Partially Screened Voltage Gap

• Partially screened voltage gap is defined as a regime where a nominal voltage or band gap is only partially canceled by screening effects from ions, carriers, or interfacial states.
• It impacts diverse systems—from photovoltaics and 2D gate stacks to bilayer graphene—by reducing effective electronic barriers and modifying recombination or tunneling behaviors.
• In astrophysics, the concept applies to pulsar inner acceleration gaps where ion screening modulates particle acceleration, influencing observable emission patterns.

Searching arXiv for the cited works and closely related uses of the term across subfields. Partially screened voltage gap denotes a class of situations in which a nominal gap or voltage-controlled energy separation is reduced, redistributed, or only partly transmitted because the surrounding medium, interface, or charge reservoir screens the underlying electrostatic or recombination barrier. In the cited literature, the phrase appears most directly in connection with interface band-gap narrowing in Cu$_2$ZnSnS$_4$/CdS solar cells, Hartree- and Fock-screened band gaps in electrically biased bilayer graphene, and the partial voltage drop across vacuum-like van der Waals gaps in 2D gate stacks; closely related pulsar literature uses the Partially Screened Gap (PSG) model for an ion-screened inner acceleration region above the polar cap (Crovetto et al., 2017, Engdahl et al., 7 May 2025, Pourfath et al., 22 Sep 2025, Szary et al., 2014).

1. Scope and unifying structure

The cited literature uses related but not identical formulations: “Partially Screened Gap” in pulsar electrodynamics, “voltage gap” in tandem photovoltaics, “screening of the band gap” in bilayer graphene, and a vdW interfacial gap that causes a partial voltage drop in 2D gate stacks. A common structural feature is incomplete cancellation: an applied or intrinsic potential difference survives, but only after renormalization by ions, carriers, interface states, or a weakly polarizable interfacial region. This suggests that “partially screened voltage gap” is best treated as a family-resemblance concept rather than a single model-specific formalism.

Domain	Gap quantity	Screening agent or mechanism
CZTS/CdS photovoltaics	Interface recombination gap affecting $V_\mathrm{oc}$	Intrinsic surface/interface states (Crovetto et al., 2017)
Bilayer graphene	Bias-induced band gap ${\cal D}$	Hartree and Fock self screening (Engdahl et al., 7 May 2025)
2D gate stacks	Voltage drop across vdW interface	Weakly polarizable vdW gap (Pourfath et al., 22 Sep 2025)
Pulsars	Inner-gap accelerating potential $\Delta V$	Thermally emitted ions (Szary et al., 2014)
Tandem perovskites	$V_{\mathrm{oc}}$ gap between tandem and single-junction sum	Added nonradiative recombination in tandem stack (Yuce-Cakir et al., 16 Oct 2025)

The main distinction is between cases where screening reduces an electronic band gap or recombination barrier, and cases where it reduces an accelerating electric field. The shared language of “partial screening” therefore encodes an intermediate regime: neither unscreened nor fully shorted out.

2. Interface-limited voltage loss in Cu$_2$ZnSnS$_4$/CdS solar cells

In Cu$_2$ZnSnS$_4$ (CZTS) solar cells, the central claim is that the large open-circuit-voltage deficit can be explained by interface band gap narrowing rather than by a large cliff-like conduction-band offset. The relevant diagnostic is the recombination-energy deficit,

$_4$0

where $_4$1 is the absorber band gap and $_4$2 is the activation energy of the dominant recombination path inferred from temperature-dependent open-circuit-voltage analysis. In state-of-the-art CZTS cells, $_4$3, larger than the $_4$4 depth of bulk tail states, implying that the recombination barrier is additionally reduced at or near the interface (Crovetto et al., 2017).

First-principles DFT+NEGF calculations reveal localized states at the CZTS/CdS interface that are absent at the analogous CZTSe/CdS interface. These states are spatially confined to the CZTS side and behave as an extension of the CZTS valence band upward by about $_4$5 into the forbidden gap. The device model represents this as a $_4$6 upward shift of the valence band over a $_4$7 nm interfacial region, leaving other material parameters unchanged. The mechanism is explicitly not a single defect level and not a usual bulk valence-band tail; it is a local narrowing of the interface band gap on the CZTS side.

The recombination consequence is a strong rise in the interfacial hole density. In the simulation, adding the interface states increases the hole density at the interface by about three orders of magnitude, bringing it into the same range as the electron density and thereby enhancing Shockley–Read–Hall interface recombination,

$_4$8

Within a SCAPS model of a CZTS/CdS/ZnO device, the baseline simulation without interface states gives an extrapolated $_4$9 intercept of about $V_\mathrm{oc}$0 V, matching the transport gap of $V_\mathrm{oc}$1. With interface states included, the intercept drops to about $V_\mathrm{oc}$2 V, consistent with a narrowed interface transport gap of about $V_\mathrm{oc}$3, and the resulting recombination-energy deficit is $V_\mathrm{oc}$4, in agreement with experimental values around $V_\mathrm{oc}$5.

This is the most explicit photovoltaic instance of a partially screened voltage gap in the cited literature: the bulk absorber gap exists, but the interface electronic structure lowers the effective barrier that sets recombination. Zn-based alternative buffer layers are advantageous because Zn can passivate the surface states. The calculations show that when ZnS replaces CdS, the localized states disappear within the resolution of the calculation, and the paper notes that Zn$V_\mathrm{oc}$6Sn$V_\mathrm{oc}$7O$V_\mathrm{oc}$8 reduced $V_\mathrm{oc}$9 down to about ${\cal D}$0, close to the bulk tail-state depth.

3. Self-screened band gaps in electrically biased bilayer graphene

In biased Bernal bilayer graphene, a perpendicular electric field opens a direct band gap at the charge-neutrality point, but the observed gap is not equal to the bare gate-induced layer potential difference. The full gap ${\cal D}$1 and half-gap ${\cal D}$2 are defined by

${\cal D}$3

with layer separation ${\cal D}$4. If screening is ignored, the external gap is

${\cal D}$5

The central result is that the electronic system self-screens this bias, so the physical gap is smaller than ${\cal D}$6 (Engdahl et al., 7 May 2025).

Hartree screening enters through the layer-density imbalance: ${\cal D}$7 with Hartree correction

${\cal D}$8

The paper states that the Hartree screening alone typically renormalizes the gap by a factor of about ${\cal D}$9 relative to $\Delta V$0. The key new result is that exchange, treated as a Fock self-energy with an RPA-screened Coulomb interaction, is comparably important and in the low-density regime even more important than Hartree screening. The complete self-consistent equation is

$\Delta V$1

The low-density behavior is especially sharp. At zero temperature, the paper finds

$\Delta V$2

with the second term in the Fock expression scaling as $\Delta V$3 and negligible at very low $\Delta V$4. Numerically, the Hartree-Fock gap is often smaller than the Hartree-only gap by roughly a factor of $\Delta V$5 across realistic dielectric environments and gate distances. At zero doping and finite temperature, thermally excited carriers also strengthen screening, and the gap exhibits a step-like reduction around

$\Delta V$6

Here the partially screened voltage gap is a gate-controlled band gap that survives screening only in renormalized form. The screening is partial rather than complete because the bilayer does not cancel the applied bias; it reduces it by a factor of about $\Delta V$7, depending on density and environment.

4. van der Waals gaps as partially screened voltage drops in 2D gate stacks

At interfaces between 2D semiconductors and dielectrics, a van der Waals gap forms when the interface is held together by weak vdW interactions rather than covalent bonding. The paper defines the vdW gap geometrically as the distance between adjacent atomic planes minus the sum of their covalent radii. Because the region is vacuum-like with weak polarization, it reduces interfacial dielectric screening, adds series electrostatic thickness, and also adds a tunneling barrier (Pourfath et al., 22 Sep 2025).

First-principles calculations yield typical vdW gaps of about $\Delta V$8 with effective dielectric constant near $\Delta V$9. Treated as a series capacitor, the added equivalent oxide thickness is

$V_{\mathrm{oc}}$0

which gives

$V_{\mathrm{oc}}$1

This is the core electrostatic form of a partially screened voltage gap: the applied gate voltage is not fully screened by the main dielectric because part of the potential drops across the weakly polarizable interfacial region. The total EOT is approximately additive,

$V_{\mathrm{oc}}$2

The same gap acts as an additional tunneling barrier. The paper finds about one to two orders of magnitude reduction in gate leakage generally; two vdW gaps in graphene–hBN–graphene give about $V_{\mathrm{oc}}$3 current suppression, and a single $V_{\mathrm{oc}}$4 gap in a MoS$V_{\mathrm{oc}}$5–STO stack reduces tunneling by about an order of magnitude. The tradeoff is therefore explicitly double-edged: electrostatically bad because it adds EOT, leakage-wise good because it suppresses tunneling.

The paper introduces an insulator figure of merit and an analogous $V_{\mathrm{oc}}$6, estimating

$V_{\mathrm{oc}}$7

Its central scaling conclusion is that the vdW gap alone contributes about $V_{\mathrm{oc}}$8, so over half of a $V_{\mathrm{oc}}$9 IRDS EOT budget is consumed before the main dielectric contribution is counted. Most currently considered insulators are therefore unlikely to scale to an EOT of $_2$0 once the vdW gap is included. As a possible alternative, “zippered” structures with quasi-covalent bonding eliminate the vdW gap; the paper reports sub-$_2$1 nm EOT, experimentally around $_2$2 nm, for such structures.

5. Pulsar inner acceleration regions: the Partially Screened Gap model

In pulsar physics, the closely related term is the Partially Screened Gap, a model of the inner acceleration region above the polar cap. The observational motivation is that thermal X-ray hot spots are much smaller than the canonical dipolar polar cap, implying strongly non-dipolar surface fields and, by magnetic-flux conservation, surface magnetic fields of order $_2$3 G (Szary et al., 2011, Szary et al., 2012, Szary et al., 2014).

The screening variable is

$_2$4

where $_2$5 is the ion charge density emitted from the hot surface and $_2$6 is the Goldreich–Julian corotation density. The accelerating potential is reduced from the vacuum-gap value according to

$_2$7

The model assumes that the actual polar-cap temperature is close to a critical temperature at which thermionic ion emission becomes important. One formulation gives

$_2$8

while another gives

$_2$9

The thermal-balance condition is

$_4$0

The cited PSG literature is not uniform on the dominant breakdown channel. One study argued that curvature radiation plays dominant role in avalanche pair production in the PSG, especially for strongly curved non-dipolar field lines (Szary et al., 2011). Later PSG work separated CR-dominated and ICS-dominated regimes, denoted PSG-off and PSG-on, respectively (Szary et al., 2012, Szary et al., 2014). In the PSG-off mode, the gap is effectively weakly screened, $_4$1, particles reach $_4$2, and curvature radiation dominates. In the PSG-on mode, the gap is more strongly screened, particles reach $_4$3, inverse Compton scattering dominates, and the secondary plasma density is at least an order of magnitude higher than in the CR scenario. The model further connects the coexistence of two gap-breakdown channels with mode changing and pulse nulling.

Although this usage is not a semiconductor band-gap problem, it is structurally a partially screened voltage gap: the accelerating field above the polar cap is neither fully vacuum-like nor fully screened out.

6. Pulsar diagnostics from subpulse drifting and polarization

Subpulse drifting provides a direct observational route to PSG parameters. For drifting periodicity $_4$4, the shielding factor can be inferred as

$_4$5

where $_4$6 is the angle of the local non-dipolar magnetic field with the rotation axis. The PSG potential drop is then written as

$_4$7

In a model with sparks tightly packed near an elliptical polar-cap boundary, the spark pattern evolves in clockwise and counter-clockwise directions around a stationary central spark, and the drift speed in the two halves of the cap is

$_4$8

For PSR J1034$_4$93224, with $_2$0, the inferred parameters include $_2$1, $_2$2, $_2$3, $_2$4 m, $_2$5 K, and $_2$6 V; for PSR J1720$_2$72933, with $_2$8, the corresponding values are $_2$9, $_4$0, $_4$1, $_4$2 m, $_4$3 K, and $_4$4 V (Basu et al., 2023).

Polarization surveys extend the PSG interpretation from timing to emission physics. A meterwavelength single-pulse polarimetric survey observed 123 pulsars with periods longer than $_4$5 s. The abstract reports rotating-vector-model fits in 68 pulsars and emission-height measurements in 34 pulsars, while the detailed summary states that 50 pulsars had usable RVM fits and emission-height estimates. In all cases the radio emission was constrained to arise below $_4$6 of the light-cylinder radius. For pulsars with $_4$7 ergs s$_4$8, the mean fractional linear polarization of individual time samples in single pulses is around $_4$9, compared with $_4$00 in the average profiles, whereas the mean fractional circular polarization of individual samples is around $_4$01, similar to the measurements from the average profiles (Mitra et al., 2023).

The interpretation invokes dense spark-associated plasma clouds with high pair multiplicity, separated by low-density inter-cloud regions dominated by iron ions. Coherent curvature radiation from charge bunches in the dense clouds escapes as linearly polarized waves near cloud boundaries, while circular polarization arises due to propagation of waves in the low pair multiplicity, ion-dominated inter-cloud regions. The separation between linearly polarized intrinsic emission and circularly polarized propagation effects is presented as an interpretive framework rather than a unique proof.

7. Related terminology, adjacent usages, and ambiguity

The phrase “voltage gap” also appears in all-perovskite tandem solar cells, where it denotes the difference between the tandem open-circuit voltage and the sum of the corresponding single-junction values: $_4$02 For the devices analyzed, the sum of the single-junction $_4$03 values is about $_4$04 mV larger than the tandem $_4$05. Absolute electroluminescence hyperspectral imaging, ERE mapping, and two-diode modeling show that the narrow-bandgap subcell contributes the most toward this voltage gap: the wide-bandgap subcell changes only slightly relative to its single-junction counterpart, whereas the narrow-bandgap subcell shows $_4$06 to $_4$07 mV in tandem form (Yuce-Cakir et al., 16 Oct 2025). The paper describes the tandem structure as a case where voltage-loss contributions can be screened, but not fully, because the recombination layer and tandem growth environment introduce new nonradiative recombination pathways.

A further adjacent usage occurs in screened modified gravity, where “partially screened” denotes an intermediate thin-shell regime rather than a voltage or band-gap phenomenon. In that setting, gas and stars can be screened differently, with the expected signature

$_4$08

The MaNGA analysis reported no significant evidence for the predicted partially screened gas–star split and constrained $_4$09 at $_4$10 for all astrophysical fifth-force ranges, tightening to $_4$11 near Compton wavelength $_4$12 Mpc in the Hu–Sawicki case (Landim et al., 2024). This is not a voltage-gap result, but it shows that “partially screened” is a broader descriptor of intermediate response regimes.

Taken together, these literatures indicate that “partially screened voltage gap” does not name a single canonical theory. It denotes a recurrent physical pattern: a nominal gap, barrier, or potential drop is present, but its operative value is set by incomplete screening from ions, carriers, interface states, or weakly polarizable interfacial space.