Hadronic flow in p Pb collisions at the Large Hadron Collider?

Transcription

1 Hadronic flow in Pb collisions at the Large Hadron Collider? originally aimed to rovide reference data for the high energy Pb Pb collisions, esecially on the cold nuclear matter effects. However, a large amount of unexected collective behaviors have been discovered by the ALICE, ALAS and CMS Collaborations. For instance, a symmetric double ridge structure on both near- and away-side has been observed in high multilicity Pb collisions by the ALICE Collaboration [38]. In addition, the CMS Collaboration has showed comatible results between multiarticle (including 4-, 6- and 8-articles) and allarticle correlations with Lee-Yang Zero s (LYZ) [39], which corresonds to {4} {6} {8} {LYZ} [4] (hese results have also been confirmed by the ALAS [41] and ALICE Collaborations [42]). Recently, the measurements of azimuthal correlations have been extended to identified hadrons [43, 44]. A mass-ordering feature, which says that the differential ellitic flow at low transverse momentum region monotonically increases with the decrease of hadron mass, has been observed among ions, kaons and rotons in high multilicity events [43]. Similarly, the CMS Collaboration found the mass-ordering between S and Λ(Λ), which showed the of S is larger than the one of Λ(Λ) at lower, followed by a crossing at 2 GeV [44]. Many of these exerimental measurements have been semi-quantitatively described by (3+1)-d hydrodynamic simulations from several grous [45 49], which suorts the exerimental claim that large collective flow has been develoed in small Pb systems. In Au Au or Pb Pb collisions at RHIC and the LHC, the collective flow mainly develos in the QGP hase since the QGP fireball has long enough lifetime to develo the momentum anisotroy until the saturation is almost reached [6, 5, 51]. Meanwhile, a certain amount of collective flow is further accumulated in the hadronic stage through the microscoic rescatarxiv: v1 [nucl-th] 24 Mar 215 You Zhou, 1, 2, Xiangrong Zhu, 1 Pengfei Li, 1 1, 3, 4, and Huichao Song 1 Deartment of Physics and State ey Laboratory of Nuclear Physics and echnology, Peking University, Beijing 871, China 2 Niels Bohr Institute, University of Coenhagen, Blegdamsvej 17, 2 Coenhagen, Denmark 3 Collaborative Innovation Center of Quantum Matter, Beijing 871, China 4 Center for High Energy Physics, Peking University, Beijing 871, China (Dated: March 27, 215) Using the Ultra-relativistic Quantum Molecular Dynamics (UrQMD) model, we investigate azimuthal correlations in Pb collisions at = 5.2 ev. Comarison with the exerimental data shows that UrQMD can not reroduce the multilicity deendence of 2- and 4-article cumulants, esecially the transition from ositive to negative values of {4} in high multilicity events, which has been taken as exerimental evidence of collectivity in Pb collisions. Meanwhile, UrQMD can not qualitatively describe the differential ellitic flow, ( ), of all charged hadrons at various multilicity classes. hese discreancies show that the simulated hadronic Pb systems can not generate enough collective flow as observed in exeriment, the associated hadron emissions are largely influenced by non-flow effects. However, the characteristic ( ) mass-ordering of ions, kaons and rotons is observed in UrQMD, which is the consequence of hadronic interactions and not necessarily associated with strong fluid-like exansions. PACS numbers: Ld, Gz, q, 24..Lx I. INRODUCION he relativistic heavy ion collisions at Relativistic Heavy Ion Collider (RHIC) and the Large Hadron Collider (LHC) have rovided strong evidences for the creation of the Quark Gluon Plasma (QGP) [1 4]. One of the crucial observables is the azimuthal anisotroy of the transverse momentum distribution for roduced hadrons [5]. As a signature of the collective flow, it rovides imortant information on the Equation of State (EoS) and the transort roerties of the QGP [6 11]. Usually, the anisotroy is characterised by the Fourier flow-coefficients [12]: v n = cos[n(ϕ Ψ n ), (1) where ϕ is the azimuthal angle of the emitted hadrons, Ψ n is the n th order articiant (symmetry) lane angle and denotes an average of all articles in all events. he second Fourier flow-coefficient is called ellitic flow, which is associated with the initial ellitic overla region of the two colliding nuclei. In the ast decades, most attention had been aid to the ellitic flow, which has been systematically measured and studied at the Suer Proton Synchrotron (SPS) [13], RHIC [14 17], and the LHC [18 2] (for a summary, lease also refer to [21, 22]). More recently, it was realised that the higher order flow-coefficients are equally imortant, which rovide information on the fluctuating initial rofiles of the created QGP [23 37]. he measurements of azimuthal correlations in snn = 5.2 ev Pb collisions at the LHC were You Zhou: you.zhou@cern.ch Huichao Song: huichaosong@ku.edu.cn

