Chiral symmetry breaking in accelerating and rotating frames Zhi-Bin Zhu, Hao-Lei Chen, Xu-Guang Huang Abstract We study chiral symmetry breaking and restoration in accelerating and rotating frames using low-energy effective models. By analyzing the chiral condensate in Rindler coordinates, we show that different renormalization schemes lead to distinct conclusions in accelerating frame: the scheme with subtracting divergences in Rindler vacuum supports an acceleration-independent critical temperatures, while the other scheme with subtracting divergences in Minkowski vacuum suggests enhanced critical temperature. We further investigate system with both rotation and acceleration. We find that the critical acceleration (see definition in Section V) for chiral symmetry restoration decreases with angular velocity, indicating cooperative effects from acceleration-induced thermalization and rotation-induced effective chemical potential. I Introduction Quantum Chromodynamics (QCD) is the fundamental theory of the strong interaction governing the dynamics of quarks and gluons. Extensive research on the QCD phase diagram aims to clarify the nature of strongly interacting quark-gluon matter under extreme conditions, including high temperature, high density, and strong external fields. In the low-temperature and low-baryon-density regime, quarks remain confined in hadrons (baryons and mesons), which corresponds to ordinary nuclear matter. Above a pseudo-critical temperature of approximately 150 MeV (at zero baryon density), quarks and gluons become deconfined and the (approximate) chiral symmetry is restored, resulting in the formation of a quark-gluon plasma (QGP) in which quarks and gluons can move over extended distances. The QCD phase diagram is conventionally presented in the temperature ( T )-baryon chemical potential ( \mu_{B} ) plane Fukushima and Hatsuda 2011 . Theoretical studies indicate a crossover transition at high temperatures with low baryon density, while a first-order phase transition emerges at high baryon densities with low temperatures. These regimes are connected by a critical endpoint (CEP) terminating the first-order phase boundary Guenther 2021 . The identification of this CEP in the T - \mu_{B} plane remains a primary objective for both theoretical investigations and heavy-ion collision experiments Bzdak et al. 2020 . Significant progress has been made in understanding the QCD phase structure through various theoretical approaches, including effective models Buballa 2005 ; Fukushima 2008 ; Buballa and Carignano 2015 ; Mannarelli 2019 , lattice QCD simulations Guenther 2021 ; Ratti 2018 ; Ding et al. 2015 ; Aarts et al. 2023 , functional renormalization group Pawlowski 2007 ; Drews and Weise 2017 ; Dupuis et al. 2021 ; Fu et al. 2020 ; Fu 2022 , Schwinger–Dyson equation Fischer 2019 ; Gao and Liu 2016 ; Gutiérrez et al. 2014 , and holographic approaches Rougemont et al. 2024 ; Zhao et al. 2024 ; Liu et al. 2024 . Recent theoretical developments have extended these studies to include additional background fields such as external electromagnetic fields Andersen et al. 2016 ; Miransky and Shovkovy 2015 ; Cao and Huang 2016 and global rotation Chen et al. 2021a . In recent years, the effects of rotation on QCD matter have attracted increasing attention, motivated by the presence of rapid rotation (or fluid vorticity) in non-central heavy-ion collisions Huang 2021 ; Huang et al. 2021 and neutron stars Lattimer and Prakash 2007 ; Paschalidis and Stergioulas 2017 . Following the early works Chen et al. 2016 ; Jiang and Liao 2016 , extensive efforts have been made to investigate the QCD phase structure under rotation Chernodub and Gongyo 2017a ; Chernodub and Gongyo 2017b ; Liu and Zahed 2018 ; Huang et al. 2018 ; Zhang et al. 2020 ; Wang et al. 2019a ; Wang et al. 2019b ; Chen et al. 2024a ; Chen et al. 2021b ; Chernodub 2021 ; Fujimoto et al. 2021 ; Jiang 2022 ; Sadooghi et al. 2021 ; Eto et al. 2022 ; Chen et al. 2022a ; Zhao et al. 2023 ; Chen et al. 2022b ; Chen et al. 2023 ; Sun et al. 2023 ; Mameda and Takizawa 2023 ; Cao 2024 ; Eto et al. 2024a ; Eto et al. 2024b ; Sun et al. 2024 ; Chen et al. 2024b ; Chen et al. 2025 ; Wang and Feng 2024 ; Singha et al. 2024 ; Morales-Tejera et al. 2025 ; Kiefer et al. 2025 ; Nunes et al. 2025 ; Ambrus 2020 ; Farias and Tavares 2025 . Effective model studies suggest that rotation tends to restore chiral symmetry, whereas more recent lattice QCD simulations indicate that it enhances chiral symmetry breaking and confinement Braguta et al. 2021 ; Chernodub et al. 2023 ; Braguta et al. 2024a ; Yang and Huang 2023 ; Braguta et al. 2024b . This apparent contradiction highlights the need for rigorous non-perturbative approaches to clarify the underlying mechanisms Chen et al. 2023 ; Jiang 2024 ; Wang et al. 2025 ; Fukushima and Shimada 2025 . Compared to rotational effects, acceleration-related phenomena have received relatively less attention in QCD studies. According to the Unruh effect Hawking 1975 ; Unruh 1976 , an observer undergoing constant proper acceleration a perceives the Minkowski vacuum as a thermal bath with temperature Crispino et al. 2008 T_{U}={a\over 2\pi}. In Ref. Kharzeev and Tuchin 2005 , the authors argued that partons in heavy-ion collisions experience characteristic deceleration on the order of the saturation scale, a\sim Q_{s}\sim 1\text{GeV} , leading to an Unruh temperature T_{U}\sim 200\,\text{MeV} (see Ref. Prokhorov et al. 2025 for a numerical simulation of Unruh temperature in heavy-ion collisions). Since this exceeds the QCD pseudo-critical temperature, it suggests that acceleration-induced thermalization could have nontrivial consequences for the thermodynamics of the produced matter. The general framework for quantum field theory in accelerating (Rindler) frame was established by Candelas et al. Candelas and Raine 1976 ; Candelas