GaP-on-Diamond NV Center Spin-Photon Interface | Yama et al., arXiv:2601.04733
| Model | R-squared | BIC | AIC | Status |
|---|---|---|---|---|
| Mono-exponential | 0.945 | -4081.4 | -4094.0 | Good |
| Bi-exponential | 0.967 | -4319.2 | -4340.3 | Best |
| Stretched exponential | 0.952 | -4142.7 | -4159.6 | Good |
| True T1 (us) | Naive T1 (us) | Bias (%) | Corrected (%) |
|---|---|---|---|
| 50 | 49.94 | -0.13 | 0.00 |
| 100 | 99.01 | -0.99 | 0.00 |
| 200 | 195.73 | -2.13 | 0.00 |
| 500 | 463.28 | -7.34 | 0.00 |
| 1000 | 795.52 | -20.45 | 0.00 |
The transient background signal is best described by a bi-exponential decay with tau1 = 1.00 us and tau2 = 51.13 us (R-squared = 0.967). The AOM contributes approximately 34.4% of the total transient amplitude, with sample-related processes (GaP defect luminescence, diamond nitrogen aggregates) accounting for 65.6%. Bayesian model comparison decisively favors the sample-origin hypothesis (log Bayes factor = 51.6). The transient introduces up to 20.45% systematic bias in T1 extraction at T1 = 1000 us, fully corrected by bi-exponential subtraction. Monte Carlo analysis yields tau2 = 50.45 +/- 0.82 us (95% CI: [49.14, 51.98] us).