Fig. 4: Exchange control in a natural-Si DQD device.
a, Energy diagram depicting two-electron spin states in the (1, 1) detuning ε regime, with pulse sequence steps overlaid. Each axis is plotted with respect to the interdot tunnel-coupling tc. b, Detuning pulse sequence including initialization (P) to the \(\left|\downarrow \uparrow \right\rangle\) state via a semi-adiabatic ramp (orange), followed by a non-adiabatic pulse (J) to near zero detuning to increase the exchange coupling J(ε) for duration τJ. c, Exchange-driven oscillations between the \(\left|\downarrow \uparrow \right\rangle\) and \(\left|\uparrow \downarrow \right\rangle\) states. The measured rf-phase response is proportional to the singlet probability. d, The corresponding fast Fourier transform amplitude AFFT. e, Ratio between exchange coupling strength J(ε) and dot-to-dot Zeeman energy difference ΔEz. f, Dephasing time \({T}_{2}^{* }\) (purple dots) extracted from the decay of the exchange oscillations and fitted with \({T}_{2}^{* }=\sqrt{2}\hslash /{\rm{\delta }}\left(h\varOmega \right)\) (orange dashed line)49,54. The shaded area shows the propagated error from the ±1 standard deviation of the fitted parameters. g, Qubit quality factor \(Q={T}_{2}^{* }\varOmega\). In e and g, each data point was extracted as a fitted parameter at a fixed detuning εJ point of c. The dataset for each fit consists of around 300 points obtained from over 64 × 103 averages. Each error bar represents the standard deviation of a fitted parameter.
