astro-ph0003060/ms.tex
1: %  1.Title Page
2: %  2.Abstract & Subject Headings
3: %  3.Text & Acknowledgments
4: %  4.Appendices (if any)
5: %  5.References (new page)
6: %  6.Figure Captions (new page), including captions for any Plates
7: %  7.Tables (as part of main manuscript - new page)
8: 
9: %------------------------------- preamble
10: 
11: %\documentstyle[12pt,aasms4]{article}
12: \documentstyle[emulateapj,epsf]{article}
13: 
14: \newcommand\yr{\%~$\rm yr^{-1}$}
15: \def\ltsima{$\; \buildrel < \over \sim \;$}
16: \def\simlt{\lower.5ex\hbox{\ltsima}}
17: \def\gtsima{$\; \buildrel > \over \sim \;$}
18: \def\simgt{\lower.5ex\hbox{\gtsima}}
19:  
20: \received{}
21: \accepted{}
22: \journalid{}{}
23: \articleid{}{}
24: \slugcomment{to appear in ApJL}
25: \lefthead{}
26: \righthead{}
27: 
28: %------------------------------- title page
29: 
30: \begin{document}
31:    
32: \title{Interaction in Abell 2256: the BeppoSAX view}
33: \author{Silvano Molendi\altaffilmark{1}, Sabrina De Grandi\altaffilmark{2} \&
34: Roberto Fusco-Femiano\altaffilmark{3} 
35: }
36: \altaffiltext{1}{Istituto di Fisica Cosmica, CNR, via Bassini 15,
37: I-20133 Milano, Italy}
38: 
39: \altaffiltext{2}{Osservatorio Astronomico di Brera, via Bianchi 46,
40: I-23807 Merate (LC), Italy}
41: 
42: \altaffiltext{3}{Istituto di Astrofisica Spaziale, CNR,
43:    via del Fosso del Cavaliere,
44:    I-00133 Roma, Italy}
45: 
46: \begin{abstract}
47: We present results from a spatially resolved spectral 
48: analysis of the merging cluster Abell 2256. 
49: The long integration time (135 ks) and the good spatial
50: resolution of the MECS onboard BeppoSAX allow us to derive a 
51: new and substantially improved measurement of the temperature 
52: structure.
53: We find that, within a central region of the cluster,
54: where the effects of the merger are visible in the ROSAT surface 
55: brightness image, the azimuthally averaged projected temperature
56: is remarkably flat; outside this region the projected temperature 
57: rapidly declines.
58: The BeppoSAX data also shows clear evidence of an azimuthal 
59: temperature gradient in the 4$^\prime$-8$^\prime$ (0.4-0.8 Mpc) 
60: radial bin, oriented in the same direction as the merger itself.
61: Our metal abundance profile shows, for the first time, firm evidence
62: of an abundance gradient in a rich merging cluster. Intriguingly 
63: the abundance map shows, in the SE sector, i.e. the one furthest
64: away from the merger, a sharp factor of two drop in abundance
65: at a radius comparable to the core radius of the cluster.
66: A possible interpretation is that, prior to the merger event, 
67: a cooling flow had already developed in the core of 
68: the infalling subgroup, as suggested by Fabian and Daines (1991).
69: The interaction between the substructures would have disrupted the 
70: cooling flow thereby re-heating and re-mixing the gas. 
71: As the merger in A2256 is still in a relatively early 
72: stage, the gas located on the side opposite to the merger event 
73: would still retain the low abundances associated with the 
74: intra-cluster medium (ICM) prior to the cooling flow disruption.
75: 
76: \end{abstract}
77: 
78: \keywords{X-rays: galaxies --- Galaxies: clusters: individual 
79:           (Abell 2256)}
80: 
81: \section {Introduction}
82: 
83: Abell 2256 (hereafter A2256) 
84: is a rich, nearby  (z$=$ 0.057, Bothun \& Schombert 1990), cluster of 
85: galaxies.
86: Studies in the optical band have shown that the velocity dispersion is 
87: quite large ($\sim $1400 km/s; Fabricant, Kent \& Kurtz 1989, 
88: Bothun \& Schombert 1990). 
89: An early ROSAT PSPC image of A2256 (Briel et al. 1991) provided clear 
90: evidence of substructure, showing two emission peaks separated by about 
91: 3.5 arcminutes. One of the two peaks is coincident with the cD galaxy 
92: while the distorted morphology of the other indicates that it is merging 
93: with the main cluster. A reanalysis, by Briel et 
94: al. (1991), of the velocity distribution of the galaxies measured by  
95: Fabricant, Kent \& Kurtz (1989) shows that it can be separated into two
96: distinct distributions coincident with the two X-ray peaks.  
