betaFeedback

WorksPhysics and Astronomy

Transverse momentum fluctuations as a probe of thermalization and collective dynamics of QGP

Abstract

We study fluctuations of mean transverse momentum per particle ([pT][p_T]) in ultrarelativistic heavy-ion collisions. We show that the steep fall in the variance of transverse momentum fluctuation in ultracentral Pb+Pb collision serves as a natural consequence of the thermalization of the QGP medium. We study the correlation between the spectra and [pT][p_T] which maps these fluctuations differentially and dubbed as v0(pT)v_0(p_T). We highlight the importance of v0(pT)v_0(p_T) showing that it plays similar role as anisotropic flow when probing the collective nature of QGP.

Introduction

Ultrarelativistic collision of two nuclei creates a hot fluid-like medium known as the Quark Gluon Plasma (QGP), the dynamics of which can be described by relativistic viscous hydrodynamics and the evidence of which has traditionally relied on the anisotropic flow of the particles driven by the initial pressure gradient [1, 2]. However, the evidence is indirect, as it depends on particles’ direction (azimuths). We show how the event-by-event fluctuations of the mean transverse momentum per particle ([pT][p_T]) in ultracentral collisions, can serve as a direct probe of the QGP formation (thermalized medium), which solely depend on the magnitude of the particles’ momenta instead of their direction of flight. Moreover, we show how the correlation between particle spectra and [pT][p_T] can map this fluctuations (σpT\sigma_{p_T}) as a function of pTp_T through a new observable v0(pT)v_0(p_T).

Mean transverse momentum fluctuations in ultracentral Pb+Pb collision

The energy deposited in a heavy-ion collision is shared by the particles produced. This is characterized by the mean transverse momentum per particle, [pT]≡∑pT/Nch[p_T] \equiv\sum p_T/N_{ch}, where NchN_{ch} is the total number of charged particles produced. [pT][p_T] fluctuates event-by-event, leading to a net dynamical fluctuations at the end. ATLAS measures the variance of this dynamical fluctuations as a function of NchN_{ch}. The relative dynamical fluctuations are small, around 1 %. Interestingly, the data encounter a sharp decrease over a narrow range of NchN_{ch} in ultracentral region. We argue this sudden decrease as a natural consequence of thermalization of the QGP medium produced in the collision. In order to model this dependence, we perform event-by-event hydro simulations of

1000 Pb+Pb collisions at fixed bb (b=0b=0) as shown in Figure 1. The figure shows that at a constant volume (fixed bb) a larger number of particle production implies a larger [pT][p_T]. We assume that this distribution can be described by a 2D correlated Gaussian distribution P([pT],Nch∣b)P([p_T],N_{ch}|b) which characterized by five parameters: Nch(b)N_{ch}(b), σNch(b)\sigma_{N_{ch}}(b), pT‾(b)\overline{p_T}(b), σpT(b)\sigma_{p_T}(b) and r(b)r(b), where r(b)r(b) denotes the Pearson correlation coefficient between [pT][p_T] and NchN_{ch} at fixed bb [3]. The distribution of pTp_T (or δpT≡[pT]−pT‾\delta p_T \equiv[p_T]-\overline{p_T}) at fixed NchN_{ch} is obtained by taking the superposition of the distributions at fixed bb and dividing by P(Nch)P(N_{ch}): P(δpT∣Nch)=(∫P(δpT,Nch∣b)P(b) db)/P(Nch)P(\delta p_T|N_{ch})=\left(\int P(\delta p_T,N_{ch}|b)P(b)\,db\right)/P(N_{ch}), from which the variance can be obtained by taking the square width,

Figure 1. Left: Two dimensional scattered plot of [pT][p_T] and NchN_{ch} at b=0b=0. The blue symbols denote the hydro results and the straight line represents the mean at fixed NchN_{ch}. The elliptic curve correspond to the 99 % confidence ellipse of the correlated Gaussian distribution. The orange color shows the similar results from a non-thermal HIJING model calculation. Right: Variance of [pT][p_T] as a function of NchN_{ch} in 20 % most central Pb+Pb collisions. The blue points and the red lines show the ATLAS measurements and our model fit to the data respectively. The blue dashed curve and the green dashed-dotted curve represent the contribution of intrinsic and impact-parameter fluctuations respectively.

Var⁡(pT∣Nch)=⟨(δpT−⟨δpT⟩)2⟩=Var⁡(pT∣Nch,b)+(⟨δpT2⟩b−⟨δpT⟩b2|Nch).(1)\operatorname{Var}(p_T|N_{ch})=\left\langle(\delta p_T-\langle\delta p_T\rangle)^2\right\rangle=\operatorname{Var}(p_T|N_{ch},b)+\left(\left\langle\delta p_T^2\right\rangle_b-\left\langle\delta p_T\right\rangle_b^2\middle|N_{ch}\right). \tag*{(1)}

Eq. 1 shows that there are two separate contributions to the variance of [pT][p_T] at fixed NchN_{ch}: i) variance at fixed NchN_{ch} and fixed bb ii) variance of [pT][p_T] due to the fluctuation of bb at fixed NchN_{ch}. The first contribution can be understood from Figure 1 (left), where even at fixed bb (=0) and fixed NchN_{ch}, there are fluctuations of [pT][p_T] which can be attributed to intrinsic fluctuations. The second contribution comes from the fluctuations of impact parameter at fixed multiplicity (see Figure 1 of [4] and Figure 2 of [3]). At fixed NchN_{ch}, a larger bb corresponds to a smaller collision volume [4], hence larger density Nch/VN_{ch}/V. From relativistic thermodynamics, a larger density results in larger energy per particle eventually producing larger [pT][p_T]. The distribution of impact parameter at fixed NchN_{ch} is obtained using Bayes’ theorem: P(b∣Nch)=P(Nch∣b)P(b)P(Nch)P(b|N_{ch})=\frac{P(N_{ch}|b)P(b)}{P(N_{ch})}. In Eq. 1 ⟨⋯ ⟩\langle\cdots\rangle denotes the mean calculated w.r.t the distribution P(δpT∣Nch)P(\delta p_T|N_{ch}) and ⟨⋯ ⟩b\langle\cdots\rangle_b denotes the mean calculated w.r.t the distribution P(b∣Nch)P(b|N_{ch}).

Figure 2. Left: Normalized charged particle spectra (top) and v0(pT)/v0v_0(p_T)/v_0 (bottom) as a function of pTp_T at b=0b=0. Results obtained from hydrodynamic simulations are shown for shear and bulk viscosity with smooth and fluctuating initial conditions. Results for b=7b=7 fm is also shown as a comparison. Right: Variance of [pT][p_T] for different pTp_T-windows as a function of multiplicity. The circular symbols denote the ATLAS measurements and the solid curves represent the reproduced results in our model taking the data in the largest window as inputs and using the correction factor CAC_A derived from v0(pT)/v0v_0(p_T)/v_0

The variance Var⁡(pT∣Nch)\operatorname{Var}(p_T|N_{ch}) calculated from Eq. 1 is fitted to ATLAS data for 20% most central events [3], shown in Figure 1. In the same figure we also show the two separate contributions in Eq. 1. It could be seen that the contribution of bb-fluctuations gradually goes to zero in the ultracentral limit, causing the sharp decrease. The values of the fit parameters are shown on Figure 11. Our model returns a very large value of rr implying a very strong correlations between NchN_{ch} and [pT][p_T] (missing in non-thermal model e.g. HIJING cf. Figure 1) stemming as a direct consequence of the thermalization of the QGP medium.

Mean transverse momentum-spectra correlation : v0(pT)v_0(p_T)

Event-by-event fluctuations of mean transverse momentum per particle result from event-by-event fluctuations of pTp_T-spectra. A softer spectra results in a [pT][p_T] smaller than ⟨[pT]⟩\langle[p_T]\rangle, and a harder spectra gives rise to a larger [pT][p_T]. Therefore, it is intriguing to explore how these two quantities are correlated and what one can learn from that. Following the proposal by Teaney et al. [5], we construct this correlation, dubbed as v0(pT)v_0(p_T) [6],

v0(pT)≡⟨δN(pT)δpT⟩N0(pT)σpTandv0=σpT⟨pT⟩(2)v_0(p_T) \equiv\frac{\langle\delta N(p_T)\delta p_T\rangle}{N_0(p_T)\sigma_{p_T}} \quad\text{and}\quad v_0 = \frac{\sigma_{p_T}}{\langle p_T\rangle} \tag*{(2)}

where N(pT)N(p_T) denotes the normalized spectra, δN(pT)=N(pT)−⟨N(pT)⟩\delta N(p_T)=N(p_T)-\langle N(p_T)\rangle, δpT=[pT]−⟨[pT]⟩\delta p_T=[p_T]-\langle[p_T]\rangle, ⟨[pT]⟩≡⟨pT⟩\langle[p_T]\rangle\equiv\langle p_T\rangle and ⟨N(pT)⟩=N0(pT)\langle N(p_T)\rangle=N_0(p_T). While v0v_0 represents the total relative fluctuations of [pT][p_T],

v0(pT)v_0(p_T) provides a differential estimate of that total fluctuations. The left panel of Figure 2 present the results for v0(pT)v_0(p_T) along with the spectra obtained from hydrodynamic simulations at fixed impact parameter. Instead of v0(pT)v_0(p_T), the scaled observables v0(pT)/v0v_0(p_T)/v_0 is presented, which is independent of centrality (or bb), as seen from the figure. Two sets of hydro results are presented: one with fluctuating initial conditions similar to what is done in experiments, and the other one is with the smooth initial conditions where we perform only two hydro simulations. In case of hydro with smooth initial conditions, the observable can be calculated as, v0(pT)v0≡δln⁡N(pT)δln⁡[pT]\frac{v_0(p_T)}{v_0} \equiv\frac{\delta\ln N(p_T)}{\delta\ln[p_T]} discussed in details in [6]. We consider different scenarios with shear and bulk viscosity, and find that v0(pT)/v0v_0(p_T)/v_0 shows negligible sensitivity to η/s\eta/s while showing a little sensitivity to ζ/s\zeta/s at large pTp_T. This sensitivity to bulk viscosity is partly due to the sensitivity of mean ⟨pT⟩\langle p_T\rangle to bulk viscosity which is washed out when the horizontal axis is scaled by ⟨pT⟩\langle p_T\rangle i.e. plotting as a function of pT/⟨pT⟩p_T/\langle p_T\rangle [7].

The observable v0(pT)v_0(p_T) is as important as the anisotropic flow in the sense that it shows similar mass hierarchy as the elliptic flow v2(pT)v_2(p_T) when studied for identified particles [6], and can be measured in experiments through a similar method [8, 9]. Therefore v0(pT)v_0(p_T) could be termed as the true radial flow. Another important feature of this observable is that using v0(pT)/v0v_0(p_T)/v_0 one could estimate ⟨pT⟩\langle p_T\rangle and σpT\sigma_{p_T} in any pTp_T-acceptance range taking inputs in a particular acceptance, through an acceptance correction factor CAC_A [6, 10]. The right panel of Figure 2 shows ck≡σpTkc_k \equiv\sigma_{p_T}^k as a function of NN for three different pTp_T windows. Taking the red points (0.5<pT<50.5 < p_T < 5 GeV) as the inputs and using CAC_A we reproduce the data at the other two pTp_T windows: 0.5<pT<20.5 < p_T < 2 GeV and 1<pT<21 < p_T < 2 GeV.

Conclusions and Outlook

We show that the fluctuations of mean transverse momentum per particles can provide direct robust probes of the QGP medium and its collective behavior. The sharp decrease of the variance in ultracentral collisions is caused by the rapid decline in impact parameter fluctuations, providing direct evidence of the underlying thermalization of QGP manifested in hydrodynamics. The observable v0(pT)v_0(p_T) characterizes these fluctuations at differential level and its similarities with anisotropic flow establish it as a true measure of the radial flow which directly probes the collective dynamics of the QGP medium. Extension of these studies to small systems such as p+Pb and p+p in future, will push our understanding of the dynamics of QGP medium.

Acknowledgments

RS acknowledges the support of the Polish National Science Centre grant 2019/35/O/ST2/00357 and 2023/51/B/ST2/01625 for the participation in the conference. RS also acknowledges the support of the organizers of ATHIC 2025.

References

  1. [1]U. Heinz and R. Snellings, Ann. Rev. Nucl. Part. Sci. 63, 123 (2013), arXiv:1301.2826 [nucl-th] .DOI
  2. [2]W. Busza, K. Rajagopal, and W. van der Schee, Ann. Rev. Nucl. Part. Sci. 68, 339 (2018), arXiv:1802.04801 [hep-ph] .DOI
  3. [3]R. Samanta, S. Bhatta, J. Jia, M. Luzum, and J.-Y. Ollitrault, Phys. Rev. C 109, L051902 (2024), arXiv:2303.15323 [nucl-th] .DOI
  4. [4]R. Samanta, J. a. P. Picchetti, M. Luzum, and J.-Y. Ollitrault, Phys. Rev. C 108, 024908 (2023), arXiv:2306.09294 [nucl-th] .DOI
  5. [5]B. Schenke, C. Shen, and D. Teaney, Phys. Rev. C 102, 034905 (2020), arXiv:2004.00690 [nucl-th] .DOI
  6. [6]T. Parida, R. Samanta, and J.-Y. Ollitrault, Phys. Lett. B 857, 138985 (2024), arXiv:2407.17133 [nucl-th] .
  7. [7]T. Parida, QM25 talk at Frankfurt, https://indico.cern.ch/event/1334113/contributions/6289775/attachments/3047185/5384425/QM2025_Tribhuban_Parida.pdf.
  8. [8]S. Acharya et al. (ALICE), (2025), arXiv:2504.04796 [nucl-ex] .
  9. [9]G. Aad et al. (ATLAS), (2025), arXiv:2503.24125 [nucl-ex] .
  10. [10]S. Bhatta, A. Dimri, and J. Jia, (2025), arXiv:2504.20008 [nucl-th] .DOI

Paper details

vlnc:W_jcv6z5h4
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents