# Fileset

[s41467-026-70231-7.pdf](https://mdr.nims.go.jp/filesets/dcb1f015-e76d-4710-aea2-ceae3ffd81bb/download)

## Creator

Hiroto Tomita, Wataru Hosoda, Takumi Taniguchi, Hirokazu Fujiwara, Noriyuki Kataoka, Taisuke Kageura, [Yoshihiko Takano](https://orcid.org/0000-0002-1541-6928), Hiroshi Kawarada, Tamio Oguchi, Takayoshi Yokoya, Tomohiro Matsushita

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Atomic imaging for hydrogen and boron aggregates in boron-doped diamond by spectro-photoelectron holography](https://mdr.nims.go.jp/datasets/22467b51-6d03-4b38-ad28-2789bc6d1d8c)

## Fulltext

Atomic imaging for hydrogen and boron aggregates in boron-doped diamond by spectro-photoelectron holographyArticle https://doi.org/10.1038/s41467-026-70231-7Atomic imaging for hydrogen and boronaggregates in boron-doped diamond byspectro-photoelectron holographyHiroto Tomita 1 , Wataru Hosoda2, Takumi Taniguchi2, Hirokazu Fujiwara 3,4,Noriyuki Kataoka2, Taisuke Kageura5, Yoshihiko Takano6, Hiroshi Kawarada 7,Tamio Oguchi 8, Takayoshi Yokoya2 & Tomohiro Matsushita1The electrical properties of diamond are modulated by impurity doping. Iso-lated substitutional boron atoms introduce holes; however, the fraction ofelectrically 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 yetto be experimentally verified. Here we show the origin of multiple chemicalstates in homoepitaxially grownboron-doped diamond thin films by analyzingthe atomic environment of boron using spectro-photoelectron holography.Our analysis identifies boron dimers and boron-hydrogen complexes, withhydrogen occupying different atomic sites that give rise to distinct chemicalshifts. These results suggest that hydrogen incorporation during growth leadsto passivation of boron acceptors. We demonstrate that photoelectron holo-graphy serves as a promising tool for imaging hydrogen as well as determiningthe atomic sites of dopants.Impurity doping is a fundamental technique for tuning the proper-ties of functional materials. Trace levels of dopants can significantlyinfluence the electrical and mechanical behavior of semiconductorsand alloys1. Accordingly, accurate knowledge of dopant positionsand their surrounding atomic arrangement is essential for enhan-cing the predictive reliability of theoretical models and acceleratingmaterials design. Nonetheless, direct observation of atomicarrangements without translational symmetry, such as isolateddopants, remains technically challenging.In contrast, unintentional hydrogen incorporation in solidsoften leads to deleterious effects. In boron-doped diamonds (BDDs),a p-type semiconductor, hydrogen leads to carrier neutralization viadopant trapping1. Additionally, hydrogen incorporation inducesembrittlement in β‑titanium alloys2. Consequently, clarifying thelocal behavior of hydrogen is necessary to understand its impact onmaterial properties.Various experimental techniques have been developed to char-acterize impurity distributions, including dopants and hydrogen.Scanning transmission electron microscopy (STEM) is a widelyemployed approach for analyzing the atomic arrangement of dopants.STEM image contrast is governed by the electron-scattering cross-sections of the constituent elements, enabling the detection ofdopants within a matrix. However, elements with low scattering cross-sections are difficult to detect. Hydrogen, as the lightest element,presents particular challenges, with successful imaging largely con-fined to specialized systems, such as periodic hydrogen columns in thecrystalline solid YH23. Atom probe tomography (APT) offers a pro-mising approach for the direct observation of three-dimensionalReceived: 5 August 2025Accepted: 23 February 2026Check for updates1Graduate School of Science and Technology, Nara Institute of Science and Technology, Nara, Japan. 2Research Institute for Interdisciplinary Science,Okayama University, Okayama, Japan. 3Department of Advanced Materials Science, Graduate School of Frontier Sciences, The University of Tokyo,Chiba, Japan. 4Material Innovation Research Center (MIRC), The University of Tokyo, Chiba, Japan. 5Sensing Technology Research Institute, National Instituteof Advanced Industrial Science and Technology, Saga, Japan. 6Research Center for Materials Nanoarchitectonics, National Institute for Materials Science,Ibaraki, Japan. 7Faculty of Science and Engineering, Waseda University, Tokyo, Japan. 