Skip to content

Radiation tolerance of a diamond radiation detector for space use

FieldValue
JournalInstitutional Repositories DataBase (IRDB)
Published2026-06-26
AuthorsYoshiyuki Ando, Shutaro Ueda, Ryota Heibatake, Kaito Ozawa, M. Arimoto
DOI10.48550/arxiv.2606.28180

Two MPCVD single-crystal diamond detectors retained X-ray spectroscopic performance after proton irradiation equivalent to at least 10 years in the assumed KSAT3-X orbit. The Element Six detector showed no significant degradation after a 100-year equivalent. Its pulse-height slope fell by 5.7% after that longer exposure.

KSAT3-X is a 3U CubeSat planned for launch in 2027, with about 1U allocated to the payload. The mission aims to observe electrons and protons at 10 to 40 keV in Earth’s magnetosphere. Its planned sun-synchronous orbit is 400 to 600 km, and its expected lifetime is six months to one year.

The dose model assumes 600 km altitude and 11.1 mm of cumulative aluminum shielding from the detector housing and satellite body. Active cooling and an onboard calibration source were not preferred for the payload. The limited payload volume requires a compact instrument.

The study compared a commercial Element Six detector with an in-house Kanazawa University detector. Both used MPCVD single-crystal diamond, but the KU plate was about one sixth as thick as the e6 plate.

DetectorDiamond materialSurface areaThicknessElectrode stack and active area
e6Element Six MPCVD single-crystal2.0 × 2.0 mm²0.5 mm10 nm Ti plus 10 nm Au; 1.5 × 1.5 mm²
KUKanazawa University MPCVD single-crystal1.5 × 1.5 mm²0.08 mm10 nm Ti plus 10 nm Au; 1 mm diameter

A 90 µm layer was grown on a 2 × 2 × 0.3 mm HPHT type-Ib [100] substrate. Growth used 1500 W, 45 kPa, 5% CH4 in H2, and a 1200°C substrate temperature. Laser cutting and polishing yielded an 80 µm free-standing plate.

SIMS on material grown under the same conditions found B below 0.006 ppm and N below 0.1 ppm. The authors noted that batch variation could have raised contamination in the sample used for measurement.

Each face received 10 nm Ti and 10 nm Au after oxygen termination in a 1:3 HNO3 and H2SO4 acid treatment and a 200°C heat step. The 20 nm metal stack was thinner than the 150 to 200 nm electrodes used in prior studies. The authors say the reduction was essential for improving low-energy charged-particle detectability.

SHIELDOSE-2 in SPENVIS modeled the orbital dose at 600 km behind the 11.1 mm aluminum shield. The calculation included trapped protons, electrons, and bremsstrahlung. It assumed proton exposure from 40 keV to 1000 MeV and found 100 MeV protons dominant.

The authors combined the SPENVIS dose rate with PSTAR stopping powers. PSTAR lacked diamond inputs, so they used amorphous carbon at 2.0 g cm⁻³ as a proxy for diamond at 3.52 g cm⁻³. They expected similar mass stopping powers because density and bonding have weak effects for the same element.

The 100 MeV equivalent annual fluence was 8.89 × 10^8 protons cm⁻² yr⁻¹ for ionizing energy loss. The non-ionizing estimate gave 7.79 × 10^8 protons cm⁻² yr⁻¹. The authors chose the larger ionizing value as conservative.

The samples received cumulative 100 MeV proton fluences equivalent to three months, one year, three years, and ten years. The e6 detector also received a 100-year equivalent. Lower-energy protons contributed to the calculated dose, but 100 MeV protons represented the dominant environment.

Both detectors operated at about 20°C under a +200 V bias in photon-counting mode. Setup 1 enabled consistent linearity comparisons, while Setup 2 was used for energy-resolution measurements. Setup 1’s noise prevented useful spectra below 20 keV, so Setup 2 used lower-noise circuitry. Its shaping time changed from 1 µs before irradiation to 3 µs after irradiation. System noise was subtracted from the reported energy resolution.

Three radioisotope sources supplied the test spectrum.

SourceCharacteristic energye6 response before irradiationKU response before irradiation
241Am59.5 keV gammaPhotoelectric peak detectedNo photoelectric peak
133Ba31.0 and 35.0 keVPhotoelectric peak detectedPhotoelectric peak detected
109Cd22.2 and 24.9 keVPhotoelectric peak detectedPhotoelectric peak detected

For e6, pulse height versus incident energy followed y = ax + b. The initial slope was 11.76 ± 0.05 ch keV⁻¹, and the intercept was -40.72 ± 2.70 ch.

FWHM was 3.39 ± 0.16 keV for 241Am. The 133Ba value was 2.56 keV, with upper and lower uncertainties of 0.36 and 0.31 keV. The 109Cd value was 3.29 keV, with upper and lower uncertainties of 0.82 and 0.41 keV.

The e6 resolution did not separate the Ba or Cd Kα and Kβ lines. Background measurements placed the detectable range above about 12 keV. A 13.9 keV feature from 241Am was visible, but its peak position was hard to determine.

The KU detector did not show the 59.5 keV 241Am peak even before irradiation. It did detect the Ba and Cd peaks. The authors infer that KU can measure signal charge from X-rays with energies up to at least 35 keV.

e6 performance after the 100-year equivalent

Section titled “e6 performance after the 100-year equivalent”

No significant spectral degradation appeared after any irradiation stage through 100 years. Spectral shapes and photoelectric peaks remained consistent despite some changes in background level. Leakage current also remained unchanged.

The linearity slope stayed at 11.76 ± 0.05 ch keV⁻¹ before irradiation and 11.74 ± 0.03 after ten years. After 100 years it was 11.09 ± 0.04, a 5.7% decrease with uncertainties of +0.5% and -0.5%. Energy resolution stayed consistent across all measurement epochs.

These data support the authors’ conclusion that the e6 detector has high tolerance under the assumed orbit and shielding. They also state that an onboard calibration source is not required under those conditions.

KU tolerance and unresolved response differences

Section titled “KU tolerance and unresolved response differences”

After the ten-year equivalent exposure, KU still showed no 59.5 keV 241Am photoelectric peak. The 133Ba and 109Cd photoelectric peaks remained detectable, so no significant degradation was observed for the available response range.

The authors discuss several possible causes for KU’s weaker baseline response. Its 80 µm thickness may increase capacitance and noise or reduce high-energy photon hit efficiency. Surface polishing damage may extend tens of micrometers, which could affect KU more than 500 µm e6. Bulk traps, nitrogen-related centers, interface states, and Schottky behavior may also reduce partial charge collection.

Both diamonds had nitrogen levels comparable to or below the SIMS detection limit, so the authors could not use nitrogen concentration to explain the difference. These mechanisms remain hypotheses.

The authors call for thicker KU and thinner e6 samples. They also recommend electrical and transient measurements and electron-versus-hole collection tests. A boron-doped contact layer is another proposed test.

The study did not pump the crystals with photon or electron irradiation before measurement. Typical KU exposure time exceeded 100 minutes.

Selection implications for space detectors

Section titled “Selection implications for space detectors”

Diamond has a 5.47 eV band gap at 300 K. Si, Ge, and CdTe have 1.12, 0.66, and 1.56 eV, respectively. The electron-hole pair creation energy in diamond is 13.1 to 13.9 eV. Si, Ge, and CdTe have 3.62, 2.96, and 4.43 eV, respectively.

The authors also cite low room-temperature leakage current and operation at higher and lower temperatures. These properties, compactness, and shielding feasibility support e6 and KU as candidates for KSAT3-X.

The results support e6 for its higher spectroscopic performance and KU within the current mission requirements. Proton irradiation and X-ray spectroscopy do not establish direct detection of 10 to 40 keV charged particles. The authors call for charged-particle irradiation and better understanding of the 1 to 100 keV spectral response. Charge-collection efficiency and charge-collection distance were deferred to future work.

Device selection should therefore separate radiation tolerance from charge transport and spectral response.

KSAT3-X needs an MPCVD single-crystal diamond detector plate with Ti and Au electrodes. 6CCVD’s diamond materials catalog lists MPCVD single-crystal plates up to 125 mm, 0.1 to 500 µm thick, Ti and Au metallization, and single-crystal polishing below Ra 1 nm. The 0.5 mm e6 plate sits at the range’s upper end, whereas the 80 µm KU plate is thinner than its lower limit, so engineering support would need to confirm detector-specific geometry.

View Original Abstract

We present a study of the radiation tolerance of two types of diamond radiation detectors for space use. We plan to launch a 3U-size CubeSat, KSAT3-X, developed by Kanazawa University in 2027. The KSAT3-X mission is aimed to observe inflows and outflows of charged particles such as electrons and protons, particularly in the 10 - 40 keV energy range, in the Earth’s magnetosphere. As the mission instrument, we have developed two diamond radiation detectors. The first is composed of a microwave plasma chemical vapor deposition (MPCVD) diamond fabricated by Element Six, and the second is based on a MPCVD diamond produced in-house at Kanazawa University. We irradiate both diamonds with 100 MeV protons and evaluate their spectroscopic performance as an indicator of radiation tolerance using characteristic X-rays from radioisotope sources. We find no significant degradation in their spectroscopic performance up to at least the 10-year equivalent irradiation under the orbital environments of KSAT3-X. We additionally irradiate the Element Six diamond with 100 MeV protons up to the 100-year equivalent. As a result, no significant degradation in the spectroscopic performance is observed. These results indicate that the two diamond radiation detectors have sufficiently high radiation tolerance. We also discuss possible physical origins of the observed difference in the spectroscopic performance between the two detectors.