2 2 terings, which leads to the mass-ordering among various hadron secies [52 55]. Comared with Au Au or Pb Pb collisions, the smaller systems created in Pb collisions have much shorter lifetime. As a result, the momentum anisotroy is not likely to reach saturation even if the QGP has been created. he measured azimuthal correlations in Pb collisions might be largely influenced by the hadronic evolution. On the other hand, non-flow effects (e.g. from hadron resonance decays) are significantly enhanced for a smaller system with much lower article yields, which also contribute to 2-article correlations [21]. With an assumtion of early thermalization for the created Pb systems, hydrodynamics simulates the evolution of both QGP and hadronic hases, and associate the azimuthal correlations of all charge and identified hadrons with the collective exansion of the systems [45 49]. In this aer, we assume that the high energy Pb collisions do not reach the threshhold of the QGP formation, only ure hadronic systems are roduced. We utilize a hadron cascade model Ultra-relativistic Quantum Molecular Dynamics (UrQMD) [56 58] to simulate the evolution of the hadronic matter and then study the azimuthal correlations of final roduced hadrons. Our research focuses on two asects: (1) investigating whether ure hadronic interactions could generate the observed flow signatures in high multilicity events; (2) studying the mass-ordering of 2-article correlations in ure hadronic Pb systems. he aer is organized as follows. Section II briefly introduces the UrQMD model. Section III outlines the 2- and 4-article Q-Cumulant method. Section IV comares exerimental measurements with the UrQMD calculations on 2- and 4- article azimuthal correlations, including centrality deendence and transverse momentum deendence. Section V summarizes and concludes this work. II. UrQMD HADRON CASCADE MODEL UrQMD is a microscoic transort model to describe hadron hadron, hadron nucleus and nucleus nucleus collisions at relativistic energies, based on the Boltzmann equations for various hadron secies [56 58]. It has successfully described the soft hysics at the AGS and SPS energies, where the created systems are dominated by strongly interacting hadrons. In UrQMD, the initial hadron roductions are modeled via the excitation and fragmentation of strings. For higher collision energies above = GeV, the PYHIA mode [59] is imlemented to describe the hard rocesses and the related hadron roductions. he classical trajectories of the roduced hadrons are then simulated through solving a large set of Boltzmann equations with flavour deendent cross sections. In the later version, UrQMD contains 55 baryon and 32 meson secies with masses u to 2.25 GeV, sulemented by the corresonding antiarticles and Event class N ch ( η < 1.) -5% >72 5-% % % % % <17 ABLE I. Event class determination in UrQMD according to the number of all charged hadrons within η < 1.. Event class N ch (2.8 < η < 5.1) -2% >88 2-4% % % % 13 ABLE II. Event class determination in UrQMD according to the number of all charged hadrons within 2.8 < η < 5.1. isosin-rojected states [58]. he elementary cross sections in the collision terms are either fitted from the exerimental data or calculated via models e.g. a modified additive quark model (AQM). For two closely roagating hadrons, whether or not a collision haens is determined by a critical distance associated with the related cross section. When all elastic and inelastic collisions cease and all unstable hadrons have decayed into stable hadrons, the system is considered to reach kinetic freeze-out. UrQMD then oututs the momentum and osition information of the final roduced hadrons. In this aer, we imlement UrQMD version 3.4 to simulate the evolution of the assumed hadronic systems created in high energy Pb collisions. he simulations are executed in the equal seed system of two colliding nucleons with = 5.2 ev. Corresondingly, the outut information for final roduced hadrons are defined in the centre-of-mass frame. In order to comare with the exerimental data in the laboratory frame, we make a transformation between the centre-of-mass frame and the laboratory frame, which shifts the raidity by.465. Following the related exerimental aers [42, 43, 6], the UrQMD oututs are divided into several multilicity classes, determined by the number of all charged hadrons N ch within a seudoraidity range η < 1 or 2.8 < η < 5.1. he N ch values in these two centrality definitions are shown able I and II. he seudoraidity density of all charged hadrons as a function of seudoraidity in minimum bias Pb collisions is resented in Fig. 1. III. ANALYSIS MEHOD AND DEFINIIONS In this aer, the azimuthal correlations are calculated using 2- and 4-article Q-Cumulant method [62,

3 3 lab dn ch /dη Pb η lab = 5.2 ev ALICE UrQMD FIG. 1. Pseudoraidity density of all charged hadrons in minimum bias Pb collisions at = 5.2 ev, measured by ALICE [61] and calculated from UrQMD. 63], which were used in exeriment at RHIC [64] and the LHC [18, 33, 4, 65]. In this method, both 2- and multi-article azimuthal correlations are analytically exressed in terms of a Q-vector, which is defined as: Q n = M e inϕi, (2) i=1 where M is the multilicity of the Reference Flow Particles (RFPs) and ϕ is their azimuthal angle. he single-event average 2- and 4-article azimuthal correlations can be calculated via: 2 = Q n 2 M M(M 1), 4 = Q n 4 + Q 2n 2 2 Re[Q 2n Q nq n] M(M 1)(M 2)(M 3) 2 2(M 2) Q n 2 M(M 3), M(M 1)(M 2) (3) here stands for the average over all articles in a single event. he 2- and 4-article cumulants could be achieved as: c n = 2, c n {4} = , (4) here denotes the average over all articles over all events. In order to roceed with the calculation of the differential flow of the Particles Of Interests (POIs), the n and q n vectors for secific kinematic range and/or for secific hadron secies are needed: m n = e inφi, i=1 m q q n = e inφi, i=1 (5) where m is the total number of articles labeled as POIs, m q is the total number of articles tagged both as RFP and POI. he single-event average differential 2- and 4- article azimuthal cumulants are calculated as: 2 = nq n m q m M m q, 4 = [ n Q n Q nq n q 2n Q nq n n Q n Q 2n 2 M n Q n 2 m q Q n q n Q n Q n q n + q 2n Q 2n + 2 n Q n + 2 m q M 6 m q ]. /[(m M 3m q )(M 1)(M 2)(M 3)] (6) For detectors with uniform azimuthal accetance the differential 2- and 4-article cumulants are given by: d n = 2, d n {4} = (7) Finally the estimated differential flow ( ) from 2- and 4-article correlations are given by: v n ( ) = d n cn, v n {4}( ) = d n{4} ( c n {4}). 3/4 (8) Unfortunately, the v n obtained from the 2-article Q-Cumulant contains contributions from so-called non-flow effects, which are additional azimuthal correlations between the articles due to e.g. resonance decays, jet fragmentation, and Bose-Einstein correlations. hey can be suressed by aroriate kinematic cuts. For instance, one can introduce a seudoraidity ga between the articles in the 2-article Q- Cumulant method [65]. Accordingly, the whole event is divided into two sub-events, A and B, which are searated by a η ga. his modifies Eq. (3) to: 2 η = QA n Q B n, (9) M A M B where Q A n and Q B n are the flow vectors from sub-event A and B, M A and M B are the corresonding multilicities. he 2-article Q-Cumulant with a η ga is given by: c n {2, η } = 2 η. ()

4 4 ] -2 N/(dm dy) [GeV 2 1/N ev 1/(2πm )d ALICE UrQMD Multilicity Class -2% π Multilicity Class 2-4% m -m m -m m -m (c) -Pb = 5.2 ev < y <.5 CMS Multilicity Class 8-% FIG. 2. m sectra of ions, kaons and rotons in Pb collisions at = 5.2 ev measured by ALICE [6] and calculated from UrQMD. Here the multilicity class determination in UrQMD is based on able II. For the calculations of differential flow with a seudoraidity ga, there is no overla of POIs and RPs if we select RPs from one subevent and POIs from the other. his modifies Eqs. (6) to: 2 η = n,aq n,b m,a M B, (11) and we get the differential 2-article cumulant as: d n {2, η } = 2 η. (12) Finally, the differential flow from 2-article cumulant can be obtained by inserting the 2-article reference flow (with η ga) to the differential 2-article cumulant: v n {2, η }( ) = d n{2, η } cn {2, η }. (13) In this aer, the second and third Fourier flowcoefficients are evaluated using above equations, by setting the n = 2 and 3, resectively. IV. RESULS AND DISCUSSIONS his section mainly investigates the second and third Fourier flow-coefficients with cumulants in Pb collisions at = 5.2 ev. Before studying the 2-article and 4-article correlations, it is imortant to check the single hadron information. Fig. 1 lots the seudoraidity density of all charged hadrons dn ch /dη in minimum bias Pb collisions. In general, UrQMD roughly describes the forward-backward asymmetry of the dn ch /dη curve within η < 2. At midraidity, dn ch /dη from UrQMD are close to the ALICE measurements, but about 5% lower than the exerimental values. Figure 2 lots the m sectra of ions, kaons and rotons in high energy Pb collisions. It is generally believed that, in the absence of radial flow, m sectra as a function of m m (m stands for the mass of the hadron and m = 2 + m2 ) satisfies the m scaling, where the sloe of the sectra is indeendent of hadron secies [66]. Such m scaling has been observed in - collisions at = 2 GeV [66, 67]. In heavy ion collisions at the SPS energies and above, the m scaling is broken, which rovides evidence for the develoment of strong radial flow in the hot QCD systems [66 7]. In high energy Pb collisions at = 5.2 ev, the ALICE measurements in Fig. 2 show that m scaling is broken at -2% and 2-4% multilicity classes where the measured rotons sectra are flatter than kaons ones 1. his rovides evidence for the develoment of radial flow in high multilicity events. he UrQMD calculations in Fig. 2 also resent a weak broken of the m scaling, but show steeer sectra for ions, kaons and rotons when comared with the ALICE curves. his indicates that the assumed hadronic -Pb systems could not accumulate sufficient radial flow as observed in exeriment. With brief investigations of the single hadron data, we now focus on studying azimuthal correlations in high energy Pb collisions. Figure 3 resents the centrality deendence of the 2-article cumulant of the second Fourier flow-coefficient, calculated from UrQMD hadron cascade model (left) and measured by the ALICE collaboration (right). For various seudoraidity gas, from UrQMD exhibit decreasing trend from eriheral (low multilicity events) to central collisions (high multilicity events), which agrees with the exectation of the azimuthal correlations not associated with the symmetry lane, the 1 he ion sectra is largely influenced by resonance decays at lower m m, which break the ion s m scaling even for the case without radial flow

5 Pb =5.2 ev (UrQMD) {2, η >.4} {2, η >1.} -Pb =5.2 ev (ALICE) {2, η >.4} {2, η >1.} Multilicity Class (%) 1 Multilicity Class (%) FIG. 3. (Color online) of all charged hadrons in Pb collisions at = 5.2 ev, calculated from UrQMD (Left) and measured by ALICE (Right) [42]. he circle, square and diamond markers reresent various seudoraidity ga cuts without η ga, with ga η >.4 and η > 1., resectively. Here the multilicity class determination in UrQMD is based on able I. {4} {4} 2 c Multilicity Class (%) 1 -Pb 5.2 ev {4}(UrQMD) {4}(ALICE) Multilicity Class (%) FIG. 4. (Color online) {4} of charged articles in Pb collisions at = 5.2 ev, calculated from UrQMD and measured by ALICE [42]. Here the multilicity class determination in UrQMD is based on able I. so-called non-flow effects. As the seudoraidity ga increases, the magnitudes of become weaker for both ALICE and UrQMD, which illustrates that nonflow effects, usually few-article correlations from resonance decays and jets, are suressed by a large seudoraidity ga. When the seudoraidity ga η is larger than 1., from ALICE show much weaker centrality deendence, which is suggested as one of the hints for collective exansion in the created Pb systems. However, from UrQMD still resent strong centrality deendence for η > 1., showing a tyical non-flow behavior. Usually, the non-flow effects between 2-article correlations, denoted as δ n, behave as δ n 1/M where M is the multilicity. he decreasing trend of with the increase of multilicity indicates that UrQMD hadronic exansion could not generate enough flow in a small -Pb system, non-flow effects are still retty large even for the case with a large seudoraidity ga cut η > 1.. o better understand the hadronic systems simulated by UrQMD, we investigate the 4-article cumulant of the second Fourier flow-coefficient {4}, which is equal to {4} 4 and exected to be less sensitive to non-flow effects. Figure 4 lots the centrality deendence of {4} of all charged hadrons in Pb collisions at = 5.2 ev. Both the UrQMD and ALICE results show that {4} increase with the decrease of multilicity from semi-central to eriheral collisions. For the most central collisions (<%), {4} from ALICE exhibits a transition from ositive to negative values, indicating the creation of flowdominated systems in the high multilicity events. However, {4} from UrQMD kees ositive for all available multilicity classes, including the most central collisions. As a result, real values of {4} can not be extracted in UrQMD for all centrality bins. his comarison further illustrates the difference between the Pb systems created in exeriment and simulated by UrQMD. he hadron emissions from UrQMD are largely influenced by non-flow effects. Without the contributions from the initial stage and/or the QGP hase, the measured flow-like 4-article correlations in high multilicity events can not be reroduced by a microscoic transort model with only hadronic scatterings and decays. In Figure 5, we further study the 2-article azimuthal correlations for the third Fourier flowcoefficient. he UrQMD calculations and the ALICE measurements with various seudoraidity gas are resectively shown in the left and right anels of Fig. 5. Similar to in Fig. 3, also decreases with the increase of η. For the ALICE measurement, kees ositive for all seudoraid-

6 Pb =5.2 ev (UrQMD) {2, η >.4} {2, η >1.} -Pb =5.2 ev (ALICE) {2, η >.4} {2, η >1.}.2 1 Multilicity Class (%) 1 Multilicity Class (%) FIG. 5. (Color online) of all charged hadrons in Pb collisions at = 5.2 ev, calculated from UrQMD(Left) and measured by ALICE (Right) [42]. Here the multilicity class determination in UrQMD is based on able I. ity gas, which leads to real values of triangular flow v 3 (v 3 = ) as measured in [42]. Considered that non-flow effects are largely suressed by a large seudoraidity ga, the measured at η > 1. (and the associated triangular flow v 3 ) is ossibly mainly caused by collective exansion and reflects initial state fluctuations of the -Pb systems. In contrast, from UrQMD turns to negative for η >.4 and η > 1., which does not roduce a real value of v 3 2. he fact that UrQMD could not generate the exerimentally observed triangular flow, together with the results shown Figs. 3 5, strongly indicates that Pb systems from UrQMD contain large non-flow azimuthal correlations. Following Eq. (13), we calculate the second Fourier flow-coefficient as a function of transverse momentum, ( ), for the UrQMD simulations at multilicity class -2%, 2-4%, 4-6% and 6-%. Figure 6 shows that ( ) monotonically increases from high to low multilicity class, which agrees with the trend of shown in Fig. 3 ( is the square of integrated ). Meanwhile, ( ) from UrQMD increases with the increase of and show strong sensitivity to the seudoraidity ga. he observed large seudoraidity ga suression of ( ) indicates that non-flow effects are large in UrQMD, as already shown in Fig Figure 6 also shows that UrQMD can not correctly reroduce the shae of the exerimental ( ) curves, when imlemented the same seudoraidity ga η > 1.. It underredicts the data at lower and overestimates the data above 1 GeV. Comared with the integrated, the differential ellitic flow ( ) contains more information on the evolving system, which reflects the interlay between radial and ellitic flow. he m sectra in Fig. 2 has 2 We also find that {4} only shows ositive values, just as {4}, which does not roduce a real value of v 3 {4}. already shown that UrQMD can not roduce sufficient radial flow as observed in exeriment. he insufficient radial flow, together with the insufficient flow anisotroy accumulation (shown in Fig. 3-5) leads to the fact that UrQMD could not reroduced the ( ) curves measured by ALICE. Figure 7 investigates azimuthal correlations of identified hadrons in high energy Pb collisions. he right anels resent the ALICE measurements with two different multilicity classes [43, 44], which show a characteristic feature of ( ) mass-ordering among ions, kaons and rotons. In the ast research, hydrodynamic simulations from several grous have systematically studied the flow data, which reroduced the mass-ordering feature of the Pb systems [46, 49]. In the hydrodynamic language, the radial flow further accumulated in hadronic stage tends to ush heavier hadrons from lower to higher, leading to an enhanced slitting between ions and rotons [53, 55]. he observation of mass-ordering is thus generally believed as a strong evidence for the collective exansion of the -Pb systems created in snn = 5.2 ev collisions. However, the left anels of Fig. 7 shows that UrQMD also generate a mass-ordering for the 2-article correlations among ions, kaons and rotons 3. Such mass-ordering attern, caused by ure hadronic interactions, qualitatively agrees with the ones from the ALICE measurement [43] and from the hydrodynamic calculations [46, 49]. In UrQMD, the unknown cross sections are calculated by the additive quark 3 Due to limited statistics, we aly η >.2 in our calculations, rather than η >.8 as used in exeriment. In fact, {2; η >.8} from our current UrQMD simulations has large error bars, esecially for rotons. However, a tendency of mass ordering among ions, kaons and rotons is still observed.

7 Pb =5.2 ev (UrQMD) {2, η >.2} {2, η >.4} {2, η >1.}.1 Multilicity Class: -2% Multilicity Class: 2-4%.4 -Pb =5.2 ev (ALICE) (c) (d).3 {2, η >.8}.2.1 Multilicity Class: 4-6% (GeV/ c ) Multilicity Class: 6-% (GeV/ c ) FIG. 6. (Color online) ( ) of all change hadrons in Pb collisions at = 5.2 ev, calculated from UrQMD and measured by ALICE [43]. Here the multilicity class determination in UrQMD is based on able II. model (AQM) through counting the number of constituent quarks within two colliding hadrons. As a result, the main meson-baryon (M-B) cross sections from AQM are about 5% larger than the mesonmeson (M-M) cross sections, leading to the slitting between mesons and baryons after the evolution of hadronic matter. Comarison simulations in aendix A (Fig. 9) also show that, with the M-B and M-M interaction channels closed in UrQMD, the mass-ordering almost disaears. he combined results in Fig. 7 and 9 illustrate that the hadronic interactions could lead to a mass-ordering in 2-article correlations among ions, kaons and rotons, even for small Pb systems without enough flow generation. V. SUMMARY Using UrQMD hadron cascade model, we studied azimuthal correlations in Pb collisions at = 5.2 ev. Comarisons with the exerimental data showed that the -Pb systems created in exeriment are not the trivial hadronic systems described by UrQMD. Here, we summarize the main results as the following: (1) With large seudoraidity gas ( η > 1.), the measured 2-article cumulant of the second Fourier flow-coefficient from ALICE shows a weak centrality deendence from central to semieriheral collisions. In contrast, the UrQMD calculations still resent a strong centrality deendence for {2, η > 1.}. (2) In the most central collisions, {4} from ALICE exhibits a transition from ositive to negative values, which indicates the develoment of strong collective flow in high multilicity events. However, {4} from UrQMD kees ositive for all multilicity classes, which does not roduce {4} with a real value. (3) For large seudoraidity gas, from UrQMD turns to negative values, which can not roduce the triangular flow as observed in exeriments. (4) UrQMD can not fit the differential flow ( ) from ALICE at various multilicity classes. More secifically, the related exerimental data of azimuthal correlations have accumulated strong evidence for the develoment of strong collective flow in high multilicity events. With the assumtion that high energy -Pb collisions do not reach the threshold for the QGP formation and only roduce trivial hadronic systems, we did hadron transort simulations with UrQMD. We found that hadronic interactions alone could not generate sufficient collective flow as observed in exeriment. Non-flow effects, e.g. from resonance decays and/or jet-like fragmentations, largely influence the hadron emissions of the UrQMD systems. In order to fit the measured azimuthal correlations of all charged hadrons in -Pb collisions at snn = 5.2 ev, the contributions from the initial stage and/or the QGP hase can not be neglected. In addition, we extended our study of azimuthal correlations to identified hadrons. he calculations of the 2-article correlations for ions, kaons and rotons showed that UrQMD can generate a massordering with the characteristic feature similar to

8 8 {2, η >.2}.4 -Pb =5.2 ev (UrQMD) π.2 {2, η >.8}.2.1 -Pb π ± ± () =5.2 ev (ALICE) Multilicity Class: -2% Multilicity Class: -2% {2, η >.2}.4.2 (c) {2, η >.8}.2.1 (d) Multilicity Class: 4-6% (GeV/c ) Multilicity Class: 4-6% (GeV/c ) FIG. 7. (Color online) ( ) of ions, kaons and rotons in Pb collisions at = 5.2 ev, calculated from UrQMD(left anels) and measured by ALICE(right anels) [43]. Here the multilicity class determination in UrQMD is based on able II. ] -2 N/(dm dy) [GeV 2 1/N ev 1/(2πm )d UrQMD with M-M(B) 2 UrQMD without M-M(B) Multilicity Class -2% π Multilicity Class 2-4% m -m m -m m -m (c) -Pb = 5.2 ev < y <.5 CMS Multilicity Class 8-% FIG. 8. m sectra of of ions, kaons and rotons in Pb collisions at = 5.2 ev, calculated from UrQMD with and without M-M and M-B collisions. Here the multilicity class determination in UrQMD is based on able II. the ALICE measurements. Comarison runs from UrQMD with main hadronic scatterings turned on and off showed that the mass-ordering in UrQMD is mainly caused by hadronic interactions. he mass-ordering alone is not necessarily a flow signature associated with strong fluid-like exansions.

9 9 UrQMD with M-M(B) collisions UrQMD without M-M(B) collisions {2, η >.2} Pb π =5.2 ev {2, η >.2} Pb π =5.2 ev (c).1.1 Multilicity Class: -2% Multilicity Class: -2% {2, η >.2} {2, η >.2} (d).1 Multilicity Class: 4-6% (GeV/c).1 Multlicity Class: 4-6% (GeV/c) FIG. 9. ( ) of ions, kaons and rotons in Pb collisions at = 5.2 ev, calculated from UrQMD with and without M-M and M-B collisions. Here the multilicity class determination in UrQMD is based on able II. ACNOWLEDGMENS We thank S. A. Bass, A. Bilandžić, J. J. Gaardhøje, U. Heinz, J. Y. Ollitrault, H. Petersen, J. Schukraft and R. Snellings for valuable discussions. YZ thanks to the Danish Council for Indeendent Research, Natural Sciences and the Danish National Research Foundation (Danmarks Grundforskningsfond) for suort, thanks to Peking University for the host. XZ, PL and HS are suorted by the NSFC and the MOS under grant Nos and 215CB8569. We gratefully acknowledge extensive comuting resources rovided to us on ianhe-1a by the National Suercomuting Center in ianjin, China. Aendix A: UrQMD comarison runs with/without M M and M B collisions his aendix exlores how hadronic interactions in UrQMD influence sectra and azimuthal correlations of identified hadrons for the hadronic Pb systems. In UrQMD, hadronic scatterings include Meson-Meson (M-M) collisions, Meson-Baryon (M-B) collisions and Baryon-Baryon (B-B) collisions. When switching off all of these collision channels, UrQMD simulations, in rincile, consist of initial hadron roductions and the succeeding resonance decays, which are mainly influenced by non-flow effects. However, not all these collision channels in the current version of UrQMD (v3.4) can be simultaneously turned off. With B-B collision channels turned off, all the secondary rotonnucleon collisions from the initial Pb collisions are automatically turned off without roceeding any further hadron roductions and decays. Considering the robability of B-B collisions is much lower than M-M and M-B collisions, we only turn off the M-M and M- B interaction channels for the UrQMD comarison runs in this aendix. Figure 8 lots the m sectra of ions, kaons and rotons in Pb collisions at = 5.2 ev, calculated from UrQMD with and without M-M and M-B interactions. In Sec. IV, we have already showed that, although the m scaling is weakly broken in UrQMD, ure hadronic interactions can not generate sufficient radial flow as observed in exeriment. Here, Fig. 8 shows that M-M and M-B collisions only slightly change the sloe of the m sectra 4. he slight broken of the m scaling in UrQMD are very ossibly caused by mechanisms of the initial hadron roduc- 4 For 2.76 A ev Pb+Pb collisions at 7-8% centrality (where dn ch /dη at mid-raidity is about 5 [55] and close to dn ch /dη in high multilicity Pb collisions), hybrid model simulations also show that hadronic scatterings almost do not change the m sectra of identified hadrons [71].

10 tions 5. Comared with the m sectra, the massordering are more sensitive to hadronic interactions. For tyical flow-dominated systems created in highenergy Au Au or Pb Pb collisions, hybrid model simulations have shown that hadronic rescatterings dramatically increase the slitting between ions and rotons, but only slightly change the m sectra [54]. Figure 9 resents ( ) of identified hadrons in high energy Pb collisions, based on UrQMD simulations in the senarios of with (left anels) and without (right anels) M-M and M-B collisions. In the cases that M-M and M-B collisions are turned off, the massordering among ions, kaons and rotons almost disaears, when comared with the cases with M-M and M-B interactions. In Sec. IV, detailed study of the 2-article and 4- article correlations has already shown that UrQMD could not generate sufficient flow as observed in exeriments, its final hadron emissions are largely influenced by non-flow effects. he comarisons runs in Fig. 9 illustrate that the mass-ordering can be exlained as the consequence of hadronic interactions, which is not necessarily associated with strong fluidlike exansions. [1] I. Arsene et al. (BRAHMS Collaboration), Nucl. Phys. A757, 1 (25); B. B. Back et al. (PHO- BOS Collaboration), ibid.,. 28; J. Adams et al. (SAR Collaboration), ibid.,. 2;. Adcox et al. (PHENIX Collaboration), ibid., [2] M. Gyulassy, in Structure and dynamics of elementary matter, edited by W. Greiner et al., NAO science series II: Mathematics, hysics and chemistry, Vol. 166 (luwer Academic, Dordrecht, 24), M. Gyulassy and L. McLerran, Nucl. Phys. A75, 3 (25); E. V. Shuryak, ibid.,. 64. [3] B. Muller and J. L. Nagle, Ann. Rev. Nucl. Part. Sci. 56, 93 (26). [4] B. Muller, J. Schukraft and B. Wyslouch, Ann. Rev. Nucl. Part. Sci. 62, 361 (212). [5] J. -Y. Ollitrault, Phys. Rev. D 46, 229 (1992). [6] P. Huovinen, in Quark Gluon Plasma 3, edited by R. C. Hwa and X. N. Wang (World Scientific, Singaore, 24),. 6 [nucl-th/3564]; P. F. olb and U. Heinz, ibid., [7] D. A. eaney, arxiv: [nucl-th]. [8] P. Romatschke, Int. J. Mod. Phys. E 19, 1 (2). [9] C. Gale, S. Jeon and B. Schenke, Int. J. Mod. Phys. A 28, (213). [] U. Heinz and R. Snellings, Ann. Rev. Nucl. Part. Sci. 63, 123 (213). [11] H. Song, Nucl. Phys. A 94-95, 114c (213); arxiv: [12] S. Voloshin and Y. Zhang, Z. Phys. C 7, 665 (1996). [13] C. Alt et al. [NA49 Collaboration], Phys. Rev. C 68, 3493 (23). [14]. H. Ackermann et al. [SAR Collaboration], Phys. Rev. Lett. 86, 42 (21); C. Adler et al. [SAR Collaboration], Phys. Rev. Lett. 87, (21); C. Adler et al. [SAR Collaboration], Phys. Rev. C 66, 3494 (22). [15] J. Adams et al. [SAR Collaboration], Phys. Rev. C 72, 1494 (25); B. I. Abelev et al. [SAR Collaboration], Phys. Rev. C 77, 5491 (28); J. Adams 5 In the UrQMD simulations for -Pb collisions at = 5.2 ev, most of the initial hadron roductions are triggered by the PYHIA mode due to large momentum transfer [56 58, 72]. Backu simulations from PYHIA with collision energy set to 5.2 ev show that the m scaling are also weakly broken for the m sectra of ions, kaons and rotons [71]. et al. [SAR Collaboration], Phys. Rev. Lett. 95, (25). [16] L. Adamczyk et al. [SAR Collaboration], Phys. Rev. C 88, 1492 (213). [17]. Adcox et al. [PHENIX Collaboration], Phys. Rev. Lett. 89, (22); S. S. Adler et al. [PHENIX Collaboration], Phys. Rev. Lett. 91, (23). [18] Aamodt et al. [ALICE Collaboration], Phys. Rev. Lett. 5, (2). [19] B. B. Abelev et al. [ALICE Collaboration], arxiv: [nucl-ex]. [2] G. Aad et al. [ALAS Collaboration], Phys. Lett. B 77, 33 (212). [21] S. A. Voloshin, A. M. Poskanzer and R. Snellings, in Landolt-Boernstein, Relativistic Heavy Ion Physics, Vol. 1/23, 5-54 (Sringer-Verlag, 2). [22] R. Snellings, New J. Phys. 13, 558 (211). [23] B. Alver and G. Roland, Phys. Rev. C 81, 5495 (2) [Erratum-ibid. C 82, 3993 (2)]. [24] R. S. Bhalerao, M. Luzum and J. Y. Ollitrault, Phys. Rev. C 84, 349 (211); R. S. Bhalerao, M. Luzum and J. Y. Ollitrault, Phys. Rev. C 84, 5491 (211); F. G. Gardim, F. Grassi, M. Luzum and J. Y. Ollitrault, Phys. Rev. C 85, 2498 (212). [25] M. Luzum and H. Petersen, J. Phys. G 41, 632 (214). [26] H. Petersen, G. Y. Qin, S. A. Bass and B. Muller, Phys. Rev. C 82, 4191 (2); G. Y. Qin, H. Petersen, S. A. Bass and B. Muller, Phys. Rev. C 82, 6493 (2). [27] H. Holoainen, H. Niemi and. J. Eskola, Phys. Rev. C 83, 3491 (211). [28] L. Pang, Q. Wang and X. -N. Wang, Phys. Rev. C 86, (212). [29] D. eaney and L. Yan, Phys. Rev. C 86, 4498 (212). [3] B. Schenke, P. ribedy and R. Venugoalan, Phys. Rev. C 86, 3498 (212); C. Gale, S. Jeon, B. Schenke, P. ribedy and R. Venugoalan, Phys. Rev. Lett. 1, 1232 (213). [31] Z. Qiu and U. Heinz, Phys. Lett. B 717, 261 (212). [32] E. Retinskaya, M. Luzum and J. Y. Ollitrault, Phys. Rev. C 89, no. 1, 1492 (214). [33]. Aamodt et al. [ALICE Collaboration], Phys. Rev. Lett. 7, 3231 (211);. Aamodt et al. [ALICE Collaboration], Phys. Lett. B 78, 249 (212). [34] G. Aad et al. [ALAS Collaboration], Phys. Rev. C 86, 1497 (212).

11 11 [35] S. Chatrchyan et al. [CMS Collaboration], Eur. Phys. J. C 72, 212 (212). [36] L. Adamczyk et al. [SAR Collaboration], Phys. Rev. C 88, no. 1, 1494 (213). [37] A. Adare et al. [PHENIX Collaboration], Phys. Rev. Lett. 7, (211). [38] B. Abelev et al. [ALICE Collaboration], Phys. Lett. B 719, 29 (213). [39] R. S. Bhalerao, N. Borghini and J. Y. Ollitrault, Nucl. Phys. A 727, 373 (23). [4] Q. Wang [CMS Collaboration], Nucl. Phys. A 931, 997 (214). [41] G. Aad et al. [ALAS Collaboration], Phys. Lett. B 725, 6 (213). [42] B. B. Abelev et al. [ALICE Collaboration], Phys. Rev. C 9, no. 5, 5491 (214). [43] B. B. Abelev et al. [ALICE Collaboration], Phys. Lett. B 726, 164 (213). [44] V. hachatryan et al. [CMS Collaboration], arxiv: [nucl-ex]. [45] P. Bozek, Phys. Rev. C 85, (212), P. Bozek and W. Broniowski, Phys. Lett. B 718, 1557 (213); Phys. Rev. C 88, 1493 (213). [46] P. Bozek, W. Broniowski and G. orrieri, Phys. Rev. Lett. 111, (213). [47] A. Bzdak, B. Schenke, P. ribedy and R. Venugoalan, Phys. Rev. C 87, 6496 (213). [48] G. Y. Qin and B. Mulle, Phys. Rev. C 89, no. 4, 4492 (214). [49]. Werner, M. Bleicher, B. Guiot, I. arenko and. Pierog, Phys. Rev. Lett. 112, no. 23, (214). [5] P. F. olb, J. Sollfrank and U. W. Heinz, Phys. Rev. C 62, 5499 (2). [51] H. Song and U. W. Heinz, Phys. Rev. C 77, 6491 (28). [52] C. Nonaka and S. A. Bass, Phys. Rev. C 75, 1492 (27). [53]. Hirano, U. W. Heinz, D. harzeev, R. Lacey and Y. Nara, Phys. Rev. C 77, 4499 (28). [54] H. Song, S. A. Bass and U. Heinz, Phys. Rev. C 83, (211). [55] H. Song, S. Bass and U. W. Heinz, Phys. Rev. C 89, no. 3, (214); X. Zhu, F. Meng, H. Song and Y. X. Liu, Phys. Rev. C 91, no. 3, 3494 (215). [56] S. A. Bass, M. Belkacem, M. Bleicher, M. Brandstetter, L. Bravina, C. Ernst, L. Gerland and M. Hofmann et al., Prog. Part. Nucl. Phys. 41, 255 (1998). [57] M. Bleicher, E. Zabrodin, C. Sieles, S. A. Bass, C. Ernst, S. Soff, L. Bravina and M. Belkacem et al., J. Phys. G 25, 1859 (1999). [58] H. Petersen, M. Bleicher, S. A. Bass and H. Stocker, arxiv: [he-h]. [59]. Sjostrand, L. Lonnblad, S. Mrenna and P. Z. Skands, he-h/ [6] B. B. Abelev et al. [ALICE Collaboration], Phys. Lett. B 728, 25 (214). [61] B. Abelev et al. [ALICE Collaboration], Phys. Rev. Lett. 1, no. 3, 3231 (213). [62] A. Bilandžić, R. Snellings and S. Voloshin, Phys. Rev. C 83, (211). [63] A. Bilandzic, C. H. Christensen,. Gulbrandsen, A. Hansen and Y. Zhou, Phys. Rev. C 89, 6494 (214). [64] G. Agakishiev et al. [SAR Collaboration], Phys. Rev. C 86, 1494 (212). [65] Y. Zhou [ALICE Collaboration], arxiv: [nucl-ex]. [66] U. W. Heinz, he-h/4736. [67] O. Y. Barannikova et al. [SAR Collaboration], Nucl. Phys. A 715, 458 (23). [68] S. V. Afanasiev,. Anticic, B. Baatar, D. Barna, J. Bartke, R. A. Barton, M. Behler and L. Betev et al., Nucl. Phys. A 715, 161 (23). [69] J. Velkovska [PHENIX Collaboration], Nucl. Phys. A 698, 57 (22);. Adcox et al. [PHENIX Collaboration], Phys. Rev. Lett. 88, (22). [7] C. Adler et al. [SAR Collaboration], Phys. Rev. Lett. 87, (21). [71] X. R. Zhu and H. Song, unublished notes. [72] S. A. Bass, rivate communications.