Performance enhancement of diamond p+–i–n+ diodes employing electron injection through heavily nitrogen–doped layers
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Diamond and Related Materials |
| Published | 2026-02-06 |
| Authors | Taichi Miyazaki, Kan Hayashi, Kazuki Kobayashi, Kimiyoshi Ichikawa, Taro Yoshikawa |
| DOI | 10.1016/j.diamond.2026.113416 |
| Citations | 1 |
The optimized nitrogen-doped diamond pin diode increased forward current density by about five orders of magnitude over the authors’ earlier nitrogen-doped devices. Deep-ultraviolet emission confirmed electron injection through the heavily nitrogen-doped n+ layer.
Device design and rationale
Section titled “Device design and rationale”Both devices used high-pressure, high-temperature diamond substrates measuring 2 mm × 2 mm × 0.3 mm with (111) orientation. The p+ p i n+ device added a heavily boron-doped p+ layer to the back of a high-resistivity p-type IIb (111) substrate. An intrinsic layer and a heavily nitrogen-doped n+ layer were grown on the front. The p+ i n+ device used a heavily boron-doped IIb (111) substrate as its p+ region. It placed its intrinsic and n+ layers on the front surface.
The authors state that thinning the intrinsic layer promotes conductivity modulation and suppresses its voltage drop. Thinning the n+ layer was intended to raise the electric field at the metal and n+ interface. That design sought lower contact resistance while retaining electron injection across the junction. Electron injection was examined through electroluminescence and current-voltage measurements.
The electroluminescence device was designed to verify electron injection through excitonic emission. The current device was designed to evaluate high-current operation. Device development targets reliable operation in high-temperature and high-radiation environments, including nuclear and space applications.
Growth conditions and material characterization
Section titled “Growth conditions and material characterization”All diamond layers were deposited by microwave plasma chemical vapor deposition in separate p+, intrinsic, and n+ chambers. Separate chambers were used to avoid impurity contamination from chamber history.
| Layer | Gas feed and ratios | Pressure | Microwave power | Temperature by structure |
|---|---|---|---|---|
| p+ | TMB, CH₄, H₂; [CH₄]/[H₂] = 0.24%; B/C = 1.5% | 30 kPa | 1050 W | 725 to 828 °C for p+ p i n+ |
| i | CH₄, H₂; [CH₄]/[H₂] = 0.50% | 30 kPa | 1300 W | 750 to 844 °C for p+ p i n+; 872 to 957 °C for p+ i n+ |
| n+ | N₂, CH₄, H₂; [N₂]/[CH₄] = 500%; [CH₄]/[H₂] = 0.20% | 30 kPa | 800 W | 653 to 730 °C for p+ p i n+; 689 to 752 °C for p+ i n+ |
The p+ layer was approximately 200 nm thick. The n+ layer grew for 15 minutes. The authors selected (111) instead of (100) to reduce the growth rate. The high nitrogen-to-methane ratio also produced an etching effect.
SIMS with Cs+ ions accelerated at 1.5 keV measured nitrogen at 3.8 × 10²⁰ cm⁻³. The same measurement gave an n+ thickness of approximately 100 nm. Roughness changed the secondary-ion yield, so the thin intrinsic layer could not be quantified reliably by SIMS.
Electron injection and deep-ultraviolet emission
Section titled “Electron injection and deep-ultraviolet emission”After growth, oxygen termination used a 1:3 HNO₃:H₂SO₄ mixture heated to 200 °C. Vacuum evaporation formed a 10 nm Ti layer under a 100 nm Au layer. Circular n+ electrodes ranged from 50 to 200 μm in diameter. The p-side electrode area was 1 mm².
The p+ p i n+ device was measured at room temperature by applying 200 V without an integrating sphere. Its spectrum contained peaks at 235 nm and 242 nm. The 235 nm peak corresponded to 5.27 eV and was assigned to free-exciton radiative recombination assisted by a TO phonon. The 242 nm peak was assigned to TO plus OΓ phonon-assisted recombination.
These spectra confirmed electron injection and bipolar operation through the heavily nitrogen-doped n+ layer. The paper reports emission peaks but not emission intensity or external quantum efficiency.
Current-voltage performance
Section titled “Current-voltage performance”The current device showed a rectification ratio of approximately ten orders of magnitude. The EL device showed approximately six orders of magnitude. The difference mainly reflected substrate resistivity. The current device used a substrate with approximately 2 × 10²⁰ cm⁻³ boron. The EL device used a p-type substrate with approximately 1 × 10¹⁷ cm⁻³ boron.
Forward current-voltage behavior in the nitrogen-doped device was comparable to the phosphorus-doped reference. The phosphorus comparison curve came from a cited reference rather than a device grown in this study. The nitrogen concentration was 3.8 × 10²⁰ cm⁻³, versus approximately 1 × 10²⁰ cm⁻³ phosphorus in the reference.
The extracted ideality factors were 1.39 for nitrogen and 1.32 for phosphorus. The authors attributed an approximately 1 V turn-on difference to a built-in potential difference arising from donor-level depths. The paper reports turn-on behavior around 3 V.
Operation on the high-resistivity p-type substrate suggests possible detector use, but the paper measured no detector response.
Temperature-dependent differential resistance
Section titled “Temperature-dependent differential resistance”A semiconductor parameter analyzer collected the current-voltage and differential-resistance data. The current device was measured from 25 to 420 °C in a vacuum chamber. Measurements began after pressure fell below 1 × 10⁻⁴ Pa.
Below 3 V, off-state leakage dominated the apparent differential resistance. Between 3 and 5 V, a voltage-dependent component appeared with weak temperature dependence. The authors considered field-emission-related carrier injection likely in this regime.
Above 5 V, the current became limited by a bulk resistive component. At around 6 V, the effective n-type contact resistance fell below the 10⁻¹ Ω cm² level. The authors state that bulk resistivity then limited further contact-resistance reduction.
The high-bias differential resistance had weak temperature dependence over the measured range. The authors expect the thinned intrinsic layer to contribute little voltage drop once the diode is fully on. An Arrhenius fit at 10 V gave an apparent activation energy of approximately 16 meV. This value is compatible with hopping conduction discussed for boron-doped diamond. It does not uniquely identify the limiting series region, and the n+ layer may also contribute.
Unresolved limits and device implications
Section titled “Unresolved limits and device implications”The main material limitation is the deep nitrogen donor level. Substitutional nitrogen lies about 1.7 eV below the conduction-band minimum, which suppresses room-temperature electron activation. Phosphorus has a reported activation energy of about 0.6 eV.
The introduction cites n-type conductivity of 9.8 × 10⁻² Ω⁻¹ cm⁻¹. Heavily p-type diamond has a cited conductivity on the order of 10² Ω⁻¹ cm⁻¹. The introduction presents this contrast as a major asymmetry in room-temperature conduction.
Prior work cited an n-type specific contact resistance of 0.9 Ω cm² near zero applied voltage. A cited p-type value was 2.8 × 10⁻⁷ Ω cm². These values came from separate studies and do not measure this device directly.
The paper does not isolate the contact contribution from every bulk contribution in the doped regions. The weak temperature dependence therefore cannot assign the full series resistance to the n+ layer. Lowering bulk resistivity is the authors’ stated route to further contact-resistance reduction.
The numerical intrinsic-layer thickness was not reported, despite its stated role in conductivity modulation. The paper states that the manuscript was a preprint and had not been peer reviewed. Further optimization is needed for higher-current bipolar operation with nitrogen-doped electron-injection layers.
At the reported concentration and thickness, nitrogen is a viable electron-injection dopant in diamond. The contact-resistance value applies under high forward bias and should not be treated as a zero-bias value. For p+ contacts, the paper uses heavily boron-doped diamond, which has the much lower cited specific contact resistance.
Materials and support
Section titled “Materials and support”For the p+ p i n+ device, the paper requires a heavily boron-doped MPCVD single-crystal layer and Ti/Au metallization for both contacts. The 6CCVD diamond materials catalog lists MPCVD-grown single-crystal diamond, boron-doped diamond, and Au and Ti metallization, which match those stated layer and contact requirements. The catalog facts do not cover nitrogen-doped diamond, so the full n+ layer requirement needs custom fabrication support through engineering support.
View Original Abstract
Nitrogen (N) is an attractive dopant candidate for n-type diamond because it can be incorporated at higher concentrations than phosphorus, yet the practical use of N-doped n-type diamond has been hindered by poor room-temperature conductivity and large resistive losses. To improve the conductivity in diamond pin diodes with heavily N-doped n + -type layer, we fabricated two device structures with intentionally thinned intrinsic and n + -type layers: a p + -p-i-n + diode for excitonic electroluminescence (EL) and a p + -i-n + diode for high-current evaluation. The heavily N-doped n + -type layer was precisely controlled to ∼100 nm, enhancing the electric field at the metal / n + -type layer interface to facilitate effective reduction of the contact resistance. Excitonic deep-ultraviolet EL measurements confirmed electron injection and bipolar operation, exhibiting free-exciton-related emission peaks at 235 and 242 nm and a turn-on behavior around 3 V. The optimized p + -i-n + diode showed forward J - V characteristics comparable to pin diodes with phosphorus-doped n + -type layer and achieved an approximately five-orders-of-magnitude increase in forward current density compared with previously reported pin diodes with N-doped layer. Differential-resistance analysis revealed that the effective n-type contact resistance decreased to ∼10 −1 Ω·cm 2 at ∼6 V, highlighting N as a practical dopant for forming an electron-injection layer in diamond and motivating further optimization toward higher-current bipolar operation.
Boron Doped Single Crystal Diamond Plate
View Specs →
Boron Doped Diamond On Titanium Bdd Electrode
View Specs →
Boron Doped Diamond Electrode On Niobium Bdd Nb Electrode
View Specs →References we cover
Section titled “References we cover”Original source
Section titled “Original source”References
Section titled “References”- 2008 - Diamond as an electronic material [Crossref]
- 2006 - Silicon carbide and diamond for high temperature device applications [Crossref]
- 2012 - High temperature application of diamond power device [Crossref]
- 2020 - Electronic properties of a synthetic single-crystal diamond exposed to high temperature and high radiation [Crossref]
- 2017 - Characterization of X-ray radiation hardness of diamond Schottky barrier diode and metal-semiconductor field-effect transistor
- 2021 - Radiation hardened H-diamond MOSFET (RADDFET) operating after 1 MGy irradiation [Crossref]
- 2021 - Initial evaluation of a multimodal diamond sensor for simultaneous temperature and radiation measurements [Crossref]
- 2019 - Heat and radiation resistances of diamond semiconductor in gamma-ray detection [Crossref]
- 2024 - High temperature X-ray and γ-ray spectroscopy with a diamond detector [Crossref]
- 2003 - Diamond detectors for space applications [Crossref]