8Center for Spintronics Research Network, The University of Osaka,Osaka, Japan. e-mail: tomita.hiroto.tf7@naist.ac.jpNature Communications |         (2026) 17:3482 11234567890():,;1234567890():,;http://orcid.org/0000-0001-5622-8715http://orcid.org/0000-0001-5622-8715http://orcid.org/0000-0001-5622-8715http://orcid.org/0000-0001-5622-8715http://orcid.org/0000-0001-5622-8715http://orcid.org/0000-0002-8113-6000http://orcid.org/0000-0002-8113-6000http://orcid.org/0000-0002-8113-6000http://orcid.org/0000-0002-8113-6000http://orcid.org/0000-0002-8113-6000http://orcid.org/0000-0001-7496-4265http://orcid.org/0000-0001-7496-4265http://orcid.org/0000-0001-7496-4265http://orcid.org/0000-0001-7496-4265http://orcid.org/0000-0001-7496-4265http://orcid.org/0000-0001-7109-2801http://orcid.org/0000-0001-7109-2801http://orcid.org/0000-0001-7109-2801http://orcid.org/0000-0001-7109-2801http://orcid.org/0000-0001-7109-2801http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-026-70231-7&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-026-70231-7&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-026-70231-7&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-026-70231-7&domain=pdfmailto:tomita.hiroto.tf7@naist.ac.jpwww.nature.com/naturecommunicationselemental distributions with sub-nanometer resolution. Recent APTstudies have reported the direct detectionof deuteriumvia deuterium-charging methods and cryogenic transfer protocols4–6. While thesereports provide valuable insights into hydrogen embrittlementmechanisms, distinguishing intrinsic hydrogen in solid solution fromresidual chamber gases requires rigorous assessment7. Althoughhydrogen behavior has been investigated using electron paramagneticresonance8, elastic recoil detection analysis9,10, secondary ion massspectrometry11, and neutron scattering12,13, these methods lack thecapability to directly observe local atomic arrangements aroundimpurities. Therefore, a technique capable of visualizing hydrogen-containing impurities within the host matrix remains a critical need.Photoelectron holography (PEH) has recently been applied as atechnique for characterizing dopants, enabled by the development ofhigh-throughput detectors and synchrotron radiation with highbrightness and low emittance14–21. These improvements enable thedetection of dopants at concentrations near 0.1 at%14,16,18. Figure 1illustrates the principle of photoelectron hologram formation. Core-level photoelectrons emitted from dopants (emitters) propagate assphericalwaves, referred to as directwaves.A fractionof thesewaves isscattered by the electrostatic potential of neighboring atoms (scat-terers), generating additional spherical waves centered on the scat-terers, referred to as scattered waves. The angular distribution ofphotoelectron intensity corresponds to the interference pattern ofthese waves. This is called a photoelectron hologram, which reflectsthe local three-dimensional atomic arrangement surrounding theemitter. In particular, the method of element- and chemical-state-selective analysis by measuring the kinetic-energy dependence ofphotoelectron holograms is called spectro-PEH. Since the scatteringpotential is primarily governed by the positive nuclear charge, lightelements such as hydrogen are also detectable. X-ray photoelectrondiffraction (XPD), a technique formeasuring the samephenomenon asPEH, has also been increasingly reported in recent years22–26. XPDpatterns also exhibit emitter-site specificity27–29. Given these simila-rities, this work adopts the terms “holography” and “hologram”throughout. Recent developments in photoelectron holography anddiffraction are comprehensively reviewed elsewhere30–32.In this study, we applied spectro-PEH to investigate dopants in theBDD. In such materials, passivation of boron by hydrogen incorpora-tion into the crystal lattice has been predicted33. Consequently, thelocal behavior of hydrogen near boron sites has attracted significantattention34–38. We observed hydrogen atoms trapped by dopants.These findings clarify the origin of multiple chemical states observedin photoemission spectra and enable detailed characterization of BDD.Results and DiscussionEvaluation of the crystallinity of the matrix by photoelectronholographyC 1s core-level photoelectron hologram measurements were per-formed to characterize the diamond lattice. The diamond unit cellcontains two crystallographically inequivalent sites: the A site at(0, 0, 0) and the B site at (0.25, 0.25, 0.25). The crystal structures asviewed from each site are presented in Fig. 2a, b. The experimen-tally obtained hologram represents the sum of contributions fromboth sites. The left image in Fig. 2c shows the experimentallyobtained C 1s hologram recorded at a kinetic energy (Ek) of~610 eV, which closely matches prior diamond C 1s hologramsreported by Küttel et al.39 and Yokoya et al.16. The right image ofFig. 2c presents a simulated hologram based on a pristine diamondcluster. Fig. 2d illustrates atomic positions as viewed from theemitter along the <111> direction. Both experimental and simulatedpatterns display prominent forward focusing peaks (FFPs) origi-nating from atoms aligned along the <001 > , <111 > , and <110>directions. The dark features marked by dashed white lines areidentified as quasi-Kikuchi lines, which arise from Bragg-like scat-tering and interference ring overlap, indicating long-range crys-tallographic order31,40. The simulation accurately reproduces boththe FFPs and the crescent-shaped fine structure near the <001>direction, confirming the high crystallinity of the BDD thin film.Spectro-photoelectron holography for B 1s core-levelB 1s core-level photoemission spectra were measured and fitted uti-lizing the parameters established by Okazaki et al.41, as displayed inFig. 3a. The lowest binding energy component (component 1) exhi-bits the narrowest peak width, consistent with prior observations41.Components 5, 6, and 7 are attributed to boron atoms near thesurface, a conclusion supported by the spectra's emission-angledependence. This paper focuses on components 1‒4. The photo-electron hologram of component 1 in Fig. 3b closely resembles theexperimental C 1s hologram, indicating boron substitution at a car-bon site, in agreement with Okazaki et al.41. Consequently, compo-nent 1 is assigned to isolated substitutional boron (BS). The simulatedhologram in Fig. 3c, based on the BS model in Fig. 3d, accuratelyreproduces the primary features of the component 1 pattern. Theseresults demonstrate that spectro-PEH effectively identifies boronatoms substituted at electrically active sites.Structural determination of boron dimerFigure 4a shows the experimental hologramof component 4 in the B 1score-level. Although prominent FFPs are preserved, the hologramdisplays pattern broadening relative to the component 1 hologram;fine structures in the <111> and <110> directions are less defined. Theline profile at the left of Fig. 4a shows the disappearance of the sharppeaks and valleys characteristic of component 1, resulting in reducedcontrast. Such broadening suggests that boron atoms are displacedfrom ideal substitutional lattice sites. The photoelectron hologram issensitive to emitter position shifts.Apioneering studybyFedchenkoetal. demonstrated the disappearance of Kikuchi lines caused by sub-angstrom-scale relaxation of the emitter27. Prior theoreticalinvestigations42–48 propose two structural candidates to account foratomic shifts: a hydrogen atom located at the center of a B‒C bond,forming a B‒Hcomplexwith bond-center configuration (Fig. 4b), and anearest-neighbor boron pair (B dimer), as shown in Fig. 4c46. Figure 4dshows the simulated hologram of the B‒H complex with bond-centerconfiguration. The simulation indicates that insertion of hydrogenwithin the B‒C bond produces a large emitter displacement (B‒H‒Fig. 1 | Principle of photoelectron hologram formation. Photoelectrons excitedby incident light propagate as spherical waves. The interference between directlydetected photoelectrons and those scattered by neighboring atoms generates thephotoelectron hologram.Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 2www.nature.com/naturecommunicationsC ≈ 2.21 Å46, 2.22 Å45), yielding a pattern that diverges drastically fromthe experimental pattern of component 4; thus, the bond-centerconfiguration is excluded from the candidate model. In contrast,relaxation calculations predict a modest displacement for the B dimer(≈0.20 Å46), and the simulated hologramof B dimer (Fig. 4e) preservesthe substitutional-site pattern while capturing the observed broad-ening. Accordingly, component 4 is attributed to boron atoms arran-ged in the B dimer configuration. The atomic arrangement yielding thehighest similarity to the experimental image was determined byminimizing the rootmean squared error (RMSE). A grid searchmethodoptimized the boron position parameters. As shown in Fig. 4f, theextracted B‒B distance is 2.04 Å, indicating an elongation relative to aC‒C bond length of 1.57Å and in agreement with theoretical values of1.94 Å49, 1.97 Å46 and 1.99 Å49.Detection of hydrogen trapped by boronFigure 5a, b demonstrate that the photoelectron holograms ofcomponents 2 and 3 closely resemble that of component 1. Tovisualize structural deviations from the BS configuration, these pat-terns were divided by the component 1 pattern. The resulting ratioimages are displayed in the left panels of Fig. 5e, f. The component 2/1 ratio image exhibits bright regions along the <001> direction(orange dashed line), whereas the component 3/1 ratio image exhi-bits bright regions along the <111> direction (red dashed lines). Horieet al. previously demonstrated that oxygen vacancies manifest asdark regions in ratio images21. Conversely, the bright regionsobserved in this study indicate the presence of additional hydrogenatoms surrounding the substitutional boron.In the component 2/1 ratio image, the bright regions along the<001>direction suggest that hydrogen atomsoccupybridging sites, asshown in Fig. 5c. A simulated hologram of an atomic cluster withoutcarbon atoms, based on the B‒H complex with bridging site config-uration proposed by a first-principles calculation46, is presented in theright panel of Fig. 5e. Thismodel utilizes a B‒Hdistanceof 1.20 Å46. Thebright regions in the experimental ratio image exhibit excellentagreement