Table of Contents
Fetching ...

Quenching through the QCD chiral phase transition

Adrien Florio, Eduardo Grossi, Aleksas Mazeliauskas, Alexander Soloviev, Derek Teaney

TL;DR

This work investigates the out-of-equilibrium dynamics of Model G, the dynamic universality class associated with the QCD chiral transition in the chiral limit. Through 3D Langevin simulations of an $O(4)$ order parameter coupled to conserved charges, the authors reveal a regime where the chiral condensate grows and drives a long-lived enhancement of soft pions, with equal-time correlators $G_{\pi\pi}(t,k)$ peaking above their equilibrium values. A scaling analysis identifies three key timescales—relaxation, ballistic transport, and diffusion—and shows that the growth is governed by the non-linear, non-dissipative dynamics of an $SU(2)_L\times SU(2)_R$ superfluid, yielding a parametrically large, persistent enhancement roughly scaling as $\sim 1/(k\xi)$ for $k\xi\ll 1$. Lattice quenches corroborate the scaling predictions, while a mean-field treatment captures the qualitative growth but misses the correct $k$-dependence and the ballistic growth phase. The results imply a potentially observable soft-pion yield enhancement in heavy-ion collisions and provide a pathway to determine hydrodynamic parameters from lattice QCD within the chiral scaling window.

Abstract

We present a detailed numerical and analytical study of the out-of-equilibrium dynamics of Model G, the dynamical universality class relevant to the chiral phase transition. We perform numerical 3D stochastic (Langevin) simulations of the $O(4)$ critical point for large lattices in the chiral limit. We quench the system from the high-temperature unbroken phase to the broken phase and study the non-equilibrium dynamics of pion fields. Strikingly, the non-equilibrium evolution of the two-point functions exhibits a regime of growth, a parametrically large enhancement, and a subsequent slow relaxation to equilibrium. We analyze our numerical results using dynamic critical scaling and mean-field theory. The growth of the two point functions is determined by the non-linear dynamics of an ideal non-abelian superfluid, which is a limit of Model G that reflects the broken chiral symmetry. We also relate the non-equilibrium two-point functions to a long-lived parametric enhancement of soft pion yields relative to thermal equilibrium following a quench.

Quenching through the QCD chiral phase transition

TL;DR

This work investigates the out-of-equilibrium dynamics of Model G, the dynamic universality class associated with the QCD chiral transition in the chiral limit. Through 3D Langevin simulations of an order parameter coupled to conserved charges, the authors reveal a regime where the chiral condensate grows and drives a long-lived enhancement of soft pions, with equal-time correlators peaking above their equilibrium values. A scaling analysis identifies three key timescales—relaxation, ballistic transport, and diffusion—and shows that the growth is governed by the non-linear, non-dissipative dynamics of an superfluid, yielding a parametrically large, persistent enhancement roughly scaling as for . Lattice quenches corroborate the scaling predictions, while a mean-field treatment captures the qualitative growth but misses the correct -dependence and the ballistic growth phase. The results imply a potentially observable soft-pion yield enhancement in heavy-ion collisions and provide a pathway to determine hydrodynamic parameters from lattice QCD within the chiral scaling window.

Abstract

We present a detailed numerical and analytical study of the out-of-equilibrium dynamics of Model G, the dynamical universality class relevant to the chiral phase transition. We perform numerical 3D stochastic (Langevin) simulations of the critical point for large lattices in the chiral limit. We quench the system from the high-temperature unbroken phase to the broken phase and study the non-equilibrium dynamics of pion fields. Strikingly, the non-equilibrium evolution of the two-point functions exhibits a regime of growth, a parametrically large enhancement, and a subsequent slow relaxation to equilibrium. We analyze our numerical results using dynamic critical scaling and mean-field theory. The growth of the two point functions is determined by the non-linear dynamics of an ideal non-abelian superfluid, which is a limit of Model G that reflects the broken chiral symmetry. We also relate the non-equilibrium two-point functions to a long-lived parametric enhancement of soft pion yields relative to thermal equilibrium following a quench.

Paper Structure

This paper contains 25 sections, 122 equations, 11 figures, 2 tables.

Figures (11)

  • Figure 1: Cartoon of the quench protocol in terms of correlation length $\xi$. The quenches studied in \ref{['sec:quenches']} transition from the high temperature phase to the low temperature phase, ${\sf t}_{\sf r}^0 \rightarrow -{\sf t}_{\sf r}^0$ with $\xi({\sf t}_{\sf r}^0)=\xi(-{\sf t}_{\sf r}^0)$ up to subleading corrections, see \ref{['tab:subleading']}. The colors from top to bottom correspond to ${\sf t}_{\sf r}^0 =0.009, 0.015, 0.023, 0.04$ and are correlated with the lines in the upper panels of \ref{['fig:VeVquench', 'fig:quench', 'fig:sgquenchk']}. In the the lower panels of \ref{['fig:VeVquench', 'fig:quench', 'fig:sgquenchk']} and \ref{['fig:quench_fixed_k']} we vary the lattice size while ${\sf t}_{\sf r}^0 = 0.04$.
  • Figure 2: (a) Schematic of a quench from the high-temperature phase to the broken phase in a spatial region of size $\ell$, which is several correlation lengths long (e.g., $\ell \sim 3\xi$). The red dashed line shows the initial potential at temperature ${\sf t}_{\sf r}^0$, and the blue line shows the post-quench potential. The ball represents a local order parameter $\bar{\phi}_a(t)$, averaged over the region. The orange Gaussian depicts the initial thermal distribution for $\bar{\phi}_a$ before the quench, with variance $\chi/\ell^d$. Over a time of order $t\sim \tau_R$, $\bar{\phi}_a$ evolves to the well’s bottom in a random direction, forming a local equilibrated domain of size $\ell$. (b) Field configuration for $t \gg \tau_R$ in a region of size $L \gg \ell \sim \xi$, showing randomly oriented domains of the chiral condensate. The domains merge over a time of order $t \sim L/v$ with dynamics governed by the ideal superfluid equations given in \ref{['eq:idealsuper']}.
  • Figure 3: A sketch of different stages in the typical evolution of the (pion) two-point function: early evolution setting the initial conditions $t\sim \tau_R$, rapid growth until the maximum, and slow decay to its equilibrium value in an oscillatory manner. See \ref{['sec:scaling']} for detailed discussion.
  • Figure 4: Time evolution of magnetization for symmetric quenches in \ref{['fig:sim_recap']}, with (top) different starting ${\sf t}_{\sf r}^0$, but same lattice size $L=192$ and (bottom) with ${\sf t}_{\sf r}^0=0.04$ and different lattice sizes. All curves are normalized by the initial equilibrium value, \ref{['eq:chi']}. Colors indicate the system size in units of correlation length $L/\xi$. The time is given in units of relaxation time $t/\tau_R\propto t/\xi^{d/2}$ and pion ballistic propagation time $L/v$. Dashed lines show the expected equilibrium value with the leading order finite-size correction \ref{['eq:Hasenfratz']}.
  • Figure 5: (top left) Enhancement of equal time pion-pion field correlator over thermal equilibrium expectation during symmetric quench for different starting ${\sf t}_{\sf r}^0$ and wavenumbers $k=2\pi/L$, where $L=192$. The time is given in units of relaxation time $t/\tau_R\propto t/\xi^{d/2}$. (top right) The same plot multiplied with $k\xi$ on $y$-axis and time given in units of pion ballistic propagation time $L/v\sim \tau_R (L/\xi)$, where $v$ is the pion velocity at corresponding ${\sf t}_{\sf r} ={\sf t}_{\sf r}^0$. (bottom left and right) Enhancement of pion-pion field correlator over thermal equilibrium expectation during symmetric quench from ${\sf t}_{\sf r} =0.04$ to ${\sf t}_{\sf r} =-0.04$ for different wavenumbers $k= 2\pi/L$, where $L=192,128,96,64$.
  • ...and 6 more figures