Electronic Raman Scattering: Principles & Applications

• Electronic Raman scattering (eRS) is an inelastic light scattering process that specifically probes electronic excitations via symmetry-resolved selection rules, distinguishing it from phonon-based Raman methods.
• The technique leverages polarization, gating, and resonance conditions to isolate interband transitions and reveal key phenomena such as superconducting pair breaking, nematic fluctuations, and collective modes.
• Practically, eRS serves as a non-destructive metrology tool for analyzing band topology, stacking orders, and electron–hole dynamics in materials like graphene, carbon nanotubes, and superconductors.

Searching arXiv for recent and foundational papers on electronic Raman scattering to ground the article in published work. arXiv search query: "electronic Raman scattering graphene nanotubes superconductors nematic" Electronic Raman scattering (eRS) is the inelastic scattering of light by electronic excitations rather than by lattice vibrations. In eRS, the Raman shift can measure the energy of an interband electron–hole excitation, a superconducting pair-breaking scale, a nematic fluctuation, or a collective mode, while the small momentum transfer of visible photons generally restricts the process to essentially vertical, near-$\mathbf q=0$ excitations. In contrast to phonon Raman spectroscopy, which tracks vibrational degrees of freedom, eRS is a symmetry-resolved two-particle probe of the electronic response, and its observable spectrum is controlled by Raman vertices, polarization geometry, resonance conditions, and many-body susceptibilities (Riccardi et al., 2016, Tanaka et al., 2019, Gallais et al., 2015).

1. Definition, scope, and distinction from vibrational Raman scattering

In the systems discussed across the literature, eRS consistently denotes inelastic light scattering from electronic degrees of freedom. In monolayer graphene, the relevant excitations are interband electron–hole pairs across the Dirac cone; in single-walled carbon nanotubes they are continua of electron–hole pairs across quantized subbands; in superconductors they appear as pair-breaking continua and symmetry-selected low-energy responses; and in correlated metals they can represent charge, nematic, spin, or collective excitations (Riccardi et al., 2016, Hu et al., 2020, Tanaka et al., 2019).

The distinction from vibrational Raman scattering is operational as well as conceptual. In graphene, the dominant ordinary Raman features are the G band near $1580~\mathrm{cm}^{-1}$, the 2D band, and substrate-related phonon peaks from Si below about $1100~\mathrm{cm}^{-1}$, whereas the electronic contribution is a weak continuum whose Raman shift corresponds to the energy $\hbar\omega$ needed to create an electron–hole pair. In nanotubes, the electronic feature is broad and tracks an electronic transition energy rather than a phonon frequency. In iron-based superconductors, the measured quantity is a symmetry-resolved electronic susceptibility rather than a phonon line, with the experimentally accessible intensity given by

$$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$

These examples establish that eRS is not a generic “continuum background” but a spectroscopic channel with its own selection rules and response functions (Riccardi et al., 2016, Farhat et al., 2011, Gallais et al., 2015).

A recurrent interpretive issue is the differentiation of eRS from fluorescence, substrate luminescence, and weak phonon combinations. The literature resolves this by combining polarization analysis, resonance conditions, gate dependence, and symmetry arguments. In graphene, the gate-dependent continuum appears only in crossed polarization; in metallic nanotubes it is observed on both Stokes and anti-Stokes sides and is suppressed by electrostatic gating; in semiconducting nanotubes it appears only when the laser satisfies a specific threshold condition. These behaviors are inconsistent with simple luminescence and support the assignment to electronic Raman processes (Riccardi et al., 2016, Farhat et al., 2011, Hu et al., 2020).

2. Microscopic formulation and symmetry structure

At the microscopic level, eRS is a two-photon process generated by coupling electrons to the electromagnetic field through minimal coupling or the Peierls substitution. The light–matter Hamiltonian therefore contains terms linear and quadratic in the vector potential, and the scattering amplitude generally decomposes into non-resonant, resonant, and mixed contributions. A unified treatment of these channels was formulated in a Schwinger–Keldysh canonical perturbation framework, in which all two-photon scattering processes from electrons, the non-resonant charge density response, elastic Rayleigh scattering, fluorescence, intrinsic energy-shift Raman scattering, and the mixed response are represented as contour-ordered correlators (Su, 2015).

In the non-resonant regime, the Raman intensity is commonly expressed in terms of the imaginary part of an electronic susceptibility. For graphene in the long-wavelength limit,

$$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$

and for a $k$-independent Raman vertex $\gamma_0$,

$$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$

At $T=0$,

$1580~\mathrm{cm}^{-1}$0

so charge neutrality yields a continuum linear in $1580~\mathrm{cm}^{-1}$1, while doping imposes a Pauli-blocking threshold at $1580~\mathrm{cm}^{-1}$2. This connects eRS directly to the imaginary part of the dynamical polarizability $1580~\mathrm{cm}^{-1}$3 (Riccardi et al., 2016).

Selection rules are central. In monolayer graphene, crossed polarizations probe $1580~\mathrm{cm}^{-1}$4 and $1580~\mathrm{cm}^{-1}$5, while parallel polarizations probe $1580~\mathrm{cm}^{-1}$6 and $1580~\mathrm{cm}^{-1}$7; the gate-dependent electronic continuum appears only in crossed polarization, consistent with the dominant non-resonant interband Raman process of $1580~\mathrm{cm}^{-1}$8 symmetry. In Bernal bilayer graphene, the two-step Raman amplitude dominates over the one-step process and is proportional to $1580~\mathrm{cm}^{-1}$9, again selecting the crossed-polarized channel. The bilayer Raman operator is effectively proportional to the $1100~\mathrm{cm}^{-1}$0-component of isospin, $1100~\mathrm{cm}^{-1}$1, which filters the response to chiral interband excitations and suppresses non-chiral channels through destructive interference in the two virtual Raman pathways (Riccardi et al., 2018).

Near resonance, the Raman amplitude need not be a simple incoherent sum of pathways. In a quantum model of surface-enhanced Raman scattering, off-resonant electronic levels can be adiabatically eliminated to yield the standard optomechanical Hamiltonian with

$1100~\mathrm{cm}^{-1}$2

whereas a near-resonant zero-phonon-line transition must be retained explicitly. The resulting resonant and off-resonant Raman amplitudes interfere coherently, enhancing or suppressing Stokes and anti-Stokes peaks by several orders of magnitude and modifying photon-correlation observables (Martínez-García et al., 2023).

A further generalization arises in inversion-broken metals with spin-orbit coupling. There the non-resonant Raman tensor acquires spin structure because the velocity operators and low-energy Hilbert space are spin-textured. In that setting, the large-$1100~\mathrm{cm}^{-1}$3 expansion of the Raman operator contains commutators of velocity operators that survive in multiband spin-orbit-coupled systems, so eRS can couple directly to spin excitations even without tuning the laser to an internal resonance (Saleh et al., 1 Apr 2026).

3. Dirac materials, nanotubes, and graphitic multilayers

Carbon-based systems provide the clearest demonstrations of eRS as a probe of band topology, chirality, gating, and moiré or stacking structure. In monolayer graphene, the central result is the direct observation of a gate-tunable, crossed-polarization continuum from interband electron–hole excitations across the Dirac cone. The experimentally defined ratio

$1100~\mathrm{cm}^{-1}$4

suppresses gate-independent background, and its threshold-like suppression with increasing $1100~\mathrm{cm}^{-1}$5 follows Pauli blocking below $1100~\mathrm{cm}^{-1}$6. The same electronic polarizability that controls this continuum also governs the G-band linewidth through $1100~\mathrm{cm}^{-1}$7, giving the relation $1100~\mathrm{cm}^{-1}$8 (Riccardi et al., 2016).

In Bernal bilayer graphene, eRS is more selective than optical absorption. The continuum is dominated by $1100~\mathrm{cm}^{-1}$9-symmetry interband chiral excitations, specifically transitions between bands symmetric with respect to $\hbar\omega$0, while non-chiral interband channels make a vanishing contribution because of destructive interference in the Raman amplitude. As in monolayer graphene, the continuum is progressively suppressed by gate-induced Pauli blocking (Riccardi et al., 2018).

In metallic single-walled carbon nanotubes, the observed eRS feature is a broad peak produced by low-energy electron–hole pairs across graphenelike linear subbands and resonantly enhanced when the scattered photon energy matches an optical transition $\hbar\omega$1. A distinctive property is that the Raman shift is highly dispersive with laser energy while the scattered photon energy remains pinned near the relevant $\hbar\omega$2. The signal is absent in semiconducting tubes under comparable conditions and is strongly reduced by electrostatic gating, consistent with Pauli blocking (Farhat et al., 2011).

Semiconducting nanotubes extend this picture. eRS in suspended semiconducting SWNTs is enabled not by low-energy metallic-like carriers but by high-energy electron–hole pairs above $\hbar\omega$3. The condition for observing eRS at a given $\hbar\omega$4 is

$\hbar\omega$5

When this is satisfied, the ERS band remains centered at the scattered-photon energy corresponding to $\hbar\omega$6, allowing determination of transition energies with $\hbar\omega$7 accuracy. This overturned the prior assumption that ERS is exclusive to metallic nanotubes (Hu et al., 2020).

Twisted and rhombohedrally stacked graphitic multilayers introduce miniband and thickness fingerprints into eRS.

System	Dominant electronic excitation	Diagnostic eRS signature
Monolayer graphene	Interband electron–hole pairs across the Dirac cone	Crossed-polarized continuum; Pauli blocking below $\hbar\omega$8
Bilayer graphene	$\hbar\omega$9-symmetry chiral interband excitations	Non-chiral channels suppressed by destructive interference
Metallic SWNTs	Low-energy electron–hole pairs across linear subbands	Broad peak fixed in scattered photon energy near $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$0
Semiconducting SWNTs	High-energy electron–hole pairs above $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$1	Threshold $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$2; peak centered at $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$3
Twisted few-layer graphene	van Hove singularities in moiré minibands	Two twist-dependent peaks below about $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$4
Rhombohedral graphite	Transitions between layer-dependent DOS edges	Thickness- and stacking-specific ERS peaks

For twistronic few-layer graphene, the eRS spectrum of two-, three-, and four-layer systems with one twisted interface is predicted to exhibit two prominent peaks originating from van Hove singularities in moiré minibands: a lower-energy peak from direct hybridization of Dirac states across the twisted interface and a higher-energy peak from moiré-induced band folding. Both peaks move strongly with twist angle and can therefore serve as non-invasive twist-angle metrology, including in hBN-encapsulated structures (Garcia-Ruiz et al., 2020).

For rhombohedral graphite, eRS is structured rather than featureless because ABC stacking creates topological low-energy surface bands and pronounced subband-edge singularities. In thin films, the peak sequence follows

$$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$5

whereas Bernal stacking gives an essentially featureless electronic Raman background (Fuentes et al., 2019). This layer dependence has been turned into a practical identification protocol: for thicknesses from 3 to 12 layers, each perfect rhombohedral structure has distinctive ERS peak positions, measurable at room temperature with visible excitation on a conventional confocal Raman spectrometer, and deviations in peak position or missing peaks reveal stacking faults (Pálinkás et al., 2024).

4. Superconductivity, nematicity, and momentum-selective electronic response

In superconductors and correlated metals, eRS is a symmetry-resolved probe of low-energy electronic structure rather than a simple continuum spectroscopy. In cuprates, the standard channels are $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$6, primarily sensitive to antinodal regions, and $$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$7, primarily sensitive to the nodal region. A direct comparison between ERS and ARPES in Bi$$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$8Sr$$S_{\mu}(\omega)=\frac{1}{\pi}\left[1+n_B(\omega)\right]\chi_{\mu}^{\prime\prime}(\mathbf{q}=0,\omega).$$9CaCu$$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$0O$$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$1 computed Raman spectra from experimental ARPES data via a Kubo formula and found good overall agreement except for the $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$2 peak energies. The $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$3 peak evolution could be reproduced from a doping-independent $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$4-wave nodal gap once momentum-dependent spectral-weight redistribution along the Fermi surface was included, whereas the $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$5 mismatch with antinodal ARPES increased with underdoping, implying that the pseudogap affects ERS and ARPES differently (Tanaka et al., 2019).

In iron-based superconductors, eRS has been used as a direct probe of charge nematic fluctuations. In a minimal model of orbital nematicity, orbital fluctuations produce strong quasi-elastic light scattering in the $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$6 channel around the nematic critical temperature $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$7, in $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$8 only below $$I_{ERS}(\omega,q)\propto (1+n(\omega,T))\,\chi''(\omega,q),$$9, and not in $k$0. The low-energy central mode follows a Lorentzian form whose width scales with distance from the transition, while at higher energies a nematic Lifshitz transition shifts $k$1 spectral weight upward (Yamase et al., 2013). A complementary review framed the same physics in terms of a $k$2-wave Pomeranchuk transition and emphasized that the $k$3 quasi-elastic peak and Curie–Weiss-like nematic susceptibility are consistent with an electronically driven structural transition. Because Raman probes the dynamical $k$4 response, it is essentially unaffected by acoustic phonons in the relevant limit and thus isolates the bare electronic nematic susceptibility (Gallais et al., 2015).

Disorder complicates the interpretation of superconducting eRS but also sharpens it. In iron pnictides, a self-consistent $k$5-matrix treatment showed that intraband disorder in an extended $k$6 state can lift accidental nodes and open a small minimum gap, whereas isotropic scattering creates impurity-band states and low-energy spectral weight. This gives a concrete spectroscopic distinction between accidental nodes and symmetry-protected $k$7-wave nodes, since node lifting is not possible in $k$8-wave (Boyd et al., 2010).

For unconventional superconductors with nodes, the Raman vertex itself controls the low-energy power law. In a two-orbital bilayer model for La$k$9Ni$\gamma_0$0O$\gamma_0$1, nodal $\gamma_0$2 and $\gamma_0$3 states show robust low-energy power-law responses distinct from fully gapped states: for $\gamma_0$4, the $\gamma_0$5 channel behaves as $\gamma_0$6 while the other channels are linear, and for $\gamma_0$7 the analogous cubic law appears in $\gamma_0$8. The same study also showed that multiorbital effects reshape peak intensities and pocket assignments in ways not captured by a purely band-additive approximation (Zhan et al., 1 Apr 2026).

A separate field-angle formulation extends superconducting eRS into the vortex state. For a 2D $\gamma_0$9 superconductor treated by the quasi-classical Doppler-shift method, the $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$0 Raman intensity depends on both magnetic-field angle and Raman shift. The notable prediction is a reversal of angular extrema with frequency: at low normalized frequency $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$1, maxima occur for field directions along the nodal angles, whereas at higher $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$2 the maxima move to antinodal directions. This suggests a spectroscopic route to gap-structure determination that is both symmetry- and energy-resolved (Okada et al., 2013).

5. Collective modes, spin-orbit coupling, and inversion breaking

A conventional expectation is that fully symmetric Raman scattering at small momentum transfer primarily probes charge excitations, and that plasmons appear only with spectral weight proportional to $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$3. Several recent works modify this expectation by showing that spin-orbit coupling and inversion breaking can qualitatively change the Raman vertex.

In inversion-broken 2D metals with Rashba, Dresselhaus, or valley-Zeeman spin-orbit coupling, broken SU(2) symmetry spin-splits the low-energy bands and also modifies the photon–matter coupling. The resulting non-resonant eRS response is sensitive to spin-flip excitations and coherent chiral-spin modes even without tuning to an internal resonance. The predicted spectra depend strongly on polarization channel and Hilbert-space structure: graphene with Rashba SOC shows a sharp resonance at $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$4 and thresholds at $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$5, whereas valley-Zeeman-dominated systems lack the sharp spin-flip resonance and instead show shifted spin-preserving continua. The same theory predicts that graphene-like Dirac systems yield much stronger signals than a Rashba 2DEG because of the large Dirac velocity and off-shell intermediate-state contributions (Saleh et al., 1 Apr 2026).

Near resonance, spin-orbit effects can even make plasmons Raman active at effectively zero momentum transfer. In a theory developed for giant Rashba systems, the conventional $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$6-weighted plasmon Raman response in the fully symmetric channel is supplemented by a spin-mediated term proportional to spin-orbit coupling. The mechanism requires both a low-energy Rashba-induced spin–charge susceptibility and a high-energy modification of interband transition matrix elements, so that the resonant Raman vertex acquires an effective spin component. In BiTeI, this framework was supported experimentally by identification of a prominent $$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$7-axis plasmon peak in the fully symmetric channel of the resonant Raman spectrum (Sarkar et al., 2023).

Collective-mode sensitivity also appears in excitonic insulators. For the bilayer iridate Sr$$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$8Ir$$\chi''(\omega)=\gamma_0^2\,\omega \Big[f\!\left(-\frac{\hbar\omega}{2}-E_F\right)-f\!\left(\frac{\hbar\omega}{2}-E_F\right)\Big].$$9O$T=0$0 under pressure, an effective mean-field Raman operator constructed from a Hubbard model yields a low-energy electronic Raman cross section whose lowest peak is a longitudinal mode of an antiferromagnetic excitonic insulator. This mode is active in $T=0$1, inactive in $T=0$2, and softens near the pressure-driven transition, providing a direct proposed Raman fingerprint of triplet exciton condensation (Suwa et al., 2024).

Disorder-driven and structurally heterogeneous systems push the concept further but also more cautiously. In silicon glass, the observed low-energy and high-energy broad emission bands were interpreted as eRS enabled by disorder, an excess density of states in the forbidden gap described as an “Urbach bridge,” and electron–photon momentum matching from confined visible photons. The same work explicitly notes that the interpretation is phenomenological and does not provide a full first-principles cross section, so its significance lies less in establishing a settled mechanism than in extending the possible domain of eRS to nanoscale-disordered semiconductors (Kharintsev et al., 2023).

6. Metrology, microscopy, and broader methodological significance

One of the most consequential developments is the use of eRS as a non-destructive metrology tool. In twistronic few-layer graphene, the two van-Hove-derived Raman peaks move strongly with twist angle, enabling optical twist-angle calibration without direct access to the moiré interface (Garcia-Ruiz et al., 2020). In rhombohedral graphite, thickness- and stacking-dependent ERS peaks provide a practical route to identifying flawless ABC stacking and detecting stacking faults when conventional 2D-peak analysis fails for thicker samples (Pálinkás et al., 2024). In monolayer graphene, visible-light Raman microscopy with submicron spatial resolution makes it possible to probe local electronic properties that are difficult to access by infrared methods (Riccardi et al., 2016).

A related extension appears in electronic pre-resonance stimulated Raman scattering microscopy. There, the stimulated Raman signal of a dye is strongly enhanced by tuning the incident laser frequency near an electronic excitation while the Raman line remains narrow because the signal is still tied to a specific vibrational mode. A displaced harmonic oscillator model was shown to reproduce experimental intensities of triple-bond-bearing probes and to rationalize scaffold-dependent signal strengths through the coupling between the electronic excitation and the targeted vibrational mode. This places epr-SRS within the broader electronic Raman framework as a stimulated, pre-resonant realization of electronically enhanced Raman scattering (Du et al., 2023).

Surface-enhanced Raman scattering provides a further methodological bridge. An open-system quantum treatment demonstrated that Raman spectra in SERS can contain coherent interference between resonant and non-resonant electronic pathways, with strong consequences for Stokes and anti-Stokes amplitudes, fluorescence backgrounds, and frequency-filtered photon correlations. A plausible implication is that future eRS analyses, especially in nanocavities and molecule–plasmon hybrids, will increasingly treat Raman signals not merely as static cross sections but as open quantum-system observables (Martínez-García et al., 2023).

Across these applications, a common theme emerges: eRS is most informative when its electronic origin is disentangled from phonons, luminescence, and instrumental background by symmetry, resonance, gating, or many-body modeling. Under those conditions it functions as a highly selective probe of interband continua, Pauli blocking, chirality, nodal and antinodal superconducting structure, nematic fluctuations, spin-orbit-enabled spin responses, collective charge modes, exciton condensation, moiré minibands, and stacking-dependent density-of-states features. That breadth is not the result of a single universal mechanism; rather, it reflects the fact that eRS measures symmetry-filtered electronic susceptibilities whose physical content depends on how photons couple to the specific low-energy Hilbert space of the material.