Particle Production via Strings and Baryon

Stopping within a Hadronic Transport Approach

J Mohs1,2,3 , S Ryu1,2 and H Elfner1,2,3
1
Frankfurt Institute For Advanced Studies, Ruth-Moufang-Strae 1, 60438 Frankfurt
am Main
2
Institute for Theoretical Physics, Goethe University, Max-von-Laue-Strasse 1,
60438 Frankfurt am Main, Germany
3
GSI Helmholtzzentrum fr Schwerionenforschung, Planckstr. 1, 64291 Darmstadt,
Germany
E-mail: mohs@[Link]

9 April 2019

Abstract. The stopping of baryons in heavy ion collisions at SPS energies is lacking
a quantitative description within theoretical calculations. In this work we implement
a string model for treating hadron-hadron interactions within a hadronic transport
approach. Free parameters of the string routine are tuned to match experimental
measurements in proton-proton collisions. The model is applied to heavy ion collisions,
where the experimentally observed change of the shape of the proton rapidity spectrum
from a single peak structure to a double peak structure with increasing beam energy
is reproduced. Heavy ion collisions provide the opportunity to study the formation
process of string fragments. By adjusting details of the interactions between string
fragments a good agreement with the measured rapidity spectra of protons and pions
is achieved while insights on the fragmentation process are obtained.

Keywords: baryon stopping, string fragmentation, hadron transport, heavy ion collisions

1. Introduction

• What are we trying to do

• Explain what is experimentally observed for baryon stopping
• Reference other models
• Importance of an initial state for hydro, especially 3D at low beam energy

The properties of strongly interacting matter are studied in heavy ion collisions.
In order to obtain a deeper understanding of the evolution of a heavy ion collision,
dynamical models need to be applied. Descriptions of heavy ion collisions within
hydrodynamic simulations prove to be very successful in the bulk observables Find
good citations [Schenke:2010nt].
Particle Production via Strings and Baryon Stopping 2

Since the experimentally observed stopping of baryons is still lacking a quantitative

description within theoretical models, we apply the hadronic transport model SMASH
to understand the stopping of baryons in the SPS energy range. Within this approach,
one does not rely on the system being in thermal equilibrium which is important for the
applicability in the early stages of a heavy ion collision. The full evolution of a heavy ion
collision is calculated within the hadronic transport approach, but also initial conditions
for a hydrodynamic evolution within a hybrid approach are provided. In the relevant
energy range for this work, it is extremely important to have three-dimensional initial
conditions for starting a hydrodynamic evolution, since the system cannot be assumed
to be boost invariant and the colliding nuclei are too slow to reasonably neglect their
longitudinal extend because of length contraction.
The paper is structured as follows: Details of the transport model with a focus
on the implementation of cross sections and the particle production at intermediate
energies are given in Section 2. Section 3 continues with calculations for proton-proton
collisions, where the influence of changing the free parameters is investigated and the
best possible set of parameters is found. In Section 4 we advance to heavy ion collisions,
where one has the opportunity to study the interactions of string fragments and the
formation process. Finally we provide calculations for the time of the collision where
the colliding nuclei overlap in Section 5.

2. Model Description

In this work we investigate baryon stopping within the transport model SMASH
[Weil:2016zrk]. The degrees of freedom within the calculation are hadrons. The
properties of the hadrons originate from the Particle Data Group [PDG], where the
more established resonances up to a mass of m ≈ 2 GeV are included. Probably too
brief and uncomprehensive
The inelastic scatterings between hadrons at low energies are described via
resonance formations and decays. Since there are no resonances with masses larger
than m ≈ 2 GeV in the calculation, the cross section for resonance formations fades out
when the center of mass energy of the interacting hadrons is too large. This can be seen
as an example for the proton-pion cross section shown in Figure 1. TODO: actually the
cross section fades out because we turn it off in this figure. Do I need to clarify that?
In order to investigate baryon stopping at higher incident energies, a different model for
calculating the final state of hadronic interactions at high energy has to be employed.
In this work, we implement a string model, where colliding hadrons are excited
to strings which then fragment producing new particles. The calculation for the string
excitation is split into hard and soft processes. The hard string processes are relevant
for very high energetic binary interactions as can be seen in Figure 1, where the cross
section for the hard processes becomes dominant going to large center of mass energies.
In the energy range where the hard string processes dominate, perturbative QCD is
applicable. For the description of the pQCD scatterings, the string excitation and the
Particle Production via Strings and Baryon Stopping 3

string fragmentation, Pythia [Andersson:1983ia, Sjostrand:2014zea] is used. The

hard string routine in described in more detail in section 2.1.
In the transition region where the energy is too large to have cross sections via
resonances but too low to apply pQCD, a phenomenological model for the excitation
of strings is implemented. In single diffractive, double diffractive and non-diffractive
processes, strings are excited in hadronic interactions. Using the calculated mass and
momentum of the string as well as the flavor of the leading quarks, Pythia is employed
only for the fragmentation of the string. Details of the string excitation at intermediate
energies can be found in Section 2.2.
Figure 1 also shows the contribution of the elastic scatterings to the total cross
section. Elastic collisions play an important role at all beam energies. In heavy ion
collisions, a large amount of nucleons will scatter elastically. Especially for baryon
stopping in heavy ion collisions the angular distribution of the final state particles in
elastic scatterings plays an important role.

70 p+π−
Total
60 Elastic
Resonance
50 Soft string
Hard string
40
σ [mb]

30

20

10

0
10 1 10 2 10 3
p
s [GeV]
Figure 1. Cross section of a proton interacting with a negatively charged pion as a
function of the center of mass energy of the colliding hadrons within SMASH. The total
cross section is split into contributions from elastic collisions, resonance formations, soft
string excitations and hard string excitations via Pythia.

2.1. Hard String Routine

Hard string excitations dominate the hadronic cross section at large center of mass
energies. In the energy range where the hard processes are mainly applied, Pythia
has proven to describe the experimental findings in proton-proton collisions very well.
The Reason for that being, that hard parton interactions, which can be calculated using
perturbative QCD methods, play the most important role at large energies. Pythia is
therefore employed for the entire calculation of each individual hadronic interaction at
Particle Production via Strings and Baryon Stopping 4

high energies. The cross section σhard for partonic interactions is calculated as follows
[Sjostrand:1987su]:
XZ Z
k
σhard = dx1 dx2 fi (x1 )fj (x2 )σi,j , (1)
i,j,k
k
where σi,j is the cross section for a subprocess k between two partons i and j. fi (x) gives
the probability of parton i carrying the momentum fraction x of the ingoing hadron.
k
While the σi,j can be calculated from pQCD, the structure functions fi (x) are determined
by soft dynamics within the colliding hadrons. Pythia includes the structure functions
for (anti-)nucleons and pions. There is not much information available on the structure
functions of other hadron species.
It is therefore not possible to directly use Pythia for the calculation of collisions
between more exotic hadrons. When investigation heavy ion collisions, it is necessary
to be able to describe the secondary interactions between other hadron species than
nucleons and pions. Hence, the assumption is made that all mesons have a similar
structure function as pions and all (anti-)baryons have a similar structure function
as (anti-)nucleons. Technically, this is achieved by mapping different hadron species to
(anti-)nucleons and pions, where the baryon number of the original and mapped particle
are the same. If possible also the electric charge of the original and mapped hadrons is
equal.
In case it is not possible to keep the electric charge, like in the case of a negatively
charged baryon being mapped to a neutron, no large error occurs due to isospin
symmetry. Before the produced strings are fragmented within the Pythia calculation,
light (anti-)quarks are exchanged with quarks of different flavor in order to restore
the original flavors that existed before the mapping of hadrons to nucleons and pions.
Changing the flavor of quarks in the intermediate state before the fragmentation in
Pythia affects the mass, and therefore also the total energy of the system. Hence, the
momenta of all particles are rescaled in order to conserve the energy of the system.
Due to annihilation processes, it is not always possible to find a quark with the
flavor of the mapped quark. In the rare case where all quarks with the flavor of the
mapped quark have annihilated, there is no possibility to restore the initial total flavor
of the system by just changing the flavor of one quark to the original flavor.. In this case
a gluon is split into a quark-antiquark pair with the flavor of the mapped quark. After
that, the produced quark is exchanged to have the flavor of the original quark. If there
is no gluon in the intermediate state, the final attempt of restoring the quark content
of the system is by replacing an anti-quark fo the original quark by an anti-quark of the
flavor of the mapped quark.

