Skip to content

Electron Beam Irradiation-Induced Transport and Recombination in P-type Diamond Diodes

FieldValue
JournalECS Journal of Solid State Science and Technology
Published2026-07-01
AuthorsGabriel Marciaga, John Olsen, Hsiao-Hsuan Wan, Chao-Ching Chiang, F. Ren
DOI10.1149/2162-8777/ae87cd

Sustained electron beam irradiation increased the minority-carrier diffusion length in boron-doped p-type diamond from 4.27 to 4.61 µm at 40 °C and from 8.27 to 8.58 µm at 120 °C over a reported 1800 s. EBIC analysis gave a 151 meV activation energy for injection-driven lengthening and a 177 meV activation energy for temperature-driven lengthening. The authors attribute this response to electron-induced passivation of deep-level, boron-related recombination centers. Cathodoluminescence tracked the competing radiative and non-radiative pathways.

The studied device was a vertical p/p+ Schottky diode built on single-crystal (100) diamond. It had a 100 µm diameter front Schottky contact and a full-area Ti/Au back contact. The back contact thicknesses were 20 nm Ti and 80 nm Au.

The 5 µm epitaxial layer contained 1.3 × 10¹⁶ cm⁻³ boron. The 250 µm substrate contained 3 × 10²⁰ cm⁻³ boron. SIMS measured nitrogen at 8 × 10¹⁵ cm⁻³ and silicon below 10¹⁵ cm⁻³.

The layer was grown by microwave plasma-assisted chemical vapor deposition using CH₄, B₂H₆, and high-purity H₂. Growth occurred between 800 and 850 °C, with 5 kW microwave power and 100 Torr chamber pressure. The deposition rate was 1 to 2 µm h⁻¹. FTIR confirmed the final thickness after polishing, and AFM measured Ra = 0.33 nm. A 30 min UV/ozone treatment produced a partially oxygen-terminated front surface. XRD showed a rocking-curve FWHM below 30 arcsec and an estimated dislocation density of 8 × 10⁶ cm⁻².

EBIC measurement and surface-recombination fit

Section titled “EBIC measurement and surface-recombination fit”

The EBIC setup used a temperature-controlled stage in a scanning electron microscope. Planar scans ran outward from the Schottky contact edge at 20 kV. The filament current was 60 µA, while absorbed current was at the nA scale. Each line scan lasted 12 s with a 66 ms dwell time.

The fit used the far field, where the lateral distance was much greater than the generation radius. The authors set α to −1/2 for negligible surface recombination. This choice was supported by the chemically polished, oxygen- and hydrogen-terminated (100) surface. The agreement between EBIC and CL activation energies supported the assumption because CL did not use the junction collection geometry.

The pre-irradiation EBIC signal decayed exponentially from the contact edge. The extracted diffusion length was 4.27 µm at 40 °C. It increased monotonically between 40 °C and 120 °C. This trend differs from Ga₂O₃ and GaN, where the paper reports decreasing diffusion length as temperature rises.

The authors modeled the temperature response with an Arrhenius relation. The fit gave a 177 meV activation energy. They attribute the increase to thermal ionization of deep traps or a temperature-dependent reduction in capture cross-section. Both processes can raise minority-carrier lifetime. The fitted value was consistent with boron acceptor ionization at higher dopant concentrations.

During sustained irradiation, the authors extracted the diffusion length every 100 s at five temperatures from 40 °C to 120 °C. The diffusion length rose near-linearly at every temperature. The authors interpret this passivation as capture of beam-generated electrons by defect states, which reduces the active recombination-center density.

The diffusion-length change rate decreased with temperature because thermal excitation competed with continued passivation. An Arrhenius analysis isolated a 151 meV activation energy for injection-driven lengthening. The close agreement with the 177 meV thermal value led the authors to assign both processes to the same boron-related defect population. The reported lengthening was less pronounced than in the other ultrawide-bandgap materials cited by the paper.

Reporting is inconsistent for the EBIC exposure. The experimental section gives 1200 s per temperature, while the EBIC results describe a 1800 s increase.

Temperature-resolved CL covered −193 °C to 100 °C at 20 kV. The A band appeared at 437 nm and persisted through 50 °C. It was nearly quenched by 100 °C. The retained line shape suggested that the emitting center remained intact while phonon-mediated processes reduced its radiative efficiency. A thermal-quenching fit gave 112 meV. The authors associated this energy with hole ionization from boron-related states or thermally activated escape from the emitting defect complex.

Under sustained irradiation, A-band intensity decreased monotonically at every measured temperature. Normalized inverse square-root intensity was linear with exposure time, consistent with intensity decreasing as the inverse square of exposure time. The decay rate increased from 40 °C to 100 °C, unlike the cited Ga₂O₃ and GaN behavior.

Arrhenius analysis found two regimes separated near −50 °C. The high-temperature regime gave 110 meV, consistent with the 112 meV thermal-quenching value. The low-temperature regime gave 11 meV and indicated a separate trap-dynamics mechanism. The high-temperature agreement supports the same boron-related non-radiative defect for both processes.

The lifetime calculation used the Einstein relation and assumed minority-electron mobility in p-type diamond equal to majority-electron mobility in n-type diamond. The assumed value was 1000 cm² V⁻¹ s⁻¹. This gave a diffusivity of 27 to 34 cm² s⁻¹ over the measured temperature range. The corresponding lifetimes were about 7 ns at 40 °C and 20 ns at 120 °C. Using a literature mobility of 422 cm² V⁻¹ s⁻¹ gave 11 ns and 47 ns at the same temperatures. These diffusivities and lifetimes were calculated rather than measured directly.

The baseline 4.27 µm diffusion length was consistent with reported lifetimes in CVD-grown diamond thin films. The calculated diffusion lengths and nanosecond lifetimes were comparable with other quality diamond samples. The paper cites a deep boron ionization energy of about 0.37 eV. It treats the 177 meV fitted value as consistent with concentration effects.

Other cited device benchmarks include a Ga₂O₃ p-n heterojunction with breakdown above 3 kV. The paper also cites an n-type monolayer MoS₂ and p-type diamond junction with 4000 A cm⁻² forward current density. A vertical diamond Schottky rectifier reached a PFOM of 30.5 MV cm⁻². A phosphorus-doped p-i-n diamond diode reached a breakdown field of 1.25 MV cm⁻¹. These are device-level reference points rather than direct comparisons with the measured diffusion lengths.

The transport results come from one p/p+ diode, so device-to-device variation was not tested. The EBIC fit assumes negligible surface recombination through α = −1/2, although CL provides a separate consistency check. The lifetime values remain estimates because minority-electron mobility was taken from n-type diamond and was not measured in this device.

The experiments infer boron-related defect control from activation energies and spectral behavior. They do not identify a specific defect species. The activation-energy match therefore supports the authors’ assignment rather than proving the microscopic structure of the trap.

Device selection should treat EBIC diffusion length and CL emission as separate criteria. Heating increased carrier collection distance but also suppressed A-band luminescence. A p-type bipolar device can therefore show longer carrier collection distance with weaker defect luminescence. Operating temperature and electron-beam history should be treated as material-state variables during selection because sustained irradiation changed both transport and recombination.

For a p-type Schottky diode application, the paper uses a single-crystal (100) boron-doped MPCVD epitaxial layer about 5 µm thick with Ti/Au metallization. 6CCVD’s diamond materials catalog lists boron-doped single-crystal diamond grown by MPCVD, thickness from 0.1 to 500 µm, Ti and Au among its metallization options, and single-crystal polishing below Ra 1 nm. The engineering support service can specify growth, polishing, and metallization requirements for a comparable device, including the study’s 5 µm layer, Ra = 0.33 nm surface, and Ti/Au contact.

View Original Abstract

This study investigates minority carrier diffusion length ( L ) and defect-mediated recombination phenomena in boron-doped p-type diamond epitaxial layers. Utilizing temperature-dependent electron beam-induced current (EBIC) and cathodoluminescence (CL), the influence of temperature and sustained electron beam irradiation was characterized. The diamond sample demonstrates a pronounced thermal enhancement to the minority carrier diffusion length. This temperature enhancement for L has an activation energy of 177 meV. Furthermore, sustained electron injection induces a near-linear elongation of the diffusion length through passivation of deep-level defects. From the EBIC injection data, an activation energy of 151 meV was extracted from fitting. Temperature-resolved CL reveals the quenching of the characteristic A-band emission at 100 °C. Additionally, a thermal activation energy of 112 meV was extracted via a decay fit of the temperature dependent CL intensity data. While sustained electron beam irradiation reveals dose-dependent quenching of the A-band emission with thermally activated decay rates. This yields activation energies of 110 meV and 11 meV associated with distinct defect generation mechanisms at high and low temperatures, respectively. This provides a complementary assessment of the defect landscape governing minority carrier lifetimes in these boron-doped diamond diodes.