and Deutsch 1977 ; Candelas and Deutsch 1978 , and Lee and collaborators Lee 1986 later demonstrating the universality of the Unruh effect for interacting quantum fields of arbitrary spin. These results provide a solid theoretical basis for studying acceleration-induced phenomena—such as chiral symmetry breaking and restoration—using effective models that incorporate thermodynamic behavior in non-inertial frames Ohsaku 2004 ; Ebert and Zhukovsky 2007 ; Castorina and Finocchiaro 2012 ; Benic and Fukushima 2015 ; Casado-Turrión and Dobado 2019 ; Basu et al. 2023 ; Kou and Chen 2024 ; Chernodub 2025 . In addition, recent developments have employed Wigner function techniques Becattini et al. 2021 ; Palermo et al. 2021 , density operator formulations Prokhorov et al. 2019a ; Prokhorov et al. 2019b ; Prokhorov et al. 2019c ; Ambrus and Chernodub 2024 , and lattice simulation Chernodub et al. 2025 to investigate the thermodynamic properties of quantum systems in accelerating frame. These studies offer complementary insights into the behavior of QCD matter under acceleration. Given that QGP created in heavy-ion collisions behaves as an ultra-vortical fluid with strong temperature gradients, in this paper we study the QCD matter subject to combined accelerating and rotational effects, with particular emphasis on acceleration-driven phenomena. We employ low-energy effective models in non-inertial frames to study the chiral condensate \langle\bar{\psi}\psi\rangle (the order parameter for chiral symmetry breaking) as a function of acceleration a and angular velocity \omega . Similar to rotational effects, the influence of acceleration on chiral symmetry is still under debate. Existing studies report conflicting conclusions: while some suggest that acceleration may restore chiral symmetry Ohsaku 2004 ; Ebert and Zhukovsky 2007 ; Casado-Turrión and Dobado 2019 , others argue it may suppress it or have little effect Unruh and Weiss 1984 ; Salluce et al. 2024 ; Chernodub 2025 . This discrepancy appears to stem primarily from the choice of renormalization schemes, as will be discussed in later sections. The paper is organized as follows. Section II discuss the general description of accelerating and rotating thermal equilibrium. Section III establishes the Rindler coordinate framework for analyzing accelerating systems. Section IV and Sec. V respectively investigate chiral symmetry breaking under pure acceleration and combined acceleration-rotation scenarios. Finally, Sec. VI concludes with a summary of the main results. We use natural units \hbar=c=k_{B}=1 and Minkowski metric \eta^{\mu\nu}=\eta_{\mu\nu}={\rm{diag}}(1,-1,-1,-1) through out this paper. II Thermal equilibrium under acceleration and rotation Before we going into detailed discussion about how the acceleration and rotation affect the chiral symmetry breaking, it is important to notify that a many-body system could maintain thermal equilibrium under acceleration and rotation. We briefly discuss such a thermal equilibrium in this section; see Refs. Becattini 2012 ; Becattini 2018 ; Buzzegoli 2021 for more discussions. The so-called local equilibrium density operator \rho_{\rm LE} is the density operator that maximizes the entropy functional S[\rho]=-{\rm Tr}\left(\rho\ln\rho\right), (1) under the constraints that the local energy-momentum and angular momentum fluxes across a given space-like hypersurface \Xi_{\mu} in the Minkowski spacetime is fixed to their true physical values, \displaystyle n_{\mu}(x){\rm Tr}\left[\rho T^{\mu\nu}(x)\right]=n_{\mu}(x)T_{\rm true}^{\mu\nu}(x), (2) \displaystyle n_{\mu}(x){\rm Tr}\left[\rho J^{\mu\rho\sigma}(x)\right]=n_{\mu}(x)J_{\rm true}^{\mu\rho\sigma}(x), (3) where n_{\mu} is the normal direction of \Xi_{\mu} , T^{\mu\nu} is the energy-momentum tensor, J^{\mu\rho\sigma}=T^{\mu\sigma}x^{\rho}-T^{\mu\rho}x^{\sigma}+\Sigma^{\mu\rho\sigma} is the angular momentum tensor with \Sigma^{\mu\rho\sigma} the spin tensor. The resultant \rho_{\rm LE} has the following form: \rho_{\rm LE}=\frac{1}{Z_{\rm LE}}\exp\left[-\int d\Xi_{\mu}\left(T^{\mu\nu}\beta_{\nu}-\frac{1}{2}J^{\mu\rho\sigma}\varpi_{\rho\sigma}\right)\right], (4) where \beta_{\nu} and \varpi_{\rho\sigma}(=-\varpi_{\sigma\rho}) are two local Lagrange multipliers which have the physical meaning of thermal flow and “chemical” potential for angular momentum and Z_{\rm Le} is the partition function. When the system reaches global thermal equilibrium, the density operator must be time independent, or more precisely, independent of the choice of the space-like hypersurface \Xi_{\mu} . This is achieved when \beta^{\mu} and \varpi^{\mu\nu} are constants as can be easily checked by noticing that \partial_{\mu}\left(T^{\mu\nu}\beta_{\nu}-J^{\mu\rho\sigma}\varpi_{\rho\sigma}/2\right)=0 owing to Gauss’ theorem and the conservation laws for energy-momentum and angular momentum. Therefore, we have \rho_{\rm eq}=\frac{1}{Z_{\rm eq}}\exp\left\{-\int d\Xi_{\mu}\left[T^{\mu\nu}\left(\beta_{\nu}+\varpi_{\nu\lambda}x^{\lambda}\right)-\frac{1}{2}\varpi_{\rho\sigma}\Sigma^{\mu\rho\sigma}\right]\right\}. (5) Let us choose n_{\mu} be along the direction of \beta_{\mu} (which is time-like) and without loss of generality, we set \beta_{\mu} to the coordinate time direction, \beta^{\mu}=(\beta,\bm{0}) , so that \displaystyle\rho_{\rm eq} \displaystyle= \displaystyle\frac{1}{Z_{\rm eq}}\exp\left[-\beta H+\varpi_{0i}K^{i}+\frac{1}{2}\varpi_{ij}J^{ij}\right], (6) where H=\int d^{3}{\mathbf{x}}T^{00} is the Hamiltonian, K^{i}=\int d^{3}{\mathbf{x}}\left(T^{0i}t-T^{00}x^{i}+\Sigma^{00i}\right) is the boost operator, and J^{ij}=\int d^{3}{\mathbf{x}}\left(T^{0j}x^{i}-T^{0i}x^{j}+\Sigma^{0ij}\right) is the angular momentum operator. If we write \varpi_{0i}=\beta a^{i} and \varpi_{ij}=\beta\epsilon^{ijk}\Omega^{k} , it becomes clear that \bm{a} is the linear acceleration and \bm{\Omega} is the angular velocity. Note that though K^{i} explicitly depends on t , it is actually time independent as it is conserved (which can be directly checked using Poincare algebra, dK^{i}/dt=0 ) Becattini 2018 . This allows us to set t=0 in \rho_{\rm eq} . Choosing both the linear acceleration and rotation along z direction, we have \displaystyle\rho_{\rm eq} \displaystyle= \displaystyle\frac{1}{Z_{\rm eq}}\exp\left[-\beta\left(H-aK^{z}-\Omega J^{z}\right)\right] (7) \displaystyle= \displaystyle\frac{1}{Z_{\rm eq}}\exp\left[-\beta\int d^{3}{\mathbf{x}}\left((1+az)T^{00}-a\Sigma^{003}-\Omega(T^{02}x-T^{01}y+\Sigma^{012})\right)\right]\bigg|_{t=0}, where J^{z}=J^{12} . We now apply the above formalism to the single-flavor Nambu–Jona-Lasinio (NJL) model, whose Lagrangian is \mathcal{L}=\bar{\psi}\left[i\gamma^{\mu}\partial_{\mu}-m_{0}\right]\psi+\frac{G_{\pi}}{2}\left[(\bar{\psi}\psi)^{2}+\left(\bar{\psi}i\gamma^{5}\psi\right)^{2}\right], (8) where \psi denotes the Dirac spinor field, m_{0} is the fermion mass, and G_{\pi} is the coupling constant. Applying the Noether theorem, we obtain the canonical energy-momentum tensor and spin tensor as \displaystyle T^{\mu\nu} \displaystyle= \displaystyle{\bar{\psi}}i\gamma^{\mu}\partial^{\nu}\psi-g^{\mu\nu}{\cal L}, (9) \displaystyle\Sigma^{\mu\rho\sigma} \displaystyle= \displaystyle\frac{1}{2}{\bar{\psi}}\gamma^{\mu}\sigma^{\rho\sigma}\psi, (10) with \sigma^{\mu\nu}=(i/2)[\gamma^{\mu},\gamma^{\nu}] so that \Sigma^{003}=(i/2){\bar{\psi}}\gamma^{3}\psi and \Sigma^{012}=(1/2)\psi^{\dagger}\sigma^{z}\psi . Then we have \displaystyle H-aK^{z}-\Omega J^{z} \displaystyle= \displaystyle\int d^{3}{\mathbf{x}}\Big\{(1+az)\bar{\psi}\left[-i\gamma^{i}\partial_{i}+m_{0}\right]\psi (11) \displaystyle-(1+az)\frac{G_{\pi}}{2}\left[(\bar{\psi}\psi)^{2}+\left(\bar{\psi}i\gamma_{5}\psi\right)^{2}\right]-a\frac{i}{2}{\bar{\psi}}\gamma^{3}\psi-\Omega\psi^{\dagger}\left[-ix\partial_{y}+iy\partial_{x}+\frac{\sigma^{z}}{2}\right]\psi\Big\}. The partition function can then be written in a path-integral form as \displaystyle Z_{\rm eq}={\rm Tr}\;e^{-\beta(H-aK^{z}-\Omega J^{z})}=\int[d\psi][d\bar{\psi}]e^{-S_{E}[\psi,\bar{\psi}]}, (12) where S_{E} is the Euclidean action which is obtained from the Minkowski action S=\int d^{4}x{\bar{\psi}}i\gamma^{0}\partial_{t}\psi-\int dt(H-aK^{z}-\Omega J^{z}) by the standard Wick rotation and imaginary-time compactification with anti-periodic boundary condition for \psi and {\bar{\psi}} in the imaginary-time direction. Notably, the Minkowski action S is completely equal to the NJL action in an accelerating and rotating coordinates with metric \displaystyle g_{\mu\nu}=\begin{pmatrix}(1+az)^{2}-\Omega^{2}(x^{2}+y^{2})&\Omega y&-\Omega x&0\\ \Omega y&-1&0&0\\ -\Omega x&0&-1&0\\ 0&0&0&-1\end{pmatrix}; (13) see Sec. V , especially Eq. ( 76 ). Therefore, we can study the accelerating and rotating thermal equilibrium by formulating the theory in the accelerating and rotating coordinates so that the techniques developed for field theory in curved spacetime can be used. This will be our strategy in the following sections and it will be shown more convenient to shift z coordinate to \xi=z+1/a coordinate (see Sec. III ). It is well known that the definitions of energy-momentum tensor and spin tensor are not unique. They suffer from the so-called pseudo-gauge ambiguity Becattini et al. 2019 ; Huang 2025 . For any differentiable local tensor \Phi^{\lambda\mu\nu}(=-\Phi^{\lambda\nu\mu}) , T^{\prime\mu\nu} and \Sigma^{\prime\mu\rho\sigma} defined through the following pseudo-gauge transformation, \displaystyle T^{\prime\mu\nu} \displaystyle= \displaystyle T^{\mu\nu}+\frac{1}{2}\partial_{\lambda}\left(\Phi^{\lambda\mu\nu}+\Phi^{\nu\mu\lambda}+\Phi^{\mu\nu\lambda}\right), (14) \displaystyle\Sigma^{\prime\mu\rho\sigma} \displaystyle= \displaystyle\Sigma^{\mu\rho\sigma}-\Phi^{\mu\rho\sigma}, (15) are also well-defined energy-momentum and spin tensors in the sense that the conservation laws and the total four-momentum and total angular momentum are unchanged by this transformation. Therefore, the global thermal equilibrium density operator \rho_{\rm eq} and partition function Z_{\rm eq} are invariant under pseudo-gauge transformation (though the local equilibrium density operator and partition function are not) and we thus have a freedom to choose a suitable pseudo-gauge to simply the action S . In particular, we could eliminate the acceleration-spin tensor coupling -a\Sigma^{003}=-ia{\bar{\psi}}\gamma^{3}\psi/2 term in Eq. ( 11 ) by choosing the following pseudo-gauge potential \Phi^{\lambda\mu\nu} : \Phi^{\lambda\mu\nu}=\frac{i}{2}\bar{\psi}\left(\eta^{\lambda\mu}\gamma^{\nu}-\eta^{\lambda\nu}\gamma^{\mu}\right)\psi. (16) This transforms Eqs. ( 9 )-( 10 ) to \displaystyle T^{\prime\mu\nu} \displaystyle= \displaystyle\frac{i}{2}[\bar{\psi}\gamma^{\mu}\partial^{\nu}\psi-\bar{\psi}\overleftarrow{\partial}^{\nu}\gamma^{\mu}\psi]-g^{\mu\nu}\mathcal{L}^{\prime}, (17) \displaystyle\Sigma^{\prime\mu\rho\sigma} \displaystyle= \displaystyle\frac{1}{4}{\bar{\psi}}\{\gamma^{\mu},\sigma^{\rho\sigma}\}\psi=-\frac{1}{2}\epsilon^{\mu\rho\sigma\nu}{\bar{\psi}}\gamma_{\nu}\gamma_{5}\psi, (18) where \mathcal{L}^{\prime} is a new Lagrangian, \mathcal{L}^{\prime}={\bar{\psi}}\left[\frac{i}{2}(\gamma^{\mu}\partial_{\mu}-\overleftarrow{\partial}_{\mu}\gamma^{\mu})-m_{0}\right]\psi+\frac{G_{\pi}}{2}\left[(\bar{\psi}\psi)^{2}+\left(\bar{\psi}i\gamma^{5}\psi\right)^{2}\right]. (19) The difference is that \Sigma^{\prime 003}=0 so that -a\Sigma^{\prime 003} term no longer appears in H^{\prime}-aK^{\prime z}-\Omega J^{\prime z} : \displaystyle H^{\prime}-aK^{\prime z}-\Omega J^{\prime z} \displaystyle= \displaystyle\int d^{3}{\mathbf{x}}\Big\{(1+az)\bar{\psi}\left[\frac{1}{2}(-i\gamma^{i}\partial_{i}+i\gamma^{i}\overleftarrow{\partial}_{i})+m_{0}\right]\psi (20) \displaystyle-(1+az)\frac{G_{\pi}}{2}\left[(\bar{\psi}\psi)^{2}+\left(\bar{\psi}i\gamma_{5}\psi\right)^{2}\right]-\Omega\psi^{\dagger}\left[-ix\partial_{y}+iy\partial_{x}+\frac{\sigma^{z}}{2}\right]\psi\Big\}. After integration by part, one finds that H^{\prime}-aK^{\prime z}-\Omega J^{\prime z} indeed equals to H-aK^{z}-\Omega J^{z} in Eq. ( 11 ) (up to surface terms which are assumed to be vanished). Similar to discussions around Eqs. ( 12 )-( 13 ), the partition function of H^{\prime}-aK^{\prime z}-\Omega J^{\prime z} can be formulated as path integral with Lagrangian ( 19 ) in accelerating and rotating frames with metric ( 13 ). III Rindler coordinates The spacetime geometry in Rindler coordinates is characterized by the metric (see Appendix A for a derivation) \mathrm{d}s^{2}=\xi^{2}\mathrm{d}t^{2}-\mathrm{d}\xi^{2}-\mathrm{d}x_{1}^{2}-\mathrm{d}x_{2}^{2}. (21) A particle at rest in Rindler coordinates corresponds to the one undergoing constant proper acceleration in Minkowski spacetime. For a particle under proper acceleration a>0 , the worldline is given by t=a\,\tau , \xi=1/a , where \tau denotes the proper time. To describe finite-temperature systems, a Euclidean formulation can be obtained via analytic continuation t_{E}=it . A system with proper acceleration a and temperature T is described by the Euclidean metric: \mathrm{d}s^{2}=-\mathrm{d}s_{E}^{2}=-\frac{\xi^{2}}{\nu^{2}}\mathrm{d}t_{E}-\mathrm{d}\xi^{2}-\mathrm{d}x_{1}^{2}-\mathrm{d}x_{2}^{2}. (22) where \nu=2\pi T/a . The Euclidean time t_{E} satisfies 00 and \mathrm{Re}(l+\mu+\nu)>-1 ), and following the Klein-Gordon inner product in Rindler coordinates Candelas and Raine 1976 ; Crispino et al. 2008 , we find that the eigenfunctions obey the orthonormality condition Candelas and Raine 1976 : (\phi_{i},\phi_{j})_{\text{KG}}=2\omega\int_{0}^{\infty}\mathrm{d}\xi\int\mathrm{d}^{2}{\mathbf{x}}\sqrt{-g}g^{00}\phi_{i}\phi_{j}=\delta(i,j), (34) where i,j=(\omega,k_{1},k_{2}) labels the eigenmodes of the field \phi . Upon performing the secondary quantization, the scalar field \phi becomes an operator: \hat{\phi}=\int^{\infty}_{0}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}{\mathbf{k}}}{2\pi}\sqrt{\frac{\sinh\pi\omega}{\pi^{2}}}K_{i\omega}(m_{\perp}\xi)(\hat{a}_{\omega,k}\mathrm{e}^{-i\omega t+i{\mathbf{k}}\cdot{\mathbf{x}}}+\hat{a}_{\omega,k}^{\dagger}\mathrm{e}^{i\omega t-i{\mathbf{k}}\cdot{\mathbf{x}}}), (35) where \hat{a}_{\omega,k} and \hat{a}_{\omega,k}^{\dagger} are the annihilation and creation operators, respectively. Particles and antiparticles in Rindler coordinates are defined with respect to the Killing vector \partial/\partial t through their frequency \omega : modes with positive (negative) frequency correspond to particles (antiparticles). It follows directly that the Euclidean propagator can be written as: \begin{split}G_{E}(x_{E},x_{E}^{\prime})&=\int^{\infty}_{0}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}\mathbf{k}}{(2\pi)^{2}}\frac{\sinh\pi\omega}{\pi^{2}}K_{i\omega}(m_{\perp}\xi)K_{i\omega}(m_{\perp}\xi^{\prime})e^{-\omega|t_{E}-t_{E}^{\prime}|+i\mathbf{k}\cdot(\mathbf{x}-\mathbf{x}^{\prime})},\end{split} (36) where t_{E} is the Euclidean Rindler time. We emphasize that the above expression for G_{E} corresponds to the Rindler vacuum at zero temperature, serving as the baseline for the construction of the thermal propagator and the renormalization scheme discussed later. In previous studies, the temperature has typically been fixed at T=T_{U} , a natural choice from the Unruh effect (see, e.g., Ref. Casado-Turrión and Dobado 2019 ). In this work, we extend the analysis by developing a generalized formalism valid for arbitrary temperatures satisfying T\geq T_{U} . It is important to note that Eq. ( 36 ) describes the Euclidean propagator without boundary condition, corresponds to a system in the Rindler vacuum at zero temperature, from which we can construct a propagator at any temperature by imposing periodic boundary condition in Euclidean time. Specifically, the finite-temperature propagator G_{\nu} should satisfy the condition G_{\nu}(t_{E},t_{E}^{\prime})=G_{\nu}(t_{E}+\beta_{R},t_{E}^{\prime}) where \beta_{R}=1/T_{R}=a/T and \nu=2\pi T/a as defined in Eq. ( 22 ). Based on the periodic boundary condition in Euclidean time, the finite-temperature propagator G_{\nu} can be constructed from the zero-temperature propagator G_{E} in Eq. ( 36 ) via \begin{split}G_{\nu}(x_{E},x^{\prime}_{E})&=\sum_{n}G_{E}(t_{E}-t_{E}^{\prime}+\beta_{R}n)\\ &=\int^{\infty}_{0}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}{\mathbf{k}}}{(2\pi)^{2}}\frac{\cosh\omega(|t_{E}-t_{E}^{\prime}|-\beta_{R}/2)}{\sinh(\beta_{R}\omega/2)}\frac{\sinh\pi\omega}{\pi^{2}}K_{i\omega}(m_{\perp}\xi)K_{i\omega}(m_{\perp}\xi^{\prime})\mathrm{e}^{i{\mathbf{k}}\cdot({\mathbf{x}}-{\mathbf{x}^{\prime}})}.\\ \end{split} (37) Alternatively, this result can be obtained by directly solving the thermal Green’s function equation in Euclidean Rindler space with periodic boundary condition in t_{E} : \left(\frac{1}{\xi^{2}}\frac{\partial^{2}}{\partial t_{E}^{2}}+\frac{\partial^{2}}{\partial\xi^{2}}+\frac{1}{\xi}\frac{\partial}{\partial\xi}+\frac{\partial^{2}}{\partial x_{E1}^{2}}+\frac{\partial^{2}}{\partial x_{E2}^{2}}-m^{2}\right)G(x_{E},x^{\prime}_{E})=-\frac{\delta^{(4)}(x_{E}-x^{\prime}_{E})}{\xi}, (38) whose solution is G_{\nu}(x_{E},x^{\prime}_{E})=\sum_{n}\int^{\infty}_{0}\frac{\mathrm{d}\omega}{\beta}\int\frac{\mathrm{d}^{2}{\mathbf{k}}}{(2\pi)^{2}}\frac{\mathrm{e}^{i\omega_{n}(t_{E}-t_{E}^{\prime})+i{\mathbf{k}}\cdot({\mathbf{x}}_{1}-{\mathbf{x}}_{2})}}{\omega_{n}^{2}+\omega^{2}}\frac{2\omega\sinh\pi\omega}{\pi^{2}}K_{i\omega}(m_{\perp}\xi)K_{i\omega}(m_{\perp}\xi^{\prime}), (39) where \omega_{n}=2\pi n/\beta_{R} is the Matsubara frequency of bosons. Performing the frequency summation in Eq. ( 39 ) yields Eq. ( 37 ). For our purpose of studying chiral phase transition, the key quantity is the diagonal part of the propagator: \begin{split}G_{\nu}(x_{E},x_{E})&=\int^{\infty}_{0}\frac{\mathrm{d}\omega}{\pi^{2}}\frac{\mathrm{d}^{2}{\mathbf{k}}}{(2\pi)^{2}}\coth{(\omega\beta_{R}/2)}\,\sinh(\pi\omega)K^{2}_{i\omega}(m_{\perp}\xi).\end{split} (40) According to the first gap equation in Eq. ( 29 ), the order parameter \sigma satisfies -\Box\sigma+\lambda\sigma=0 in the chiral limit ( M_{\pi}\to 0 ). Although both \lambda and \sigma generally depend on position due to the inhomogeneous nature of acceleration, we focus on phase transitions evaluated at the observer’s instantaneous location (with a fixed \xi ). Under the slow-variation approximation ( \Box\sigma\approx 0 ), the gap equation simplifies to \lambda=0 and \sigma^{2}=f_{\pi}^{2}-NG_{\nu}(x_{E},x_{E};m=0) , and the phase structure can be determined by evaluating the diagonal part of the massless scalar propagator at the observer’s location. Applying the integral identity, \displaystyle\int_{0}^{\infty}\mathrm{d}xK_{i\mu}(\alpha x)K_{i\mu}(\alpha x)x=\frac{\pi\mu}{2\alpha^{2}\sinh(\pi\mu)}, (41) valid for -1<\mathrm{Im}(\mu)<1 and \mathrm{Re}(\alpha)>0 , we can derive the diagonal part of the massless propagator as: \displaystyle G_{\nu}(x_{E},x_{E};m=0) \displaystyle=\int^{\infty}_{0}\frac{\mathrm{d}\omega}{\pi^{2}}\frac{1}{(2\pi)}\coth{(\omega\beta_{R}/2)}\frac{\pi\omega}{2\xi^{2}} (42) \displaystyle=\int^{\Lambda}_{0}{\mathrm{d}\omega}\frac{\omega}{4\pi^{2}a^{2}\xi^{2}}+\int^{\infty}_{0}{\mathrm{d}\omega}\frac{\omega}{4\pi^{2}a^{2}\xi^{2}}\frac{2}{e^{\omega\beta}-1}. In the last line, we have applied a change of variables \omega\to\omega/a , which facilitates ultraviolet regularization via an a independent energy cutoff \Lambda . To obtain a finite result, we subtract the divergent, temperature-independent part of G_{\nu} (the first term in the second line of Eq. ( 42 )), corresponding to the vacuum contribution. This accounts to the a subtraction G_{\nu}^{ren}=G_{\nu}-G_{\nu=0} . A more detailed discussion of the renormalization scheme can be found in Sec. IV.3 . Accordingly, the renormalized diagonal part of the propagator takes the form: G^{ren}_{\nu}(x_{E},x_{E};m=0)=\frac{\nu^{2}}{48\pi^{2}\xi^{2}}. (43) Note that this subtraction procedure can also be viewed as a renormalization procedure to the phenomenological constant, f^{2}_{\pi}\to f^{2}_{\pi}+\frac{N\Lambda^{2}}{8\pi^{2}a^{2}\xi^{2}} . Substituting this into the gap equation ( 29 ), we obtain the expression for the order parameter \sigma : \begin{split}\sigma^{2}=f_{\pi}^{2}-\frac{T^{2}}{a^{2}\xi^{2}}\frac{N}{12}.\end{split} (44) This shows that, as the temperature increases, the chiral condensate \sigma decreases and eventually vanishes. We define the critical temperature at which \sigma reaches 0 as \begin{split}T_{c}(\xi)=a\xi\sqrt{\frac{12f_{\pi}^{2}}{N}}.\end{split} (45) This result is consistent with earlier study Casado-Turrión and Dobado 2019 . As emphasized previously, our analysis focuses on the observer’s instantaneous local position. The worldline of an observer of constant proper acceleration a is given by t=a\tau and \xi=1/a , as discussed in Sec. III . Thus we find that the critical temperature observed by this observer becomes independent of its proper acceleration: T_{c}(\xi=1/a)=\sqrt{12f_{\pi}^{2}/N}\equiv T_{c0} , indicating that the phase transition depends solely on the local temperature experienced by the accelerated observer, not on the proper acceleration of the observer itself. Note that thus Eq. ( 45 ) can be rewritten as T_{c}(\xi)/a\xi=T_{c0} with the left-hand side representing the Tolman-Ehrenfest critical temperature. Therefore the chiral phase transition happens at the same Tolman-Ehrenfest temperature showing that local observers at different \xi agree with the critical temperature for chiral phase transition. We will revisit and further interpret this result after evaluating the fermionic case in the next subsection. IV.2 NJL model analysis We next turn to the quark degree of freedom to investigate how acceleration affects fermions. In contrast to the scalar case, a key distinction lies in the coupling between acceleration and fermion spin for some choices of pseudo-gauge, analogous to the spin-rotation coupling in a rotating system. Here we start with the single-flavor NJL model, whose dynamics in a general spacetime are governed by the curved-spacetime version of Lagrangian ( 8 ): \mathcal{L}_{NJL}=\bar{\psi}\left[i\gamma^{\mu}\nabla_{\mu}-m_{0}\right]\psi+\frac{G_{\pi}}{2}\left[(\bar{\psi}\psi)^{2}+\left(\bar{\psi}i\gamma^{5}\psi\right)^{2}\right]. (46) The covariant derivative \nabla_{\mu} , which encodes the effects of curved spacetime on fermions, will be defined explicitly in the following discussion. From the Lagrangian in Eq. ( 46 ), the corresponding generating functional or partition function is given by: \displaystyle Z \displaystyle=\int\mathcal{D}[\bar{\psi},\psi]\exp\left(i\int\mathrm{d}^{4}x\sqrt{-g}\mathcal{L}_{\mathrm{NJL}}\right) (47) \displaystyle=\int\mathcal{D}[\bar{\psi},\psi,\sigma,\pi]\exp\left\{i\int\mathrm{d}^{4}x\sqrt{-g}\left[\bar{\psi}\left(i\gamma^{\mu}\nabla_{\mu}-m-i\gamma^{5}\pi\right)\psi-\frac{\sigma^{2}+\pi^{2}}{2G_{\pi}}\right]\right\}, where the Hubbard-Stratonovich transformation introduces auxiliary fields \sigma\equiv-G_{\pi}\bar{\psi}\psi and \pi\equiv-G_{\pi}\bar{\psi}i\gamma^{5}\psi , with the constituent quark mass defined as m=m_{0}+\sigma . Within the mean-field approximation, we derive the spacetime-dependent effective potential density, which is given by: V_{eff}(x)=\frac{\sigma^{2}+\pi^{2}}{2G_{\pi}}-\frac{1}{i}\operatorname{Tr}\ln\left(i\gamma^{\nu}\nabla_{\nu}-m-i\gamma^{5}\pi\right). (48) Minimizing the effective potential density leads to the gap equations: \frac{m-m_{0}}{G_{\pi}}=i\operatorname{Tr}(S),\quad\frac{\pi}{G_{\pi}}=i\operatorname{Tr}(i\gamma^{5}S), (49) where the fermion propagator S satisfies: \left(i\gamma^{\nu}\nabla_{\nu}-m-i\gamma^{5}\pi\right)_{x}S(x,x^{\prime})=\frac{1}{\sqrt{-g}}\delta^{4}(x,x^{\prime}). (50) Without loss of generality, here we can set \pi=0 due to the U(1) chiral symmetry and focus solely on \sigma . Following general formalism of quantum field theory in curved spacetime, \gamma^{\mu}=e^{\mu}_{\hat{m}}\gamma^{\hat{m}} represent the spacetime-dependent gamma matrices and \gamma^{\hat{m}} represent the gamma matrices in Minkowski spacetime. And we choose the vierbein e^{\mu}_{\hat{m}} as e^{0}_{\hat{0}}=\frac{1}{\xi}\,,\,e^{i}_{\hat{0}}=0,e^{0}_{\hat{j}}=0,e^{i}_{\hat{j}}=\delta^{i}_{j}. (51) where i,j=1,2,3 with 3 for \xi coordinate. The covariant derivative \nabla_{\mu} and spin connection \Gamma_{\mu} satisfy: \displaystyle\nabla_{\mu}=\partial_{\mu}+\Gamma_{\mu},\quad\Gamma_{\mu}=-\frac{i}{4}\omega_{\mu\hat{\nu}\hat{\rho}}\sigma^{\hat{\nu}\hat{\rho}},\quad\sigma^{\mu\nu}=\frac{i}{2}[\gamma^{\hat{\mu}},\gamma^{\hat{\nu}}],\quad\omega_{\mu\hat{\kappa}\hat{\eta}}=g_{\nu\rho}e^{\nu}_{\hat{\kappa}}\nabla_{\mu}e^{\rho}_{\hat{\eta}}. (52) For the metric specified in Eq. ( 21 ), the corresponding gamma matrices take the form: \begin{gathered}\gamma^{0}(x)=\frac{1}{\xi}{\gamma^{\hat{0}}},\quad\gamma^{i}(x)=\gamma^{\hat{i}},\end{gathered} (53) with the non-zero spin connection component: \Gamma_{0}=\frac{1}{2}\gamma^{\hat{0}}\gamma^{\hat{3}}. (54) Note that the spin connection introduces a spin-dependent term in the partition function [Eq. ( 47 )] and in the propagator [see Eq. ( 50 ) and also Eq. ( 55 ) below]. Such a term does not appear in Becattini 2018 ; Prokhorov et al. 2019c , where accelerating fermions are treated via the density operator. In fact, this extra contribution can be removed through a pseudo-gauge transformation, as we have shown in Sec. II . The resultant theory is equivalent to replacing \mathcal{L}_{NJL} in Eq. ( 46 ) by \mathcal{L}^{\prime}_{NJL}=(\mathcal{L}_{NJL}+\mathcal{L}_{NJL}^{\dagger})/2 . Note that the equation of motion is unchanged by the pseudo-gauge transformation, and therefore both \mathcal{L}^{\prime}_{NJL} and \mathcal{L}_{NJL} yield the same propagator. We will hance focus solely on \mathcal{L}_{NJL} in the following. We now return to the discussion of the propagator. Based on the above setup, the explicit expression of Eq. ( 50 ) is: {\left[\frac{\gamma^{\hat{0}}}{\xi}\left(i\partial_{0}+\frac{i}{2}\gamma^{\hat{0}}\gamma^{\hat{3}}\right)+i\gamma^{\hat{i}}\partial_{i}-m\right]S(x,x^{\prime})\equiv(\hat{D}-m)S(x,x^{\prime})=\frac{1}{\sqrt{-g}}\delta^{4}(x,x^{\prime}).} (55) Since both \mathcal{L}^{\prime}_{NJL} and \mathcal{L}_{NJL} yield the same equations of motion, S(x,x^{\prime}) is also the propagator for \mathcal{L}^{\prime}_{NJL} . To solve this equation, we adopt the ansatz that the propagator can be expressed as S(x,x^{\prime})=(\hat{D}+m)G(x,x^{\prime}), (56) with G(x,x^{\prime}) an auxiliary propagator which satisfies \displaystyle\left(\hat{D}^{2}-m^{2}\right)G(x,x^{\prime})=\frac{1}{\sqrt{-g}}\delta(x,x^{\prime}), (57) where the squared Dirac operator \hat{D}^{2} explicitly takes the form: \hat{D}^{2}=-\frac{1}{\xi^{2}}\left(i(i\partial_{0})-\frac{\gamma^{\hat{0}}\gamma^{\hat{3}}}{2}\right)^{2}+\partial_{3}^{2}+\frac{1}{\xi}\partial_{3}+\partial_{1}^{2}+\partial_{2}^{2}. (58) In comparison to the scalar case, a only distinguishing feature is the presence of the \gamma^{\hat{0}}\gamma^{\hat{3}} term, which reflects the contribution from spin. Therefore, the propagator can be constructed by the eigenmodes of Klein-Gordon operator with i\omega\rightarrow i\omega-{\gamma^{\hat{0}}\gamma^{\hat{3}}}/{2} : e^{-i\omega t}e^{i\mathbb{k}\cdot\mathbb{x}}K_{i\omega-{\gamma^{\hat{0}}\gamma^{\hat{3}}}/{2}}(m_{\perp}\xi). (59) For later use, it is useful to notice that, for a function containing \gamma^{0}\gamma^{3} , it can always be decomposed as f(\gamma^{0}\gamma^{3})=P^{+}f(1)+P^{-}f(-1), (60) where P^{\pm}=(1\pm\gamma^{0}\gamma^{3})/2 are two projections. Therefore, the auxiliary propagator in Euclidean Rindler spacetime is directly obtained from Eq. ( 36 ). The result is G_{E}(x_{E},x_{E}^{\prime})=\int^{\infty}_{0}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}{\mathbf{k}}}{(2\pi)^{2}}\frac{\sinh\pi(\omega+i\gamma^{\hat{0}}\gamma^{\hat{3}}/2)}{\pi^{2}}K_{i\omega-\gamma^{\hat{0}}\gamma^{\hat{3}}/2}(m_{\perp}\xi)K_{i\omega-\gamma^{\hat{0}}\gamma^{\hat{3}}/2}(m_{\perp}\xi^{\prime})e^{-\omega|t_{E}-t_{E}^{\prime}|+i{\mathbf{k}}\cdot({\mathbf{x}}-{\mathbf{x}}^{\prime})}. (61) By applying the fermionic anti-periodic boundary condition G_{\nu}(t_{E},t_{E}^{\prime})=-G_{\nu}(t_{E}+\beta_{R},t_{E}^{\prime}) , the thermal propagator G_{\nu}(x_{E},x_{E}^{\prime}) can be constructed in a manner analogous to the scalar case: \begin{split}G_{\nu}&=\sum_{n}(-1)^{n}G_{E}(t_{E}-t_{E}^{\prime}+\beta_{R}n)\\ &=\int^{\infty}_{0}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}{\mathbf{k}}}{(2\pi)^{2}}\frac{\sinh(\beta_{R}\omega/2-\omega|t_{E}-t_{E}^{\prime}|)}{\cosh(\beta_{R}\omega/2)}\frac{\sinh\pi(\omega+i\gamma^{\hat{0}}\gamma^{\hat{3}}/2)}{\pi^{2}}K_{i\omega-\gamma^{\hat{0}}\gamma^{\hat{3}}/2}(m_{\perp}\xi)K_{i\omega-\gamma^{\hat{0}}\gamma^{\hat{3}}/2}(m_{\perp}\xi^{\prime})e^{i{\mathbf{k}}\cdot({\mathbf{x}}-{\mathbf{x}}^{\prime})}.\end{split} (62) The diagonal part of the propagator becomes: G_{\nu}(x_{E},x_{E})=\sum_{s=\pm}\int^{\infty}_{0}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}{\mathbf{k}}}{(2\pi)^{2}}\frac{\sinh({\beta_{R}}\omega/2)}{\cosh({\beta_{R}}\omega/2)}\frac{is\cosh\pi\omega}{\pi^{2}}K^{2}_{i\omega-s/2}(m_{\perp}\xi). (63) As before, we consider the chiral limit and focus on the phase boundary where m=0 , \displaystyle G_{\nu}(x_{E},x_{E};m=0) \displaystyle=\int_{0}^{\infty}\mathrm{d}\omega\frac{\omega}{\pi^{2}\xi^{2}}\tanh({\beta_{R}}\omega/2) (64) \displaystyle=\int_{0}^{\Lambda}\mathrm{d}\omega\frac{\omega}{\pi^{2}a^{2}\xi^{2}}+\int_{0}^{\infty}\mathrm{d}\omega\frac{\omega}{\pi^{2}a^{2}\xi^{2}}\frac{2}{e^{\beta\omega}+1} \displaystyle=\frac{1}{6a^{2}\xi^{2}}\left(\frac{3\Lambda^{2}}{\pi^{2}}-T^{2}\right). In the second line, we have performed a variable substitution \omega\to\omega/a and introduced an acceleration-independent cutoff \Lambda to regularize the divergent term (the first term), which corresponds to the Rindler vacuum contribution, G_{\nu=0}(x_{E},x_{E};m=0) . Substituting this result into the gap equation Eq. ( 49 ) and using the properties that the trace of odd-number gamma matrices vanishes, we have \frac{1}{G_{\pi}}=\frac{1}{6a^{2}\xi^{2}}\left(\frac{3\Lambda^{2}}{\pi^{2}}-T^{2}\right) (65) Therefore, we get the critical temperature: T_{c}(\xi)=a\xi\sqrt{\frac{3\Lambda^{2}}{\pi^{2}}-\frac{6}{G_{\pi}}}. (66) This expression agrees with the result in Ref. Ohsaku 2004 . Furthermore, by substituting the accelerated observer’s worldline condition \xi=1/a , one finds that the critical temperature is independent of the acceleration a , T_{c}(\xi=1/a)=\sqrt{{3\Lambda^{2}}/{\pi^{2}}-{6}/{G_{\pi}}}\equiv T_{c0} . As discussed in Section III , renormalization schemes are not unique. In Sec. IV.3 , we present a supplemental alternative scheme that yields results different from those in this subsection and in Sec. IV.1 . Figure 1 shows the phase diagram calculated from Eqs. ( 45 ) and ( 66 ). The blue region corresponds to the phase with spontaneously broken chiral symmetry, the white region indicates restored symmetry, and the grey region marks the unphysical domain where TT_{U} . As shown in Eqs. ( 45 ) and ( 66 ), the phase transition is governed solely by the temperature perceived by an accelerated observer with respect to the Rindler vacuum. Acceleration manifests its influence on phase transitions only indirectly via the temperature dependence. For instance, if one assumes the accelerated observer’s temperature to be fixed to the Unruh relation T=T_{U}=a/2\pi , as has been considered in Refs. Ohsaku 2004 ; Ebert and Zhukovsky 2007 ; Castorina and Finocchiaro 2012 ; Casado-Turrión and Dobado 2019 ; Basu et al. 2023 , one can determine a critical acceleration a_{c}=2\pi T_{c} , above which( a>a_{c} ) chiral symmetry is restored. Figure 1: The T-a phase diagram based on Eq. ( 45 )(for bosons) and Eq. ( 66 )(for fermions). T_{c0} denotes the critical the critical temperature in a=0 . Blue region denotes the chiral symmetry breaking phase, while the white region denotes the chiral symmetry restored phase. Grey region denotes area of T0 . Therefore, \displaystyle a=\frac{du^{z}_{M}}{dt_{M}}. (90) Because \alpha^{2}_{M} is a Lorentz scalar (so does a ), we can go into the instantaneous inertial frame comoving with the particle at each instant where a is just the acceleration measured in that reference frame. For this reason, a is called the proper acceleration of the particle. Let us find out the trajectory in Minkowski coordinates of a particle under constant proper acceleration a>0 . From Eq. ( 90 ), we find that \gamma_{M}v^{z}_{m}=at_{M} (supposing that v_{M}^{z}=0 at t_{M}=0 ) and thus \displaystyle v^{z}_{M}(t_{M})=\frac{at_{M}}{\sqrt{1+(at_{M})^{2}}}, (91) and \gamma_{M}(t_{M})=\sqrt{1+(at_{M})^{2}} . Integrating v_{M}^{z}(t_{M}) over t_{M} , we obtain \displaystyle z_{M}(t_{M})=\frac{1}{a}\sqrt{1+(at_{M})^{2}}+z_{M}(0)-\frac{1}{a}. (92) The particle is initially at z_{M}(0) and then accelerated towards the positive z_{M} direction (such that z_{M}(t_{M})\geq z_{M}(0) ). The trajectory is half of a hyperbola (another half of the hyperbola corresponds to a particle under the same proper acceleration a but moving in -z_{M} direction), \displaystyle t_{M}^{2}-\left(z_{M}-z_{M}(0)+\frac{1}{a}\right)^{2}=-\frac{1}{a^{2}},\quad z_{M}>z_{M}(0). (93) Now suppose that the particle is an observer, so that its reference frame is the accelerating frame associated with the particle. Its path is the above hyperbola in the Minkowski coordinates. Evidently, it is more convenient to make a coordinate transformation from (t_{M},z_{M}) to (\tau,z) so that the observer is static in this coordinate system. Let \displaystyle t_{M} \displaystyle= \displaystyle\left(z+\frac{1}{a}\right)\sinh(a\tau), (94) \displaystyle z_{M} \displaystyle= \displaystyle\left(z+\frac{1}{a}\right)\cosh(a\tau)+z_{M}(0)-\frac{1}{a}. (95) We have t_{M}^{2}-(z_{M}-z_{M}(0)+1/a)^{2}=-(z+1/a)^{2} . Thus, the observer at z=0 in the new coordinates (\tau,z) is accelerated by a proper acceleration a and \tau is its proper time. An observer at an arbitrary z is accelerated by a proper acceleration a/(1+az) . The line element ds^{2} is then \displaystyle ds^{2}=dt_{M}^{2}-dz_{M}^{2}=(1+az)^{2}d\tau^{2}-dz^{2}. (96) This coordinates is called Rindler coordinates. It is important to notice that (\tau,z) does not cover the whole Minkowski spacetime. In fact, for z>-1/a , z_{M}-z_{M}(0)+1/a>|t_{M}| (called the right Rindler wedge ), while for z<-1/a , z_{M}-z_{M}(0)+1/a<-|t_{M}| (called the left Rindler wedge). At z=-1/a , we have z_{M}-z_{M}(0)+1/a=\pm t_{M} which are two light cones called Rinder horizons. Sometimes it is more convenient to use another coordinates (t,\xi) with t=a\tau,\xi=z+1/a , so that \displaystyle ds^{2}=\xi^{2}dt^{2}-d\xi^{2}. (97) The observer at \xi has constant proper acceleration 1/\xi . This is the coordinates in Eq. ( 21 ). Appendix B Chiral limit Gap equation under acceleration and rotation In this appendix, we will provide a detailed derivation process for Eq. ( 87 ). We start from Eq. ( 84 ), where \displaystyle F_{0}=\int_{0}^{\infty}\mathrm{d}\omega\int\frac{\mathrm{d}^{2}p_{t}}{(2\pi)^{2}}\frac{1}{a}\frac{4\cosh(\pi\omega/a)}{\pi^{2}}{\rm{Im}}\left[K^{2}_{\frac{i\omega}{a}+\frac{1}{2}}\left({m_{\perp}\xi}\right)\right]=\frac{\Lambda^{2}}{2\pi^{2}}, (98) and F_{\Omega}=-\sum_{l}\int_{0}^{\infty}\mathrm{d}\omega\int_{0}^{\infty}\frac{p_{t}\mathrm{d}p_{t}}{2\pi}\frac{2\cosh(\pi\omega/a)}{a\pi^{2}}\left[\frac{1}{e^{\beta(\omega+\Omega j)}+1}+\frac{1}{e^{\beta(\omega-\Omega j)}+1}\right]{\rm{Im}}\left[K^{2}_{\frac{i\omega}{a}+\frac{1}{2}}\left({m_{\perp}\xi}\right)\right]\left[J^{2}_{l}(p_{t}r)+J^{2}_{l+1}(p_{t}r)\right], (99) where we have used a variable substitution \omega\to\omega/a . We handle rotational effects through analytic continuation from real to imaginary angular velocity ( \Omega=i\Omega_{I} ). This yields the F_{\Omega} as: \displaystyle F_{\Omega}=\int_{0}^{\infty}\mathrm{d}\omega\int_{0}^{\infty}\frac{p_{t}\mathrm{d}p_{t}}{2\pi}\frac{2\cosh(\pi\omega/a)}{a\pi^{2}}\sum_{n=1}^{\infty}(-1)^{n}e^{-n\beta\omega}{\rm{Im}}\left[K^{2}_{\frac{i\omega}{a}+\frac{1}{2}}\left({m_{\perp}\xi}\right)\right]4J_{0}(p_{t}r\sqrt{2-2\cos(n\beta\Omega_{I})})\cos\left(\frac{n\beta\Omega_{I}}{2}\right), (100) where the Fermi-Dirac distribution 1/(\mathrm{e}^{\beta(\omega\pm\Omega j)}+1) has been expanded via -\sum_{n=1}^{\infty}(-1)^{n}\mathrm{e}^{-n\beta(\omega\pm i\Omega_{I}j)} . The angular momentum summation over l is resolved through Bessel function identities Chen et al. 2023 : \sum_{n\in\mathbb{Z}}\mathrm{e}^{in\theta}J_{n}^{2}(x)=J_{0}(x\sqrt{2-2\cos\theta}). (101) We can deal with the K_{\frac{i\omega}{a}+\frac{1}{2}}\left({m_{\perp}\xi}\right) by using Candelas and Raine 1976 K_{i\nu}(\mu\xi)K_{i\nu}(\mu\xi^{\prime})=\frac{1}{2}\int_{-\infty}^{\infty}\mathrm{d}\lambda\mathrm{e}^{i\nu\lambda}K_{0}(\mu\gamma_{1}), (102) where \gamma^{2}_{1}=\xi^{2}+{\xi^{\prime}}^{2}+2\xi\xi^{\prime}\cosh(\lambda) . Thus \displaystyle{\rm{Im}}\left[K^{2}_{\frac{i\omega}{a}+\frac{1}{2}}\left({m_{\perp}\xi}\right)\right]=-\frac{1}{4}i\left[\int^{\infty}_{-\infty}\mathrm{d}\lambda\exp({i\frac{\omega}{a}\lambda})\left(2\sinh(\frac{\lambda}{2})\right)K_{0}\left(m_{\perp}\xi\sqrt{2+2\cosh(\lambda)}\right)\right]. (103) To perform the p_{t} integration, we expand K_{\nu} by using the integral representation K_{\nu}(t)=\frac{1}{2}(\frac{1}{2}t)^{\nu}\int_{0}^{\infty}\exp\left(-z-\frac{t^{2}}{4z}\right)\frac{\mathrm{d}z}{z^{\nu+1}}. (104) Applying this formula and setting \xi=1/a , we have \displaystyle\int_{0}^{\infty}\frac{\mathrm{d}z}{2z}e^{-\frac{1}{2}z}\int_{0}^{\infty}\frac{p_{t}\mathrm{d}p_{t}}{2\pi}\exp\left[{-\frac{1+\cosh(\lambda)}{z}\frac{m^{2}+p_{t}^{2}}{a^{2}}}\right]J_{0}(p_{t}r\sqrt{2-2\cos(n\beta\Omega_{I})}) (105) \displaystyle= \displaystyle\int_{0}^{\infty}\frac{\mathrm{d}z}{2z}e^{-\frac{1}{2}z}\int_{0}^{\infty}\frac{p_{t}\mathrm{d}p_{t}}{2\pi}\exp\left[-\frac{A}{z}(m^{2}+p_{t}^{2})\right]J_{0}(p_{t}B)=\int_{0}^{\infty}\frac{\mathrm{d}z}{2z}e^{-\frac{1}{2}z}\frac{z}{2\pi}exp\left[-\frac{zB^{2}}{4A}-\frac{Am^{2}}{z}\right]/(2A) \displaystyle= \displaystyle\frac{1}{2\pi}\frac{m}{\sqrt{(2+2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))}}K_{1}(m\sqrt{(2+2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))}), where auxiliary parameters A=\frac{1+\cosh(\lambda)}{a^{2}} and B=r\sqrt{2-2\cos(n\beta\Omega_{I})} . The integral over \omega can be evaluated using \int_{0}^{\infty}\mathrm{d}\omega\cosh(\frac{\pi\omega}{a})\exp[-n\beta\omega+i\frac{\omega}{a}\lambda]=\frac{ia}{2(ia\beta n+\lambda-i\pi)}+\frac{ia}{2(ia\beta n+\lambda+i\pi)}. (106) Now F_{\Omega} is simplified into \displaystyle F_{\Omega}= \displaystyle\frac{1}{\pi^{3}}\sum_{n=1}(-1)^{n}\int_{-\infty}^{\infty}\mathrm{d}\lambda\left(\frac{1}{ia\beta n+\lambda-i\pi}+\frac{1}{ia\beta n+\lambda+i\pi}\right)\frac{m}{\sqrt{(2+2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))}} (107) \displaystyle\times K_{1}(m\sqrt{(2+2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))})\sinh(\frac{\lambda}{2})\cos(\frac{n\beta\Omega_{I}}{2}) \displaystyle= \displaystyle\frac{1}{\pi^{3}}\sum_{n=1}(-1)^{n}\int_{C}\mathrm{d}\lambda\left(\frac{1}{ia\beta n+\lambda}\right)i\cosh(\frac{\lambda}{2})\frac{m}{\sqrt{(2-2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))}} \displaystyle\times K_{1}(m\sqrt{(2-2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))})\cos(\frac{n\beta\Omega_{I}}{2}), where the contour C is -\infty-i\pi\to\infty-i\pi and \infty+i\pi\to-\infty+i\pi . Based on the results obtained above, we can get the critical acceleration for chiral symmetry restoration a_{c}(\Omega) as a function of angular velocity \Omega analytically, by setting m\to 0 . Following the procedure outlined, the expression for F_{\Omega} in Eq. ( 107 ) consequently takes the form \displaystyle F_{\Omega} \displaystyle=i\frac{1}{\pi^{3}}\sum_{n=1}(-1)^{n}\int_{C}\mathrm{d}\lambda\left(\frac{1}{ia\beta n+\lambda}\right)\cosh(\frac{\lambda}{2})\frac{1}{{(2-2\cosh(\lambda))/a^{2}+r^{2}(2-2\cos(n\beta\Omega_{I}))}}\cos(\frac{n\beta\Omega_{I}}{2}) (108) \displaystyle=\sum_{n=1}(-1)^{n}\frac{2a\cot\left(\frac{\beta n\Omega_{I}}{2}\right)\sinh^{-1}\left(ar\sin\left(\frac{\beta n\Omega_{I}}{2}\right)\right)}{\pi^{2}a^{2}\beta^{2}n^{2}r+4\pi^{2}r\sinh^{-1}\left(ar\sin\left(\frac{\beta n\Omega_{I}}{2}\right)\right)^{2}}, where the \lambda integration is analytically computed through the residue theorem. To perform the n summation, we apply the fermionic analytic distillation theorem in Palermo et al. 2021 , yielding F_{\Omega}=-\left(\frac{\Omega_{I}^{2}\left(a^{2}r^{4}\Omega_{I}^{2}-a^{2}r^{2}-r^{2}\Omega_{I}^{2}-3\right)}{24\pi^{2}(r^{2}\Omega_{I}^{2}+1)^{2}}+\frac{1}{6\beta^{2}(r^{2}\Omega_{I}^{2}+1)}\right). (109) By implementing the condition \beta=2\pi/a and \Omega_{I}\to-i\Omega , F_{\Omega} is finally simplified into F_{\Omega}=-\frac{a^{2}\left(r^{4}\Omega^{4}+1\right)-r^{2}\Omega^{4}+3\Omega^{2}}{24\pi^{2}\left(r^{2}\Omega^{2}-1\right)^{2}}. (110) This directly leads to the gap equation G_{\pi}\left(\frac{\Lambda^{2}}{2\pi^{2}}-\frac{a^{2}\left(r^{4}\Omega^{4}+1\right)-r^{2}\Omega^{4}+3\Omega^{2}}{24\pi^{2}\left(r^{2}\Omega^{2}-1\right)^{2}}\right)=1, (111) as given by Eq. ( 87 ) in the main text.