sign in sign up
<<
Home , Fermi liquid theory

Propagation of shear stress in strongly interacting metallic Fermi enhances transmission of terahertz radiation

1,2 3 1, D. Valentinis , J. Zaanen , and D. van der Marel ∗ 1Department of Quantum Matter , University of Geneva, 24 Quai Ernest-Ansermet, 1211 Geneva 4, Switzerland 2Institute for Theoretical , Karlsruhe Institute of Technology, Wolfgang-Gaede Straße 1, 76131 Karlsruhe, Germany and 3Institute-Lorentz for Theoretical Physics, Leiden University,

PO Box 9506, NL-2300 RA Leiden, The Netherlands∗ (Dated: March 29, 2021) A highlight of Fermi- phenomenology, as explored in neutral 3He, is the observation that in the collisionless regime shear stress propagates as if one is dealing with the transverse of a . The existence of this “transverse zero sound” requires that the mass enhancement exceeds a critical value. Could such a propagating shear stress also exist in strongly correlated systems? Despite some noticeable differences with the neutral case in the Galilean continuum, we arrive at the verdict that transverse zero sound should be generic. We present an experimental setup that should be exquisitely sensitive in this regard: the transmission of terahertz radiation through a thin slab of heavy- material will be strongly enhanced at low temperature and accompanied by giant oscillations, which reflect the interference between light itself and the “material photon” being the actual manifestation of transverse zero sound in the charged Fermi liquid.

I. INTRODUCTION lations, with the ramification that a solid exhibits a prop- agating shear mode (the transverse acoustic phonon, TA) The elucidation of the Fermi liquid as a unique state of and a liquid a relaxational response instead. The spatial matter is a highlight of twentieth century physics 1 . It angular “-1” Landau parameter regulates [ ] S has a precise identity only at strictly zero temperature. At the mass enhancement of , m∗/m = 1 + F1 /3. Abrikosov and Khalatnikov 3, 4 predicted that for F S > 6 times large compared to ħh/(kB T) it can be adiabatically [ ] 1 continued to the high-temperature limit and it is therefore a new propagating mode forms - see Fig.1a. This mode indistinguishable from a classical fluid - the “collision-full propagates in fact shear stress (“transverse zero sound”), en- tailing a coherent shear deformation of the Fermi sphere 4 . regime” . However, at ħhω > kB T (the “collisionless [ ] regime”) the unique nature of the zero-temperature state can This may be confusing at first sight: the Fermi liquid is a be discerned, being different from either the non-interacting liquid also at zero temperature. The resolution is that the Fermi or a thermal fluid. Instead, the closest analogy is static shear rigidity of a solid is associated with the spin-2 with ordered matter characterized by spontaneous symmetry channel involving two space directions (e.g. Refs. 5 and breaking: the Fermi surface takes the role of order parameter, 6). Propagating shear requires only one space direction: acting however very differently from its “bosonic” analogues. it is spin-1 and can therefore be reconciled with a system S that is hydrostatically a liquid. Hence, shear-stress propa- A case in point is zero sound. The F0 Landau parameter en- capsulates the effects of interactions on the . gation in a Fermi liquid is unique in the regard that it is When this is finite, a propagating collective mode splits off an inherently dynamical phenomenon. As for longitudinal from the Lindhard continuum of particle-hole excitations. zero sound, temperature dictates the parameter space in The Fermi surface “hardens” and zero sound is the coherent which transverse zero sound propagates. The Fermi liquid s-wave breathing motion of this “membrane” corresponding exhibits a “viscoelastic-like” behavior [7], in the sense that to an oscillation of fermion density in space and time. Histor- at short times (collisional regime) shear is relaxing while ically the neutral Fermi liquid realized in 3He has formed the it propagates in the long-time collisionless regime. This experimental theater to test these general notions. The ob- prediction was confirmed in the 1970’s by transverse ultra- sound measurements in 3He 8 . in also servation of the sound attenuation maximum at ħhω kB T [ ] separating the hydrodynamic protection of first (thermal)≈ form Fermi liquids. We address here the following question: arXiv:2010.11554v2 [cond-mat.str-el] 20 May 2021 sound and the “Fermi-surface” protected zero sound has could it be that propagating shear is ubiquitous in the large been seminal in this regard [2]. variety of heavy Fermi liquid [9]? We will present The strongly interacting Fermi liquid has yet another, less here the case that it appears to be natural for such collective well known zero-temperature property. This revolves around modes to exist in these systems - see Fig.1b. Propagating the response to external shear stress. Shear rigidity is sup- shear in solids has been overlooked up to now for the simple posed to be uniquely associated with the breaking of trans- reason that it is quite difficult to measure. The difficulty is that one can only exert shear forces on electrons via elec- tromagnetic (EM) fields. To this end, gradients have to be applied and special experimental conditions are required to ∗ To whom correspondence should be addressed. E-mail: [email protected] overcome the kinematic mismatch between the electron and 2

S ω (a) Uncharged ω Charged, F1 > 6 (b) - shear mode, ωp S F1 > 6 photon: shear- ω v q polariton = F∗ ω = cq q ωp/c ≈

Lindhard continuum ω = vF∗q

S shear mode, F1 < 6 Lindhard continuum O Req O Req

Figure 1. Transverse collective mode spectrum of neutral and charged Fermi liquids. (a) Sketch of the real part of the dispersion S relation of transverse sound in a neutral Fermi liquid, corresponding to a pole in the transverse susceptibility. For F1 > 6 (solid line) S transverse sound propagates at frequencies above the Lindhard continuum (purple-shaded area); for F1 < 6 transverse sound is Landau- S damped (dashed line), and the pole disappears deep inside the continuum at an F1 -dependent frequency marked by the red dot. (b) Sketch of the real part of the dispersion relation of shear collective modes () in a charged Fermi liquid, in the long-wavelength S propagating-shear regime with F1 > 6, obtained from the Fermi-liquid transverse dielectric functin in linear response. The interaction of the photon root (dashed golden line) with the transverse sound of the neutral case (panel a) generates the usual plasmon-polariton (solid orange line), which propagates above the frequency ωp, and the shear-polariton (red solid line), which bears the signatures of Fermi-surface shear rigidity in the charged case. The shear-polariton is repelled by the interaction with the nearby photon root at low frequency (red-shaded area), so that it bends with quadratic dispersion and submerges into the continuum at q ωp/c. ∝ light velocities. ory applied to the Fermi Liquid (the Landau kinetic equa- tion [3]) we will present here a quantitative phenomenology for the transverse optical response of heavy Fermi-liquid II. RESULTS metals in the Galilean continuum. We will firsts re-derive the Abrikosov-Khalatnikov results for the neutral Fermi liq- uid, which form a useful template for the comparison with We will explain an experimental setup where propagat- the charged Fermi liquid. We will then turn to the charged ing shear should give rise to spectacular, counter-intuitive Fermi liquid, focusing on the implications of shear-stress signals. The configuration is conceptually straightforward: propagation for the optical properties associated with the it involves the transmittance of light through a thin metallic transverse response of heavy-mass quasiparticles. We will layer (see Fig.2a) as a function of radiation frequency and subsequently compute quantitative predictions for the trans- for a fixed layer thickness. Dealing with a weakly interacting mission experiment. We will finish with a discussion of how Fermi-liquid , the incident light is heavily attenuated such computations generalize to heavy-mass Fermi liquids inside the layer, which becomes essentially nontransparent: realized in solids. We will suggest an experimental protocol the EM field strength is damped exponentially on a length to observe propagating shear modes, concluding with the scale set by the skin depth. This behavior continues up to potential to use this as a new form of high precision “Fermi- the plasma frequency ω . However, upon increasing the p surface spectroscopy” . The latter may be used to shed a quasiparticle mass, a critical point is reached for m 3m: ∗ new light on the mysteries of strongly correlated electron above this threshold a collective mode of coherent≥ shear systems. deformation of the Fermi sphere propagates inside the ma- terial (see Fig.2b), and the transmittance starts to severely oscillate as function of frequency (see Fig.2c-f). In the ideal case of extremely clean samples and very low temper- atures, the transmittance can become so large that at the Let us start out reproducing the Abrikosov and Khalatnikov maximum of the oscillation the metallic layer may become results [3, 4] for the neutral Fermi liquid. The Landau kinetic quasi-transparent, at a thickness where the weakly interact- equation contains all the information required to analyze the ing metal would block all the incident light (see Fig.2c-d)! linear response to shear perturbations of given momentum Recently Khoo et al. studied shear sound of interacting q and frequency ω. This information is contained in the electrons in a different setting, namely in two-dimensional transverse susceptibility χT (q, ω), while the quasiparticle systems [10] and in narrow strips [11]. interactions are parametrized in terms of the momentum- 2 2 The plan for the remainder of this paper is as follows. conserving collision time τc ħhEF /[(ħhω) + (πkB T) ] in On the basis of the semiclassical Boltzmann kinetic the- the Galilean continuum [12] (≈cf. sec.IV). 3

0.06 (a) Fermi liquid

2 (d) 0.05 10−

~ | E z, ω q (ω) 4 ( ) + ) 0.04 10− ω ( t | | )

ω 6 ( 0.03 10

t − | ~ ET (z, ω) Drude 0.02 10 8 q (ω) − 5 4 3 2 1 − 10− 10− 10− 10− 10− ω/ωp O z 0.01 T = 1 K d T = 300 K (c) 0 0.006 1 10− (f) 0.005 3 10− | 0.004 )

ω 5

( 10− t | | ) ω

( 0.003 7 t 10 | −

0.002 10 9 − 5 4 3 2 1 10− 10− 10− 10− 10− ω/ωp 0.001 T = 1 K T = 300 K (e) 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014

ω/ωp

Figure 2. Simulations of optical transmission experiments at normal incidence on a charged Fermi liquid slab. (a) Schematics of the transmission experiments on a charged Fermi-liquid slab of thickness d. The incident electric field E(z, ω) produces the plasmon- polariton and the shear-polariton in the slab, which correspond to wave vectors q+(ω) and q (ω) for each frequency. Each mode is the − sum of two counterpropagating rays from Fabry-Perot internal reflections, which lead to the transmitted field ET (z, ω). (b) Schematic depiction of the Fermi-surface shear deformation induced by transverse sound. (c) Slab transmission modulus t(ω) = ET (z, ω)/E(z, ω) , | | | | as a function of frequency ω/ωp for a Fermi liquid in the Galilean continuum at temperatures T = 300 K (red line) and T = 1 K (blue 23 3 S S line). We assume a slab thickness d = 10c/ωp ( 0.1µm for a density n = 10 cm− ), Landau parameters F0 = 1 and F1 = 20, and the ∼ 2 slip length corresponding to specular interface scattering with hh0kF = 3.8 (cf. sec.IV). The shear mode submerges into the continuum below the dashed green line. (d) Same data as in panel c on logarithmic scale. The gray dashed line shows the result for a standard Ohmic conductor. (e) Slab transmission modulus t(ω) for a charged Fermi liquid including momentum relaxation due to Umklapp | | scattering with αU = 0.5, and acoustic with Debye temperature ħhωD/kB =500 K and electron-phonon coupling constant λ = 0.1. All other parameters are as in panels c-d. (f) Same data as in panel e on logarithmic scale. The gray dashed line is the result for an Ohmic conductor.

III. DISCUSSION associated with entering the Lindhard electron-hole contin- S uum [3]. We notice that for F1 < 6 a remnant of the shear mode still exists at small momenta, but it is critically damped Transverse zero sound appears as a pole of χT (q, ω) due to the decay in the Lindhard continuum (dashed line S above the critical F1 , representing the oscillating shear de- in Fig.1a), and it ceases to exist altogether deep inside the formation of the Fermi surface (see Fig.2b). This repro- continuum (dot in Fig.1a) [4]. duces the main Abrikosov-Khalatnikov result as indicated For future purposes it is convenient to sketch a “ in Fig.1a: the shear mode appears as an anti-bound state diagram” of the shear response in the neutral Fermi liquid as formed from the Lindhard continuum. In the collisionless S a S a function of ω and F1 , both for zero- ( ) and finite tempera- limit τc + and at large F1 , this mode has a veloc- ture (b): Fig.3. This will later serve as a basis to understand →q∞ S ity vs = vF (1 + F1 /3)/5 [13], which is enhanced with the transverse response of quasiparticles endowed with elec- respect to the velocity vF∗ of Fermi-surface quasiparticles. tric charge, where the latter allows for the interaction of The quasiparticle Fermi velocity is renormalized by inter- Fermi-surface shear modes with photons. S actions according to vF∗ = ħhkF /m∗ = vF /(1 + F1 /3) where In essence, the response diagram of Fig.3a-b outlines the 2 1/3 S vF = ħhkF /m is the bare Fermi velocity and kF = (3π n) is conditions on ω and F1 for transverse sound to develop in the Fermi wave vector in a three-dimensional Fermi liquid, the neutral system, as testified by the emergence of a pole S with n electron density. Therefore, for small q and F1 > 6 in the susceptibility χT (q, ω) (cf. sec.IV) in the presence transverse sound is protected against the Landau damping of the Fermi-liquid collision time τc [4, 8, 14]. An analysis 4

3 (a) (c) 2 10− 10− Propagating shear

10 4 Propagating shear − Lindhard 3 10−

5 10− 4 T = 0 10− ħhω/EF ħhω/EF 6 F A F A 0, F S 1 10− 0 = 1 = 0 = 10 5 Anomalous skin effect (Lindhard) regime 7 − 10− EF = 7.9 eV, T = 0 6 8 10− 10−

9 7 10− 10−

3 Propagating shear 3 (b) (d) 10− 10− 10 5 Propagating shear − k T/E 4 Lindhard B F 10− 7 10− 5 kB T/EF 10 9 Anomalous skin effect (Lindhard) regime 10− − 11 ħhω/EF ħhω/EF 10− EF = 7.9 eV, T = 3 K 6 10 EF = 7.9 eV, T = 3 K − 13 A A S 10− F0 = F1 = 0, F0 = 1 7 10 15 10− − 17 10− 10 8 − Viscous 19 Viscous 10− 10 9 10 21 − 0 2 4 6 8 10 12 14 16 18 20 − 0 5 10 15 20 25 30 35 40 45 50 S S F1 F1

Figure 3. Diagram of the shear response in neutral and charged Fermi liquids. (a) Shear-response diagram of the zero-temperature S neutral Fermi liquid as a function of excitation ħhω/EF and F1 , calculated from the transverse-sound velocity vs(ω) = ω(q)/q, and using the Fermi-liquid collision time τc . Notice that the zero-temperature phase diagram is valid at any electron density n, since at T = 0 S the collision time τc becomes a universal function of ħhω/EF . The threshold F1 = 6 separates the Lindhard regime, where the Lindhard continuum dominates the response and the transverse susceptibility has no pole, and the propagating shear regime, where transverse 23 3 S sound propagates. (b) Neutral shear-response diagram at T = 3 K and EF = 7.6 eV (electron density n = 10 cm− ). For F1 < 6, the red solid line marks the crossover between the viscous regime of hydrodynamic damped transverse sound and the Lindhard regime. S For F1 > 6, the condition ω = vF∗ Req (dashed golden line) identifies the crossover between viscous and propagating shear regimes. (c) Long-wavelength shear-response diagram of the charged Fermi liquid at T = 0 and EF = 7.6 eV,calculated from the dispersion relation of the shear-polariton and the Fermi-liquid optical conductivity in Kubo formalism. The shear-polariton emerges from the continuum at the solid blue line. Deep inside the continuum, the response is dominated by the Lindhard continuum and yields the phenomenology of anomalous skin effect. (d) Charged shear-response diagram at T = 3 K and EF = 7.6 eV.The red line shows the condition δ0 = lc , with δ0 the EM skin depth in anomalous regime and lc the collisional mean free path, which indicates the crossover between the anomalous skin effect and viscous regimes, with the latter implying hydrodynamic response.

S of the mode dispersion relation ω(q) for F1 > 6 confirms Fig.3a-b. that the real (i.e. reactive) part of ω(q) prevails over its imaginary (i.e. dissipative) part for Reω q v q: indeed At zero temperature (Fig.3a) the physical meaning of ( ) F∗ S transverse sound propagates nearly undamped≥ outside of the the above analysis is very simple: above the critical F1 the Lindhard continuum, consistently with the sketch in Fig.1a. response is governed by the propagating shear mode, and otherwise one recovers the incoherent response of the Fermi Such propagation occurs in the regime ωτc 1, and in the S  gas, devoid of collective modes. In the finite-temperature large-F1 collisionless limit the mode velocity reaches vs∞. case (Fig.3b) the crossover from the high-frequency zero It is instructive to translate the qualitative criterion sound to the low-frequency thermal regime occurs at a fre- Reω q v q, which characterizes the crossover from over- 3 ( ) = F∗ quency ħhω kB T , as known in the He community: such damped to propagating shear at the edge of the contin- frequency is so low because we need many collisions dur- uum, into a condition Reω(q) vF∗q on the mode velocity ing an excitation cycle, i.e. ωτc 1, to establish local S ≥ 3  vs(ω, F1 ) = ω(q)/q as commonly defined in the He liter- equilibrium among quasiparticles and enable hydrodynamic ature [4, 8]. We determine such condition numerically, as behaviour. In thermal regime there is no longer a critical S marked by the dashed orange line in Fig.3a-b. Conversely, change associated with F1 : we are dealing with the Galilean- S for F1 < 6 the pole in χT (q, ω) is heavily Landau-damped invariant continuum, which is always governed by Navier- at low frequency, in the hydrodynamical/collisional regime Stokes hydrodynamics characterized by the viscosity of the S 2 ωτc 1, and it disappears at an F1 -dependent frequency finite-temperature Fermi liquid ν(0) = vF τc, as follows from evaluated numerically and marked by the red solid line in the kinetic theory [3, 15] (cf. sec.IV). With regard to hydro- 5 dynamics, it will become clear that it is largely irrelevant turn assumes the , departing from a free Fermi for the electron systems: extremely clean systems such as gas, characterized by an overall scattering time τK due to graphene are required to observe hydrodynamic flow [16– momentum relaxation. In the Galilean continuum this is 21], while the latter is actually very difficult to detect by absent, and one has to reconsider the skin effect mechanism. S radiative means. The answer is well known at low temperature and small F1 Let us now turn to the less well-charted charged Fermi in the form of the anomalous skin effect, elucidated a long liquid - to the best of our knowledge the transverse response time ago by Reuter and Sondheimer [26]. Upon lowering S temperature, the mean free path of the electrons l v τ for large F1 was only addressed in the book by Nozieres mf = F and Pines [1] in the form of an exercise. A first ramifica- (either τc, τK or both) exceeds the length over which the tion of the Coulomb interactions is a of the transverse electric field penetrates the metal. The screen- Landau parameters [22, 23]. More importantly, one has ing is now entirely due to the Lindhard continuum, with now to address the linear response to EM sources, which the outcome that the skin depth saturates as a function of 2 2 1/3 can be handled by time-dependent mean-field (RPA) using temperature at a value δ0 = (4c vF )/(3πωωp) . the Kubo formalism (cf. sec.IV). The way this works in the This implies in turn that in{ the Galilean continuum} the longitudinal channel is well known. The zero (longitudinal) finite-temperature hydrodynamical fluid (e.g. Fig.3b) is ac- sound mode gets “dressed” by the Coulomb interaction, with tually hidden from observation in a large frequency regime! the effect that it is promoted to a plasmon with dispersion Hydrodynamics sets in at a scale larger than the collision q 2 2 2 p 2 large ωp + vF q where ωp = ne /ε0m is the electron length lc = vF∗τc, but we just learned that the screening is plasma∼ frequency. complete at the length scale δ0 < lc. In this regime the hy- However, light is transversally polarized. This means drodynamical fluid responds therefore like a free . that, in the presence of a propagating transverse mode in Upon raising temperature lc decreases, and when δ0 lc a the material, the latter blends with light through radiation- crossover occurs to a regime where hydrodynamical currents≈ matter interaction, which can be rationalized in terms of take over the screening (cf. sec.IV). In an earlier paper we analyzed the ramifications of such screening 15 . These simple linear mode coupling [24]. As for the plasmon, this [ ] is a universal affair and one may as well first consider a observations may be of relevance in the context of graphene- like systems, but we leave a detailed analysis to a future Wigner crystal formed by electrical charges [5]. The shear modes of such ensemble are the electronic analogues of trans- publication. verse acoustic phonons in ionic lattices, and such “phonons” As for the neutral case, we can sketch an electromagnetic couple to photons possessing a much larger velocity (of response diagram: Fig.3c-d. In the presence of electric light). A first product of this coupling is the usual “light- charge, the role of χT (q, ω) is taken by the transverse di- like” transversally polarized plasmon-polariton, which is electric function εT (q, ω) , with the two quantities linked by universal: indeed, at q = 0 both transverse and longitu- the Kubo formula (cf. sec.IV). The fact that the Fermi liquid dinal “” precisely match. However, its partner is is interrogated by photons, with dispersion relation ω =c the phonon-like “shear-polariton”: this mode continues to q (see Fig.1b), fixes momentum so that the only indepen- 2 S be massless, but it acquires a quadratic dispersion ω q dent parameters are ω and F1 . Polaritons are self-consistent at small momenta 5, 6 (a similar mechanism occurs∼ in a solutions of Maxwell’s equations in the Fermi liquid, i.e. [ ] 2 2 2 particular viscoelastic-like holographic strange metal, see q c /ω = εT (q, ω): in other words, we look at the photon reference 25.) The shear mode of the heavy-mass Fermi self-energy to deduce the response of the matter. liquid shares an analogous fate, see Fig.1b. The only quali- In the following, we describe the EM phenomenology tative difference with the neutral case of Fig.1a is that, for through the analytical solutions for polaritons obtained deep very small momenta being less than q ωp/c (cf. sec.IV), in the low- and high-frequency regimes (cf. sec.IV), which the shear mode dips below the upper≈ bound of the Lind- are sufficient to understand the essential physics. We leave hard continuum at ω vF q, thus falling prey to Landau the full numerical computation of polaritons near crossovers damping. Actually, according≈ to the kinetic theory, Landau between different regimes for a subsequent work, but here damping further modifies the mode dispersion, which be- we provide robust qualitative estimates on where such 4 comes ω q at vanishing frequencies [13]. crossovers occur. In the high-frequency, long-wavelength ∝ Once again, the only way to exert forces at finite frequency regime ω vF∗Req, at leading order we find two polariton on the charged Fermi liquid is by electromagnetic radiation. branches (cf. sec.IV): one is the usual plasmon-polariton This in turn implies restrictions on which intrinsic properties root, which propagates above the plasma frequency. The of this system can actually be measured: a primary conse- other, lower-frequency root is the “shear-polariton” , a gen- quence is that the response to the EM field is inherently a uine product of Fermi-surface shear interacting with light. boundary phenomenon. The reason is that an oscillating As such, the shear-polariton is a collective mode rooted in EM field is screened by the metal, with the ramification time-dependent mean field, being in this regard in the same that the transverse response corresponding to the screening category as the plasmon or even a TA phonon. Via continu- currents decreases exponentially away from the boundary, ity, the shear-polariton is associated with a coherent shear the classical skin effect. In an Ohmic conductor the skin current, which takes over the screening from the Lindhard p depth is δs = 2ρ/µω where ρ and µ are the DC resistiv- excitations as soon as it springs into existence. ity and permeability of the conductor, respectively. This in Fig.3c-d shows that, compared to the neutral case, 6 charged propagating shear takes over above a critical, fi- tons, one already anticipates that these oscillations in the nite frequency. This expresses the fact that at large wave- transmittance of the slab may be very large, as confirmed lengths the shear mode becomes overdamped because of by our computations: Fig.2c-d. its dispersion bending from linear to quadratic due to EM We addressed a similar set up some time ago, dealing forces at q ωp/c (see Fig.1b), which causes the mode to instead with a Navier-Stokes liquid characterized by its re- dive into the≤ Lindhard continuum [1]. The qualitative crite- laxational shear mode [15]. This computation is trivially rion for the shear-polariton to emerge from the continuum adapted to the present propagating shear – it is just an exer- is that its dispersion obeys ω vF∗Req. We estimate the cise involving the transmission and reflection of two linearly latter condition analytically (cf.sec.IV), as shown by the coupled evanescent waves in the slab (cf. sec.IV). In the blue line in Fig.3c-d. Notice that the shear-polariton never hydrodynamical case the shear mode is overdamped, with a S propagates for F1 < 6: this is because light needs to couple complex momentum dominated by its imaginary part while to propagating transverse sound in the material to gener- it has still a finite real (“propagating” ) part. We found out ate the propagating shear-polariton, and transverse sound that this sufficed to form similar oscillations as function of S propagation necessitates F1 > 6 as previously explained. frequency, being however very small while the transmission On the other hand, in the low-frequency, short-wavelength as a whole is strongly attenuated [15]. Dealing with the propagating shear mode, its imaginary momentum is small regime ω vF∗Req and at finite τc, the leading-order expan- sion of the Fermi-liquid dielectric function yields anomalous compared to its real one, which entails that the slab becomes skin effect (cf. sec.IV) due to the incoherent response of quite transparent and characterized by large oscillations in the continuum to EM fields. A final difference with the neu- the transmittance as a function of frequency. tral case is that the apparent onset of the hydrodynamical Using the results of the kinetic theory as an input, we “viscous” regime is at a much lower frequency than in the can realistically simulate the transmission of slabs contain- neutral case. But as we explained, this is in a way a decep- ing Galilean-invariant Fermi liquids (cf. sec.IV). For the tion: although in principle the bulk may well already be results in Fig.2 we assumed a slab thickness d = 10c/ωp 23 3 in the thermal regime, the boundary layer where screening (= 0.1 µm for a density n = 10 cm− ), Landau parameters ∼S S currents are running is still too thin for the fluid to reach F0 = 1, and F1 = 20, and the slip length corresponding to local thermal equilibrium, given that the collision length is 2 specular interface scattering with hh0kF = 3.8 allowing for large compared to the skin depth, and the ensuing Fermi-gas a maximal transmission of t(ω) 0.06 in the frequency response (anomalous skin effect) is insensitive to the dif- 3 | | ≈ range between 10− ωp and 0.1ωp. As we just discussed, ference between the thermal- and zero-temperature liquid. at small frequencies up to the green dashed line the shear We qualitatively estimate the crossover between Lindhard mode is overdamped, while spectacular transmittance oscil- and hydrodynamical regimes by solving δ0 = lc numerically lations set in upon entering the propagation regime (see blue with the Fermi-liquid collision time τc: this gives the red line in Fig.2d). One infers that the overall transmission is curve in Fig.3d. much larger than the standard result for an ohmic conductor Having established where to look for the propagating (dashed gray line in Fig.2d). Up to this point we ignored shear in the Fermi-liquid metal, the next question is how the lifetime of the shear mode. This is accounted for by the to find it. In this regard we wish to propose an experiment kinetic theory, parametrized by the frequency dependence 2 yielding a unique, qualitative signal that will unambiguously of the collision time, τc ω− . This attenuates the shear confirm or falsify the existence of the shear mode. This mode at high frequencies,∝ with the consequence that the revolves around shining light through a thin, micrometer- enhanced transmission decreases rapidly as illustrated in the sized slab of metal. In a normal metal, the field strength of log-log plot (Fig.2). Raising temperature has a very similar 2 the radiation decays on the length scale set by the (anoma- effect since τc(ω = 0, T) T − . (red line in Fig.2d). lous) skin depth, which is small compared to the width of The conclusion is that,∝ in order to detect the propagating the slab. In these conditions, the slab is opaque. However, shear oscillations in a heavy-mass Fermi liquid, temperature the propagating shear mode is in essence a photon “made should be small compared to EF , while small frequencies (of 3 from matter”, being only different in the regard that its order of 10− of the plasma frequency) are required, just velocity is much smaller than the light velocity. The cou- above the point where the propagating shear escapes the pled light-matter system is therefore characterized by two long-wavelength Landau damping sketched in Fig.1b. propagating modes: the light-like plasmon- and matter-like As illustrated in the above, the Landau kinetic theory shear-polariton, characterized by wave vectors q+(ω) and has remarkable powers in its capacity to deliver a quan- q (ω) -see Fig.2a. These modes propagate with different titative phenomenology. However, it requires Galilean in- velocities,− while both are in part reflected back at each inter- variance, which is badly broken in the experimentally avail- face. Given the difference in the real part of their momenta, able charged Fermi-liquids: the metallic state of electrons their phases evolve differently along the path through the formed in solids. Such breaking of Galilean invariance has material. Accordingly – pending the frequency of the in- the fundamental ramification that total momentum is no coming light – the two polaritons interfere constructively or longer conserved. Considering the longitudinal response destructively at the slab boundaries, giving rise to a transmit- of the Fermi-liquid, total momentum conservation entails ted field that exhibits oscillations as function of frequency that the sound mode (or the plasmon) is infinitely long- (see Fig.2b). Given the propagating nature of both polari- lived both in the zero- and finite-temperature cases in the 7 long-wavelength limit. The same principle applies to trans- of electron systems with a large quasiparticle mass. In turn, verse zero sound, although it needs in addition the sharp these strongly renormalized Fermi liquids are rooted in a zero-temperature Fermi surface. strongly interacting microscopic physics, which is typically The role of momentum (non-)conservation is thereby iden- described by Kondo lattice models, by the proximity to quan- tical regardless of whether one is dealing with the longitudi- tum phase transitions, and so forth [29]. As a matter of nal or transverse electromagnetic response, and we know fact, it is not at all understood why a heavy Fermi liquid precisely how this works for the former. In the Fermi liq- emerges in the deep infrared, but this is the domain where uid, the large Umklapp momentum scattering associated the experimental study of transverse zero sound may have with a perfectly periodic lattice is irrelevant at long wave- considerable potential as a spectroscopic technique, due to length, and what remains is elastic scattering against im- its exquisite sensitivity to the characteristics of the Fermi purities. Furthermore, at finite temperature (and energy) surface. momentum relaxation also acts through electron-electron There is leeway for theoretical guidance in the form of the interactions and electron-phonon coupling. Good metals tradition on determining the renormalized band structure are “nearly hydrodynamical” , in the sense that the width of governing the low-temperature Fermi-surface topology. This the Drude peak, which represents the momentum relaxation is typically a rather complex affair for real solids, with a vari- rate, is small. Such scale governs the decay of transverse ety of sheets characterized by different effective masses [30]. sound as well. In a Fermi liquid, the main consequence In such case, the shape of the Fermi surface, as well as the of a finite-momentum lifetime is that the hydrodynamical quasiparticle mass and transverse sound, are associated with flow of the continuum is in any but the most perfect solids projections based on the point-group symmetries of the crys- (like graphene) immediately destroyed and replaced by the tal. This implies that, for instance, the velocity and the damp- Drude (Ohmic) regime, where momentum is damped in the ing of transverse sound become anisotropic, but this can be lattice by the individual quasiparticles [27]. Consequently, dealt with in principle, as demonstrated in the theory of the hydrodynamic regime in Fig.3d is replaced by the Ohmic elasticity and quite recently also in hydrodynamics [31, 32]. skin effect. Can transverse zero sound be measured in the laboratory, However, the propagating shear is only affected to be exploited as a spectroscopic tool to study the mysteries “perturbatively” : it will decay faster compared to the contin- of electron systems with heavy quasiparticle mass? From uum case due to momentum relaxation. To offer some intu- the above we have established the required conditions. First, ition regarding these numbers, we show in Fig.2e-f a realistic one needs good metals characterized by a small resistivity at example (cf. sec.IV). Compared to Fig.2c-d we introduce low temperatures, to assure that transverse sound does not two additional scattering channels: the temperature- and decay too rapidly through momentum relaxation. To avoid complications associated with extreme anisotropy, the ideal frequency-dependent momentum relaxation time τK (T, ω) which parametrizes the T 2 contribution rooted in a realistic candidate materials for transverse-shear optical spectroscopy should be preferably rather isotropic and three-dimensional, “Umklapp efficiency” αU = 0.5 [13, 15], and the coupling to phonons, which is particularly efficient in relaxing momen- with a band structure as simple as possible. Most importantly, tum. Transverse zero sound will couple to transverse lattice they should be characterized as Fermi liquids with large S vibrations, and we consider such acoustic phonons with a De- quasiparticle mass enhancement: since a large F1 protects transverse sound from Landau damping, the higher the mass bye temperature ħhΩD/kB = 500 K and an electron-phonon coupling constant λ = 0.1. The outcome is that phonons the better. Examples of candidate materials fulfilling these and Umklapp effectively conceal any signal of shear prop- criteria are α-cerium [33, 34], MnSi [35], and UPd2Al3 [36]. agation at high temperatures, but at low temperatures the These should be tailored into submicron-width thin layers oscillations are still clearly discernible. However, even for a and then cooled to low temperature, while subsequently the relatively weak electron-phonon coupling, the phonons rep- transmission should be measured as function of frequency resent an efficient decay channel for transverse zero sound: in the terahertz range: the sweet-spot frequency window to as the figure suggests, one better looks for transverse shear look for lies in between the low-frequency regime, where at frequencies below the typical phonon frequency scale in the traditional skin effect takes over, and the regime where a given material. phonons damp transverse sound. It is a beneficial circum- In addition to slab transmission, surface impedance stance that the technology to realize these conditions in the measurements provide in principle simpler access to the laboratory has just become available [37]. EM skin depth which discerns the different regimes in Fig.3d [13, 15]. However, explicit computation [28] shows that the differences between the signals in propagating-shear, ACKNOWLEDGMENTS Ohmic and anomalous skin effect regimes are minimal, once momentum relaxation is taken into account. Hence surface We acknowledge stimulating discussions with A. Mor- impedance experiments on bulk heavy-mass Fermi-liquids purgo, C. Berthod, J. Schmalian, A. B. Kuzmenko, G. A. are only subsidiary with respect to thin-film optical spec- Inkof, M. Baggioli and F. Pientka. This work was supported troscopy in identifying Fermi-surface propagating shear. by the Swiss National Science Foundation through grants no. A real complication is that invariably strong periodic back- 200020 179157 (D.v.d.M.), and the Early Postdoc Mobility ground potentials are required to endow the Fermi surface grant P2GEP2− 181450 (D.V.). 8

s Author contributions ized by the Landau parameters Fl . The result is DV, JZ and DvdM conceived the project and discussed the model, DV performed calculations, DV,JZ and DvdM wrote  qvk,σ cos θ ω εk (q, ω) the paper. − Z 0 + Competing interests dΩ X∞ s + qvk,σ cos θ F ℘l (cos θ 0 )εk (q, ω) None. 4π l 0 l=0 Materials & Correspondence e Not applicable. = Icoll (q, ω)+Ir (q, ω) qvk,σ cos θ [k A(q, ω)] , − mvF∗ · (2)

where θ is the angle between the wave vector q and the quasiparticle velocity vk,σ, and dΩ is the solid angle element IV. METHODS in three dimensions. We further expand the Fermi-surface displacement function εk (q, ω) in the spin-symmetric chan- nel in terms of spherical harmonics m θ, φ 13 : In this section we summarize the main features of the Yl ( ) [ ] model used to obtain all results discussed in the main text. + +l Additional details of the formalism are described in Ref. 13. X∞ X s m εk (q, ω) = εl,mYl (θ, φ). (3) l=0 m= l − In the following we consider the first transverse mode with m = 1, generating shear currents in the first spin-symmetric interaction channel. Therefore we truncate the sums over l in Eqs. (2) and (3) to l 1, so that P+ F s℘ cos α F S = l=∞0 l l ( ) 0 + A. Landau kinetic equation F S cos α, and ε q, ω P+ εs Y 1 θ, φ εs ≡θ eiφ, 1 k ( ) = l=∞0 l,1 l ( ) ( ) s ≡ where ε (θ) collects the θ-dependent part of the displace- ment. Integration over the angles θ and φ in Eq. (2) with We start from the Landau kinetic equation, which deter- 0 0 the displacement εs θ eiφ gives null result for terms involv- mines the first-order deviation (or “displacement") ε q, ω ( ) k ( ) ing F S 13 , and we are left with of the quasiparticle distribution function, with momentum k 0 [ ] on the Fermi surface, with respect to global thermodynamic s iφ equilibrium, driven by external perturbations of momentum (cos θ s) ε (θ)e − π q and frequency ω. Perturbation sources include quasiparti- Z sin θ dθ cle interactions, collisions, momentum-relaxing scattering cos θ 0 0 F S sin θ sin θ πeiφεs θ + 4π 1 0 ( 0) and external driving forces like electromagnetic fields. In 0 general, the kinetic equation reads 1,3, 13 1 1 e [ ] = Icoll (q, ω)+ Ir (q, ω)+ cos θ [k A(q, ω)] , qvF∗ qvF∗ mvF∗ · (4)  q vk,σ ω εk (q, ω) + q vk,σδ [E (k, σ) µ] · ( − · − ) where X   fk,σ,k ,σ εk (q, ω)δ En(k0, σ0) µ ω 0 0 0 s , (5) · k ,σ − = 0 0 qvF∗ = Icoll (q, ω) + Ir (q, ω) S e  and vF∗ = vF /(1 + F1 /3) is the velocity of quasiparticles on q vk,σ δ [E (k, σ) µ] k A(q, ω), (1) mv the Fermi surface – i.e. at the Fermi wave vector k kF – − F∗ · − · ≡ renormalized with respect to vF = ħhkF /m and kF by quasi- particle interactions. where vk,σ is the quasiparticle velocity for momentum k and Eq. (4) is the starting point of our analysis: it describes spin σ , , fk,σ,k ,σ encode short-ranged interac- coherent shear vibrations of the Fermi surface with the dis- = 0 0 tions between{|↑〉 quasiparticles,|↓〉} Icoll (q, ω) is the momentum- persion relation (5) in response to collisions, interactions, conserving collision integral, Ir (q, ω) is the integral for momentum-relaxing scattering and applied vector potentials. momentum-relaxing scattering, A q, is the applied vec- S ( ω) Notice that Eq. (4) does not depend on F0 , since shear de- tor potential, and the delta functions δ [En(k, σ) µ] select formations generate transverse currents with no net density quasiparticle states at the µ−. Consid- flow. ering a three-dimensional system, we standardly expand In the following sections we solve Eq. (4) for the the interaction term in angular harmonics using Legendre transverse susceptibility in the Galilean continuum by polynomials ℘l (cos θ 0 ) [3, 7, 13, 38], and we focus on the explicitating the form of the collision integral as prescribed spin-symmetric channel for density excitations, character- by Fermi-liquid theory [4]. Then, we analyze the effect of 9 momentum-relaxing scattering in relaxation-time approxi- 1 β = . (10b) mation [7] considering impurities, Umklapp processes and iωτc 1 acoustic phonons as independent scattering sources. − To include damping of transverse currents due to momentum relaxation, we have to utilize the full kinetic equation (4) with the integral Ir (q, ω). A convenient way to relate the momentum-relaxing transverse susceptibility to the result B. Fermi-liquid transverse susceptibility (9) without relaxation is provided by reference 7: we define the momentum-relaxing integral in single relaxation time approximation as Starting from the kinetic equation (4), we can obtain the transverse (current) susceptibility of the Fermi liquid in s r 1 ε (θ) ε (θ) iφ closed form thanks to the assumption of only two harmonic Ir q, ω se . (11) qv ( ) = iωτ− components of short-range interactions. We first neglect F∗ − K both the collision and the momentum-relaxing integrals, According to Eq. (11), momentum relaxation tends to re- i.e. Icoll (q, ω) = Ir (q, ω) = 0, and we assume no interac- store a “locally relaxed" equilibrium distribution function, tions. In this case, from the expression of the paramagnetic r characterized by the displacement ε (θ) [7], which would current density one derives the noninteracting transverse be in equilibrium without relaxation but in the presence of paramagnetic susceptibility 13 : [ ] an effective vector potential Ar (q, ω) modified by relaxation. The scattering processes described by the time τ conserve nm K χ0 q, ω 3 ∗ s , (6) particle number but not current 7 , and may be ω- and T ( ) = 2 ( ) [ ] m I T-dependent. This way, we obtain the momentum-relaxing  S  transverse susceptibility 13 : F1 [ ] where m∗ = m 1 + 3 is the effective mass of Landau ˜ quasiparticles, and n ω 1 + β [1 + 3 (ζ)] χ q, ω . (12) T ( ) =  F S  I  ‹ m ω + i/τK 1 ˜ ˜ 1 1 2 s 2 s + 1 1 3 3 β (ζ) + β (s) = + s + 1 s ln . (7) − − I I −3 2 4 − s 1 − where We now allow for momentum-conserving collisions in the  ‹ Galilean continuum, i.e. with no momentum relaxation. The i i i ζ = ξ + s = s 1 + + , (13) conservation of particle number, energy and momentum in ωτK ωτc ωτK collisions imposes constraints on the form of the collision and β˜ βξ/ζ. We recognize a pole at the denominator of integral Icoll (q, ω) [3]. The latter is written is terms of a = Eqs. (12) and (9), in the presence or absence of relaxation re- single collision time τc = τc(ω, T), which may depend on spectively: this collective mode is transverse sound, studied frequency ω and temperature T, as [4, 13] and observed in the electrically neutral Fermi liquid realized in 3He 4, 8 . The dispersion relation of transverse sound 1  s iφ [ ] Icoll (q, ω) = Icoll ε (θ)e stems from the condition that the susceptibility diverges for qvF∗ s s s a given real frequency ω and complex momentum q: ε (θ) [ε (θ)]av 3 [ε (θ) sin θ]av sin θ iφ = − − se (8) 1 − iωτc χT (q, ω)− = 0 (14)

π 2 where the notation R (sin θ) dθ denotes the angu- Eq. (14) determines the velocity of the transverse collective [ ]av = 0 4 S lar average with respect· to θ. Notice that· for purely shear mode with first Landau parameter F1 and collisions encoded s by τ τ ω, T, F S . This mode has been labeled “transverse stresses [ε (θ)]av = 0 [4]. Again the definition of the param- c = ( 1 ) agnetic current density allows one to derive the interacting sound” in the liquid- literature, although physically transverse paramagnetic susceptibility [13]. We have it amounts to a shear oscillation of the Fermi surface. The numerical solutions of Eqs. (9) and (14) are shown schemat- X 0 ξ ically in Fig.1a. T ( ) S S χT ξ For F < 6, there is an F -dependent value of ω above ( ) =  S  2 1 1 th F1 m 0 1 3 3 β 3nm X T (ξ) + β which Eq. (14) does not have solutions. Physically, this − − ∗ means that no Fermi-surface collective modes are allowed  F S  3 1 1 ξ n + 3 ( ) for ω > ωth, and the only possible excitations are incoherent I . (9) electron-hole pairs. Knowing the dependence τ ω, T, F S , = m F1s  c( 1 ) 1 3 3 β (ξ) + β in the Fermi-liquid model – i.e. Eq. (24) – one finds the − − I S numerical solution ω(T, F1 ) which is displayed with the red where S solid curves in Fig.3a,b. For F1 > 6 Eq. (14) has solutions  i ‹ for all frequencies, but above a certain threshold frequency ξ = s 1 + , (10a) ωτc ωth the solutions are outside the particle-hole continuum 10 and thus free from Landau damping. This threshold is set tions inside the Fermi liquid [1]: by solving Eq. (14) together with the condition q2c2 " q, ω (17) ω2 = T ( ) 1 ωthReq = v (15) − F∗ Analytical solutions of Eqs. (16) and (17) are available in the regimes q 0 and q : the leading-term solution The orange dashed curves in Fig.3a,b represent the in the q 0 limit→ reproduces→ ∞ the phenomenology of viscous S charged→ fluids 15 , while in the q regime we retrieve numerical solution of ωth as a function of F1 . [ ] the phenomenology of anomalous→ skin ∞ effect [13, 26, 39]. In the regime ω vF∗ Req, the leading-order expansion of  Eq. (17) for ω/(vF∗q) gives [13] → ∞ 2 (ωp) " q, ω 1 , (18) T ( ) = 2 2 − ω + iω/τK + iων˜(ω)q C. Dielectric function and polaritons where

 S  2 F1 (vF∗) τc The effect of the long range Coulomb interaction is de- ν˜(ω) = 1 + . (19) 3 5 (1 + τc/τK iωτc) scribed by the transverse dielectric function, which can be − obtained from the paramagnetic transverse susceptibility In the limit τc τK , that is when momentum-conserving (12) through the Kubo formula. We have collisions are much more frequent than momentum-relaxing scattering, Eq. (18) becomes [13]

2 2 e (ωp) "T (q, ω) = 1 χT (q, ω) ε q, ω 1 " ω2 T ( ) = 2 2 − 0 − ω + iω/τK + i (ω + i/τK ) ν(ω)q 2 ˜ 2 (ωp) 1 + β [1 + 3I (ζ)] (ωp) 1 . (16) 1 , (20) =  F S  2 2 − ω (ω + i/τK ) 1 ˜ ˜ ω + iω/τK + iων(ω)q 1 3 3 β I (ζ) + β ≈ − − − where we have defined the generalized shear modulus of the isotropic Fermi liquid [7, 40, 41] p 2 where ωp ne / m"0 is the electron plasma frequency. = ( ) S 2 Notice that the poles of the dielectric function (16) cor- 1 + F1 /3 (vF∗) respond to the transverse sound dispersion relation (14), ν (ω) = . (21) 5 1/τc iω i.e. the shear resonance of the Fermi surface. Due to the − charge of the electrons, these resonances couple to radiation Eqs. (20) and (21) can be derived from the combination of through the shear stresses exerted by photons. The radiation- the linearized Navier-Stokes equation for transverse currents matter interaction gives rise to new collective modes, con- and of Maxwell’s equations under the assumption that τc sisting of the combined shear response of photons and Fermi- τK . In the propagating shear regime Eqs. (17) and (18) give surface quasiparticles: the polaritons. Mathematically, po- rise to two degenerate polariton branches for each frequency, laritons satisfy the self-consistent solutions of Maxwell equa- characterized by the complex wave vectors

( q ) 2 τ ω c2 iν ω ω 2 4ic2ν ω ω 2 i c2 iν ω ω 2 2 ω iω K [ + ( ) ] ( )( p) + [ + ( ) ] q 1 − , (22) = 2 + 2p 2c 2ν(ω) ± c ωτK + i

One solution of Eq. (22) is the usual plasmon-polariton mation of the latter condition by reverting to Eq. (15) and S which propagates above ωp, as also found in the standard noticing that, for F1 > 6, the shear-polariton emerges from Drude model, and the other slower-dispersing root is the the continuum in the regime ωτc 1, Req Imq. This shear-polariton. The two roots are schematically shown in allows us to find the analytical solution  Fig.1b. The shear-polariton root of Eq. (22) emerges from the electron-hole continuum, thus ceasing to be Landau- ωp 3vF∗ ωth = r (23) damped, for ω > ωth. We can obtain a quantitative esti- c S  ” 2— F1 6 1 vF∗/c − − 11

1 Eq. (23) is shown by the blue solid curves in Fig.3c,d. Z t  1 1 2 ‹ x, t t3 duu2 . It confirms the early order-of-magnitude estimate by S ( ) = u x + u x + u e t 1 e + t 1 e 1 Nozières and Pines 1 , that is Req ω/v ω /c at the 0 − + + [ ] = F∗ p −(26c) boundary with the electron-hole continuum.≈ In the region where λ is the electron-phonon coupling constant, hω is between the blue solid lines and the red dashed lines in ħ D the Debye energy, and ψ z is the first derivative of the Fig.3c,d, the shear-polariton propagates but it cannot 0( ) Digamma function. For the impurity scattering rate, we be described by the viscous-liquid phenomenology of Eq. assume a positive constant τ consistently with first Born (20). The dispersion of the shear-polariton is robust against i approximation. momentum relaxation, as it is negligibly affected by the value of τK in the propagating regime.

E. Thin-film optical transmission in propagating-shear regime

We now turn to the computational method for the optical D. Collision and momentum relaxation rates transmission of a Fermi-liquid thin film of thickness d in the propagating shear regime. The transmission coefficients at the vacuum/material interfaces follow from the condition For the electron-electron momentum-conserving collision that the electromagnetic wave in the vacuum has to match time τc = τc(ω, T) we employ the Fermi-liquid result [14, the frequency-degenerate polariton modes, Eq. (22), inside 29, 42] the film. To calculate these coefficients, we exploit the anal-

ogy between Fermi-liquid electrodynamics for ω vF∗Req and the electromagnetic response of viscous charged flu- h • 7 5 5 ˜ ħ π S 2 S 2 S S ids 15 : appropriate additional constitutive relations at the = (A1) + (A0) A0A1 [ ] τc 12 24 8 − 12 × slab boundaries can be formulated in terms of the slip length hω 2 πk T 2 λ, in accordance with Ref. 15, which closes the system of (ħ ) + ( B ) , (24) E equations for the reflection and transmission coefficients F∗ of each polariton mode at the slab boundaries. A total of six linear equations stem from the continuity of the electric j j ” j — field and of its derivative, and the constitutive relation from where Al = Fl / 1 + Fl / (2l + 1) . Eq. (24) holds for an S S the linearized Navier-Stokes equation. The solution of such interaction comprising the Landau parameters F0 and F1 , linear equations gives the transmission coefficient of the film consistently with the assumptions made in obtaining the tfilm = tfilm(ω) relative to the transmission in vacuum tv: kinetic equation (4). The momentum transfer rate, 1/τK , results from impurity scattering, 1/τ , electron-phonon scat- i iωd/c  iq+ d iq+ d iq d tfilm (ω) /tv = e− t+e + θ+e− + t e − tering, 1/τe ph, and Umklapp scattering. The Umklapp term − − iq d  is given by a fraction αU (0, 1) of the electron-electron +θ e− − , (27) collisions. Typical values in∈ transition-metal compounds are − where the indices refer to the branches of Eq. (22), t and αU 0.5 [42]. Consequently v ≈ the amplitudes t , θ are determined± numerically through Eq. (9) of Ref. 15±. ± 1 αU 1 1 We microscopically determine the slip length following = + + (25) τK τc τe ph τi Ref. 43: − 1 45π2 3 λs nmν 0 v τ , We calculate the electron-phonon relaxation rate from the = h2 h 2 k 4 k 4h ( ) = h2 h 2 k 4 F∗ c ( 0) ( F ) ( F ) ħ ( 0) ( F ) many-body self-energy associated with scattering between (28a) electrons and an Einstein phonon branch [13]. Explicitly, 8π2 8 λd nmν 0 v τ , (28b) = k 4h ( ) = 15 F∗ c  ‹ ( F ) ħ ħh 2 ω kB T = πλħhωDS , , λhωD € kB T Š where h and h are amplitude and correlation length of τe ph(ω, T) 1 ħ Q ħhωD ħhωD 0 − πkB T ħhωD the interface roughness, k is the Fermi wave vector, and − (26a) F λs (λd ) refer to specular (diffuse) scattering. In Fig.2c-f we compare the results for t(ω) = tfilm (ω) /tv from the Z 1 §  ‹ª propagating-shear model, Eq. (27), with the standard Drude 2 1 u Q(t) = duu Im ψ0 + i (26b) model (Ohmic conductor) for thin-film transmission. 0 2 2πt 12

[1] P. Nozières and D. Pines, Theory Of Quantum Liquids, Ad- [20] J. A. Sulpizio, L. Ella, A. Rozen, J. Birkbeck, D. J. Perello, vanced Books Classics Series (Westview Press, Nashville, D. Dutta, M. Ben-Shalom, T. Taniguchi, K. Watanabe, 1999). T. Holder, R. Queiroz, A. Principi, A. Stern, T. Scaffidi, A. K. [2] W. R. Abel, A. C. Anderson, and J. C. Wheatley, Phys. Rev. Geim, and S. Ilani, Nature 576, 75 (2019). Lett. 17, 74 (1966). [21] L. Ella, A. Rozen, J. Birkbeck, M. Ben-Shalom, D. Perello, [3] A. A. Abrikosov and I. M. Khalatnikov, Rep. Prog. Phys. 22, J. Zultak, T. Taniguchi, K. Watanabe, A. K. Geim, S. Ilani, and 329 (1959). J. A. Sulpizio, Nat. Nanotechnol. , 1 (2019). [4] M. J. Lea, A. R. Birks, P.M. Lee, and E. R. Dobbs, J. Phys. C: [22] V.P.Silin, Sov. Phys. JETP-USSR 6, 387 (1958). Solid State 6, L226 (1973). [23] V.Silin, Sov. Phys. JETP-USSR 6, 985 (1958). [5] A. J. Beekman, J. Nissinen, K. Wu, K. Liu, R.-J. Slager, Z. Nussi- [24] D. Forcella, A. Mezzalira, and D. Musso, J. High Energy Phys. nov, V.Cvetkovic, and J. Zaanen, Phys. Rep. 683, 1 (2017). 2014, 1 (2014). [6] A. J. Beekman, J. Nissinen, K. Wu, and J. Zaanen, Phys. Rev. [25] M. Baggioli, U. Gran, and M. Tornsö, J. High Energy Phys. B 96, 165115 (2017). 2020, 1 (2020). [7] S. Conti and G. Vignale, Phys. Rev. B 60, 7966 (1999). [26] G. E. H. Reuter and E. H. Sondheimer, Proc. R. Soc. A 195, [8] P.R. Roach and J. B. Ketterson, Phys. Rev. Lett. 36, 736 (1976). 336 (1948). [9] D. Stricker, J. Mravlje, C. Berthod, R. Fittipaldi, A. Vecchione, [27] R. N. Gurzhi, Sov. Phys. Uspekhi 11, 255 (1968). A. Georges, and D. van der Marel, Phys. Rev. Lett. 113, [28] D. Valentinis, J. Zaanen, and D. van der Marel, (2020), 087404 (2014). unpublished. [10] J. Y. Khoo and I. S. Villadiego, Phys. Rev. B 99, 075434 (2019). [29] P. Coleman, Introduction to many body physics (Cambridge [11] J. Y. Khoo, P.-Y. Chang, F. Pientka, and I. Sodemann, Phys. University Press, Cambridge, 2010). Rev. B 102, 085437 (2020). [30] G. Kotliar, S. Y. Savrasov, K. Haule, V.S. Oudovenko, O. Par- [12] For all calculations in this work we employed the momentum- collet, and C. A. Marianetti, Rev. Mod. Phys. 78, 865 (2006). conserving collision time τc (a single-particle quantity) in [31] B. Bradlyn, M. Goldstein, and N. Read, Phys. Rev. B 86, accordance with Eq. (24) in sec.IV. This is different from the 245309 (2012). sound-attenuation or optical scattering rates (two-particle [32] C. Q. Cook and A. Lucas, Phys. Rev. B 99, 235148 (2019). 2 quantities), which for a Fermi liquid are ħhEF /[(ħhω) + [33] J. W. van der Eb, A. B. Kuzmenko, and D. van der Marel, 2 ∝ (2πkB T) ] [44]. Phys. Rev. Lett. 86, 3407 (2001). [13] D. Valentinis, arXiv:2003.06619 (2020). [34] K. Haule, V.Oudovenko, S. Y. Savrasov, and G. Kotliar, Phys. [14] D. van der Marel, J. L. M. van Mechelen, and I. I. Mazin, Rev. Lett. 94, 036401 (2005). Phys. Rev. B 84, 205111 (2011). [35] F. P. Mena, D. van der Marel, A. Damascelli, M. Fäth, A. A. [15] D. Forcella, J. Zaanen, D. Valentinis, and D. van der Marel, Menovsky, and J. A. Mydosh, Phys. Rev. B 67, 241101 (2003). Phys. Rev. B 90, 035143 (2014). [36] Scheffler, M. and Dressel, M. and Jourdan, M. and Adrian, H., [16] J. Crossno, J. K. Shi, K. Wang, X. Liu, A. Harzheim, A. Lucas, Nature 438, 1135 (2005). S. Sachdev, P.Kim, T. Taniguchi, K. Watanabe, et al., Science [37] L. Prochaska, X. Li, D. C. MacFarland, A. M. Andrews, 351, 1058 (2016). M. Bonta, E. F. Bianco, S. Yazdi, W. Schrenk, H. Detz, A. Lim- [17] D. A. Bandurin, I. Torre, R. Krishna Kumar, M. Ben Shalom, beck, Q. Si, E. Ringe, G. Strasser3, J. Kono, and S. Paschen, A. Tomadin, A. Principi, G. Auton, E. Khestanova, K. S. Science 367, 285 (2020). Novoselov, I. V.Grigorieva, L. A. Ponomarenko, A. K. Geim, [38] G. Giuliani and G. Vignale, Quantum Theory of the Electron and M. Polini, Science 351, 1055 (2016). Liquid (Cambridge University Press, Cambridge, 2005). [18] K. R. Kumar, D. Bandurin, F. Pellegrino, Y. Cao, A. Prin- [39] E. H. Sondheimer, Ad. Phys. 50, 499 (2001). cipi, H. Guo, G. Auton, M. B. Shalom, L. A. Ponomarenko, [40] G. Giuliani and G. Vignale, Quantum Theory of the Electron G. Falkovich, K. Watanabe, T.Taniguchi, I. V.Grigorieva, L. Lev- Liquid (Cambridge University Press, Cambridge, 2005). itov, M. Polini, and A. Geim, Nat. Phys. (2017). [41] K. Bedell and C. J. Pethick, J. Low Temp. Phys. 49, 213 (1982). [19] J. Gooth, F. Menges, N. Kumar, V. Süβ, C. Shekhar, Y. Sun, [42] W.E. Lawrence and J. W.Wilkins, Phys. Rev. B 7, 2317 (1973). U. Drechsler, R. Zierold, C. Felser, and B. Gotsmann, Nat. [43] E. I. Kiselev and J. Schmalian, Phys. Rev. B 99, 035430 (2019). Comm. 9 (2018). [44] C. Berthod, J. Mravlje, X. Deng, R. Žitko, D. van der Marel, and A. Georges, Phys. Rev. B 87, 115109 (2013).