Photoelectric detection of single spins in diamond by optically controlled discharge of long-lived trap states

TL;DR

This work tackles the challenge of electrically reading out solid-state spins in wide-bandgap materials by storing spin information as long-lived interface-trapped charge and retrieving it through photoinduced electrode currents. The authors demonstrate CCDMR, where spin-dependent photoionisation of a single NV center creates trapped charges at a diamond–metal interface and a subsequent readout laser releases these charges as a photocurrent, yielding a spin-sensitive signal. They report a zero-field resonance at $2.87 GHz$ with a magnetic-field–induced splitting of $20 MHz$ under $4 G$, a Rabi period of $221.1 ± 4.6 ns$, and a Hahn-echo time of $T_2 = 24.90 ± 3.29 μs$, alongside spatially resolved mappings of charge generation and trapping. The results establish CCDMR as a viable, scalable photoelectrical spin-readout method that leverages long-lived interface traps and suggests pathways toward higher bandwidth and integrated quantum sensing in diamond and related wide-bandgap materials.

Abstract

Electrical detection methods for solid-state spins are attractive for quantum technologies, being readily chip-scalable and not subject to the small photon budgets of single emitters. However, realising electrical spin readout in wide-bandgap materials with similar fidelity and bandwidth to optical approaches remains challenging. Here, we introduce a photoelectrical spin readout scheme that detects spin information stored long-term as trapped electrical charges. Using nitrogen-vacancy (NV) centres in diamond as a model system, spin-dependent photoionisation generates charge carriers that are stored in long-lived trap states at a diamond-metal Schottky junction. On-demand illumination of the junction under electrical bias releases stored charge, yielding a photocurrent transient proportional to the amount of trapped charge and hence spin state. Spin readout after coherent control of single NVs is demonstrated using charge readout in a protocol we call charge-capture detected magnetic resonance (CCDMR), and we use charge-based imaging to identify charge carrier generation and trapping processes. Our results establish CCDMR as a new technique for solid-state spin qubit readout, combining attaractive features of electrical detection with the stability of long-lived charge traps in wide-bandgap materials.

Figures (4)

• Figure 1: Experimental schematic and charge readout procedure. a) A diamond sample with metal surface contacts is connected to a voltage source and a current preamplifier for photocurrent measurements. The sample is positioned in a scanning confocal microscope for optical excitation. Inset: A PL map of a region containing 5 NV centres which were employed in this work. b) Trapped charge detection mechanism. i) 532 nm laser excitation generates a diffusing distribution of electrons and holes from a single NV centre. ii) Photo-generated charges become trapped in long-lived states residing within the diamond and at diamond-metal interfaces. iii) A bias voltage is applied between the electrodes while a laser excites the diamond-metal interface, generating a photocurrent transient proportional to the total number of trapped photocarriers. c) Photocurrent transients for laser pumping a single NV centre for 3 s at 3.5 mW (blue) compared to the same photo-pumping parameters at an adjacent empty spot (black). d) Photocurrent transients for wait times after pumping of 1 minute, 1 hour and 24 hours (turquoise, purple, and blue respectively). Inset: change in total integrated charge measured for increasing delay periods. e) The excess photocurrent is integrated to obtain a measure of the total trapped charge produced by the NV centre during laser pumping. Performing this measurement repeatedly as the pump laser spot is scanned across the sample enables photoelectric imaging of single NV centres.
• Figure 2: Spin measurement and coherent control with CCDMR. a) Schematic showing measurement process, a laser polarisation pulse (300 $\upmu$W for 2.5 $\upmu$s) initiates an $m = 0$ spin state before a microwave pulse is applied, and the resulting spin state is converted to charges via an ionisation pulse (3.5 mW for 500 ns). The whole sequence is repeated for 1.5 s before readout of the charges is performed at the electrode. b) (left): CCDMR showing characteristic NV zero-field resonance at 2.87 GHz and (right) CCDMR under a bias magnetic field of 4 G oriented along the NV axis. c) Rabi oscillation measured at 2.87 GHz under zero bias field. d) Hahn-echo signal measured for the lower energy resonance ($f = 2860\,$MHz) shown in panel b). Shaded regions are one standard deviation in measurement outcomes across 10 repeats.
• Figure 3: Optical power, wavelength and spatial dependence of charge capture and readout scheme. a) Integrated charge as a function of space over the positively-biased electrode pad at 2.2 V. The inset shows an x-z slice along the contact left edge. b) Left axis demonstrates integrated charge as a function of green illumination power at the contact to induce EGPC. Right axis depicts the EGPC limiting time constant. c) Integrated charge measured as a function of applied bias and contact illumination wavelength for a 532 nm NV pump of 1 s at 3.5 mW. d) Schematic of three target NV centres and their corresponding integrated charge as a function of pump time on the NV.
• Figure 4: Spatial mapping of hole current density. a) Decay rate from NV$^-$ to NV$^0$ measured as a function of the shortest distance from the electrode edge to each NV centre. Inset: Decay rates measured as a function of radial distance from the contact illumination spot. b) Maps of average photon counts obtained from each measured NV centre (circled areas) as a function of contact illumination time under a 1 V bias. PL image was acquired with green light, and the electrode boundaries are shown in dashed white lines. Transition from green (NV$^-$) to blue (NV$^0$) indicates increased probability of hole capture. The laser illumination spot is indicated by the green cone. c) Decay rate from NV$^-$ to NV$^0$ for the case where the NV is first pumped for 1 s to generate photocarriers prior to EGPC initiation (left) and the case where no photocarriers are introduced before EGPC initiation (right).