Readout of strongly coupled NV center-pair spin states with deep neural networks

TL;DR

This paper demonstrates that deep neural networks can effectively decode collective spin states of strongly coupled NV center pairs from photon count data, enabling individual readout in regimes where traditional microscopy fails.

Contribution

The authors introduce a novel method combining spin to charge conversion with deep neural networks for single-shot readout of strongly coupled NV center pairs, overcoming resolution limitations.

Findings

Successful decoding of collective spin states using neural networks

Enhanced readout capability for strongly coupled NV centers

Proof of concept measurement of correlated classical signals

Abstract

Optically addressable electron spin clusters are of interest for quantum computation, simulation and sensing. However, with interaction length scales of a few tens of nanometers in the strong coupling regime, they are unresolved in conventional confocal microscopy, making individual readout problematic. Here we show that when using a single shot readout technique, collective states of the combined register space become accessible. By using spin to charge conversion of the defects we draw the connection between the intricate photon count statistics with spin state tomography using deep neural networks. This approach is particularly versatile with further scaling the number of constituent spins in a cluster due to complexity of the analytical treatment. We perform a proof of concept measurement of the correlated classical signal, paving the way for using our technique in realistic applications.