Unusual paraelectric system, KBr:Li+

We have measured the paraelectric resonance absorption of lithium-doped KBr over the frequency range 42-217 GHz with electric fields applied along the $〈100〉$, $〈111〉$, and $〈110〉$ orientations. Many lines are observed, with zero-field splittings (ZFS) ranging from 52 to 161 GHz for the ${}_{}{}^7{}_{}{}^{}\mathrm{Li}_{}^{+}{}_{}{}^{}$ isotope. When ${}_{}{}^6{}_{}{}^{}\mathrm{Li}_{}^{+}{}_{}{}^{}$ replaces ${}_{}{}^7{}_{}{}^{}\mathrm{Li}_{}^{+}{}_{}{}^{}$, a large isotope shift of the ZFS (typically about 40%) is observed, indicating that ${\mathrm{Li}}^{+}$ is tunneling in some multiwell potential. This paraelectric system is unusual in several respects: no significant changes in the relative line intensities are observed between 4.2 and 1.4 K, impurities such as O${\mathrm{H}}^{-}$ can suppress the observed signal, and quenching low-concentration samples increases the signal intensity by factors of 10-1000. We attempted to fit the data to the standard tunneling models and had limited success with the $〈110〉$ model; several (but not all) strong lines in each field orientation could be fit very well. However, when both the weak and strong lines are considered, too many ZFS are observed, which suggests that this ${\mathrm{Li}}^{+}$ system is far too complex to be explained by any of the standard tunneling models. A comparison with other experiments and some suggestions for future work are included.