Skip to content

Self‐Powered Diamond Homojunction Photodetectors for Stable Ultraviolet Sensing Under High Temperatures

FieldValue
JournalElectron
Published2026-05-16
AuthorsRunlei Zhang, Xiaohua Chen, Pengyang Zhang, Saibin Han, Yingnan Wang
DOI10.1002/elt2.70049

An unintentionally doped and heavily boron-doped diamond homojunction maintained self-powered ultraviolet detection at 250 °C. Under 213 nm illumination at 0 V, it produced a 1.8 × 10^9 rectification ratio and a 273 photocurrent-to-dark-current ratio. Responsivity was 0.37 mA/W, and specific detectivity was 1.7 × 10^12 Jones. Response times were 3.45 ms for rise and 2.68 ms for decay. After 110 on-off cycles at both 25 °C and 250 °C, the photocurrent showed no reported degradation. Single-pixel images remained distinct at 250 °C.

Here, self-powered operation means that the detector produced a photocurrent at 0 V bias. At 213 nm and 175 W/m², the room-temperature responsivity reached 0.41 mA/W. Specific detectivity at that condition was 2.0 × 10^12 Jones. The measured rise and decay times were 5.64 ms and 5.12 ms.

Photocurrent increased linearly with incident optical power at 0 V. The fit gave R² = 0.9964. The authors associated this behavior with low defect density and a mainly photoconductive response. Above 175 W/m², light absorption approached saturation, so responsivity and detectivity declined.

The reported temperature-dependent metrics are summarized below. The photodetection metrics apply to 213 nm illumination at 175 W/m² and 0 V bias.

Metric25 °C250 °C
Rectification ratio6.8 × 10^81.8 × 10^9
Photocurrent-to-dark-current ratio387273
Responsivity0.41 mA/W0.37 mA/W
Specific detectivity2.0 × 10^12 Jones1.7 × 10^12 Jones
Rise time / decay time5.64 / 5.12 ms3.45 / 2.68 ms

The rise time covers the transition from 10% to 90% of maximum current. The decay time covers the transition from 90% to 10%. A control detector without the unintentionally doped layer showed poorer ultraviolet performance in Supporting Information S1.

The detector placed an unintentionally doped diamond layer on heavily boron-doped diamond, termed p+ diamond in the paper. The unintentionally doped layer was grown for 60 min by microwave plasma chemical vapor deposition.

The boron-doped film was deposited with H2 as carrier gas. CH4 and trimethylboron served as the carbon and boron precursors. The boron-doped films were mechanically polished and chemically treated in a 3:1 H2SO4 to HNO3 mixture at 120 °C. The unintentionally doped layer was then grown on the treated boron-doped surface.

Thermal evaporation deposited Au for the top and bottom electrodes. The control device contained p+ diamond alone and showed ohmic contact behavior. Device 1, which included the unintentionally doped p+ junction, showed strong rectification. This comparison supports the authors’ conclusion that the built-in field at the homojunction caused the rectification.

High-resolution X-ray diffraction gave full width at half maximum values of 0.019° for unintentionally doped diamond and 0.021° for boron-doped diamond. Raman spectroscopy placed the main boron-doped peak at 1332.01 cm⁻¹ and the unintentionally doped peak at 1332.33 cm⁻¹. Additional peaks at 588, 920, and 1042 cm⁻¹ were attributed to scattering from boron atoms in the diamond lattice.

Both films had an ultraviolet-visible absorption edge at 225 nm. Tauc analysis gave an optical band gap of 5.5 eV for each material.

Ultraviolet photoelectron spectroscopy measured work functions of 4.08 eV for unintentionally doped diamond and 5.18 eV for p+ diamond. The Fermi level of the boron-doped material lay near the valence-band maximum. The authors inferred that Fermi-level alignment causes band bending at the junction and creates a built-in electric field directed from the unintentionally doped region toward the p+ region. Under ultraviolet illumination, this field separates photogenerated electron-hole pairs and supports carrier extraction without an applied voltage.

The temperature study covered 25 °C to 250 °C under 213 nm illumination. At 0 V and 175 W/m², the photocurrent decreased only slightly as temperature rose. Responsivity fell from 0.41 mA/W to 0.37 mA/W. Specific detectivity fell from 2.0 × 10^12 Jones to 1.7 × 10^12 Jones.

The authors attribute the small current reduction to stronger phonon scattering at elevated temperatures. Diamond’s wide band gap suppresses intrinsic carrier generation and limits thermally activated carrier leakage. The rectification ratio nevertheless increased from 6.8 × 10^8 at 25 °C to 1.8 × 10^9 at 250 °C.

