Forbidden Photocurrent Selection Rule

• The Forbidden Photocurrent Selection Rule is a symmetry constraint that forces complete cancellation of photocurrent under time-reversal symmetry.
• It utilizes bichromatic ω–2ω fields to isolate TRS effects from spatial mirror symmetry, yielding an exact zero current in materials like graphene.
• The rule serves as a diagnostic tool to detect TRS breaking and topological transitions, and is validated through nonperturbative TDDFT simulations and experimental setups.

Searching arXiv for the cited work and closely related papers to ground the article in current literature. A forbidden photocurrent selection rule is a symmetry constraint that forces a specified photocurrent channel to vanish identically rather than merely suppressing it parametrically. In its most explicit recent formulation, the rule concerns bulk photogalvanic-effect photocurrent in a time-reversal-symmetric material driven by a bichromatic $\omega$–$2\omega$ field that itself preserves time-reversal symmetry: even when the tailored light breaks all in-plane spatial point-group symmetries, the net light-field-driven current must still be zero because time-reversal symmetry enforces pairwise cancellation between $k$ and $-k$ contributions (Lesko et al., 8 Jul 2025). More broadly, the same phrase is used for symmetry-enforced zeros in helicity-dependent photocurrent, magneto-photocurrent, and dark-exciton photocurrent channels, with closely related analogues in Floquet photonic frequency conversion (Quereda et al., 2020, Kim et al., 31 Mar 2026, Quereda et al., 2017, Camacho, 11 Jun 2026).

1. Symmetry content of the rule

In the tailored-light formulation, the relevant symmetries are mirror symmetry, inversion symmetry, and time-reversal symmetry (TRS) of the combined light–matter system. The central case is a TRS-preserving bichromatic drive satisfying

$\mathbf{E}(t)=\mathbf{E}(-t),$

applied to a TRS- and inversion-symmetric material such as graphene. Because graphene has degenerate states at $k$ and $-k$ and zero Berry curvature, the driven carrier distribution obeys

$n_c(k)=n_c(-k),$

while band velocities satisfy

$\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$

The Brillouin-zone current therefore cancels pairwise,

$$\mathbf{J}\propto \sum_c \int_{\mathrm{BZ}} d^2k\, \mathbf{v}_c(k)\,n_c(k)=0,$$

so the photocurrent is symmetry-forbidden for all field strengths (Lesko et al., 8 Jul 2025).

This formulation is distinct from a conventional spatial selection rule that nulls only one tensor component. The vanishing is global: the relevant injection-current tensor elements, or equivalently the full Brillouin-zone integral defining the current, vanish identically. The rule therefore isolates TRS itself as the protecting symmetry.

The tailored field supplies two independent control parameters: the relative polarization angle $2\omega$0 between the $2\omega$1 and $2\omega$2 components, and the two-color phase $2\omega$3. These parameters allow mirror symmetry and TRS in the light to be independently preserved or broken. For $2\omega$4 and $2\omega$5, the field obeys the dynamical mirror relation

$2\omega$6

which forbids $2\omega$7 through mirror symmetry. By contrast, for

$2\omega$8

and generic $2\omega$9, the waveform is fully two-dimensional, breaks mirror symmetry, but preserves TRS. In that regime, any vanishing photocurrent cannot be attributed to remaining spatial symmetry and is therefore a purely TRS-induced forbidden photocurrent selection rule (Lesko et al., 8 Jul 2025).

2. Nonlinear photocurrent mechanism and exact cancellation

In graphene, the relevant bulk photogalvanic response is injection current rather than shift current. The reason given is that graphene, being both inversion- and TRS-symmetric, has zero Berry curvature throughout the Brillouin zone, so shift-current-type terms tied to Berry curvature vanish. Any bulk photogalvanic effect must therefore arise from asymmetric conduction-band population, or higher-order but still population-based processes (Lesko et al., 8 Jul 2025).

The treatment is explicitly nonperturbative. The current is computed in real-time TDDFT from the Kohn–Sham states, and the injected photocurrent is defined as the single-cycle average after the pulse,

$k$0

Conceptually, however, the current retains the standard structure

$k$1

with the occupation $k$2 a highly nonlinear function of the bichromatic field (Lesko et al., 8 Jul 2025).

The symmetry argument is then immediate. TRS maps $k$3, complex conjugates the Bloch states, and reverses velocities: $k$4 If the field also preserves TRS, then the drive does not distinguish the $k$5 and $k$6 sectors, and the populations satisfy

$k$7

Each $k$8 pair contributes equal and opposite current, so $k$9 at all times up to numerical noise, and the cycle-averaged $-k$0 vanishes as well. The selection rule is therefore an exact cancellation of all contributions to the nonlinear current operator, not a perturbatively small coefficient (Lesko et al., 8 Jul 2025).

A frequent misconception is to identify any zero-current line in a polarization scan with a mirror or point-group selection rule. The tailored-light analysis shows that this is incomplete. At $-k$1, vanishing current can indeed be caused by dynamical mirror symmetry. At $-k$2 and generic $-k$3, however, the field breaks mirror symmetry while preserving TRS, so the zero is enforced solely by TRS.

3. Tailored-light implementation and graphene demonstration

The bichromatic probe consists of linearly polarized $-k$4 and $-k$5 beams with adjustable relative phase and polarization angle. In the simulations, the vector potential uses a fixed amplitude ratio $-k$6, with $-k$7 and $-k$8 as the control parameters. The corresponding electric field is $-k$9. For $\mathbf{E}(t)=\mathbf{E}(-t),$0, the drive preserves TRS for any $\mathbf{E}(t)=\mathbf{E}(-t),$1, while mirror symmetry is generically broken for $\mathbf{E}(t)=\mathbf{E}(-t),$2. For $\mathbf{E}(t)=\mathbf{E}(-t),$3 and $\mathbf{E}(t)=\mathbf{E}(-t),$4, the field instead realizes the dynamical mirror symmetry described above (Lesko et al., 8 Jul 2025).

The experimental implementation used a 1550 nm erbium fiber laser frequency-doubled to 775 nm in BiBO, yielding synchronized $\mathbf{E}(t)=\mathbf{E}(-t),$5 and $\mathbf{E}(t)=\mathbf{E}(-t),$6 pulses. The relative phase $\mathbf{E}(t)=\mathbf{E}(-t),$7 was controlled by a calcite inline interferometer, and the relative polarization angle $\mathbf{E}(t)=\mathbf{E}(-t),$8 by a dichroic two-color half-wave plate. The pulses were focused onto monolayer epitaxial graphene on SiC in vacuum, and currents were measured along $\mathbf{E}(t)=\mathbf{E}(-t),$9 through gold electrodes (Lesko et al., 8 Jul 2025).

The measured and simulated map $k$0 exhibits two symmetry-enforced zero-current lines. First, $k$1 for all $k$2 at $k$3, consistent with the dynamical mirror rule. Second, $k$4 for all $k$5 at $k$6, identifying the forbidden TRS selection rule. Away from these loci, when both TRS and mirror symmetry are broken, the current is nonzero and strongly phase- and angle-dependent (Lesko et al., 8 Jul 2025).

The significance of the graphene result is not merely that a node exists, but that the node persists even when all in-plane spatial symmetries are broken. That feature makes the selection rule a direct diagnostic of whether the material itself preserves TRS.

4. Lifting the rule in time-reversal-broken and topological phases

The same tailored probe becomes diagnostically useful when TRS is broken in the material while the probe itself remains TRS-preserving. In that case, any current measured at the “forbidden” probe configuration is a direct signature that the material no longer obeys the TRS cancellation argument (Lesko et al., 8 Jul 2025).

The first example is monolayer CrI$k$7, treated as a 2D hexagonal ferromagnet with finite local magnetic moments on Cr atoms. TDDFT simulations at $k$8 show that graphene retains a clear node at $k$9, whereas CrI$-k$0 has no exact node: the minimum is shifted from $-k$1 to $-k$2 and never reaches zero. The interpretation given is that broken TRS lifts the degeneracy between $-k$3 and $-k$4, so $-k$5 and/or $-k$6, and the Brillouin-zone cancellation fails. This creates a background-free signal for magnetism at the nominally forbidden operating point (Lesko et al., 8 Jul 2025).

The second example is a Floquet topological insulator in graphene produced by dressing with a circularly polarized mid-infrared field at 2200 nm. The dressing field creates a Floquet Chern insulator with a small topological gap of about $-k$7 eV, while the probe remains the same TRS-preserving two-color field. In the dressed state, the photocurrent is nonzero for all $-k$8, including $-k$9, and the minimum shifts from $n_c(k)=n_c(-k),$0 to about $n_c(k)=n_c(-k),$1. As the dressing intensity increases, the photocurrent magnitude grows and the minimum evolves nontrivially, tracking the increasing gap and the spreading of Berry curvature from $n_c(k)=n_c(-k),$2 (Lesko et al., 8 Jul 2025).

The paper explicitly interprets the disappearance of the forbidden node as sensitivity to TRS breaking, Berry curvature, and Chern physics. The forbidden configuration therefore functions as a reference point: in a TRS-invariant system it yields exactly zero current; in a TRS-broken system it yields a finite response whose magnitude, sign, and phase dependence encode magnetic or Floquet-topological information.

A closely related symmetry-lowering mechanism appears in chiral cubic Bi$n_c(k)=n_c(-k),$3SiO$n_c(k)=n_c(-k),$4, where global crystal symmetry forbids a longitudinal odd-in-$n_c(k)=n_c(-k),$5 magneto-photocurrent in Voigt geometry. There, oxygen vacancies create in-gap bound states and localized magnetic moments, and an applied magnetic field selects a time-reversal-broken sector of the defect ensemble, lowering the effective magnetic symmetry and lifting the longitudinal selection rule. The resulting circular and linear magneto-photocurrent channels correlate spatially in $n_c(k)=n_c(-k),$6-space with Berry-curvature-rich and quantum-metric-rich regions, respectively (Kim et al., 31 Mar 2026).

5. Related forbidden and allowed photocurrent channels

The broader photocurrent literature shows that forbiddenness depends on the actual symmetry of the device or effective electronic state, not only on the ideal bulk crystal.

In monolayer MoSe$n_c(k)=n_c(-k),$7 with $n_c(k)=n_c(-k),$8 symmetry, the circular photogalvanic effect is forbidden, and the circular photon drag effect contributes only terms proportional to $n_c(k)=n_c(-k),$9, which vanish at normal incidence. Direct metal–MoSe$\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$0 Schottky contacts, however, generate strong local electric fields that break the horizontal mirror and reduce the symmetry “to at most a single mirror plane,” allowing additional circular photocurrent contributions. Experimentally, h-BN tunnel-barrier devices show $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$1, consistent with the forbidden rule, whereas direct-contact devices show nonzero helicity-dependent current at normal incidence, indicating that the device environment lifts the nominal selection rule (Quereda et al., 2020).

A different relaxation of optical forbiddenness occurs in monolayer MoSe$\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$2 photocurrent spectroscopy. The bright $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$3 exciton appears at $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$4 eV, while a second peak at $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$5 eV emerges only when gate voltage brings the Fermi level near the bottom of the conduction band. The peak is assigned to the spin-forbidden dark exciton $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$6, with a splitting of about $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$7 meV relative to $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$8, matching the conduction-band spin–orbit splitting. The interpretation is that gating activates many-body pathways involving conduction electrons, so a transition that is dark in ordinary optics becomes visible in photocurrent because excitation and dissociation need not obey the same effective constraints as radiative recombination (Quereda et al., 2017).

The symmetry logic also has a noise counterpart. In second-order DC photocurrent shot noise, inversion symmetry still forbids the average DC photocurrent, but does not forbid the DC shot noise, because the noise susceptibility tensors are inversion-even. In a time-reversal-invariant system, the allowed channels are polarization-selective: shift DSN is allowed for circularly polarized light, and injection DSN is allowed for linearly polarized light, whereas the complementary combinations are time-reversal-odd and therefore forbidden (Xiang et al., 2023). This establishes that the conventional “inversion forbids photocurrent” rule does not automatically extend to fluctuations of the photocurrent operator.

6. Analogues, interpretation, and significance

A photonic analogue is provided by temporal glide symmetry in a time-modulated trilayer waveguide. There, the combined operation of transverse reflection and half-period time translation enforces an exact parity–sideband selection rule: $\mathbf{v}_c(k)=-\mathbf{v}_c(-k).$9 For a parity-pure incident mode, scattering into sideband $$\mathbf{J}\propto \sum_c \int_{\mathrm{BZ}} d^2k\, \mathbf{v}_c(k)\,n_c(k)=0,$$0 is allowed only into output modes of parity $$\mathbf{J}\propto \sum_c \int_{\mathrm{BZ}} d^2k\, \mathbf{v}_c(k)\,n_c(k)=0,$$1; the complementary channels are symmetry-forbidden and are suppressed to numerically negligible values (Camacho, 11 Jun 2026). Although this is not a photocurrent problem, it is structurally the same kind of symmetry-protected zero.

The general lesson is that forbidden photocurrent selection rules are statements about the symmetry of the full driven system. They may be enforced by TRS of the light–matter Hamiltonian, by crystal point-group operations, by magnetic point groups in a finite field, or by spatiotemporal Floquet symmetries. They are lifted not only by explicit symmetry breaking in the lattice, but also by effective symmetry reduction through contacts, built-in fields, carrier populations, defects, or external dressing fields.

Within that framework, the tailored-light rule in graphene is distinctive for two reasons. First, it isolates TRS from other suppressing symmetries by using a bichromatic field that preserves TRS while breaking spatial symmetries. Second, it yields an exactly vanishing background in a TRS-invariant reference system, so any finite signal at the same operating point is a direct marker of TRS breaking (Lesko et al., 8 Jul 2025). This makes the forbidden photocurrent selection rule not merely a constraint on nonlinear response, but a symmetry-resolved spectroscopy principle for ultrafast magnetism and Chern physics.