2.2. Soft String Routine

• Emphasize why we don’t want to use pythia directly in this energy region
– pQCD processes in Pythia
– performance?
Particle Production via Strings and Baryon Stopping 5

– What else?
• Explain subprocesses
• Highlight differences to UrQMD
– Fragmentation function
– Formation times depend on mass
– what else?

In the transition region from low energies where resonances are formed to high
energies where pQCD applies, a phenomenological model for the soft excitation of strings
is employed for the calculation of hadronic interactions. The soft string excitations are
the most abundant processes in the intermediate energy range as can be seen in Figure
1 where the total cross section is dominated by the contribution from the soft string
routine. As in UrQMD [Bass:1998ca, Bleicher:1999xi], the excitation of a soft string
can be performed according to one of the three subprocesses. The first subprocess
is the single diffractive case. In this picture, two ingoing hadrons interact, exchange
momentum and one of the two hadrons is excited to form a string. Details on this
subprocess are gathered in Section 2.2.1. Another possibility is the double diffractive
case which is described in Section 2.2.2. In this scenario, the two colliding hadrons are
both excited to form a string after momentum exchange. Finally, the most common
subprocess is the non-diffractive case. Here, the interacting hadrons exchange one of
their valence quarks including the momentum that this quark carries. Again, both
hadrons are excited to form a string as described in 2.2.3.

2.2.1. Single Diffractive The single diffractive process describes the interaction
between two hadrons, where exactly one of the two colliding hadrons A and B is excited
to form a string X:
A+B →A+X or A + B → X + B (2)
The excited string X has a larger mass than the ingoing hadron. The string mass MX
has to be determined in order to be able to calculate the dynamics of the collision. In
single diffractive processes, the string mass is distributed as follows [Ingelman:1984ns]:
dN 1
2
∝ 2 (3)
dMX MX
For the momentum transfer between the two colliding hadrons [Ingelman:1984ns]
suggests a distribution which only slightly deviates from an exponential for the
Mandelstam variable t. As in UrQMD [Bass:1998ca], since −t is the squared
momentum transfer, the transverse momentum transfer between the incoming hadrons
is also assumed to follow a Gaussian distribution in this approach:
d2 N
 2
p
∝ exp − T2 (4)
dpx dpy σT
Particle Production via Strings and Baryon Stopping 6

Here, σT is a free parameter which regulates the transverse momentum that is transferred
between hadrons in diffractive processes. The influence of this parameter on the
observables in proton-proton collisions is investigated in Section 3.2.
In the center of mass frame of the colliding hadrons, the squared energy s is given
by:
 q q 
0 2 2 2 2 2 2 2 2
s = (pA + pX ) = mA + mX + 2 p + mA + p mX + p , (5)

where p0A and pX are the momenta of hadron A after the collision and the string X.
mA denotes the mass of hadron A. p is the absolute momentum of both A and X in
the center of mass frame, where the momenta of the two particles are equal but point
into opposite directions. Equation 5 can be solved for p in order to determine the
longitudinal momentum pz of both A and X using p2z = p2 − p2T . Knowing the mass
and the momenta of the reaction products, one can calculate the velocity of the string
in order to boost to the rest frame, where one can apply the fragmentation machinery
from Pythia to obtain the particles in the final state of the interaction.

2.2.2. Double Diffractive The double diffractive subprocess describes a collision of two
hadrons A and B which are both excited to strings:
A+B →X +X (6)
The dynamics of the interaction are calculated in the center of mass frame of the two
colliding hadrons, where the collision axis is set to be the longitudinal direction. The
longitudinal momentum transfer is modeled via the exchange of gluons. Each hadron
emits one gluon, which the other hadron absorbs. The light cone momentum fraction
x of each gluon is sampled from the parton distribution function for gluons, which is
assumed to be of the form:
1
PDF ∝ (1 − x)β+1 , (7)
x
where β is set to be 0.5. The light cone momenta p± of the hadrons are given by
E ± pz
p± = √ , (8)
2
where E is the energy of the hadron and pz is the projection of the momentum on
the collision axis of the colliding hadrons. The light cone momenta of the exchanged
gluons are calculated as p± ± ±
g = x p . The distribution for the momentum transfer in each
transverse direction px and py between the colliding hadrons is taken to be the same as in
the single diffractive case, see Equation 4. Also the width of the distribution is the same
in both subprocesses. With the transverse momentum transfer pT between the hadrons,
the light cone momenutm transfer is given by [Bass:1998ca, Bleicher:1999xi]:
p2T
Q+ = − (9)
2x+
B pB
+

p2T
Q− = (10)
2x−
A pA
−
Particle Production via Strings and Baryon Stopping 7

The transfered energy ∆E from B to A and longitudinal momentum ∆pT follows as:
Q+ + Q− Q+ − Q−
∆E = √ ; ∆pz = √ (11)
2 2
The mass of both excited strings can be calculated individually using the energy-
momentum relation. Each string is then fragmented in the rest frame of that string
using the implementation of the fragmentation within Pythia.

2.2.3. Non-Diffractive The last of the three subprocesses of the soft string routine is the
soft non-diffractive case. The non-diffractive string excitation is the most probable soft
process, and therefore has th largest impact on the dynamics of the produced particles
in the SPS energy region. During the interaction, each hadron emits one valence quark,
which is adopted by the other hadron. The exchanged valence quark carries a fraction of
the longitudinal momentum of the hadron it is emitted from. The light cone momentum
fraction carried by the exchanged quark is sampled according to the parton distribution
function for quarks, which is assumed to have the following functional form:
PDF ∝ xα−1 (1 − x)β−1 , (12)
α and β are in principal free parameters. Since the assumption of the model is to always
exchange a valence quark, the parton distribution function must have a mean value close
to 1/3. Therefore we set the value of the parameters to α = 2.0 and β = 5.0, leading
to the PDF shown in Figure 2. The momentum transfer in the transverse direction is

0.25
α = 2, β = 5
0.20

0.15
PDF

0.10

0.05

0.00
0.0 0.2 0.4 0.6 0.8 1.0
x
Figure 2. Parton distribution function used for sampling the light cone momentum
fraction of valence quarks.

sampled according to the same Gaussian as in the single diffractive and double diffractive
case, using Equation 4. With the light cone momentum fraction each exchanged quark
carries and the transferred transverse momentum, the light cone momentum transfer is
written as follows [Bass:1998ca, Bleicher:1999xi]:
p2T
Q+ = −x+ +
A pA + (13)
2x−
B pB
−
Particle Production via Strings and Baryon Stopping 8
p2T
Q− = x− −
B pB − + , (14)
2x+
A pA

where xA and xB are the light cone momentum fractions for the exchanged quarks, p± A
and p±B are the light cone momenta of the colliding hadrons before the collision and
pT is the transferred transverse momentum. The exchanged energy and longitudinal
momentum can be calculated using Equation 11. The masses of the strings are
obtained using the relativistic energy-momentum relation. and each string is fragmented
individually in the rest frame of the string using Pythia.

2.3. Cross Sections for String Processes

During a calculation within SMASH, not only the total cross section for each pair of
hadrons is necessary. The cross section of each individual string subprocess is needed
separately in order to be able to randomly choose which of the processes contributing to
the total cross section is actually performed. The total cross section σtot , which is shown
for p-π − collisions in figure 1, is parameterized to fit the experimental data. The same
goes for the elastic cross section σel . The inelastic cross section σinel is the difference of
the two:
σinel = σtot − σel (15)
For the process selection, cross sections for both the single σSD and double diffractive
σDD processes are necessary. They are estimated in [Schuler:1993wr] and implemented
in Pythia. From σSD and σDD the non-diffractive cross section σND is derived:
σND = σinel − σSD − σDD (16)
The non diffractive cross section includes the hard and soft non-diffractive processes
however. The cross section for hard non-diffractive processes is based on partonic corss
sections and is given in equation 1. The sum in this calculation takes each possible
combination of one parton from each ingoing hadron into account. This cross section
can therefore be larger than σND , incorporating the information of multiple partonic
scatterings occurring.
For the selection of string processes, the number of partonic scatterings is not
meaningful. The question is rather whether the hadrons interact or not. In order
to extract this information from σhard , the number of parton interactions is assumed
to follow a Poissonian distribution, where the average is given by σhard /σND . The
probability of having no hard interaction is calculated according to the Poissonian
distribution as:
 
σhard
P (0) = exp − (17)
σND
In this case, the process is assumed to be soft non-diffractive, leading to a soft non-
diffractive cross section σND,soft of:
 
σhard
σND,soft = σND exp − (18)
σND
Particle Production via Strings and Baryon Stopping 9

Finally, the cross section σND,hard for the hard string process follows as:
σND,hard = σND − σND,soft (19)

2.4. Particle Formation

• I don’t know if we want to have Figure 3 (If we do, we need to change it so it
doesn’t imply the formation times to be equal)
√
• Explain how formation times are calculated hτ i = 2 m
κ
• Mention that they are not fluctuating
• Introduce scaling factors and how they can increase with time
• Explain initial scaling factors of leading fragments.
 α
t − tprod
fσ (t) = (1 − f0 ) + f0 (20)
tform − tprod

3.5

3.0

2.5
t [fm]

2.0
τ = 1 fm
1.5

1.0

0.5

0.0
2 1 0 1 2
x [fm]
Figure 3. Image of a string fragmenting within the yoyo model.

2.5. Elastic Collisions

3. Proton-Proton Collisions

Necessary Subsections:
• Fragmentation Function
– Leading frag mean
– leading frag width
• Lund fragmentation a
• Lund Fragmentation b
• Transverse momentum transfer
Particle Production via Strings and Baryon Stopping 10

1.2
α = 0. 5
1.0 α = 1. 0
α = 2. 0
α = 3. 0
0.8 α = 1000. 0
0.6

fσ
0.4

0.2

0.0
tprod tform
t
Figure 4. Cross section scaling factor fσ as a function of time for different powers α
with which the cross section grows in time. In this example, the initial cross section
scaling factor is set to be fσ (tprod ) = 0.

• Fragmentation transverse momentum

• Strangeness suppression
• Diquark suppression
• Result collection

3.1. Fragmentation Function

p
1.4 p sNN = 17. 27 GeV
SMASH
1.2 SMASH Lund fragmentation
NA49 data
1.0
dN/dxF

0.8

0.6

0.4

0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0
xF
Figure 5.

3.2. Transverse Momentum

4. Heavy Ion Collisions

5. Initial State Calculations

Particle Production via Strings and Baryon Stopping 11
p
p sNN = 17. 27 GeV π + p sNN = 17. 27 GeV
0.5 0.5

0.4 0.4
pT [GeV]

pT [GeV]
0.3 0.3
®

®
0.2 0.2
­

­
SMASH SMASH
0.1 0.1
SMASH Lund fragmentation SMASH Lund fragmentation
NA49 data NA49 data
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
xF xF
Figure 6.

p+p→p+X
0.25
− mp [GeV]

0.20

0.15

0.10
®

UrQMD 3. 4, 0 < y < 0. 2

mT

HSD 2. 5, 0 < y < 0. 2

0.05
­

NA61, − 0. 4 < y < − 0. 2

SMASH, − 0. 4 < y < − 0. 2
0.00
5 10 15 20
p
sNN [GeV]
Figure 7.

p 0.08 p̄
6 Lines from SMASH Lines from SMASH
NA 61 plab = 20 GeV × 1 0.07 NA 61 plab = 30. 9 GeV × 1
5 NA 61 plab = 30. 9 GeV × 2 NA 61 plab = 40 GeV × 2
NA 61 plab = 40 GeV × 4 0.06 NA 61 plab = 80 GeV × 4
NA 61 plab = 80 GeV × 8 NA 61 plab = 158 GeV × 4
4 NA 61 plab = 158 GeV × 16 0.05
dN/dy

dN/dy

3 0.04

0.03
2
0.02
1
0.01

0 0.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
y y
Figure 8.
Particle Production via Strings and Baryon Stopping 12

14 π+ 14 π−
Lines from SMASH Lines from SMASH
12 NA 61 plab = 20 GeV × 1 12 NA 61 plab = 20 GeV × 1
NA 61 plab = 30. 9 GeV × 2 NA 61 plab = 30. 9 GeV × 2
10 NA 61 plab = 40 GeV × 4 10 NA 61 plab = 40 GeV × 4
NA 61 plab = 80 GeV × 8 NA 61 plab = 80 GeV × 8
NA 61 plab = 158 GeV × 16 NA 61 plab = 158 GeV × 16
dN/dy

dN/dy
8 8

6 6

4 4

2 2

0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
y y
Figure 9.

1.2 K+ 0.06 K−
Lines from SMASH Lines from SMASH
NA 61 plab = 20 GeV × 1 NA 61 plab = 20 GeV × 1
1.0 0.05
NA 61 plab = 30. 9 GeV × 2 NA 61 plab = 30. 9 GeV × 1
NA 61 plab = 40 GeV × 4 NA 61 plab = 40 GeV × 1
0.8 NA 61 plab = 80 GeV × 8 0.04 NA 61 plab = 80 GeV × 1
NA 61 plab = 158 GeV × 16 NA 61 plab = 158 GeV × 1
dN/dy

dN/dy

0.6 0.03

0.4 0.02

0.2 0.01

0.0 0.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
y y
Figure 10.

p
45 Pb + Pb, sNN = 17. 27 GeV
40
bform = 0. 8
bform = 1. 0
bform = 1. 5
p
35
NA49
30
dN/dy

25
20
15
10
5
0
4 3 2 1 0 1 2 3 4
y
Figure 11.
Particle Production via Strings and Baryon Stopping 13
p
200 Pb + Pb, sNN = 17. 27 GeV
bform = 0. 8
bform = 1. 0
bform = 1. 5
π−
150 NA49
dN/dy
100

50

0
4 3 2 1 0 1 2 3 4
y
Figure 12. Need errorbars for data!!

p
50 Pb + Pb, sNN = 8. 765 GeV

40
αform = 1. 0
αform = 2. 0
αform = − 1. 0
p
NA49

30
dN/dy

20

10

0
3 2 1 0 1 2 3
y
Figure 13.
Particle Production via Strings and Baryon Stopping 14
p
140 Pb + Pb, sNN = 8. 765 GeV
120
αform = 1. 0
αform = 2. 0
αform = − 1. 0
π−
100 NA49
dN/dy
80

60

40

20

0
3 2 1 0 1 2 3
y
Figure 14.

60
p
p, Pb + Pb
sNN = 17. 27 GeV
p
sNN = 8. 765 GeV
50 p
sNN = 6. 27 GeV
SMASH
40 UrQMDv. 2. 3
NA49data
dN/dy

30

20

10

0
6 4 2 0 2 4 6
y
Figure 15.
Particle Production via Strings and Baryon Stopping 15

200
p
π − , Pb + Pb
sNN = 17. 27 GeV
p
sNN = 8. 765 GeV
p
sNN = 6. 27 GeV
150 SMASH
UrQMDv. 2. 3
NA49data
dN/dy
100

50

0
6 4 2 0 2 4 6
y
Figure 16.

p
10 Pb + Pb at sNN = 17. 27 GeV, b = 2. 0 fm
27
24
5
21
² [GeV/fm 3 ]

18
y [fm]

0 15
12
9
5
6
3
10 0
10 5 0 5 10
x [fm]