Speeding up Adiabatic Quantum State Transfer by Using Dressed States

We develop new pulse schemes to significantly speed up adiabatic state transfer protocols. Our general strategy involves adding corrections to an initial control Hamiltonian that harness nonadiabatic transitions. These corrections define a set of dressed states that the system follows exactly during the state transfer. We apply this approach to stimulated Raman adiabatic passage protocols and show that a suitable choice of dressed states allows one to design fast protocols that do not require additional couplings, while simultaneously minimizing the occupancy of the “intermediate” level.

Figure 1

(a) Schematic of a composite quantum system where the source and the target systems (qubits in this schematic) are coupled via some intermediate system. (b) Schematic of the possible evolutions: (red line) perfect adiabatic evolution; (blue line) speeding up the evolution results in nonadiabatic errors leading to an imperfect state transfer; (green line) by dressing the adiabatic eigenstates it is possible to design an evolution that leads to a perfect state transfer.

Figure 2

(a) Comparison of the residual error between STIRAP [Eq. (19)], SATD [Eq. (20)], and the modified SATD [Eq. (21)] as a function of the effective protocol duration $\tau$ in units of ${\tau }_{\min }$. (b) Comparison of the integrated population in $|B⟩$ over the whole protocol time between SATD [Eq. (20)] and our new dressed state approach [Eq. (21)] as a function of $\tau$ in units of ${\tau }_{\min }$. Inset: ratio of those two quantities. The integrated population is reduced by at least 21% and at most 26% with our new protocol. Plot of the corrected pump pulse for SATD (c) and the modified SATD (d) for different values of $\tau$ as a function of time $(t-t_i)$ in units of the total protocol time $(t_f-t_i)$.

Figure 3

(a) Comparison of the residual error for STIRAP with Gaussian densities [Eq. (23)] and the modified SATD [Eq. (24)] as a function of the effective protocol duration $\tau$ in units of ${\tau }_{\min }$. The residual error is reduced by several orders of magnitude in the nonadiabatic regime. (b) Corrected pump pulse for different values of $\tau$ as a function of time $(t-t_i)$ in units of the total protocol time $(t_f-t_i)$.