Reversing annealing-induced optical loss in diamond microcavities
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | APL Photonics |
| Published | 2026-03-01 |
| Authors | Vinaya K. Kavatamane, Natalia C. Carvalho, Ahmas El-Hamamsy, Elham Zohari, Paul E. Barclay |
| DOI | 10.1063/5.0305434 |
A 1200 °C anneal in a chamber evacuated to 10⁻⁹ mbar reduced the average intrinsic quality factor of diamond microdisks to about 1,000. A boiling tri-acid clean restored their quality factors to preannealing values between roughly 1,000 and 50,000. Raman spectroscopy identified the temporary loss source as an amorphous carbon surface layer. These results support annealing pre-fabricated cavities for color-center formation when a postannealing clean is included.
Process and measurement
Section titled “Process and measurement”The devices were microdisks patterned in optical-grade single-crystal (100)-orientation CVD diamond from Element Six. Quasi-isotropic etching with oxygen plasma created undercut structures and thin pedestals. Disk diameters ranged from 5 to 8 μm, and their thickness was about 940 nm.
The experiment tracked each chip as cleaned, annealed, and cleaned again after annealing. An evanescently coupled dimpled fiber taper and a 1480 to 1640 nm swept-wavelength laser measured transmission resonances. Line fits gave optical and intrinsic quality factors, with the intrinsic value removing fiber-taper linewidth effects. Singlets used one Lorentzian, while doublets used a coherent superposition of Lorentzians.
The vacuum chamber was evacuated to 10⁻⁹ mbar before heating. The heater rose from room temperature to 1200 °C at 1 °C per minute. The temperature was held for 120 minutes at 200, 400, 600, 800, 1000, and 1200 °C. The final ramp from 1200 °C to room temperature took 120 minutes.
Before and after annealing, the sample was cleaned at about 250 °C for about 120 minutes. The solution was an equal mixture of H₂SO₄, HNO₃, and HClO₄. Before annealing, cleaning removed residue that could burn and damage the optical properties. After annealing, cleaning targeted the non-diamond layer.
Optical quality factor changes
Section titled “Optical quality factor changes”The optical measurements separated three states of the same fabricated devices.
| Device state | Processing condition | Measured result |
|---|---|---|
| Cleaned, C | Tri-acid clean before annealing | Intrinsic Q between about 1,000 and 50,000 |
| Annealed, C-A | 1200 °C anneal after evacuation to 10⁻⁹ mbar | Average intrinsic Q of about 1,000 |
| Cleaned after annealing, C-A-C | Tri-acid clean after annealing | Intrinsic Q returned to preannealing values |
After annealing, the average across all measured modes fell to about 1,000. Cleaning after annealing returned the average to its preannealing level. The recovery shows that annealing caused no permanent optical damage in the tested disks.
The authors note that individual modes respond differently, so the average gives a qualitative trend. Ref. 45 reported optimized diamond microdisks with Q above 300,000, higher than this study’s preannealing range. The demonstrated recovery returns each device to its starting condition, rather than to that earlier value.
Cleaning narrowed the resonances, but a blue shift of about 0.2 nm remained relative to the preannealing state. The authors explained this shift with removal of a thin amorphized diamond layer no greater than 1 nm.
Loss mechanisms and geometry
Section titled “Loss mechanisms and geometry”The fabrication made pedestals wider as disk diameter increased because the undercut distance decreased. Contact between the pedestal and whispering-gallery field increased geometry-limited loss in the largest disks. Pedestal loss therefore dominated the largest clean disks over scattering, absorption, and radiation losses.
Annealed disks showed much less dependence of intrinsic quality factor on diameter. The authors attribute this behavior to absorption or surface scattering becoming the dominant loss channels. Local heating shifted resonances as input power changed across each resonance.
Modified Lorentzian fits to photothermally shifted resonances estimated absorption and scattering contributions. Thermal conductance values came from simulations that included each pedestal geometry. These fits predicted surface scattering as the main limit before and after cleaning. The authors interpret the reversible loss as a surface effect rather than permanent bulk degradation.
Raman evidence and morphology
Section titled “Raman evidence and morphology”Raman spectra were collected with 635 nm excitation over 500 to 2500 cm⁻¹ at constant laser power. A 635 nm excitation was chosen to reduce overlap with zero-phonon emission from neutral nitrogen-vacancy defects.
Freshly annealed diamond showed the native diamond peak near 1350 cm⁻¹ and a broad non-diamond peak near 1600 cm⁻¹. The broad peak increased with acquisition time, and the displayed annealed and cleaned spectra each used 120 seconds. Three-Lorentzian decomposition assigned the broad feature to the D and G bands of amorphous carbon. Tri-acid cleaning removed both amorphous-carbon contributions.