with the simulation.On the other hand, the bright regions along the <111> direction inthe component 3/1 ratio image suggest the presence of hydrogen atthe anti-bonding site, as shown in Fig. 5d. The right panel of Fig. 5fdisplays a simulated hologram based on the B‒H complex with anti-bonding configuration, incorporating a B‒H distance of 1.17 Å46. TheFFPs in the simulation accurately reproduce the experimental featuresalong the <111> direction.In the B 1s core-level photoemission spectrum, components 2 and3 exhibit a chemical shift toward higher binding energy compared tocomponent 1 of the BS configuration, indicating that the boron in the B‒H complex possesses a greater positive charge, while the neighboringhydrogen behaves in a proton-like state. The peak areas of componentsFig. 2 | Photoelectron hologram of the diamond matrix. Diamond crystalstructures with carbon atoms at (a) the A site and (b) theB site as emitters. Blue andgreen spheres represent the local atomic environment for emitters at the A and Bsites, respectively. c Comparison of an experimental C 1s core-level photoelectronhologram obtained from the boron-doped diamond (111) thin film (left) and asimulated hologram based on pristine diamond (right). The centers of the holo-grams correspond to the <111> direction.d Projected atomic positionswithin an8Åradius of the emitters. Blue and green circles correspond to views from the A and Bsites, respectively; the circle size scales with proximity to the emitter. *Source dataare provided as a source data file.Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 3www.nature.com/naturecommunications2 and 3 indicate that boron at the B‒H complex with anti-bonding siteconfiguration is approximately 3.8 times more abundant than in thebridging site configuration. Most theoretical studies conclude that thebridge site model is more stable45–47,50–52. However, the small differencein formation energies between the two configurations, coupled withthe fact that the experimental abundance ratios were obtained underlimited temperature conditions, suggests that the results are not sig-nificantly inconsistent with theoretical predictions.Spectro-PEH revealed the atomic arrangement surrounding thedopants in the boron-doped diamond thin film. The chemical shiftobserved in the B 1 s core-level photoemission spectra is attributed tothe formation of boron aggregates and B‒H complexes, as confirmedby photoelectron holograms. Ratio images identify the trapping sitesof hydrogen atoms neighboring boron. These defect structures arisefrom hydrogen introduced during crystal growth and from hydrogenetching. In conclusion, spectro-PEH effectively correlates core-levelspectral shifts with atomic arrangement and presents a promising toolfor imaging hydrogen atoms.MethodsSample preparationA boron-doped diamond thin film was homoepitaxially grown on ahigh-pressure and high-temperature (HPHT) synthetic type Ib (111)-oriented single-crystal diamond substrate using an ASTeX microwaveplasma-assisted chemical vapor deposition (MPCVD) apparatus53–58.Trimethylboron [B(CH3)3], diluted in hydrogen andmethane, served asthe dopant source. The gas phase composition was controlled with amethane concentration of 5% and a trimethylboron/methane ratio([B(CH3)3]/[CH4]) of 9000 ppm, while the total gas flow rate wasmaintained at 100 sccm. Based on the growth rate, the film thicknesswas calculated to be approximately 300 nm. Synthesis pressure andsubstrate temperature were maintained at 110 Torr and 800 °C,respectively. Following synthesis, the sample was boiled in a 1:3 mix-ture of HNO3 and H2SO4 at 200 °C for 30minutes to ensure surfaceoxygen termination. The boron concentration, estimated by SIMS, was8‒10 × 1021cm−3 (4.5 at%). The superconducting transition temperaturedetermined via the four-point probe method was 10.0 K. The samplewas annealed at 500 °C under ultrahigh vacuum to reduce oxygen-related contaminations on the surface.Spectro-photoelectron holography experimentsSpectro-photoelectron holography experiments were conducted atthe soft X-ray beamline BL25SU in SPring-8, Japan, utilizing a ScientaOmicron DA30 electron energy analyzer. Measurements wereacquired at room temperature with unpolarized synchrotron radia-tion. The Fermi-edge of the molybdenum sample holder served asthe calibration standard for binding energy. The total energy reso-lution was set to be ~200meV. The angle between the beam axis andthe sample surface was set at 5 °. Supplementary Fig. 1 illustratesexperimental geometry. The DA30 analyzer simultaneously capturesphotoelectron angular dispersion and kinetic energy (Ek) alongthe longitudinal slit direction (θx) within an acceptance angle of±15 °. Furthermore, a built-in deflector permits analysis of the angulardispersion in the transverse direction (θy) up to a deflection angle of±10 °. Muro et al. provide a comprehensive description of theexperimental