Surface roughness noise analysis and comprehensive noise effects on depth-dependent coherence time of NV centers in diamond

TL;DR

This paper analyzes various surface and bulk noise sources affecting the coherence time of NV centers in diamond, highlighting the dominant noise mechanisms and their impact on quantum sensing performance.

Contribution

It provides a comprehensive analysis of surface roughness-induced noise and its effects on NV center coherence, integrating multiple noise sources for accurate T2 predictions.

Findings

Surface charge density noise dominates across frequency range.

Magnetic nuclei near surface significantly reduce T2.

Shallower NV layers improve signal-to-noise ratio in magnetometry.

Abstract

Noise is a detrimental issue for nitrogen-vacancy (NV) centers in diamond, causing line broadening and decreasing the coherence time (T2). Following our previous electric and magnetic field noise work, we investigate noise caused by the diamond surface roughness, which is a source for charge density fluctuations and incoherent photon scattering. We find that the varying surface charge density noise source is prevalent throughout the entire NV dynamical decoupling frequency range, while the photon scattering noise is almost negligible. Next, we combine the results from various noise sources to perform comprehensive analyses on T2 and how it varies with NV depth. At a given NV depth of 5 nm below a hydrogen- or fluorine-terminated surface, we find that these magnetic nuclei reduce the NV coherence time the most, followed by the surface electric field noise sources. The photon scattering and bulk magnetic field noise effects on T2 are weak compared to the varying charge density, electric dipole, and surface impurity noise. However, with oxygen surface termination, the surface electric field noise becomes comparable to the surface magnetic field noise. Our calculated values of T2,Hahn (few microseconds to ten microseconds) are in good agreement with the experimental values reported elsewhere. Finally, we calculate an anticipated signal-to-noise ratio (SNR) for NV AC magnetometry of external nuclear spins. In our simplified assessment, where some depth-dependent parameters (e.g. NV conversion efficiency) are held constant, we find that shallower NV layers should yield the best SNR, which is consistent with experimental findings.