1:
2: % aa.dem
3: % AA vers. 6.1, LaTeX class for Astronomy & Astrophysics
4: % demonstration file
5: % (c) Springer-Verlag HD
6: % revised by EDP Sciences
7: %-----------------------------------------------------------------------
8: %
9: %\documentclass[referee]{aa} % for a referee version
10: %\documentclass[onecolumn]{aa} % for a paper on 1 column
11: %\documentclass[longauth]{aa} % for the long lists of affiliations
12: %\documentclass[rnote]{aa} % for the research notes
13: %\documentclass[letter]{aa} % for the letters
14: %
15: \documentclass{aa}
16: %\documentclass[referee]{aa}
17: \usepackage{graphicx}
18: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
19: \usepackage{txfonts}
20: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
21: %
22: \def\simlt{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}
23: \def\simgt{\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}}
24: \def\simpropto{\lower.2ex\hbox{$\; \buildrel \propto \over \sim \;$}}
25:
26:
27:
28: \begin{document}
29:
30: \title{Phase-Space Evolution of Dark Matter Halos}
31:
32:
33: \author{S. Peirani\inst{1,}\inst{2} \and J. A. de Freitas Pacheco\inst{3} }
34:
35: \institute{
36: Institut d'Astrophysique de Paris, 98 bis Bd Arago, 75014
37: Paris, France - UMR 7095 CNRS
38: - Universit\'e Pierre et Marie Curie\\
39: \email{peirani@iap.fr}
40: \and
41: Department of Physics, University of Oxford,
42: Denys Wilkinson Building,
43: Keble Road, Oxford OX1 3RH, UK
44: \and
45: Observatoire de la C\^ote d'Azur -
46: Laboratoire Cassiop\'ee - UMR 6202 -
47: BP 4229 - 063
48: 04 - Nice Cedex 4 - France\\
49: \email{pacheco@oca.eu}
50: }
51:
52:
53:
54: \date{Received ..., ...; accepted ..., ...}
55:
56: % \abstract{}{}{}{}{}
57: % 5 {} token are mandatory
58:
59: \abstract
60: % context heading (optional)
61: % {} leave it empty if necessary
62: {
63: In a Universe dominated by dark matter, halos are continuously accreting
64: mass (violently or not) and such mechanism affects their dynamical state.
65: }
66: % aims heading (mandatory)
67: {
68: The evolution of dark matter halos in phase-space, and using
69: the phase-space density indicator $Q=\rho/\sigma^3$ as a tracer, is discussed.
70: }
71: % methods heading (mandatory)
72: {
73: We have performed cosmological $N$-body simulations from which we have carried
74: a detailed study of the evolution of $\sim 35$ dark halos in the interval
75: $0\leq z \leq 10$.
76: }
77: % results heading (mandatory)
78: {
79: The follow up of individual halos indicates two distinct evolutionary
80: phases. First, an early and fast decrease
81: of $Q$ associated to virialization after the gravitational collapse
82: takes place. The nice
83: agreement between simulated data and theoretical expectations based on
84: the spherical collapse model support such a conjecture. The late and long period where
85: a slow decrease of the phase-space density occurs is related to accretion and
86: merger episodes. The study of some merger events in
87: the phase-space (radial velocity versus radial distance)
88: reveals the formation of structures quite similar to caustics generated in
89: secondary infall models of halo formation.
90: After mixing in phase-space, halos
91: in quasi-equilibrium have flat-topped velocity
92: distributions (negative kurtosis) with respect
93: to Gaussians. The effect is more noticiable for
94: captured satellites and/or substructures
95: than for the host halo.
96: }
97: % conclusions heading (optional), leave it empty if necessary
98: {}
99:
100:
101:
102:
103:
104: \keywords{dark matter -- Galaxies: halos -- Galaxies: interactions --
105: Methods: N-body Simulations
106: }
107:
108:
109: \maketitle
110: %
111: %________________________________________________________________
112:
113:
114:
115: \section{Introduction}
116:
117: In the cold dark matter paradigm, galaxies are formed when baryonic gas
118: falls into the gravitational
119: potential well of cold dark matter (CDM) halos. These halos evolve by
120: accreting mass either by quasi continuous
121: processes or by merger events, when their masses vary sudden and significantly. Mass
122: accretion is also an important mechanism by which halos acquire angular momentum after
123: turnaround, since during the collapse phase tidal torques become inefficient
124: (Peirani et al. 2004). Moreover, Wechsler et al. (2002)
125: found an important correlation between accretion and the halo concentration.
126: Halos in a state of high accretion have central densities related
127: to the background density while those with low infall rates have approximately
128: constant central densities.
129:
130: After a merger episode, the resulting halo is not in equilibrium. Rapid
131: variations in the gravitational potential
132: contribute to the relaxation of the resulting system through a well known
133: mechanism dubbed ``violent
134: relaxation'' (Lynden-Bell, 1967). This process, after a few dynamical time
135: scales ($t_{dyn} \sim 1/\sqrt{G\bar\rho}$),
136: produces a smooth mass-independent distribution function (DF) as a result of the
137: gravitational scattering of particles.
138: Violent relaxation leads to a more mixed
139: system (Tremaine et al. 1986), reducing the value of the
140: coarse-grained DF. Mixing effects were also considered by Dehnen (2005), who
141: has investigated the behavior of the {\it excess-mass function $D(f)$} defined
142: in a phase-space volume where the coarse-grained distribution
143: function is greater than a given value $f$. For central density profiles represented
144: by a power law ($\rho\propto r^{-\gamma}$), Dehnen (2005) found that
145: $D \propto f^{-2(3-\gamma)/(6-\gamma)}$, suggesting that steeper cusps are less
146: mixed that shallower ones. As a corollary, if halos having different power law
147: density profiles merge, the resulting halo cannot have a cusp steeper than those
148: of the progenitors. A similar approach was adopted by Arad et al. (2004), who
149: defined instead the function $v(f)$ such as $v(f)df$ be the phase-space volume occupied
150: by phase-space elements where the density lies in the range $f, f+df$. They have
151: found from cosmological simulations that $v(f)$ is quite well described by a
152: power law, e.g., $v(f) \propto f^{-2.5\pm 0.05}$, over three to five decades in $f$.
153: According to them, such a power law behavior reflects the halo substructure, consequence
154: of the hierarchical clustering and not the result of violent relaxation.
155:
156: Investigations on the resulting equilibrium state after a merger episode, based
157: on numerical simulations, use
158: in general a more simple estimator for the coarse-grained DF defined
159: as $Q = \rho/\sigma^3$, where $\rho$ and $\sigma$
160: are respectively the density and the cube of the 1-D velocity dispersion of
161: dark matter particles inside a considered
162: volume. High resolution simulations of galaxy-size CDM halos indicate an
163: increase of $Q$ towards the center
164: (Taylor \& Navarro, 2001) and similar results were obtained
165: by Boylan-Kolchin \& Ma (2004), Rasia et al. (2004), Ascasibar et al. (2004),
166: Dehnen \& McLaughlin (2005), Hoffman et
167: al. (2007), Ascasibar \& Gottl{\"o}ber (2008), Vass et al (2008a, 2008b),
168: who have
169: also obtained a power-law variation, e.g., $Q \propto r^{-\beta}$ for
170: cluster-size halos, with $\beta$ quite close
171: to the value found by Taylor \& Navarro, namely, $\beta \approx$ 1.87.
172: Moreover, similar trends have been found by Knollmann, Knebe \& Hoffman (2008)
173: considering different cosmogonies although they found that the slope $\beta$
174: depends on the concentration parameter of dark matter halos.
175: No adequate explanation presently exists
176: for such a power-law profile which describes the phase-space density profile in dark
177: halos. However, Austin et al. (2005)
178: used semi analytic extended secondary infall models to show that such a behavior
179: is a robust feature of
180: virialized halos, which have attained equilibrium via violent relaxation
181: and not the result of hierarchical merging, a conclusion in contradiction with
182: that by Arad et al. (2004).
183:
184: Analyses of the phase-space density in the core of dwarf spheroidal galaxies, rotating
185: dwarfs, low surface brightness
186: galaxies and clusters suggest another
187: scaling law, e.g., $Q \propto \sigma^{-n}$, with $n \sim$ 3-4 (Dalcanton
188: \& Hogan, 2001). This scaling can be understood if the merging halos were
189: initially close to equilibrium and
190: if the fusion process preserves approximately the physical density as
191: each layer is homologously added to
192: form the new system (Dalcanton \& Hogan, 2001). A scaling relation
193: close to $Q \propto \sigma^{-3}$ was
194: obtained from cosmological simulations by Dav\'e at al. (2001) for collisionless
195: as well as
196: for self-interacting dark matter (SIDM). This result for SIDM is unexpected since
197: in this case the material should be
198: compressed to higher densities during a merger event, sinking to where it
199: reaches the local pressure
200: equilibrium and where the specific entropy matches. Therefore mergers should
201: occur at nearly constant $Q$ rather than constant density.
202:
203: In a previous investigation (Peirani et al. 2006,
204: hereafter PDP06), we have
205: reported cosmological simulations aiming to study the evolution
206: of the phase-space density
207: indicator $Q$ in core of dark matter halos. Halos were classed in two
208: different catalogs: the {\it accretion} , comprising
209: 781 objects which have never undergone a major merger event and whose
210: masses varied continously and smoothly. The
211: {\it merger catalog} contains 567 halos which had at least one major
212: merger event, corresponding to an
213: increase of their masses at least by a factor 1/3. These simulations
214: indicate that the phase-space density
215: decreases continuously in time, scaling with the velocity dispersion
216: as $Q \propto \sigma^{-2.1}$ and with the
217: halo mass as $Q \propto M^{-0.82}$. No differences in these scaling relations
218: were seen between ``cold" and ``warm"
219: dark matter models but halos which have underwent
220: important merger events are, on the average, more relaxed
221: having $Q$ values lower than halos of
222: the accretion class. The follow up of individual halos indicates an early and
223: fast phase in which $Q$
224: decreases on the average by a factor of 40 followed by a long period in
225: which $Q$ further decreases by about factor
226: of 20. The decrease of $Q$ (or the {\it increase} in the entropy) in the
227: first phase is probably a consequence
228: of the randomization of bulk motions during the first shell crossing while
229: accretion and merger events are responsible
230: for the slow decrease observed in late epochs.
231:
232:
233:
234: In the present work we report a detailed investigation of 35 CDM halos whose
235: evolution was followed from $z \sim 10$
236: up today. For each of these halos, we have estimated the redshift at which
237: the first shell crossing occurs and which
238: we assume to coincide with ``virialization". The phase-space density
239: indicator $Q$ in core of these halos was estimated
240: at that moment and compared with theoretical estimates performed through
241: the spherical model. The agreement between
242: theoretical and numerical estimates confirms our previous conjecture
243: concerning the rapid decrease of $Q$ observed
244: in the early evolutionary phases. We have also studied the late behavior
245: in some specific examples to show
246: the evolution of the velocity distribution during a merger event and
247: the evolution of structures in the phase-space.
248: This paper is organized as follows: in Section 2 we describe
249: briefly the simulations, in Section 3 the phase-space
250: density at ``virialization" is discussed, in Section 4 we describe
251: some merger episodes and the behavior of the
252: velocity distribution as well as of structures in phase-space
253: and, finally, in Section 5 we present our
254: main conclusions.
255:
256: \section{The simulations}
257:
258: In the present work we have used the same halo catalogs as in PDP06, including
259: objects with masses in the
260: range $10^{10}-10^{13}\,M_{\odot}$. For the sake of completeness, we summarize
261: here the main steps
262: performed to prepare these catalogs.
263:
264: The N-body simulation uses the adaptive particle-particle/particle-mesh ($AP^3M$)
265: code HYDRA (Couchman et al. 1995). The adopted cosmological parameters
266: were h = 0.65, $\Omega_m = 0.3$ and
267: $\Omega_{\Lambda} = 0.7$, with the power spectrum
268: normalization $\sigma_8 = 0.9$. The simulation was
269: performed in a box of side 30$h^{-1}$ Mpc, including 256$^3$
270: particles, corresponding to a mass resolution
271: of $2.05\times 10^8\,M_{\odot}$. The simulation started at $z=50$ and
272: ended at the present time ($z=0$).
273: Halos were initially detected by using a friends-of-friends (FOF)
274: algorithm (Davis et al. 1985) and, in a second step, unbound
275: particles were removed by an iterative procedure. Thus, our selected halos
276: are all gravitationally bound
277: objects. In the total, 1348 halos were detected, constituting the two aforementioned
278: catalogs (PDP06).
279: The evolution of these objects was chased from $z=3.5$ until $z=0$, but
280: about 40 halos with initially enough
281: particles (N$\geq$50) were followed during a longer time
282: interval ($10\leq z \leq 0$). For each halo, we have monitored the evolution of the
283: mass, density, virial ratio and velocity dispersion.
284:
285: \section{The phase-space density at virialization}
286:
287: According to the analysis by PDP06, the phase-space density $Q$ decreases
288: continuously as halos accrete matter. The derivative $dQ/dt$ shows the
289: presence of ``valleys" which coincide with peaks
290: in the derivative $d\sigma/dt$ of the velocity dispersion. These structures are
291: always associated
292: with sudden variations in the mass, induced by merger events and leading
293: to a slight
294: ``heating", followed by a relaxation of the system, measured by the decrease in
295: the phase-space density indicator $Q$, which is equivalent to say that the
296: entropy of the system
297: increased. If this behavior characterizes in general the late evolution of
298: halos, the first and
299: strong peak in the derivative of the velocity dispersion, correlated with the
300: first and deep valley in
301: the derivative of the phase-space density (see Fig.~4 in PDP06), is not a
302: consequence of merging, but probably a consequence of the gravitational
303: collapse. If this interpretation is correct, an important
304: energy transfer from bulk to random motions should occur due to collective
305: effects, heating all the particles, regardless their initial energies.
306:
307: In order to test our conjecture, we have first derived for a sample of 35 halos,
308: which have been followed in the interval $10\leq z \leq 0$, which includes the redshift
309: at which the first shell crossing
310: occurs. We assumed here, as in Sugerman et al. (2000), that
311: the first shell crossing episode coincides
312: with the time of maximum velocity dispersion as well as with ``virialization''. In
313: fact, as halos
314: accrete mass, the virial ratio $2T/\mid W\mid$ approaches asymptotically the
315: unity and, just
316: after the first shell crossing, they are only in quasi dynamical equilibrium. For
317: our subsequent
318: analysis, we assume for simplicity that the relaxation process was almost achieved.
319:
320:
321:
322: \begin{figure}
323: \rotatebox{0}{\includegraphics[width=\columnwidth]{f1.eps}}
324: \caption{The halo mass as a function of the redshift $z_v$ at which ``virialization''
325: occurs (defined as the time of maximum velocity dispersion). The solid line
326: represents the best fit solution.}
327: \end{figure}
328:
329:
330: \begin{figure}
331: \rotatebox{0}{\includegraphics[width=\columnwidth]{f2.eps}}
332: \caption{The phase-space density Q as a function of the halo mass. The points
333: represent values derived from simulated data whereas the solid line is obtained
334: from relation (8).}
335: \end{figure}
336:
337:
338:
339: In fig. 1, the halo mass is plotted as a function of the redshift $z_v$ at
340: which ``virialization'' occurs.
341: Notice that, as expected in the hierarchical picture, less massive halos collapse
342: first and that
343: large galaxy-size halos ($\sim 10^{12}\,M_{\odot}$) are being
344: virialized around $z_v \sim 1-2$.
345: A best fit of these simulated data gives
346: \begin{equation}
347: M_{11}=\frac{185}{(1+z_v)^{2.86}}
348: \end{equation}
349: where $M_{11}$ stands for the halo mass in units of $10^{11}\,M_{\odot}$.
350:
351: The phase-space density $Q$ at ``virialization'' was calculated
352: as follows: for each halo, the gravitational
353: radius $r_g = GM/\mid W\mid$ was computed as well as the mean
354: density $\bar\rho = 3M/(4\pi r_g^3)$. The
355: 1-D velocity dispersion $\sigma$ was calculated by assuming isotropy, e.g.,
356: $\sigma = \sqrt{(\sigma_x^2 + \sigma^2_y + \sigma^2_z)/3}$ and,
357: finally the phase-space density indicator, $Q=\bar\rho/\sigma^3$.
358:
359: In order to perform a theoretical estimate of the value of the phase-space
360: density $Q$ just after
361: virialization, we have appealed to the spherical collapse model, including
362: effects of a cosmological constant, in order to be consistent with the cosmology
363: adopted in our simulations.
364:
365: The expected density contrast at virialization, using the results of
366: Bryan \& Norman (1998), is
367: approximately given by
368: \begin{equation}
369: \Delta_v = 18\pi^2+82\lbrack \Omega_m(z_v)-1\rbrack - 39\lbrack \Omega_m(z_v)-1\rbrack^2
370: \end{equation}
371: Thus, the mean halo density at virialization is
372: \begin{equation}
373: \rho_v =\Delta_v\frac{3H_0^2}{8\pi G}\Omega_m^0(1+z_v)^3=\frac{3M}{4\pi R_v^3}
374: \end{equation}
375: where $\Omega_m^0$ is the present matter (dark+baryonic) density parameter and $R_v$ is
376: the radius at virial equilibrium.
377:
378: Defining the ratio between the virialization and the turnaround
379: radii as $\eta=R_v/R_0$ and using the
380: energy conservation as well as the virial relation, one obtains (Lahav et al. 1991)
381: \begin{equation}
382: \lambda\eta^3-2(2+\lambda)\eta+2=0
383: \end{equation}
384: where we have introduced the parameter $\lambda = \Omega_vH^2_0R^3_0/GM$.
385: The solution of this cubic equation can approximately be expressed by
386: \begin{equation}
387: \eta = 0.5 - 0.20936\lambda + 0.04949\lambda^2
388: \end{equation}
389: Since the ratio between densities at turnaround and
390: virialization is $\rho_{ta}/\rho_v = \eta^3$,
391: using eq.(3) and the definition of the parameter $\lambda$ one obtains
392: \begin{equation}
393: \lambda\eta^3= \frac{3\Omega_vH^2_0}{4\pi G\rho_v}
394: \end{equation}
395:
396: On the other hand, since $3\sigma^2=-2E$, where $E$ is the total energy, we
397: can express the
398: 1-D velocity dispersion as
399: \begin{equation}
400: \sigma^3=\lbrack\frac{2}{3\lambda^{1/3}}+
401: \frac{\lambda^{2/3}}{3}\rbrack^{3/2}(\Omega_vH_0^2)^{1/2}GM
402: \end{equation}
403:
404: Therefore, the procedure adopted to estimate $Q$ was: for a given virialization
405: redshift $z_v$, one computes
406: the virialization density $\rho_v$ from eqs.~(2) and (3). Then, using eqs.~(5) and (6)
407: the value of $\lambda$
408: at the considered $z_v$ is derived. This value is used in eq.~(7), together with
409: the mass $M$ virializing
410: at $z_v$ and estimated from eq.~(1), to obtain the velocity
411: dispersion. The resulting
412: values of the phase-space density $Q$ are quite well fitted by the relation
413: \begin{equation}
414: Q \approx \frac{3.51\times 10^{-9}}{M_{11}^{1.54}} \,\, M_{\odot}pc^{-3}km^{-3}s^{-3}
415: \end{equation}
416:
417: Figure 2 compares the expected values of the phase-space density $Q$
418: as a function of the halo mass, computed from the equation above and
419: values derived from simulated data. The agreement is quite good in spite of the
420: fact that the model
421: assumes a constant halo mass, but confirms our previous conjecture that the
422: early and fast decrease
423: of the phase-space density observed in the halo history reflects
424: the ``thermalization'' of bulk motions.
425:
426: \section{Phase-Space Mixing in Mergers}
427:
428: In the late evolutionary phases of dark matter halos, the rate at which the
429: phase-space density
430: decreases depends on the frequency of mergers and on the amount of mass accreted in
431: these events. Subhalos captured in quite eccentric orbits are generally
432: disrupted completely by tidal forces, transferring angular momentum to
433: the host halo and thus increasing its spin (Peirani et al. 2004) but forming,
434: as we shall see latter, streams detectable
435: in phase-space diagrams before to mix completely with the background material.
436:
437: An attempt to verify whether the merging history of our Galaxy left ``finger-prints'' in the
438: phase-space structure of nearby stars was made by Helmi \& White (1999). In
439: that investigation,
440: the infall of satellites onto a fixed potential was followed numerically as well as the
441: evolution of the debris in phase-space. Helmi \& White found that
442: after $\sim$ 10 Gyr, stars
443: having a common origin are distributed smoothly in space, but form clumps
444: in velocity space.
445:
446: We have searched in our simulation for merger episodes involving massive
447: halos, e.g., having
448: a mass greater than $10^{12}\,M_{\odot}$ at $z=0$, in order to perform evolutionary studies
449: in the phase-space.
450:
451: %%%%%%%%%%%%%%%%%%%%
452: %%%% EXAMPLE 1 %%%%%
453: %%%%%%%%%%%%%%%%%%%%
454: The first example consists of a main halo having presently a mass of
455: about $4.0\times 10^{12}\,M_{\odot}$,
456: which was caught in the act of capture of two subhalos (at $z = 0.92$) with masses at
457: that moment respectively
458: equal to $1.6\times 10^{11}\,M_{\odot}$ (subhalo-1)
459: and $2.8\times 10^{11}\,M_{\odot}$ (subhalo-2).
460: The orbits of each subhalo are sketched in fig. 3. The trajectory of the
461: center of mass of subhalo-1
462: is represented both in the orbital and in the transversal planes in the first
463: two panels. The
464: trajectory of subhalo-2 is represented in the orbital plane up to $z = 0.48$
465: only, since its identity
466: is completely lost at late times. Subhalo-1 has an impact parameter or
467: orbital angular momentum
468: higher than the subhalo-2, illustrating the well known effect of this dynamical parameter
469: on the tidal stripping of captured objects.
470:
471:
472: \begin{figure*}
473: %\rotatebox{0}{\includegraphics[width=6.0cm]{f3a.eps}}
474: %\rotatebox{0}{\includegraphics[width=6.0cm]{f3b.eps}}
475: %\rotatebox{0}{\includegraphics[width=6.0cm]{f3c.eps}}
476: \rotatebox{0}{\includegraphics[width=18.0cm]{f3.eps}}
477: \caption{Sketch of the orbits of satellites in 3D-space for example 1. In the left
478: and right panels, the motion in the orbital plane is represented for both satellites,
479: while in the middle panel the motion of satellite-1 is represented in the
480: transversal plane. The cross indicates the position of the center of mass
481: of the system.}
482: \end{figure*}
483:
484:
485: Successive snapshots of the phase-space evolution (radial velocity versus
486: radial distance) are
487: shown in fig.~4. The subhalo-1 has an orbital period of about 6 Gyr and its
488: periastron is about
489: 100 kpc. Tidal forces strip off only the outer and loosely bound particles, the
490: object preserving
491: most of its identity after 8 Gyr. However, already at $z \approx$ 0.67
492: (near periastron), stripped
493: particles with high positive velocities can be seen, while the bulk has
494: negative values. Effects
495: in the velocity distribution are better visualized in fig.~5, where the
496: corresponding radial
497: velocity distributions are shown for each snapshot of fig.~4. Initially ($z = 0.92$)
498: three distinct velocity distributions can be identified, associated respectively to the
499: main halo and the two subhalos, with velocity
500: dispersions of $\sigma_H$ = 202 km/s, $\sigma_1$ = 51 km/s
501: and $\sigma_2$ = 78 km/s. For comparison, the best fitted Gaussians are also shown
502: as solid lines.
503: In spite of subhalo-1 to have preserved its identity, stripped
504: particles mix in the velocity space and
505: this process can be followed in the different snapshots
506: either in fig.~4 or fig.~5. Notice in particular
507: the appearance of extended tails in the velocity distribution of the
508: captured halos. Subhalo-2
509: has a short orbital period ($\sim$ 2 Gyr) and a quite eccentric orbit.
510: Extended tidal arms are developed after the various periastron passage,
511: forming structures spatially displaced by the orbital decay resulting from
512: dynamical friction. In this case, the mixing
513: in the velocity space (see, fig.~5) is quite efficient and after 3.4 Gyr
514: the particle velocity distribution of subhalo-2 coincides practically with
515: that of the main halo.
516:
517:
518: \begin{figure*}
519: \rotatebox{0}{\includegraphics[width=18cm]{f4.eps}}
520: \caption{The phase-space evolution (radial velocity versus radial distance)
521: for the host halo and satellites of example 1. Upper right labels
522: indicate the redshift.
523: Satellite-1 and Satellite-2 are respectively represented by red and black points.}
524: \end{figure*}
525:
526:
527: In spite of Gaussians be able to fit reasonable well the radial velocity
528: distribution of the main halo (mean determination coefficient $r = 0.97$)
529: and satellites (but $r = 0.86$ for subhalo-2),
530: we have also calculated the kurtosis, $k = (<v^4>/<v^2>^2)-3$, for both the
531: main halo and the subhalo-2 at $z=0$. For the main halo, the kurtosis was
532: evaluated for ten shells with 2800 particles each, while for subhalo-2 we have
533: considered only three shells with about 400 particles each. No particular trend
534: with the radial distance was found, with the main halo having a mean
535: kurtosis $k = -0.57$ and the subhalo-2 $k = -0.73$. These values indicate
536: that the radial velocity distribution is flat-topped with respect to a Gaussian
537: and that the effect in more accentuated in the resulting substructure.
538: %This trend was already noticed by Diemand et al. (2004), who concluded from
539: %their simulations that subhalos have flat-topped velocity distributions with
540: %a typical kurtosis $k = -0.7$.
541:
542: %%%%%%%%%%%%%%%%%%%%
543: %%%% EXAMPLE 2 %%%%%
544: %%%%%%%%%%%%%%%%%%%%
545: Another interesting case of capture is illustrated by our second example. Here the main halo
546: has a mass of about $3.3\times10^{12}$ M$_{\odot}$ and the capture of a satellite with a
547: mass of $5.2\times10^{11}$ M$_{\odot}$ occurs at $z = 0.92$. Here again masses correspond
548: to that instant, since at the end, due to the continuous infall process, masses are
549: substantially higher. Fig.~6 shows snapshots of the phase space
550: at different redshifts. Structures are gradually formed as the satellite passes by periastron. The
551: observed streams in the phase-space are very similar to those derived from models of halo
552: formation by secondary infall (see, for instance, Sikivie et al. 1997 and references therein).
553: However, structures seen in the latter case are formed by successive passages of infalling
554: (or outgoing) particles
555: and are associated to ``caustics''. From a mathematical point of view, caustics have an infinite
556: density. In physical systems, the density is limited by the fact that the motion of DM particles
557: is not purely radial (non-zero angular momentum) and by a non-zero velocity dispersion
558: (Sikivie et al. 1997). If the nature of DM particles is specified, then other constraints
559: can be imposed. For instance, neutralinos and anti-neutralinos may annihilate with a rate proportional
560: to the square of their density, a mechanism which imposes upper bounds to the density. Moreover,
561: neutralinos are fermions and one would expect that a ``classical'' behavior must be abandoned
562: when their de Broglie wavelength will be greater than the mean inter-particle separation. This occurs
563: for densities {\bf n} satisfying $n > (m_{\chi}v/h)^3$, when repulsive forces of quantum nature
564: due to the Pauli principle become effective, introducing a further physical mechanism able
565: to impose limits on the density. In a given instant and for a fixed value of the
566: radial distance, a certain number of velocity peaks, corresponding
567: to different streams can be seen in the phase-space diagram (Fig.~6), which depends on the
568: initial angular momentum, as simulations suggest. In the extreme limit of circular orbits only one
569: stream should be observed, a case comparable to that of satellite-1
570: in the first example. The width $\delta v_m$ of the
571: m-velocity peak associated to the m-caustic is generally estimated
572: from the Liouville theorem,
573: using the velocity dispersion and density at the moment of their first turnaround.
574: In general and in the present case, this procedure can not be
575: applied. The fine-grained DF is
576: strictly conserved in the phase-space flow but this is not the case of the coarse-grained DF,
577: since phase-space elements of high density are stretched out and folded
578: with elements of low density because of mixing (Lyndell-Bell 1967).
579:
580:
581: \begin{figure*}
582: \rotatebox{0}{\includegraphics[width=18cm]{f5.eps}}
583: \caption{The evolution of the radial velocity distribution for the host halo
584: (light blue), satellite-1 (red) and satellite-2 (black) for example 1. At
585: $z=0.92$, the best fitted Gaussian distribution is superimposed to histograms.
586: Notice that the velocity distribution of satellite-1 is preserved but it develops
587: high velocity tails.}
588: \end{figure*}
589:
590: \begin{figure*}
591: \rotatebox{0}{\includegraphics[width=18cm]{f6_wb.eps}}
592: \caption{The phase-space evolution of the host halo and satellite (black points)
593: for example 2. The latter is completely disrupted by tidal forces. The different
594: snapshots show the formation of streams in phase-space.}
595: \end{figure*}
596:
597: \begin{figure}
598: \rotatebox{0}{\includegraphics[width=\columnwidth]{f7.eps}}
599: \caption{The evolution of the virial ratio $2T/|W|$ as a a function of the redshift (example 2).
600: Each maximum corresponds to the passage of the subhalo by periastron.}
601: \end{figure}
602:
603: \begin{figure}
604: \rotatebox{0}{\includegraphics[width=\columnwidth]{f8.eps}}
605: \caption{The radial profile of the specific angular momentum $j$ for the
606: merger example 2. Consequence of a "head-on" collision, no significant
607: modifications in the profile of $j$ is observed after the merger event
608: ($z=0$) which occurred at $z=1.08$.
609: }
610: \end{figure}
611:
612: \begin{figure*}
613: \rotatebox{0}{\includegraphics[width=18cm]{f9.eps}}
614: \caption{(Top) The phase-space evolution of the main halo and a satellite (black points)
615: completely disrupted by tidal forces, in example 3.
616: The other panels show respectively the evolution of
617: the radial velocity distribution for all particles of both host and satellite (second line),
618: and the radial velocity distribution at different shells from the center (all particles, third line
619: and for the satellite, fourth line). Solid lines represent the best Gaussian fit.}
620: \end{figure*}
621:
622:
623:
624: It's worth mentioning that the time evolution of the virial ratio $2T/|W|$
625: of the whole system
626: has clear maxima corresponding to successive pericenter passages of the center of mass of
627: the sub-halo (see Fig.~7), in agreement with merger simulation
628: by Valluri et al. (2007). However, in
629: this particular case, there is no expressive redistribution of the specific angular
630: momentum {\bf j} after the merger event, the
631: radial profiles of {\bf j} being practically the same before
632: and after merging (see fig.~8), since
633: the collision orbit is almost radial and, consequently, no
634: significant transfer of angular momentum occurs.
635: As in the previous example, the analysis of the kurtosis at $z = 0$ indicates
636: top-flatted velocity
637: distributions with the satellite-1 having a flatter
638: distribution (mean kurtosis $k = -0.85$ in
639: comparison with $k = -0.64$ for the main halo).
640:
641: %%%%%%%%%%%%%%%%%%%%
642: %%%% EXAMPLE 3 %%%%%
643: %%%%%%%%%%%%%%%%%%%%
644: In the third example of a merger episode, the main halo has a
645: mass of $4.3\times 10^{12}\,M_{\odot}$ and underwent
646: an important fusion at z=0.92, since the captured object has a
647: comparable mass ($\sim 1.2\times 10^{12}\,M_{\odot}$).
648: In the three upper panels of fig.~9, the phase-space evolution is
649: shown at three distinct redshifts. At $z=0.48$, structures formed
650: at different periastron passages can be seen, while at $z=0$ the
651: captured halo is completely disrupted and its particles are well mixed
652: in phase-space. This behavior is consistent with the simulations by
653: Valluri et al. (2007), which indicate that mixing in the space of the dynamical
654: variables $(E,J)$ occurs primarily during periastron passages and are
655: driven probably by compressive tidal shocks.
656: The subsequent panels show the evolution of the radial
657: velocity distribution. The second line shows the evolution for all
658: particles of the main halo and satellite, which at $z=0.92$ has a velocity dispersion
659: of 130 km/s. At $z=0.48$ and $z=0$ both velocity distributions practically match.
660: The average kurtosis are $k= -0.57$ and $k = -0.65$ for the main halo and satellite
661: respectively, confirming the trend that substructures have top-flatted velocity
662: distributions more accentuated than those of the main halos.
663:
664: Notice that the fact that all selected merger examples
665: occur at $z = 0.92$ is completely fortuitous and is probably associated to
666: a maximum rate of mergers around $z \sim 1$ as well as to the time resolution
667: adopted in the present study.
668:
669: The main halos in the aforementioned examples have typically $(3-4)\times 10^4$
670: particles while for satellites the numbers are about one order of magnitude lower. In the simulations by Wojtak et al. (2005), cluster size halos with few times $10^4$ particles (numbers comparable to those of our selected examples)
671: were considered. These authors concluded that the velocity distribution of
672: main halos are only Gaussian to a good approximation near the center, but more and more flat-topped with respect to Gaussians when approaching the
673: virial radius.
674: Numerical experiments of higher resolution were performed by Diemand et al. (2004) and Kazantzidis et al. (2004). The former authors considered galactic
675: and cluster size halos extracted from cosmological simulations, including
676: several millions of particles. They concluded that subhalos have flat-topped
677: velocity distributions with a typical kurtosis $k \simeq -0.7$. Equilibrium
678: halos with about $10^5$ particles were generated by Kazantzidis et al. (2004)
679: using the Eddington method. They found also that the velocity distribution of such systems are less peaked than Gaussians.
680:
681: In spite of the different employed resolutions, all these
682: investigations (includind our own) lead to consistent results, e.g., the
683: velocity distribution of halos are flat-topped with kurtosis in the
684: range $-0.4 > k > -0.9$.
685:
686:
687:
688: \section{Conclusions}
689:
690: The dynamical evolution of dark matter halos after the gravitational
691: collapse depends on their accretion history. The phase-space density
692: indicator $Q$ decreases by a large factor ($\sim 40$)
693: during the first shell crossing
694: as a consequence of the randomization process of initial bulk motions. Our
695: theoretical estimates of the
696: phase-space density $Q$ based on the spherical collapse model are in agreement
697: with our simulated data, suggesting
698: that at ``virialization" the phase-space density scales with the halo
699: mass as $Q \propto M^{-1.5}$. The analysis
700: by Armarzguioui \& Gron (2005) indicates that the entropy of an ideal gas
701: increases to the first order in
702: the gravitational collapse within the context of the FRW cosmology. This entropy
703: increase is a consequence of
704: the transfer of gravitational potential energy to thermal energy during the
705: collapse, a process similar
706: to that observed for colisionless matter. Effects of shell crossing
707: play probably a non negligible role in the way halos reach equilibrium
708: and should be taken into account in future studies, since they
709: affect the resulting ``virial" radius of the system (S\'anchez-Conde et al. 2006).
710:
711: Non-disrupted satellites develop high velocity tails in their radial
712: velocity distribution. The mecanism (or mechanisms) responsible
713: for such a heating is not well
714: established. Funato, Makino \& Ebisuzaki (1992)
715: based on N-body simulations of collisionless systems argued
716: that wave-particle interactions, e.g., collective
717: effects driven by large fluctuations of the gravitational potential are
718: able to produce a substantial
719: heating in a few dynamical time scales, regardless the particle initial
720: energy. Another possibility
721: is the heating produced by compressive tides that arise when one
722: collisionless gravitating system passes
723: through another on a time scale shorter than the internal dynamical time
724: of the infalling object. Such
725: tides can impulsively heat particles as a result of the transient deepening
726: of the net gravitational
727: potential (Valluri et al. 2007). Impulsive compressive tidal shocks are
728: well known to produce changes in the internal structure of globular
729: clusters (Gnedin et al. 1999) or in subhalos orbiting within
730: massive hosts (Kravtsov et al. 2004).
731:
732: Disrupted satellites leave ``finger-prints'' in phase-space, generating streams
733: which depend on the initial
734: orbital angular momentum and which remind caustic structures seen in secondary
735: infall models of halo
736: formation.
737: Future high resolution simulations will certainly give a better
738: insight on the evolution of these structures.
739: It is worth mentioning that extended secondary infall models are also
740: able to explain the scaling-free
741: $Q$ profile (Austin et al. 2005). Stars behave as a colisionless fluid similar to
742: dark matter. In this
743: case, one should expect that disrupted satellites will form not only dark matter
744: streams but also
745: stellar streams, which could eventually be detected in phase-space by the
746: forthcoming space mission GAIA as also suggested by Brown et al. (2005).
747:
748: Our previous study (PDP06) and the present work support the view that
749: violent variations in the gravitational
750: potential, which occur in important merger events, lead to more mixed
751: systems as the temporal behaviour of the
752: phase-space density suggests. The study of some merger episodes indicates that
753: the radial velocity of
754: completely disrupted satellites and that of the main halo fuse in a common
755: distribution in a few dynamical time
756: scales. Quasi-relaxed halos have top-flatted velocity distributions, which
757: are observed not only for the bulk of particles but also for particles inside shells
758: at different distances from the center. The significance of these top-flatted profiles is
759: measured by the kurtosis (always negative) and this effect is more accentuated
760: for captured subhalos and/or substructures in comparison with the velocity distribution
761: of the main halo. Kazantzidis et al. (2004) have developed an algorithm for
762: constructing N-body realizations of equilibrium spherical systems, which differs
763: from the usual assumption that the local
764: velocity distribution is Maxwellian. Equilibrium halo models built through their procedure
765: when evolved, develop velocity distributions less peaked than Gaussians and the disruption
766: time scale of captured satellites becomes significantly longer in comparison with initial
767: Maxwellian equilibrium models. Cosmological simulations by Diemand et al. (2004)
768: lead also to the conclusion that substructures of either galactic-size or cluster-size
769: halos have top-flatted velocity profiles, in agreement with our results.
770: %{\bf Since values of the kurtosis for both main halo and disrupted satellite
771: % are quite similar to the ones obtained from these past studies,
772: %the relative low resolution used in our simulation (10 to 100 times less particles)
773: % seems to not affect our results and main conclusions.}
774: Moreover, an
775: investigation of structures resulting from head-on collisions of halos having a
776: Navarro-Frenk-White density profile was performed by Hansen et al. (2006). They concluded
777: that the phase-space density indicator $Q$ has a power-law distribution and that
778: the resulting velocity distribution is not Gaussian, but can be represented by a function
779: derived from the Tsallis statistics (Tsallis 1988). Thus, deviations from Gaussianity
780: seems to be a quite general result and, in this sense, the present simulations
781: represent a contribution to establish an empirical basis for future theoretical developements
782: in this area.
783:
784: \acknowledgements
785: S.\,P.\, acknowledges the financial support from a ANR grant.
786: S.\,P.\, is grateful to R. Mohayaee for discussions on the early phase of this work.
787: We also thank S. Hansen and F. Durier for their
788: useful comments which have significantly improved this paper.
789:
790:
791:
792: \begin{thebibliography}{(dummy)}
793: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
794: \bibitem[]{arad04}
795: Arad, I., Dekel, A. and Klypin, A., 2004, MNRAS 353, 15
796: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
797: %\bibitem[Arad \& Lynden-Bell(2005)]{2005MNRAS.361..385A}
798: %Arad, I., and Lynden-Bell, D.\ 2005, MNRAS, 361, 385
799: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
800: \bibitem[]{agron05}
801: Armarzguioui, M. and Gron, O., 2005, PRD 71, 083011
802: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
803: \bibitem[]{asc04}
804: Ascasibar, Y., Yepes, G., Gottl{\"o}ber, S., M\"uller, V.\ 2004, MNRAS, 352, 1109
805: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
806: \bibitem[]{asc08}
807: Ascasibar, Y., \& Gottl{\"o}ber, S.\ 2008, MNRAS, 386, 2022
808: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
809: \bibitem[]{aust05}
810: Austin, C.G., Williams, L.L.R., Barnes, E.I., Babul, A. and Dalcanton, J.J., 2005, ApJ 634, 756
811: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
812: \bibitem[]{boy04}
813: Boylan-Kolchin, M., \& Ma, C.-P.\ 2004, MNRAS, 349, 1117
814: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
815: \bibitem[]{bryan98}
816: Bryan, G.L. and Norman, M.L., 1998, ApJ 495, 80
817: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
818: \bibitem[]{brow05}
819: Brown, A.~G.~A., Vel{\'a}zquez, H.~M., \& Aguilar, L.~A.\ 2005, MNRAS, 359, 1287
820: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
821: \bibitem[]{cou95}
822: Couchman, H.M.P., Thomas, P.A. and Pearce, F.R., 1995, ApJ 452, 797
823: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
824: \bibitem[]{dal01}
825: Dalcanton, J.J. and Hogan, C.J., 2001, ApJ 561, 35
826: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
827: \bibitem[]{dave01}
828: Dav\'e, R., Spergel, D.N., Steinhardt, P.J. and Wandelt, D., 2001, ApJ 547, 574
829: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
830: \bibitem[Davis et al.(1985)]{1985ApJ...292..371D}
831: Davis, M., Efstathiou, G., Frenk, C.~S., and White, S.~D.~M.\ 1985, ApJ, 292, 371
832: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
833: \bibitem[]{deh05}
834: Dehnen, W., 2005, MNRAS 360, 892
835: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
836: \bibitem[]{dehn05}
837: Dehnen, W., \& McLaughlin, D.~E.\ 2005, MNRAS, 363, 1057
838: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
839: \bibitem[]{die04}
840: Diemand, J., Moore, B. and Stadel, J., 2004, MNRAS 352, 535
841: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
842: \bibitem[]{fme92}
843: Funato, Y., Makino, J. and Ebisuzaki, T., 1992, PASJ 44, 613
844: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
845: \bibitem[]{gne99}
846: Gnedin, O.Y., Lee, H.M. and Ostriker, J.P, 1999, ApJ 522, 935
847: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
848: \bibitem[]{han06}
849: Hansen, S.H., Moore, B., Zemp, M. and Stadel, J., 2006, JCAP 0601, 014
850: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
851: \bibitem[]{helmi99}
852: Helmi, A. and White, S.D.M., 1999, MNRAS 307, 495
853: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
854: \bibitem[]{hof07}
855: Hoffman, Y., Romano-D{\'{\i}}az, E., Shlosman, I., \& Heller, C.\ 2007, ApJ, 671, 1108
856: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
857: \bibitem[]{kaz04}
858: Kazantzidis, S., Magorrian, J. and Moore, B., 2004, ApJ 601, 37
859: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
860: \bibitem[]{kol08}
861: Knollmann, S.~R., Knebe, A., \& Hoffman, Y.\ 2008, MNRAS, 391, 559
862: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
863: \bibitem[]{kra04}
864: Kravtsov, A.V., Gnedin, O.Y. and Klypin, A.A., 2004, ApJ 609, 482
865: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
866: \bibitem[]{lahav91}
867: Lahav, O., Lilje, P.B., Primack, J.R. and Rees, MJ., 1991, MNRAS 251, 128
868: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
869: \bibitem[]{lynd67}
870: Lynden-Bell, D., 1967, MNRAS 136, 101
871: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
872: %\bibitem[]{naka00}
873: %Nakamura, T.K., 2000, ApJ 531, 739
874: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
875: \bibitem[]{pmp04}
876: Peirani, S., Mohayaee, R. and de Freitas Pacheco, J.A., 2004, MNRAS 348, 921
877: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
878: \bibitem[]{pdp06}
879: Peirani, S., Durier, F. and de Freitas Pacheco, J.A., 2006, MNRAS 367, 1011 (PDP06)
880: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
881: \bibitem[]{rasia04}
882: Rasia, E., Tormen, G. and Moscardini, L., 2004, MNRAS, 351, 237
883: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
884: \bibitem[]{sanchez06}
885: S{\'a}nchez-Conde, M.~A., Betancort-Rijo, J., \& Prada, F.\ 2007, MNRAS,
886: 378, 339
887: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
888: \bibitem[]{siki97}
889: Sikivie, P., Tkachev, I.I. and Wang, Y., 1997, Phys.Rev. D56, 1863
890: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
891: \bibitem[]{suger00}
892: Sugerman, B., Summers, F.J. and Kamionkowski, M., 2000, MNRAS 311, 762
893: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
894: \bibitem[]{tayna01}
895: Taylor, J.E. and Navarro, J.F., 2001, ApJ 563, 483
896: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
897: \bibitem[]{thl86}
898: Tremaine, S., H\'enon, M. and Lynden-Bell, D., 1986, MNRAS 219, 285
899: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
900: \bibitem[]{tas88}
901: Tsallis, C., 1988, J.Stat.Phys. 52, 479
902: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
903: \bibitem[]{val06}
904: Valluri, M., Vass, I.M., Kazantzidis, S., Kravtsov, A.V. and Bohn, C.L., 2007,
905: ApJ 658, 731
906: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
907: \bibitem[]{vas08}
908: Vass, I.~M., Valluri, M., Kravtsov, A., \& Kazantzidis, S.\ 2008a, arXiv:0810.0277
909: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
910: \bibitem[]{}
911: Vass, I.~M., Kazantzidis, S., Valluri, M., \& Kravtsov, A.~V.\ 2008b, arXiv:0812.3659
912: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
913: \bibitem[]{wec02}
914: Wechsler, R.H., Bullock, J.S., Primack, J.R., Kravtsov, A.V. and Dekel, A., 2002, ApJ, 568, 52
915: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
916: \bibitem[]{woj05}
917: Wojtak, R., Lokas, E.L., Gottloeber, S. and Mamon, G., 2005, MNRAS
918: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
919: \end{thebibliography}
920:
921:
922:
923:
924:
925:
926:
927:
928:
929:
930:
931:
932: \end{document}
933:
934:
935:
936:
937:
938:
939:
940:
941:
942:
943:
944:
945:
946: