betaFeedback

WorksPhysics and Astronomy

Thermodynamic Susceptibilities for a Unitary Fermi Gas

Abstract

The unitary Fermi gas provides a unique window into both cold atom experiments and neutron star properties. There are major challenges in determining the physical properties within a neutron star, both experimentally and theoretically. However, there is a region within the crust of a neutron star that resembles a gas of fermions that is recreated in a cold atom laboratory. This is the so-called unitary Fermi gas characterized by a large negative scattering length and small effective range. We set out to calculate from first principles certain transport coefficients that appear at second order in the hydrodynamic expansion. These particular transport coefficients are obtained from two point correlation functions in flat space that do not require analytic continuation from Euclidean to Minkowski space. We are motivated by the potential to learn about neutron star properties from cold atom experiments.

Introduction

Currently, one of the prominent unresolved challenges in physics is understanding the properties of matter under conditions of ultra-high density and low temperature. Neutron stars (NS) are an embodiment of such a physical regime, yet the conditions within a NS are far removed from terrestrial experiments and observations of such objects are limited. Since their discovery in 1967 [1] over 3,000 neutron stars have been cataloged, the vast majority detected as pulsars [2]. This relatively small set of data, which mostly relies on a particular alignment of Earth with the axis of the NS beacon, puts into perspective the challenges experimentalists face.

Theoretical obstacles are no less significant. Calculating properties of nuclear matter properties at high densities from first principles requires solving the equations of quantum chromodynamics (QCD), the accepted theory of nuclear interactions. Unfortunately, while efficient numerical methods exist for calculating QCD matter properties in equilibrium at small densities, these so-called lattice gauge theory calculations cannot be applied to high densities due to the infamous sign problem. In other words, the equations for calculating neutron star matter from first principles are known, but no one knows how to solve them.

The purpose of this paper is to calculate NS properties from first principles that can be measured in the lab. We are motivated by the fact that a region within the crust of a NS is composed of neutron-rich nuclei surrounded by a sea of free neutrons. Although free, these neutrons are strongly interacting, characterized by a large s-wave scattering length as≈−18.5 fma_s \approx-18.5\ \mathrm{fm} and a small effective range reff≈2.7 fmr_{\mathrm{eff}} \approx2.7\ \mathrm{fm} [3, 4]. When the scattering length is sent to infinity and the effective range to zero, all scales drop out and the system exhibits universal behavior since it is characterized only by its temperature, chemical potential, and density. Such a system is called a unitary Fermi gas (UFG) and the neutron sea we are interested in is close to this limit. Whether the fermions are atoms, electrons, quarks or something else, the way these systems interact in bulk will always be similar. UFGs have been studied extensively in cold atom laboratories where the s-wave scattering length of atomic fermions can be varied by applying an external magnetic field [5]. Utilizing Feshbach resonances, the s-wave scattering length of certain atomic systems can be made very large such that effectively kF∣as∣→∞k_F|a_s| \to\infty, and sufficiently dilute such that kFrs→0k_Fr_s \to0, where the Fermi momentum is related to the density of the system via kF∼n−1/3k_F \sim n^{-1/3} [6]. With this in mind, it may be possible to probe NS properties using table-top experiments. In particular, we are interested in transport properties of the UFG.

When a system is perturbed out of equilibrium, transport coefficients govern how the system relaxes back to equilibrium. Well-known examples include conductivities, viscosities, and diffusion constants [7]. Over the past several years many experimental and theoretical studies have been devoted to hydrodynamic properties of strongly interacting quantum fluids, such as the UFG or the quark gluon plasma (QGP). It is interesting that these two systems are considered the most perfect fluids in the universe [8], meaning their shear viscosity to entropy density ratio is close to the conjectured quantum lower bound η/s≲ℏ4πkB\eta/s \lesssim\frac{\hbar}{4\pi k_B} [9]. Transport coefficients such as the shear viscosity are, in general, difficult to calculate. However, in a remarkable paper Kovtun and Shukla [10] showed that a certain family of transport coefficients, so-called thermodynamic transport coefficients, are considerably easier to calculate because they do not require analytic continuation from Euclidean to Minkowski spacetime. These quantities are classified as “thermodynamic” because they can be obtained by calculating certain thermodynamic susceptibilities in equilibrium.

Hydrodynamics is an effective low energy theory of systems near thermal equilibrium [11] [12]. The expressions are valid on length scales much larger than the characteristic microscopic sizes of the system, such as the mean free path or correlation length. Hydrodynamics admits a derivative expansion in terms of the fluid variables such as temperature, chemical potential, and fluid velocity. At each order in this expansion are various transport coefficients quantifying the response of the system to external perturbations. Certain transport coefficients exhibit dissipation, leading to the generation of entropy in non-equilibrium conditions, while others remain non-dissipative and persist even in the limit of thermal equilibrium. Thermodynamic transport coefficients appear at second order in the gradient expansion and are non-dissipative in nature [13, 14].

These coefficients have been calculated for free uncharged relativistic scalar, vector, and fermionic quantum field theories [10, 15, 16] as well as free Dirac fermions at finite density [17]. For interacting theories, results have been obtained for large NN, N=4\mathcal{N}=4 super Yang-Mills theory at infinite coupling for both zero and finite density utilizing the AdS/CFT conjecture [13, 18, 19, 20, 21]. For coupling strengths between the weak and strong limits, analytic results were obtained in the large NN limit for the O(N)O(N) model [22, 23] and for cold non-relativistic fermions at finite density [24]. In this latter paper, Lawrence and Romatschke calculated the second order transport coefficient κ\kappa for a UFG and we set out to calculate one more.

We work in natural units, ℏ=c=kB=1\hbar=c=k_B=1.

Calculation

A system of fermions at unitarity begins with the Hamiltonian density in three spatial dimensions,

H=∑s=↑,↓ψs†(x)(−∇22m)ψs(x)+4πasmψ↑†(x)ψ↓†(x)ψ↓(x)ψ↑(x),(1)\mathcal{H}=\sum_{s=\uparrow,\downarrow}\psi_s^\dagger(\mathbf{x})\left(\frac{-\nabla^2}{2m}\right)\psi_s(\mathbf{x})+\frac{4\pi a_s}{m}\psi_\uparrow^\dagger(\mathbf{x})\psi_\downarrow^\dagger(\mathbf{x})\psi_\downarrow(\mathbf{x})\psi_\uparrow(\mathbf{x}), \tag*{(1)}

where ψs†\psi_s^\dagger, ψs\psi_s are the fermionic creation and annihilation operators for states s∈{↑,↓}s\in\{\uparrow,\downarrow\}, respectively, mm is the fermion mass, and as<0a_s<0 is the s-wave scattering length. Thermodynamic properties of this system are derived from its partition function at finite temperature, TT and chemical potential, μ\mu. The partition function in the path integral formulation is,

Z(T,μ)=∫Dψs†Dψse−SE,(2)Z(T,\mu)=\int\mathcal{D}\psi_s^\dagger\mathcal{D}\psi_s e^{-S_E}, \tag*{(2)}
SE=∫0βdτ∫d3x[ψs†(x)(∂τ−∇22m−μ)ψs(x)+4πasmψ↑†(x)ψ↓†(x)ψ↓(x)ψ↑(x)],S_E=\int_0^\beta d\tau\int d^3x\left[\psi_s^\dagger(\mathbf{x})\left(\partial_\tau-\frac{\nabla^2}{2m}-\mu\right)\psi_s(\mathbf{x})+\frac{4\pi a_s}{m}\psi_\uparrow^\dagger(\mathbf{x})\psi_\downarrow^\dagger(\mathbf{x})\psi_\downarrow(\mathbf{x})\psi_\uparrow(\mathbf{x})\right],

where β≡1/T\beta\equiv1/T is the inverse temperature and we utilize the Einstein summation convention to sum over the repeated indices ss. We can make the path integral tractable by decoupling the four-fermi interaction with a Hubbard-Stratonovich transformation and introducing a complex auxiliary field ζ(x)\zeta(\mathbf{x}),

exp⁡(−∫τ,x4πasmψ↑†ψ↓†ψ↓ψ↑)=∫Dζexp⁡(∫τ,x[m4πasζζ∗−iψ↓ψ↑ζ∗+iψ↑†ψ↓†ζ])(3)\exp\left(-\int_{\tau,\mathbf{x}}\frac{4\pi a_s}{m}\psi_\uparrow^\dagger\psi_\downarrow^\dagger\psi_\downarrow\psi_\uparrow\right)=\int\mathcal{D}\zeta\exp\left(\int_{\tau,\mathbf{x}}\left[\frac{m}{4\pi a_s}\zeta\zeta^*-i\psi_\downarrow\psi_\uparrow\zeta^*+i\psi_\uparrow^\dagger\psi_\downarrow^\dagger\zeta\right]\right) \tag*{(3)}

where we ignore an irrelevant normalization constant. Our partition function then reads,

Z(T,μ)=∫Dψs†DψsDζe−SE,eff,(4)Z(T,\mu)=\int\mathcal{D}\psi_s^\dagger\mathcal{D}\psi_s\mathcal{D}\zeta e^{-S_{E,\mathrm{eff}}}, \tag*{(4)}
SE,eff=∫τ,x[ψs†(∂τ−∇22m−μ)ψs+iψ↓ψ↑ζ∗−iψ↑†ψ↓†ζ−m4πasζζ∗].S_{E,\mathrm{eff}}=\int_{\tau,\mathbf{x}}\left[\psi_s^\dagger\left(\partial_\tau-\frac{\nabla^2}{2m}-\mu\right)\psi_s+i\psi_\downarrow\psi_\uparrow\zeta^*-i\psi_\uparrow^\dagger\psi_\downarrow^\dagger\zeta-\frac{m}{4\pi a_s}\zeta\zeta^*\right].

where all path integrals are now taken with respect to the effective Euclidean action SE,effS_{E,\mathrm{eff}}.

Thermodynamics

We may repackage the fermion bilinears into matrix form using Nambu–Gor’kov spinors, Ψ=(ψ↑,ψ↓†)\Psi=(\psi_\uparrow,\psi_\downarrow^\dagger), such that the effective Euclidean action becomes,

SE,eff=∫τ,x[Ψ†(∂τ1−∇22mσz−μσz+iζ∗σ−−iζσ+)Ψ−mζζ∗4πas],(5)S_{E,\mathrm{eff}}=\int_{\tau,\mathbf{x}}\left[\Psi^\dagger\left(\partial_\tau\mathbb{1}-\frac{\nabla^2}{2m}\sigma_z-\mu\sigma_z+i\zeta^*\sigma_- -i\zeta\sigma_+\right)\Psi-\frac{m\zeta\zeta^*}{4\pi a_s}\right], \tag*{(5)}

where 1\mathbb{1} is the 2×22\times2 identity matrix, σi\sigma_i denote Pauli matrices, and σ±≡12(σx±iσy)\sigma_\pm\equiv\frac{1}{2}(\sigma_x\pm i\sigma_y). The path integral is now quadratic in fermion fields and can be integrated out exactly,

Z(T,μ)=∫Dζe∫τ,x[mζζ∗4πas+ln⁡(−det⁡G−1)](6)Z(T,\mu)=\int\mathcal{D}\zeta e^{\int_{\tau,\mathbf{x}}\left[\frac{m\zeta\zeta^*}{4\pi a_s}+\ln(-\det G^{-1})\right]} \tag*{(6)}

where G−1(ζ,ζ∗)=(∂τ−∇22m−μ−iζiζ∗∂τ+∇22m+μ)G^{-1}(\zeta,\zeta^*)=\begin{pmatrix}\partial_\tau-\frac{\nabla^2}{2m}-\mu& -i\zeta\\ i\zeta^* & \partial_\tau+\frac{\nabla^2}{2m}+\mu\end{pmatrix} and the minus sign in front of the determinant comes from the Jacobian ∫Dψs†Dψs=−∫DΨ†DΨ\int D\psi_s^\dagger D\psi_s=-\int D\Psi^\dagger D\Psi.

Up to this point we have not made any approximations. To perform the path integral over the ζ\zeta fields we employ the non-perturbative resummation method introduced by Romatschke [25] by expanding the auxiliary field around the global zero mode, ζ(τ,x)=iΔ+ζ′(τ,x)\zeta(\tau,\mathbf{x})=i\Delta+\zeta'(\tau,\mathbf{x}). In the R0R0 approximation (or equivalently the leading large NN expansion [26, 27]) we neglect the field fluctuations ζ′\zeta' and assume the zero mode Δ\Delta is real,

ZR0(T,μ)=∫−∞∞dΔexp⁡[βV(mΔ24πas+T∑ωn∫pln⁡(−det⁡G~−1))](7)Z_{R0}(T,\mu)=\int_{-\infty}^{\infty}d\Delta\exp\left[\beta V\left(\frac{m\Delta^2}{4\pi a_s}+T\sum_{\omega_n}\int_{\mathbf{p}}\ln(-\det\widetilde{G}^{-1})\right)\right] \tag*{(7)}

where the sum is over fermionic Matsubara frequencies ωn=2πT(n+1/2)\omega_n=2\pi T(n+1/2) with n∈Zn\in\mathbb{Z}, VV is the volume of the system, and the inverse propagator in frequency-momentum space is

G~−1(ωn,p,Δ)=(iωn+ϵp−μΔΔiωn−ϵp+μ)(8)\widetilde{G}^{-1}(\omega_n,\mathbf{p},\Delta)=\begin{pmatrix}i\omega_n+\epsilon_{\mathbf{p}}-\mu& \Delta\\ \Delta& i\omega_n-\epsilon_{\mathbf{p}}+\mu\end{pmatrix} \tag*{(8)}

and ϵp=p22m\epsilon_{\mathbf{p}}=\frac{\mathbf{p}^2}{2m}.

In the large volume limit, ZR0Z_{R0} is solved exactly by the saddle point condition with respect to Δ\Delta,

mΔ4πas+T∑ωn∫d3p(2π)3ΔΔ2+(ϵp−μ)2+ωn2=0.(9)\frac{m\Delta}{4\pi a_s}+T\sum_{\omega_n}\int\frac{d^3\mathbf{p}}{(2\pi)^3}\frac{\Delta}{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2+\omega_n^2}=0. \tag*{(9)}

We see that there is the trivial solution Δ=0\Delta=0, but we proceed to look for non-trivial solutions. After performing the sum over fermionic Matsubara frequencies we obtain the condition

−m4πas=12∫d3p(2π)3tanh⁡(Δ2+(ϵp−μ)22T)Δ2+(ϵp−μ)2.(10)-\frac{m}{4\pi a_s}=\frac{1}{2}\int\frac{d^3\mathbf{p}}{(2\pi)^3}\frac{\tanh\left(\frac{\sqrt{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2}}{2T}\right)}{\sqrt{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2}}. \tag*{(10)}

We note here that the poles of the Green’s function, whose inverse is given in (8), define the spectrum located at iωn=Δ2+(ϵp−μ)2i\omega_n=\sqrt{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2}. We see that excitations require a minimum energy input of Δ\Delta which characterizes a gapped spectrum. We refer to (10) as the gap equation and such systems are characteristic of superfluids.

At zero temperature the Matsubara sum becomes an integral, T∑ωn→∫dω2πT\sum_{\omega_n}\to\int\frac{d\omega}{2\pi}, and the pressure is given by

p=ln⁡ZR0βV=mΔ24πas+∫dωd3p(2π)4ln⁡(Δ2+(ϵp−μ)2+ω2).(11)p=\frac{\ln Z_{R0}}{\beta V}=\frac{m\Delta^2}{4\pi a_s}+\int\frac{d\omega d^3\mathbf{p}}{(2\pi)^4}\ln\left(\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2+\omega^2\right). \tag*{(11)}

The argument inside the the logarithm needs to be unitless so we regulate by subtracting ∫dωln⁡ω2\int d\omega\ln\omega^2. This ensures the integral over ω\omega converges since in dimensional regularization ∫dωln⁡ω2=0\int d\omega\ln\omega^2=0. Thus, in dimensional regularization the pressure is

p=mΔ24πas+∫d3p(2π)3Δ2+(ϵp−μ)2.(12)p=\frac{m\Delta^2}{4\pi a_s}+\int\frac{d^3\mathbf{p}}{(2\pi)^3}\sqrt{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2}. \tag*{(12)}

The integral over momenta is now divergent but this is taken care of using a well-known technique in dimensional regularization. We can expand the integrand in (12) as a binomial series

Δ2+(ϵp−μ)2=∑n=0∞(12n)(−2μϵp)n(Δ2+μ2+ϵp2)n−12.(13)\sqrt{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2}=\sum_{n=0}^{\infty}\binom{\frac{1}{2}}{n}\frac{(-2\mu\epsilon_{\mathbf{p}})^n}{(\Delta^2+\mu^2+\epsilon_{\mathbf{p}}^2)^{n-\frac{1}{2}}}. \tag*{(13)}

and each integral is solved in dimensional regularization using [28],

∫dDp(2π)Dϵpα(ϵp2+A2)β2=Γ(D+2α4)Γ(2β−2α−D4)2Γ(D2)Γ(β2)(mA2π)D2Aα−β.(14)\int\frac{d^D\mathbf{p}}{(2\pi)^D}\frac{\epsilon_{\mathbf{p}}^\alpha}{(\epsilon_{\mathbf{p}}^2+A^2)^{\frac{\beta}{2}}}=\frac{\Gamma\left(\frac{D+2\alpha}{4}\right)\Gamma\left(\frac{2\beta-2\alpha-D}{4}\right)}{2\Gamma\left(\frac{D}{2}\right)\Gamma\left(\frac{\beta}{2}\right)}\left(\frac{mA}{2\pi}\right)^{\frac{D}{2}}A^{\alpha-\beta}. \tag*{(14)}

We find that in the limit D→3D\to3,

∫d3p(2π)3Δ2+(ϵp−μ)2=25μ(2mμ)3/23π2g(y),(15)\int\frac{d^3\mathbf{p}}{(2\pi)^3}\sqrt{\Delta^2+(\epsilon_{\mathbf{p}}-\mu)^2}=\frac{2}{5}\frac{\mu(2m\mu)^{3/2}}{3\pi^2}g(y), \tag*{(15)}

where y=μΔ2+μ2y=\frac{\mu}{\sqrt{\Delta^2+\mu^2}}, the function is g(y)=y−5/2[(4y2−3)E(1+y2)+3+y−4y22K(1+y2)]g(y)=y^{-5/2}\left[(4y^2-3)E\left(\frac{1+y}{2}\right)+\frac{3+y-4y^2}{2}K\left(\frac{1+y}{2}\right)\right], and E(x)E(x), K(x)K(x), are the complete elliptic integrals of the first and second kind, respectively. At zero temperature, the non-trivial solution to (10) fulfills

2E(1+y2)−K(1+y2)=πy1/222mμas(16)2E\left(\frac{1+y}{2}\right)-K\left(\frac{1+y}{2}\right)=\frac{\pi y^{1/2}}{2\sqrt{2m\mu}a_s} \tag*{(16)}

One can check that the non-trivial solution to (10) has lower free energy than the trivial solution Δ=0\Delta=0 for all as<0a_s<0. For the unitary fermi gas as→−∞a_s\to-\infty and the gap equation becomes K(1+y2)=2E(1+y2)K\left(\frac{1+y}{2}\right)=2E\left(\frac{1+y}{2}\right) corresponding to the numerical solution y≈0.652y\approx0.652 or Δ≈1.162μ\Delta\approx1.162\mu. It is noted in [24] that for small scattering lengths the solution is close to y≈1y\approx1 and (16) can be expanded around this point to obtain the analytic result

Δ=μe−m2πasν−2+3ln⁡2,∣as∣≪1(17)\Delta=\mu e^{-\frac{m}{2\pi a_s\nu}-2+3\ln2},\qquad|a_s|\ll1 \tag*{(17)}

Lawrence and Romatschke point out that this analytic solution is within ten percent of the numerical solution of (16) for all as<0a_s<0. For the cold unitary Fermi gas,

lim⁡as→−∞Δ=μe−2+3ln⁡2≈1.083μ(18)\lim_{a_s\to-\infty}\Delta=\mu e^{-2+3\ln2}\approx1.083\mu \tag*{(18)}

we see that the R0 approximation gives the correct order of magnitude: Δ/μ=O(1)\Delta/\mu=O(1) [29].

At unitarity, g(y)=y−3/2E(1+y2)g(y)=y^{-3/2}E\left(\frac{1+y}{2}\right) and so the pressure is given by,

p=25μ(2mμ)3/23π2E(1+y2)y3/2∣y=(1+Δ2μ2)−1/2≈2.2pfree(19)p=\left.\frac{2}{5}\frac{\mu(2m\mu)^{3/2}}{3\pi^2}\frac{E\left(\frac{1+y}{2}\right)}{y^{3/2}}\right|_{y=\left(1+\frac{\Delta^2}{\mu^2}\right)^{-1/2}}\approx2.2p_{\mathrm{free}} \tag*{(19)}

where pfree=25μ(2mμ)3/23π2p_{\mathrm{free}}=\frac{2}{5}\frac{\mu(2m\mu)^{3/2}}{3\pi^2} is the pressure of a cold non-interacting fermi gas. The universality of a unitary fermi gas dictates p=ξ−3/2pfreep=\xi^{-3/2}p_{\mathrm{free}} where ξ\xi is a constant known as the Bertsch parameter. In our approximation we find ξR0≈0.59\xi_{\mathrm{R0}}\approx0.59 (which is what one obtains from a mean-field theory calculation [26]) which should be compared to the measured value of ξ≈0.38\xi\approx0.38 [30, 31].

From the pressure we can obtain the number density,

lim⁡as→−∞n(μ)=∂p∂μ=(2mμ)3/23π2ξ−3/2(20)\lim_{a_s\to-\infty}n(\mu)=\frac{\partial p}{\partial\mu}=\frac{(2m\mu)^{3/2}}{3\pi^2}\xi^{-3/2} \tag*{(20)}

which is used to define the Fermi momentum

kF=(3π2n)1/3.(21)k_F=(3\pi^2n)^{1/3}. \tag*{(21)}

Transport Coefficients

Hydrodynamics provides an effective low energy description of physical systems near thermal equilibrium [11]. Hydrodynamics describes the dynamics of conserved macroscopic quantities of an underlying microscopic theory. The conserved hydrodynamic variable TμνT^{\mu\nu} (the energy-momentum tensor) is related to the symmetries of the microscopic theory via Noether’s theorem. Further, we assume that these variables can be expressed as functions of a local temperature T(x)T(x), a local chemical potential μ(x)\mu(x), and a local fluid four-velocity uμ(x)u_\mu(x) normalized to u2=−1u^2=-1.

The energy-momentum tensor can be decomposed in the following way [12],

Tμν=Euμuν+PΔμν+Qμuν+Qνuμ+Tμν(22)T^{\mu\nu}=\mathcal{E}u^\mu u^\nu+\mathcal{P}\Delta^{\mu\nu}+Q^\mu u^\nu+Q^\nu u^\mu+\mathcal{T}^{\mu\nu} \tag*{(22)}

where Δμν=gμν+uμuν\Delta_{\mu\nu}=g_{\mu\nu}+u_\mu u_\nu is a projector transverse to the fluid four-velocity (e.g. uμΔμν=0u^\mu\Delta_{\mu\nu}=0), E=uμuνTμν\mathcal{E}=u_\mu u_\nu T^{\mu\nu} is the energy density, P=13ΔμνTμν\mathcal{P}=\frac{1}{3}\Delta_{\mu\nu}T^{\mu\nu} is the pressure, the energy flux Qμ=−ΔμαTαβuβQ^\mu=-\Delta^{\mu\alpha}T_{\alpha\beta}u^\beta is transverse to uμu_\mu, and the stress Tμν=T⟨μν⟩\mathcal{T}^{\mu\nu}=T^{\langle\mu\nu\rangle} is transverse to uμu_\mu, symmetric, and traceless. The angled brackets denote the symmetric transverse traceless part of a tensor, X⟨μν⟩=12(ΔμαΔνβ+ΔμβΔνα−23ΔμνΔαβ)XαβX^{\langle\mu\nu\rangle}=\frac{1}{2}\left(\Delta^{\mu\alpha}\Delta^{\nu\beta}+\Delta^{\mu\beta}\Delta^{\nu\alpha}-\frac{2}{3}\Delta^{\mu\nu}\Delta^{\alpha\beta}\right)X_{\alpha\beta}.

Hydrodynamics assumes that the quantities E,P,Qμ\mathcal{E},\mathcal{P},Q^\mu, and Tμν\mathcal{T}^{\mu\nu} can be expanded in terms of the hydrodynamic variables uμu^\mu, TT, and μ\mu and their derivatives. These expressions are called the constitutive relations [12]. Gradients of hydrodynamic variables correspond to perturbations out of equilibrium and each term is coupled to a quantity called a transport coefficient. Transport coefficients dictate the real-time evolution for equilibrium to be restored, familiar examples being conductivities, viscosities, and diffusion coefficients. These well-known quantities, however, control the linear (first-order) response of the system to a perturbation whereas non-linear contributions (second-order, third-order, etc.) will also be present.

In general, transport coefficients of a strongly interacting quantum many-body system are difficult to calculate. However, for a particular subset of second-order transport coefficients, so-called thermodynamic transport coefficients, Kovtun & Shukla showed such quantities are considerably easier to calculate from two-point correlation functions in flat space without the need to analytically continue from Euclidean to Minkowski spacetime [10]. These transport coefficients are non-dissipative in nature and exist even in the limit of thermal equilibrium [17]. As first noted in [14, 32], these non-dissipative transport coefficients follow from an equilibrium generating functional which itself admits a derivative expansion, with various thermodynamic susceptibilities appearing as coefficients of different terms in the expansion. The variation of the equilibrium generating functional with respect to external sources produces the constitutive relations for the conserved currents in the system, which contain various transport coefficients that do not vanish in the equilibrium limit. Thus, these second-order thermodynamic transport coefficients are linear combinations of the thermodynamic susceptibilities. So we proceed to calculate susceptibilities and use the dictionary provided in [10] to obtain transport coefficients. Susceptibilities are denoted as fif_i, i={1,2,3,…}i=\{1,2,3,\ldots\} and the constitutive relations are,

E=ϵ+(f1′−f1)R+(4f1′+2f1′′−f2−f2′)a2+(f1′−f2−3f3+f3′)Ω2−2(f1+f1′−f2)uαRαβuβ,(23)\begin{aligned} \mathcal{E} &= \epsilon+(f'_1-f_1)R +(4f'_1+2f''_1-f_2-f'_2)a^2 \\ &\quad+(f'_1-f_2-3f_3+f'_3)\Omega^2-2(f_1+f'_1-f_2)u^\alpha R_{\alpha\beta}u^\beta, \tag*{(23)} \end{aligned}
P=p+13f1R−13(2f1′+f3)Ω2−13(2f1′+4f1′′−f2)a2+23(2f1′−f1)uαRαβuβ,(24)\mathcal{P}=p+\frac{1}{3}f_1R-\frac{1}{3}(2f'_1+f_3)\Omega^2-\frac{1}{3}(2f'_1+4f''_1-f_2)a^2+\frac{2}{3}(2f'_1-f_1)u^\alpha R_{\alpha\beta}u^\beta, \tag*{(24)}
Qμ=(f1′+2f3′)ϵμλρσaλuρΩσ+(2f1+4f3)Δμ ρRρσuσ,(25)Q_\mu=(f'_1+2f'_3)\epsilon_{\mu\lambda\rho\sigma}a^\lambda u^\rho\Omega^\sigma+(2f_1+4f_3)\Delta_\mu^{\ \rho}R_{\rho\sigma}u^\sigma, \tag*{(25)}
Tμν=(4f1′+2f1′′−2f2)a⟨μaν⟩−12(f1′−4f3)Ω⟨μΩν⟩+2f1′uαRα⟨μν⟩βuβ−2f1R⟨μν⟩,(26)\mathcal{T}^{\mu\nu}=(4f'_1+2f''_1-2f_2)a_{\langle\mu}a_{\nu\rangle}-\frac{1}{2}(f'_1-4f_3)\Omega_{\langle\mu}\Omega_{\nu\rangle}+2f'_1u^\alpha R_{\alpha\langle\mu\nu\rangle\beta}u^\beta-2f_1R_{\langle\mu\nu\rangle}, \tag*{(26)}

where

fn′=Tfn,T+μfn,μf'_n=T f_{n,T}+\mu f_{n,\mu}
fn′′=T2fn,T,T+2μTfn,T,μ+μ2fn,μ,μf''_n=T^2f_{n,T,T}+2\mu T f_{n,T,\mu}+\mu^2f_{n,\mu,\mu}

and the comma denotes the partial derivative with respect to the argument that follows. Terms appearing in 23 include the Riemann scalar RR, Riemann tensor RμνR^{\mu\nu}, acceleration aμ=uλ∇λuμa^\mu=u^\lambda\nabla_\lambda u^\mu, and vorticity Ωμ≡ϵμναβuν∇αuβ\Omega^\mu\equiv\epsilon^{\mu\nu\alpha\beta}u_\nu\nabla_\alpha u_\beta.1

We set out to calculate f3f_3 given by [10],

f3=14lim⁡k→0∂2∂k2(−⟨TE01TE01⟩(k)+⟨TE12TE12⟩(k)).(27)f_3=\frac{1}{4}\lim_{k\to0}\frac{\partial^2}{\partial k^2}\left(-\langle T_E^{01}T_E^{01}\rangle(\mathbf{k})+\langle T_E^{12}T_E^{12}\rangle(\mathbf{k})\right). \tag*{(27)}

where TEμνT_E^{\mu\nu} is the energy-momentum tensor of the underlying quantum system in Euclidean space. Without loss of generality we assume the external momentum is oriented in the zz-direction, k⃗=kz^\vec{k}=k\hat{z}. We wish to calculate the Euclidean correlator

CE(ω,kz^)=−⟨TE01TE01⟩(ω,kz^)+⟨TE12TE12⟩(ω,kz^).(28)C_E(\omega,k\hat{z})=-\langle T_E^{01}T_E^{01}\rangle(\omega,k\hat{z})+\langle T_E^{12}T_E^{12}\rangle(\omega,k\hat{z}). \tag*{(28)}

In particular, we will calculate

f3=14lim⁡k→0∂2∂k2CE(ωn=0,k)∣k=0(29)f_3=\frac{1}{4}\lim_{k\to0}\left.\frac{\partial^2}{\partial k^2}C_E(\omega_n=0,k)\right|_{k=0} \tag*{(29)}

Stress-Tensor Correlators in the Cold Superfluid Fermi Gas

The Euclidean stress-tensor two-point correlator in the path integral formulation is

CE(x−y)=Z−1∫Dψ†Dψe−SE[−TE01(x)TE01(y)+TE12(x)TE12(y)].(30)C_E(x-y)=Z^{-1}\int\mathcal{D}\psi^\dagger\mathcal{D}\psi e^{-S_E}\left[-T_E^{01}(x)T_E^{01}(y)+T_E^{12}(x)T_E^{12}(y)\right]. \tag*{(30)}

We obtain TEμν(x)T_E^{\mu\nu}(x) by starting with the energy-momentum tensor of relativistic Dirac fermions and taking the non-relativistic limit (e.g. where the energy mm is much higher than any other energy scale in our system). We find

TE01(x)=14[ψ↑†(∂1ψ↑)+ψ↓†(∂1ψ↓)−(∂1ψ↑†)ψ↑−(∂1ψ↓†)ψ↓]+O(1/m)(31)T_E^{01}(x)=\frac{1}{4}\left[\psi_\uparrow^\dagger(\partial_1\psi_\uparrow)+\psi_\downarrow^\dagger(\partial_1\psi_\downarrow)-(\partial_1\psi_\uparrow^\dagger)\psi_\uparrow-(\partial_1\psi_\downarrow^\dagger)\psi_\downarrow\right]+\mathcal{O}(1/m) \tag*{(31)}

and TE12(x)∝1mT_E^{12}(x)\propto\frac{1}{m}, where we ignore all O(1/m)\mathcal{O}(1/m) and higher terms. Expressing the stress-tensor component using Nambu-Gor'kov spinors, Ψ=(ψ↑ψ↓†)\Psi=\begin{pmatrix}\psi_\uparrow\\\psi_\downarrow^\dagger\end{pmatrix}, our path integral becomes straight-forward to calculate. Since a vanishing total divergence of fields implies (∂1ψs†)ψs=−ψs†(∂1ψs)(\partial_1\psi_s^\dagger)\psi_s=-\psi_s^\dagger(\partial_1\psi_s), TE01T_E^{01} simplifies to

TE01(x)=12Ψ†(x)∂1Ψ(x)+O(1/m).(32)T_E^{01}(x)=\frac{1}{2}\Psi^\dagger(x)\partial_1\Psi(x)+\mathcal{O}(1/m). \tag*{(32)}

Using the same auxiliary fields as in (4) and to leading order in 1/m1/m, our path integral is

CE(x−y)=14Z∫DΨ†DΨDζ Tr⁡[∂1yG(y−x)∂1xG(x−y)](33)C_E(x-y)=\frac{1}{4Z}\int\mathcal{D}\Psi^\dagger\mathcal{D}\Psi\mathcal{D}\zeta\ \operatorname{Tr}\left[\partial_1^yG(y-x)\partial_1^xG(x-y)\right] \tag*{(33)}

where the trace is over spinor indices and

⟨Ψ(x)Ψ†(y)⟩=G(x−y).(34)\left\langle\Psi(x)\Psi^\dagger(y)\right\rangle= G(x-y). \tag*{(34)}

In the R0 approximation we restrict the auxiliary fields to only the global zero mode. The R0-propagator from (8) is

G(ωn,p,Δ)=1(ϵp−μ)2+ωn2+Δ2(iωn+ϵp−μΔΔiωn−ϵp+μ)(35)G(\omega_n,\mathbf{p},\Delta)=\frac{1}{(\epsilon_p-\mu)^2+\omega_n^2+\Delta^2} \begin{pmatrix} i\omega_n+\epsilon_p-\mu& \Delta\\ \Delta& i\omega_n-\epsilon_p+\mu \end{pmatrix} \tag*{(35)}

and the stress-tensor correlator in frequency-momentum space CE(kE,k)C_E(k_E,\mathbf{k}) at zero temperature is

CE(kE,k)=−∫−∞∞dΔeβVp4Z∫d4p(2π)4p12Tr⁡[G(kE+ω,k+p)G(ω,p)](36)C_E(k_E,\mathbf{k})=-\int_{-\infty}^{\infty}d\Delta\frac{e^{\beta V p}}{4Z}\int\frac{d^4p}{(2\pi)^4}p_1^2\operatorname{Tr}\left[G(k_E+\omega,\mathbf{k}+\mathbf{p})G(\omega,\mathbf{p})\right] \tag*{(36)}

where ∫dΔeβVp=ZR0\int d\Delta e^{\beta V p}=Z_{R0} and the integral over Δ\Delta will once again restrict its value to the saddle point condition approximated in (18). Upon performing the trace over spinor components and restricting to zero external frequency kE=0k_E=0, our correlator is

CE(kE=0,k)=−12∫dω2πd3p(2π)3p12(ϵk+p−μ)(ϵp−μ)+Δ2−ω2[(ϵk+p−μ)2+Δ2+ω2][(ϵp−μ)2+Δ2+ω2](37)C_E(k_E=0,\mathbf{k})=-\frac{1}{2}\int\frac{d\omega}{2\pi}\frac{d^3p}{(2\pi)^3}p_1^2\frac{(\epsilon_{\mathbf{k}+\mathbf{p}}-\mu)(\epsilon_p-\mu)+\Delta^2-\omega^2}{\left[(\epsilon_{\mathbf{k}+\mathbf{p}}-\mu)^2+\Delta^2+\omega^2\right]\left[(\epsilon_p-\mu)^2+\Delta^2+\omega^2\right]} \tag*{(37)}

After lim⁡k→0∂2∂k2CE\lim_{\mathbf{k}\to0}\frac{\partial^2}{\partial k^2}C_E and averaging over angles,

CE′′(K=0)=−∫P{32ϵp2(ϵp−μ)415[Δ2+(ϵp−μ)2+ω2]4−4ϵp(13ϵp−5μ)(ϵp−μ)215[Δ2+(ϵp−μ)2+ω2]3+ϵp(19ϵp−15μ)15[Δ2+(ϵp−μ)2+ω2]2}−Δ2∫P{32ϵp2(ϵp−μ)215[Δ2+(ϵp−μ)2+ω2]4−4ϵp(7ϵp−5μ)15[Δ2+(ϵp−μ)2+ω2]3}(38)C_E^{\prime\prime}(K=0)=-\int_P\left\{\frac{32\epsilon_p^2(\epsilon_p-\mu)^4}{15\left[\Delta^2+(\epsilon_p-\mu)^2+\omega^2\right]^4}-\frac{4\epsilon_p(13\epsilon_p-5\mu)(\epsilon_p-\mu)^2}{15\left[\Delta^2+(\epsilon_p-\mu)^2+\omega^2\right]^3}+\frac{\epsilon_p(19\epsilon_p-15\mu)}{15\left[\Delta^2+(\epsilon_p-\mu)^2+\omega^2\right]^2}\right\} -\Delta^2\int_P\left\{\frac{32\epsilon_p^2(\epsilon_p-\mu)^2}{15\left[\Delta^2+(\epsilon_p-\mu)^2+\omega^2\right]^4}-\frac{4\epsilon_p(7\epsilon_p-5\mu)}{15\left[\Delta^2+(\epsilon_p-\mu)^2+\omega^2\right]^3}\right\} \tag*{(38)}
≡CEa′′+CEb′′(39)\equiv C_E^{a^{\prime\prime}}+C_E^{b^{\prime\prime}} \tag*{(39)}

where the top line of (38) is CEa′′C_E^{a^{\prime\prime}} and the bottom line proportional to Δ2\Delta^2 is CEb′′C_E^{b^{\prime\prime}}. Note that the integral in the bottom line of (38) is finite so that CEb′′C_E^{b^{\prime\prime}} vanishes in the limit Δ→0\Delta\to0. We can rewrite the denominator of each term as

1[Δ2+(ϵp−μ)2+ω2]n=(−1)n−1(n−1)!∂n−1∂(Δ2)n−11Δ2+(ϵp−μ)2+ω2(40)\frac{1}{\left[\Delta^2+(\epsilon_p-\mu)^2+\omega^2\right]^n} = \frac{(-1)^{n-1}}{(n-1)!}\frac{\partial^{n-1}}{\partial(\Delta^2)^{n-1}} \frac{1}{\Delta^2+(\epsilon_p-\mu)^2+\omega^2} \tag*{(40)}

for integers n≥1n \ge1 so that we are left with an integral over ω\omega with simple poles at ±iΔ2+(ϵp−μ)2\pm i\sqrt{\Delta^2+(\epsilon_p-\mu)^2}. The next step is to integrate over momenta utilizing equations (13) and (14) which, in the unitary limit as→−∞a_s \to-\infty, leads to

CEa′′(K=0)=−(2mμ)3/248π2μE(1+y2)y3/2y2=−n(μ)16μy2(41)C_E^{a''}(K=0)=-\frac{(2m\mu)^{3/2}}{48\pi^2\mu}\frac{E\left(\frac{1+y}{2}\right)}{y^{3/2}}y^2=-\frac{n(\mu)}{16\mu}y^2 \tag*{(41)}
CEb′′(K=0)=(2mμ)3/2720π2μE(1+y2)y3/28μ2+23Δ2Δ2y2=n(μ)16μ8μ2+23Δ215Δ2y2(42)C_E^{b''}(K=0)=\frac{(2m\mu)^{3/2}}{720\pi^2\mu}\frac{E\left(\frac{1+y}{2}\right)}{y^{3/2}}\frac{8\mu^2+23\Delta^2}{\Delta^2}y^2=\frac{n(\mu)}{16\mu}\frac{8\mu^2+23\Delta^2}{15\Delta^2}y^2 \tag*{(42)}
CE′′(K=0)=n(μ)30μ(μ2Δ2)(43)C_E^{\prime\prime}(K=0)=\frac{n(\mu)}{30\mu}\left(\frac{\mu^2}{\Delta^2}\right) \tag*{(43)}
f3=14CE′′(K=0)=n(μ)120μ(μ2Δ2).(44)f_3=\frac{1}{4}C_E^{\prime\prime}(K=0)=\frac{n(\mu)}{120\mu}\left(\frac{\mu^2}{\Delta^2}\right). \tag*{(44)}

This is our final result for the thermodynamic susceptibility f3f_3. Second order thermodynamic transport coefficients are linear combinations of second order thermodynamic susceptibilities and their derivatives. These transport coefficients, as formulated in [11, 13], and their relation to the susceptibilities are given in [10]. For example, the second order transport coefficient λ3\lambda_3 couples to the vorticity and, to leading order in the UFG, is given by λ3≈−8f3\lambda_3 \approx-8f_3.

Conclusion

The unitarity Fermi gas is an important example of a strongly correlated quantum fluid. Measurements of both equilibrium and non-equilibrium properties serves as a crucial reference point for a diverse range of physical systems, spanning from the study of dilute neutron matter in neutron stars to the investigation of the quark gluon plasma in relativistic heavy ion collisions.

Neutron star matter cannot be recreated in the lab and cannot be studied using lattice gauge theory because of the sign problem. Our result, equation (44), is an analytic solution from first principles and attempts to shed light on these issues. Because thermodynamic transport coefficients (which are linear combinations of thermodynamic susceptibilities as illustrated in equations (23) do not require analytic continuation from Euclidean to Minkowski spacetime, our calculation is amenable to lattice calculations. Such calculations could further our understanding of such transport properties in NS crusts. Ultracold atomic experiments are known for their extraordinarily precise measurements. Since they are capable of creating UFGs in the lab, perhaps there is (or will be in the future) a means of measuring second-order transport coefficients.

It has been suggested that superfluid vortices are relevant to NS glitches. Since f3f_3 couples to the vorticity (at second order) further study as to its relation to NS glitches might be worthwhile, although the relation between vorticity and superfluid vortices is unclear.

Our results may be of assistance to those studying the NS equation of state (EoS). The authors of [33] propose a “unitary gas conjecture” which states that the energy of a UFG serves as a lower bound for pure neutron matter. This conjecture imposes constraints on EoS parameters.

There are six other thermodynamic susceptibilities that can be calculated for the unitary fermi gas [17], one of which has already been calculated [24]. Calculating the rest of these transport coefficients would be the logical next step to this paper. The susceptibilities f1f_1, f2f_2, and f3f_3 characterize the constitutive relations in equations (23), where f2f_2 has not been calculated for the UFG. The reason for this is that the two-point correlation functions used to determine f2f_2 consist of diagonal components of TμνT^{\mu\nu} which make it much more complicated. We leave this for future work.

Acknowledgments

This work was supported by the Department of Energy, DOE award No DE-SC0017905.

I would like to thank Paul Romatschke for suggesting such an exciting and rich research topic as well as the enormous amount of help to make this work possible.

Free and Non-Relativistic Limit

We compare our result in the free limit (Δ→0\Delta\to0) with the non-relativistic result in [17]. In that paper, Shukla calculated all thermodynamic transport coefficients for a relativistic system of free fermions. There he found,

f3=−∣μr∣96π2μr2−m2(45)f_3=-\frac{|\mu_r|}{96\pi^2}\sqrt{\mu_r^2-m^2} \tag*{(45)}

where we distinguish between the relativistic chemical potential μr\mu_r and the non-relativistic chemical potential used in this paper, μ\mu. In the non-relativistic limit, the chemical potential is only slightly larger than the mass,

μr≈m+μ(46)\mu_r \approx m+\mu \tag*{(46)}

where μ≪m\mu\ll m. We can then expand f3f_3 around μ=0\mu=0,

f3≈m3/2μ482π2=−(2mμ)3/2192π2μ=−nfree64μ(47)f_3 \approx\frac{m^{3/2}\sqrt{\mu}}{48\sqrt{2}\pi^2}=-\frac{(2m\mu)^{3/2}}{192\pi^2\mu}=-\frac{n_{\mathrm{free}}}{64\mu} \tag*{(47)}

where nfree=(2mμ)3/23π2n_{\mathrm{free}}=\frac{(2m\mu)^{3/2}}{3\pi^2} is the density of a degenerate Fermi gas. Looking at Eqns (41) & (44) and noting that lim⁡Δ→0\lim_{\Delta\to0} is equivalent to lim⁡y→1\lim_{y\to1},

f3=14lim⁡Δ→0CEa′′(K=0)=lim⁡y→1(2mμ)3/2192π2μE(1+y2)y3/2=−(2mμ)3/2192π2μ(48)f_3=\frac{1}{4}\lim_{\Delta\to0}C_{E}^{a''}(K=0)=\lim_{y\to1}\frac{(2m\mu)^{3/2}}{192\pi^2\mu}\frac{E\left(\frac{1+y}{2}\right)}{y^{3/2}}=-\frac{(2m\mu)^{3/2}}{192\pi^2\mu} \tag*{(48)}

which agrees with the non-relativistic limit obtained in [17]. In order that our results agree with free fermions in the non-relativistic limit, we must also show that lim⁡Δ→0CEb′′(K=0)=0\lim_{\Delta\to0}C_{E}^{b''}(K=0)=0.

To this end, we must look to [34, 35] which conclude calculations (at zero temperature and finite chemical potential) that involve integrals over momenta and frequencies do not commute. That is, the physically correct results are obtained by integrating over momenta first, and not vice versa as is commonly done. To this point we also note that when one sets Δ=0\Delta=0 at the outset and goes on to calculate ⟨T01T01⟩\langle T^{01}T^{01}\rangle to leading order, integrating over frequencies first produces zero, which disagrees with [17]. It is straightforward from eqn (38) that lim⁡Δ→0CEb′′(K=0)=0\lim_{\Delta\to0}C_{E}^{b''}(K=0)=0 since it is the product of a finite integral and Δ2\Delta^2, so it must vanish in this limit.

References

References

  1. [1]A. Hewish, S. J. Bell, J. D. H. Pilkington, P. F. Scott, and R. A. Collins, Observation of a Rapidly Pulsating Radio Source, Nature (London) 217, 709 (1968).
  2. [2]R. N. Manchester, G. B. Hobbs, A. Teoh, and M. Hobbs, The australia telescope national facility pulsar catalogue, The Astronomical Journal 129, 1993 (2005).DOI
  3. [3]Q. Chen, C. R. Howell, T. S. Carman, W. R. Gibbs, B. F. Gibson, A. Hussein, M. R. Kiser, G. Mertens, C. F. Moore, C. Morris, A. Obst, E. Pasyuk, C. D. Roper, F. Salinas, H. R. Setze, I. Slaus, S. Sterbenz, W. Tornow, R. L. Walter, C. R. Whiteley, and M. Whitton, Measurement of the neutron-neutron scattering length using the π⁻d capture reaction, Phys. Rev. C 77, 054002 (2008).DOI
  4. [4]D. E. González Trotter, F. Salinas, Q. Chen, A. S. Crowell, W. Glöckle, C. R. Howell, C. D. Roper, D. Schmidt, I. Šlaus, H. Tang, W. Tornow, R. L. Walter, H. Witala, and Z. Zhou, New Measurement of the ¹S₀ Neutron-Neutron Scattering Length Using the Neutron-Proton Scattering Length as a Standard, Phys. Rev. Lett. 83, 3788 (1999), arXiv:nucl-ex/9904011 [nucl-ex].DOI
  5. [5]W. Ketterle and M. W. Zwierlein, Making, probing and understanding ultracold Fermi gases, Nuovo Cimento Rivista Serie 31, 247 (2008), arXiv:0801.2500 [cond-mat.other].DOI
  6. [6]V. Gurarie and L. Radzihovsky, Resonantly paired fermionic superfluids, Annals of Physics 322, 2 (2007), arXiv:cond-mat/0611022 [cond-mat.supr-con].DOI
  7. [7]L. D. Landau and E. M. Lifshitz, Fluid mechanics (1959).
  8. [8]T. Schäfer and D. Teaney, Nearly perfect fluidity: from cold atomic gases to hot quark gluon plasmas, Reports on Progress in Physics 72, 126001 (2009), arXiv:0904.3107 [hep-ph].
  9. [9]P. K. Kovtun, D. T. Son, and A. O. Starinets, Viscosity in Strongly Interacting Quantum Field Theories from Black Hole Physics, Phys. Rev. Lett. 94, 111601 (2005), arXiv:hep-th/0405231 [hep-th].pubmed.ncbi.nlm.nih.gov/15903845
  10. [10]P. Kovtun and A. Shukla, Kubo formulas for thermodynamic transport coefficients, Journal of High Energy Physics 2018, 7 (2018), arXiv:1806.05774 [hep-th].DOI
  11. [11]P. Romatschke and U. Romatschke, Relativistic Fluid Dynamics In and Out of Equilibrium, Cambridge Monographs on Mathematical Physics (Cambridge University Press, 2019) P. Kovtun, Lectures on hydrodynamic fluctuations in relativistic theories, Journal of Physics A Mathematical General 45, 473001 (2012), arXiv:1205.5040 [hep-th].
  12. [13]R. Baier, P. Romatschke, D. Thanh Son, A. O. Starinets, and M. A. Stephanov, Relativistic viscous hydrodynamics, conformal invariance, and holography, Journal of High Energy Physics 2008, 100 (2008), arXiv:0712.2451 [hep-th].DOI
  13. [14]K. Jensen, M. Kaminski, P. Kovtun, R. Meyer, A. Ritz, and A. Yaron, Towards Hydrodynamics without an Entropy Current, Phys. Rev. Lett. 109, 101601 (2012), arXiv:1203.3556 [hep-th].DOI
  14. [15]P. Romatschke and D. T. Son, Spectral sum rules for the quark-gluon plasma, Phys. Rev. D 80, 065021 (2009), arXiv:0903.3946 [hep-ph].DOI
  15. [16]G. D. Moore and K. A. Sohrabi, Thermodynamical second-order hydrodynamic coefficients, Journal of High Energy Physics 2012, 148 (2012), arXiv:1210.3340 [hep-ph].DOI
  16. [17]A. Shukla, Equilibrium thermodynamic susceptibilities for a dense degenerate Dirac field, Phys. Rev. D 100, 096010 (2019), arXiv:1906.02334 [hep-th].DOI
  17. [18]S. Bhattacharyya, S. Minwalla, V. E. Hubeny, and M. Rangamani, Nonlinear fluid dynamics from gravity, Journal of High Energy Physics 2008, 045 (2008), arXiv:0712.2456 [hep-th].DOI
  18. [19]S. I. Finazzo, R. Rougemont, H. Marrochio, and J. Noronha, Hydrodynamic transport coefficients for the non-conformal quark-gluon plasma from holography, Journal of High Energy Physics 2015, 51 (2015), arXiv:1412.2968 [hep-ph].DOI
  19. [20]S. Grozdanov and A. O. Starinets, On the universal identity in second order hydrodynamics, Journal of High Energy Physics 2015, 7 (2015), arXiv:1412.5685 [hep-th].DOI
  20. [21]S. Grieninger and A. Shukla, Second order equilibrium transport in strongly coupled N = 4 supersymmetric SU(Nc) Yang-Mills plasma via holography, Journal of High Energy Physics 2021, 108 (2021), arXiv:2105.08673 [hep-th].DOI
  21. [22]P. Romatschke, Analytic transport from weak to strong coupling in the O (N ) model, Phys. Rev. D 100, 054029 (2019), arXiv:1905.09290 [hep-th].DOI
  22. [23]M. Weiner and P. Romatschke, Determining all thermodynamic transport coefficients for an interacting large (2022), arXiv:2208.10502.DOI
  23. [24]S. Lawrence and P. Romatschke, Gravitational-wave-to-matter coupling of superfluid Fermi gases near unitarity, Phys. Rev. A 107, 033327 (2023), arXiv:2206.04765 [cond-mat.str-el.]. P. Romatschke, Simple non-perturbative resummation schemes beyond mean-field: case study for scalar ϕ⁴ theory in 1+1 dimensions, Journal of High Energy Physics 2019, 149 (2019), arXiv:1901.05483 [hep-th].
  24. [26]M. Y. Veillette, D. E. Sheehy, and L. Radzihovsky, Large- N expansion for unitary superfluid Fermi gases, Phys. Rev. A 75, 043614 (2007), arXiv:cond-mat/0610798 [cond-mat.other].DOI
  25. [27]P. Nikolić and S. Sachdev, Renormalization-group fixed points, universal phase diagram, and 1/N expansion for quantum liquids with interactions near the unitarity limit, Phys. Rev. A 75, 033608 (2007), arXiv:cond-mat/0609106 [cond-mat.supr-con].DOI
  26. [28]Y. Nishida and D. T. Son, Fermi gas near unitarity around four and two spatial dimensions, Phys. Rev. A 75, 063617 (2007), arXiv:cond-mat/0607835 [cond-mat.other].DOI
  27. [29]S. Gandolfi, A. Gezerlis, and J. Carlson, Neutron Matter from Low to High Density, Annual Review of Nuclear and Particle Science 65, 303 (2015), arXiv:1501.05675 [nucl-th].DOI
  28. [30]M. J. H. Ku, A. T. Sommer, L. W. Cheuk, and M. W. Zwierlein, Revealing the Superfluid Lambda Transition in the Universal Thermodynamics of a Unitary Fermi Gas, Science 335, 563 (2012), arXiv:1110.3309 [cond-mat.quant-gas].DOI
  29. [31]M. G. Endres, D. B. Kaplan, J.-W. Lee, and A. N. Nicholson, Lattice Monte Carlo calculations for unitary fermions in a finite box, Phys. Rev. A 87, 023615 (2013), arXiv:1203.3169 [hep-lat].DOI
  30. [32]N. Banerjee, J. Bhattacharya, S. Bhattacharyya, S. Jain, S. Minwalla, and T. Sharma, Constraints on fluid dynamics from equilibrium partition functions, Journal of High Energy Physics 2012, 46 (2012), arXiv:1203.3544 [hep-th].DOI
  31. [33]I. Tews, J. M. Lattimer, A. Ohnishi, and E. E. Kolomeitsev, Symmetry Parameter Constraints from a Lower Bound on Neutron-matter Energy, Astrophys. J. 848, 105 (2017), arXiv:1611.07133 [nucl-th].DOI
  32. [34]T. Gorda, J. Österman, and S. Säppi, Augmenting the residue theorem with boundary terms in finite-density calculations, Phys. Rev. D 106, 105026 (2022), arXiv:2208.14479 [hep-th].DOI
  33. [35]J. Österman, P. Schicho, and A. Vuorinen, Integrating by parts at finite density, arXiv e-prints, arXiv:2304.05427 (2023), arXiv:2304.05427 [hep-ph].DOI

Paper details

vlnc:W_6jfym5vu
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