1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\documentclass{emulateapj}
4: %\usepackage{apjfonts}
5: \def\({\left(}
6: \def\){\right)}
7: \def\[{\left[}
8: \def\]{\right]}
9: \def\msun{{\rm ~M}_{\odot}}
10: \def\myr{{\rm ~Myr}}
11: \def\mdot{\dot M}
12: \def\mpy{{\rm ~M}_{\odot} {\rm ~yr}^{-1}}
13: \def\medd{{\rm M_{Edd}}}
14: \def\ledd{{L_{\rm Edd}}}
15: \usepackage{graphicx}
16: \usepackage{natbib}
17: \usepackage[usenames,dvips]{color}
18: \newcommand{\ron}[1]{{\color{red}\bf Ron: #1}}
19: \begin{document}
20: \title{ON THE PROPERTIES OF INNER COOL DISKS IN THE HARD STATE OF BLACK HOLE
21: X-RAY TRANSIENT SYSTEMS}
22:
23: \author{Ronald E. Taam}
24: \affil{Northwestern University, Department of Physics and Astronomy,
25: 2131 Tech Drive, Evanston, IL 60208; ASIAA/TIARA, P.O. Box 23-141, Taipei,
26: 10617 Taiwan; ASIAA/National Tsing Hua University - TIARA, Hsinchu, Taiwan}
27: \email{r-taam@northwestern.edu}
28:
29: \author{B. F. Liu}
30: \affil{National Astronomical Observatories/Yunnan Observatory, Chinese
31: Academy of Sciences, P.O. Box 110, Kunming 650011, China}
32: \email{bfliu@ynao.ac.cn}
33:
34:
35: \author{F. Meyer}
36: \affil{Max-Planck-Institut f\"ur Astrophysik, Karl Schwarzschildstr. 1,
37: D-85740, Garching, Germany}
38: \email{frm@mpa-garching.mpg.de}
39:
40: \and
41:
42: \author{E. Meyer-Hofmeister}
43: \affil{Max-Planck-Institut f\"ur Astrophysik, Karl Schwarzschildstr. 1,
44: D-85740, Garching, Germany}
45: \email{emm@mpa-garching.mpg.de}
46: \begin{abstract}
47: The formation of a cool disk in the innermost regions of black hole X-ray transient systems in
48: the low hard state is investigated. Taking into account the combined cooling associated with
49: the Compton and conductive energy transport processes in a corona, the radial structure of a
50: disk is described for a range of mass accretion rates. The mass flow in an optically thick
51: inner region can be maintained by the condensation of matter from a corona with the disk
52: temperature and luminosity varying continuously as a function of the accretion rate. Although such a
53: disk component can be present, the contribution of the optically thick disk component to the
54: total luminosity can be small since the mass flow due to condensation in the optically thick
55: disk underlying the corona can be significantly less than the mass flow rate in the corona. The
56: model is applied to the observations of the low quiescent state of the black hole source GX 339-4
57: at luminosities of $\sim 0.01 L_{Edd}$ and is able to explain the temperature of the thermal component
58: at the observed luminosities. Since conductive cooling dominates Compton cooling at low mass
59: accretion rates, the luminosity corresponding to the critical mass accretion rate above which a weak
60: thermal disk component can be present in the low hard state is estimated to be as low as $0.001 L_{Edd}$.
61: \end{abstract}
62:
63: \keywords{accretion, accretion disks --- black hole physics ---
64: X-rays: stars --- X-rays: binaries --- stars:individual (GX 339-4)}
65:
66: \section{Introduction}
67:
68: The class of objects known as transient X-ray binary systems containing black holes is one of
69: the few examples of astrophysical laboratories available for investigating the nature of accretion
70: flows and the structure of accretion disks surrounding compact objects. Their luminosities provide
71: us with an opportunity to investigate a given system over a wide range of mass
72: accretion rates onto the central black hole. Indeed, these systems have been observed in distinct
73: states generally categorized as low/hard and high/soft states (see McClintock \& Remillard 2006). In this
74: categorization, the high/soft state is characterized by a strong thermal component
75: corresponding to a color temperature of order 1 keV. These states are observed during the outburst
76: phase in which the luminosity can approach the Eddington luminosity, $L_{\rm Edd} \sim 10^{39}$ ergs
77: s$^{-1}$ for a $10 \msun$ black hole. On the other hand, the low/hard state is typically
78: described by a power law with a photon power law index $\sim 1.7$. This state is observed during the
79: quiescent stages, where the luminosities can be as low as $10^{-7} L_{\rm Edd}$ (Corbel et al.
80: 2006). Hence, it is generally accepted that these states are distinguished by
81: the spectrum and a range of luminosity with a transition at $\sim1-4\%$ of
82: the Eddington luminosity (Maccarone 2003) from the low/hard state to the high/soft state.
83:
84: Given these observational diagnostic clues, progress in our theoretical understanding has advanced in
85: parallel. Specifically, the spectrally soft state is believed to represent the emission from a cool optically
86: thick and geometrically thin accretion disk extending to the innermost stable circular orbit ($3 R_S$
87: for a non-rotating black hole, where $R_S = 2GM/c^2$ is the Schwarzshild radius) as described in the
88: seminal study by Shakura \& Sunyaev (1973, hereafter SS). In contrast, the spectrally hard state is
89: thought to arise from a geometrically thick, hot optically thin coronal or advectively dominated
90: accretion flow (ADAF; Narayan \& Yi 1994; 1995a, b) located in the innermost regions of the disk near
91: the black hole. Here, the SS accretion disk is truncated at a distance of several hundred Schwarzschild
92: radii from the black hole, leading to a transition from a disk dominated by a geometrically thin outer
93: structure to a geometrically thick inner structure Esin et al. (1997, 1998, 2001). Although
94: a complete description of the truncation
95: of the optically thick disk has yet to be worked out in detail, the mechanisms suggested to facilitate
96: this transition have been investigated in the studies by Honma (1996) and Manmoto \& Kato (2000)
97: based on a radial conductive energy transport process and by Meyer et al. (2000), Spruit \& Deufel
98: (2002), and Dullemond \& Spruit (2005) based on a vertical evaporative process.
99:
100: Recently, additional observational clues concerning the nature of the accretion flow and geometry
101: in the low quiescent state have been found pointing to the possible existence of an inner cool
102: optically thick disk structure in black hole transient X-ray binary systems. Although the
103: observational signatures are difficult to extract because of the low X-ray count rates, the evidence
104: for such a structure has been inferred from the source GX 339-4 (see Miller et al. 2006a, b; Tomsick
105: et al. 2008) where a weak, soft, thermal emission component ($\lesssim 0.3$ keV) has been detected
106: and blurred reflection and iron features are indicated in fits to the X-ray spectrum for luminosities
107: as low as $\sim 0.01 L_{\rm Edd}$. One of the goals of current observational studies has been to
108: extend such studies to other sources (e.g., J 1817-330, see Rykoff et al. 2007; and J1753.5-0127,
109: see Miller et al. 2006a; Ramadevi \& Seetha 2007; Soleri et al. 2008) and to determine the
110: luminosity range for which such a thermal component exists so that a full description of the
111: accretion flow geometry in the various states of transient X-ray binary systems can be constructed.
112:
113: Given the existence of this weak thermal component, the accretion disk structure in the quiescent state
114: would be described by three distinct regions. Specifically, a small inner cool, optically thick disk,
115: a geometrically thick hot coronal region at intermediate radial scales, and a very extensive geometrically
116: thin disk nearly filling the Roche lobe of the black hole would contribute to the weak thermal X-ray
117: component, the power law X-ray component, and the infrared and optical emission respectively. Such a
118: configuration was already contemplated in early work by R\`o\.za\`nska \& Czerny (2000) and more
119: recently by Liu et al. (2006), Meyer et al. (2007) and Mayer \& Pringle (2007). To maintain such a remnant
120: cool disk in the innermost regions (in the presence of mass transport by turbulent viscous stresses)
121: gas must condense from an overlying corona.
122:
123: In this study we focus on the general description of the structure and properties of the disk in its
124: innermost regions during the quiescent state of a black hole X-ray transient binary system, taking into
125: account the effects of both thermal bremsstrahlung and comptonization. This study differs from our recent
126: study (Liu et al. 2007) where only one such process was considered dominant for any given structure. In
127: \S 2, we describe the theoretical model and the modifications required in considering
128: more than one cooling process in any given region of the disk. The solutions to the modified set of
129: equations are presented for a range of mass accretion rates and viscosity parameters in \S 3.
