Noise Suppression via Pulsed All-Optical Magnetometry with Nitrogen-Vacancy Ensembles

Technical Summary: Noise Suppression via Pulsed All-Optical Magnetometry with Nitrogen-Vacancy Ensembles

Problem Statement
All-optical (AO), microwave-free magnetometry using nitrogen-vacancy (NV) centers in diamond offers experimental simplifications and broader sample compatibility compared to conventional microwave-driven schemes. However, standard continuous-wave (cw) AO detection of photoluminescence (PL) changes is highly susceptible to systematic fluctuations, particularly optical intensity noise and mechanical instability. While microwave-based protocols effectively suppress low-frequency common-mode noise through normalized on/off-resonance measurements, cw AO methods typically rely on lock-in detection with amplitude-modulated magnetic fields to improve signal-to-noise ratio (SNR). This approach introduces design complexity and risks perturbing sensitive samples, such as superconducting materials. There is a need for an AO protocol that suppresses common-mode noise without requiring modulated magnetic fields or microwave irradiation.

Methodology
The authors introduce a pulsed AO protocol, termed the "single-τ sequence," designed to resolve AO-PL contrast features at near-zero magnetic fields. The method employs two PL readouts within a single optical pulse: a signal readout ($I_{sig}$) taken near the beginning of the pulse and a reference readout ($I_{ref}$) taken near the end. By calculating a normalized intensity ratio ($n = I_{sig}/I_{ref}$), the protocol suppresses common-mode noise (e.g., slow optical intensity variations) that equally affects both measurements.

The experiments utilize three CVD-grown diamond samples (D1, D2, D3) with varying NV concentrations ($\approx$3.8, 2, and 0.3 ppm, respectively). The setup involves 532 nm laser excitation focused onto the diamond, with magnetic fields controlled by three-axis Helmholtz coils. The researchers investigate the dependence of AO-PL contrast on:

1. Readout timing ($t_r$): The delay of the signal readout from the start of the optical pulse.
2. Dark time ($\tau$): The duration between optical pulses.
3. Laser power: Comparing 57 mW and 275 mW.

The pulsed results are benchmarked against conventional cw AO measurements using power-spectral-density (PSD) analysis, both with and without added artificial optical intensity noise.

Key Results

• Contrast Optimization: The pulsed AO-PL contrast is maximized when the signal readout occurs shortly after optical excitation ($t_r \approx 1$ $\mu$s) and the reference readout occurs near steady-state. This timing captures the transient reduction in PL intensity caused by NV-NV cross-relaxation at near-zero fields.
• Dark Time Dependence: The optimal dark time ($\tau$) for maximum contrast is dependent on NV concentration. For sample D1 (3.8 ppm), the contrast peaks at $\tau \approx 2.5$ ms, while sample D2 (2 ppm) peaks at a longer $\tau$. The data fits a bi-exponential model characterized by two timescales: $T_{cr}$ (associated with NV-NV cross-relaxation) and $T_{in}$ (intrinsic depolarization). Higher NV concentrations lead to stronger dipolar interactions, shortening both timescales and allowing for higher contrast at shorter $\tau$.
• Noise Suppression: In the absence of added noise, the pulsed protocol achieves a $\sim$10 dB (approx. 10$\times$) improvement in the low-frequency noise floor (near DC) compared to cw AO. This advantage diminishes above 10 Hz, where the cw method exhibits a lower noise floor due to its significantly higher sampling rate (100 kS/s vs. 400 S/s for pulsed).
• Robustness to Intensity Noise: When 0.2% amplitude-modulated optical intensity noise is introduced, the pulsed protocol maintains its $\sim$10 dB low-frequency advantage and additionally exhibits a 4–5 dB lower noise floor across the 10–200 Hz range compared to cw AO. This demonstrates effective rejection of common-mode noise without the need for lock-in detection or modulated magnetic fields.

Significance and Claims
The paper claims that the proposed pulsed AO protocol provides a significant improvement in low-frequency noise performance for NV-ensemble magnetometry near zero magnetic fields. By utilizing normalized measurements within a single optical pulse, the method effectively suppresses common-mode noise, enabling sensitive DC magnetometry without the complexity of amplitude-modulated magnetic fields or the perturbation risks associated with them.

The authors note that while the current millisecond-scale dark time limits the duty cycle and sampling rate, the findings offer critical insights for optimizing NV-diamond samples. Specifically, using samples with higher nitrogen concentrations could shorten the required dark times, potentially enabling higher sampling rates for sensitive pulsed AO measurements of higher-frequency magnetic signals. The work positions this approach as a viable path toward low size, weight, and power (low-SWaP) quantum sensors capable of operating in challenging environments where microwave irradiation or magnetic field modulation is impractical.