Efficient Analysis of Photoluminescence Images for the Classification of Single-Photon Emitters

TL;DR

This paper introduces an automated image analysis method that efficiently identifies and classifies single-photon emitters in photoluminescence images, significantly speeding up the discovery process in quantum materials research.

Contribution

The authors develop a quantitative, regression-based approach for automatic detection and classification of SPEs in PL images, reducing reliance on manual identification.

Findings

Effective classification of SPEs in nanodiamond and boron nitride samples.

Adaptive criteria enable detection despite variations in emitter properties.

Method can be tailored to different material systems and experimental conditions.

Abstract

Solid-state single-photon emitters (SPE) are a basis for emerging technologies such as quantum communication and quantum sensing. SPE based on fluorescent point defects are ubiquitous in semiconductors and insulators, and new systems with desirable properties for quantum information science may exist amongst the vast number of unexplored defects. However, the characterization of new SPE typically relies on time-consuming techniques for identifying point source emitters by eye in photoluminescence (PL) images. This manual strategy is a bottleneck for discovering new SPE, motivating a more efficient method for characterizing emitters in PL images. Here we present a quantitative method using image analysis and regression fitting to automatically identify Gaussian emitters in PL images and classify them according to their stability, shape, and intensity relative to the background. We demonstrate efficient emitter classification for SPEs in nanodiamond arrays and hexagonal boron nitride flakes. Adaptive criteria detect SPE in both samples despite variation in emitter intensity, stability, and background features. The detection criteria can be tuned for specific material systems and experimental setups to accommodate the diverse properties of SPE.