Exact Rotating Wave Approximation

TL;DR

This paper develops a systematic method to improve the rotating wave approximation for driven two-level quantum systems, enabling accurate analysis of strong, time-dependent drives by including higher-order corrections and derivatives of the drive envelope.

Contribution

The authors introduce the Magnus-Taylor expansion to derive effective Hamiltonians and kick operators, extending RWA accuracy to strong, time-dependent drives in quantum systems.

Findings

Derived a recurrence relation for effective Hamiltonians.

Included first derivative of the drive envelope in corrections.

Achieved accurate time evolution matching exact dynamics.

Abstract

The Hamiltonian of a linearly driven two-level system, or qubit, in the standard rotating frame contains non-commuting terms that oscillate at twice the drive frequency, $\omega$, rendering the task of analytically finding the qubit's time evolution nontrivial. The application of the rotating wave approximation (RWA), which is suitable only for drives whose amplitude, or envelope, $H_1(t)$, is small compared to $\omega$ and varies slowly on the time scale of $1/\omega$, yields a simple Hamiltonian that can be integrated relatively easily. We present a series of corrections to the RWA Hamiltonian in $1/\omega$, resulting in an effective Hamiltonian whose time evolution is accurate also for time-dependent drive envelopes in the regime of strong driving, i.e., for $|H_1(t)| \lesssim \omega$. By extending the Magnus expansion with the use of a Taylor series we introduce a method that we call the Magnus-Taylor expansion, which we use to derive a recurrence relation for computing the effective Hamiltonian. We then employ the same method to derive kick operators, which complete our theory for non-smooth drives. The time evolution generated by our kick operators and effective Hamiltonian, both of which depend explicitly on the envelope and its time derivatives, agrees with the exact time evolution at periodic points in time. For the leading Hamiltonian correction we obtain a term proportional to the first derivative of the envelope, which competes with the Bloch-Siegert shift.