97: Fabian \& Daines (1991), from the ROSAT PSPC surface brightness distribution, 
98: have estimated cooling times of 2$\times 10^{10}$ years and 5$\times 10^{9}$ 
99: years at the center of the main cluster and of the infalling subcluster 
100: respectively. The above authors imply that, prior to the merger event, 
101: a cooling flow had already developed in the core of the infalling subgroup
102: and that the merger may have interrupted the cooling flow and stirred up the 
103: gas within it.
104: 
105: Various attempts have been made to measure the temperature structure
106: of A2256. Briel \& Henry (1994), using ROSAT PSPC data, find evidence 
107: that the infalling group has a lower temperature than the main peak.
108: They also find evidence 
109: of two hot spots opposite each other and perpendicular to 
110: the presumed infall direction of the subgroup, however this result 
111: was not  confirmed by Markevitch \& Vikhlinin (1997) who reanalyzed 
112: the same data. 
113: %the rest of the cluster is essentially isothermal with  
114: %a small indication of a temperature drop in the outskirts.
115: Markevitch (1996, hereafter M96), from ASCA data, finds evidence of a 
116: smoothly declining radial temperature profile, going from $\sim$8.7 keV near 
117: the core to $\sim$4 keV in the outskirts.
118: His temperature map shows that the subgroup has a smaller 
119: temperature than the main peak.  
120: Irwin, Bregman \& Evrard (1999), from ROSAT PSPC hardness ratios, 
121: find a  radial profile consistent with a constant
122: temperature out to 15$^{\prime}$ from the cluster core. 
123: Their hardness ratio two dimensional map is in general agreement with 
124: the one of Briel \& Henry (1994).
125: White (1999, hereafter W99), from a  reanalysis of the ASCA data,
126: finds a radial temperature profile consistent with being 
127: constant out to 18$^{\prime}$ from the cluster core. 
128: 
129: In this Letter we report BeppoSAX observations of A2256.
130: We use our data to perform an independent measurement of the 
131: temperature profile and two-dimensional map of A2256.
132: We also present the  abundance profile and the first abundance map of 
133: A2256. 
134: The outline of the Letter is as follows.
135: In section 2 we give some information on the BeppoSAX observation
136: of A2256 and on the data preparation.
137: In section 3  we present spatially resolved measurements
138: of the temperature and metal abundance.
139: In section 4 we discuss our results and compare them to previous
140: findings.
141: %In section 5 we summarize our main results.
142: Throughout this Letter we assume H$_{o}$=50 km s$^{-1}$Mpc$^{-1}$
143: and q$_{o}$=0.5.
144:  
145: \section {Observation and Data Preparation}
146: The cluster A2256  was observed by the BeppoSAX 
147: satellite (Boella et al. 1997a) at two different epochs;
148: between the 11$^{th}$ and the 12$^{th}$ of February 1998
149: and between the 25$^{th}$ and the 26$^{st}$ of February 1999.
150: We will discuss here data from the
151: MECS instrument onboard BeppoSAX; a joint analysis of the 
152: MECS and PDS spectra of A2256 is presented in Fusco-Femiano
153: et al. (2000).
154: The MECS (Boella et al. 1997b) is presently composed 
155: of two units working in the 1--10 keV 
156: energy range. At 6~keV, the energy resolution is  $\sim$8\%  and the 
157: angular resolution 
158: is $\sim$0.7$^{\prime}$ (FWHM). 
159: Standard reduction procedures and screening criteria have been
160: adopted to produce linearized and equalized event files.
161: Data preparation and linearization was performed using 
162: the {\sc Saxdas} package under {\sc Ftools} environment.
163: The total effective exposure time for the two observation was
164: 1.3$\times$10$^5$ s. 
165: All spectral fits have been performed using XSPEC Ver. 10.00.
166: Quoted confidence 
167: intervals are 68\% for 1 interesting parameter (i.e. $\Delta \chi^2 =1$), 
168: unless otherwise stated.
169: 
170: 
171: \section{Spatially Resolved Spectral Analysis} 
172: Spectral distortions introduced by the energy dependent PSF
173: must be accounted for when performing spatially resolved
174: spectroscopy of galaxy clusters.
175: As for the analysis of other BeppoSAX observations of clusters
176: (e.g. A2319, Molendi et al. 1999), we
177: have taken them into account using the {\sc Effarea} program
178: publicly available within the latest {\sc Saxdas} release.
179: We remark that we fit spectra individually. This is not what 
180: is typically done when performing spatially resolved spectroscopy 
181: of clusters with ASCA data.  Here spectra accumulated from 
182: different regions are typically analyzed simultaneously, 
183: the reason being that the correction to be applied to a 
184: given region depends on the temperature of all the others.
185: The lack of a strong dependence of the MECS PSF on energy allows
186: us to avoid such complications.
187: 
188: \subsection{Radial Profiles}
189: For each of the two observations
190: we have accumulated spectra from 6 annular regions 
191: centered on the main X-ray emission peak of A2256, 
192: with inner and outer radii of 0$^{\prime}$-2$^{\prime}$, 
193: 2$^{\prime}$-4$^{\prime}$, 4$^{\prime}$-6$^{\prime}$, 
194: 6$^{\prime}$-8$^{\prime}$, 8$^{\prime}$-12$^{\prime}$ and  
195: 12$^{\prime}$-16$^{\prime}$.
196: We have also accumulated a global spectrum from a circle
197: with radius 16$^{\prime}$.
198: The background subtraction has been performed
199: using spectra extracted from blank sky event files in the same region
200: of the detector as the source.
201: A correction for the absorption caused by the strongback supporting
202: the detector window has been applied for  the  
203: 8$^{\prime}$-12$^{\prime}$ annulus, where the annular part of 
204: the strongback is contained. For the 6$^{\prime}$-8$^{\prime}$ and 
205: 12$^{\prime}$-16$^{\prime}$ annuli, where the strongback
206: covers only a small fraction of the available area, we have chosen
207: to exclude the regions shadowed by the strongback.
208: For the 5 innermost annuli the energy range considered 
209: for spectral fitting was 2-10 keV; 
210: %for the 8$^{\prime}$-12$^{\prime}$ 
211: %annulus the fit was restricted to the 2.7-10 keV band to limit
212: %spectral distortions associated to the degradation of the detector PSF 
213: %at low energies in the region covered by the strongback;
214: for the outermost annulus,
215: the fit was restricted to the  2-8 keV energy range 
216: to limit spectral distortions which could be caused by an incorrect 
217: background subtraction (see De Grandi \& Molendi 1999a for details).
218: Source and background spectra accumulated for each of the 
219: two observations have then been summed together. 
220: 
221:  
222: We have fitted each spectrum with a MEKAL model absorbed 
223: by the Galactic line of sight equivalent hydrogen 
224: column density, $N_H$, of 4.1$\times 10^{20}$ cm$^{-2}$. 
225: The temperature and abundance we derive from the global
226: spectrum are respectively 7.5$\pm 0.1$ keV and 0.25$\pm 0.02$, 
227: solar units.
228: In figure 1  we show the temperature and abundance 
229: profiles obtained from our six annular regions.  
230: %By fitting the temperature and abundance profiles with a 
231: %constant we derive the following average values: 
232: %$6.9\pm$0.1 keV and 0.25$\pm$0.02, solar units.
233: A constant does not provide a good fit to the temperature or the
234: abundance profile (see table 1).
235: %Using the $\chi^2$ statistics we find: $\chi^2 =$ 31.5 for 5 d.o.f., 
236: %corresponding to a probability of 7$\times 10^{-6}$ for the observed 
237: %distribution to be drawn from a constant parent distribution.
238: %The temperature profile is remarkably flat out to 
239: %8$^{\prime}$ from the cluster core and  decreases by a factor
240: %of almost 2 in the subsequent 8$^{\prime}$. 
241: %A constant provides a poor fit to the 
242: %abundance profile, $\chi^2 =$ 13.86 for 5 d.o.f. (Prob.$=$0.02). 
243: %the abundance profile appears to be substantially flat, 
244: %Ab $\sim 0.28$, solar units, out to 12$^{\prime}$ from the cluster 
245: %core, while in the outermost bin it is consistent with being 0.
246:    
247: As in Molendi et al. (1999), we have used the Fe K$_{\alpha}$ line 
248: as an independent estimator of the ICM temperature. 
249: Considering the limited number of counts available in the line, we 
250: have performed the analysis on 2 annuli with bounding radii,
251: 0$^{\prime}$-8$^{\prime}$ and 8$^{\prime}$-12$^{\prime}$, the very small 
252: Fe abundance measured in the  12$^{\prime}$-16$^{\prime}$ annulus
253: prevents us from deriving a reliable line centroid for this region.
254: We have fitted each spectrum with a bremsstrahlung model plus a 
255: line, both at a redshift of z=0.057 (ZBREMSS and ZGAUSS models 
256: in XSPEC), absorbed by the galactic  $N_{H}$.
257: A systematic negative shift of 40 eV has been included in the
258: centroid energy to account for a slight misscalibration of 
259: the energy pulseheight-channel relationship near the Fe line.  
260: To convert the energy centroid  into a temperature 
261: we have derived an energy centroid vs. temperature relationship. 
262: This has been done by simulating thermal spectra, using the MEKAL 
263: model and the MECS response matrix, and fitting them with the same 
264: model, which has been used to fit the real data.
265: We derive a temperature of 8.0$^{+0.9}_{-1.0}$ keV for the inner
266: radial bin and of 3.2 $^{+2.8}_{-1.7}$ keV for the outer one.
267: Thus, our two independent measurements of the temperature 
268: profile are in good agreement with each other. 
269: 
270: %In figure 1 we have overlaid the temperatures derived from the 
271: %centroid analysis on those previously obtained through the 
272: %thermal continuum fitting. The two measurements of the temperature 
273: %profile are in good agreement with each other. 
274: 
275: \subsection{Maps}
276: 
277: As shown in figure 2, we have divided the MECS image of A2256 into 
278: 4 sectors: NW, SW, SE and NE, each sector has been divided into 4 
279: annuli with bounding radii,
280: 2$^{\prime}$-4$^{\prime}$, 4$^{\prime}$-8$^{\prime}$, 
281: 8$^{\prime}$-12$^{\prime}$ and 12$^{\prime}$-16$^{\prime}$. 
282: The background subtraction has been performed
283: using spectra extracted from blank sky event files in the same region
284: of the detector as the source.
285: Correction or exclusion of the regions shadowed by the 
286: strongback supporting the detector window have been 
287: performed as in the previous subsection.
288: The energy ranges and the spectral models adopted for fitting are the same 
289: used for the azimuthally averaged spectra.
290: 
291: In figures 3 and 4 we show respectively  the temperature and abundance 
292: profiles obtained from the spectral fits for each of the 4 sectors.
293: In table 1 we report the best fitting constant temperatures and 
294: abundances for the profiles shown in figures 3 and 4.
295: Note that in all the profiles we have included the  
296: measure obtained for the central circular region with 
297: radius 2$^{\prime}$. 
298: %Fitting each radial profile with a constant temperature we derive 
299: %the following average sector temperatures: 
300: %6.6$\pm$0.2 keV for the NW sector, 
301: %7.1$\pm$0.1 keV for the SW sector, 
302: %7.2$\pm$0.2 keV for the SE sector and 
303: %6.8$\pm$0.2 keV for the NE sector.
304: All sectors, except for the SW sector, show a statistically significant 
305: temperature decrease with increasing radius.
306: %From the $\chi^2$ statistics we find:
307: %$\chi^2 =11.4$ for 4 d.o.f. (Prob.$= 2.2\times 10^{-2}$) for the NW sector, 
308: %$\chi^2 =4.6$  for 4 d.o.f. (Prob.$= 3.3\times 10^{-1}$) for the SW sector, 
309: %$\chi^2 =11.0$ for 4 d.o.f. (Prob.$= 2.7\times 10^{-2}$) for the SE sector and 
310: %$\chi^2 =46.9$ for 4 d.o.f. (Prob.$= 1.6\times 10^{-9}$) for the NE sector. 
311: In the NW sector
312: the temperature decreases continuously as the distance from the cluster
313: center increases. In the SE and NE sectors the temperature first increases, 
314: reaching a maximum in either the second (NE sector) or third (SE sector) 
315: annulus, and then decreases. 
316: Interestingly, a fit to the temperatures of the  4 sectors in the third 
317: annulus (bounding radii 4$^\prime$-8$^\prime$) with a constant, yields
318: $\chi^2=19.2$ for 3 d.o.f., with an associated probability for 
319: the temperature to be constant of $2.5\times 10^{-3}$, indicating that 
320: an azimuthal temperature gradient is present near the core of the cluster.
321: More specifically the NW sector of the cluster is  
322: the coldest, 6$\pm 0.3$ keV, and the SE sector the hottest, 8.4$\pm 0.5$ keV.
323: %
324: %In figure 4 we show the abundance profiles for each of the 4 sectors.
325: %In table 1 we report the best fitting constant temperature for 
326: %the profiles shown in figure 3.
327: %In all profiles we have included the  
328: %abundance measure obtained for the central circular region with 
329: %bounding radius 2$^{\prime}$. 
330: %Fitting each radial profile with a constant abundance we derive 
331: %the following sector averaged abundances: 
332: %0.30$\pm$0.03 for the NW sector, 
333: %0.27$\pm$0.03 for the SW sector, 
334: %0.21$\pm$0.03 for the SE sector and 
335: %0.30$\pm$0.03 for the NE sector.
336: %The $\chi^2$ values and probabilities associated to the above fits are:
337: %$\chi^2 =3.77$ for 4 d.o.f. (Prob.$= 0.44             $) for the NW sector, 
338: %$\chi^2 =8.02$ for 4 d.o.f. (Prob.$= 0.09             $) for the SW sector, 
339: %$\chi^2 =29.1$ for 4 d.o.f. (Prob.$= 7.6\times 10^{-6}$) for the SE sector and 
340: %$\chi^2 =2.01$ for 4 d.o.f. (Prob.$= 0.73             $) for the NE sector. 
341: The SE sector is the only one to show clear evidence of an abundance decline 
342: with increasing radius, all other sectors have abundance profiles which
343: are consistent with being flat.
344: 
345: \begin{table*}
346: \caption{Best fit constant values for the temperature and
347: the abundance for the azimuthally averaged radial
348: profiles (All) and for the profiles of the 4 sectors (NW, SW, SE and NE)}
349: \centerline{
350: \begin{tabular}{lccccccc}
351: \hline
352: Sector &    kT       & $\chi^2$ & Prob.$^a$         & Abundance   & $\chi^2$ & Prob.$^a$         &d.o.f. \cr
353:        &  (keV)      &          &                   &(Solar Units)&          &                   &    \cr
354: \hline
355: \hline
356: All    & 6.9$\pm$0.1 &   36.4   &$7.9\times 10^{-7}$& 0.25$\pm$0.02&   13.9  &$1.6\times 10^{-2}$& 5  \cr  
357: \hline
358: NW     & 6.6$\pm$0.2 &   11.3   &$2.4\times 10^{-2}$& 0.30$\pm$0.03&    3.8  & 0.43              & 4  \cr
359: SW     & 7.0$\pm$0.2 &    7.4   & 0.12              & 0.27$\pm$0.03&    8.1  &$9.0\times 10^{-2}$& 4  \cr
360: SE     & 7.2$\pm$0.2 &   11.3   &$2.4\times 10^{-2}$& 0.21$\pm$0.03&   29.0  &$7.8\times 10^{-6}$& 4  \cr
361: NE     & 6.7$\pm$0.2 &   52.2   &$1.3\times10^{-10}$& 0.30$\pm$0.03&    2.0  & 0.74              & 4  \cr
362: \hline
363: \end{tabular}
364: }
365: \par
366: $^a$Probability for  the observed distribution to be drawn from a constant parent distribution.
367: \end{table*}
368: 
369: 
370: 
371: 
372: \section{Discussion}
373: 
374: Previous measurements of the temperature structure of A2256
375: have been performed by Briel \& Henry (1994) and Irwin et al. (1999),
376: using ROSAT data, by  M96 and by W99 using ASCA data. 
377: We have performed a detailed comparison of our radial temperature 
378: profile with the ones based on the ASCA satellite (M96 and W99) 
379: which covers an energy  range similar to ours.
380: %The results of Irwin et al. (1999) cannot be readily compared with 
381: %ours as they are presented in the form of hardness ratios.
382: %As far as the profile presented by Briel \& Henry (1994) is concerned it
383: %is in broad agreement with ours.
384: In figure 1 we have overlaid the temperature profile obtained by  
385: M96 and by W99 on our profile.
386: The higher quality of the BeppoSAX measurement, due in part
387: to the much longer exposure time and in part to the better
388: angular resolution of our instrument, is quite evident.
389: The innermost bin in the M96 profile, 0$^\prime$-6$^\prime$ has
390: a temperature that is inconsistent with the temperature 
391: we measure from our three innermost bins spanning the same radial range.
392: It must be noted that, while our profile is azimuthally averaged over 
393: all angles, the M96 measurement has been obtained excluding the region 
394: presumably contaminated by the softer emission of the infalling group.
395: To obtain a direct comparison between our measurement and the one reported
396: in M96, we have derived the temperature from a circular region with 
397: radius 6$^\prime$ excluding the NW sector containing the infalling 
398: group. Our measurement, 7.5$\pm$0.2 keV, although somewhat higher than the 
399: one obtained by averaging over all directions, is still incompatible at more
400: than the 3$\sigma$ level with the one reported by M96.
401: The second radial bin reported in M96
402:  (6$^\prime$-11$^\prime$) is characterized by a temperature 
403: apparently larger than the mean temperature for our corresponding bins 
404: (i.e.  6$^\prime$-8$^\prime$ and 8$^\prime$-12$^\prime$).
405: However this difference is only apparent, indeed if we simultaneously
406: fit the BeppoSAX spectra for the 6$^\prime$-8$^\prime$ and 
407: 8$^\prime$-12$^\prime$ bins, which is equivalent to fitting data
408: from the 6$^\prime$-12$^\prime$ bin, we derive a temperature of 
409: 6.6$\pm$0.3 keV, which is consistent with the one derived by M96.
410: The temperature for the outermost bin in the M96 profile is in  
411: agreement with our own measurement. 
412: The  W99 measurement, which comes from a
413: different analysis of the same ASCA observation used by M96,
414: is in agreement with ours for radii smaller than 6$^\prime$.
415: % the mean temperature
416: %of the W99 profile, within this region, is $6.9\pm 0.5$ keV. Interestingly, 
417: %this value differs at more than the 3$\sigma$ level from the one derived
418: %by M96 for the same region.  Thereby showing how a different treatment 
419: %of the same ASCA observation may lead to significant different 
420: %temperatures. 
421: The outermost bin reported in W99 appears to have a temperature 
422: substantially larger than the mean temperature for our corresponding bins 
423: (i.e.  6$^\prime$-8$^\prime$, 8$^\prime$-12$^\prime$ and 
424: 12$^\prime$-16$^\prime$). This difference is only apparent, if we 
425: simultaneously fit the BeppoSAX spectra for the 6$^\prime$-8$^\prime$, 
426: the 8$^\prime$-12$^\prime$ and the 12$^\prime$-16$^\prime$ bins we derive 
427: a temperature of 6.5$\pm$0.3,
428: which is consistent with the one derived by W99.
429: The apparent difference is related to the strong gradient in the surface 
430: brightness profile when going from 6$^\prime$ to 16$^\prime$, which causes 
431: the emission from the entire region to be dominated 
432: by the contribution of the innermost annuli.
433: In summary: for radii larger than 6$^\prime$, our profile is in agreement 
434: with the M96 and W99 profile, while for  radii smaller than 6$^\prime$,
435: our profile is in agreement with the  W99 profile and
436: in disagreement with the M96 profile.
437: 
438: The most striking feature of our radial temperature profile is the 
439: presence of  a relatively localized gradient. The temperature is 
440: flat out to 8$^\prime$ and decreases by almost a factor
441: two within the following 8$^\prime$.
442: The radius at which the temperature starts to decline
443: $\sim 8^\prime$ (0.8 Mpc) is comparable to the radius at which the 
444: X-ray isophotes are no longer disturbed by the interaction of the two
445: subclusters, which is clearly seen at smaller scales in the ROSAT PSPC 
446: image (e.g. figure 2 of Briel et al. 1991).
447: Thus the presence of a hot almost isothermal region in the core is most 
448: likely related  to the on-going merger between the main cluster and the 
449: group. 
450: The BeppoSAX temperature map shows clear evidence of an azimuthal gradient 
451: in the 8$^\prime$-12$^\prime$ radial bin. The NW sector is found to be the 
452: coldest while the SE sector appears to be the hottest, thus the 
453: gradient appears to be oriented in the same direction as the merger itself. 
454: Interstingly, in a previous work (De Grandi \& Molendi 1999a), the merging 
455: cluster A3266 was found to have a similar temperature structure.
456: No evidence of the two hot spots reported by Briel \& Henry (1994)
457: is found in our map. 
458: %In summary the interaction between the main cluster and the 
459: %infalling group has resulted in a hot almost isothermal central
460: %region of 8$^\prime$ (0.8 Mpc) radius, and in a temperature 
461: %gradient, on similar scales, oriented in the direction of the 
462: %on-going merger.
463:  
464: The metal abundance in A2256 appears to decrease with increasing radius
465: (see figure 1). 
466: This is the first firm case, to our knowledge, of an abundance 
467: gradient in a rich non cooling flow cluster.
468: Evidence of an abundance gradient has been found in the poor cluster 
469: MKW4 (Finoguenov et al. 1999), while marginal evidence has been
470: found in A399 (Fujita et al. 1996) and A1060 (Finoguenov et al.
471: 1999). 
472: In A2256 the abundance averaged over a central region of 0.2 Mpc 
473: radius is 
474: $\sim$ 0.3, solar units, a  value which, although higher than the average 
475: abundance for non cooling flow cluster, $\sim$ 0.20 (Allen \& Fabian 1998),
476: is smaller than those commonly observed in the core of cooling flow cluster 
477: (see for example Finoguenov et al. 1999, for an
478: analysis of abundance profiles from ASCA data, and our own BeppoSAX 
479: results on Abell 2029, Molendi \& De Grandi 1999, and PKS 0745-191,
480: De Grandi \& Molendi 1999b).
481: Furthermore, the abundance map (see figure 4) shows that the SE sector, 
482: i.e. the one furthest away from the on-going merger, presents a highly 
483: significant abundance decline  (probability $= 7.6\times 10^{-6}$)
484: localized at a radius comparable to the core radius of the cluster.
485: A possible interpretation is that, prior to the merger event, 
486: a cooling flow had already developed in the core
487: of the infalling subgroup, as suggested by Fabian \& Daines (1991). 
488: The above authors, from the gas densities
489: at the center of the main cluster and of the infalling subcluster
490: compute cooling times of 2$\times 10^{10}$ years and 
491: 5$\times 10^{9}$ years respectively, implying that the 
492: infalling subcluster must have had a cooling flow. 
493: The interaction between the substructures would have disrupted the 
494: cooling flow thereby re-heating and re-mixing the gas. 
495: As the merger in A2256 is still in a relatively early 
496: stage, the gas located on the side opposite to the merger event 
497: may still retain the low abundances associated to the 
498: ICM prior to the cooling flow disruption.
499: It seems unlikely that a contribution to the metallicity 
500: enhancement has come from the main cluster as its core 
501: density implies a cooling time that is larger than the age of the
502: Universe. 
503: Finally we speculate that other rich merging clusters,
504: similar to A2256, may present metallicity gradients produced 
505: by disrupted cooling flows. BeppoSAX and future XMM
506: observations of merging clusters will certainly contribute in clarifying 
507: this issue.
508: 
509: 
510: %\section{Summary}
511: 
512: \acknowledgments
513: We acknowledge support from the BeppoSAX Science Data Center.
514: We thank the referee for useful comments and D. Lazzati for 
515: help in producing the contour plot.  
516: 
517: %--------------------------------- references
518: 
519: \begin{thebibliography}{}
520: 
521: %   \bibitem{} Baker N., 1966,
522: %      in: Stellar Evolution,
523: %      eds.\ R. F. Stein, A. G. W. Cameron,
524: %      Plenum, New York, p.\ 333
525: 
526: \bibitem{} Allen, S., W., \& Fabian, A., C. 1998, MNRAS, 297, L63 
527: %\bibitem{} Arnaud, M., \& Raymond, J. 1992, ApJ, 398, 394
528: \bibitem{} Boella, G., Butler, R. C., Perola, G. C., Piro, L., Scarsi, L., 
529: \& Bleeker, J. A. M. 1997a, A\&AS, 122, 299 
530: \bibitem{} Boella, G., et al. 1997b, A\&AS, 122, 327 
531: \bibitem{} Bothun, G., D.,\& Schombert J., M. 1990, 360, 436
532: \bibitem{} Briel, U. G., Henry, J. P., Schwarz, R. A., et al. 1991, A\&A, 246, L10
533: \bibitem{} Briel, U. G., \& Henry, J. P. 1994, Nature, 372, 439 
534: %\bibitem{} Chiappetti, L., et al. 1998
535: %{\it http://sax.ifctr.mi.cnr.it/Sax/Mecs/tour.html}
536: %\bibitem{} D'Acri, F., De Grandi, S., \& Molendi S. 1998, Nuclear Physics,
537: %69/1-3, 581 (astro-ph/9802070)
538: \bibitem{} David, L. P., Slyz, A., Jones, C., Forman, W., Vrtilek, S. D., 
539: \& Arnaud, K. A. 1993, ApJ, 412, 479
540: \bibitem{} Dickey, J. M., \& Lockman, F., J. 1990, ARA\&A, 28, 215 
541: \bibitem{} De Grandi, S., Molendi, S. 1999a, ApJ, 527, L25
542: \bibitem{} De Grandi, S., Molendi, S. 1999b, A\&A, 351, L45
543: %\bibitem{} Dupke, R. A. \& Raymond 1999 ApJ submitted (astro-ph/9902112) 
544: %\bibitem{} Ezawa, H., Fukazawa, Y., Makishima, K., Ohashi, T., Takahara, F.,
545: \bibitem{} Fabian, A. C., \& Daines, S. J. 1991, MNRAS, 252, P17
546: \bibitem{} Fabricant, D., G., Kent, S. M., \& Kurtz, M. J. 1989, ApJ, 336, 77
547: \bibitem{} Finoguenov, A., David, L. P., Ponman, T. J. 1999, ApJ submitted 
548: (astro-ph/9908150)
549: \bibitem{} Fujita, Y.; Koyama, K.; Tsuru, T.; Matsumoto, H. 1996, PASJ, 48, 191
550: \bibitem{} Fusco-Femiano, R. et al. 2000, in preparation 
551: %\bibitem{} Fukazawa Y. et al. 1994, PASJL, 46, 55
552: %\bibitem{} Fukazawa, Y. et al. 1998, PASJ, 50, 187
553: %\bibitem{} Fusco-Femiano, R., Dal Fiume, D., Feretti, L., Giovannini, G.,
554: % Grandi, P., Matt, G., Molendi, S., \& Santangelo, A. 1999 ApJL, 513, 21
555: \bibitem{} Irwin, J. A., Bregman J. N., \& Evrard A. E. 1999, ApJ, 519, 518 
556: %\bibitem{} Lockman, F. J., Savage, Blair D. 1995, ApJS, 97, 1
557: \bibitem{} Markevitch, M. 1996, ApJL, 465, 1
558: \bibitem{} Markevitch, M., \& Vikhlinin, A. 1997, ApJ, 474, 84 
559: %\bibitem{} Mohr, J. J., Fabricant, D.G., \& Geller, M.J. 1993, ApJ, 413, 492
560: %\bibitem{} Mohr, J. J., Mathiesen, B., \& Evrard, A.E. 1999, ApJ, 517, 627
561: \bibitem{} Molendi, S., De Grandi, S., Fusco-Femiano, R. et al. 1999 ApJ, 525, L73 
562: \bibitem{} Molendi, S., \& De Grandi, S. 1999, A\&A, 351, L41
563: %\bibitem{} Mushotzky, R. F. 1984, Phys Scripta T7, 157
564: %\bibitem{} Molendi, S., Matt, G., Antonelli, L. A., et al. 1998, ApJ, 499, 608
565: %\bibitem{} Oegerle, W. R., Hill, J. M., \& Fitchett, M. J. 1995, AJ, 110, 32
566: %\bibitem{} Peres, C. B., Fabian, A. C., Edge, A. C., Allen, S. W., 
567: %Johnstone, R. M., \& White, D. A. 1998, MNRAS, 298, 416
568: %\bibitem{} Schindler, S., \&  M\"uller, E. 1993, A\&A, 272, 137
569: %\bibitem{} Sarazin, C. L.  1988, X-ray emission from clusters of
570: %           galaxies (Cambridge: Cambridge University Press) 
571: \end{thebibliography}
572: 
573: \clearpage
574: 
575: %---------------------------------- figure captions
576: 
577: {\begin{figure}
578: \epsfxsize=\textwidth
579: %\vskip 3cm
580: \epsffile{fig1_co.ps}
581: \vskip -11cm
582: \figcaption
583: {{\bf Top Panel}: Projected radial temperature profile for A2256.
584: Filled circles indicate temperatures obtained from BeppoSAX MECS data.
585: %Open circles indicate temperatures estimated
586: %by the position of the centroid of the Fe K$_{\alpha}$ line. 
587: Filled and open triangles represent the temperature profile derived 
588: from ASCA data, respectively by M96 and W99.
589: All the uncertainties on the
590: temperature measurements are at the 68$\%$ confidence level (we have
591: converted the 90$\%$ confidence errors reported in figure 4 of
592: Markevitch et al. 1998 into 68$\%$ confidence errors by dividing
593: them by 1.65, for further details see Markevitch \& Vikhlinin 1997).  
594: {\bf Bottom Panel}: projected radial abundance profile from BeppoSAX MECS
595: data. 
596: }
597: \end{figure}}
598: \clearpage
599: 
600: \begin{figure}
601: \vskip -6cm
602: \plotone{fig2_bw.ps}
603: \vskip 2cm
604: %\figcaption
605: \caption
606: {BeppoSAX MECS image of A2256. Contour levels are indicated
607: by the thick lines. The thin lines show how the cluster
608: has been divided to obtain temperature and abundance maps.
609: }
610: \end{figure}
611: \clearpage
612: 
613: \begin{figure}
614: \vskip -1cm
615: \plotone{fig3_co.ps}
616: \vskip -1cm
617: %\figcaption
618: \caption
619: {Radial temperature profiles for the NW sector (first panel), the SW sector
620: (second panel), the SE sector (third panel) and the NE sector (forth
621: panel).
622: The temperature for the leftmost bin is derived from the entire circle,
623: rather than from each sector. The dashed lines indicate the average 
624: temperature derived from the MECS temperature profile reported in 
625: figure 1.
626: }
627: \end{figure}
628: \clearpage
629: 
630: \begin{figure}
631: \vskip -1cm
632: \plotone{fig4_co.ps}
633: \vskip -1cm
634: %\figcaption
635: \caption
636: {Radial abundance profiles for the NW sector (first panel), the SW sector
637: (second panel), the SE sector (third panel) and the NE sector (forth
638: panel).
639: The abundance for the leftmost bin is derived from the entire circle,
640: rather than from each sector. 
641: The dashed lines indicate the average 
642: abundance derived from the profile reported in 
643: figure 1.
644: The abundance measure for the 8$^\prime$-12$^\prime$ bin in the NE
645: sector is 0.89$^{+1.40}_{-0.64}$. 
646: }
647: \end{figure}
648: 
649: %\clearpage
650: 
651: \end{document}
652: