A laboratory preprint reports two location patterns for quantum emitters in gallium nitride, or GaN. In the studied depth maps, every imaged natural emitter aligned with the boundary between GaN and sapphire, regardless of the wavelength used to plot the map. In a separate sapphire design targeting 5 micrometres below the surface, the researchers observed an emitter at the depth of a low-temperature GaN layer. The generated density at that layer was approximately 2 x 10^7 cm-2, compared with approximately 3 x 10^8 cm-2 at the interface over the 520-780 nanometre range.
That interface result describes the studied maps, not a complete census: the total mapped area and total number of mapped emitters were not reported. The density estimates for the depth-targeted design were restricted to the stated wavelength range, and no formal precision estimate was reported.
Optical signals measured at room temperature
The natural emitters also produced narrow spectral features. Representative spectra from 15 emitters showed sharp zero-phonon lines, or ZPLs, with a median full width at half maximum, a measure of line width, of 4.2 nanometres. Because the set was representative, that median describes the measured group rather than the entire population.
Among the reported natural emitters, the second-order correlation at zero delay, written as g2(0), was 0.06 at room temperature. Saturation counts exceeded 2.5 megahertz, and the degree of linear polarization was above 90 per cent. These figures were reported for highlighted or representative emitters, and the study gives no formal uncertainty range or full distribution.
The measurements used room-temperature micro-photoluminescence, which records light from microscopic regions. Individual emitters were addressed with a continuous-wave 488 nanometre monomode laser and an oil-immersion objective with a numerical aperture of 1.45. To estimate depth, the researchers moved the objective's focal point along the GaN c-axis with a piezo stage at each emitter's in-plane position.
Growth-step comparisons
The observations differed across the low-temperature growth steps examined. Emitters appeared in both annealed and as-deposited nucleation layers, but those in the as-deposited layer were less bright. No emitters were detected in nitridated sapphire. These are descriptive comparisons, and the paper does not provide formal uncertainty estimates for them.
The intentionally introduced emitters showed zero-phonon lines across 530-920 nanometres and a mean linewidth of 5.4 nanometres. Their Debye-Waller factors, which indicate how much emission is concentrated in the narrow line, ranged from 0.69 to 0.98, with a mean of 0.81. The paper describes the emission as bright, antibunched and linearly polarized, although some introduced emitters had lower signal-to-background ratios than the cited natural-emitter comparison.
The samples were grown by metal-organic vapour-phase epitaxy on commercial c-plane sapphire and free-standing GaN. In the free-standing-GaN sample, the researchers also detected an emitter at the low-temperature-layer depth, and the paper reports substrate-independent defect introduction. The evidence covers those two substrate types only, with no formal reproducibility analysis reported.
Etching provided an indirect depth check
The depth assignment was tested by removing material and repeating the emitter maps. The emitters disappeared after complete removal of the low-temperature layer. No emitters were detected in high-temperature GaN, whose upper-limit density was three orders of magnitude lower than in low-temperature GaN. From the etching and surface measurements, the researchers report a depth accuracy of around 60 nanometres; that figure was inferred rather than obtained from direct nanometre-resolved imaging.
After the emitter-forming layer and a 250 nanometre cap, atomic-force microscopy confirmed that the growth structure remained fully coalesced. It measured a root-mean-square surface roughness of 0.94 nanometres across a 5 micrometre by 5 micrometre scan. That measurement describes the stated scan area, not broader wafer-scale uniformity.
Secondary-ion mass spectrometry, or SIMS, found carbon, oxygen and hydrogen concentrations greater than 10^19 cm-3 in both the low-temperature layer and the GaN-sapphire interface. Oxygen was one order lower in the low-temperature layer than in the nucleation layer and was associated with defect density, but the relationship was not one-to-one. The oxygen profile near the sapphire had a gradient and acknowledged measurement uncertainty. These observations do not identify the emitter defect or establish that oxygen causes the emission.
A separate electrical measurement placed an upper bound of 1 x 10^16 cm-3 on net doping in the low-temperature layer. The result came from capacitance-voltage profiling and Poisson-equation analysis and was consistent with full depletion at 1 volt forward bias. The upper bound depends on the study's abrupt-junction and full-depletion assumptions.
Publication status
The manuscript is a preprint identified as arXiv:2608.26057v1 and dated 26 August 2026. The work was supported by the Swiss National Science Foundation under grant number 200020-215633.
Paper data and sources
Original title: Depth Control of Room-Temperature Quantum Emitters in Gallium Nitride
Authors: Alexandros Bampis, Johann Stachurski, Anna Schwab et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text