Skip to content

Cavity-enhanced emission and absorption of color centers in a diamond membrane with selectable strain

MetadataDetails
Publication Date2025-03-20
JournalPhysical Review Applied
AuthorsRobert Berghaus, Selene Sachero, Gregor Bayer, Julia Heupel, Tobias Herzig
InstitutionsUniversity of Kassel, Leipzig University
Citations2
AnalysisFull AI Review Included
  • A single-crystal diamond membrane containing an implanted Si-V− ensemble was coupled to a tunable open Fabry-Perot microcavity for cavity-enhanced emission, resonant absorption, and lifetime-control studies.
  • Local strain increased the Si-V− ground-state splitting from approximately 48 GHz to as much as 550 GHz—about an 11-fold increase—potentially enabling millisecond spin coherence near 1.5 K.
  • The fabricated membrane had an approximately 1 mm × 1 mm thinned region, a minimum thickness of 850 nm, and a surface roughness of 0.44 nm, providing low optical scattering and strong van der Waals bonding to a flat DBR mirror.
  • Ion implantation used 28Si at 72 keV with a fluence of 3 × 1011 cm−2 and a mean depth of 50 nm. Annealing and chemical cleaning produced an estimated Si-V− density of 1.5 × 1010 cm−2.
  • The cavity reached an experimental finesse of 2,800 ± 120, linewidth of 5.4 ± 0.3 GHz, mode volume of 45 ± 5 λlas3, and quality factor of 76,000 ± 4,000.
  • Cavity-assisted spectroscopy yielded an ensemble absorption cross section of 4.9 ± 0.9 × 10−11 cm2. A single-emitter cross section of approximately 3.2 × 10−13 cm2 was estimated from approximately 150 coupled emitters.
  • The emitter lifetime decreased from 1.74 ± 0.27 ns in free space to 0.80 ± 0.04 ns at 4 K, corresponding to an effective Purcell factor of 2.18 ± 0.11 and cooperativity of 1.18 ± 0.06.
ParameterValueUnitContext
Diamond materialElectronic-grade single-crystal diamond—Host material for implanted Si-V− centers
Crystal orientation⟨100⟩—Used for the Si-V− orientation and dipole analysis
Initial bulk dimensions20 µm × 2 mm × 2 mmµm, mmDiamond before membrane fabrication
Thinned membrane areaApproximately 1 mm × 1 mmmmOptical membrane region
Minimum membrane thickness850nmProfilometer measurement; approximately 880 nm along the second scan
Typical membrane thickness in cavity region1.2µmEstimated from cavity measurements
Frame thickness20µmSupporting frame used for handling
Surface roughness0.44nm RMSAtomic-force-microscopy measurement
Round-trip scattering loss<160ppmAssociated with the smooth membrane surface
Silicon isotope28Si—Implanted species
Implantation energy72keVMean implantation depth of 50 nm
Implantation fluence3 × 1011cm−2Dense Si-V− ensemble
Implantation depth50nmMean depth inside the membrane
Implantation distribution25-75nmRange containing 95% of ions
Estimated emitter creation yield5%Used for ensemble-density estimation
Estimated Si-V− density1.5 × 1010cm−2Based on fluence and estimated yield
Annealing temperature1200 and 1500°CTwo high-temperature annealing stages
Annealing duration1h per stagePost-implantation processing
CleaningHydrofluoric acid followed by triacid—Used to remove glass and surface residues
Membrane bondingvan der Waals bonding to flat DBR—No heating applied during bonding
Cavity typeOpen Fabry-Perot microcavity—Plane mirror plus curved mirror
Curved-mirror radius19.7 ± 2.1µmOverall fitted radius of curvature
Mirror ellipticity0.19—Measured cavity-mirror geometry
Cavity lengthApproximately 10-10.7µmEffective cavity length
Cavity finesse, theoretical3,140—Bare-cavity estimate
Cavity finesse, experimental2,800 ± 120—Airlike modes
Cavity finesse, diamondlike modes500 ± 100—Demonstrates preference for airlike operation
Cavity linewidth5.4 ± 0.3GHzEffective linewidth at approximately 10 µm length
Cavity quality factor76,000 ± 4,000—Derived from finesse and mode number
Cavity beam waist1.52 ± 0.09µmGaussian fundamental mode
Cavity mode volume45 ± 5λlas3At approximately 737 nm
Cavity mode area1.81 ± 0.20µm2Fundamental transverse mode
Mirror reflectivity99.9%At 737 nm
Mirror transmission1,000ppmAt 737 nm
Cavity mode numberApproximately 27-30—Used during resonant measurements
Operating temperature4, 70, 80, 120 K and room temperatureKCryogenic and temperature-dependent measurements
Excitation wavelength712.3nmOff-resonant cavity-mode excitation
Si-V− zero-phonon-line regionApproximately 736-739nmStrain-dependent optical transitions
Single-emitter linewidth171 ± 22MHzResonant PLE measurement at 4 K
Ensemble inhomogeneous linewidth, free space25.1 ± 1.9GHzAt 4 K
Ensemble inhomogeneous linewidth, cavity66 ± 7GHzAt 4 K; broader because the cavity samples more strain regions
Ground-state splitting, unstrainedApproximately 48GHzTypical Si-V− spin-orbit splitting
Ground-state splitting, maximum strained550GHzLocally selected region
Ground-state splitting increaseApproximately 11-fold—Relative to unstrained Si-V−
Expected coherence benefitMillisecond scale—Projected for highly strained emitters at 1.5 K
Number of emitters contributing to cavity signalUp to approximately 150—Estimated from cavity-mode area and emitter density
Ensemble absorption cross section4.9 ± 0.9 × 10−11cm2Cavity-assisted absorption at the C-transition
Single-emitter absorption cross sectionApproximately 3.2 × 10−13cm2Estimated from ensemble value and emitter number
Emitter-induced cavity loss2,700 ± 400ppmAt the C-transition
Maximum reduced finesse820 ± 270—At the Si-V− C-transition
Free-space lifetime, 4 K1.74 ± 0.27nsOff-resonant excitation
Cavity lifetime, 4 K0.80 ± 0.04nsResonant C-transition excitation
Cavity lifetime, 70 K1.44 ± 0.08nsCavity-integrated measurement
Free-space lifetime, 70 K1.71 ± 0.01nsTemperature-dependent comparison
Free-space lifetime, room temperature1.21 ± 0.01nsReduced by thermally activated nonradiative decay
Effective Purcell factor, 4 K2.18 ± 0.11—Cavity-induced lifetime reduction
Effective Purcell factor, 70 K1.19 ± 0.07—Reduced spectral overlap
Theoretical Purcell factor9.2 ± 0.4—Based on ideal cavity parameters
Beta factor0.540 ± 0.024—Fractional emission into the cavity mode
Cooperativity1.18 ± 0.06—At 4 K
Coupling rate0.38GHzReported lower-bound estimate for g / 2π
Cavity decay rate5.4GHzReported κ / 2π
Emitter decay rate0.091GHzReported γ / 2π
Cavity-length stability160pmStability above a one-minute timescale
Second-order correlation at zero delay0.07+0.18−0.07—Single-emitter photon antibunching measurement
  1. Starting material preparation

    • Used electronic-grade single-crystal diamond with dimensions of 20 µm × 2 mm × 2 mm and a ⟨100⟩ facet.
    • Cleaned the diamond in piranha solution using a 3:1 mixture for 1 h.
    • Applied oxygen-plasma treatment at 150 W and 0.7 mbar for 2 min.
  2. Membrane fabrication

    • Protected the sample with a bulk-diamond mask containing an angled inlet to limit overetching near the membrane edges.
    • Used an Ar/Cl2 + O2 cyclic ICP-RIE recipe.
    • Applied Ar/Cl2 steps for surface smoothing and O2 steps for rapid deep etching.
    • Reduced the membrane thickness to a minimum of 850 nm over an approximately 1 mm × 1 mm area.
    • Measured the surface with AFM, obtaining 0.44 nm RMS roughness.
  3. Si-V− center formation

    • Implanted 28Si ions at 72 keV.
    • Used a fluence of 3 × 1011 cm−2.
    • Obtained a mean implantation depth of 50 nm, with 95% of ions between 25 nm and 75 nm.
    • Annealed for 1 h at 1200 °C and for 1 h at 1500 °C.
    • Removed glass residue and contamination using hydrofluoric acid and triacid cleaning.
    • Estimated a 5% Si-V− creation yield and an emitter density of 1.5 × 1010 cm−2.
  4. Membrane-mirror integration

    • Bonded the membrane directly to a flat distributed Bragg reflector through van der Waals forces.
    • Avoided intentional heating during bonding to prevent adding uncontrolled thermal strain.
    • Selected the membrane position laterally to access regions with different strain values and therefore different Si-V− transition frequencies.
  5. Cavity assembly

