1: In this section we present our methods of classifying the evolving morphologies of our simulated mergers. We shall find that as our mergers progress, they exist primarily in one of three states: compact cool core (CCC), compact warm core (CWC) and extended warm core (EWC) morphologies. In the sections which then follow, we shall discuss the temporal evolution of our mergers through these three states and relate their occurrence to the changing structure of our merger remnants' cores.
2:
3: \subsection{Observational Datasets}
4:
5: To compare our classified systems to observations, we will employ the 20
6: cluster statistical sample presented by \citet{Sandersonetal06}. We
7: have also augmented the results of this study by fitting 1 dimensional
8: $\beta$-models (\cf\ Equation 1 below) to each cluster. To do so, we took the
9: APEC spectral normalization per unit area as a measure of projected
10: emissivity, using the more finely binned, projected annular spectral
11: profiles of \citet{Sandersonetal06} to construct azimuthally-averaged radial
12: surface brightness profiles. Eqn. 1 of \citet{Sandersonetal06} was fitted using the
13: $\chi^2$ statistic to each profile across the full radial range to
14: determine the core radius; no attempt was made to excise or model any
15: excess emission associated with a cool core. The use of projected
16: emissivity instead of actual X-ray surface brightness is advantageous,
17: since it allows for the effects of galactic absorption, which is a
18: free-fitted parameter in each annular spectrum for most of the
19: clusters analyzed \citep[but fixed at the level expected from HI
20: surveys in the few clusters where this parameter was not well
21: constrained; see][for details]{Sandersonetal06}. For example, Abell 478,
22: suffers from strong galactic absorption with significant spatial
23: variation \citep{Sandersonetal05} that could bias fits based on the surface brightness.
24:
25: In later sections we also employ the catalogue of ROSAT and ASCA observations presented by \citet{Horner_thesis}. This catalogue is best suited to comparisons of the global X-ray properties of our simulations to observations because it is large, complete and presents results in which the cores of systems are not excised (a common procedure in other catalogues performed to reduce scatter in scaling relations for cosmological studies).
26:
27: \subsection{Classification of Morphological States}\label{analysis-morphology}
28:
29: We seek to determine how our evolving merger systems compare with recent \Chandra\ and \XMM\ observations
30: of the ICM: at which points they would manifest as compact or
31: extended core systems, when they would appear to have been significantly
32: affected by radiative cooling based on their entropy and projected
33: temperature profiles, and when they would exhibit evidence of merger
34: activity under typical observational circumstances. To minimize
35: biases in our comparison to observations, we have developed a
36: classification scheme based on a set of easily observable properties
37: of the system which can be readily related to its physical state.
38:
39: It has been noted that projection effects can have a significant
40: influence on the surface brightness structure and projected
41: temperatures of a merging system \citep{RT02}. For this reason, we have
42: considered three orthogonal projections in our study: two in the plane of the orbit along the line initially connecting
43: the systems (x) and in the direction of initial tangential velocity
44: (y) and one normal to the plane of the orbit (z). However, given the rarity of
45: observing systems in a configuration corresponding to our x-projections, we shall
46: focus only on the y and z-projections in this paper. The only significant effect of this omission
47: is to exclude those rare and peculiar states produced when the systems are
48: separate and still relaxed but appear as one system in projection.
49:
50: \subsubsection{surface brightness}
51:
52: We first classify our systems according to the shape of their
53: projected surface brightness profiles. To do so, we utilize the
54: simple 3-parameter $\beta$-model commonly employed in analysis of
55: observations. This model characterizes the azimuthally averaged
56: surface brightness profile of a system at a projected
57: radius$^{7}$\footnotetext[7]{Throughout this paper, we adopt the peak
58: of the projected bolometric X-ray surface brightness as the center
59: of the system.} $r_p$ by the equation
60:
61: \begin{center}
62: \begin{equation}\label{eqn-Sigma}
63: \Sigma_x(r_p)=\frac{\Sigma_o}{{\left( 1+ {\left( r_p/r_c \right)}^{2} \right)}^{(6\beta-1)/2}}
64: \end{equation}
65: \end{center}
66:
67: \noindent where $r_c$ is the profile's core radius, $\beta$
68: characterizes its asymptotic logarithmic slope at large radius and
69: $\Sigma_o$ its normalization. We exclude projected radii less than
70: $30$\kpc\ from this analysis and have verified
71: that varying this radius from $20$ to $50$\kpc\ has a minimal effect
72: and does not influence the conclusions of our study.
73:
74: Although more sophisticated parameterizations of the X-ray surface brightness
75: profiles of clusters have been developed \citep[\eg][]{Vikhlininetal06, Mahdavietal07}, we utilize this one
76: to take advantage of the large statistical samples of $\beta$-model
77: fits presently available.
78:
79: To separate ``compact'' cores from ``extended'' cores we shall use a cut of $100$\kpc. This corresponds approximately to the unpopulated region of the observed $r_c$ distribution reported by \citet{OtaMitsuda04} separating observed compact and extended core systems.
80:
81:
82: \subsubsection{central entropy and temperature gradient}
83:
84: To judge the relevance of radiative cooling for our evolving cluster
85: cores we will use the central entropy of our systems. We do this by
86: computing radially binned entropy profiles, denoted as $S(r)$, and
87: interpolate from them to obtain ``central'' values at $40$\kpc\ (denoted as $S_{40}$). We use
88: entropy for this purpose rather than the system's projected
89: temperature because it is numerically robust
90: in our simulations and unambiguously tied to the
91: relative importance of heating and cooling for both our observed and
92: simulated systems. Furthermore, as pointed out previously by several
93: authors
94: \citep[\eg][]{Voitetal02,Voitetal03,KaiserandBinney03,McCarthyetal07},
95: variations in the core properties of X-ray clusters are most insightfully
96: understood in terms of this quantity.
97:
98: To separate ``cool'' core systems from ``warm'' core systems, we shall use a cut in $S_{40}$ of $50$\keVcmsq\ and refer to systems with $S_{40}<50$\keVcmsq\ as ``cool core'' systems and systems with $S_{40}>50$\keVcmsq\ as ``warm core'' systems. This choice is motivated by an examination of the entropy profiles of the \citet{Sandersonetal06} catalogue (these will be presented in a future publication); central entropies lower than this value tend to occur in cool core systems while higher central entropies tend to occur in warm cores. Furthermore, recent examinations of BCGs in galaxy clusters find a break at this entropy level such that those in clusters with $S_{40} \lesssim 50$\keVcmsq\ show
99: evidence of star formation while those in systems with warmer cores do not \citep{Bildfelletal08, Raffertyetal08}. For reference, the cooling time of material with $S=50$\keVcmsq\ is $\sim2$\Gyrs\ for our systems.
100:
101: However, temperature is an intuitively appealing quantity and can be more directly
102: compared to observations. For this reason we have also measured
103: the shape of our systems' evolving projected central temperature
104: profiles.
105: \Chandra~ and \XMM~ observations have shown that projected cluster temperature
106: profiles have a universal shape outside of the projected radius $0.1R_{200}$; a broad peak between $0.1R_{200}$
107: and $0.2R_{200}$ and a continuous decline outside of $0.2R_{200}$
108: \citep{Markevitchetal98,Sandersonetal06,Prattetal06}. For projected radii within $0.1R_{200}$, a great
109: deal of system-to-system variation in temperature profiles is seen.
110: This radius roughly corresponds to the cooling radius of compact core
111: clusters and it is within this region that cool core systems exhibit
112: relatively low temperatures \citep{Piffarettietal04}.
113:
114: Motivated by these observations, we have chosen to characterize the
115: apparent temperature structure of the cores of our merging systems by
116: measuring the strength of their central temperature gradients with the
117: quantity
118:
119: \begin{center}
120: \begin{equation}\label{eqn-DeltaT}
121: \Delta_T=\frac{T_{outer}-T_{inner}}{T_{outer}}
122: \end{equation}
123: \end{center}
124:
125: \noindent Here, $T_{inner}$ is the spectrally fit temperature (see
126: Poole07 for details regarding our method of simulating observed
127: cluster temperatures) of material in projected radii
128: from $30$\kpc\ to $0.1R_{200}$ ($0$ to $0.1R_{200}$ for the observations; this
129: represents a negligible difference however) and $T_{outer}$ the spectrally
130: fit temperature for material in the range of projected radii
131: $0.1R_{200}$ to $0.2R_{200}$. This quantity is thus a
132: measure of the temperature gradient within the core region, where
133: system-to-system variations are observed and cool core systems are
134: seen to have low temperatures. Defined in this way, $\Delta_T$ is
135: positive for systems with core temperatures cooler than isothermal and
136: negative for systems with core temperatures warmer than isothermal.
137:
138: The entropy cut we have chosen ($S_{40}=50$\keVcmsq) for dividing cool
139: cores from warm cores roughly translates to a cut in central temperature gradients of \DT$=0.05$ for relaxed systems.
140:
141: \subsubsection{apparent dynamical state}
142:
143: Each projection for each system is also classified into instances
144: where it would exhibit no substructure and regular isophots to an
145: observer, significant substructure, or so much substructure as to
146: render the system indescribable by a $\beta$-model (generally due to
147: the close proximity of the secondary or various other brief but extreme
148: distortions). We refer to these situations as undisturbed, disturbed
149: and strongly disturbed respectively. Classification is performed
150: through visual inspection of simulated $50$ks \Chandra~ observations
151: for each output (with the system at $z=0.1$), in each projection.
152: Our simulated observations were produced following the procedure detailed in Poole06.
153:
154: Since instances of strongly disturbed morphology can not be assigned meaningful core radii, they are omitted from this morphological discussion. Such instances are rare and their exclusion does not represent a significant omission, although it is important to keep in mind that this exceptional class does manifest occasionally at early stages of the interaction.
155:
156: \begin{figure*}
157: \begin{minipage}{175mm}
158: \begin{center}
159: \leavevmode \hspace*{-1cm} \epsfysize=19cm \epsfbox{delta_rc_astroph.ps}
160: \caption[Central temperature gradient (\DT) plotted against core radius (\rc)]{Central temperature gradient (\DT) plotted against core radius (\rc) for each output of our nine simulations in y and z-projections. The cuts in core radius we use to separate compact-core systems ($r_c<100$\kpc) from extended-core systems ($r_c>100$\kpc) as well as the cut in \DT\ which roughly separates cool core ($S_{40}<50$\keVcmsq) from warm core ($S_{40}>50$\keVcmsq) systems are shown with dashed lines. Compact core systems with low/high central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with low/high central entropies (\ie\ ECC and EWC systems) are in magenta and green respectively. The quadrants of this plane corresponding to each morphology are labelled in the central pannel. Each population is further subdivided into systems showing obvious evidence of substructure (crosses) and systems with regular isophots (solid circles). Data from the statistically representative catalogue of \citet{Sandersonetal06} is shown in black with error bars (closed circles and crosses indicate systems noted in their study to be relaxed or visibly disturbed respectively). Black text around the boundary indicate the mass ratio and $v_t/V_c$ of the simulations presented in each panel.}
161: \label{fig-delta_rc}
162: \end{center}
163: \end{minipage}
164: \end{figure*}
165:
166: \begin{figure*}
167: \begin{minipage}{175mm}
168: \begin{center}
169: \leavevmode \hspace*{-1cm} \epsfysize=19cm \epsfbox{tcool_DeltaT_astroph.ps}
170: \caption[Central cooling time plotted against central temperature decrement]{Central cooling time ($t_{cool}$) plotted against central temperature decrement (\DT) for each output of our nine simulations in y and z-projections (following the same format as Fig. \ref{fig-delta_rc}). The cuts in \DT\ and $t_{cool}$ which roughly separate cool core systems from warm core systems are shown with dashed lines (the quadrants corresponding to warm and cool morphologies are labelled in the central panel) while the $t_{cool}=5$\Gyr\ cut used by \citet{Sandersonetal06} to divide short cooling time systems from long cooling time systems is shown with dotted lines. Compact core systems with low ($S_{40}<50$\keVcmsq) and high ($S_{40}>50$\keVcmsq) central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with low/high central entropies (\ie\ ECC and EWC systems) are in magenta and green respectively. Each population is further subdivided into systems showing obvious evidence of substructure (crosses) and systems with regular isophots (solid circles). Data from the statistically representative catalogue of \citet{Sandersonetal06} is shown in black with error bars (closed circles and crosses indicate systems noted in their study to be relaxed or visibly disturbed respectively). Black text around the boundary indicate the mass ratio and $v_t/V_c$ of the simulations presented in each panel.}
171: \label{fig-tcool_DeltaT}
172: \end{center}
173: \end{minipage}
174: \end{figure*}
175:
176: \subsection{Three Dominant Morphologies}\label{analysis-dominant}
177:
178: In Fig. \ref{fig-delta_rc} we present a \DTrc\ plot for our nine simulations. Projections in y and z are displayed for each of the 120 outputs (spanning $12$\Gyr\ in $100$\Myr\ intervals) of our simulations with instances of undisturbed and disturbed morphology indicated by filled circles or crosses respectively. Our cuts in \rc\ and \Sforty\ separate four potential populations which we identify in this plot using colours: compact cool cores (CCCs) are shown in blue, compact warm cores (CWCs) are in red, extended cool cores are in magenta and extended warm cores (EWCs) are green. We also plot the catalogue of \citet{Sandersonetal06} in black for comparison.
179:
180: Fig. \ref{fig-delta_rc} shows that very few instances of extended cool cores occur. Looking carefully at those circumstances which have been identified as an extended cool core, we find that in many cases it is a result of one of two effects. In some instances when the system appears disturbed, the passage of a merging core near to the primary core (either during first or second pericentric passage) can enlarge estimates of $r_c$ while the primary core remains sufficiently undisturbed to be cool. The system is very obviously disturbed at these times but not so much that a $\beta$- model can not be reasonably fit to its surface brightness. The only case in which significant numbers of relaxed extended cool core states occur is the 10:1 head-on merger. The system has relaxed to a state with core radii only slightly larger and central entropies only slightly less than our morphology cuts in this case.
181:
182: With exceptional instances such as these aside, we see that throughout the duration of a merger an observer would classify the system as existing predominantly in one of three states: compact systems with cool or warm cores (CCC or CWC) or extended warm cores (EWC). In each case, the system may or may not also possess obvious substructure and/or disturbed isophots, meaning that members of each class could be found which appear to be disturbed or undisturbed. Including the approximate cut in \DT\ which results from our cut in $S_{40}$, we see that these three populations (especially the undisturbed set) map well to three separate regions in the \DT$-r_c$ plane. Encouragingly, we also find good agreement between the regions of this plane occupied by our simulations and by the observations. In Section \ref{analysis-states} we will examine the evolution of our systems through this plane and in Section \ref{analysis-structure}, examine the processes principally responsible for setting a system's morphology.
183:
184: It has been noted that clusters with warm cores can have cooling times less than their median formation age \citep{Donahueetal05,Sandersonetal06}. In Fig. \ref{fig-tcool_DeltaT} we plot the central cooling time (measured at $40$\kpc) against $\Delta_T$ from the \citet{Sandersonetal06} catalogue and for each of our simulations. We can see from this figure that our simulations capture the relationship between cooling time, core temperature and core size observed by \citet{Sandersonetal06}: warm cores can have short cooling times if they are compact. This is a straightforward consequence of the strong dependence of cooling on density and its weak dependence on temperature.
185:
186: Although compact cores are not necessarily cool they nevertheless have
187: short cooling times (even when disturbed), preserving the importance
188: of radiative cooling in establishing their structure and evolution
189: during and after a merger. Furthermore, although extended core
190: morphologies generally (but not necessarily) have longer cooling times
191: than compact core systems, only one of the cases we have studied produces
192: states with cooling times greater than $5$\Gyrs\ for any significant length of time:
193: our 3:1 $v_t/V_c=0.15$ case.
194:
195: \subsection{Temporal Evolution of Morphology}\label{analysis-states}
196:
197: To see how the morphologies of our simulations progress through these
198: three states, we present the temporal evolution of \rc~ and \DT~ in
199: Figs. \ref{fig-rc_t} and \ref{fig-Delta_t}. In both cases, a black
200: curve traces the system's evolution as seen in the $z$-projection with
201: coloured points plotted (following the format of
202: Fig. \ref{fig-delta_rc}) to indicate each interval's morphological
203: state. These curves change very little when viewed in $y$-projections
204: and differ only at early times in the $x$-projection due to the
205: apparently close proximity of the two cluster cores prior to
206: interaction (the system can appear as a ECC system in x-projections prior to \tclosest).
207: To clarify these plots, we show only the $z$-projection. We also
208: note \tclosest\ and \trelax\ for each simulation to indicate the interval over
209: which the system will tend to look disturbed.
210:
211: \subsubsection{core size}\label{analysis-states-core_size}
212:
213: In Fig. \ref{fig-rc_t} we consider the evolution of $r_c$ for our simulations.
214: We see that in all cases the system starts with a compact core (by
215: construction) and experiences a period of substantially increased core
216: size (often $r_c>300$\kpc) during first pericentric passage.
217: For the head-on mergers, we see that in the 1:1 and 3:1
218: cases this lasts (with varying amplitude) until roughly \taccrete\
219: while in the 10:1 case it lasts for only $\sim$$0.5$\Gyr. These
220: elevated values of $r_c$ are produced by the explosive expansion of
221: the core after the secondary's impact.
222:
223: In off-axis cases, these elevated values are primarily a product of
224: the close proximity of the cluster cores during pericentric passage.
225: In these cases there is an interval between \tclosest~ and \taccrete~
226: during which the system returns to a compact core state. This
227: indicates that the initial interaction of the cores fails to disrupt
228: the primary system to an extended core state in any of our off-axis
229: mergers.
230:
231: Following \taccrete\ when the remainder of the secondary core accretes
232: to the remnant core, many of our merger remnants are in, or nearly in,
233: extended core states. They subsequently evolve back towards their
234: initial core size (rapidly in some cases) with only the 3:1 head-on
235: and 1:1 and 3:1 $v_t/V_c=0.15$ cases spending prolonged periods of time in
236: extended core states. Of these, only the 3:1
237: $v_t/V_c=0.15$ case fails to formally recover a compact core by the end of the simulation. The remnant in this case reaches an extended core state with $r_c\sim140$\kpc\ following
238: \taccrete~ and evolves very little afterward.
239:
240: In Section \ref{analysis-structure} we will discuss the physical processes which shape the
241: diversity in remnant core states.
242:
243: \subsubsection{central temperature gradients}\label{analysis-states-temperature}
244:
245: In Fig.\ref{fig-Delta_t} we plot the evolution of $\Delta_T$ for our
246: simulations following the format of Fig. \ref{fig-rc_t}. In all
247: cases our systems begin with strong positive central temperature
248: gradients (by construction) which remains intact until $t_o$, modulo a very slow decline due to
249: the numerical effects discussed in Section \ref{numerics-cores}.
250:
251: \begin{figure*}
252: \begin{minipage}{175mm}
253: \begin{center}
254: \leavevmode \epsfysize=8.4cm \epsfbox{rc_t_astroph.ps}
255: \caption[Temporal evolution of core radius]{The temporal evolution of
256: core radius (\rc) for our simulations. Black curves illustrate
257: $r_c(t)$ and coloured points indicate the morphology of the system measured in the
258: $z$-projection. Compact core systems with low ($S_{40}<50$\keVcmsq) and high ($S_{40}>50$\keVcmsq) central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with low/high central entropies (\ie\ ECC and EWC systems) are in magenta and green respectively. The cut in $r_c$ used to separate compact
259: core systems from extended core systems is shown with dashed lines (see labels in central panel).
260: Vertical dotted lines indicate (from left to right) \tclosest,
261: \taccrete\ and \trelax\ for each case. The interval between \tclosest\ and \trelax\ indicates the time during which the
262: system will show signs of being disturbed. Black text around the boundary indicate the mass ratio and $v_t/V_c$ of the simulations presented in each
263: panel.}
264: \label{fig-rc_t}
265: \end{center}
266: \end{minipage}
267: \end{figure*}
268: \begin{figure*}
269: \begin{minipage}{175mm}
270: \begin{center}
271: \leavevmode \epsfysize=8.4cm \epsfbox{delta_t_astroph.ps}
272: \caption[Temporal evolution of central temperature gradient]{The
273: temporal evolution of the central temperature gradient (\DT)
274: following the same format as Fig. \ref{fig-rc_t}. Black curves
275: illustrate $\Delta_T(t)$ and coloured points indicate the morphology
276: of the system
277: measured in the $z$-projection. Compact core systems with low ($S_{40}<50$\keVcmsq) and high ($S_{40}>50$\keVcmsq) central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with low/high central entropies (\ie\ ECC and EWC systems) are in magenta and green respectively. The \DT$=0.05$ line roughly separating cool core systems from
278: warm core systems (classified by entropy) is shown with dashed lines (see labels in central panel). Vertical dotted lines
279: indicate (from left to right) \tclosest, \taccrete\ and
280: \trelax\ for each case. The interval between \tclosest\ and \trelax\ indicates the time during which the
281: system will show signs of being disturbed. Black text around the boundary indicate the
282: mass ratio and $v_t/V_c$ of the simulations presented in each panel.}
283: \label{fig-Delta_t}
284: \end{center}
285: \end{minipage}
286: \end{figure*}
287:
288: Considering the head-on mergers first, we see that in the rare 1:1
289: head-on case, $\Delta_T$ remains above $0$ at nearly all times.
290: In the more common 3:1 and 10:1 head-on cases, the system briefly develops a
291: negative central temperature gradient after \tclosest\ and oscillates
292: with a warm core two or three times on dynamical timescales.
293:
294: In off-axis cases, the primary core's positive central temperature
295: gradient survives until \taccrete. This indicates that the first
296: interaction of the cores in off-axis cases does not heat the primary
297: core to a warm state. At (or shortly following) \taccrete\ however, all of our
298: 1:1 and 3:1 mergers experience a rapid decline in \DT. In the 10:1 cases there is
299: little evolution in \DT\ following \taccrete\ due to the substantial disruption of the secondary system
300: prior to second pericentric passage in these cases.
301:
302: Hence, we find that cool cores resiliently maintain their
303: integrity until a merging system's core makes direct impact, either at
304: \tclosest\ in head-on cases or during second pericentric passage at
305: \taccrete\ in 1:1 and 3:1 off-axis cases. Extended cores are not generically
306: produced (except as very short lived transient states during which the
307: system would appear visibly disturbed) and the cores generally recover cool states once relaxed
308: (the most significant exceptions are the 3:1 head-on and $v_t/V_c=0.15$ cases).
309:
310: We explore the physical processes which shape the diversity of remnant core states in Section \ref{analysis-structure}.
311:
312: \subsection{The recovery of compact cool cores}\label{analysis-structure}
313:
314: As we can see from both Figs. \ref{fig-rc_t} and \ref{fig-Delta_t}, a
315: pattern emerges for the durations of EWC states (illustrated in green)
316: produced in our mergers: near-axis (head-on and $v_t/V_c=0.15$) and
317: intermediate mass ratios (3:1) are the most efficient cases for
318: producing this morphology. To understand this, we now seek to
319: explain two things: how the initial conditions of the merger (namely, the mass
320: ratio and orbits of the clusters) dictate the final remnant state
321: which emerge at \trelax\ and what properties of the remnant core's
322: structure principally govern the recovery of CCC morphologies.
323:
324: In essence, this reduces to an examination of the final balance
325: between shock heating and radiative cooling for material destined to
326: end-up in the core of the remnant. Thus, the pattern of sustained EWC
327: morphologies seen in Figs. \ref{fig-rc_t} and \ref{fig-Delta_t} can be
328: understood as the result of a ``Goldie Locks'' scenario where the
329: circumstances necessary for effective core heating are optimized while
330: the formation of dense structures capable of offsetting heating
331: through radiative cooling are prevented.
332:
333: These competing processes are both sensitive to the evolving dynamical
334: state of the initial systems' cores and the relative densities and
335: mach numbers of gaseous material forced to interact during the event.
336: We find that a great variety of processes are involved in establishing
337: this balance, with the effects of both shock heating and radiative
338: cooling being a sensitive function of the mass ratio and initial
339: orbits of the interacting clusters.
340:
341: For the remainder of this section we shall examine this variety.
342: However, as discussed in Section \ref{numerics-statistics}, our sparse
343: sampling of $M_p:M_s$ and $v_t/V_c$ prevents us from systematically
344: studying many of these processes in a quantitative way. We will thus
345: restrict ourselves to a heuristic discussion in what follows.
346:
347: \begin{figure*}
348: \begin{minipage}{175mm}
349: \begin{center}
350: \leavevmode \epsfysize=10cm \epsfbox{entropy_t_astroph.ps}
351: \caption[Evolution of central entropy]{The
352: temporal evolution of the central entropy (measured at $40$\kpc)
353: following the same format as Fig. \ref{fig-rc_t}. Black curves
354: illustrate $S_{40}(t)$ and coloured points indicate the morphology
355: of the system measured in the $z$-projection. Compact core systems with low ($S_{40}<50$\keVcmsq) and high ($S_{40}>50$\keVcmsq) central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with low/high central entropies (\ie\ ECC and EWC systems) are in magenta and green respectively. The cut in $S_{40}$ used in our morphological classification to separate cool core systems from
356: warm core systems is shown with dashed lines (see labels in central panel). Vertical dotted lines
357: indicate (from left to right) \tclosest, \taccrete\ and
358: \trelax\ for each case. The interval between \tclosest\ and \trelax\ indicates the time during which the
359: system will show signs of being disturbed. Black text around the boundary indicate the
360: mass ratio and $v_t/V_c$ of the simulations presented in each panel.}
361: \label{fig-entropy_t}
362: \end{center}
363: \end{minipage}
364: \end{figure*}
365:
366: \subsubsection{entropy and the formation of extended cores}\label{analysis-entropy}
367:
368: The temperature structure of a relaxed cluster's ICM is primarily set
369: by its gravitational potential. Because of this and the strong
370: $\rho_g^2$ dependence of the ICM's X-ray emissivity, variations in the
371: central X-ray surface brightnesses of clusters (and hence, variations
372: in $r_c$) are primarily a product of variations in central gas
373: densities. Because it is directly tied to the mass accretion history of the system and the
374: relative balance between heating (from accretion shocks and feedback) and radiative cooling,
375: these density variations are most straightforwardly interpreted in terms of variations in central (or minimum) gas entropies.
376:
377: In Fig. \ref{fig-entropy_t} we present the temporal evolution of the
378: central entropies of our simulations where we see a good
379: correlation between $S_{40}(t)$ and $r_c(t)$ following \trelax. Hence, the
380: surface brightness core radii of our merger remnants are tightly related to their
381: central entropy. We now address the question: what processes set this entropy?
382:
383: As in the plots of $r_c(t)$ and $\Delta_T(t)$, we observe important
384: differences in $S_{40}(t)$ between head-on and off-axis cases.
385: In the head-on cases, one very dramatic increase in $S_{40}$ occurs at
386: \tclosest\ followed by several small and erratic oscillations which
387: end $\sim1$\Gyr\ before \trelax. The details of this evolution depend
388: quite sensitively on the mass ratio of the event. Examining the
389: dynamical evolution of these mergers in the spatial distribution of
390: their surface brightness, temperature and entropy (Figs. 1 to 9 in
391: Poole06) we can see why this is so. In the 3:1 and 10:1 cases, the
392: secondary core fully penetrates the primary core triggering explosive shocks which
393: heat the entire primary core from the inside out. In the
394: 1:1 case, the merging cores inelastically collide and remain in very
395: dense structures following \tclosest. This has the dual effect of
396: reducing the efficiency of entropy production through shocks and
397: increasing the rate of radiative cooling. Together, these effects
398: lead to a counter intuitive result: the 3:1 and 10:1 head-on mergers
399: ultimately heat the primary core more than the 1:1 head-on merger.
400:
401: In off-axis cases, there are slight and brief increases in central
402: entropy at \tclosest\ in some cases, but there is generally very
403: little sustained increase in $S_{40}$ until \taccrete. This is
404: because the initial interaction of the systems has very little effect
405: on the primary core and is consistent with what we noted in Section
406: \ref{analysis-states}: the primary core retains its initial CCC
407: morphology until second pericentric passage in off-axis cases. In the
408: 1:1 cases, the secondary core remains significantly intact until
409: \taccrete\ and we thus see a significant jump in entropy. In other words: the evolution of the 1:1 $v_t/V_c=0.15$ case following \taccrete\
410: is very similar to the 3:1 head-on case following \tclosest. In the 10:1
411: off-axis cases, the secondary core is so significantly disrupted by
412: \taccrete\ that essentially no evolution in central entropy is seen
413: throughout the entire simulation.
414:
415: In the 3:1 off-axis cases, an evolution distinctly different from the 1:1 and 10:1
416: off-axis cases is observed. In Poole06 we used our 3:1 $v_t/V_c=0.15$ case to illustrate how an
417: off-axis merger of a cold-core system can generate low-entropy streams
418: of material which accrete to the remnant core with high velocities.
419: Fig. \ref{fig-entropy_t} illustrates that the accretion of these
420: streams can have a significant impact on the central entropy of
421: off-axis merger remnants. In both off-axis 3:1 cases we see a very
422: low amplitude jump at \taccrete\ when the surviving portion of the
423: secondary core returns for its second pericentric passage. It is not
424: until significantly after \taccrete\ that peak levels of $S_{40}$ are
425: reached. These peaks occur roughly when the streams formed from the
426: secondary cores are accreting most significantly onto the remnant.
427:
428: The heating effects of an accreting stream depend sensitively on its
429: total mass, density, and mach number as well as the state of the
430: remnant core onto which it falls. The ram pressure and tidal forces
431: experienced by the cores of interacting systems, which are responsible
432: for stripping the secondary, forming the stream and disturbing the
433: primary core, depend very sensitively on the initial mass ratio and
434: impact parameter of the merger. The stream's mach number during
435: accretion also depends on the amount of dynamical friction experienced
436: by the secondary system during its orbit, and thus, the mass ratio of
437: the interacting systems. Hence, the heating effects of stream
438: accretion are very sensitive to the initial mass ratio and orbits of
439: the interacting clusters.
440:
441: Thus, the central entropy of the remnant, which is a very sensitive
442: function of the mass ratio and initial orbit of the merger, plays a
443: key role in setting its core size. In the cases
444: we have studied, the only long lived extended core remnant results from a slightly
445: off-axis 3:1 merger for which stream heating is the dominant mode of
446: core entropy production.
447:
448: \subsubsection{cooling timescales and cool core recovery}\label{analysis-timescales}
449:
450: Radiative emission can reduce the temperature of material, but must it do so?
451: Imagine a disturbed parcel of low-entropy gas radiatively cooling as it falls to the center of the system. As it cools, its
452: entropy will fall on its cooling timescale, $t_{cool}\propto
453: S^{1.5}T^{-1}$. However, because of its dynamically disturbed state,
454: its temperature can drop only if the cooling timescale is sufficiently
455: short with respect to its local dynamical time, $t_{dyn}\propto (G
456: \Delta \rho_c)^{-0.5}$. If $t_{cool}\gg t_{dyn}$, temperature increases from
457: compression dominate over radiative losses. Under these
458: circumstances, the temperature can only fall once the material reaches a quasi-static state.
459: It will then do so on timescales governed by the cooling time.
460:
461: At all times following \taccrete\ and for all the cases we have studied, we find that
462: $t_{cool}\gg t_{dyn}$ within $r_c$ (by approximately an order of magnitude) for our disturbed but relaxing remnants. As a
463: result, cooling does not lower our remnant cores' temperatures until
464: they have returned to quasi-stable states of hydrostatic equilibrium and a substantial fraction of a
465: cooling time has passed. In Poole06 we examined how a system returns to hydrostatic equilibrium
466: following a merger. We showed that our remnants generally return to
467: within $10$\% of hydrostatic equilibrium at $R_{500}$ $0.5-1.0$\Gyrs\ following
468: the system's apparent relaxation at \trelax. Examining this behavior at $R_{2500}$
469: (roughly $r_c$) we find that the system returns to within $10$\% of hydrostatic equilibrium
470: a little sooner and generally within $0.5$\Gyrs\ of when it looks relaxed at \trelax. Hence, cooling
471: generally will not have a significant effect on central remnant temperatures until a time $t_{cool}$
472: (\ie\ $\sim2-5$\Gyrs) after appearing relaxed at \trelax. This accounts for the occurrence of compact
473: warm core (CWC) morphologies as well as the timing of the late increases in \DT\ occurring in several of our 1:1 and 3:1 merger
474: remnants (whose central cooling times are short).
475:
476: To conclude, we find that following a disturbance from a merger, the initially dense cores
477: of CCC systems generally recover much faster (roughly a few dynamical times following $t_o$)
478: than their cool temperatures (roughly $t_{cool}$ afterward). The existence
479: of compact warm core (CWC) systems observed by \Chandra\ and \XMM\ \citep{Sandersonetal06,Prattetal06} can
480: thus be understood naturally within the framework of hierarchical structure formation without appealing to other heating
481: sources such as AGN. Their existence is merely a natural consequence of hierarchical structure formation
482: and the fact that $t_{cool}\gg t_{dyn}$ in cluster cores.
483:
484: \subsection{Morphological segregation in scaling relations}\label{analysis-segregation}
485:
486: Scaling relations between mass, X-ray temperature and luminosity are
487: powerful tools for studying the processes that shaped the structure of
488: the ICM. It has been known for some time
489: \citep{Fabianetal94,Markevitch98} that a significant fraction of the
490: scatter in these relations is due to variations in the core properties
491: of clusters. This has been supported by recent theoretical
492: \citep[][Poole07, McCarthy et al. 2007]{OHaraetal05} and observational studies
493: \citep{Sandersonetal06,Prattetal06}. For this reason, and because the
494: generation of accurate mass functions from scaling relations has been a
495: significant motivator for their study, past analyzes have typically
496: been conducted with the central cores of systems excised to reduce scatter.
497:
498: However, it has been shown by M04 that there is interesting structure
499: in these relations which correlate with the core properties of
500: clusters. They use the ACC catalogue of \citet{Horner_thesis} to
501: compare their analytic entropy injection model to published \Chandra~
502: and \XMM~ temperature profiles, classifying them by the presence or
503: absence of central temperature gradients. They further separate these
504: systems into classes showing evidence of significant or little
505: substructure. Despite the small number of published \Chandra~ and
506: \XMM~ temperature profiles available to them at the time, they found
507: that these classes clearly (but not necessarily distinctly) segregate
508: in both the \LT~ and \betarc~ planes.
509:
510: In Poole07 we examined the evolving global properties (including
511: X-ray temperature and luminosity) of our simulations, as well as the
512: scaling relations generated from them. Having quantified the evolving
513: morphologies of our cluster cores, we now examine how the three
514: morphology classes we have identified segregate in these planes. In
515: Poole07 we found that the evolution in the $L_x-T_x$, $M_t-L_x$,
516: $M_t-T_x$, $SZ-L_x$ and $SZ-T_x$ planes exhibited a generic pattern
517: and we find the same for morphological segregation in these planes as
518: well. For this reason we will only examine the $L_x-T_x$ and $\beta-r_c$ relations here.
519:
520: \subsubsection{the $L_x-T_x$ plane}\label{analysis-LT_segregation}
521:
522: \begin{figure*}
523: \begin{minipage}{175mm}
524: \begin{center}
525: \leavevmode \hspace*{-1cm} \epsfysize=19cm \epsfbox{L_T_morphology_astroph.ps}
526: \caption[$L_x$ against $T_x$ illustrating morphological
527: segregation in this plane]{A plot of $L_x$ against $T_x$ illustrating
528: the morphological segregation in this plane.
529: Compact core systems with low ($S_{40}<50$\keVcmsq) and high ($S_{40}>50$\keVcmsq) central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with high central entropies (\ie\ EWC systems) are in green. Filled circles and crosses mark systems which appear relaxed
530: or unrelaxed (respectively) in simulated $50$\ks\ \Chandra\
531: observations. Black dots are the
532: observed catalogue of \citet{Horner_thesis} with open circles
533: marking systems with extended cores ($r_c>100$\kpc) and filled
534: circles marking systems with compact cores ($r_c<100$\kpc). The
535: black dashed line is a fiducial entropy injection model with
536: $S_o=150$\keVcmsq~ (see Poole07 for details). Black text around
537: the boundary indicate the mass ratio and $v_t/V_c$ depicted by each
538: panel.}
539: \label{fig-L_T}
540: \end{center}
541: \end{minipage}
542: \end{figure*}
543: \begin{figure*}
544: \begin{minipage}{175mm}
545: \begin{center}
546: \leavevmode \hspace*{-1cm} \epsfysize=19cm \epsfbox{beta_rc_astroph.ps}
547: \caption[$\beta$ against $r_c$ illustrating the morphological
548: segregation in this plane]{A plot of $\beta$ against $r_c$
549: illustrating the segregation of morphology and of long and short cooling time systems
550: in this plane. Compact core systems with low ($S_{40}<50$\keVcmsq) and high ($S_{40}>50$\keVcmsq) central entropies (\ie\ CCC and CWC systems) are labeled in blue and red respectively while extended core systems with low/high central entropies (\ie\ ECC and EWC systems) are in magenta and green respectively. Filled circles and crosses mark systems
551: which appear relaxed or unrelaxed (respectively) in simulated
552: $50$\ks~ \Chandra~ observations with the system placed at $z=0.1$.
553: Black dots
554: are the observed catalogue of \citet{Horner_thesis} with open
555: circles marking systems with long cooling times ($t_{cool}>5$\Gyrs)
556: and filled circles marking systems with short cooling times
557: ($t_{cool}<5$\Gyrs). Dashed lines mark our $r_c=100$\kpc\ cut separating compact from extended core systems (see labels in central panel).
558: Black text around the boundary indicate the
559: mass ratio and $v_t/V_c$ depicted by each panel.}
560: \label{fig-beta_rc}
561: \end{center}
562: \end{minipage}
563: \end{figure*}
564:
565: In Fig. \ref{fig-L_T} we present the positions of our systems in the
566: $L_x-T_x$ plane, at each time in y and z-projections,
567: when they appear as CCC (blue), CWC (red) or EWC (green) systems. The
568: initial and final states of the systems are indicated with black four
569: and five point stars respectively. These are compared against the
570: observed catalogue of \citet[][in black]{Horner_thesis} for which we use the catalog of ROSAT 1-D $\beta$-model fits from \citet{ReiprichandBohringer02}
571: to identify systems with compact cores ($r_c<100$\kpc) using filled
572: circles and systems with extended cores ($r_c>100$\kpc) using open
573: circles.
574:
575: This plot reinforces what was found in Poole07: the low-luminosity
576: side of the $L_x-T_x$ scaling relation can not be significantly
577: populated with the mergers we have studied. It is comforting to note
578: that our systems remain confined within the scatter of the
579: observations at all times, but instances of systems lying near the low
580: luminosity side of this scatter are rare and short lived.
581:
582: Comparing the regions occupied by systems which exhibit compact and
583: extended cores, we find a good agreement between the observations and
584: our simulations (with the exception of the rare 1:1 off-axis mergers,
585: which yield much more under-luminous compact core systems). Furthermore, there is a very
586: clean divide between warm and cool cores in this plane. We have plotted a fiducial
587: entropy injection model with a minimum entropy of $S_o=150$\keVcmsq\
588: \citep[computed following the general approach of][see Poole07 for
589: some additional details]{MBBPH} to roughly quantify the location of this divide.
590:
591: For our simulations, instances of disturbed morphology generally lie
592: near or above this fiducial line while relaxed morphologies lie below it. This is consistent
593: with common anecdotal observations that extended core systems
594: generally appear disturbed. However, our results suggest that disturbed
595: morphologies may correlate better with the presence of warm cores than
596: with the presence of extended cores. It is important to emphasize however
597: that our simulations do not populate the low-luminosity side of the scatter in this plane nor do they yield relaxed extended core systems such as Abell 2034 or Abell 2631. Trends such as these suggest that the observed scatter in the $L_x-T_x$ plane is a result of physical processes other than mergers.
598:
599: We have similarly examined the other scaling relations discussed in
600: Poole07 and find the same qualitative behavior in all
601: cases.
602:
603: \subsubsection{the $\beta-r_c$ plane}\label{analysis-beta_rc_segregation}
604:
605: In Fig. \ref{fig-beta_rc} we present the positions our simulations occupy in
606: the $\beta-r_c$ plane, at each time in y and z-projections,
607: when they appear as CCC (blue), CWC (red), ECC (magenta) or EWC (green)
608: systems. Instances of relaxed-looking morphology are marked with solid circles and disturbed morphology with crosses.
609: The initial and final states of the systems are indicated
610: with black four and five point stars respectively. These are compared
611: against the observed catalogue of \citet{ReiprichandBohringer02} (black dots).
612: We identify systems with short cooling times ($t_{cool}<5$\Gyrs)
613: with large filled circles and systems with long cooling times
614: ($t_{cool}>5$\Gyrs) with large open circles. These cooling times are obtained by combining temperatures from \citet{Horner_thesis} with denisties derived from \citet{ReiprichandBohringer02}.
615:
616: We can see from this plot that $\beta$ and $r_c$ express a clear
617: correlation in the observations and that our mergers
618: reproduce some of its features. In particular, our simulations account
619: for the high-$\beta$ envelope of the observed scatter in this plane. Furthermore,
620: our simulations populate the narrow region occupied by the observations
621: at the largest values of $r_c$. While they significantly populate this region, they
622: are always visibly disturbed. Observed systems in this region for which we have
623: dynamical classifications \citep{MBBPH} are similarly all disturbed.
624:
625: However, our simulations fail to produce states which last for significant lengths of time in the
626: region of this plane occupied by the vast majority of observed extended core systems. Most
627: observed extended core systems have values of $\beta$ substantially less than the states
628: produced by our mergers.
629:
630: In Fig. \ref{fig-tcool_DeltaT} we show that our mergers rarely produce
631: systems with cooling times in excess of $5$\Gyrs. This is primarily a
632: consequence of the cores of our systems remaining compact for most of
633: their evolution. We can see in this plane that observed systems with short cooling times
634: are generally found along the high-$\beta$ side of the observed scatter in this plane, in
635: good accord with our simulations. Observed systems with $t_{cool}>5$\Gyrs\ lie in the
636: low-$\beta$/large-$r_c$ region of the plane which our simulations fail to significantly populate.
637:
638: To conclude, our simulations are able to account for the high-$\beta$ envelope occupied by
639: observed systems with short cooling times as well as the tail of disturbed systems at the highest values
640: of \rc. They do not account for the bulk of observed extended core systems with long cooling times.
641:
642: \begin{figure*}
643: \begin{minipage}{175mm}
644: \begin{center}
645: \leavevmode \epsfysize=10cm \epsfbox{Z_r.ps}
646: \caption[Initial and final metallicity gradients]{A plot comparing the
647: initial and final metallicity gradients of our simulations. Initial
648: profiles of the primary are shown with thick red lines and the final
649: remnant profile is shown with a thick black line. The final
650: distribution of material originating in the secondary is shown with
651: a thin blue line. In all cases, the metallicity profiles of our
652: systems are initialized to have the metallicity profiles observed by
653: \citet{DeGrandiandMolendi02}.}
654: \label{fig-Z_r}
655: \end{center}
656: \end{minipage}
657: \end{figure*}
658:
659: \section{Metallicity gradients}\label{analysis-mixing}
660:
661: It is a well established fact that extended core systems tend to lack
662: metallicity gradients (with mean amplitudes of $0.1-0.2Z_\odot$) while
663: CCC systems routinely exhibit centrally peaked metallicity gradients
664: within $0.2R_{180}$, often reaching central values which approach
665: solar metallicity \citep{DeGrandiandMolendi02}. This, along with
666: anecdotal evidence suggesting that extended core systems tend to be
667: preferentially disturbed has been used to claim, as a result, that
668: mergers effectively mix the ICM.
669:
670: To examine this hypothesis, we have initialized our systems to have
671: the metallicity profiles presented by \citet{DeGrandiandMolendi02} and
672: examined the resulting metallicity profiles of our merger remnants. We
673: present these results in Fig. \ref{fig-Z_r} where the initial
674: metallicity profiles of our primary systems are presented in blue and
675: the final metallicity profiles of our remnants in red.
676:
677: It is clear from this figure that mergers do not efficiently mix the
678: material of the initial systems. This makes sense given the fact that
679: the enhanced metallicity is initially tied to the lowest entropy
680: material in the clusters. This material is not preferentially heated
681: above the entropy of the surrounding material and thus, remains the
682: low entropy material constituting the remnant. Once the system
683: relaxes, the low-entropy/high-metallicity material sinks to the center
684: of the system due to a lack of buoyant support, reestablishing a
685: central metallicity peak with only a slightly lower amplitude and slightly broader
686: distribution.
687:
688: