution is

intimately tied to the QPO time evolution. The QPO evolution at the beginning of all the

hard dips is similar, each beginning with the sudden appearance of a QPO at 6 10 Hz.

This QPO will smoothly drop to between 2 3 Hz within seconds of the initiation of the

hard dip. After the X-ray trigger spike, the primary U-shaped QPO feature will disappear,

replaced by occasional power peaks. In Table 7-3 I list the minimum QPO frequency

during the dip and the time spent near that frequency. Based on the observable variations,

I then divide the light curves into three broad groups:

Group 1 Figure 7-5 shows an example of the 3-class light curves in Group 1. A

series of X-ray oscillations calm into a low, spectrally-hard dip within about 100 seconds.

During this time, a QPO arises at ~ 6 8 Hz. Following the intensity drop of the light

curve and more specifically the power law flux, the QPO falls steadily to ~ 2 Hz. It

remains at this frequency for over 150 seconds (see Table 7-3), after which the power law

flux and QPO frequency begin a slow rise. The U-shaped QPO vanishes after the X-ray

trigger spike. The total length of the dip is on the order of 600 seconds, and ~ :nl of

that time is spent at the minimum frequency.

Group 2 -This group is also composed of p-class light curves with similar spectral

behavior to Group 1 (see Fig. 7-6). However, the QPO behavior in this group is somewhat

different. While in Group 1, the QPO falls off to ~ 2 Hz and lingers, in Group 2 the

QPO immediately starts to rise again. In Table 7-3, I show that while the Group 1 events

remain near the minimum frequency for > 150 seconds, the Group 2 events remain for

< 100 seconds (about 15'. of the dip length). The following rise in frequency is the start









of the hysteresis in the QPO frequency -power law flux scatter plot (Fig. 7-2) and likely

indicates that the QPO and power law flux have decoupled. In addition, these cases see

the rise of a low frequency noise component above 2 Hz as the QPO weakens in amplitude

and rises rapidly in frequency. While these low frequency points are excluded from

correlation analysis as being associated with low frequency noise, it is possible that they

represent an increase in rapid, unresolved QPO oscillations. The difference in behaviors

in the first two groups is most remarkable because their X-ray light curves and spectral

behaviors are so similar.

Group 3 The final group contains the a-class light curves represented in Figure

7-7. In these events, the hard dip is surrounded by X-ray oscillations but no independent

terminal spike is observed. The total length of the dip is ~ 1200 seconds and the QPO

disappears when the dip ends. This case is similar in shape to Group 1 QPO evolution,

though twice as long. The a-class differs in that on entering the dip the QPO frequency

levels off at ~ 3 Hz. Overall, the frequency varies over a smaller range than that seen in

the other two groups.

7.2.3 Differing Trigger Spike Morphology in p-class Light Curves

The separation of the 3-class curves into two groups of QPO evolution inspires the

search for other differences in the light curve behavior. I focus here on the trigger spike

which appears at the end of the hard dip and coincides with the start of the infrared

flare and note that the intensity and morphology of the spikes vary between Group 1 and

Group 2 events. Figure 7-8 shows a one-second resolution light curve for each of the nine

observed spikes in a time range of 55 seconds before and after the maximum.

The light curves are fit with a double Gaussian plus a polynomial of the form:


f Aoe-z + Boe-A + Co + CIt

where ZA (t TA)/IA and ZB = (t TB)/aB. is the mean (center) Gaussian value and

a is the standard deviation.









The second Gaussian term fits a small rise proceeding the X-ray spike -a feature

more distinct in the 0-3 through 3-7 light curves. This second Gaussian is treated with

the polynomial as part of the background emission. A normalized peak amplitude for

the spike is calculated as A = Ao/Abg where Abg is the polynomial and second Gaussian

evaluated at the primary Gaussian peak time, t = TA. It should be noted that the

apparent symmetry of the spike is tied to the time resolution. The one-second time

resolution allows a fair sampling of data points for the range while smoothing the rapid

X-ray variability seen in an eight-millisecond curve. Broader time intervals will leave the

spike under-sampled and reduce the accuracy of the fit.

The relevant numerical fits are listed in Table 7-4. It is interesting to note that the

integrated count rate (fit) over the full-width-half-max of the peak is similar for all data

sets, but in most other aspects of the fit, the August (Group 1) and September (Group 2)

data have different properties.

1. The Group 1 data have a higher normalized amplitude: A0'P > AGrp2

2. The Group 1 data are more symmetric while the Group 2 data shows a sharp cut-off

after the peak flux (see Fig. 7-8).

3. The Group 1 data spikes are narrower: ca pl < aGrp

4. The underlying slope of the light curve is positive in the Group 1 data and flat or

negative in the Group 2 data: C'"p > 0 while CGrp2 < 0.

While all Group 2 events show a negative underlying light curve slope, it is interesting

that two of them (3-6 and 3-7) have a nearly flat slope while the other two (3-8 and 3-9)

have a decidedly negative slope. From Figure 7-2 I note that the two with nearly flat

slopes have a slightly more visible hysteresis because they have more QPO detections

above 4 Hz. The low frequency noise component does not rise as strongly above 2 Hz and

the QPO frequency shows a stronger correlation to blackbody temperature. In addition,

the terminating spikes of 3-8 and 3-9 look much more disturbed than those of other light

curves (Fig. 7-8). The differences in trigger spike morphology -i-i-. -1 that the variation









in QPO behavior is not an artifact of the detection method. The /-6 event is particularly

interesting because it can be considered a crossover between Group 1 and Group 2 events.

The QPO frequency -power law flux relation in /-6 shows a significant correlation,

r = 0.78, despite the apparent hysteresis. From Figure 7-2, I see that this event seems

to be a bridge between the sharp linear correlations of Group 1 and the divergent shapes

of Group 2. I classify the /-6 light curve as Group 2 because the QPO frequency clearly

deviates from the initial regression and because of its trigger spike morphology. The

trigger spike of j3-6 is wider and more .-i-iiiii,. ii than Group 1 events. In addition, the

flat underlying light curve of /-6 si,--l: that it is more appropriately associated with

Group 2 than Group 1 events. This being said, I acknowledge that the groups are not

absolute, but are likely part of a continuum of behaviors.

7.2.4 Associated Infrared Flaring

Simultaneous infrared coverage is available for four out of five of my Group 1 data

sets. Eikenberry et al. (1998) show that these dip/spike pairs are usually followed by large

infrared flares. The events observed ranged from ~ 60 to 200 mJy. Although the rise and

fall of the flare does not correspond to the period of X-ray oscillation, a weak infrared

excess which lasts throughout the period of the X-ray oscillations is observed. Eikenberry

et al. (2000) and Rothstein et al. (2005) explain this excess as the superposition of

many faint infrared flares each on the order of ~ 10 mJy. The dominant infrared flare is

associated with the X-ray trigger spike.

Mirabel et al. (1998) observed a single infrared flare event associated with my /-7

curve, which reached an amplitude of ~ 30 mJy. Simultaneous infrared coverage is not

available for the other hard dips in Group 2. In all three a-class light curves of Group 3,

the hard dip is followed by a ~ 30 mJy infrared flare (Rothstein et al., 2005). In Mirabel

et al. (1998), an a-class event is observed to be followed by a flare that peaked at ~ 10

mJy. Rothstein et al. (2005) showed that the ~ 30 mJy flares can be explained as a

summation of Class C sub-flares, each associated with a soft X-ray flare.









7.3 Discussion

7.3.1 QPO Correlation with Spectral Features

Previous studies have shown that the QPO is most strongly tied to the thermal disk

component (\! iil.wardt et al., 1999; Feroci et al., 1999; Muno et al., 1999). Using the

September 1997 data (my 3-6 through 3-9), Markwardt et al. (1999) pointed out that the

correlation to disk flux is strongest when the QPO frequency is above 4 Hz. While this

is true, the QPO frequency is in this range less than 25'. of the time, and mostly falls

into this range when entering or exiting the hard dip. During the course of the hard dip,

the QPO will change significantly while the blackbody flux remains relatively constant.

Markwardt et al. (1999) also i- that at lower frequencies, there is an apparent broad

correlation with power law flux. I confirm that this correlation may exist, and show that

it is most apparent at lower frequencies. Because of a hysteresis effect, the deviation in

the QPO frequency -power law flux relationship is more apparent at high frequencies. I

-,l.:. -1 that an initially tight correlation is broken as the QPO begins to rise in the latter

half of the dip. Two out of four of the observed September events are strongly correlated

to power law flux and a different two out of four are correlated to blackbody temperature.

In contrast, the August 1997 (my /3-1 through 3-5) events show a strong correlation

with both power law and blackbody features specifically the power law flux and

blackbody temperature. A partial correlation analysis shows that if either the power

law flux or blackbody temperature is removed, the correlation to the other is weakened, so

it is not likely the effects of these components can be untangled. I do, however, argue that

the correlation to the power law component may be more fundamental, especially since

the power law flux is more strongly tied to the X-ray emission at this point and the disk

component is vanishingly small. In addition, the a-class curves show a consistently strong

correlation to the power law flux and less consistent correlation to blackbody features.

As mentioned before, the bulk of the QPO change occurs during the entry into and

exit from the dip. While in the dip, the spectral features remain fairly stable. Thus much









of the correlation strength is tied to QPO stability during the dip and the relatively short

motion at the start and end of the dip. This being the case, the rise in low-frequency

noise which greatly affects QPO detections in 3-8 and 3-9 is also likely to be affecting

the correlation coefficients. It is reasonable to assume that the number of non-detections

late in the dip is reducing the number of QPO detections above 4 Hz and thus artificially

lowering the correlation to blackbody features in these two cases. Short time-scale

variations (where the QPO suddenly dips and recovers) also tend to create outliers to

the correlation because the change in frequency is not accompanied by spectral variation.

It is uncertain if these jumps are real as the shape of the variation is indeterminant on

four-second time scales. These outliers are omitted from my correlation analysis.

7.3.2 Infrared Flaring Behavior

When comparing these events, I cannot ignore the fact that all these hard dips are

followed by an infrared flare, the strength of which is tied to the presence or absence of a

terminating spike. Furthermore, the shape of the spike is tied to the behavior of the QPO

during the preceding dip. Therefore, the properties of the QPO appear to be fundamental

for determining the subsequent jet activity in GRS 1915+105, and the question becomes,

"If all hard dips start relatively the same (with the appearance of a QPO), why don't all

hard dips end the same?" What causes the QPO to quickly rise in Group 2 or to have a

higher minimum frequency in Group 3? Is the energy that would be fed into the QPO and

subsequent infrared flare for a Group 1 event being leaked out before the dip termination

in Group 2 events, or is the energy missing from the system altogether? The Group 1

events have ..'i.- symmetric tri -:. r spikes and strong ~ 100 mJy infrared flares. The

relatively weak tri :._ r spike and ~ 30 mJy flare observed in September 1997 may be

evidence that a weaker trigger spike would be associated with a weaker primary flare.

On a final note, what I have referred to as the I 11i--. i spike" of the 3-class is so

named because it coincides with the start of the infrared flare. However, observations of

several 3-class events by Eikenberry et al. (1998) were unable to conclusively determine









whether the infrared flare began simultaneously with the tri -. -.-r spike. Also, in looking

at the Mirabel et al. (1998) event corresponding to my 3-7, one might believe that the

infrared flare starts 100 200 seconds prior to the spike (see Mirabel et al., 1998, their

Figure 3). Noting that the QPO also significantly weakens compared to the low frequency

noise component 100 200 seconds prior to the spike (see Fig. 6-5) sl-.-. -1- that the

origin of the infrared flare may be causally linked to the mechanism powering the QPO,

and not tied initially to the soft X-ray flaring or trigger spike. This behavior seir-.-- -1-

that the QPO is tied to a multi-wavelength energy release. The decoupling of the QPO

from X-ray spectral features in this case supports this hypothesis. In all other cases, the

disappearance of the QPO coincided with the trigger spike or the first X-ray oscillation

(and thus the infrared flare), so this picture would be consistent. Eikenberry et al.

(2000) observed a series of class C infrared flares preceding the X-ray soft flares in 0-class

dip-flare cycles, so this sequence would not be entirely unprecedented.

7.3.3 A Cause and Effect Summary

While the nature of the QPO is still uncertain, it does reflect and may possibly

be used to predict observable outcomes. In all of these cases, a QPO appears when the

dominant X-ray flux changes from soft to hard. While the change in hardness occurs very

quickly, the flux drops slowly for ~ 100 seconds. The QPO, initially around 6 Hz, falls

with similar smoothness (see Fig. 7-1). This region is where the tracks diverge. I believe

that the QPO behavior at the divergent point can ultimately be used to predict how

the dip will end. Consider the three groups I identify, summarized in terms of cause and

effect:

Group 1 A /-class light curve enters a hard dip phase. The QPO falls off to 2 Hz

and maintains that frequency for > 150 seconds.

End result:

The QPO frequency is tightly correlated to both blackbody and power law spectral

features and the correlation lasts the length of the dip.









The dip lasts 500 700 seconds and terminates with a -1 i-i.- narrow, symmetric

spike and increasing underlying flux.

A strong class B infrared flare of strength ~ 100 mJy follows with infrared excess

explained by pursuant X-ray oscillations.

Group 2 A 3-class light curve enters a hard dip phase. The QPO falls off to 2 Hz,

but begins increasing after < 100 seconds.

End result:

The QPO frequency decouples from the spectral features and the strength of the

correlation is related to the degree of hysteresis.

The QPO significantly weakens before the terminating spike.

The dip lasts 500 700 seconds and terminates with a weak, wide, ..i-mmetric spike

and flat or decreasing underlying flux.

A slightly weaker class B infrared flare of strength ~ 30 mJy follows, possibly starting

with the weakening of the QPO.

Group 3 An a-class light curve enters a hard dip phase. The QPO falls off to 3 Hz

and maintains that frequency for several hundred seconds.

End result: The QPO is correlated to power law features but not to blackbody

features.

The dip lasts ~ 1200 seconds and then immediately enters a period of oscillation.

A series of class C infrared flares with summed amplitude ~ 10 30 mJy follows and

is associated with the duration of the X-ray oscillations.

7.4 Summary

I have divided the 3-class into two sub-groups, identified by spectral behavior,

QPO frequency evolution, trigger spike morphology, and infrared flare strength. I compare

these events to a-class events. Each hard dip is associated with a variable low frequency

QPO and is followed by an infrared flare, indicating an ejection event. Similarities in the









X-ray light curve, X-ray hardness, and infrared flaring -,i-'-i -1 that a similar physical

mechanism is responsible for these behaviors.

While it is easy to expect different behavior from a- and 3-class light curves, it is

surprising to find differences within the 3-class itself. Most interesting is the spectrum

of QPO time evolution behaviors seen in my small set of observations and the fact that

only a slight variation is necessary to affect the end result. The correlation of the QPO

to X-ray spectral features hinges on the time evolution of the QPO. The time evolution

is also intimately tied to the trigger spike morphology and subsequent infrared flaring.

Most interesting is the possibility that the infrared flare begins with the disappearance

of the QPO and is not solely tied to the trigger spike or X-ray flares. It is clear that this

study only scratches the surface of the spectrum of behaviors exhibited by GRS 1915+105.

However, by studying these events together, I may better understand the underlying

mechanism. In the next chapter, I combine the a- and 3- class data and compare their

behavior to the O-class, which is associated with a smaller, class C infrared flare. I then

discuss the distinguishing behavors of these classes as associated with a QPO/jet ejection

model.





























200 400 600 200 400 600 200 400 600


0 500 1000 0 500 1000 0 500 1000
Time Time Time


Figure 7-1.


An overlay of the 1-second resolution X-ray light curve (line) and fine-binned
4-second QPO frequency (circles). Time is in seconds and QPO frequency
is in Hz. The open circles are low frequency power peaks observed when
the QPO is not detected (see Section 7.2). They are shown for illustrative
purposes. Note that for 3-6 through 3-9, the QPO spends less time near the
minimum frequency.














U

O-


0
0_
02


U
u
, 6
"3

L 4
O-





o
02




S2
U




c


t-4
ED 6









a-
u_ 4
0
a-



o 2




o 2


Figure 7-2.


2 4 6 2 4 6 2 4 6
Flux-PL Flux-PL Flux-PL


Scatter plots of QPO frequency (in Hz) with power law flux, Flux-PL, (in
10-sery cm-2s-1) for a- and 3-class light curves. Triangles indicate detections
in the first 100 seconds of entering the dip. Squares are points within the last
100 seconds before exiting the dip. The line is a fourth-order best fit to the
time evolution of the features. Note that for /3-1 through 3-5, a strong linear
correlation is apparent. For 3-6, 3-7, and 3-9, the correlation weakens and
I see different degrees of hysteresis. The open circles (non-detections of the
QPO as defined in Figure 7-1) do not contribute to the apparent hysteresis
pattern. In 3-8, the range of frequencies is significantly less, probably due to
increased non-detections. This is evidence that a continuum of QPO behaviors
exists within the /3-class light curves. In a-class light curves, a correlation
exists over a smaller range of frequencies as well.



















'6 / jA A4 4A AA



0
o- .4
a- -o 7 P-2 P -3
O 2

G6










1 2 3 1 2 -5 1 2 3
A

















10-8g cm-2 -1) for a- and 3-class light curves. Symbols are as in Figure
4 -












7-2. A correlation is observed above 4 Hz, but below 4 Hz, the QPO can
change significantly while the blackbody flux remains relatively constant.
o 2

1 2 3 1 2 3 1 2 3
Flux-BB Flux-BB Flux-BB

Figure 7-3. Scatter plots of QPO frequency (in Hz) with blackbody flux, Flux-BB, (in
10-s"er cm-2s-1) for a- and ,-class light curves. Symbols are as in Figure
7-2. A correlation is observed above 4 Hz, but below 4 Hz, the QPO can
change significantly while the blackbody flux remains relatively constant.






















CE
03

14
o
02
0



U


a2
o 2
>a




U
LZ 4
0









0-
0 2




0-
a
Lu 4

02
a2





u- 4
0
o2


1.0
Temp-BB


1.5 1.0 1.5
Temp-BB


Figure 7-4.


Scatter plots of QPO frequency (in Hz) with blackbody temperature,
Temp-BB, (in keV) for a- and 3-class light curves. Symbols are as in Figure
7-2. The a-class light curves clearly show no correlation. In most of the
/-class light curves, a correlation is seen but with a wider dispersion than
that of the power law flux relation. Hysteresis is particularly apparent in 3-7
and 3-9.


1.0
Temp-BB


















O
a
U)











(0
N








-J
o

o.






at
0
_3


L







E
I
-J






Qu
4)


200 400
Time in seconds


600


Figure 7-5.


The time evolution of a Group 1 event. The top panel shows a 1-second
resolution X-ray light curve. The next panel shows the fine-binned
QPO frequency and the open circles are low frequency power peaks where the
QPO is not detected. The X-ray power law flux (in units of 10-"erg cm-2s-1)
and blackbody disk temperature are shown at four-second resolution. The
QPO frequency is most strongly correlated to the power law flux during the
hard dip, but is also correlated to the blackbody disk temperature.


P-2

Group 1





I I I I
*
S



S*I *







----------- ^"


800


















10


S6 Group 2

24-

I 2



6- *-
o
n, 0

4 c 0
2 o0

X
3
4"- 4


W 2
0.


oI I I


1- -



0
0 200 400 600 800
Time in seconds


Figure 7-6. The time evolution of a Group 2 event. Panels are as described in Figure 7-5.
In this case, the QPO frequency begins rising while the power law flux and the
blackbody temperature are relatively constant.









132























a.
U
U)

o








U
Cr


o.
N






-.
E


I-
a3


0
0







E
I
,uc
Q.


Figure 7-7.


0 500 1000 1500
Time in seconds


The time evolution of a Group 3 event. Panels are as described in Figure 7-5.
The hard dip lasts ~ 1200 seconds twice the length of the Group 1 and 2
events. The QPO frequency behavior is similar in morphology to Group 1, and
while strongly correlated to the power law features, there is no accompanying
correlation to blackbody features.





















61-1 1-2 1-3


4


2

0 --- -- --

1P-4 1P-5 f-6
6

4


2




6 -7 -8 -9

4
=I

2

0
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
Relative Time (sec) Relative Time (sec) Relative Time (sec)


Figure 7-8.


The tri- v.-r spike of the /-class light curves for 1997 August (/3-1 through /-5)
and 1997 September (3-6 through /-9) data at one-second time resolution.
The solid lines are a double Gaussian plus polynomial fit to the data points. I
classify the /-1 through /-5 events as Group 1. These have a -1 i .i:- narrow,
symmetric spike and the underlying flux has a positive slope. The /-6 through
3-9 data I classify as Group 2 events. These have weaker, wider, .,-v iiii I lic
spikes and the underlying slope is flat or negative. The /3-8 and /-9 light
curves have more disturbed spike morphologies on this time scale.


a-

t

U


a-
N.
U


uU


a..
u
6)
u
U)
N.
U
10
>




















Table 7-1. Linear Pearson Correlation Coefficients


ID Group TF BBN BBF BBT PLN


/3-1
/3-2
/3-3
/3-4
/3-5
/3-6
/-7
/3-8
/3-9
a-10
a-11
n-12


0.96
0.93
0.96
0.96
0.96
0.90
0.76
0.84
0.87
0.90
0.87
0.87


-0.62
-0.63
-0.62
-0.66
-0.57
-0.53
-0.42
-0.40
-0.36
-0.10
-0.11
-0.11


0.84
0.76
0.84
0.89
0.86
0.89
0.78
0.54
0.67
0.73
0.57
0.63


0.87
0.86
0.90
0.89
0.85
0.88
0.75
0.63
0.63
0.48
0.31
0.43


0.87
0.87
0.81
0.88
0.90
0.79
0.63
0.68
0.76
0.67
0.74
0.64


PLF PLI


0.92
0.91
0.88
0.94
0.93
0.78
0.59
0.75
0.74
0.77
0.83
0.75


0.52
0.58
0.39
0.44
0.62
0.53
0.56
0.40
0.39
0.47
0.46
0.41


Note. -Correlation of spectral features to the
QPO frequency. The existence of a correlation is believable
for values of r|>0.40 and highly significant for values of
|r|>0.70. In most cases, the correlation to the power law flux
is strongest. The abbreviations are as follows. TF: total flux;
BBN: blackbody normalization; BBF: blackbody flux; BBT:
blackbody temperature; PLN: power law normalization; PLF:
power law flux; PLI: power law index. The Groups are defined
in the text.



















ID Group


/3-1
0-2
0-3
/-4
0-5
/-6
/-7
/-8
/-9
a-10
a-11
a-12


Table 7-2. Partial Correlation Coefficients
PLF BBF PLF BBT BBF PLF BBT -PLF


0.84
0.84
0.87
0.82
0.84
0.62
0.72
0.80
0.81
0.80
0.80
0.81


0.67
0.65
0.49
0.74
0.75
0.16
0.08
0.58
0.52
0.79
0.83
0.78


0.67
0.54
0.82
0.66
0.68
0.81
0.84
0.64
0.77
0.77
0.45
0.72


0.37
0.37
0.58
0.48
0.39
0.64
0.58
0.31
0.16
0.54
0.30
0.53


Note. -Partial correlation of spectral features to the QPO frequency.
The first column (PLF BBF) is the partial of the QPO frequency and
power law flux with the effect of blackbody flux removed. Note that the
correlations are strong in most of the cases. The second column (PLF -
BBT) relates the QPO frequency to power law flux, removing blackbody
temperature. These correlations are weaker, but significant in most Group 1
cases. The third column (BBF PLF) shows the QPO frequency correlation
to blackbody flux, removing power law flux. Most are believable, -i.--. lii-; a
complex interplay between the power law and blackbody features. The fourth
column (BBT -PLF) shows the QPO frequency correlation to blackbody
temperature, removing power law flux. In this case, nearly all correlations
drop below significance showing that power law flux may trace QPO behavior
better than blackbody temperature.


















Table 7-3. Length of Frequen
Min Freq Time at Min
ID Group (Hz) (s)


3-1
/-2
/-3
/-4
/-5
/-6
/-7
/-8
/-9
a-10
a-11
a-12


180
240
160
160(50)
180
80(20)
60
100
130(0)
560(0)
520(0)
800(0)


:y Dip


Fraction at Min

0.31
0.34
0.24
0.27
0.34
0.15
0.12
0.15
0.21
0.50
0.44
0.63


Note. -Approximate length of time QPO I1 I i- near lowest
frequency. The third column shows the minimum QPO frequency.
The fourth column shows the length of time the QPO 1 li within
0.5 Hz of the minimum frequency. The number in parentheses
is the time during which the QPO frequency is below 2.5 Hz if
it is not equal to the listed time. The fifth column shows the
approximate fraction of the dip length spent at the minimum
frequency. In four cases, the frequency does not drop below
2.5 Hz. In general, Group 2 events spend less time at the
minimum frequency than Group 1 events.





















Table 7-4. Gaussian Fit Parameters


ID Group A


0.70
1.10
0.90
1.09
0.86
0.80
0.62
0.58
0.56


JA C1 fint


2.20
2.39
2.56
1.85
2.23
3.64
3.26
5.16
6.11


18.24
9.98
10.32
12.14
16.12
-2.56
-1.61
-18.07
-16.62


Note. -Gaussian fit parameters
calculated for the 11 i.-.- i spike" in 3-class
light curves. The column labels are as
follows: A = the normalized amplitude; aA
Sthe width (standard deviation) of the
gaussian; C1 = the slope of the light curve
background; fit = the integrated count
rate. Their quantities are grouped by a
combination of spike strength, spike width,
and the sign of the underlying slope.


82.10
85.96
94.75
89.26
91.85
94.49
92.97
97.10
75.23









CHAPTER 8
GRS 1915+105: THE ACCRETION-EJECTION INSTABILITY MODEL AND THE
RADIUS DILEMMA4

8.1 Introduction

Despite the number of multi-wavelength observations of discrete jet ejections in

GRS 1915+105, the source of the low frequency QPO and the mechanism for launching

the relativistic jet remain unknown. The physical model for a microquasar starts with an

X-ray binary system. A mass-donating companion contributes primarily to the optical

and infrared light of the source. An accretion disk around a compact object adds light in

multiple wavebands, ranging from the optical to the gamma-ray. There are two primary

models for accretion in microquasars: the truncated disk model and the disk-corona

model. In the truncated disk model, a cool (<1 keV) disk extends to a transition radius,

inside of which accretion continues in a radiatively inefficient manner, such as by an

Advection-dominated Accretion Flow (N oi li-, and Yi, 1995, see also ('! ipter 1). In the

disk-corona model, a cool disk extends to the innermost stable orbit around the black

hole, while hard X-rays are produced in a magnetically heated corona above the disk.

In Figure 8-1, I show a cartoon model of how jet ejection might occur in the

disk-corona model. In the first phase (left), the accretion disk extends to the innermost

stable radius (a few Schwarzschild radii). The disk is visible through a transparent corona.

In the second phase (middle), some perturbation in the system causes pre-jet material to

accumulate and the corona to become opaque. This is when the QPO becomes apparent

in the X-ray light curve and evolves. Because part of the disk is obscured, the system

will appear spectrally hard. In the final phase (right), the inner disk is evacuated via jet

ejection. Once the ejection has occurred, the inner disk refills, returning the system to

phase 1. The evacuation of the disk means that whatever process that was causing the



4 This work was prepared in collaboration with P. Varniere, S. S. Eikenberry, D. M.
Rothstein.









QPO has ceased. Because the accretion disk is no longer obscured by the corona, the

system immediately appears softer and the soft component strengthens as the disk refills.

This model makes several assumptions. For example, it assumes that the accretion disk

extends to the innermost stable orbit and is obscured by the corona in the pre-ejection

phase. This model also implies that the QPO may be an observational signature of the

perturbation that causes the state change. These assumptions are not necessarily the case

and they have great implications for where and how the QPO arises.

In this chapter I discuss a specific disk-corona model called the Accretion-Fi. I i',

Instability (AEI) model, which explains low frequency QPOs observed in low-mass XRBs,

specifically those with black hole primaries like GRS 1915+105 (Varnikre et al., 2002).

The model ties the QPO frequency to an AEI frequency (T ,--. r" et al., 2004) and makes

specific predictions about X-ray spectral behavior and jet ejection strength. The results

presented here are a preliminary, but promising, report of an ongoing collaborative work.

8.2 X-ray QPO Phenomenology in the 0-, a-, and 3-classes

The X-ray spectrum of GRS 1915+105 is modeled with a thermal blackbody

component that traces the radiating hot disk. The non-thermal component is fit with

a simple power law. I studied three classes of X-ray light-curves, each with extended,

spectrally hard dips. In previous chapters, I focused on 2-10 Hz QPOs in the X-ray

light-curve of the a- and 3-classes and showed that the QPO frequency evolution can be

used to predict subsequent X-ray and infrared flaring behavior. Following the analysis of

the previous chapter, I extended my work to include the O-class. The inclusion of this data

will allow us to test multiple regimes of the AEI model (see below). In each of these three

classes, a QPO appears with the onset of a spectrally hard dip and an infrared flare begins

as the hard dip ends. The infrared flares associated with each dip/flare cycle are ~ 100

mJy for the /3-class, ~ 10 mJy for the a-class, and ~ 5 mJy for the O-class. The X-ray

events discussed are detailed in Table 6-1. The X-ray coverage of the 0-class events does









not encompass a full dip cycle, so although I can apply correlation analyses, I am unable

to perform specific timing analyses analogous to those performed in the previous chapter.

As shown in Figure 6-6, the QPO is clearly present in the 0-class above 4 Hz. I use

this apparent criterion to select the coherent QPO frequency and distinguish it from

low-frequency noise in subsequent analyses. In Figures 8-2 through 8-6, I show scatter

plots of the QPO frequency versus total flux, power-law flux, blackbody flux, blackbody

temperature, and color radius. Instead of examining the events individually, I compare the

a, 3, and 0 classes as a whole, focusing on detections above 4 Hz. Immediately apparent

in Figure 8-2 is that the 0-class observations occur at a higher total flux than the a- and

/-class observations, and in Figures 8-3 and 8-4, I note that the difference is apparent in

both the power law and blackbody flux. However, the total flux observed in the 0-class

increases with QPO frequency at twice the rate of that observed in the a- and 3-classes.

This trend is mirrored in the blackbody flux (Fig. 8-4), but not in the power law flux,

where the QPO frequency power law flux trend in the 0-class runs approximately parallel

to that of the a- and p-classes (Fig. 8-3). In Table 8-1, I calculate the Linear Pearson

Correlation Coefficient (see previous chapter) between the QPO frequency and various

spectral features and show that in the 0-class the QPO frequency is strongly correlated to

the power law, blackbody and total fluxes. While the blackbody temperature is relatively

higher in the 0-class, it follows a trend with QPO frequency that runs parallel to the

relationship shown in the a and 3 classes (Figure 8-5). The blackbody temperature is also

well-correlated to the QPO frequency.

Finally, the 0-class observations show a lower blackbody normalization overall, and

are consequently associated with a lower average radii (Fig 8-6). The radius (also called

the color radius) is calculated from the blackbody normalization, NBB, the distance, and

the inclination. From Fender et al. (1999), recall that the distance to GRS 1915+105 is

d = 11.2 0.8 kpc. The inclination of the system is 66 2. The color radius, Rco is









calculated such that
d NBB
1lkpc cos(i) (8

Due to the nature of the model, the color radius measures the innermost observed

radius of the accretion disk, RIN. The average color radius value is r ~ 21km in the

0-class observations, as compared to ~ 30km in the a-class, ~ 34km in the 32-class,

and liii in the 31-class (see previous chapter for description of classes). None of the

three classes show apparent correlation between the radius and the QPO frequency, but

the apparent distinction in the range of radii occupied by each of the light curve classes

becomes important when applying these observations to tests of the AEI model.

8.3 QPO Models and the Radius Dilemma

QPO models tend to tie the QPO frequency to a magnetoacoustical frequency (e.g.,

Titarchuk and Fiorito, 2004) or a Keplerian frequency (e.g., Merloni et al., 2000). The

Keplerian frequency 2K describes the motion of a particle in an accretion disk due to

gravity. The Keplerian frequency associated with the inner-most accretion radius is

GM R-3/2
wK (8-2)
27 IN

The general problem with this is that the QPO frequency is too low to be a

Keplerian frequency. The Keplerian frequency associated with the Schwarzschild radius of

GRS 1915+105 (M = 14Me; Greiner et al., 2001) is QK = 2000Hz, whereas the observed

QPO frequency is 2 10 Hz. Thus the QPO frequency is often monotonically scaled in

order to reconcile it to the Keplerian frequency. In that way, the QPO frequency can be

directly related to Keplerian dynamical motion, which is well understood.

As mentioned in C'i lpter 7 and again in the previous section, the color radius

associated with the a-, 3-, and 0-classes is very poorly correlated to the QPO frequency,

as compared to other spectral features. Thus, either the QPO frequency does not trace the

Keplerian frequency or the color radius output from the XSpec model does not trace the

inner disk radius. I refer to this contradiction as the Radius Dilemma.









The dissociation of the color radius and the inner disk radius is not a surprising

discrepancy. When I fit the X-ray spectra using XSpec 11.3, the blackbody normalization

reveals a radius too small to be the innermost stable orbit. The compact object in

GRS 1915+105 is estimated to be 14Me, placing the Schwarzschild radius at Rs ~ 42km

(\! ,rti et al., 2000). My XSpec fits use blackbody normalization to estimate a radius and

find that values ranging from 10 80km, which is often inside the Schwarzschild radius.

This is a known problem that seems to be often ignored in current literature. Merloni

et al. (2000) performed a reliability study of using XSpec in an attempt to resolve this

dilemma. They created a series of model X-ray spectra, added noise and accounted for

detection quirks, and found that the XSpec radius fits systematically underestimated the

radius values for a standard Shakura-Sunyaev disk. For more complex systems in which

the accretion rate or the energy dissipated in the corona are allowed to vary, the inferred

radius from the model appeared to change, even when the actual radius was fixed at the

innermost stable orbit. In their study, Merloni et al. (2000) found that there is no single

correction factor that can be applied to correct the measured radius to the true radius,

especially in a volatile system.

Despite the known caveats, the belief persists that the QPO frequency, VQPO, and the

color radius, Rco, can be associated with the Keplerian frequency. This is in part because

in a subset of observations, the QPO frequency and the color radius show an apparent

trend of vQPo o Ro /2. Of the 17 light-curves I examine for this work, only three show

this apparent trend. In Table 8-2, I estimate the power law value, n, of a VQPo oc Rxol

fit. I fit only data taken within the first 200 s of each light curve's hard dip, as the

blackbody flux is still relatively strong here and the trace of the blackbody normalization

is probably more reliable. In examining 9 3-class light curves, 3 a-class light-curves,

and 5 0-class light curves, I find that the instance of a "K. III. ii i relationship between

the QPO frequency and color radius is in the 3-class. Even then, the frc equ-ii' --radius

relationship is inconsistent. Both the a- and 0-class show relatively flat relationships. In









fact, when considering the radius and QPO frequency evolution over the entire dip, the

0-class shows a relationship inverse of what one would expect. In Figure 8-7, I plot the

relationship of the color radius to the QPO frequency when considering the full length of

the dip. The radius relationship is singularly inverted in the case of the 0-class.

8.4 The Accretion-ejection Instability Model

The AEI model was developed by T r.--. r and Pellat (1999) in order to explain the

pattern of low-frequency QPOs in microquasars. The AEI predicts a clear observational

signature -the turnover of the QPO frequency inner radius relationship (Varnikre et al.,

2002; T .-.- r et al., 2004). In Figure 8-8, I show the predictions of the AEI model against

the observational data presented above. At higher color radii (Reo, > 40km), the model

predicts a Keplerian relationship vQPo o R--32. At lower radii (Rcol < 40km), the model

deviates from a Keplerian relationship. The variation of the QPO frequency with the disk

inner radius is thus one of the few potential observational tests of this theory. Varnikre

et al. (2002) use this predicted turnover to explain discrepancies in the behavior of the

QPO in the microquasars GRO J1655-40 and GRS 1915+105.

Thus, while the inversion observed between the 0- and 0-classes in the QPO

frequency radius relationship negates any monotonic solution for reconciling the

Keplerian frequency to the QPO frequency, it is not an issue for the AEI model because

there exists a regime in the AEI model in which the QPO frequency-radius relation

exhibits an inversion (Varnikre et al., 2002). In the AEI model, a Magneto-Rotational

Instability (\ilRI) causes disk turbulence that suppresses the QPO during the soft state.

As soon as the AEI appears, the disk cools down and the power density spectrum changes.

The appearance of the QPO causes the transition into the low-hard state by stopping

the heating of the accretion disk and funneling energy into the corona (T.--. I- et al.,

2004). Using a single spiral arm model, the source of the AEI is associated outward to

a co-rotation radius which is (in some magnetic regimes) associated with the Keplerian

Frequency and hence the Keplerian radius. In the low-hard state, the disk accumulates a









critical amount of vertical magnetic flux in its inner region and only a reconnection event

can decrease the magnetic field. For a long dip, the QPO is the observational signature of

a physical mechanism that quenches MRI turbulence. Observationally, energy that would

normally be lost via heat radiation is funneled into the mechanism powering the QPO

which may explain the initial strong correlation of the QPO frequency to the blackbody

flux and disk temperature. Once initiated, the QPO transfers energy to the corona, hence

establishing a correlation to the power law flux.

Relativistic effects as the accretion disk approaches the central object affect the

physics of the disk instability and thus the QPO frequency inner radius relationship. As

the inner radius of the disk approaches the last stable orbit, the QPO frequency associated

with the AEI deviates from the scaled Keplerian frequency. This is the turnover point

in the model, as shown in Figure 8-8. The model assumes that QPO frequency and the

radius are directly connected, which, according to early literature, is true (\I!,i et al.,

1999). In the previous chapter, I show that the blackbody normalization, and hence the

color radius, is correlated to the QPO frequency, with correlation coefficients ranging

from R = -0.36 -0.66 in the 3-class (see Table 7-1). Recall from ('!i lter 7 that

R| = 0.4 0.6 are believable, but note that I generally disregarded these correlation

values when more statistically significant correlations were present. The correlations to

0-class data show an inverted, but still correlated relationship R = 0.51 0.65 (Table 8-1).

Note these are linear correlation coefficients, and thus would naturally be less statistically

significant for a non-linear correlation.

It would not be unreasonable to assume that the magnetic field conditions are

sufficiently different between the a-, 3-, and 0-classes so as to push each group onto

different regimes of this curve. In fact, the discovery of this inversion in the QPO

fre',-I, i- T,-i-I.m-r radius trend is the first record of an inversion within a single source

and lends credence to the AEI model. The AEI model is currently one of the only

QPO frequency models that accounts for this heretofore unexplained reversal in the









frc, -uT.- -, -radius trend. Additionally, the AEI predicts that the energy in the jet ejecta

will vary depending on the range of the fre qT-- -radius curve that is observed in an

event. Finding evidence of such a turnover within a single source is a very exciting step

toward validating this model.

Acknowledging that all X-ray states of GRS 1915+105 do not fall within the

Keplerian regime of the AEI model resolves part of the "Radius Dil ii. n, i described

in the previous section. However, it is reasonable to assume that all observations of a

given class of X-ray light curves will operate within a single regime of the model. In

Figure 8-7, I show that the 3-class falls within the "K. 1I1, i ', regime. Assuming that

the AEI model can accurately predict the color radius based on the QPO frequency, I

perform the following test to see if the blackbody normalization (and related color radius)

determined by fitting the RXTE data with XSpec models can be improved.

First, using a /-class light curve, I find the QPO frequency as a function of time.

By using the /3-1 data set, in which the QPO frequency and color radius are shown to be

nearly Keplerian, I establish a scaling factor such that

VQPO coi
log o= -1.5 x log
2.5H 58km

In this way, I can then calculate a theoretical blackbody normalization based on

the QPO frequency. I allow for an error of 0.6Hz based on the standard scatter of the

QPO frequency. I emphasize that there is a distinction in this blackbody normalization

determined by a QPO frequency relationship ( NBB,K) and the blackbody normalization

predicted by XSpec model fits (NBB,x). I then input the theoretical blackbody normalization,

NBB,K, into the XSpec model fits, allowing the XSpec parameter NBB,X to vary over the

approximate error related to the spread in QPO frequency. Other XSpec model fitting

parameters (e.g., absorption, power law flux, and blackbody temperature) are left as free

parameters. Attempts to absolutely fix the XSpec blackbody normalization parameter

NBB,X to the Keplerian prediction of NBB,K yielded no good fits. Even allowing for









the error associated with 0.6Hz and a loose restriction of X2 < 8, I find that 5(0'. of

my spectra failed to make any fit. In contrast, when allowing the XSpec blackbody

normalization NBB,X to be completely free and placing a tighter constraint of X2 < 2, I

find good fits for 91 '. of the spectra. I compare the results of the two spectral fitting tests

in Figure 8-9. This test sI--l:. -1- that the XSpec blackbody normalization NBB,x cannot

be forcibly conformed to the predictions of the Keplerian regime, NBB,K, in the 3-class.

This may -r.-.ii. -1 that the instantaneous color radius of a spectrum is not associated with

the QPO frequency in a Keplerian manner. It is more likely that the system prior to a jet

ejection event is dissipating energy into the corona and/or the accretion rate is (I i1,ilr,

and thus the radius fits are unreliable (\I. I. h.i et al., 2000).

Still, the AEI model is an improvement over models that predict simple Keplerian

scaling in the QPO frequency radius relationship. Despite inherent problems in radius

estimation, it is interesting (and promising) that the 0-class data, which have lower radii

than the 3-class data, fall within the same 2 10 Hz frequency range and are hence

consistent with the turnover predicted in the AEI model. It gives hope that while the

instantaneous estimates of the color radius may be off, the trend predicted by the XSpec

model fits is statistically accurate and useful for testing models.

8.5 Caveats, Conclusions, and Future Work

The Radius Dilemma, as I have called it, is a multi-l iv-r dilemma arising from

attempts to associate the poorly understood QPO frequency with the better understood

Keplerian frequency.I have sought to investigate a few oft-overlooked issues and point out

a few more caveats that need to be addressed when modeling jet ejection. In summary, I

have shown:

1. The standard XSpec blackbody model produces erroneously low radius fits,

often inside the presumed Schwarzschild radius of the source, even for a standard

Shakura-Sunyaev disk. Merloni et al. (2000) have previously studied this issue and









shown that the radius fits cannot be monotonically adjusted to an associated inner disk

radius.

2. If the radius trend traces the inner disk radius, then no monotonic scaling factor

can rectify the discrepancy in the relation between the QPO frequency to the Keplerian

frequency. QPO models must account for an apparent inversion in the QPO frequency -

color radius relationship.

3. The AEI model does have a potential solution for the inverted QPO frequency -

radius relation seen in the O-class. However, attempts to predict the color radius using

Keplerian scaling of the QPO and using this radius to constrain model fits to X-ray data

yields poor fits for the /3-class data (which the AEI model predicts to be in a Keplerian

regime). These poor fits are likely indicative of a volatile accretion system that is not

accurately port i 'i-, by this generally used XSpec model.

The dilemma is exacerbated by problems inherent to the data I have collected.

First, it is difficult to accurately detect the blackbody contribution of the spectrum in a

spectrally hard state. Also, in these particular states, I believe that the accretion disk is

undergoing serious disruption, so there is no reason to assume a blackbody model will be

accurate. It is also likely that even though the blackbody plus power law model that I

use produces good fits to the data, it is too simplistic to convey meaningful information

about this system. Finally, XSpec is known to return erroneous outputs depending on the

current behavior of the system (. I !in et al., 2000). These caveats make it difficult to

-,v confidently at this juncture that the QPO frequency is not associated to the Keplerian

frequency.

To end on a positive note, however, if one assumes that the trend I observe in

the QPO frequency radius relation of GRS 1915+105 can be believed, then this set

of observations captures a very important observational prediction of the AEI model

and behaviorally links GRS 1915+105 to the microquasar GRO 1655-40, taking an

exciting step toward validating this model. I can further refine this result by gathering









archival data of GRS 1915+105 in different X-ray spectral states, specifically targeting

times when multi-wavelength observations are available. In this way, I can test the

validity of the model in predicting jet ejection strength as well as the QPO behavior.

Additionally, I can expand the analysis to multiple microquasar systems and test the

applicability of the model. For example, the microquasar GRO 1655-40 has been observed

in the non-Keplerian regime, but additional observations may show that this source, like

GRS 1915+105, can display behaviors consistent with multiple regimes. In comparing

multiple sources exhibiting QPOs, we can determine whether most XRB sources stay

confined to a single part of the fre qu, -ii --radius curve or whether all microquasars have

a range of variability (and associated jet ejections) similar to GRS 1915+105. Ultimately,

in testing observations against theory, we can test whether we are able to accurately

predict larger jet ejections via QPO observations so that we can study ejection events

more closely.










Disk Evacuation Cartoon Model


Figure 8-1.


A cartoon model of jet ejection in the disk-corona scenario. 1. A full accretion
disk extends to the event horizon and is visible through a transparent corona.
2. The inner disk begins to evacuate and the corona becomes opaque. 3. The
evacuated disk component is funneled into the jet. After the jet ejection, the
disk fills in again.














































Figure 8-2. QPO frequency vs. total flux for a-class (diamonds), 3-class (crosses), and
0-class (triangles) above 4 Hz. The flux is measured in erg/s; the frequency
is measured in Hz. The O-class observations have a higher total flux, and the
trend is twice as steep as that observed in the a- and 3-classes.















































+ ( +


1x10-8 -




4 5 6 7 8 9 10
QPO-Frequency


Figure 8-3. QPO frequency vs. power flux for a-class (diamonds), p-class (crosses), and
0-class (triangles) above 4 Hz. The flux is measured in erg/s; the frequency is
measured in Hz. The 0-class observations have a higher overall flux, and the
trend is parallel to that observed in the a- and p-classes.














































I X' I 0v8
A-





4 5 6 7 8 9 10
QPO-Frequency

Figure 8-4. QPO frequency vs. total flux for a-class (diamonds), p-class (crosses), and
0-class (triangles) above 4 Hz. The flux is measured in erg/s; the frequency
is measured in Hz. The 0-class observations have a higher blackbody flux,
and the trend, like that of the total flux in Fig. 8-2, is twice as steep as that
observed in the a- and /-classes.

































A +


-4 +
++ +

^ ++4
+ 0
.. ;. +

















Figure 8-5. QPO frequency vs. blackbody temperature for a-class (diamonds), +-class
(crosses), and 0-class (triangles) above 4 Hz. The temperature is measured in
OQ







4 5 6 7 8 9 10
QPO Frequency


Figure 8-5. QPO frequency vs. blackbody temperature for a-class (diamonds), /3-class
(crosses), and 0-class (triangles) above 4 Hz. The temperature is measured in
keV; the frequency is measured in Hz. The O-class observations have a higher
overall temperature, and the trend is parallel to that observed in the a- and
3-classes.














































Figure 8-6. QPO frequency vs. inner disk radius for a-class (diamonds), 3-class (crosses),
and O-class (triangles) above 4 Hz. The radius is measured in kilometers; the
frequency is measured in Hz. The O-class observations have a lower overall
radius. No obvious trend is apparent in either class.













155














I I



0


I )





(14/08/97)


I I I


Alpho (27/07/02)

S. I .. I I


0 20 40 60
Color Radius (km)


I I


K Theta (10/07/98)
STheto (05/09/97)

0 20 40 60
Color Radius (km)


Figure 8-7.


The color radius vs. QPO frequency. The bars on these points represent the
dispersion of frequencies observed in each radius bin. Although the average
points for the 3-class -i-.-, -1 a trend reminiscent of a Keplerian frequency -
radius relation, there is high dispersion at the lower radii. The a-class shows
no apparent trend between color radius and QPO frequency. The 0-class shows


an inverted relationship.
hard dip events.


These points represent averages among 17 different


Table 8-1. Linear Pearson Correlation Coefficients
ID TF BBN BBF BBT PLN PLF PLI


0-13 0.86
0-14 0.91
0-15 0.86
0-16 0.86


0.65
0.57
0.07
0.51


0.90
0.91
0.81
0.85


0.61
0.60
0.77
0.65


0.61
0.10
0.36
0.49


0.72
0.74
0.76
0.76


-0.13
-0.64
-0.34
-0.37


Note. -Correlation of spectral features to the
QPO frequency for 0-class observations. The existence
of a correlation is believable for values of |r1>0.40
and highly significant for values of |r >0.70. The
abbreviations are as in Table 7-1.


. Beto


Beto (09/09/97)
SI I I *


0 20 40 60
Color Radius (kIn)






















U :
-. -









-m '
m
m1 MUE :" *,*


*










m
U*


.. q


E
ME -
* U UEM
M ME =U
mn U%".


a OPO theo
. OPO (Hz)
. beta 1: freq
. theta_13:freq


r int 1


Figure 8-8.


Radius (in km) vs. QPO frequency (in Hz) on a log-log scale. The theoretical
predictions of the AEI model are shown as large blue circles while the
observational data is shown as small squares. The /-class data is shown in red,
and occupies the Keplerian regime of the model. The c-class data is shown in
pink and occurs near the turnover. The 0-class data is shown in black, where
the trend is inverted. This preliminary comparison of the data and model was
contributed by collaborator Peggy Varniere.

















































Figure 8-9.


Radius fits for the 3-2 event. The open squares are the original radius fits
with XSpec 11.3 (X2<2, radius is a free parameter). The black lines are
theoretical estimates based on the Magnetic Flood Model and allowing for
a 0.6 Hz frequency error. The filled squares are the revised XSpec fits using
the restricted theoretical radius. In order to get a reasonable number of fits,
I allowed X2<8. Even with the relaxed restriction on X2, only 5 !'. of the
spectra can be fit (compared to my original criteria, which fit 911 of the
spectra).












Table 8-2. QPO Frequency- Radius Fits
Set ID Power Law Index

3-1 -1.40
0-2 -1.42
0-3 -0.98
0-4 -0.69
0-5 -1.53
0-6 -0.92
/P-7 -0.74
0-8 -0.48
0-9 -0.29
a-10 +0.26
t-11 -0.21
a-12 -0.02
0-13 +0.106
0-14 +0.24
0-15 +0.09
0-16
0-17 +0.06


Note. -Observation IDs of
the 17 epochs, employing the
naming scheme from Table 6-1.
I fit the QPO-Frequency and
the color radius with a power
law. In a Keplerian case, the
Power Law Index would be
-1.5. These fits only consider
the first 200 seconds of the dip
where the correlation of the
QPO frequency and blackbody
features are more certain (see
('!i lpter 7). No value is given
for 0-16 because the start of
the dip is not visible. Note
how the a and 0 classes appear
relatively flat here.









CHAPTER 9
CONCLUSIONS AND FUTURE WORK

Uncommon sources like GRS 1915+105 and Edd-1 open the door to studying a

myriad of physical processes including the evolution of massive stars and the physics

at the inner edge of an accretion disk. This dissertation as a whole is composed of

several smaller projects designed to understand such exotic and 1i rI i. us X-ray

sources. In ('! Ilpters 2-5, I identify IR counterparts to the Galactic X-ray-emitting

population, I perform an extensive study on the source Edd-1, and and I describe my

efforts to construct a database of known massive binary systems. In Chapters 6-8, I

describe my investigation of the archetype microquasar GRS 1915+105, specifically the

phenomenological behavior proceeding a jet ejection.

9.1 The Galactic X-ray Population

The demographics of the stellar population of the Galactic Center are only recently

coming to light with the advent of high-resolution X-ray and IR astronomy. X-ray sources

require careful study at multiple wavelengths in order to identify their nature with

similar accuracy to that achieved using optically-based classification systems. For this

dissertation, I have performed spectroscopic observations to find infrared counterparts to

X-ray sources, specifically seeking signatures to high energy processes.

In C!i lpter 2, I detail the nature of my IR survey. IR spectroscopic follow-up is

imperative for confirming counterparts to X-ray sources in the GC and resolving their

nature. For the bright sources targeted in our survey, I did not find a significantly large

fraction of massive star or emission line counterparts. In 2005, I identified Edd-1 as the

first spectroscopically confirmed IR counterpart to a low luminosity Chandra source.

The IR survey begun here will meet quick advancement as the Flamingos-2

instrument comes online and the Flamingos-2 Galactic Center Survey begins (Eikenberry

et al., 2005). The unprecedented spectroscopic capabilities and the specific targeting of

X-ray source counterparts will yield a significant increase in the IR spectroscopic database









of compact objects in the Galactic Center. This will aid in statistically evaluating our

method of selecting counterparts to X-ray sources. Additionally, I expect a number of

interesting counterparts will be targeted for follow-up observations.

In ('!i lpters 3 and 4, I discuss follow-up observations of a strong emission-line

source identified during this survey as the infrared counterpart to an X-ray source.

C'!i lpter 3 focuses on the IR discovery spectrum and the IR/X-ray spectral analysis of

the source, while in C'! Ilpter 4 I use a detailed timing analysis to place constraints on

the nature of the system. Edd-1 shows strong Hel (2.114-pm), Br6 (1.945-pm), and

Br7 (2.166-pm) emission lines, typical of both accretion-powered binaries and Colliding

Wind Binaries (CWBs). Additionally, I observe Brackett series, Hel, Hell, CIII, and NIII

line emission. P Cygni profiles are visible in several Hell lines, -ir-'-. -I ii.; wind activity

around a massive star. The X-ray spectrum of this source is particularly intriguing, having

prominent Fe-XXV emission centered at 6.7 keV with an equivalent width of 2.2 keV. This

is one of the highest equivalent width Fe-XXV lines ever seen. Combining a long-term

Chandra monitoring campaign with archival XMM data, I searched for periodicity in the

X-ray light curve and found a period of 189 6 di -. I am currently using combined X-ray

and IR spectral and timing information to place constraints on the nature of the source.

My analysis indicates that Edd-1 is a binary system containing at least one massive star

- either a CWB or a HMXB. If the X-ray flux variation is interpreted as an X-ray eclipse

associated with an orbital period, then Edd-1 is likely a HMXB.

Future IR and X-ray observations of Edd-1 should target the source at times

of expected X-ray minima and maxima so that we can understand more fully the

evolution of the spectrum over the 189 d period. If Edd-1 is an accreting binary source,

IR observations of the source at times of low X-ray flux may reveal the nature of the

companion.

In C'! lpter 5, I discuss my preliminary work on a multi-wavelength spectral and

photometric Atlas designed for the rapid characterization of X-ray/IR sources. The









compilation of multi-wavelength statistics is vital for distinguishing XRBs and CWBs in

the Galactic Center. By establishing a spectral resource of canonical archetypes, I can

better and more quickly characterize new GC sources. In the immediate future, I plan

to extend the number of sources in the Atlas and refine the user interface. Identifying

new high mass binary systems and establishing a multi-wavelength Atlas will enhance our

knowledge of the processes at work in these sources.

9.2 Microquasars

Because their energetic jets evolve over short timescales, microquasars are ideal

laboratories for studying jet formation. GRS 1915+105 is particularly interesting because

it has been in a state of continuous outburst for sixteen years and thus allows extended

study of the phenomenon. In C'i plter 6, I introduce GRS 1915+105 and in C'! plter

7, I discuss the X-ray evolution proceeding a jet ejection in the context of subsequent

infrared flaring. I show that the X-ray QPO can be used to predict subsequent X-ray and

IR flaring behavior. In the three X-ray light curve classes that I observe (a, 3, and 0),

simultaneous X-ray and infrared observations show infrared flares rising as the hard X-ray

dip ends and the QPO vanishes. Similarities in the X-ray light curve, X-ray hardness, and

infrared flaring I-,-. -1 that a similar mechanism is responsible for these behaviors in each

class.

In C'! lpter 8, I reconcile my observations to one of the predominant jet ejection

models, the AEI. The AEI predicts a clear observational signature the turnover of

the QPO frcqu'- i-_, -inner radius curve. One imperative in validating this model will be

better understanding the evolution of the inner disk radius. Previous observations with

the Rossi X-ray Timing Explorer (RXTE) allowed us to study the QPO evolution with

exceptional time resolution. However, the QPO is visible in spectrally hard states, thus it

is often difficult to distinguish the soft (0.5-4 keV) blackbody flux from the total flux. To

better study the soft component, one could propose to do observations of GRS 1915+105

on X1 1-I-N. .-ton and Suzaku. Both the XMM and Suzaku instruments have a wider









coverage area than RXTE in the soft X-ray band, allowing more detailed soft X-ray

spectra. Consequently they are better equipped to probe the disk contribution to the

overall flux. By observing GRS 1915+105 with more sensitive soft X-ray coverage using

XMM and Suzaku, we can constrain the nature of the the disk contribution in ejection

events.

The formation of relativistic jets is an unusual and counterintuitive process.

Microquasars give us the opportunity to examine this process at close range and at

high speeds. The complex disk-jet interaction probes the warping of space-time in the

vicinity of the black hole testing the limits of General Relativity in cases of extreme

gravity.









APPENDIX
ABBREVIATIONS
2MASS Two-Micron All Sky Survey

AEI Accretion-F!i i .n Instability

AGN Active Galactic Nuclei

CV Cataclysmic Variable

CWB Colliding Wind Binary

Ci',i h i. Chandra X-ray Observatory

F2GCS The Flamingos-2 Galactic Center Survey

FFT Fast Fourier Transform

GC Galactic Center

GP Galactic Plane

HMXB High-mass X-ray binary

IR infrared

IRTF Infrared Telescope Facility

LMXB Low-mass X-ray binary

MRI Magneto-Rotational Instability

QPO Quasiperiodic Oscillation

RXTE Rossi X-ray Timing Explorer

XRB X-ray Binary

XMM X111--N. i-ton, derived from X-ray Multi-Mirror

mission









REFERENCES

Arnaud, K. A. 1996, in Astronomical Society of the Pacific Conference Series, 101,
Astronomical Data Analysis Software and Systems V, Edited by G. H. Jacoby &
J. Barnes, 17

Balbus, S. A., & Hawley, J. F. 1991, ApJ, 376, 214

Balbus, S. A., & Hawley, J. F. 1998, Reviews of Modern Physics, 70, 1

Bandyopadhyay, R., Shahbaz, T., Charles, P. A., van Kerkwijk, M. H., & N lilor, T. 1997,
MNRAS, 285, 718

Bandyopadhyay, R. M. 2008, private communication

Bandyopadhyay, R. M., Miller-Jones, J. C. A., Blundell, K. M., Bauer, F. E.,
Podsiadlowski, P., Gosling, A. J., Wang, Q. D., Pfahl, E., & Rappaport, S. 2005,
MNRAS, 364, 1195

Beckmann, V., Kennea, J. A., Markwardt, C., Paizis, A., Soldi, S., Rodriguez, J.,
Barthelmy, S. D., Burrows, D. N., C'! -I. r, M., Gehrels, N., Mowlavi, N., & Nousek, J.
2005, ApJ, 631, 506

Belloni, T., Klein-Wolt, M., M6ndez, M., van der Klis, M., & van P i ldiis, J. 2000, A&A,
355, 271

Blandford, R. D., & Begelman, M. C. 1999, MNRAS, 303, L1

Boirin, L., Parmar, A. N., Oosterbroek, T., Lumb, D., Orlandini, M., & Schartel, N. 2002,
A&A, 394, 205

Cardelli, J. A., Cl?-vton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245

Cassinelli, J. P., Waldron, W. L., Sanders, W. T., Harnden, F. R., Jr., Rosner, R., &
Vaiana, G. S. 1981, ApJ, 250, 677

Castro-Tirado, A. J., Brandt, S., & Lund, N. 1992, IAU Circ., 5590, 2

Choi, C. S., Dotani, T., Day, C. S. R., & N I, -., F. 1996, ApJ, 471, 447

Clark, J. S., CI' i l!. s, P. A., Clarkson, W. I., & Coe, M. J. 2003, A&A, 400, 655

Clark, J. S., Miroshnichenko, A. S., Larionov, V. M., Lyuty, V. M., Hynes, R. I., Pooley,
G. G., Coe, M. J., McCollough, M., Dieters, S., Efimov, Y. S., Fabregat, J., Goranskii,
V. P., Haswell, C. A., Metlova, N. V., Robinson, E. L., Roche, P., Shenavrin, V. I., &
Welsh, W. F. 2000, A&A, 356, 50

Clark, L. L., & Dolan, J. F. 1999, A&A, 350, 1085

Coe, M. J., Buckley, D. A. H., Fabregat, J., Steele, L. A., Still, M. D., & Torrejon, J. M.
1997, A&AS, 126, 237









Cohen, D. H. 2000, in Astronomical Society of the Pacific Conference Series, 214, IAU
Colloq. 175: The Be Phenomenon in Early-Type Stars, Edited by M. A. Smith, H. F.
Henrichs, & J. Fabregat, 156

Corcoran, M. F., Hamaguchi, K., Pollock, A. M. T., Pittard, J. M., Stevens, I. R., Henley,
D. B., Moffat, A. F. J., & Marchenko, S. 2006, in IAU Symposium, 230, Populations of
High Energy Sources in Galaxies, Edited by E. J. A. Meurs & G. Fabbiano, 35

Cox, A. N. 2000, Allen's A-i i. .1piv-ical Quantities (New York: AIP Press)

Crowther, P. A. 2007, ARA&A, 45, 177

Cushing, M. C., Vacca, W. D., & Rayner, J. T. 2004, PASP, 116, 362

Cutri, R. M., Skrutskie, M. F., van Dyk, S., Beichman, C. A., Carpenter, J. M.,
C'! -I. r, T., Cambresy, L., Evans, T., Fowler, J., Gizis, J., Howard, E., Huchra,
J., Jarrett, T., Kopan, E. L., Kirkpatrick, J. D., Light, R. M., Marsh, K. A.,
McCallon, H., Schneider, S., Stiening, R., Sykes, M., Weinberg, M., Wheaton,
W. A., Wheelock, S., & Zacarias, N. 2003, 2MASS All Sky Catalog of point sources.
(The IRSA 2MASS All-Sky Point Source Catalog, NASA/IPAC Infrared Science
Archive. http://irsa.ipac.caltech.edu/applications/Gator/)

Davidson, K., Ishibashi, K., & Corcoran, M. F. 1998, New A-l i.i .,iniv, 3, 241

De Becker, M., Rauw, G., Sana, H., Pollock, A. M. T., Pittard, J. M., Blomme, R.,
Stevens, I. R., & van Loo, S. 2006, MNRAS, 371, 1280

Dupree, A. K., Gursky, H., Black, J. H., Davis, R. J., Hartmann, L., Matilsky, T.,
Raymond, J. C., Hammerschlag-Hensberge, G., van den Heuvel, E. P. J., Burger,
M., Lamers, H. J. G. L. M., Vanden Bout, P. A., Morton, D. C., de Loore, C., van
Dessel, E. L., Menzies, J. W., Whitelock, P. A., Watson, M., Sanford, P. W., & Pollard,
G. S. G. 1980, ApJ, 238, 969

Dutra, C. M., Santiago, B. X., Bica, E. L. D., & Barbuy, B. 2003, MNRAS, 338, 253

Eikenberry, S., Blum, R., Bandyopadhyay, R., Muno, M., Sellgren, K., Olsen, K., &
F2GCS Team. 2005, in Bulletin of the American Astronomical Society, 37, 1462

Eikenberry, S. S., Matthews, K., Morgan, E. H., Remillard, R. A., & Nelson, R. W. 1998,
ApJ, 494, L61

Eikenberry, S. S., Matthews, K., Muno, M., Blanco, P. R., Morgan, E. H., & Remillard,
R. A. 2000, ApJ, 532, L33

Evans, N. R., Seward, F. D., Krauss, M. I., Isobe, T., Nichols, J., Schlegel, E. M., & Wolk,
S. J. 2003, ApJ, 589, 509

Fender, R., Belloni, T., & Gallo, E. 2005, Ap&SS, 300, 1









Fender, R. P., Garrington, S. T., McKay, D. J., Muxlow, T. W. B., Pooley, G. G.,
Spencer, R. E., Stirling, A. M., & Waltman, E. B. 1999, MNRAS, 304, 865

Fender, R. P., & Pooley, G. G. 1998, MNRAS, 300, 573

Feroci, M., Matt, G., P...I. i, G., Costa, E., Tavani, M., & Belloni, T. 1999, A&A, 351,
985

Figer, D. F., McLean, I. S., & N i -ro, F. 1997, ApJ, 486, 420

Figer, D. F., N li -ro, F., Morris, M., McLean, I. S., Geballe, T. R., Ghez, A. M., &
Langer, N. 1998, ApJ, 506, 384

Filliatre, P., & ('! I .,, S. 2004, ApJ, 616, 469

Forman, W., Jones, C., Cominsky, L., Julien, P., Murray, S., Peters, G., Tananbaum, H.,
& Giacconi, R. 1978, ApJS, 38, 357

Frank, J., King, A., & Raine, D. J. 2002, Accretion Power in A-i i. .1,,! -i. Third Edition
(Cambridge, UK: Cambridge University Press, February 2002.)

Giacconi, R., Bechtold, J., Branduardi, G., Forman, W., Henry, J. P., Jones, C., K. 11, .--
E., van der Laan, H., Liller, W., Marshall, H., Murray, S. S., Pye, J., Schreier, E.,
Sargent, W. L. W., Seward, F., & Tananbaum, H. 1979, ApJ, 234, L1

Giacconi, R., Gursky, H., Paolini, F. R., & Rossi, B. B. 1962, Physical Review Letters, 9,
439

Giacconi, R., Murray, S., Gursky, H., K. 11 .-.- E., Schreier, E., & Tananbaum, H. 1972,
ApJ, 178, 281

Gies, D. R., & Bolton, C. T. 1986, ApJ, 304, 389

Girardi, L., Bertelli, G., Bressan, A., Chiosi, C., Groenewegen, M. A. T., Marigo, P.,
Salasnich, B., & Weiss, A. 2002, A&A, 391, 195

Gosling, A. J., Bandyopadhyay, R. M., & Blundell, K. M. 2007, in American Institute of
Physics Conference Series, 924, 893

G6oig, E., Patel, S. K., Wilson, C. A., Woods, P. M., Finger, M. H., & Kouveliotou, C.
2005, ApJ, 632, 1069

Greiner, J., Cuby, J. G., McCaughrean, M. J., Castro-Tirado, A. J., & Mennickent, R. E.
2001, A&A, 373, L37

Groote, D., & Hunger, K. 1982, A&A, 116, 64

Groote, D., & Schmitt, J. H. M. M. 2004, A&A, 418, 235

Gursky, H. 1972, ApJ, 175, L141









Hanson, M. M., Conti, P. S., & Rieke, M. J. 1996, ApJS, 107, 281

Heras, A. M., Shipman, R. F., Price, S. D., de Graauw, T., Walker, H. J., Jourdain
de Muizon, M., Kessler, M. F., Prusti, T., Decin, L., Vandenbussche, B., & Waters,
L. B. F. M. 2002, A&A, 394, 539

Hillier, D. J., Davidson, K., Ishibashi, K., & Gull, T. 2001, ApJ, 553, 837

Home, J. H., & Baliunas, S. L. 1986, ApJ, 302, 757

Hyland, A. R., & Mould, J. R. 1973, ApJ, 186, 993

Iaria, R., Di Salvo, T., Robba, N. R., Burderi, L., Lavagetto, G., & Riggio, A. 2005, ApJ,
634, L161

Klein-Wolt, M., Fender, R. P., Pooley, G. G., Belloni, T., Migliari, S., Morgan, E. H., &
van der Klis, M. 2002, MNRAS, 331, 745

Kroupa, P. 2001, MNRAS, 322, 231

Lada, C. J. 2006, ApJ, 640, L63

Laycock, S., Grindlay, J., van den Berg, M., Zhao, P., Hong, J., Koenig, X., Schlegel,
E. M., & Persson, S. E. 2005, ApJ, 634, L53

L.- .-tt, S. K., Allard, F., Berriman, G., Dahn, C. C., & Hauschildt, P. H. 1996, ApJS,
104, 117

Leitherer, C., & Wolf, B. 1984, A&A, 132, 151

L6pine, S., Wallace, D., Shara, M. M., Moffat, A. F. J., & Niemela, V. S. 2001, AJ, 122,
3407

Lewin, W. H. G., & van der Klis, M. 2006, Compact Stellar X-ray Sources (Cambridge,
UK: Cambridge University Press)

Lewin, W. H. G., van Paradijs, J., & van den Heuvel, E. P. J. 1997, X-ray Binaries
(Cambridge, UK: Cambridge University Press)

Liu, Q. Z., van Paradijs, J., & van den Heuvel, E. P. J. 2006, A&A, 455, 1165

Luo, D., McCray, R., & Mac Low, M.-M. 1990, ApJ, 362, 267

Mafz-Apelldniz, J., Walborn, N. R., Galu6, H. A., & Wei, L. H. 2004, ApJS, 151, 103

Margon, B., Anderson, S. F., Aller, L. H., Downes, R. A., & Keyes, C. D. 1984, ApJ, 281,
313

Markwardt, C. B., Swank, J. H., & Taam, R. E. 1999, ApJ, 513, L37

Marti, J., Mirabel, I. F., C(!I .r,, S., & Rodrfguez, L. F. 2000, A&A, 356, 943









Martin, J. C., Davidson, K., Humphreys, R. M., Hillier, D. J., & Ishibashi, K. 2006, ApJ,
640, 474

McNamara, D. H., Madsen, J. B., Barnes, J., & Ericksen, B. F. 2000, PASP, 112, 202

Merloni, A., Fabian, A. C., & Ross, R. R. 2000, MNRAS, 313, 193

Mikles, V. J., Eikenberry, S. S., Muno, M. P., Bandyopadhyay, R. M., & Patel, S. 2006,
ApJ, 651, 408

Mineshige, S., Ouchi, N. B., & Nishimori, H. 1994, PASJ, 46, 97

Mirabel, I. F., Dhawan, V., C'! ily, S., Rodriguez, L. F., Marti, J., Robinson, C. R.,
Swank, J., & Geballe, T. 1998, A&A, 330, L9

Mirabel, I. F., & Rodriguez, L. F. 1994, Nature, 371, 46

Moffat, A. F. J. 1969, A&A, 3, 455

Mokiem, M. R., de Koter, A., Evans, C. J., Puls, J., Smartt, S. J., Crowther, P. A.,
Herrero, A., Langer, N., Lennon, D. J., N ii ,ro, F., Villamariz, M. R., & Vink, J. S.
2007, A&A, 465, 1003

Morel, T., & Grosdidier, Y. 2005, MNRAS, 356, 665

Morris, P. W., Eenens, P. R. J., Hanson, M. M., Conti, P. S., & Blum, R. D. 1996, ApJ,
470, 597

Mouchet, M., Ilovaisky, S. A., & C'h. v .!. r, C. 1980, A&A, 90, 113

Muno, M. P., Ai i .,dij, J. S., Baganoff, F. K., Bautz, M. W., Brandt, W. N., Broos,
P. S., Feigelson, E. D., Garmire, G. P., Morris, M. R., & Ricker, G. R. 2004a, ApJ, 613,
1179

Muno, M. P., Baganoff, F. K., Bautz, M. W., Brandt, W. N., Broos, P. S., Feigelson,
E. D., Garmire, G. P., Morris, M. R., Ricker, G. R., & Townsley, L. K. 2003, ApJ, 589,
225

Muno, M. P., Baganoff, F. K., Bautz, M. W., Feigelson, E. D., Garmire, G. P., Morris,
M. R., Park, S., Ricker, G. R., & Townsley, L. K. 2004b, ApJ, 613, 326

Muno, M. P., Bauer, F. E., Bandyopadhyay, R. M., & Wang, Q. D. 2006a, ApJS, 165, 173

Muno, M. P., Bower, G. C., Burgasser, A. J., Baganoff, F. K., Morris, M. R., & Brandt,
W. N. 2006b, ApJ, 638, 183

Muno, M. P., Morgan, E. H., & Remillard, R. A. 1999, ApJ, 527, 321

N ,-,. F., Corbet, R. H. D., Day, C. S. R., Inoue, H., Takeshima, T., Yoshida, K., &
Mihara, T. 1992, ApJ, 396, 147









Narayan, R., & Yi, I. 1995, ApJ, 452, 710

Narayan, R., Yi, I., & Mahadevan, R. 1995, Nature, 374, 623

Nazi, Y., Rauw, G., Vreux, J.-M., & De Becker, M. 2004, A&A, 417, 667

Negueruela, I., & Schurch, M. P. E. 2007, A&A, 461, 631

Ogilvie, G. I., & Dubus, G. 2001, MNRAS, 320, 485

Predehl, P., & Schmitt, J. H. M. M. 1995, A&A, 293, 889

Ransom, S. M., Eikenberry, S. S., & Middleditch, J. 2002, AJ, 124, 1788

Rayner, J. T., Toomey, D. W., Onaka, P. M., Denault, A. J., Stahlberger, W. E., Vacca,
W. D., Cushing, M. C., & W ,1.- S. 2003, PASP, 115, 362

Reiners, A., Stahl, O., Wolf, B., Kaufer, A., & Rivinius, T. 2000, A&A, 363, 585

Rieke, G. H., & Lebofsky, M. J. 1985, ApJ, 288, 618

Rothstein, D. M., Eikenberry, S. S., & Matthews, K. 2005, ApJ, 626, 991

Sana, H., Stevens, I. R., Gosset, E., Rauw, G., & Vreux, J.-M. 2004, MNRAS, 350, 809

Sanz-Forcada, J., Franciosini, E., & Pallavicini, R. 2004, A&A, 421, 715

Schild, H., Giidel, M., Mewe, R., Schmutz, W., Raassen, A. J. J., Audard, M., Dumm, T.,
van der Hucht, K. A., Leutenegger, M. A., & Skinner, S. L. 2004, A&A, 422, 177

Schultz, A. 2005, NIC \!OS Instrument Handbook, Version 8.0 (Baltimore: STScI)

Schulz, N. S., Canizares, C. R., Lee, J. C., & Sako, M. 2002a, ApJ, 564, L21

Schulz, N. S., Cui, W., Canizares, C. R., Marshall, H. L., Lee, J. C., Miller, J. M., &
Lewin, W. H. G. 2002b, ApJ, 565, 1141

Seward, F. D., Butt, Y. M., Karovska, M., Prestwich, A., Schlegel, E. M., & Corcoran, M.
2001, ApJ, 553, 832

Shakura, N. I., & Sunyaev, R. A. 1976, MNRAS, 175, 613

Sidoli, L., Mereghetti, S., Favata, F., Oosterbroek, T., & Parmar, A. N. 2006, A&A, 456,
287

Skinner, S. L., Giidel, M., Schmutz, W., & Stevens, I. R. 2001, ApJ, 558, L113

Skinner, S. L., Zhekov, S. A., Palla, F., & Barbosa, C. L. D. R. 2005, MNRAS, 361, 191

Steiner, J. E., & Damineli, A. 2004, ApJ, 612, L133

Ti.-.- r, M., & Pellat, R. 1999, A&A, 349, 1003









Ti.-.- r, M., Varniere, P., Rodriguez, J., & Pellat, R. 2004, ApJ, 607, 410

Terrell, D., & Wilson, R. E. 2005, Ap&SS, 296, 221

Timmer, J., & Koenig, M. 1995, A&A, 300, 707

Titarchuk, L., & Fiorito, R. 2004, ApJ, 612, 988

Titarchuk, L., Shaposhnikov, N., & Arefiev, V. 2007, ApJ, 660, 556

Tomsick, J. A., C'! I .i, S., Rodriguez, J., Foschini, L., Walter, R., & Kaaret, P. 2006, ApJ,
647, 1309

Tousey, R., Watanabe, K., & Purcell, J. D. 1951, Physical Review, 83, 792

Townsley, L. K., Broos, P. S., Nousek, J. A., & Garmire, G. P. 2002, Nuclear Instruments
and Methods in Physics Research A, 486, 751

Turner, D. G. 1985, A&A, 144, 241

Usov, V. V. 1992, ApJ, 389, 635

Vacca, W. D., Cushing, M. C., & Rayner, J. T. 2003, PASP, 115, 389

van der Hucht, K. A. 2001, New Astronomy Review, 45, 135

van Genderen, A. M., de Groot, M. J. H., & The, P. S. 1994, A&A, 283, 89

Vanbeveren, D., De Loore, C., & Van Rensbergen, W. 1998, A&A Rev., 9, 63

Varniere, P., Rodriguez, J., & Ti.-.- r, M. 2002, A&A, 387, 497

Varricatt, W. P., Williams, P. M., & Ashok, N. M. 2004, MNRAS, 351, 1307

Viotti, R. F., Antonelli, L. A., Rossi, C., & Rebecchi, S. 2004, A&A, 420, 527

Wallace, L., & Hinkle, K. 1997, ApJS, 111, 445

Wallace, L., M. i,. r, M. R., Hinkle, K., & Edwards, S. 2000, ApJ, 535, 325

Walter, R., Zurita Heras, J., Bassani, L., Bazzano, A., Bodaghee, A., Dean, A., Dubath,
P., Parmar, A. N., Renaud, M., & Ubertini, P. 2006, A&A, 453, 133

Westbrook, 0., Evans, N. R., Wolk, S. J., Nichols, J., K ,i-1, p V. L., Mendygral, P. J., &
Waldron, W. L. 2006, in Bulletin of the American Astronomical Society, 38, 1117

Williams, P. M., van der Hucht, K. A., Sandell, G., & The, P. S. 1990, MNRAS, 244, 101

Wilson, C. A., Weisskopf, M. C., Finger, M. H., Coe, M. J., Greiner, J., R. i. P., &
Papamastorakis, G. 2005, ApJ, 622, 1024

Yadav, J. S. 2006, ApJ, 646, 385









Zickgraf, F.-J., Wolf, B., Leitherer, C., Appenzeller, I., & Stahl, 0. 1986, A&A, 163, 119









BIOGRAPHICAL SKETCH

I was born in Norfolk, Virginia, and raised in Baltimore, Maryland by a musician and

a sound engineer. The fine arts were my first education and I was prone to trying out any

instrument sitting on the coffee table. Never much of a stargazer, my first fascination with

astronomy began with an 8th grade science project researching stars. I was fascinated

by the fact that stars not only live and evolve, but they die and warp space time. Still, I

never considered astronomy as a career until I read a book by Kip Thorne called "Black

Holes and Time Warps." It was then that I decided to pursue a degree in .-1 .!1i, --ics

not only because black holes captivated my imagination, but because the intellectual

community port ,i-,, d in the book was something I wanted to be a part of.

I was fortunate my junior year of high school to have a physics teacher who was

also an astronomer, so I knew physics was vital to the career. I entered Johns Hopkins

University as a physics 1ni, i r and set out immediately to find work as a research assistant,

and a brilliant scientist named Dr. Wei Zheng gave me a chance. I worked for him for

three and a half years and completed my senior thesis under his advisement. During that

time, I participated in two other summer research experiences and sank roots into the

science community.

I came to the University of Florida for graduate school because of all the places

I visited and applied, I loved the community here the most. The graduate students

immediately took me in, teaching me the ropes, and compelling me to succeed. In my

time here, I have sought to make others feel as welcome as I did that first d-,v.

For my master's project, I worked with Vicki Sarajedini a fine mentor and an

excellent role model for women in science. We searched for super-massive black holes

in the core of distant galaxies. Good times. For my PhD thesis, I worked for Steve

Eikenberry. The black holes were smaller, but close enough to see and fiddle with the

physics.





PAGE 4

4

PAGE 5

page ACKNOWLEDGMENTS ................................. 4 LISTOFTABLES ..................................... 8 LISTOFFIGURES .................................... 9 ABSTRACT ........................................ 11 CHAPTER 1INTRODUCTION .................................. 13 1.1TheGalacticX-rayPopulation ........................ 14 1.2Wind-collidingSystems ............................. 14 1.3AccretingBinarySystems ........................... 15 1.3.1Persistentvs.TransientSources .................... 16 1.3.2Multi-wavelengthStudyofX-rayBinarySources ........... 17 1.3.3ModelingAccretionFlowsinX-rayBinarySources .......... 17 1.3.4JetsinX-rayBinarySources ...................... 20 1.3.5VariabilityinX-rayBinarySources .................. 20 1.4ThisWork .................................... 22 2THESEARCHFORCOMPACTOBJECTSINTHEGALACTICCENTER .. 28 2.1Motivation .................................... 28 2.2TargetSelection ................................. 30 2.3ObservationsandAnalysis ........................... 31 2.4Summary .................................... 33 3THEINFRAREDCOUNTERPARTTOCXOGCJ174536.1-285638(EDD-1) 38 3.1Introduction ................................... 38 3.2Analysis ..................................... 38 3.2.1Infrared ................................. 38 3.2.2X-ray ................................... 40 3.3Discussion .................................... 42 3.3.1IsEdd-1anIsolatedStar? ....................... 43 3.3.2IsEdd-1AHigh-MassX-rayBinary? ................. 44 3.3.3IsEdd-1AColliding-WindBinary? .................. 45 3.4Summary .................................... 47 4THEX-RAYPERIODOFCXOGCJ174536.1-285638(EDD-1) ......... 59 4.1Introduction ................................... 59 4.2ObservationsandAnalysis ........................... 59 4.2.1Infrared ................................. 59 5

PAGE 6

................................ 60 4.3Discussion .................................... 62 4.3.1InfraredVariability ........................... 62 4.3.2X-rayVariability ............................ 63 4.3.3TheOrbitalPeriodAssumption .................... 64 4.3.3.1WindObscurationScenario ................. 65 4.3.3.2TheEclipsingBinaryScenario ................ 67 4.3.3.3Edd-1asaWind-AccretingHMXB ............. 68 4.4Summary .................................... 70 4.5ConcludingRemarksAboutEdd-1 ...................... 71 5ANATLASOFKNOWNHIGH-MASSOBJECTSINTHEGALAXY ..... 84 5.1BackgroundandMotivation .......................... 84 5.2Typesofsourcesincluded/criteria ...................... 86 5.3Constructionofthedatabase .......................... 87 5.4StatisticsandParametersintheAtlas .................... 88 5.4.1ObjectNameandGeneralInformation ................ 88 5.4.2InfraredInformation .......................... 88 5.4.3Reddening ................................ 88 5.4.4Distance ................................. 89 5.4.5X-rayInformation ............................ 89 5.4.5.1QuiescentandBurstX-rayLuminosity ........... 89 5.4.5.2X-rayEnergyWaveband ................... 90 5.4.5.3X-rayModel,thermaltemperature,andNH 90 5.4.6SpectralInformation .......................... 90 5.4.7CalculatedLuminosity ......................... 91 5.4.8References ................................ 91 5.5TheWebInterface ............................... 92 5.6ScienceApplications .............................. 93 5.6.1CharacterizinganUnclassiedSourceSpectrum ........... 93 5.6.2ComparingClassesofSources ..................... 93 5.6.3PopulationStudyofaSourceClass .................. 94 5.7Summary .................................... 94 6THEMICROQUASARGRS1915+105 ....................... 99 6.1Introduction ................................... 99 6.2X-rayLightCurvesandQuasiperiodicOscillations .............. 100 6.3InfraredFlaringBehavior ........................... 103 6.4Summary .................................... 104 7THELOWFREQUENCYX-RAYQPOBEHAVIOROFGRS1915+105 .... 112 7.1Introduction ................................... 112 7.2ReningQPODetection ............................ 114 7.2.1QPOFrequency-CorrelationtoSpectralFeatures .......... 115 6

PAGE 7

.......................... 118 7.2.3DieringTriggerSpikeMorphologyin-classLightCurves ..... 119 7.2.4AssociatedInfraredFlaring ....................... 121 7.3Discussion .................................... 122 7.3.1QPOCorrelationwithSpectralFeatures ............... 122 7.3.2InfraredFlaringBehavior ........................ 123 7.3.3ACauseandEectSummary ..................... 124 7.4Summary .................................... 125 8GRS1915+105ANDTHEACCRETION-EJECTIONINSTABILITYMODEL 139 8.1Introduction ................................... 139 8.2X-rayQPOPhenomenologyinthe-,-,and-classes ........... 140 8.3QPOModelsandtheRadiusDilemma .................... 142 8.4TheAccretion-ejectionInstabilityModel ................... 144 8.5Caveats,Conclusions,andFutureWork .................... 147 9CONCLUSIONSANDFUTUREWORK ...................... 160 9.1TheGalacticX-rayPopulation ........................ 160 9.2Microquasars .................................. 162 APPENDIX ABBREVIATIONS .................................. 164 REFERENCES ....................................... 165 BIOGRAPHICALSKETCH ................................ 173 7

PAGE 8

Table page 2-1SpeXObservingLog 37 2-2TableofIRLines 37 3-1IdentiedLines 54 3-2PCygniLineVelocity 55 3-3ChandraSpectrumofEdd-1 56 3-4X-rayandIRSourceComparison .......................... 57 4-1ObservingLog:IRSpectra 78 4-2ObservingLog:Chandra 79 4-3ObservingLog:XMM-Newton 80 4-4Massratioestimationsfortheeclipsingscenario 80 4-5InfraredLineRatios 81 4-6MassratioestimationsforthewindobscurationscenariointhecaseofaHMXB 82 4-7Summaryofscenarios 83 5-1AtlasInformation ................................... 96 6-1RXTEObservations 111 7-1LinearPearsonCorrelationCoecients 135 7-2PartialCorrelationCoecients 136 7-3LengthofFrequencyDip 137 7-4GaussianFitParameters 138 8-1LinearPearsonCorrelationCoecients 156 8-2QPOFrequency-RadiusFits 159 8

PAGE 9

Figure page 1-1Wind-collidingbinarysystem ............................ 24 1-2RocheLobeOverowinabinarysystem ...................... 25 1-3Schematicdrawingofamicroquasar ........................ 26 1-4RadiojetsinthemicroquasarGRS1915+105 ................... 27 2-1K-bandspectraofsurveyedsources. ......................... 35 2-2ExampleofspectralimagetracewithSpeX. .................... 36 3-1K-bandspectrumofEdd-1 .............................. 48 3-2H-bandspectrumofEdd-1 .............................. 49 3-3J-bandspectrumofEdd-1 .............................. 50 3-4PCygniprolesintheK-bandspectrumofEdd-1 ................. 51 3-5X-rayspectrumofEdd-1 ............................... 52 3-6ComparisonoftheX-rayandIRluminosityofEdd-1tootherknownX-raysystemscontainingmassivestars. ............................... 53 4-1K-bandspectraofEdd-1in2005and2006. ..................... 72 4-2BrregionofselectEdd-1spectra. ......................... 73 4-3TworepresentativeXMMspectraofEdd-1separatedby0.7inphase. ...... 74 4-4X-raylight-curveandfoldedlight-curveofEdd-1. ................. 75 4-5PeriodogramanalysisofEdd-1'sX-raylightcurve. ................ 76 4-6Theoreticalmassratioasafunctionofprimarymassforbinarieswitha189dayperiod. ...................................... 77 5-1Screenshotofdatabasestatisticselectionpage. .................. 95 6-1Typical-classlightcurveat1sresolution. .................... 105 6-2Typical-classlightcurveat1sresolution. .................... 106 6-3Typical-classlightcurveat1sresolution. .................... 107 6-4EightmslightcurveofGRS1915+105duringaspectrallyharddipandthepowerspectrumofselect4sregionsofthelightcurve. .............. 108 6-5PDSand1-secondlightcurvesforthe-and-class. ............... 109 9

PAGE 10

................... 110 6-7AsamplelightcurveshowingtheX-ray,theinfrared,andtheQPObehaviorduringajetejection. ................................. 110 7-1Overlayofthe1-secondresolutionX-raylightcurveandne-binned4-secondQPOfrequencyinGR1915+105. .......................... 127 7-2ScatterplotsofQPOfrequencywithpowerlawuxfor-and-classlightcurves. 128 7-3ScatterplotsofQPOfrequencywithblackbodyuxfor-and-classlightcurves. ......................................... 129 7-4ScatterplotsofQPOfrequencywithblackbodytemperaturefor-and-classlightcurves. ...................................... 130 7-5TimeevolutionofaGroup1eventinGRS1915+105. ............... 131 7-6TimeevolutionofaGroup2eventinGRS1915+105. ............... 132 7-7TimeevolutionofaGroup3eventinGRS1915+105. ............... 133 7-8Triggerspikeofthe-classlightcurvesinGRS1915+105. ............ 134 8-1CartoonmodelofjetejectioninanX-raybinary. ................. 150 8-2QPOfrequencyvs.totaluxfor-,-,and-classesabove4Hz. ........ 151 8-3QPOfrequencyvs.powerlawuxfor-,-,and-classesabove4Hz. ..... 152 8-4QPOfrequencyvs.blackbodyuxfor-,-,and-classesabove4Hz. ..... 153 8-5QPOfrequencyvs.blackbodytemperaturefor-,-,and-classesabove4Hz. 154 8-6QPOfrequencyvs.innerdiskradiusfor-,-,and-classesabove4Hz. .... 155 8-7Observations:Colorradiusvs.QPOfrequency. .................. 156 8-8Theory:Radiusvs.QPOfrequency. ......................... 157 8-9Radiustsforthe-2event. ............................. 158 10

PAGE 11

11

PAGE 12

12

PAGE 13

Touseyetal. 1951 ).TheSun'senergysignaturewasfoundtobesoweak,itwasassumedthatnoothercelestialobjectwouldbevisibletousintheX-rays.Despitetheimplausibilityofndingadditionalcelestialsources,asmallX-raydetectorwaslaunchedin1962,andScoX-1becametherstknownnon-solarcelestialX-raysource( Giacconietal. 1962 ).Atthetime,nooneguessedatthewealthofextremeenvironmentsandsourcesradiatingX-raysandllingthesky.UsingX-rayuxvariabilityfromidentiedpointsourcestoconstrainthesizeofanX-rayemittingregionanddynamicalconstraintsfromobservingcounterpartsatopticalandinfrared(IR)wavelengths,ledeventuallytotheidenticationoftherstcompactobjectsand,moreimportantly,conrmationoftheexistenceofblackholes.TheidenticationofopticalandIRcounterpartstothesehighenergyX-raysourcesallowsustotargetandstudysystemsexhibitingtheextremephysicalprocessesthatproduceX-rays,inthecontextoftheirlocalenvironments.Inthepastdecade,ourknowledgeofX-rayemittingsourceshasadvancedsignicantlyduetotheobservationsofspace-basedX-rayobservatoriessuchastheRossiX-rayTimingExplorer(RXTE),theChandraX-rayObservatory,andXMM-Newton.Theunprecedentedangularresolution,energyresolution,andtimingresolutionoftheseinstrumentsletusstudyheretoforeunobservablepropertiesofX-raysources,includingtheshockfrontproducedinawind-collisionregionbetweentwomassivestars(x1.2)andrapidvariabilityphenomenathatoccurnearthesurfaceofacompactobject(x1.3.5).Inaddition,multi-wavelengthobservationsofthesesourcesfromradiotogamma-rayshaveresultedinthediscoveryofrelativisticjetsinaccretingstellar-masssources(x1.3.4).TheincredibleinuxofX-raydatahasledtosignicantadvancesintheoreticalmodelingofbothwind-collidingsystemsandaccretingbinarysystems. 13

PAGE 14

Kroupa 2001 ; Lada 2006 ).Albeitrare,massivestarsgreatlyaecttheGalacticenvironmentandthechemicalevolutionofourUniverse.Thegalacticandintergalacticmediumisenrichedbytheproductsofthenucleosynthesisthatoccursinthecoresofmassivestars.Theseproductsaredistributedtothecircumstellarenvironmentbywind-drivenmassloss,andbynovaeorsupernovaeevents.ThestudyofthestellarX-raysourcepopulationinourGalaxyhasincreasedourknowledgeoftheevolutionofthesesourcesandhowtheycontributetotheGalaxy'sevolution. Corcoranetal. 2006 ).ThehighresolutionandsensitivityofChandraandXMM-Newton,aswellasthemonitoringcapabilitiesofRXTE,haveallowedadetailedexaminationofsomeofthebrighter(LX1036erg=s)CWBsandidenticationofthefainterCWBpopulationdowntoLX1032erg=s( Munoetal. 2006a ).Inthesesources,X-rayemissionprimarilyarisesintheshockfrontwherethewindsfromthestarscollide(see,e.g. Luoetal. 1990 ; Sanaetal. 2004 ; DeBeckeretal. 2006 ).Thephysicsoftheheatingprocessintheshocklayerispoorlyunderstood,butithasbeenshownthatthethermalX-rayluminosityscalesas_M2( Luoetal. 1990 ).Theforcebalancebetweenthetwocontributingwindsdetermineswhichwinddominates.InFigure 1-1 ,IshowanexampleofacollidingwindsystemcontainingaWolf-RayetstarandanO-typecompanionasderivedmathematically 14

PAGE 15

Usov ( 1992 ).TheWolf-RayetstarcontributesthedominantwindbecausethetypicalmasslossrateofaWolf-RayetisexpectedtobeanorderofmagnitudehigherthanthatofanO-typestar( Usov 1992 ).TheshockregionthusoccursmuchclosertotheO-typestar.TheX-rayspectralenergydistributionofaCWBisthatofacollisionallyionizedplasma,withelectronsacceleratedbyshockheating( Luoetal. 1990 ).TheluminosityandabsorbingcolumnoftheplasmaaremeasuredbyttingmodelstotheX-rayspectrum( Sanaetal. 2004 ).Boththeluminosityandtheabsorptioncolumnareknowntovarywithtime,oftenasafunctionoftheorbitalphase( Luoetal. 1990 ).Windparameterssuchasthemasslossratesandwindvelocitiescanbecharacterizedbyradiointensityandultravioletlineproperties.Stellarparameterssuchasmass,temperature,radii,androtationalvelocitycanoftenbeestimatedviaspectraltypingintheopticalwavelengths( Corcoranetal. 2006 ).TheemissionfromCWBsprovidesanextremelyusefulprobeofthenatureoftheshockandisvaluableforinvestigatingtheunderlyingphysicsofparticleaccelerationinconditionstooextremetobereproducedinalaboratory.Additionally,wecaninferfromCWBsthenatureofthemasslossprocessesinmassivestars,whichcanthenleadtobetterunderstandingoftheevolutionandendproductsofmassivestars. LewinandvanderKlis 2006 ).Therearetwocommonmodesofmassaccretion:RocheLobeOverowandWind-fedAccretion.IntheexampleofRocheLobeOverow,massisaccretedontoacompactobjectwhenthecompanion(i.e.,thelessevolvedstar)llsitsRocheLobeandmassthentransfersthroughtheinnerLagrangianpointasshowninFigure 1-2 .The 15

PAGE 16

LewinandvanderKlis 2006 ).Thisiscalledalow-massX-raybinary(LMXB).Duetothelow-massnatureofthecompanion,theopticallightofLMXBsisduemostlytoreprocessingoftheX-rayuxfromtheouteraccretionowsandLMXBsaresignicantlybrighterintheX-raythanintheopticalwavelengths( LewinandvanderKlis 2006 ).AcompactobjectcanalsoaccretematterfromacompanionstarthatdoesnotllitsRocheLobeifthecompanionstarislosingmassintheformofstellarwind.Wind-fedaccretiondependsonstrongstellarwinds(_M106M=yr),suchasthosepresentinthemassiveO-typestarsdiscussedabove.Systemswithmassivecompanionstarsarecalledhigh-massX-raybinaries(HMXBs).TheopticalluminosityinthesesystemsisdominatedbythecompanionstarandtheX-rayluminositydependsontherateofmasstransfer( Franketal. 2002 ).IwillelaborateuponandcalculatetheX-rayluminosityresultantfromawind-fedaccretionsysteminChapter4,x4.3.2.2. LewinandvanderKlis 2006 ).Mosttransientsourcesdisplayoutburstslastingfromafewweekstomonths,separatedbylongperiodsofinactivitylastingmonthstodecades.Duringanoutburstinatransientsource,theluminositychangesbyseveralordersofmagnitude.Thelengthoftheoutburstisconsistentwithviscoustimescalesofaccretiondisks.LessthanhalfofknownX-raysourcesaretransient( LewinandvanderKlis 2006 ). 16

PAGE 17

Lewinetal. 1997 ).IfthesourceisX-raytransient,wecanusetimesof\quiescence"(atimeoflowX-rayluminosity,presumablycausedbyreducedaccretionrate)tolearnmoreaboutthecompanionstar.Intimesofquiescence,theopticalspectrumisdominatedbythecompanionstar,whichallowsustomoreaccuratelyspectraltypethecompanionanddynamicallymeasurethemassandbinaryorbitofthesystem.DetailedstudiesofXRBsatIR,optical,andUVwavelengthsarealsocrucialinordertomeasurethetemperatureproles,ionizationfractions,andabundancesofelementsintheaccretiondisk( LewinandvanderKlis 2006 ). 17

PAGE 18

LewinandvanderKlis 2006 ).RegardlessofthemethodofmasstransferinanXRB,themotionofaparticleinanaccretiondiskisdeterminedbytheamountofangularmomentumitpossesses,thephysicalprocessesbywhichitlosesangularmomentum,andtheradiationprocessesbywhichitcools.Inordertoobtainadisksolution,wesimultaneouslysolvefourconservationequations:mass,angularmomentum,verticalmomentum,andenergy.Ifweassumetheaccretiondiskisgeometricallythin,thenthereisnoverticalmotioninthediskandtheparticlesareinhydrostaticequilibriuminthatdimension.Thestandardmodeldevelopedby ShakuraandSunyaev ( 1976 )assumesthatangularmomentumislostathighratesduetoanunknownprocessthatisproportionaltothepressure.Thismodelalsoassumesthatradiationprocessesareecientmeaningthataccretionowisrelativelycool.Theresultisthattheaccretiondiskisgeometricallythin(thethicknessismuchlessthanradius).Ultimately,diskstructureiscontrolledbythemassofthesystemM,theaccretionrate_M,theradiusR,theviscosity,andthepressureratio( ShakuraandSunyaev 1976 ).Oncedened,mostotherparametersoftheaccretinggas(suchastheluminosityofthedisk,thedensityoftheregion,andthespeedandtemperatureoftheparticles)canbecalculated.Moresophistocatedmodelshavesincedevelopedbecauseoftheidenticationandmodelingofthemagneto-rotationalinstability(MRI),whichhasaectedourunderstandingofhowturbulencearisesandtransportsangularmomentuminastrophysicalaccretiondisks( BalbusandHawley 1991 1998 ).IntheMRImodel,aseedmagneticeldintheowstartingwithinnitesimalstrengthgetsenhancedandtangled.Theresultingturbulenceisanecientmechanismforangularmomentumtransport. 18

PAGE 19

NarayanandYi 1995 ).Advectiondominatedaccretionows(ADAFs)werelatershowntobecapableoflaunchingpowerfuloutows( BlandfordandBegelman 1999 ).ThestudyofADAFsbeganwhenindependentmeasurementsofthemassandaccretionrateintheseunder-luminousaccretingsystemspredictedluminositiesmuchbrighterthanobserved( Narayanetal. 1995 ).Theapparentradiativeineciencyoftheseobjectsintroducedanenergyimbalanceinthestandardaccretionmodel.Advectionwassuggestedasameanstodissipatetheexcessheat.Advectivecoolingassumesaradiativelyinecientenvironmentwhereenergyisstoredinaccretinggasasentropy( NarayanandYi 1995 ).Thesystemishotbecauseitisnotradiating,anditisgeometricallythickbecauseitishot.Inablackholesystem,advectionresultsincoolingbecauseoncethegasisaccretedacrosstheeventhorizon,theenergyitcarriesislostfromthesystem.IntraditionalADAFmodels,thegashasapositiveBernoulliparameterandisgravitationallyunbound,meaningthatspontaneouswindscandevelopandproduceoutows.TheBernoulliparameterBisthesumofthekineticenergy,thepotentialenergy,andtheenthalpyoftheaccretinggas.Itmeasurestheliklihoodofspontaneouswindsandoutows.The\AdvectionDominatedInow/OutowSolutions"orADIOSwasdevelopedby BlandfordandBegelman ( 1999 ),whocreatedaself-consistentsolutionbyaddingapowerfulwindwhichtransportsawaymass,energy,andangularmomentuminsuchawayastokeeptheBernoullinumberlocallynegativeandthusallowmasstoaccrete. 19

PAGE 20

LewinandvanderKlis 2006 ).Radioobservationsshowlargeuxratiosbetweentheapproachingandrecedingsidesofthejetsasexpectedinrelativisticows( MirabelandRodriguez 1994 ).Thepropagationspeedsofjetscanbemeasuredonimagesbyfollowingthekinematicsofindividualradioknots.XRBsthathaverelativisticjetsarecalledmicroquasars.InFigure 1-3 ,Ishowaschematicdrawingofamicroquasar.In1979, Margonetal. ( 1984 )observedtherstmicroquasar,SS433,ahigh-energysourcewithprecessing,mildlyrelativisticjetsejectingmaterialatavelocityof0.26c.In1992,theWATCH/GRANATsatellitedetectedthemicroquasarGRS1915+105,whichhadhighlyrelativisticjets,measuredatavelocityof0.92c.TheradioimageoftheejectaisshowninFigure 1-4 ( MirabelandRodriguez 1994 ).Whilejetejectionstakedecadestoobserveinextragalacticquasars,microquasarsshowusasimilareventinamatterofhoursordays.Thus,microquasarsareanexcellentlaboratorythroughwhichwestudyaccretionphysics,relativisticjetformation,anddisk-jetinteractions.Becausedisk-jetinteractionsoccurclosetothecentralcompactobject,understandingtheinteractionwillhelpusprobethelimitsofGeneralRelativityandanswerthequestionofwhathappensattheedgeofablackhole. 20

PAGE 21

LewinandvanderKlis 2006 ).Forexample,a1.4Mneutronstarhasadynamicaltimescale(r3=GM)1=2of0:1msnearthesurfaceofthecompactobject(15km),and2msatadistanceof100km( LewinandvanderKlis 2006 ).A14Mblackhole,likeGRS1915+105( Greineretal. 2001 ),hasadynamicaltimescaleof1msatadistanceof3RS.Variabilitycanbeusedtoprobeaccretionowdynamics.Fora14Mblackhole,mostofthegravitationalenergyisreleasedwithintheinner100kmofthesystem,whichischaracterizedbymilliseconddynamicaltimescales( LewinandvanderKlis 2006 ).Sincethetemperatureoftheinneraccretiondiskishigh(>107K),mostoftheemissionisexpectedtobeintheX-rays.FactorsaectingtheX-rayluminosityofanaccretionowincludethemass-transferratefromthecompaniontothecompactobject,theclumpinessoftheaccretedmaterial,theinwarddiusionofmatterintheaccretionow,andtheinteractionoftheaccretionowwiththecentralobject.PerhapsmosttellingabouttheinnerworkingsoftheaccretiondiskandthelaunchingofoutowsareX-rayquasi-periodicoscillations(QPOs).QPOshavehighlyvariablefrequenciesandmaylosecoherenceaftertensorhundredsofseconds( LewinandvanderKlis 2006 ).TheobservedQPOfrequencies(fromafewhertztotensofkilohertz)areattributedtothecharacteristicdynamicandhydrodynamicfrequenciesattheinnerpartofthedisk.Therapidvariabilitypropertiesofaccretionowsprovideusefulprobesintothephysicalconditionsclosetocompactobjectsandthegeometryofaccretionows.Lowfrequency(2-10Hz)QPOsarecommonlyobservedinXRBs,buttheeorttotiethemtotheaccretion-owgeometryiscomplicatedbecausetheyaremuchlowerthanthedynamicalfrequenciesassociatedwithparticlesorbitingintheinneraccretiondisk.ThephysicalmechanismthatgeneratestheselowfrequencyQPOsisnotknown,though 21

PAGE 22

LewinandvanderKlis 2006 ).InChapter8,IdiscusstheAccretionEjectionInstability(AEI)modelwhichcombinesmagneticinstabilityandKeplerianmotiontoexplaintheobservedlowfrequencyQPOs.TheQPObehaviorandtheAEImodelarediscussedextensivelyinChapters6-8. Giacconietal. 1972 ; Gursky 1972 ; Formanetal. 1978 ).IdentifyingandstudyingthepopulationofcompactobjectsintheGCgivesimportantinsightsintothemassivestarformationhistoryofthisregion.Withthesub-arcsecondresolutionofmodernIRandX-rayinstruments,Iamabletopenetratetheextinguishingveiloftheinterstellarmediumandexplorethisexcitingregion.IntheIRspectroscopicsurvey,Iseekemissionsignatures,particularlyintheBrlineoftheIRspectrum,whichareindicativeofhighenergyprocesses.InChapter2,Idetailthenatureofmysurvey.InChapters3and4,Idiscussfollow-upobservationsofastrong 22

PAGE 23

23

PAGE 24

ExampleofawindcollidingbinarysystemcontainingamassiveO-typestarandaWolf-Rayet.S1andS2aretheshockwavesandCisthecontactsurface.Theregionofstellarwindcollisionishatched.DistheorbitalseparationandrWRandrOBarethedistancesofthecontactsurfacetotheWolf-RayetandO-starrespectively.SincethewindfromtheWolf-Rayetisstronger,thecontactsurfaceisclosertotheO-typestar.From Usov ( 1992 ). 24

PAGE 25

TheRocheLobeofastarisheregionsurroundingastarinabinarysysteminsideofwhichthestar'smaterialisgravitationallyboundtothestar.Ifthestar'senvelopellstheRocheLobe,mattercanbetransferredtoitscompanionstar.ThisisacommonmethodofaccretioninLow-MassX-rayBinaries. 25

PAGE 26

Schematicdrawingofamicroquasarshowingthecompactobject,accretiondisk,companionstar,andrelativisticjets.Thecompanionstardominatestheopticalandinfraredemission,theaccretiondiskdominatestheX-rayemission,andtheendpointsoftherelativisticjetsdominateradioemission. 26

PAGE 27

JetprogressionimagesmadefromofaseriesofVLAradioimagesofthemicroquasarGRS1915+105takenatwavelengthof3.5cm,depictingtherapidejectionofjetsoverthespanofamonth(March27toApril30,1994).From MirabelandRodriguez ( 1994 ). 27

PAGE 28

Giacconietal. 1979 ).TheGalacticCenter(GC)regionhasbeenwell-knownasahomeforX-rayemittingcompactobjectssincethebeginningofsuchstudies( Giacconietal. 1972 ).IdentifyingandstudyingthepopulationofcompactobjectsintheGCgivesimportantinsightsintothemassivestarformationhistoryoftheregionaswellasprovidingindividualtargetsformulti-wavelengthfollow-up.Despitethis,thecrowdedstellarelds,highvisualextinction,andpoorimageresolutionofpreviousX-raymissionshavepreventedidenticationofmanyopticalandinfraredcounterpartstothesecompactobjects.SteadyimprovementsindetectortechnologyhaveledtobetterresolutionandmoredetailedanalysesofX-rayobjects.TheChandraX-rayObservatoryhastakenarevolutionaryleapoveritspredecessorswithangularresolutioncomparabletoopticaltelescopes(1arcsec).ThisallowsustoidentifyX-raysourcesandmatchthemtocounterpartsatotherwavelengthswithunprecedentedprecision.X-raybinaries(XRBs),whichgenerallyappearindistinctfromstarsinopticalandinfrared(IR)surveys,shinebrightlyintheX-ray,makingChandraanidealinstrumentfortargetingandstudyingXRBsatmultiplewavelengths.Theadventofthehigh-resolutioncapabilityofChan-dra,coupledwithsensitiveIRimagersandspectrographscapableofpenetratingtheextinguishingveiloftheinterstellarmedium,oersanewopportunitytoexploretheGCregion. Munoetal. ( 2003 )presentedacatalogof2357X-raypointsourcesina17'x17'eldaroundtheGC,completetoLX1031ergs1.About80%ofthesesourcesaredetectedat<3inthehard2-8keVenergyrangeandarevirtuallyinvisiblebelow2.5keV( Laycocketal. 2005 ).Thespectraarecutoatthelowerenergiesduetohighinterstellarextinction( Laycocketal. 2005 ; NegueruelaandSchurch 2007 ).The2-8keV 28

PAGE 29

Laycocketal. 2005 ).However,sourcessuchasCataclysmicVariables(CVs),pulsars,CollidingWindBinaries(CWBs),andXRBsareexpectedtolieintheluminosityrangeoftheGCsources( Munoetal. 2003 ; Miklesetal. 2006 ).WithanaverageopticalextinctionAV25mag,searchesforcounterpartstotheX-raysourcesintheGCregionarelimitedtoIRwavelengths.Unfortunately,theinfraredskyissignicantlymorecrowdedthantheX-rayskyanddominatedbynon-X-rayemittingsources.The2MASSsurveycatalogsover12,000sourcesinthe0:3o0:3oregioncoveredbythe Munoetal. ( 2003 )X-raysurvey( Cutrietal. 2003 ).Ananalysisby Dutraetal. ( 2003 )quotesanaverageof60,000starsperr=1oeldwith8mag
PAGE 30

Bandyopadhyayetal. 2005 ; Laycocketal. 2005 ; Miklesetal. 2006 ).SuchsurveysforcounterpartsarefurtherfueledbythelaunchofobservatoriessuchasINTEGRALandSwift.Sinceitslaunch,INTEGRALhasfoundoridentiedmoresupergiantXRBsthanwerepreviouslyknown( Walteretal. 2006 ).Thisdiscoverychallengescurrentbinarystarpopulationsynthesismodels,makingcounterpartdiscoveriesanimportantmatter( NegueruelaandSchurch 2007 ). Bandyopadhyayetal. ( 2005 )performedaspectroscopicsurveyand,followingtheirinitialwork,analyzedtheeectivenessofndingcounterpartsusingtheirtechnique.Thesuccessrateislowowingtothenumberofchancecoincidencespossibleincrowdedregionsandlarge(0:5arcsec)astrometricerrors( Bandyopadhyay 2008 ).Theremayalsobephysicalconditionsresponsiblefortheabsenceofanaccretionsignature,suchasthesystembeinginaquiescentphase.Whilesomeofthesephysicallimitationscanbeovercomebydeeperormulti-epochimaging,single-look(orsnapshot)spectroscopicsurveysremainthemostecientmethodforidentifyingcounterpartsandeliminatingfalsematches. Munoetal. ( 2004b )identify400serendipitousX-raysourceswithLX[D=8kpc]210311035ergs1within10-arcminofSgrA*,or24pcoftheGCwith<1-arcsecpositionalaccuracy.Patelcross-correlatedtheseChandraobservationswiththe2MASScatalog,identifyingpossibleIRcounterpartstotheX-raysources,andin2004beganacampaigntolocateandidentifycompactobjectsintheGCviaIRspectroscopy.TheX-ray/IRanalysisheemploystoidentifycandidatecompactobjectsisdesignedtoselectsourcesconsistentwithreddenedaccretingcompactobjectssuchascataclysmicvariables,relativisticjetsources,backgroundquasars,andblackholebinaries.Takingtheastrometrically 30

PAGE 31

31

PAGE 32

Rayneretal. 2003 ).Thisresultedinatotaloftentargetspectra.Table 2-1 showstheobservinglogforthesesources.Noddingalongthelengthoftheslit,Iobtainedsix120sexposuresforatotalexposuretimeof720s.Intheshort-wavelength,cross-dispersedmode,IattainedaresolutionofR1200overtheJHKbandpass.TheresolutionestimationisconrmedbymeasurementsofOHskylines.TargetobservationswerefollowedbyobservationsoftheG0VstarHR6836atsimilarairmassforremovaloftelluricabsorptionfeatures.UsingthestandardSpeXmacrocal sxd 0.5,Iobtainedateldsandwavelengthcalibration.IextractedspectrausingthestandardSpeXToolprocedureforABnoddeddata,resultinginaseriesofsky-subtracted,wavelength-calibratedspectra(describedbelow; Vaccaetal. 2003 ; Cushingetal. 2004 ).ExtractingapointsourceusingSpeXToolishighlyautomated.Tosetupthegeneralprocedure,onemustrstsetupasampleofasingleABpair.Inputtingtheappropriatecalibrationframe,SpeXToolwillloadasubtractedimageandthenallowtheusertoselectthelocationofaspatialproleofthespectrumonthedetector.Inordertoextractspectraforthewavebandsseparately,thespatialprolesmustbeselectedseparately.AlthoughSpeXToolwillattempttodothisautomatically,oneisabletoadjusttheaperturepositionandmanuallydeneanextractionregionifdesired.Thisfeatureishelpfulifobservingmultiplestarsontheslitduringasingleobservation,asIdidforourstars1and2.InFigure 2-2 ,IshowtheABsubtractedimageofstars1and2fromoursample.TheHandKbandareclearlyvisibleforbothsourcesastheparallellightanddarklinesontheimage.Ifthesourceorbandisnotbright(asinthecaseoftheJ-bandspectraofthesestars),thentheautomatedselectionmayfail.Manualplacementoftheapertureallowsonetoattempttoextractthesespectra.IobtainedJ,H,andK-bandspectraforbothsourcesinthisimage.Oncethespectrawereextracted,Imedian-combinedthemusingxcombspec.Iloadedaseriesofsourcelesintotheprogramandviewedthemasagroupbeforewritingthe 32

PAGE 33

RiekeandLebofsky ( 1985 ).TheresultantspectraareplottedinFigure 2-1 .Amajorityappeartoberedstars,oftypeKorcoolerasisapparentbytheCOabsorptionbandsat>2:1m( WallaceandHinkle 1997 ).InTable 2-2 ,IlistthewavelengthcentersandidenticationoflinesIobserveinthespectra.IalsoseeevidenceofCaItripletabsorptionandthesodiumdoublet,commontothesetypesofsources( Leggettetal. 1996 ).Source3appearstobeabluestar(typeForhotter; WallaceandHinkle 1997 ),andisrelativelyfeaturelessintheinfrared.Onlysource1hasstrongBremission.Idescribetheextendedanalysisofsource1(nicknamedEdd-1)inthenexttwochapters. 33

PAGE 34

Munoetal. ( 2006a )suggestthatalargefractionoflowluminosityChandrasources(upto90%)maybeCVsorLMXBs.(LMXBshaveLX=LK>>1.)IfthesourcestargetedinthesurveyareLMXBs,theirIRcounterpartsmaybemorethanvemagnitudesfainterthanthetargetrange{beyondtheconfusionlimitof2MASSandtoofainttobepracticallyobservedwithIRTF.Inthatinstance,futuresurveysliketheFlamingos-2GalacticCenterSurvey( Eikenberryetal. 2005 )wouldbenecessarytoresolvethenatureofthosesources. 34

PAGE 35

Spectraobtainedfromthe2005July01-03observingrun.TheobservinglogisdetailedinTable 2-1 .ThelocationofBr(2:166m)ismarkedwithaverticalline.OthervisiblelinesarelistedinTable 2-2 35

PAGE 36

Thecalibrated,subtractedspectralimagetraceforobjects1and2,obtainedsimultaneouslywithSpeX.TheK-bandspectrumofobject1isoutlinedingreen. 36

PAGE 37

TargetIDDateTimeExp.TimeMag.R.A.Dec.(UT)(min) 12005-07-0111:17810.3317h45m36.139s-28d56'38.47"22005-07-0111:17832005-07-0112:02811.1217h45m52.142s-29d07'43.92"42005-07-0210:20810.6317h45m17.359s-29d06'25.03"52005-07-0210:56811.6417h46m9.953s-29d03'21.64"62005-07-0211:30811.0417h45m8.532s-28d56'51.55"72005-07-0212:01811.3317h45m56.974s-28d55'19.71"82005-07-0310:481211.8517h45m35.002s-29d07'51.29"92005-07-0311:301211.9717h45m49.214s-29d01'37.71"102005-07-0312:151211.1717h46m8.393s-29d06'23.81" Note.|AllmagnitudesareKSfrom2MASS. Table2-2. LineWavelength(m) Note.|SelectK-bandIRlinesidentiedinsourcespectra.Wavecenterstakenfrom Bandyopadhyayetal. ( 1997 ). 37

PAGE 38

Munoetal. 2004b ).Usingthecriteriadiscussedinthepreviouschapter,IidentifyCXOGCJ174536.1-285638(theBremissionsourcenotedinChapter2;hereafter,Edd-1)asapotentialcompactobject.Inthischapter,IdiscussthedenitiveidenticationofEdd-1asanIRcounterparttooneofthesenewChandrasources. 3.2.1InfraredAccordingtothe2MASScatalog,Edd-1hasIRmagnitudesof:J=15:560:08mag,H=12:110:06mag,andKS=10:330:07mag.On2005July1UT,Iobtained1:12:4mspectraofEdd-1usingSpeXonIRTF( Rayneretal. 2003 ).IdiscussthedetailsofmyobservationandinitialanalysisinChapter2,butIemployastrictermethodherefordeterminingthereddeningofthesource.ToestimatethereddeningtowardEdd-1,IassumeaGCdistanceof8kpc( McNamaraetal. 2000 ).IestimatetheinfraredextinctiontowardEdd-1intwoways.First,Iassumea Cardellietal. ( 1989 )reddeninglawandanintrinsic(HK)0=0forahotstar/disk,whichgivesavalueofAV=29mag.Asasecondmethod,IestimatethereddeningbyttingtheK-bandspectralcontinuumtotheRayleigh-JeanstailofaT>104Kblackbody.ThetcorrespondstoAV=33mag.WhilebothvaluesaretypicaloftheGC,IadoptthemoreconservativevalueofAV= 38

PAGE 39

3-1 3-2 ,and 3-3 showthespectrafortheK,H,andJbandsrespectively,dereddenedbyAV=29mag.Thespectraaredominatedbystronghydrogenemissionlines,includingPaschen-,Brackett-,andBrackettserieslinesBr10-Br14.TheBr13lineisnotdistinguishableinmyspectrum.IobservetwoneutralHeliumlines(1.701and2.113m)andsixHeIItransitions(1.163,1.736,1.772,2.189,2.038,and2.348m).TheHeII1.736isblendedwiththeBr10line.IntheK-bandIalsoobservemetallinesfromCIIIandNIII,consistentwithanaccretionsignatureoracollidingwindsystem.ItaGaussianfunctiontothelineproletodeterminethelinecentersandFWHM.IestimatethespectralresolutionoftheinstrumentbymeasuringthewidthofOHskylinesandcorrectmymeasuredlinewidthsaccordingly.Ipresentthelinecenters,equivalentwidths,andfull-widthvelocityinTable 3-1 .Mostoftheemissionlinesarebroadwithafull-widthvelocityabove300km/s.TheBrlineisstrongestandhasafull-widthvelocityof710km/s.Giventhespectralresolutionhere,itisnottrivialtodeconvolvetheBrlinefromneighboringHeIemissionat2:1622:166m.DetailedmodelingisrequiredtoaccuratelyassesstheBrlineequivalentwidthindependentlyoftheHeIcontributionandsuchanalysisisbeyondthescopeofthiswork.Both Morrisetal. ( 1996 )and Hansonetal. ( 1996 )studytheIRspectraofmassivestarswithoutquantifyingtherelativecontributions. Morrisetal. ( 1996 )notethattheapparentasymmetryoftheBrlineislikelycausedbyHeIcontribution,butsaysthecontributionisrelativelyweakinthecontextofLBVandOfpe/WN9stars.SincemydeterminationofspectralclassicationisbasedonbroadX-rayandIRspectralfeaturesandnotonspeciclineratios,itisunlikelythatthecompositenatureofthelineaectsmyresultshere.ThreeHeIIlinesinEdd-1showPCygniprolesat2.034m,2.189m,and2.348m(Fig. 3-4 ,Table 3-2 ).Icalculate 39

PAGE 40

( 2004a )examinedthespectrumandvariabilityofthe2-8keVX-rayemissionfromEdd-1aspartofastudyof2000X-raysourcesdetectedtowardtheGalacticCenter.Here,IsummarizethepropertiesoftheX-raysourcederivedfromthatstudy,basedon626ksofChandraobservationstakenbetween1999Septemberand2002June.TheinitialX-rayanalysiswasperformedbymycollaboratorMichaelP.Munoandisdescribedindetailin Munoetal. ( 2004b ).First,thepulseheightsofeacheventwerecorrectedtoaccountforposition-dependentcharge-transferineciency(CTI; Townsleyetal. 2002 ),andthelistswerecleanedusingstandardtoolsinCIAOversion3.2toremovethosethatdidnotpassthestandardASCAgradelters,thatdidnotfallwithinthegoodtimeintervalsdenedbytheChandraX-raycenter,orthatoccurredduringintervalswhenbackgroundratearedto3abovethemeanlevel.Countswerethenextractedfromacontourenclosing90%ofthepointspreadfunctionaroundEdd-1,andbinnedthemasafunctionoftimetocreatelightcurves,andasafunctionofenergytocreatespectra.Thebackgroundwasestimatedfromanannularregionsurroundingthesource.Foreachobservation,heobtainedtheinstrumentalresponsefunctionsfrom Townsleyetal. ( 2002 )andcomputedeectiveareafunctionsusingtheCIAOtoolmkarf,andaveragedthese,weightedbythenumbersofcountsineachobservation.TheX-rayemissioninEdd-1variedinintensitybyafactorof3inthe2-8keVrange.Thethreeobservationsinwhichthesourcewasfaintlastedonly30ksintotal,whichisasmallfractionofthetotalexposure.Thesedidnotprovideenoughsignaltotestwhetherthespectrumvariedalongwiththeux,soonlytheaveragespectrumisexaminedindetail.TheaverageX-rayspectrumofEdd-1isdisplayedinFigure 3-5 .Themostprominentfeatureislineemissioncenteredat6.7keVfromthen=2{1transitionofHe-likeFewith 40

PAGE 41

Arnaud 1996 )whoseemissionhasbeenabsorbedbyinterstellargasandscatteredbyinterstellardust.ThespectrumisconsistentwithakT=2:0+0:50:2keVthermalplasmaabsorbedbyacolumndensityofNH=3:5+0:30:41022cm2.Despitetheprominenceofthelineemission,theabundancesofS,Ar,Ca,andFeareconsistentwiththesolarvalues(seeTable 3-3 ).However,theabsorptioncolumncorrespondstoareddeningofonlyAV=20mag,whichismuchlowerthanthatinferredfromtheIRspectrum,AV=(2933).AlthoughitiscommonfortheabsorptioncolumninferredfromtheX-rayspectrumtobelargerthanwouldbederivedfromtheIRspectrum(e.g.,iftheX-raysareproducedbyaneutronstarembeddedinthewindofitscompanion),thereisnoknownphysicalsituationinwhichonewouldexpectthecolumnofmaterialabsorbingtheX-raystobesmaller.Instead,themodelfortheX-rayprobablyunder-estimatestheamountofuxproducedbelow2keVbyEdd-1,whichmeansthattheabsorptionisunder-estimatedaswell.Therefore,mycollaboratorhasaddedasecond,coolerplasmacomponenttothemodeloftheX-rayspectrum,andxedtheextinctiontowardtheX-raysourceatNH=5:21022cm2(from PredehlandSchmitt 1995 ,forAV=29).ThespectrumcanbeadequatelymodeledwithtwoplasmacomponentsoftemperatureskTs=0:70:1keVandkTh=4:60:7keV,withthecoolercomponentproducing30%oftheobserved2{8keVux.However,whenthespectrumisdereddened,theadditionalsoftcomponentproducesanenormousamountofuxbetween0.5{2keV,raisingthetotalinferredluminositybyafactorof20to(1:10:3)1035ergs.ThiswouldmakeEdd-1eitheroneofthemostluminousknowncollidingwindbinaries,oramoderatelybrightaccretingblackholeorneutron 41

PAGE 42

3-3 3-6 ,IplottheX-rayandIRluminosityofEdd-1againstknownhigh-massstarsandmassivebinarysystems,includingmassiveOB-stars,LuminousBlueVariables(LBVs),HMXBs,andCWBs.IrecognizethattheX-rayux(andtheIRtoalesserdegree)tendstobevariableinbinarysystemsandthatmyphotometricdataarenotsimultaneous.Thus,thespecicpointsarenotasusefulastheregionsubtendedbythegeneralclasses.Inaddition,forthesesources,thereisoftenuncertaintyinbothcolumndensityanddistancethatcanshifttheindividualpoints.IidentifythedistancesandextinctionusedinplacingthesesourcesaswellastheindividualsourcenamesinTable 3-4 .IdonotplotLowMassX-rayBinaries(LMXBs)inFigure 3-6 becauseLMXBstendtohaveLX=LK>>1.ForEdd-1,LX=LK104.EvenifmostoftheX-rayemissionfromEdd-1isobscured,anditsintrinsicluminosityis100timeslargerthanwhatIhaveinferred,thevalueofLX=LKismoreconsistentwithanHMXBthanaLMXB.InadditiontoX-rayandIRcolor,Iobserveseveralinterestingspectralfeatureswhichmayhelpidentifythenatureofthissource.BecauseEdd-1hasPCygniprolesinseveralHeIIlines,IhavesearchedintheliteratureforobjectswithPCygniprolesinHeIIintheopticalandIR.PCygniprolestendtoappearintheHeIlinesofHMXBsandCWBs(e.g.,CygX-1: GiesandBolton 1986 ;IGRJ16318-4848: FilliatreandChaty 2004 ;andCarinae: Hillieretal. 2001 ).PCygniprolesinHeIIlines,suchasthoseobservedinEdd-1,arerare.IhavealsosearchedforsimilarX-rayfeaturessuchasstrongFe-XXVemissioninEdd-1.Below,IcompareEdd-1todierenttypesofsystemscontainingmassivestars,focusingonsimilaritiestoEdd-1'sdistinguishingspectralfeatures. 42

PAGE 43

3-6 withsimilarcolortoEdd-1,includingthepeculiarOe-starHD108. Nazeetal. ( 2004 )observedHD108intheopticalandX-ray.TheyobserveweakFeemissionintheX-rayat6.6keV.Whentwithatwo-temperatureplasma,theyndkT10:2keVandkT21.4keV,coolerthanthatobservedinEdd-1whenusingatwo-temperaturemodel,butconsistentwiththelowAVmodel.AlthoughsomehavesuggestedthatHD108isabinarybecauseofitsstrongX-rayemission,long-termobservationsby Nazeetal. ( 2004 )suggestthatHD108doesnotexhibitthesamebehaviorasclassicalshort-orlong-termbinaries.TheHandHeIlinesinHD108havebeenobservedtochangefromstrongPCygniprolestosimpleabsorptions( Nazeetal. 2004 ).ThefactthatEdd-1showsPCygniprolesinHeII,notHeI,suggeststhewindinEdd-1isarisinginahotterregion.TheK-bandspectrumofHD108showsprimarilyemission,includingBrandHeI2:114m( Morrisetal. 1996 ),consistentwiththeIRspectrumofEdd-1.OverallsimilaritiesintheX-rayplasmatemperatures,infraredemissionfeatures,andthepresenceofaHeliumwindsuggestthatEdd-1andHD108maybesimilarobjects.TheBpstarOrionisE,alsosharessomeofEdd-1'sdistinguishingcharacteristics.Inquiescence,itsX-raythermaltemperatureismeasuredbetween0.3and1.1keVandreaches3.5keVinanoutburst,consistentwiththeEdd-1models.WhileFe-XXV(6.7keV)isweaklypresent,theFeK(6.4keV)appearsinexcessduringaare( Sanz-Forcadaetal. 2004 ).Incontrast,Edd-1showsweakFeKemissionandstrongFe-XXVemission.TheHeliumlinesinOrionisEvarystrongly( GrooteandHunger 1982 ),butthisvariationisinterpretedasinhomogeneouschemicalabundancesatthestellarsurface( Reinersetal. 2000 ).LBVshavesimilarspectratoB[e]starsandareknownforstrong,variableBrlinesaswellasHIandHeIemission.IplotthepositionsoftheLBVPCygniandtheLBVcandidateknownasthePistolstarinFigure 3-6 .TheX-rayluminositiesthatIcitefor 43

PAGE 44

Munoetal. 2006b ).BrightX-rayemission,suchasthatobservedinEdd-1,wouldnotbeexpectedfromanisolatedLBV.Inshort,manyknownisolatedLBVsdonotshowstrongX-rayemission.ThetwoisolatedOBstarsIhaveconsideredarebothmodeledashavingacoolerthermalX-rayplasmathanEdd-1.Inaddition,thestrongFe-XXVfeatureinEdd-1isobservedtobeweakinisolatedstars.VariableHandHelinesareobservedinisolatedstarsandPCygniprolescanoccur,butarerarelyobservedinHeIIasisseeninEdd-1.ThuswhileIcannotruleoutthisclassicationbasedonthespectralandphotometricdatapresentedhere,IbelievetheisolatedstarscenarioislesslikelythanabinarynatureforEdd-1.Inthenextchapter,IuseX-rayvariabilitydatatoruleouttheisolatedstarscenario. Schulzetal. 2002a ; Nagaseetal. 1992 ).Itisrare,however,tondironlineswithequivalentwidths>1keV.InCenX-3,bothFeK(6.40keV)andtheFe-XXVtripletareobserved( Iariaetal. 2005 ).TheequivalentwidthsoftheselinesaremeasuredatonlyafeweVwhenoutofeclipse.Incontrast,stronglineemissionisseenduringtheeclipsesofVelaX-1,withFeKequivalentwidthaslargeas1.3keVindeepeclipse(e.g. Choietal. 1996 ; Schulzetal. 2002a ).NeitheroftheseshowdominantFe-XXVemission.BothoftheseHMXBshaveasupergiantOBcompanion.TheiropticalspectrashowHydrogenandHeliuminbothabsorptionandemissionassociatedwiththestar( Mouchetetal. 1980 ; Dupreeetal. 1980 ).AmorerareHMXBcompanion,asupergiant-B[e]star(sgB[e]),hasbeenobservedinCICamelopardalis(CICam).WhileseveralsgB[e]starshavebeenobservedintheMagellanicClouds( Zickgrafetal. 1986 ),theyarerarelyobservedintheMilkyWay.LikeEdd-1,CICamhasprominentlineemissioninitsIRspectrum,butadditionallyCICam 44

PAGE 45

FilliatreandChaty 2004 ),butitdoesnotexhibitstrongFe-XXVemission.RecentlyIGRJ16318{4848hasbeententativelyidentiedashavingasgB[e]companion.WhiletheX-rayFeKemissionfromIGRJ16318{4848isnegligible,aHeIwindhasbeenobservedwithavelocityof41040km/s( FilliatreandChaty 2004 ).TheHeIIlinesinEdd-1indicateamuchweakerwindwithavelocityof17070km/s.PCygniprolesarefairlycommonamonghigh-masssystems;however,theyaremoreoftenobservedinHorHeIthanHeII(e.g.Vela-X-1,IGRJ16318-4848).Todate,onlyoneHMXBhasbeenobservedtohaveaHeIIwind:CygX-1.CygX-1isablack-holebinarywithanO-starcompanion.ItshowsweakFeKwithanequivalentwidthofonly16eV. GiesandBolton ( 1986 )foundthatthePCygniproleofHeII4686AinCygX-1variedwithorbitalphaseandthemaximumPCygnivelocitywas100km/s,similartothatofEdd-1.ThevaryingPCygniproleinCygX-1ismodeledasafocusedstellarwindwhoseuxrelatedtothemasstransferrateinthesystem.ThepresenceofHeliumwindandtheprominentHandHeemissionintheIRspectrumofHMXBslendssupporttothisscenarioforEdd-1.However,althoughstrongFeKlinesareseeninHMXBs,thisfeatureisrarelytheFe-XXVlineobservedinEdd-1.Finally,whiletheinferredX-rayluminosityofEdd-1ismoretypicalofanHMXBthananisolatedstar,theIRluminosityisbrighterthantheclassicalHMXBslistedinTable 3-4 .SoalthoughanHMXBscenarioissupported,Icannotmakeaconclusiveclassicationonthebasisofthesedata.Inthenextchapter,IusetheX-rayvariabilityofEdd-1toconstrainphysicalscenariosinwhichEdd-1isconsistentwithaHMXB. 45

PAGE 46

Sanaetal. 2004 ; Skinneretal. 2005 ).Forexample,2Velorum(hereafter2Vel),aWC8+O7binary,hasameasuredtemperaturekT=1:5keV( Skinneretal. 2001 ).BecausetheWRstardominateslineemission,2VelappearsHelium-rich.InWR+Obinaries,weakBrackettseriesemissionmaybepresent,butbroadHeliumlinesdominateemission( Skinneretal. 2001 ; Varricattetal. 2004 )ItispossibletheHe-windobservedinEdd-1comesfromaneclipsedorobscuredWRcompanion.However,becauseBrackettseriesemissiondominatesEdd-1,Indthisscenariolesslikely.IfurtherconsidertheunlikelihoodofaWR+Obinarysysteminthenextchapter.BasedontheX-rayspectrum,Edd-1holdsthegreatestsimilaritytoEtaCarinae.EtaCarinaeisanunusuallybrightX-ray/IRsourcewhosenatureisnotdenitive,butmodelssuggestaCWBcontaininganLBV+Ostarpair.ThegreatestdierencebetweenEdd-1andEtaCarinaeistheinfraredluminosity.EtaCarinaeisintrinsicallymuchbrighterintheIRbyovertwoordersofmagnitude.ThesimilaritiesbetweenEdd-1andEtaCarinaeareprimarilyintheX-ray. Viottietal. ( 2004 )ndthatthethermalcomponentofEtaCarinaehasatemperatureof5.5keVwithNH=4:81022cm2,comparabletothetwo-temperaturemodelofEdd-1.Inaddition,EtaCarinaehasasizeableFe-XXVlinevaryingbetween0.9and1.5keV( Viottietal. 2004 ).ThisistheonlysourceIhavefoundintheliteraturewithaFe-XXVlineofsimilarequivalentwidthtothatobservedinEdd-1.TypicallyassumedtobeanLBVinabinarysystem,theX-rayemissionofEtaCarinaeismodeledasarisingfromthecollidingwind.DetectionofPCygniprolesinEtaCarinae'sopticalspectrumisconsistentwithaCWBclassication. SteinerandDamineli ( 2004 )observedvariableHeIIemission.Itisbelievedtooriginateinadensestellarwind,howevertheenergysupplyisstilldebated( Martinetal. 2006 ).BecausewindfromtheLBVinEtaCarinaeismostlycool,theHeIIemissionisbelievedtobefromthecompanion,suggestedtobeanO-star.ThepresenceofstrongFe-XXVemissioninEtaCarinae,thesimilarX-raythermaltemperaturetoCWBs,andthepresenceofHeliumwindsinCWBslendstrongsupporttoEdd-1beingaCWB. 46

PAGE 47

47

PAGE 48

TheK-bandspectrumofEdd-1showsstrongBr,Br,andHeIemission.PCygniprolesareseeninseveraloftheHeliumlines,suggestingaHeliumwindaroundamassivestar. 48

PAGE 49

BrackettseriesemissiondominatestheH-bandspectrumofEdd-1. 49

PAGE 50

ThePalineisclearlyvisibleintheJ-bandspectrumofEdd-1.Atmosphericnoisedistortsthespectrumat=1:351:42mandat<1:15m. 50

PAGE 51

PCygniprolesfortheHeliumlines.Thedottedlineshowstheapproximatecontinuumlevel.Theverticallineisplacedatthevacuumcenterwavelength. 51

PAGE 52

TheaverageChandraX-rayspectrumofEdd-1.Thesourcedisplaysprominentlineemissionfromthen=2{1transitionsofHe-likeSat2.4keV,Arat3.1keV,Caat3.9keV,andFeat6.7keV. 52

PAGE 53

ComparisonoftheX-rayandIRluminosityofEdd-1tootherknownX-raysystemscontainingmassivestars.Whenasinglesourceisobservedatvaryingluminosities,thetwopointsareconnectedwithaline.Thelightersymbolindicatesaquiescentstate.BecausesimultaneousX-rayandIRdataisnottypicallyavailable,IgenerallyhaveonlyoneK-banddatapoint.FurtherinformationonthesesourcesisavailableinTable 3-4 53

PAGE 54

JHI5-3(Pa)1.282192310blend? Hunknown1.5031:90:4340 HHI14-4(Br14)1.5892:30:3280 HHI12-4(Br12)1.6416:60:4470 HHI11-4(Br11)1.6815:90:6570 HHeI4D-3P,3D-3P01.7012:70:4370 HHI10-4,HeII20-81.736blblblend HHeII19-81.77210:00:52160 KHI8-4(Br)1.94536:20:6640HeIblend? KCIII/NIII2.104blblblend KHeI4S-3P,1S-1P02.11413:80:5590blend?CIII/NIII KHI7-4(Br)2.16636:60:3710 KNIII2.2471:70:2310 Note.|Theidentiedlinestransitions,vacuumwavelength,equivalentwidth,andfull-widthvelocity.Thefull-widthvelocityhasbeencorrectedfortheintrinsiclinewidthoftheinstrument.Vacuumlinecentersareobtainedfrom Hansonetal. ( 1996 ); Morrisetal. ( 1996 ); Figeretal. ( 1997 ); Wallaceetal. ( 2000 ); Schultz ( 2005 ).Uncertaintyinvelocityisupto70km/sduetouncertaintiesinmeasuredlinewidth. 54

PAGE 55

KHeII10-72.1891150 KHeII13-82.3464180 Note.|Thevelocityiscalculatedbasedonthedierencebetweentheblueedgeandvaccumcentralwavelength.LinecentersareasreferencedinTable 3-1 .BecauseofdiscrepanciesinthelocationoftheHeII13-8line,Iestimatethecentralwavelengthusingtheequation0:0911138Z2n2in2j=(n2jn2i).UncertaintiesduetopixelsizeandpeaklocationgiveaV=70km=s. 55

PAGE 56

NH(1022cm2)3:5+0:30:45.2akTs(keV)0:7+0:10:1KEM;s(1056cm3)32+92kTh(keV)2:0+0:50:24:6+0:70:7KEM;h(1056cm3)5+111:4+0:40:2ZS=ZS;1:1+0:50:40:9+0:20:2ZAr=ZAr;1:1+0:70:71:3+0:70:4ZCa=ZCa;1:5+0:80:82:2+1:30:7ZFe=ZFe;1:4+0:30:31:5+0:20:2FFeXXV(107photoncm2s1)32212=125.2/107105.6/105FX(1013ergcm2s1)1:21.3LX;s[D/8kpc]2(1033ergs1)110LX;h[D/8kpc]2(1033ergs1)63 56

PAGE 57

dSourceClasslog(LX)alog(Lk)(kpc)AKReferences CICamsgB[e]+X33.54(q)39.2350.3[3],[4] IGRJ16283-4838Be+NS34(q)b35.6650.9[5]IGRJ16283-4838Be+NS35(b)b35.8451.4[5] XTEJ1906+090Be+P34.84(q)36.0641.8[6]XTEJ1906+090Be+P36.84(b)36.0641.8[6] GROJ2058+42Be+X33.4737.559.01.2[7]GROJ2058+42Be+X33.9537.559.01.2[7] X1908+075OBI+NS36c37.9672.3[8] CygX-1O9I+BH36.80(q)37.862.50.36[9],[10]CygX-1O9I+BH37.20(b)37.862.50.36[9],[10] CenX-3OI+NS37.70d37.8181.8[11],[12] VelaX-1BI+NS33.3237.941.90.28[13],[14] HD152248O8I+O32.90(q)37.871.7570.15[10],[15],[16]HD152248O8I+O33.04(b)37.871.7570.15[10],[15],[16] HD150136O3+O6V33.3837.851.320.20[10],[17] PCygniLBV<31.0e39.022.10.2[20],[21] 57

PAGE 58

HD108O6f33.037.602.10.2[10],[28],[29] HD152408O8Iaf<31.7e38.292.160.16[21],[28] HD151804O8I31.9e38.071.660.13[21],[28] X174516.1O?33.339.2882.7[21] H2O?33.139.4584.5[21] 3-6 .Here,Ispeciythesourceclassications(whenavailable)aswellasthedistanceandK-bandextinctionusedincalculatingluminosity.Theluminositiesareinerg=s.Thereferencenumbersareasfollows:[1]thiswork;[2] Munoetal. ( 2003 );[3] Boirinetal. ( 2002 );[4] Clarketal. ( 2000 );[5] Beckmannetal. ( 2005 );[6] Gogusetal. ( 2005 );[7] Wilsonetal. ( 2005 );[8] MorelandGrosdidier ( 2005 );[9] Schulzetal. ( 2002b );[10] Maz-Apellanizetal. ( 2004 );[11] Coeetal. ( 1997 );[12] Nagaseetal. ( 1992 );[13] Schulzetal. ( 2002a );[14] HylandandMould ( 1973 );[15] Cassinellietal. ( 1981 );[16] Sanaetal. ( 2004 );[17] Skinneretal. ( 2005 );[18] Schildetal. ( 2004 );[19] Williamsetal. ( 1990 );[20] Turner ( 1985 );[21] Munoetal. ( 2006b );[22] vanGenderenetal. ( 1994 );[23] Sewardetal. ( 2001 );[24] Evansetal. ( 2003 );[25] Figeretal. ( 1998 );[26] GrooteandHunger ( 1982 );[27] GrooteandSchmitt ( 2004 );[28] LeithererandWolf ( 1984 );[29] Nazeetal. ( 2004 ) 58

PAGE 59

4.2.1InfraredOn2006Aug02-04UT,IobtainedJ,H,andKband(1.1-2.4m)spectraofEdd-1usingSpeXonIRTF( Rayneretal. 2003 ).Ditheringalongthe0.5arcsecslit,Iobtained184exposuresof120seachoverthecourseofthreehalf-nights,givingusatimebaselineof3-4hourspernight.Intheshort-wavelength,cross-dispersedmode,IgetaresolutionofR1200overtheJHKbandpass.TargetobservationswerefollowedbyobservationsoftheG0VstarHR6836atsimilarairmassforremovaloftelluricabsorptionfeatures.IextractspectrausingthestandardSpexToolprocedureforABnoddeddata(detailedinChapter2),resultinginaseriesofsky-subtracted,wavelength-calibratedspectra( Vacca 59

PAGE 60

, 2003 ; Cushingetal. 2004 ).Usingmypreviousobservationstakenon2005July1UT(Chapter3),IadoptareddeningvalueofAV=29magandapplythiscorrectiontoalldata.Icombinedthespectrafromeachnighttotestvariabilityonmultipletimescales.Figure 4-1 showstheseriesoftwenty-oneK-bandspectratakenoverthecourseofourobservations,withintegrationtimesbetween8and20minutesperspectrum.IlistthespecictimestampsandexposuretimesofthesespectrainTable 4-1 .Tosearchforradialvelocityvariationsintheemissionlines,Itrackthelinecenterswithtwomethods:rstbytakingastatisticalmeanofthewavelengtharoundthelinecenter,weightedbyux,andsecondbyttingaGaussiantotheline.Indnoradialvelocityvariations,nordoIndsignicantuxvariationsinthelines.IcheckedforIRvariabilityon1year,3day,3hour,1hour,and30minutebaselinesandfoundnoevidenceofperiodicvariabilityoraresinthissample.TheonlyapparentvariationisinthestructureoftheBrlinecomplex(seeFigure 4-2 ),butthisdoesnotoftenvarymorethan5timestheRMSspectraldierenceinthevicinityofthe2:164Heliumcomponent.FurtherInotethatthisregionisaectedbyourdatareductionprocess(i.e.,theremovaloftheintrinsicBrackettabsorptionintheG0V). ( 2004a )examinedthespectrumandvariabilityoftheX-rayemissionfromEdd-1aspartofastudyof2000X-raysourcesdetectedtowardtheGC.Theanalysisisdescribedindetailin Munoetal. ( 2004b )andsummarizedinChapter3.IlisttheChandradatausedinouranalysisinTable 4-2 .DuringtheChandraobservations,theX-rayemissioninEdd-1variedinintensitybyafactorof3inthe2-8keVrange.InTable 4-3 ,IlistXMMarchivaldatausedinouranalysis.Whilethe2001-2002XMMdataencompassfairlyshortobservations(exposuretime<7h),in2004Ihavefourobservationsof40consecutivehourseach.IfollowedthestandardXMMreductionroutineforodfdata,rstbuildingacif-file,thenusingodfingestandepchaintocreate 60

PAGE 61

4-3 .Theuxvariesbyafactorofvebetweentheseobservations.Becauseoftheextremelylowcountrate,IcannottthefainterspectrumwithXSpecmodels.Iextractedlightcurvesatone-hourintervalsoverthefull2-8keVband,aswellasfromthe\soft"2-4keVband,andthe\hard"4-8keVbandseparately.Usingtheseone-hourresolutionlight-curves,Iperformaperiodanalysissearchingforperiodicitiesintherangeof0.1-40hours,butndnosignicantperiodsinthisrange.IndthattheXMMX-rayuxisconstantwithinPoissonerrorsduringasingleobservation(aslongas40hours).Idonda4variationinconsecutiveobservationsseparatedbyvemonths(seeFig. 4-3 ).ThusIcalculateasingleuxvalueforeachXMMobservationepochandcombinethesemeasurementswiththeChandralightcurveinFigure 4-4 .Usingthecombinedlightcurve,Icantestforthepresenceorabsenceofperiodicuxvariationsontimescaleslongerthan40hours.Usingthemethodof HorneandBaliunas ( 1986 ),Iperformaperiodogramanalysisofthecombinedlightcurveandndaperiodof1896days.InFigure 4-5 ,Ishowtheresultantperiodogramwhichtestsforperiodicityonscalesof1-1500days.Thepeakat189disclearlydistinctandadditionalpeaksarevisibleatintegermultiplesoftheperiod.InFigure 4-4 ,IplottheX-raylightcurvefoldedonthe189dperiod.Analyticallyestimatingthesignicanceofasignalinnon-uniformlysampleddataisnon-trivial.Thus,inordertoestimatethecondenceofthisdetection,IperformaMonteCarlosimulationasfollows.Itaketheexistingdatasetandmaintainthesamesamplingintervalsthroughout.ForeachMonteCarlorealization,Irandomlyreassigntheobserveduxvaluestothetimesamples,eectivelyscramblingthelightcurve.In30,000trials,Idonotachieveapeakpowerapproachingthepowerofouroriginalperiodogram,implying 61

PAGE 62

Titarchuketal. 2007 ).Rednoiseisauxvariationinthepowerspectrumthatcanbeparameterizedwithafrequencydependencef.Awhitenoiseprocesswillgenerateaatpowerspectrumsuchthat0;avalueof2describesrandomwalknoise( TimmerandKoenig 1995 ).A1dependencehasbeennotedinstellar-massblackholecandidatesandmaybestronglyrelatedtoaccretionphysicsinthesystem( Mineshigeetal. 1994 ; TimmerandKoenig 1995 ; Titarchuketal. 2007 ).Followingthemethodof TimmerandKoenig ( 1995 ),Itestthepossibilityofrednoisecreatingafalsesignalwhichmatchesthestrengthofourperiodogram.Bysimulatinganumberofrednoisedominatedlightcurvesofvaryingpowerlawslope,,Indthatasincreases,morenoisegetsshuntedneartheperiodfrequency,andthesignicanceofourdetectiondecreases.Indthesignicanceofourperioddetectionremainsabove3forvaluesof1:0andabove2:5for1:5showingthatthesignicancedecreasesslowlyasrednoiseisincreased. 4.3.1InfraredVariabilityIntheinitialdiscoveryspectrum,IidentiedthreeHeIIlineswithPCygniproles:2.0379m,2.1891m,and2.3464m(Chapter3).Intheanalysisofour2005data,IestimatedthePCygnivelocityat17070km/s.Irepeatouranalysisonthe2006datatosearchforvariationsandndtheapproximatevelocityofthewindis20070km/s.(Theerrorisdominatedbythespectralresolution.)IndnoevidenceofchangesinthePCygniproleorvelocityoverourthreedayobservations.Also,the2005and2006spectrahaveconsistentPCygniprolesandvelocities.InFigure 4-2 ,Ishowseveralclose-upsoftheBrackett-regionofEdd-1'sspectrumoverthecourseofourthreenightIRTFrunin2006.Totheleftofthe2.164mmarker, 62

PAGE 63

4-4 ),wheretheobjectistransitioningfromanapparentlow-uxstatetoahigh-uxstate.BecauseIhavenoIRdataconsistentwiththelowestX-rayuxevents,itisimpossibletoknowfromthesedataiftheapparentHeliumvariabilityat2:164mIobserveisassociatedwiththisX-rayuxtransition.Becausetheinfraredspectrafrom2005and2006aretakenatnearlythesamephaseintheX-raylight-curve,Idonotexpecttondradialvelocityvariations.NorcanIobservevariationsiftheorbitalvelocityislessthan70km/s. Cohen 2000 ),andingwhichfavorsabinaryinterpretation.InthestandardmodelsforCWBs,X-rayemissionarisesfromtheshockfrontofcollidingwindsintwomassivestars(see,e.g. Luoetal. 1990 ; Sanaetal. 2004 ; DeBeckeretal. 2006 ).Observedvariabilityisoftenattributedtophase-lockeduxmodulations 63

PAGE 64

Davidsonetal. 1998 ).Inthesesituations,themodulationoftheX-rayuxiscorrelatedtorecurrentbehavioraectingthewindemission,butnotrelatedtotheorbitalperiod.However,inHMXBs,periodicX-rayuxchangescanbetheresultofeitherorbitalorsuperorbitalmotion.SeveralHMXBsinthe Liuetal. ( 2006 )cataloghaveorbitalperiodsof>100d.Theselong-periodsystemsareoftenX-raytransientswithBestarcounterparts( Lewinetal. 1997 ).CenX-3andCygX-1arebothhigh-massbinarysystemsshowingbothorbitalandsuperorbitalperiods.Thesuperorbitalperiodsofthesesystemsare140dand142drespectively,valueswhichareassociatedwithaprecessingaccretiondisk.Theirorbitalperiodsare2.1dand5.6d( OgilvieandDubus 2001 ).However,ifEdd-1hasasimilarshortorbitalperiodinadditiontothe189dperiod,IdonotdetectthatshortperiodinourcurrentX-raydata(Fig. 4-5 ).ThusIrestrictmyselftoexploringthepossibilitythatthe189dperiodisorbitalratherthansuperorbital. Vanbeverenetal. 1998 ; LewinandvanderKlis 2006 ).InChapter3,IdetermineanabsolutemagnitudeMK=7:60:3forEdd-1usingadistanceof8kpc,reddeningofAK=3:4,anda2MASSmagnitudeofKS=10:33.IcanuseEdd-1'sexceptionalbrightnessandtheX-rayperiodtoplaceconstraintsonthenatureofthesystem.Forourpurposes,the\primary"star(mass,MOB)willrefertothemassiveOB-starandthe\secondary"star(mass,M2)willrefertothecompanionwhosenaturehasyettobeidentied. 64

PAGE 65

Cox 2000 ; Girardietal. 2002 ).InFigure 4-6 ,Iplotthemassratioasafunctionoftheinferredorbitalvelocityforarangeofprimarymassesandnotethattheorbitalvelocityislessthanourinfraredspectralresolutionof70km/sforcasesofmassratioq<0:5.Evenforhighermassratios,aradialvelocityvariationwouldhavelowsignal-to-noisewithourcurrentobservations.Thus,Irequirehigherresolutionspectroscopyinordertoobserveradialvelocityvariationsintheinfraredassociatedwiththisperiodicity.InthenexttwosectionsIdiscussthepossibilitythatthemodulationsinX-rayuxarecausedby(1)obscurationoftheX-raysourcebystellarwind;and(2)eclipseoftheX-raysource. 65

PAGE 66

r2=_M Mokiemetal. 2007 ),thisbecomes Rbp 4{4 shouldbeconsideredalowerlimitofthe_MrequiredtoproducetheabsorptioncolumnthateectstheuxchangeinEdd-1.Iestimatethemasslossfortwospecialcases.Intherstcase,Ipostulatethattheinfraredlightisdominatedbyasinglebrightsource.InHMXBs,thestarisexpectedtocontributemoreheavilytoopticalandinfraredemissionthantheaccretiondisk( Lewinetal. 1997 ).IncertainCWBcases,especiallyoflowermassratios,itispossiblethatasinglesourcedominatesemission( Lepineetal. 2001 ).ThusforCWBandHMXB 66

PAGE 67

Girardietal. ( 2002 )andndthatastarwithMK7:6willlikelyhavearadiusROB80Rvalidforarangeofmasses20100M.Usingequation 4{4 ,Igetamass-lossrateof_M4105Myr1.Inthesecondcase,IconsiderasystemthatcontainstwomassivestarseachcontributinghalfoftheinfraredluminositywhichisonlyconsistentforCWBscontainingtwostarsofsimilarbolometricluminosity.ThesestarswouldhaveROB20Rand_M1105Myr1.TypicalmassiveO-starsarereportedtohavemass-lossratesof106105Myr1( Mokiemetal. 2007 ).Thus,whilearelativelywindystarisnecessarytoproducetheuxvariationsthatIobserve,themass-lossrateisnotunreasonable.Thereforeawind-obscurationscenarioisconsistentwithbothCWBandHMXBinterpretationsofEdd-1. (qsini)3 (=50d)3r3OB 4-4 ,Ilistaseriesofmassratios,q,associatedwithvaryingfractionsr3OB=mOB.Asanexample,IcanexaminethetwocasesasIdidabove.Tocompletethisnumericalexercise,IchooseamedianprimarymassMOB=40M(whileacknowledgingthatawiderangeofmassesispossible).IftwomassivestarseachcontributehalfoftheIRluminosity,thenROB20R,r3OB=mOB=200,andthemassratioisq0:05.Thisresultingmassratio 67

PAGE 68

TerrellandWilson 2005 ).ThusifthesystemisaCWB,itwouldhavetohavearelativelylowmassratiowiththeinfraredemissiondominatedbyasinglesource.Thisimpliesthatthewindemissionofonesourceoverwhelmsthatofitscompanion( Luoetal. 1990 ).ItispossibleforCWBstohavelowermassratiosifthesecondaryisaWolf-Rayet(WR)star.BythetimeamassivestarreachestheWRstage,itmayhavearelativelysmallmass,butstillhaveenormouslypowerfulwinds( Crowther 2007 ).Forexample,2VelorumisaWR+Ostarwithamassratioq0:35( vanderHucht 2001 ).Inthecaseof2Velorum,theWRstardominatestheIRemission,sothesourceappearsHeliumrich( Crowther 2007 ).ItispossiblethattheHeliumemissionIobserveinEdd-1isevidenceofanobscuredWRcompanion.However,becauseBrackettseriesemissionratherthanHeliumemissiondominatestheIRspectrum,Indthisscenariolesslikely.InTable 4-5 ,IlistlineratiosofBrtoHeI2.114mandBrtoHeII2.189mforknownCWBsandXRBs.InknownWR+Obinaries,theHeII2.189misnotablystrongerthanBr.Comparatively,Edd-1hasmuchstrongerBremission,andhenceaquitedierentBr/HeIIlineratiofromwhatisobservedinWR+Osystems.Infact,InotetheBr/HeIandBr/HeIIlineratiosinEdd-1aremoreconsistentwithHMXBsthaneitherO+OorO+WRCWBs.ThusifEdd-1isaWR+OCWB,itisveryunusual.Intheeclipsingbinaryscenario,Edd-1wouldmorelikelybeanHMXB. Tomsicketal. 2006 ; Sidolietal. 2006 ) 68

PAGE 69

Nagaseetal. 1992 ; Schulzetal. 2002b ).IexploretheimplicationsoftheobservedperiodforthecasewhereEdd-1isanaccretingbinarysystemwithacompactobject.SincetheIRdatasuggestthatEdd-1containsahigh-massstar,Ifocusonthecaseofwind-fedaccretion.Takingthestandardaccretionluminosityas Franketal. ( 2002 ). Franketal. ( 2002 )estimatetheaccretioneciency,_M=_Mwind,bycomparingthemassuxwithinanaccretioncylindertothetotalmasslossofthedonorstar.Theaccretioncylinderisestimatedfromthegravitationalpotentialofthecompactobject,giving_M 4a2vwind(a)whereracc2GM2=v2wind,vwind(2GM1=R1)1=2,andaistheorbitalseparation.Thisgivesus _M 4M2 69

PAGE 70

4{10 andndthatthemassratioofthesystemisq0:01.ThissuggestsamassiveM>80Mdonorforatypicalneutronstarcompanion.Byrelaxingtheestimateofthemassivestarradius,ROB,,Indthatqwillincreaseandmorecompactobjectsolutionsexistoverawiderrangeofprimarymasses.TheestimateofROB=80Risderivedfromthegravitationalpotentialasestimatedinthe Girardietal. ( 2002 )isochrones.InTable 4-6 ,IlistaseriesofsolutionsforEquation 4{10 .Becausetheeclipsingscenariocaseplacesrmconstraintsonthemassratioofthesystem,IuseEquation 4{10 tocalculatethemasslossratesassociatedwithvariousscenarios.IlistthosevaluesinTable 4-4 .Forcasewherethemassratioisq0:2,theassociatedmasslossrateislow_Mwind2107M=yr,foraneciency0:1.Thisisnotunreasonableformassivestars( Mokiemetal. 2007 ).Interestingly,inboththewindobscurationandtheeclispingbinaryscenario,theX-rayluminosityisconsistentwithalowmassratioforthesystem. 70

PAGE 71

4-7 .Iftheobservedperiodisorbitalinnature,andtheX-raymodulationiscausedbyobscurationoftheX-raysourceduetoadensewind,thenEdd-1isconsistentwithbothCWBandHMXBinterpretations.ThefurtherconstraintoftheX-rayluminosityisconsistentwithamassive(MOB>80M)donorwithaneutronstarcompanion.IfX-raymodulationiscausedbyaneclipse,themassratioislowandEdd-1ismoreconsistentwithanHMXBinterpretation. 71

PAGE 72

TheK-bandspectraofEdd-1.Ishowtheoriginal2005spectrumatthebottomandthetwenty-minutecombinationsofthe2006spectraoverthethreenights.Theseareosetbytimeofobservation,suchthattheearliestspectraarelowerandlaterarehigher.TherelativetimesofthesespectraarelistedinTable 4-1 72

PAGE 73

TheBrregionofselectEdd-1spectratakenfrom2006Aug02-04.Theregionshowsapparentnon-periodicvariation,mostlyaroundthe2:164mHeliumcontribution.Thesevariationsareonlyoccasionallygreaterthan5-timestheRMSspectraldierence.HigherresolutionspectroscopyisneededtoshowwhetherthisisintrinsictoEdd-1oranartifactofthedatareduction.TherelativetimesofthesespectraarelistedinTable 4-1 73

PAGE 74

TworepresentativeXMMspectraseparatedby0.7inphase.WhilethestrengthoftheFe-XXVlineisconsistentbetweenthetwoobservations,thecontinuumleveldropssignicantly.Ifsuchvariationwerecausedentirelybycolumnabsorptionduetoastellarwind,thenNHwouldincreaseby2:51023cm2. 74

PAGE 75

TheX-raylightcurve(top)andfoldedlightcurve(bottom)ofEdd-1.Thelightcurveisfoldedona189dperiod.ThesquaresareXMMdata;thediamondsareChandradata. 75

PAGE 76

AperiodogramanalysisoftheX-raylightcurve.Thepreferredperiodis1896days.Subsequentpeaksappearatintegermultiplesofthisperiod. 76

PAGE 77

Usingthemassfunctionandtheputativeperiodof189days,Icalculatetheexpectedmassratio,q=M2=MOB,forprimarymassesMOB=20100M.Theprimarymassisindicatedtotheleftofeachline.TheverticaldashedlinerepresentsthelimitingIRspectralresolution. 77

PAGE 78

a2006-08-026:3820b2006-08-027:0520c2006-08-028:2520d2006-08-028:5516e2006-08-029:468f2006-08-035:2716g2006-08-035:5216h2006-08-036:3720i2006-08-037:0720j2006-08-037:4920k2006-08-038:2020l2006-08-039:0220m2006-08-039:3212n2006-08-045:5120o2006-08-046:2716p2006-08-046:4616q2006-08-047:2320r2006-08-047:5816s2006-08-048:1118t2006-08-048:4720u2006-08-049:1020 Note.|TheseobservationIDsareassociatedwithFigures1and2.Thedaysalignwithdays2404-2406onourX-raylightcurves. 78

PAGE 79

2000-10-2618:15:111561a35.7266.41344-29.01281264.72001-07-1401:51:101561b13.5266.41344-29.01281264.72001-07-1814:25:48228410.6266.40415-28.94090283.82002-05-2222:59:15294334.7266.41991-29.0040775.52002-02-1914:27:32295112.4266.41867-29.0033591.52002-03-2312:25:04295211.9266.41897-29.0034388.22002-04-1910:39:01295311.7266.41923-29.0034985.22002-05-0709:25:07295412.5266.41938-29.0037482.12002-05-2515:16:033392165.8266.41992-29.0040875.52002-05-2805:34:443393157.1266.41992-29.0040775.52003-06-1918:28:55354924.8266.42092-29.01052346.82002-05-2411:50:13366338.0266.41993-29.0040775.52002-06-0301:24:37366589.9266.41992-29.0040775.52004-07-0522:33:11468349.5266.41605-29.01238286.22004-07-0622:29:57468449.5266.41597-29.01236285.42004-08-2812:03:5953605.1266.41477-29.01211271.02005-07-2419:58:27595048.5266.41519-29.01222276.72005-07-2719:08:16595144.6266.41512-29.01219276.02005-07-2919:51:11595243.1266.41508-29.01219275.52005-07-3019:38:31595345.4266.41506-29.01218275.32005-08-0119:54:13595418.1266.41502-29.01215274.92005-02-2706:26:0461134.9266.41870-29.0035390.62006-07-1703:58:28636329.8266.41541-29.01228279.52006-04-1105:33:2066394.5266.41890-29.0036986.22006-05-0322:26:2666405.1266.41935-29.0038382.82006-06-0116:07:5266415.1266.42018-29.0044069.72006-07-0411:01:3566425.1266.41633-29.01237288.42006-07-3014:30:2666435.0266.41510-29.01218275.42006-08-2205:54:3466445.0266.41484-29.01202271.72006-09-2513:50:3566455.1266.41448-29.01195268.32006-10-2903:28:2066465.1266.41425-29.01178264.4 79

PAGE 80

01129721012001-09-0401:19:347.501113501012002-02-2603:11:271401113503012002-10-0306:36:49502026705012004-03-2814:37:164002026706012004-03-3014:29:074002026707012004-08-3102:54:314002026708012004-09-0202:44:0840 Table4-4. 4{10 whichisonlyvalidfortheHMXBcase. 80

PAGE 81

Edd-113.836.6<212.65>18.3 Note.|Infraredlineratios.IcomparetherelativestrengthofHeIandHeIIlinestoBrinEdd-1andaselectionsofHMXBsandCWBs.NotethattheHeII2.189mlineinEdd-1hasaPCygniprole.IgroupO+OandO+WRbinariesseparatelyastheformersystemsarelesslikelytoproducelowmassratios.InknownWR+Osystems,theBr/HeIIlineratioissignicantlydierentthanthatobservedinEdd-1.REFERENCES-(1)Chapter3;(2) ClarkandDolan ( 1999 );(3) Clarketal. ( 2003 );(4) Figeretal. ( 1997 );(5) FilliatreandChaty ( 2004 );(6) Hansonetal. ( 1996 ). 81

PAGE 82

80600.0110.6 801000.0161.6 50200.0150.3 50600.0221.3 501000.0262.6 20200.030.6 20600.063.6 201000.077.0 Note.|Themassratioandcompactobjectmassexpectedforaprimaryofthegivenmasstoradiusratiointhewindobscurationscenario,validfortheHMXBcase.TheestimationofqisbasedonEquation 4{10 .IuseLX=1.11035ergs1andassumeaneciency=0.1,masslossrate_M=4105Myr1.ThevalueROB=80Rismostconsistentwithourobservedinfraredluminosity( Girardietal. 2002 ). 82

PAGE 83

WINDOBSCURATIONSCENARIOECLIPSINGBINARYSCENARIO 4{4 ) 4{6 )consistent inconsistentwithinitialassumptions 4{4 ) 4{6 )consistent IRlineratiosinconsistentwithknownWR+Osystems 4{4 ) 4{6 )q0:01(Eq. 4{10 ) _M=2107M=yr(Eq. 4{10 )radiusconstraintsuggestsMOB>80M 83

PAGE 84

Herasetal. 2002 ).Additionally,( Herasetal. 2002 )showitispossibletospectrallytypelow-mass,KandMstarsduetoabsorptionbandspresentintheIR.TheIR/opticalspectraltypingoftheX-raysourcecounterpartisavitalsteptowarddistinguishingCWBsandHMXBs.However,the\canonical"X-raystellarsourcesdening 84

PAGE 85

Munoetal. 2006a )andthecomplementaryinuxofIRspectroscopiccounterpartsexpectedinIRsurveys(e.g.,theFlamingos-2GalacticCenterSurvey(F2GCS): Eikenberryetal. 2005 ),IconcludedthatitwouldbeusefultoestablishacompleteIR/X-rayspectralatlasofarchivedobservationsofknownCWBsandHMXBstobetterandmorequicklycharacterizenewGCsourcesobservedinthesewavebands.ThegoalistocreateasearchabledatabaseofspectroscopicandphotometricX-ray,IR,andopticalinformationonsourcesthatarealreadyclassiedaseitherCWBsorHMXBs.Thisprovidesastartingpointfromwhichtocompareunclassied,newsourcestowell-studiedsystemsthatmaysharesimilarfeaturestoseeiffurthercluesastothenatureofnewsourcescanbediscernedwithlessdirectobservationalinformationaboutthem.CompilingsourcesintoanAtlasoersconsiderablescienticbenetsaswell.ThecharacteristicsofXRBsandCWBsdierwithintheindividualclasses,thusmanynewlydiscoveredsourcesareonlycomparedtooneortwowell-studiedsourcesatatime.Additionally,muchofthecompiledinformationcomparestheX-rayperiodicityofasourceratherthanluminosityorlinefeatures.Thisisdue,inpart,totheknownvariablenatureofX-rayemittingbinariesandtheusefulnessofvariabilityinclassifyingasystem,andinpartduetopracticalconsiderations(e.g.,timeallocation)involvedwithtakenmulti-epochspectroscopyofasource.AlthoughthisAtlasisnotintendedtodescribethefullrangeofvariabilityandbehaviorsofasinglesource,bycollectinginformationfromIRandX-rayspectraatoneormoreepochs,Icanexaminecommonalitiesamongawidercross-sectionofsourcesthatarenotoftenconsideredside-by-side. 85

PAGE 86

Westbrooketal. 2006 ).MyAtlasmakesuseofexistent,publishedIRandX-rayspectra,buthopefullywillinspirenewobservationsatbothwavelengths.Thepreliminaryinteractiveonlineversionfeatures(1)atableofcommonstatisticsonthecatalogedsourcesincludingdistance,reddening,X-rayluminosity,andIRluminosity;(2)theabilitytoselectstatisticsonewantsinanoutputtable;and(3)theabilitytobrowseallcollectedinformationonanindividualsourceinthecatalog.TheAtlasishostedat:http://www.astro.ufl.edu/mikles/vjm scrapbook/index.html(user:web-access,pswd:nadamucho).Thenumberofisolatedmassivestarswilllikelyremainsmallcomparedtothenumberofbinarysourcesbecausemymainfocusismassiveinteractingbinarystars.Also,fewisolatedmassivestarshavedetectableX-rayemission.However,sourcesmaybeaddedtoensurethatthestatisticalinformationonisolatedmassivestarsisrepresentativeofthegroup.ThelistofHMXBsandCWBswillbeexpandedpreferentially,sothatweensureabroadbaselineofstatisticsontheseobjects.Periodicitymaybeincludedinthenotes, 86

PAGE 87

Liuetal. ( 2006 )arediligentlymaintainingcatalogsofperiodicityforXRBs.Asthenumberofcompiledobjectsexpands,theobjectclassescurrentlygroupedtogethermaybefurthersub-categorized.Forexample,theCWBclassmaybedividedintobinariescontainingtwomassiveO-typestars(O+O)andbinariescontaininganO-typestarwithaWolf-Rayetcompanion(O+WR).Thesesourcesalreadyhavenoticeablespectraldistinctions(seepreviouschapter).Also,HMXBscanbedividedintocompactobjectswithsupergiantcompanions(sg+X),compactobjectswithBe-typecompanions(Be+X),SupergiantFastX-rayTransients(SFXTs),andmicroquasars.Thecompilationofstatisticsenablingustosearchforcommonsignaturesforagivensourcetype(e.g.lineratios)canopennewwindowstoourunderstandingofthesesources. 5-1 .Thegeneralhopeistohaveacompletesourceofinformationforeachsource,thoughlikelynotuniformacrosstheAtlas,duetothevagariesofwhathasbeenpublishedinliterature.Whileitisusefultobrowseindividualrecords,thepurposeandpoweroftheAtlasisultimatelytoprovidegeneralstatisticalinformationforclassesofobjects. 87

PAGE 88

5.4.1GeneralInformation:ObjectName,AlternateNames,RA,Dec,SpectralType,ObjectClassTheAtlasstoresaprimarynameforeachsource,aswellasothercatalogidenticationstoaidinanamesearchingfeature(notyetdeveloped).EachobjectisthenclassiedasaHMXB,CWB,orIsolatedMassiveStar(abbreviatedasOB).Additionalobjectclassesmaybeaddedinthefuture.The\spectraltype"eldfurtherrenestheobjectclassbyidentifyingthespectraltypeoftheisolatedstarorthetwocomponentstarsinthebinarysystem.Forinstance,ifthecompactobjectinaHMXBsystemisunknown,thespectraltypeislistedasOB+X.Ifthecompactobjectisknownandisaneutronstar,thespectraltypeislistedasOB+NS.Theletters\OB"indicatethatthedonorisanunclassiedmassivestar.Theexactspectraltypecanbeplacedinthateldaswell. 88

PAGE 89

Cardellietal. ( 1989 )reddeninglawassumingRV=3:1.IfanX-rayvalueofNHisknown,Iusetheestimateof PredehlandSchmitt ( 1995 )toconverttheX-rayvaluesofNHtoopticalextinction,AV. 89

PAGE 90

90

PAGE 91

Cox 2000 ).Thisnormalizationisreasonablebecausetheaccretiondiskandwindsareaminorcomponenttotheoverallinfraredluminosityofsystemscontainingmassivestars. 91

PAGE 92

5-1 ).UsingtheHTMLform,theviewercanthenselectavarietyofparameterstoincludeinatableorplot.Forexample,byselectingX-rayandK-bandluminosity,onecanquicklymakeatableoftherequestedstatisticsthatincludeseachobjectinthecatalog.Itisalsopossibletorestricttheoutputtabletoasingleclassofobjects.The\viewcommonstatistics"catalogbypassesthestepofselectingstatisticsandproducesatablethatincludesallofthecatalogobjectsandcontainsbasic,X-ray,IR,andreddeningdataforthesources. 92

PAGE 93

5.6.1CharacterizinganUnclassiedSourceSpectrumTheAtlaswasinspiredbyanddesignedtoputthesourceEdd-1intocontext,andisbasedontheinformationcollectedintheprevioustwochapterstoachievesaidgoal(seespecically,Fig. 3-6 andTable 3-4 .TheAtlasisthusimmediatelyapplicabletocharacterizingnewsourcespectrainsimilarmanners.OnecaneasilycomparetheX-rayandIRluminosityofanewhigh-masscandidatestellarsourcetoothersofknownclassications,orndsourcesthatshareaspeciccharacteristic,suchasHeliumwindorprominentIronemission.Thecharacterizationofanunclassiedsourceisapowerfulsteptowardunderstandingthenatureoftheunderlyingphysicalprocessesatworkinthesystem. LewinandvanderKlis ( 2006 ),whoconcludedthattheemissionisrelatedtotheaccretionprocessneartheinnermoststablecircularorbit.TheFe-XXVemissionintheCWBCarinaeisdescribedasarisinginthewindcollisionregionofthesource(seeChapter3).However,emissionasstrongasthatobservedinEdd-1(orCar)isnotcommonamongeitherHMXBsorCWBs.BysimplyrecordingthestrengthoftheIronemissionlineineachX-raysource,theAtlasallowsustotargetobjectswithsimilarIroninordertocompareenvironmentsofsystemsthatproduceprominentIronlinesandthosethatproduceweakemissionornone 93

PAGE 94

3-6 ,IcomparedtheX-rayandIRluminosityofvariousHMXBs,CWBs,andmassivestarsthatarethefoundationofthisAtlas.Withtheexceptionof-Carinae,theCWBsobservedtodatesubtendarelativelycompactrangeofX-rayandIRluminositiescomparedtotheHMXBs.Thisleadsonetowonder:Isthisanauthenticrepresentation?Whywouldthatbe?MyhopeisthattheAtlaswillleadtomanyareasofinquirysuchasthis. 94

PAGE 95

Screenshotofdatabasestatisticselectionpage. 95

PAGE 96

StatisticDescription BasicInformation InfraredData JMagMagnitudeinJ-bandHMagMagnitudeinH-bandKMagMagnitudeinK-orKS-bandAlternateMagnitudesIfthevariabilityrangeisknownorifcontradictingreportsexistReference1(IR)ReferencefortheInfrareddataReference2(IR)AdditionalreferencefortheInfrareddata ReddeningData 96

PAGE 97

X-rayData log(LX;burst)(ergcm2s1)ThecalculatedX-rayluminosityinahigh-uxstatelog(LX;qui)(ergcm2s1)ThecalculatedX-rayluminosityinalow-uxstate(ortheknownluminosityofonlyoneobservationexists)X-rayEnergyrange(keV)TheenergyrangeoftheX-rayspectrumthatwastfortheabovemodelingX-rayModelThemodelusedtoestimatetheluminosityX-rayThermalTemperature(keV)Ifathermalblackbodywasused,thetemperatureofthemodelX-rayNH(cm2)TheabsorptionvalueusedinthemodelComments(X-ray)CommentsrelevanttoX-raymodelingorvariabilityReference1(X-ray)ReferencefortheX-raydataReference2(X-ray)AdditionalreferencefortheX-raydata SpectralInfo DoestheIRSpectrumExist?(yesorno)Reference:IRSpectrumReferencefortheIRspectrumshoulditexistDoestheX-raySpectrumExist?(yesorno)Reference1:X-raySpectrumReferencefortheX-rayspectrumshoulditexistReference2:X-raySpectrumAlternatedreferencefortheIRspectrumshoulditexistDoestheOpticalSpectrumExist?(yesorno)Reference:OpticalSpectrumReferencefortheopticalspectrumshoulditexistDoestheAnotherSpectrumExist?(yesorno) SpectralLineInfo DoesthespectrumhaveHydrogen?yesorno;relevanttoanywavebandDoesthespectrumhaveHelium?yesorno;relevanttoanywavebandDoesthespectrumhaveFeI?(equivalentwidth?)yesorno;relevanttoX-raywaveband 97

PAGE 98

Wind/PCygniprole Iswindobserved?yesorno;relevanttoanywavebandElementsinwindPrimarilyexpectHIorHeIVelocityofwindThePCygnivelocitymeasuredinkm/sReference(wind)ReferenceforthewindelementsandvelocityshoulditexistCommentsonspectraGeneralcommentsrelatedtothespectraavailableonthesource 98

PAGE 99

Fenderetal. 1999 ).AmicroquasarisanX-raybinarythatexhibitsbothdiskaccretionandrelativisticjetejections.GRS1915+105appearedintheX-rayskyin1992asanX-raytransient( Castro-Tiradoetal. 1992 ),butunlikeanyX-raytransientknownbeforeitorsince,ithascompletelyfailedtodisappear.Itslongevityhasallowedextendedstudyofthesourceandjetproductionevents.GRS1915+105isclassiedasaLow-MassX-rayBinary(LMXB)withablackholecounterpart.Themassofthecompactobjectisestimatedat14MandthecompanionisK-MIIIstar( Greineretal. 2001 ).Thelow-masscompanionwasidentiedduetoabsorptionbandspresentintheinfraredspectrum.LMXBssuchasGRS1915+105accreteprimarilyviaRocheLobeoverow.RadioimagingofGRS1915+105showninFigure 1-4 ,showsthatjetejectaassociatedwiththesourcemoveawayfromthecentralobjectat0:98+0:20:5c( MirabelandRodriguez 1994 ; Fenderetal. 1999 ).Continuedradio,infrared,andX-rayobservationshaveledtotheclassicationofGRS1915+105'sradiojetsintothreetypes:(1)\steady"radiojets,(2)discreteplasmaejectioneventsof20-40minutedurationintheinfraredandradio,and(3)largesuperluminaljetsakintothe1994eventthatearnedGRS1915+105thename\microquasar."ThesteadyjetisassociatedwitharadioandX-ray\plateau"state,meaningthereisrelativelylittlevariationinthelightcurvesintheseregimes.Thesteadyjetsareopticallythick,extendfor<200AUfromthecentralsource,andhaveavelocityof0:10:4c( Yadav 2006 ).Thediscretejetsareassociatedwithradio,infrared,andX-rayoscillatorybehavior.Thediscretejetsarecompact(extendingafewhundredAU),buthasavelocitycloserto90%ofthespeedoflight.Theyrangeinstrengthform5-200mJyintheinfraredandradio(e.g., Eikenberryetal. 1998 ; Rothsteinetal. 2005 ; 99

PAGE 100

, 2006 ).Thesuperluminalradiojetscanbeupto1JyintheradioandextendseveralthousandAU( MirabelandRodriguez 1994 ; Fenderetal. 1999 ).OnlyahandfulofX-raybinariesareclassiedasmicroquasarsdisplayingjetactivity,though Fenderetal. ( 2005 )havesuggestedthatallXRBswillemit\superluminal"jetsatsomepointduringtheirevolution.ThepresenceofajetcanbeinferredbyobservationsinthehardX-ray,IR,orradioregimes.Whilejetejectionstakedecadestoobserveinextragalacticquasars,microquasarsshowsimilareventsinamatterofhoursordays.Thus,microquasarsareanexcellentlaboratorythroughwhichwecanstudyaccretionphysics,relativisticjetformation,anddisk-jetinteractions.Becausedisk-jetinteractionsoccurclosetothecentralcompactobject,understandingtheinteractionwillhelpusprobethelimitsofGeneralRelativityandanswerthequestionofwhathappensattheedgeofablackhole. Bellonietal. ( 2000 )deneatotaloftwelveX-raylightcurveclassesforGRS1915+105,distinguishablebygeneralappearance,countrate,andX-raycolor(relatedtohardness).InFigures 6-1 6-2 ,and 6-3 ,Ishowthe1-secondX-raylightcurvesofthreeclasses:,,and.IfocusonthesethreeclassesofeventsbecausesimultaneousX-rayandinfraredanalysisrevealthepresenceofdiscreteoscillatingjetsrecurringon30minute 100

PAGE 101

Eikenberryetal. 2000 ; Rothsteinetal. 2005 ).The-classX-raylightcurveischaracterizedbyM-shapedintervalslastingafewhundredseconds,andspectrallyharderdipslasting100-200seconds(Fig 6-1 ).Boththe-and-classlightcurveshaveextendedspectrallyharddips,butthelengthofthe-classharddipisnearlytwicethelengthofthe-classone.The-classlightcurveshaveadipof1200secondsfollowedbystrongX-rayoscillations,markingtheendofthehardstate(seeFig. 6-2 ).The-classlightcurvesaremorecomplex.Thediplasts500700secondsanditsendingismarkedbyaspectrallysoftspikefollowedbyanearlymonotonicrise.Afterreachingapeakux,theX-raysbeginrapid,large-amplitudeoscillations(seeFig. 6-3 ).Itiswellestablishedthatthespectrally-harddipsthatfrequentlyappearintheX-raylightcurvesofGRS1915+105areassociatedwithinfraredandradioares( Eikenberryetal. 1998 ; Mirabeletal. 1998 ; FenderandPooley 1998 ; Klein-Woltetal. 2002 ).Thesespectrally-harddipsarealsoassociatedwith210Hzvariablequasi-periodicoscillations(QPOs).QPOsareaformofrapid,aperiodicvariabilityobservedinbothblackholeandneutronstarbinaries.QPOsarenotrelatedtoperiodicvariabilitysuchaspulsations,dips,oreclipses( Lewinetal. 1997 ).ItisbelievedthatQPOsprobetheaccretionowdynamicsattheinnermostpartoftheaccretiondisk.IseektounderstandhowandwhytheQPOisrelatedtothejetejection.AQPOcannotalwaysbeidentiedwithinanX-raylightcurvewiththenakedeye.Toresolvetheoscillations,exceptionaltimingresolutionisrequired;suchresolutionisavailablethroughtheRossiX-rayTimingExplorer(RXTE)ProportionalCounterArray(PCA).With8-millisecondresolutionlightcurves,Iamabletoobservetheevolutionofthepowerspectrumandtrackthefrequencyofanoscillationasitchangeswithtime.Thepowerspectrumofeachsampledintervaliscomprisedofanumberofvariabilitycomponents.BroadfeaturesarecallednoiseandnarrowfeaturesarecalledQPOs.For 101

PAGE 102

6-4 ,Ishowaportionofthe8-millisecondlightcurveofGRS1915+105duringaspectrallyharddipandtheFFToftworegionseachlasting4secondsandseparatedby4seconds.TheQPOisclearlyvisibleineachpowerspectrumandthefrequencyoftheoscillationshiftsnoticeably.Thus,itisclearthatestimatingtheQPOfrequencyoverlongertimesamplings(>8s)canblurtheoscillationfrequencyandmaketheQPOmorediculttoidentifyandmodel.ByapplyingFourieranalysistotheentirelengthoftheharddip,IcanstudyQPObehaviorwithoutmakingassumptionsabouttheunderlyingprocess.FromseveralRXTEobservationsofGRS1915+105takenon14-15August1997,9September1997,and27-28July2002(seeTable 6-1 ),IextractProportionalCounterArray(PCA)Standard-1lightcurvesusingFTOOLSv5.3andidentifyregionsshowingahardX-raydip.Iexamine17regionshere:nineofthemare-class,threeare-class,andveare-class( Bellonietal. 2000 ).Theyarenumberedsequentially-1through-9,-10through-12,and-13through-17.Fortheseventeenregions,Iextractbinnedmode8-millisecondlightcurvesinthe213keVrangeand4-secondresolutionbinnedandeventX-rayspectrainthe225keVrange.UsingXSPECv11.3.0,Itthespectrawithacombinationofabsorbedmulti-temperaturediskblackbodyandpowerlawmodels.ThisisacommoncombinationofmodelsfordescribinganX-raybinarysystem.Themulti-temperatureblackbodymodelsthethermalemissionfromageometricallythinaccretiondiskaroundacompactobject,recognizingthatthetemperatureofthediskincreasesatlowradiiclosetothecompactobject.ThismodelischaracterizedbytheX-rayux,andthetemperatureandradiusattheinnerpartoftheaccretiondisk.Thepowerlawcomponentdominatesthehard(>4keV)X-rays,andisattributedtosynchrotronemissionorothernon-thermalprocesses.Inouranalyses,weonlyusetswhichhavea2<2. 102

PAGE 103

Ransometal. 2002 ).Finebinningallowsmetocalculateahigher-resolutionFourierresponsebyinterpolatingresponsesatnon-integerfrequencies.InFigures 6-5 and 6-6 ,Ishowagray-scaledne-binnedPDSandoverplotaone-secondresolutionlightcurve.ThePDSisgray-scaledaccordingtotherelativepowerofthespectrumsuchthatdarkerregionsindicatepowerpeaks.TheQPOpeakfrequencyappearstoevolvesmoothlyduringtheprogressionoftheharddip,developingaU-shapeinthe-classlightcurveswhiletracingthetotaluxinthe-and-classes.IdeterminetheQPOfrequencybyttingaMoatfunctiontothePDSinthe210Hzfrequencyrange.TheMoatfunctionisaLorentzianmodiedwithavariablepowerlawindex( Moat 1969 ).TheQPOisconsideredtobedetectedifithasaqualityfactorQ==FWHM>2,whereisthecentroidfrequencyoftheLorentzianandFWHMisthefull-widthathalfmax.ThismethodyieldsaconsistentandrepeatableQPOdetection.However,low-frequencynoiseinthepowerdensityspectrummaystillleadtofalsedetections.Inthenextchapter,Idiscussthene-tuningrequiredtoselectthecoherentQPOineachlightcurveclass. Eikenberryetal. 1998 ; Mirabeletal. 1998 ; Rothsteinetal. 2005 ). Eikenberryetal. ( 1998 )observedaone-to-onecorrespondenceofX-raydipstoinfraredaresat30-minuteintervals.IshowasampleX-raylightcurvewithsimultaneousinfrareddatainFigure 6-7 withtheQPOfrequencyover-plotted.TheX-rayobservationswerecarriedoutusingtheRXTEProportionalCounterArrayandsimultaneousinfrareddatawasobtainedatthePalomar200-inchtelescope.Inthegure,thesolidlineistheX-rayux.Fortherst500seconds 103

PAGE 104

Eikenberryetal. ( 2000 )asclassCandclassB,withtheuxofthearesrangingfrom30200mJywhendereddenedby3.3magnitudes.IndividualclassCinfraredaresaresmaller(510mJy)andhavebeenobservedtobeindividuallyassociatedwiththe-class( Eikenberryetal. 2000 ).Idiscussthedistinctaringbehaviorsoftheseclassesmorefullyinthenextchapter. 104

PAGE 105

Atypical-classlightcurveat1sresolution.ThecurvehasacharacteristicM-shapeandbecomesspectrallyhardatlowux.TheobservationIDisderivedfromthenamingschemeinTable 6-1 105

PAGE 106

Atypical-classlightcurveat1sresolution.The-classcurvehasnotablylongerharddipsandshortersoftaringtimes.TheobservationIDisderivedfromthenamingschemeinTable 6-1 106

PAGE 107

Atypical-classlightcurveat1sresolution.Theharddipendswithasoftspikethatleadsintorapidoscillations.TheobservationIDisderivedfromthenamingschemeinTable 6-1 107

PAGE 108

(top)An8mslightcurveofGRS1915+105duringaspectrallyharddip.(bottom)Thepowerspectrumofselect4sregionsofthelightcurve.TheQPOfrequencycanchangenoticablyovershorttimescales. 108

PAGE 109

Agray-scaled,ne-binnedpowerdensityspectrumforthe-and-classlightcurves.Iplotthe1-secondresolutionlightcurveoverthePDS. 109

PAGE 110

Agray-scaled,ne-binnedpowerdensityspectrumforthe-classlightcurves.Iplotthe1-secondresolutionlightcurveoverthePDS. Figure6-7. AsamplelightcurveshowingtheX-ray,theinfrared,andtheQPObehaviorduringajetejection.ThesolidlineistheX-raylightcurve.Thedotsaretheinfrared,andthecrossesaretheQPOfrequency.TheQPOevolveswhiletheX-raysareinaspectrallyharddip.AfterthespikeintheX-ray,theQPOdropsoandaninfraredarebegins.Thisisinterpretedasaplasmaejection. 110

PAGE 111

SetIDRXTEDATAIDDateObservedStartTime Note.|ObservationIDsanddatesofthe17epochs.TheStartTimeindicateswherespectralttingbegan.TheIDsarebasedonthe Bellonietal. ( 2000 )classications. 111

PAGE 112

Eikenberryetal. 1998 ; Mirabeletal. 1998 ; Rothsteinetal. 2005 ).WhiletheX-raysarerapidlyoscillating,theinfraredarepeaksanddecayswithouttherapid,largeamplitudevariationsseenintheX-ray.Theinfraredaresassociatedwiththeseclassesaredenedby Eikenberryetal. ( 2000 )asclassBandclassC.WhenseveralclassCinfraredaresoccurinrapidsuccession,theycanappearasalargerare. Rothsteinetal. ( 2005 )showthistobethecaseforthe-classlightcurvesinmysample,wheretheassociatedinfraredaresrangefrom1030mJy. Rothsteinetal. ( 2005 )showedthatifeachsoftX-rayarewereassociatedwitha510mJy(classC)sub-are,thenthedurationandstrengthoftheoverallinfraredarewouldbeexplained.Bythatanalysis,thepredictedcollectionofinfraredsub-aresassociatedwitha-classlightcurvewouldcontributeonlyasmallfractionoftheoverallobserved60200mJyinfrareduxfromGRS1915+105.Thissmallcontributionwasobservedby Eikenberryetal. ( 2000 )asan\infraredexcess."AlthoughtheperiodoftheriseanddecayoftheprimaryclassBinfraredareisnotcoupledtotheperiodoftheX-rayoscillations,theoveralldurationoftheinfraredexcessiscoupled( Eikenberryetal. 2000 ; Rothsteinetal. 2005 ).Thedierencebetweenthedip/arecyclesassociatedwith-vs.-classcurvesbecomesthepresenceofalargeprimary 112

PAGE 113

6-7 ).Itisbelievedthatbotharelinkedtotheunderlyingcauseoflargerplasmaejections.Severalauthorshaveshownthatinfraredarestendtobefollowedbyradioaresofsimilarmorphology( FenderandPooley 1998 ; Mirabeletal. 1998 ).Thesequenceofaspectrally-harddip,aninfraredare,andthenaradioare,isgenerallyassociatedwithaplasmaejectionfromthesource.TheejectionisobservedastheX-raystransitionfromaspectrally-hardtoaspectrally-softstate.Duringtheharddipthatprecedesanejectionevent,a210HzQPOisalwaysobserved( Bellonietal. 2000 ).WhiletheQPOispresent,thediskX-rayemissionisgreatlyreducedinthe225keVrangeandtheX-rayluminosityisdominatedbytheharderpowerlawux.SeveralgroupshavefoundtheQPOpeakfrequencypositivelycorrelatedtobothpowerlawandthermaldiskcomponents,suggestingthattheQPOarisesinthesamelocationastherelatedemissionoriscausallyrelatedtoit( Munoetal. 1999 ; Ferocietal. 1999 ). Markwardtetal. ( 1999 )observedthatatfrequenciesabove4Hz,theQPOismoststronglycorrelatedtothethermaldiskcomponent,specicallytheblackbodydiskux.Atlowerfrequencies,theQPOshowsabroadassociationwiththepowerlawux. Munoetal. ( 1999 )alsoshowtheQPOisstronglyassociatedwithdisktemperatureduringtheharddip.Afterthetriggerspike,thepowerdensityspectrumbecomessmooth( Markwardtetal. 1999 ).Inthischapter,IexaminethebehavioroftheQPOduringtheharddipandtherelationshipofitsbehaviortosubsequentinfraredares.InSection7.2,Idiscussmyobservationsandanalysistechnique.IcalculateanddiscussthecorrelationoftheQPOpeakfrequencytospectralfeaturesinSection7.2.1andcommentonthetimeevolutionoftheQPOduringthedipinSection7.2.2.InSection7.2.3,Iexaminethemorphologyofthetriggerspikeinthe-classlightcurveandtheindicationofacontinuumofbehaviorsfromthisclasstothe-class,whichdoesnotshowatriggerspike.InSection7.2.4,I 113

PAGE 114

6-1 )andusethereductionproceduredescribedinthepreviouschapter.IrenemyQPOdetectionmethodtoaccountforlowfrequencynoiseintheobservations.In-classcurves,IrequiretheQPOfrequencytobebetween2.6and10Hztoavoidcontaminationbylowfrequencynoise.In-classcurves,Iallowdetectionsinthefull210Hzrange.However,alsovisibleinthePDSofthe-classisalow-frequencynoisecomponent,themaximumfrequencyofwhichtendstoriseabove2Hzatthebeginningandendofthedip.Toavoidspuriousdetectionsfromthiscomponent,IrequirethateachdetectedQPOdoesnotvarytoosharplyfromtheQPOsatsurroundingtimes.Imakethefollowing\by-eye"assessments,keepingmyrenementsasgeneralaspossibleforthesakeofrepeatability.Forpointsattimest>350seconds,Itakeanaverageoftheprior10detectedfrequencies,avg.Thenextdetectionisrequiredtobegreaterthan75%ofavg.Fort<150seconds,QPOsaretreatedsimilarly,thoughavgisdeterminedbyusingpointsfollowingthereferencepoint,asopposedtothosepreceedingit.Althoughomittingpointsasnon-detectionscanleavepartoftheQPOevolutionunder-sampled,IbelieveIhavesucientrepresentationfromtheregionstoensureasoundqualitativeresultandareasonablequantitativeone.InadditionthismethodyieldsaconsistentandrepeatableQPOdetection.Itispossiblethatadjustingtherangeinthismannercausesmetoomitreal,unresolvedoscillationsintheQPO,butvisualinspectionshowsthatthesepointsarenotpartoftheprimaryU-shapedQPOfeatureIwishtofocuson.ThisshapeisapparentinthePDSshowninFigure 6-5 .InFigure 7-1 ,IplotthedetectedQPOsaslledcirclesagainsta1-secondresolutionlightcurve.Forillustrativepurposes,IshowpowerpeaksinthePDSthatdonotmeetthedetectioncriteriaasopencircles. 114

PAGE 115

7-1 .Inallofthecases,IobserveahighlysignicantcorrelationbetweentheQPOfrequencyandthetotalux,whichisgenerallystrongerthanthosetoindividualblackbodyorpowerlawfeatures.Whenconsideringmodel-specicspectralfeatures,elevenoutoftwelveshowstrongcorrelationstopowerlawux,andonlysixoutofthoseelevenshowstrongcorrelationstoblackbodyfeatures.Innineoutofthoseelevencases,thecorrelationtothepowerlawuxisstrongerthanthecorrelationtoanyblackbodyfeature.ThisgivestheapparentresultthattheQPOfrequencyismorefundamentallytiedtothepowerlawcomponentthantheblackbodycomponent.Figure 7-2 showsascatterplotoftheQPOfrequencyversusthepowerlawuxforthetwelvecases.Thepointsmarkedwithtrianglesaredetectionswithintherst100secondsofthedipandthesquaresarefromthelast100secondsofthedip.Thelineisbasedonafourth-orderttothetimeevolutionoftheQPOfrequencyandthepowerlawuxandtracestheapproximatepathoftheevolution.Theopencirclesarepowerpeaksbetween2and10Hz,whereaQPOwasnotdetected(seeabove).Thesepointsaregenerallynotapartoftheobservedtrends.In-1through-5(August1997)theevolutionoftheQPO-powerlawuxrelationisgenerallytighterandthecorrelationsstronger(r0:80).The-10through-12curvesalsoshowatightcorrelation,thoughtheQPOfrequencyvariesoverasmallerrangeoffrequencies.Likethe-classcurves,Ionlyhavedetectionsfor-8overasmallrangeof 115

PAGE 116

7-2 ,thehysteresisisextremeandlowersthecorrelationcoecientdrasticallydespitetheinitiallylinearcorrelationatthedipentrance.Thesquareswhichtracethelast100secondsofthedipshowtheQPOfrequencyrisingsharplywhilethepowerlawuxisrelativelyconstant,suggestingadecouplingofthefeatures.In-1through-7,thecorrelationbetweentheQPOfrequencyandthepowerlawuxisaccompaniedbyasimilarlystrongcorrelationtotheblackbodyuxandblackbodydisktemperature.InFigure 7-3 ,Iseeastrongcorrelationtotheblackbodyuxismostprominentabove4Hz,apointalsoobservedby Markwardtetal. ( 1999 ).However,thecorrelationdoesnotlastoverthebroadpartoftheharddipwheretheQPOcanchangesignicantlywhileatrelativelyconstantblackbodyux.In-8and-9,wheretheblackbodyuxappearsuncorrelated,therearefewerQPOdetectionsabove4Hzduetotheriseinlowfrequencynoiseatthedipexit.IshowtheQPOfrequency-blackbodytemperaturescatterplotsinFigure 7-4 .ClearlytheQPOinthe-classcurvesisnotbelievablycorrelatedtotheblackbodytemperature.In-1through-5and-8,thecorrelationstotheblackbodytemperaturefollowasteadytrend,butsueraslightlywiderdispersionintheirevolutionthanthepowerlawux.In-6,-7,and-9,Iseeasimilarhysteresistothatobservedinthepowerlawux.Tountanglethepossibleinterplayoftheblackbodytemperatureandpowerlawux,Iapplyapartialcorrelationanalysis.ThepartialcorrelationcoecientsarelistedinTable 7-2 .Inthetable,Icalculatecoecientsforfourscenarios: 1. TheQPOfrequency|powerlawuxcorrelationremovingtheeectofblackbodyux. 2. TheQPOfrequency|powerlawuxcorrelationremovingtheeectofblackbodytemperature. 116

PAGE 117

TheQPOfrequency|blackbodyuxcorrelationremovingtheeectofpowerlawux. 4. TheQPOfrequency|blackbodytemperaturecorrelationremovingtheeectofpowerlawux.Intherstscenario,Iseethatwiththeblackbodyuxremoved,thecorrelationtopowerlawuxisstillstrong.Thisisexpectedbecausetheblackbodyuxpoorlyexplainsvariancesatfrequenciesbelow4Hz(seeFig. 7-3 andTable 7-1 ).Inthesecondcase,Indtheremovalofblackbodytemperaturehasamoresignicanteect.Inmostcases,theQPOfrequency|powerlawuxcorrelationisweak,thoughbelievable(r>0:4).ThismeansthatafterremovingvariationsintheQPOfrequencyandpowerlawuxthatcanbeexplainedbyvariationsinblackbodytemperature,variationsintheQPOfrequencyremainthatcanbeatleastpartiallyexplainedbyvariationsinthepowerlawux.Highsignicanceisseenin-classcases,whichisexpectedduetotheirweakerdependenceonblackbodytemperature.Inthethirdcase,Itestthecorrelationtotheblackbodyuxafterremovingthepowerlawux.Thesecorrelationsarebelievableandoccasionallystrong,suggestingthatacombinationoftheblackbodyandpowerlawsourcesisrequiredtoexplaintheQPOfrequency.Inthenalcase,theremovalofpowerlawuxfromtheQPOfrequency|blackbodytemperaturecorrelation,Iseethatpartialcorrelationcoecientdropsbelowsignicanceinhalfofthecases.ThismeansthatoncevariationsintheQPOfrequencythatcanbeexplainedbyvariationsinthepowerlawuxareremoved,novariationremainsthatcanbeexplainedbyblackbodytemperature.Insummary,theQPOfrequencyisoftencorrelatedtothepowerlawux.For-classlightcurves,whenthiscorrelationisstrongest,Itendtondacorrelationtotheblackbodyuxandblackbodytemperatureaswell.Incaseswherethereisaslightlyweakercorrelationtothepowerlawux,thecorrelationtotheblackbodyfeaturesislesspredictableandahysteresiseectisvisibleinthepowerlawuxandblackbodytemperaturerelations.Incontrast,-classQPOstendtohaveastrongcorrelationtothe 117

PAGE 118

7-2 ),itislikelythatthespectralevolutionisintimatelytiedtotheQPOtimeevolution.TheQPOevolutionatthebeginningofalltheharddipsissimilar,eachbeginningwiththesuddenappearanceofaQPOat610Hz.ThisQPOwillsmoothlydroptobetween23Hzwithinsecondsoftheinitiationoftheharddip.AftertheX-raytriggerspike,theprimaryU-shapedQPOfeaturewilldisappear,replacedbyoccasionalpowerpeaks.InTable 7-3 IlisttheminimumQPOfrequencyduringthedipandthetimespentnearthatfrequency.Basedontheobservablevariations,Ithendividethelightcurvesintothreebroadgroups:Group1{Figure 7-5 showsanexampleofthe-classlightcurvesinGroup1.AseriesofX-rayoscillationscalmintoalow,spectrally-harddipwithinabout100seconds.Duringthistime,aQPOarisesat68Hz.Followingtheintensitydropofthelightcurveandmorespecicallythepowerlawux,theQPOfallssteadilyto2Hz.Itremainsatthisfrequencyforover150seconds(seeTable 7-3 ),afterwhichthepowerlawuxandQPOfrequencybeginaslowrise.TheU-shapedQPOvanishesaftertheX-raytriggerspike.Thetotallengthofthedipisontheorderof600seconds,and30%ofthattimeisspentattheminimumfrequency.Group2{Thisgroupisalsocomposedof-classlightcurveswithsimilarspectralbehaviortoGroup1(seeFig. 7-6 ).However,theQPObehaviorinthisgroupissomewhatdierent.WhileinGroup1,theQPOfallsoto2Hzandlingers,inGroup2theQPOimmediatelystartstoriseagain.InTable 7-3 ,IshowthatwhiletheGroup1eventsremainneartheminimumfrequencyfor>150seconds,theGroup2eventsremainfor<100seconds(about15%ofthediplength).Thefollowingriseinfrequencyisthestart 118

PAGE 119

7-2 )andlikelyindicatesthattheQPOandpowerlawuxhavedecoupled.Inaddition,thesecasesseetheriseofalowfrequencynoisecomponentabove2HzastheQPOweakensinamplitudeandrisesrapidlyinfrequency.Whiletheselowfrequencypointsareexcludedfromcorrelationanalysisasbeingassociatedwithlowfrequencynoise,itispossiblethattheyrepresentanincreaseinrapid,unresolvedQPOoscillations.ThedierenceinbehaviorsinthersttwogroupsismostremarkablebecausetheirX-raylightcurvesandspectralbehaviorsaresosimilar.Group3{Thenalgroupcontainsthe-classlightcurvesrepresentedinFigure 7-7 .Intheseevents,theharddipissurroundedbyX-rayoscillationsbutnoindependentterminalspikeisobserved.Thetotallengthofthedipis1200secondsandtheQPOdisappearswhenthedipends.ThiscaseissimilarinshapetoGroup1QPOevolution,thoughtwiceaslong.The-classdiersinthatonenteringthediptheQPOfrequencylevelsoat3Hz.Overall,thefrequencyvariesoverasmallerrangethanthatseenintheothertwogroups. 7-8 showsaone-secondresolutionlightcurveforeachofthenineobservedspikesinatimerangeof55secondsbeforeandafterthemaximum.ThelightcurvesaretwithadoubleGaussianplusapolynomialoftheform:f=A0ez2A+B0ez2B+C0+C1twherezA=(tA)=AandzB=(tB)=B.isthemean(center)Gaussianvalueandisthestandarddeviation. 119

PAGE 120

7-4 .Itisinterestingtonotethattheintegratedcountrate(fint)overthefull-width-half-maxofthepeakissimilarforalldatasets,butinmostotheraspectsofthet,theAugust(Group1)andSeptember(Group2)datahavedierentproperties. 1. TheGroup1datahaveahighernormalizedamplitude:AGrp1>AGrp2. 2. TheGroup1dataaremoresymmetricwhiletheGroup2datashowsasharpcut-oafterthepeakux(seeFig. 7-8 ). 3. TheGroup1dataspikesarenarrower:Grp1A0whileCGrp21<0.WhileallGroup2eventsshowanegativeunderlyinglightcurveslope,itisinterestingthattwoofthem(-6and-7)haveanearlyatslopewhiletheothertwo(-8and-9)haveadecidedlynegativeslope.FromFigure 7-2 InotethatthetwowithnearlyatslopeshaveaslightlymorevisiblehysteresisbecausetheyhavemoreQPOdetectionsabove4Hz.Thelowfrequencynoisecomponentdoesnotriseasstronglyabove2HzandtheQPOfrequencyshowsastrongercorrelationtoblackbodytemperature.Inaddition,theterminatingspikesof-8and-9lookmuchmoredisturbedthanthoseofotherlightcurves(Fig. 7-8 ).Thedierencesintriggerspikemorphologysuggestthatthevariation 120

PAGE 121

7-2 ,IseethatthiseventseemstobeabridgebetweenthesharplinearcorrelationsofGroup1andthedivergentshapesofGroup2.Iclassifythe-6lightcurveasGroup2becausetheQPOfrequencyclearlydeviatesfromtheinitialregressionandbecauseofitstriggerspikemorphology.Thetriggerspikeof-6iswiderandmoreasymmetricthanGroup1events.Inaddition,theatunderlyinglightcurveof-6suggeststhatitismoreappropriatelyassociatedwithGroup2thanGroup1events.Thisbeingsaid,Iacknowledgethatthegroupsarenotabsolute,butarelikelypartofacontinuumofbehaviors. Eikenberryetal. ( 1998 )showthatthesedip/spikepairsareusuallyfollowedbylargeinfraredares.Theeventsobservedrangedfrom60to200mJy.AlthoughtheriseandfallofthearedoesnotcorrespondtotheperiodofX-rayoscillation,aweakinfraredexcesswhichlaststhroughouttheperiodoftheX-rayoscillationsisobserved. Eikenberryetal. ( 2000 )and Rothsteinetal. ( 2005 )explainthisexcessasthesuperpositionofmanyfaintinfraredareseachontheorderof10mJy.ThedominantinfraredareisassociatedwiththeX-raytriggerspike. Mirabeletal. ( 1998 )observedasingleinfraredareeventassociatedwithmy-7curve,whichreachedanamplitudeof30mJy.SimultaneousinfraredcoverageisnotavailablefortheotherharddipsinGroup2.Inallthree-classlightcurvesofGroup3,theharddipisfollowedbya30mJyinfraredare( Rothsteinetal. 2005 ).In Mirabeletal. ( 1998 ),an-classeventisobservedtobefollowedbyaarethatpeakedat10mJy. Rothsteinetal. ( 2005 )showedthatthe30mJyarescanbeexplainedasasummationofClassCsub-ares,eachassociatedwithasoftX-rayare. 121

PAGE 122

7.3.1QPOCorrelationwithSpectralFeaturesPreviousstudieshaveshownthattheQPOismoststronglytiedtothethermaldiskcomponent( Markwardtetal. 1999 ; Ferocietal. 1999 ; Munoetal. 1999 ).UsingtheSeptember1997data(my-6through-9), Markwardtetal. ( 1999 )pointedoutthatthecorrelationtodiskuxisstrongestwhentheQPOfrequencyisabove4Hz.Whilethisistrue,theQPOfrequencyisinthisrangelessthan25%ofthetime,andmostlyfallsintothisrangewhenenteringorexitingtheharddip.Duringthecourseoftheharddip,theQPOwillchangesignicantlywhiletheblackbodyuxremainsrelativelyconstant. Markwardtetal. ( 1999 )alsosaythatatlowerfrequencies,thereisanapparentbroadcorrelationwithpowerlawux.Iconrmthatthiscorrelationmayexist,andshowthatitismostapparentatlowerfrequencies.Becauseofahysteresiseect,thedeviationintheQPOfrequency|powerlawuxrelationshipismoreapparentathighfrequencies.IsuggestthataninitiallytightcorrelationisbrokenastheQPObeginstoriseinthelatterhalfofthedip.TwooutoffouroftheobservedSeptembereventsarestronglycorrelatedtopowerlawuxandadierenttwooutoffourarecorrelatedtoblackbodytemperature.Incontrast,theAugust1997(my-1through-5)eventsshowastrongcorrelationwithbothpowerlawandblackbodyfeatures-specicallythepowerlawuxandblackbodytemperature.Apartialcorrelationanalysisshowsthatifeitherthepowerlawuxorblackbodytemperatureisremoved,thecorrelationtotheotherisweakened,soitisnotlikelytheeectsofthesecomponentscanbeuntangled.Ido,however,arguethatthecorrelationtothepowerlawcomponentmaybemorefundamental,especiallysincethepowerlawuxismorestronglytiedtotheX-rayemissionatthispointandthediskcomponentisvanishinglysmall.Inaddition,the-classcurvesshowaconsistentlystrongcorrelationtothepowerlawuxandlessconsistentcorrelationtoblackbodyfeatures.Asmentionedbefore,thebulkoftheQPOchangeoccursduringtheentryintoandexitfromthedip.Whileinthedip,thespectralfeaturesremainfairlystable.Thusmuch 122

PAGE 123

Eikenberryetal. ( 1998 )wereunabletoconclusivelydetermine 123

PAGE 124

Mirabeletal. ( 1998 )eventcorrespondingtomy-7,onemightbelievethattheinfraredarestarts100200secondspriortothespike(see Mirabeletal. 1998 ,theirFigure3).NotingthattheQPOalsosignicantlyweakenscomparedtothelowfrequencynoisecomponent100200secondspriortothespike(seeFig. 6-5 )suggeststhattheoriginoftheinfraredaremaybecausallylinkedtothemechanismpoweringtheQPO,andnottiedinitiallytothesoftX-rayaringortriggerspike.ThisbehaviorsuggeststhattheQPOistiedtoamulti-wavelengthenergyrelease.ThedecouplingoftheQPOfromX-rayspectralfeaturesinthiscasesupportsthishypothesis.Inallothercases,thedisappearanceoftheQPOcoincidedwiththetriggerspikeortherstX-rayoscillation(andthustheinfraredare),sothispicturewouldbeconsistent. Eikenberryetal. ( 2000 )observedaseriesofclassCinfraredaresprecedingtheX-raysoftaresin-classdip-arecycles,sothissequencewouldnotbeentirelyunprecedented. 7-1 ).Thisregioniswherethetracksdiverge.IbelievethattheQPObehavioratthedivergentpointcanultimatelybeusedtopredicthowthedipwillend.ConsiderthethreegroupsIidentify,summarizedintermsofcauseandeect:Group1{A-classlightcurveentersaharddipphase.TheQPOfallsoto2Hzandmaintainsthatfrequencyfor>150seconds.Endresult:TheQPOfrequencyistightlycorrelatedtobothblackbodyandpowerlawspectralfeaturesandthecorrelationlaststhelengthofthedip. 124

PAGE 125

125

PAGE 126

126

PAGE 127

Anoverlayofthe1-secondresolutionX-raylightcurve(line)andne-binned4-secondQPOfrequency(circles).TimeisinsecondsandQPOfrequencyisinHz.TheopencirclesarelowfrequencypowerpeaksobservedwhentheQPOisnotdetected(seeSection7.2).Theyareshownforillustrativepurposes.Notethatfor-6through-9,theQPOspendslesstimeneartheminimumfrequency. 127

PAGE 128

ScatterplotsofQPOfrequency(inHz)withpowerlawux,Flux-PL,(in108ergcm2s1)for-and-classlightcurves.Trianglesindicatedetectionsintherst100secondsofenteringthedip.Squaresarepointswithinthelast100secondsbeforeexitingthedip.Thelineisafourth-orderbestttothetimeevolutionofthefeatures.Notethatfor-1through-5,astronglinearcorrelationisapparent.For-6,-7,and-9,thecorrelationweakensandIseedierentdegreesofhysteresis.Theopencircles(non-detectionsoftheQPOasdenedinFigure 7-1 )donotcontributetotheapparenthysteresispattern.In-8,therangeoffrequenciesissignicantlyless,probablyduetoincreasednon-detections.ThisisevidencethatacontinuumofQPObehaviorsexistswithinthe-classlightcurves.In-classlightcurves,acorrelationexistsoverasmallerrangeoffrequenciesaswell. 128

PAGE 129

ScatterplotsofQPOfrequency(inHz)withblackbodyux,Flux-BB,(in108ergcm2s1)for-and-classlightcurves.SymbolsareasinFigure 7-2 .Acorrelationisobservedabove4Hz,butbelow4Hz,theQPOcanchangesignicantlywhiletheblackbodyuxremainsrelativelyconstant. 129

PAGE 130

ScatterplotsofQPOfrequency(inHz)withblackbodytemperature,Temp-BB,(inkeV)for-and-classlightcurves.SymbolsareasinFigure 7-2 .The-classlightcurvesclearlyshownocorrelation.Inmostofthe-classlightcurves,acorrelationisseenbutwithawiderdispersionthanthatofthepowerlawuxrelation.Hysteresisisparticularlyapparentin-7and-9. 130

PAGE 131

ThetimeevolutionofaGroup1event.Thetoppanelshowsa1-secondresolutionX-raylightcurve.Thenextpanelshowsthene-binnedQPOfrequencyandtheopencirclesarelowfrequencypowerpeakswheretheQPOisnotdetected.TheX-raypowerlawux(inunitsof108ergcm2s1)andblackbodydisktemperatureareshownatfour-secondresolution.TheQPOfrequencyismoststronglycorrelatedtothepowerlawuxduringtheharddip,butisalsocorrelatedtotheblackbodydisktemperature. 131

PAGE 132

ThetimeevolutionofaGroup2event.PanelsareasdescribedinFigure 7-5 .Inthiscase,theQPOfrequencybeginsrisingwhilethepowerlawuxandtheblackbodytemperaturearerelativelyconstant. 132

PAGE 133

ThetimeevolutionofaGroup3event.PanelsareasdescribedinFigure 7-5 .Theharddiplasts1200seconds{twicethelengthoftheGroup1and2events.TheQPOfrequencybehaviorissimilarinmorphologytoGroup1,andwhilestronglycorrelatedtothepowerlawfeatures,thereisnoaccompanyingcorrelationtoblackbodyfeatures. 133

PAGE 134

Thetriggerspikeofthe-classlightcurvesfor1997August(-1through-5)and1997September(-6through-9)dataatone-secondtimeresolution.ThesolidlinesareadoubleGaussianpluspolynomialttothedatapoints.Iclassifythe-1through-5eventsasGroup1.Thesehaveastrong,narrow,symmetricspikeandtheunderlyinguxhasapositiveslope.The-6through-9dataIclassifyasGroup2events.Thesehaveweaker,wider,asymmetricspikesandtheunderlyingslopeisatornegative.The-8and-9lightcurveshavemoredisturbedspikemorphologiesonthistimescale. 134

PAGE 135

Note.|CorrelationofspectralfeaturestotheQPOfrequency.Theexistenceofacorrelationisbelievableforvaluesofjrj>0.40andhighlysignicantforvaluesofjrj>0.70.Inmostcases,thecorrelationtothepowerlawuxisstrongest.Theabbreviationsareasfollows.TF:totalux;BBN:blackbodynormalization;BBF:blackbodyux;BBT:blackbodytemperature;PLN:powerlawnormalization;PLF:powerlawux;PLI:powerlawindex.TheGroupsaredenedinthetext. 135

PAGE 136

Note.|PartialcorrelationofspectralfeaturestotheQPOfrequency.Therstcolumn(PLF|BBF)isthepartialoftheQPOfrequencyandpowerlawuxwiththeeectofblackbodyuxremoved.Notethatthecorrelationsarestronginmostofthecases.Thesecondcolumn(PLF|BBT)relatestheQPOfrequencytopowerlawux,removingblackbodytemperature.Thesecorrelationsareweaker,butsignicantinmostGroup1cases.Thethirdcolumn(BBF|PLF)showstheQPOfrequencycorrelationtoblackbodyux,removingpowerlawux.Mostarebelievable,suggestingacomplexinterplaybetweenthepowerlawandblackbodyfeatures.Thefourthcolumn(BBT|PLF)showstheQPOfrequencycorrelationtoblackbodytemperature,removingpowerlawux.Inthiscase,nearlyallcorrelationsdropbelowsignicanceshowingthatpowerlawuxmaytraceQPObehaviorbetterthanblackbodytemperature. 136

PAGE 137

MinFreqTimeatMinIDGroup(Hz)(s)FractionatMin Note.|ApproximatelengthoftimeQPOstaysnearlowestfrequency.ThethirdcolumnshowstheminimumQPOfrequency.ThefourthcolumnshowsthelengthoftimetheQPOstayswithin0.5Hzoftheminimumfrequency.ThenumberinparenthesesisthetimeduringwhichtheQPOfrequencyisbelow2.5Hzifitisnotequaltothelistedtime.Thefthcolumnshowstheapproximatefractionofthediplengthspentattheminimumfrequency.Infourcases,thefrequencydoesnotdropbelow2.5Hz.Ingeneral,Group2eventsspendlesstimeattheminimumfrequencythanGroup1events. 137

PAGE 138

Note.|Gaussiantparameterscalculatedforthe\triggerspike"in-classlightcurves.Thecolumnlabelsareasfollows:A=thenormalizedamplitude;A=thewidth(standarddeviation)ofthegaussian;C1=theslopeofthelightcurvebackground;fint=theintegratedcountrate.Theirquantitiesaregroupedbyacombinationofspikestrength,spikewidth,andthesignoftheunderlyingslope. 138

PAGE 139

NarayanandYi 1995 ,seealsoChapter1).Inthedisk-coronamodel,acooldiskextendstotheinnermoststableorbitaroundtheblackhole,whilehardX-raysareproducedinamagneticallyheatedcoronaabovethedisk.InFigure 8-1 ,Ishowacartoonmodelofhowjetejectionmightoccurinthedisk-coronamodel.Intherstphase(left),theaccretiondiskextendstotheinnermoststableradius(afewSchwarzschildradii).Thediskisvisiblethroughatransparentcorona.Inthesecondphase(middle),someperturbationinthesystemcausespre-jetmaterialtoaccumulateandthecoronatobecomeopaque.ThisiswhentheQPObecomesapparentintheX-raylightcurveandevolves.Becausepartofthediskisobscured,thesystemwillappearspectrallyhard.Inthenalphase(right),theinnerdiskisevacuatedviajetejection.Oncetheejectionhasoccurred,theinnerdiskrells,returningthesystemtophase1.Theevacuationofthediskmeansthatwhateverprocessthatwascausingthe 139

PAGE 140

Varniereetal. 2002 ).ThemodeltiestheQPOfrequencytoanAEIfrequency( Taggeretal. 2004 )andmakesspecicpredictionsaboutX-rayspectralbehaviorandjetejectionstrength.Theresultspresentedhereareapreliminary,butpromising,reportofanongoingcollaborativework. 6-1 .TheX-raycoverageofthe-classeventsdoes 140

PAGE 141

6-6 ,theQPOisclearlypresentinthe-classabove4Hz.IusethisapparentcriteriontoselectthecoherentQPOfrequencyanddistinguishitfromlow-frequencynoiseinsubsequentanalyses.InFigures 8-2 through 8-6 ,IshowscatterplotsoftheQPOfrequencyversustotalux,power-lawux,blackbodyux,blackbodytemperature,andcolorradius.Insteadofexaminingtheeventsindividually,Icomparethe,,andclassesasawhole,focusingondetectionsabove4Hz.ImmediatelyapparentinFigure 8-2 isthatthe-classobservationsoccuratahighertotaluxthanthe-and-classobservations,andinFigures 8-3 and 8-4 ,Inotethatthedierenceisapparentinboththepowerlawandblackbodyux.However,thetotaluxobservedinthe-classincreaseswithQPOfrequencyattwicetherateofthatobservedinthe-and-classes.Thistrendismirroredintheblackbodyux(Fig. 8-4 ),butnotinthepowerlawux,wheretheQPOfrequency-powerlawuxtrendinthe-classrunsapproximatelyparalleltothatofthe-and-classes(Fig. 8-3 ).InTable 8-1 ,IcalculatetheLinearPearsonCorrelationCoecient(seepreviouschapter)betweentheQPOfrequencyandvariousspectralfeaturesandshowthatinthe-classtheQPOfrequencyisstronglycorrelatedtothepowerlaw,blackbodyandtotaluxes.Whiletheblackbodytemperatureisrelativelyhigherinthe-class,itfollowsatrendwithQPOfrequencythatrunsparalleltotherelationshipshownintheandclasses(Figure 8-5 ).Theblackbodytemperatureisalsowell-correlatedtotheQPOfrequency.Finally,the-classobservationsshowalowerblackbodynormalizationoverall,andareconsequentlyassociatedwithaloweraverageradii(Fig 8-6 ).Theradius(alsocalledthecolorradius)iscalculatedfromtheblackbodynormalization,NBB,thedistance,andtheinclination.From Fenderetal. ( 1999 ),recallthatthedistancetoGRS1915+105isd=11:20:8kpc.Theinclinationofthesystemis66o2o.Thecolorradius,Rcolis 141

PAGE 142

TitarchukandFiorito 2004 )oraKeplerianfrequency(e.g., Merlonietal. 2000 ).TheKeplerianfrequencyKdescribesthemotionofaparticleinanaccretiondiskduetogravity.TheKeplerianfrequencyassociatedwiththeinner-mostaccretionradiusis K=p Greineretal. 2001 )isK=2000Hz,whereastheobservedQPOfrequencyis2-10Hz.ThustheQPOfrequencyisoftenmonotonicallyscaledinordertoreconcileittotheKeplerianfrequency.Inthatway,theQPOfrequencycanbedirectlyrelatedtoKepleriandynamicalmotion,whichiswellunderstood.AsmentionedinChapter7andagainintheprevioussection,thecolorradiusassociatedwiththe-,-,and-classesisverypoorlycorrelatedtotheQPOfrequency,ascomparedtootherspectralfeatures.Thus,eithertheQPOfrequencydoesnottracetheKeplerianfrequencyorthecolorradiusoutputfromtheXSpecmodeldoesnottracetheinnerdiskradius.IrefertothiscontradictionastheRadiusDilemma. 142

PAGE 143

Martetal. 2000 ).MyXSpectsuseblackbodynormalizationtoestimatearadiusandndthatvaluesrangingfrom1080km,whichisofteninsidetheSchwarzschildradius.Thisisaknownproblemthatseemstobeoftenignoredincurrentliterature. Merlonietal. ( 2000 )performedareliabilitystudyofusingXSpecinanattempttoresolvethisdilemma.TheycreatedaseriesofmodelX-rayspectra,addednoiseandaccountedfordetectionquirks,andfoundthattheXSpecradiustssystematicallyunderestimatedtheradiusvaluesforastandardShakura-Sunyaevdisk.Formorecomplexsystemsinwhichtheaccretionrateortheenergydissipatedinthecoronaareallowedtovary,theinferredradiusfromthemodelappearedtochange,evenwhentheactualradiuswasxedattheinnermoststableorbit.Intheirstudy, Merlonietal. ( 2000 )foundthatthereisnosinglecorrectionfactorthatcanbeappliedtocorrectthemeasuredradiustothetrueradius,especiallyinavolatilesystem.Despitetheknowncaveats,thebeliefpersiststhattheQPOfrequency,QPO,andthecolorradius,Rcol,canbeassociatedwiththeKeplerianfrequency.Thisisinpartbecauseinasubsetofobservations,theQPOfrequencyandthecolorradiusshowanapparenttrendofQPO/R3=2col.Ofthe17light-curvesIexamineforthiswork,onlythreeshowthisapparenttrend.InTable 8-2 ,Iestimatethepowerlawvalue,n,ofaQPO/Rncolt.Itonlydatatakenwithintherst200sofeachlightcurve'sharddip,astheblackbodyuxisstillrelativelystronghereandthetraceoftheblackbodynormalizationisprobablymorereliable.Inexamining9-classlightcurves,3-classlight-curves,and5-classlightcurves,Indthattheinstanceofa\Keplerian"relationshipbetweentheQPOfrequencyandcolorradiusisinthe-class.Eventhen,thefrequency-radiusrelationshipisinconsistent.Boththe-and-classshowrelativelyatrelationships.In 143

PAGE 144

8-7 ,IplottherelationshipofthecolorradiustotheQPOfrequencywhenconsideringthefulllengthofthedip.Theradiusrelationshipissingularlyinvertedinthecaseofthe-class. TaggerandPellat ( 1999 )inordertoexplainthepatternoflow-frequencyQPOsinmicroquasars.TheAEIpredictsaclearobservationalsignature{theturnoveroftheQPOfrequency-innerradiusrelationship( Varniereetal. 2002 ; Taggeretal. 2004 ).InFigure 8-8 ,IshowthepredictionsoftheAEImodelagainsttheobservationaldatapresentedabove.Athighercolorradii(Rcol>40km),themodelpredictsaKeplerianrelationshipQPO/R3=2col.Atlowerradii(Rcol<40km),themodeldeviatesfromaKeplerianrelationship.ThevariationoftheQPOfrequencywiththediskinnerradiusisthusoneofthefewpotentialobservationaltestsofthistheory. Varniereetal. ( 2002 )usethispredictedturnovertoexplaindiscrepanciesinthebehavioroftheQPOinthemicroquasarsGROJ1655-40andGRS1915+105.Thus,whiletheinversionobservedbetweenthe-and-classesintheQPOfrequency-radiusrelationshipnegatesanymonotonicsolutionforreconcilingtheKeplerianfrequencytotheQPOfrequency,itisnotanissuefortheAEImodelbecausethereexistsaregimeintheAEImodelinwhichtheQPOfrequency-radiusrelationexhibitsaninversion( Varniereetal. 2002 ).IntheAEImodel,aMagneto-RotationalInstability(MRI)causesdiskturbulencethatsuppressestheQPOduringthesoftstate.AssoonastheAEIappears,thediskcoolsdownandthepowerdensityspectrumchanges.TheappearanceoftheQPOcausesthetransitionintothelow-hardstatebystoppingtheheatingoftheaccretiondiskandfunnelingenergyintothecorona( Taggeretal. 2004 ).Usingasinglespiralarmmodel,thesourceoftheAEIisassociatedoutwardtoaco-rotationradiuswhichis(insomemagneticregimes)associatedwiththeKeplerianFrequencyandhencetheKeplerianradius.Inthelow-hardstate,thediskaccumulatesa 144

PAGE 145

8-8 .ThemodelassumesthatQPOfrequencyandtheradiusaredirectlyconnected,which,accordingtoearlyliterature,istrue( Munoetal. 1999 ).Inthepreviouschapter,Ishowthattheblackbodynormalization,andhencethecolorradius,iscorrelatedtotheQPOfrequency,withcorrelationcoecientsrangingfromR=0:360:66inthe-class(seeTable 7-1 ).RecallfromChapter7thatjRj=0:40:6arebelievable,butnotethatIgenerallydisregardedthesecorrelationvalueswhenmorestatisticallysignicantcorrelationswerepresent.Thecorrelationsto-classdatashowaninverted,butstillcorrelatedrelationshipR=0:510:65(Table 8-1 ).Notethesearelinearcorrelationcoecients,andthuswouldnaturallybelessstatisticallysignicantforanon-linearcorrelation.Itwouldnotbeunreasonabletoassumethatthemagneticeldconditionsaresucientlydierentbetweenthe-,-,and-classessoastopusheachgroupontodierentregimesofthiscurve.Infact,thediscoveryofthisinversionintheQPOfrequency-innerradiustrendistherstrecordofaninversionwithinasinglesourceandlendscredencetotheAEImodel.TheAEImodeliscurrentlyoneoftheonlyQPOfrequencymodelsthataccountsforthisheretoforeunexplainedreversalinthe 145

PAGE 146

8-7 ,Ishowthatthe-classfallswithinthe\Keplerian"regime.AssumingthattheAEImodelcanaccuratelypredictthecolorradiusbasedontheQPOfrequency,Iperformthefollowingtesttoseeiftheblackbodynormalization(andrelatedcolorradius)determinedbyttingtheRXTEdatawithXSpecmodelscanbeimproved.First,usinga-classlightcurve,IndtheQPOfrequencyasafunctionoftime.Byusingthe-1dataset,inwhichtheQPOfrequencyandcolorradiusareshowntobenearlyKeplerian,IestablishascalingfactorsuchthatlogQPO 146

PAGE 147

8-9 .ThistestsuggeststhattheXSpecblackbodynormalizationNBB;XcannotbeforciblyconformedtothepredictionsoftheKeplerianregime,NBB;K,inthe-class.ThismaysuggestthattheinstantaneouscolorradiusofaspectrumisnotassociatedwiththeQPOfrequencyinaKeplerianmanner.Itismorelikelythatthesystempriortoajetejectioneventisdissipatingenergyintothecoronaand/ortheaccretionrateischanging,andthustheradiustsareunreliable( Merlonietal. 2000 ).Still,theAEImodelisanimprovementovermodelsthatpredictsimpleKeplerianscalingintheQPOfrequency-radiusrelationship.Despiteinherentproblemsinradiusestimation,itisinteresting(andpromising)thatthe-classdata,whichhavelowerradiithanthe-classdata,fallwithinthesame2-10HzfrequencyrangeandarehenceconsistentwiththeturnoverpredictedintheAEImodel.Itgiveshopethatwhiletheinstantaneousestimatesofthecolorradiusmaybeo,thetrendpredictedbytheXSpecmodeltsisstatisticallyaccurateandusefulfortestingmodels. Merlonietal. ( 2000 )havepreviouslystudiedthisissueand 147

PAGE 148

Merlonietal. 2000 ).ThesecaveatsmakeitdiculttosaycondentlyatthisjuncturethattheQPOfrequencyisnotassociatedtotheKeplerianfrequency.Toendonapositivenote,however,ifoneassumesthatthetrendIobserveintheQPOfrequency-radiusrelationofGRS1915+105canbebelieved,thenthissetofobservationscapturesaveryimportantobservationalpredictionoftheAEImodelandbehaviorallylinksGRS1915+105tothemicroquasarGRO1655-40,takinganexcitingsteptowardvalidatingthismodel.Icanfurtherrenethisresultbygathering 148

PAGE 149

149

PAGE 150

Acartoonmodelofjetejectioninthedisk-coronascenario.1.Afullaccretiondiskextendstotheeventhorizonandisvisiblethroughatransparentcorona.2.Theinnerdiskbeginstoevacuateandthecoronabecomesopaque.3.Theevacuateddiskcomponentisfunneledintothejet.Afterthejetejection,thediskllsinagain. 150

PAGE 151

QPOfrequencyvs.totaluxfor-class(diamonds),-class(crosses),and-class(triangles)above4Hz.Theuxismeasuredinerg/s;thefrequencyismeasuredinHz.The-classobservationshaveahighertotalux,andthetrendistwiceassteepasthatobservedinthe-and-classes. 151

PAGE 152

QPOfrequencyvs.poweruxfor-class(diamonds),-class(crosses),and-class(triangles)above4Hz.Theuxismeasuredinerg/s;thefrequencyismeasuredinHz.The-classobservationshaveahigheroverallux,andthetrendisparalleltothatobservedinthe-and-classes. 152

PAGE 153

QPOfrequencyvs.totaluxfor-class(diamonds),-class(crosses),and-class(triangles)above4Hz.Theuxismeasuredinerg/s;thefrequencyismeasuredinHz.The-classobservationshaveahigherblackbodyux,andthetrend,likethatofthetotaluxinFig. 8-2 ,istwiceassteepasthatobservedinthe-and-classes. 153

PAGE 154

QPOfrequencyvs.blackbodytemperaturefor-class(diamonds),-class(crosses),and-class(triangles)above4Hz.ThetemperatureismeasuredinkeV;thefrequencyismeasuredinHz.The-classobservationshaveahigheroveralltemperature,andthetrendisparalleltothatobservedinthe-and-classes. 154

PAGE 155

QPOfrequencyvs.innerdiskradiusfor-class(diamonds),-class(crosses),and-class(triangles)above4Hz.Theradiusismeasuredinkilometers;thefrequencyismeasuredinHz.The-classobservationshavealoweroverallradius.Noobvioustrendisapparentineitherclass. 155

PAGE 156

Thecolorradiusvs.QPOfrequency.Thebarsonthesepointsrepresentthedispersionoffrequenciesobservedineachradiusbin.Althoughtheaveragepointsforthe-classsuggestatrendreminiscentofaKeplerianfrequency-radiusrelation,thereishighdispersionatthelowerradii.The-classshowsnoapparenttrendbetweencolorradiusandQPOfrequency.The-classshowsaninvertedrelationship.Thesepointsrepresentaveragesamong17dierentharddipevents. Table8-1. Note.|CorrelationofspectralfeaturestotheQPOfrequencyfor-classobservations.Theexistenceofacorrelationisbelievableforvaluesofjrj>0.40andhighlysignicantforvaluesofjrj>0.70.TheabbreviationsareasinTable 7-1 156

PAGE 157

Radius(inkm)vs.QPOfrequency(inHz)onalog-logscale.ThetheoreticalpredictionsoftheAEImodelareshownaslargebluecircleswhiletheobservationaldataisshownassmallsquares.The-classdataisshowninred,andoccupiestheKeplerianregimeofthemodel.The-classdataisshowninpinkandoccursneartheturnover.The-classdataisshowninblack,wherethetrendisinverted.ThispreliminarycomparisonofthedataandmodelwascontributedbycollaboratorPeggyVarniere. 157

PAGE 158

Radiustsforthe-2event.TheopensquaresaretheoriginalradiustswithXSpec11.3(2<2,radiusisafreeparameter).TheblacklinesaretheoreticalestimatesbasedontheMagneticFloodModelandallowingfora0.6Hzfrequencyerror.ThelledsquaresaretherevisedXSpectsusingtherestrictedtheoreticalradius.Inordertogetareasonablenumberofts,Iallowed2<8.Evenwiththerelaxedrestrictionon2,only54%ofthespectracanbet(comparedtomyoriginalcriteria,whicht90%ofthespectra). 158

PAGE 159

SetIDPowerLawIndex Note.|ObservationIDsofthe17epochs,employingthenamingschemefromTable 6-1 .IttheQPO-Frequencyandthecolorradiuswithapowerlaw.InaKepleriancase,thePowerLawIndexwouldbe-1.5.Thesetsonlyconsidertherst200secondsofthedipwherethecorrelationoftheQPOfrequencyandblackbodyfeaturesaremorecertain(seeChapter7).Novalueisgivenfor-16becausethestartofthedipisnotvisible.Notehowtheandclassesappearrelativelyathere. 159

PAGE 160

Eikenberryetal. 2005 ).TheunprecedentedspectroscopiccapabilitiesandthespecictargetingofX-raysourcecounterpartswillyieldasignicantincreaseintheIRspectroscopicdatabase 160

PAGE 161

161

PAGE 162

162

PAGE 163

163

PAGE 164

164

PAGE 165

Arnaud,K.A.1996,inAstronomicalSocietyofthePacicConferenceSeries,101,AstronomicalDataAnalysisSoftwareandSystemsV,EditedbyG.H.Jacoby&J.Barnes,17 Balbus,S.A.,&Hawley,J.F.1991,ApJ,376,214 Balbus,S.A.,&Hawley,J.F.1998,ReviewsofModernPhysics,70,1 Bandyopadhyay,R.,Shahbaz,T.,Charles,P.A.,vanKerkwijk,M.H.,&Naylor,T.1997,MNRAS,285,718 Bandyopadhyay,R.M.2008,privatecommunication Bandyopadhyay,R.M.,Miller-Jones,J.C.A.,Blundell,K.M.,Bauer,F.E.,Podsiadlowski,P.,Gosling,A.J.,Wang,Q.D.,Pfahl,E.,&Rappaport,S.2005,MNRAS,364,1195 Beckmann,V.,Kennea,J.A.,Markwardt,C.,Paizis,A.,Soldi,S.,Rodriguez,J.,Barthelmy,S.D.,Burrows,D.N.,Chester,M.,Gehrels,N.,Mowlavi,N.,&Nousek,J.2005,ApJ,631,506 Belloni,T.,Klein-Wolt,M.,Mendez,M.,vanderKlis,M.,&vanParadijs,J.2000,A&A,355,271 Blandford,R.D.,&Begelman,M.C.1999,MNRAS,303,L1 Boirin,L.,Parmar,A.N.,Oosterbroek,T.,Lumb,D.,Orlandini,M.,&Schartel,N.2002,A&A,394,205 Cardelli,J.A.,Clayton,G.C.,&Mathis,J.S.1989,ApJ,345,245 Cassinelli,J.P.,Waldron,W.L.,Sanders,W.T.,Harnden,F.R.,Jr.,Rosner,R.,&Vaiana,G.S.1981,ApJ,250,677 Castro-Tirado,A.J.,Brandt,S.,&Lund,N.1992,IAUCirc.,5590,2 Choi,C.S.,Dotani,T.,Day,C.S.R.,&Nagase,F.1996,ApJ,471,447 Clark,J.S.,Charles,P.A.,Clarkson,W.I.,&Coe,M.J.2003,A&A,400,655 Clark,J.S.,Miroshnichenko,A.S.,Larionov,V.M.,Lyuty,V.M.,Hynes,R.I.,Pooley,G.G.,Coe,M.J.,McCollough,M.,Dieters,S.,Emov,Y.S.,Fabregat,J.,Goranskii,V.P.,Haswell,C.A.,Metlova,N.V.,Robinson,E.L.,Roche,P.,Shenavrin,V.I.,&Welsh,W.F.2000,A&A,356,50 Clark,L.L.,&Dolan,J.F.1999,A&A,350,1085 Coe,M.J.,Buckley,D.A.H.,Fabregat,J.,Steele,L.A.,Still,M.D.,&Torrejon,J.M.1997,A&AS,126,237 165

PAGE 166

Corcoran,M.F.,Hamaguchi,K.,Pollock,A.M.T.,Pittard,J.M.,Stevens,I.R.,Henley,D.B.,Moat,A.F.J.,&Marchenko,S.2006,inIAUSymposium,230,PopulationsofHighEnergySourcesinGalaxies,EditedbyE.J.A.Meurs&G.Fabbiano,35 Cox,A.N.2000,Allen'sAstrophysicalQuantities(NewYork:AIPPress) Crowther,P.A.2007,ARA&A,45,177 Cushing,M.C.,Vacca,W.D.,&Rayner,J.T.2004,PASP,116,362 Cutri,R.M.,Skrutskie,M.F.,vanDyk,S.,Beichman,C.A.,Carpenter,J.M.,Chester,T.,Cambresy,L.,Evans,T.,Fowler,J.,Gizis,J.,Howard,E.,Huchra,J.,Jarrett,T.,Kopan,E.L.,Kirkpatrick,J.D.,Light,R.M.,Marsh,K.A.,McCallon,H.,Schneider,S.,Stiening,R.,Sykes,M.,Weinberg,M.,Wheaton,W.A.,Wheelock,S.,&Zacarias,N.2003,2MASSAllSkyCatalogofpointsources.(TheIRSA2MASSAll-SkyPointSourceCatalog,NASA/IPACInfraredScienceArchive.http://irsa.ipac.caltech.edu/applications/Gator/) Davidson,K.,Ishibashi,K.,&Corcoran,M.F.1998,NewAstronomy,3,241 DeBecker,M.,Rauw,G.,Sana,H.,Pollock,A.M.T.,Pittard,J.M.,Blomme,R.,Stevens,I.R.,&vanLoo,S.2006,MNRAS,371,1280 Dupree,A.K.,Gursky,H.,Black,J.H.,Davis,R.J.,Hartmann,L.,Matilsky,T.,Raymond,J.C.,Hammerschlag-Hensberge,G.,vandenHeuvel,E.P.J.,Burger,M.,Lamers,H.J.G.L.M.,VandenBout,P.A.,Morton,D.C.,deLoore,C.,vanDessel,E.L.,Menzies,J.W.,Whitelock,P.A.,Watson,M.,Sanford,P.W.,&Pollard,G.S.G.1980,ApJ,238,969 Dutra,C.M.,Santiago,B.X.,Bica,E.L.D.,&Barbuy,B.2003,MNRAS,338,253 Eikenberry,S.,Blum,R.,Bandyopadhyay,R.,Muno,M.,Sellgren,K.,Olsen,K.,&F2GCSTeam.2005,inBulletinoftheAmericanAstronomicalSociety,37,1462 Eikenberry,S.S.,Matthews,K.,Morgan,E.H.,Remillard,R.A.,&Nelson,R.W.1998,ApJ,494,L61 Eikenberry,S.S.,Matthews,K.,Muno,M.,Blanco,P.R.,Morgan,E.H.,&Remillard,R.A.2000,ApJ,532,L33 Evans,N.R.,Seward,F.D.,Krauss,M.I.,Isobe,T.,Nichols,J.,Schlegel,E.M.,&Wolk,S.J.2003,ApJ,589,509 Fender,R.,Belloni,T.,&Gallo,E.2005,Ap&SS,300,1 166

PAGE 167

Fender,R.P.,&Pooley,G.G.1998,MNRAS,300,573 Feroci,M.,Matt,G.,Pooley,G.,Costa,E.,Tavani,M.,&Belloni,T.1999,A&A,351,985 Figer,D.F.,McLean,I.S.,&Najarro,F.1997,ApJ,486,420 Figer,D.F.,Najarro,F.,Morris,M.,McLean,I.S.,Geballe,T.R.,Ghez,A.M.,&Langer,N.1998,ApJ,506,384 Filliatre,P.,&Chaty,S.2004,ApJ,616,469 Forman,W.,Jones,C.,Cominsky,L.,Julien,P.,Murray,S.,Peters,G.,Tananbaum,H.,&Giacconi,R.1978,ApJS,38,357 Frank,J.,King,A.,&Raine,D.J.2002,AccretionPowerinAstrophysics:ThirdEdition(Cambridge,UK:CambridgeUniversityPress,February2002.) Giacconi,R.,Bechtold,J.,Branduardi,G.,Forman,W.,Henry,J.P.,Jones,C.,Kellogg,E.,vanderLaan,H.,Liller,W.,Marshall,H.,Murray,S.S.,Pye,J.,Schreier,E.,Sargent,W.L.W.,Seward,F.,&Tananbaum,H.1979,ApJ,234,L1 Giacconi,R.,Gursky,H.,Paolini,F.R.,&Rossi,B.B.1962,PhysicalReviewLetters,9,439 Giacconi,R.,Murray,S.,Gursky,H.,Kellogg,E.,Schreier,E.,&Tananbaum,H.1972,ApJ,178,281 Gies,D.R.,&Bolton,C.T.1986,ApJ,304,389 Girardi,L.,Bertelli,G.,Bressan,A.,Chiosi,C.,Groenewegen,M.A.T.,Marigo,P.,Salasnich,B.,&Weiss,A.2002,A&A,391,195 Gosling,A.J.,Bandyopadhyay,R.M.,&Blundell,K.M.2007,inAmericanInstituteofPhysicsConferenceSeries,924,893 Gogus,E.,Patel,S.K.,Wilson,C.A.,Woods,P.M.,Finger,M.H.,&Kouveliotou,C.2005,ApJ,632,1069 Greiner,J.,Cuby,J.G.,McCaughrean,M.J.,Castro-Tirado,A.J.,&Mennickent,R.E.2001,A&A,373,L37 Groote,D.,&Hunger,K.1982,A&A,116,64 Groote,D.,&Schmitt,J.H.M.M.2004,A&A,418,235 Gursky,H.1972,ApJ,175,L141 167

PAGE 168

Heras,A.M.,Shipman,R.F.,Price,S.D.,deGraauw,T.,Walker,H.J.,JourdaindeMuizon,M.,Kessler,M.F.,Prusti,T.,Decin,L.,Vandenbussche,B.,&Waters,L.B.F.M.2002,A&A,394,539 Hillier,D.J.,Davidson,K.,Ishibashi,K.,&Gull,T.2001,ApJ,553,837 Horne,J.H.,&Baliunas,S.L.1986,ApJ,302,757 Hyland,A.R.,&Mould,J.R.1973,ApJ,186,993 Iaria,R.,DiSalvo,T.,Robba,N.R.,Burderi,L.,Lavagetto,G.,&Riggio,A.2005,ApJ,634,L161 Klein-Wolt,M.,Fender,R.P.,Pooley,G.G.,Belloni,T.,Migliari,S.,Morgan,E.H.,&vanderKlis,M.2002,MNRAS,331,745 Kroupa,P.2001,MNRAS,322,231 Lada,C.J.2006,ApJ,640,L63 Laycock,S.,Grindlay,J.,vandenBerg,M.,Zhao,P.,Hong,J.,Koenig,X.,Schlegel,E.M.,&Persson,S.E.2005,ApJ,634,L53 Leggett,S.K.,Allard,F.,Berriman,G.,Dahn,C.C.,&Hauschildt,P.H.1996,ApJS,104,117 Leitherer,C.,&Wolf,B.1984,A&A,132,151 Lepine,S.,Wallace,D.,Shara,M.M.,Moat,A.F.J.,&Niemela,V.S.2001,AJ,122,3407 Lewin,W.H.G.,&vanderKlis,M.2006,CompactStellarX-raySources(Cambridge,UK:CambridgeUniversityPress) Lewin,W.H.G.,vanParadijs,J.,&vandenHeuvel,E.P.J.1997,X-rayBinaries(Cambridge,UK:CambridgeUniversityPress) Liu,Q.Z.,vanParadijs,J.,&vandenHeuvel,E.P.J.2006,A&A,455,1165 Luo,D.,McCray,R.,&MacLow,M.-M.1990,ApJ,362,267 Maz-Apellaniz,J.,Walborn,N.R.,Galue,H.A.,&Wei,L.H.2004,ApJS,151,103 Margon,B.,Anderson,S.F.,Aller,L.H.,Downes,R.A.,&Keyes,C.D.1984,ApJ,281,313 Markwardt,C.B.,Swank,J.H.,&Taam,R.E.1999,ApJ,513,L37 Mart,J.,Mirabel,I.F.,Chaty,S.,&Rodrguez,L.F.2000,A&A,356,943 168

PAGE 169

McNamara,D.H.,Madsen,J.B.,Barnes,J.,&Ericksen,B.F.2000,PASP,112,202 Merloni,A.,Fabian,A.C.,&Ross,R.R.2000,MNRAS,313,193 Mikles,V.J.,Eikenberry,S.S.,Muno,M.P.,Bandyopadhyay,R.M.,&Patel,S.2006,ApJ,651,408 Mineshige,S.,Ouchi,N.B.,&Nishimori,H.1994,PASJ,46,97 Mirabel,I.F.,Dhawan,V.,Chaty,S.,Rodriguez,L.F.,Marti,J.,Robinson,C.R.,Swank,J.,&Geballe,T.1998,A&A,330,L9 Mirabel,I.F.,&Rodriguez,L.F.1994,Nature,371,46 Moat,A.F.J.1969,A&A,3,455 Mokiem,M.R.,deKoter,A.,Evans,C.J.,Puls,J.,Smartt,S.J.,Crowther,P.A.,Herrero,A.,Langer,N.,Lennon,D.J.,Najarro,F.,Villamariz,M.R.,&Vink,J.S.2007,A&A,465,1003 Morel,T.,&Grosdidier,Y.2005,MNRAS,356,665 Morris,P.W.,Eenens,P.R.J.,Hanson,M.M.,Conti,P.S.,&Blum,R.D.1996,ApJ,470,597 Mouchet,M.,Ilovaisky,S.A.,&Chevalier,C.1980,A&A,90,113 Muno,M.P.,Arabadjis,J.S.,Bagano,F.K.,Bautz,M.W.,Brandt,W.N.,Broos,P.S.,Feigelson,E.D.,Garmire,G.P.,Morris,M.R.,&Ricker,G.R.2004a,ApJ,613,1179 Muno,M.P.,Bagano,F.K.,Bautz,M.W.,Brandt,W.N.,Broos,P.S.,Feigelson,E.D.,Garmire,G.P.,Morris,M.R.,Ricker,G.R.,&Townsley,L.K.2003,ApJ,589,225 Muno,M.P.,Bagano,F.K.,Bautz,M.W.,Feigelson,E.D.,Garmire,G.P.,Morris,M.R.,Park,S.,Ricker,G.R.,&Townsley,L.K.2004b,ApJ,613,326 Muno,M.P.,Bauer,F.E.,Bandyopadhyay,R.M.,&Wang,Q.D.2006a,ApJS,165,173 Muno,M.P.,Bower,G.C.,Burgasser,A.J.,Bagano,F.K.,Morris,M.R.,&Brandt,W.N.2006b,ApJ,638,183 Muno,M.P.,Morgan,E.H.,&Remillard,R.A.1999,ApJ,527,321 Nagase,F.,Corbet,R.H.D.,Day,C.S.R.,Inoue,H.,Takeshima,T.,Yoshida,K.,&Mihara,T.1992,ApJ,396,147 169

PAGE 170

Narayan,R.,Yi,I.,&Mahadevan,R.1995,Nature,374,623 Naze,Y.,Rauw,G.,Vreux,J.-M.,&DeBecker,M.2004,A&A,417,667 Negueruela,I.,&Schurch,M.P.E.2007,A&A,461,631 Ogilvie,G.I.,&Dubus,G.2001,MNRAS,320,485 Predehl,P.,&Schmitt,J.H.M.M.1995,A&A,293,889 Ransom,S.M.,Eikenberry,S.S.,&Middleditch,J.2002,AJ,124,1788 Rayner,J.T.,Toomey,D.W.,Onaka,P.M.,Denault,A.J.,Stahlberger,W.E.,Vacca,W.D.,Cushing,M.C.,&Wang,S.2003,PASP,115,362 Reiners,A.,Stahl,O.,Wolf,B.,Kaufer,A.,&Rivinius,T.2000,A&A,363,585 Rieke,G.H.,&Lebofsky,M.J.1985,ApJ,288,618 Rothstein,D.M.,Eikenberry,S.S.,&Matthews,K.2005,ApJ,626,991 Sana,H.,Stevens,I.R.,Gosset,E.,Rauw,G.,&Vreux,J.-M.2004,MNRAS,350,809 Sanz-Forcada,J.,Franciosini,E.,&Pallavicini,R.2004,A&A,421,715 Schild,H.,Gudel,M.,Mewe,R.,Schmutz,W.,Raassen,A.J.J.,Audard,M.,Dumm,T.,vanderHucht,K.A.,Leutenegger,M.A.,&Skinner,S.L.2004,A&A,422,177 Schultz,A.2005,NICMOSInstrumentHandbook,Version8.0(Baltimore:STScI) Schulz,N.S.,Canizares,C.R.,Lee,J.C.,&Sako,M.2002a,ApJ,564,L21 Schulz,N.S.,Cui,W.,Canizares,C.R.,Marshall,H.L.,Lee,J.C.,Miller,J.M.,&Lewin,W.H.G.2002b,ApJ,565,1141 Seward,F.D.,Butt,Y.M.,Karovska,M.,Prestwich,A.,Schlegel,E.M.,&Corcoran,M.2001,ApJ,553,832 Shakura,N.I.,&Sunyaev,R.A.1976,MNRAS,175,613 Sidoli,L.,Mereghetti,S.,Favata,F.,Oosterbroek,T.,&Parmar,A.N.2006,A&A,456,287 Skinner,S.L.,Gudel,M.,Schmutz,W.,&Stevens,I.R.2001,ApJ,558,L113 Skinner,S.L.,Zhekov,S.A.,Palla,F.,&Barbosa,C.L.D.R.2005,MNRAS,361,191 Steiner,J.E.,&Damineli,A.2004,ApJ,612,L133 Tagger,M.,&Pellat,R.1999,A&A,349,1003 170

PAGE 171

Terrell,D.,&Wilson,R.E.2005,Ap&SS,296,221 Timmer,J.,&Koenig,M.1995,A&A,300,707 Titarchuk,L.,&Fiorito,R.2004,ApJ,612,988 Titarchuk,L.,Shaposhnikov,N.,&Areev,V.2007,ApJ,660,556 Tomsick,J.A.,Chaty,S.,Rodriguez,J.,Foschini,L.,Walter,R.,&Kaaret,P.2006,ApJ,647,1309 Tousey,R.,Watanabe,K.,&Purcell,J.D.1951,PhysicalReview,83,792 Townsley,L.K.,Broos,P.S.,Nousek,J.A.,&Garmire,G.P.2002,NuclearInstrumentsandMethodsinPhysicsResearchA,486,751 Turner,D.G.1985,A&A,144,241 Usov,V.V.1992,ApJ,389,635 Vacca,W.D.,Cushing,M.C.,&Rayner,J.T.2003,PASP,115,389 vanderHucht,K.A.2001,NewAstronomyReview,45,135 vanGenderen,A.M.,deGroot,M.J.H.,&The,P.S.1994,A&A,283,89 Vanbeveren,D.,DeLoore,C.,&VanRensbergen,W.1998,A&ARev.,9,63 Varniere,P.,Rodriguez,J.,&Tagger,M.2002,A&A,387,497 Varricatt,W.P.,Williams,P.M.,&Ashok,N.M.2004,MNRAS,351,1307 Viotti,R.F.,Antonelli,L.A.,Rossi,C.,&Rebecchi,S.2004,A&A,420,527 Wallace,L.,&Hinkle,K.1997,ApJS,111,445 Wallace,L.,Meyer,M.R.,Hinkle,K.,&Edwards,S.2000,ApJ,535,325 Walter,R.,ZuritaHeras,J.,Bassani,L.,Bazzano,A.,Bodaghee,A.,Dean,A.,Dubath,P.,Parmar,A.N.,Renaud,M.,&Ubertini,P.2006,A&A,453,133 Westbrook,O.,Evans,N.R.,Wolk,S.J.,Nichols,J.,Kashyap,V.L.,Mendygral,P.J.,&Waldron,W.L.2006,inBulletinoftheAmericanAstronomicalSociety,38,1117 Williams,P.M.,vanderHucht,K.A.,Sandell,G.,&The,P.S.1990,MNRAS,244,101 Wilson,C.A.,Weisskopf,M.C.,Finger,M.H.,Coe,M.J.,Greiner,J.,Reig,P.,&Papamastorakis,G.2005,ApJ,622,1024 Yadav,J.S.2006,ApJ,646,385 171

PAGE 172

172

PAGE 173

IwasborninNorfolk,Virginia,andraisedinBaltimore,Marylandbyamusicianandasoundengineer.TheneartsweremyrsteducationandIwaspronetotryingoutanyinstrumentsittingonthecoeetable.Nevermuchofastargazer,myrstfascinationwithastronomybeganwithan8thgradescienceprojectresearchingstars.Iwasfascinatedbythefactthatstarsnotonlyliveandevolve,buttheydieandwarpspacetime.Still,IneverconsideredastronomyasacareeruntilIreadabookbyKipThornecalled\BlackHolesandTimeWarps."ItwasthenthatIdecidedtopursueadegreeinastrophysicsnotonlybecauseblackholescaptivatedmyimagination,butbecausetheintellectualcommunityportrayedinthebookwassomethingIwantedtobeapartof.Iwasfortunatemyjunioryearofhighschooltohaveaphysicsteacherwhowasalsoanastronomer,soIknewphysicswasvitaltothecareer.IenteredJohnsHopkinsUniversityasaphysicsmajorandsetoutimmediatelytondworkasaresearchassistant,andabrilliantscientistnamedDr.WeiZhenggavemeachance.Iworkedforhimforthreeandahalfyearsandcompletedmyseniorthesisunderhisadvisement.Duringthattime,Iparticipatedintwoothersummerresearchexperiencesandsankrootsintothesciencecommunity.IcametotheUniversityofFloridaforgraduateschoolbecauseofalltheplacesIvisitedandapplied,Ilovedthecommunityherethemost.Thegraduatestudentsimmediatelytookmein,teachingmetheropes,andcompellingmetosucceed.Inmytimehere,IhavesoughttomakeothersfeelaswelcomeasIdidthatrstday.Formymaster'sproject,IworkedwithVickiSarajedini-anementorandanexcellentrolemodelforwomeninscience.Wesearchedforsuper-massiveblackholesinthecoreofdistantgalaxies.Goodtimes.FormyPhDthesis,IworkedforSteveEikenberry.Theblackholesweresmaller,butcloseenoughtoseeandddlewiththephysics. 173