Forward-biased operation produced much higher temperature-dependent currents. At 6 V, the dark current rose from 8.79 mA to 100 mA between 25 °C and 250 °C. The photocurrent increased from 8.86 mA to 100 mA over the same interval. The authors associated this change with increased boron ionization as temperature rose. Boron has an ionization energy of about 0.36 eV, so the higher free-carrier density reduced resistance and increased forward current.

Comparison with reported diamond detectors

Section titled “Comparison with reported diamond detectors”

The paper reports a dark current of 4.4 × 10⁻¹¹ A at 250 °C. Its selected comparison with other diamond-based photodetectors is shown below.

PhotodetectorTemperatureBiasPDCRRise / decay
SiC/diamond planar MEM200 °C15 V2.78Not reported
Ga2O3/diamond heterostructure200 °C20 VUnder 10^2495 / 539 ms
Pt/H-diamond/Au SBD166 °CNot reportedNot reportedNot reported
MSM300 °C-10 V1.63112 / 7380 ms
3-D SBD300 °C-10 V~1.0185 / 2100 ms
P-BDD/n-WO3 detector300 °C-5 V0.151200 / 1130 ms
Diamond homojunction250 °C0 V2733.45 / 2.68 ms

The homojunction combined zero-bias operation with a higher PDCR than every listed comparator. It also had shorter listed response times than the comparators with reported values. The authors describe its PDCR and response speed as among the highest reported for elevated-temperature diamond ultraviolet detectors.

The comparison is not controlled because the devices use different structures, temperatures, and applied voltages. It supports comparison with published devices rather than attribution to one material parameter.

The detector maintained stable on-off current under 213 nm illumination at 175 W/m² and 0 V as the temperature increased from 25 °C to 250 °C. A separate cycling test covered 110 switching cycles at both 25 °C and 250 °C. The paper reports stable switching behavior and no photocurrent degradation.

The imaging system used one detector as a single pixel. The device was fixed on a heating stage behind a metal mask with hollow letters S, D, and U. Output current was converted into spatial coordinates to reconstruct two-dimensional images.

Images were recorded at 25 °C, 150 °C, and 250 °C under 213 nm illumination at 0 V. The letter patterns remained distinct at 250 °C. The authors report no obvious resolution degradation relative to the room-temperature result.

The measured evidence supports 213 nm, zero-bias operation between 25 °C and 250 °C. The paper does not report spectral response at other wavelengths or operation above 250 °C. The authors also identify controllable doping of diamond as a barrier to systematic temperature-dependent studies.

Although the paper describes long-term stability, the reported endurance test contains 110 switching cycles. It does not establish lifetime under continuous ultraviolet exposure or repeated thermal cycling. The imaging result uses a single pixel rather than an array, and no numerical resolution metric is reported.

The text also leaves the unintentionally doped layer thickness, quantitative boron concentration, TMB flow, effective area, and device-to-device variability unspecified. Those details affect junction reproduction and scaling. For material selection, the result supports boron-doped diamond homojunctions when zero-bias UV sensing and thermal stability are priorities.

The paper’s photodetector needs an MPCVD-grown unintentionally doped and heavily boron-doped diamond junction with Au electrodes for 213 nm sensing tested to 250 °C. 6CCVD’s diamond materials catalog includes boron-doped diamond, Au metallization, and MPCVD-grown single-crystal and polycrystalline plates and wafers up to 125 mm. The paper’s specific UID/p+ junction and Au electrode pattern require custom fabrication, which can be discussed through the linked engineering support.

View Original Abstract

ABSTRACT Ultraviolet photodetectors (UVPDs) operating at high temperatures are critical for sensing in extreme environments such as aerospace, nuclear electronic systems, and flame warning. However, most existing UVPDs suffer from severe performance degradation, increased dark current, and the need for external bias as the operating temperature rises. Here, we report a self‐powered UVPD based on an unintentionally doped/boron‐doped diamond homojunction. Benefitting from the ultra‐wide bandgap and excellent thermal stability of the diamond, the device has achieved high optoelectrical properties under UV illumination with negligible performance degradation even at an operating temperature of up to 250°C. The device demonstrates a high rectification ratio of 1.8 × 10 9 , high specific detectivity of over 10 12 Jones, and fast response times of 3.45 ms/2.68 ms under 0 V bias. Moreover, the device shows excellent long‐term operating stability and achieves high‐contrast imaging applications with excellent temperature tolerance. These results highlight the promise of diamond homojunctions for reliable UV detection and imaging in harsh operating environments.