A novel coherent Raman effect induced by a laser and a radio-frequency (rf) field is used to detect cw and pulsed nuclear magnetic resonance (NMR) in ground and excited electronic states. The effect is illustrated in the impurity-ion solid
Pr3+: La
F3 at 1.6 K utilizing the
Pr3+ optical transition $^{3}H_{4}({\ensuremath{\Gamma}}_{1})\ensuremath{\rightarrow}^{1}D_{2}({\ensuremath{\Gamma}}_{1})$. The laser field of frequency ${\ensuremath{\omega}}_{E}$ and the rf field (${\ensuremath{\omega}}_{H}$) induce a light wave at the sum ${\ensuremath{\omega}}_{E}+{\ensuremath{\omega}}_{H}$ (anti-Stokes) and difference ${\ensuremath{\omega}}_{E}\ensuremath{-}{\ensuremath{\omega}}_{H}$ (Stokes) frequencies, generating an absorptive or dispersive heterodyne beat signal (${\ensuremath{\omega}}_{H}$) with the laser field at a photodetector. The theory of this effect is characterized in a new three-level perturbation calculation which requires, unlike the usual stimulated Raman effect, that all three transitions be electric- or magnetic-dipole allowed. Detailed predictions are confirmed by cw measurements of the
Pr3+: La
F3 hyperfine splittings where the optical heterodyne signals are shot-noise limited. The
Pr3+ nuclear quadrupole parameters are obtained for the
3H4 and
1D2 states where the line centers are determined with kilohertz precision. The corresponding wave functions show significant hyperfine-state mixing, as required for all three transitions to be dipole allowed. The cw line shapes are narrow (30-160 kHz), inhomogeneously broadened by nuclear magnetic interactions, and reveal either a Gaussian or an anomalous second-derivative---like line shape. The spin-echo measurements for the
3H4 and
1D2 hyperfine transitions yield homogeneous line shapes which are Lorentzian, and rather surprisingly, linewidths in the narrow range 10-20 kHz, a result which tests current line-broadening theories.
An optical heterodyne technique based on the coherent Raman effect is demonstrated for detecting nuclear magnetic resonances (NMR) and coherent spin transients in normal and optically excited impurity-ion solids at low temperature. Initial measurements on
Pr3+: La
F3 provide the first spin-echo studies of an electronically excited state, and cw resonances yield NMR line centers and shapes with kilohertz precision.
Laser and microwave fields can be mixed in paramagnetic samples in a coherent Raman process. The radiation scattered in the forward direction contains frequency components at the sum and difference frequency. We consider the details of this scattering process in a static magnetic field, where two types of signals appear: magnetic-resonance transitions give rise to signals at fixed magnetic fields, independent of the laser frequency. Another set of resonances is determined by the optical transition frequencies. These resonances are particularly prominent in samples with narrow optical absorption lines. We derive a comprehensive theory that describes both contributions and compare it with experimental data from the R lines of ruby.
Raman heterodyne detection of nuclear magnetic resonance is a sensitive technique for optical detection of an NMR transition, which is driven by a resonant radio-frequency field. The observed signal is linear in the transition matrix elements of the magnetic resonance transition and two optical transitions. This linearity is the reason that signal contributions from different sites or different rf transitions can interfere, in many cases destructively. In this article, we discuss such an interference effect, which appears to have been overlooked in the past. It occurs between signal contributions that originate from the same magnetic resonance transition of atoms with different positions within the inhomogeneously broadened optical resonance line. These atoms contribute to the coherent Raman scattering through different scattering paths, which involve optical transitions to different nuclear spin states of the same electronic state. We show that the interference between all possible scattering paths leads to complete signal cancellation, if the atoms that are involved in the different scattering paths are equally polarized. To study the effect and to eliminate the interference, we used a pump-and-probe technique with two laser beams. With two independent laser frequencies, it becomes possible to separate the individual scattering paths. We calculate the dependence of the signal on the frequency of both laser beams as well as on the radio frequency and compare the results to experimental data from
Pr3+:
YAlO3. Our results show that the interference reduces the signal amplitude of the conventional Raman heterodyne experiment but can be eliminated in the new experiment.
Raman-heterodyne-detected multipulse rf nuclear-spin-echo decay of
Pr3+ in
YAlO3 is measured with continuous-detection light throughout the spin-locking pulse train and with the light on during a short period centered at the echo time only for every sixth echo. The echo decay shows three components: an initial fast decay (\ensuremath{\simeq}400 \ensuremath{\mu}s), an intermediate decay (5.4\ifmmode\pm\else\textpm\fi{}0.6 ms), and a slower decay of 22.5\ifmmode\pm\else\textpm\fi{}1.7 ms for gated echoes. For cw light, the decay is described by the same initial fast decay, but a single decay at 7.6\ifmmode\pm\else\textpm\fi{}0.6 ms is observed at longer times. The difference in the observed echo decay is attributed to optical pumping effects. A rate-equation model is constructed, consisting of three ground-state levels and three excited-state levels. It includes radiative decay, spin-lattice relaxation, the optical pumping and rf pulses, and rf coherence loss due to excited-state occupancy. The model shows three decay times, fast (\ensuremath{\simeq}400 \ensuremath{\mu}s), intermediate (6.08\ifmmode\pm\else\textpm\fi{}0.07 ms), and slow (21.0\ifmmode\pm\else\textpm\fi{}0.1 ms) decay for gated echoes. The last two rates are affected by the strength and duration of the optical field. For continuous light, the calculated echo is seen to decay at \ensuremath{\simeq}400 \ensuremath{\mu}s, 2.04\ifmmode\pm\else\textpm\fi{}0.07 ms, and 3.43\ifmmode\pm\else\textpm\fi{}0.02 ms and then increases at a much slower rate (100 ms). This is only an apparent increase, because significant signal cancellation occurs in the sum of the six ion groups contributing to the signal. If a reduced light level and faster spin-lattice relaxation rates are assumed, the calculated echo-decay times for cw light are 2.44\ifmmode\pm\else\textpm\fi{}0.03 and 7.37\ifmmode\pm\else\textpm\fi{}0.04 ms. This corresponds to a homogeneous linewidth increase of 2.5 times and a spin-lattice relaxation rate increase of 2.5 times equivalent to crystal heating by the cw detecting light of 0.35 K over that of the gated light.