1: \documentclass{appolb}
2: \usepackage{epsfig}
3: \preprint{}
4: % epsfig package included for placing EPS figures in the text
5: %------------------------------------------------------
6:
7: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
8: % %
9: % BEGINNING OF TEXT %
10: % %
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12: \begin{document}
13: \bibliographystyle{num}
14: % \eqsec % uncomment this line to get equations numbered by (sec.num)
15: \title{Size of the thermal source in relativistic heavy-ion collisions%
16: \thanks{Research supported in part by the Polish State Committee for
17: Scientific Research, grant 2 P03B 059 25}%
18: % you can use '\\' to break lines
19: }
20: \author{Piotr Bo\.zek
21: \address{The H. Niewodnicza\'nski Institute of Nuclear Physics, Polish Academy
22: of Sciences, PL-31342 Krak\'ow, Poland}
23: }
24: \maketitle
25: \begin{abstract}
26: The dependence of the size of the thermal source on the centrality
27: in ultrarelativistic heavy-ion collisions is studied.
28: The interaction region consists of a well defined thermalized core, and of
29: an outer mantle where the production scales with the number of participants.
30: The thermal source builds up in the region with the largest
31: density of participants
32: in the transverse plane.
33: Particle production in the thermalized core is enhanced in
34: comparison to the wounded nucleon model.
35: The change of the degree of strangeness saturation with centrality is
36: also discussed. We perform an estimate of high $p_\perp$ jet absorption
37: finding that an increase of the absorption in the thermal core is compatible
38: with the data.
39: \end{abstract}
40: \PACS{25.75.-q; 24.85.+p}
41:
42: \section{Introduction}
43: Relativistic heavy-ion collision experiments at RHIC energies
44: $\sqrt{s}=130$GeV
45: and $200$GeV reveal a large degree of thermalization in the produced system.
46: Particle spectra have thermal shapes, with a substantial amount of collective
47: flow. The presence of a strong transverse and elliptic collective flow
48: demonstrates the creation of a strongly interacting phase in the history
49: of the
50: system. This stage can be described to a large extent by
51: an ideal hydrodynamical
52: evolution. Particle production in heavy-ion collisions is characterized by
53: the thermal parameters of the source~: its temperature, transverse flow and
54: possibly the degree of strangeness saturation. The temperature,
55: the chemical potentials and the
56: strangeness saturation at the chemical freeze-out can be extracted by fits
57: of a statistical hadronization model to
58: the ratios of abundances of particles produced in
59: the central rapidity region.
60: The description of particle spectra in transverse momentum
61: requires the knowledge of the collective flow profile in the source and of
62: the temperature at the kinetic freeze-out.
63:
64: Usually it is assumed that all the produced particles are
65: emitted from a single thermal source.
66: One finds that the parameters of the thermal
67: freeze-out depend on the centrality. For the chemical freeze-out one observes
68: a significant variation of the strangeness saturation factor in peripheral
69: collisions \cite{xu,rafelski}. For the kinetic freeze-out a stronger
70: collective flow and a
71: lower freeze-out temperature comes out from the fits to the
72: spectra of particles in central collisions \cite{starspectra}.
73: % This assumes that the
74: %entire interaction region can be approximated by one roughly homogenous
75: % source with a single temperature on the decoupling surface.
76: A different approach allows for spatial variations of the parameters of the
77: freeze-out surface surface, i.e. temperature and chemical potential gradients
78: \cite{budalund}.
79:
80: The extraction of the thermal parameters of the source and the description
81: of its dynamics in
82: hydrodynamical models are useful in the description of the evolution
83: of the dense matter produced in the collision. Such studies
84: include the modeling of the elliptic flow in non-central collisions.
85: Quantitative results require the knowledge of the initial conditions
86: in the simulation. Therefore the understanding of the variation of the size
87: and of the nature of the thermal source with centrality is crucial.
88:
89:
90:
91: \section{Size of the thermal source}
92:
93:
94: Experimental data at the highest RHIC energies demonstrate the formation of a
95: thermalized system in $Au+Au$ collisions~: the spectra are thermal with a
96: collective transverse flow component, the elliptic flow is compatible
97: with an ideal fluid dynamics, particle ratios can be described by universal
98: chemical freeze-out parameters. From the theoretical side we expect that at
99: high energy densities formed in the collision the system undergoes a phase
100: transition to the quark-gluon plasma. This phase transition could facilitate
101: a complete chemical equilibration. If the partonic plasma is strongly
102: interacting near the critical temperature, the system would behave as
103: an ideal fluid. These feature should not vary with centrality if the
104: thermalization and the evolution of the thermal source are dominated by the
105: phase transition. The kinetic freeze-out is given by the local density and
106: flow gradient and should not vary strongly with the impact parameter of the
107: collision either.
108:
109: Therefore a reasonable assumption would be to have a thermal source with
110: the same thermal parameters at all centralities; the variation of the
111: particle production coming
112: only from the change of the size of the source.
113: It has been noticed that particle production per
114: participant in heavy-ion collisions at RHIC energies is
115: significantly larger than in $p+p$ or $d+Au$ collisions
116: \cite{Back:2004dy,Back:2004mr,Back:2002uc,Adler:2004zn}.
117: In the analysis of the experiments at RHIC energies
118: Glauber Monte Carlo calculations based on the wounded nucleon picture
119: permit to extract the number of participants (or
120: equivalently wounded nucleons) used to quantify
121: the difference of the production per participant in $Au+Au$ or $d+Au$
122: collisions
123: with respect to $p+p$ interactions.
124: Geometrical scaling predicts that the bulk of the particle production,
125: dominated by soft particles, scales with the number of wounded nucleons
126: \cite{abc}. It has been noticed that the density of
127: charged particles at $\eta=0$ and the total number of charged
128: particles produced
129: in $d+Au$ collisions
130: scales with the number of participants~\cite{Back:2004mr}~:
131: \begin{equation}
132: \frac{dN_{dAu}}{d\eta}=N_{part}\frac{dN_{pp}}{d\eta} \ .
133: \end{equation}
134: Remarkably, this scaling is
135: fulfilled in a more general way by the distributions of charged particles
136: in a broad range of pseudorapidity \cite{bcscaling}.
137: On the other hand particle production in $Au+Au$ collisions deviates from
138: a simple superposition of independent production from all the wounded
139: nucleons. The experimental data show a larger multiplicity than predicted by
140: the scaling of the $p+p$ production by $N_{part}/2$. This and the
141: evidence of the thermalization discussed above point to the importance of
142: the second stage of the collision. The main characteristic of this
143: stage is the existence of a
144: thermally equilibrated source. Particle production
145: from the thermal source occurs late
146: and the production mechanism is far from the superposition of elementary
147: collisions. In a system undergoing Bjorken longitudinal expansion
148: the number of particles emitted at central rapidity is proportional to the
149: transverse extension of the source. If the density of elementary collisions
150: in the transverse plane is high one expect fast thermalization and eventually
151: a phase transition to the quark-gluon plasma. In a geometrical Glauber model
152: the decisive parameter is the density of wounded nucleons in
153: the transverse plane. In peripheral collisions the density of colliding
154: nucleons is small, and the production would be similar as in $d+Au$ collisions.
155: In semi-central and central collisions, at
156: the center of the interaction region
157: the density of wounded nucleons is large and the transverse energy deposited
158: is high enough to guaranty a fast thermalization. The boundary of the
159: thermal source
160: is given by the limiting density of wounded nucleons in the transverse plane
161: sufficient for the fast thermalization processes to occur.
162: The observation of
163: universal chemical freeze-out parameters in $Au+Au$ collisions
164: and the indication of the formation of a strongly interacting plasma
165: at near critical conditions, lead to the assumption that the boundary of the
166: thermal source and of the quark-gluon plasma phase are the same.
167: In the outer part of the interaction region the density of participants and
168: therfore the density of the deposited energy is not
169: high enough to guaranty the
170: approach of the critical point an the resulting fast thermalization.
171: At RHIC
172: energies the boundary in
173: the interaction region between the thermalized
174: source and the outer mantle is well defined by the criterion of the phase
175: transition. It turns out that the density of
176: participants in the outer mantle is low, they undergo at most $2.5$
177: binary collisions per wounded nucleon pair. A similar number of
178: binary collision is predicted for central $d+Au$ collisions. The number of
179: particles
180: produced in $d+Au$ interactions (even in the most central collisions)
181: scales with the number of participants. The low number of binary collisions in
182: the outer mantle justifies the assumption that the number of charged
183: particles produced in this way is proportional to the number of wounded
184: nucleons in the mantle $N_{part}^{m}$.
185:
186: The number of particles emitted
187: from the source depends on the source size at the freeze-out. Since the
188: freeze-out conditions are universal, the final number of particles produced
189: should scale with the transverse energy deposited in the thermal source.
190: The deposited energy is proportional to the
191: number of wounded nucleons $N_{part}^{c}$
192: in this high density core of the
193: interaction region.
194: The density of charged particles produced in the central
195: pseudorapidity region is
196: \begin{equation}
197: \frac{dN_{ch}}{d\eta}=\frac{N_{part}^{m}}{2}\frac{dN_{ch}^{pp}}
198: {d\eta}+
199: \frac{N_{part}^{c}}{2}\frac{dN_{ch}^{th}}{d\eta} \ ,
200: \end{equation}
201: where $\frac{dN_{ch}^{pp}}{d\eta}$ is the density of particles produced
202: produced in $p+p$ collisions and $\frac{dN_{ch}^{th}}{d\eta}$
203: is the density per participant pair from the
204: thermal source.
205: The enhancement of
206: the particle production in the thermal source
207: can be effectively described by one parameter $\alpha$
208: \begin{equation}
209: \frac{dN_{ch}^{th}}{d\eta}=(1+\alpha)\frac{dN_{ch}^{pp}}{d\eta} \ .
210: \end{equation}
211:
212:
213:
214: \begin{figure}
215: \centering
216: \includegraphics*[width=0.7\textwidth]{dn.eps}
217: \caption{Density of charged particles produced at central
218: pseudorapidity in $Au+Au$
219: collisions scaled by the density of particles in $p+p$ collisions from
220: the BRAHMS \cite{Bearden:2001xw,Bearden:2001qq},
221: PHENIX \cite{Adler:2004zn}
222: and PHOBOS \cite{Back:2004dy,Back:2002uc} experiments
223: compared to the predictions of the core-mantle model
224: (solid line).
225: The dashed line represents the number of participant pairs.}
226: \label{specas}
227: \end{figure}
228:
229:
230: The number of wounded nucleons in the mantle is calculated from the formula
231: \begin{equation}
232: N_{part}^{m}=\int d^2s \frac{d\phi}{2\pi} \frac{d^2N_{part}}{ds^2}
233: \theta(d_{cut}-\frac{d^2N_{part}}{ds^2})\Theta(\phi) \ ,
234: \end{equation}
235: where $\frac{d^2N_{part}}{ds^2}$ is the density of participants in the
236: transverse plane, calculated from the
237: Glauber Monte Carlo model. The integration
238: is restricted to densities of participants lower than $d_{cut}$ and involves
239: only particles emitted in directions other than the dense core (the function
240: $\Theta(\phi)$ is one if the emission direction
241: is not shadowed by the core and zero
242: otherwise). The number of participants in the core is calculated
243: analogously
244: \begin{eqnarray}
245: N_{part}^{c}& =&\int d^2s \frac{d\phi}{2\pi} \frac{d^2N_{part}}{ds^2}
246: \theta(d_{cut}-\frac{d^2N_{part}}{ds^2})\left(1-\Theta(\phi)\right)
247: \nonumber \\
248: &+& \int d^2s \frac{d\phi}{2\pi} \frac{d^2N_{part}}{ds^2}
249: \theta(\frac{d^2N_{part}}{ds^2}-d_{cut}) \ ,
250: \end{eqnarray}
251: the main contribution comes from the region with the density of participants
252: larger than $d_{cut}$. For the calculation of the density of wounded nucleons
253: we use Woods-Saxon nuclear density profiles
254: ${1}/\left({1+\exp((r-R_{A})/a)}\right)$,
255: with parameters $a=0.535$fm and $R_{A}=1.12A^{1/3}$fm.
256: The cross sections are $\sigma=42$mb and $41$mb for $\sqrt{s}=200$GeV and
257: $130$GeV respectively.
258:
259: The results of the calculations for the scaled particle pseudorapidity density
260: \begin{equation}
261: \frac{dN_{AuAu}}{d\eta}/\frac{dN_{pp}}{d\eta}
262: \end{equation}
263: are presented in Figure \ref{specas}.
264: Published experimental data from the BRAHMS, PHENIX and PHOBOS collaborations
265: are shown. The density of charged particles divided by the $p+p$ density
266: increases faster than the number of participant pairs
267: shown by the dashed line. Moreover this increase is not exactly linear in
268: $N_{part}$. There is a strong increase of $dN/d\eta$ around $30-100$
269: participants and then a gradual increase of the
270: slope up to the highest centralities. We can reproduce this effect in
271: the core-mantle model with the parameters $d_{cut}=2$fm$^{-2}$ and
272: $\alpha=0.65$ at $\sqrt{s}=200$GeV and $d_{cut}=2.2$fm$^{-2}$ and
273: $\alpha=0.5$ at $\sqrt{s}=130$GeV. At centralities corresponding to
274: $N_{part}=40$ the thermalized core appears in the interaction region,
275: causing a fast increase of the scaled particle density.
276: For more central collisions the contribution of the
277: thermal core in the total
278: production increases gradually up to $95\%$ at zero impact parameter. This
279: leads to a slightly faster than linear increase of the particle density
280: with the number of participants, similar as seen in the data.
281: %A precise fit of
282: %the parameters of the model is not possible due to large discrepancies between
283: %data sets and the role of possible fluctuations of the size of the source
284: %which is not taken into account in this paper.
285: The increase of the production per participant pair in the core $1+\alpha$ and
286: the decrease of the cutoff density of participants $d_{cut}$ with energy
287: indicate that the wounded nucleons get effectively ``fatter'' in the the
288: transverse plane and deposit a larger transverse energy at higher
289: collision energy. At SPS energies the mantle part of the interaction region
290: is dominant. The number of binary collision in the mantle increases and the
291: assumption of the $N_{part}^{m}$ scaling of the production in the mantle
292: breaks down.
293:
294:
295: \section{Strangeness}
296:
297: Following the arguments from the previous section
298: one can argue that the production of identified strange particles
299: in heavy-ion collisions at RHIC energies occurs by two different mechanisms
300: in the core and in the mantle of the interaction region.
301: In studies with
302: a single thermal source different degree of strangeness equilibration
303: as function of centrality is seen
304: \cite{xu,rafelski}.
305: In the core-mantle model we choose a different approach. Since the physics of
306: the thermal core is dominated by the phase transition we assume
307: a complete equilibration of strangeness in the thermal source, this is
308: consistent with the data on particle ratios for the most central collisions.
309:
310: \begin{figure}
311: \centering
312: \includegraphics*[width=0.7\textwidth]{strange.eps}
313: \caption{$K^-/\pi^-$ ratio as a function of centrality. The
314: data are from the PHENIX
315: collaboration \cite{phenixstrange,phenixmata}.
316: The solid line denotes the sum of the emission from the thermal core
317: and of a contribution from $p+p$ collisions scaled with the
318: number of collisions in the mantle (equation \ref{scaledau}).
319: The dashed line represents the
320: results of a calculation with a thermal source and a scaled
321: $p+p$ contribution
322: for the production in the mantle.
323: }
324: \label{strangefig}
325: \end{figure}
326:
327:
328: Unlike the charged particle density which
329: scales with the number of wounded nucleons in $d+Au$
330: collisions, the ratio $K/\pi$ depends on the centrality. Therefore, we cannot
331: assume that in the mantle the particle ratios
332: are similar as in $p+p$ collisions. Preliminary data from the PHENIX
333: collaboration \cite{phenixmata} show an increase of the $K/\pi$ ratio with the
334: mean number of binary collisions $\nu=\frac{ N_{bin}}{N_{part}/2}$.
335: This dependence can be fitted for the $K^-/\pi^-$ ratio by the formula
336: \begin{equation}
337: \label{scaledau}
338: \left(\frac{K^-}{\pi^-}\right)_{dAu}(\nu)=0.062+0.025 \times \nu \ .
339: \end{equation}
340: Within the core-mantle model the number of produced particles can be written
341: as
342: \begin{eqnarray}
343: N_K=\frac{N_{part}^{m}}{2}n_{K}^{el}+\frac{N_{part}^{c}}{2}n_{K}^{th}
344: \nonumber \\
345: N_\pi=\frac{N_{part}^{m}}{2}n_{\pi}^{el}+\frac{N_{part}^{c}}{2}n_{\pi}^{th}
346: \ ,
347: \end{eqnarray}
348: where $n^{el}$ and $n^{th}$ represent the production of particles per
349: participant pair in the mantle and in the core respectively.
350: Using $n_\pi^{th}\simeq (1+\alpha)n_\pi^{el}$, i.e. the pion
351: multiplicity follows the charged particles scaling, we have
352: \begin{equation}
353: \frac{N_K}{N_\pi}\simeq \frac{N_{part}^{m}\frac{n_K^{el}}{n_\pi^{el}}+
354: (1+\alpha)N_{part}^{c}\frac{n_K^{th}}{n_\pi^{th}}}
355: {N_{part}^{m}+(1+\alpha)N_{part}^{c}} \ .
356: \end{equation}
357: The $K^-/\pi^-$ ratio in the thermal core should reproduce the experimental
358: data for the most central collisions.
359: We take $\frac{n_{K^-}^{th}}{n_{\pi^-}^{th}}=0.16$ from a statistical
360: hadronization model \cite{share} with parameters fitted to particle ratios
361: in central collisions. The ratio
362: in the mantle $\frac{n_K^{el}}{n_\pi^{el}}$ is taken according to
363: the formula (\ref{scaledau}), where $\nu$ is the average number of collisions
364: in the mantle part of the interaction region of a $Au+Au$ collision.
365: The results reproduce well the experimental data \cite{phenixstrange}
366: on $K^-/\pi^-$ ratio
367: as a function of centrality (Figure \ref{strangefig}).
368: The increase of strange particle production
369: comes from the increase of the number of collisions in the mantle and from the
370: increase of the contribution of the thermal core to the total emission.
371: Neglecting the dependence of the ratio $\frac{n_K^{el}}{n_\pi^{el}}$
372: or $\left(\frac{K^-}{\pi^-}\right)_{dAu}$
373: on the number of collisions we cannot reproduce the centrality dependence
374: in $d+Au$ and $Au+Au$ interactions (dashed line in Figure \ref{strangefig}).
375: We note that the smooth scaling of the production from peripheral $d+Au$ up to
376: central $Au+Au$ collisions (solid line
377: in figure \ref{strangefig}) cannot be extended to the data for $K^+/\pi^+$.
378:
379:
380:
381:
382: \section{Jet absorption}
383:
384: Jet quenching has been advocated as a tool for the tomography of the dense
385: phase in the collision \cite{jet2}. Experimental data show a significant
386: suppression of jets in $Au+Au$ collisions but not in $d+Au$ interactions,
387: suggesting that the reduction of the jet rate is due to the dense medium
388: created in heavy-ion collisions \cite{phobosjet,phenixjet,starjet,brahmsjet}.
389: These results were supported by the observation of the disappearance of back
390: to back azimuthal jet
391: correlations in central $Au+Au$ collisions \cite{starjet}.
392: The origin of the jet suppression is the energy loss of jets or the absorption
393: of jets
394: in the dense matter. In the following we use a model of jet
395: absorption \cite{jetabs} which can reproduce the jet suppression in the high
396: $p_\perp$ range.
397:
398: \begin{figure}
399: \centering
400: \includegraphics*[width=0.7\textwidth]{raa.eps}
401: \caption{$R_{AA}$ for charged hadrons and for $\pi^0$ integrated for
402: $p_\perp>4.5$GeV
403: as a function of centrality
404: \cite{phenixpt}. The absorption only in the thermalized source is
405: denoted by the dotted line. Absorption proportional to the
406: wounded nucleon density
407: is represented by the dashed line, the same is for the solid line but with an
408: absorption increased by $65\%$ in the thermal core.}
409: \label{raafig}
410: \end{figure}
411:
412: The survival probability
413: \begin{equation}
414: \label{abspro}
415: f=e^{-kI}
416: \end{equation} of a jet originating from
417: a binary collision at $(x,y)$ and traveling in the direction $(n_x,n_y)$
418: depends on the density of the matter along its
419: trajectory
420: \begin{equation}
421: \label{absint}
422: I=\int_\tau^\infty dl \rho(x+l n_x, y+l n_y) \ ,
423: \end{equation}
424: modifications due to the longitudinal expansion and the increase
425: of the absorption with $l$ largely cancel out and are not taken into account.
426: Taking for the density the wounded nucleon density
427: $\rho=d^2N_{part}/ds^2$, the jet suppression
428: at high transverse momentum can be reproduced \cite{jetabs}.
429: Using for the formation length of the jet $\tau=0.2$fm
430: and for the absorption
431: coefficient $k=0.12$fm, the PHENIX data on jet suppression
432: \cite{phenixpt} can be reproduced (dashed line in figure \ref{raafig}).
433:
434: In a different scenario, it can be assumed that the absorption occurs mainly
435: within the dense thermal core of the collision. To illustrate the effect
436: we use the extreme
437: limit with a large absorption within the deconfined thermal core
438: and no absorption in the mantle. This corresponds to changing $\rho$ to
439: $\theta(d^2N_{part}/ds^2-d_{cut})$
440: in equation (\ref{absint}) and taking
441: $k=0.8$fm$^{-1}$ in equation (\ref{abspro}). We use a larger value of
442: $\tau=0.8$fm, which
443: corresponds in this scenario to the formation time of the thermalized
444: opaque core. The jet suppression in the most central collisions
445: is reproduced (dotted line in figure \ref{raafig}).
446: In peripheral collisions the thermal
447: core disappears and the jet suppression ratio $R_{AA}$ goes to one, unlike
448: in the
449: data. If we believe that the dense phase is not
450: created in peripheral collisions, this
451: means that the jet suppression must occur in the lower density mantle part of
452: the interaction region as well. This observation confirms the theoretical
453: calculation of the jet suppression which increases smoothly with the density,
454: irrespectively of the nature of the absorbing medium \cite{baier}.
455:
456: Finally we perform a calculation taking into account the increased density
457: within the thermalized core. Following
458: the increase of the particle production we
459: take
460: \begin{equation}
461: \rho=\frac{d^2N_{part}}{ds^2}\theta
462: \left(d_{cut}-\frac{d^2N_{part}}{ds^2}\right)+(1+\alpha)
463: \frac{d^2N_{part}}{ds^2}\theta
464: \left(\frac{d^2N_{part}}{ds^2}-d_{cut}\right)
465: \end{equation}
466: Using the absorption coefficient $k=0.09$fm ($\tau=0.2$fm)
467: we reproduce the data in the
468: whole centrality range. This result shows that the existence of a
469: dense core within the
470: interaction region is compatible with the data on the jet suppression,
471: if the absorption is proportional to the (increased) density in the core.
472:
473:
474: \section{Conclusions}
475:
476: We study a model based on the separation of the interaction region
477: in $Au+Au$ collisions in two
478: distinct parts~: a thermalized core in the most dense region and an
479: outer mantle which is not thermalized. The inner core is rapidly
480: thermalized when passing the region of the critical temperature and above.
481: This includes both the kinetic and chemical equilibration. The
482: decay of the thermal core leads to particle production
483: per participant pair larger by 65\% than in $p+p$ interactions
484: at $\sqrt{s}=200$GeV. The charged particle production in the mantle
485: can be taken as in a $p+p$ collision scaled by the number of participant
486: pairs.
487: The thermalized core appears at around $N_{part}=40$, we expect
488: modifications of the density of produced particle and increased fluctuations
489: in this range of centralities due to fluctuations in the size
490: of the thermal source; these are not taken into account in this study.
491: $K^-/\pi^-$ ratio as a function of centrality can be reproduced as a sum of a
492: contribution of a chemically equilibrated thermal source and
493: of the production from the mantle. In the mantle an increased strangeness
494: per participant must be taken into account, following the $d+Au$ data.
495: High $p_\perp$ jet suppression is studied in a jet absorption model.
496: In order to reproduce the results in peripheral collisions, absorption must
497: occur in the dense core as well as in the mantle. The increase of the
498: density in the thermal core resulting in an increased absorption of jets
499: is compatible with the data. Applying the model to $Cu+Cu$ collisions at
500: $\sqrt{s}=200$GeV we find that the charged particle production
501: depends on the number of participants
502: similarly as in $Au+Au$ collisions, up to $N_{part}=120$. However, we expect
503: that in this case the physics could be modified by fluctuations of the
504: relative size of the core and of the mantle.
505:
506:
507:
508: \bibliography{../rhic}
509:
510:
511:
512: \end{document}
513:
514: