Diamond Schottky PIN Diodes (SPIND)

• Diamond Schottky PIN Diodes (SPIND) are diamond semiconductor devices that feature Schottky contacts coupled with p–i–n-like injection regions for controlled carrier recombination.
• They are engineered in varied configurations—such as pseudo-vertical p⁺⁺/i/n stacks and metal/n-diamond structures—leveraging hydrogen passivation and defect management.
• SPINDs enable promising applications in extreme-environment power electronics and quantum optoelectronics through precise barrier engineering and advanced transport control.

Searching arXiv for recent and relevant papers on diamond Schottky PIN diodes and closely related diamond Schottky/p-i-n device physics. A diamond Schottky PIN diode (SPIND) denotes a class of diamond devices in which a Schottky contact supplies the rectifying function while the internal electrostatics and carrier transport either incorporate a physical p–i–n stack or emulate p–i–n-like injection and recombination behavior. In the recent literature, this designation spans at least two closely related realizations: a pseudo-vertical p$^{++}$/i/n diamond diode with a Schottky cathode developed for extreme-environment power electronics, and a metal/n-diamond Schottky structure on hydrogen-passivated phosphorus-doped diamond that forms, under forward bias, a high-field region containing both electrons and holes and can electrically excite color centers. Taken together, these studies position SPINDs at the intersection of diamond power electronics, quantum optoelectronics, and defect-engineered device physics (Sledz et al., 2024, Surdi et al., 1 Jul 2026).

1. Nomenclature and device class

The term “Schottky PIN diode” is not used uniformly across the cited works. In one usage, the device is structurally a conventional p–i–n stack in which the rectifying contact is Schottky-like on the n-type top layer; in another, the device is formally metal–n-diamond, but its forward-bias band bending creates a depleted injection zone that behaves functionally like the intrinsic region of a p–i–n diode (Sledz et al., 2024, Surdi et al., 1 Jul 2026, Kaintz et al., 6 May 2026).

Work	Physical structure	SPIND interpretation
"Electrical excitation of color centers in phosphorus-doped diamond Schottky diodes" (Sledz et al., 2024)	Au/H-terminated n-type phosphorus-doped diamond with lateral ohmic ring	Schottky-on-n device with an effective PIN-like injection region
"Diamond Diode for Extreme Venus Environments" (Surdi et al., 1 Jul 2026)	p$^{++}$ substrate / i-layer / n-layer with Schottky cathode	Hybrid p–i–n stack whose current-limiting junction is Schottky-like
"From Defects to Devices: Design Guidelines for High-Performance Diamond-Based Solar Cells and Single-Dopant Diodes" (Kaintz et al., 6 May 2026)	p–i–n or p–n defect-engineered stacks with one Schottky side as an adaptation	Generalized hybrid of a p–i–n junction and a Schottky interface

In the pseudo-vertical power-diode realization, the device is explicitly described as a diamond p–i–n diode in which the rectifying contact is a Schottky contact on the n-type diamond layer. In the color-center realization, the n-type homoepitaxial layer and high-work-function metal contact create, under forward bias, a depletion region that acts like an “i-region,” while the metal injects minority holes and the n-diamond supplies electrons. This suggests that “SPIND” is presently best understood as a functional descriptor for diamond devices that combine Schottky boundary conditions with p–i–n-like carrier injection, rather than as a single fully standardized topology.

2. Materials systems, layer stacks, and interfaces

One experimentally demonstrated SPIND-like architecture is fabricated on a single-crystal HPHT (111) diamond substrate with a homoepitaxial phosphorus-doped n-type layer grown by MW-PECVD. The n-layer thickness is approximately $2.2~\mu\text{m}$, the donor concentration is $N_D = 10^{18}~\text{cm}^{-3}$, and the compensation ratio is about $10\%$. Color centers are introduced into this layer: SiV centers by Si implantation into the top $\sim 200$ nm followed by annealing at $1200^\circ\text{C}$, NV centers from vacancies plus nitrogen during the same anneal, and H$_3$ centers in the nitrogen-rich HPHT substrate and possibly the overlayer. The ohmic contact is Ti/Au, formed together with a high-temperature hydrogen plasma treatment that anneals Ti to TiC and hydrogen-terminates the diamond surface; the Schottky contact is Au directly on hydrogen-terminated diamond, patterned as circular pads of diameter $\approx 70~\mu\text{m}$ בתוך a lateral geometry (Sledz et al., 2024).

