Oscillatory free-induction decay

Large-amplitude oscillations are predicted in the free-induction decay (FID) of atomic or spin two-level quantum systems that are coherently prepared by a resonant electromagnetic field in the form of a square amplitude pulse of duration $T$. The effect should be detectable in nuclear magnetic resonance (NMR) or in optical resonance experiments when the transition line shape is inhomogeneously broadened ($T_2^{*}\ll T_2$) and the applied field is intense enough that the pulse area $\chi T\ge 2\pi$, where $\chi$ is the Rabi frequency. Physically, the effect is due to Rabi oscillations which are generated in the preparative stage and are then reproduced because of atomic memory in the radiative period which follows. The atomic polarization derived from the Bloch equations assumes the form of an integral over the inhomogeneous line shape, and can be evaluated numerically for arbitrary linewidth $\sigma$ or analytically when $\sigma \to \infty$. The characteristics of the oscillation are unusual since the oscillation frequency, which is of order $\chi$, increases with time while the oscillation envelope vanishes identically for times $t\ge 2T$. Coherent emission is therefore confined to one pulse width $T$ immediately following the pulse. Previous NMR and infrared FID experiments with small-area pulses ($\chi T\le 4\pi$), interpreted erroneously in terms of an edge echo effect, support the oscillatory FID behavior described here. For the case $\chi T>\mathrm{}4\mathrm{}\pi$, an experimental test would be even more decisive, as the oscillations are more pronounced.