Conductivity of a Non-Galilean–Invariant Fermi Liquid: Exact Solution of the Kinetic Equation
Tatia Kiliptari, Vladimir I. Yudson, Dmitrii L. Maslov · 2026
Read Conductivity of a Non-Galilean–Invariant Fermi Liquid: Exact Solution of the Kinetic Equation by Tatia Kiliptari, Vladimir I. Yudson, Dmitrii L. Maslov (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-2605-21774.txt.
Introduction
Tatia Kiliptari, Vladimir Yudson, and Dmitrii Maslov solve for the conductivity of a non-Galilean Fermi liquid. They solve the kinetic equation with Coulomb and z-equals-three Pomeranchuk forces. Electron-electron forces act via the quasiparticle scattering time. The collisionless-hydrodynamic shift occurs when its inverse tops the impurity rate or the probe frequency. The result gives the optical response in that regime. Near such a critical point, kinetic and Kubo methods agree only with mass renormalization in the Eliashberg approach, placing the optical shift at the Planckian scale.
Catalog record
| CREATOR | Tatia Kiliptari, Vladimir I. Yudson, Dmitrii L. Maslov |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39178 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.21774
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Conductivity of a Non-Galilean–Invariant Fermi Liquid: Exact Solution of the Kinetic Equation Tatia Kiliptari, Vladimir I. Yudson, Dmitrii L. Maslov Abstract We obtain an exact expression for the conductivity of a non-Galilean-invariant Fermi liquid with impurities by solving the kinetic equation with both screened Coulomb and z = 3 z=3 Pomeranchuk critical interactions. While consistent with previous asymptotic results, our solution shows that electron-electron interactions enter the conductivity solely via the quasiparticle scattering time, τ ee \tau_{\mathrm{ee}} . Accordingly, the crossovers between the collisionless and hydrodynamic regimes occur when 1 / τ ee 1/\tau_{\mathrm{ee}} becomes comparable to the larger of the impurity scattering rate and the probe frequency, Ω \Omega . In addition, the exact solution yields the optical response in the hydrodynamic regime, Ω ≪ 1 / τ ee \Omega\ll 1/\tau_{\mathrm{ee}} , which is inaccessible within perturbation theory. Near a z = 3 z=3 Pomeranchuk quantum critical point, consistency between the kinetic-equation and Kubo approaches requires proper inclusion of mass renormalization within the Eliashberg approximation, which also ensures that the crossover between the collisionless and hydrodynamic regimes in the optical conductivity occurs at the Planckian scale Ω ∼ T \Omega\sim T . Introduction. The temperature ( T ) and frequency ( \Omega ) dependencies of the electronic conductivity provide key insights into the scattering mechanisms that govern electron dynamics. Traditionally, a Fermi liquid (FL) is expected to exhibit a T^{2} dc resistivity and Gurzhi-like optical conductivity, \mathrm{Re}\,\sigma(\Omega,T)\propto(\Omega^{2}+4\pi^{2}T^{2})/\Omega^{2} 12 , while the deviations are interpreted as signatures of non-FL physics. However, the T^{2} resistivity requires either umklapp scattering 18 or compensated electron-hole pockets 3 . If these conditions are not met, normal electron-electron ( ee ) collisions still affect the conductivity of a non-Galilean-invariant FL, provided that momentum is relaxed by another mechanism, e.g., impurity scattering. However, the ee contribution to the conductivity exhibits non-canonical scaling with T and \Omega . In the low-temperature limit, the dc resistivity of a isotropic, non-Galilean–invariant FL with impurities is given by \rho(T)=\rho_{\rm i}+\delta\rho(T) , where \rho_{\rm i} is the residual resistivity and \delta\rho(T)\propto T^{4} (up to a logarithmic factor in two dimensions) is independent of impurity scattering 24 ; 28 ; 17 . This form appears to obey Matthiessen’s rule and thus suggests a suppressed “current relaxation rate” due to ee interactions, 1/\tau_{J}\propto T^{4}\ll 1/\tau_{\mathrm{ee}}\propto T^{2} , where \tau_{\mathrm{ee}} is the quasiparticle scattering time. The collisionless limit of the optical conductivity \mathrm{Re}\,\sigma(\Omega,T)\propto\max\{\Omega^{4},T^{4}\}/\Omega^{2} 31 ; 32 ; 33 ; 11 , reinforces this interpretation: viewed as a Drude tail, it implies that 1/\tau_{J}\propto\max\{\Omega^{4},T^{4}\} . The same interpretation would place the low- T /high- T crossover at \delta\rho(T)\sim\rho_{\rm i} , implying an \mathcal{O}(1) ee contribution to the resistivity, large enough to attribute sizable experimental variations of \rho to normal ee collisions. In this Letter, we obtain an exact result for the conductivity by solving the kinetic equation both at zero and finite \Omega . It shows that the low- T /high- T crossover in the dc resistivity and the collisionless/hydrodynamic crossover in the optical conductivity are controlled by \tau_{\mathrm{ee}}/\tau_{\rm i} and \Omega\tau_{\mathrm{ee}} , respectively, where \tau_{\rm i} is the impurity scattering time. Therefore, interpreting the previously known asymptotic results in terms of a current relaxation rate is misleading. In particular, in the dc case we find that \delta\rho(T)\ll\rho_{\rm i} throughout the temperature range of the Fermi-liquid regime. Model. As a concrete realization of a non-Galilean–invariant FL, we consider a 2D system with a linear dispersion and an energy-independent \tau_{i} . While the linear dispersion is just a simple example of a non-parabolic isotropic band, the same analysis applies to a generic isotropic dispersion, with the magnitude of the effect controlled by the degree of non-parabolicity. We also employ two models of the ee interaction. The first model is the screened Coulomb (C) interaction, U_{\mathrm{C}}(q,\omega) , treated within the random-phase approximation (RPA). We assume weak coupling, which justifies the RPA treatment and, at the level of the kinetic equation, allows us to neglect Fermi-liquid renormalizations of the effective mass and the deviation of the Z-factor from unity. We focus on the regime of \mathcal{E}\equiv\max\{|\Omega,T\}\ll\Omega_{\rm p}\lesssim\varepsilon_{\rm F} , where \Omega_{\mathrm{p}}=v_{\mathrm{F}}\kappa is the characteristic plasma frequency, v_{\mathrm{F}} is the Fermi velocity, \kappa is the screening wavevector, and \varepsilon_{\rm F} is the Fermi energy; in this regime screening is effectively static. The extension to dynamic screening does not modify the main conclusions and will be discussed elsewhere. The second model is a FL near the z=3 quantum critical point (QCP), with the effective Hertz-Millis (HM) ee interaction 13 ; 25 , U_{\rm HM}(q,\omega)=\frac{g}{q^{2}+\xi^{-2}-i\alpha\omega/q},\quad\alpha=g\nu_{\rm F}/v_{\rm F}, (1) where \xi is the correlation length of the order parameter and \nu_{\mathrm{F}} is the density of states. For simplicity, we assume that the instability is in the \ell=0 channel. The FL regime corresponds to \mathcal{E}\ll\mathcal{E}_{\mathrm{FL}}=1/\alpha\xi^{3} . We adopt the Eliashberg approximation, i.e., we assume that in the FL regime the self-energy \Sigma^{R} is local 5 and vertex corrections are small. In contrast to the Coulomb case, the FL renormalizations cannot be neglected close to the QCP. Main results: General. For the two models specified above, we solve the kinetic equation (KE) exactly–by mapping it onto an inhomogeneous Legendre-type differential equation. The resulting conductivity is given by \displaystyle\sigma(\Omega,T) \displaystyle= \displaystyle\sigma_{\mathrm{i}}(\Omega)+\sigma_{\mathrm{ee}}(\Omega,T), (2a) \displaystyle\sigma_{\mathrm{ee}}(\Omega,T) \displaystyle= \displaystyle e^{2}\varepsilon_{\rm F}\tau_{\mathrm{i}}(\Omega)\Phi\left(\Omega,T\right)\mathcal{F}\!\left(\frac{\tau_{\rm ee}(\Omega,T)}{\tau_{\mathrm{i}}(\Omega)}\right), (2b) where \tau_{\mathrm{i}}(\Omega)=\tau_{\mathrm{i}}/(1-i\Omega\tau_{\mathrm{i}}) , \sigma_{\mathrm{i}}(\Omega)=e^{2}\nu_{F}v_{F}^{2}\tau_{\mathrm{i}}(\Omega) is the residual conductivity, \nu_{F}=\varepsilon_{\rm F}/2\pi v_{F}^{2} is the density of states per spin, and \Phi(\Omega,T)=(T^{2}+\Omega^{2}/4\pi^{2})/\varepsilon_{\rm F}^{2} . The quasiparticle scattering rate, 1/\tau_{\mathrm{ee}}(\Omega,T) , related to the imaginary part of the on-shell self-energy, is given by \tau_{\mathrm{ee}}^{-1}=a_{\rm M}\mathcal{E}^{2}L_{\rm M}(\mathcal{E}), where {\rm M}={\rm C,HM} indicates the model and L_{\rm M}(\mathcal{E})\equiv\ln\left(\Lambda_{\rm M}/\mathcal{E}\right) . Whether the FL parameters are taken as bare or renormalized also depends on the model, i.e., C vs. HM. The function \mathcal{F}(z) is expressed in terms of the associated Legendre functions, with the following asymptotic limits \mathcal{F}(|x|\gg 1)=-4\pi/15x and \mathcal{F}(|x|\ll 1)=-\pi/6+\pi x/8 . Main results: Coulomb interaction. For this model, all Fermi-liquid parameters are bare, with a_{\rm C}\sim 1/\varepsilon_{\rm F} and \Lambda_{\rm C}\sim\Omega_{\rm p} . In the dc limit ( \Omega=0 ), Eqs. ( 2a ) and ( 2b ) yield the known low- T behavior \sigma_{ee}(0,T)\propto T^{4}\ln T 28 ; 33 ; 17 and the high- T saturation of the conductivity, up to a Sommerfeld-type correction of order (T/\varepsilon)^{2} 27 ; 15 . Beyond reproducing these limits, the exact solution also establishes that the low- T /high- T crossover occurs when \gamma_{T}\sim 1/\tau_{\mathrm{i}} , with \gamma_{T}\equiv 1/\tau_{\mathrm{ee}}(0,T) . Moreover, since |\sigma_{\mathrm{ee}}|/\sigma_{\rm i}(0)\lesssim(T/\varepsilon_{\rm F})^{2}\ll 1 , the ee contribution remains small, rather than order one, in the degenerate regime. Both the T^{4}\ln T scaling and the low- T /high- T crossover originate from the change in total velocity in a normal ee collision, \displaystyle\Delta\bm{v}=\bm{v}_{\mathbf{k}}+\bm{v}_{\mathbf{p}}-\bm{v}_{\mathbf{k}-\mathbf{q}}-\bm{v}_{\mathbf{p}+\mathbf{q}}. (3) For a general isotropic, non-parabolic dispersion, \Delta\bm{v}\neq 0 . Upon projection onto the Fermi surface (FS), however, \bm{v}_{\mathbf{k}}\propto\mathbf{k} and \Delta\bm{v}=0 by momentum conservation, as in a Galilean-invariant system. The leading FS contribution therefore cancels, and the conductivity is controlled by states close to, but not exactly on, the FS. Thermal smearing then supplies a factor \propto T^{2} , represented by \Phi(0,T) in Eq. ( 2b ). Combined with the quasiparticle rate \gamma_{T}\propto T^{2}\ln T , this yields the T^{4}\ln T scaling. Since \mathcal{F} depends only on