Skip to content

Quantum Electrometry of Non‐Volatile Space Charges in Diamond

FieldValue
JournalAdvanced Optical Materials
Published2026-03-05
AuthorsR. M. Goldblatt, Nikolai Dontschuk, Daniel J. McCloskey, Andy M Martin, Alexander A. Wood
DOI10.1002/adom.202600008
Citations1

A 25 μs, 6 mW green pulse at 200 V produced a space charge that completely cancelled the applied field over nearly 25 μm. NV-center electrometry mapped the screening. The measurements point to carrier transport and capture by widely dispersed substitutional nitrogen.

Continuous-wave electrometry kept the electrode bias on while green laser and microwave pulses alternated. The 1.5 μs, 2.5 mW readout pulse photogenerated carriers that drifted under the bias and were recaptured. Each repetition increased the opposing field until the microwave pulse sampled a zero net field.

Pulsing the bias only during the microwave prevented accumulation, confirming that optical exposure and bias together caused screening. Reversing the bias polarity between optical generation and readout produced a measured difference nearly twice the applied field. NV electrometry cannot determine field sign from splitting magnitude alone, so this reversal established the opposing direction. The nonlinear response at the largest total fields in this sequence remains unexplained.

The sample was CVD diamond with (100) faces, approximately 1 ppm nitrogen, and 0.01 ppm NV centers. Cr/Au electrodes were deposited on its surface, and the confocal focus was 7.5 μm below that surface. All measurements were performed at room temperature.

Green illumination used 532 nm at up to 8 mW, while orange charge-state readout used 594 nm at 7 μW. A transverse 8.3 G field produced Stark-sensitive NV eigenstates. ODMR splitting then measured the local transverse field through d⊥ = 17 Hz V⁻¹ cm⁻¹. The axial coefficient was d∥ = 0.3 Hz V⁻¹ cm⁻¹, which the analysis neglected.

Above 10 kV cm⁻¹, the splitting was approximated by 2d⊥η multiplied by the applied-field magnitude. The lateral projection factor was η ≈ 0.89. Below 10 kV cm⁻¹, the transverse magnetic field made the response slightly nonlinear.

Without a reset, the space charge grew cumulatively between repetitions because the short readout pulse could not restore a neutral charge environment. A 9 mW, 25 μs green reset without bias restored that environment before each measurement. A variable-power generation pulse was then applied with 200 V, with the probe 10 μm from the electrode edge.

Pulses of 50 μs produced a power-independent saturation splitting of 0.39(2) MHz, equivalent to about 20 kV cm⁻¹. Time-dependent splitting followed a logistic curve with a dimensionless rise parameter near 10. The fitted growth rate Γ increased linearly with green power.

This power law is consistent with a single-photon process, but it does not identify the carrier source by itself. The authors could not exclude vacancy clusters, substitutional boron, surface defects, or continued NV charge cycling. In the dual-beam protocol, the cumulative state typically reached a steady state within 100 to 200 repetitions, of order milliseconds. That result supports stability during the protocol, but it does not establish the dark storage lifetime.

A dual-beam microscope separated space-charge generation from NV readout. Both 532 nm beams had diffraction-limited 1/e² diameters of 545 nm. Beam 2 generated charge with a 25 μs, 6 mW pulse at 200 V, while fixed beam 1 read the field. Fixed beam 1 then used a 1.5 μs, 2.2 mW readout pulse.

The mapped space-charge region extended nearly 25 μm along both electrode and transverse directions. This scale was much larger than the 545 nm illumination spot and the nanometre-scale charge neighborhood of an NV center.

During generation pulses of 1 to 40 μs, charge first accumulated on the electron drift side. The region then broadened and became nearly symmetric as hole-related effects appeared. Charge-state imaging used a 500 μs, 7 μW orange pulse and collected NV⁻ fluorescence from 691 to 720 nm. NV⁻ converts to NV⁰ when it captures a photogenerated hole, so fluorescence tracked hole capture.

Hole capture became apparent only after generation pulses longer than 20 μs. Prior work cited by the authors found the highest hole-capture cross section near |E| ≈ 0. The authors explain this by a zero-field region that lets holes diffuse against the applied-field direction.

The evidence points to substitutional nitrogen as the dominant trap, but it does not directly image nitrogen charge states. Nitrogen is abundant in the sample and has stable neutral and positive states at room temperature. A distribution of neutral and positive traps can form a space charge without a negative charge state.

The authors propose two carrier sources with different spatial behavior. In their model, one-way ionization of neutral nitrogen supplies electrons and produces a directional electron-mediated field. NV charge cycling supplies both carriers and makes the profile more symmetric as its contribution grows.

In their simplified model, a zero-field core causes isotropic diffusion, which erodes polarized traps at its centre. Polarized trapping at the edges then rebuilds the screening field over a growing region. The authors state that coupled drift-diffusion, rate, and Poisson equations are needed to test this model.

Earlier NV electrometry work emphasized nearby surfaces, Debye screening, and near-surface band bending. This experiment instead tracked lateral screening from a 545 nm optical spot across nearly 25 μm. The authors report that a defect charge configuration 10 to 20 μm away can affect an NV spin state and local charge dynamics.

Direct measurement of the three-dimensional charge configuration was outside the experiment. The galvo mirror scanned only the lateral plane, so the full depth dependence remained unresolved. The nonlinear response at the largest fields in the polarity-reversal sequence also has no explanation in the paper.

Space-charge lifetime was not measured. The authors expect stored fields to persist as long as the relevant deep-level states, based on NV states that remain stable for days or longer. That expectation is not a measured lifetime for the space-charge configuration. The carrier source is not uniquely identified, and the exact charge configuration was not measured.

Space-charge screening can make an NV electrometry sequence insensitive to the applied field. The ground-state response was already weak because d⊥ was only 17 Hz V⁻¹ cm⁻¹ and d∥ was 0.3 Hz V⁻¹ cm⁻¹. Dense NV ensembles with higher nitrogen densities are likely to suffer more screening, according to the authors.

In photoelectric detection, trapping reduces photocurrent yield and removes the bias needed to sweep carriers toward collection electrodes. Within a screened region, the total field can vanish, allowing diffusion rather than directed drift. Device design must therefore control nitrogen concentration, defect density, illumination timing, and charge reset strength.

The measured fields of order 10 to 40 kV cm⁻¹ can be written into diamond, but their storage time is unknown. Material selection based only on nominal diamond purity would miss this charge-stability issue.

For NV electrometry and photoelectric detection, the paper requires a diamond plate with a (100) orientation, about 1 ppm nitrogen, 0.01 ppm NV centers, and Cr/Au electrodes. The diamond materials catalog includes MPCVD-grown single-crystal diamond plates and lists Au among its metallization options. The engineering support page can be used to confirm the defect specification and electrode stack, which the stated catalog facts do not establish.

View Original Abstract

ABSTRACT We use electrometry of quantum defects embedded within a diamond to observe stable, micron‐scale space charge distributions formed from trapped photogenerated charges. These space charges grow under optical illumination in the presence of an applied electric field, eventually screening the applied electric field entirely over a spatial extent of tens of microns due to charge carrier drift and capture. Our measurements suggest that these space charge fields originate from widely‐dispersed spatial configurations of nitrogen charges. Our results have important consequences for electrometry and photoelectric detection using qubits in wide‐bandgap semiconductors.