Persistent Photoconductivity (PPC) Overview

• Persistent Photoconductivity (PPC) is a long-lasting photo-induced state characterized by a prolonged increase in conductivity after light exposure.
• It involves various mechanisms such as metastable defect configurations, adsorption/desorption processes, and disorder-induced localization across differing material platforms.
• PPC is significant for applications including high-gain photodetection, optically written memory, and in situ tuning of carrier density in advanced semiconductor research.

Searching arXiv for relevant papers on persistent photoconductivity to ground the article in current and canonical literature. arXiv search query: persistent photoconductivity semiconductor oxide MoS2 YBCO diamond GeSbTe SrTiO3 Persistent photoconductivity (PPC) is a long-lived increase in electrical conductivity that persists after light illumination is turned off. Across the literature, PPC is reported as a metastable photo-induced conducting state in systems as different as individual oxide nanostructures, highly doped III–V alloys, oxide heterostructures, chalcogenide films, monolayer transition-metal dichalcogenides, hydrogen-terminated diamond, and oxygen-deficient cuprates. Depending on the material and defect landscape, the persistence ranges from seconds to hours, weeks, up to a month, and in room-temperature SrTiO$_3$ even over a year, so the term denotes a kinetic regime rather than a single microscopic mechanism (Luo et al., 1 Dec 2025, Misuraca et al., 2014, Dhara et al., 2020, Zhang et al., 2019).

1. Definition, observables, and operational criteria

PPC is operationally identified when the post-illumination conductance or current remains substantially above the dark value on timescales far exceeding ordinary carrier lifetimes. In the oxide-heterostructure literature it is described as a persistent perturbation induced by light; in semiconductor transport studies it is treated as a persistent excess conductance $\Delta G(t)$, excess current $\Delta I(t)$, or a long-lived resistance change $\Delta R(t)$; and in cuprates it can be tracked through a persistent decrease of $\rho_{xx}(T)$, together with a light-induced increase of $T_c$ (Gennaro et al., 2013, Ovadyahu, 2018, Hage et al., 2023).

The same label therefore covers several experimental observables. In SnO$_2$ nanobelts, PPC is the slow decay of photocurrent after 403 nm illumination is removed, with measurable lifetimes up to $1.20\times10^5$ s in vacuum at 300 K (Viana et al., 2012). In highly Si-doped Al$_{0.3}$Ga$_{0.7}$As, PPC is a persistent photodoping state that survives for “weeks” at low temperature and enables in situ tuning of carrier density across the metal–insulator transition (Misuraca et al., 2014). In monolayer MoS$\Delta G(t)$0, it appears as a long-lived enhancement of channel current after illumination, with decay times exceeding $\Delta G(t)$1 s in one study and a bi-exponential giant PPC with a $\Delta G(t)$2-day component in another (Bartolomeo et al., 2017, George et al., 2020). In hydrogenated diamond, PPC is the slow, strongly asymmetric photocurrent decay after 400 nm sub-bandgap excitation, with room-temperature decay times that decrease from $\Delta G(t)$3 s to $\Delta G(t)$4 s under progressive oxygen termination (Sulthana et al., 9 Jul 2025).

An important terminological boundary concerns sign. Most work uses PPC to denote a positive conductance change. A related but distinct regime, negative persistent photoconductance, was reported in BP–MoS$\Delta G(t)$5 heterostructures, where visible-light-driven interlayer recombination reduces the majority-hole density in BP and yields a slow negative conductance component with $\Delta G(t)$6 s and $\Delta G(t)$7 s (Jawa et al., 2021).

2. Microscopic mechanisms

No single microscopic mechanism accounts for all PPC. The reported mechanisms fall into several recurrent classes: metastable defect configurations with recombination barriers, adsorption/desorption-controlled surface charging, random local potential fluctuations, and coupled deep/shallow trap kinetics.

In highly Si-doped Al$\Delta G(t)$8Ga$\Delta G(t)$9As with $\Delta I(t)$0, PPC is associated with the standard deep-donor DX-center regime. Illumination raises electrons out of deep donor states, and a barrier for recapture leaves the photoexcited electrons mobile at low temperature; the material therefore functions as a persistent photodoping medium rather than merely a transient photoconductor (Misuraca et al., 2014). In room-temperature SrTiO$\Delta I(t)$1, hybrid-DFT calculations identify a different metastable-defect route: sub-bandgap excitation of substitutional hydrogen $\Delta I(t)$2 to $\Delta I(t)$3, followed by a low-barrier transformation to $\Delta I(t)$4. The full proposed reaction,

$\Delta I(t)$5

provides both the large conductivity increase and the extreme persistence (Zhang et al., 2019).

In surface-dominated oxides, PPC can instead be governed by adsorption/desorption and surface electrostatics. For individual SnO$\Delta I(t)$6 nanobelts under 403 nm sub-bandgap illumination, the central reactions are

$\Delta I(t)$7

and

$\Delta I(t)$8

The reported conclusion is explicit: the molecular-oxygen recombination with holes is the origin of PPC in this system, and the effect is not related to oxygen vacancies as commonly presented in the literature (Viana et al., 2012). By contrast, in STO/Al heterostructures, the PPC state is attributed to slow re-trapping of photoexcited carriers into deep OV-related states, while a distinct gate-voltage-induced trapping channel involves shallower states with much faster $\Delta I(t)$9–$\Delta R(t)$0 s dynamics (Luo et al., 1 Dec 2025).

A third class invokes disorder-induced potential landscapes. In monolayer MoS$\Delta R(t)$1 on OTS/SiO$\Delta R(t)$2, PPC was attributed to random localized potential fluctuations of predominantly extrinsic origin, with a direct correlation to percolation transport and strong substrate dependence; suspended devices showed negligible PPC (Ovadyahu, 2015). In a later study on CVD monolayer MoS$\Delta R(t)$3 in high vacuum, giant PPC was instead linked mainly to intrinsic sulfur-vacancy and strain-induced localized states, with STS and HRTEM supporting a random-potential description characterized by $\Delta R(t)$4 eV and $\Delta R(t)$5 nm (George et al., 2020). Taken together, these results suggest that “intrinsic” and “extrinsic” labels are device- and growth-dependent rather than universally transferable across MoS$\Delta R(t)$6 platforms.

Hydrogen-terminated diamond provides a closely related but chemically distinct example. There, PPC is assigned to random local potential fluctuations created by inhomogeneous hydrogen termination and surface adsorbates, reinforced by Coulomb interactions between the two-dimensional hole gas and the negatively charged adsorbate layer; the reported transport is percolative rather than bulk-defect dominated (Sulthana et al., 9 Jul 2025). In amorphous ZTO TFTs, giant PPC is assigned specifically to sub-gap tail states near the conduction band, whereas deeper states produce only mild PPC; discharge-current analysis was used to separate these two contributions (Dhara et al., 2020).

3. Representative material platforms

The breadth of PPC is best appreciated by comparing the regimes in which it has been reported.

Platform	Reported PPC signature	Emphasized mechanism
SnO$\Delta R(t)$7 nanobelts (Viana et al., 2012)	$\Delta R(t)$8 s at 300 K	Surface molecular oxygen and hole recombination
Si-doped Al$\Delta R(t)$9Ga$\rho_{xx}(T)$0As (Misuraca et al., 2014)	Persistent photodoping for “weeks” at low $\rho_{xx}(T)$1	Deep DX-center donors
STO/Al heterostructures (Luo et al., 1 Dec 2025)	$\rho_{xx}(T)$2 h at 4 K	Deep-level re-trapping plus shallow gate traps
GeBiTe and GeSbTe films (Ovadyahu, 2018, Ovadyahu, 2015)	IR-induced excess conductance with $\rho_{xx}(T)$3–$\rho_{xx}(T)$4 s fits	Metastable defect kinetics in disordered chalcogenides
Monolayer MoS$\rho_{xx}(T)$5 (Bartolomeo et al., 2017, George et al., 2020)	$\rho_{xx}(T)$6 s decay; $\rho_{xx}(T)$7 days in GPPC	Interface/defect trapping; random-potential localization
Hydrogenated diamond (Sulthana et al., 9 Jul 2025)	$\rho_{xx}(T)$8 from $\rho_{xx}(T)$9 s to $T_c$0 s with oxygen termination	RLPF, percolation, adsorbate-coupled surface states
SrTiO$T_c$1 and related oxides (Zhang et al., 2019)	Resistance drop by three orders of magnitude, lasting over a year	Photoinduced instability of $T_c$2
Polar/non-polar oxide interfaces (Gennaro et al., 2013, Uccio et al., 2012)	Ultra-slow metastable photoresponse, including sub-gap excitation	Long-lived electron–hole pairs in interface electric fields
Oxygen-deficient YBCO systems (Hage et al., 2023, Kawashima et al., 2013)	Persistent decrease of $T_c$3 and increase of $T_c$4	Photodoping plus scattering-rate changes

Within this spectrum, several platforms illustrate qualitatively different roles for PPC. In Al$T_c$5Ga$T_c$6As, PPC is primarily an experimental control parameter: the channel density can be tuned in situ from insulating through $T_c$7 to the metallic side on one sample, without electrostatic gating (Misuraca et al., 2014). In STO/Al, PPC is itself the slowly relaxing non-equilibrium state that modulates the amplitude of fast gate-induced trapping, with the trap amplitudes following the same $T_c$8 h time constant as the PPC baseline (Luo et al., 1 Dec 2025). In GeBiTe and GeSbTe, PPC is prominent even in high-carrier-density chalcogenides at 4.1 K, while electron-glass behavior emerges only in sufficiently localized samples (Ovadyahu, 2018, Ovadyahu, 2015).

The MoS$T_c$9 literature shows the strongest spread in persistence. One room-temperature study on monolayer back-gated FETs reported photogating from hole traps at the MoS$_2$0/SiO$_2$1 interface and in MoS$_2$2 defects, with $_2$3 s and $_2$4 s after long illumination (Bartolomeo et al., 2017). Another reported ultraviolet-induced giant PPC in CVD monolayers, with conductivity enhanced by up to a factor of $_2$5 and a slow time constant of $_2$6 days in high vacuum, attributed mainly to intrinsic sulfur-vacancy and strain disorder (George et al., 2020).

4. Kinetics, energetics, and transport formalisms

PPC decay is rarely a single-exponential process. A canonical representation is the stretched exponential

$_2$7

used for SnO$_2$8 nanobelts, where $_2$9 increases slowly with temperature and the lifetime $1.20\times10^5$0 decreases strongly as $1.20\times10^5$1 rises (Viana et al., 2012). The same Kohlrausch form is also reported for GeBiTe and GeSbTe, with $1.20\times10^5$2 and $1.20\times10^5$3–$1.20\times10^5$4 s at 4.1 K, and for hydrogenated diamond, where room-temperature fits give $1.20\times10^5$5 s and $1.20\times10^5$6 in pristine H-terminated material (Ovadyahu, 2018, Sulthana et al., 9 Jul 2025).

Thermally activated PPC is often discussed through Arrhenius kinetics. In SnO$1.20\times10^5$7,

$1.20\times10^5$8

and the fitted activation energy $1.20\times10^5$9 meV matches the PL-derived acceptor level at $_{0.3}$0 meV, linking PPC decay to thermal ionization of holes from acceptor states (Viana et al., 2012). In hydrogenated diamond, the reported high-temperature behavior follows

$_{0.3}$1

with recombination barriers decreasing from $_{0.3}$2 meV in pristine HD to $_{0.3}$3 meV after stronger oxygen termination (Sulthana et al., 9 Jul 2025).

Other platforms require multi-component kinetics. In STO/Al, the post-illumination sheet-resistance recovery is fitted with

$_{0.3}$4

where the exponential term represents refilling of deep states and the logarithmic term captures thermally activated redistribution; $_{0.3}$5 h at 4 K (Luo et al., 1 Dec 2025). In monolayer MoS$_{0.3}$6 GPPC, the drain-current decay is described by

$_{0.3}$7

with $_{0.3}$8 day and $_{0.3}$9 days (George et al., 2020). In BP–MoS$_{0.7}$0, the persistent negative component follows

$_{0.7}$1

again emphasizing the coexistence of faster and slower relaxation channels (Jawa et al., 2021).

Several papers connect PPC to transport crossover phenomena. In hydrogenated diamond, the photocurrent build-up obeys

$_{0.7}$2

with $_{0.7}$3 K for pristine HD and $_{0.7}$4 K after 60 s ozonation, supporting a percolative interpretation of the onset of strong PPC (Sulthana et al., 9 Jul 2025). In monolayer MoS$_{0.7}$5 on OTS/SiO$_{0.7}$6, the PPC build-up level scales as $_{0.7}$7, with $_{0.7}$8 in the $_{0.7}$9–$\Delta G(t)$00 K range depending on the device, paralleling the percolation threshold extracted from transport (Ovadyahu, 2015).

5. Relation to other non-equilibrium phenomena

PPC often coexists with, but should not be conflated with, photodoping, electron-glass relaxation, photosuperconductivity, or field-effect trapping. The distinction is most explicit in oxygen-deficient YBCO. There, photodoping is measured through the Hall number $\Delta G(t)$01, whereas PPC and photosuperconductivity track changes in Hall mobility $\Delta G(t)$02 and $\Delta G(t)$03. The reported conclusion is that persistent conductivity enhancement and photosuperconductivity are linked to a photo-induced decrease of the electronic scattering rate, not to the concomitant carrier-density increase, because $\Delta G(t)$04 correlates with $\Delta G(t)$05 but not with $\Delta G(t)$06 (Hage et al., 2023).

In GeBiTe and GeSbTe, PPC and the electron-glass phase are clearly separated by disorder dependence and relaxation law. PPC is observable even for $\Delta G(t)$07 of the order of $\Delta G(t)$08 k$\Delta G(t)$09, whereas electron-glass memory dips require strongly localized films, with $\Delta G(t)$10 in the M$\Delta G(t)$11 range. PPC relaxes as a stretched exponential, while electron-glass relaxation follows a logarithmic law. Yet the two can coexist, and the memory dip can be enhanced in the PPC state, a result discussed in terms of mesoscopic compositional disorder and an increased interaction-to-disorder ratio (Ovadyahu, 2018, Ovadyahu, 2015).

In STO/Al, PPC does not merely add to gate-voltage trapping; it modulates it. The two shallow-trap amplitudes $\Delta G(t)$12 and $\Delta G(t)$13 decay with the same $\Delta G(t)$14 as the PPC baseline, which identifies deep-level occupancy as the slowly evolving background on which the faster gate-sensitive processes operate (Luo et al., 1 Dec 2025). In Al$\Delta G(t)$15Ga$\Delta G(t)$16As, PPC serves as an in situ density-control mechanism for Hanle and spin-drift-diffusion measurements, yielding spin lifetimes on the order of nanoseconds that vary across the metal–insulator transition (Misuraca et al., 2014).

A related boundary case appears in van der Waals heterostructures where the sign of photoconductance can switch. BP–MoS$\Delta G(t)$17 demonstrates visible-to-near-infrared switching between negative and positive photoconductance, with a gate-dependent negative persistent component that was explicitly connected to optosynaptic behavior (Jawa et al., 2021). This does not redefine PPC, but it shows that persistence can attach to either sign once interlayer recombination and trapping are sufficiently slow.

6. Control parameters, applications, and recurrent controversies

The dominant control parameters are repeatedly the same: temperature, defect chemistry, surface chemistry, oxygen partial pressure, gate bias, carrier density, illumination wavelength, and illumination history. In SnO$\Delta G(t)$18, both the steady photocurrent and the PPC lifetime are strongly atmosphere dependent, with $\Delta G(t)$19 and $\Delta G(t)$20 s, $\Delta G(t)$21 s, $\Delta G(t)$22 s at 300 K (Viana et al., 2012). In hydrogenated diamond, partial oxygen termination suppresses PPC by reducing the 2DHG density and lowering the recombination barrier (Sulthana et al., 9 Jul 2025). In monolayer MoS$\Delta G(t)$23, substrate engineering can suppress or amplify PPC, ranging from negligible persistence in suspended devices to slow photoresponse on SiO$\Delta G(t)$24 and GPPC in defect-rich CVD monolayers (Ovadyahu, 2015, George et al., 2020).

This tunability underlies several application domains already identified in the literature. PPC has been treated as a challenge for fast photodetectors because of long recovery times, but also as an opportunity for high-gain photodetection, optically written memory, and history-dependent switching (Bartolomeo et al., 2017, Dhara et al., 2020). In STO/Al, the coupling between PPC and gate-induced trapping was presented as an optical gating mechanism for non-volatile, optically programmable oxide electronics (Luo et al., 1 Dec 2025). In Al$\Delta G(t)$25Ga$\Delta G(t)$26As, PPC permits carrier-density tuning in situ without the electric fields associated with electrostatic gating, which is advantageous for spin-transport studies (Misuraca et al., 2014). In YBCO, persistent photoinduced changes provide a route to optically modify both normal-state conductivity and superconducting $\Delta G(t)$27 (Hage et al., 2023).

Three controversies recur. First, oxygen vacancies are not a universal explanation. In SnO$\Delta G(t)$28, the paper explicitly rejects the common vacancy-centered account in favor of surface molecular oxygen and hole kinetics (Viana et al., 2012). Second, “intrinsic” versus “extrinsic” PPC can depend on sample structure rather than on the material name. Monolayer MoS$\Delta G(t)$29 studies support both substrate-driven random localized potential fluctuations and intrinsic defect/strain-driven giant PPC, and oxide-interface work shows that the difference between transient and persistent photoresponse can hinge on growth method and interfacial charge-transfer permeability (Ovadyahu, 2015, George et al., 2020, Kawashima et al., 2013). Third, PPC is not always reducible to photodoping. Oxygen-deficient cuprates show that persistent conductivity enhancement can instead be governed primarily by a light-induced decrease of scattering rate (Hage et al., 2023).

Across these systems, the common denominator is kinetic asymmetry between photogeneration and recombination. The asymmetry may arise from metastable defect configurations, surface adsorption cycles, spatially separated carriers in an internal field, or disorder-driven distributions of trap depths and percolation barriers. The specific microscopic implementation varies, but the phenomenology remains recognizable: illumination writes a non-equilibrium electronic state whose erasure is controlled not by ordinary carrier lifetime, but by a much slower structural, electrostatic, or configurational relaxation landscape.