Atomic force microscopy with a 1 nm tip found RMS roughness below 100 pm in most unpatterned regions. Within those regions, AFM found no marked difference between clean and annealed samples. AFM did not measure microdisk sidewalls, top surfaces, or bottom surfaces. Annealed linewidths prevented a mode-splitting estimate of patterned-surface roughness.
Limits and unresolved questions
Section titled “Limits and unresolved questions”Several limits restrict how far these results can be applied. Color centers formed during annealing were not studied directly. AFM covered only unpatterned regions and did not resolve the patterned device surfaces.
Cleaning changed the disk diameter and thickness, which can affect photonic-crystal designs with tight geometry tolerances. The remaining resonance shift shows that the cleaned geometry did not exactly match the original geometry. The study cannot separate the effect of cleaning from the oxygen-terminated surface produced by the tri-acid treatment. Tests using O₂, F₂, and N₂ termination were proposed but not performed.
Current disks were limited by surface roughness, so smoother higher-Q cavities could reveal bulk absorption effects. The effect of ion implantation on resonator optical loss also remains unresolved. Prior focused-ion-beam work used air annealing, which differs from this high-vacuum treatment. Whether high-vacuum annealing can repair nanofabrication damage remains an open question. Mechanical loss and surface-termination effects were proposed for systematic study.
Device engineering implications
Section titled “Device engineering implications”Device patterning before localized implantation and annealing did not permanently reduce the measured cavity quality factor. A tri-acid postannealing step is part of the demonstrated recovery sequence. Geometry-sensitive designs still require checks of resonance shift and dimensional change after cleaning.
Process planning must include 1200 °C vacuum annealing, surface cleaning, and controls for patterned dimensions. The sequence supports color-center integration in high-Q cavities used for Purcell emission and spin-photon-phonon interfaces. The authors identify quantum transduction and quantum networking as target applications.
For material selection, optical-grade single-crystal diamond must remain compatible with both vacuum annealing and boiling tri-acid cleaning. Surface condition and resonator geometry remain separate constraints on cavity quality.
Materials and support
Section titled “Materials and support”The paper calls for optical-grade single-crystal diamond as stock for a patterned quantum-photonic microcavity with 940 nm device thickness. The diamond materials catalog includes MPCVD-grown single-crystal plates and wafers for this single-crystal requirement. The device still needs patterning, 1200 °C vacuum annealing, and tri-acid cleaning, making 6CCVD’s engineering support the relevant contact point.
View Original Abstract
A key challenge for quantum photonic technologies based on spin qubits is the creation of optically active defects in photonic resonators. Several of the most promising defects for quantum applications are hosted in diamond and are commonly created through ion implantation and annealing at high temperatures and high vacuum. However, the impact of annealing on photonic resonator quality factor, a critical parameter governing their coupling to defects, has not been reported. In this work, we characterize the effect of annealing at temperatures up to 1200 °C in high vacuum on the quality factors of diamond microdisk resonators. We investigate the optical losses associated with a non-diamond layer formed during annealing and use Raman spectroscopy to analyze the resonator surface morphology and demonstrate that tri-acid cleaning can restore their optical quality factors. These results show a viable method for creating defects in pre-fabricated diamond microdisk resonators without degrading their optical properties.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
High Concentration Quantum Grade Single Crystal Diamond
View Specs →
Single Crystal Cvd Diamond Plate
View Specs →References we cover
Section titled “References we cover”- Nanophotonic quantum sensing with engineered spin-optic coupling
- Realization of a multinode quantum network of remote solid-state qubits
- Coupling spins to nanomechanical resonators - Toward quantum spin-mechanics
- Statistically modeling optical linewidths of nitrogen vacancy centers in microstructures
- Cavity quantum electrodynamics with color centers in diamond
Original source
Section titled “Original source”References
Section titled “References”- 2003 - Quantum technology: The second quantum revolution [Crossref]
- 2019 - Integrated photonic quantum technologies [Crossref]
- 2021 - Development of quantum interconnects (QuICs) for next-generation information technologies [Crossref]
- 2014 - Large-scale modular quantum-computer architecture with atomic memory and photonic interconnects [Crossref]
- 2016 - Optimal architectures for long distance quantum communication [Crossref]
- 2014 - A quantum network of clocks [Crossref]
- 2023 - Nanophotonic quantum sensing with engineered spin-optic coupling [Crossref]
- 2008 - The quantum internet [Crossref]
- 2020 - Cavity quantum electrodynamics with color centers in diamond [Crossref]
- 2022 - Diamond integrated quantum nanophotonics: Spins, photons and phonons [Crossref]