Hydrogen passivation is central in that device. It produces a negative electron affinity surface, with $\chi_e$ as low as $^{++}$0, suppresses metal-induced surface states and Fermi-level pinning, and thereby enables a large electron barrier and a comparatively smaller hole barrier. From temperature-dependent current-voltage analysis, the electron Schottky barrier height is extracted as $^{++}$1. With diamond’s bandgap $^{++}$2, this corresponds to a hole barrier $^{++}$3. Simulations in the same work explore even lower $^{++}$4 values, down to $^{++}$5–$^{++}$6, as an optimization target for efficient hole injection (Sledz et al., 2024).

A second experimentally demonstrated SPIND architecture is a pseudo-vertical p–i–n device grown on heavily boron-doped p-type HPHT diamond. Its layer sequence is a p$^{++}$7 substrate of thickness $^{++}$8 and doping $^{++}$9, a nominally intrinsic $2.2~\mu\text{m}$0 PECVD diamond i-layer with background doping $2.2~\mu\text{m}$1, and a phosphorus-doped n-layer of thickness $2.2~\mu\text{m}$2 and doping $2.2~\mu\text{m}$3. The cathode is Ti/Pt/Au on the n-layer and forms the Schottky contact; the anode, also Ti/Pt/Au, is formed on the nearly metallic p$2.2~\mu\text{m}$4 substrate by etched trenches adjacent to the mesa. The nominal diode mesa diameter is $2.2~\mu\text{m}$5, the cathode diameter is $2.2~\mu\text{m}$6, and the anode is offset laterally by $2.2~\mu\text{m}$7, so current flows mainly vertically through the p–i–n stack and then laterally in the substrate (Surdi et al., 1 Jul 2026).

Related vertical diamond Schottky diodes without an n-layer show how strongly surface morphology and substrate quality affect Schottky performance. In those devices, a highly boron-doped $2.2~\mu\text{m}$8 HPHT substrate supports a homoepitaxial lightly doped p-type drift layer of thickness $2.2~\mu\text{m}$9–$N_D = 10^{18}~\text{cm}^{-3}$0, with top Ti/Pt/Au Schottky contact and backside ohmic Ti/Pt/Au contact. Smooth as-grown, hillock-rich, and polished surfaces show distinct barrier homogeneity, leakage, and breakdown behavior, establishing interface morphology as a first-order design variable for diamond Schottky structures (Reinke et al., 2020).

3. Transport physics and PIN-like behavior

The unifying transport picture of SPINDs is that the Schottky contact controls injection while the diamond bulk sustains strong space-charge fields and forms the active region. For an n-type Schottky diode, the depletion width on the semiconductor side obeys

$N_D = 10^{18}~\text{cm}^{-3}$1

so a high-doping n-layer can have a narrow but intense depletion region. In the Au/H-terminated n-diamond device, forward bias lowers the effective barrier for carrier exchange, while hole injection from the metal over the smaller hole barrier and electron supply from the n-diamond bulk produce a region beneath the Schottky contact where both carrier species coexist at appreciable density. That depleted, high-field zone acts as a quasi-intrinsic injection region; in functional terms, the metal behaves as an effective p-side reservoir, the phosphorus-doped diamond behaves as the n-region, and recombination occurs in the intervening depleted zone (Sledz et al., 2024).

For the pseudo-vertical p$N_D = 10^{18}~\text{cm}^{-3}$2/i/n SPIND, the transport description explicitly combines thermionic emission at the Schottky cathode with space-charge-limited current in the low-doped interior. The Schottky contribution is modeled with the Richardson equation

$N_D = 10^{18}~\text{cm}^{-3}$3

while the high-injection limit is described by the trap-free Mott–Gurney form

$N_D = 10^{18}~\text{cm}^{-3}$4

The analytical picture is that low forward bias is contact-limited, whereas higher bias progressively fills the i-layer with injected carriers until the current approaches an SCLC regime modified by traps, contact resistance, and real-device nonidealities (Surdi et al., 1 Jul 2026).

The broader diamond p–i–n literature adds a second essential concept: superinjection. In all-ohmic diamond homojunction p–i–n diodes, the deep donor level of phosphorus implies that the equilibrium free-electron density in the n-region can be as low as $N_D = 10^{18}~\text{cm}^{-3}$5 even when $N_D = 10^{18}~\text{cm}^{-3}$6. Under forward bias, a self-consistent potential well can form in the i-region near the p–i junction, allowing the electron density in the active region to exceed the n-side equilibrium density by up to four orders of magnitude. One study further shows that replacing an ohmic n-contact with a Schottky contact of $N_D = 10^{18}~\text{cm}^{-3}$7 or $N_D = 10^{18}~\text{cm}^{-3}$8 leaves the dependence of maximum single-photon electroluminescence rate on current density essentially unchanged; the main penalty is added voltage drop rather than loss of the internal superinjection mechanism (Khramtsov et al., 2018, Khramtsov et al., 2018).

Taken together, these results identify two related SPIND operating modes. One is barrier-engineered minority injection at a Schottky/n-diamond interface; the other is Schottky-assisted access to bulk p–i–n superinjection physics. A plausible implication is that practical SPIND optimization depends less on the presence or absence of a formal p-layer than on the ability to shape the internal field profile so that high local $N_D = 10^{18}~\text{cm}^{-3}$9 and $10\%$0 overlap in the intended active volume.

4. Color-center electroluminescence and single-photon operation

The most developed quantum-optical SPIND-like realization is the Au/H-terminated phosphorus-doped diamond Schottky device that electrically excites color centers at ambient conditions and up to $10\%$1. Electroluminescence turns on at approximately $10\%$2, spatially localizes beneath the Schottky contact, and decays radially into the n-layer, consistent with a minority-hole density that decreases with distance from the contact. At $10\%$3 and $10\%$4, peak counts of $10\%$5 in $10\%$6 integration were reported. The dominant spectral features are NV$10\%$7 emission with ZPL at $10\%$8 nm and a phonon sideband centered near $10\%$9 nm, H$\sim 200$0 emission around $\sim 200$1 nm, and broad Band-A emission around $\sim 200$2 nm. SiV ZPL at $\sim 200$3–$\sim 200$4 nm is weak in the main device region and more visible only in more heavily implanted regions or earlier device generations, where it remains small relative to NV and H$\sim 200$5 emission (Sledz et al., 2024).

The same work reports strong temperature dependence. At $\sim 200$6, electroluminescence is dominated by H$\sim 200$7, with NV$\sim 200$8 hardly visible. At $\sim 200$9, the NV$1200^\circ\text{C}$0 ZPL becomes visible; between $1200^\circ\text{C}$1 and $1200^\circ\text{C}$2, NV$1200^\circ\text{C}$3 increases strongly with temperature, tracking the increase in free-electron density as phosphorus donors ionize more effectively. Conductance measurements yield an activation energy $1200^\circ\text{C}$4, consistent with phosphorus donors. The extracted Schottky barrier and the temperature scaling of the spectrum support the interpretation that, in the demonstrated device, excitation of NV$1200^\circ\text{C}$5 is electron-capture limited rather than limited by the internal radiative lifetime (Sledz et al., 2024).

The electroluminescence rate of an electrically driven color center is modeled in the related p–i–n superinjection literature by a capture-limited expression,

$1200^\circ\text{C}$6

where $1200^\circ\text{C}$7 and $1200^\circ\text{C}$8 are the local carrier densities, $1200^\circ\text{C}$9 and $_3$0 are the electron and hole capture rate constants, $_3$1 is the radiative lifetime, and $_3$2 is the radiative quantum efficiency. The cited works use $_3$3 in one case and $_3$4 in another, with $_3$5 or $_3$6. Because hole capture is faster, electron capture is often the bottleneck in diamond (Khramtsov et al., 2018, Sledz et al., 2024).

This rate model explains why barrier engineering is decisive. In the Schottky/n-diamond simulations, a single emitter with $_3$7 at $_3$8 is predicted to reach $_3$9 for $\approx 70~\mu\text{m}$0 and only $\approx 70~\mu\text{m}$1 for $\approx 70~\mu\text{m}$2. In the all-ohmic p–i–n superinjection case, the same capture-limited physics leads to much larger predicted rates: $\approx 70~\mu\text{m}$3 per center at room temperature, and $\approx 70~\mu\text{m}$4 for a $\approx 70~\mu\text{m}$5 i-region with the center positioned about $\approx 70~\mu\text{m}$6 from the p–i junction at $\approx 70~\mu\text{m}$7 (Sledz et al., 2024, Khramtsov et al., 2018).

A common misconception is that electrical color-center emission in a diamond Schottky device is already equivalent to a demonstrated single-photon source. The Schottky color-center study explicitly does not report external quantum efficiency, internal quantum efficiency, or $\approx 70~\mu\text{m}$8. It therefore demonstrates electrical excitation of color centers, not antibunched single-photon operation as a measured device characteristic (Sledz et al., 2024).

5. Power rectification, surface morphology, and extreme-environment operation

The power-electronics branch of SPIND research is represented by the pseudo-vertical p$\approx 70~\mu\text{m}$9/i/n Schottky PIN diode operated from $\chi_e$0 to $\chi_e$1 and in a simulated Venus atmosphere. That device reaches a maximum current density of $\chi_e$2, carries a total current of $\chi_e$3 through a $\chi_e$4-wide pseudo-vertical structure, and exhibits a maximum power handling capacity of $\chi_e$5. At a forward bias of $\chi_e$6, the specific on-resistance is $\chi_e$7, and the current on-off ratio is $\chi_e$8. The turn-on voltage decreases from $\chi_e$9 at $^{++}$00 to $^{++}$01 at $^{++}$02, while the ideality factor changes from $^{++}$03 at room temperature to $^{++}$04 at $^{++}$05, with a dip to $^{++}$06 near $^{++}$07 that is interpreted as contact annealing (Surdi et al., 1 Jul 2026).

The same device was operated for 15 days in Venus-like conditions in the Glenn Extreme Environments Rig. Under continuous reverse-bias stress, reverse current increased with temperature, stabilized at about $^{++}$08 for the specified bias condition, and returned to noise level as the chamber cooled. I–V measurements during idle, ramp-up, Venus atmosphere, and ramp-down showed that the device remained functional and retained an on/off ratio of $^{++}$09 during the GEER run. Post-test XPS indicated trace sulfur incorporation near Ni contact pads, minor increases in oxidized carbon species, and reacted Ni-containing surface species, but no electrical degradation; the reported behavior slightly improved, likely because of contact annealing (Surdi et al., 1 Jul 2026).

Surface morphology studies on high-voltage vertical diamond Schottky diodes clarify why SPIND performance is so sensitive to interface preparation and substrate quality. Smooth as-grown devices show $^{++}$10 up to $^{++}$11, breakdown at $^{++}$12, and ideality factor $^{++}$13. Polished samples show similar breakdown voltage and reverse current density, with ideality factor $^{++}$14. Hillock-rich devices block similar voltages but show $^{++}$15, a $^{++}$16 reduction in calculated 1D breakdown field, and a secondary Schottky barrier that can be fitted with a modified thermionic-emission model employing the Lambert W-function. Reported Baliga figures of merit are $^{++}$17, $^{++}$18, and $^{++}$19 for hillock-rich, smooth as-grown, and polished samples, respectively (Reinke et al., 2020).

These results also correct a second common simplification: polishing alone does not remove the influence of “killer defects.” The polished sample demonstrates per-device performance comparable to the smooth sample, but the statistical analysis shows that high defect density still reduces the feasible device area and increases the probability of elevated leakage. For SPINDs intended for multi-kV blocking or large-area integration, substrate and epitaxial defect density remain as important as the Schottky metal stack itself (Reinke et al., 2020).

6. Modeling frameworks, defect-engineered extensions, and unresolved design issues

The experimentally oriented SPIND studies rely on self-consistent electrostatic and transport simulation. For the Schottky color-center diode, Silvaco TCAD solves Poisson, drift–diffusion, and continuity equations with Schottky boundary conditions defined by $^{++}$20 and $^{++}$21, using $^{++}$22, $^{++}$23, and effective mobilities $^{++}$24 for the heavily doped $^{++}$25 film. For the extreme-environment SPIND, Silvaco ATLAS is combined with an analytical model including thermionic emission, SCLC, multiple single trap levels, and other physical models emulating a real device. In both cases, simulation identifies barrier engineering, defect density, and contact resistance as decisive parameters (Sledz et al., 2024, Surdi et al., 1 Jul 2026).

First-principles work extends the SPIND design space into defect-engineered diamond. One proposed route uses a boron–vacancy–boron intermediate-band absorber in a p–i–n stack and a phosphorus–vacancy impurity-band conductor for degenerate p-type functionality. Practical guidelines from that study include using graded junctions to mitigate tunnelling losses at abrupt interfaces, targeting an absorber thickness of $^{++}$26, aligning incident light in the $^{++}$27-plane to exploit anisotropic absorption, and leveraging the high transparency of both contact layers for bifacial device configurations. The same work describes a SPIND structurally as a conventional p–i–n diode with one Schottky contact and proposes that PV-doped diamond can provide high conductivity at room temperature through impurity-band transport while BVB-doped diamond preserves high carrier mobility and thermal conductivity (Kaintz et al., 6 May 2026).

Across the cited literature, the main optimization levers recur. For quantum-light SPINDs, they are lower $^{++}$28, improved hydrogen termination or alternative surface terminations, higher-work-function metals or surface-dipole engineering, reduced lateral separation between Schottky and ohmic contacts to $^{++}$29 or less, low-nitrogen electronic-grade substrates or buffer layers, controlled placement of single emitters $^{++}$30–$^{++}$31 below the surface, and optical-extraction structures such as planar Yagi–Uda nano-antennas. For power SPINDs, the dominant levers are reduced trap density, reduced contact resistance, improved Schottky homogeneity, and geometry choices that move operation toward the Mott–Gurney SCLC limit (Sledz et al., 2024, Surdi et al., 1 Jul 2026).

Several objective limitations remain. Efficient NV$^{++}$32 electroluminescence in the demonstrated Schottky color-center diode requires elevated temperature up to $^{++}$33; the reported electron barrier remains pinned near $^{++}$34–$^{++}$35, implying incomplete suppression of interface states; nitrogen-rich substrates generate strong H$^{++}$36 background; and single-photon metrics are not yet reported. In the power-diode branch, the record current density and low $^{++}$37 do not yet establish the ultimate limit, since the analytical and TCAD studies indicate that lower turn-on voltage and lower $^{++}$38 should be achievable by reducing defects and contact resistance (Sledz et al., 2024, Surdi et al., 1 Jul 2026).

The present state of the field therefore supports a precise synthesis. Diamond SPINDs are not a single device but a family of Schottky-controlled diamond structures in which the internal active region behaves as, or physically is, a p–i–n system. Their demonstrated capabilities already include color-center electroluminescence, record diamond-diode current density, multi-kV Schottky blocking, and long-duration operation in Venus-analog conditions. Their remaining bottlenecks are equally clear: interface-state control, defect density, contact resistance, and the translation from ensemble electroluminescence or high-current rectification to fully validated, room-temperature, electrically driven single-photon and large-area high-voltage platforms.