University of Amsterdam
MSc Physics and Astronomy
Track: Astronomy & Astrophysics
Master Thesis
Probing Jet Physics Using
Multi-wavelength Data of Stellar-Mass
Black Holes
by
Zheng Cao
Student ID: 12282294 (UVA)
August 28, 202060 ECTS
September 2019 - August 2020
Supervisors:
Prof. Dr. Sera Markoff Matteo Lucchini (MSc)
Examiners Prof. Dr. Sera Markoff Dr. Phil Uttley
Anton Pannekoek
Institute
Abstract
One of the main question in black hole astrophysics is how the black hole jet is coupled to the accretion flow over time. Observations have shown similarities in the accretion process of both stellar-mass black holes in Black hole Low-Mass X-ray Bina-ries (BH-LMXBs) and their supermassive counterparts in the Active Galactic Nuclei (AGN), indicating a common jet production process. This thesis aims to study the jet changes with respect to different accretion states of the black hole, by conduct-ing a systematic study on multiple BH-LMXBs in different spectral states, which are analogous to different AGN species in a much larger evolution timescale. It is one of the first few attempts to combine the multi-wavelength spectral analysis with the timing power color method in the study of the accretion process, which we have applied to observations both in hard and, unprecedentedly, in intermediate spectral states of BH-LMXBs. We have selected and analyzed three BH-LMXB sources in their epochs with quasi-simultaneous multi-wavelength observations available. Our main result suggests a possible systematic change in the size of the jet base during the state transition, providing the evidence for how the jet changes with respect to different accretion states. From the results, this thesis points out a possible geo-metrical link behind the empirical spectral-timing correlation discovered in previous LMXBs studies, which can help us understand better about the evolution of the ac-cretion process in BH-LMXB outbursts. Our study might also help the study of jets in AGN species from an analogous point of view. In this thesis, we have shown that the combination of multi-wavelength spectral analysis with power color analysis can help isolate the phenomenological changes in BH-LMXB state transitions, and test the main driver behind them.
Popular Summary
One of the most mysterious kinds of celestial bodies is the black hole. The gravity of a black hole is so strong that even light cannot escape from its event horizon. The accretion process refers to the process in which the matter outside falls towards the black hole under such strong gravity. For supermassive black holes in galactic centers, the accretion is usually fueled by interstellar medium, and for stellar-mass black holes in binary systems the accreted matter comes from the companion star.
Sometimes the accretion process will generate outflows, one kind of which is a black hole jet. Jets are collimated, relativistic outflows of plasma and fields launched from the accretion region close to the accreting black hole, extending out to a very large distance. Therefore, by studying the jets we can infer how the black holes impact the space environment far away, and how this impact might change the evolution of our universe. One of the main questions in black hole astrophysics is how the jet changes over time. Previous studies on the binary systems in our Galaxy have found evidence that the black hole jet’s turning on and off is closely related to the physical properties in the accretion region very close to the black hole where it is launched, and changes could happen in timescales of days to weeks. On the other hand, supermassive black holes evolve over timescales of millions of years, so we do not see any jet changes like in those stellar-mass black holes but we only observe two species of accreting supermassive black holes, with or without the jet. Therefore, by studying the jets in stellar-mass black holes we can also probe the jets in their supermassive counterparts as well assuming the latter ones follow the same physics in jets turning on/off but just evolve much more slowly.
This thesis aims at probing the jet physics in the accreting stellar-mass black holes with various methods. We want to search for evidence in data from radio to X-rays that could tell us how the jets change over time in multiple stellar-mass black hole sources in binary systems. It could help us know more about how the jet is controlled by the physical properties in the region close to the black hole, and eventually how this leads to the jets turning on/off. We are one of the first few studies that combine multi-wavelength spectral modeling with timing analysis that reveals the short-term variability of the binary sources.
Our results show that the changes in the size of the jet close to the black hole, the jet base, might have the same trend over time among different stellar-mass black hole sources, providing a valuable piece of evidence on how the jet changes in accreting
black holes. Furthermore, it has been frequently observed that in those sources their X-ray short-term variability is well related to their spectral behaviour, but the physics behind this connection is still an open question. Our results in this thesis suggest that sources with the size of the jet base in similar scale with respect to the black hole radius, would have similar X-ray short-term variability, that it is the jet base connecting the spectral behaviour of our sources to their X-ray short-term variability.
Contents
I Research 1
1 Introduction 2
1.1 Black Hole Low-Mass X-ray Binaries . . . 2
1.1.1 Accretion . . . 5
1.1.2 Jets . . . 7
1.1.3 LMXB Outbursts and canonical spectral states . . . 8
1.2 Analysis Techniques . . . 11
1.2.1 Broadband Spectral Analysis . . . 11
1.2.2 X-ray Timing Analysis . . . 12
1.3 Motivation . . . 13 1.4 Outline . . . 14 2 Methods 16 2.1 Jet model . . . 16 2.2 Power colors . . . 21 2.3 Targets Selection . . . 24
3 Observations and Data Reduction 26 3.1 XTE J1752-223 . . . 28
3.2 MAXI J1659-152 . . . 30
3.3 XTE J1650-500 . . . 35
4 Results 40 4.1 Power color diagram . . . 40
4.2 Spectral modelling . . . 43
4.2.1 XTE J1752-223 high/low hard states . . . 49
4.2.2 MAXI J659-152 Epoch A/B/C . . . 50
4.2.3 The B Cluster . . . 53
5 Discussion 57
6 Conclusion 64
List of Figures
1.1 Artist’s impression of an X-ray Binary . . . 3
1.2 Fundamental plane of black hole Lradio∼ LX relation . . . 4
1.3 An example of a LMXB X-ray spectrum . . . 6
1.4 An example of the jet spectrum . . . 8
1.5 Hardness-intensity diagram of GX 339-4 . . . 9
1.6 Representative energy spectra of BH-LMXB during the hard/soft states 10 1.7 Examples of power spectra with different types of QPOs . . . 13
2.1 Schematic of the jet emission model . . . 19
2.2 Power color and hue angle . . . 22
2.3 Power colors for 12 XRB systems . . . 23
3.1 An example of region selection in Swift/UVOT data reduction . . . . 28
3.2 X-ray hardness-intensity diagram for XTE J1752-223 2009 outburst . 29 3.3 X-ray hardness-intensity diagram for MAXI J1659-152 2010 outburst . 31 3.4 X-ray hardness-intensity diagram for XTE J1650-500 2001 outburst . 36 4.1 Power colors of XTE J1752-223, MAXI J1659-152, and XTE J1650-500 data . . . 41
4.2 Hue values in each selected epoch of XTE J1752-223, MAXI J1659-152, and XTE J1650-500 . . . 42
4.3 Change of hue values in MAXI J1659-152 and XTE J1650-500 over time 42 4.4 Residuals of XTE J1752-223 epochs fitted with power law . . . 43
4.5 Residuals of MAXI J1659-152 epochs fitted with power law . . . 44
4.6 Residuals of XTE J1650-500 epochs fitted with power law . . . 45
4.7 Unfolded spectrum and residuals of Epoch A of XTE J1752-223 fitted with Model 2 . . . 46
4.8 Multi-wavelength fitting of XTE J1752-223 high/low hard states with jet model . . . 50
4.9 Multi-wavelength fitting of MAXI J1752 epochs with jet model . . . . 52
4.10 Multi-wavelength fitting of the B cluster with jet model . . . 54
5.1 R0 vs. hue and Nj vs. R0 in the joint-fits . . . 58
5.2 R0 vs. hue relation in XTE J1752-223, MAXI J1659-152, XTE J1650-500, and MAXI J1836-194 . . . 59
List of Figures
5.3 Rin vs. hue in our joint-fits . . . 60
List of Tables
2.1 List of parameters and descriptions in the jet model . . . 20
3.1 List of known parameters for selected sources . . . 27
3.2 List of radio observations of XTE J1752-223 . . . 30
3.3 List of X-ray observations of XTE J1752-223 . . . 30
3.4 List of radio observations of MAXI J1659-152 . . . 32
3.5 List of X-ray observations of MAXI J1659-152 . . . 33
3.6 List of IR/UV observations of MAXI J1659-152 . . . 34
3.7 List of radio observations of XTE J1650-500 . . . 35
3.8 List of X-ray observations of XTE J1650-500 . . . 37
3.9 List of IR observations of XTE J1650-500 . . . 38
4.1 Parameter table of XTE J1752-223 epochs fitted with Model 2 . . . . 47
4.2 Parameter table of MAXI J1659-152 epochs fitted with Model 2 . . . . 47
4.3 Parameter table of XTE J1650-500 epochs fitted with Model 2 . . . . 48
4.4 Parameter table of XTE J1752-223 high/low hard states fitted with jet model . . . 51
4.5 Parameter table of MAXI J1659-152 epochs fitted with jet model . . . 53
List of Abbreviations
ADAF Advection-Dominated Accretion Flow AGN Active Galactic Nuclei
ATCA Australia Telescope Compact Array BH Black Hole
FP Fundamental Plane
GRMHD General Relativistic Megnetohydrodynamic
HEASARC High Energy Astrophysics Science Archive Research Center HHS High Hard State
HID Hardness-Intensity Diagram HIMS Hard-Intermediate State HMXB High-Mass X-ray Binary HS Hard State
ISCO Innermost Stable Circular Orbit LHS Low Hard State
LMXB Low-Mass X-ray Binary
MAXI Monitor of All-sky X-ray Image MCMC Monte-Carlo Markov Chain
NICER Neutron star Interior Composition ExploreR NS Neutron Star
PC1 Power Colour Ratio 1 PC2 Power Colour Ratio 2 QPO Quasi-Periodic Oscillation RXTE Rossi X-ray Timing Explorer SED Spectral Energy Distribution SIMS Soft-Intermediate State SS Soft State
List of Tables VLA Very Large Array
Part I
Chapter 1
Introduction
1.1
Black Hole Low-Mass X-ray Binaries
Most of the stellar systems in our galaxy consist of two or more stars (Eggleton & Tokovinin 2008). An X-ray Binary (XRB) can form when one of the stars in a binary system reaches its end of life and forms a compact object as the primary object of the binary system. This formation of a compact object, either a Black Hole (BH) or a Neutron Star (NS), only happens for high mass stars and/or low metallicity stars. (see Benacquista (2012), for a review). Via multiple channels, the primary object, which is more massive than the secondary star, can accrete matter from its companion during which process a large amount of the gravitational energy of those stellar matters is released, giving rise to kinetic energy and strong X-ray emission, and hence the system is called X-ray binaries.
Based on the mass of the secondary companion star, an XRB can be classified either as a High-Mass X-ray Binary (HMXB) or a Low-Mass X-ray Binary (LMXB). The former refers to the binary system that has a companion star of high mass (≥10M ), while a LMXB has a companion star of low mass (≤1M ). The
interme-diate cases (1M -10M ) are rarely detected due to their inefficiency in all channels
of mass transfer that is responsible for starting the accretion process that generates the X-ray emission (Van Paradijs & Van der Klis 2001). The focus of this thesis lies with the LMXB systems with a stellar-mass black hole as the primary object. In a BH-LMXB system, the black hole accretes matter from its companion star via the Roche-Lobe overflow (see §1.1.1), when the companion star fills its Roche lobe either due to its expansion after the main sequence or due to a short orbital separation of the binary. Similar to their super-massive counterparts which are often found in galactic centers, the stellar-mass black holes in binary systems can also have outflows driven by the mass accretion process. For example, highly collimated streams of plasma or jets can be launched in the vicinity of the black hole (e.g. Fender et al.
Black Hole Low-Mass X-ray Binaries
;
Fig. 1.1 – An artist’s impression of an XRB, showing the matter from a companion star on the
right streaming onto an accretion disk (red and orange on the picture) via Roche-Lobe overflow. Jets, or streams of high energy particles, are launched from the inner part of the accretion region. Image credit: NASA/CXC/M.Weiss.
2004; Falcke & Biermann 1996) and the particles inside the jets can be accelerated to a high speed very close to the speed of light c (Mirabel & Rodriguez 1999; Rosswog & Bruggen 2003), which together with the strong gravitational background posed by the black hole, requires a relativistic treatment. This connection among black holes See Fig. 1.1 for an artistic impression of a LMXB showing the accretion via the Roche-Lobe overflow and the jet.
LMXB/AGN Connections
Active Galactic Nuclei (AGN) are accreting supermassive black holes at the center of galaxies. Some pioneering works on the discovery of AGN could be found in Fath (1909); Seyfert (1943); Schmidt (1963). People discover that some galactic centers are much more luminous than the other, and it is the on-going accretion process of supermassive black hole that gives rise to this extra luminosity, and hence the name AGN (Salpeter 1964; Zel’Dovich 1989). The classification of AGN is based on ob-served signatures. For example, based on the ratio of flux between the radio (5 GHz) to optical (B-band) F5GHz/FB, AGN can be divided into radio-loud (F5GHz/FB≥ 10)
and radio-quiet (F5GHz/FB< 10) classes (Kellermann et al. 1989). More sub-classes
can be divided based on differences of either the luminosity or spectral lines in var-ious energy bands (see Padovani et al. (2017) for a summary). Typical unification models can explain the different AGN species by their differences in orientations (e.g.
Black Hole Low-Mass X-ray Binaries
Antonucci 1993; Urry & Padovani 1995), jet presence (e.g. Padovani 2016), radiative efficiency and host galaxy properties. Shortly after the discovery of an empirical cor-relation between the radio (5 GHz) and X-ray (0.5-10 keV) luminosities of LMXBs during the hard states (Corbel et al. 2000; Gallo et al. 2003), two groups (Merloni et al. 2003; Falcke et al. 2004) independently extended this correlation to a large pop-ulation of AGNs that could explained the jet production across a large range of black hole masses (the Fundamental Plane, Fig. 1.2). Besides the radio/X-ray luminosities, LMXBs and AGNs have many other observational similarities supporting the idea of a common jet production process , summarized in Fender (2016). Therefore, it might be the same reason behind the absence of jets in some spectral states of LMXBs and in some AGN classes. However, the evolution timescale of a supermassive black hole might be millions of years, preventing us to study the long-term variability of individ-ual sources. So, the study of the LMXB outbursts that have a much shorter evolution timescale (weeks or months) might hold the key to the question of why the jets are present in some accretion system but not in the others, indicating the explanation to different AGN species distinguished by whether they have jets or not. In this thesis we want to study this question of jet presence by looking at the physical changes in jets during those LMXB systems transiting between accretion states with/without a compact jet.
30 35 40
log Lradio (erg s−1) 35 40 45 50 55 log Lxray − ξM log MBH (erg s −1 ) GBH (10 MΟ) Sgr A* (106 M Ο) LLAGN (107−8 M Ο) FR I (108−9 M Ο) SDSS HBLs (108−9 M Ο) • • • • •
log LX = (1.45 ± 0.04) log LR - (0.88 ± 0.06) log MBH - 6.07 ± 1.10
;
Fig. 1.2 – Fundamental plane of black hole Lradio vs. LX relation. The X-axis is the radio
luminosity Lradio at 5 GHz and the Y-axis is the X-ray luminosity LX in the 0.5-10 keV band,
shifted by the corresponding black hole mass MBH. Despite the huge difference in mass, black holes
with a jet presence in radio have a similar relation among LX, Lradio, and black hole mass MBH.
This empirical relation indicates a similarity in how the jet(radio & X-ray) and the disk(mainly X-ray) are tightly coupled across the mass and power scales of jet-launching black holes. Adopted from Plotkin et al. 2012.
Black Hole Low-Mass X-ray Binaries
1.1.1 Accretion
In BH-LMXBs, the mass transfer from the companion star to the black hole occurs due to Roche-Lobe overflow. When the companion star fills its Roche lobe, the area in which the stellar matter is gravitationally bound to the star, the matter from the outer envelope of the star starts to fall into the black hole through the first Lagrange point and start spiraling towards the compact object (Frank et al. 2002). Eventually the orbits of falling-in matter circularise due to the energy dissipation and an accretion disk is formed (see Fig.1.1 for the illustration). The high specific angular momentum possessed by the matter falling inward will be transported outward by the viscous torques, to keep the angular momentum conserved in the system as matter keeps falling into the black hole. One of the simplest and most straight-forward descriptions of the radial structure of an accretion disk is a geometrically-thin, optically-thick, and steady disk (Shakura & Sunyaev 1973, 1976).
The scale of the accretion disk for black holes is typically described in the unit of the gravitational radius Rg of the central black hole, which is the characterized
scale of a black hole system:
Rg =
GM
c2 , (1.1)
where G is the gravitational constant and M is the mass of the black hole. In General Relativity, the Innermost Stable Circular Orbit (ISCO) (see Thorne et al. 2000 for a derivation) is scaled with Rg. Closer than the ISCO radius, there are no circular
geodesic lines for a test particle, and hence the name. Therefore, RISCO poses a
theoretical limit on how close to the black hole the inner edge of the accretion disk can be. For a non-spinning black hole the ISCO radius is 6Rg and for an extreme
spinning black hole the ISCO radius could reach 1Rg and coincides with the event
horizon (Bambi 2017). For a standard Shakura-Sunyaev disk, inside the inner edge of the disk, matter plunges into the black hole quickly in the form of a hot, optically-thick flow and does not contribute much to the electromagnetic emissions. In BH-LMXBs, the standard accretion disk can be truncated before reaching the ISCO and the inner region is dominated by such a hot flow (Narayan & Yi 1994). The outer radius of a standard Shakura-Sunyaev disk is usually set by the Bondi accretion radius or the orbital compactness of the binary (Frank et al. 2002).
The Eddington Luminosity LEdd sets the time when the radiation outward is
so strong that the radiation pressure force balances the gravitational force of the in-falling matter. It is defined as:
LEdd = 4πGMBHmpc σT = 1.26 × 1038 MBH M erg/s, (1.2) where MBH is the mass of the black hole, mp is the mass of a proton, and σT =
Black Hole Low-Mass X-ray Binaries
LEdd, then the accretion process is stopped and matter will be pushed away from the
black hole by radiation pressure. Therefore, LEdd sets an idealised upper limit of the
luminosity that an accreting system can reach if the accretion process is stable. The total energy Laccr released in the accretion process is given by:
Laccr = η
GMBHM˙
R∗
(1.3) (Frank et al. 2002). Here η is the efficiency of the energy of the in-falling matters mpc2 converting to radiation, ˙M is the mass accretion rate and R∗is the inner radius
where the accretion ends, which in a black hole system can be assumed to be the event horizon or the ISCO. A typical value for η for a BH is ∼ 10%. Compared to η ∼ 0.7% for the nuclear fusion p-p chain inside main sequence stars, black hole accretion systems are much more powerful engines of converting other forms of energy into light.
;
Fig. 1.3–An example of the X-ray spectrum of BH-LMXB 4U 1543-475 (upper panel), observed by
space-borne instruments RXTE /PCA (black) and RXTE /HEXTE (red). A phenomenological model of disk emission + power law is fitted to the continuum, in order to best illustrate the broadened iron K-α line around 6.4 keV, in the bottom ratio(=data/model) panel. Adopted from Miller et al. 2009
Fig. 1.3 shows an example of a LMXB spectrum. It is generally considered that the corona, which refers to a population of hot, optically-thin electrons, is located close to the accretion disk (Haardt & Maraschi 1993). The corona electrons can up-scatter the thermal photons from the accretion disk to an X-ray power law1
1Flux in photon counts F = AE−Γ
, where A is the normalization, E is the photon energy and Γ is the photon index parameter that captures the slope of the log-scaled spectrum
Black Hole Low-Mass X-ray Binaries continuum (Rybicki & Lightman 2008). However, the geometry and physical origin of the corona are still open questions (see, e.g. Nowak et al. 2011). Some suggest it could be extended as the Advection-Dominated Accretion Flow (ADAF) surrounding the accretion disk (e.g. Esin et al. 1997), while others suggest it could be a localized source like the base of the black hole jet (e.g. Markoff et al. 2005).
Secondary features superimposed on the continuum, such as broadened iron K-α line at 6.4 keV and Compton hump over 10 keV, are the signatures of the disk reprocessing the primary X-ray emissions from the corona (see Fabian & Ross 2010 for a review). This so-called ”disk reflection” of the corona spectrum can be used to probe the disk-corona interactions or the space-time geometry in the vicinity of the black holes (see, e.g. Cao et al. 2018, Kara et al. 2019).
1.1.2 Jets
Outflows are almost always associated with accretion process, and jets are one the two main types of outflows happening in BH-LMXB systems (besides disk winds). Jets are collimated relativistic outflows of plasma and fields launched from the region close to the accreting black hole, with some suggesting the jet base and the corona may be the same thing (Markoff et al. 2005). A Jet is a collimated outflow that can be accelerated to speeds comparable to the speed of light (Rosswog & Bruggen 2003; Fender et al. 2004). It can produce an approximatively flat Spectral Energy Distribution (SED) extending from the radio to the mid-IR (e.g. Fender et al. 2000; Fender 2001; Corbel & Fender 2002). There are still many debates concerning the jet physics (e.g. Romero et al. 2017), and one of them is the mechanism powering the particle acceleration and/or the collimation of jets (e.g. Spruit et al. 1997; McKinney 2006; Narayan et al. 2014). Jets could be powered by either the energy tapped from the rotation of the black hole (Blandford & Znajek 1977), or by the energy released from the accretion flow (Blandford & Payne 1982). In either scenario, the pres-ence of strong magnetic field is required to accelerate and collimate particles to the direction perpendicular to the accretion disk plane. General Relativistic Megneto-hydrodynamic (GRMHD) simulations of jet formation suggest that both mechanism may play a role (McKinney et al. 2012).
Fig.1.4 shows an example of the multi-wavelength BH-LMXB SED, fitted by a jet model. A natural product of jets being magnetized is the synchrotron emission by particles in the jet, whose optically-thick part makes the radio spectrum flat/inverted and becomes one of the marks of the presence of the jet (Blandford & K¨onigl 1979; Fender et al. 2000; Corbel & Fender 2002). In the example shown here, the optically-thin tail of the synchrotron emission extends to X-ray bands and dominates the emission there. Another important ingredient in jet spectrum is the the Inverse-Comptonization (IC) of the synchrotron photons (Also see §2.1). The seed photons can be the disk thermal photons or the synchrotron photons inside the jet itself, or
Black Hole Low-Mass X-ray Binaries
;
Fig. 1.4–An example of the jet broadband Spectral Energy Distribution (SED), with the
schemat-ics of the jet model on the top right. It is the photon intensity in mJy vs. photon frequency in GHz, both in log scales. See the text for more details about the different jet emissions across the whole spectrum. Adopted from Markoff et al. 2001.
any other external photons (e.g. companion star, cosmic background).
The long term variability of jets in BH-LMXBs is controlled by the status of the accretion process. Jets can be switched on/off between different X-ray spectral states of XRBs (§1.1.3), indicating that jet launching is closely related to the status of the inner accretion region and the accretion disk.
1.1.3 LMXB Outbursts and canonical spectral states
Unlike supermassive black holes at galactic centers, the long term variability (weeks or years) of BH-LMXB systems is controlled not by the accretion from ambient medium but by the mass accretion ˙M via the Roche-lobe overflow, connected to the evolution of the companion star. Most of BH-LMXBs are transient X-ray sources, that have outbursts when the X-ray luminosity increases by several orders of magnitude which is triggered by the instabilities in the accretion disk and caused by an increase in
Black Hole Low-Mass X-ray Binaries
Spectral Hardness
(spectral slope, soft=steep, hard=flat)
X-ray Luminosity Hard Soft Thermal Nonthermal Γ < 2 Γ > 2 ? ? Jet line ?
?
~LEdd ~1% LEdd ;Fig. 1.5 – The hardness-intensity diagram for an outburst of source GX 339-4, with schematics
of the corresponding geometry of the accretion system. See text for more details. Adopted from Romero et al. 2017.
mass accretion rate M ((Done et al. 2007)). As a consequence, during the outburst
LMXB source will go through a sequence of X-ray spectral states that is suggested to be driven by disk instabilities built-up over time (?).
Fig.1.5 shows the schematic view of this state-evolution cycle on the Hardness-Intensity Diagram (HID) of the canonical LMXB GX 339-4 using the actual data, adopted from Romero et al. (2017). The cartoons around Fig.1.5 shows the geometry of the jet/disk system during the different states. In X-ray spectra, the hardness of the spectrum is defined as the photon counts ratio between a high energy band and a low energy band, e.g., Hardness=F3kev−10kev/F0.5kev−3kev. The exact ranges of
energy bands are not defined strictly; hardness is a simple proxy to characterize the slope of the X-ray continuum. A Soft State (SS) means the X-ray has more low-energy photons and a Hard State (HS) means more high-low-energy photons. Based on the luminosity, there are High Hard State (HHS) and Low Hard State (LHS). During a typical outburst, the source first rises out of quiescence and increases its luminosity from the LHS to the HHS (the top right corner), and then it will change from HHS
Black Hole Low-Mass X-ray Binaries
;
Fig. 1.6 – Representative energy spectra of BH-LMXB during the hard (orange) and soft (grey)
states, taken from the observations of MAXI J1535-571 2017 outburst. Adopted from C´uneo et al. 2020.
to Hard-Intermediate State (HIMS), Soft-Intermediate State (SIMS), and eventually to the SS on the left of the HID when no radio jet has been detected. Instead of photons that have higher energy coming from other parts of the system (e.g. the corona, or the non-thermal particle acceleration process inside jet (Markoff et al. 2001) in the HS, in the soft state the X-ray is dominated by the thermal photons from the accretion disk (Fig.1.6). During this outburst, the source experienced state transitions between the soft and hard states. A more detailed category of the spectral states can be found in Homan & Belloni (2005).
As illustrated in Fig. 1.5, in LMXBs the radio signal from the synchrotron emis-sion in the continuous jet is typically detected in the HS, regardless of the source luminosity, and in soft states this part of the components in radio is gone (e.g. Fender et al. 1999; Gallo et al. 2003). When the source is during the intermediate states approaching towards the soft state, instead of a collimated jet, the dominant channel of the accretion outflows becomes the disk wind (Ponti et al. 2012) that can be indicated by the absorption lines observed in the spectra (e.g. Balakrishnan et al. 2020). The idea is that the jet is quenched at soft states but the physics behind this change is still unclear (e.g. Fender et al. 2004; Remillard & McClintock 2006; Fender et al. 2009; Gallo 2015). This project is focused on investigating hard and intermediate states close to the state transitions of XRBs and finding evidence in data that may help us understand how the jet changes with the accretion process and why the jet is quenched during the soft states.
Analysis Techniques Another open question in the field is the accretion disk geometry during the state transitions of LMXBs. Based on the X-ray reflection spectroscopy and also the short-term variability studies, some suggest that disk can be truncated in the HSs (Plant et al. 2014, Ingram & Done 2011). Disk truncation means that the switch between the standard Shakura-Sunyaev disk and the inner hot flow is not at the ISCO radius but farther away. As the outburst goes towards SS, the optically thick, geometrically-thin region gradually develops inwards, and with the radius of the inner edge getting smaller and smaller (and eventually to the ISCO radius), the disk becomes hotter and more luminous, causing the X-ray spectra to be more thermally-dominated (Esin et al. 1997). However, there is also some evidence suggesting a non-truncated or mildly-truncated (6 20Rg) disk in the HSs, and instead, it is the
corona geometry that differs between soft and hard states (Kara et al. 2019, Jiang et al. 2020). Either way, the geometry changes in the inner accretion region play an important part in the spectral state transitions, which might also be related to the launching and shutting down of jets.
Therefore, it is essential to combine the broadband spectral analysis (§??) with X-ray variability studies (or so-called ”X-ray timing analysis”, §1.2.2) to understand how the jet-disk coupling changes throughout the outburst of BH-LMXBs, especially during the state transitions when we know there is a major change in the jet and the disk physics based on the evidence mentioned in this section. While there are studies focused on only the hard states of individual sources, attempting to find links between the spectral and timing properties (e.g. Connors et al. 2019), a systematic study of multiple sources in a variety of states, with a physical model, has not been performed. This is necessary in order to find similarities in different BH-LMXBs, and to identify the main driver causing the state transitions, that might also be responsible for the jet differences between states. This thesis aims to perform such an innovative spectral-timing study, to probe the physical changes in jet-disk systems during the state transitions, that might also be behind any correlations between spectral and timing properties during the spectral evolution of BH-LMXB in outbursts (§1.2.2).
1.2
Analysis Techniques
1.2.1 Broadband Spectral Analysis
The spectra of XRBs are generated by different mechanisms and from different parts of the system, such as synchrotron photons from the jet dominate the radio or ther-mal disk photons dominate the soft X-ray spectra. From the study of spectrum, i.e. the energy (photon frequency) distribution of photons received from the source, much information can be extracted about the physics of the source by comparing the observed data with our theoretical model.
Analysis Techniques
Fig.1.4 shows an example of the jet model fitting to the real SED of the BH-LMXB XTE J1118+480 (Markoff et al. 2001), with the schematics of jet model on the top right corner. Multiple radiative mechanisms contribute to the emission across a wide range of frequencies (see §2.1). In order to understand the relative importance of these various components at different frequencies, it is essential to analyze multi-wavelength data, from radio such as VLA observations (Thompson et al. 1980), to X-rays such as RXTE observations (Jahoda et al. 1996).
1.2.2 X-ray Timing Analysis
Another powerful method in the studies of XRBs is analyzing the source variability; this is typically done with X-ray instruments. Instead of looking at the energy distri-bution of photons, X-ray timing analysis focuses on extracting information from the data over time, often using Fourier techniques that tells you what is the frequency distribution in the lightcurve, and hence the variabilities over different timescales. The lightcurves of observations can be Fourier transformed into power spectra2 that reveal the X-ray variability in the lightcurves. The shape of the power spec-tra and the Quasi-Periodic Oscillation (QPO)s superimposed on the power specspec-tra (Fig.1.7) are correlated to the spectral states of the LMXB (Belloni 2010), which suggests that different spectral states have different characteristic short-term vari-ability. This close relation between the spectral and the timing property in the X-ray behaviour of XRBs is still an open question in the field (Ingram & Motta 2019).
The work in this thesis also relies on X-ray timing analysis to find similarities in the state evolutions of different XRB systems, following the idea that the timing properties of X-ray observation indicate geometric effects. The shape of the power spectrum can be characterized in power color indices that are described in §2.2. The correlation between power color indices and the spectral states is another remark-able evidence that the spectral and timing properties of XRB outbursts are closely associated (e.g. Homan et al. 2001), and this thesis is also trying to find evidence for the physical links behind this correlation. See §2 for more about this correlation and how it is explored in this work.
2More precisely speaking, a power spectrum is the modulus squared of the Fourier transform of
Motivation 0.01 0.1 1 10 10−10 10−8 10−6 10−4 10−2 100 Frequency (Hz) Squared rms / Hz 1 2 3 4 5 ;
Fig. 1.7 – Examples of power spectra at different evolution phases during the outburst, with
different types of QPOs as peaks on top of the power spectra. Adopted from Belloni (2010). See the original paper for the detection and classification of QPOs. Spectra are shifted in power for clarity.
1.3
Motivation
As the previous sections show, one of the main questions in black hole astrophysics is how the jet is coupled to the accretion flow over time. This question of interaction leads to many unexplained phenomena including why jets cease to exist after the hard-to-soft state transitions during BH-LMXB outbursts, and why the presence of jets distinguishes different AGN species. While there are a lot of X-ray studies using phenomenological models such as a broken power-law for the inverse-comptonization inside corona, and several hard-state-only broadband studies on individual sources (e.g. Lucchini et al. 2019, Connors et al. 2019), attempting to find links between the timing and spectral properties, a systematic spectral-timing study of multiple BH-LMXB outbursts, during the state transitions with broadband physical models, is not present but needed in order to help identify what is/are changing in the physics during the state transitions concerning the jet-disk coupling. This is the goal driving the explorations in this thesis. Furthermore, we want to search in the data for any trends that might help to isolate the physics behind the correlation between the
Outline
spectral and timing properties of BH-LMXB. In summary, this thesis aims to test whether data can be interpreted as there are any trends during the hard-to-soft states spectral evolution of XRBs, that mark the jet-disk changes and potentially link the spectral and timing evolution of the sources, by performing a systematic analysis on multiple XRB sources at different outburst stages, and correlating the parameters, with ”hue” angles / the shape of the power spectra (§2.2).
1.4
Outline
This thesis is structured as follows: Chapter 2 introduces the jet model (§2.1) used in this project to fit the XRB spectra and extract the geometric information, as well as the empirical method of power colors to capture the X-ray timing properties of sources (§2.2). Chapter 2 also has a brief description of the selection criteria for the sources and epochs considered in this study, while the detail of all the observations used in this project is presented in Chapter 3. Results are presented in Chapter 4, with interpretations and further diskussions provided in Chapter ??. The research of this thesis is concluded in Chapter 6. Bibliography is provided in the Appendix as the final part of this thesis.
Chapter 2
Methods
In order to understand the coupling between the jet and accretion flow, we studied several XRB outbursts during the hard states and during the hard-to-soft transitions after which the jet shuts down. We want to find evidence in the data and identify what physical changes are happening during the state transitions, by fitting and comparing the the multi-wavelength (Radio to X-rays) data of XRBs at different epochs during the spectral evolution. In this thesis we focus on testing whether we can find trends in how the jet-disk system is changing when the sources are moving towards the soft states, and to see whether these spectral and timing changes are mostly driven by a change in the geometry of the system, by using systematic analysis of multiple targets.
Sources were selected selected based on the availability of multi-wavelength data (§2.3). We used bljet, a multi-wavelength jet model described in §2.1, to fit the broadband SED and estimate the source geometry. The timing properties of each epoch are estimated by calculating the power spectral ”power colors”, following the empirical approach of (Heil et al. 2015) (§2.2). Eventually we used the results from both the spectral and timing analysis to probe the changes of the jet-disk system with respect to the different phases in XRB outbursts, that might be behind the spectral-timing correlation during the LMXB outbursts.
2.1
Jet model
The jet model bljet used in this work to fit the broadband spectra (Radio to X-rays) of BH-LMXBs is a reworking of the model presented first in Markoff et al. (2005), and further developed in Maitra et al. (2009); Connors et al. (2017, 2019). The previous versions (called Agnjet ) have mainly been used in studies of HSs and the quiescence of LMXBs (e.g. Markoff et al. 2005; Gallo et al. 2007; Connors et al. 2017). It can
Jet model also reproduce the broadband SEDs of some AGN of low luminosity (e.g. Maitra et al. 2011; Markoff et al. 2015; Connors et al. 2017). Model bljet shares the same treatment of particle acceleration and radiation with its predecessor. Additionally it includes higher jet bulk speed possible in the model and proton-electron pair content that better parameterizes the GRMHD simulation results. (Lucchini et al. (2019), and further changes discussed in Lucchini et al. (submitted )). The details of the model can be found in the papers mentioned. Briefly, bljet is a steady-state, multi-zone, and semi-analytical model, that parameterizing the GRMHD works, attempting to combine simulation studies with a model that could be used in multi-wavelength data analysis.
Besides a thermal Shakura-Sunyaev accretion disk (Fig.2.1), the model param-eterizes the power of the jet as a percentage Nj of the black hole’s Eddington
lumi-nosity LEdd, divided among hot electrons, cold protons and magnetic fields, injected
into the jet at the jet base. The base of the jet is a simple cylinder perpendicular to the accretion disk, and can be thought of as a lamp-post corona (e.g. Miller et al. 2015; F¨urst et al. 2015), starting from 2Rg. In this work, the radius of the jet base
R0 is a free parameter measured in units of Rg while the height of this cylinder h is
fixed to be 2 ∗ R0. The jet bulk acceleration starts at the top of the jet base, tuning
the initial magnetic field content of the jet into bulk kinetic energy up to a distance zacc from the black hole. The jet magnetization parameter σ is defined as:
σ(z) = Pb(z) + Ub(z) Up(z) + Ue(z) + Pe(z)
(2.1) where Pb(z) and Ub(z) are the pressure and the internal energy of the magnetic
field, Ue(z) and Pe(z) are those of the electrons, and Up(z) the energy density of the
protons which, since they are cold, is npmpc2. σ captures the relative strength of
the magnetic field during the bulk acceleration, in which process the conservation of energy holds. At a distance zdiss, which in this study is taken to be equal to
zacc, a dissipation region occurs, and a fraction pl=10% of the thermal electrons
are channeled to a non-thermal, law particle distribution tail with a power-law index of p, which roughly mimics the behaviour of shock acceleration seen in particle-in-cell (PIC) simulations (Sironi & Spitkovsky 2010). This dissipation region could represent the shock region in the jet that accelerates electrons to produce high energy electrons responsible for the synchrotron radiation observed in the broadband data of LMXBs (Blandford & K¨onigl 1979; Boettcher & Dermer 2010; Malzac 2013), which region might be caused by fluctuations of the outflow velocity (Malzac 2014). As we go further out from zacc, the percentage of accelerated particles along the jet
are reduced with respect to the distance z by a factor (log10(zacc)/log10(z))fpl to
characterize the dissipation of jet particles using fpl parameter. A short list of the
important parameters and descriptions concerning the work in this thesis are listed in Table 2.1, while the full list of parameters in bljet could be found in Lucchini et al. (submitted ).
Jet model
Our jet model only considers leptonic processes (Rybicki & Lightman 2008); the protons carried in the jet are always assumed to be cold and therefore do not con-tribute to the SEDs. We compute the synchrotron and inverse-Compton emission from both thermal and non-thermal leptons throughout the outflow. The accretion disk emits thermal photons, which then could be inverse-Comptonized by the syn-chrotron inside the jet (e.g. Romero et al. 2017). Note, here one of the most crucial assumptions is that the jet base serves as the black hole corona re-illuminating the disk, as suggested by both the jet SED studies (Markoff et al. 2005) and X-ray re-verberation studies assuming a corona with a lamp-post geometry (e.g. Martocchia & Matt 1996; Middleton 2016). So our best-fits to the SEDs can also help elucidate to what extent our multi-wavelength (Radio to X-rays) data tolerate the ”corona = jet base” assumption. Hereafter the terms ”corona” and ”jet base” are used inter-changeably. The corona then re-illuminates the accretion disk and generates the disk reflection spectrum (Garcia et al. 2013). Meanwhile, following the standard scenario of jet emission studies (Romero et al. 2017), inside the jet our model considered syn-chrotron radiation and its self-comptonization (SSC), and inverse-comptonization of disk thermal photons (Fig.2.1).
In most cases, the multi-wavelength data available in the archive is not sufficient to constrain all the parameters (Chapter 3), and thus we freeze several parameters in our SED modelling. XRB jets are only moderately relativistic (Fender et al. 2004), and hence we take the terminal bulk Lorenz factor of the jet to be Γ = 3. The final magnetization parameter σf after the bulk acceleration at zacc is set to
be 0.1, in broad agreement both with the GRMHD simulations ( Chatterjee et al. 2019, McKinney 2006) and physical modelling of AGN jets (Lucchini et al. 2019). In agreement with the estimates of Gandhi et al. 2008 for the canonical XRB GX 339-4, p and zacc are fixed to 2.3 and 2000Rg. Other parameter values fixed among fits can
be found in Table 2.1 as well.
The two primary geometrical parameters we study in this work are the size of the jet base R0, and the radius of the inner edge of the Shakura-Sunyaev disk Rin
(§1.1.3). R0 mainly sets the radiation energy density and optic depth of the corona,
and thus it has large impacts on the flux of the thermal synchrotron emission inside the corona, as well as on the inverse-comptonization of the disk and synchrotron photons. Rin controls the X-ray flux from the outer disk since the disk has a higher
temperature at smaller radius (Frank et al. 2002). Rin sets the radius minimum the
Shakura-Sunyaev disk could reach, and inside Rinwe assume there is ADAF (§1.1.1)
Jet model
;
Fig. 2.1 – Schematic of the jet emission model used in this work. The thermal photons from the
accretion disk are inverse-comptonized (IC) by the synchrotron electrons inside the jet. The jet base serves as the black hole corona to re-illuminate the disk and generate disk reflection spectrum. Also jet synchrotron is subject to self-comptonization (SSC). See the text for more details on the model.
Jet model
Table 2.1A full list of parameters and descriptions in the jet model used in this work, with values of
all the freezed parameters given in the table. Full details about the model parameters could be found in Lucchini et al. (2019) and Lucchini et al. (submitted ). Note that there is no free parameters for normalization, and the normalization is achieved only through physical parameters such as distance d.
Parameter Description
MBH Mass of the black hole in units of M .
θi Viewing angle between the observer and the jet direction.
d Distance of the black hole to the observer in units of kpc. Nj Total power channeled into the jet base, normalized by the
Eddington luminosity LEdd of the black hole.
R0 Radius of the cylindrical jet base/corona in units of Rg.
Te Temperature of the relativistic electrons in the cylindrical
jet base/corona in units of keV .
fpl Reduces particle temperature and percentage of
acceler-ated particles along the jet after the shocks by a factor of (log10(zacc)/log10(z))fpl, parameterizing the effect of
adia-batic cooling in the outer regions of the jet.
Rin Radius of the inner edge of the accretion disk in units of Rg.
Ldisk Luminosity of the disk in units of LEdd.
zacc=2000Rg Distance along the jet direction where the jet bulk
accel-eration ends and particle accelaccel-eration by magnetic shocks happen.
zmax=1 × 109Rg The maximum length of the jet considered.
pl=0.1 The number fraction of thermal particles accelerated into a
power-law tail.
p=2.3 Slope of the power-law tail of the non-thermal particle dis-tribution.
fb=0.1 Sets the effective adiabatic cooling timescale tad = R(z)·c
fb
(R(z) is the jet radius at distance z), which determines the cooling break energy in the particle distribution.
βp=0.0315 Defined as βp = UUeb(z)(z), together with σ (eq.2.1) determines
the number ratio between protons and electrons, which will be at order of ∼ 10 in our cases, see Lucchini et al. (submit-ted ) and Crumley et al. (2016) for the choices of βp values.
fsc=0.1 Efficiency of the shock acceleration, which sets the maximum
energy cutoff of the particle distribution.
σf = 0.1 The magnetization parameter after the jet bulk acceleration
stops at zacc
Rout=100000Rg Outer radius of the accretion disk.
γ0=1.09 Initial bulk Lorentz factor of the jet.
γ =3 Final bulk Lorentz factor of the jet after the bulk accelera-20
Power colors
2.2
Power colors
Power color is an empirical description capturing the shape of the XRB’s X-ray power spectra and linking the X-ray timing properties with the spectral states of XRB outbursts (Heil et al. 2015). It is clear that there is an evolution of broadband variability properties in X-rays through different spectral states during the XRB outbursts (see e.g. Dunn et al. 2010), and power color provides an empirical way to summarize the timing evolution with respect to different spectral states. Inspired by the rms-intensity relations of XRB data (Heil et al. 2012), power colors are defined as the ratio amount of variance in different frequency bands of the power spectra, effectively measuring the ”colors” of the power spectra. Fig. 2.2 (left panel) sketches how those ratios of variance values, or ”power color ratios”, are calculated. Following Heil et al. 2015, Power Colour Ratio 1 (PC1) is defined as the variance in 0.25-2.0 Hz/0.0039-0.031 Hz, and Power Colour Ratio 2 (PC2) as the variance in 0.031-0.25 Hz/2.0-16.0 Hz. A power-color diagram can be drawn with PC1 on the X-axis and PC2 on the Y-axis, both in log-scale (Fig.2.2(right)). To distinguish observations with different power colors, one can define ”power spectral hue”, which is the angle between the point marking two power color ratios of a certain observation, and the central point [4.51920, 0.453724] on the diagram, where hue=0◦ corresponds to the semi-major axis at 45◦ to the x-and-y axes.
Heil et al. (2015) studied 12 transient XRB sources (Fig. 2.3) and found out that different hue ranges on the power color diagram correspond to different spectral states (Fig. 2.2 (right)). Power colors/hue angles can thus be used to trace the outburst evolution and identify the source spectral state. During an outburst, a BH-XRB will start from the top of the power color diagram and evolve clockwise, as the spectral state evolves from the hard state to the soft state, and then counter-clockwise to get back to the hard state or into quiescence (§1.1.3). It is clear that the evolution of the power color is closely related to the HID spectral evolution, but the physical reason behind this link between timing and spectral properties is still an open question.
This thesis aims to exploit this spectral-timing correlation and to test the impact of the geometry of the jet-disk system on both the timing and spectral properties, by gathering epochs of quasi-simultaneous multi-wavelength (Radio to X-rays) data from multiple BH-LMXB sources from hard states to the hard-to-soft transitions (§2.3), fitting the epochs with our advanced jet model to extract the values of geometric parameters (R0&Rin), and correlating them with the hue angle derived from the
Power colors
;
Fig. 2.2 – Left: A conceptual diagram representing the power spectrum of one XRB X-ray
ob-servation, divided into frequency bins in log space. Power color ratios PC1 and PC2 are defined as the ratio between two integral power across defined frequency bands A, B, C and D. See text for more details. Right: The power color diagram and the introduction of hue angles, where hue=0◦ corresponds to the semi-major axis at 45◦ to the x-and-y axes. Different hue ranges correspond to different spectral states (hard, hard intermediate HIMS, soft intermediate SIMS, and soft). In the overlap region on the top left of the diagram, both soft and hard states might be found. See text for more details on the evolution of hue angles and its correspondence to the spectral states. Both left and right figures adopted from Connors et al. 2019.
Power colors
;
Fig. 2.3 – Power colors for all the X-ray observations of 12 BH-LMXB sources. color schemes
corresponds to different value regions for hue angles (1-18). Adopted from Heil et al. 2015. Once the spectral state of the XRB source is known, the area on this plot in which the associated power colors could be found, providing us a correlation between the hue and the X-ray spectral states. The outburst evolution of XRBs (§1.1.3) can also be traced by the clockwise and counter-clockwise evolution of hue angle.
Targets Selection
2.3
Targets Selection
This section introduces the selection criteria that are used in this work to find suitable BH-LMXB sources for further broadband spectral analysis and X-ray timing analysis. In order to capture both the behaviour of the jet and the accretion disk, ideally radio, infrared/optical/UV and X-ray data are needed. Due to the short timescale of jet-disk interaction (Fender et al. 2004), our work requires that the radio, X-ray, and infrared/optical/UV/ observations to be quasi-simultaneous, which we defined as observations performed within two days of each other, to make it a suitable multi-wavelength epoch for our study. The IR/optical/UV data are essential to minimize the degeneracy in our jet model. Furthermore, suitable targets should have Rossi X-ray Timing Explorer (RXTE) observations available because RXTE was the best timing instrument until a few years ago, and its energy range is what the power colors were first defined for. In practice this requirement does not narrow our scope of targets since most of XRB sources have RXTE observations due to its long on-duty time (1995-2012).
Other selection criteria are posed by the applicability of the jet model used. The presence of a compact jet requires the XRB radio SED to be flat or inverted (Fender et al. 2004) in order to avoid epochs of radio flaring/ejections. We also require that epochs should have power-law dominated X-ray spectra that have spectral and timing characteristics consistent with a HS, HIMS, or close to a transition. Therefore, the selection of epochs are restricted in those states.
The database used in this work is the Whole-sky Alberta Time-resolved Compre-hensive black-Hole Database Of the Galaxy, or WATCHDOG (Tetarenko et al. 2016). This comprehensive database includes 47 transient and 10 persistent XRB sources with confirmed black holes or black hole candidates as the primary objects. The total number of transient outbursts recorded is more than 130 over the two decades con-sidered in this database (1996-2015). The WATCHDOG website1 provides machine readable formats of all the data, making it very convenient to perform selection with customized conditions over the entire BH-LMXB species. We have considered 42 XRBs with black holes (black hole candidates) that have reported radio detections in HSs or HIMSs. We then searched the literature to to check each of those candi-date sources if any one is suitable for our further analysis according to the selection criteria of this thesis (see above in this section). Eventually three optimal sources are selected in this work, and they are: XTE J1752-223, MAXI J1659-152, and XTE J1650-500. All three sources have more than one epoch when multi-wavelength, quasi-simultaneous data are available. All the observations used in this thesis and epoch details for all three sources are listed in the next chapter of Observations and Data Reduction.
1
Chapter 3
Observations and Data
Reduction
The timing analysis of all the RXTE observations selected in this work was performed using the Chromos1pipline (Gardenier & Uttley 2018). It is developed for automating the extraction of spectra and lightcurves from the RXTE data archive, as well as the calculation of power spectra and power colors. Data from X-ray instruments other than RXTE (if available) will only be subject to spectral analysis. Meanwhile, for RXTE spectral analysis, this work uses the standard data product produced by the RXTE pipeline from the archive2, where the data reduction follows the standard procedure for RXTE data. The energy range considered in this work for RXTE/PCA is 3keV-22keV, and 20keV-200keV for RXTE/HEXTE if available. Data from both instruments are grouped into energy bins to meet a minimum signal-to-noise ratio to be 4.5 in each bin in order to use the χ2-statistics. Finally, a systematic error of 0.5% is introduced to RXTE/PCA before spectral fittings to account for uncertainties in the detector response as suggested in Shaposhnikov et al. (2012).
All the X-ray data reduction in this thesis is performed using HEASoft tools (version 6.26.1), including Chromos RXTE pipeline and online Swift xrtpipeline. Radio and IR data in this work is taken from the literature on each individual source. We supplement our RXTE hard X-ray data with soft X-ray data from the XRT instrument on board Swift, and optical/UV data from Swift/UVOT. X-ray spectra from Swift/XRT are extracted using the online xrtpipeline provided by Evans et al. (2009)3, and the energy range considered is 1.0keV-10keV to avoid detector features below 1.0keV. Spectra are grouped to reach a minimum signal-to-noise ratio of 20 before spectral analysis. Finally, we included any available Swift/UVOT observations
1
https://github.com/davidgardenier/chromos
2https://heasarc.gsfc.nasa.gov/docs/archive.html 3
Chapter 3. Observations and Data Reduction taken together with the Swift/XRT data. All the optical/UV images of Swift/UVOT, if available, are taken directly from the archive, and uvotsource is used to extract the photometry of selected sources from the image. The source regions are chosen to be circular, with the centers the same as the coordinates of our sources provided in astronomy database4, and a radius of 5 arcsec as suggested by Swift team5. The backgrounds are chosen to be a larger circular area without any other sources and close to our targets, in order to minimize the impact from background fluctuations. The radii of the background regions vary among targets depending on the sky where they locate. Fig.3.1 shows an example of region selection during the flux extraction of Swift/UVOT data.
The de-reddening of IR/optical/UV data was performed manually. The hydro-gen column density NH in cm−2 related to the galactic extinction for each target
is adopted from existing publications (Table 3.1, see following sections). We esti-mate the extinction using the AV ∼ NH relation in Foight et al. (2016): NH =
(2.87 ± 0.12) × 1021AV. Finally, we re-scale the extinction AV to various
IR/op-tical/UV bands Aω using the online tool6 based on the extinction law studied in
Cardelli et al. (1989). The measured IR/optical/UV flux in mJy at wavelength ω band is de-reddened by a factor of 100.4Aω. This allows us to use consistent
ex-tinction/absorption in the IR/optical/UV and X-ray bands, and thus during the broadband spectral fittings the NH in the X-ray extinction model TBabs (Wilms
et al. 2000).
For each source, we fixed the BH mass, inclination, and distance based on ex-isting estimates in the literature. These are reported in Table 3.1, with associated references mentioned in the following sections.
Table 3.1 List of known parameters for selected sources in this work. MBH is the black hole mass
in units of M , d and θi the distance and inclination angle of the system, and NH the hydrogen
column density for interstellar extinction. We take these adopted values as known in our spectral analysis. See the text for the references for these values.
MBH (M ) d (kpc) θi (◦) NH (1022cm−2) XTE J1752-223 9.6 6 35 1.0 MAXI J1659-152 6 6 75 0.32 XTE J1650-500 5.1 2.6 45 0.5 4 http://simbad.u-strasbg.fr/simbad/ 5https://www.swift.ac.uk/analysis/uvot/mag.php 6 http://www.dougwelch.org/Acurve.html
XTE J1752-223
;
Fig. 3.1 – An example of the region selections in Swift/UVOT data reduction. The small circle
is the source region with 5 arcsec radius and the larger is the background region. The image is taken from UVW1(∼ 0.26µm) band, Swift/UVOT observation on MAXI J1659-152. Observation ID:00434928011.
3.1
XTE J1752-223
The X-ray transient XTE J1752-223 was discovered by RXTE in 2009 October (Mark-wardt et al. 2009) when it went into an outburst, lasting for almost 8 months; the evolution of the outburst was typical of that of a black hole system, as shown in Fig. 3.2. The radio emission is consistent with a typical compact jet in the hard state, and shows optically thin flares in the soft states (Brocksopp et al. 2013). The mass of the black hole is estimated to be 9.6 ± 0.8M , using correlations between
spectral and variability properties with GRO J1655-40 and XTE J1550-564 (Sha-poshnikov et al. 2010), while there is a lack of dynamical mass constraint. From detailed modelling of the reflection signatures in the X-ray spectra for the month-long high-hard state with stable X-ray luminosity and hardness, the inclination angle of XTE J1752-223 is estimated to be 35◦± 4◦, and the galactic extinction NH to be
1.0 × 1022cm−2 (Garc´ıa et al. 2018).
The distance d to XTE J1752-223 is not well constrained. A distance of 3.5 ± 0.4 kpc was determined by Shaposhnikov et al. (2010) using the same technique for mass estimation. However, modelling of X-ray photoelectric absorption edges in the source (Chun et al. 2013) suggests a larger distance (>5 kpc), and the optical detection on the companion star during the quiescence of XTE J1752-223 (Ratti et al. 2012) also favors a larger distance up to 8 kpc. A distance larger than 3.5 kpc also agrees with the high column density NH from X-ray spectroscopy; at a distance of 3.5 kpc on
XTE J1752-223
;
Fig. 3.2–X-ray HID for XTE J1752-223 2009 outburst from RXTE data, adopted from Brocksopp
et al. (2013). The Y-axis is the net photon counts per second observed by RXTE/PCA unit PCU2. The four epochs studied in this work are highlighted by red circles and labeled in A/B/C/D. Epoch A was during the month-long high-hard state when hardness and luminosity were stable; Epoch B coincided with the radio peak detected by ATCA; Epoch C marks the return from soft state and Epoch D marks the re-brightening in radio. All four epochs are in HSs or HIMSs with radio presence (but not optically-thin radio flares).
of what is suggested by X-ray studies (Chun et al. 2013). Therefore, in this work we take 6 kpc as the distance of XTE J1752-223, a median value from the estimated range of the distance in Ratti et al. 2012.
RXTE monitored XTE J1752-223 throughout the entire 2009-2010 outburst. In this work we only downloaded, reduced, and analyzed RXTE data taken on four dates (year-month-day: 2009-11-05, 2010-01-21, 2010-04-14, 2010-06-03), highlighted by red circles in the HID in Fig.3.2. These four dates correspond to epochs in which the compact jet is detected by Australia Telescope Compact Array (ATCA) (Brocksopp et al. 2009). All the radio and X-ray observations of XTE J1752-223 used in this work are listed in Table 3.2 and Table 3.3, and the quasi-simultaneous UV data for XTE J1752-223 is provided by Swift/UVOT, while Swift/XRT provides the soft X-ray spectra down to 1 keV.
MAXI J1659-152
Table 3.2List of radio observations of XTE J1752-223 analyzed in this work, detected by ATCA.
Flux data in units of mJy in two frequency bands are adopted from Brocksopp et al. (2013) and labeled by according to their corresponding X-ray pointings.
Date MJD 5.5 GHz (mJy) 9 GHz (mJy) Epoch 2009-11-05 551 40.4 1.87 ± 0.07 2.05 ± 0.07 A 2010-01-21 552 17.9 20.00 ± 0.06 21.71 ± 0.04 B 2010-04-14 553 00.9 1.07 ± 0.09 0.09 ± 0.05 C 2010-06-03 553 50.5 0.20 ± 0.03 0.18 ± 0.03 D
Table 3.3 List of X-ray observations of XTE J1752-223 analyzed in this work, with associated
epochs labeled. The UV flux of each epoch is derived from SwiftUVOT images taken during the same observations listed. On 2009-11-05, RXTE took three successive observations and in this work we only study the first one for spectral analysis, which is the closest to the radio observation in time on the same day, but include the rest two in power-color timing analysis. Swift has no observations coinciding Epoch B, due to the orbital restraints of the satellite.
Satellite Date ObsID Epoch RXTE 2009-11-05 94331-01-02-11 A 94331-01-02-13 94331-01-02-17 2010-01-21 94331-01-06-02 B 2010-04-13 95360-01-12-03 C 2010-06-03 95702-01-07-03 D Swift 2009-11-03 00031532009 A 2010-04-14 00031688002 C 2010-06-05 00031688022 D
3.2
MAXI J1659-152
MAXI J1659-152 was detected by Swift and Monitor of All-sky X-ray Image (MAXI) during its 2010 outburst (Mangano et al. 2010, Negoro et al. 2010). Its orbital period of ∼ 2.4hr (Kuulkers et al. 2013) makes it the shortest period BH-LMXB source known (Kennea et al. 2010). In this work we take 6+1.8−1.3 M as the black hole mass,
estimated by Molla et al. (2016). The inclination angle of the system is constrained to be 65◦< i < 80◦by using the cyclical absorption dips in X-ray lightcurves (Kuulkers et al. 2013). In this work we take i = 75◦.
MAXI J1659-152
;
Fig. 3.3–X-ray HID for MAXI J1659-152 2010 outburst from RXTE data, adopted from Kuulkers
et al. (2013). The three epochs analyzed in spectra in this work are highlighted by red circles and labeled in A/B/C.
Like XTE J1752-223, the distance to MAXI J1659-152 is poorly constrained, ranging from 1.6 to 8.6 kpc (Miller-Jones et al. 2011, Kennea et al. 2010, Jonker et al. 2012, Kong 2012, Kuulkers et al. 2013). Here we take it to be 6 kpc. This value reconciles the works above and is in agreement with Kong 2012 who suggest the companion star is an M2 dwarf.
In van der Horst et al. (2013) they presented and analyzed multi-wavelength data of the entire outburst, lasting about 40 days. They estimate a line-of-sight column density NH = (0.319±0.009)×1022cm−2. The multi-wavelength data are presented in
Table 3.4 and Table 3.6. The three epochs in which radio, IR/optical/UV and X-ray data are available and our jet model is applicable (§2.3) are selected and highlighted by red circles in Fig.3.3. In this work, we focus on the transition from the HS to the SS, and therefore we analyzed observations from 2010-09-29 to 2010-10-15 only. These are reported in Table 3.5. Note that we choose only RXTE observations with scientific event observation modes (i.e. Event and Good Xenon modes) during the selected period of time, because they have higher time resolution than the standard data mode (i.e binned mode), which allows the power color analysis to be performed (Heil et al. 2015).
MAXI J1659-152
Table 3.4 List of radio observations of MAXI J1659-152 analyzed in this work, detected by Very
Large Array (VLA). Flux data adopted from van der Horst et al. (2013), labeled with Epoch A/B/C. Date MJD 4.9 GHz (mJy) 8.5 GHz (mJy) 22 GHz (mJy) 43 GHz (mJy) Epoch 2010-09-29 554 68.05 9.88 ± 0.30 10.03 ± 0.31 11.81 ± 0.71 11.19 ± 0.59 A 2010-10-01 554 70.06 10.29 ± 0.32 9.74 ± 0.30 8.84 ± 0.49 4.84 ± 0.35 B 2010-10-03 554 61.98 9.23 ± 0.28 7.55 ± 0.42 7.88 ± 0.42 3.74 ± 0.40 C
MAXI J1659-152
Table 3.5 List of X-ray observations of MAXI J1659-152 downloaded, reduced, and analyzed in
this work. From 2010-09-29 to 2010-10-15 only a fraction of the RXTE observations has science event observation modes available for power color analysis (see text), which is all included in this work. Timing analysis has been performed on all the listed observations in order to track the power color evolution, while spectral analysis only considers the observations that coincide with the radio/IR/UV observations (the closest if there are multiple observations on the same day), labeled by Epoch A/B/C.
Satellite Date ObsID Epoch RXTE 2010-09-29 95358-01-02-01 A 2010-10-01 95358-01-03-00 2010-10-01 95108-01-02-00 B 2010-10-02 95358-01-03-01 2010-10-02 95108-01-03-00 2010-10-02 95108-01-04-00 2010-10-03 95108-01-05-00 C 2010-10-03 95358-01-03-02 2010-10-03 95108-01-06-00 2010-10-03 95108-01-07-00 2010-10-04 95108-01-08-00 2010-10-04 95108-01-09-00 2010-10-04 95108-01-10-00 2010-10-05 95108-01-11-00 2010-10-05 95108-01-12-00 2010-10-06 95108-01-13-00 2010-10-06 95108-01-14-00 2010-10-07 95108-01-15-00 2010-10-07 95108-01-16-00 2010-10-07 95108-01-17-00 2010-10-07 95108-01-18-00 2010-10-08 95108-01-18-01 2010-10-15 95118-01-01-00 2010-10-15 95118-01-01-01 Swift 2010-09-29 00434928007 A 2010-10-01 00434928009 B 2010-10-03 00434928011 C
MAXI J1659-152
Table 3.6 List of IR/UV observations of MAXI J1659-152 used in this work. The magnitude of
the data in each filter without de-reddening is adopted from van der Horst et al. (2013). Data are de-reddened (See text above) before the spectral analysis.
Date MJD Instrument Filter Magnitude Error Epoch 2010-09-29 55468.044 SMARTS J 15.13 0.13 A 55468.05 SMARTS H 14.7 0.13 A 55468.494 UVOT U 15.84 0.025 A 2010-10-01 55469.79 BOOTES-2 R 16.59 0.06 B 55469.993 SMARTS J 15.26 0.1 B 55469.999 SMARTS H 14.88 0.21 B 55470.481 UVOT UVM2 16.702 0.031 B 2010-10-03 55471.996 SMARTS J 15.26 0.16 C 55472.002 SMARTS H 15.02 0.08 C 55472.12 UVOT UVW1 16.317 0.027 C
XTE J1650-500
3.3
XTE J1650-500
XTE J1650-500 was detected by RXTE during its 2001 outburst (Remillard 2001). We adopt 5.1 M for the mass of the black hole, as estimated by Slan`y & Stuchlik
(2008) and within the mass limit estimated by Orosz et al. (2004) (2.7 < MBH < 7.3
M ). We take an intermediate inclination i=45◦, following the studies of modelling
the reflection X-ray spectra of the source (Miller et al. 2002, Miniutti et al. 2004). The distance of XTE J1650-500 is estimated to be 2.6 ± 0.7 kpc by empirically studying the X-ray luminosity of XRBs during the state transitions (Homan et al. 2006). NH estimations vary among studies but all favor a moderate absorption (e.g.
Miller et al. 2002, Montanari et al. 2009); here we take NH = (0.5 ± 0.1) × 1022cm−2
from (Miniutti et al. 2004) as the column density for XTE J1650-500.
Fig. 3.4 shows the path of the source on the HID during its outburst, using RXTE data. In total there are eight radio observations simultaneous with X-ray observations, but only two epochs have simultaneous IR data in the HS and HIMS, indicated by red circles in Fig.3.4. It is essential that at least one more energy band other than radio or X-ray is available to minimize the degeneracy in our jet model (§2.3). And thus multi-wavelength spectral analysis is only performed on these two epochs. See Table 3.7, Table 3.8, and Table 3.9 for the lists of radio, X-ray, and IR observations used in this work respectively. Note that we downloaded, reduced and analyzed all RXTE observations between the first and the second epoch mentioned above, that are in proper observe mode (Heil et al. 2015) to allow power color analysis. This allows us to track the evolution of power colors during this transition of hardness states.
Table 3.7 List of radio observations of XTE J1650-500 used in this work. All flux density data is
in units of mJy, adopted from Corbel et al. (2004) and labeled by Epoch A/B.
Date MJD 1384 MHz 2496 MHz 4800 MHz 8640 MHz Epoch 2001-09-08 521 60.81 4.08 ± 0.20 5.30 ± 0.15 5.28 ± 0.10 4.48 ± 0.10 A 2001-09-24 521 77.01 - - 0.83 ± 0.10 0.77 ± 0.10 B
XTE J1650-500
;
Fig. 3.4 – X-ray HID for XTE J1650-500 2001 outburst from RXTE data, adopted from Corbel
et al. (2004). The two epochs analyzed in spectra in this work are highlighted by red circles and labeled in A/B. Both epochs are intermediate states, together with X-ray data in between, capturing the spectral changes during the transition. The eight epochs with only simultaneous radio/X-ray data are highlighted with numbers 1-8, but only Epochs 2 & 3 are analyzed due to the availability of IR observations, and due to the model’s inapplicability in soft states.
