Radiative recombination of near-surface donor–acceptor pairs in hydrogenated homoepitaxial diamond nanolayers
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Applied Physics Reviews |
| Published | 2026-10-09 |
| Authors | Alexey M. Romshin, A. P. Bolshakov, Alexander A. Zhivopistsev, Alexander V. Gritsienko, Oleg S. Kudryavtsev |
| DOI | 10.1063/5.0346046 |
Hydrogen-terminated homoepitaxial diamond nanolayers produced narrow donor-acceptor pair (DAP) emission with an intensity maximum at an interlayer thickness of 1.7 nm. The same thickness gave a mean fluorescence lifetime of 1.9 ns. A representative individual line gave (g^{(2)}(0)=0.37), while the brightest ensemble reached 82% detection-polarization contrast.
Layer geometry sets pair separation
Section titled “Layer geometry sets pair separation”The emitters formed near a boundary between nitrogen-rich bulk donors and hydrogen-induced surface acceptors. A nominally undoped chemical vapor deposited (CVD) interlayer set their characteristic separation.
A strongly yellow high-pressure, high-temperature (HPHT) single-crystal diamond plate with a (100) surface served as the main substrate. Fourier-transform infrared (FTIR) analysis gave a substitutional nitrogen concentration of (2.0 \times 10^{19}\ \mathrm{cm}^{-3}), or approximately 110 ppm. The estimate used the nitrogen C-center band at 1135 cm^-1.
An approximately 1 × 3 mm diamond mask covered part of the surface during microwave plasma chemical vapor deposition (MPCVD) growth. Masking produced a monotonic local thickness gradient through lateral changes in radical flux and possibly substrate temperature.
Contact atomic force microscopy (AFM) mapped the topography before confocal photoluminescence (PL) recorded the same area. One profile region varied by approximately 6 nm over approximately 25 µm. A steeper region varied by approximately 30 nm over approximately 3 µm.
Films were grown in an ARDIS-300 MPCVD reactor at 2.45 GHz. Chamber pressure was 60 to 70 Torr, microwave power was 2 to 2.3 kW, methane content was 0.5%, and substrate temperature was 800 to 830 °C. Growth lasted 30 to 80 s.
The synthesis produced a nominally undoped 2 to 5 nm film, with the mask edge supplying thinner and thicker local regions. After growth, hydrogen plasma at 750 °C for 10 min terminated the film and exposed substrate surface.
Earlier electrical work placed donor-surface acceptor recombination within 0 to 12 nm of the surface, but the optical behavior of those pairs had not been investigated.
Thickness selects the emission maximum
Section titled “Thickness selects the emission maximum”Confocal PL used 532 nm excitation at 500 µW. Bright narrow-band emission appeared from 545 to 650 nm, except for the 570 to 575 nm diamond Raman line.
Integrated emission was strongest for interlayers between 1 and 4.5 nm. Its maximum occurred near 1.7 nm, where maps also showed the highest density of luminescent spots. The probability of resolving individual narrow lines followed the same thickness dependence.
Emission weakened for very thin layers with direct surface termination. It also weakened above the optimum.
The authors attribute the thin-layer loss to short-range pair configurations and associated non-radiative quenching. At larger separations, they attribute the lower intensity to weaker donor and acceptor wave-function overlap.
Their hydrogen-like model treats radiative relaxation as decreasing exponentially with separation. Non-radiative quenching remains short ranged in the model.
Excitation energy selects pair distance
Section titled “Excitation energy selects pair distance”The thickness maximum shifted with excitation wavelength. It was 1.48 nm under 457 nm, 1.46 nm under 473 nm, 1.67 nm under 532 nm, and 2.55 nm under 633 nm.
The authors connect this shift to selective resonant excitation. Shorter wavelengths address smaller separations with larger Coulomb contributions, while longer wavelengths address larger separations.
The detected band was a phonon replica shifted by about 180 meV from the resonant transition. The broad envelope at 570 to 590 nm had detuning of 170 to 190 meV. That detuning agrees with DAP values reported for natural diamond.
A fitted thickness-energy relation gave (\Delta E) of approximately 1.75 eV and (d_{\min}) of approximately 1.3 nm. Using 2.2 eV for positively charged nitrogen, the authors estimated 1.45 eV for valence-band bending near the surface. This agrees with earlier hydrogen-terminated diamond estimates.
Spectral and time-resolved signatures
Section titled “Spectral and time-resolved signatures”At layer thicknesses of 1 to 3 nm, narrow lines were so closely spaced that they formed a quasi-continuous ensemble. The broad band came from overlapping phonon replicas, not direct emission from one resonantly excited pair.
Time-resolved measurements used pulsed excitation and a 575 to 585 nm detection window. For layers below 1 nm, mean lifetimes were 1.0 to 1.3 ns, and the decay was biexponential.
Raising thickness from 1 to 2 nm increased mean lifetime. It reached 1.9 ns at 1.7 nm, where the decay was closest to monoexponential. Above 6 nm, mean lifetime fell to approximately 0.6 ns.
The lifetime trend tracked PL intensity. The authors associate short lifetimes near the surface with short-range non-radiative quenching. They associate the 1.9 ns lifetime with stronger radiative recombination. Beyond the optimum, reduced overlap and resonant detuning shorten the observed decay.
Single photons and planar polarization
Section titled “Single photons and planar polarization”Single-photon acquisition used 532 nm excitation at 0.5 mW. A representative narrow line gave a second-order autocorrelation value of (g^{(2)}(0)=0.37). A Hanbury Brown Twiss interferometer with two avalanche photodiodes measured the value.
This result supports single-photon emission from the selected line. The paper does not report a photon-counting rate or an emitter-array yield.
Polarization-resolved PL was measured for ensembles with maximum intensity at 1.7 nm. Rotating the detection analyzer produced 82% polarization contrast. Changing excitation polarization produced no noticeable intensity change from 545 to 700 nm.
A common analyzer minimum indicated dipoles oriented mainly perpendicular to the surface. The orientation spread did not exceed ±20°. The authors associate the residual spread with phonon replicas, electron-phonon coupling, and optical-component polarization sensitivity.
Similar behavior occurred in two other DAP ensembles on the sample.
Controls for hydrogen termination and nitrogen
Section titled “Controls for hydrogen termination and nitrogen”Local electron-beam treatment tested whether surface hydrogen was required. Irradiation used 15 kV, 1 nA, and 30 s over 3 × 3 µm square rasters.
Narrow-band emission was completely suppressed inside the exposed region. Electrostatic force microscopy showed an abrupt change in near-surface electrostatic contrast. Under a +10 V probe bias, the treated region appeared dark with a well-defined boundary.
The authors interpret the contrast as partial binding or compensation of surface acceptors by donor-acceptor pairs. They also observed bright halos extending several micrometers, with widths of 5 to 8 µm, near treated boundaries. They associate the halos with local charge redistribution on the heterogeneously terminated surface.
A reference plate had substitutional nitrogen below 20 ppb. It received a thin CVD layer and the same hydrogen-plasma treatment. No comparable narrow-band PL appeared. The control shows that hydrogen termination and a CVD overlayer alone did not produce the emitters. Nitrogen-rich bulk material was also required.
Engineering implications and unresolved questions
Section titled “Engineering implications and unresolved questions”The measured thickness dependence gives a direct design variable for DAP quantum emitters. A 1.7 nm interlayer gave the strongest 532 nm response and the longest mean lifetime in this sample.
Excitation wavelength can select the favored separation without changing the layer. The 633 nm maximum at 2.55 nm provides a concrete example.
Surface-state control also matters. Electron-beam hydrogen desorption removed the emission locally, but the paper did not demonstrate reversible writing or readout.
The results provide an optical signature of hydrogen-induced surface acceptors coupled to substitutional nitrogen. The paper proposes ordered arrays of similarly oriented emitters. It does not demonstrate an array fabrication method.
The acceptor species was not chemically identified, and its assignment follows the transfer-doping model. Long-term emission stability, array uniformity, and photonic coupling were not tested.
Materials and support
Section titled “Materials and support”The platform needs a hydrogen-terminated single-crystal diamond nanolayer on a nitrogen-rich (100) HPHT substrate, with a 2 to 5 nm homoepitaxial overlayer setting the donor-acceptor separation. 6CCVD’s diamond materials catalog lists MPCVD-grown single-crystal diamond plates and wafers up to 125 mm, with thickness from 0.1 to 500 µm. These entries provide a relevant plate format, but the study’s nitrogen content, orientation, hydrogen termination, and nanometer epitaxy require review through engineering support.
View Original Abstract
Hydrogen-terminated diamond is known for its p-type surface conductivity, which arises from a near-surface hole accumulation layer induced by adsorbed acceptor species. Here, we demonstrate that these surface acceptors also form optically active donor-acceptor pairs (DAPs) with substitutional nitrogen donors in diamond. The insertion of a nominally undoped CVD interlayer between a nitrogen-rich high-pressure high-temperature substrate and a hydrogen-terminated surface enables the precise tuning of the donor-acceptor separation with nanometer precision. Radiative DAP recombination appears as bright, spectrally narrow lines whose intensity, energy, and decay dynamics depend systematically on interlayer thickness. Individual lines show single-photon statistics, while ensembles exhibit strong polarization anisotropy reflecting the planar donor-acceptor geometry. These findings reveal an optical counterpart of hydrogen-induced surface transfer doping in diamond and establish a surface-defined, nanometer-tunable platform for engineering DAP-based quantum emitters.
Single Crystal Cvd Diamond Plate
View Specs →
Customized Single Crystal Cvd Diamond Plate
View Specs →