Fig. 6: Optimization of the excitation pulses used for single-shot readout.

Inset: The spin is prepared in the bright (orange) or dark (blue) state and the electric field amplitude of the excitation pulses is varied. When averaging the fluorescence in a time interval of 3 μs after the first ten pulses, coherent Rabi oscillations are obtained. Assuming 100% population transfer at the maximum, one can calibrate the excitation probability as a function of the pulse area. Data points are larger than 1 SD. Main panel: The decay of the fluorescence with the number of applied pulses is measured using the same sequence as in Fig. 5a for different excitation pulse areas. The constant N0 extracted from a fit to the observed exponential decay is then multiplied by the calibrated excitation probability for each pulse amplitude to determine the cyclicity ζ. For both initializations, a reduction of ζ with increasing pulse area is observed, which indicates the presence of an excitation pulse induced spin-flip mechanism that is independent of the initial spin state (blue and orange triangles). Thus, also the single-shot readout fidelity (red, right axis) depends on the used pulse area. The highest value is obtained at higher ζ, i.e. lower power. However, also the fluorescence signal is reduced with lower power, which entails a longer readout duration and a stronger contribution of detector dark counts. Error bars: 1 SD.