Figure 3: Rabi oscillations.

(a) Microwave and optical pulses scheme: the frequency of the microwave is fixed at one of the microwave resonances and its duration t_MW is varied. The time interval between initialization and read-out pulses is kept constant at 210 ns, not to include the effect of the spin population decay into the measured signal. (b) Rabi oscillations between states and (green dots) and between states and (orange dots) (corresponding to the green and orange circular arrows in Fig. 2b, respectively): the measured peak ratio between the leading edges of the read-out and initialization pulses oscillates as a function of the microwave pulse duration. The error bars correspond to the s.d. of the peak ratio (the difference in error bars between the two graphs is due to different integration times (see Methods)). The green (respectively orange) solid curve is a fit based on an eight-level master equation model (Supplementary Note 5). (c) Evolution of the Rabi frequency as a function of the square root of the input microwave power P_MW, measured between states and . The error bars are smaller than dots. The grey line is a linear fit. (d) Effective Rabi frequency as a function of microwave frequency detuning, measured on transition and (orange dots). The error bars are the s.d. of the Rabi frequencies extracted from fits. The grey line is a fit of the form , where Ω is the bare Rabi frequency and δ is the detuning.