Valley Photovoltaics: 2D Materials & III–V Devices

• Valley photovoltaics are phenomena where inequivalent momentum valleys generate selective photocurrents via Berry curvature and optical selection rules.
• Key studies in materials like monolayer BiAsI2 and germanene demonstrate gate-tunable, helicity-controlled switching of valley-polarized currents.
• Device models in III–V semiconductors reveal that intervalley scattering and hot-carrier dynamics limit efficiency despite engineered valley-selective contacts.

Searching arXiv for the papers on arXiv and closely related valley photovoltaics work to ground the article with current citations. Valley photovoltaics denotes photovoltaic and photogalvanic phenomena in which inequivalent valleys in momentum space participate differently in optical excitation, carrier transport, and electrical readout. In the cited literature, the term is used in two related but distinct senses. One concerns valley-polarized photocurrent generation in two-dimensional materials with valley-contrasting Berry curvature, optical selection rules, and nonlinear bulk photovoltaic responses. The other denotes a hot-carrier solar-cell concept in multivalley III–V semiconductors, where intervalley scattering into satellite valleys is intended to support high-voltage extraction through valley-selective contacts (Yang et al., 2024, Xia et al., 26 Jul 2025). In both usages, the central question is how valley-indexed populations can be created, preserved, switched, and converted into measurable dc signals.

1. Scope and representative material platforms

In hexagonal two-dimensional crystals, the relevant valleys are typically the inequivalent $K$ and $K'$ points of the Brillouin zone, often labeled by a valley index $\tau_z=\pm 1$. Valley photovoltaics in this setting is usually concerned with valley-polarized photocurrent, meaning that photoexcited carriers originate predominantly from one valley rather than equally from both. In multivalley III–V absorbers, by contrast, the operative valleys are the primary $\Gamma$ minimum and higher-energy satellite valleys such as $L$, and the photovoltaic objective is to use intervalley scattering as a hot-carrier management mechanism (Yang et al., 2024, Xia et al., 26 Jul 2025).

The literature represented here spans electrically switchable circular bulk photovoltaic effects in buckled honeycomb monolayers, hidden valley-polarized bulk photovoltaic responses in antiferromagnets, excitonic valley Hall transport in transition-metal dichalcogenide heterostructures, photonic valley analogues, and device-scale drift-diffusion modeling of multivalley solar cells (Xue et al., 2023, Huang et al., 2019, Chen et al., 2016, Dai et al., 28 Mar 2026).

Platform	Mechanism	Reported result
Monolayer BiAsI$_2$, germanene	Circular BPVE / CPGE with gate-tuned Berry curvature	Fully valley-polarized photocurrent switched between $K$ and $K'$
MnPSe$_3$ monolayer	PT-symmetric valley-contrasting BPV	Hidden valley-polarized photoconductivity of $\sim 1350\,\mu\text{A/V}^2$
MoS$K'$0/WSe$K'$1	Interlayer exciton valley Hall transport	Room-temperature valley Hall separation of $K'$2
InGaAs/InAlAs VPV device	Intervalley-scattering hot-carrier concept	S-shaped $K'$3–$K'$4 curves, low fill factor, and low efficiency

2. Geometric and optical foundations

For valley-polarized photovoltaic responses in nonmagnetic two-dimensional materials, the essential structure is a valley-contrasting Berry curvature field. In inversion-broken but time-reversal-symmetric systems, the Berry curvature obeys $K'$5, so the two valleys carry opposite geometric response. This underlies valley Hall transport, helicity-dependent optical selection rules, and valley-selective nonlinear photocurrents (Yang et al., 2024).

A minimal description for buckled honeycomb systems is the two-band Dirac Hamiltonian

$K'$6

where $K'$7 is a staggered sublattice potential controlled by an out-of-plane electric field. The corresponding band energies are

$K'$8

and the valence-band Berry curvature in a given spin-valley sector is

$K'$9

Because the sign is controlled by $\tau_z=\pm 1$0, an out-of-plane electric field can continuously tune and reverse the Berry curvature at a given valley (Yang et al., 2024).

The optical selectivity is encoded by the degree of circular polarization,

$\tau_z=\pm 1$1

with $\tau_z=\pm 1$2 for exclusive coupling to $\tau_z=\pm 1$3 light and $\tau_z=\pm 1$4 for exclusive coupling to $\tau_z=\pm 1$5. The circular bulk photovoltaic effect enters through the valley-resolved CPGE tensor,

$\tau_z=\pm 1$6

which makes the valley photocurrent explicitly proportional to the Berry curvature. A sign reversal of $\tau_z=\pm 1$7 therefore reverses both the optical selection rule and the valley-resolved CPGE response (Yang et al., 2024).

3. Electrically switchable valley-polarized photocurrent in two-dimensional materials

The clearest explicit realization of valley photovoltaic switching in the present corpus is the electric-field-controlled circular bulk electro-photovoltaic effect in monolayer BiAsI$\tau_z=\pm 1$8 and germanene. In the two-band Dirac model, the regime $\tau_z=\pm 1$9 corresponds to a quantum spin Hall effect phase, while $\Gamma$0 corresponds to a quantum valley Hall effect phase. Crossing $\Gamma$1 closes and reopens the gap, producing a topological phase transition and a sign flip of the Berry curvature at each valley (Yang et al., 2024).

Monolayer BiAsI$\Gamma$2 is a buckled hexagonal lattice with broken inversion symmetry, strong spin–orbit coupling, and a direct gap at $\Gamma$3 of $\Gamma$4 meV with SOC. At zero out-of-plane field it is in a QSHE phase. Under a gate-induced out-of-plane electric field, density-functional calculations show that the bandgap decreases with increasing $\Gamma$5, closes at a critical gate, and then reopens, signaling a QSHE $\Gamma$6 QVHE transition. The Berry-curvature maps show that $\Gamma$7 at $\Gamma$8 flips sign after the transition, and the degree of circular polarization at $\Gamma$9 switches from $L$0 to $L$1 as $L$2 crosses the transition (Yang et al., 2024).

The first-principles CPGE tensor component $L$3, computed at $L$4 eV, exhibits a dipole distribution near each valley and a nonzero net valley-integrated response with opposite signs in $L$5 and $L$6. As the gate voltage is increased from $L$7 to $L$8 V, $L$9 switches sign from positive to negative, while $_2$0 switches from negative to positive. In a two-probe NEGF-DFT device with five unit cells of monolayer BiAsI$_2$1 in the central region, semi-infinite BiAsI$_2$2 leads, and circularly polarized illumination incident along $_2$3, the calculated normalized valley photoresponse shows fully valley-polarized switching. Under $_2$4 light, $_2$5 yields $_2$6 and $_2$7, with $_2$8; at $_2$9 V, $K$0 and $K$1 is finite, giving $K$2. Under $K$3 light, the pattern is reversed (Yang et al., 2024).

Germanene realizes a related but symmetry-distinct case. In pristine form it has inversion symmetry, so at $K$4 the Berry curvature vanishes everywhere and the circular BPVE is zero. Applying an out-of-plane electric field breaks inversion symmetry, generates nonzero valley-contrasting Berry curvature, and allows the sign of $K$5 to be controlled by the sign of $K$6. DFT calculations show that as $K$7 is swept from $K$8 to $K$9 V, the band gap closes twice, indicating topological transitions between different QVHE phases. At $K'$0, both $K'$1 and $K'$2 vanish. For $K'$3 light, $K'$4 V gives only $K'$5 response with $K'$6, while $K'$7 V gives only $K'$8 response with $K'$9; for $_3$0 light the pattern is reversed. Because germanene is symmetric at zero gate, the gate-voltage dependence is symmetric in $_3$1, unlike the asymmetric case in BiAsI$_3$2 (Yang et al., 2024).

4. Hidden, excitonic, and magnetic valley photoresponses

Valley photovoltaics is not restricted to a net dc current flowing in a uniform sample. In PT-symmetric antiferromagnetic MnPSe$_3$3 monolayer, the total bulk photovoltaic current can be symmetry-forbidden while the valley-resolved currents are large and opposite, producing what is termed hidden valley-polarized photoconductivity. The relevant nonlinear response is the magnetic injection current under linearly polarized light,

$_3$4

so the response is governed by the velocity difference and the quantum metric. In MnPSe$_3$5, the Néel vector controls the magnetic point group at $_3$6, the allowed current direction, and the valley-local symmetry reduction at $_3$7. Even when the global current component vanishes, the valley-resolved photoconductivity can remain sizable, reaching $_3$8. The same symmetry control is tied to a predicted two-dimensional ferrotoroidic response, again depending on the Néel vector direction (Xue et al., 2023).

A second route is excitonic rather than electronic. In near-zero-twist MoS$_3$9/WSe$\sim 1350\,\mu\text{A/V}^2$0, type-II band alignment drives electrons to the MoS$\sim 1350\,\mu\text{A/V}^2$1 conduction band and holes to the WSe$\sim 1350\,\mu\text{A/V}^2$2 valence band, forming interlayer excitons with an out-of-plane permanent dipole moment and a long lifetime. Polarization-resolved photoluminescence shows interlayer exciton valley polarization of $\sim 1350\,\mu\text{A/V}^2$3 at low temperature and $\sim 1350\,\mu\text{A/V}^2$4 at room temperature, while time-resolved photoluminescence shows lifetimes on the order of nanoseconds. Because the excitons inherit valley-contrasting Berry curvature, they exhibit a valley Hall effect under a potential gradient. The observed transverse separations between $\sim 1350\,\mu\text{A/V}^2$5 and $\sim 1350\,\mu\text{A/V}^2$6 interlayer-exciton PL maxima are $\sim 1350\,\mu\text{A/V}^2$7 and $\sim 1350\,\mu\text{A/V}^2$8 at $\sim 1350\,\mu\text{A/V}^2$9 K, and $K'$00 at room temperature (Huang et al., 2019).

The semiclassical description of this excitonic transport uses the anomalous velocity term associated with Berry curvature. For electrons,

$K'$01

and the supplementary phenomenology for interlayer excitons adopts

$K'$02

This makes the photovoltaic relevance direct: a built-in or engineered potential gradient can convert optically prepared valley populations into spatially separated exciton flows, which in turn suggests lateral valley-selective collection schemes if dissociation into free carriers is provided (Huang et al., 2019).

5. Topology, photonic analogues, and the status of valley Chern numbers

Photonic valley systems furnish an important analogue because they isolate the geometric structure of valley degrees of freedom, Berry curvature, and chirality without electronic transport complications. In all-dielectric photonic valley crystals consisting of two interlaced triangular sublattices of silicon rods in air, inversion asymmetry opens a band gap at $K'$03 and $K'$04. A $K'$05 description takes the form

$K'$06

with Berry curvature

$K'$07

and valley Chern numbers $K'$08, $K'$09 within the simplified Dirac picture. These structures exhibit valley-contrasting orbital angular momentum, selective excitation of $K'$10 or $K'$11 bulk states by sources carrying orbital angular momentum of matching chirality, and valley-dependent edge states supporting broadband robust transmission through a Z-shaped bend over a bandwidth of about $K'$12 (Chen et al., 2016).

The more recent photonic reassessment complicates this standard picture. In time-reversal-symmetric valley photonic crystals, the global Chern number vanishes, and valley Chern numbers are defined only by integrating Berry curvature over a half-Brillouin zone. Systematic calculations across continuous families of photonic-crystal designs show that these valley Chern numbers are generically unquantized and vary continuously with structural parameters. The Berry-curvature distribution develops fine structures described as “triangle,” “petal,” and “island,” and the sign of the valley Chern number need not match the sign of the Berry curvature at the valley point itself because off-$K'$13 regions can dominate the integral. The paper attributes the unquantized values to intra- and inter-valley cancellation of Berry curvature and argues that there is no protecting mechanism for quantization (Dai et al., 28 Mar 2026).

For valley photovoltaics, this establishes an important interpretive limit. The photonic work explicitly concludes that valley-dependent responses in time-reversal-symmetric systems are geometric rather than strictly topological, and a plausible implication is that photovoltaic observables should be analyzed in terms of the detailed Berry-curvature distribution over the optically active $K'$14-space region rather than by assigning a protected half-integer topological charge to each valley. This aligns with the explicit dependence of CPGE, magnetic injection current, and excitonic anomalous velocity on Berry curvature, quantum metric, and occupation-weighted integrals rather than on a global topological invariant (Dai et al., 28 Mar 2026).

6. Device architectures, modeling, and practical limits

The device concepts emerging from the two-dimensional literature are opto-valleytronic rather than conventional p–n-junction photovoltaics. In BiAsI$K'$15 and germanene, the proposed architecture is a monolayer channel between leads, a back gate controlling the out-of-plane electric field and thus the staggered potential $K'$16, and circularly polarized light incident normal to the plane. The intended write-read-flip cycle uses gate voltage to choose which valley is active under a given helicity, optical illumination to generate the valley-polarized photocurrent, and the lead current as the electrical readout. Because the state depends on the instantaneous gate voltage and light, the operation is volatile (Yang et al., 2024). In MnPSe$K'$17, a domain wall between antiferromagnetic domains related by time reversal is proposed as a way to convert hidden valley-polarized currents into a measurable signal, since carriers from a given valley accumulate at the wall (Xue et al., 2023). In MoS$K'$18/WSe$K'$19, the out-of-plane dipole of interlayer excitons suggests electrical control over transport and dissociation, although no photocurrent device is implemented in the reported experiment (Huang et al., 2019).

A different usage of “valley photovoltaics” appears in multivalley III–V semiconductors. Here the concept is a proposed hot-carrier solar cell using the $K'$20 valley and higher-energy satellite valleys such as $K'$21 in In$K'$22Ga$K'$23As. The modeled device consists of front n$K'$24 In$K'$25Al$K'$26As, n-In$K'$27Ga$K'$28As as the absorber and valley-scattering region, and back p$K'$29 In$K'$30Al$K'$31As. At $K'$32 K, the paper uses $K'$33 eV and $K'$34 eV, with a built-in electric field of about $K'$35 kV/cm in the original device. The device-scale model introduces separate quasi-Fermi levels for the $K'$36 and $K'$37 valleys, drift-diffusion currents

$K'$38

Poisson’s equation, and a valley-scattering source term

$K'$39

To embed ensemble Monte Carlo scattering data into a heterogeneous device, the model replaces the homogeneous electric field by a quasi-electric field $K'$40, so that detailed balance is recovered at equilibrium (Xia et al., 26 Jul 2025).

The principal result of that modeling is negative. Nonequilibrium carrier populations in satellite valleys are not enough for valley photovoltaics to achieve high efficiency, and increasing the built-in electric field of the valley-scattering region does not improve efficiency. The simulations reproduce the experimentally observed S-shaped current-voltage curves and low fill factor without invoking explicit extraction barriers or nonradiative losses as the dominant explanation. The S-shape arises because valley scattering that is helpful in reverse bias becomes adverse in forward bias: the net transfer into the extracting $K'$41 valley decreases and can reverse sign as the bias raises $K'$42 relative to $K'$43. For the best fit to experiment, the low-field $K'$44 scattering rate is taken as $K'$45. Thinning the InGaAs layer improves fill factor, but the model attributes this not to improved valley scattering in the power-generating regime; rather, more light reaches the bottom InAlAs, and the extra InAlAs photocurrent improves the $K'$46–$K'$47 curve. Under the quasi-equilibrium assumptions of the model, the open-circuit voltage remains limited by the $K'$48-valley gap, not by the larger $K'$49-valley energy scale (Xia et al., 26 Jul 2025).

Taken together, these results define the present encyclopedic meaning of valley photovoltaics. In the two-dimensional valleytronic literature, it refers to Berry-curvature-driven photovoltaic and photogalvanic phenomena that can be helicity-selective, gate-switchable, magnetic-order-controlled, or excitonic. In the multivalley III–V device literature, it denotes a hot-carrier strategy based on intervalley scattering and valley-selective extraction, whose first device-scale modeling identifies strong limitations under quasi-equilibrium transport. The common thread is the attempt to convert valley-resolved nonequilibrium populations into useful electrical output; the major open issue is not whether valleys can be optically addressed, but under what symmetry, scattering, and device conditions that addressability can survive as a robust, measurable, and efficient photovoltaic function (Yang et al., 2024, Xia et al., 26 Jul 2025).