Emergence of large thermal noise close to a temperature-driven metal-insulator transition

Abstract

We report that close to a Mott transition there is an emergence of large thermal noise (${S}_{\mathrm{th}}$) which occurs concomitantly with large correlated flicker noise ($\frac{1}{f}$ noise) with significant non-Gaussian content. This was observed in films of ${\text{NdNiO}}_{3}$ (thickness 15 nm) grown on crystalline ${\text{SrTiO}}_{3}$ substrates with different crystallographic orientations that show a hysteretic transition from a high temperature metallic phase to a low temperature insulating phase in the temperature range 160 to 211 K depending on the substrate orientation and the heating and cooling cycle. The thermal noise, which is distinct from the flicker noise, deviates from the canonical Johnson-Nyquist value of $4{k}_{B}TR$ as measured through the ratio $\ensuremath{\zeta}(T)(\ensuremath{\equiv}\frac{{S}_{\mathrm{th}}(T)}{4{k}_{B}TR})$. The ratio $\ensuremath{\zeta}$ reaches a maximum value of ${\ensuremath{\zeta}}_{M}$ at a temperature ${T}^{*}$ that is close to but distinct from the metal-insulator transition (MIT) temperature ${T}_{\mathrm{MI}}$. In all the films near ${T}^{*}$, the scaled thermal noise maxima ${\ensuremath{\zeta}}_{M}\ensuremath{\gg}1$. The films were found to be largely strain relaxed with residual in-plane and out-of-plane strain as measured by x-ray reciprocal space mapping. It has been observed that the ratio $\frac{{T}^{*}}{{T}_{\mathrm{MI}}}$ as well as ${\ensuremath{\zeta}}_{M}$ have a close dependence on the in-plane-strain in the film. The enhanced thermal noise that occurs along with large correlated flicker noise both arise from slow kinetics of relaxation as established from temperature dependence of the correlation time ($\ensuremath{\tau}$) that gets significantly larger in the temperature range around ${T}^{*}$, reaching a maxima at $T={T}^{*}$. It has been proposed that the existence of large noise (both thermal and flicker noise) owes its origin to electronic phase separation (EPS) that exists near the MIT. A physical model has been suggested that EPS near MIT temperature can give rise to a sparse phase of nanometric small pockets of metallic phases (nanopuddles) that are surrounded by and embedded within the minority insulating phase. The nanopuddles act as a source of charge fluctuations and are coupled weakly to the majority metallic phase by tunneling through the layer of the minority insulating phase. Such isolated metallic nanopuddles can be Coulomb charged if the charging energy ${E}_{C}\ensuremath{\ge}{k}_{B}T$ and can have slow relaxation of fluctuations acting as a source of large noise. It has been argued that the size distribution of the nanopuddles, their average size $\ensuremath{\langle}d\ensuremath{\rangle}$, as well as the temperature dependence of their number density ${N}_{d}$ can determine the temperature ${T}^{*}$.