---
title: Second-Order Nonlinear Electrical Response
url: https://www.emergentmind.com/topics/second-order-nonlinear-electrical-response-nler
type: topic
---

# Second-Order Nonlinear Electrical Response

Searching arXiv for the cited papers and related NLER work to ground the article in current literature.
Second-order nonlinear electrical response (NLER) is the component of an electrical or optical response that is quadratic in the applied field. At the molecular level it appears through the first hyperpolarizability tensor \(\beta_{ijk}\) in \(\mu_i = \mu_i^{(0)} + \alpha_{ij}E_j + \frac{1}{2}\beta_{ijk}E_jE_k + \cdots\), while in macroscopic media it is encoded in a second-order conductivity or susceptibility tensor, as in \(j_i = \sigma^{(1)}_{ij}E_j + \sigma^{(2)}_{ijk}E_jE_k + \cdots\) or \(P_i^{(2)} = \varepsilon_0 \chi^{(2)}_{ijk}E_jE_k\) [1912.02174]. Although the conventional electric-dipole \(\chi^{(2)}\) vanishes in centrosymmetric media for spatially uniform fields, the broader NLER landscape includes gradient-induced bulk response in centrosymmetric hyperbolic media, extrinsic skew-scattering and side-jump mechanisms in inversion-broken conductors, quantum-geometric contributions governed by Berry curvature dipoles and quantum metric dipoles, and formal many-body constructions based on Matsubara response theory [2605.24184].

## 1. Definition and constitutive structure

Second-order NLER is the part of the response described by a term quadratic in the driving electric field. In nonlinear optics and response theory, the dipole moment of a molecule in an external field is expanded as
\[
\mu_i = \mu_i^{(0)} + \sum_j \alpha_{ij}E_j + \frac{1}{2}\sum_{jk}\beta_{ijk}E_jE_k + \frac{1}{6}\sum_{jkl}\gamma_{ijkl}E_jE_kE_l + \cdots,
\]
with \(\beta_{ijk}\) the first hyperpolarizability tensor and therefore the molecular second-order response [1912.02174]. In condensed-matter transport, the analogous expansion is
\[
j_\alpha = \sigma^{(1)}_{\alpha\beta}E_\beta + \chi^{(2)}_{\alpha\beta\gamma}E_\beta E_\gamma + \dots,
\]
or equivalently \(j_i = \sigma^{(1)}_{ij}E_j + \sigma^{(2)}_{ijk}E_jE_k + \dots\), depending on notation [2201.09274].

This quadratic response underlies second-harmonic generation, sum-frequency generation, difference-frequency generation, the nonlinear Hall effect, bilinear magnetoelectric response, and the Pockels electro-optic effect [1912.02174]. In integrated photonics and nonlinear optics, the relevant macroscopic constitutive law is often written as
\[
P_i^{(2)}(2\omega)=\varepsilon_0\sum_{jk}\chi^{(2)}_{ijk}(2\omega;\omega,\omega)E_j(\omega)E_k(\omega),
\]
while in transport the observable is a second-order conductivity tensor entering \(J_\alpha(\omega_1+\omega_2)=\sigma^{(2)}_{\alpha\beta\gamma}(\omega_1,\omega_2)E_\beta(\omega_1)E_\gamma(\omega_2)\) [2605.24184].

A recurring symmetry statement is that inversion symmetry forbids the ordinary electric-dipole second-order response for spatially uniform fields. In a centrosymmetric molecule at zero field, \(\beta_{ijk}=0\); in a centrosymmetric bulk medium, \(\chi^{(2)}_{ijk}\) vanishes in the dipole approximation [1912.02174]. The data block, however, also documents several mechanisms by which a measurable second-order response survives or is engineered despite nominal centrosymmetry: externally induced symmetry breaking in molecules, electrical poling in amorphous dielectrics, and gradient-induced response in hyperbolic media [2210.09374].

## 2. Symmetry, inversion breaking, and routes to finite response

In conventional nonlinear optics, inversion symmetry eliminates the uniform-field electric-dipole term because under \(\mathbf{r}\to-\mathbf{r}\), both \(\mathbf{E}\) and \(\mathbf{P}\) change sign, so a term proportional to \(E_jE_k\) is incompatible with centrosymmetry. This is the standard reason why a bulk \(\chi^{(2)}\) is absent in centrosymmetric crystals and why inversion breaking is a prerequisite for ordinary second-order response [2605.24184].

One route to finite NLER is direct symmetry breaking by an external field. Hexalithiobenzene, \( \mathrm{C}_6\mathrm{Li}_6 \), is planar and highly symmetric at zero field, with \(\mu_Z = 0.002\) D and \(\beta_0 = 0.5\) a.u., so its second-order response is effectively negligible. Under an oriented external electric field applied along the molecular \(z\)-axis, the symmetry is broken, a finite dipole develops, and the first mean hyperpolarizability rises to \(33870.4\) a.u. at \(50\times10^{-4}\) a.u. [1912.02174]. This is a field-induced activation of second-order response through controlled symmetry lowering.

A second route is structural reconfiguration. Stoichiometric \(\mathrm{Si}_3\mathrm{N}_4\) is amorphous and statistically centrosymmetric, so its bulk \(\chi^{(2)}\) vanishes. Electrical poling at high temperature, with an applied DC field of approximately \(0.676\ \mathrm{MV/cm}\), aligns Si–N bonds or local polar units and freezes in a noncentrosymmetric configuration. The resulting poled material exhibits a bulk \(\chi^{(2)}\), inferred from electro-optic measurements to be approximately \(0.24\ \mathrm{pm/V}\) [2210.09374]. In that case the symmetry class is modeled as analogous to \(C_{\infty v}\), and the induced electro-optic response is consistent with tensor components \(r_{33}\) and \(r_{13}\).

A third route does not require breaking structural centrosymmetry at all. In centrosymmetric hyperbolic media, the gradient term \(\nabla\mathbf{E}\,\mathbf{E}\) is inversion-even, because both \(\nabla\) and \(\mathbf{E}\) pick up a minus sign under inversion. Consequently, a second-order polarization proportional to \((\nabla\mathbf{E})\mathbf{E}\) is symmetry-allowed even in a centrosymmetric bulk medium [2605.24184]. This mechanism is central to recent work on hyperbolic media and is distinct from ordinary bulk \(\chi^{(2)}\) in noncentrosymmetric crystals.

## 3. Microscopic mechanisms and quantum-geometric origins

A central distinction in the literature is between intrinsic and extrinsic mechanisms. Intrinsic mechanisms are tied to equilibrium band geometry or linear response tensors. Extrinsic mechanisms arise from disorder-mediated dynamics such as skew scattering and side jump.

In inversion-broken but time-reversal-symmetric conductors, the intrinsic nonlinear Hall effect is closely related to the Berry curvature dipole. The semiclassical framework summarized in the data block defines a Berry curvature dipole tensor
\[
D_{db} = \sum_n \int_{\mathrm{BZ}} \frac{d^dk}{(2\pi)^d}\,\frac{\partial f_{n\mathbf{k}}}{\partial k_d}\,\Omega_b^n(\mathbf{k}),
\]
and the second-harmonic Hall current at lowest order in the field is controlled by this dipole [2110.06582]. In the Matsubara framework, the corresponding second-order Hall conductivity appears through derivatives of an interband Berry curvature object \(F^{ab}_{\mu\nu}\), yielding \(\sigma^{(1,0)}\), \(\sigma^{(0,1)}\), and the first-order SHG conductivity \(\sigma^{(1)}\) in terms of Berry-curvature-dipole structures [2507.08035].

The quantum metric provides a distinct intrinsic channel. In the Matsubara theory of nonlinear Ohmic electromagnetic response, the order-\(\tau^0\) Ohmic SHG conductivity is
\[
\sigma^{(O,0)}_{\mu\nu\gamma}
= \frac{2 e^3}{V}\sum_{\mathbf{k},a\neq b} f_a(\mathbf{k})
\left[
\partial_\nu\!\left(\frac{g_{\mu\gamma}}{\varepsilon_{ab}}\right)
+\partial_\gamma\!\left(\frac{g_{\mu\nu}}{\varepsilon_{ab}}\right)
+\partial_\mu\!\left(\frac{g_{\nu\gamma}}{\varepsilon_{ab}}\right)
\right],
\]
so the intrinsic nonlinear Ohmic response is governed by the fully symmetrized normalized quantum metric dipole [2605.16367]. In a related non-Hermitian framework, the narrow-wavepacket second-order DC conductivity contains an intrinsic, scattering-time-independent term
\[
\sigma^{\mathrm{int}}_{\theta\mu\nu}
= -e^3\int_{\mathbf{k}} f_0\,
\mathrm{Re}\Big[
2\partial_\theta G^{LR(0)}_{\mu\nu}
-\frac{\partial_\nu G^{LR(0)}_{\mu\theta}+\partial_\mu G^{LR(0)}_{\nu\theta}}{2}
\Big],
\]
so the real part of the non-Hermitian quantum metric controls an intrinsic nonlinear conductivity even in open systems with a spectral line gap [2509.11765].

Extrinsic mechanisms dominate in several moiré materials. In twisted bilayer graphene aligned with hBN, the \(C_3\) point-group symmetry forbids the intrinsic Berry-curvature-dipole contribution, so the observed giant nonlinear Hall effect is extrinsic. The measured second-order Hall conductivity reaches \(8.76~\mu\mathrm{m\,S\,V}^{-1}\), and the effect is attributed to skew scattering from static Coulomb impurities at low temperature together with phonon skew scattering at elevated temperature [2201.09274]. Twisted double bilayer graphene shows an even larger extrinsic second-order conductivity, \(|\sigma^{(2)}_{xxx}|\sim 70\ \mu\mathrm{m\,V}^{-1}\Omega^{-1}\), near a mid-band van Hove singularity, again tied to extrinsic side-jump and skew-scattering channels [2507.05969].

Strongly correlated systems add another microscopic layer. In a noncentrosymmetric Kondo lattice with Rashba-type spin-orbit coupling, ferromagnetism is required for a finite second-order conductivity, and the response becomes finite only perpendicular to the ferromagnetic magnetization. In the low-density approximation, the second-order conductivity matrix scales as
\[
\sigma^{(2)} \simeq
\frac{6 m \mu}{\pi}(\tau e)^3 J_1J_2
\begin{pmatrix}
M_y & M_y\\
-M_x & -M_x
\end{pmatrix},
\]
so for \(\mathbf{M}\parallel x\) only the \(y\)-direction nonlinear response survives [2110.10496].

## 4. Centrosymmetric hyperbolic media and gradient-induced bulk response

The recent hyperbolic-media mechanism is unusual because it produces a bulk second-order response in a centrosymmetric material without invoking crystal-potential anharmonicity. The starting point is a more complete expansion of the polarization,
\[
\mathbf{P}^{(2)}(2\omega)
=
\hat{\mathbf{e}}_\alpha\,
\Lambda_{\alpha\beta\gamma\nu}
\frac{\partial E_\gamma}{\partial x_\beta}E_\nu,
\]
which is explicitly gradient-dependent [2605.24184]. The tensor \(\Lambda_{\alpha\beta\gamma\nu}\) is determined entirely by linear susceptibilities through
\[
\Lambda_{\alpha\beta\gamma\nu}
=
\frac{\chi_{\alpha\gamma}(2\omega)\chi_{\beta\nu}(\omega)}{Ne},
\]
with \(N\) the atomic number density and \(e\) the electron charge [2605.24184]. The paper explicitly states that this mechanism does not rely on any anharmonicity of the crystalline potential and is entirely governed by the linear response of the medium.

Hyperbolic media enable this response because their dielectric tensor has opposite signs along different principal directions, for example
\[
\epsilon_x=\epsilon_y=\epsilon_\perp,\qquad \epsilon_z,
\]
with \(\epsilon_\perp\) and \(\epsilon_z\) of opposite sign. The resulting TM isofrequency surface,
\[
\frac{k_z^2}{\epsilon_\perp}+\frac{k_\perp^2}{\epsilon_z}=\frac{\omega^2}{c^2},
\]
is a hyperboloid, and large-\(k\) modes are propagating rather than evanescent [2605.24184]. Consequently, the internal field can vary on deeply subwavelength scales, so \(|\nabla\mathbf{E}|\) is no longer limited by \(|\mathbf{E}|/\lambda_0\). This activates the otherwise negligible \(E\,\nabla E\) nonlinearity.

For a structured hyperbolic slab, the far field sees an effective bulk second-order susceptibility defined by
\[
P^{(2)}=\chi^{(2)}_{\mathrm{eff}}E_0^2.
\]
For the dominant \(zz\) component, the explicit result is
\[
\begin{aligned}
\bigl(\chi^{(2)}_{\mathrm{eff}}\bigr)_{zz}
&=
\frac{\epsilon_\perp(\omega)}{\epsilon_z(\omega)}
\frac{\chi_{zz}(\omega)\chi_{zz}(2\omega)\alpha_{\rm eff}^2(\omega)}
{\left(1+i\sqrt{-\epsilon_z(\omega)\epsilon_\perp(\omega)}/\epsilon_0\right)^2}
\frac{\pi a}{eNbd}\\
&=\mathcal{O}\!\left(\frac{a}{eNbd}\right),
\end{aligned}
\]
with defect size \(a\), spacing \(b\), and slab thickness \(d\) [2605.24184]. Using \(a\sim10\ \mathrm{nm}\), \(b\sim5a\), \(d\sim10a\), the estimated magnitude is
\[
\chi^{(2)}_{\mathrm{eff}}\sim 10^{-9}\ \mathrm{esu},
\]
which is of the same order as ADP and KDP values quoted in the paper [2605.24184]. This supports the claim that centrosymmetric hyperbolic media such as hBN can exhibit bulk second-harmonic, sum-frequency, and difference-frequency generation efficiencies comparable to established nonlinear crystals.

A plausible implication is that this mechanism broadens the meaning of “bulk” second-order response. The effective \(\chi^{(2)}_{\mathrm{eff}}\) is not a bare crystal parameter but an emergent property of the hyperbolic medium, subwavelength couplers, and the resulting field distribution [2605.24184].

## 5. Materials platforms and representative magnitudes

The data block documents several experimentally or theoretically important material classes in which second-order NLER is large, tunable, or unusually structured.

| Platform | Mechanism emphasized | Representative magnitude |
|---|---|---|
| hBN hyperbolic medium | Gradient-induced effective bulk \(\chi^{(2)}\) in centrosymmetric medium | \(\chi^{(2)}_{\mathrm{eff}}\sim10^{-9}\ \mathrm{esu}\) [2605.24184] |
| Poled \(\mathrm{Si}_3\mathrm{N}_4\) | Electrical poling and bond alignment | \(\chi^{(2)}\approx0.24\ \mathrm{pm/V}\) [2210.09374] |
| \( \mathrm{C}_6\mathrm{Li}_6 \) under OEEF | Field-induced symmetry breaking and gap reduction | \(\beta_0=33870.4\) a.u. at \(50\times10^{-4}\) a.u. [1912.02174] |
| TBG/hBN | Extrinsic skew scattering | \(\chi_{yxx}=8.76~\mu\mathrm{m\,S\,V}^{-1}\) [2201.09274] |
| tDBLG | Extrinsic NLER near vHSs | \(|\sigma^{(2)}_{xxx}|\sim70\ \mu\mathrm{m\,V}^{-1}\Omega^{-1}\) [2507.05969] |
| BLG | Lifshitz-transition-sensitive NLER | \(\sigma^{(2)}\) exceeds \(30~\mu\mathrm{m\,V}^{-1}\Omega^{-1}\) at 3 K [2507.05871] |

In poled \(\mathrm{Si}_3\mathrm{N}_4\), the induced \(\chi^{(2)}\) manifests as a high-speed Pockels response with \(r_{33}\) up to approximately \(30\ \mathrm{fm/V}\) and a 3 dB electro-optic bandwidth of at least \(15\ \mathrm{GHz}\), in contrast to the \(3\ \mathrm{GHz}\) carrier-dominated response of non-poled devices [2210.09374]. In \( \mathrm{C}_6\mathrm{Li}_6 \), the same external field that enhances \(\beta_0\) also reduces the HOMO–LUMO gap from \(1.894\) eV to \(0.859\) eV, consistent with the paper’s two-level argument \(\beta_0\propto \Delta\mu\,f_0/(\Delta E)^3\) [1912.02174].

In bilayer graphene, NLER becomes a probe of Fermi-surface topology. The second-order conductivity changes sign near Lifshitz transitions and remains diagnostic even at \(T\gtrsim 10\) K [2507.05871]. In twisted double bilayer graphene, a “cascade of singularities” appears as sharp sign reversals and extrema in \(\sigma^{(2)}\) near multiple van Hove singularities across several moiré bands [2507.05969]. These observations suggest that second-order conductivity is exceptionally sensitive to Fermi-surface reconstructions.

## 6. Many-body, Matsubara, and coarse-grained response theory

A general many-body formulation of second-order NLER requires a causal three-point response function. The imaginary-time formalism paper proves that fully causal \(n\)-th order response functions can be obtained by analytic continuation of Matsubara \(n\)-point functions, with the main theorem
\[
\chi^{(n)}_{AB}(\{\omega_i^+\}) = (-1)^n \bar{\chi}^{(n)}_{AB}(\{i\nu_i=\omega_i^+\}),
\]
implemented at the level of the Lehmann representation [2506.21428]. For \(n=2\), this means the physical causal second-order conductivity or susceptibility can be computed from an imaginary-time three-point correlator and analytically continued to \((\omega_1+i0^+,\omega_2+i0^+)\).

The same work provides the time-domain Kubo structure through nested commutators,
\[
\tilde{\chi}^{(2)}_{AB}(t-t_1,t-t_2)
=
-\,\Theta(t-t_1)\Theta(t-t_2)
\langle [[A(t),B(t_1)],B(t_2)]\rangle_0\,e^{\epsilon(t_1+t_2-t)},
\]
and derives a high-frequency sum rule for \(n\)-th harmonic generation,
\[
\lim_{\omega\to\infty}\omega^n\chi^{(n)}_{AB}(\omega)
=
\frac{1}{n!}\langle [\cdots[[A,B],B],\dots,B]\rangle_0,
\]
which for \(n=2\) constrains the asymptotic second-harmonic response [2506.21428]. This provides a rigorous route from Matsubara diagrammatics to measurable second-order NLER in interacting and disordered systems.

At the level of metamaterial homogenization, second-order NLER is richer than a purely electric \(\chi^{(2)}\). In a bi-anisotropic medium one may define
\[
\chi^{(2)}_{abc,\alpha\beta\gamma},
\qquad a,b,c\in\{\mathrm e,\mathrm m\},
\]
to include electric, magnetic, and magneto-electric second-order couplings. The total number of effective second-order susceptibilities is \(2^3\times 3^3=216\), and a nonlinear transfer-matrix retrieval scheme can recover them from a sufficient set of sum-frequency-generation measurements or simulations [1712.04962]. This is a formal generalization of NLER from conventional dielectric polarization to full effective electromagnetic response.

A different formal generalization arises in coarse-grained response theory. For equilibrium stochastic systems, the second-order response of an observable can be written as
\[
O^{(2)}=-\langle S'_h D'_h\,O\rangle_{\rm eq},
\]
where \(S'_h\) is the time-antisymmetric entropy-production part and \(D'_h\) is the time-symmetric dynamical-activity part [2005.05169]. The related path-space formulation emphasizes that linear response depends only on entropy production, while second-order response explicitly probes frenetic, dynamical details of the system [1410.7450]. This suggests that in coarse-grained electrical systems, second-order response can reveal kinetic structure invisible in linear transport.

## 7. Experimental signatures, selection rules, and open directions

A standard experimental signature of second-order NLER under AC drive \(E(t)\propto \sin\omega t\) is a voltage or current at \(2\omega\). In twisted bilayer graphene, the transverse voltage \(V_y^{2\omega}\) scales linearly with \(V_x^2\) and does not change sign when current direction is reversed, consistent with a second-order signal [2201.09274]. In bilayer graphene and tDBLG, the second-harmonic longitudinal and transverse voltages are extracted to obtain \(\sigma^{(2)}_{xxx}\) and \(\sigma^{(2)}_{yxx}\), with sign reversals near Lifshitz transitions or van Hove singularities [2507.05871].

Selection rules depend sharply on time-reversal and inversion symmetry. In the semiclassical harmonic-order framework, with time-reversal symmetry preserved and inversion symmetry broken, even harmonics of the charge current are purely transverse while odd harmonics are longitudinal [2110.06582]. With both time-reversal and inversion preserved, even harmonics vanish. With both broken, all harmonic orders of charge and spin currents can in principle appear [2110.06582]. These results organize when second-order NLER is allowed and whether it should be Hall-like or longitudinal.

The recent literature also separates Hall-like and Ohmic second-order responses. In the Matsubara theory of nonlinear Ohmic electromagnetic response, the order-\(\tau\) Ohmic conductivity vanishes for both SHG and the bilinear magnetoelectric effect, while the intrinsic order-\(\tau^0\) contribution survives and is governed by the normalized quantum metric dipole [2605.16367]. In the broader quantum theory of nonlinear electromagnetic response, Berry-curvature-dipole terms dominate time-reversal-symmetric nonlinear Hall effects, while quantum-metric-dipole terms govern intrinsic nonlinear Hall effects in time-reversal-breaking systems [2507.08035].

Several open directions are explicit in the source material. The hyperbolic-media work notes that once an effective \(\chi^{(2)}_{\mathrm{eff}}\) is established, conventional phase matching can be used to enhance output [2605.24184]. The poled \(\mathrm{Si}_3\mathrm{N}_4\) work points to higher poling fields and lower-loss geometries as routes toward larger \(\chi^{(2)}\) [2210.09374]. The moiré-material studies suggest that NLER is a reliable tool for locating Lifshitz transitions and Fermi-surface reconstructions under time-reversal-symmetric conditions [2507.05871]. A plausible implication is that second-order conductivity will continue to function both as a spectroscopy of electronic quantum geometry and as a device figure of merit for rectification, electro-optic modulation, and frequency conversion.

Source: https://www.emergentmind.com/topics/second-order-nonlinear-electrical-response-nler