    • Formed a hemispherical open Fabry-Perot cavity using the flat DBR mirror and a curved mirror.
    • The curved mirror had a radius of curvature of 19.7 ± 2.1 µm and ellipticity of 0.19.
    • Controlled the cavity length using a piezo-driven z nanopositioner.
    • Adjusted the membrane position with x and y nanopositioners.
    • Operated primarily in airlike modes, which minimized diamond-related scattering losses.
  6. Optical spectroscopy

    • Used a tunable titanium-sapphire laser for excitation.
    • Performed off-resonant excitation at 712.3 nm and resonant excitation near the strain-shifted Si-V− zero-phonon line.
    • Tuned the cavity resonance to the selected C-transition and collected cavity-modulated photoluminescence and photoluminescence excitation.
    • Compared cavity and free-space spectra at 120 K, 70 K, and 4 K.
  7. Absorption and lifetime measurements

    • Measured the reduction in cavity finesse caused by resonant absorption of the Si-V− ensemble.
    • Calculated the ensemble absorption cross section from emitter-induced losses and cavity-mode area.
    • Compared free-space and cavity-coupled fluorescence lifetimes.
    • Determined Purcell enhancement, beta factor, and cooperativity from the lifetime reduction and cavity parameters.
6CCVD materialSuitability for this researchRecommended use
Optical Grade SCDPrimary material for the single-crystal membrane platformReproduce the low-roughness, high-optical-quality membrane and support cavity-enhanced Si-V− spectroscopy
Implantable SCDSuitable starting material for controlled Si ion implantationProduce shallow Si-V− ensembles for resonant absorption, Purcell, and spin-photon-interface studies
Low-strain SCDUseful reference material for baseline measurementsDistinguish fabrication-induced and bond-induced strain from deliberately applied strain
Strained SCDAppropriate for applications requiring selectable local strainTarget regions with increased ground-state splitting and reduced phonon sensitivity
Polycrystalline diamondNot preferred for coherent single-emitter spectroscopyMore suitable for broadband optical, thermal, electrode, or high-current applications
Boron-doped diamondNot the primary material for this Si-V− cavity experimentRelevant where conductive diamond, electrochemical functionality, or integrated electrical control is required

For the closest replication of the paper, 6CCVD recommends implantable Optical Grade SCD with controlled thickness, low surface roughness, and a well-characterized ⟨100⟩ orientation. A strained SCD membrane is preferable when the objective is to reproduce the reported selectable-strain behavior and access ground-state splittings substantially above the unstrained Si-V− value.

  • 6CCVD can provide custom SCD and PCD plates and wafers up to the listed dimensional range, including PCD dimensions up to 125 mm.
  • SCD and PCD can be supplied in thicknesses from 0.1 µm to 500 µm. For cavity integration, 6CCVD should be consulted regarding membrane thickness, thickness uniformity, and backside-etch compatibility.
  • Optical-grade SCD polishing is available down to Ra < 1 nm, supporting the low-scattering surface requirements of the reported Fabry-Perot geometry.
  • Inch-size PCD can be polished to Ra < 5 nm, although PCD is not the preferred material for narrow-line single-defect spectroscopy.
  • Custom dimensions, membrane windows, laser cutting, and thickness control can be evaluated for compatibility with DBR bonding, strain transfer, and cryogenic packaging.
  • In-house metalization is available in Au, Pt, Pd, Ti, W, and Cu for electrodes, bonding layers, reflective features, electrical contacts, and integrated device structures.
  • For this application, Ti or Ti-containing adhesion layers followed by Au or Pt may be considered for electrical contacting or mirror-compatible device integration. The final metal stack should be selected to avoid introducing optical loss, uncontrolled thermal-expansion mismatch, or additional strain.
  • 6CCVD can support the development of thin SCD windows and patterned membranes, but cavity-specific curvature, DBR deposition, ion implantation, and high-temperature annealing should be specified as separate process steps.
  • 6CCVD’s in-house PhD team can assist with:
    • Selecting Optical Grade SCD, implantable SCD, or strained SCD.
    • Defining membrane dimensions, thickness, orientation, and surface-finish requirements.
    • Comparing nominal membrane thickness with the active optical thickness required by the cavity.
    • Evaluating metalization stacks for bonding, electrodes, and optical-loss control.
    • Designing strain-compatible mounting and packaging approaches.
    • Assessing the trade-offs between low surface roughness, membrane strength, ion-implantation depth, and cavity mode overlap.
    • Supporting applications involving strain-engineered Si-V− diamond membranes for cavity quantum electrodynamics and spin-photon interfaces.

The reported system requires more than a generic diamond plate: membrane flatness, implantation depth, surface roughness, crystal orientation, local strain uniformity, and optical-mode overlap are all critical to reproducing the observed linewidths, Purcell enhancement, and absorption response.

Section titled “Recommended Specification for a 6CCVD Development Build”
Specification areaRecommended starting pointEngineering rationale
MaterialImplantable Optical Grade SCDSupports low-loss optical measurements and controlled defect formation
Crystal orientation⟨100⟩ or orientation specified by the implantation designImportant for Si-V− dipole classes and strain response
Membrane thickness0.85-1.2 µm active regionMatches the paper’s submicrometer-to-micrometer membrane regime
Surface roughnessRa < 1 nm targetMinimizes cavity scattering and improves DBR bonding
Membrane areaApproximately 1 mm × 1 mm or customizedProvides room for lateral selection of strain regions
Implant species28SiMatches the reported defect-formation approach
Implant energy72 keV starting pointProduces a mean depth near 50 nm
Implant fluence3 × 1011 cm−2 starting pointProduces a dense ensemble for cavity absorption
Defect concentrationApproximately 1.5 × 1010 cm−2 estimatedSupports ensemble spectroscopy and measurable absorption
Annealing1200 °C and 1500 °C process developmentRequired for Si-V− formation and activation
Cavity operating pointAirlike mode where possibleProvides lower scattering loss than diamondlike modes
Optical targetApproximately 737 nm with strain tuningMatches the reported Si-V− optical transition region
MetalizationCustom Au, Pt, Pd, Ti, W, or CuAvailable for contacts and device integration
Surface preparationLow-roughness polish and contamination-controlled cleaningProtects cavity finesse and bonding quality

For custom specifications or material consultation, visit 6ccvd.com or contact our engineering team directly.

View Original Abstract

Group IV color centers in diamond are among the most promising optically active electron-spin systems with strong optical transitions and long spin coherences. The orbital ground-state splitting of the center is particularly important to suppress the interaction with coherence-limiting phonons. A higher orbital ground-state splitting improves the spin-coherence properties and sets the upper limit for the operating temperature. Negatively charged silicon-vacancy centers have an ordinary ground-state splitting from spin-orbit coupling of only <a:math xmlns:a=“http://www.w3.org/1998/Math/MathML” display=“inline” overflow=“scroll”><a:mn>48</a:mn><a:mspace width=“0.1em”/><a:mrow><a:mi mathvariant=“normal”>G</a:mi><a:mi>Hz</a:mi></a:mrow></a:math>, resulting in required temperatures below 1 K, which can only be achieved by dilution refrigerators. Here, we increase the orbital ground-state splitting by up to an order of magnitude by induced strain in a single-crystal diamond membrane. The characterized orbital ground-state splitting in highly strained regions could allow for electron-spin-coherence times of milliseconds at <f:math xmlns:f=“http://www.w3.org/1998/Math/MathML” display=“inline” overflow=“scroll”><f:mrow><f:mn>1.5</f:mn></f:mrow><f:mspace width=“0.1em”/><f:mrow><f:mrow><f:mi mathvariant=“normal”>K</f:mi></f:mrow></f:mrow></f:math>. We demonstrate cavity-assisted spectroscopy enabled by coupling the emitter ensemble with a selectable strain to the mode of a Fabry-Perot microcavity. Calculation of the absorption cross section yields <k:math xmlns:k=“http://www.w3.org/1998/Math/MathML” display=“inline” overflow=“scroll”><k:msubsup><k:mi>σ</k:mi><k:mi>abs</k:mi><k:mi>ens</k:mi></k:msubsup><k:mo>≈</k:mo><k:mn>4.9</k:mn><k:mo>⋅</k:mo><k:msup><k:mn>10</k:mn><k:mrow><k:mo>−</k:mo><k:mn>11</k:mn></k:mrow></k:msup><k:mspace width=“0.1em”/><k:msup><k:mi>cm</k:mi><k:mn>2</k:mn></k:msup></k:math>. Furthermore, temperature-dependent lifetime measurements reveal a Purcell enhanced twofold reduction in emitter lifetime below <o:math xmlns:o=“http://www.w3.org/1998/Math/MathML” display=“inline” overflow=“scroll”><o:mn>1</o:mn><o:mspace width=“0.1em”/><o:mi>ns</o:mi></o:math> at <s:math xmlns:s=“http://www.w3.org/1998/Math/MathML” display=“inline” overflow=“scroll”><s:mn>4</s:mn><s:mspace width=“0.1em”/><s:mrow><s:mrow><s:mi mathvariant=“normal”>K</s:mi></s:mrow></s:mrow></s:math>.