Robust quantification of the diamond nitrogen-vacancy center charge state via photoluminescence spectroscopy

TL;DR

The paper tackles the challenge of quantifying the NV$^0$/NV$^-$ charge-state ratio in photoluminescence spectra, where overlapping phonon sidebands complicate analysis. It introduces a dual excitation protocol (DEP) that derives sample- and setup-specific pure NV$^0$ and NV$^-$ spectra using blue and green excitation, and then uses linear least-squares fitting to determine charge-state contributions in subsequent measurements. Benchmarking against ZPL fitting, skewed Gaussian fitting, and NNMF across bulk diamond and 20–100 nm nanodiamonds demonstrates that DEP yields robust, quantitative charge-state ratios and often outperforms alternative methods in challenging samples. DEP thus provides a versatile, platform-agnostic approach for NV charge-state quantification, enabling more reliable deployment of NV-based quantum sensing and related technologies.

Abstract

Nitrogen vacancy (NV) centers in diamond are at the heart of many emerging quantum technologies, all of which require control over the NV charge state. Hence, methods for quantification of the relative photoluminescence (PL) intensities of the NV$^0$ and NV$^-$ charge state, i.e., a charge state ratio, are vital. Several approaches to quantify NV charge state ratios have been reported but are either limited to bulk-like NV diamond samples or yield qualitative results. We propose an NV charge state quantification protocol based on the determination of sample- and experimental setup-specific NV$^0$ and NV$^-$ reference spectra. The approach employs blue (400-470 nm) and green (480-570 nm) excitation to infer pure NV$^0$ and NV$^-$ spectra, which are then used to quantify NV charge state ratios in subsequent experiments via least squares fitting. We test our dual excitation protocol (DEP) for a bulk diamond NV sample, 20 and 100 nm nanodiamond particles and compare results with those obtained via other commonly used techniques such as zero-phonon line fitting and non-negative matrix factorization. We find that DEP can be employed across different samples and experimental setups and yields consistent and quantitative results for NV charge state ratios that are in agreement with our understanding of NV photophysics. By providing robust NV charge state quantification across sample types and measurement platforms, DEP will support the development of NV-based quantum technologies.

Figures (5)

• Figure 1: NV center photodynamics. A: Fluorescence spectra of NV$^{0}$ and NV$^{-}$ with characteristic ZPLs at 575 and 637 nm (marked via vertical lines), respectively, showing significant overlap between 600850 nm. B: Energy level diagram of NV$^{0}$ and NV$^{-}$ within the diamond band gap. C-E: Dynamics of the NV center under different excitation wavelengths. Violet-blue excitation (C) leads to single-photon photoionization of NV$^{-}$ into NV$^{0}$, while still exciting NV$^{0}$, which relaxes down to the ground state via fluorescence. Green illumination (D) leads to efficient excitation of both charge states and allows for dynamic charge state switching via ionization and re-pumping processes. Red light (E) has insufficient energy to promote NV$^{0}$ to the excited state, leading exclusively to NV$^{-}$ emission. Horizontal, black lines indicate charge state switching, with relative length indicative of the rate in a particular direction. Vertical, short grey lines represent phonon relaxations.
• Figure 2: Dual excitation protocol on 140 nm FNDs suspended in water. A: Reference spectra are acquired using 400 and 520 nm excitation to obtain a pure NV$^{0}$ signal ($S_{400}^n(\lambda)$) and a mixed NV$^{0}$/NV$^{-}$ spectrum ($S_{520}^n(\lambda)$), respectively, which are then normalized for the same NV$^{0}$ contribution between 550600 nm (shaded region). B: A pure NV$^{-}$ spectrum is obtained by taking the difference between the two reference spectra in (A), which is then compared to experiment under 633 nm illumination ($S_{633}(\lambda)$). C: The reference NV$^{0}$ and NV$^{-}$ spectra can then be used to fit experimental data using \ref{['eq:nv_fitting']} to determine the charge state ratio.
• Figure 3: NV charge state ratio protocol comparison. A: Fluorescence spectra of bulk diamond for excitation wavelengths ($\lambda_{\text{ex}}$) ranging 405520 nm under a confocal microscope. B: NV$^{-}$ contribution of the bulk diamond using our dual-excitation protocol (DEP), ZPL, Skewed Gaussian (SG), and NNMF fitting. The vertical line roughly indicates the wavelength we expect to see single-photon photoionization of NV$^{-}$. C: Fluorescence spectra of 20 nm FNDs suspended in DI water, excited using 520 nm for excitation powers ($P_{\text{ex}}$) between 0.6422.30 mW. D: NV$^{-}$ contribution of the 20 nm FNDs as a function of excitation power. E: Fluorescence spectra of dispersed 100 nm FNDs under a confocal microscope, ordered by increasing NV$^{-}$ content. F: NV$^{-}$ contribution for all 100 nm particles using each protocol. All spectra (A, C, E) were normalized between 550600 nm (shaded regions), with vertical lines marking the positions of the NV ZPLs.
• Figure 4: Dual excitation protocol workflow. DEP requires two initial measurements taken with $<470$ nm and 480570 nm to obtain reference NV$^{0}$ and mixed NV$^{0}$/NV$^{-}$ spectra. Background signals must be removed from the two measurements before normalization between 550600 nm. Once normalized, the difference between the signals can be calculated to obtain the NV$^{-}$ reference spectrum. Once both reference spectra are determined, they can be used to fit data to determine the relative contribution of each charge state.
• Figure 5: NV fluorescence spectra measured using an excitation wavelength of 430 nm (A) and 630 nm (B) with and without the presence of an external $\vec{B}$-field, with the difference shown in (C) and (D), respectively. The change in fluorescence under 430 nm is fit with the NV reference spectra, with residuals shown in (E). F: original and revised NV reference spectra.