NLHE-induced Electro-optic Kerr Effect in WTe₂

• NLHE-induced EOKE is a phenomenon where nonlinear Hall conductivity in non-magnetic materials alters optical anisotropy, leading to a Kerr-like rotation of reflected light.
• The effect in materials like monolayer WTe₂ is predominantly governed by the Berry curvature dipole, distinguishing it from conventional magneto-optic Kerr effects.
• Accurate characterization of NLHE-induced EOKE requires precise control of optical geometry and transport symmetry to separate intrinsic nonlinear Hall contributions from cascaded or bias-induced electro-optic effects.

NLHE-induced electro-optic Kerr effect (EOKE) denotes a Kerr rotation and ellipticity of reflected light that arises in non-magnetic, time-reversal-symmetric materials because a nonlinear Hall response modifies the optical anisotropy of the medium. In the formulation developed for monolayer WTe$_2$, the relevant object is the second-order nonlinear Hall conductivity $\sigma_{(2)}$, which alters the dielectric tensor, splits the optical polarization eigenmodes, and thereby produces a reflected-light polarization rotation analogous to the magneto-optical Kerr effect (MOKE), but without magnetization or a linear Hall conductivity (Li et al., 14 Jul 2025). The term sits at the intersection of nonlinear transport and electro-optics, and adjacent literatures use related “Kerr-like” or EOKE language for distinct mechanisms, including classical DC Kerr birefringence, cascaded $\chi^{(2)}$-electro-optic processes, and THz-driven surface-electronic responses; the topic is therefore defined as much by what is excluded as by what is included.

1. Conceptual definition and relation to earlier Kerr phenomena

In conventional MOKE, the Kerr angle is tied to the linear Hall conductivity $\sigma_{xy}$ and therefore to magnetic order or other forms of time-reversal breaking. The NLHE-induced version generalizes that logic to non-magnetic systems: the linear Hall response vanishes by time-reversal symmetry, but a second-order nonlinear Hall conductivity remains nonzero and feeds into the optical response. The resulting effect is a Kerr-like rotation of reflected light caused by nonlinear transport rather than by magnetization (Li et al., 14 Jul 2025).

This usage differs from the classical DC Kerr effect. In the DC Kerr setting, a propagating high-frequency beam interacts with a strong stationary electric field in a medium with a cubic Kerr nonlinearity, producing a polarization-dependent phase shift and therefore elliptical polarization after propagation. In that picture, the operative constitutive ingredient is $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$, and the external field creates an effective birefringence through $\chi^{(3)}$ rather than through nonlinear Hall transport (Eptaminitakis et al., 2 May 2025). The silicon-photonics realization follows the same logic: silicon lacks an intrinsic $\chi^{(2)}$, but a static bias converts the third-order nonlinearity into an effective linear electro-optic response proportional to $F_{DC}F_{RF}$, enabling Kerr-dominated high-speed modulation in reverse-biased PIN junction waveguides (Peltier et al., 2023).

The terminological overlap matters because “electro-optic Kerr effect” can therefore refer to at least three distinct classes of phenomena: a classical field-induced $\chi^{(3)}$ birefringence, a transport-driven Kerr rotation associated with the NLHE, and Kerr-like electro-optic responses produced by cascaded or THz-mediated processes. A precise use of the term reserves “NLHE-induced EOKE” for the second class.

2. Transport-to-optics framework

The theoretical construction for NLHE-induced EOKE starts from Maxwell’s equation in a medium,

$$\nabla(\nabla\cdot\boldsymbol{E})-\nabla^2\boldsymbol{E} =-\mu\frac{\partial(\boldsymbol{J}+\partial_t\boldsymbol{D})}{\partial t}, \qquad \boldsymbol{D}=\varepsilon\boldsymbol{E},$$

together with a current expanded to second order in the optical field. In the notation used for monolayer WTe$\sigma_{(2)}$0,

$\sigma_{(2)}$1

Substitution into the wave equation yields an optical eigenvalue problem in which the nonlinear conductivity enters through

$\sigma_{(2)}$2

The nonlinear contribution survives only when $\sigma_{(2)}$3; because $\sigma_{(2)}$4, the relevant surviving case is $\sigma_{(2)}$5 (Li et al., 14 Jul 2025).

For a 2D thin film at normal incidence, the optical response is described by two eigenmodes $\sigma_{(2)}$6 with complex refractive indices $\sigma_{(2)}$7. Their reflection coefficients are

$\sigma_{(2)}$8

The Kerr angle is then defined from the phase difference between the reflected eigenmodes,

$\sigma_{(2)}$9

In this framework, EOKE is the optical manifestation of a nonlinear-transport-induced splitting of polarization eigenmodes rather than a direct consequence of magnetization (Li et al., 14 Jul 2025).

The second-order conductivity is decomposed into Berry curvature dipole (BCD), Drude, injection, and shift contributions. The BCD term is written as

$\chi^{(2)}$0

while the Drude, injection, and shift terms are treated on the same footing through quantum kinetics. For monolayer WTe$\chi^{(2)}$1, the calculated EOKE is predominantly governed by the BCD; the other nonlinear mechanisms are present but much smaller (Li et al., 14 Jul 2025).

3. Monolayer WTe$\chi^{(2)}$2 as the prototypical NLHE-induced EOKE platform

Monolayer WTe$\chi^{(2)}$3 is the canonical material platform because it is non-magnetic, can preserve time-reversal symmetry, lacks inversion symmetry in the relevant phase, and combines strong spin-orbit coupling, electron-hole compensation, and high mobility with a large Berry curvature dipole. The analysis focuses on the $\chi^{(2)}$4 phase and employs a six-band $\chi^{(2)}$5 model from Shi and Song, with an out-of-plane electric field incorporated via a term $\chi^{(2)}$6 containing Rashba- and Ising-like couplings $\chi^{(2)}$7 and $\chi^{(2)}$8, so that $\chi^{(2)}$9 (Li et al., 14 Jul 2025).

The central theoretical result is that the EOKE signal is overwhelmingly dominated by the BCD. The Drude, injection, and shift terms are retained in the formalism, but the BCD contribution is the only significant nonlinear modification once the substantial intrinsic in-plane optical anisotropy background, $\sigma_{xy}$0, is included. In this sense, the Kerr rotation in WTe$\sigma_{xy}$1 is described as the linear anisotropy background plus a BCD-induced nonlinear correction (Li et al., 14 Jul 2025).

Several dynamical signatures are emphasized. Over the optical-frequency range studied, the nonlinear conductivity changes only weakly with frequency, and the EOKE is correspondingly approximately frequency-independent in magnitude. By contrast, the Kerr angle exhibits time-dependent oscillations tied to second-harmonic timing, with period $\sigma_{xy}$2; at $\sigma_{xy}$3, the period is $\sigma_{xy}$4, and across $\sigma_{xy}$5 Hz to $\sigma_{xy}$6 Hz the oscillation period decreases as expected while remaining locked to the second-harmonic timescale. The Kerr amplitude grows strongly with electric field in the range $\sigma_{xy}$7 to $\sigma_{xy}$8, for $\sigma_{xy}$9 ps and $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$0 eV, so the nonlinear contribution becomes progressively more visible relative to the linear background (Li et al., 14 Jul 2025).

These features motivate time-resolved Kerr effect (TRKE) as an optical probe of the NLHE. The proposed observables are a Kerr rotation whose amplitude grows with driving field, millisecond-scale temporal oscillations, and a dominant BCD fingerprint. Within the paper’s logic, the effect functions as the nonlinear-transport analogue of MOKE: nonlinear Hall conductivity modifies the dielectric tensor and refractive indices, and the ensuing mode-dependent reflection phases generate the Kerr rotation (Li et al., 14 Jul 2025).

4. Experimental analogues and Kerr-like electro-optic responses beyond the strict NLHE definition

The closest experimentally established neighboring class is the cascaded Kerr-like response in LiNbO$P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$1. There, femtosecond optical pump-probe measurements show that the near-zero-delay signal is not due to a single mechanism but to the sum of an instantaneous third-order Kerr response and a cascaded second-order process in which the pump first generates THz radiation by optical rectification and the THz field then changes the optical refractive index through the electro-optic effect. In the paper’s language, this is a THz-generation-mediated electro-optic Kerr effect in the cascaded sense, not a pure intrinsic $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$2 response (Wang et al., 2019).

The distinction is formal as well as phenomenological. For the ordinary Kerr contribution,

$P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$3

so the signal has no negative part and its Fourier transform has a finite dc component. For the Kerr-like channel, the THz field is generated through $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$4 and acts through the electro-optic tensor $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$5, so that the measured spectrum becomes

$P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$6

A major result is that phase mismatch between the THz wave and the optical pulse reshapes the Kerr-like spectrum without eliminating it. At $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$7, the waveform is a one-cycle pulse of about $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$8 ps with a negative part and a spectrum extending up to $P=\epsilon_0\chi^{(1)}E+\epsilon_0\chi^{(3)}|E|^2E$9 THz with interference structure; at $\chi^{(3)}$0, the interaction length is shorter, the interference structure becomes less obvious, and high-frequency components become more prominent. The same measurements also show a long oscillatory tail attributed to low-frequency phonon polaritons with extracted frequency $\chi^{(3)}$1 THz (Wang et al., 2019).

A second experimental analogue is the THz electro-optic effect in Bi$\chi^{(3)}$2Se$\chi^{(3)}$3, where an intense single-cycle THz pump and a weak $\chi^{(3)}$4 nm femtosecond probe reveal a transient change in the polarization state of the reflected probe. For small incidence angles, the signal is essentially proportional to the rotation angle, $\chi^{(3)}$5, and the measured waveform is fit by

$\chi^{(3)}$6

The linear and quadratic components are of comparable magnitude; the former shows three-fold rotational symmetry consistent with the $\chi^{(3)}$7 surface symmetry, whereas the latter is nearly independent of crystal orientation and is naturally identified as the Kerr-like or EOKE-like component. The inferred response delay is sub-100 fs to about $\chi^{(3)}$8 fs, the effect vanishes in the non-topological metal state at indium concentration $\chi^{(3)}$9, and it can be quenched by a femtosecond pre-pulse with recovery on the order of $\chi^{(2)}$0 ps. The authors associate the effect primarily with surface Dirac electronic states and discuss the quadratic term in terms of intraband nonlinear Hall or scattering-driven anisotropy of the surface electrons (Melnikov et al., 2024).

These two systems clarify an important boundary condition. A Kerr-like electro-optic signal may scale as $\chi^{(2)}$1 and yet originate from a cascaded $\chi^{(2)}$2+EO process, or from prompt THz-driven surface-electronic dynamics, rather than from the specific Berry-curvature-dipole-driven nonlinear Hall conductivity used to define NLHE-induced EOKE in WTe$\chi^{(2)}$3.

5. Distinct mechanisms often grouped under electro-optic Kerr terminology

The literature uses closely related language for several physically different effects. The following comparison isolates the mechanism rather than the nomenclature.

Mechanism class	Key driver	Relation to NLHE-induced EOKE
Monolayer WTe$\chi^{(2)}$4 EOKE	Nonlinear Hall conductivity, predominantly BCD	Direct realization
LiNbO$\chi^{(2)}$5 Kerr-like response	Optical rectification $\chi^{(2)}$6 electro-optic cascade	Kerr-like, but not intrinsic NLHE
Bi$\chi^{(2)}$7Se$\chi^{(2)}$8 THz EO effect	Linear $\chi^{(2)}$9 and quadratic $F_{DC}F_{RF}$0 surface-electronic response	EOKE-like quadratic term; not formulated as BCD-driven NLHE
Intrinsic electro-optic Kerr rotation in metals	$F_{DC}F_{RF}$1 mixed conductivity $F_{DC}F_{RF}$2	Explicitly distinct from usual NLHE-induced EOKE
Gapped bilayer graphene metallic Kerr EO	Non-Drude “Snap” term	Explicitly not an NLHE/Berry-curvature-dipole Kerr mechanism
Silicon DC Kerr modulation	Bias-enabled $F_{DC}F_{RF}$3 effective linear EO response	Classical EOKE usage rather than transport NLHE

The paper on intrinsic electro-optic Kerr rotation makes the distinction especially explicit. It identifies a previously overlooked contribution in isotropic nonmagnetic homogeneous metals arising from the interplay of matter, a static homogeneous electric field $F_{DC}F_{RF}$4, and the magnetic component of light $F_{DC}F_{RF}$5, with current

$F_{DC}F_{RF}$6

The mechanism is described as an “unconventional optical Hall effect,” but the paper states that it is not the nonlinear Hall effect itself, and stresses the distinction $F_{DC}F_{RF}$7. In the 2D and bulk geometries derived there, $F_{DC}F_{RF}$8-polarized oblique incidence is required and $F_{DC}F_{RF}$9 (Syljuåsen et al., 24 Sep 2025).

The metallic electro-optics literature on gapped bilayer graphene provides a second comparator. There, the dominant Kerr contribution in clean metals is a non-Drude “Snap” term,

$\chi^{(3)}$0

with $\chi^{(3)}$1 involving a third derivative of the velocity. The paper is explicit that this Kerr mechanism is not geometric and is not based on Berry curvature or the BCD; skew scattering contributes to the Pockels effect, not the Kerr effect, and in gapped bilayer graphene the BCD is symmetry-forbidden by $\chi^{(3)}$2 (Ma et al., 2024).

The classical DC Kerr and silicon-modulation results form a third contrast class. In the rigorous Maxwell model of the DC Kerr effect, the strong stationary field produces birefringence through a cubic Kerr nonlinearity, while the beam propagates along linear characteristics and accumulates a polarization-dependent phase shift. In silicon PIN-junction modulators, the measured refractive-index change under combined DC and RF fields contains a term proportional to $\chi^{(3)}$3, which acts like a Pockels response generated from $\chi^{(3)}$4; in the PIN3 structure, the DC Kerr effect contributes up to $\chi^{(3)}$5 of the total index change, becomes larger than the plasma-dispersion contribution above $\chi^{(3)}$6 V reverse bias, and supports eye diagrams up to $\chi^{(3)}$7 Gbits/s in NRZ format (Eptaminitakis et al., 2 May 2025, Peltier et al., 2023).

6. Measurement, attribution, and interpretive challenges

A central methodological issue is that Kerr-like time dependence is not always a direct record of intrinsic nonlinear dynamics. In anisotropic, dispersive solids, two-color ultrafast optical Kerr effect experiments can be dominated by group-delay differences, phase mismatch, and polarization mixing. The corresponding four-wave-mixing treatment shows that even an instantaneous $\chi^{(3)}$8 can generate oscillations, branch splitting, and delayed temporal structure because the propagating fields acquire frequency-dependent phases and experience different group indices along the fast and slow axes. The detected modulation frequencies are linked to birefringence and group-velocity mismatch rather than necessarily to intrinsic modes, and the paper gives practical extraction formulas for group indices from the reflection times $\chi^{(3)}$9 and $$\nabla(\nabla\cdot\boldsymbol{E})-\nabla^2\boldsymbol{E} =-\mu\frac{\partial(\boldsymbol{J}+\partial_t\boldsymbol{D})}{\partial t}, \qquad \boldsymbol{D}=\varepsilon\boldsymbol{E},$$0 (Huber et al., 2020).

This caution is highly relevant for NLHE-induced EOKE experiments because the sought-after signal is itself a small polarization rotation embedded in an anisotropic optical background. The LiNbO$$\nabla(\nabla\cdot\boldsymbol{E})-\nabla^2\boldsymbol{E} =-\mu\frac{\partial(\boldsymbol{J}+\partial_t\boldsymbol{D})}{\partial t}, \qquad \boldsymbol{D}=\varepsilon\boldsymbol{E},$$1 results add a second cautionary layer: phase mismatch can reshape the Kerr-like spectrum and alter its relative spectral weight without extinguishing it, so geometry-dependent spectral changes are not, by themselves, decisive evidence for or against a Hall-driven origin (Wang et al., 2019). A third warning comes from blue-phase liquid crystals, where a fast Kerr response can be accompanied by a slow component caused by coupling to lattice deformation, and polymer stabilization removes the slow response by inhibiting electrostriction. This shows that structural relaxation can enter nominally Kerr-type electro-optic measurements on timescales vastly longer than the local optical response (Yoshida et al., 2013).

A plausible implication is that convincing identification of NLHE-induced EOKE requires simultaneous control of transport symmetry, optical geometry, and competing bias-induced electro-optic channels. The WTe$$\nabla(\nabla\cdot\boldsymbol{E})-\nabla^2\boldsymbol{E} =-\mu\frac{\partial(\boldsymbol{J}+\partial_t\boldsymbol{D})}{\partial t}, \qquad \boldsymbol{D}=\varepsilon\boldsymbol{E},$$2 proposal points to symmetry-sensitive, field-enhanced, time-resolved Kerr detection as a route to isolating the NLHE contribution (Li et al., 14 Jul 2025). At the same time, neighboring developments in biased electro-optics show that low-symmetry nonequilibrium systems can also enter nonreciprocal and non-Hermitian regimes through free–bound charge interactions, with gain and indefinite anti-Hermitian response under bias (Lannebère et al., 12 Mar 2025). This suggests that the broader electro-optic Kerr landscape is expanding from passive birefringence and conservative rotation toward a taxonomy in which Hall-driven, cascaded, free-carrier, and active bias-linearized mechanisms must be disentangled at the level of tensors, symmetry, and measurement protocol.