1: %\documentclass[12pt,preprint]{aastex}

2: %\usepackage{emulateapj5,epsfig,graphicx,ifthen}

3: \documentclass[apj]{emulateapj}

4: \usepackage{natbib,apjfonts}

5: 

6: \def\hh{^{\rm h}}

7: \def\mm{^{\rm m}}

8: \def\ss{^{\rm s}}

9: 

10: %%%%%%%%%%%%%%%%%%%  PAPER BEGINS HERE  %%%%%%%%%%%%%%%%%%%%%%%%%

11: \shortauthors{KEMPNER \& DAVID}

12: \shorttitle{CHANDRA STUDY OF THE CORE OF A576}

13: 

14: \slugcomment{Accepted for publication in the Astrophysical Journal.}

15: 

16: \begin{document}

17: 

18: \title{A {\it Chandra} Study of the Core of the Nearby Cluster Abell 576}

19: 

20: \author{Joshua C. Kempner and Laurence P. David}

21: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St.,

22: Cambridge, MA 02138}

23: \email{jkempner@head-cfa.harvard.edu,

24: ldavid@head-cfa.harvard.edu}

25: 

26: \begin{abstract}

27: We present data from a {\it Chandra} observation of the nearby cluster of

28: galaxies Abell 576.  The core of the cluster shows a significant departure

29: from dynamical equilibrium.  We show that this core gas is most likely the

30: remnant of a merging subcluster, which has been stripped of much of its

31: gas, depositing a stream of gas behind it in the main cluster.  The

32: unstripped remnant of the subcluster is characterized by a different

33: temperature, density and metalicity than that of the surrounding main

34: cluster, suggesting its distinct origin.  Continual dissipation of the

35: kinetic energy of this minor merger may be sufficient to counteract most

36: cooling in the main cluster over the lifetime of the merger event.

37: \end{abstract}

38: 

39: \keywords{

40: cooling flows ---

41: galaxies: clusters: individual (Abell 576) ---

42: intergalactic medium ---

43: X-rays: galaxies: clusters

44: }

45: 

46: \section{Introduction}

47: \label{sec:intro}

48: 

49: The study of the cores of clusters of galaxies has undergone a renaissance

50: in the past few years with the launch of the {\it Chandra} and {\it

51: XMM-Newton} observatories.  While previous observatories lacked the spatial

52: resolution necessary to resolve structure within the cores of clusters,

53: this new generation of telescopes has revealed an astonishing level of

54: complexity in the structure of the intracluster medium (ICM).  Many

55: clusters previously thought to be relaxed, regular systems have proven to

56: be far from dynamical equilibrium, particularly in their cores

57: \citep*[e.g.][]{mvm01,mem03}.  Abell 576 is no exception.  Earlier data,

58: especially optical spectra of the cluster's galaxy population, hinted at

59: dynamical complexity in the cluster core \citep{mgf+96}, but earlier X-ray

60: observations showed the cluster to be quite regular and even cooler in the

61: center, suggesting either a cooling flow or multi-phase gas with a very

62: cool component \citep{rvm+84,mgf+96}.

63: 

64: With its low redshift, Abell 576 makes excellent use of the capabilities of

65: {\it Chandra}, allowing us to examine in detail the very core of the

66: cluster.  In this paper, we focus on the dynamical activity in the core of

67: cluster.  The cluster shows strong evidence, first suggested by

68: \citet{mgf+96} from an analysis of the galaxy population, of the remnant

69: core of a small merged subcluster.  We demonstrate that the X-ray data are

70: consistent with this picture, and even suggest it as the most likely origin

71: for the non-equilibrium gas at the center of Abell 576.  In fact, the

72: subcluster may still be in the process of settling into the center of the

73: main cluster's potential.

74: 

75: Throughout this paper, we use the cosmological parameters derived from the

76: first release WMAP results \citep{bhh+03}, so $1\arcsec=0.738$ kpc at

77: $z=0.0377$.  All errors are quoted at 90\% confidence unless otherwise

78: stated.

79: 

80: \section{Observation and Data Reduction}

81: \label{sec:obs}

82: 

83: The data were obtained during {\it Chandra} Cycle 3 in a single exposure of

84: 38.6 kiloseconds.  The focus was set on the back-illuminated S3 chip,

85: although significant flux from the cluster is detected on the adjacent

86: front-illuminated S2 chip.  The standard background reduction for Very

87: Faint (VF) mode data was applied using the CIAO tool acis\_process\_events.

88: In addition, the data were corrected for charge transfer inefficiency (CTI)

89: using the CXC/MIT CTI-corrector in CIAO version

90: 2.3\footnote{http://asc.harvard.edu/ciao/}, although this only corrects the

91: front-illuminated chips such as the S2.  They were then filtered on the

92: standard {\it ASCA} grades 0, 2, 3, 4, and 6.  The data were processed

93: using version 2.18 of the {\it Chandra} CALDB.  CTI-corrected blank sky

94: background files\footnote{http://hea-www.harvard.edu/$\sim$maxim/axaf/acisbg/}

95: provided by Maxim Markevitch were used for the background correction.  The

96: VF mode background reduction was also applied to the background files.

97: 

98: The final 10.7 ksec of the observation was contaminated by a background

99: flare.  The flare was of the hard spectrum variety, which affects both the

100: front- and back-illuminated chips.  Unlike the soft flares that {\it

101: Chandra} sometimes experiences, which affect only the back-illuminated

102: chips, the spectra of these hard flares are not consistent from flare to

103: flare, and therefore cannot be modeled \citep{mar02}.  We therefore excised

104: the contaminated data, and only consider here the first 27.9 ksec of the

105: observation.

106: 

107: Because of the small field of view of the S3 chip and the low redshift of

108: Abell 576, our analysis is mostly restricted to the central $\sim$200 kpc

109: of the cluster.

110: 

111: For the spectroscopic analysis we considered only data in the range

112: 0.5--8 keV.  Below this range the calibration of the ACIS CCDs is less

113: certain, and above this range the data contain few photons and are

114: dominated by particle background.  Spectral response matrices were

115: corrected for the reduction in quantum efficiency at low energies using

116: the time-dependent but spatially invariant {\it acisabs} model provided

117: by the Chandra X-ray Center.  Because of spatial variations in the

118: contaminant, the correction is not exact, especially below 1.2 keV

119: where the effect increases.  For this reason, we kept the absorption

120: column fixed to the Galactic value of $5.71\times10^{20}$~cm$^{-2}$

121: \citep{dl90} in the spectral fits presented here.  As a test, we

122: allowed the absorption column to vary, and measured a column density

123: nearly identical to the Galactic value, but with substantial errors,

124: typically on the order of 10\%.  The only effect of allowing the

125: absorption to be a variable parameter was to increase the errors in our

126: determinations of the other free parameters.  Thus, we concluded that

127: it was preferable to leave the absorption fixed at the Galactic value.

128: 

129: \begin{figure}

130: \epsscale{0.65}

131: \plotone{f1.eps}

132: \epsscale{1.0}

133: \caption{Gaussian smoothed, exposure-corrected, 0.3--6.0 keV image of

134: Abell 576.  Both the S2 and S3 chips are shown.  The north and southeast

135: surface brightness edges are clearly visible; the west edge is less

136: distinct.

137: \label{fig:img}}

138: \end{figure}

139: 

140: \section{Brightness Edges}

141: \label{sec:edges}

142: 

143: Figure~\ref{fig:img} shows a Gaussian smoothed, exposure-corrected

144: image of the cluster.  At least two, and perhaps more, surface

145: brightness edges are visible within the central 50 kpc.  As shown

146: below, they encompass a region of cool, high-metalicity gas.  This cool

147: gas also extends in a finger to the north of the cluster core, slightly

148: west of center.  As we will discuss in the remainder of this section,

149: we believe this finger of gas to have originated in a small subcluster

150: which is currently accreting into the center of the main cluster.  The

151: orientation of the edges are not consistent with gas simply sloshing

152: back and forth in a more or less fixed potential, which would create

153: parallel edges \citep*{qbb01} as opposed to the roughly triangular

154: configuration observed.  The observed edges are more consistent with

155: being the outer edges of a wake of stripped gas left behind by a

156: merging subcluster.  In this picture, the subcluster initially fell in

157: from the north, slightly to the west of the main cluster's center,

158: passed the main cluster center once, and is now making it's second pass

159: of the cluster's center.  The west and southeast edges describe the

160: outer edges of the wake of stripped gas from it's previous and current

161: pass of the main cluster's center.  This hypothesis also neatly

162: explains the finger of gas to the north, which cannot be easily

163: explained by simple sloshing.

164: 

165: \begin{figure}

166: \plotone{f2.eps}

167: \caption{Regions used for the spectral and surface brightness profiles.

168: The exact regions used for the spectral profiles are indicated.  The

169: surface brightness profiles used the same regions, but subdivided into

170: smaller radial bins.

171: \label{fig:regions}}

172: \end{figure}

173: 

174: \begin{figure}

175: \plotone{f3.eps}

176: \caption{Surface brightness profile across the north edge.  The solid line

177: outside 30 kpc is a beta model fit to the data using the parameters derived

178: by \protect\citet{mef+95} using the full field of view of the {\it

179: Einstein} IPC.  Inside 30 kpc, three models for the surface brightness jump

180: are shown: the solid line indicates the same beta model after adjusting for

181: the enhancement due to both the density and abundance increases inside the

182: edge.  The abundance only and density only components are shown as the

183: dashed and dotted lines, respectively .  For the $\beta$-model, $\beta =

184: 0.64$ and $r_c = 169$ kpc.  The core radius of the model is indicated as

185: ``${\rm r_c}$'' for reference.  The errors bars are 1$\sigma$.

186: \label{fig:north_sbr}}

187: \end{figure}

188: 

189: \begin{deluxetable*}{cccccccc}

190: \tablecaption{Spectral fit parameters outside the brightness edges

191: \label{tab:spec}}

192: \tablehead{

193: \colhead{} &

194: \colhead{$kT$} &

195: \colhead{($Z$)} &

196: \colhead{$\dot{M}$}&

197: \colhead{$kT_{\rm min}$} &

198: \colhead{$kT_2$} &

199: \colhead{} &

200: \colhead{} \\

201: \colhead{model} &

202: \colhead{(keV)} &

203: \colhead{($Z_\sun$)} &

204: \colhead{($M_\sun$ yr$^{-1}$)}&

205: \colhead{(keV)} &

206: \colhead{(keV)} &

207: \colhead{$\chi^2$} &

208: \colhead{d.o.f.}

209: }

210: \startdata

211: {\sc mekal}           & $4.05^{+0.17}_{-0.16}$ & $0.32^{+0.06}_{-0.07}$ & \nodata             & \nodata                & \nodata              & 357 & 297 \\

212: {\sc mekal + mkcflow} & $4.91^{+0.28}_{-0.51}$ & $0.37^{+0.08}_{-0.08}$ & $5.0^{+1.1}_{-1.0}$ & $0.01^{+0.20}_{-0.01}$ & \nodata              & 322 & 295 \\

213: {\sc mekal + mekal}   & $4.47^{+0.19}_{-0.19}$ & $0.47^{+0.09}_{-0.09}$ & \nodata             & \nodata                & $0.25^{+0.4}_{-0.2}$ & 289 & 295

214: \enddata

215: \end{deluxetable*}

216: 

217: We extracted surface brightness and spectral profiles across all three

218: edges using the regions shown in Figure~\ref{fig:regions}.  The

219: brightness edge 40\arcsec\ north of the peak of the X-ray emission

220: shows by far the largest jump in surface brightness: a factor of

221: $1.8\pm0.15$ (1$\sigma$) increase across the discontinuity (see

222: Figure~\ref{fig:north_sbr}).  A large jump in the abundance is also

223: visible across the discontinuity, while the temperature does not change

224: significantly (see the points with red error bars in

225: Figure~\ref{fig:spec_profile}).  This jump in abundance is significant

226: at more than 90\% confidence across the north edge.  At the low

227: temperature of Abell 576, the increased abundance across the edge has a

228: non-negligible effect on the emissivity of the gas.  This is

229: illustrated in Figure~\ref{fig:north_sbr}, which shows the surface

230: brightness profile across the north edge.  Outside the brightness edge,

231: the solid line is a $\beta$-model fit to the data using the core radius

232: and slope determined from observations with {\it Einstein}

233: \citep{mgf+96}.  Inside the edge, we use the same $\beta$-model, but we

234: increase the emissivity by an amount expected from each of three

235: different models: the dotted line indicates the increased surface

236: brightness due to the increase in density alone; the dashed line

237: indicates the increased emissivity due to the higher abundance; and the

238: solid line is the increased brightness due to both effects.  For all

239: three models we assume spherical symmetry for consistency with the

240: deprojection analysis.  As the figure demonstrates, neither the added

241: emissivity from the higher abundance nor that from the increased

242: density can account entirely for the observed increase in surface

243: brightness, but the two effects combined reproduce the overall

244: normalization of the central brightness quite well.  We note that had

245: we kept the abundance in our spectral models fixed, we would have

246: underestimated the emissivity of the gas inside the brightness edge,

247: and would therefore have overestimated its density.  This would then

248: overestimate the density contrast across the edge, which would

249: negatively affect the analysis that follows.

250: 

251: \begin{figure}

252: \plotone{f4.eps}

253: \caption{Spectral profiles in three sectors to the north, southeast, and

254: west from the cluster center.  The top set of points are temperatures; the

255: bottom set are abundances.  All error bars are 90\% confidence.

256: \label{fig:spec_profile}}

257: \end{figure}

258: 

259: We deprojected the surface brightness, which, when combined with the

260: temperature and abundance profiles, allowed us to determine deprojected

261: density and pressure profiles across the edge, under the assumption of

262: spherical symmetry.  We find a small jump in both the density and the

263: pressure across the north edge.  The density increases by a factor of

264: $2.8^{+0.8}_{-1.2}$ while the pressure increases by a factor of $2.4\pm0.8$

265: (both 1$\sigma$).  In order for the higher density gas to remain confined,

266: the pressure difference across the edge must be balanced by ram pressure

267: from motion of the high density gas through the lower density gas.  The

268: observed pressure difference implies that the higher density gas is moving

269: through the lower density gas with a velocity of $750 \pm 270$ km s$^{-1}$,

270: or Mach $0.9 \pm 0.3$ at the sound speed of the lower density gas (both

271: errors are 1$\sigma$).  We note that if we had failed to account for the

272: increase in abundance across the edge, that is, if we had assumed a

273: constant abundance on both sides of the edge, we would have overestimated

274: the density jump across the edge and consequently would have overestimated

275: the velocity of the dense gas cloud.  The velocity we measure is consistent

276: with velocities of both merging/accreting subclusters measured in other

277: clusters \citep[e.g.][]{mpn+00} and with velocities measured for some

278: ``sloshing'' edges in the cores of relaxed cluster \citep{mar03}.  On its

279: own, then, the measured velocity of the north edge is incapable of

280: distinguishing between these two scenarios for the creation of the

281: non-hydrostatic features in the cluster core.

282: 

283: To the southeast of the cluster center, a fainter edge is visible in the

284: image (see Figure~\ref{fig:img}).  Another yet fainter edge appears to the

285: west.  Both of these edges display the same abundance gradient as the north

286: edge, though in both cases the abundance jump appears to be more of a

287: gradient than a sharp edge, and is measured with much less significance

288: (see Figure~\ref{fig:spec_profile}).

289: 

290: A combined spectral analysis of all 3 of sectors yields improved

291: statistics at the cost of spatial resolution.  This is plotted as the

292: solid grey boxes in Figure~\ref{fig:spec_profile}.  The best-fit models

293: for the combined spectra outside the edges are given in

294: Table~\ref{tab:spec}.  The columns are as follows: (1) spectral model;

295: (2) \& (3) temperature and abundance, respectively, of primary MEKAL

296: model; (4) Cooling rate of MKCFLOW model; (5) minimum temperature of

297: cool gas in MKCFLOW model; (6) temperature of second MEKAL model, where

298: the abundance of the second model is tied to that of the primary MEKAL

299: model; (7) \& (8) $\chi^2$ and number of degrees of freedom for the

300: fit.  From this combined analysis, we find that the central abundance

301: is different from the abundance outside the edges at greater than 90\%

302: confidence.  As expected, the temperature measured outside the edges is

303: approximately equal to the mean temperature measured with higher

304: spatial resolution in the individual sectors.  However, when fit with a

305: two temperature model, the best-fit high temperature is only slightly

306: higher than for the single temperature model, while the cool

307: temperature is $0.25^{+0.4}_{-0.2}$ keV and contributes only 5\% to

308: normalization of the model spectrum.  (Because of the errors on this

309: measurement and given that our spectral fit only includes data at

310: energies $>0.5$ keV, this cool component should be considered an upper

311: limit of $0.65$ keV.)  Similarly, a single temperature plus cooling

312: flow model yields a relatively small mass accretion rate of $5\pm1$

313: ${\rm M_\sun~yr^{-1}}$ but to a very low temperature of less than $0.2$

314: keV.  An F-test shows that the two temperature model is a better fit

315: than the single temperature model at about the 3$\sigma$ level, while

316: the cooling flow model is better than the single temperature by only

317: $\sim$1$\sigma$.  In both multi-phase models, the cooler component

318: contributes a very small fraction of the total emission.

319: 

320: \begin{figure*}

321: \plottwo{f5a.eps}{f5b.eps}

322: \caption{{\it (a)} Heating rate and luminosity of radiative cooling as a

323: function of radius.  The exact definition of the heating rate is given in

324: the text.  The dotted curves indicate the minimum and maximum allowed rates

325: given the errors on our measurement of the subcluster's velocity.  {\it

326: (b)} Integrated cooling rate as a function of radius.  The solid line is

327: the rate in the absence of heating.  The dotted lines are the resulting

328: integrated cooling rate if the kinetic energy of the cool core is

329: dissipated over the given timescales.  The dissipated energy is assumed to

330: heat the gas at a constant rate per volume over the entire region.

331: \label{fig:cooling}}

332: \end{figure*}

333: 

334: \begin{deluxetable*}{cccccccc}

335: \tablecaption{Spectral fit parameters for cool core \label{tab:cflow}}

336: \tablehead{

337: \colhead{} &

338: \colhead{$kT$} &

339: \colhead{($Z$)} &

340: \colhead{$\dot{M}$}&

341: \colhead{$kT_{\rm min}$} &

342: \colhead{$kT_2$} &

343: \colhead{} &

344: \colhead{} \\

345: \colhead{model} &

346: \colhead{(keV)} &

347: \colhead{($Z_\sun$)} &

348: \colhead{($M_\sun$ yr$^{-1}$)}&

349: \colhead{(keV)} &

350: \colhead{(keV)} &

351: \colhead{$\chi^2$} &

352: \colhead{d.o.f.}

353: }

354: \startdata

355: {\sc mekal}           & $3.48^{+0.16}_{-0.18}$ & $0.56^{+0.13}_{-0.11}$ & \nodata     & \nodata               & \nodata                & 186 & 188 \\

356: {\sc mekal + mkcflow} & $3.77^{+0.25}_{-0.18}$ & $0.64^{+0.13}_{-0.11}$ & $1.1\pm0.5$ & $0.01^{+0.9}_{-0.01}$ & \nodata                & 183 & 186 \\

357: {\sc mekal + mekal}   & $3.62^{+0.25}_{-0.10}$ & $0.66^{+0.16}_{-0.14}$ & \nodata     & \nodata               & $0.65^{+2.11}_{-0.45}$ & 181 & 186 \\

358: {\sc mekal + foreground}& $3.21^{+0.27}_{-0.28}$ & $0.70^{+0.22}_{-0.19}$ & \nodata     & \nodata               & $4.05$\tablenotemark{a} & 186 & 188

359: \enddata

360: \tablenotetext{a}{\centering Temperature and abundance of foreground component taken from single-temperature fit in Table~\protect{\ref{tab:spec}}.}

361: \end{deluxetable*}

362: 

363: While a radial temperature gradient across the region used for our

364: analysis ($\sim$30--150 kpc) would show qualitatively similar effects,

365: we would expect such a gradient to cover a relatively small range of

366: temperatures, say, a factor of 2 as is seen in the most relaxed cooling

367: flow systems \citetext{e.g. \citealp{ppk+01}; \citealp*{bsm03}}.  To

368: find gas at a temperature less than $1/4$ of the ambient temperature is

369: quite unusual.  The merging subcluster hypothesis mentioned above

370: provides an attractive explanation for this extremely cool gas, if the

371: cool gas has been stripped from the subcluster during its infall.  The

372: small radius of curvature of the north edge suggests that this remnant

373: core is physically quite small and that the original subcluster was

374: also relatively small and therefore cool.  This  provides a consistent

375: qualitative explanation for the presence of the cool component.

376: Unfortunately, given the large amount of gas likely to have been

377: stripped from the subcluster during infall, it is not possible for us

378: to determine the subcluster's original mass and virial temperature and

379: is therefore not possible to conclude whether or not the cooler gas is

380: consistent with having originated in the undisturbed subcluster.

381: We note, however, that the ``sloshing'' picture for the brightness

382: edges cannot explain the existence of the cool component.  Sloshing

383: would create the same general appearance of a cold front

384: \citep[e.g.][]{mvm01}, but would not cause the higher density gas to

385: cool any more than it would in the absense of the sloshing.

386: 

387: An alternative explanation for the presence of gas at less than $1/4$

388: of the ambient temperature is that the usual method of preventing a

389: classical cooling flow from developing, whether that be AGN heating,

390: conduction, or some other mechanism, is absent or suppressed in A576.

391: Indeed the absence of the characteristic ``bubbles'' seen in other

392: clusters \citep[e.g.][]{fse+00,mwn+00,bsm+01} suggests there has been

393: no recent strong AGN activity, for one.  However, we find this

394: explanation to be unsatisfactory, since it would require the existence

395: of a strong cooling flow, which is explicitly ruled out by the small

396: cooling rate determined above.

397: 

398: \section{Cooling Flow}

399: \label{sec:cflow}

400: 

401: As Figure~\ref{fig:spec_profile} shows, the temperature drops in the very

402: core of the cluster, that is, inside the brightness edge.  It is natural to

403: ask, then, if this gas shows any evidence of being multi-phase.  To test

404: this, we fit a spectrum of the gas in the core with a single-temperature

405: absorbed MEKAL model \citep{kaa92}, with a MEKAL model plus a multi-phase

406: MEKAL model (MKCFLOW model), and with the sum of two MEKAL models.  The

407: salient parameters of the fits are shown in Table~\ref{tab:cflow}.  The

408: columns are as follows: (1) spectral model; (2) \& (3) temperature and

409: abundance, respectively, of primary MEKAL model; (4) Cooling rate of

410: MKCFLOW model; (5) minimum temperature of cool gas in MKCFLOW model; (6)

411: temperature of second MEKAL model, where the abundance of the second model

412: is tied to that of the primary MEKAL model; (7) \& (8) $\chi^2$ and number

413: of degrees of freedom for the fit.  As shown by the similar fit statistics,

414: the goodness of fit of all three models are essentially identical.

415: 

416: The spectroscopic cooling rate we measure is quite low: an order of

417: magnitude smaller than the ``classical cooling flow'' accretion rate of

418: $\dot{M} = M / t_{\rm cool} = 11$ M$_\sun$ yr$^{-1}$ derived from the gas

419: mass and cooling time in the inner 30 kpc (see Figure~\ref{fig:cooling} and

420: \S\ref{sec:suppress} below).  We note that the cool component in the

421: two-temperature model only contributes $\sim$0.1\% to the overall

422: normalization of the model.  This is consistent with the extremely small

423: cooling rate measured for the cooling flow model.

424: 

425: When the second temperature component of the two-MEKAL model was set to the

426: parameters of the fit to the outer gas (the equivalent of deprojection),

427: the fit was no better than for the single-temperature model.  This is

428: perhaps not too surprising, however, since the contribution of the

429: projected foreground component to the overall normalization of the model is

430: only about 2\%.

431: 

432: Surprisingly, the best-fit low temperature in the cooling flow model, that

433: is, the temperature to which the gas cools, is equal to the minimum

434: temperature allowed by the model, $kT_{\rm low} = 0.01$ keV.  However, the

435: error on the fit allows for $kT_{\rm low}$ to be as high as 0.9~keV.  This

436: is about 1/4 of the ambient temperature, which is consistent with minimum

437: temperatures found in stronger cooling flows \citep[e.g.][]{ppk+01}.  The

438: low temperature of this gas is consistent with that of the cool gas found

439: at larger radii.

440: 

441: In short, we find essentially no evidence of multi-phase gas in the

442: central cool core.  Of the models described above, the cooling flow

443: model contributes the largest contribution from multi-phase gas to the

444: overall emission, with only 5\% of the emission coming from the

445: multi-phase gas.

446: 

447: \section{Suppression of Cooling}

448: \label{sec:suppress}

449: 

450: Given that we observe essentially no cooling in either the core or outside

451: the core in what is otherwise a quite relaxed cluster, we now explore

452: whether the dissipation of the kinetic energy of the remnant core is

453: capable of suppressing cooling at the observed level.  In the rest of this

454: section we take the merging subcluster hypothesis to be the correct

455: explanation of the brightness edges.

456: 

457: We determined the kinetic energy of the gas inside the north edge using the

458: velocity of the edge we measured above plus a gas mass determined from the

459: deprojected density profile inside the edge ($\sim2\times10^{10}~{\rm

460: M_\sun}$).  We then calculated the rate of energy input from the

461: dissipation of this kinetic energy over a variety of timescales.

462: Figure~\ref{fig:cooling}a shows this energy dissipation rate compared to

463: the luminosity due to radiative cooling as a function of radius.  The

464: timescale used for calculating the heating rate is three crossing times of

465: the cluster to the given radius at the current velocity of the north edge.

466: In three crossing times, the moving cool core will have swept up its mass

467: in gas, reducing its kinetic energy by $3/4$.  We therefore assume

468: perfectly efficient thermalization of $3/4$ of the kinetic energy over a

469: timescale equal to three crossing times by the core at its current velocity

470: in calculating the heating rate.

471: 

472: If we take the point at which the west and southeast edges converge as

473: indicative of the current orbital radius of the subcluster ($\sim$100 kpc

474: from the cluster center), the dissipation timescale derived using the above

475: method is $4 \times 10^8$ years.  As can be seen from

476: Figure~\ref{fig:cooling}b, the heating simply from the dissipation of the

477: kinetic energy of the subcluster is capable of suppressing cooling by a

478: factor of more than 4 in the inner 100 kpc over this timescale.  If the

479: dissipation of the core's kinetic energy is spread out over $10^9$ years,

480: cooling can still be suppressed at the level observed in the inner 30 kpc.

481: Numerical simulations have shown that mergers of unequal mass clusters

482: thermalize their kinetic energy on timescales of one to a few times $10^9$

483: years \citep{rs01}, although the mass ratio involved in A576 is probably

484: much larger than has generally been tested in these simulations.

485: 

486: So far, we have omitted any discussion of the role of turbulence in

487: suppression of cooling.  Turbulence generated by the motion of such a small

488: subcluster would eventually dissipate and be thermalized on a timescale $<

489: 10^8$ yr \citep*{fts03}, which is short compared to the timescale we are

490: considering for dissipation of the kinetic energy.  We therefore consider

491: it reasonable to assume close to 100\% thermalization efficiency.  We have

492: also neglected the influence of the dark matter on the total energy budget.

493: This is not so easily dismissed, as the dark matter halo of the subcluster

494: will take longer to dissipate its kinetic energy than will the gas.  The

495: dark matter oscillations about the center of mass of the system will induce

496: additional motion in the gas, transferring kinetic energy from the dark

497: matter to the gas and thereby increasing the total energy available for

498: heating the gas.  This will only serve to amplify the effect we have

499: discussed.  Our na\"ively calculated heating rate should therefore be taken

500: as a lower limit to the heating resulting from the subcluster merger.

501: 

502: Other alternatives for balancing cooling in clusters have been discussed.

503: These include heating from central AGN \citep{csf+02,bk02,fsa+03}, from

504: conduction \citep[e.g.][]{nm01}, and from turbulent mixing \citep{kn03}

505: induced by AGN activity.  Our data show no evidence for the X-ray cavities

506: typical of clusters with strong central AGN, and indeed the cluster has one

507: extremely weak AGN at its center with no non-nuclear radio emission.  Since

508: these cavities tend to persist for several AGN duty cycles

509: \citep[e.g. Perseus,][]{fse+00} and none are observed in A576, AGN heating

510: can be ruled out, at least over the last few times $10^8$ years.

511: 

512: \section{Conclusions}

513: \label{sec:conclusions}

514: 

515: While earlier X-ray observations of Abell 576 have shown it to be quite

516: regular on large scales, we have demonstrated that the core of the cluster

517: is far from dynamical equilibrium.  We found multiple surface brightness

518: edges in the cluster center which we have demonstrated to be indicative of

519: jumps in both density and abundance.  Of the two most likely explanations

520: for the existence of these edges, our analysis favors the hypothesis that

521: they are formed by gas stripped from a merging subcluster.  Most of the gas

522: appears to have been stripped from the subcluster, leaving a core only

523: $\sim$30 kpc in radius.  The stripped gas has been found to have both a

524: lower temperature and a higher abundance than the gas in the rest of Abell

525: 576.

526: 

527: We find no evidence of gas cooling from the ambient temperature of the main

528: cluster, but do find some suggestion of very cool gas at a temperature

529: expected of gas that had condensed out of the ICM of the subcluster.  The

530: simple cooling rate derived from the gas mass and cooling time is an order

531: of magnitude larger than the spectroscopically measured $\dot{M}$.  We have

532: demonstrated that dissipation of the kinetic energy of the observed remnant

533: core of an infalling subgroup may be sufficient to reduce cooling to the

534: observed rate, if that energy is dissipated over a timescale of $\lesssim

535: 10^9$ years.

536: 

537: \acknowledgements

538: Support for this work was provided by the National Aeronautics and Space

539: Administration through {\it Chandra} Award Number G01-2131X issued by the {\it

540: Chandra} X-ray Observatory Center, which is operated by the Smithsonian

541: Astrophysical Observatory for and on behalf of NASA under contract

542: NAS8-39073, and by NASA contract NAG5-12933.

543: 

544: \begin{thebibliography}{21}

545: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

546: 

547: \bibitem[{{Bennett} {et~al.}(2003){Bennett}, {Halpern}, {Hinshaw}, {Jarosik},

548:   {Kogut}, {Meyer}, {Page}, {Spergel}, {Tucker}, {Wollack}, {Wright}, {Barnes},

549:   {Greason}, {Hill}, {Nolta}, {Odegard}, {Peirs}, {Verde}, \&

550:   {Weiland}}]{bhh+03}

551: {Bennett}, C.~L., {Halpern}, M., {Hinshaw}, G., {Jarosik}, N., {Kogut},

552:   A.~{Limon}, M., {Meyer}, S.~S., {Page}, L., {Spergel}, D.~N., {Tucker},

553:   G.~S., {Wollack}, E., {Wright}, E.~L., {Barnes}, C., {Greason}, M.~R.,

554:   {Hill}, R.~S.~{Komatsu}, E., {Nolta}, M.~R., {Odegard}, N., {Peirs}, H.~V.,

555:   {Verde}, L., \& {Weiland}, J.~L. 2003, \apjs, 148, 1

556: 

557: \bibitem[{{Blanton} {et~al.}(2001){Blanton}, {Sarazin}, {McNamara}, \&

558:   {Wise}}]{bsm+01}

559: {Blanton}, E.~L., {Sarazin}, C.~L., {McNamara}, B.~R., \& {Wise}, M.~W. 2001,

560:   \apjl, 558, L15

561: 

562: \bibitem[{{Blanton} {et~al.}(2003){Blanton}, {Sarazin}, \& {McNamara}}]{bsm03}

563: {Blanton}, E.~L., {Sarazin}, C.~L., \& {McNamara}, B.~R. 2003, \apj, 585, 227

564: 

565: \bibitem[{{Br{\" u}ggen} \& {Kaiser}(2002)}]{bk02}

566: {Br{\" u}ggen}, M. \& {Kaiser}, C.~R. 2002, \nat, 418, 301

567: 

568: \bibitem[{{Churazov} {et~al.}(2002){Churazov}, {Sunyaev}, {Forman}, \& {B{\"

569:   o}hringer}}]{csf+02}

570: {Churazov}, E., {Sunyaev}, R., {Forman}, W., \& {B{\" o}hringer}, H. 2002,

571:   \mnras, 332, 729

572: 

573: \bibitem[{{Dickey} \& {Lockman}(1990)}]{dl90}

574: {Dickey}, J.~M. \& {Lockman}, F.~J. 1990, \araa, 28, 215

575: 

576: \bibitem[{{Fabian} {et~al.}(2003){Fabian}, {Sanders}, {Allen}, {Crawford},

577:   {Iwasawa}, {Johnstone}, {Schmidt}, \& {Taylor}}]{fsa+03}

578: {Fabian}, A.~C., {Sanders}, J.~S., {Allen}, S.~W., {Crawford}, C.~S.,

579:   {Iwasawa}, K., {Johnstone}, R.~M., {Schmidt}, R.~W., \& {Taylor}, G.~B. 2003,

580:   \mnras, in press (astro-ph/0306036)

581: 

582: \bibitem[{{Fabian} {et~al.}(2000){Fabian}, {Sanders}, {Ettori}, {Taylor},

583:   {Allen}, {Crawford}, {Iwasawa}, {Johnstone}, \& {Ogle}}]{fse+00}

584: {Fabian}, A.~C., {Sanders}, J.~S., {Ettori}, S., {Taylor}, G.~B., {Allen},

585:   S.~W., {Crawford}, C.~S., {Iwasawa}, K., {Johnstone}, R.~M., \& {Ogle}, P.~M.

586:   2000, \mnras, 318, L65

587: 

588: \bibitem[{{Fujita} {et~al.}(2003){Fujita}, {Takizawa}, \& {Sarazin}}]{fts03}

589: {Fujita}, Y., {Takizawa}, M., \& {Sarazin}, C.~L. 2003, \apj, {584}, 190

590: 

591: \bibitem[{{Kaastra}(1992)}]{kaa92}

592: {Kaastra}, J.~S. 1992, {An X-Ray Spectral Code for Optically Thin Plasmas},

593:   Tech. rep., {Internal SRON-Leiden Report, updated version 2.0}

594: 

595: \bibitem[{{Kim} \& {Narayan}(2003)}]{kn03}

596: {Kim}, W.-T. \& {Narayan}, R. 2003, \apjl, submitted (astro-ph/0308376)

597: 

598: \bibitem[{{Markevitch} {et~al.}(2000){Markevitch}, {Ponman}, {Nulsen}, {Bautz},

599:   {Burke}, {David}, {Davis}, {Donnelly}, {Forman}, {Jones}, {Kaastra},

600:   {Kellogg}, {Kim}, {Kolodziejczak}, {Mazzotta}, {Pagliaro}, {Patel}, {Van

601:   Speybroeck}, {Vikhlinin}, {Vrtilek}, {Wise}, \& {Zhao}}]{mpn+00}

602: {Markevitch}, M., {Ponman}, T.~J., {Nulsen}, P.~E.~J., {Bautz}, M.~W., {Burke},

603:   D.~J., {David}, L.~P., {Davis}, D., {Donnelly}, R.~H., {Forman}, W.~R.,

604:   {Jones}, C., {Kaastra}, J., {Kellogg}, E., {Kim}, D.-W., {Kolodziejczak}, J.,

605:   {Mazzotta}, P., {Pagliaro}, A., {Patel}, S., {Van Speybroeck}, L.,

606:   {Vikhlinin}, A., {Vrtilek}, J., {Wise}, M., \& {Zhao}, P. 2000, \apj, 541,

607:   542

608: 

609: \bibitem[{{Markevitch} {et~al.}(2001){Markevitch}, {Vikhlinin}, \&

610:   {Mazzotta}}]{mvm01}

611: {Markevitch}, M., {Vikhlinin}, A., \& {Mazzotta}, P. 2001, \apjl, 562, L153

612: 

613: \bibitem[{{Markevtich}(2002)}]{mar02}

614: {Markevtich}, M. 2002, in Proceedings of the Chandra Calibration Workshop

615:   (Cambridge: Chardra X-ray Center), http://asc.harvard.edu/ccw/proceedings/

616: 

617: \bibitem[{{Markevtich}(2003)}]{mar03}

618: {Markevtich}, M. 2003, in The Riddle of Cooling Flows in Galaxies and Clusters

619:   of Galaxies (Charlottesville: University of Virginia), in press

620:   (http://www.astro.virginia.edu/coolflow/)

621: 

622: \bibitem[{{Mazzotta} {et~al.}(2003){Mazzotta}, {Edge}, \& {Markevitch}}]{mem03}

623: {Mazzotta}, P., {Edge}, A., \& {Markevitch}, M. 2003, \apj, in press

624:   (astro-ph/0303314)

625: 

626: \bibitem[{{McNamara} {et~al.}(2000){McNamara}, {Wise}, {Nulsen}, {David},

627:   {Sarazin}, {Bautz}, {Markevitch}, {Vikhlinin}, {Forman}, {Jones}, \&

628:   {Harris}}]{mwn+00}

629: {McNamara}, B.~R., {Wise}, M., {Nulsen}, P.~E.~J., {David}, L.~P., {Sarazin},

630:   C.~L., {Bautz}, M., {Markevitch}, M., {Vikhlinin}, A., {Forman}, W.~R.,

631:   {Jones}, C., \& {Harris}, D.~E. 2000, \apjl, 534, L135

632: 

633: \bibitem[{{Mohr} {et~al.}(1995){Mohr}, {Evrard}, {Fabricant}, \&

634:   {Geller}}]{mef+95}

635: {Mohr}, J.~J., {Evrard}, A.~E., {Fabricant}, D.~G., \& {Geller}, M.~J. 1995,

636:   \apj, 447, 8

637: 

638: \bibitem[{{Mohr} {et~al.}(1996){Mohr}, {Geller}, {Fabricant}, {Wegner},

639:   {Thorstensen}, \& {Richstone}}]{mgf+96}

640: {Mohr}, J.~J., {Geller}, M.~J., {Fabricant}, D.~G., {Wegner}, G.,

641:   {Thorstensen}, J., \& {Richstone}, D.~O. 1996, \apj, 470, 724

642: 

643: \bibitem[{{Narayan} \& {Medvedev}(2001)}]{nm01}

644: {Narayan}, R. \& {Medvedev}, M.~V. 2001, \apjl, 562, L129

645: 

646: \bibitem[{{Peterson} {et~al.}(2001){Peterson}, {Paerels}, {Kaastra}, {Arnaud},

647:   {Reiprich}, {Fabian}, {Mushotzky}, {Jernigan}, \& {Sakelliou}}]{ppk+01}

648: {Peterson}, J.~R., {Paerels}, F.~B.~S., {Kaastra}, J.~S., {Arnaud}, M.,

649:   {Reiprich}, T.~H., {Fabian}, A.~C., {Mushotzky}, R.~F., {Jernigan}, J.~G., \&

650:   {Sakelliou}, I. 2001, \aap, 365, L104

651: 

652: \bibitem[{{Ricker} \& {Sarazin}(2001)}]{rs01}

653: {Ricker}, P.~M. \& {Sarazin}, C.~L. 2001, \apj, 561, 621

654: 

655: \bibitem[{{Rothenflug} {et~al.}(1984){Rothenflug}, {Vigroux}, {Mushotzky}, \&

656:   {Holt}}]{rvm+84}

657: {Rothenflug}, R., {Vigroux}, L., {Mushotzky}, R.~F., \& {Holt}, S.~S. 1984,

658:   \apj, 279, 53

659: 

660: \bibitem[{Quilis} {et~al.}(2001){Quilis}, {Bower}, \& {Balogh}]{qbb01}

661: {Quilis}, V., {Bower}, R.~G., \& {Balogh}, M.~L. 2001, \mnras, 328, 1091

662: 

663: \end{thebibliography}

664: 

665: \end{document}

666: