Finite-density equation of state of hot QCD using the complex Langevin equation

Michael Mandl, Dénes Sexty, Daniel Unterhuber · 2026

Read Finite-density equation of state of hot QCD using the complex Langevin equation by Michael Mandl, Dénes Sexty, Daniel Unterhuber (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-19649.txt.

Introduction

Mandl, Sexty, and Unterhuber present lattice simulations at the continuum limit of quantum chromodynamics above the crossover temperature and at newly high baryon levels at physical masses, using the complex Langevin method. They pin down the QCD equation of state from baryon density and pressure as functions of baryon chemical potential and temperature. Wrong turns of complex dynamics are under control. Results agree with previous lattice studies at smaller values and with hard-thermal-loop results at high heat.

Catalog record

CREATORMichael Mandl, Dénes Sexty, Daniel Unterhuber
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24806
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.19649

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Finite-density equation of state of hot QCD using the complex Langevin equation Michael Mandl, Dénes Sexty, Daniel Unterhuber Abstract We present the results of continuum-extrapolated lattice simulations of quantum chromodynamics (QCD) above the crossover temperature and for unprecedentedly high baryon densities at the physical point, employing the complex Langevin equation. In particular, we determine the QCD equation of state by computing the baryon density as well as the pressure as functions of the baryon chemical potential and the temperature. Potential issues with wrong convergence of complex Langevin dynamics are under control and we indeed find agreement with previous lattice studies working at smaller chemical potentials, as well as with perturbative hard-thermal-loop calculations at high temperatures. I Introduction Quantum chromodynamics (QCD) is the theory describing the strong interaction between quarks and gluons. Besides governing the confinement of these fundamental constituents into bound states like protons and neutrons, QCD encodes the physics of various other phenomena of contemporary interest. For instance, the strong interaction plays a major role in collision experiments of heavy ions, e.g. at RHIC in Brookhaven or the LHC at CERN, in compact stellar objects such as neutron stars, and in the early stages of the universe like the nucleosynthesis era. For the theoretical description of these systems, a proper understanding of QCD under extreme conditions such as high temperature and baryon density is of fundamental importance Borsányi and Parotto 2025 . By now, the temperature axis of the QCD phase diagram is well understood owing to ab initio lattice simulations, which have predicted, for instance, that the high-temperature transition of QCD is really an analytic crossover Aoki et al. 2006 . For non-zero baryon chemical potentials, on the other hand, lattice simulations face the infamous sign problem Troyer and Wiese 2005 , severely restricting the range of baryon densities one can reliably access. The most prominent approaches based on reweighting, Taylor expansion or imaginary chemical potential are currently limited to baryon chemical potentials about two to four times the temperature Allton et al. 2005 ; Borsányi et al. 2012a ; Bazavov et al. 2017 ; D’Elia et al. 2017 ; A. Bazavov et al. (2019) HotQCD Collaboration ; Giordano et al. 2020 ; Borsányi et al. 2021 ; Borsányi et al. 2022a ; Borsányi et al. 2022b ; D. Bollweg et al. (2022) HotQCD Collaboration ; Borsányi et al. 2023 ; D. Bollweg et al. (2023) HotQCD Collaboration ; Borsányi et al. 2024 ; Abuali et al. 2025 . The presence of the sign problem thus calls for alternative approaches not based on importance sampling. While various ideas for such alternatives exist, none of them has been able to fully circumvent the sign problem of QCD for arbitrary baryon densities. One particularly appealing solution to the sign problem is the complex Langevin approach Parisi 1983 ; Klauder 1983 , which is based on the complexification of the underlying field degrees of freedom and their stochastic evolution in an artificial time dimension – a generalization of stochastic quantization Parisi and Wu 1981 to complex variables; see, e.g., Berger et al. 2021 for a review. Since the introduction of the gauge cooling technique for stabilizing simulations Seiler et al. 2013 ; Nagata et al. 2016a , its application to finite-density QCD has shown promising results Sexty 2014a ; Sexty 2014b ; Aarts et al. 2014 ; Fodor et al. 2015 ; Aarts et al. 2017 ; Nagata et al. 2018 ; Sexty 2019 ; Kogut and Sinclair 2019 ; Ito et al. 2020 ; Scherzer et al. 2020a ; Attanasio et al. 2020 ; Attanasio et al. 2022 ; Hansen and Sexty 2025 ; Tsutsui et al. 2025 and even produced a full phase diagram in the heavy-dense limit Aarts et al. 2016 . Moreover, important progress has been made in the theoretical understanding of complex Langevin dynamics and its limitations – in particular, the notorious wrong-convergence problem –, resulting in several diagnostic tools that can be used to assess the correctness of simulation results Aarts et al. 2010a ; Aarts et al. 2011 ; Nagata et al. 2016b ; Scherzer et al. 2019 ; Scherzer et al. 2020b ; Seiler et al. 2024 ; Mandl et al. 2025a ; Joseph and Kumar 2025 ; Mandl 2026 , further establishing the reliability of the method. Up to this date, however, complex Langevin simulations of QCD were restricted to unphysical pion masses. In this work, we present the first complex Langevin results for QCD at the physical point, including a continuum extrapolation. We reach unprecedentedly high baryon chemical potentials \mu_{B} as compared to approaches relying on importance sampling, but are restricted to temperatures T above the crossover. Here, we discuss our main results for the QCD equation of state (see, e.g., Philipsen 2013 ) in the (\mu_{B},T) plane including the baryon density as well as the pressure, while a more detailed analysis, covering also additional thermodynamic quantities, will be published elsewhere Mandl et al. . II Lattice setup We have performed large-scale lattice simulations using complex Langevin dynamics. In the context of QCD, this entails the complexification of the usual \mathrm{SU}(3) gauge links to non-compact \mathrm{SL}(3,\mathbb{C}) -valued ones. The stochastic evolution of the complexified gauge links then proceeds via the discrete update equation \displaystyle\begin{aligned} \tilde{U}_{x\mu}^{(n)}=\exp\left[\mathrm{i}\left(\varepsilon K_{x\mu}[U^{(n)}]+\sqrt{\alpha\varepsilon}\eta_{x\mu}^{(n)}\right)\right]U_{x\mu}^{(n)}\;,\\ U_{x\mu}^{(n+1)}=\exp\left[\mathrm{i}\left(\varepsilon\gamma\tilde{K}_{x\mu}^{(n)}+\sqrt{\alpha\varepsilon}\eta_{x\mu}^{(n)}\right)\right]U_{x\mu}^{(n)}\;,\end{aligned} (1) with \alpha=2-3\varepsilon , \gamma=\frac{1}{2}+\frac{\varepsilon}{4} , K_{x\mu}[U]=\sum_{a}\lambda_{a}K_{x\mu a}[U]\;,\quad\eta_{x\mu}^{(n)}=\sum_{a}\lambda_{a}\eta_{x\mu a}^{(n)}\;, (2) and \tilde{K}_{x\mu}^{(n)}=K_{x\mu}[U^{(n)}]+K_{x\mu}[\tilde{U}^{(n)}]\;. (3) In ( 1 ) and ( 2 ), U_{x\mu}^{(n)} denotes the link connecting the lattice point x with its nearest neighbor in the \mu -direction after the n -th update, \lambda_{a} are the Gell-Mann matrices and \varepsilon denotes the discrete Langevin-time step-size. Moreover, K_{x\mu a} , the so-called drift term, when employing staggered fermions, reads \displaystyle\begin{aligned} K_{x\mu a}[U]&=\\ -D_{x\mu a}&S_{g}[U]+\frac{1}{4}\sum_{f=u,d,s,c}\mathrm{tr}\left[M_{f}^{-1}[U]D_{x\mu a}M_{f}[U]\right]\;,\end{aligned} (4) where S_{g} denotes the gauge action, M_{f} is the Dirac matrix for the fermion flavor f , and we consider 2+1+1 flavors. The trace in ( 4 ), involving the left derivative D_{x\mu a}f[U]=\partial_{\alpha}f\left[e^{\mathrm{i}\alpha\lambda_{a}}U_{x\mu}\right]_{\alpha=0}\;, (5) is evaluated stochastically Sexty 2014a . Finally, the \eta_{x\mu a}^{(n)} in ( 1 ) denote independent real Gaussian noise variables with zero mean and unit variance. We note that the prescription ( 1 ), which was introduced in Ukawa and Fukugita 1985 , is partially \mathcal{O}(\varepsilon) -improved. This, together with our choice of keeping \varepsilon\leq 10^{-4} in the employed adaptive-step-size algorithm Aarts et al. 2010b , ensures that finite-Langevin-step-size effects are smaller than our statistical errors, as comfirmed by comparisons to hybrid Monte-Carlo (HMC) simulations at zero chemical potential. We use a tree-level improved Symanzik gauge action and four steps of stout-smearing Morningstar and Peardon 2004 (generalized to \mathrm{SL}(3,\mathbb{C}) in Sexty 2019 ) with a smearing parameter \rho=0.125 . Denoting the numbers of lattice points in the spatial and temporal directions as N_{s} and N_{t} , respectively, our lattices have an aspect ratio of N_{s}/N_{t}=2 and we consider N_{t}=8 , 10 , 12 , and 16 . We change the temperature via the inverse gauge coupling \beta and work at physical pion masses. Indeed, our setup follows the one used in Bellwied et al. 2015 , from which we also take the line of constant physics (LCP) along which we perform the continuum extrapolation 1/N_{t}^{2}\to 0 (omitting N_{t}=8 ). As a consistency check, we

Read the complete text