Atomic imaging for hydrogen and boron aggregates in boron-doped diamond by spectro-photoelectron holography
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Nature Communications |
| Published | 2026-03-05 |
| Authors | Hiroto Tomita, Wataru Hosoda, Takumi Taniguchi, Hirokazu Fujiwara, Noriyuki Kataoka |
| DOI | 10.1038/s41467-026-70231-7 |
Spectro-photoelectron holography reconstructed a boron dimer with a 2.04 Å B-B separation in a boron-doped diamond film. Four analyzed B 1s states also resolved isolated boron and two boron-hydrogen configurations. The authors connected the boron-hydrogen assignments to boron acceptor passivation.
Sample growth and electrical state
Section titled “Sample growth and electrical state”The sample was a homoepitaxial film grown on HPHT synthetic type Ib (111)-oriented single-crystal diamond by MPCVD. Trimethylboron was diluted in hydrogen and methane as the boron source.
| Parameter | Condition or method | Reported value |
|---|---|---|
| Substrate | Homoepitaxial growth | HPHT synthetic type Ib (111)-oriented single-crystal diamond |
| Film thickness | Calculated from growth rate | approximately 300 nm |
| Synthesis pressure | MPCVD | 110 Torr |
| Substrate temperature | MPCVD | 800 °C |
| Methane concentration | Growth gas | 5% |
| Trimethylboron to methane ratio | Growth gas | 9000 ppm |
| Total gas flow | Growth gas | 100 sccm |
| Boron concentration | SIMS | 8 to 10 × 10²¹ cm⁻³ (4.5 at%) |
| Superconducting transition | Four-point probe | 10.0 K |
After growth, the film was boiled in a 1:3 HNO3:H2SO4 mixture at 200 °C for 30 minutes to ensure oxygen termination of the surface. It was then annealed at 500 °C under ultrahigh vacuum to reduce oxygen-related surface contamination.
Measuring and reconstructing the holograms
Section titled “Measuring and reconstructing the holograms”Photoelectron holography records interference between direct photoelectron waves and waves scattered by neighboring atoms. The angular intensity pattern encodes the local three-dimensional atomic arrangement around the emitter.
Energy-resolved B 1s spectra separate chemically distinct components, allowing each component to produce a separate hologram. Measurements were performed at room temperature with unpolarized soft X-rays at SPring-8 BL25SU.
A Scienta Omicron DA30 analyzer captured the angular dispersion. Binding energy was calibrated with the Fermi edge of a molybdenum sample holder. The total energy resolution was approximately 200 meV. The beam axis was set at 5° to the sample surface.
The analyzer acceptance field covered ±15° × ±10°. Polar scans covered 0° to 70°, and in-plane rotation covered 0° to 120°. Each hologram required 158,400 spectra and about 66 hours.
Shirley-type background subtraction preceded symmetric Voigt fitting of each spectrum. The fitted fragments were mapped onto an equidistant cylindrical grid with 0.5° spacing. Angular calibration used pristine-diamond C 1s simulations and was applied to the B 1s fragments. Symmetrization used the diamond lattice symmetry.
TMSP simulations used partial-wave multiple scattering with a 20 Å cluster radius. Finite clusters allowed analysis without periodic boundary conditions. The selected inelastic mean free path was 14 Å at 610 eV, close to the calculated 13.95 Å value. Backscattering was excluded because its contribution was negligible over the modeled energy range.
Assigning isolated boron and boron dimers
Section titled “Assigning isolated boron and boron dimers”The carbon C 1s hologram recorded at approximately 610 eV matched earlier diamond results from Küttel et al. and Yokoya et al. Forward focusing peaks followed the <001>, <111>, and <110> directions. Quasi-Kikuchi lines indicated long-range crystallographic order. A pristine-diamond simulation reproduced the peaks and crescent-shaped structure near <001>, confirming high film crystallinity.
Component 1 closely resembled the carbon C 1s hologram. A simulation with isolated substitutional boron reproduced its main features, so the authors assigned component 1 to isolated substitutional boron at an electrically active site.
Component 4 retained forward focusing peaks but showed broader lines and reduced contrast. This indicated that its boron emitters had shifted from ideal substitutional sites.
The bond-center boron-hydrogen model predicted B-H-C distances of about 2.21 or 2.22 Å. Its large emitter displacement produced a hologram unlike component 4, so the model was excluded. Calculations for the boron dimer predicted a much smaller relaxation of approximately 0.20 Å.
A grid search minimized RMSE between simulated and measured component 4 holograms. It gave the best B-B separation of 2.04 Å.
Locating hydrogen around boron
Section titled “Locating hydrogen around boron”Components 2 and 3 also resembled component 1, so their holograms were divided by the component 1 pattern. Component 2 produced bright ratio features along <001>. Component 3 produced bright features along <111>. Bright regions indicated hydrogen surrounding substitutional boron.
These directions matched simulations for hydrogen at a bridging site and an anti-bonding site, respectively. The bridging model used a B-H distance of 1.20 Å. The anti-bonding model used 1.17 Å. Both simulations omitted carbon atoms to isolate the hydrogen contribution.
Both components had higher binding energy than component 1. The authors interpreted this shift as a more positive boron charge and protonlike neighboring hydrogen.
| Component | Assignment | Main holographic evidence | Structural value |
|---|---|---|---|
| 1 | Isolated substitutional boron | Resembles the C 1s pattern | Boron at a carbon site |
| 2 | Bridging-site boron-hydrogen | Bright component 2/1 features along <001> | B-H distance of 1.20 Å in the model |
| 3 | Anti-bonding-site boron-hydrogen | Bright component 3/1 features along <111> | B-H distance of 1.17 Å in the model |
| 4 | Boron dimer | Broadening with retained forward focusing peaks | B-B separation of 2.04 Å from reconstruction |
Comparison with theory and earlier methods
Section titled “Comparison with theory and earlier methods”The reconstructed dimer separation exceeded the 1.57 Å C-C bond length in diamond. It also agreed with calculated dimer distances of 1.94, 1.97, and 1.99 Å.
The fitted peak areas indicated that the anti-bonding configuration was approximately 3.8 times more abundant than the bridging configuration. Most theoretical studies instead predict that the bridging configuration is more stable. The authors noted that the formation energies differed little and that the measured ratio came from limited temperature conditions. They therefore did not regard the disagreement as significant.
STEM has difficulty detecting hydrogen because of its low scattering cross-section. Atom probe tomography must separate intrinsic hydrogen from residual chamber gases. Photoelectron holography instead uses emitter-specific angular interference to examine local atomic sites.
Limits and unresolved questions
Section titled “Limits and unresolved questions”The results describe one homoepitaxial (111) film made with one gas recipe. The study does not establish defect fractions across other MPCVD conditions, crystal orientations, or boron concentrations.
The 3.8 ratio came from fitted spectral areas rather than direct counts of individual atoms. The authors also note that the abundance comparison was obtained under limited temperature conditions.
Components 5, 6, and 7 were attributed to near-surface boron and were outside the detailed analysis. The reported atomic assignments depend on simulations of candidate clusters and their agreement with measured angular patterns.
Each hologram required about 66 hours at a synchrotron beamline. The total energy resolution was approximately 200 meV. The supplied manuscript is an unedited early-access version, so the paper warns that errors may remain.
The study did not connect defect populations quantitatively to carrier density, mobility, or device performance. It also did not test whether changing growth conditions reduced passivation.
Implications for diamond material selection
Section titled “Implications for diamond material selection”Total SIMS boron concentration alone cannot distinguish the electrically active fraction in this film. The paper measured 4.5 at% boron, while the B 1s analysis separated isolated boron from dimers and two hydrogen-trapping configurations.
The authors suggest that hydrogen introduced during growth and hydrogen etching produces the passivating complexes. Substitutional boron introduces holes, while hydrogen passivation neutralizes the associated carrier activity. Growth chemistry and hydrogen management therefore affect dopant activation as well as the boron dose.
Spectro-PEH can distinguish whether a boron signal corresponds to active substitutional sites or inactive aggregates. For p-type material selection, site chemistry should be checked alongside nominal concentration. The measured 10.0 K transition confirms superconductivity, but the paper did not relate it to the fraction of isolated substitutional boron.
Materials and support
Section titled “Materials and support”The paper studies boron-doped single-crystal diamond for electronic-state characterization and superconductivity. Its sample preparation requires MPCVD-grown material and a controlled oxygen-terminated surface before annealing. The 6CCVD diamond materials catalog includes single-crystal and polycrystalline diamond grown by MPCVD, and boron-doped diamond. The engineering support link provides custom fabrication support for application-specific preparation.
View Original Abstract
The electrical properties of diamond are modulated by impurity doping. Isolated substitutional boron atoms introduce holes; however, the fraction of electrically inactive boron atoms increases at higher doping concentrations. This has been attributed to boron aggregation and hydrogen passivation, although their structural identification based on atomic arrangement has yet to be experimentally verified. Here we show the origin of multiple chemical states in homoepitaxially grown boron-doped diamond thin films by analyzing the atomic environment of boron using spectro-photoelectron holography. Our analysis identifies boron dimers and boron-hydrogen complexes, with hydrogen occupying different atomic sites that give rise to distinct chemical shifts. These results suggest that hydrogen incorporation during growth leads to passivation of boron acceptors. We demonstrate that photoelectron holography serves as a promising tool for imaging hydrogen as well as determining the atomic sites of dopants. Using spectro-photoelectron holography, researchers visualized hydrogen trapping sites in boron-doped diamond. The results reveal the atomic-scale origin of acceptor passivation mechanisms in semiconductors.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
Free Standing Boron Doped Diamond Pcd Film
View Specs →