130: The numerical results are compared to the observations of GX 339-4 in \S 4 and their implications
131: are discussed in the last section.
132:
133: \section{Theoretical Model}
134:
135: Early investigations (Meyer \& Meyer-Hofmeister 1994, Meyer, Liu, \& Meyer-Hofmeister
136: 2000; Liu et al. 2002) have shown that a hot corona overlying a cool disk leads to the
137: evaporation of cool gas into a corona, through which mass is continuously supplied for
138: accretion onto a central compact object. The evaporation is driven by vertical thermal
139: conduction, and the accretion of evaporated mass through the corona is a result of
140: viscous transport. The properties of the evaporation fed corona are a function of the
141: distance from the accreting compact object. In the case of a black hole transient
142: X-ray binary system in the quiescent state, the mass flow rate in the corona reaches a
143: maximum at a distance of $\sim 300 R_S$ by efficient evaporation
144: and decreases inward as a result of condensation of coronal gas to a cooler disk.
145:
146: When the mass supply rate is sufficiently low, the optically thick outer cool disk is
147: truncated in the region where the integral of the evaporation rate over the radial extent
148: is greater than the mass inflow rate, leading to the formation of a gap between outer and
149: inner disk, filled by a coronal flow/ADAF. The radial extent of the coronal
150: gap region is dependent on the mass supply rate to the outer disk with a larger coronal gap
151: region corresponding to a lower mass accretion rate. For black hole X-ray binary systems in
152: the quiescent state, the accretion rate can be very low, and the outer disk is sufficiently distant
153: from the central black hole that it can be neglected in describing the innermost disk regions.
154: In this latter region, the dominant accretion flow lies within the corona/ADAF, accompanied
155: by a weak mass inflow within an underlying cool disk. The condensation of coronal gas to the
156: disk follows from energy (and pressure) balance between this inner cool disk and its overlying
157: corona, providing the only source of mass in the inner optically thick disk. As shown by Liu
158: et al. (2007) such a configuration can be maintained for a range of accretion rates and, hence,
159: luminosities.
160:
161: In this work, we generalize the description of the inner disk and corona configuration by
162: considering Compton scattering and conduction, contributing simultaneously, to the
163: cooling of the disk. The treatment of both cooling processes leads to a modification of
164: the temperature of the electrons in the corona, and, hence, the condensation rate. In addition,
165: we also include contributions from both inverse Compton emission from the corona and
166: bremsstrahlung emission from the radiating/transition layer to the total luminosities. This
167: treatment of the cooling processes is in contrast to our previous study (Liu et al. 2007) where
168: always only one of the dominant emission mechanism was taken into account (but similar in our
169: neglect of non-thermal cooling processes). As a consequence of the new treatment, the
170: physical characteristics of the inner regions (e.g., the effective temperature of the inner
171: cool disk) vary continuously with variations in the luminosity. To facilitate comparisons
172: with Liu et al. (2007), the effect of X-ray irradiation of the hot coronal gas on the cool disk
173: is neglected; however, an approximate estimate of its effect is given and its implications discussed
174: in \S5.
175:
176: \subsection{Condensation caused by either conductive cooling or
177: Compton cooling}
178:
179: The energy balance in the transition layer between the disk and corona depends
180: on the ADAF lying above the disk (Narayan \& Yi 1995b). As derived in our
181: earlier work (Meyer et al. 2007; Liu et al. 2007)
182: the condensation rate (mass flow rate in vertical direction)
183: per unit area is determined as
184:
185: \begin{equation}\label{cnd-general}
186: \dot m_z={\gamma-1\over \gamma} \beta{-F_{\rm c}^{\rm ADAF}\over {\Re T_{i}\over
187: \mu_i}}\(\sqrt{C}-1\),
188: \end{equation}
189: where
190: \begin{eqnarray}\label{C}
191: C \equiv\kappa{_0} b
192: \left(\frac{0.25\beta^2 p_0^2}{k^2}\right)
193: \left(\frac{T_{\rm {cpl}}}{F_c^{\rm{ADAF}}}\right)^2 \nonumber
194: \end{eqnarray}
195: with $F_{\rm c}^{\rm ADAF}$ the conductive flux from the upper ADAF
196: arriving
197: at the interface of the transition layer, $T_{\rm cpl}$ the coupling
198: temperature and
199: $p_0=(2/\sqrt \pi)p$ the pressure in the transition layer,
200: \begin{eqnarray}\label{scaled}
201: &F_{\rm c}^{\rm ADAF}= -(\kappa_0 K n_i n_e T_i)^{1/2}T_{\rm em},\nonumber\\
202: &T_{\rm {cpl}}=1.98\times 10^9 \alpha^{-4/3}\dot m^{2/3}\rm K.
203: \end{eqnarray}
204: Pressure, ion temperature and electron number density, $p$, $T_i$ and $n_e$
205: are quantities of the ADAF,
206: \begin{eqnarray}\label{scaledadd}
207: &T_i+1.077 T_e =1.98\times10^{12} r^{-1}\rm K,\\\nonumber
208: &p=1.87\times 10^{16}\alpha^{-1}m^{-1}\dot m r^{-5/2}
209: \, \rm{g cm^{-1} s^{-2}},\\
210: &n_e=5.91\times10^{19}\alpha^{-1}m^{-1}\dot m r^{-3/2} \, \rm{cm^{-3}} \nonumber .
211: \end{eqnarray}
212: The meaning of the further quantities in the equations is:
213: $\beta$ is the ratio of gas
214: pressure to total pressure, $\gamma$ the ratio of specific heats, $\gamma=(8-3\beta)/(6-3\beta)$,
215: $\mu_i$ the molecular weight of ions
216: (1.23 for the assumed standard chemical abundance) and $k$ is the
217: Boltzmann constant.
218: $\kappa{_0} =10^{-6}{\rm erg\,s^{-1}cm^{-1}K^{-7/2}}$ is the
219: thermal conductivity
220: coefficient in the relation for the conductive flux,
221: $F_{\rm c}=-\kappa_0 T_e^{5/2}dT_e/dz$ (Spitzer 1962). The coefficient
222: $b=10^{-26.56}\rm g\,cm^5\,s^{-3}\,K^{-1/2}$ is used for the description
223: of the free-free radiation (Sutherland \& Dopita 1993). The energy transfer
224: rate from ions to electrons is taken as $q_{\rm ie}=Kn_in_eT_iT_e^{-3/2}$
225: with $K= 1.64\times 10^{-17} \rm {g cm^{5}s^{-3}deg^{1/2}}$ (Stepney
226: 1983) and $n_i=n_e/1.077$ is the ion number density. Following
227: convention, we take $\alpha$ as the viscosity coefficient. The quantities $m$,
228: $\dot m$, and $r$ are expressed in units of a solar mass, the Eddington rate,
229: $\dot M_{\rm Edd}=L_{\rm Edd}/0.1c^2$, and the Schwarzschild
230: radius, $R_S$, respectively. $T_{\rm em}$ is the maximal electron temperature in
231: the corona, to be determined for either conduction or Compton cooling,
232: as derived in the earlier work (Liu et al. 2007).
233:
234: Once $T_{\rm em}$ is known the condensation rate
235: can be determined by Eqs.(\ref{cnd-general})--(\ref{scaledadd}) at any
236: radius for a given black hole mass, mass accretion rate and
237: viscosity.
238:
239: In the case of inefficient Compton cooling in the corona,
240: vertical conduction is the dominant cooling mechanism.
241: Here, by collisional heat input from the ions (${dF_c/dz}=q_{\rm ie}$),
242: a vertical temperature gradient forms which results in a maximal
243: temperature at the upper boundary of the ADAF and a conductive flux at the
244: lower boundary,
245:
246: \begin{eqnarray}\label{T_em}
247: &T_{\rm em}=2.01\times 10^{10}\alpha^{-2/5}\dot m^{2/5} r^{-2/5}\rm K,\\
248: &F_{\rm c}^{\rm ADAF}=-6.52\times 10^{24} \alpha^{-7/5}m^{-1}\dot m^{7/5} r^{-12/5}\nonumber.
249: \end{eqnarray}
250: On the other hand, for a sufficiently high radiation flux from the underlying disk, the disk
251: radiation propagating through the corona can lead to efficient Compton cooling of the corona. The
252: electrons are cooled by Compton scattering to a temperature (determined by $q_{\rm ie}=
253: q_{\rm cmp}$) lower than the value given by conductive cooling (Eq.(\ref{T_em})). As a result,
254: the heat flux at the transition layer decreases and the condensation rate increases, as
255: determined by the energy balance in the transition layer between cooling by bremsstrahlung
256: emission and heating by the conductive flux from the upper ADAF and by the enthalpy flux of the
257: condensing gas. The maximal temperature in the corona and the corresponding conductive flux to
258: the transition/radiating layer are expressed as
259: \begin{eqnarray}\label{T_ec}
260: & T_{\rm em}=3.025\times 10^9\alpha^{-2/5} m^{-2/5}\dot m^{2/5}r^{1/5}\[1-\({3\over r}\)^{1/2}\]^{-2/5}\({T_{\rm eff,max}\over 0.3keV}\)^{-8/5}\rm K,\\
261: & F_c^{\rm ADAF}=-9.816\times 10^{23}\alpha^{-7/5} m^{-7/5}\dot m^{7/5}r^{-9/5}
262: \[1-\({3\over r}\)^{1/2}\]^{-2/5}
263: \({T_{\rm eff,max}\over 0.3keV}\)^{-8/5}
264: \rm erg\,s^{-1}\,cm^{-2}.\nonumber
265: \end{eqnarray}
266: The effective temperature of the disk, $T_{\rm eff}$, corresponds to the
267: mass flow in the disk accumulated by the condensation of matter from the corona.
268: $T_{\rm eff}$ reaches its maximum value $ T_{\rm eff, max}=1.335\times
269: 10^7m^{-1/4}\dot m_{\rm cnd}^{1/4}$ at a distance
270: $r_{\rm tmax}=3\,\frac{49}{36} R_s$ (for details see Liu et al. 2007).
271:
272: \subsection{Condensation caused by the two processes: conductive cooling and
273: Compton cooling operating simultaneously}
274:
275: The above describes the formulation of our model as outlined in a previous study (Liu et al.
276: 2007) when one of the cooling processes (either thermal conduction or Compton scattering)
277: is dominant. However, in general, both processes can contribute to
278: the cooling simultaneously. In particular,
279: for a certain range of mass accretion rates, the rate of cooling associated with the
280: conduction and Compton processes are comparable. In this case, neither the Compton dominated
281: model nor the conduction dominated model provides an adequate description for the disk/corona
282: configuration. In this case, both mechanisms must be treated together. However,
283: construction of the detailed vertical structure for the condensation region cannot be expressed
284: in analytical form. In the spirit of the analysis of Liu et
285: al. (2007), we modify the formulae describing the
286: conduction dominant and Compton dominant cases in such a way that we
287: obtain a smooth transition when both mechanisms are of comparable
288: importance. Our approach is described below.
289:
290: For given values of $\alpha$, $m$, and $\dot m$ at a given distance, $r$, the importance of
291: Compton or conduction cooling is determined by comparing the electron temperature given by
292: considering the operation of Compton cooling alone (i.e. $q_{\rm ie}=q_{\rm cmp}$) and
293: conductive cooling alone (i.e. $q_{\rm ie}=dF_c/dz$). The process that results in a lower
294: temperature is the more efficient cooling mechanism and, thus, is taken as the dominant one.
295: To determine the strength of the neglected cooling process compared to that of the dominant process,
296: we define the ratio, $\lambda$, between the Compton cooling rate, $q_{\rm cmp}$, and the
297: conduction cooling rate, $dF_c/dz$,
298: \begin{equation}\label{lambda-d}
299: \lambda\equiv {q_{\rm cmp}\over dF_c/dz}.
300: \end{equation}
301: The Compton cooling rate is given by
302: \begin{equation}
303: q_{\rm cmp} = {4kT_e\over m_ec^2} n_e \sigma_T c {aT_{\rm eff}^4(r)\over 2}
304: =f_1(\alpha, m,\dot m) \dot m_{\rm cnd} T_e r^{-9/2}\[1-\(\frac{3}{r}\)^{1/2}\],
305: \end{equation}
306: where $a$ is radiation density constant, $\sigma=\frac{ac}{4}$, $\sigma_T$ Thomson cross section,
307: $\sigma T_{\rm eff}^4(r)=(3GM \dot M_{\rm cnd}/8\pi R^3) \[1-(3R_S/R)^{1/2}\]$, and $f_1$ is some
308: function only depending on the ADAF parameters.
309:
310: $dF_c/dz$ can
311: be approximated by taking the value of the flux at the upper boundary ($F_c=0$) and at the
312: lower boundary, $F_c^{\rm ADAF}=-(k_0Kn_in_eT_i)^{1/2}T_{\rm em}$, to give
313: $$dF_c/dz\propto -F_c^{\rm ADAF}/H\propto -F_c^{\rm ADAF}r^{-1}.$$
314: Replacing $F_c^{\rm ADAF}$ by Eqs.(\ref{scaled}) and (\ref{scaledadd}) we have
315: \begin{equation}\label{Fc-apprx}
316: dF_c/dz= f_2\(\alpha, m,\dot m\)r^{-3} T_{\rm em}.
317: \end{equation}
318: Thus, from Eqs.(\ref{lambda-d})--(\ref{Fc-apprx}) and $T_e=T_{\rm em}$, $\lambda$ can be
319: approximately expressed as
320: \begin{equation}\label{lambda0}
321: \lambda=f\(\alpha, m,\dot m\)\dot m_{\rm cnd} r^{-3/2}\[1-\(\frac{3}{r}\)^{1/2}\].
322: \end{equation}
323: The function $f$ again only depends on the ADAF parameters.
324: The value of $\lambda$ is very small when conduction is the overwhelmingly dominant cooling
325: process, reaching unity when the Compton process is equally important, and is much
326: larger than unity for cooling dominated by Compton scattering. In the critical case when
327: the cooling process switches from conduction dominant to Compton dominant, $\lambda_c=1$,
328: the temperature inferred from the Compton cooling process and the conduction cooling process
329: is the same, which gives (by Eq.(\ref{T_em})=Eq.(\ref{T_ec})),
330: \begin{equation}
331: \dot m_{\rm cnd}r_{\rm c}^{-3/2}\[{1-\({3\over r_{\rm c}}\)^{1/2}}\]={1\over 2.464\times 10^4}.
332: \end{equation}
333: This determines $f\(\alpha, m,\dot m\)=2.464\times 10^4$ by the conditions defining the critical
334: case from Eq.(\ref{lambda0}). Thus, $\lambda$ can be written, in general, as
335: \begin{equation}\label{lambda}
336: \lambda=2.464\times 10^4\dot m_{\rm cnd}r^{-3/2}\[1-\(\frac{3}{r}\)^{1/2}\]
337: \end{equation}
338: and used in calculating the condensation rate. To obtain a solution that includes both effects,
339: thermal conduction and Compton cooling, we modify the energy equation, leading to a change in the
340: electron temperature and, hence, a change in the condensation
341: rate and luminosities from both the corona and the disk. In the case of conduction dominant
342: cooling, the energy equation is modified as
343: \begin{equation}
344: q_{\rm ie}=(1+\lambda)dF_c/dz.
345: \end{equation}
346: Since $q_{\rm ie}\propto T_e^{-3/2}$ and $dF_c/dz\approx -F_c^{\rm ADAF}r^{-1}\propto T_{\rm em}$,
347: the energy equation leads to a modified temperature
348: \begin{equation}
349: T_{\rm em}=2.01\times 10^{10}\alpha^{-2/5}\dot m^{2/5} r^{-2/5}(1+\lambda)^{-2/5}\rm K.
350: \end{equation}
351: Similarly, in the case of Compton dominant cooling, the energy equation should be modified as
352: \begin{equation}
353: q_{\rm ie}=(1+{1\over\lambda})q_{\rm cmp}.
354: \end{equation}
355: Note that $q_{\rm cmp}\propto T_e$, and the modified energy equation results in a
356: modification factor $(1+{1\over\lambda})^{-2/5}$ to the temperature,
357: \begin{equation}
358: T_{\rm em}=3.025\times 10^9\alpha^{-2/5} m^{-2/5}\dot m^{2/5}r^{1/5}\[1-\({3\over r}\)^{1/2}\]^{-2/5}(1+{1\over\lambda})^{-2/5} \({T_{\rm eff,max}\over 0.3keV}\)^{-8/5}\rm K.
359: \end{equation}
360: With the modified temperature, the heat flux and the resulting integrated condensation rate
361: in a region between $R_i$ and $R_o$ can be calculated for given $\alpha$, $m$ and $\dot m$,
362: \begin{equation}\nonumber
363: \dot m_{\rm cnd}=\int_{R_i}^{R_o} {4\pi R \over \dot M_{\rm Edd}}\dot m_zdR.
364: \end{equation}
365: The condensation rate for Compton dominant cooling is
366: \begin{equation}\label{condensation-cmp}
367: \dot m_{\rm cnd}=A\left\{2B\[\({r_{\rm o}\over r_i}\)^{1/2}-1\]-\int_{r_i/3}^{r_{\rm o}/3}(1+{1\over\lambda})^{-2/5}x^{1/5}\(1-x^{-1/2}\)^{-2/5}dx \right\},
368: \end{equation}
369: and for conduction dominant cooling
370: \begin{equation}\label{condensation_cnd}
371: \dot m_{\rm cnd}= 3.23\times 10^{-3}\alpha^{-7}\dot m^3 f(x)
372: \end{equation}
373: with
374: \begin{equation}\label{functionx}
375: f(x)=\({r_o\over r_1}\)^{3/5}\[{6\({r_1\over r_o}\)^{1/10}-6{\({r_1\over r_o}\)^{1/10} x^{1/2}-3 \int_{r_i/r_o}^1 \(1+\lambda\)^{-2/5} x^{-2/5}}dx}\],
376: \end{equation}
377: where $r_{1}=0.815\alpha^{-28/3}\dot m^{8/3}$ is the condensation radius determined by $C=1$ in the case of conduction cooling
378: before modification. The new factor $r_o\over r_1$ is a consequence of generalizing the
379: condensation formula to reflect the condensation rate at any condensation region. When
380: $\lambda=0$ and $r_o= r_1$, Eq.(\ref{functionx}) reverts to our old form. When $\lambda\rightarrow
381: \infty$, Eq.(\ref{condensation-cmp}) also reverts to the previous form for the case of Compton
382: cooling.
383:
384: Note that since $\lambda$ depends on the condensation rate, iteration is required until a
385: a self-consistent condensation rate is obtained. In practice, the modification of the
386: condensation rate is very small at low accretion rate since the conduction process dominates.
387: Similarly, the modification is also small if Compton cooling is very
388: strong compared to conduction. We note, however, that these conditions are not necessarily
389: satisfied for black hole X-ray binary systems in their low hard or intermediate states.
390: In this regime, the modification factors are important since conduction and Compton
391: scattering processes can be of comparable strength in cooling the corona.
392:
393: \subsection{The radial extent of Compton and conduction dominated regions}
394: In describing the thermal state of the corona above an inner disk, (i) cooling in the corona
395: could be conduction dominant throughout the corona, (ii) Compton dominant throughout, or (iii)
396: conduction dominant in the outer regions and Compton dominant in the inner region, depending
397: on the binary system parameters. Therefore, for given values of $\alpha$, $m$ and $\dot m$,
398: an estimate of the radial distance from the black hole for the Compton dominant region is
399: obtained by assuming that the Compton process cools the electrons to the same temperature as the
400: conduction process,
401: \begin{equation}\label{r_cmp}
402: r_{\rm cmp}\[{1-\({3\over r_{\rm cmp}}\)^{1/2}}\]^{-2/3}= 846.71\dot m_{\rm cnd}^{2/3},
403: \end{equation}
404: The above equation either admits two solutions or no solution for $r_{\rm cmp}$. In
405: case no solution exists, Compton cooling never dominates in the corona and the modified
406: conduction model is an adequate description for the inner disk and corona. The existence of the
407: two solutions $r_{\rm cmp1}$ and $r_{\rm cmp2}$ determines the Compton dominant region $r_{\rm
408: cmp1}<r<r_{\rm cmp2}$. If $r_{\rm cmp2}$ is larger than the condensation radius calculated
409: by the conduction model, which is modified as $r_{d}=0.815\alpha^{-28/3}\dot m^{8/3}
410: \[1+\lambda(r_d)\]^4$, the cooling is dominated by the Compton process at all
411: radii, and we use the modified Compton model for the numerical calculations. On the
412: other hand, if $r_{\rm cmp2}$ is smaller than $r_{d}$, the coronal region is divided into
413: a conduction and Compton dominant region. In Figure \ref{cmpregion} the radial extent of
414: Compton and conduction dominant regions is presented. The left panel delineates the regions
415: dominated by Compton or conduction cooling as a function of the total condensation rate.
416: At the lowest condensation rates, conduction is dominant throughout the disk. On the other hand, at higher
417: condensation rates,
418: the Compton dominant region always lies interior to the conduction dominant region. The right panel reveals the relative strength of the Compton cooling rate
419: as compared to conduction as a function of distance at given condensation rates. The modified
420: condensation rate is calculated employing the relevant models and values of $\lambda$ in each
421: region. Their contributions are summed to determine the modified total condensation rate of the system.
422:
423: The luminosity from bremsstrahlung radiation in the radiating/transition layer does not
424: vary with the electron temperature in the corona since the density and coupling temperature
425: are determined by $\alpha$, $m$, and $\dot m$, independent of the coronal temperature. Thus,
426: it has the same form for both the Compton and the conduction dominant regions
427: and is unaffected by the modification when both cooling mechanisms are included. We integrate
428: the bremsstrahlung emission throughout the disk-corona region yielding,
429: \begin{equation}\label{L_cnd}
430: {L_{\rm Brem}\over L_{\rm Edd}}=
431: 0.0642 \alpha^{-7/3}\dot m^{5/3}\[1-\({3\over r_d}\)^{1/2}\].
432: \end{equation}
433: Since the Compton luminosity from the corona depends on the electron temperature, its
434: determination must include the contributions from both cooling processes. In the Compton
435: cooling dominant region, a correction factor $(1+{1\over\lambda})^{-2/5}$ is added to the
436: integrand,
437: \begin{equation}\label{L_cmp1}
438: {L_{\rm cmp}\over L_{\rm Edd}}=0.392\alpha^{-7/5}m^{3/5}\dot m^{7/5}\({T_{\rm eff, max}\over 0.3keV}\)^{12/5}\int_{r_{\rm i}/3}^{r_{\rm o}/3}(1+{1\over\lambda})^{-2/5}x^{-23/10}\(1-x^{-1/2}\)^{3/5}dx.
439: \end{equation}
440: In the conduction dominant region, the luminosity from Compton scattering was neglected in
441: our previous work. To take this contribution into account, the electron temperature is replaced by
442: Eq.(\ref{T_em}) in the Compton cooling rate, leading to a Compton luminosity from
443: the conduction dominant region from $r_i$ to $r_o$,
444: \begin{equation}\label{L_cmp2}
445: {L_{\rm cmp}\over L_{\rm Edd}}=1.349\alpha^{-7/5}m\dot m^{7/5}\({T_{\rm eff, max}\over 0.3keV}\)^4\int_{r_{\rm i}/3}^{r_{\rm o}/3}(1+\lambda)^{-2/5}x^{-29/10}\(1-x^{-1/2}\)dx.
446: \end{equation}
447: The total corona luminosity is, thus, the sum of bremsstrahlung radiation
448: (Eq.(\ref{L_cnd})) in the radiation/transition layers and Compton scattering
449: in the ADAF above
450: (Eqs. (\ref{L_cmp1}) and (\ref{L_cmp2})) when both regions exist.
451:
452: The disk luminosity is determined under the assumption of emission by a local blackbody and
453: calculated as an integral over the disk as,
454: \begin{equation}\label{Ldisk}
455: {L_{d}\over L_{\rm Edd}}=\int_{3R_s}^{R_d}\sigma T_{\rm eff}^4(R){4\pi R\over L_{\rm Edd}} dR=\int_{3R_s}^{R_d}{3GM\dot M\[1-({3R_s\over R})^{1/2}\]\over 8 \pi R^3}{4\pi R\over L_{\rm Edd}} dR.
456: \end{equation}
457: Note that the mass flowing in the inner disk through any radius $R$ originates from the condensed gas
458: integrated from the critical condensation radius ($R_d$) to $R$. Hence, the mass accretion rate in the disk is
459: a function of radius, which increases with decreasing radii, representing the cumulative effect of
460: the condensation rate of gas onto the disk. The disk luminosity is then
461: calculated from all radii with the corresponding accretion rate profile.
462:
463: In our procedure, the condensation rate is first calculated assuming conduction as the only cooling
464: mechanism. If this condensation rate heats the disk to a temperature sufficiently high that
465: Compton cooling is more efficient than conduction in some regions, we adopt the Compton
466: cooling model in the Compton dominant region but retain the conduction model in the conduction
467: region. The condensation rate is recalculated from these different regions as iterative
468: calculations are required for the Compton dominant region.
469: These results are modified taking into account the neglected cooling by calculating the
470: value of $\lambda$ from the inferred condensation rate. As the modified condensation rate
471: results in a new value of $\lambda$, the process is repeated until the presumed condensation
472: rate approaches the derived condensation rate. Finally, the contributions to the total luminosity
473: from Compton scattering and bremsstrahlung throughout the coronal region and multi-colour blackbody radiation from the inner disk are
474: calculated with the derived condensation rate and correction factor $\lambda$ by equations
475: (\ref{L_cnd}), (\ref{L_cmp1}), (\ref{L_cmp2}) and (\ref{Ldisk}).
476:
477:
478: \section{Numerical Results}
479:
480: To investigate the properties of the inner optically thick disk in X-ray transient binary systems,
481: we adopt a black hole mass of $10M_\odot$ and the low mass accretion rate range corresponding to
482: the transition to an ADAF. In the following, we describe the numerical results for the inner disk temperature,
483: size, and luminosity from the inner disk as determined from the condensation rates from the corona
484: as a function of the total mass accretion rate. To determine their sensitivity to the viscosity in
485: the disk, we also examine their dependence on the viscosity parameter.
486:
487: The luminosity from the disk and corona is illustrated as a function of accretion rate in Figure
488: \ref{f:mdot-L}.
489: It can be seen that the Eddington scaled luminosity does not scale linearly with the mass accretion
490: rate, indicating that the conversion of gravitational energy to radiative energy by an ADAF+disk in
491: low states is not constant nor as efficient ($\eta=0.10$) as in a SS accretion disk. The deviation
492: of the $L/L_{\rm Edd}-\dot M/\dot M_{\rm Edd}$ curve from a linear relation indicates that the energy
493: conversion efficiency decreases with lower mass accretion rates, qualitatively consistent with an
494: ADAF model.
495:
496: The size of the inner disk is taken to be limited by the critical radius where neither evaporation
497: nor condensation occurs. Externally, gas from the disk evaporates into the corona, while interior
498: to this radius the coronal gas condenses onto the disk as a consequence of the disk corona interaction.
499: For greater mass flow rates in the corona, more gas condenses onto the disk and the
500: critical radius of the inner disk increases. To illustrate this relationship, the size is
501: calculated for a range of mass accretion rates. The radial extent of the inner disk is
502: modified by the inclusion of the two cooling processes since the additional cooling results in a lower
503: heat flux down to the transition layer. However, the over-cooling by bremsstrahlung radiation is
504: partially compensated by the heating associated with the enthalpy flux from the condensation.
505: As a consequence, the size of the inner disk is larger than the case for which one cooling
506: process was considered. In the left panel of Figure \ref{f:r_d-L}, the inner disk
507: size is shown as a function of the luminosity. It can be seen that the inner disk extends
508: to greater distances with increasing luminosities or mass accretion rates. In the right
509: panel of Figure \ref{f:r_d-L}, the critical radii of the optically thick disk components are illustrated
510: as a function of the mass accretion rate. For a given accretion rate, the smaller radius corresponds
511: to the outer edge of the inner disk and the larger radius corresponds to the inner edge of the outer
512: disk, with their difference representing the spatial extent of the coronal gap region.
513: Hence, the radial extent of the coronal gap region between the inner and
514: outer optically thick disk shrinks as the luminosity increases, in agreement with the presence of
515: a radially continuous cold disk for sufficiently high luminosities. In this case, condensation
516: is no longer the only source for disk accretion, and the contribution of the disk to
517: the total luminosity is much greater than for a disk fed by condensation alone.
518:
519: The size of the inner disk is a function of the viscosity parameter as also seen in Figure
520: \ref{f:r_d-L} where, at a given luminosity, the disk is smaller for larger viscosity parameters.
521: This trend results from the (vertical) pressure and energy balance between the disk and overlying
522: corona. Specifically, for a given accretion rate, the surface density in the corona decreases as
523: $\alpha$ is increased. With a lower density the heat flux to the transition layer decreases, but
524: the bremsstrahlung radiation rate is reduced even more. The energy balance between cooling and heating
525: results in a reduced heating rate associated with a lower enthalpy flux. The net effect results in a
526: decreased condensation rate and, hence, a smaller disk size. We note that larger viscosity
527: parameters, yielding smaller disk radii, would be preferable to be consistent with our neglect
528: of X-ray irradiation effects in our present level of approximation.
529:
530: The total condensation rate, which is an integral of the rate of locally condensing mass from the critical
531: condensation radius to $r_{\rm tmax}=3{49\over 36}$, relative to the mass accretion rate is displayed as a function of
532: the mass accretion rate in Figure \ref{f:ratiomdot-mdot}. It is evident that a large fraction
533: (up to 20\%) of the coronal flow condenses onto an inner disk at high accretion rates. As a consequence,
534: the rate of mass flow in the corona decreases with decreasing distance. On the other hand, all the gas
535: accretes through the corona without condensation if the mass flow rate in the corona is very low.
536: The relative mass flow rate in the inner disk depends sensitively on the viscosity value.
537:
538: The maximal disk temperature as calculated from the rate of condensation integrated over the
539: entire inner disk is illustrated as a function of luminosity in Figure \ref{f:Teff-L}. It is
540: evident that the temperature increases as the total luminosity increases. Moreover, the luminosity
541: from the disk component increases at higher accretion rates and luminosities as shown in Figure
542: \ref{f:ratioL-mdot} and Figure \ref{f:ratioL-L}. This feature, together with the increasing
543: disk size, implies that the corona/ADAF dominant flow in the low state naturally evolves to a disk
544: dominant flow at high states consistent with ``standard'' picture developed for the different spectral
545: states of X-ray binaries, although the exact transition is likely to be affected by the inclusion
546: of X-ray irradiation of the hot coronal gas on the cool disk.
547:
548: The increasing condensation rate with accretion rate or luminosity is a direct result of the
549: efficient cooling at high accretion rates. Specifically for higher accretion rates, the density
550: correspondingly increases ($n \propto \dot m$), which greatly enhances the bremsstrahlung
551: cooling rate ($\propto n^2$). In the transition layer, the conductive heating rate approximately
552: increases with the viscous heating rate ($\propto n$). To maintain energy balance, additional
553: heating, i.e., the condensation enthalpy rate, must increase with density more rapidly than
554: $\propto n^2$. As the condensation rate affects the luminosity of the inner disk, while the total
555: luminosity consists of contributions from bremsstrahlung, Compton and disk radiation, the disk
556: luminosity increases more rapidly than the total luminosity. Thus, their ratio increases with the
557: luminosity.
558:
559: The numerical results reveal that condensation occurs only for accretion rates exceeding a critical
560: value, which is found to depend on the viscosity parameter. In Figures 3-7, the effect of
561: condensation was plotted from the lowest accretion rates found in the numerical calculations.
562: In fact, the corresponding luminosity at the critical accretion rate is from a pure ADAF rather than
563: from the small disk region as shown in the figures. Therefore, the lower limit to the luminosity for
564: which an inner disk could exist may well be as low as 0.001 $L_{\rm Edd}$, depending on the radiative
565: efficiency of the ADAF.
566:
567: \section{Comparison to Observations}
568:
569: The geometry of the accretion flow in the low hard states of X-ray transient systems has yet
570: to be clarified since observational indications exist that cool material can be present in the
571: inner regions where the reflection of X-rays and the formation of broad iron lines can
572: contribute to the spectrum. For the bright portion of hard-state systems near the transition
573: from a spectrally hard to soft state, continuous emission and reflection components from a
574: geometrically thin, optically thick inner disk are theoretically expected. Although the observational
575: signatures for such a disk component
576: are difficult to detect, evidence for such a feature has been inferred from the source GX 339-4
577: during its 2004 outburst (Miller et al. 2006a, b) by the presence of a weak, soft, thermal component
578: and by the suggestion of blurred reflection and iron lines. SWIFT J1753.5-0127 also reveals such a
579: weak disk component (Miller et al. 2006a; Ramadevi \& Seetha 2007; Soleri et al. 2008), however
580: iron features were not detected (Miller et al. 2006a). For systems at very low luminosities, the
581: general theoretical consensus favors a disk described by a pure ADAF solution in the inner region.
582: In support of this picture, high quality spectra for several systems in quiescence at luminosities
583: $L/L_{\rm Edd} \lesssim 10^{-6}$ have not revealed any evidence for a thermal component or iron
584: features (e.g. Bradley et al. 2007; Corbel et al. 2006).
585:
586: Recently, Tomsick et al. (2008) have detected a weak thermal component from the source GX 339-4
587: at low luminosities ($L \lesssim 0.03 L_{\rm Edd}$) based on observations they had obtained using SWIFT and
588: RXTE. In addition, broad features due to iron K transitions that are likely produced by reflection
589: of hard X-rays from the optically thick material near the ISCO have also been inferred from the
590: spectrum. Here, we compare the results from our model predictions with these observational data. Of
591: particular relevance for comparison with our theoretical analysis are the two observations of GX 339-4
592: after the source underwent a transition to the hard state carried out by Tomsick et al. (2008).
593:
594: Specifically, Tomsick et al. (2008) carried out two observations of GX 339-4 after the source had
595: undergone a transition to the hard state. One spectrum was obtained a few days after the transition
596: (called spectrum \#1) and another spectrum was obtained 2-3 weeks later (called spectrum \#2). The
597: inferred luminosity of spectrum \#1 in the energy range from 1 to 100 keV was 0.023 times the Eddington
598: value based on simultaneous observations obtained with the XRT, PCA and HEXTE instruments. Spectral
599: fitting indicated the presence of a cool, optically thick disk located near the ISCO with a temperature
600: of 0.201 keV. A similar analysis of spectrum \#2 yielded a luminosity of 0.008 $L_{\rm Edd}$ and a temperature
601: of the inner disk of 0.165 keV. Here, for convenient comparison we adopt a source distance of 8 kpc and
602: a black hole mass of $5.8 \msun$, which were used by Tomsick et al. (2008) and Miller et al. (2006b).
603:
604: In order to fit the observed luminosity and disk temperature, the accretion rate $\dot m$ and viscosity
605: parameter $\alpha$ were varied until agreement between the theoretical
606: modelling and the data deduced from observations was achieved. In general, the disk can be described
607: in one of four different configurations depending upon the chosen values of the observables. In particular,
608: for systems characterized by very low luminosities and disk temperatures (e.g. with coronal luminosities,
609: $L_{\rm c}/L_{\rm Edd}=0.001$ and $T_{\rm eff}=0.04$ keV) in the quiescent state, condensation is
610: unimportant and the disk is described by an ADAF without the presence of an optically thick inner disk.
611: At the other extreme, for systems at very high luminosities, say, $L_{\rm c}/L_{\rm Edd}\sim 0.1$, the
612: mass accretion rate exceeds the critical rate above which an ADAF does not exist. The third configuration
613: corresponds to the case where the spatial extent of the inner optically thick disk connects to the outer
614: optically thick disk indicating that the disk is geometrically thin throughout and is no longer
615: condensation-fed (a limiting case of configuration 4, see below). This occurs as a consequence
616: of a relatively high mass accretion rate and low viscosity for systems with intermediate luminosities
617: and relatively high temperatures. For instance, taking
618: $L_{\rm c}/L_{\rm Edd}=0.03$ and $T_{\rm eff}=0.26$ keV, it is found that the accretion rate and viscous
619: parameter are 0.035 and 0.2 respectively. However, the derived size of the disk is $417R_{\rm S}$,
620: implying that the coronal gap region is absent. Only for values of $\dot m$ and $\alpha$ in the appropriate
621: range where the derived size of the inner disk is not very large (the fourth configuration) is the fit
622: regarded as consistent with
623: observations.
624:
625: Our fitting procedure leads to the resulting parameters for spectrum \#1 and \#2 as listed
626: in Table 1. Given $\alpha=0.31$ and $\dot m=0.0687$, we derive a coronal luminosity of 0.023 $\ledd$, a
627: disk temperature of 0.203 keV, and an inner disk size of 71 $R_S$. Similarly, given $\alpha=0.23$ and
628: $\dot m=0.0299$, we derive a coronal luminosity of 0.008 $\ledd$, a disk temperature of 0.164 keV, and an
629: inner disk size of 82 $R_S$. These results are in good agreement with the inferred disk parameters
630: based on the two observations of Tomsick et al. (2008). With these two fits the relative strength of the
631: cool disk to the hot corona can be estimated. As listed in Table 1, the ratio of the luminosities
632: associated with the thermal component, $L_{\rm d}$, to the total luminosity, $L$, is $\sim 10\%$,
633: indicating a weak disk in the two observations. In
634: addition, the radii delineating the regions where Compton cooling or conduction is important are listed in
635: Table 1. For the quantitative comparison of
636: cooling via Compton scattering and thermal conduction the ratio $\lambda$
637: is shown in Figure \ref{f:lambda-r} as a function of radius for these two cases. As the radial
638: distribution of $\lambda$ only depends on the condensation rate (see Eq.\ref{lambda}), i.e., the
639: disk temperature listed in Table 1, the values of $\lambda$ in the two fits differ by a factor of
640: $\sim 2.3$ at any given distance. Since it seems plausible to have
641: the same value of $\alpha$ for both cases
642: we have also explored solutions for the same value of $\alpha$ (see Table 1). Here, the observed
643: luminosities are fit, but we permit small differences between the predicted and observed temperatures. It
644: can be seen that the fit can still be taken as consistent with observations. The distinction between the
645: fits for $\alpha = 0.25$ in the two spectra can be seen from the extent of the Compton dominated region
646: as shown in Table 1. Specifically, the cooling in the corona for the fit to spectrum \#2 is dominated by
647: conduction throughout the inner disk, whereas the fit to spectrum \#1 is dominated by Compton scattering
648: in the inner region. This is a consequence of the differing disk temperatures as listed in Table 1.
649: This also leads to the greater sensitivity of the luminosities for a small change in accretion rates in
650: the two fits. Since the disk temperature determines the energy density of the soft photon field for
651: Compton scattering ($u\propto T_{\rm eff}^4$), the Compton luminosity in the fit to spectrum \#1 is much
652: larger than in spectrum \#2, hence, becoming the dominant contribution to the coronal luminosity. Therefore,
653: the radiation losses are more efficient when Compton cooling becomes important.
654:
655: The fits to the two observations reveal that the temperature of the inner disk is positively correlated
656: with the luminosity. We note the inferred reflection covering factors of Tomsick et al. (2008) as well
657: as Miller et al. (2006b) are comparable (i.e., the covering factor is inferred as 0.22 from spectrum \#1
658: and 0.24 from spectrum \#2 and 0.22 from the 2004 observations (Miller et al. 2006b)), which may suggest a comparable spatial
659: extent of the inner disk. Such an insensitivity to the luminosity could be understood if the viscosity
660: parameter varied (see Table 1).
661:
662: \section{Discussion}
663:
664: The structure of the inner disk region of accretion disks have been investigated for black hole
665: X-ray transient systems in their low luminosity quiescent state within the context of an
666: $\alpha$ viscosity model. Particular attention has been
667: focused on the conditions under which a weak, soft thermal component would be present in this state.
668: By generalizing the study of Liu et al. (2007) to include the cooling associated with both conduction
669: and Compton processes together, we have examined the properties of an optically thick inner disk
670: fed by the condensation of gas from a corona for a range of mass accretion rates. It has
671: been shown that such a disk component can be present with its spatial extent, temperature, and
672: luminosity varying continuously and decreasing with a decreasing mass accretion rate.
673: Despite its existence and the higher conversion efficiency of gravitational energy into radiation in
674: the cool disk, the contribution of the disk component to the total luminosity can be very low due to
675: the low condensation rate. Throughout the quiescent state, the mass flow rate to the central black
676: hole from the corona always exceeds the mass flow rate in the inner cool disk. Hence, the disk
677: structure remains geometrically thick even in the innermost regions although a soft thermal component
678: is present in the spectrum.
679:
680: In our analysis, we have neglected the effect of X-ray irradiation to facilitate comparisons
681: with our earlier work. Although the qualitative features of our model will not change, the parameters
682: describing the quantitative comparisons will be modified. In particular, the inclusion of X-ray
683: irradiation can affect the fitting parameters $\alpha$ and $\dot m$ since the observed disk
684: temperature will not solely be determined by the mass flow in the inner disk. For example, for ADAF
685: emission from the innermost region at height $H=R_{\rm ISCO}=3R_s$, the flux irradiating the disk
686: is given by
687: \begin{equation}
688: F_{\rm ir}(r)={{1\over 2}L_c\over 4\pi R^2}{H\over R}=1.716\times 10^{26} {\rm erg\,s^{-1}\,cm^{-2}}
689: m^{-1}r^{-3}{L_c\over L_{\rm Edd}}.
690: \end{equation}
691: Adopting a mean photon energy of 50keV, about one tenth of the irradiation flux (the ratio of the
692: mean photon energy to electron rest mass energy) is absorbed by the disk through Compton scattering
693: collisions with electrons in the disk. Although the K-shell absorption of the metals is negligible
694: at 50 keV, it can become significant at lower energies since the absorption cross section is
695: proportional to $(m_ec^2/h\nu)^{7/2}$. For a photon distribution of $n_\nu\propto \nu^{-1.6}$, the
696: flux absorbed by K-shell absorption is found to be $\sim 30\%$ of the total irradiation flux. Thus,
697: about 40\% of the total irradiation flux is absorbed by the K-shell absorption and Compton scattering.
698: The re-radiation of the absorbed flux at the inner disk surface yields an effective temperature of
699: \begin{equation}
700: T_{\rm ir}(r)=\left({0.4F_{\rm ir}\over \sigma}\right)^{1/4}=2.86keVm^{-1/4}r^{-3/4}
701: \left({L_c\over L_{\rm Edd}}\right)^{1/4}.
702: \end{equation}
703: At the distance of maximal temperature, $r_{tmax}=3{49\over 36}$, the effective temperature caused
704: by irradiation alone for spectrum\#1 is 0.25keV and for Spectrum\#2 is 0.19keV (see \S4). These
705: temperatures are slightly higher than the temperatures generated by disk accretion and could contribute
706: to the observationally inferred temperature. We note that this theoretical estimate is very
707: approximate since an accurate calculation depends on the region of origin of the ADAF flux and
708: the size of the coronal gap region. From Eqs.\ref{L_cmp1} and \ref{L_cmp2} it can be seen that
709: the corona/ADAF radiation is dominant from a region outside the ISCO. In this case, the irradiation
710: flux incident on the innermost disk decreases with distance of the source, $F_{\rm ir}
711: \propto r^{-2}$. For instance, assuming the ADAF radiation is dominant from $5R_s$ rather than $3R_s$,
712: the irradiation flux absorbed by the disk around the ISCO is 7\% of the total flux. On the other hand,
713: many coronal photons escape without interception by the disk for a large coronal gap, thereby, reducing
714: the thermal effects associated with the irradiation. In the case for Spectrum \#1 and \#2
715: of GX339-4, the reflection covering factor inferred from observations is $\sim 20\%$ of the half
716: sky, indicating that, indeed, a very large fraction of the coronal photons escape, in contrast to
717: the case of Haardt \& Maraschi (1991). Hence, these estimates indicate that our results are
718: indicative and that additional freedom exists in determining $\dot m$ and $\alpha$ when irradiation
719: is taken into account.
720:
721: For sufficiently low mass accretion rates, the cool inner disk vanishes. An estimate of the mass
722: accretion rate for which this occurs depends on whether condensation occurs in the vicinity of the
723: ISCO. If condensation does not occur, accretion and/or evaporation can deplete the inner disk within
724: a viscous time. In the case of cooling dominated by conduction, the critical condensation radius is
725: given by $r_d=0.815\alpha^{-28/3}\dot m^{8/3}\ge 3$. Thus, the lower limit to the accretion rate is
726: \begin{equation}\label{llimit1}
727: \dot m = 0.006 \({\alpha\over 0.2}\)^{7/2}.
728: \end {equation}
729: For an accretion rate just above this lower limit, both the size of the inner disk and the
730: condensation rate are very small, leading to a very low disk temperature. In this case, the Compton
731: cooling is negligible, and the lower limit predicted by the conduction model is self-consistent.
732:
733: For disk temperatures exceeding 0.14 keV, on the other hand, Compton cooling dominates in a corona
734: around a $10 M_\odot$ black hole. If this temperature is produced by the accretion of condensed gas,
735: the condensation rate would be greater than 0.002 $\dot M_{\rm Edd}$. However, this occurs at high
736: accretion rates and is greater than the lower limit given by conduction model for $0.1 \le \alpha\le
737: 0.4$ (King et al. 2007). Thus, the lower limit to the mass accretion rate for maintenance of an inner
738: disk is set by the conduction model, i.e. equation (\ref{llimit1}). Notwithstanding the possible
739: importance of irradiation of the inner disk by the corona in providing some latitude in disk parameters,
740: $\alpha$ and $\dot m$ (see above), when fitting the observational data, a determination of the mass
741: flow rate below which the cool disk component vanishes could provide a constraint, within the
742: framework of the condensation model, on the $\alpha$ viscosity parameter during the quiescent state.
743:
744: In the future, global calculations of the full radial and vertical disk structure should be considered
745: taking into account the effect of X-ray irradiation of the hot coronal gas on the inner disk and
746: of the cooling associated with both the Compton and conductive processes in order to quantitatively
747: confirm the picture presented in this paper. With such models in hand, a detailed comparison of
748: the spectra from such disks with observations can be undertaken.
749:
750: \acknowledgments
751: Financial support for this work is provided by the Theoretical Institute
752: for Advanced Research in Astrophysics (TIARA) operated under
753: Academia Sinica and the National Science Council Excellence
754: Projects program in Taiwan administered through grant number NSC
755: 96-2752-M-007-007-PAE. In addition, B.F. Liu acknowledges the support by
756: the National Natural Science Foundation of China (grants 10533050 and 10773028).
757:
758: \begin{thebibliography}{}
759: \bibitem[]{711} Bradley, C.K., Hynes, R.I., Kong, A.K.H. et al. 2007, \apj, 667, 427
760: \bibitem[]{712} Corbel, S., Tomsick, J. A., \& Kaaret, P. 2006, \apj, 636, 971
761: \bibitem[]{713} Dullemond, C. P., \& Spruit, H. C. 2005, \aap, 434, 415
762: \bibitem[]{756} Esin, A.A., McClintock, J.E., \& Narayan, R. 1997, \apj, 489,865
763: \bibitem[]{757} Esin, A.A., Narayan, R., \& Cui, W. et al. 1998, \apj, 505, 854
764: \bibitem[]{758} Esin, A.A., McClintock, J.E., \& Drake, J.J. et al. 2001, \apj, 555, 483
765: \bibitem[]{} Haardt, F. \& Maraschi, L. 1991, \apj, 380, L51
766: \bibitem[]{714} Honma, F. 1996, \pasj, 48, 77
767: \bibitem[]{715} King, A. R., Pringle, J. E., \& Livio, M. 2007, \mnras, 376, 1740
768: \bibitem[]{716} Liu, B. F., Mineshige, S., \& Meyer, F., Meyer-Hofmeister,E., \&
769: Kawaguchi, T. 2002, \apj, 575, 117
770: \bibitem[]{718} Liu, B. F., Meyer, F., \& Meyer-Hofmeister, E. 2006, \aap, 454, L9
771: \bibitem[]{719} Liu, B. F., Taam, R. E., Meyer, F., \& Meyer-Hofmeister, E. 2007, \apj, 671, 695
772: \bibitem[]{765} Maccarone, T. J. 2003, \aap, 409, 697
773: \bibitem[]{720} Manmoto, T., \& Kato, S. 2000, \apj, 538, 295
774: \bibitem[]{721} Mayer, M. \& Pringle, J. E., 2007, \mnras, 376, 435
775: \bibitem[]{722} McClintock, J.E., \& Remillard, R.A., 2006, In: Compact Stellar X-ray Sources,
776: eds. W.H.G. Lewin, M. van der Klis, Cambridge Univ. Press
777: \bibitem[]{724} Meyer, F., Liu, B. F., \& Meyer-Hofmeister, E. 2000, \aap, 361, 175
778: \bibitem[]{725} Meyer, F., Liu, B. F., \& Meyer-Hofmeister, E. 2007, \aap, 463, 1
779: \bibitem[]{726} Meyer, F., \& Meyer-Hofmeister, E. 1994, \aap, 288, 175
780: \bibitem[]{727} Miller, J. M., Homan, J., \& Miniutti, G. 2006a, \apj, 652, L113
781: \bibitem[]{728} Miller, J. M., Homan, J., Steeghs, D., Rupen, M., Hunstead,
782: R. W., Wijnands, R., Charles, P. A., \& Fabian, A. C. 2006b, \apj, 653, 525
783: \bibitem[]{730} Narayan, R., \& Yi, I. 1994, \apj, 428, L13
784: \bibitem[]{731} Narayan, R., \& Yi, I. 1995a, \apj, 444, 231
785: \bibitem[]{732} Narayan, R., \& Yi, I. 1995b, \apj, 452, 710
786: \bibitem[]{733} Ramadevi, M. C., \& Seetha, S. 2007, \mnras, 378, 182
787: \bibitem[]{734} R\`o\.za\`nska, A., \& Czerny, B. 2000, \aap, 360, 1170
788: \bibitem[]{735} Rykoff, E. S., Miller, J. M., Steeghs, D., \& Torres, M. A. P. 2007, \apj, 666, 1129
789: \bibitem[]{736} Shakura, N. I., \& Sunyaev, R. A. 1973, \aap, 24, 337
790: \bibitem[]{737} Soleri, P., Altamirano, D., Fender, R. et al. 2008, astro-ph/0803.1735
791: \bibitem[]{738} Spitzer L., 1962, Physics of Fully Ionized Gases,
792: 2nd edition, Interscience Publ., New York, London
793: \bibitem[]{740} Spruit, H. C., \& Deufel, B. 2002, \aap, 387, 918
794: \bibitem[]{741} Stepney, S. 1983, \mnras, 202, 467
795: \bibitem[]{742} Sutherland, R.S. \& Dopita, M.A. 1993,\apjs, 88, 253
796: \bibitem[]{743} Tomsick, J. A. et al. 2008, \apj, in press (astro-ph/0802.3357)
797: \end{thebibliography}
798:
799: \clearpage
800:
801: \begin{table}[h]
802: \caption{\label{t:fits}Fitting results for GX 339-4}
803: \begin{tabular}{cccccccc}
804: \tableline\tableline
805: $\alpha$&$\dot m$&$L_{\rm c}/L_{\rm Edd}$&$L/L_{\rm Edd}$&$L_{\rm d}/L$(\%)&$T_{\rm eff}$(keV)&$r_d$&Compton region\\
806: \tableline
807: \multicolumn{8}{c}{Variable viscous parameter}\\
808: \tableline
809: 0.31&0.0687&0.023&0.0253&9.3&0.203&71&$3.3\la r\la 19$\\
810: 0.23&0.0299&0.008&0.009&11.2&0.164&82&$4.0\la r \la 8$\\
811: \tableline
812: \multicolumn{8}{c}{ Fixed viscous parameter}\\
813: \tableline
814: 0.25&0.0471&0.023&0.0269&14.4&0.221&142&$3.2\la r\la 25$\\
815: 0.25&0.0350&0.008&0.0088&9.0&0.156&61&no Compton region\\
816: \tableline
817: \end{tabular}
818: \tablecomments{Here $L_{\rm c}$, $L$, $L_{\rm d}$, $T_{\rm eff}$, and $r_d$ are the model predictions
819: for the corona luminosity, total luminosity, disk luminosity, disk temperature and disk size for given
820: viscous parameter $\alpha$ and accretion rate $\dot m$.}
821: \end{table}
822:
823: \clearpage
824:
825: \begin{figure}
826: \plottwo{f1a.eps}{f1b.eps}
827: \caption{\label{cmpregion}The left panel illustrates regions dominated by the Compton or conduction
828: cooling process as a function of the total condensation rate. The Compton scattering is dominant
829: at higher condensation rates (or higher accretion rates). The right panel shows the regions where Compton
830: cooling is important and its relative strength, $\lambda$, as a function of distance for condensation
831: rates of $\dot m_{\rm cnd}=0.001$ (dotted curve) and $\dot m_{\rm cnd}=0.01$ (solid curve). Between
832: these two condensation rates, inclusion of both cooling mechanisms is necessary, especially at distances
833: where $\lambda \sim 1$.}
834: \end{figure}
835:
836: \begin{figure}
837: \plotone{f2.eps}
838: \caption{\label{f:mdot-L}The luminosity as a function of the mass accretion rate. The total luminosity
839: increases with accretion rate and the energy conversion efficiency in the accretion flow, as measured
840: by ${\eta\over 0.1}={L\over L_{\rm Edd}}/{\dot M\over \dot M_{\rm Edd}}$, is greater at higher accretion
841: rates and lower values of the viscosity parameter $\alpha$.}
842: \end{figure}
843:
844: \begin{figure}
845: \plottwo{f3a.eps}{f3b.eps}
846: \caption{\label{f:r_d-L}Left panel: The inner disk size, which is modified by inclusion of two cooling
847: processes is shown as a function of the total luminosity. The size of the inner disk is large
848: at high luminosity. At a given luminosity, the inner disk size increases with decreasing values
849: of the viscosity parameter $\alpha$. Right panel: The outer extent of the inner cool disk and
850: the inner extent of the outer cool disk are illustrated as a function of the mass accretion rate
851: for a given viscosity parameter, $\alpha=0.2$. For a given accretion rate, the difference between these
852: two radii yields the spatial extent of the coronal gap region. The curve combines the condensation
853: model described in \S 2 for the inner disk and the disk evaporation model for the inward extent
854: of the outer disk based on Liu et al. (2002).}
855: \end{figure}
856:
857: \begin{figure}
858: \plotone{f4.eps}
859: \caption{\label{f:ratiomdot-mdot}The fraction of integrated condensation flow as a function of
860: mass accretion rate for three values of the viscosity parameter $\alpha$. A greater fraction of the coronal
861: flow condenses onto an inner disk at higher accretion rates, whereas no condensation occurs at
862: very low accretion rates. This indicates a strong coronal flow can become weaker toward
863: the central black hole with gas continuously condensing and accreting through a cool disk.}
864: \end{figure}
865:
866: \begin{figure}
867: \plotone{f5.eps}
868: \caption{\label{f:Teff-L}The maximal disk effective temperature as a function of luminosity
869: for three values of the viscosity parameter $\alpha$. The disk is heated by the accretion of the
870: condensed gas to a high temperature at a high luminosity. At a given luminosity, the
871: temperatures are insensitive to the viscosity parameters with only a slight increase
872: for lower viscosity parameters $\alpha$.}
873: \end{figure}
874:
875: \begin{figure}
876: \plotone{f6.eps}
877: \caption{\label{f:ratioL-mdot}The fraction of luminosity produced in the inner disk to the total
878: luminosity as a function of the mass accretion rate for three values of the viscosity parameter $\alpha$.
879: The contribution of the disk strengthens at higher accretion rates.}
880: \end{figure}
881:
882:
883: \begin{figure}
884: \plotone{f7.eps}
885: \caption{\label{f:ratioL-L}The fraction of disk luminosity to the total luminosity as a
886: function of the total luminosity for three values of the viscosity parameter $\alpha$. At low luminosities
887: ($L/L_{\rm Edd}\la 0.01$), the relative strength of disk component increases rapidly with the total
888: luminosity. The increase is less rapid and the ratio of luminosities tends to nearly a constant value
889: at higher luminosities.}
890: \end{figure}
891:
892:
893: \begin{figure}
894: \plotone{f8.eps}
895: \caption{\label{f:lambda-r}The cooling rate of the Compton process relative to that of the conduction process
896: as a function of distance in the fits to spectrum \#1 and \#2. It can be seen that Compton cooling is
897: stronger than conduction cooling in the model for spectrum \#1 where a disk thermal component
898: of 0.2 keV is fit. In the model for spectrum \#2 the Compton cooling is comparable with conduction only
899: in a small region and is much lower in the outer region.}
900: \end{figure}
901: \end{document}
902:
903:
904:
905:
906: