astro-ph0305354/ms.tex
1: % A2256/XTE ApJ paper; revised version, May 9, 2003
2: \documentstyle[12pt,aas2pp4,epsfig]{article}
3: %\documentstyle[12pt,aas2pp4,epsfig]{preprint}
4: 
5: \def\be{\begin{equation}}
6: \def\ee{\end{equation}}
7: \def\ea{{\it et al.}\,}
8: \def\np{\newpage}
9: \def\eg{{\it e.g.},\,}
10: \def\c{\ {\cdot}}
11: \def\aa{{\it A\&A}\,}
12: \def\apj{{\it ApJ}\,}
13: \def\apjs{{\it ApJS}\,}
14: \def\n{{\it Nature}\,}
15: \def\mn{{\it MNRAS}\,}
16: \def\aj{{\it AJ}\,}
17: \def\rel{relativistic \,}
18: 
19: \begin{document}
20: 
21: \hfill\today
22: \title{RXTE Observations of A2256}
23: 
24: \author{Yoel Rephaeli\altaffilmark{1,2}, and Duane 
25: Gruber\altaffilmark{3}}
26: 
27: \affil{$^1$Center for Astrophysics and Space Sciences, 
28: University  of California, San Diego,  La Jolla, CA\,92093-0424}
29: 
30: \affil{$^2$School of Physics and Astronomy, 
31: Tel Aviv University, Tel Aviv, 69978, Israel}
32: 
33: \affil{$^3$4789 Panorama Drive, San Diego CA 92116}
34: 
35: \begin{abstract}
36: The cluster of galaxies A2256 was observed by the PCA and HEXTE 
37: experiments aboard the RXTE satellite during the period July 2001 - 
38: January 2002, for a total of $\sim$343\,ks and $\sim$88\,ks, 
39: respectively. Most of the emission is thermal, but the data analysis 
40: yields evidence for two components in the spectrum. Based on 
41: statistical likelihood alone, the secondary component can be either 
42: thermal or power-law. Inclusion in the analysis of data from ASCA 
43: measurements leads to a more definite need for a second component. 
44: Joint analysis of the combined RXTE-ASCA data sets yields $kT_1 = 
45: 7.9^{+0.5}_{-0.2}$ and $kT_2 = 1.5^{+1.0}_{-0.4}$, when the second 
46: component is also thermal, and $kT = 7.7^{+0.3}_{-0.4}$ and $\alpha = 
47: 2.2^{+0.9}_{-0.3}$, if the second component is fit by a power-law with 
48: (photon) index $\alpha$; all errors are at 90\% confidence. Given the 
49: observed extended regions of radio emission in A2256, it is reasonable 
50: to interpret the deduced power-law secondary emission as due to Compton 
51: scattering of the radio producing relativistic electrons by the cosmic 
52: microwave background radiation. If so, then the {\it effective, mean 
53: volume-averaged} value of the magnetic field in the central 1$^{o}$ 
54: region of the cluster -- which contains both the `halo' and `relic' 
55: radio sources -- is $B \sim 0.2^{+1.0}_{-0.1}$ $\mu G$.
56: \end{abstract}
57: 
58: \keywords{Galaxies: clusters: general --- galaxies: clusters: individual 
59: (A2256) --- galaxies: magnetic fields --- radiation mechanisms: 
60: non-thermal} 
61: 
62: \section{Introduction} 
63: 
64: The improved spatial resolution and wider spectral coverage 
65: of current X-ray satellites provide further motivation for 
66: a less simplified description of the properties of intracluster 
67: (IC) gas, and the consideration of additional cluster phenomena,  
68: such as non-thermal (NT) processes. An obvious generalization of 
69: the simple isothermal model for the gas is made by allowing for a 
70: more realistic temperature structure. This has, in fact, been 
71: considered in quite a few analyses of cluster X-ray data, leading 
72: to clear evidence for radial variation of the gas temperature in 
73: some clusters (\eg, Honda \ea 1996, Watanabe \ea 1999, Markevitch 
74: \ea 1998). Somewhat less obvious is the need to include a NT 
75: component in the X-ray spectra of (at least some) clusters. Such 
76: emission has long been predicted (\eg Rephaeli 1977; for a review, 
77: see Rephaeli 2001), and recent observations give appreciable evidence 
78: for its likely detection in a few clusters.
79: 
80: It is of considerable interest to know if IC gas within the 
81: central ($\sim 1$ Mpc) cluster region is non-isothermal. In addition 
82: to insight gained from the form of the temperature profile on 
83: physical processes in the gas, and its cosmological evolution, 
84: knowledge of the density and temperature distributions is clearly 
85: very important not only for the determination of basic cluster 
86: properties, such as the gas and dark matter masses, but also for 
87: the use of phenomena in clusters to determine cosmological parameters 
88: (\eg, from measurements of the Sunyaev-Zeldovich effect and 
89: gravitational lensing). On the other hand, measurement of a NT 
90: spectral component -- especially in clusters in which extended 
91: regions of radio emission have been measured --  is essential for 
92: the characterization of NT quantities and phenomena in clusters, 
93: such as the strength and morphology of magnetic fields, densities 
94: and energy content of \rel electrons and protons, and the interaction 
95: of these particles with the gas.
96: 
97: Measurements with RXTE and BeppoSAX satellites are particularly 
98: useful in the search for NT spectral components in cluster spectra.
99: First attempts to detect NT emission from a few clusters 
100: with the HEAO-1, CGRO, and ASCA satellites were unsuccessful (Rephaeli, 
101: Gruber \& Rothschild 1987, Rephaeli \& Gruber 1988, Rephaeli, Ulmer 
102: \& Gruber 1994, Henriksen 1998). The improved sensitivity and wide 
103: spectral band of the RXTE and BeppoSAX seem to have resulted in the 
104: detections of NT emission in Coma (Rephaeli, Gruber \& Blanco 1999, 
105: Fusco-Femiano \ea 1999, Rephaeli \& Gruber 2002), and A2319 
106: (Gruber \& Rephaeli 2002). A NT component was also deduced in the 
107: BeppoSAX spectra of A2256, A119, and A754 (Fusco-Femiano \ea 2002). 
108: Here we report the results from a long RXTE observation of A2256, 
109: possibly a merging cluster with complex X-ray (Sun \ea 2002) and 
110: radio (Giovannini, G., \ea 1999) morphologies.
111: 
112: \section{Observations and Data Reduction}
113: 
114: A2256 is a rich radio and X-ray bright cluster at $z=0.0581$, with 
115: complex morphology. Extended radio emission from A2256 was measured 
116: by Bridle \& Fomalont (1976), Bridle \ea (1979), Giovannini \ea
117: (1999), and most recently by Clarke \& Ensslin (2001). In addition to 
118: emission from several strong radio sources in the central region of 
119: the cluster, there is a centrally located extended emission region, 
120: as well as regions of extended emission (located in the northern side 
121: of the cluster) that are thought to be radio relics. The centrally 
122: located emission is characterized by a spectral (energy) index 
123: $\alpha_C \sim 2.0$ in the $\sim 1.4 - 5$ GHz band, while the main 
124: relic has a flatter spectrum with index $\alpha_R \sim 1.0$ (Clarke 
125: \& Ensslin (2001).   
126: 
127: The cluster was observed by most previous X-ray satellites, most 
128: recently by XMM and {\it Chandra}. Several emission regions were 
129: resolved by {\it Chandra}, yielding further evidence for the view 
130: that the cluster is undergoing a merger (Sun \ea 2002). These 
131: measurements also indicate considerable variation of the temperature 
132: in the central region, with a mean value of $\simeq 6.7$ keV, but 
133: with hotter ($\sim 10$ keV) and colder ($\sim 5$ keV) regions. Of 
134: particular interest to us here are BeppoSAX observations with the MECS 
135: and PDS detectors: From an analysis of these measurements 
136: Fusco-Femiano \ea (2000) deduced the presence of a NT spectral 
137: component at a $\sim 4.6\sigma$ confidence, with a (photon) index 
138: roughly in the range $0.3 - 1.7$, in addition to the main thermal 
139: component with a temperature of $7.4 \pm 0.2$ keV. A short, $\sim 30$
140: ks, observation of A2256 by RXTE yielded only an upper limit on a NT 
141: flux, $\sim 2.3\times 10^{-7}$ cm$^{-2}\,s^{-1}\,keV^{-1}$ at 30 keV, 
142: and a lower limit of $\sim 0.36\, \mu$G on the mean magnetic field 
143: (Henriksen 1999).
144: 
145: A2256 was observed with RXTE for a nominal total observation
146: time of 400 ks between July 2001 and January 2002. After the
147: application of data selection criteria recommended by the RXTE
148: project, 343 ks of screened data were collected with the PCA in
149: 111 one-orbit observations, spaced irregularly over the seven-month
150: campaign. For the HEXTE, which beam-switches observations with 32-second 
151: dwells between source and background fields, and has in addition about
152: 50\% detector dead time, the net observation time was 88 ks with
153: each of the two clusters. On time scales of two weeks or longer,
154: the limit to variability observed with PCA was less than 1\%.
155: Because of the much lower signal to background, corresponding
156: HEXTE limits to variability are larger, about 20\%.
157: 
158: Following project practice at the time, PCA data were collected in two
159: of the 5 detectors.  One of these, PCU 0, had lost its propane guard
160: layer by the time of the A2256 campaign, but the net flux and spectral
161: shape differed negligibly from those obtained with PCU 2, which indicates
162: that the project has very successfully produced modifications for the
163: response matrix and background estimation tool for PCU 0.
164: 
165: ASCA observations of A2256 have been archived for several
166: observations, the longest of which was 36 ks on 22 July 1993, and
167: the next-longest, 26 ks, was carried out on 8 April 1993. Standard
168: archival GIS and SIS spectra and matrices were provided. Preliminary
169: spectral study of the two observations gave very similar best fits
170: of an isothermal spectral model. However, there was large 
171: non-statistical scatter of about 10\% in the SIS data of the longer 
172: July observation. For this reason the July data were considered less 
173: reliable, and the April data alone were employed in the joint analysis 
174: with the RXTE data.
175: 
176: \section{Spectral Analysis}
177: 
178: The combined ASCA and RXTE data provide spectral information on the 
179: rather broad energy range of 0.6 to 100 keV. The four ASCA detectors, 
180: two PCA detectors and two HEXTE clusters, with thousands of energy 
181: channels combined, rather heavily oversample this range. After pilot 
182: spectral studies revealed no evidence for sharp features in the raw 
183: data, we proceeded to reduce the spectral oversampling to a reasonable 
184: level by combining counts from similar detectors and by summing counts 
185: in adjacent spectral channels into groups whose energy width was set 
186: at about one half of the FWHM detector energy resolution at the given 
187: energy.  Appropriate response matrices were also generated, and 
188: standard ftools were employed. The final spectral set contained 44 
189: energy bands for the ASCA data and 44 for the RXTE data.
190: 
191: An additional systematic error of 0.5\% per energy channel was added 
192: in quadrature to the statistical error of the PCA data (e.g Wilms 
193: et al. 1999). No systematic error was used with HEXTE data, and 2\% 
194: systematic error was used for both the ASCA SIS and GIS data. 
195: Spectral analysis was performed separately on the RXTE data, jointly 
196: on the RXTE and ASCA data sets, and in a restricted analysis, also 
197: on just the HEXTE data above 15 keV. 
198: 
199: Spectral models were limited to three cases: an isothermal thermal 
200: spectrum (based on a Raymond-Smith emission code), two-temperature 
201: thermal, and a thermal plus a power-law. The RXTE data by themselves 
202: provide only weak evidence of the need for an extra component beyond 
203: isothermal. The $\chi^2$ of 46.8 (40 degrees of freedom [dof]) for an 
204: isothermal fit is acceptable; however inclusion of a second component 
205: reduces $\chi^2$ modestly to 40.2 (38 dof) with an extra 0.9 keV thermal 
206: component, or to 40.4 (38 dof) with an extra power-law, whose best-fit 
207: photon index is a rather steep 4.0. For four ``interesting'' parameters, 
208: the change in $\chi^2$ (Lampton \ea 1976) gives 90\% error limits for 
209: the 4-20 keV power-law flux of $(0.5-4.8)\times 10^{-11}$ 
210: erg-cm$^{-2}$ s$^{-1}$. The temperature for the main spectral component 
211: is in the rnage $\sim 7.7-7.9$ keV for all three cases, with formal 
212: ($1\sigma$) errors of $\sim 0.1-0.3$ keV.
213: 
214: By fitting jointly with the ASCA data one obtains much more decisive
215: results.  An isothermal fit is ruled out both by a high $\chi^2$
216: of 155.4 for 82 dof, and by much improved fits with a second component,
217: $\chi^2$ = 96.9 with a second thermal at 1.4 keV, and $\chi^2$ = 104.5
218: with a power-law component with best-fit photon index 2.2.  For 
219: both of these cases the $\chi^2$ is somewhat high for 80 dof, but
220: this may reflect slight under-correction for systematic errors of
221: background subtraction and the response matrices.  With four 
222: interesting parameters (kT, abundance, power-law flux and index) 90\% 
223: error bounds for the power-law flux, now given for the interval 
224: 0.8--40 keV, are $(2.5-19.1)\times 10^{-11}$ erg-cm$^{-2}$ s$^{-1}$.  
225: Best fit parameters and 90\% confidence errors for the joint fits to 
226: the three spectral models are listed in Table 1. (For each combination 
227: of detectors the energy range for the power-law flux has been chosen 
228: to provide parameter and error estimates which are nearly independent 
229: of the other parameters. This is approximately equal to the energy 
230: span of the joint data set.) In Figure 1, we show the spectrum of the 
231: best-fit isothermal plus power-law model (data and model components 
232: are displayed) in the upper panel, and residuals to the fit in the 
233: lower panel.
234: 
235: Most of the statistical weight in parameter estimation comes from
236: data at the lowest energies.  Of special interest for the thermal
237: plus power-law case is whether the HEXTE data favor the presence
238: of a power-law component. We tested the HEXTE data against a model
239: in which the thermal parameters are set by the joint fit. With
240: no second component the HEXTE data give a marginally acceptable
241: ($P<0.06$) $\chi^2$ of 27.1 for 19 dof. When the power-law flux
242: is allowed to float to a best fit value the $\chi^2$ is dropped
243: by 9.3 to 17.8.  Allowing also the index to vary gives a best-fit
244: value of 1.8, which -- within errors -- is consistent with the value 
245: of 2.2 obtained in the joint fit. For one interesting parameter, 
246: the error bounds on the 15--40 keV flux are $(1.2-4.3)\times 10^{-12}$ 
247: erg-cm$^{-2}$ s$^{-1}$.
248: 
249: The NT 20-80 keV flux (of interest for a direct comparison with 
250: the BeppoSAX rasults) computed from the best-fit parameters resulting 
251: from the full (ASCA, PCA, and HEXTE) dataset is 
252: $(0.7-8.6)\times 10^{-12}$ erg-cm$^{-2}$ s$^{-1}$, formally 
253: significant at the 2.9 $\sigma$ level. The signficance of this 
254: flux is lower if HEXTE data are not included in the analysis. 
255: Note also that similar but less significant results are obtained 
256: when we first find the (poorly determined) best-fit isothermal to 
257: the PCA and ASCA data, and then determine a net high energy flux 
258: from the HEXTE data (a procedure adopted in the corresponding 
259: BeppoSAX MECS/PDS analysis). Doing so results in a 20-80 keV flux 
260: error bounds of $(0.2-8.0)\times 10^{-12}$ erg-cm$^{-2}$ s$^{-1}$, 
261: formally significant at 2.2 $\sigma$.
262: 
263: \begin{table*}        
264: \caption{Results of the spectral analysis}
265: 
266: \begin{tabular}{|l|ccc|} 
267: \hline 
268: Parameter & Single Thermal & Double Thermal & Thermal + Power-law \\                   
269: \hline   
270: $kT_1$ (\rm{keV}) &$7.66\pm 0.12$ & $7.91^{+0.48}_{-0.20}$ & $7.67^{+0.28}_{-0.39}$ \\      
271:                   &               &                        &                  \\
272: $kT_2$ (\rm{keV}) &               & $1.45^{+0.98}_{-0.35}$ &                  \\
273:                   &               &                        &                  \\            
274: $\alpha$          &               &                        &  $2.16^{+0.86}_{-.30}$ \\      
275: Secondary flux fraction &         &                        &                      \\       
276: \,\,\, 0.5-2 keV   &              & $0.084^{+0.069}_{-0.035}$  &                   \\
277: \,\,\, 2-10 keV    &              & $0.015^{+0.130}_{-0.060}$  &                   \\
278: \,\,\, 0.8-40 keV &             &                       & $0.101^{+0.083}_{-0.077}$ \\   
279:                   &               &                        &                  \\       
280: Abundance (solar)  & 0.194$\pm$0.018 & 0.208$\pm$0.028 & 0.218$\pm$0.030   \\                
281: \hline
282: \end{tabular} 
283:          
284: \tablenotetext{}{All quoted errors are at the 90\% confidence level. } 
285: \end{table*}               
286: 
287: Fusco-Femiano et al. (2000), reporting results of A2256 measurements 
288: with the BeppoSAX satellite, have claimed detection of a 20-80 keV 
289: NT flux of $1.2\times 10^{-11}$ ergs cm$^{-2}$ s$^{-1}$ at 4.6$\sigma$ 
290: significance. Assuming normal statistics we convert this 
291: to 90\% error bounds of [0.74, 1.66] in the same units.  Using the  
292: joint RXTE/ASCA dataset we obtain a comparable best-fit value for 
293: 20-80 keV flux of 0.26 and 90\% confidence error bounds of 
294: [0.01, 0.79] in these units. Thus, while our best fit value is a 
295: factor of 4.6 smaller than that obtained by Fusco-Femiano et al. 
296: (2000), it is not in strong conflict. Fusco-Femiano have discussed 
297: the possibility that the radio and X-ray NT components are complex, 
298: with more than one index. In this case, our joint RXTE/ASCA NT flux 
299: may be sensitive largely to the steeper index visible to ASCA. Thus 
300: a better comparison may be with our HEXTE-only analysis. Indeed, 
301: this analysis corresponds rather closely to the approach adopted in 
302: the analysis of the BeppoSAX data. The HEXTE result gives a best-fit 
303: 20-80 keV flux of $4.3 \cdot 10^{-12}$ ergs cm$^{-2}$ s$^{-1}$ and a 
304: 90\% confidence error interval of $(0.3-10.0)\times 10^{-12}$ 
305: ergs cm$^{-2}$ s$^{-1}$. While smaller than the reported BeppoSAX value 
306: by almost a factor of three, this result is not in a great conflict. 
307: 
308: \section{Discussion}
309: 
310: A2256 is the third cluster with extended regions of radio emission 
311: that has been observed by the RXTE for more than 100 ks. The results 
312: of the analysis reported here are qualitatively similar to those we 
313: have previously reported on Coma (Rephaeli, Gruber \& Blanco 1999, 
314: Rephaeli \& Gruber 2002) and A2319 (Gruber \& Rephaeli 2002). In all 
315: three clusters the RXTE measurements yield evidence that the spectra 
316: in the combined PCA and HEXTE bands contain a secondary component which 
317: is either thermal or power-law. In the case of A2256 this evidence is 
318: much stronger when the ASCA data are included in the analysis. However, 
319: the spectral analysis alone does not yield sufficient statistical 
320: preference for the nature of the second component. We invoke other 
321: considerations in an attempt to determine the nature of this 
322: component.                     
323: 
324: Consider first thermal emission from IC gas at a lower temperature than 
325: that of the main emission component, as listed in Table 1. That deep 
326: obervations over a wide spectral range require a more realistic emission 
327: model than a single temperature gas is, of course, not unexpected. 
328: Indeed, recent mapping of the temperature in the central $\sim 1/2$ 
329: Mpc ($H_0$ = 70 km s$^-1$ Mpc$^-1$) radial region of A2256 by {\it 
330: Chandra} shows an emitting region (in the NE side of the cluster) 
331: with a fractional projected area of roughly $\sim 1/10$ where the gas 
332: temperature is $kT \sim 4$ keV. This value is still significantly higher 
333: than the value at the upper end of the 90\% significance interval, 
334: $kT_2 \simeq 2.4$ keV. Emission at this temperature with a fractional 
335: 0.5-2 keV flux contribution of more than $\sim 5\%$ would have been 
336: measured, especially by ROSAT whose PSPC detector energy band matched 
337: this spectral range. Even so, we cannot formally rule out (also because 
338: of differences in collecting areas of the various detectors) the 
339: possibility that the second component is thermal at the above 
340: (relatively) low temperature. This is particularly so given the more 
341: realistic expectation that a two-temperature gas model is just a 
342: simplified representation of a more realistic continuous temperature 
343: distribution, as we have previously argued in the interpretation of 
344: RXTE measurements of Coma and A2319 (Rephaeli, Gruber \& Blanco 1999, 
345: Gruber \& Rephaeli 2002, Rephaeli \& Gruber 2002).
346:                                       
347: The main motivation for selecting A2256 for the long RXTE observation is, 
348: of course, the presence of extended regions of radio emission in the 
349: cluster. We first note that since significant emission from an AGN in the 
350: FOV is unlikely -- emission from QSO 4C +79, near the edge of the RXTE
351: FOV, was estimated to be negligible (Fusco-Femiano 2000) -- and given
352: no evidence for flux variability, we naturally associate the emission
353: with the cluster. As is well known (\eg, Rephaeli 1977), Compton
354: scattering of the radio producing relativistic electrons off the
355: cosmic microwave background boosts photon energies to the X-ray 
356: region. From the measured radio and X-ray fluxes the magnetic field can 
357: then be inferred. However, in the case of A2256 this is not 
358: straightforward because of the complex structure of the radio emission 
359: which is dominated by a few extended sources with spectral (energy) 
360: indices in the wide range, $\sim 0.3 - 1.1$, with substantial errors. 
361: Given no clear expectation on the predicted spectral index of the NT 
362: X-ray emission -- and the complex spatial morphology of the radio 
363: emission -- we use the measured total flux at 1.4 GHz, 397 mJy 
364: (Giovannini \ea 1999), and the deduced range of the X-ray power-law index 
365: to compute an {\it effective} value of the magnetic field across the 
366: large field of view of the RXTE, which includes all the dominant radio 
367: sources. Doing so, we determine the relatively wide 90\% confidence range 
368: for the {\it mean, volume-averaged} field, $B_{rx} \simeq 
369: 0.2^{+1.0}_{-0.1} \,\mu$G. (The very high value at the 
370: upper end of this interval results from our conservatively estimated 
371: lower limit on the flux.) We emphasize that this value has only limited 
372: meaning: Since there is no spatial information on the power-law X-ray 
373: emission, the implicit assumption made here -- and in all similar 
374: analyses of cluster magnetic fields from radio and X-ray measurements -- 
375: is that these emissions occur over the same volume. If so, we can also 
376: {\it estimate} the mean \rel energy density within the emitting region, 
377: radius R, by integrating the electron energy distribution over 
378: energies in the observed radio and X-ray bands. Doing so, we obtain 
379: $\rho_{e} \simeq 5^{+1.0}_{-4}\times 10^{-14} (R/1 Mpc)^{-3}$ 
380: erg\,cm$^{-3}$.
381: 
382: Although the estimated mean field has limited meaning, it is comparable 
383: to the values we deduced for the mean field in Coma (Rephaeli \& Gruber 
384: 2002) and A2319 (Gruber \& Rephaeli 2002). Values of the field deduced 
385: from radio and X-ray measurements, $B_{rx}$, are generally much lower 
386: than those obtained from Faraday rotation measurements (\eg, Clarke, 
387: Kronberg, and B\"ohringer 2001, and the review by Carilli \& Taylor 2002) 
388: of background radio sources seen through clusters, $B_{fr}$. The mean 
389: strength of IC fields has direct implications on the range of electron 
390: energies that are deduced from radio measurements, and therefore on the 
391: electron (Compton-synchrotron) loss time. The higher the electron energy, 
392: the shorter is the energy loss time; a short loss time would have 
393: immediate consequences on \rel electron models (\eg, Rephaeli 1979, 
394: Sarazin 1999, Ensslin \ea 1999, Brunetti \ea 2001, Petrosian 2001). 
395: Reliable estimates of the field are therefore quite essential.
396: 
397: Differences between $B_{rx}$ and $B_{fr}$ could, however, be due to the 
398: fact that the former is a volume-weighted measure of the field, whereas 
399: the latter is an average along the line of sight, weighted by the 
400: electron density. In addition, the field and \rel electron density would 
401: generally have different spatial profiles that could lead to very 
402: different spatial averages (Goldshmidt \& Rephaeli 1993). Various 
403: statistical and physical uncertainties in the Faraday rotation 
404: measurements, and their impact on deduced values of IC fields, were 
405: investigated recently by Newman, Newman \& Rephaeli (2002); their work 
406: strengthens the conclusion that a simple comparison of values of 
407: $B_{rx}$ and $B_{fr}$ is meaningless. More importantly, Rudnick \& 
408: Blundell (2003) have recently shown very clearly that the estimation 
409: of cluster fields from Faraday rotation measurements is very uncertain 
410: due to the inclusion in the sample of {\it cluster} radio sources whose 
411: large contributions to the rotation measures originate from their 
412: {\it intrinsic} fields, not the cluster-wide fields that they were 
413: presumed to sample.
414: 
415: RXTE and BeppoSAX measurements yielded evidence for NT X-ray 
416: emission in 5 clusters. It is important to continue the search 
417: for NT emission in other clusters with extended regions of radio 
418: emission. In particular, it is essential to obtain spatial information on 
419: this emission. This will likely be done for the first time by the IBIS 
420: instrument on the INTEGRAL satellite during a planned 500 ks observation 
421: of the Coma cluster. With the moderate $\sim 12'$ spatial resolution of 
422: IBIS, it should be feasible to determine the location of the region where 
423: the secondary emission is produced in this cluster.
424: 
425: \acknowledgments
426: We thank the referee, Dr. Roberto Fusco-Femiano, for his suggestions. 
427: This project has been supported by a NASA grant at UCSD.
428: 
429: \parskip=0.02in
430: \def\ref{\par\noindent\hangindent 20pt}
431: \noindent
432: 
433: \begin{references}
434: \ref{Bridle, A.H., \& Fomalont, E.B. 1976, A\&A, 52, 107}
435: \ref{Bridle, A.H., \ea 1979, A\&A, 80, 201}
436: \ref{Brunetti, G., \ea 2001, MN, 320, 365}
437: \ref{Carilli, C.L., \& Taylor, G.B. 2002, ARAA, in press 
438: (astro-ph/0110655)}
439: \ref{Clarke, T.E., Ensslin, T.A. 2001, Proceedings of XXI Moriond 
440: Meeting, astro-ph/0106137}
441: \ref{Clarke, T.E., Kronberg,  P.P., \& B\"ohringer, H. 2001, ApJ, 547, 
442: L111}
443: \ref{Ensslin T.A., \ea 1999, A\&A, 344, 409}
444: \ref{Fusco-Femiano, R., \ea 1999, ApJ, 513, L21}
445: \ref{Fusco-Femiano, R. \ea 2000, ApJL, 534, L7}
446: \ref{Fusco-Femiano, R. \ea 2003, Proceedings of `Matter and Energy in 
447:  Clusters of Galaxies', ASP, Conf. Ser., eds: S.Bowyer \& C-Y.Hwang, 
448:  astro-ph/0207241}
449: \ref{Giovannini, G., \ea 1993, ApJ, 406, 399}
450: \ref{Goldshmidt, O., \& Rephaeli, Y., 1993, ApJ, 411, 518}
451: \ref{Gruber, D.E. \ea 2001, ApJ 562, 499.}
452: \ref{Gruber, D.E., \& Rephaeli, Y. 2002, ApJ, 565, 877}
453: \ref{Henriksen, M., \ea 1999, ApJ, 511, 666}
454: \ref{Honda, H., \ea 1996, ApJ, 473, L71}
455: \ref{Lampton, M., Margon, B., \& Bowyer, S. 1976, ApJ, 208, 177}
456: \ref{Markevitch, M. \ea 1998, ApJ, 503, 77}
457: \ref{Newman, W.I., Newman, A.L., \& Rephaeli, Y. 2002, ApJ, 575, 755}
458: \ref{Petrosian, V. 2001, ApJ, 557, 560}
459: \ref{Rephaeli, Y. 1977, ApJ, 212, 608}
460: \ref{Rephaeli, Y. 1979, ApJ, 227, 364}
461: \ref{Rephaeli, Y. 2001, in `Astrophysical Sources of High Energy
462: Particles \& Radiation', NATO ASI, edited by Shapiro \ea, Kluwer, p.143}
463: \ref{Rephaeli, Y., \& Goldshmidt, O. 1992, ApJ, 397, 438}
464: \ref{Rephaeli, Y., \& Gruber, D.E. 1988, ApJ, 333, 133}
465: \ref{Rephaeli, Y., \& Gruber, D.E. 2002, ApJ, 579, 587}
466: \ref{Rephaeli, Y., Gruber, D.E., \& Blanco, P.R. 1999, ApJ, 511, L21}
467: \ref{Rephaeli, Y., Ulmer, M., \& Gruber, D.E. 1994, ApJ, 429, 554}
468: \ref{Rudnick, L, \& Blundell, K. 2003, ApJ, in press, astro-ph/0301260}
469: \ref{Sarazin, C.L. 1999, ApJ, 520, 529}
470: \ref{Sarazin, C.L., \& Kempner, J.C. 2000, ApJ, 533, 73}
471: \ref{Sun, M. 2002, ApJ, 565, 867}
472: \ref{Watanabe, M., \ea 1999, ApJ, 527, 80}
473: \ref{Wilms, J. \ea 1999, ApJ 522, 460}
474: 
475: \end{references}
476: \np
477: \begin{figure}  
478: \centerline{\psfig{file=fig1.ps, width=7cm, angle=270}}
479: \figcaption{Joint RXTE-ASCA (photon) spectrum of the A2256 with folded 
480: Raymond-Smith ($kT \simeq 7.7$) and power-law (index $=2.2$) models. 
481: ASCA data are shown in green and blue circles; crosses are PCA data, and 
482: HEXTE data points are marked with red circles (with 68\% error bars). 
483: The total fitted spectrum is shown with a histogram, while the lower 
484: histogram shows the power-law portion of the best fit. The quality of 
485: the fit is demonstrated in the lower panel, which displays the observed 
486: difference normalized to the standard error of the data point.} 
487: \end{figure} 
488: 
489: \end{document}
490: 
491: 
492: 
493: 
494: 
495: 
496: 
497: 
498: 
499: 
500: 
501: 
502: 
503: 
504: 
505: 
506: 
507: