Quark Coalescence and Hadronic Equilibrium

T. S. Biró∗
CERN Theory Division, CH-1211 Geneve 23, Switzerland

Abstract In a simplified model we study how close can the result of a fast non-equilibrium process, like hadronization of
quark matter in an expanding fireball, come to hadronic composition in equilibrium. In a chemical approach of in-medium
quark fusion (A + B → C) we find that this depends on the time-integrated rate per volume multiplied by the conserved
number of constituents. Ideal hadron gas equilibrium constrained by conservation laws, the fugacity parametrization, as well
as linear and non-linear coalescence approaches are recognized as different approximations to this chemistry. It is shown that
color confinement requires a dependence of the hadronization cross section on quark density.
arXiv:hep-ph/0005067v2 27 Jun 2000

Since strangeness enhancement had been suggested as a possible signature of quark-gluon plasma formation in
heavy-ion collisions, there is a particular attention to the question of hadronic and quark level equilibrium in these
reactions [1–4]. Recent claims of describing data at AGS as well as at CERN SPS by constrained hadronic equilibrium
(i.e. using a common volume, temperature and only chemical potentials of conserved quantities for the computation
of hadron numbers) “quite” satisfactorily [5–9] led to some astounishment in the community: the total reaction time
at SPS energy is namely too short for achieving hadronic equilibrium by hadronic conversion processes alone. Pushing
aside the faux pais, that this “satisfactory” equilibrium description agrees with experimental data for a few ratios
only within a factor of two while the general agreement is on the 30% level, one faces the question, how can this
simple idea work even so well. The time argument above would exclude this.

Meanwhile there is a number of non-equilibrium approaches to this problem ranging from the use of phenomeno-
logical parameters describing the deviation on the hadronic level (fugacities) [10,11] through linear [12] and nonlinear
(constrained) coalescence models [13,14], and mixed quark-hadron chemistry [16] up to very detailed microscopic sim-
ulations of the hadronization (parton cascades [17–20] VENUS [21], HIJING [22,23], RQMD [24,25], UrQMD [26–29],
string model [30–33], chromodielectric model [34]). The basic physical concepts of these approaches oppose hadronic
equilibrium; it might be reached in course of time, but for some components it is actually not. For recent transport
simulations at SPS energies see [35]. There are also recent doubts whether chemical equilibrium can be reached at
RHIC [36].

The purpose of this letter is to identify a useful quantitative factor in describing the approach to hadronic equilibrium
states with quark level processes instead of hadronic ones in a fast expanding fireball. In order to concentrate on this
quantifying effort of the time evolution of a complicated system we shall deal with a simple fusion process on the
constituent level:

A + B ←→ C.

An example is given by the fusion of an s quark and a ud diquark to Λ particle. In this case baryon charge and
strangeness conservation constrains the equilibrium.

In such a process the energy-impulse balance is assumed to be satisfied by the surrounding medium, we are interested
in the time evolution of the number of final state products NC (t). The equilibrium number of precursors NA,eq and
NB,eq - in a quark matter scenario - will eventually be taken as zero. The initial state is assumed to be far from
this equilibrium, it is oversaturated by A and B. The simplest is to consider no C particles at all in the beginning.
As the reaction proceeds there will be C type particles at the end (or equilbrium will never be reached), they are
hadrons. The cross section for such a reaction - also as a simplification - can be taken as a constant parameter, but
temperature and reaction volume must be considered as time dependent quantities. We shall point out, however,
that the assumption of constant cross sections actually contradicts to the confinement principle in quark matter

∗
On leave from KFKI Budapest, Hungary

1
scenarios, so in that case the hadronization cross section has to be quark density dependent. This extra dependency,
not derivable from perturbative or lattice QCD so far, will be referred as “dynamical confinement” in this paper.

In the followings – after a brief presentation of general chemical equations – we shall consider the simple model
of hadronization due to coalescence (fusion in the presence of a medium). Comparison of different limiting cases
(allowing or excluding the precursors A and B in equilibrium) and estimates of the characteristic deviation from
hadronic equilibrium with and without dynamical confinement will close this study.

In the chemical approach to this process all particle numbers are time dependent. The entropy grows during a
(r)
spontaneous approach towards equilibrium. A general reaction can be characterized by stochiometric coefficients αi
in a fashion arranged to zero:
X (r)
αi Ni ←→ 0
i

(r)
for each reaction r. In this arrangement the αi -s can be negative as well as positive. An important quantity with
respect to approaching chemical equilibrium is the activity of a reaction,
X (r) µi
x(r) = αi . (1)
i
T

It describes the (un)balance of gain and loss terms. Rate equations summarize contributions from each reaction
channel,
X (r)  (r)

Ṅi = αi R(r) 1 − ex . (2)
(r)

Here the factors R(r) depend on numbers of other particles, cross sections and volume. Due to known (e.g. adiabatic)
expansion a relation between volume V and temperature T is given.

In simple scenarios of expansion V T 3/2 = const. (massive ideal gas expands adiabatically, no back effect from
chemistry is taken into account). Now

Ni = Ni,eq eµi /T (3)

and
 3/2
mi T
Ni,eq = V di e−mi /T . (4)
2π
In simple scenarios like this the time dependence of particle numbers, Ni (t), is entirely due to that of the
(µi (t) − mi )/T (t). The equilibrium ratio,
 3/2
NA,eq NB,eq dA dB mA mB
Req = = V T 3/2 e(mC −mA −mB )/T (5)
NC,eq dC mC

is nevertheless constant for adiabatic expansion (V T 3/2 =const.) and composite final products (mC = mA + mB ).

All rate equations drive the system towards chemical equilibrium. The total entropy production can be obtained
from

T dS + µi dNi = dE + pdV = 0 (6)

for adiabatic flow of an ideal (“dry”) fluid. One obtains

X µi X  (r) 
Ṡ = − Ṅi = R(r) x(r) ex − 1 ≥ 0. (7)
i
T r

2
In chemical equilibrium x(r) = 0 for each reaction, i.e.
X (r)
αi µi = 0. (8)
i

There are as many linear constraints for the chemical potentials µi as many reactions. If this is less than the number
of particle sorts, the equilibrium state is undetermined. If this is greater than the number of particle sorts, the
reactions cannot be all independent: there are conserved quantities (e.g. from Ṅ1 = Ṅ2 it follows that B = N1 − N2 is
conserved). In this latter case all µi can be expressed by a few numbers corresponding to these conserved quantities.
This is the constrained equilibrium.

Let us denote the conserved quantities by Qa (a = baryon, strangeness, etc.). The corresponding charge of the i-th
particle sort let be qi,a . We have
X
qi,a Ṅi = 0 (9)
i

Substituting the rate equations (2) we get

 (r)

(r)
XX
αi qi,a R(r) 1 − ex =0 (10)
i r

Since conservation is fulfilled even out of equilibrium, we conclude that

X (r)
αi qi,a = Qa(r) = 0, (11)
i

i.e. there is no charge unbalance in any of the reactions. This fact allows us to redefine the chemical potentials as
X
µi = qi,a µa + µ̃i . (12)
a

The first contribution, depending on only as many “chemical potentials” as conserved charges we have, does not
change the activities:

(r) qi,a µa (r) µ̃i

XX X
x(r) = αi + αi . (13)
i a
T i
T

Since the first part is zero due to eq.(11), chemical equilibrium (vanishing activities) means

(r) µ̃i
X
αi = 0. (14)
i
T

For a well-determined or overdetermined system the only equilibrium solution is given by µ̃i = 0. In this (constrained)
equilibrium state the original chemical potentials are a linear combination of a few µa -s, associated to conserved
charges,
X
µi,eq = qi,a µa . (15)
a

Whether this chemical equilibrium state can be reached in a finite time depends on the rates R(r) and activities
(r)
x . In this respect whether a quantity is conserved or not is determined by the set of rate equations. This way it
happens that not only real conserved charges but also the number of each constituent quark may be conserved in a
sudden hadronization process, like ALCOR assumes this.

3
 For ideal massive particles (Boltzmann distribution)
Ni
µi = T log . (16)
Ni,eq

The activities hence contain ratios of equilibrium numbers in each reaction (mass action law, detailed balance). For a
simplified expansion picture these ratios are (nearly) constant. The factors R(r) contain precursor numbers, averaged
cross sections (fluxes) and the reaction volume. In an expanding system its time-dependence is far from trivial.

For a general set of reactions and particle sorts the time evolution can only be followed numerically. In order to
understand, however, possible deviations from equilibrium in a final state - even after infinitely long time - we consider
an oversimplified system in the followings.

In the case of only one reaction channel, A + B ←→ C, we denote the equilibrium ratio by Req (cf. eq.(5)). Since
αA = αB = 1 and αC = −1 for this very reaction, we have ṄA = ṄB = −ṄC . From this two conservation laws follow,

NA + NC = NA (0) = N0 , (17)

NB + NC = NB (0) = N0 . (18)

For considering an evolution towards equilibrium the specific rate,

hσvi
ν= , (19)
V
is a useful quantity. It occurs in the rate equation

ṄC = ν (NA NB − Req NC ) . (20)

Substituting the conservation laws (17,18) we arrive at

ṄC = ν ((N0 − NC )(N0 − NC ) − Req NC ) = ν(NC − N+ )(NC − N− ). (21)

This rate is vanishing for the particle numbers NC = N± with

q
2 /4 + R N .
N± = Req /2 + N0 ± Req (22)
eq 0

Here N− is a stable, N+ is an unstable equilibrium point. For a non-relativistic, adiabatic expansion

V T 3/2 = const. (23)

and hence Req = const. for constituent hadrons (those with mass mC = mA + mB ). A realistic limit is Req = 0 when
no A or B particles are allowed in equilibrium. In particular this is the case for quark confinement.

The simple rate equation (21) can be solved analytically giving

NC (∞) tbr
dNC
Z Z
= νdt = r. (24)
NC (0) (NC − N+ )(NC − N− ) t0

Here t0 is the initial time, when the fireball contains no hadrons, NC (0) = 0. tbr is the break-up time, beyond which
no chemical reactions (even no particle collisions) occur. In some cases it is large compared to characteristic times
and can be approximated by infinity. Now using the identity
 
1 1 1 1
= − (25)
(NC − N+ )(NC − N− ) N+ − N− NC − N+ NC − N−
we arrive at

4
 log(NC − N+ ) − log(NC − N− )|0NC (∞) = (N+ − N− )r. (26)

This leads to the solution

1−K
NC (∞) = N+ N− (27)
N+ − KN−
with

K = exp (−(N+ − N− )r) . (28)

Here N− is the constrained equilibrium value, achieved for K = 0. There are two alternative forms of this relation
derived by using N+ N− = N02 . The first one is the coalescence form,
1−K
NC (∞) = N02 = C(K, N0 )N02 = C(K, N0 )NA (0)NB (0), (29)
N+ − KN−
the second one is the equilibrium form,
1−K
NC (∞) = N− = γC N− = γC NC,eq . (30)
1 − K(N− /N0 )2

In particular NC (∞) = N− in the K = 0 (infinite rate) case. On the other hand in the case of confinement Req = 0,
from which N+ = N− = N0 follows.

This agrees with the ALCOR assumption: after long enough time all quarks A and B are converted into hadrons
C. This means N− = N0 , satisfied for Req = 0, i.e. when the quark equilibrium numbers are zero (confinement in
equilibrium). Then ALCOR ends up with the equilibrium number. (Note: ALCOR does not take into account the
gain/loss balance as if it were infinite activity in favor of hadrons - against quarks. This feature is equivalent to
putting Req = 0.)

Without quark confinement, but (N+ − N− ) small (O(Req )), i.e. when precursors are suppressed in the equilibrium
state, we consider a negligible, but not zero Req . In the limit of Req → 0 the precursors A and B are suppressed in
equilibrium, according to the static confinement principle. This is, however, insufficient for hadronizing all quarks in
an expanding matter. We show this in the followings.

We expand K in terms of (N+ − N− ) which is in the same order of magnitude. We get

K = exp(−(N+ − N− )r) = 1 − r(N+ − N− ) (31)

and the final hadron number becomes

r
NC (∞) = N02 . (32)
1 + rN0

Eq. (32) – valid if the precursors are suppressed in chemical equilibrium (and this is all we know from lattice QCD)
– has different readings. For r small it comes close to the linear coalescence equation

NC (∞) = rN02 = rNA (0)NB (0). (33)

(Note that NA (0) = NB (0) = N0 .) ALCOR uses extra modification factors bA and bB by assuming the modified
coalescence rate

NC (∞) = rbA bB N02 . (34)

To start with it seems to be rather similar to the coalescence picture. Due to conservation, however, ALCOR obtains
r
NC (∞) = rbA bB N02 = N2 (35)
1 + rN0 0

5
and determines the factors bi accordingly (depending on N0 ). It leads exactly to the constrained equilibrium result
in the limit rN0 → ∞:

NC (∞) = N0 . (36)

So, reversing the argument, we conclude that for quark matter hadronization r has to diverge in order to reflect
confinement in the end stage.

Finally the chemistry approach gives the interpretation of the fugacity due to
r
NC (∞) = N02 = γC N 0 . (37)
1 + rN0
In fact the difference between these approaches is more pronounced by inspecting chemical potentials instead of final
numbers. In the discussed simple case the chemical potential of the (final state) hadron is given by

µC NC (∞)
= log . (38)
T NC,eq

Substituting our result for NC (∞) it splits to contributions with different physical meanings:

µC N− NC (∞) qC,1 µ1 + qC,2 µ2

= log + log = + log γC . (39)
T NC,eq N− T

The conserved quantities (charges) carried by the hadron C are those involved in the identities respected by the rate
equations, NA (t) + NC (t) = NA (0) = N0 and NB (t) + NC (t) = NB (0) = N0 . In this case qC,1 = 1 and qC,2 = 1,
respectively. The fugacity parameter (as used by Rafelski et al.) is the remaining part of the formula, in equilibrium
it is γC = 1.

Finally we note that the ALCOR approach (constrained coalescence) leads to equilibrium result only if a single
reaction channel is considered, like in the case discussed in the present paper. For several, competing channels the
relative hadron ratios are different for ALCOR and constrained equilibrium. This best can be seen by inspecting the
corresponding µi chemical potentials to the pre resonance decay hadron numbers as a function of strangeness in the
ALCOR simulation. Slight deviations from the perfect straight lines, which would mark the constrained equilibrium
in the ideal case, can be observed [37].

In case of confinement for particles A and B the equilibrium ratio Req vanishes and N− has to be replaced by N0
in the above formula. Exactly this way are usually the chemical potentials associated to conserved charges, µ1 and
µ2 , determined. The fugacity factor
1−K
γC = , (40)
1 − K(N− /N0 )2
in only static confining scenarios reduces to
rN0
γC = < 1. (41)
1 + rN0
This must be viewed as an indication for the necessity of considering dynamic confinement in hadronization.

There is no difference between ALCOR and constrained equilibrium in this simple model if confinement is im-
plemented: both lead to γC = 1 eventually. The linear coalescence picture on the other hand assumes rN0 ≪ 1,
corresponding to a rather impressive deviation from equilibrium. Therefore it is interesting to estimate the factor rN0
for different heavy-ion experiments.

We continue by estimating the integrated specific rate r. Its value depends not only on cross section(s) – first
assumed to be constant and later density dependent – but also on the expansion and cooling of the reaction zone.

6
We consider here two limiting cases: that of the one dimensional expansion with a rather relativistic equation of
state (ideal gas with particles having negligible masses) and that of three dimensional expansion of a massive, non-
relativistic ideal Boltzmann gas. The first case is thought to be rather characteristic for RHIC and LHC, the latter
rather for SP S and AGS energies.

During a one-dimensional expansion the volume grows linearly in time, V = πR02 t, and the (thermal) average of
relative velocities is near to the light speed hvi = 1. In this case we obtain
tbr
σ dt hσv0 i tbr
Z
r= = t0 log . (42)
t0 πR02 t V0 t0

The time-integrated specific rate grows with the break-up time tbr monotonically, it can eventually be infinite, if there
is no thermal break-up of the expanding system. In that case γC = 1, equilibrium is achieved. For a three-dimensional
3/2
expansion one gets V = V0 (t/t0 )3 . Furthermore a massive, non-relativistic ideal gas satisfiespV T 3/2 = V0 T0 during
an adiabatic expansion. The average relative velocity is temperature dependent like hvi ∼ T /m ∼ 1/t, with some
reduced mass of the fusing pair m. This together leads to the estimate

hσv0 i tbr 4 dt
Z
r= t0 4 . (43)
V0 t0 t

Evaluating the integral we arrive at

 3 !
1 hσv0 i t0
r= t0 1 − . (44)
3 V0 tbr

It is interesting to note that r is finite even without break-up, i.e. in the tbr → ∞ limit. This also means that the non-
relativistic, 3-dimensional spherical expansion scenario with constant hadronization cross section cannot comply with
quark confinement. In this case σ has to be time dependent at least as σ ∼ t3 ∼ V ∼ 1/n, i.e. the hadronization cross
section has to be inversely proportional to the density of quarks. Exactly this assumption was made in transchemistry
[16].

In orther to simplify the further discussion we relate the thermal average of the rate at the beginning of the
expansion to estimates of the collision time due to mean free path,
1
tc = , (45)
hσv0 iρ0

with density ρ0 = 2N0 /V0 . This leads to

hσv0 i 1
= . (46)
V0 2N0 tc
The quantity describing the deviation from constrained equilibrium for hadrons then is given as follows:
t0 tbr
rN0 = log (47)
2tc t0
for one-dimensional relativistic flow and
 3 !
t0 t0
rN0 = 1− (48)
6tc tbr

for three-dimesional non-relativistic flow of massive constituents. The general dependence of the fugacity on the
estimated collision time is in both cases
1
γC = , (49)
1 + β(tc /t0 )

7
just the constant β differs. For the realtivistic, one-dimensional flow case it is
2
β= (50)
log ttbr0

and for three-dimensional spherical flow of massive particles it is given by

6
β=  3 . (51)
t0
1− tbr

From these formuli one realizes that for tbr = t0 , i.e. in case of immediate break-up the chemical equilibrium cannot
be approached, no hadron C will be formed and γC remains zero. The opposite limit, tbr → ∞ is less uniform: while
in the relativistic one-dimensional scenario equilibrium is always reached if enough time is given (γC (∞) = 1), this is
not so for a stronger, more spherical expansion. In the latter case γC tends to a finite value,
1
γC (∞) = , (52)
1 + 6tc /t0
which is in general less than one. It means that chemical equilibrium cannot be really reached even in infinite time.
(An analysis of proton – deuteron mixture with a similar conclusion was given in Ref. [38].) For early break-up,
tbr − t0 ≪ t0 , both scenarios lead to
2t0
β= . (53)
tbr − t0

In quark matter hadronization scenarios eq.(52) contradicts to the quark confinement principle. In order to reach
γC (∞) = 1 a dynamical confinement mechanism, causing the quark fusion cross section in medium to be (at least)
inversely proportional to the quark density, has to be taken into account besides the equilibrium suppression of quark
number (referred in this paper as “static confinement”).

Let us assume that A and B are quark level (colored) objects, and the fusion cross section scales as
 1+ǫ
V 2N0
σ = σ0 (54)
V0 NA + NB

with some extra power ǫ. The solution of the rate equation in the Req = 0 case (both static and dynamic confinement)
becomes
 
NC = N0 1 − (1 − δ)1/ǫ , (55)

with
 3ǫ !
t0 t
δ= −1 . (56)
6tc t0

For ǫ = −1 (without dynamical confinement) one gets back the result discussed so far. In the limiting case ǫ = 0
(minimal dynamical confinement) one obtains
  2tt0 !
t0 c
NC = N0 1 − . (57)
t

Fig.1 plots the dependence of the non-equilibrium fugacity factor γC on the scaled break-up time tbr /t0 for these
scenarii: one and three dimensional scaling expansion, in the three-dimensional case both for static and dynamic con-
finement (color density dependent quark fusion cross section). The uppermost curve, showing the fastest equilibration

8
belongs to a 3-dimensional expansion scenario with a cross section proportional to the inverse of the quark density
(cf. eq.(57)). The lowest curve, saturating well below the equilibrium value, corresponds to the same expansion with
static confinement only. The middle curve shows the static RHIC scenario, reaching equilibrium (vanishing quark
number) eventually, but quite slowly. For all calculations we used the ratio t0 /tc = 3 for the sake of simplicity. (The
smaller t0 , the higher density, therefore the shorter inter-collision time tc initially.)

Equilibration
1

0.9 SPS dynamic

0.8
RHIC static
0.7

0.6
NC / N 0

0.5

0.4
SPS static

0.3

0.2

0.1

0
1 2 3 64 5
7 8 9 10
t br / t 0
FIG. 1. Degree of the chemical equilibrium reached for different break-up times tbr /t0 and initial inter-collison times tc /t0
for SPS and RHIC expansion scenarios with static and dynamic confinement.

Finally we compare characteristic values for SPS and RHIC. In the SPS scenarios we assume t0 = 1 fm/c as
the starting time of the expansion. The collision time changes somewhat due to different temperatures (and effective
quark masses) presumed at the beginning of hadronization. We consider a value characteristic for quark-gluon plasma,
tc ≈ 0.33 fm/c. We take over the ratio t0 /tc = 3 for the RHIC scenario, too. The break-up time is the most unknown
factor, therefore the equilibration has been plotted as function of tbr /t0 . Fortunately, once the stauration value of
γC has been approached, there is no strong dependence on it. The 1 − e−1 ≈ 0.63 fraction of equilibrium value is
reached after 2 fm/c for the dynamic SPS, after 3.5 fm/c for the static RHIC scenario. In the static SPS scenario it
will be never reached, γC saturates near 0.3 already after 2 fm/c. This last scenario, quark coalescence with density
independent probability, can be excluded on the basis of quark confinement.

In conclusion 1) the results of the statistical model can be interpreted not only by assuming hadronic equilibrium,
but as well by quark – hadron mixture equilibrium with quarks eventually suppressed due to confinement. This
interpretation avoids the collison time problem: the hadrons need not transmutate into each other directly. 2) The
ALCOR and constrained hadronic equilibrium leads to the same result in this simple scenario due to too many
constraints on conserved quantities. Quark coalescence, which takes conservation into account, therefore mimics
hadrons in chemical equilibrium. This feature certainly does not not generalize for several competing reaction channels,
but the overall picture can be similar. ALCOR actually assumes more conserved quantities, than usual: quark and

9
antiquark numbers in any flavor are conserved separately due to the assumption of rapid hadronization. 3) γC ≈ 0.3 is
typical for CERN SPS without considering dynamical confinement, even for an arbitrary long hadronization process.
This leads to a factor two to three characteristic error on equilibrium model results. The experimental evidence,
reaching at least 70-80% of equilibrium ratios for almost all hadrons in a short time in the order of several fm/c,
excludes this possibility. It also underlines the quark matter hadronization scenario with dynamical confinement
effect, as transchemistry describes it. 4) An expansion scenario typical for RHIC and LHC energies in principle
allows for reaching the equilibrium even with static confinement, but the process is rather slow and an early break-up
may prevent from hadronizing all quarks. Dynamical confinement has to be in work for one-dimensional expansion
scenarios, too.

One is tempted to speculate about the origin of dynamical confinement, looking for a cause of the density dependence
of the hadronization cross section. Since this dependence makes the cross section stronger with diminishing quark
density, it cannot have perturbative origin; on the contrary it has to do with formation of strings or color ropes. As
the color density drops, strings, follow-ups the chromoelectric flux lines, become longer before ending on a partner
color. The volume, in which possible partner to form a hadron occur, hence increases, so does the hadronization cross
−1/3
section, σ ∼ ℓ3 . The characteristic string length on the other hand depends on the color density as ∼ ncolor . Such a
mechanism would lead to σ = σ0 ncolor (0)/ncolor .

Acknowledgements Discussions with J. Zimányi, K. Redlich and U. Heinz are gratefully acknowledged. This
work was supported by the US-Hungarian Joint Fund TéT 649, by the Deutsche Forschungsgemeinschaft DFG-MTA
101 and by the Hungarian National Research Fund OTKA T029158.

[1] J. Rafelski and B. Müller, “Strangeness Production In The Quark - Gluon Plasma,” Phys. Rev. Lett. 48 (1982) 1066.
[2] T. Biró and J. Zimányi, “Quarkochemistry In Relativistic Heavy Ion Collisions,” Phys. Lett. B113 (1982) 6.
[3] T. S. Biró and J. Zimányi, “Quark Gluon Plasma Formation In Heavy Ion Collisions And Quarkochemistry,” Nucl. Phys.
A395 (1983) 525.
[4] P. Koch, B. Müller and J. Rafelski, “Strangeness In Relativistic Heavy Ion Collisions,” Phys. Rept. 142 (1986) 167.
[5] P. Braun-Münzinger, I. Heppe and J. Stachel, “Chemical equilibration in Pb + Pb collisions at the SPS,” Phys. Lett.
B465 (1999) 15 [nucl-th/9903010].
[6] F. Becattini, M. Gazdzicki and J. Sollfrank, “Thermal fits of hadron abundances from p p to A A collisions,” Nucl. Phys.
A638 (1998) 403.
[7] F. Becattini, M. Gazdzicki and J. Sollfrank, “On chemical equilibrium in nuclear collisions,” Eur. Phys. J. C5 (1998) 143
[hep-ph/9710529].
[8] J. Cleymans, H. Oeschler and K. Redlich, “Particle ratios at SPS, AGS and SIS,” J. Phys. G G25 (1999) 281 [nucl-
th/9809031].
[9] N. J. Davidson, H. G. Miller, D. W. von Oertzen and K. Redlich, “Thermal gas models and particle production in heavy
ion collisions,” Z. Phys. C56 (1992) 319.
[10] J. Rafelski, J. Letessier and A. Tounsi, “Thermal strangeness and charm in QGP,” Heavy Ion Phys. 4 (1996) 181.
[11] J. Letessier and J. Rafelski, “Chemical non-equilibrium and deconfinement in 200-A-GeV sulphur induced reactions,” Phys.
Rev. C59 (1999) 947 [hep-ph/9806386].
[12] A. Bialas, “Quark model and strange baryon production in heavy ion collisions,” Phys. Lett. B442 (1998) 449 [hep-
ph/9808434].
[13] T. S. Biró, P. Lévai and J. Zimányi, “ALCOR: A Dynamic model for hadronization,” Phys. Lett. B347 (1995) 6.
[14] J. Zimányi, T. S. Biró and P. Lévai, “The dependence of strange hadron multiplicities on the speed of hadronization,” J.
Phys. G G23 (1997) 1941 [hep-ph/9712352].
[15] J. Zimányi, T. S. Biró and P. Lévai, “The mischievous linear coalescence model and the correct quark counting in heavy
ion collisions,” hep-ph/9904501.
[16] T. S. Biró, P. Lévai and J. Zimányi, “Hadronization with a confining equation of state,” Phys. Rev. C59 (1999) 1574.
[17] K. Geiger and B. Müller, “Parton cascades in highly relativistic nuclear collisions,” Nucl. Phys. A544 (1992) 467C.
[18] K. Geiger, “Particle production in high-energy nuclear collisions: A Parton cascade cluster hadronization model,” Phys.
Rev. D47 (1993) 133.

10
[19] K. Geiger, “Non-equilibrium QCD: Interplay of hard and soft dynamics in high-energy multi-gluon beams,” Phys. Rev.
D56 (1997) 2665 [hep-ph/9611400].
[20] K. Geiger, R. Longacre and D. K. Srivastava, “VNI version 4.1: Simulation of high-energy particle collisions in QCD:
Space-time evolution of e+ e- –¿ A + B collisions with parton cascades, parton hadron conversion, final-state hadron
cascades,” nucl-th/9806102.
[21] K. Werner, J. Hufner, M. Kutschera and O. Nachtmann, “Particle Production In Hadron Nucleus Collisions At 100-Gev
From A Parton String Model,” Z. Phys. C37 (1987) 57.
[22] K. J. Eskola, B. Müller and X. Wang, “Screening of initial parton production in ultrarelativistic heavy-ion collisions,”
Phys. Lett. B374 (1996) 20 [hep-ph/9509285].
[23] X. Wang, “Dynamical screening and radiative parton energy loss in a quark-gluon plasma,” nucl-th/0003033.
[24] H. Sorge, “Xi and Anti-xi production in central S (200-A-GeV) on W collisions from RQMD,” Phys. Lett. B344 (1995)
35.
[25] J. Sollfrank, U. Heinz, H. Sorge and N. Xu, “Thermal analysis of hadron multiplicities from RQMD,” J. Phys. G G25
(1999) 363 [nucl-th/9811011].
[26] F. Antinori et al. [WA97 Collaboration], “Strangeness enhancement at midrapidity in Pb Pb collisions at 158-A-GeV/c:
A comparison with VENUS and RQMD models,” Eur. Phys. J. C11 (1999) 79.
[27] S. A. Bass et al., “URQMD: A new molecular dynamics model from GANIL to CERN energies,” Prepared for International
Conference on Nuclear Physics at the Turn of Millennium: Structure of Vacuum and Elementary Matter, Wilderness /
George, South Africa, 10-16 Mar 1996.
[28] S. A. Bass, M. Gyulassy, H. Stöcker and W. Greiner, “Signatures of quark-gluon-plasma formation in high energy heavy-ion
collisions: A critical review,” J. Phys. G G25 (1999) R1 [hep-ph/9810281].
[29] M. Bleicher, S. A. Bass, H. Stöcker and W. Greiner, “RHIC predictions from the UrQMD model,” hep-ph/9906398.
[30] N. S. Amelin and L. V. Bravina, “The Monte Carlo Realization Of Quark - Gluon String Model For Description Of
High-Energy Hadron-Hadron Interactions,” Sov. J. Nucl. Phys. 51 (1990) 133.
[31] D. J. Dean, M. Gyulassy, B. Müller, E. A. Remler, M. R. Strayer, A. S. Umar and J. S. Wu, “Multiparticle production in
lepton nucleus collisions and relativistic string models,” Phys. Rev. C46 (1992) 2066.
[32] S. A. Bass et al., “Microscopic models for ultrarelativistic heavy ion collisions,” Prog. Part. Nucl. Phys. 41 (1998) 225
[nucl-th/9803035].
[33] P. Csizmadia, P. Lévai, S. E. Vance, T. S. Biró, M. Gyulassy and J. Zimányi, “Strange hyperon and antihyperon production
from quark and string-rope matter,” J. Phys. G G25 (1999) 321 [hep-ph/9809456].
[34] C. T. Traxler, U. Mosel and T. S. Biró, “Hadronization of a quark-gluon plasma in the chromodielectric model,” Phys.
Rev. C59 (1999) 1620 [hep-ph/9808298].
[35] E. L. Bratkovskaya, W. Cassing, C. Greiner, M. Effenberger, U. Mosel and A. Sibirtsev, “Aspects of thermal and chemical
equilibration of hadronic matter,” nucl-th/0001008.
[36] S. J. Lindenbaum and R. S. Longacre, “Possible striking signals for a quark-gluon plasma at RHIC,” nucl-th/0004019.
[37] T. S. Biró, P. Lévai, J. Zimányi, “Strange Hadrons from Quark and Hadron Matter: a Comparison”, contribution to S’2000
Conference, Berkeley, July 20-25, 2000.
[38] T. S. Biró, “Asymptotic Hadrochemistry,” Phys. Lett. B143 (1984) 50.

11