1: %%MS#52136
2: \documentstyle[12pt,aasms4]{article}
3: %\documentstyle[10pt,emulateapj]{article}
4:
5: \received{June 16 2000}
6: \accepted{July 26 2000}
7: %\journalid{337}{00 October 1996}
8: %\articleid{11}{14}
9:
10: %
11:
12: \begin{document}
13:
14: \title{A comparison of central temperatures of the intracluster gas
15: determined from X-ray and SZ measurements}
16:
17: \author{Tong-Jie Zhang}
18: \affil{Department of Astronomy, Beijing Normal University,
19: Beijing, 100875; and
20: National Astronomical Observatories,
21: Chinese Academy of Sciences, Beijing 100012, China}
22:
23: \and
24:
25: \author{Xiang-Ping Wu}
26: \affil{Beijing Astronomical Observatory and
27: National Astronomical Observatories,
28: Chinese Academy of Sciences, Beijing 100012, China}
29:
30: \begin{abstract}
31: A combination of the X-ray imaging and SZ measurements of clusters
32: permits an indirect determination
33: of the radial temperature profiles of intracluster gas, which requires no
34: assumption about the dynamical properties and the equation of state
35: for clusters. A comparison of such a derived gas temperature with that
36: given by the X-ray spectral analysis constitutes an effective probe of
37: properties of intracluster gas.
38: Using the available data of 31 clusters in literature, we have performed
39: the first comparison of the central gas temperatures provided by the
40: two methods. The good agreement between these two
41: independent temperature estimates suggests that the distribution of
42: intracluster gas is essentially consistent with isothermality
43: characterized by a mean polytropic index of $\gamma=0.9\pm0.1$.
44: \end{abstract}
45:
46: \keywords{galaxies: clusters: general ---
47: intergalactic medium --- X-rays: galaxies}
48:
49:
50: \section{Introduction}
51:
52: While the X-ray spectroscopic measurement is a powerful and unique
53: way nowadays to obtain the temperature of hot, diffuse gas contained
54: within clusters of galaxies,
55: it yields an emission-weighted temperature rather than
56: the true temperature distribution. The lack of detailed information
57: on the temperature profiles in clusters is probably the major source of
58: uncertainties in the present determinations of the total masses and
59: baryon fractions of clusters, which further hinders the dynamical
60: properties of clusters from the cosmological applications such as the
61: estimates of the cosmological parameters and the test of various
62: models of structure formation in the universe.
63: An independent, complementary method
64: of measuring the distribution of intracluster gas is thus desirable.
65: Silk \& White (1978) were the first to suggest the utilization of
66: the nonparametric reconstruction of the radial profiles of density
67: [$n_e(r)$] and temperature [$T(r)$] of intracluster gas
68: by combining the X-ray and SZ measurements. The basic idea is to
69: inverse the observed X-ray and SZ temperature surface brightness profiles of
70: clusters which have the functional forms of $n_e^2(r)T^{1/2}(r)$ and
71: $n_e(r)T(r)$, respectively. And a simple combination of the two
72: functions gives straightforwardly the gas density $n_e(r)$ and temperature
73: $T(r)$. This method requires no assumption about the dynamical
74: properties of clusters and the equation of state for intracluster
75: gas, and can therefore be regarded as an ideal and ultimate tool to probe
76: the gas distribution in clusters under spherical approximation.
77:
78:
79: Despite its elegant mathematical treatment further developed by
80: Yoshikaya \& Suto (1999) based on some theoretical models,
81: the pioneering suggestion of Silk \& White (1978)
82: has not yet been put into practice to date.
83: This has been primarily limited by the instrumental sensitivity and
84: resolution of detecting the temperature variations of the
85: cosmic background radiation (CBR) behind clusters.
86: So far, the marginal detections of the radial SZ temperature distributions
87: have been reported only for a few clusters (see Birkinshaw 1999),
88: which can hardly be used for the purpose of reconstructing
89: the gas temperature profiles because of the sparse data points
90: and the large associated uncertainties.
91: Nevertheless, we notice that the central measured or estimated
92: SZ effects have been available for a great number of clusters
93: with the past two decades' efforts, and these central CBR temperature
94: decrements in the Rayleigh-Jeans limit are usually used for setting
95: constraints on the cosmological parameters ($H_0$, $\Omega_M$ and
96: $\Omega_{\Lambda}$) in conjunction with the X-ray imaging and
97: spectral measurements. Yet, another application of
98: the central SZ data is instead to estimate the central temperature of
99: intracluster gas when combined with the X-ray imaging observation,
100: although this does not
101: completely achieve the original goal of Silk \& White (1978).
102: It provides an indirect measurement of the X-ray temperature at cluster
103: centers, which can be directly compared with the result given by the X-ray
104: spectral analysis. Additionally, because both X-ray emission and SZ
105: effect result from the gas distribution projected along the
106: light of sight, a comparison of the central temperatures indicated
107: by the two methods may also reveal valuable information about the
108: radial temperature variations. Using the published X-ray and SZ data
109: in literature, we present for the first time such a comparison
110: in this paper and discuss its implications for dynamical
111: properties of clusters. Throughout the
112: paper we assume $H_0=50$ km s$^{-1}$ Mpc$^{-1}$ and $\Omega_0=1$.
113:
114:
115:
116: \section{The model}
117:
118: The X-ray imaging observation provides a reliable measurement of the
119: X-ray azimuthally-averaged surface brightness profile of a cluster
120: which is usually described by the conventional $\beta$ model
121: (Cavaliere \& Fusco-Femiano 1976)
122: %1
123: \begin{equation}
124: S_x(r)=S_0\left(1+\frac{r^2}{r_c^2}\right)^{-3\beta+1/2}.
125: \end{equation}
126: In the scenario of an optically-thin, thermal bremsstrahlung emission,
127: the above form of $S_x(r)$ indicates (Cowie, Henriksen \& Mushotzky 1987)
128: %2
129: \begin{equation}
130: n_e(r)T^{1/4}(r)=n_{e0}T_0^{1/4}
131: \left(1+\frac{r^2}{r_c^2}\right)^{-3\beta/2},
132: \end{equation}
133: where $n_e$ and $T$ are the electron number density and temperature,
134: respectively. If we assume an equation of state,
135: $T(r)=T_0[n_e(r)/n_{e0}]^{\gamma-1}$, we can write the
136: electron number density as $n_e(r)=n_{e0}(1+r^2/r_c^2)^{-\delta}$,
137: in which $\delta=6\beta/(3+\gamma)$. The central electron number
138: density $n_{e0}$, temperature $T_0$ and the X-ray surface brightness $S_0$
139: are connected by
140: %3
141: \begin{equation}
142: n_{e0}^2= \left(\frac{3m_e\hbar c^2}{2^4e^6}\right)
143: \left(\frac{3m_ec^2}{2\pi kT_0}\right)^{1/2}
144: \frac{4\pi^{1/2}}{\mu_e g}
145: \frac{\Gamma(3\beta)}{\Gamma(3\beta-1/2)}
146: \frac{S_0(1+z)^4}{r_c},
147: \end{equation}
148: where $\mu_e=2/(1+X)$, $X=0.768$ is the primordial hydrogen mass fraction,
149: $g\approx1.2$ is the average Gaunt factor, $\Gamma$ is the gamma function,
150: and $z$ is the cluster redshift.
151:
152:
153:
154: On the other hand, the CBR temperature decrement/increment at the cluster
155: center predicted by the SZ effect is
156: %4,5
157: \begin{eqnarray}
158: \frac{\Delta T_{sz}}{T_{\rm CBR}}=-4\xi(x)\int_{0}^{+\infty}
159: \frac{kT(r)}{m_ec^2}\sigma_T n_e(r) dr,\\
160: \xi(x)=\frac{x^2e^x}{2(e^x-1)^2}\left(4-x\coth \frac{x}{2}\right),
161: \end{eqnarray}
162: where $T_{\rm CBR}$ is the temperature of the present CBR, and
163: $x=h\nu/kT_{\rm CBR}$ is the dimensionless frequency.
164: A straightforward computation using eqs.(2)-(5) as well as
165: the polytropic equation of state gives
166: %6
167: \begin{eqnarray}
168: \left(\frac{kT_0}{m_ec^2}\right)^{3/2}= &
169: \left(\frac{\Delta T_{sz}}{T_{\rm CBR}}\right)^2
170: \left(\frac{\Gamma(3\beta^{\prime}/2)}
171: {\Gamma(3\beta^{\prime}/2-1/2)}\right)^2
172: \frac{\Gamma(3\beta-1/2)}{\Gamma(3\beta)} \nonumber\\
173: & \;\cdot\; \frac{\sqrt{2} \alpha \mu_e g m_e c^3 }
174: {8\pi^2\sqrt{3}(3\beta-3/2)\xi^2\sigma_T(1+z)^4S_0r_c}
175: \end{eqnarray}
176: in which $\beta^{\prime}=3\beta\gamma/(3+\gamma)$,
177: and $\alpha$ is the fine structure
178: constant. The above expression is often used in the determination of
179: the Hubble constant for an isothermal gas distribution
180: if we write $r_c=d_A(z)\theta_c$ where $d_A$ is the
181: angular diameter distance to the cluster at $z$. On the other hand,
182: given the Hubble constant and the polytropic index $\gamma$,
183: all the parameters at the right-hand of eq.(6)
184: are measurable from the X-ray imaging and SZ measurements, which
185: will instead allow us to derive the central temperature of
186: the X-ray emitting gas.
187:
188:
189: \section{Application to X-ray clusters}
190:
191: Our cluster sample consists primarily of 21 nearby clusters
192: studied extensively by Mason \& Myers (2000) plus 10 high-redshift
193: clusters with recent, reliable SZ data in the literature.
194: Inclusion of a cluster is also subject to whether or not
195: the good spectral temperature data are available
196: in order to facilitate our comparison between different methods.
197: We take the best-fit values of $\beta$, $r_c$ and $S_0$ (or $L_x$)
198: as well as the emission-weighted temperature data
199: directly from the literature. For the latter, we adopt
200: the temperature given by a single-phase (s.p.) model and
201: the value by excluding the cooling flows (c.l.) (White 2000),
202: respectively. Since the X-ray spectrum is always
203: observed over a finite bandpass, we assume an optically-thin,
204: thermal bremsstrahlung emission model with the metal abundances of
205: $30\%$ solar to convert the observed $S_0$ (or $L_x$) over the
206: corresponding energy band into the bolometric value.
207: We use the central SZ decrement/increment of $\Delta T_{sz}$
208: after the correction of the finite beam in each observation.
209: All the input data of the X-ray and SZ measurements are summarized in
210: Table 1.
211:
212:
213: For an isothermal gas distribution, the jointly determined central
214: temperature, $T_{est}$, from the X-ray imaging and SZ measurements
215: (eq.[6]) can be directly compared with the X-ray spectral result, $T_{spec}$.
216: We display in Fig.1 our derived $T_{est}$ versus the observed
217: $T_{spec}$ (s.p. model)
218: for the 31 clusters in Table 1. A glimpse of Fig.1 reveals
219: that the two sets of data are roughly consistent with each other.
220: The best-fit $T_{spec}$-$T_{est}$ relation reads
221: %7,8
222: \begin{eqnarray}
223: T_{est}=10^{-0.22\pm0.18}T_{spec}^{1.22\pm0.20}, &
224: {\rm s.p.\; model};\\
225: T_{set}=10^{-0.36\pm0.26}T_{spec}^{1.29\pm0.29}, &
226: {\rm c.l.\; corrected},
227: \end{eqnarray}
228: in which the error bars are the $90\%$ confidence limits which have
229: taken the measurement uncertainties into account.
230: We have also tested the dependence of the best-fit
231: $T_{spec}$-$T_{est}$ relation on the Hubble constant by adopting a larger
232: value of $H_0=75$ km s$^{-1}$ Mpc$^{-1}$. This only leads to a minor
233: modification to the above results. For instance, the best-fit relation for
234: the s.p. model now reads $T_{est}=10^{-0.34\pm0.11}T_{spec}^{1.22\pm0.12}$.
235:
236:
237: \placefigure{fig1}
238:
239: If the intracluster gas deviates from isothermality,
240: our comparison of $T_{est}$ and $T_{spec}$
241: becomes to be a little complex because we need to recover the central
242: temperature from the emission-weighted spectral fit. On the
243: other hand, we can place a useful constraint on the polytropic index
244: $\gamma$ if we demand that our derived central temperature from eq.(6)
245: identify the spectral value after a proper correction to
246: the emission-weighted temperature $\overline{T}$ predicted by
247: %8
248: \begin{equation}
249: \frac{dE}{d\nu}\propto \overline{T}^{-1/2} g_{ff}(\overline{T},\nu)
250: \exp(-h\nu/k\overline{T})
251: \int n_e^2 dV,
252: \end{equation}
253: where $g_{ff}$ is the Gaunt factor of the free-free emission, $\nu$ is
254: the X-ray observing frequency, and the integration is performed over the whole
255: cluster volume if we neglect the finite detection aperture.
256: This compares with the theoretically expected spectrum for
257: the hot intracluster gas
258: %9
259: \begin{equation}
260: \frac{dE}{d\nu}\propto \int n_e^2 T^{-1/2} g_{ff}(T,\nu)
261: \exp(-h\nu/k{T}) dV.
262: \end{equation}
263: For each cluster with the reported $\overline{T}=T_{spec}$,
264: we use the Monte Carlo simulations over the typical energy band
265: $0.5$ - $10$ keV to obtain the central temperature $T_0$ and the
266: corresponding $\gamma$ with the restriction of eq.(6). The
267: resultant $\gamma$ value for each cluster is listed in Table 1,
268: and a combined analysis of the results among the 31 clusters yields
269: %10
270: \begin{eqnarray}
271: \langle \gamma\rangle =0.91\pm0.14, & {\rm s.p. \; model};\\
272: \langle \gamma\rangle =0.93\pm0.14, & {\rm c.l.\; corrected},
273: \end{eqnarray}
274: which indicates that clusters are essentially consistent with isothermality
275: at $90\%$ significance level.
276:
277:
278: \section{Discussion and conclusions}
279:
280: Our joint analysis of the indirect and direct temperature measurements
281: from the X-ray and SZ observations provides a useful clue to resolving
282: the temperature profile discrepancy raised in recent years
283: particularly by Markevitch et al. (1998), Irwin, Bregman \& Evrard (1999),
284: White (2000) and Irwin \& Bregman (2000). Markevitch et al. (1998) claimed
285: a significant temperature decline with increasing radius represented
286: by a polytropic index of 1.2-1.3, while the other investigators have
287: essentially detected an isothermal temperature profile.
288: Our present study gives an independent yet convincing evidence
289: for the assumption of isothermality. The advantage of our algorithm
290: over the previous analysis is that we have obtained the central temperatures
291: of clusters without utilizing the X-ray spectral data, while a combination
292: of both has allowed us to set a valuable constraint on the radial temperature
293: profiles of the hot X-ray emitting gas.
294:
295:
296: A joint determination of the central temperature of intracluster gas
297: from the X-ray imaging and SZ measurements can be used at present
298: as an independent and complementary method to examine the reliability and
299: accuracy of the temperature
300: measurement provided by the conventional X-ray spectral analysis,
301: although our ultimate goal is to obtain the detailed radial profiles of
302: electron number density and temperature when the SZ surface brightness
303: can be precisely mapped (Silk \& White, 1978; Yoshikawa \& Suto 1999).
304: Yet, we should point out that it may not be easy to identify
305: the temperature structure of the hot gas from such a combined analysis,
306: in the sense that the dominant component in both $S_x$ and $\Delta T_{sz}$
307: is characterized by the sharp decrease of
308: the electron number density with outward radius, while it is unlikely that
309: the temperature profiles can demonstrate a similar variation.
310: An accurate measurement of $S_x$ and $\Delta T_{sz}$ at outermost radii
311: will thus be needed if one wants to apply the present technique to
312: the reconstruction of temperature profiles of intracluster gas.
313: Alternatively, the present technique requires that the X-ray surface
314: brightness is well represented by the $\beta$ model. Presence of cooling
315: flows in the central regions of many X-ray clusters may yield an
316: underestimate of the $\beta$ parameter and the core radius $r_c$ and
317: an overestimate of the central surface brightness $S_0$,
318: which will in turn affect our estimate of $T_{est}$ through eq.(6).
319: Indeed, an examination of Fig.1 reveals that all the massive
320: cooling flow clusters with $\dot{M}>600$ $M_{\odot}$ yr$^{-1}$
321: (A478, A1835, A2204 and Zw3146) demonstrate rather large scatters around
322: the $T_{est}=T_{spec}$ relation. Therefore, the present method is most
323: applicable to morphologically simple clusters without cooling flows.
324:
325:
326:
327:
328: \acknowledgments
329: This work was supported by
330: the National Science Foundation of China, under Grant 19725311.
331:
332:
333: \clearpage
334:
335:
336: \begin{references}
337:
338: \reference{}Arnaud, M., \& Evrard, A. E., 1999, \mnras, 305, 631
339: \reference{}Birkinshaw, M., 1999, \physrep, 310, 97
340: \reference{}Cavaliere, A., \& Fusco-Femiano, R., 1976, \aap, 49, 137
341: \reference{}Cooray, A. R., 1999, \mnras, 307, 841
342: \reference{}Cowie, L. L., Henriksen, M., \& Mushotzky, R. F., 1987,
343: \apj, 317, 593
344: \reference{}Ettori, S., \& Fabian, A. C., 1999, \mnras, 305, 834
345: \reference{}Irwin, J. A., \& Bregman, J. N., 2000, \apj, in press
346: \reference{}Irwin, J. A., Bregman, J. N., \& Evrard, A. E., 1999,
347: \apj, 519, 518
348: \reference{}Markevitch, M., Vikhlinin, A., Forman, W. R., \&
349: Sarazin, C. L., 1998, \apj, 527, 545
350: \reference{}Mason, B. S., \& Myers, S. T., 2000, \apj, in press
351: \reference{}Mohr, J. J., Mathiesen, B., \& Evrard, A. E., 1999,
352: \apj, 517,627
353: \reference{}Reese, E.D., et al., 2000, \apj, 533, 38
354: \reference{}Rizza, E., \& Burns, J. O., 1998, \mnras, 301, 328
355: \reference{}Silk, J., \& White, S. D. M., 1978, \apj, 226, L103
356: \reference{}White, D. A., 2000, \mnras, 312, 663
357: \reference{}Wu, X.-P., Xue, Y.-J., \&
358: Fang, L.-Z., 1999, \apj, 524, 22
359: \reference{}Yoshikawa, K., \& Suto, Y., 1999, \apj, 513, 549
360: \end{references}
361:
362:
363: \clearpage
364:
365:
366:
367: \figcaption{The derived central temperatures are plotted against
368: the spectral results (single-phase model) for a sample of 31 clusters.
369: Solid line shows the best-fit relation, and dotted
370: line indicates $T_{est}=T_{spec}$. The high-redshift ($z>0.1$) and
371: low-redshift ($z<0.1$)
372: clusters are represented by the filled squares and triangles,
373: respectively. The marked clusters are the massive cooling flow ones with
374: $\dot{M}>600$ $M_{\odot}$ yr$^{-1}$ (White 2000).
375: \label{fig1}}
376:
377:
378: \end{document}
379:
380:
381:
382:
383:
384:
385:
386:
387:
388:
389: