Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Higgs Mode in Superconductors

Higgs Mode in Superconductors When a continuous symmetry of a physical system is spontaneously broken, two types of collec- tive modes typically emerge: the amplitude and phase modes of the order-parameter uctuation. For superconductors, the amplitude mode is recently referred to as the \Higgs mode" as it is a condensed-matter analogue of a Higgs boson in particle physics. Higgs mode is a scalar excita- tion of the order parameter, distinct from charge or spin uctuations, and thus does not couple to electromagnetic elds linearly. This is why the Higgs mode in superconductors has evaded experimental observations over a half century after the initial theoretical prediction, except for a charge-density-wave coexisting system. With the advance of nonlinear and time-resolved terahertz spectroscopy techniques, however, it has become possible to study the Higgs mode through the nonlinear light-Higgs coupling. In this review, we overview recent progresses on the study of the arXiv:1906.09401v1 [cond-mat.supr-con] 22 Jun 2019 I. INTRODUCTION AND A BRIEF HISTORICAL OVERVIEW Spontaneous breaking of a continuous symmetry is a fundamental concept of phase transi- tion phenomena in various physical systems ranging from condensed matter to high-energy physics [1]. For instance, a ferromagnetic transition is characterized by the spontaneous breaking of spin rotational symmetry. Likewise, superconductivity is characterized by the spontaneous breaking of U(1) rotational symmetry with respect to the phase of a macro- scopic wavefunction [2]. When a continuous symmetry is spontaneously broken, there emerge two types of col- lective modes in general: uctuations of the phase and amplitude of the order parameter as schematically shown in Figure 1a. The phase mode (also termed Nambu-Goldstone (NG) mode) is primarily a gapless (massless) mode [3{8] as required by the symmetry. An acoustic phonon is such an example that is associated with spontaneous breaking of trans- lational symmetry in a crystal lattice. The amplitude mode, on the other hand, is generally a gapped (massive) mode since its excitation costs an energy due to the curvature of the potential (Figure 1a ). The amplitude mode (especially for superconductors) is often called the Higgs mode according to the close analogy with the Higgs boson in particle physics. It may sound strange that one talks about the Higgs particle in condensed matter systems, but in fact the origin of the idea of Higgs physics can be found in the course of the study of superconductivity [3{9]. Namely, the standard model in particle physics can in some sense be viewed as a relativistic version of the Ginzburg-Landau theory [2], i.e., a low-energy e ective theory of superconductors. The emergence of the Higgs mode is a universal and fundamental phenomenon in systems with spontaneous symmetry breaking. In the case of superconductors where the order parameter couples to gauge elds, the Higgs mode occupies a special status since it is the lowest collective excitation mode (Fig- ure 1b ). The massless phase mode is absorbed into the longitudinal component of electro- magnetic elds, and is lifted up to high energy in the scale of the plasma frequency due to the Anderson-Higgs mechanism [4{6, 9{13]. As a result, the amplitude mode becomes stable against the decay to the phase mode in superconductors. The existence of the amplitude mode in superconductors was suggested by Anderson [5, 14] soon after the development of a microscopic theory of superconductivity by Bardeen, Cooper, and Schrie er (BCS) [15]. In an s-wave (BCS-type) superconductor, the Higgs-mode gap energy ! coincides with the 2 a b <latexit sha1_base64="QHK9fgKuka7iN1RctYFrmwXDUbk=">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</latexit> Phys. Scr. 92 (2017) 024003 R Matsunaga and R Shimano Free energy quasiparticle Anderson-Higgs excitations mechanism Higgs mode Higgs mode <latexit sha1_base64="tkrKJiURdjBMrxZsmPrJbCrt0gw=">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</latexit>Re Nambu-Goldstone NG mode <latexit sha1_base64="IsdAcr1U4kit6BDS167S+m063Cs=">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</latexit>Im (NG) mode <latexit sha1_base64="HlEL285e3sXISJRF4mVDzB64A44=">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</latexit> <latexit sha1_base64="9tneBTFFq1wSKGVGcADajPcqCYM=">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</latexit>0 Figure 1. (a) A schematic picture of the phase mode (blue arrow) and the amplitude mode (red arrow) represented by the Ginzburg–Landau free energy potential. (b) A schematic picture for collective modes and single particle excitation in a symmetry-broken system with the Anderson–Higgs mechanism. the Higgs mode have been demonstrated by a random-phase based on the time-dependent GL theory or the Boltzmann Figure 1. (a) A schematic picture of the Higgs (red) and Nambu-Goldstone (blue) modes rep- approximation theory [37]. equation is not justified [38]. To describe the dynamics of One main difficulty for the observation of the Higgs order parameter in non-equilibrium superconducting states in mode in superconductors is attributed to the fact that the such a nonadiabatic regime, the Anderson’s pseudospin resented on the Mexican-hat free-energy potential as a function of the complex order parameter representation [32] has been used as an useful microscopic Higgs mode does not have a charge nor electric dipole and treatment and nontrivial dynamics of the order parameter has therefore it does not couple directly to the electromagnetic been investigated in a wide range of quench parameters based . (b) A schematic field. This excitation explains why thesp experimental ectrum observation of an of the s-wave superconductor. Due to the Anderson-Higgs on the BCS mean-field theory [38–41]. Subsequently, the Higgs mode has been limited for a long time to the case of subject has been applied to the case of metallic super- Raman spectroscopy for NbSe in which a charge density mechanism, the Nambu-Goldstone mode acquires conducting an energy materials [42 gap –46] in inwhich thequasiparticle order of dis- the plasma fre- wave (CDW) order coexists with the superconducting phase tribution can be suddenly perturbed by an ultrashort laser [48]. Initially the Raman peak observed at the super- pulse. conducting gap energy of 2Δ in this material was interpreted quency ! , while the Higgs mode remains in low energy with an energy gap 2, above which the In the pseudospin model, up or down pseudospins for a as the single-particle excitation across the gap [48]. Soon wavenumber k means both of the two electron states at k and later, Littlewood and Varma have theoretically elucidated that −k being occupied or unoccupied, respectively. While in the the peak corresponds to the scalar excitation of the amplitude quasiparticle excitation continuum overlaps. normal state at T = 0 all the pseudospins are up for |k|<k mode (Higgs mode) [36, 37]. The stability and the visibility and down for |k|>k (k is the Fermi wavenumber), in the F F of the Higgs mode in Raman spectroscopy were further superconducting state the BCS ground state is expressed as a revealed recently in this strong electron-phonon coupling superposition of up and down pseudospins for k. This system in which CDW and superconductivity coexist representation makes the BCS Hamiltonian [49, 50]. By contrast, it has been unclear whether the Higgs x y z where s = ss s is the pseu- H =⋅2, b s k ( ) BCS å kk k k k mode is observable or not in superconductors without such a dospin operator for k, and b is represented as CDW order, and indeed the Raman peak of the Higgs mode superconducting gap energy 2 [5, 16{19] (see also [20, 21]). This is eventually consistent was absent in superconducting NbS which is a material 2 b = -D¢ - D e.1() k() k similar to NbSe but has no CDW order [49]. ε is the energy dispersion measured from the Fermi energy. It has been shown theoretically that the Higgs mode 2 2 2 with Nambu's conjectural sum rule [21, 22]: ! +! = 4 with the NG mode gap energy Δ is the order parameter obtained by the sum of the pseu- appears in the dynamics of a complex order parameter after a H NG dospins: sudden perturbation [34]. Recently, such an artificial control of the pairing interaction has become feasible in the experi- D=D¢+iU D= () ss +i,2() ! = 0. Since the Higgs mode lies at the lower bound of the quasiparticle k k continuum near NG ments in the field of ultracold atomic systems [51], and the k dynamics of symmetry-broken ordered states after a sudden where U is the attractive interaction between the two electron quench of the interaction has gained renewed interests [38– zero momentum, the decay of the Higgs mode to single-particle excitations is suppressed states. s is parallel to b in equilibrium for each k. Then the k k 41]. It should be noted here that free energy is in general well equation of motion, namely the Bogoliubov–de Gennes defined only in a static regime and the time-dependent GL equation, is given by and becomes a much slower power law [16]. For nonzero momentum, the Higgs mode can theory is valid only when the change of the order parameter is slow enough. When the order parameter rapidly changes in ss== iH[],2b´s, (3) kk BCS k k ¶t the timescale of τ =h/Δ, the phenomenological model decay by transferring its energy to quasiparticle excitations. The energy dispersion and the damping rate of the Higgs mode has been derived within the random phase approximation [18, 19]. Provided that the theory of the Higgs particle originates from that of superconductivity and that the Higgs particle has been discovered in LHC experiments in 2012 [23, 24], it would be rather surprising that an experimental observation of the Higgs mode in superconductors, a home ground of Higgs physics, has been elusive for a long time. There are, however, good reasons for that: The Higgs mode does not have any electric charge, electric dipole, magnetic moment, and other quantum numbers. In other words, the Higgs mode is a scalar excitation (which is distinct from, e.g., charge uctuations). Therefore it does not couple to external 3 probes such as electromagnetic elds in the linear-response regime. Another reason is that the energy scale of the Higgs mode lies in that of the superconducting gap, which is in the order of millielectron volts in typical metallic superconductors. To excite the Higgs mode, one needs an intense terahertz (THz) light source (1 THz  4 meV  300 m), which has become available only in the last decade. One exception (and the rst case) for the observation of the Higgs mode in superconduc- tors in the early stage was the Raman scattering experiment for a superconductor 2H-NbSe [25, 26]. It is exceptional in the sense that superconductivity and charge density wave (CDW) coexist in a single material. The Raman peak observed near the superconducting gap energy 2 in 2H-NbSe was rst interpreted as a single-particle excitation across the gap [25, 26]. Soon later, Littlewood and Varma have theoretically elucidated that the peak corresponds to the scalar excitation of the amplitude mode (i.e., the Higgs mode) [18, 19] (see also Reference [27]). A renewed interest in the Raman signal has recently revealed the importance of the coexisting CDW order for the Higgs mode to be visible in the Raman response [28{30]. Indeed, the Raman peak of the Higgs mode has been shown to be absent in a superconducting NbS , which is a material similar to NbSe but has no CDW order 2 2 [28]. It has been a long-standing issue as to whether the Higgs mode can be observed in superconductors without CDW order. On the theoretical side, there have been various proposals for the excitation of the Higgs mode, including a quench dynamics [16, 31{37] as well as a laser excitation [38{44]. It is only after the development of ultrafast pump-probe spectroscopy techniques in the low-energy terahertz-frequency region that a clear observation of the Higgs-mode oscillation has been reported in a pure s-wave superconductor NbN [45] (with `pure' meaning no other long-range order such as CDW). Subsequently, it has been experimentally demonstrated through the THz pump-probe and third harmonic generation (THG) measurements [46] that the Higgs mode can couple to electromagnetic elds in a nonlinear way. In the THG experiment, a resonant enhancement of THG was discovered [46] at the condition 2! = 2 with ! the frequency of the incident THz light, indicating the nonlinear light-Higgs coupling in a two-photon process [42]. Meanwhile, it has been theoretically pointed out that not only the Higgs mode but also charge density uctuation (CDF) can contribute to the THG signal [47]. It was shown that the CDF gives a much larger contribution than the Higgs mode in a clean-limit super- 4 conductor within the BCS mean- eld approximation. While the relative magnitude of the Higgs-mode and CDF contributions to the THG has been under debate [47{50], theoretical progresses have recently been made on the theory of the light-Higgs coupling, showing that both the e ects of phonon retardation [48] and nonmagnetic impurity scattering [51{54] drastically enhance the Higgs-mode contribution to nonlinear optical responses, far exceed- ing the CDF contribution. The experiments have been extended to high-T cuprate superconductors, showing the presence of the Higgs-mode contribution to the pump-probe signal in d-wave superconductors [55]. Recently the THG has also been identi ed in various high-T cuprates [56]. The Higgs mode can be made visible in a linear optical response if dc supercurrent is owing in a superconductor [57]. The infrared activated Higgs mode in the presence of supercurrents has recently been observed in a superconductor NbN [58]. In this review article, we overview these recent progresses in the study of the Higgs mode in superconductors (where the Anderson-Higgs mechanism is taking place), with an empha- sis on the experimental aspects. In a broader context, collective amplitude modes are not limited to superconductors but are ubiquitous in condensed matter systems that may not be coupled to gauge elds (hence without the Anderson-Higgs mechanism). Those include squashing modes in super uid He [59], amplitude modes in bosonic and fermionic conden- sates of ultracold-atom systems [60, 61], amplitude modes in quantum antiferromagnets [62], and so on. A comprehensive review including these topics can be found in Reference [63]. II. NONLINEAR LIGHT-HIGGS COUPLING In this section, we review from a theoretical point of view how the Higgs mode in supercon- ductors couples to electromagnetic elds in a nonlinear way. It is important to understand the mechanism of the nonlinear light-Higgs coupling for observing the Higgs mode in ex- periments. In fact, the Higgs mode does not have a linear response against electromagnetic elds in usual situations, which has been an obstacle for experiments to detect the Higgs mode in superconductors for a long time. 5 A. A phenomenological view Let us rst take a phenomenological point of view based on the Ginzburg-Landau (GL) theory [2], which provides us with a quick look at the nonlinear light-Higgs coupling. For early developments based on the time-dependent GL theory, we refer to References [64{ 68]. In the GL theory, the free energy density f is assumed to be a function of a complex order-parameter eld (r), b 1 2 4  2 f [ ] = f + aj (r)j + j (r)j + j(ir e A) (r)j ; (1) 2 2m where a = a (T T ), a and b are some constants, m and e are the e ective mass and 0 c 0 the e ective charge of the Cooper pair condensate, and A is the vector potential that rep- resents the external light eld (E(t) = @A(t)=@t). The amplitude of the order parameter corresponds to the super uid density (i.e., j j = n ), while the phase corresponds to that of the condensate. The free energy density (1) is invariant under the global U(1) phase rotation (r) ! i' ie (r) e (r) and more generally under the gauge transformation (r) ! e (r), A(r) ! A(r) + r(r) for an arbitrary scalar eld (r). In addition, the GL free energy density (1) is invariant against the particle-hole transformation (r) ! (r). The presence of the particle-hole symmetry (including the time derivative terms in the action) is crucial [27, 63, 69] in decoupling the Higgs amplitude mode from the phase mode (Nambu-Goldstone mode) in electrically neutral systems such as super uid ultracold atoms. Without the particle-hole symmetry, the Higgs mode quickly decays into the phase mode and becomes short-lived. In superconductors (which consist of electrically charged electrons), on the other hand, the phase mode acquires an energy gap in the order of the plasma frequency due to the Anderson- Higgs mechanism. This prevents the Higgs mode from decaying into the phase mode. The particle-hole symmetry itself is realized as an approximate symmetry in the Bogoliubov-de Gennes Hamiltonian, which describes microscopic low-energy physics of superconductors around the Fermi energy at the mean- eld level. When the temperature goes below the critical temperature (i.e., a < 0), the global U(1) symmetry is spontaneously broken, and the system turns into a superconducting state. The order-parameter uctuation from the ground state ( (r) = ) can be decomposed into 6 amplitude H (r) and phase (r) components, i(r) (r) = [ + H (r)]e : (2) Then, the GL potential can be expressed up to the second order of the uctuation as 1 e 1 2 2 2 f = 2aH + (rH ) + A r ( + H ) + : (3) 2m 2m e The rst term on the right hand side of Equation 3 indicates that the Higgs mode has an 1=2 1=2 energy gap (Figure 1b ) proportional to (a) / (T T ) . The microscopic BCS mean- eld theory predicts that the Higgs gap is actually identical to the superconducting gap 2 [5, 16{19]. The phase eld (r) does not have such a mass term, so that at rst sight one would expect to have a massless Nambu-Goldstone mode. However, (r) always appears in the form of (A r=e ) in Equation 3, which allows one to eliminate (r) from the GL 0  0 potential by taking a unitary gauge A = Ar=e . As a result, we obtain (denoting A as A) 2 2 2 1 e e 2 2 0 2 2 f = 2aH + (rH ) + A + A H + : (4) 2m 2m m One can see that the phase mode is absorbed into the longitudinal component of the elec- tromagnetic eld. At the same time, there appears a mass term for photons (the third term on the right hand side of Equation 4) with the mass proportional to e =m (Anderson-Higgs mechanism). Due to this, electromagnetic waves cannot propagate freely inside superconductors but decay exponentially with a nite penetration length (Meissner e ect). From Equation 4, one immediately sees that there is no linear coupling between the Higgs and electromagnetic eld. This is consistent with the fact that the Higgs mode does not have an electric charge, magnetic moment, and other quantum numbers, which has been a main obstacle in observing the Higgs mode by an external probe for a long time. On the other hand, there is a nonlinear coupling term A H in Equation 4, which is responsible for the Higgs mode to contribute in various nonlinear processes. In Figure 2a , we illustrate the diagrammatic representation of the nonlinear light-Higgs coupling, where two photons are coming in with frequencies ! and ! and one Higgs is emitted with frequency ! + ! . 1 2 1 2 This is analogous to the elementary process of the Higgs particle decaying into W bosons that had played a role in the discovery of the Higgs particle in LHC experiments [23, 24]. 7 a b c pump pump <latexit sha1_base64="WqGUujrvJd5IYb6/HOCxgg1RtDg=">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</latexit><latexit sha1_base64="WqGUujrvJd5IYb6/HOCxgg1RtDg=">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</latexit> A<latexit sha1_base64="87zEWZbfw8D7sDpQWGXnBY7M9Zw=">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</latexit> <latexit sha1_base64="GJryKaYbRvDRPnIjY5gpWeresPY=">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</latexit> <latexit sha1_base64="87zEWZbfw8D7sDpQWGXnBY7M9Zw=">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</latexit> <latexit sha1_base64="2AbfLqPhMXj7VY8ewmTjZx30iZM=">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</latexit>H 3<latexit sha1_base64="uFd62FUVMaSs53WnbF4+c2WNxHk=">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</latexit> <latexit sha1_base64="2AbfLqPhMXj7VY8ewmTjZx30iZM=">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</latexit>H <latexit sha1_base64="2AbfLqPhMXj7VY8ewmTjZx30iZM=">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</latexit>H <latexit sha1_base64="87zEWZbfw8D7sDpQWGXnBY7M9Zw=">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</latexit> <latexit sha1_base64="GJryKaYbRvDRPnIjY5gpWeresPY=">AAACynicfVFNTxsxEHW2fIbv9sjFIkLiFO0CEhyBVhWHooJoIFKyQl7vbGLhj5XtBVbW3nrj2v6S/pr+G7xJVJGAOpLl5zfvaWY8Sc6ZsWH4txF8mJtfWFxabq6srq1vbG59vDGq0BQ6VHGluwkxwJmEjmWWQzfXQETC4Ta5/1znbx9AG6bkD1vmEAsykCxjlFhPXZ3ebbbCdjgK/BZEE9BCk7i822r86aeKFgKkpZwY04vC3MaOaMsoh6rZLwzkhN6TAfQ8lESAid2o0wrveibFmdL+SItH7GuHI8KYUiReKYgdmtlcTb6X6xU2O44dk3lhQdJxoazg2Cpcj41TpoFaXnpAqGa+V0yHRBNq/ef8t+epFurC2mRmmrVP9TQz5ECTfMhozX4B/1Marv1gin/1UmdG0FTu+8V15agwZeVE5eS7Yv8cwCtHt3bA0z+HhEeqhCAydf1EVL0o9rfi6diCW1HV9GuOZpf6Ftzst6ODdnR12Do5myx8CW2jHbSHInSETtA5ukQdRBGgZ/QL/Q6+BTooAzeWBo2J5xOaiuDnCyoy5qw=</latexit>A probe probe <latexit sha1_base64="pq9cDiS1JYBZf4U/jWRZ5OVcPTw=">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</latexit><latexit sha1_base64="pq9cDiS1JYBZf4U/jWRZ5OVcPTw=">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</latexit> Figure 2. (a) Diagrammatic representation of the nonlinear light-Higgs coupling A H in Equa- tion 4. (b) Third harmonic generation mediated by the Higgs mode. (c) Pump-probe spectroscopy mediated by the Higgs mode. Due to the nonlinear interaction between the Higgs mode and electromagnetic elds, one can induce a nonlinear current given by @F ie e y y y j = = [ r (r ) ] A : (5) @A 2m m Again, we expand around the ground state in Equation 5 and collect leading terms in H , which results in 2e j = AH: (6) This is nothing but the leading part of the London equation j = (e n =m )A (let us recall that the amplitude of the order parameter corresponds to the super uid density, j j = n ). When a monochromatic laser with frequency ! drives a superconductor, the amplitude of the order parameter oscillates with frequency 2! due the nonlinear coupling A H (Fig- ure 2a ). Together with the oscillation of A with frequency ! in Equation 6, the induced nonlinear current shows an oscillation with frequency 3!. As a result, one obtains the third harmonic generation mediated by the Higgs mode (Figure 2b ). Since the Higgs mode has an energy 2 at low momentum, one can resonantly excite the Higgs mode by a THz laser with a resonance condition 2! = 2. Given that the nonlinear current is proportional to the Higgs-mode amplitude (Equation 6), the third harmonic generation can be resonantly enhanced through the excitation of the Higgs mode. The Higgs-mode resonance in third harmonic generation has been experimentally observed for a conventional superconductor [46]. Another example of nonlinear processes to which the Higgs mode can contribute is a pump-probe spectroscopy where pump and probe light is simultaneously applied (Fig- ure 2c). In this case, photons with di erent frequencies ! and ! are injected, and pump probe 8 σ C D a z b BCS state normal state (T=0) 1 1 k k F x F Figure 3. (a) Momentum distribution of a superconducting state (above) and its pseudospin representation due to Anderson (below). (b) Pseudospin precession induced by a laser eld. Figures are taken from Reference [46]. those with the same respective frequencies are emitted. There are three possibilities of the frequency carried by the Higgs mode: ! ! and ! ! = ! ! = 0. pump probe pump pump probe probe If one uses a THz pump and optical probe, i.e., ! . 2 and !  2, the rst two pump probe channels do not contribute since the excitation of the Higgs mode is far o -resonant. The remaining zero-frequency excitation channel contributes to the pump-probe spectroscopy. This process has been elaborated in the experimental study of the Higgs mode in d-wave superconductors [55]. B. A microscopic view In the previous subsection, we reviewed the phenomenology for the nonlinear light-Higgs coupling using the Ginzburg-Landau theory. However, strictly speaking, the application of the Ginzburg-Landau theory to nonequilibrium problems is not microscopically justi ed in the case of gapped superconductors [70{72]. The reason is that in a usual situation one cannot neglect the e ect of quasiparticle excitations whose relaxation time is longer than the time scale of the order-parameter variation. Moreover, the Higgs mode and quasipar- ticle excitations are energetically degenerate at low momentum (Figure 1b ), so that it is inevitable to excite quasiparticles at the same time when one excites the Higgs mode. This motivates us to take a microscopic approach for further understanding. Let us adopt the time-dependent BCS or Bogoliubov-de Gennes equation, which is e- y y y ciently represented by the Anderson pseudospin  = h  i [5]. Here = (c ; c ) k k k# k k k" is the Nambu spinor,  = ( ;  ;  ) are Pauli matrices, c (c ) is the creation (annihila- x y z k Distribution Distribution tion) operator of electrons with momentum k and spin , and hi denotes the statistical average. Physically, the x and y components of the Anderson pseudospin correspond to the real and imaginary parts of the Cooper pair density, respectively, and the z component corresponds to the momentum occupation distribution of electrons (Figure 3a ). With this notation, the BCS mean- eld Hamiltonian is written as H = 2 b (t)  ; (7) BCS k k 0 00 b (t) =  (t); (t); ( +  ) : (8) k k+A(t) kA(t) 0 00 Here b (t) is a pseudomagnetic eld acting on pseudospins,  and  are the real and imaginary parts of the superconducting gap function, respectively, and  is the band dis- persion of the system. The equation of motion for the pseudospins is given by the Bloch-type equation, (t) = 2b (t)  (t); (9) k k k @t 0 00 V x supplemented by the mean- eld condition  (t) + i (t) = [ (t) + i (t)] with V k k k the attractive interaction and N the number of k points. The time evolution of the super- conducting state is thus translated into the precession dynamics of Anderson pseudospins (Figure 3b ). The coupling to the electromagnetic eld appears in the z component of the pseudo- z z magnetic eld, b = ( +  ). If we expand b in terms of A(t), we obtain k+A(t) kA(t) k k 1 @ z k 4 b =  + A A + O(A ). The linear coupling term vanishes, which is consistent k i j k ij 2 @k @k i j with the phenomenology that we have reviewed in the previous subsection. The leading nonlinear coupling term is quadratic in A(t). In Figure 4a , we show a numerical result for the time evolution of the gap function (t) when the system is driven by a multi-cycle electric- eld pulse. One can see that the 2! oscillation of the gap is generated during the pulse irradiation, after which a free gap oscillation continues with the frequency 2 and the 1=2 amplitude slowly damping as t [16]. For a monochromatic wave A(t) = A sin !t, one can solve the equation of motion semi- analytically by linearizing with respect to the quadratic coupling A (t)A (t) around the i j equilibrium solution. The result shows that the amplitude of the gap function varies as [42] (t) / j2! 2j cos(2! ) (10) 10 1 (a) (b) 1.5 E(t) -1 1.4 Δ(t) 1.3 0 10 20 30 40 50 60 0 1 2 2ω/2Δ Figure 4. (a) A numerical result for the time evolution of the superconducting gap (t) driven by an electric eld pulse E(t) with the frequency ! = 2=5  1:26. Here we take a two dimen- sional square lattice with a bandwidth 8 at half lling. The interaction is V = 4, and the initial temperature is T = 0:02. The polarization of the electric eld is parallel to x axis. The left (right) vertical axis represents the value of (t) (E(t)) as indicated by the arrows. (b) The amplitude of the 2! oscillation of the superconducting gap as a function of 2!=2. (b) is reproduced from Reference [42]. for !  , where  is a phase shift. The induced oscillation frequency is 2!, re ecting the nonlinear coupling between the Higgs mode and electromagnetic elds, A H . The oscillation amplitude diverges when the condition 2! = 2 is ful lled (Figure 4b ). This is much the same as the spin resonance phenomenon: The forced collective precession of the pseudospins with frequency 2! due to the nonlinear coupling resonates with the Higgs mode whose energy coincides with 2. The power of the divergence in Equation 10 is smaller than that of a resonance with an in nitely long-lived mode ( j2! ! j ). The reason is that the 2 2 precession gradually dephases due to the momentum-dependent frequency ! = 2  + for each pseudospin so that the average of the pseudospins (t) =  (t) shows a k k power-law damping of oscillations. The phase shift  exhibits a jump as one goes across the resonance by changing the drive frequency !. In the case of the square lattice and the polarization of the electric eld parallel to the diagonal direction in the xy plane, for example, the current is given by j(t) / A(t)(t) [42, 46], which corresponds to the phenomenological relation j(t) / A(t)H (t) (Equation 6). As Δ(t) E(t) |δΔ | 2 ω expected, the Higgs mode induces a resonant enhancement of the third harmonic generation. Later, it has been pointed out [47] that the current relation is valid only in rather special situations such as the square lattice and the polarization parallel to the diagonal direction. In general, the BCS mean- eld treatment (in the clean limit) suggests that the Higgs- mode contribution to the third harmonic generation is subdominant as compared to that of individual quasiparticle excitations. Since the energy scales of the Higgs mode and the quasiparticle pair excitations are in the same order ( 2, see Figure 1b ), the competition between the two contributions always matters. This point will be further examined in the next subsection. C. Impurity and phonon assisting As we have seen in the previous subsection, the Higgs-mode contribution to the third harmonic generation is generally subdominant in the BCS clean limit. However, there is a growing understanding from recent studies [48, 49, 51{54] that if one takes into account e ects beyond the BCS mean- eld theory in the clean limit (such as impurity scattering and phonon retardation) the light-Higgs coupling strength drastically changes. To understand the impact of these e ects, let us consider the linear optical response. The current induced by a laser eld is decomposed into the paramagnetic and diamagnetic components j = j + j with [73] para dia j = hc c i; (11) para k @k k y j = A hc c i: (12) dia i k @k@k ki As we have seen in the pseudospin picture in the previous subsection, only the diamagnetic coupling contributes to the optical response in the mean- eld level with no disorder, resulting in Re (!) / (!) and Im (!) / 1=! for the superconducting state. For nite frequency (! 6= 0), the real part of the optical conductivity vanishes in the BCS clean limit. This does not necessarily mean that the real part of the optical conductivity is suppressed in real experimental situations. For example, in superconductors with a disorder the real part of the optical conductivity is nonzero and not even small for ! 6= 0. In fact, a superconductor NbN used in the experiment of the third harmonic generation turns out to be close to the 12 dirty regime (  2, where is the impurity scattering rate) as con rmed from the measurement of the optical conductivity [45]. (3) TABLE I. Relative order of magnitudes of the third-order current j in general situations [53]. mode channel clean ! dirty 2 2 Higgs dia (A ) (= ) 2 2 2 para (p A) ( = ) ! ( = ) F F Quasiparticles dia (A ) 1 2 2 2 para (p A) ( = ) ! ( = ) F F In much the same way as the optical conductivity, the nonlinear response intensity is strongly a ected by the impurity e ect. The dependence of the relative order of magnitudes of the third-order current is summarized in Table I, where  is the Fermi energy and the (3) unit of j is taken such that the diamagnetic-coupling contribution of the quasiparticles is set to be of order 1 (which does not signi cantly depend on ). In the BCS clean limit ( ! 0), we have only the diamagnetic coupling, for which the Higgs-mode contribution is subleading in the order of (= ) compared to the quasiparticles. However, in the presence of impurities ( 6= 0) the paramagnetic coupling generally emerges to contribute to the third 2 2 2 harmonic generation. Its relative order of magnitude changes as ( = ) ! ( = ) from F F the clean to dirty limit. The maximum strength of the paramagnetic coupling is achieved around  , where the relative order reaches ( =) for both the Higgs mode and quasiparticles. Thus, the impurity scattering drastically enhances the light-Higgs coupling: in the clean limit the Higgs-mode contribution is subleading to quasiparticles, whereas in the dirty regime it becomes comparable to or even larger than the quasiparticle contribution. The precise ratio between the Higgs and quasiparticle contributions may depend on details of the system. Recent studies [52{54] suggest that in the dirty regime the Higgs-mode contribution to THG is an order of magnitude larger than the quasiparticle contribution. An enhancement of the paramagnetic coupling also occurs for strongly coupled superconductors due to the phonon retardation e ect [48]. 13 D. Further developments So far, we have reviewed the phenomenological and mean- eld treatments of the nonlinear light-Higgs coupling for conventional s-wave superconductors with or without disorder. For strongly coupled superconductors with an electron-phonon coupling constant  & 1 (which is the case for NbN superconductors [74{76]), one has to take account of strong correlation e ects. One useful approach for this is the nonequilibrium dynamical mean- eld theory [77], which takes into account local dynamical correlations by mapping a lattice model into a local impurity model embedded in an e ective mean eld. With this, the Higgs mode in strongly coupled superconductors modeled by the Holstein model has been analyzed [42, 48, 78, 79]. A closely related nonequilibrium Keldysh method has been employed to study the Higgs mode in a time-resolved photoemission spectroscopy [43, 80] and time-resolved optical conductivity [81]. Another approach is to use the gauge-invariant kinetic equation [82{85], where the shift of the center-of-mass momentum of Cooper pairs (\drive e ect") caused by a laser pulse has been emphasized. For multi-band superconductors with multiple gaps such as MgB , there are not only Higgs modes corresponding to amplitude oscillations of multiple order parameters but also the so-called Leggett mode [86], which corresponds to a collective oscillation of the relative phase between di erent order parameters. Based on the mean- eld theory, possible laser excitations of the Higgs and Leggett modes in multiband superconductors have been studied [53, 87{90]. For non-s-wave superconductors, one can expect even richer dynamics of the order pa- rameters. For example, d-wave superconductors allow for various di erent symmetries of amplitude modes [91]. The quench dynamics of the d-wave [92] and p + ip-wave [93] Higgs modes have been explored. For recent experimental progresses on pump-probe spectro- scopies and third harmonic generation in d-wave superconductors, we refer to Section III C. It has also been proposed [94] that by measuring Higgs modes for various quench symmetries one can obtain information on the symmetry of the gap function (\Higgs spectroscopy"). 14 III. EXPERIMENTS In this section, we overview recent progresses on experimental observations of the Higgs mode in superconductors. Those include Raman experiments in CDW-coexisting super- conductors 2H-NbSe and TaS (Section III A), THz spectroscopies and third harmonic 2 2 generation in a pure s-wave superconductor NbN (Section III B) and in high-T cuprates (Section III C), and THz transmittance experiments in NbN under supercurrent injection (Section III D). A. Higgs mode in a superconductor with CDW The pioneering work on the observation of the Higgs mode in superconductors dates back to 1980 when a Raman experiment was performed in 2H-NbSe , a transition metal dichalco- genide in which superconductivity coexists with CDW [25, 26]. A new peak was observed below T , distinct from the amplitude mode associated with the CDW order. Although it was initially recognized as a pair breaking peak, soon after it was identi ed as the collec- tive amplitude mode of the superconducting order (i.e., the Higgs mode) [18, 19] (see also [27, 63]). A renewed Raman experiment has been conducted recently, revealing the transfer of the oscillator strength from the amplitude mode of CDW (amplitudon) to the Higgs mode [28]. Subsequently the Raman experiment under hydrostatic pressure has been performed in 2H-NbSe [30] (Figure 5). With applying the hydrostatic pressure the CDW order is suppressed and concomitantly the Raman peak identi ed as the Higgs mode was shown to disappear with leaving only the Cooper pair breaking peak. A more clear separation between the in-gap Higgs mode and the pair breaking peak was recently reported in a similar CDW- coexisting superconductor, TaS [95]. These results indicate that the existence of CDW plays an important role on the visibility of the Higgs mode in the Raman spectrum. A theoretical study has demonstrated that, due to the coupling of the superconducting order to the coexisting CDW order, the Higgs mode energy is pushed down below the superconducting gap 2 [29]. Accordingly, the Higgs mode becomes stable due to the disappearance of the decay channel into the quasiparticle continuum. 15 4 0.009 NbSe P>P 2 c 2H-NbSe NbS 0 GPa 2 2 0.04 0.006 2g 0.003 T<T 0.03 012 3 ω / 2∆ SC 0.02 P>P T>T c c (c) P>P T<T c c P<P T>T c c P<P T<T c c 0.01 0.00 020 40 60 80 100 -1 Raman shift (cm ) FIG. 2. Collapse of the superconducting Higgs mode in the pure superconducting state of 2H-NbSe , measurements and Figure 5. Raman spectra in E symmetry of 2H-NbSe measured at various temperatures and 2g 2 theoretical predictions. (a) Raman spectra in the E symmetry measured at various (P,T) positions as identified Fig. 1(a): in 2g the coexisting SC+CDW (green) and pure CDW (blue) states at ambient pressure and in the pure SC (brown) and paramagnetic pressures. Under the ambient pressure (P < P ) and in the superconducting phase (T < T ) c c (red) state at high pressure. Both the CDW amplitudon and the SC Higgs modes disappear at high pressure. A small Cooper- pairs breaking peak remains at 2 .2 is marked by the grey band ranging from 2 measured by STM [39]atambient SC SC SC (green), a sharp peak identi ed as the Higgs mode appears below the superconducting gap 2  20 pressure to the value we extrapolate at high pressure accordingly to the increase of T with pressure [26]. Inset: Raman spectra 1 in the pure SC state of 2H-NbSe (above 4 GPa) and non-CDW NbS (0 GPa) versus the Raman shift normalized to 2 2 2 SC cm marked by the gray vertical line. In the CDW phase (T > T ) (blue), only the CDW [39, 40]. (b) Theoretical Raman responses calculated in a microscopic model (see text and Appendix A) in the four phases (SC+CDW, CDW, SC,PM) for comparison with the experimental spectra in (a). t is the hopping term. The parameters amplitude mode is observed around 40 cm . Under the pressure (P > P ) where CDW collapses, are: /t=0.025, g /t=0.14 and 0.12 in the SC+CDW and SC phases, respectively. The spectra are well-reproduced SC CDW in all phases. (c) Raman response of 2H-NbSe in the pure superconducting phase above P in the E and A symmetries. 2 c 2g 1g only a pair breaking peak is discerned at T < T (brown), and no peak is identi ed at T > T c c The black line is the theoretical Raman response of a Cooper-pairs breaking peak in a two-gaps (or anisotropic gap) s-wave superconductor in the BCS regime with an additional electronic background . The form of is : (!)= a!/ b + c! .It (red). Inset: Raman spectra in the superconducting state without CDW in 2H-NbSe (above 4 barely a↵ ects the shape of the Cooper-pairs breaking peak. GPa) and non-CDW-coexisting NbS (0 GPa), both of which show only the pair breaking peak. of the SC peak above P , i.e. a Cooper-pairs breaking IV. COMPARISON WITH A MICROSCOPIC Frequency is normalized by 2. Figure is taken from Reference [30]. MODEL peak, with insight into the energy scale of the supercon- ducting gap. Consistently with this assignment, in the A symmetry there is no signature of the pure super- 1g In 2H-NbSe the phonon coupled to the CDW be- conducting state reached in 2H-NbSe above the critical 2 longs to an acoustic branch, so the single-phonon mode is B. Higgs mode in a pure s-wave superconductor pressure, due to Coulomb screening e↵ ect [48–50](see not visible as a finite-energy peak in q⇠ 0 Raman spec- Fig. 2(c)). troscopy above T .Below T the intermediate CDW CDW electron-hole excitations which couple directly to light 1. Nonadiabatic quench with a single-cycle THz pump pulse allow to make the phonon mode at Q Raman visi- CDW ble at q = 0. This gives rise to the soft phonon modes at ⇠ 40 cm . In a general approach, the Raman response One way to excite the Higgs mode is a nonadiabatic quench of the superconducting state, below T can be schematically written as CDW which can be induced by, for example, a sudden change of the pairing interaction that ph ”(!)= Z (T, ) (1) eff CDW 2 2 2 2 The disappearance of the sharp SC mode below 2 SC (! ⌦ (T )) + generates the order-parameter dynamics, ph in the pure superconducting phase demonstrates unam- biguously its intimate link with the coexisting charge- where the soft mode frequency ⌦ (T ) and damping 0 ph j(t)j cos(2 t + ) density-wave order. These findings, consisten1 tly with the are both determined by the CDW amplitude fluctuations ' 1 + a p ; (13) theory discussed below, support the Higgs type assign- and the prefactor Z ⇠ grows proportionally to eff t CDW ment of the sharp SC mode below P . the CDW order parameter[4]. The frequency ⌦ (T ) also c 0 χ’’ (a.u.) χ’’ (u.a.) b 0.8 T=4 K 0.6 nJ/cm 9.6 8.5 0.4 7.9 7.2 6.4 5.6 0.2 4.8 2Δ 4.0 0 0.0 5 10 -4 -2 0 2 4 6 8 t (ps) Pump Energy ( nJ /cm ) pp Figure 6. Higgs-mode oscillation after the nonadiabatic excitation of quasiparticles in NbN in- duced by a monocycle THz pump. (a) The temporal evolution of the change of the probe electric eld, E , as a function of the pump-probe delay time t at various pump intensities. The probe pp solid curves represent the results tted by a damped oscillation with a power-law decay. (b) The oscillation frequency f obtained from the ts and the asymptotic gap energy 2 as a function of the pump intensity. Figures are taken from Reference [45]. at long time. Here 2 is the long-time asymptotic superconducting gap, a is some constant, and  is a phase shift. The behavior (Equation 13) has been theoretically derived on the basis of the time-dependent BCS mean- eld theory [16, 33{35]. A similar dynamics has also been investigated after a laser-pulse excitation [38{41, 43, 44]. Despite intensive theoretical studies on the Higgs mode in superconductors, the observa- tion of the Higgs mode has long been elusive except for the case of 2H-NbSe , until when the ultrafast THz-pump and THz-probe spectroscopy was performed in a conventional s- wave superconductor Nb Ti N [45]. In this experiment, quasiparticles are instantaneously 1x x injected at the superconducting-gap edge to quench the superconducting order parameter nonadiabatically. A strong single-cycle THz pump pulse with the center frequency around 1 THz ( 4 meV) was generated from a LiNbO crystal by using the tilted-pulse-front method E (t =t) (arb. units) probe gate 0 f (THz) [96{98], in order to excite high density quasiparticles just above the gap of Nb Ti N thin 1x x lms with 2 = 0.72-1.3 THz ( 3.0-5.4 meV) at 4 K. The details of the THz pump-THz probe scheme were given in Ref. [99]. The subsequent dynamics of the superconducting order parameter was probed by a weaker probe THz pulse in transmission geometry. Fig- ure 6 shows pump-probe delay dependence of the transmitted probe THz electric eld at a xed point of the probe pulse waveform which re ects the order parameter dynamics. A clear oscillation is identi ed after the pump with the oscillation frequency given by the asymptotic value of the superconducting gap 2 caused by the quasiparticle injection. The decay of the oscillation is well tted by a power-law decay predicted for the Higgs-mode oscillation. With increasing the pump intensity, the oscillation period shows a softening, as the asymptotic value of 2 becomes smaller with higher quasiparticle densities. By recording the probe THz waveform at each pump-probe delay, one can extract the dynamics of the complex optical conductivity (!) =  (!) +i (!) (not shown). The spectral-weight 1 2 oscillation with frequency 2 was clearly identi ed both in the real and imaginary parts [100], the latter of which re ects the oscillation of the super uid density (i.e., the amplitude of the superconducting order parameter). The intense single-cycle THz pump pulse was crucial to excite the quasiparticles instanta- neously at the gap edge which acts as a nonadiabatic quench of the order parameter, thereby inducing a free oscillation of the Higgs mode. On the contrary, when a near-infrared optical pulse was used as a pump, a relatively slow rise time ( 20 ps) was observed for the suppres- sion of the order parameter, though dependent on the excitation uence [99, 101]. In this case, the photoexcited hot electrons (holes) emit a large amount of high frequency phonons (! > 2) (HFP) which subsequently induces the Cooper pair-breaking until quasiparticles and HFP reach the quasiequilibrium. This process was well described by the Rothwarf- Taylor phonon-bottleneck model [101], and usually takes a time longer than (2) , which breaks the nonadiabatic excitation condition needed for the quench experiment. It should be noted here that even if one uses the near-infrared or visible optical excitation, the in- stantaneous quasiparticle excitation can occur if the stimulated Raman process is ecient [102]. The decay of the Higgs mode is also an important issue. Since the Higgs-mode energy 2 coincides with the onset of quasiparticle continuum, the Higgs mode inevitably decays into the quasiparticle continuum even without any collisions [16, 17]. However, this does not 18 mean that the Higgs mode is an overdamped mode. In fact, the decay due to the collisionless 1=2 energy transfer is expected to be proportional to t in the BCS superconductor, ensuring a much longer lifetime compared with the oscillation period. The energy dispersion of the Higgs mode for nite wavenumber q has been calculated as 2 2 v  v F 2 ! ' 2 + q i q (14) 12 24 where v is the Fermi velocity [19]. From Equation 14 one can estimate that as the wavenum- ber q increases and the mode energy enters into the quasiparticle continuum, the lifetime of Higgs mode is shortened and turns into an overdamped mode. For the case of NbN,   0.6 THz and v  2 10 m/s [76] and the imaginary part exceeds the real part in Equation 14 at q > =v  0.3 (m) . In the single-cycle THz-pump experiment in a thin lm of NbN [45], the value of in-plane wavenumber q estimated from the spot size of the pump ( 1 mm) is at most  1 (mm) , which is far smaller than =v . In such a small q limit, the Higgs mode is considered as a well-de ned collective mode [19]. 2. Multicycle THz driving with subgap frequency and third harmonic generation To excite the Higgs mode in an on-resonance condition, one can use a narrowband mul- ticycle THz pump pulse with the photon energy tuned below the superconducting gap 2. With this, the amplitude of the superconducting order parameter was shown to oscillate with twice the frequency of the incident pump-pulse frequency ! during the pulse irradia- tion (Figure 7a -d ) [46]. The observed 2! modulation of the order parameter was attributed to the forced oscillation of the Higgs mode caused by electromagnetic elds. In fact, the time-dependent Ginzburg-Landau theory shows that there is a coupling term between the Higgs (H ) and eld (A) as expressed by A H (Section II A), which results in the 2!-order parameter modulation. The nonlinear electromagnetic response of superconductors has been studied in the frequency range far below 2 and near T where the Ginzburg-Landau theory is applicable [64, 70, 103{105]. In the theory developed by Gor'kov and Eliashberg (GE), a closed set of equations for the order parameter  and the eld A was derived, which predicts the 2! modulation of  and the THG from the induced supercurrent j / A [70]. The GE theory was experimentally demonstrated through the observation of THG of a microwave eld in a paramagnetically doped superconductor [104]. 19 b Pump 2ω=0.6 THz WGP NbN/MgO WGP BPF 4 8 12 16 Probe Temperature (K) c e 10 K pump 15.5 K ω=0.6 THz |E | -2 pump 0 2 4 6 8 10 Delay Time (ps) -4 14 K 0 1 2 14.5 K ω>2Δ(T) Frequency (THz) 14.8 K 15.5 K ω<2Δ(T) 13 K 12 K 0.1 10 K 4 K 0.01 0 2 4 6 8 10 2 3 4 Pump-Probe Delay Time t (ps) E (kV/cm) pp pump Figure 7. Forced Higgs oscillation and third harmonic generation from a superconducting NbN lm under the multicycle THz pump. (a) Schematic setup for the multicycle THz pump and THz probe spectroscopy. BPF: a bandpass lter, WGP: a wire grid polarizer. (b) Temperature dependence of the superconducting gap energy. Horizontal line indicates the center frequency of the pump pulse, ! = 0:6 THz. (c) Waveform of the pump THz electric eld E and the squared pump one, jE j . (d) The change in the transmitted probe THz electric eld E as a function of pump probe the pump-probe delay time t at the tempareture range ! > 2(T ) (top panel) and ! < 2(T ) pp (bottom panel). Increase of E corresponds to the reduction of the order parameter. (e) Power probe spectra of the transmitted pump THz pulse above and below T = 15 K. (f) THG intensity as a function of the pump THz eld strength. Figures are taken from Reference [46]. I |E | pump δE (arb. units) probe Intensity (arb. units) Intensity (arb. units) 2Δ(T) (THz) a c 2ω=1.6 THz 1.2 THz 1.0 THz 0.8 THz Temperature (K) Tempreature (K) 0.8 THz 0.6 THz 0.5 THz 0.4 THz 2ω/2Δ(T) Temperature (K) Figure 8. Temperature dependence of THG signal from NbN. (a) Temperature dependence of the supercondcuting gap 2(T ). Twice the THz pump frequencies 2! are shown by horizontal lines. (b) Measured THG intensities at each frequencies as a function of temperature. (c) THG intensities normalized by the temperature dependent internal eld inside the superconducting lm (3) 2 and the transmittance of the THG signal, which corresponds to the susceptibility of THG, j j . (3) 2 4 (d) Further normalized value of j j by the temperature dependent gap (T ) as a function of normalized frequency 2!=2(T ). Relative intensities between di erent frequencies are arbitrary. Data are taken from Reference [100]. The THz pump experiments have shed a new light on the investigation of the nonlinear electromagnetic response of superconductors at the gap frequency region. Subsequent to the observation of the 2! modulation of the order parameter, namely the forced oscillation of the Higgs modes, the THG was observed under the irradiation of a subgap multicycle THz pump pulse in NbN thin lms as shown in Figure 7e , f . Importantly, the THG was shown to be resonantly enhanced when a condition 2! = 2 is satis ed [46]. The temperature de- pendence of THG intensity for various frequencies in a NbN lm is shown in Figure 8: the temperature dependence of the gap 2 and the incident frequencies (Figure 8a ), the raw THG Intensity (arb. units) 2Δ (THz) (arb. units) / Δ (arb. units) data of THG intensities (Figure 8b ), those normalized by the internal eld inside the super- (3) 2 conducting lm which is proportional to the third-order susceptibility j j (Figure 8c), and the THG intensity further normalized by the proportional factor of the THG intensity (T ) versus the frequency normalized to the gap, 2!=2(T ) (Figure 8d ). These results clearly show the two-photon resonances of the Higgs mode with the eld. As previously described in Section II C, however, it was theoretically pointed out that there is a contribution also from the quasiparticle excitation (or called charge density uctu- ations) in THG, which also exhibits the 2! = 2 resonance [47]. Within the BCS mean- eld approximation in the clean limit, the quasiparticle term was shown to exceed the Higgs mode contribution in THG. At the same time, it was predicted that the THG from the quasipar- ticle term should exhibit polarization dependence with respect to the crystal axis while the Higgs term is totally isotropic in a square lattice [47]. In the experiments in NbN, the THG intensity was shown to be independent from the incident polarization angle with respect to the crystal orientation and notably the emitted THG does not have orthogonal components with respect to the incident polarization of the fundamental (!) wave [49]. Such a totally isotropic nature of THG cannot be well accounted for by the CDF term even if one consid- ers the crystal symmetry of NbN [50] and indicates the Higgs mode as the origin of THG. Recently, the importance of p  A term on the coupling between the Higgs mode and the gauge eld was elucidated. In particular, the p  A term contribution is largely enhanced when one consider the phonon retardation e ect beyond the BCS approximation [48] and more prominently when one considers the nonmagnetic impurity scattering e ect [52{54], dominating the quasiparticle term by an order of magnitude as described in Section II C. This paramagnetic term was also shown to dominate the diamagnetic term, although its ratio depends on the ratio between the coherence length and the mean free path. The NbN superconductor is in the dirty regime as manifested by a clear superconducting gap struc- ture observed in the conductivity spectrum, and the paramagnetic term should signi cantly contribute to THG. C. Higgs mode in high-T cuprates Higgs modes can exist not only in s-wave superconductors but also in unconventional superconductors. For example, in d-wave superconductors such as high-T cuprates one can 22 Figure 9. Forced oscillation of Higgs mode in an optimally doped (OP90) Bi Sr CaCu O 2 2 2 8+x measured by THz pump and optical re ection probe experiment. Temperature dependences of re ectivity change R=R as a function of pump-probe delay time in (a) A and (b) B compo- 1g 1g nents. Red dashed lines indicate T . (c) The A and B components against the pump-probe c 1g 1g delay time at 10 K. Fitting curves are also shown by solid lines. (d) Temperature dependences of the A decaying component (blue) (incoherent quasiparticle excitation), the A oscillatory 1g 1g component (red) (forced Higgs oscillation), and the B oscillatory component (green) (likely the 1g charge density uctuation). Figures are taken from Reference [55]. expect various types of collective modes, including not only the totally isotropic oscillation of the gap function in relative momentum space (A mode) but also those that oscillate 1g anisotropically (e.g., A , B , B modes) [91]. These are reminiscent of Bardasis-Schrie er 2g 1g 2g mode for the case of s-wave superconductors [106]. Possible collective modes have been classi ed based on the point group symmetry of the lattice [91], and the relaxation behavior of the A mode induced by a quench has been discussed [92]. A theoretical proposal has 1g been made to observe the Higgs mode in a d-wave superconductor through a time- and angle-resolved photoemission spectroscopy [80]. 23 The observation of the Higgs mode in d-wave superconductors was recently made in THz- pump and optical-probe experiments in high-T cuprates, Bi Sr CaCu O (Bi2212) [55]. c 2 2 2 8+x An oscillatory signal of the optical re ectivity that follows the squared THz electric eld was observed, which is markedly enhanced below T as shown in Figure 9. This signal was interpreted as the THz-pump induced optical Kerr e ect, namely the third-order nonlinear e ect induced by the intense THz pump pulse. In the Bi2212 system, both A and B 1g 1g symmetry components with respect to the polarization-angle dependence (which should be distinguished from the relative-momentum symmetry classi cation mentioned above) were observed in the THz-Kerr signal. The doping dependence shows that the A component 1g is dominant in all the measured samples from under to near-optimal region. From the comparison with the BCS calculation of the nonlinear susceptibility, the A component was 1g assigned to be the d-wave Higgs-mode contribution. The observed oscillatory signal corresponds to the 2! modulation of the order parameter as observed in an s-wave NbN superconductor under the subgap multicycle THz pump irradiation. Accordingly, like in the case of s-wave systems, one can expect the THG signal from high-T cuprates. Recently, THG was indeed observed ubiquitously in La Sr CuO , c 1:84 0:16 4 DyBa Cu O , YBa Cu O , and overdoped Bi Sr CaCu O [56] by using 0.7 THz 2 3 7x 2 3 7x 2 2 2 8+x coherent and intense light source at the TELBE beamline at HZDR. D. Higgs mode in the presence of supercurrents So far, we have seen the Higgs mode excitation by the nonadiabatic quench with instan- taneous quasiparticle injection or through the quadratic coupling with the electromagnetic wave. Recently, it has been theoretically shown that when there is a condensate ow (i.e., supercurrent), the Higgs mode linearly interacts with electromagnetic waves with the elec- tric eld polarized along the direction of the supercurrent ow [57]. Namely, the Higgs mode should be observed in the optical conductivity spectrum at the superconducting gap ! = 2 under the supercurrent injection. This e ect is explained by the momentum term 2 2 in the action, S / Q (t)j(t)j dtdr, where Q(t) = Q +Q (t) is the gauge-invariant mo- mentum of the condensate, Q is the dc supercurrent part, and Q (t) = Re[Q exp(i t)] is the time-dependent part driven by the ac probe eld, respectively. (t) =  + (t) is the time-dependent superconducting order parameter. The action S includes the integral of 24 0 σ (ω) / σ 1 N 0 -1 -1 σ (ω) (Ω cm ) 200 6×10 ab Parallel, 5 K 2.6 A 2.5 A 2.3 A 1.9 A 0 A 0 0 2Δ MB -50 4 5 6 7 8 Photon Energy (meV) 200 0.03 0.5 Perpendicular, 5 K 0.4 2.6 A 0.02 2.3 A 0.3 1.9 A 0 A 0.2 0.01 2Δ 0.1 Parallel -0.1 Perpendicular -50 -0.01 4 5 6 7 8 4 5 6 7 8 Photon Energy (meV) Photon Energy (meV) Figure 10. (a) Schematic view of the terahertz transmittance experiments under supercurrent injection. The change of the optical conductivity induced by supercurrent injection taken with the THz probe electric eld (b) parallel or (c) perpendicular to the current direction. The optical conductivity spectra measured without the current is also plotted in (b). The dashed line represents a t by the Mattis-Bardeen model [107, 108], and the superconducting gap estimated from the t is indicated by the vertical arrow. (d) Theoretically expected changes of optical conductivity induced by the supercurrent injection with respect to the normal state conductivity  . Figures are taken from Reference [58]. Q and  Q Q where  is the Fourier component of the oscillating order (2 parameter. The rst term corresponds to the quadratic coupling of the Higgs mode to the gauge eld as described Section II. The second term indicates that the Higgs mode linearly couples with the gauge eld under the supercurrent with the condensate momentum Q parallel to the probe electric eld. Recently, in accordance with the theoretical prediction, the Higgs mode was observed in the optical conductivity in an s-wave superconductor NbN thin lm under the supercurrent injection by using THz time-domain spectroscopy [58]. Figure 10a shows the schematic experimental setup. Figures 10b and c show the change of the real part of the optical conductivity induced by the supercurrent injection with the probe polarization (b ) parallel -1 -1 δσ (ω) (Ω cm ) δσ (ω) / σ -1 -1 1 N δσ (ω) (Ω cm ) 1 and (c) orthogonal to the current direction. A clear peak is observed at the gap edge ! = 2 in the parallel geometry, showing a good agreement with the spectra calculated on the basis of the theory developed by Moor et al. [57] as represented in Figure 10d . This result indicates that, without resorting to the sophisticated nonlinear THz spectroscopy, the Higgs mode can be observed in the linear response function when the condensate has a nite momentum, expanding the feasibility for the detection of the Higgs mode in a wider range of materials. The visibility of the Higgs mode in the optical conductivity has also been addressed in a two-dimensional system near a quantum critical point [109, 110] (see also [111, 112]). Ex- perimentally, an extra spectral weight in the optical conductivity below the superconducting gap was observed in a strongly disordered superconducting lm of NbN, which was inter- preted in terms of the Higgs mode with its mass pushed below the pair breaking gap due to the strong disorder [113]. However, the assignment of the extra spectral weight below the gap remains under debates. It can also be described by the Nambu-Goldstone mode which acquires the electric dipole in strongly disordered superconductors [114{117]. The extra optical conductivity was also accounted for by the disorder induced broadening of the quasiparticle density of states below the gap [118]. IV. FUTURE PERSPECTIVE Having seen the Higgs mode in s-wave and d-wave superconductors, it is fascinating to extend the observation of the Higgs mode to a variety of conventional/unconventional superconductors. The study of the Higgs mode in multiband superconductors would be important, as it would give deeper insights into the interband couplings. For instance, the application of the nonlinear THz spectroscopy to iron-based superconductors is highly intriguing, as it may provide information on the interband interactions and pairing symmetry [87, 119, 120]. The Leggett mode, namely the collective mode associated with the relative phase of two condensates [86] is also expected to be present in multiband superconductors. The Leggett mode has been observed in a multiband superconductor MgB by Raman spectroscopy [121]. The nonlinear coupling between the Leggett mode and THz light has also been discussed [53, 87{90]. Recently, a THz-pump THz probe study in MgB has been reported, and 26 the results were interpreted in terms of the Leggett mode [122]. However, the dominant contribution of the Higgs mode over the Leggett mode in the nonlinear THz responses was theoretically pointed out in the case of dirty-limit superconductors [53]. Therefore, further experimental veri cation of the Higgs and Leggett modes is left as a future problem. The ultimate fate of the behavior of the Higgs mode in a strongly correlated regime and, what is more, in a BCS-BEC crossover regime is an intriguing problem. In particular, the decay pro le of the Higgs mode has been predicted to change from the BCS to BEC regime [36, 37, 123{125]. Experimentally, such a study has been realized in a cold-atom system, showing a broadening of the Higgs mode in the BEC regime [61]. Further investigation on the Higgs mode in the BCS-BEC crossover regime will be a future challenge in solid-state systems. The Higgs mode in spin-triplet p-wave superconductors (whose candidate materials in- clude Sr RuO ) is another issue to be explored in the future. Like in the case of super uid 2 4 He [22, 59, 126], multiple Higgs modes are expected to appear due to the spontaneous symmetry breaking in the spin channel, and its comparison with high-energy physics would be interesting. As yet another new paradigm, the observation of the Higgs mode would pave a new pathway for the study of nonequilibrium phenomena, in particular for the photoinduced superconductivity [127{130]. Being a ngerprint of the order parameter with a picosecond time resolution, the observation of the Higgs mode in photoinduced states should o er a di- rect evidence for nonequilibrium superconductivity. In strongly correlated electron systems exempli ed by the unconventional superconductors, elucidation of the interplay between competing and/or coexisting orders with superconductivity is an important issue to under- stand the emergent phases. The time-domain study of collective modes associated with those orders is expected to provide new insights for their interplay. V. DISCLOSURE STATEMENT The authors are not aware of any aliations, memberships, funding, or nancial holdings that might be perceived as a ecting the objectivity of this review. 27 ACKNOWLEDGMENTS We wish to acknowledge valuable discussions with Hideo Aoki, Yann Gallais, Dirk Manske, Stefan Kaiser, and Seiji Miyashita. We also acknowledge coworkers, Yuta Murotani, Keisuke Tomita, Kota Katsumi, Sachiko Nakamura, Naotaka Yoshikawa, and in particular Ryusuke Matsunaga for fruitful discussions and their support for preparing the manuscript. [1] Nambu Y. 2011. BCS: 50 Years, edited by L. N. Cooper and D. Feldman (World Scienti c, Singapore) [2] Ginzburg VL, Landau LD. 1950. Zh. Eksp. Teor. Fiz. 20:1064 [3] Bogoliubov NN. 1958. J. Exptl. Theoret. Phys. 34:58 [1958. Sov. Phys. 34:41] [4] Anderson PW. 1958. Phys. Rev. 110:827 [5] Anderson PW. 1958. Phys. Rev. 112:1900 [6] Nambu Y. 1960. Phys. Rev. 117:648 [7] Goldstone J. 1961. Nuovo Cim. 19:154 [8] Goldstone J, Salam A, Weinberg S. 1962. Phys. Rev. 127:965 [9] Anderson PW. 1963. Phys. Rev. 130:439 [10] Englert F, Brout R. 1964. Phys. Rev. Lett. 13:321. [11] Higgs PW. 1964. Phys. Lett. 12:132 [12] Higgs PW. 1964. Phys. Rev. Lett. 13:508 [13] Guralnik GS, Hagen CR, Kibble TWB. 1964. Phys. Rev. Lett. 13:585 [14] Anderson PW. 2015. Nat. Phys. 11:93 [15] Bardeen J, Cooper LN, Schrie er JR. 1957. Phys. Rev. 108:1175 [16] Volkov AF, Kogan SM. 1974. Sov. Phys. JETP 38:1018 [17] Kulik IO, Entin-Wohlman O, Orbach R. 1981. J. Low Temp. Phys. 43:591 [18] Littlewood PB, Varma CM. 1981. Phys. Rev. Lett. 47:811 [19] Littlewood PB, Varma CM. 1982. Phys. Rev. B 26:4883 [20] Nambu Y, Jona-Lasinio G. 1961. Phys. Rev. 122:345 [21] Nambu Y. 1985. Physica 15D:147 [22] Volovik GE, Zubkov MA. 2014. J. Low Temp. Phys. 175:486 28 [23] ATLAS Collaboration. 2012. Phys. Lett. B 716:1 [24] CMS Collaboration. 2012. Phys. Lett. B 716:30 [25] Sooryakumar R, Klein MV. 1980. Phys. Rev. Lett. 45:660 [26] Sooryakumar R, Klein MV. 1981. Phys. Rev. B 23:3213 [27] Varma C. 2002. J. Low Temp. Phys. 126:901 [28] M easson MA, Gallais Y, Cazayous M, Clair B, Rodi ere P, Cario L, Sacuto A. 2014. Phys. Rev. B 89:060503(R) [29] Cea T, Benfatto L. 2014. Phys. Rev. B 90:224515 [30] Grasset R, Cea T, Gallais Y, Cazayous M, Sacuto A, Cario L, Benfatto L, M easson MA. 2018. Phys. Rev. B 97:094502 [31] Barankov RA, Levitov LS, Spivak BZ. 2004. Phys. Rev. Lett. 93:160401 [32] Yuzbashyan EA, Altshuler BL, Kuznetsov VB, Enolskii VZ. 2005. Phys. Rev. B 72:220503 [33] Barankov RA, Levitov LS. 2006. Phys. Rev. Lett. 96:230403 [34] Yuzbashyan EA, Tsyplyatyev O, Altshuler BL. 2006. Phys. Rev. Lett. 96:097005 [35] Yuzbashyan EA, Dzero M. 2006. Phys. Rev. Lett. 96:230404 [36] Gurarie V. 2009. Phys. Rev. Lett. 103:075301 [37] Tsuji N, Eckstein M, Werner P. 2013. Phys. Rev. Lett. 110:136404 [38] Papenkort T, Axt VM, Kuhn T. 2007. Phys. Rev. B 76:224522 [39] Papenkort T, Kuhn T, Axt VM. 2008. Phys. Rev. B 78:132505 [40] Schnyder AP, Manske D, Avella A. 2011. Phys. Rev. B 84:214513 [41] Krull H, Manske D, Uhrig GS, Schnyder AP. 2014. Phys. Rev. B 90:014515 [42] Tsuji N, Aoki H. 2015. Phys. Rev. B 92:064508 [43] Kemper AF, Sentef MA, Moritz B, Freericks JK, Devereaux TP. 2015. Phys. Rev. B 92:224517 [44] Chou Y-Z, Liao Y, Foster M. S. 2017. Phys. Rev. B95:104507 [45] Matsunaga R, Hamada YI, Makise K, Uzawa Y, Terai H, Wang Z, Shimano R. 2013. Phys. Rev. Lett. 111:057002 [46] Matsunaga R, Tsuji N, Fujita H, Sugioka A, Makise K, Uzawa Y, Terai H, Wang Z, Aoki H, Shimano R. 2014. Science 345:1145 [47] Cea T, Castellani C, Benfatto L. 2016. Phys. Rev. B 93:180507(R) [48] Tsuji N, Murakami Y, Aoki H. 2016. Phys. Rev. B 94:224519 29 [49] Matsunaga R, Tsuji N, Makise K, Terai H, Aoki H, Shimano R. 2017 Phys. Rev. B 96:020505(R) [50] Cea T, Barone P, Castellani C, Benfatto L. 2018. Phys. Rev. B 97:094516 [51] Jujo T. 2015. J. Phys. Soc. Jpn. 84:114711 [52] Jujo T. 2018. J. Phys. Soc. Jpn. 87:024704 [53] Murotani Y, Shimano R. 2019. arXiv:1902.01104 [54] Silaev M. 2019. arXiv:1902.01666 [55] Katsumi K, Tsuji N, Hamada YI, Matsunaga R, Schneeloch J, Zhong RD, Gu GD, Aoki H, Gallais Y, Shimano R. 2018. Phys. Rev. Lett. 120:117001 [56] Chu H, Kim M-J, Katsumi K, Kovalev S, Dawson RD, Schwarz L, Yoshikawa N, Kim G, Putzky D, Li ZZ, Ra y H, Germanskiy S, Deinert J-C, Awari N, Ilyakov I, Green B, Chen M, Bawatna M, Christiani G, Logvenov G, Gallais Y, Boris AV, Keimer B, Schnyder A, Manske D, Gensch M, Wang Z, Shimano R, Kaiser S. 2019. arXiv:1901.06675 [57] Moor A, Volkov AF, Efetov KB. 2017. Phys. Rev. Lett. 118:047001 [58] Nakamura S, Iida Y, Murotani Y, Matsunaga R, Terai H, Shimano R. 2018. arXiv:1809.10335 [59] Vollhardt D, W ol e P. 1990. The Super uid Phases of Helium 3 (Taylor & Francis) [60] Endres M, Fukuhara T, Pekker D, Cheneau M, Schau P, Gross C, Demler E, Kuhr S, Bloch I. 2012. Nature 487:454 [61] Behrle A, Harrison T, Kombe J, Gao K, Link M, Bernier J-S, Kollath C, K ohl M. 2018. Nat. Phys. 14:781 [62] Ruegg  C, Normand B, Matsumoto M, Furrer A, McMorrow DF, et al. 2008. Phys. Rev. Lett. 100:205701 [63] Pekker D, Varma CM. 2015. Annu. Rev. Condens. Matter Phys. 6:269 [64] Abraham E, Tsuneto T. 1966. Phys. Rev. 152:416 [65] Schmid A. 1966. Phys. Kond. Mater. 5:302 [66] Caroli C, Maki K. 1967. Phys. Rev. 159:306 [67] Ebisawa H, Fukuyama H. 1971. Prog. Theor. Phys. 46:1042 [68] S a de Melo CAR, Randeria M, Engelbrecht JR. 1993. Phys. Rev. Lett. 71:3202 [69] Tsuchiya S, Yamamoto D, Yoshii R, Nitta M. 2018. Phys. Rev. B 98:094503 [70] Gor'kov LP, Eliashberg GM. 1968. Sov. Phys. JETP 27:328 30 [71] Gulian AM, Zharkov GF. 1999. Nonequilibrium electrons and phonons in superconductors (Kluwer Academic/Plenum Publishers) [72] Kopnin N. 2001. Theory of Nonequilibrium Superconductivity (Oxford University Press) [73] Schrie er JR. 1999. Theory of Superconductivity (Westview Press) [74] Kihlstrom KE, Simon RW, Wolf SA. 1985. Phys. Rev. B 32:1843 [75] Brorson SD, Kazeroonian A, Moodera JS, Face DW, Cheng TK, Ippen EP, Dresselhaus MS, Dresselhaus G. 1990. Phys. Rev. Lett. 64:2172 [76] Chockalingam SP, Chand M, Jesudasan J, Tripathi V, Raychaudhuri P. 2008. Phys. Rev. B 77:214503 [77] Aoki H, Tsuji N, Eckstein M, Kollar M, Oka T, Werner P. 2014. Rev. Mod. Phys. 86:779 [78] Murakami Y, Werner P, Tsuji N, Aoki H. 2016. Phys. Rev. B 93:094509 [79] Murakami Y, Werner P, Tsuji N, Aoki H. 2016. Phys. Rev. B 94:115126 [80] Nosarzewski B, Moritz B, Freericks JK, Kemper AF, Devereaux TP. 2017. Phys. Rev. B 96:184518 [81] Kumar A, Kemper AF. 2019. arXiv:1902.09549 [82] Yu T, Wu MW. 2017. Phys. Rev. B 96:155311 [83] Yu T, Wu MW. 2017. Phys. Rev. B 96:155312 [84] Yang F, Wu MW. 2018. Phys. Rev. B 98:094507 [85] Yang F, Wu MW. 2018. arXiv:1812.06622 [86] Leggett AJ. 1966. Prog. Theor. Phys. 36:901 [87] Akbari A, Schnyder AP, Manske D, Eremin I. 2013. Europhys. Lett. 101:17002 [88] Krull H, Bittner N, Uhrig GS, Manske D, Schnyder AP. 2016. Nat. Commun. 7:11921 [89] Cea T, Benfatto L. 2016. Phys. Rev. B 94:064512 [90] Murotani Y, Tsuji N, Aoki H. 2017. Phys. Rev. B 95:104503 [91] Barlas Y, Varma CM. 2013. Phys. Rev. B 87:054503 [92] Peronaci F, Schir o M, Capone M. 2015. Phys. Rev. Lett. 115:257001 [93] Foster MS, Dzero M, Gurarie V, Yuzbashyan EA. 2013. Phys. Rev. B 88:104511 [94] Fauseweh B, Schwarz L, Tsuji N, Cheng N, Bittner N, Krull H, Berciu M, Uhrig GS, Schnyder AP, Kaiser S, Manske D. 2017. arXiv:1712.07989 [95] Grasset R, Gallais Y, Sacuto A, Cazayous M, Manas-V ~ alero S, Coronado E, M easson MA. 2019. Phys. Rev. Lett. 122:127001 31 [96] Hebling J, Yeh KL, Ho mann MC, Bartal B, Nelson KA. 2008. J. Opt. Soc. Am. B 25:B6 [97] Watanabe S, Minami N, Shimano R. 2011. Optics Express 19:1528 [98] Shimano R, Watanabe S, Matsunaga R. 2012. J. Infrared Millim. Terahz Waves 33:861 [99] Matsunaga R, Shimano R. 2012. Phys. Rev. Lett. 109:187002 [100] Matsunaga R, Shimano R. 2017. Phys. Scr. 92:024003 [101] Beck M, Klammer M, Lang S, Leiderer P, Kabanov VV, Gol'tsman GN, Demsar J. 2011. Phys. Rev. Lett. 107:177007 [102] Mansart B, Lorenzana J, Mann A, Odeh A, Scarongella M, Chergui M, Carbone F. 2013. Proc. Natl. Acad. Sci. USA 110:4539 [103] Gor'kov LP, Eliashberg GM. 1969. Sov. Phys. JETP 29:698 [104] Amato JC, McLean WL. 1976. Phys. Rev. Lett. 37:930 [105] Entin-Wohlman O. 1978. Phys. Rev. B 18:4762 [106] Bardasis A, Schrie er JR. 1961. Phys. Rev. 121:1050 [107] Mattis DC, Bardeen J. 1958. Phys. Rev. 111:412 [108] Zimmermann W, Brandt EH, Bauer M, Seider E, Genzel L. 1991. Physica C: Supercond. 183:99 [109] Podolsky D, Auerbach A, Arovas DP. 2011. Phys. Rev. B 84:174522 [110] Gazit S, Podolsky D, Auerbach A. 2013. Phys. Rev. Lett. 110:140401 [111] Sachdev S. 1999. Phys. Rev. B 59:14054 [112] Zwerger W. 2004. Phys. Rev. Lett. 92:027203 [113] Sherman D, Pracht US, Gorshunov B, Poran S, Jesudasan J, Chand M, Raychaudhuri P, Swanson M, Trivedi N, Auerbach A, Scheer M, Frydman A, Dressel M. 2015. Nat. Phys. 11:188 [114] Cea T, Bucheli D, Seibold G, Benfatto L, Lorenzana J, Castellani C. 2014. Phys. Rev. B 89:174506 [115] Cea T, Castellani C, Seibold G, Benfatto L. 2015. Phys. Rev. Lett. 115:157002 [116] Pracht US, Cea T, Bachar N, Deutscher G, Farber E, Dressel M, Schefer M, Castellani C, Garc a-Garc a AM, Benfatto L. 2017. Phys. Rev. B 96:094514 [117] Seibold G, Benfatto L, Castellani C. 2017. Phys. Rev. B 96:144507 [118] Cheng B, Wu L, Laurita NJ, Singh H, Chand M, Raychaudhuri P, Armitage NP. 2016. Phys. Rev. B 93:180511 32 [119] Maiti S, Hirschfeld PJ. 2015. Phys. Rev. B 92:094506 [120] Muller  MA, Shen P, Dzero M, Eremin I. 2018. Phys. Rev. B 98:024522 [121] Blumberg G, Mialitsin A, Dennis BS, Klein MV, Zhigadlo ND, Karpinski J. 2007. Phys. Rev. Lett. 99:227002 [122] Giorgianni F, Cea T, Vicario C, Hauri CP, Withanage WK, Xi X, Benfatto L. 2019. Nat. Phys. online publication [123] Scott RG, Dalfovo F, Pitaevskii LP, Stringari S. 2012. Phys. Rev. A 86:053604 [124] Yuzbashyan EA, Dzero M, Gurarie V, Foster MS. 2015. Phys. Rev. A 91:033628 [125] Tokimoto J, Tsuchiya S, Nikuni T. 2019. J. Phys. Soc. Jpn. 88:023601 [126] W ol e P. 1977. Physica B 90:96 [127] Fausti D, Tobey RI, Dean N, Kaiser S, Dienst A, Ho mann MC, Pyon S, Takayama T, Takagi H, Cavalleri A. 2011. Science 331:189 [128] Kaiser S, Hunt CR, Nicoletti D, Hu W, Gierz I, Liu HY, Le Tacon M, Loew T, Haug D, Keimer B, Cavalleri A. 2014. Phys. Rev. B 89:184516 [129] Hu W, Kaiser S, Nicoletti D, Hunt CR, Gierz I, Ho mann MC, Le Tacon M, Loew T, Keimer B, Cavalleri A. 2014. Nat. Mater. 13:705 [130] Mitrano M, Cantaluppi A, Nicoletti D, Kaiser S, Perucchi A, Lupi S, Di Pietro P, Pontiroli D, Ricc o M, Clark SR, Jaksch D, Cavalleri A. 2016. Nature 530:461 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Higgs Mode in Superconductors

Condensed Matter , Volume 2019 (1906) – Jun 22, 2019

Loading next page...
 
/lp/arxiv-cornell-university/higgs-mode-in-superconductors-Eh3nfmKsRV

References (5)

ISSN
1947-5454
eISSN
ARCH-3331
DOI
10.1146/annurev-conmatphys-031119-050813
Publisher site
See Article on Publisher Site

Abstract

When a continuous symmetry of a physical system is spontaneously broken, two types of collec- tive modes typically emerge: the amplitude and phase modes of the order-parameter uctuation. For superconductors, the amplitude mode is recently referred to as the \Higgs mode" as it is a condensed-matter analogue of a Higgs boson in particle physics. Higgs mode is a scalar excita- tion of the order parameter, distinct from charge or spin uctuations, and thus does not couple to electromagnetic elds linearly. This is why the Higgs mode in superconductors has evaded experimental observations over a half century after the initial theoretical prediction, except for a charge-density-wave coexisting system. With the advance of nonlinear and time-resolved terahertz spectroscopy techniques, however, it has become possible to study the Higgs mode through the nonlinear light-Higgs coupling. In this review, we overview recent progresses on the study of the arXiv:1906.09401v1 [cond-mat.supr-con] 22 Jun 2019 I. INTRODUCTION AND A BRIEF HISTORICAL OVERVIEW Spontaneous breaking of a continuous symmetry is a fundamental concept of phase transi- tion phenomena in various physical systems ranging from condensed matter to high-energy physics [1]. For instance, a ferromagnetic transition is characterized by the spontaneous breaking of spin rotational symmetry. Likewise, superconductivity is characterized by the spontaneous breaking of U(1) rotational symmetry with respect to the phase of a macro- scopic wavefunction [2]. When a continuous symmetry is spontaneously broken, there emerge two types of col- lective modes in general: uctuations of the phase and amplitude of the order parameter as schematically shown in Figure 1a. The phase mode (also termed Nambu-Goldstone (NG) mode) is primarily a gapless (massless) mode [3{8] as required by the symmetry. An acoustic phonon is such an example that is associated with spontaneous breaking of trans- lational symmetry in a crystal lattice. The amplitude mode, on the other hand, is generally a gapped (massive) mode since its excitation costs an energy due to the curvature of the potential (Figure 1a ). The amplitude mode (especially for superconductors) is often called the Higgs mode according to the close analogy with the Higgs boson in particle physics. It may sound strange that one talks about the Higgs particle in condensed matter systems, but in fact the origin of the idea of Higgs physics can be found in the course of the study of superconductivity [3{9]. Namely, the standard model in particle physics can in some sense be viewed as a relativistic version of the Ginzburg-Landau theory [2], i.e., a low-energy e ective theory of superconductors. The emergence of the Higgs mode is a universal and fundamental phenomenon in systems with spontaneous symmetry breaking. In the case of superconductors where the order parameter couples to gauge elds, the Higgs mode occupies a special status since it is the lowest collective excitation mode (Fig- ure 1b ). The massless phase mode is absorbed into the longitudinal component of electro- magnetic elds, and is lifted up to high energy in the scale of the plasma frequency due to the Anderson-Higgs mechanism [4{6, 9{13]. As a result, the amplitude mode becomes stable against the decay to the phase mode in superconductors. The existence of the amplitude mode in superconductors was suggested by Anderson [5, 14] soon after the development of a microscopic theory of superconductivity by Bardeen, Cooper, and Schrie er (BCS) [15]. In an s-wave (BCS-type) superconductor, the Higgs-mode gap energy ! coincides with the 2 a b <latexit sha1_base64="QHK9fgKuka7iN1RctYFrmwXDUbk=">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</latexit> Phys. Scr. 92 (2017) 024003 R Matsunaga and R Shimano Free energy quasiparticle Anderson-Higgs excitations mechanism Higgs mode Higgs mode <latexit sha1_base64="tkrKJiURdjBMrxZsmPrJbCrt0gw=">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</latexit>Re Nambu-Goldstone NG mode <latexit sha1_base64="IsdAcr1U4kit6BDS167S+m063Cs=">AAAC13icfVHLbhMxFHWmPEp49MGSjUWExAJFMwWpiFUlqgoWFalK2qDMUHk8dxKrfoxsT9vIstghtt2xhV/ga/ibepIINWnFlSwfn3uO7vW9ecWZsXH8txWt3Ll77/7qg/bDR4+frK1vbB4ZVWsKfaq40oOcGOBMQt8yy2FQaSAi53Ccn75v8sdnoA1T8rOdVJAJMpKsZJTYQH11qRb4o/Dpq7Rn2Ml6J+7G08A3QTIHHTSP3slG609aKFoLkJZyYswwiSubOaItoxx8O60NVISekhEMA5REgMnctG2PXwSmwKXS4UiLp+x1hyPCmInIg1IQOzbLuYa8LTesbfk2c0xWtQVJZ4XKmmOrcDMDXDAN1PJJAIRqFnrFdEw0oTZM6r89L7TQFNamNIusvWh+s0SONKnGjDbsLoRJaTgMH1N8L0idmULj3af9Q++oMBPvhHfyVnF4juCaY9A44OKfQ8I5VUIQWbg0F36YZOFWvJhZcCfx7bDmZHmpN8HRVjd53d06eNPZeTdf+Cp6hp6jlyhB22gHfUA91EcUafQT/UK/oy/Rt+h79GMmjVpzz1O0ENHlFdRE6/8=</latexit>Im (NG) mode <latexit sha1_base64="HlEL285e3sXISJRF4mVDzB64A44=">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</latexit> <latexit sha1_base64="9tneBTFFq1wSKGVGcADajPcqCYM=">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</latexit>0 Figure 1. (a) A schematic picture of the phase mode (blue arrow) and the amplitude mode (red arrow) represented by the Ginzburg–Landau free energy potential. (b) A schematic picture for collective modes and single particle excitation in a symmetry-broken system with the Anderson–Higgs mechanism. the Higgs mode have been demonstrated by a random-phase based on the time-dependent GL theory or the Boltzmann Figure 1. (a) A schematic picture of the Higgs (red) and Nambu-Goldstone (blue) modes rep- approximation theory [37]. equation is not justified [38]. To describe the dynamics of One main difficulty for the observation of the Higgs order parameter in non-equilibrium superconducting states in mode in superconductors is attributed to the fact that the such a nonadiabatic regime, the Anderson’s pseudospin resented on the Mexican-hat free-energy potential as a function of the complex order parameter representation [32] has been used as an useful microscopic Higgs mode does not have a charge nor electric dipole and treatment and nontrivial dynamics of the order parameter has therefore it does not couple directly to the electromagnetic been investigated in a wide range of quench parameters based . (b) A schematic field. This excitation explains why thesp experimental ectrum observation of an of the s-wave superconductor. Due to the Anderson-Higgs on the BCS mean-field theory [38–41]. Subsequently, the Higgs mode has been limited for a long time to the case of subject has been applied to the case of metallic super- Raman spectroscopy for NbSe in which a charge density mechanism, the Nambu-Goldstone mode acquires conducting an energy materials [42 gap –46] in inwhich thequasiparticle order of dis- the plasma fre- wave (CDW) order coexists with the superconducting phase tribution can be suddenly perturbed by an ultrashort laser [48]. Initially the Raman peak observed at the super- pulse. conducting gap energy of 2Δ in this material was interpreted quency ! , while the Higgs mode remains in low energy with an energy gap 2, above which the In the pseudospin model, up or down pseudospins for a as the single-particle excitation across the gap [48]. Soon wavenumber k means both of the two electron states at k and later, Littlewood and Varma have theoretically elucidated that −k being occupied or unoccupied, respectively. While in the the peak corresponds to the scalar excitation of the amplitude quasiparticle excitation continuum overlaps. normal state at T = 0 all the pseudospins are up for |k|<k mode (Higgs mode) [36, 37]. The stability and the visibility and down for |k|>k (k is the Fermi wavenumber), in the F F of the Higgs mode in Raman spectroscopy were further superconducting state the BCS ground state is expressed as a revealed recently in this strong electron-phonon coupling superposition of up and down pseudospins for k. This system in which CDW and superconductivity coexist representation makes the BCS Hamiltonian [49, 50]. By contrast, it has been unclear whether the Higgs x y z where s = ss s is the pseu- H =⋅2, b s k ( ) BCS å kk k k k mode is observable or not in superconductors without such a dospin operator for k, and b is represented as CDW order, and indeed the Raman peak of the Higgs mode superconducting gap energy 2 [5, 16{19] (see also [20, 21]). This is eventually consistent was absent in superconducting NbS which is a material 2 b = -D¢ - D e.1() k() k similar to NbSe but has no CDW order [49]. ε is the energy dispersion measured from the Fermi energy. It has been shown theoretically that the Higgs mode 2 2 2 with Nambu's conjectural sum rule [21, 22]: ! +! = 4 with the NG mode gap energy Δ is the order parameter obtained by the sum of the pseu- appears in the dynamics of a complex order parameter after a H NG dospins: sudden perturbation [34]. Recently, such an artificial control of the pairing interaction has become feasible in the experi- D=D¢+iU D= () ss +i,2() ! = 0. Since the Higgs mode lies at the lower bound of the quasiparticle k k continuum near NG ments in the field of ultracold atomic systems [51], and the k dynamics of symmetry-broken ordered states after a sudden where U is the attractive interaction between the two electron quench of the interaction has gained renewed interests [38– zero momentum, the decay of the Higgs mode to single-particle excitations is suppressed states. s is parallel to b in equilibrium for each k. Then the k k 41]. It should be noted here that free energy is in general well equation of motion, namely the Bogoliubov–de Gennes defined only in a static regime and the time-dependent GL equation, is given by and becomes a much slower power law [16]. For nonzero momentum, the Higgs mode can theory is valid only when the change of the order parameter is slow enough. When the order parameter rapidly changes in ss== iH[],2b´s, (3) kk BCS k k ¶t the timescale of τ =h/Δ, the phenomenological model decay by transferring its energy to quasiparticle excitations. The energy dispersion and the damping rate of the Higgs mode has been derived within the random phase approximation [18, 19]. Provided that the theory of the Higgs particle originates from that of superconductivity and that the Higgs particle has been discovered in LHC experiments in 2012 [23, 24], it would be rather surprising that an experimental observation of the Higgs mode in superconductors, a home ground of Higgs physics, has been elusive for a long time. There are, however, good reasons for that: The Higgs mode does not have any electric charge, electric dipole, magnetic moment, and other quantum numbers. In other words, the Higgs mode is a scalar excitation (which is distinct from, e.g., charge uctuations). Therefore it does not couple to external 3 probes such as electromagnetic elds in the linear-response regime. Another reason is that the energy scale of the Higgs mode lies in that of the superconducting gap, which is in the order of millielectron volts in typical metallic superconductors. To excite the Higgs mode, one needs an intense terahertz (THz) light source (1 THz  4 meV  300 m), which has become available only in the last decade. One exception (and the rst case) for the observation of the Higgs mode in superconduc- tors in the early stage was the Raman scattering experiment for a superconductor 2H-NbSe [25, 26]. It is exceptional in the sense that superconductivity and charge density wave (CDW) coexist in a single material. The Raman peak observed near the superconducting gap energy 2 in 2H-NbSe was rst interpreted as a single-particle excitation across the gap [25, 26]. Soon later, Littlewood and Varma have theoretically elucidated that the peak corresponds to the scalar excitation of the amplitude mode (i.e., the Higgs mode) [18, 19] (see also Reference [27]). A renewed interest in the Raman signal has recently revealed the importance of the coexisting CDW order for the Higgs mode to be visible in the Raman response [28{30]. Indeed, the Raman peak of the Higgs mode has been shown to be absent in a superconducting NbS , which is a material similar to NbSe but has no CDW order 2 2 [28]. It has been a long-standing issue as to whether the Higgs mode can be observed in superconductors without CDW order. On the theoretical side, there have been various proposals for the excitation of the Higgs mode, including a quench dynamics [16, 31{37] as well as a laser excitation [38{44]. It is only after the development of ultrafast pump-probe spectroscopy techniques in the low-energy terahertz-frequency region that a clear observation of the Higgs-mode oscillation has been reported in a pure s-wave superconductor NbN [45] (with `pure' meaning no other long-range order such as CDW). Subsequently, it has been experimentally demonstrated through the THz pump-probe and third harmonic generation (THG) measurements [46] that the Higgs mode can couple to electromagnetic elds in a nonlinear way. In the THG experiment, a resonant enhancement of THG was discovered [46] at the condition 2! = 2 with ! the frequency of the incident THz light, indicating the nonlinear light-Higgs coupling in a two-photon process [42]. Meanwhile, it has been theoretically pointed out that not only the Higgs mode but also charge density uctuation (CDF) can contribute to the THG signal [47]. It was shown that the CDF gives a much larger contribution than the Higgs mode in a clean-limit super- 4 conductor within the BCS mean- eld approximation. While the relative magnitude of the Higgs-mode and CDF contributions to the THG has been under debate [47{50], theoretical progresses have recently been made on the theory of the light-Higgs coupling, showing that both the e ects of phonon retardation [48] and nonmagnetic impurity scattering [51{54] drastically enhance the Higgs-mode contribution to nonlinear optical responses, far exceed- ing the CDF contribution. The experiments have been extended to high-T cuprate superconductors, showing the presence of the Higgs-mode contribution to the pump-probe signal in d-wave superconductors [55]. Recently the THG has also been identi ed in various high-T cuprates [56]. The Higgs mode can be made visible in a linear optical response if dc supercurrent is owing in a superconductor [57]. The infrared activated Higgs mode in the presence of supercurrents has recently been observed in a superconductor NbN [58]. In this review article, we overview these recent progresses in the study of the Higgs mode in superconductors (where the Anderson-Higgs mechanism is taking place), with an empha- sis on the experimental aspects. In a broader context, collective amplitude modes are not limited to superconductors but are ubiquitous in condensed matter systems that may not be coupled to gauge elds (hence without the Anderson-Higgs mechanism). Those include squashing modes in super uid He [59], amplitude modes in bosonic and fermionic conden- sates of ultracold-atom systems [60, 61], amplitude modes in quantum antiferromagnets [62], and so on. A comprehensive review including these topics can be found in Reference [63]. II. NONLINEAR LIGHT-HIGGS COUPLING In this section, we review from a theoretical point of view how the Higgs mode in supercon- ductors couples to electromagnetic elds in a nonlinear way. It is important to understand the mechanism of the nonlinear light-Higgs coupling for observing the Higgs mode in ex- periments. In fact, the Higgs mode does not have a linear response against electromagnetic elds in usual situations, which has been an obstacle for experiments to detect the Higgs mode in superconductors for a long time. 5 A. A phenomenological view Let us rst take a phenomenological point of view based on the Ginzburg-Landau (GL) theory [2], which provides us with a quick look at the nonlinear light-Higgs coupling. For early developments based on the time-dependent GL theory, we refer to References [64{ 68]. In the GL theory, the free energy density f is assumed to be a function of a complex order-parameter eld (r), b 1 2 4  2 f [ ] = f + aj (r)j + j (r)j + j(ir e A) (r)j ; (1) 2 2m where a = a (T T ), a and b are some constants, m and e are the e ective mass and 0 c 0 the e ective charge of the Cooper pair condensate, and A is the vector potential that rep- resents the external light eld (E(t) = @A(t)=@t). The amplitude of the order parameter corresponds to the super uid density (i.e., j j = n ), while the phase corresponds to that of the condensate. The free energy density (1) is invariant under the global U(1) phase rotation (r) ! i' ie (r) e (r) and more generally under the gauge transformation (r) ! e (r), A(r) ! A(r) + r(r) for an arbitrary scalar eld (r). In addition, the GL free energy density (1) is invariant against the particle-hole transformation (r) ! (r). The presence of the particle-hole symmetry (including the time derivative terms in the action) is crucial [27, 63, 69] in decoupling the Higgs amplitude mode from the phase mode (Nambu-Goldstone mode) in electrically neutral systems such as super uid ultracold atoms. Without the particle-hole symmetry, the Higgs mode quickly decays into the phase mode and becomes short-lived. In superconductors (which consist of electrically charged electrons), on the other hand, the phase mode acquires an energy gap in the order of the plasma frequency due to the Anderson- Higgs mechanism. This prevents the Higgs mode from decaying into the phase mode. The particle-hole symmetry itself is realized as an approximate symmetry in the Bogoliubov-de Gennes Hamiltonian, which describes microscopic low-energy physics of superconductors around the Fermi energy at the mean- eld level. When the temperature goes below the critical temperature (i.e., a < 0), the global U(1) symmetry is spontaneously broken, and the system turns into a superconducting state. The order-parameter uctuation from the ground state ( (r) = ) can be decomposed into 6 amplitude H (r) and phase (r) components, i(r) (r) = [ + H (r)]e : (2) Then, the GL potential can be expressed up to the second order of the uctuation as 1 e 1 2 2 2 f = 2aH + (rH ) + A r ( + H ) + : (3) 2m 2m e The rst term on the right hand side of Equation 3 indicates that the Higgs mode has an 1=2 1=2 energy gap (Figure 1b ) proportional to (a) / (T T ) . The microscopic BCS mean- eld theory predicts that the Higgs gap is actually identical to the superconducting gap 2 [5, 16{19]. The phase eld (r) does not have such a mass term, so that at rst sight one would expect to have a massless Nambu-Goldstone mode. However, (r) always appears in the form of (A r=e ) in Equation 3, which allows one to eliminate (r) from the GL 0  0 potential by taking a unitary gauge A = Ar=e . As a result, we obtain (denoting A as A) 2 2 2 1 e e 2 2 0 2 2 f = 2aH + (rH ) + A + A H + : (4) 2m 2m m One can see that the phase mode is absorbed into the longitudinal component of the elec- tromagnetic eld. At the same time, there appears a mass term for photons (the third term on the right hand side of Equation 4) with the mass proportional to e =m (Anderson-Higgs mechanism). Due to this, electromagnetic waves cannot propagate freely inside superconductors but decay exponentially with a nite penetration length (Meissner e ect). From Equation 4, one immediately sees that there is no linear coupling between the Higgs and electromagnetic eld. This is consistent with the fact that the Higgs mode does not have an electric charge, magnetic moment, and other quantum numbers, which has been a main obstacle in observing the Higgs mode by an external probe for a long time. On the other hand, there is a nonlinear coupling term A H in Equation 4, which is responsible for the Higgs mode to contribute in various nonlinear processes. In Figure 2a , we illustrate the diagrammatic representation of the nonlinear light-Higgs coupling, where two photons are coming in with frequencies ! and ! and one Higgs is emitted with frequency ! + ! . 1 2 1 2 This is analogous to the elementary process of the Higgs particle decaying into W bosons that had played a role in the discovery of the Higgs particle in LHC experiments [23, 24]. 7 a b c pump pump <latexit sha1_base64="WqGUujrvJd5IYb6/HOCxgg1RtDg=">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</latexit><latexit sha1_base64="WqGUujrvJd5IYb6/HOCxgg1RtDg=">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</latexit> A<latexit sha1_base64="87zEWZbfw8D7sDpQWGXnBY7M9Zw=">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</latexit> <latexit sha1_base64="GJryKaYbRvDRPnIjY5gpWeresPY=">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</latexit> <latexit sha1_base64="87zEWZbfw8D7sDpQWGXnBY7M9Zw=">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</latexit> <latexit sha1_base64="2AbfLqPhMXj7VY8ewmTjZx30iZM=">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</latexit>H 3<latexit sha1_base64="uFd62FUVMaSs53WnbF4+c2WNxHk=">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</latexit> <latexit sha1_base64="2AbfLqPhMXj7VY8ewmTjZx30iZM=">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</latexit>H <latexit sha1_base64="2AbfLqPhMXj7VY8ewmTjZx30iZM=">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</latexit>H <latexit sha1_base64="87zEWZbfw8D7sDpQWGXnBY7M9Zw=">AAACz3icfVFNbxMxEHWWj5bw1cKRi0WExCnaBSQ4VZVAqBdEq5I0UrKqZr2zG7f+WNle2sjaiiu3Xsvf4Nfwb/AmEWrSipEsP795TzPjySrBrYvjP53ozt179zc2H3QfPnr85OnW9rOh1bVhOGBaaDPKwKLgCgeOO4GjyiDITOBRdvqxzR99R2O5Vt/crMJUQql4wRm4QA0nWmIJx1u9uB/Pg94EyRL0yDL2j7c7vye5ZrVE5ZgAa8dJXLnUg3GcCWy6k9piBewUShwHqECiTf283Ya+CkxOC23CUY7O2esOD9LamcyCUoKb2vVcS96WG9eu+JB6rqraoWKLQkUtqNO0nZ3m3CBzYhYAMMNDr5RNwQBz4Yf+2/NKC21hYwu7yrrzdpo1sjRQTTlr2U8YfsrgYRhMi89B6u0c2sZ//XLYeCbtrPGy8epWcXiWeM0xah14/s+h8IxpKUHlfpLJZpyk4dYiX1hoL2m6Yc3J+lJvguGbfvK2nxy86+3uLBe+SV6Ql+Q1Sch7skv2yD4ZEEZOyCW5Ir+ig+gsuoh+LKRRZ+l5TlYi+vkXDqLo/g==</latexit> <latexit sha1_base64="GJryKaYbRvDRPnIjY5gpWeresPY=">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</latexit>A probe probe <latexit sha1_base64="pq9cDiS1JYBZf4U/jWRZ5OVcPTw=">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</latexit><latexit sha1_base64="pq9cDiS1JYBZf4U/jWRZ5OVcPTw=">AAAC3XicfVFNbxMxEHWWrxK+0nLkYhEhcYp2AQlOqBIIcUEUlbSRsquV1zu7seqPle2FRJaPvVVcuXGFO7+m/6beJEJNWjGS5ec372nGM0XDmbFxfN6Lbty8dfvOzt3+vfsPHj4a7O4dGdVqCmOquNKTghjgTMLYMsth0mggouBwXJy86/LH30AbpuRXu2ggE6SWrGKU2EDlg71UCahJ7lItcKNVAT4fDONRvAx8FSRrMETrOMh3e3/TUtFWgLSUE2OmSdzYzBFtGeXg+2lroCH0hNQwDVASASZzy+Y9fhaYEldKhyMtXrKXHY4IYxaiCEpB7Mxs5zryuty0tdWbzDHZtBYkXRWqWo6twt0kcMk0UMsXARCqWegV0xnRhNowr//2vNFCV1ibymyydt79ZousNWlmjHbsewiT0nAYPqb4hyB1ZgmNd58/HXpHhVl4J7yT14rDs4ZLjknngPk/h4TvVAlBZOnSQvhpkoVb8XJlwcPE98Oak+2lXgVHL0bJy1Hy5dVw/+164TvoCXqKnqMEvUb76CM6QGNE0Rz9Qr/RnyiPTqOz6MdKGvXWnsdoI6KfFxw47nE=</latexit> Figure 2. (a) Diagrammatic representation of the nonlinear light-Higgs coupling A H in Equa- tion 4. (b) Third harmonic generation mediated by the Higgs mode. (c) Pump-probe spectroscopy mediated by the Higgs mode. Due to the nonlinear interaction between the Higgs mode and electromagnetic elds, one can induce a nonlinear current given by @F ie e y y y j = = [ r (r ) ] A : (5) @A 2m m Again, we expand around the ground state in Equation 5 and collect leading terms in H , which results in 2e j = AH: (6) This is nothing but the leading part of the London equation j = (e n =m )A (let us recall that the amplitude of the order parameter corresponds to the super uid density, j j = n ). When a monochromatic laser with frequency ! drives a superconductor, the amplitude of the order parameter oscillates with frequency 2! due the nonlinear coupling A H (Fig- ure 2a ). Together with the oscillation of A with frequency ! in Equation 6, the induced nonlinear current shows an oscillation with frequency 3!. As a result, one obtains the third harmonic generation mediated by the Higgs mode (Figure 2b ). Since the Higgs mode has an energy 2 at low momentum, one can resonantly excite the Higgs mode by a THz laser with a resonance condition 2! = 2. Given that the nonlinear current is proportional to the Higgs-mode amplitude (Equation 6), the third harmonic generation can be resonantly enhanced through the excitation of the Higgs mode. The Higgs-mode resonance in third harmonic generation has been experimentally observed for a conventional superconductor [46]. Another example of nonlinear processes to which the Higgs mode can contribute is a pump-probe spectroscopy where pump and probe light is simultaneously applied (Fig- ure 2c). In this case, photons with di erent frequencies ! and ! are injected, and pump probe 8 σ C D a z b BCS state normal state (T=0) 1 1 k k F x F Figure 3. (a) Momentum distribution of a superconducting state (above) and its pseudospin representation due to Anderson (below). (b) Pseudospin precession induced by a laser eld. Figures are taken from Reference [46]. those with the same respective frequencies are emitted. There are three possibilities of the frequency carried by the Higgs mode: ! ! and ! ! = ! ! = 0. pump probe pump pump probe probe If one uses a THz pump and optical probe, i.e., ! . 2 and !  2, the rst two pump probe channels do not contribute since the excitation of the Higgs mode is far o -resonant. The remaining zero-frequency excitation channel contributes to the pump-probe spectroscopy. This process has been elaborated in the experimental study of the Higgs mode in d-wave superconductors [55]. B. A microscopic view In the previous subsection, we reviewed the phenomenology for the nonlinear light-Higgs coupling using the Ginzburg-Landau theory. However, strictly speaking, the application of the Ginzburg-Landau theory to nonequilibrium problems is not microscopically justi ed in the case of gapped superconductors [70{72]. The reason is that in a usual situation one cannot neglect the e ect of quasiparticle excitations whose relaxation time is longer than the time scale of the order-parameter variation. Moreover, the Higgs mode and quasipar- ticle excitations are energetically degenerate at low momentum (Figure 1b ), so that it is inevitable to excite quasiparticles at the same time when one excites the Higgs mode. This motivates us to take a microscopic approach for further understanding. Let us adopt the time-dependent BCS or Bogoliubov-de Gennes equation, which is e- y y y ciently represented by the Anderson pseudospin  = h  i [5]. Here = (c ; c ) k k k# k k k" is the Nambu spinor,  = ( ;  ;  ) are Pauli matrices, c (c ) is the creation (annihila- x y z k Distribution Distribution tion) operator of electrons with momentum k and spin , and hi denotes the statistical average. Physically, the x and y components of the Anderson pseudospin correspond to the real and imaginary parts of the Cooper pair density, respectively, and the z component corresponds to the momentum occupation distribution of electrons (Figure 3a ). With this notation, the BCS mean- eld Hamiltonian is written as H = 2 b (t)  ; (7) BCS k k 0 00 b (t) =  (t); (t); ( +  ) : (8) k k+A(t) kA(t) 0 00 Here b (t) is a pseudomagnetic eld acting on pseudospins,  and  are the real and imaginary parts of the superconducting gap function, respectively, and  is the band dis- persion of the system. The equation of motion for the pseudospins is given by the Bloch-type equation, (t) = 2b (t)  (t); (9) k k k @t 0 00 V x supplemented by the mean- eld condition  (t) + i (t) = [ (t) + i (t)] with V k k k the attractive interaction and N the number of k points. The time evolution of the super- conducting state is thus translated into the precession dynamics of Anderson pseudospins (Figure 3b ). The coupling to the electromagnetic eld appears in the z component of the pseudo- z z magnetic eld, b = ( +  ). If we expand b in terms of A(t), we obtain k+A(t) kA(t) k k 1 @ z k 4 b =  + A A + O(A ). The linear coupling term vanishes, which is consistent k i j k ij 2 @k @k i j with the phenomenology that we have reviewed in the previous subsection. The leading nonlinear coupling term is quadratic in A(t). In Figure 4a , we show a numerical result for the time evolution of the gap function (t) when the system is driven by a multi-cycle electric- eld pulse. One can see that the 2! oscillation of the gap is generated during the pulse irradiation, after which a free gap oscillation continues with the frequency 2 and the 1=2 amplitude slowly damping as t [16]. For a monochromatic wave A(t) = A sin !t, one can solve the equation of motion semi- analytically by linearizing with respect to the quadratic coupling A (t)A (t) around the i j equilibrium solution. The result shows that the amplitude of the gap function varies as [42] (t) / j2! 2j cos(2! ) (10) 10 1 (a) (b) 1.5 E(t) -1 1.4 Δ(t) 1.3 0 10 20 30 40 50 60 0 1 2 2ω/2Δ Figure 4. (a) A numerical result for the time evolution of the superconducting gap (t) driven by an electric eld pulse E(t) with the frequency ! = 2=5  1:26. Here we take a two dimen- sional square lattice with a bandwidth 8 at half lling. The interaction is V = 4, and the initial temperature is T = 0:02. The polarization of the electric eld is parallel to x axis. The left (right) vertical axis represents the value of (t) (E(t)) as indicated by the arrows. (b) The amplitude of the 2! oscillation of the superconducting gap as a function of 2!=2. (b) is reproduced from Reference [42]. for !  , where  is a phase shift. The induced oscillation frequency is 2!, re ecting the nonlinear coupling between the Higgs mode and electromagnetic elds, A H . The oscillation amplitude diverges when the condition 2! = 2 is ful lled (Figure 4b ). This is much the same as the spin resonance phenomenon: The forced collective precession of the pseudospins with frequency 2! due to the nonlinear coupling resonates with the Higgs mode whose energy coincides with 2. The power of the divergence in Equation 10 is smaller than that of a resonance with an in nitely long-lived mode ( j2! ! j ). The reason is that the 2 2 precession gradually dephases due to the momentum-dependent frequency ! = 2  + for each pseudospin so that the average of the pseudospins (t) =  (t) shows a k k power-law damping of oscillations. The phase shift  exhibits a jump as one goes across the resonance by changing the drive frequency !. In the case of the square lattice and the polarization of the electric eld parallel to the diagonal direction in the xy plane, for example, the current is given by j(t) / A(t)(t) [42, 46], which corresponds to the phenomenological relation j(t) / A(t)H (t) (Equation 6). As Δ(t) E(t) |δΔ | 2 ω expected, the Higgs mode induces a resonant enhancement of the third harmonic generation. Later, it has been pointed out [47] that the current relation is valid only in rather special situations such as the square lattice and the polarization parallel to the diagonal direction. In general, the BCS mean- eld treatment (in the clean limit) suggests that the Higgs- mode contribution to the third harmonic generation is subdominant as compared to that of individual quasiparticle excitations. Since the energy scales of the Higgs mode and the quasiparticle pair excitations are in the same order ( 2, see Figure 1b ), the competition between the two contributions always matters. This point will be further examined in the next subsection. C. Impurity and phonon assisting As we have seen in the previous subsection, the Higgs-mode contribution to the third harmonic generation is generally subdominant in the BCS clean limit. However, there is a growing understanding from recent studies [48, 49, 51{54] that if one takes into account e ects beyond the BCS mean- eld theory in the clean limit (such as impurity scattering and phonon retardation) the light-Higgs coupling strength drastically changes. To understand the impact of these e ects, let us consider the linear optical response. The current induced by a laser eld is decomposed into the paramagnetic and diamagnetic components j = j + j with [73] para dia j = hc c i; (11) para k @k k y j = A hc c i: (12) dia i k @k@k ki As we have seen in the pseudospin picture in the previous subsection, only the diamagnetic coupling contributes to the optical response in the mean- eld level with no disorder, resulting in Re (!) / (!) and Im (!) / 1=! for the superconducting state. For nite frequency (! 6= 0), the real part of the optical conductivity vanishes in the BCS clean limit. This does not necessarily mean that the real part of the optical conductivity is suppressed in real experimental situations. For example, in superconductors with a disorder the real part of the optical conductivity is nonzero and not even small for ! 6= 0. In fact, a superconductor NbN used in the experiment of the third harmonic generation turns out to be close to the 12 dirty regime (  2, where is the impurity scattering rate) as con rmed from the measurement of the optical conductivity [45]. (3) TABLE I. Relative order of magnitudes of the third-order current j in general situations [53]. mode channel clean ! dirty 2 2 Higgs dia (A ) (= ) 2 2 2 para (p A) ( = ) ! ( = ) F F Quasiparticles dia (A ) 1 2 2 2 para (p A) ( = ) ! ( = ) F F In much the same way as the optical conductivity, the nonlinear response intensity is strongly a ected by the impurity e ect. The dependence of the relative order of magnitudes of the third-order current is summarized in Table I, where  is the Fermi energy and the (3) unit of j is taken such that the diamagnetic-coupling contribution of the quasiparticles is set to be of order 1 (which does not signi cantly depend on ). In the BCS clean limit ( ! 0), we have only the diamagnetic coupling, for which the Higgs-mode contribution is subleading in the order of (= ) compared to the quasiparticles. However, in the presence of impurities ( 6= 0) the paramagnetic coupling generally emerges to contribute to the third 2 2 2 harmonic generation. Its relative order of magnitude changes as ( = ) ! ( = ) from F F the clean to dirty limit. The maximum strength of the paramagnetic coupling is achieved around  , where the relative order reaches ( =) for both the Higgs mode and quasiparticles. Thus, the impurity scattering drastically enhances the light-Higgs coupling: in the clean limit the Higgs-mode contribution is subleading to quasiparticles, whereas in the dirty regime it becomes comparable to or even larger than the quasiparticle contribution. The precise ratio between the Higgs and quasiparticle contributions may depend on details of the system. Recent studies [52{54] suggest that in the dirty regime the Higgs-mode contribution to THG is an order of magnitude larger than the quasiparticle contribution. An enhancement of the paramagnetic coupling also occurs for strongly coupled superconductors due to the phonon retardation e ect [48]. 13 D. Further developments So far, we have reviewed the phenomenological and mean- eld treatments of the nonlinear light-Higgs coupling for conventional s-wave superconductors with or without disorder. For strongly coupled superconductors with an electron-phonon coupling constant  & 1 (which is the case for NbN superconductors [74{76]), one has to take account of strong correlation e ects. One useful approach for this is the nonequilibrium dynamical mean- eld theory [77], which takes into account local dynamical correlations by mapping a lattice model into a local impurity model embedded in an e ective mean eld. With this, the Higgs mode in strongly coupled superconductors modeled by the Holstein model has been analyzed [42, 48, 78, 79]. A closely related nonequilibrium Keldysh method has been employed to study the Higgs mode in a time-resolved photoemission spectroscopy [43, 80] and time-resolved optical conductivity [81]. Another approach is to use the gauge-invariant kinetic equation [82{85], where the shift of the center-of-mass momentum of Cooper pairs (\drive e ect") caused by a laser pulse has been emphasized. For multi-band superconductors with multiple gaps such as MgB , there are not only Higgs modes corresponding to amplitude oscillations of multiple order parameters but also the so-called Leggett mode [86], which corresponds to a collective oscillation of the relative phase between di erent order parameters. Based on the mean- eld theory, possible laser excitations of the Higgs and Leggett modes in multiband superconductors have been studied [53, 87{90]. For non-s-wave superconductors, one can expect even richer dynamics of the order pa- rameters. For example, d-wave superconductors allow for various di erent symmetries of amplitude modes [91]. The quench dynamics of the d-wave [92] and p + ip-wave [93] Higgs modes have been explored. For recent experimental progresses on pump-probe spectro- scopies and third harmonic generation in d-wave superconductors, we refer to Section III C. It has also been proposed [94] that by measuring Higgs modes for various quench symmetries one can obtain information on the symmetry of the gap function (\Higgs spectroscopy"). 14 III. EXPERIMENTS In this section, we overview recent progresses on experimental observations of the Higgs mode in superconductors. Those include Raman experiments in CDW-coexisting super- conductors 2H-NbSe and TaS (Section III A), THz spectroscopies and third harmonic 2 2 generation in a pure s-wave superconductor NbN (Section III B) and in high-T cuprates (Section III C), and THz transmittance experiments in NbN under supercurrent injection (Section III D). A. Higgs mode in a superconductor with CDW The pioneering work on the observation of the Higgs mode in superconductors dates back to 1980 when a Raman experiment was performed in 2H-NbSe , a transition metal dichalco- genide in which superconductivity coexists with CDW [25, 26]. A new peak was observed below T , distinct from the amplitude mode associated with the CDW order. Although it was initially recognized as a pair breaking peak, soon after it was identi ed as the collec- tive amplitude mode of the superconducting order (i.e., the Higgs mode) [18, 19] (see also [27, 63]). A renewed Raman experiment has been conducted recently, revealing the transfer of the oscillator strength from the amplitude mode of CDW (amplitudon) to the Higgs mode [28]. Subsequently the Raman experiment under hydrostatic pressure has been performed in 2H-NbSe [30] (Figure 5). With applying the hydrostatic pressure the CDW order is suppressed and concomitantly the Raman peak identi ed as the Higgs mode was shown to disappear with leaving only the Cooper pair breaking peak. A more clear separation between the in-gap Higgs mode and the pair breaking peak was recently reported in a similar CDW- coexisting superconductor, TaS [95]. These results indicate that the existence of CDW plays an important role on the visibility of the Higgs mode in the Raman spectrum. A theoretical study has demonstrated that, due to the coupling of the superconducting order to the coexisting CDW order, the Higgs mode energy is pushed down below the superconducting gap 2 [29]. Accordingly, the Higgs mode becomes stable due to the disappearance of the decay channel into the quasiparticle continuum. 15 4 0.009 NbSe P>P 2 c 2H-NbSe NbS 0 GPa 2 2 0.04 0.006 2g 0.003 T<T 0.03 012 3 ω / 2∆ SC 0.02 P>P T>T c c (c) P>P T<T c c P<P T>T c c P<P T<T c c 0.01 0.00 020 40 60 80 100 -1 Raman shift (cm ) FIG. 2. Collapse of the superconducting Higgs mode in the pure superconducting state of 2H-NbSe , measurements and Figure 5. Raman spectra in E symmetry of 2H-NbSe measured at various temperatures and 2g 2 theoretical predictions. (a) Raman spectra in the E symmetry measured at various (P,T) positions as identified Fig. 1(a): in 2g the coexisting SC+CDW (green) and pure CDW (blue) states at ambient pressure and in the pure SC (brown) and paramagnetic pressures. Under the ambient pressure (P < P ) and in the superconducting phase (T < T ) c c (red) state at high pressure. Both the CDW amplitudon and the SC Higgs modes disappear at high pressure. A small Cooper- pairs breaking peak remains at 2 .2 is marked by the grey band ranging from 2 measured by STM [39]atambient SC SC SC (green), a sharp peak identi ed as the Higgs mode appears below the superconducting gap 2  20 pressure to the value we extrapolate at high pressure accordingly to the increase of T with pressure [26]. Inset: Raman spectra 1 in the pure SC state of 2H-NbSe (above 4 GPa) and non-CDW NbS (0 GPa) versus the Raman shift normalized to 2 2 2 SC cm marked by the gray vertical line. In the CDW phase (T > T ) (blue), only the CDW [39, 40]. (b) Theoretical Raman responses calculated in a microscopic model (see text and Appendix A) in the four phases (SC+CDW, CDW, SC,PM) for comparison with the experimental spectra in (a). t is the hopping term. The parameters amplitude mode is observed around 40 cm . Under the pressure (P > P ) where CDW collapses, are: /t=0.025, g /t=0.14 and 0.12 in the SC+CDW and SC phases, respectively. The spectra are well-reproduced SC CDW in all phases. (c) Raman response of 2H-NbSe in the pure superconducting phase above P in the E and A symmetries. 2 c 2g 1g only a pair breaking peak is discerned at T < T (brown), and no peak is identi ed at T > T c c The black line is the theoretical Raman response of a Cooper-pairs breaking peak in a two-gaps (or anisotropic gap) s-wave superconductor in the BCS regime with an additional electronic background . The form of is : (!)= a!/ b + c! .It (red). Inset: Raman spectra in the superconducting state without CDW in 2H-NbSe (above 4 barely a↵ ects the shape of the Cooper-pairs breaking peak. GPa) and non-CDW-coexisting NbS (0 GPa), both of which show only the pair breaking peak. of the SC peak above P , i.e. a Cooper-pairs breaking IV. COMPARISON WITH A MICROSCOPIC Frequency is normalized by 2. Figure is taken from Reference [30]. MODEL peak, with insight into the energy scale of the supercon- ducting gap. Consistently with this assignment, in the A symmetry there is no signature of the pure super- 1g In 2H-NbSe the phonon coupled to the CDW be- conducting state reached in 2H-NbSe above the critical 2 longs to an acoustic branch, so the single-phonon mode is B. Higgs mode in a pure s-wave superconductor pressure, due to Coulomb screening e↵ ect [48–50](see not visible as a finite-energy peak in q⇠ 0 Raman spec- Fig. 2(c)). troscopy above T .Below T the intermediate CDW CDW electron-hole excitations which couple directly to light 1. Nonadiabatic quench with a single-cycle THz pump pulse allow to make the phonon mode at Q Raman visi- CDW ble at q = 0. This gives rise to the soft phonon modes at ⇠ 40 cm . In a general approach, the Raman response One way to excite the Higgs mode is a nonadiabatic quench of the superconducting state, below T can be schematically written as CDW which can be induced by, for example, a sudden change of the pairing interaction that ph ”(!)= Z (T, ) (1) eff CDW 2 2 2 2 The disappearance of the sharp SC mode below 2 SC (! ⌦ (T )) + generates the order-parameter dynamics, ph in the pure superconducting phase demonstrates unam- biguously its intimate link with the coexisting charge- where the soft mode frequency ⌦ (T ) and damping 0 ph j(t)j cos(2 t + ) density-wave order. These findings, consisten1 tly with the are both determined by the CDW amplitude fluctuations ' 1 + a p ; (13) theory discussed below, support the Higgs type assign- and the prefactor Z ⇠ grows proportionally to eff t CDW ment of the sharp SC mode below P . the CDW order parameter[4]. The frequency ⌦ (T ) also c 0 χ’’ (a.u.) χ’’ (u.a.) b 0.8 T=4 K 0.6 nJ/cm 9.6 8.5 0.4 7.9 7.2 6.4 5.6 0.2 4.8 2Δ 4.0 0 0.0 5 10 -4 -2 0 2 4 6 8 t (ps) Pump Energy ( nJ /cm ) pp Figure 6. Higgs-mode oscillation after the nonadiabatic excitation of quasiparticles in NbN in- duced by a monocycle THz pump. (a) The temporal evolution of the change of the probe electric eld, E , as a function of the pump-probe delay time t at various pump intensities. The probe pp solid curves represent the results tted by a damped oscillation with a power-law decay. (b) The oscillation frequency f obtained from the ts and the asymptotic gap energy 2 as a function of the pump intensity. Figures are taken from Reference [45]. at long time. Here 2 is the long-time asymptotic superconducting gap, a is some constant, and  is a phase shift. The behavior (Equation 13) has been theoretically derived on the basis of the time-dependent BCS mean- eld theory [16, 33{35]. A similar dynamics has also been investigated after a laser-pulse excitation [38{41, 43, 44]. Despite intensive theoretical studies on the Higgs mode in superconductors, the observa- tion of the Higgs mode has long been elusive except for the case of 2H-NbSe , until when the ultrafast THz-pump and THz-probe spectroscopy was performed in a conventional s- wave superconductor Nb Ti N [45]. In this experiment, quasiparticles are instantaneously 1x x injected at the superconducting-gap edge to quench the superconducting order parameter nonadiabatically. A strong single-cycle THz pump pulse with the center frequency around 1 THz ( 4 meV) was generated from a LiNbO crystal by using the tilted-pulse-front method E (t =t) (arb. units) probe gate 0 f (THz) [96{98], in order to excite high density quasiparticles just above the gap of Nb Ti N thin 1x x lms with 2 = 0.72-1.3 THz ( 3.0-5.4 meV) at 4 K. The details of the THz pump-THz probe scheme were given in Ref. [99]. The subsequent dynamics of the superconducting order parameter was probed by a weaker probe THz pulse in transmission geometry. Fig- ure 6 shows pump-probe delay dependence of the transmitted probe THz electric eld at a xed point of the probe pulse waveform which re ects the order parameter dynamics. A clear oscillation is identi ed after the pump with the oscillation frequency given by the asymptotic value of the superconducting gap 2 caused by the quasiparticle injection. The decay of the oscillation is well tted by a power-law decay predicted for the Higgs-mode oscillation. With increasing the pump intensity, the oscillation period shows a softening, as the asymptotic value of 2 becomes smaller with higher quasiparticle densities. By recording the probe THz waveform at each pump-probe delay, one can extract the dynamics of the complex optical conductivity (!) =  (!) +i (!) (not shown). The spectral-weight 1 2 oscillation with frequency 2 was clearly identi ed both in the real and imaginary parts [100], the latter of which re ects the oscillation of the super uid density (i.e., the amplitude of the superconducting order parameter). The intense single-cycle THz pump pulse was crucial to excite the quasiparticles instanta- neously at the gap edge which acts as a nonadiabatic quench of the order parameter, thereby inducing a free oscillation of the Higgs mode. On the contrary, when a near-infrared optical pulse was used as a pump, a relatively slow rise time ( 20 ps) was observed for the suppres- sion of the order parameter, though dependent on the excitation uence [99, 101]. In this case, the photoexcited hot electrons (holes) emit a large amount of high frequency phonons (! > 2) (HFP) which subsequently induces the Cooper pair-breaking until quasiparticles and HFP reach the quasiequilibrium. This process was well described by the Rothwarf- Taylor phonon-bottleneck model [101], and usually takes a time longer than (2) , which breaks the nonadiabatic excitation condition needed for the quench experiment. It should be noted here that even if one uses the near-infrared or visible optical excitation, the in- stantaneous quasiparticle excitation can occur if the stimulated Raman process is ecient [102]. The decay of the Higgs mode is also an important issue. Since the Higgs-mode energy 2 coincides with the onset of quasiparticle continuum, the Higgs mode inevitably decays into the quasiparticle continuum even without any collisions [16, 17]. However, this does not 18 mean that the Higgs mode is an overdamped mode. In fact, the decay due to the collisionless 1=2 energy transfer is expected to be proportional to t in the BCS superconductor, ensuring a much longer lifetime compared with the oscillation period. The energy dispersion of the Higgs mode for nite wavenumber q has been calculated as 2 2 v  v F 2 ! ' 2 + q i q (14) 12 24 where v is the Fermi velocity [19]. From Equation 14 one can estimate that as the wavenum- ber q increases and the mode energy enters into the quasiparticle continuum, the lifetime of Higgs mode is shortened and turns into an overdamped mode. For the case of NbN,   0.6 THz and v  2 10 m/s [76] and the imaginary part exceeds the real part in Equation 14 at q > =v  0.3 (m) . In the single-cycle THz-pump experiment in a thin lm of NbN [45], the value of in-plane wavenumber q estimated from the spot size of the pump ( 1 mm) is at most  1 (mm) , which is far smaller than =v . In such a small q limit, the Higgs mode is considered as a well-de ned collective mode [19]. 2. Multicycle THz driving with subgap frequency and third harmonic generation To excite the Higgs mode in an on-resonance condition, one can use a narrowband mul- ticycle THz pump pulse with the photon energy tuned below the superconducting gap 2. With this, the amplitude of the superconducting order parameter was shown to oscillate with twice the frequency of the incident pump-pulse frequency ! during the pulse irradia- tion (Figure 7a -d ) [46]. The observed 2! modulation of the order parameter was attributed to the forced oscillation of the Higgs mode caused by electromagnetic elds. In fact, the time-dependent Ginzburg-Landau theory shows that there is a coupling term between the Higgs (H ) and eld (A) as expressed by A H (Section II A), which results in the 2!-order parameter modulation. The nonlinear electromagnetic response of superconductors has been studied in the frequency range far below 2 and near T where the Ginzburg-Landau theory is applicable [64, 70, 103{105]. In the theory developed by Gor'kov and Eliashberg (GE), a closed set of equations for the order parameter  and the eld A was derived, which predicts the 2! modulation of  and the THG from the induced supercurrent j / A [70]. The GE theory was experimentally demonstrated through the observation of THG of a microwave eld in a paramagnetically doped superconductor [104]. 19 b Pump 2ω=0.6 THz WGP NbN/MgO WGP BPF 4 8 12 16 Probe Temperature (K) c e 10 K pump 15.5 K ω=0.6 THz |E | -2 pump 0 2 4 6 8 10 Delay Time (ps) -4 14 K 0 1 2 14.5 K ω>2Δ(T) Frequency (THz) 14.8 K 15.5 K ω<2Δ(T) 13 K 12 K 0.1 10 K 4 K 0.01 0 2 4 6 8 10 2 3 4 Pump-Probe Delay Time t (ps) E (kV/cm) pp pump Figure 7. Forced Higgs oscillation and third harmonic generation from a superconducting NbN lm under the multicycle THz pump. (a) Schematic setup for the multicycle THz pump and THz probe spectroscopy. BPF: a bandpass lter, WGP: a wire grid polarizer. (b) Temperature dependence of the superconducting gap energy. Horizontal line indicates the center frequency of the pump pulse, ! = 0:6 THz. (c) Waveform of the pump THz electric eld E and the squared pump one, jE j . (d) The change in the transmitted probe THz electric eld E as a function of pump probe the pump-probe delay time t at the tempareture range ! > 2(T ) (top panel) and ! < 2(T ) pp (bottom panel). Increase of E corresponds to the reduction of the order parameter. (e) Power probe spectra of the transmitted pump THz pulse above and below T = 15 K. (f) THG intensity as a function of the pump THz eld strength. Figures are taken from Reference [46]. I |E | pump δE (arb. units) probe Intensity (arb. units) Intensity (arb. units) 2Δ(T) (THz) a c 2ω=1.6 THz 1.2 THz 1.0 THz 0.8 THz Temperature (K) Tempreature (K) 0.8 THz 0.6 THz 0.5 THz 0.4 THz 2ω/2Δ(T) Temperature (K) Figure 8. Temperature dependence of THG signal from NbN. (a) Temperature dependence of the supercondcuting gap 2(T ). Twice the THz pump frequencies 2! are shown by horizontal lines. (b) Measured THG intensities at each frequencies as a function of temperature. (c) THG intensities normalized by the temperature dependent internal eld inside the superconducting lm (3) 2 and the transmittance of the THG signal, which corresponds to the susceptibility of THG, j j . (3) 2 4 (d) Further normalized value of j j by the temperature dependent gap (T ) as a function of normalized frequency 2!=2(T ). Relative intensities between di erent frequencies are arbitrary. Data are taken from Reference [100]. The THz pump experiments have shed a new light on the investigation of the nonlinear electromagnetic response of superconductors at the gap frequency region. Subsequent to the observation of the 2! modulation of the order parameter, namely the forced oscillation of the Higgs modes, the THG was observed under the irradiation of a subgap multicycle THz pump pulse in NbN thin lms as shown in Figure 7e , f . Importantly, the THG was shown to be resonantly enhanced when a condition 2! = 2 is satis ed [46]. The temperature de- pendence of THG intensity for various frequencies in a NbN lm is shown in Figure 8: the temperature dependence of the gap 2 and the incident frequencies (Figure 8a ), the raw THG Intensity (arb. units) 2Δ (THz) (arb. units) / Δ (arb. units) data of THG intensities (Figure 8b ), those normalized by the internal eld inside the super- (3) 2 conducting lm which is proportional to the third-order susceptibility j j (Figure 8c), and the THG intensity further normalized by the proportional factor of the THG intensity (T ) versus the frequency normalized to the gap, 2!=2(T ) (Figure 8d ). These results clearly show the two-photon resonances of the Higgs mode with the eld. As previously described in Section II C, however, it was theoretically pointed out that there is a contribution also from the quasiparticle excitation (or called charge density uctu- ations) in THG, which also exhibits the 2! = 2 resonance [47]. Within the BCS mean- eld approximation in the clean limit, the quasiparticle term was shown to exceed the Higgs mode contribution in THG. At the same time, it was predicted that the THG from the quasipar- ticle term should exhibit polarization dependence with respect to the crystal axis while the Higgs term is totally isotropic in a square lattice [47]. In the experiments in NbN, the THG intensity was shown to be independent from the incident polarization angle with respect to the crystal orientation and notably the emitted THG does not have orthogonal components with respect to the incident polarization of the fundamental (!) wave [49]. Such a totally isotropic nature of THG cannot be well accounted for by the CDF term even if one consid- ers the crystal symmetry of NbN [50] and indicates the Higgs mode as the origin of THG. Recently, the importance of p  A term on the coupling between the Higgs mode and the gauge eld was elucidated. In particular, the p  A term contribution is largely enhanced when one consider the phonon retardation e ect beyond the BCS approximation [48] and more prominently when one considers the nonmagnetic impurity scattering e ect [52{54], dominating the quasiparticle term by an order of magnitude as described in Section II C. This paramagnetic term was also shown to dominate the diamagnetic term, although its ratio depends on the ratio between the coherence length and the mean free path. The NbN superconductor is in the dirty regime as manifested by a clear superconducting gap struc- ture observed in the conductivity spectrum, and the paramagnetic term should signi cantly contribute to THG. C. Higgs mode in high-T cuprates Higgs modes can exist not only in s-wave superconductors but also in unconventional superconductors. For example, in d-wave superconductors such as high-T cuprates one can 22 Figure 9. Forced oscillation of Higgs mode in an optimally doped (OP90) Bi Sr CaCu O 2 2 2 8+x measured by THz pump and optical re ection probe experiment. Temperature dependences of re ectivity change R=R as a function of pump-probe delay time in (a) A and (b) B compo- 1g 1g nents. Red dashed lines indicate T . (c) The A and B components against the pump-probe c 1g 1g delay time at 10 K. Fitting curves are also shown by solid lines. (d) Temperature dependences of the A decaying component (blue) (incoherent quasiparticle excitation), the A oscillatory 1g 1g component (red) (forced Higgs oscillation), and the B oscillatory component (green) (likely the 1g charge density uctuation). Figures are taken from Reference [55]. expect various types of collective modes, including not only the totally isotropic oscillation of the gap function in relative momentum space (A mode) but also those that oscillate 1g anisotropically (e.g., A , B , B modes) [91]. These are reminiscent of Bardasis-Schrie er 2g 1g 2g mode for the case of s-wave superconductors [106]. Possible collective modes have been classi ed based on the point group symmetry of the lattice [91], and the relaxation behavior of the A mode induced by a quench has been discussed [92]. A theoretical proposal has 1g been made to observe the Higgs mode in a d-wave superconductor through a time- and angle-resolved photoemission spectroscopy [80]. 23 The observation of the Higgs mode in d-wave superconductors was recently made in THz- pump and optical-probe experiments in high-T cuprates, Bi Sr CaCu O (Bi2212) [55]. c 2 2 2 8+x An oscillatory signal of the optical re ectivity that follows the squared THz electric eld was observed, which is markedly enhanced below T as shown in Figure 9. This signal was interpreted as the THz-pump induced optical Kerr e ect, namely the third-order nonlinear e ect induced by the intense THz pump pulse. In the Bi2212 system, both A and B 1g 1g symmetry components with respect to the polarization-angle dependence (which should be distinguished from the relative-momentum symmetry classi cation mentioned above) were observed in the THz-Kerr signal. The doping dependence shows that the A component 1g is dominant in all the measured samples from under to near-optimal region. From the comparison with the BCS calculation of the nonlinear susceptibility, the A component was 1g assigned to be the d-wave Higgs-mode contribution. The observed oscillatory signal corresponds to the 2! modulation of the order parameter as observed in an s-wave NbN superconductor under the subgap multicycle THz pump irradiation. Accordingly, like in the case of s-wave systems, one can expect the THG signal from high-T cuprates. Recently, THG was indeed observed ubiquitously in La Sr CuO , c 1:84 0:16 4 DyBa Cu O , YBa Cu O , and overdoped Bi Sr CaCu O [56] by using 0.7 THz 2 3 7x 2 3 7x 2 2 2 8+x coherent and intense light source at the TELBE beamline at HZDR. D. Higgs mode in the presence of supercurrents So far, we have seen the Higgs mode excitation by the nonadiabatic quench with instan- taneous quasiparticle injection or through the quadratic coupling with the electromagnetic wave. Recently, it has been theoretically shown that when there is a condensate ow (i.e., supercurrent), the Higgs mode linearly interacts with electromagnetic waves with the elec- tric eld polarized along the direction of the supercurrent ow [57]. Namely, the Higgs mode should be observed in the optical conductivity spectrum at the superconducting gap ! = 2 under the supercurrent injection. This e ect is explained by the momentum term 2 2 in the action, S / Q (t)j(t)j dtdr, where Q(t) = Q +Q (t) is the gauge-invariant mo- mentum of the condensate, Q is the dc supercurrent part, and Q (t) = Re[Q exp(i t)] is the time-dependent part driven by the ac probe eld, respectively. (t) =  + (t) is the time-dependent superconducting order parameter. The action S includes the integral of 24 0 σ (ω) / σ 1 N 0 -1 -1 σ (ω) (Ω cm ) 200 6×10 ab Parallel, 5 K 2.6 A 2.5 A 2.3 A 1.9 A 0 A 0 0 2Δ MB -50 4 5 6 7 8 Photon Energy (meV) 200 0.03 0.5 Perpendicular, 5 K 0.4 2.6 A 0.02 2.3 A 0.3 1.9 A 0 A 0.2 0.01 2Δ 0.1 Parallel -0.1 Perpendicular -50 -0.01 4 5 6 7 8 4 5 6 7 8 Photon Energy (meV) Photon Energy (meV) Figure 10. (a) Schematic view of the terahertz transmittance experiments under supercurrent injection. The change of the optical conductivity induced by supercurrent injection taken with the THz probe electric eld (b) parallel or (c) perpendicular to the current direction. The optical conductivity spectra measured without the current is also plotted in (b). The dashed line represents a t by the Mattis-Bardeen model [107, 108], and the superconducting gap estimated from the t is indicated by the vertical arrow. (d) Theoretically expected changes of optical conductivity induced by the supercurrent injection with respect to the normal state conductivity  . Figures are taken from Reference [58]. Q and  Q Q where  is the Fourier component of the oscillating order (2 parameter. The rst term corresponds to the quadratic coupling of the Higgs mode to the gauge eld as described Section II. The second term indicates that the Higgs mode linearly couples with the gauge eld under the supercurrent with the condensate momentum Q parallel to the probe electric eld. Recently, in accordance with the theoretical prediction, the Higgs mode was observed in the optical conductivity in an s-wave superconductor NbN thin lm under the supercurrent injection by using THz time-domain spectroscopy [58]. Figure 10a shows the schematic experimental setup. Figures 10b and c show the change of the real part of the optical conductivity induced by the supercurrent injection with the probe polarization (b ) parallel -1 -1 δσ (ω) (Ω cm ) δσ (ω) / σ -1 -1 1 N δσ (ω) (Ω cm ) 1 and (c) orthogonal to the current direction. A clear peak is observed at the gap edge ! = 2 in the parallel geometry, showing a good agreement with the spectra calculated on the basis of the theory developed by Moor et al. [57] as represented in Figure 10d . This result indicates that, without resorting to the sophisticated nonlinear THz spectroscopy, the Higgs mode can be observed in the linear response function when the condensate has a nite momentum, expanding the feasibility for the detection of the Higgs mode in a wider range of materials. The visibility of the Higgs mode in the optical conductivity has also been addressed in a two-dimensional system near a quantum critical point [109, 110] (see also [111, 112]). Ex- perimentally, an extra spectral weight in the optical conductivity below the superconducting gap was observed in a strongly disordered superconducting lm of NbN, which was inter- preted in terms of the Higgs mode with its mass pushed below the pair breaking gap due to the strong disorder [113]. However, the assignment of the extra spectral weight below the gap remains under debates. It can also be described by the Nambu-Goldstone mode which acquires the electric dipole in strongly disordered superconductors [114{117]. The extra optical conductivity was also accounted for by the disorder induced broadening of the quasiparticle density of states below the gap [118]. IV. FUTURE PERSPECTIVE Having seen the Higgs mode in s-wave and d-wave superconductors, it is fascinating to extend the observation of the Higgs mode to a variety of conventional/unconventional superconductors. The study of the Higgs mode in multiband superconductors would be important, as it would give deeper insights into the interband couplings. For instance, the application of the nonlinear THz spectroscopy to iron-based superconductors is highly intriguing, as it may provide information on the interband interactions and pairing symmetry [87, 119, 120]. The Leggett mode, namely the collective mode associated with the relative phase of two condensates [86] is also expected to be present in multiband superconductors. The Leggett mode has been observed in a multiband superconductor MgB by Raman spectroscopy [121]. The nonlinear coupling between the Leggett mode and THz light has also been discussed [53, 87{90]. Recently, a THz-pump THz probe study in MgB has been reported, and 26 the results were interpreted in terms of the Leggett mode [122]. However, the dominant contribution of the Higgs mode over the Leggett mode in the nonlinear THz responses was theoretically pointed out in the case of dirty-limit superconductors [53]. Therefore, further experimental veri cation of the Higgs and Leggett modes is left as a future problem. The ultimate fate of the behavior of the Higgs mode in a strongly correlated regime and, what is more, in a BCS-BEC crossover regime is an intriguing problem. In particular, the decay pro le of the Higgs mode has been predicted to change from the BCS to BEC regime [36, 37, 123{125]. Experimentally, such a study has been realized in a cold-atom system, showing a broadening of the Higgs mode in the BEC regime [61]. Further investigation on the Higgs mode in the BCS-BEC crossover regime will be a future challenge in solid-state systems. The Higgs mode in spin-triplet p-wave superconductors (whose candidate materials in- clude Sr RuO ) is another issue to be explored in the future. Like in the case of super uid 2 4 He [22, 59, 126], multiple Higgs modes are expected to appear due to the spontaneous symmetry breaking in the spin channel, and its comparison with high-energy physics would be interesting. As yet another new paradigm, the observation of the Higgs mode would pave a new pathway for the study of nonequilibrium phenomena, in particular for the photoinduced superconductivity [127{130]. Being a ngerprint of the order parameter with a picosecond time resolution, the observation of the Higgs mode in photoinduced states should o er a di- rect evidence for nonequilibrium superconductivity. In strongly correlated electron systems exempli ed by the unconventional superconductors, elucidation of the interplay between competing and/or coexisting orders with superconductivity is an important issue to under- stand the emergent phases. The time-domain study of collective modes associated with those orders is expected to provide new insights for their interplay. V. DISCLOSURE STATEMENT The authors are not aware of any aliations, memberships, funding, or nancial holdings that might be perceived as a ecting the objectivity of this review. 27 ACKNOWLEDGMENTS We wish to acknowledge valuable discussions with Hideo Aoki, Yann Gallais, Dirk Manske, Stefan Kaiser, and Seiji Miyashita. We also acknowledge coworkers, Yuta Murotani, Keisuke Tomita, Kota Katsumi, Sachiko Nakamura, Naotaka Yoshikawa, and in particular Ryusuke Matsunaga for fruitful discussions and their support for preparing the manuscript. [1] Nambu Y. 2011. BCS: 50 Years, edited by L. N. Cooper and D. Feldman (World Scienti c, Singapore) [2] Ginzburg VL, Landau LD. 1950. Zh. Eksp. Teor. Fiz. 20:1064 [3] Bogoliubov NN. 1958. J. Exptl. Theoret. Phys. 34:58 [1958. Sov. Phys. 34:41] [4] Anderson PW. 1958. Phys. Rev. 110:827 [5] Anderson PW. 1958. Phys. Rev. 112:1900 [6] Nambu Y. 1960. Phys. Rev. 117:648 [7] Goldstone J. 1961. Nuovo Cim. 19:154 [8] Goldstone J, Salam A, Weinberg S. 1962. Phys. Rev. 127:965 [9] Anderson PW. 1963. Phys. Rev. 130:439 [10] Englert F, Brout R. 1964. Phys. Rev. Lett. 13:321. [11] Higgs PW. 1964. Phys. Lett. 12:132 [12] Higgs PW. 1964. Phys. Rev. Lett. 13:508 [13] Guralnik GS, Hagen CR, Kibble TWB. 1964. Phys. Rev. Lett. 13:585 [14] Anderson PW. 2015. Nat. Phys. 11:93 [15] Bardeen J, Cooper LN, Schrie er JR. 1957. Phys. Rev. 108:1175 [16] Volkov AF, Kogan SM. 1974. Sov. Phys. JETP 38:1018 [17] Kulik IO, Entin-Wohlman O, Orbach R. 1981. J. Low Temp. Phys. 43:591 [18] Littlewood PB, Varma CM. 1981. Phys. Rev. Lett. 47:811 [19] Littlewood PB, Varma CM. 1982. Phys. Rev. B 26:4883 [20] Nambu Y, Jona-Lasinio G. 1961. Phys. Rev. 122:345 [21] Nambu Y. 1985. Physica 15D:147 [22] Volovik GE, Zubkov MA. 2014. J. Low Temp. Phys. 175:486 28 [23] ATLAS Collaboration. 2012. Phys. Lett. B 716:1 [24] CMS Collaboration. 2012. Phys. Lett. B 716:30 [25] Sooryakumar R, Klein MV. 1980. Phys. Rev. Lett. 45:660 [26] Sooryakumar R, Klein MV. 1981. Phys. Rev. B 23:3213 [27] Varma C. 2002. J. Low Temp. Phys. 126:901 [28] M easson MA, Gallais Y, Cazayous M, Clair B, Rodi ere P, Cario L, Sacuto A. 2014. Phys. Rev. B 89:060503(R) [29] Cea T, Benfatto L. 2014. Phys. Rev. B 90:224515 [30] Grasset R, Cea T, Gallais Y, Cazayous M, Sacuto A, Cario L, Benfatto L, M easson MA. 2018. Phys. Rev. B 97:094502 [31] Barankov RA, Levitov LS, Spivak BZ. 2004. Phys. Rev. Lett. 93:160401 [32] Yuzbashyan EA, Altshuler BL, Kuznetsov VB, Enolskii VZ. 2005. Phys. Rev. B 72:220503 [33] Barankov RA, Levitov LS. 2006. Phys. Rev. Lett. 96:230403 [34] Yuzbashyan EA, Tsyplyatyev O, Altshuler BL. 2006. Phys. Rev. Lett. 96:097005 [35] Yuzbashyan EA, Dzero M. 2006. Phys. Rev. Lett. 96:230404 [36] Gurarie V. 2009. Phys. Rev. Lett. 103:075301 [37] Tsuji N, Eckstein M, Werner P. 2013. Phys. Rev. Lett. 110:136404 [38] Papenkort T, Axt VM, Kuhn T. 2007. Phys. Rev. B 76:224522 [39] Papenkort T, Kuhn T, Axt VM. 2008. Phys. Rev. B 78:132505 [40] Schnyder AP, Manske D, Avella A. 2011. Phys. Rev. B 84:214513 [41] Krull H, Manske D, Uhrig GS, Schnyder AP. 2014. Phys. Rev. B 90:014515 [42] Tsuji N, Aoki H. 2015. Phys. Rev. B 92:064508 [43] Kemper AF, Sentef MA, Moritz B, Freericks JK, Devereaux TP. 2015. Phys. Rev. B 92:224517 [44] Chou Y-Z, Liao Y, Foster M. S. 2017. Phys. Rev. B95:104507 [45] Matsunaga R, Hamada YI, Makise K, Uzawa Y, Terai H, Wang Z, Shimano R. 2013. Phys. Rev. Lett. 111:057002 [46] Matsunaga R, Tsuji N, Fujita H, Sugioka A, Makise K, Uzawa Y, Terai H, Wang Z, Aoki H, Shimano R. 2014. Science 345:1145 [47] Cea T, Castellani C, Benfatto L. 2016. Phys. Rev. B 93:180507(R) [48] Tsuji N, Murakami Y, Aoki H. 2016. Phys. Rev. B 94:224519 29 [49] Matsunaga R, Tsuji N, Makise K, Terai H, Aoki H, Shimano R. 2017 Phys. Rev. B 96:020505(R) [50] Cea T, Barone P, Castellani C, Benfatto L. 2018. Phys. Rev. B 97:094516 [51] Jujo T. 2015. J. Phys. Soc. Jpn. 84:114711 [52] Jujo T. 2018. J. Phys. Soc. Jpn. 87:024704 [53] Murotani Y, Shimano R. 2019. arXiv:1902.01104 [54] Silaev M. 2019. arXiv:1902.01666 [55] Katsumi K, Tsuji N, Hamada YI, Matsunaga R, Schneeloch J, Zhong RD, Gu GD, Aoki H, Gallais Y, Shimano R. 2018. Phys. Rev. Lett. 120:117001 [56] Chu H, Kim M-J, Katsumi K, Kovalev S, Dawson RD, Schwarz L, Yoshikawa N, Kim G, Putzky D, Li ZZ, Ra y H, Germanskiy S, Deinert J-C, Awari N, Ilyakov I, Green B, Chen M, Bawatna M, Christiani G, Logvenov G, Gallais Y, Boris AV, Keimer B, Schnyder A, Manske D, Gensch M, Wang Z, Shimano R, Kaiser S. 2019. arXiv:1901.06675 [57] Moor A, Volkov AF, Efetov KB. 2017. Phys. Rev. Lett. 118:047001 [58] Nakamura S, Iida Y, Murotani Y, Matsunaga R, Terai H, Shimano R. 2018. arXiv:1809.10335 [59] Vollhardt D, W ol e P. 1990. The Super uid Phases of Helium 3 (Taylor & Francis) [60] Endres M, Fukuhara T, Pekker D, Cheneau M, Schau P, Gross C, Demler E, Kuhr S, Bloch I. 2012. Nature 487:454 [61] Behrle A, Harrison T, Kombe J, Gao K, Link M, Bernier J-S, Kollath C, K ohl M. 2018. Nat. Phys. 14:781 [62] Ruegg  C, Normand B, Matsumoto M, Furrer A, McMorrow DF, et al. 2008. Phys. Rev. Lett. 100:205701 [63] Pekker D, Varma CM. 2015. Annu. Rev. Condens. Matter Phys. 6:269 [64] Abraham E, Tsuneto T. 1966. Phys. Rev. 152:416 [65] Schmid A. 1966. Phys. Kond. Mater. 5:302 [66] Caroli C, Maki K. 1967. Phys. Rev. 159:306 [67] Ebisawa H, Fukuyama H. 1971. Prog. Theor. Phys. 46:1042 [68] S a de Melo CAR, Randeria M, Engelbrecht JR. 1993. Phys. Rev. Lett. 71:3202 [69] Tsuchiya S, Yamamoto D, Yoshii R, Nitta M. 2018. Phys. Rev. B 98:094503 [70] Gor'kov LP, Eliashberg GM. 1968. Sov. Phys. JETP 27:328 30 [71] Gulian AM, Zharkov GF. 1999. Nonequilibrium electrons and phonons in superconductors (Kluwer Academic/Plenum Publishers) [72] Kopnin N. 2001. Theory of Nonequilibrium Superconductivity (Oxford University Press) [73] Schrie er JR. 1999. Theory of Superconductivity (Westview Press) [74] Kihlstrom KE, Simon RW, Wolf SA. 1985. Phys. Rev. B 32:1843 [75] Brorson SD, Kazeroonian A, Moodera JS, Face DW, Cheng TK, Ippen EP, Dresselhaus MS, Dresselhaus G. 1990. Phys. Rev. Lett. 64:2172 [76] Chockalingam SP, Chand M, Jesudasan J, Tripathi V, Raychaudhuri P. 2008. Phys. Rev. B 77:214503 [77] Aoki H, Tsuji N, Eckstein M, Kollar M, Oka T, Werner P. 2014. Rev. Mod. Phys. 86:779 [78] Murakami Y, Werner P, Tsuji N, Aoki H. 2016. Phys. Rev. B 93:094509 [79] Murakami Y, Werner P, Tsuji N, Aoki H. 2016. Phys. Rev. B 94:115126 [80] Nosarzewski B, Moritz B, Freericks JK, Kemper AF, Devereaux TP. 2017. Phys. Rev. B 96:184518 [81] Kumar A, Kemper AF. 2019. arXiv:1902.09549 [82] Yu T, Wu MW. 2017. Phys. Rev. B 96:155311 [83] Yu T, Wu MW. 2017. Phys. Rev. B 96:155312 [84] Yang F, Wu MW. 2018. Phys. Rev. B 98:094507 [85] Yang F, Wu MW. 2018. arXiv:1812.06622 [86] Leggett AJ. 1966. Prog. Theor. Phys. 36:901 [87] Akbari A, Schnyder AP, Manske D, Eremin I. 2013. Europhys. Lett. 101:17002 [88] Krull H, Bittner N, Uhrig GS, Manske D, Schnyder AP. 2016. Nat. Commun. 7:11921 [89] Cea T, Benfatto L. 2016. Phys. Rev. B 94:064512 [90] Murotani Y, Tsuji N, Aoki H. 2017. Phys. Rev. B 95:104503 [91] Barlas Y, Varma CM. 2013. Phys. Rev. B 87:054503 [92] Peronaci F, Schir o M, Capone M. 2015. Phys. Rev. Lett. 115:257001 [93] Foster MS, Dzero M, Gurarie V, Yuzbashyan EA. 2013. Phys. Rev. B 88:104511 [94] Fauseweh B, Schwarz L, Tsuji N, Cheng N, Bittner N, Krull H, Berciu M, Uhrig GS, Schnyder AP, Kaiser S, Manske D. 2017. arXiv:1712.07989 [95] Grasset R, Gallais Y, Sacuto A, Cazayous M, Manas-V ~ alero S, Coronado E, M easson MA. 2019. Phys. Rev. Lett. 122:127001 31 [96] Hebling J, Yeh KL, Ho mann MC, Bartal B, Nelson KA. 2008. J. Opt. Soc. Am. B 25:B6 [97] Watanabe S, Minami N, Shimano R. 2011. Optics Express 19:1528 [98] Shimano R, Watanabe S, Matsunaga R. 2012. J. Infrared Millim. Terahz Waves 33:861 [99] Matsunaga R, Shimano R. 2012. Phys. Rev. Lett. 109:187002 [100] Matsunaga R, Shimano R. 2017. Phys. Scr. 92:024003 [101] Beck M, Klammer M, Lang S, Leiderer P, Kabanov VV, Gol'tsman GN, Demsar J. 2011. Phys. Rev. Lett. 107:177007 [102] Mansart B, Lorenzana J, Mann A, Odeh A, Scarongella M, Chergui M, Carbone F. 2013. Proc. Natl. Acad. Sci. USA 110:4539 [103] Gor'kov LP, Eliashberg GM. 1969. Sov. Phys. JETP 29:698 [104] Amato JC, McLean WL. 1976. Phys. Rev. Lett. 37:930 [105] Entin-Wohlman O. 1978. Phys. Rev. B 18:4762 [106] Bardasis A, Schrie er JR. 1961. Phys. Rev. 121:1050 [107] Mattis DC, Bardeen J. 1958. Phys. Rev. 111:412 [108] Zimmermann W, Brandt EH, Bauer M, Seider E, Genzel L. 1991. Physica C: Supercond. 183:99 [109] Podolsky D, Auerbach A, Arovas DP. 2011. Phys. Rev. B 84:174522 [110] Gazit S, Podolsky D, Auerbach A. 2013. Phys. Rev. Lett. 110:140401 [111] Sachdev S. 1999. Phys. Rev. B 59:14054 [112] Zwerger W. 2004. Phys. Rev. Lett. 92:027203 [113] Sherman D, Pracht US, Gorshunov B, Poran S, Jesudasan J, Chand M, Raychaudhuri P, Swanson M, Trivedi N, Auerbach A, Scheer M, Frydman A, Dressel M. 2015. Nat. Phys. 11:188 [114] Cea T, Bucheli D, Seibold G, Benfatto L, Lorenzana J, Castellani C. 2014. Phys. Rev. B 89:174506 [115] Cea T, Castellani C, Seibold G, Benfatto L. 2015. Phys. Rev. Lett. 115:157002 [116] Pracht US, Cea T, Bachar N, Deutscher G, Farber E, Dressel M, Schefer M, Castellani C, Garc a-Garc a AM, Benfatto L. 2017. Phys. Rev. B 96:094514 [117] Seibold G, Benfatto L, Castellani C. 2017. Phys. Rev. B 96:144507 [118] Cheng B, Wu L, Laurita NJ, Singh H, Chand M, Raychaudhuri P, Armitage NP. 2016. Phys. Rev. B 93:180511 32 [119] Maiti S, Hirschfeld PJ. 2015. Phys. Rev. B 92:094506 [120] Muller  MA, Shen P, Dzero M, Eremin I. 2018. Phys. Rev. B 98:024522 [121] Blumberg G, Mialitsin A, Dennis BS, Klein MV, Zhigadlo ND, Karpinski J. 2007. Phys. Rev. Lett. 99:227002 [122] Giorgianni F, Cea T, Vicario C, Hauri CP, Withanage WK, Xi X, Benfatto L. 2019. Nat. Phys. online publication [123] Scott RG, Dalfovo F, Pitaevskii LP, Stringari S. 2012. Phys. Rev. A 86:053604 [124] Yuzbashyan EA, Dzero M, Gurarie V, Foster MS. 2015. Phys. Rev. A 91:033628 [125] Tokimoto J, Tsuchiya S, Nikuni T. 2019. J. Phys. Soc. Jpn. 88:023601 [126] W ol e P. 1977. Physica B 90:96 [127] Fausti D, Tobey RI, Dean N, Kaiser S, Dienst A, Ho mann MC, Pyon S, Takayama T, Takagi H, Cavalleri A. 2011. Science 331:189 [128] Kaiser S, Hunt CR, Nicoletti D, Hu W, Gierz I, Liu HY, Le Tacon M, Loew T, Haug D, Keimer B, Cavalleri A. 2014. Phys. Rev. B 89:184516 [129] Hu W, Kaiser S, Nicoletti D, Hunt CR, Gierz I, Ho mann MC, Le Tacon M, Loew T, Keimer B, Cavalleri A. 2014. Nat. Mater. 13:705 [130] Mitrano M, Cantaluppi A, Nicoletti D, Kaiser S, Perucchi A, Lupi S, Di Pietro P, Pontiroli D, Ricc o M, Clark SR, Jaksch D, Cavalleri A. 2016. Nature 530:461

Journal

Condensed MatterarXiv (Cornell University)

Published: Jun 22, 2019

There are no references for this article.