apparatus59.The kinetic energy dependence of the photoemission intensityangular distribution (θx, θy, Ek) was acquired by scanning the sample’spolar angle (θ = 0‒70 °) and in-plane rotation angle (φ =0‒120 °). Inthis three-dimensional dataset, the Ek axis represents the photoelec-tron spectrum. The acquisition of 158,400 photoemission spectra for asingle hologram required approximately 66 hours.Data processing for spectro-photoelectron holographyThe following data processing was performed based on the metho-dology described in Matsushita et al.60. Peak fitting was applied to themeasured spectra to retrieve two-dimensional intensity distributions.Following Shirley-type background subtraction, the spectra werefitted using symmetric Voigt functions (a convolution of Gaussianand Lorentzian). Refer to the source data for detailed fitting para-meters. The resulting images correspond to the two-dimensionalFig. 3 | Spectro-photoelectron holography analysis of the B 1s core-level. a B 1score-level photoemission spectrum obtained from the boron-doped diamond thinfilm. Following Shirley-type background subtraction, the spectrumwas fitted usingsymmetric Voigt functions. Relative energy positions with respect to component 1are indicated at the top. For details, refer to the source data. b An experimentalhologramof component 1 in theB 1s core-level. The left graphdisplays intensity lineprofiles comparing the emission-angle dependenceof the C 1s (solid black line) andcomponent 1 inB 1s (dashed red line). cA simulated hologrambasedon the isolatedsubstitutional boron model shown in d. *Source data are provided as a sourcedata file.Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 4www.nature.com/naturecommunicationsFig. 5 | Photoelectron holograms of the B‒H complexes. a An experimentallyobtained photoelectron hologram of component 2 in the B 1 s core-level. The leftgraph displays intensity line profiles comparing the emission-angle dependence ofcomponent 1 (solid black line) and component 2 (dashed red line). b An experi-mentally obtained photoelectron hologram of component 3 in the B 1 s core-level.The left graph displays intensity line profiles comparing the emission-angledependence of component 1 (solid black line) and component 3 (dashed red line).Schematics of the B‒H complex with (c) bridging site and (d) anti-bonding siteconfiguration. eAcomponent 2/1 ratio image (left) and a simulatedhologrambasedon a B‒H complex model with bridging site configuration without carbon atoms,where hydrogen is aligned in the <001> direction (right). The orange dashed circlehighlights the bright region in the <001> direction. f A component 3/1 ratio image(left) and a simulated hologram based on a B‒H complex model with anti-bondingconfiguration without carbon atoms, where hydrogen is aligned in the <111>direction (right). The red dashed circle highlights the bright region in the <111>direction. *Source data are provided as a source data file.Fig. 4 | Pattern broadening induced by atomic position shifts in the B dimer.a An experimentally obtained photoelectron hologram of component 4 in the B 1score-level. The left graph displays intensity line profiles comparing the emission angledependence of component 1 (solid black line) and component 4 (dashed red line). b,c illustrate schematics of the B‒H complex with bond-center configuration and the Bdimer, respectively. d, e are simulated holograms using the models in (b) and (c),respectively. f Root mean square error (RMSE) values between the component 4hologram and simulated holograms based on the B dimer, with variations in boronatomic positions. Black dots indicate sampling points; the red star denotes the con-figuration with the highest similarity. *Source data are provided as a source data file.Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 5www.nature.com/naturecommunicationsphotoemission intensity angular distribution within the range of±15 ° × ±10 ° (hologram fragment).The hologram fragments were mapped to equidistant cylindricalprojection with a 0.5 ° grid. Angular calibration of the C 1s hologramfragments was performed referencing simulated patterns based onpristine diamond. Obtained calibration parameter utilized for the B 1shologram fragments. Then we composed a hologram from all frag-ments. The final hologram was obtained by symmetrizing the com-posited image to take into account the symmetry of diamondstructure. All photoelectron holograms shown in this paper are pro-jected using an azimuthal equidistant projection.Photoelectron hologram simulationsHologram calculations were performed using the total analysis multi-ple scattering pattern simulation code (TMSP), which utilizes thepartial-wave expansion method for spherical waves61. The calculationsincorporate the electron inelastic mean free path (IMFP) and theDebye-Waller factor40. Although calculating scattering from individualatoms within a finite cluster is computationally intensive, thisapproach enables the analysis of aperiodic systems without relying onperiodic boundary conditions. For photoelectron kinetic energiesbetween 100 eV and several keV, backscattering makes a negligiblecontribution and is excluded from the simulations.Simulation parameters were optimized using the experimentallyobtained C 1s hologram. A grid search identified the cluster size andIMFP that minimized the RMSE, as shown in Supplementary Fig. 2. Theoptimal IMFP was determined to be 14 Å, consistent with the theore-tical value of 13.95 Å derived from the JPT formula for an electron witha kinetic energy of 610 eV in diamond62. This parameter accounts forsignal attenuation due to distant scatterers40. Regarding cluster size,the RMSE plateaued near a radius of 17 Å. Therefore, to balance com-putational efficiency with reconstruction accuracy, a cluster radius of20Å and an IMFP of 14 Å were adopted for all simulations.Reporting summaryFurther information on research design is available in the NaturePortfolio Reporting Summary linked to this article.Data availabilityThe source data for Figs. 2–5 are provided as a source data file. All rawdata that supports the findings of this study are available from thecorresponding author upon request. Source data are provided withthis paper.References1. Hayashi, K. et al. Investigationof theeffect of hydrogenonelectricaland optical properties in chemical vapor deposited on homo-epitaxial diamond films. J. Appl. Phys. 81, 744–753 (1997).2. Froes, F. H., Senkov, O. N. & Qazi, J. I. Hydrogen as a temporaryalloying element in titanium alloys: Thermohydrogen processing.Int. Mater. Rev. 49, 227–245 (2004).3. Ishikawa, R. et al. Direct imaging of hydrogen-atom columns in acrystal by annular bright-field electron microscopy. Nat. Mater. 10,278–281 (2011).4. Takahashi, J., Kawakami, K. & Tarui, T. Direct observation ofhydrogen-trapping sites in vanadium carbide precipitation steel byatom probe tomography. Scr. Mater. 67, 213–216 (2012).5. Chen, Y.-S. et al. Direct observation of individual hydrogen atoms attrapping sites in a ferritic steel. Science 355, 1196–1199 (2017).6. Chen, Y.-S., Bagot, P. A. J., Moody, M. P. & Haley, D. Observinghydrogen in steel using cryogenic atom probe tomography: Asimplified approach. Int. J. Hydrog. Energy44, 32280–32291 (2019).7. Chang, Y. et al. Characterizing solute hydrogen and hydrides inpure and alloyed titanium at the atomic scale. Acta Mater 150,273–280 (2018).8. Fernadi Lukman, M. & Pöppl, A. Electron paramagnetic resonancespectroscopy: Toward thepath of dihydrogen isotopologuedetectionin porous materials. https://doi.org/10.1039/D4CC06430E (2025).9. Torrisi, L. & Cutroneo, M. Elastic recoil detection analysis (ERDA) inhydrogenated samples for TNSA laser irradiation. Surf. InterfaceAnal. 48, 10–16 (2016).10. Topić,M.,Halindintwali, S.,Mtshali, C., Nsengiyumva, S.&Khumalo,Z. M. Hydrogen storage in Ti-based metal hydrides investigated byelastic recoil detection analysis (ERDA). Nucl. Instrum. MethodsPhys. Res. Sect. B Beam interact. Mater.450, 239–243 (2019).11. Tarzimoghadam, Z. et al. Multi-scale and spatially resolved hydrogenmapping in a Ni–Nb model alloy reveals the role of the δ phase inhydrogen embrittlement of alloy 718. Acta Mater 109, 69–81 (2016).12. Amann-Winkel, K. et al. X-ray and neutron scattering of water.Chem. Rev. 116, 7570–7589 (2016).13. Parker, B. S. F. & Collier, P. Applications of neutron scattering incatalysis: Where atoms are and how they move. Johns. MattheyTechnol. Rev. 60, 132–144 (2016).14. Tsutsui, K. et al. Individual atomic imaging of multiple dopant sitesin As-doped Si using spectro-photoelectron holography. Nano Lett17, 7533–7538 (2017).15. Matsui, F., Matsushita, T. & Daimon, H. Holographic reconstructionof photoelectron diffraction and its circular dichroism for localstructure probing. J. Phys. Soc. Jpn. 87, 061004 (2018).16. Yokoya, T. et al. Asymmetric phosphorus incorporation in homo-epitaxial P-doped (111) diamond revealed by photoelectron holo-graphy. Nano Lett 19, 5915–5919 (2019).17. Uenuma, M. et al. Atomic structure analysis of gallium oxide at theAl2O3/GaN interface using photoelectron holography. Appl. Phys.Express 15, 085501 (2022).18. Tang, J. et al. Direct observation of atomic structures and chemicalstates of active and inactive dopant sites in Mg-doped GaN. ACSAppl. Electron. Mater. 4, 4719–4723 (2022).19. Takeuchi, S., Hashimoto, Y., Daimon, H. & Matsushita, T. High-precision atomic image reconstruction from photoelectron holo-gram of O on W(110) by SPEA-L1. J. Electron Spectrosc. Relat. Phe-nom. 256, 147177 (2022).20. Fujii, M. N. et al. Atomic imaging of interface defects in an insulatingfilm on diamond. Nano Lett 23, 1189–1194 (2023).21. Horie, R. et al. Origin of unexpected Ir3+ in a superconductingcandidate Sr2IrO4 system analyzed by photoelectron holography.Inorg. Chem. 62, 10897–10904 (2023).22. Saldin, D. K., Harp, G. R. & Chen, X. Concentric-shell algorithm forauger and core-level photoelectron diffraction: Theory and appli-cations. Phys. Rev. B 48, 8234–8244 (1993).23. Chen, Y. et al. Convergence and reliability of the rehr-albers form-alism in multiple-scattering calculations of photoelectron diffrac-tion. Phys. Rev. B 58, 13121–13131 (1998).24. García De Abajo, F. J., Van Hove, M. A. & Fadley, C. S. Multiplescattering of electrons in solids and molecules: A cluster-modelapproach. Phys. Rev. B 63, 075404 (2001).25. Viana, M. L., Díez, Muiño, R., Soares, E. A., Van Hove, M. A. & DeCarvalho, V. E. Global search in photoelectron diffraction structuredetermination using genetic algorithms. J. Phys. Condens. Matter19, 446002 (2007).26. Winkelmann, A., Fadley, C. S. & Garcia De Abajo, F. J. High-energyphotoelectron diffraction: model calculations and future possibi-lities. New J. Phys. 10, 113002 (2008).27. Fedchenko, O. et al. Emitter-site specificity of hard x-ray photo-electron Kikuchi-diffraction. New J. Phys. 22, 103002 (2020).28. Medjanik, K. et al. Site-specific atomic order and band structuretailoring in the dilutedmagnetic semiconductor (In,Ga,Mn)As. Phys.Rev. B 103, 075107 (2021).29. Hoesch,M. et al. Active sites of Te-hyperdoped silicon by hard x-rayphotoelectron spectroscopy. Appl. Phys. Lett. 122, 252108 (2023).Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 6https://doi.org/10.1039/D4CC06430Ewww.nature.com/naturecommunications30. Kuznetsov, M. V. et al. Photoelectron diffraction and holographystudies of 2D materials and interfaces. J. Phys. Soc. Jpn. 87,061005 (2018).31. Fedchenko, O., Winkelmann, A. & Schönhense, G. Structure ana-lysis using time-of-flight momentum microscopy with hard X-rays:Status and prospects. J. Phys. Soc. Jpn. 91, 091006 (2022).32. Yokoya, T. Photoelectron diffraction and holography studies ondopant local structures. J. Phys. Soc. Jpn. 91, 091007 (2022).33. Chevallier, J. et al. Hydrogen-boron interactions in p -type diamond.Phys. Rev. B 58, 7966–7969 (1998).34. Chevallier, J. et al. Hydrogen in monocrystalline CVD boron-dopeddiamond. Phys. Status Solidi A 174, 73–81 (1999).35. Teukam, Z. et al. Shallow donors with high n-type electrical con-ductivity in homoepitaxial deuterated boron-doped diamond lay-ers. Nat. Mater. 2, 482–486 (2003).36. Mukuda, H. et al. 11B-NMR study in boron-dopeddiamond films.Sci.Technol. Adv. Mater. 7, S37 (2006).37. Murakami, M., Shimizu, T., Tansho, M. & Takano, Y. 11B nuclearmagnetic resonance in boron-doped diamond. Sci. Technol. Adv.Mater. 9, 044103 (2009).38. Barjon, J. et al. Hydrogen-inducedpassivation of boron acceptors inmonocrystalline and polycrystalline diamond. Phys. Chem. Chem.Phys. 13, 11511 (2011).39. Küttel, O. M., Agostino, R. G., Fasel, R., Osterwalder, J. & Schlap-bach, L. X-ray photoelectron andAuger electrondiffraction studyofdiamond and graphite surfaces. Surf. Sci. 312, 131–142 (1994).40. Matsushita, T. et al. Theory for high-angular-resolution photoelec-tron holograms considering the inelastic mean free path and theformation mechanism of quasi-kikuchi band. Phys. Status Solidi B257, 2000117 (2020).41. Okazaki, H. et al. Soft X-ray core-level photoemission studyof boronsites in heavily boron-doped diamond films. J. Phys. Soc. Jpn. 78,034703 (2009).42. Goss, J. P. et al. Deep hydrogen traps in heavily B-doped diamond.Phys. Rev. B 68, 235209 (2003).43. Goss, J. P. & Briddon, P. R. Theory of boron aggregates in diamond:First-principles calculations. Phys. Rev. B 73, 085204 (2006).44. Goss, J. P., Eyre, R. J. & Briddon, P. R. Theoretical models for dopingdiamond for semiconductor applications. Phys. Status Solidi B 245,1679–1700 (2008).45. Kumar, A., Pernot, J., Deneuville, A. & Magaud, L. Ab initio study ofboron-hydrogen complexes in diamond and their effect on elec-tronic properties. Phys. Rev. B 78, 235114 (2008).46. Oguchi, T. Electronic structure of boron-doped diamond with B–Hcomplex and B pair. Sci. Technol. Adv. Mater. 9, 044211 (2008).47. Upadhyay, A., Singh, A. K. & Kumar, A. Electronic structure andstability of hydrogen defects in diamond and boron doped dia-mond: A density functional theory study. Comput. Mater. Sci. 89,257–263 (2014).48. Watanabe, T. et al. The local structure in heavily boron-doped dia-mond and the effect this has on its electrochemical properties.Carbon 137, 333–342 (2018).49. Long, R. et al. Effect of B-complexes on lattice structure and elec-tronic properties in heavily boron-doped diamond. Diam. Relat.Mater. 17, 234–239 (2008).50. Dai, Y., Dai, D., Liu, D., Han, S. & Huang, B. Mechanism of p-type-to-n-type conductivity conversion in boron-doped diamond. Appl.Phys. Lett. 84, 1895–1897 (2004).51. Goss, J. P., Briddon, P. R., Sque, S. J. & Jones, R. Boron-hydrogencomplexes in diamond. Phys. Rev. B 69, 165215 (2004).52. Lombardi, E. B., Mainwood, A. & Osuch, K. Interaction of hydrogenwith boron, phosphorus, and sulfur in diamond. Phys. Rev. B 70,205201 (2004).53. Takano, Y. et al. Superconductivity in diamond thin filmswell aboveliquid helium temperature. Appl. Phys. Lett. 85, 2851–2853 (2004).54. Takano, Y. et al. Superconductivity in polycrystalline diamond thinfilms. Diam. Relat. Mater. 14, 1936–1938 (2005).55. Yokoya, T. et al. Origin of the metallic properties of heavily boron-doped superconducting diamond. Nature 438, 647–650 (2005).56. Takano, Y. et al. Superconducting properties of homoepitaxial CVDdiamond. Diam. Relat. Mater. 16, 911–914 (2007).57. Kawano, A. et al. Superconductor-to-insulator transition in boron-doped diamond films grown using chemical vapor deposition.Phys. Rev. B 82, 085318 (2010).58. Okazaki, H. et al. Signature of high Tc above 25 K in high qualitysuperconducting diamond. Appl. Phys. Lett. 106, 052601 (2015).59. Muro, T. et al. Soft X-ray ARPES for three-dimensional crystals in themicrometre region. J. Synchrotron Radiat. 28, 1631–1638 (2021).60. Matsushita, T., Muro, T., Matsui, F., Happo, N. & Hayashi, K. Dataprocessing for atomic resolution holography. Jpn. J. Appl. Phys. 59,020502 (2020).61. Matsushita, T., Matsui, F., Daimon, H. & Hayashi, K. Photoelectronholography with improved image reconstruction. J. ElectronSpectrosc. Relat. Phenom. 178–179, 195–220 (2010).62. Jablonski, A., Tanuma, S. & Powell, C. J. Calculations of electroninelastic mean free paths (IMFPs). XIV. Calculated IMFPs for LiF andSi3N4 and development of an improved predictive IMFP formula.Surf. Interface anal. 55, 609–637 (2023).AcknowledgementsThe authors thank Dr. Muro for his support in developing the experi-mental apparatus and for his support during the experiments. Thesynchrotron radiation experiments were performed with the approvalof JASRI proposal number 2018A1161 (T.Y.). This work was supportedby JSPS Grants-in-Aid for Transformative Research Area (A) “Hyper-Ordered Structures Science” grant numbers 20H05882 (T.Y.) and20H05884 (T.M.); JSPS KAKENHI grant numbers 20H01841 (T.M.) and21K18184 (T.M.); JSPS Program for Forming Japan’s Peak ResearchUniversities (J-PEAKS) grant number JPJS00420230010. This studywas conducted with support from Nara Institute of Science andTechnology, Data Science Center, RX Platform Organizational Devel-opment Project, funded by MEXT’s Education and Research Organi-zation Reform Project (T.M.).Author contributionsW.H., T.T., H.F., and N.K. performed measurements. T.K. and Y.T. pre-pared the sample. H.T. analyzed the experimental results. T.M. providedsimulation and analysis software. T.O., H.K., and T.Y. supervised theproject. All authors contributed to the writing of the manuscript.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-026-70231-7.Correspondence and requests for materials should be addressed toHiroto Tomita.Peer review information Nature Communications thanks the anon-ymous reviewers for their contribution to the peer review of this work. Apeer review file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 7https://doi.org/10.1038/s41467-026-70231-7http://www.nature.com/reprintswww.nature.com/naturecommunicationsOpen Access This article is licensed under a Creative CommonsAttribution-NonCommercial-NoDerivatives 4.0 International License,which permits any non-commercial use, sharing, distribution andreproduction in any medium or format, as long as you give appropriatecredit to the original author(s) and the source, provide a link to theCreative Commons licence, and indicate if you modified the licensedmaterial. Youdonot havepermissionunder this licence toshare adaptedmaterial derived from this article or parts of it. The images or other thirdparty material in this article are included in the article’s CreativeCommons licence, unless indicated otherwise in a credit line to thematerial. If material is not included in the article’s Creative Commonslicence and your intended use is not permitted by statutory regulation orexceeds the permitted use, you will need to obtain permission directlyfrom the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.© The Author(s) 2026Article https://doi.org/10.1038/s41467-026-70231-7Nature Communications |         (2026) 17:3482 8http://creativecommons.org/licenses/by-nc-nd/4.0/http://creativecommons.org/licenses/by-nc-nd/4.0/www.nature.com/naturecommunications Atomic imaging for hydrogen and boron aggregates in boron-doped diamond by spectro-photoelectron holography Results and Discussion Evaluation of the crystallinity of the matrix by photoelectron holography Spectro-photoelectron holography for B 1s core-level Structural determination of boron dimer Detection of hydrogen trapped by boron Methods Sample preparation Spectro-photoelectron holography experiments Data processing for spectro-photoelectron holography Photoelectron hologram simulations Reporting summary Data availability References Acknowledgements Author contributions Competing interests Additional information