Low-Energy Electronic Properties of Clean ${\mathrm{CaRuO}}_3$: Elusive Landau Quasiparticles

We have prepared high-quality epitaxial thin films of ${\mathrm{CaRuO}}_3$ with residual resistivity ratios up to 55. Shubnikov–de Haas oscillations in the magnetoresistance and a $T^2$ temperature dependence in the electrical resistivity only below 1.5 K, the coefficient of which is substantially suppressed in large magnetic fields, establish ${\mathrm{CaRuO}}_3$ as a Fermi liquid (FL) with an anomalously low coherence scale. At $T>1.5\mathrm{K}$ non-Fermi-liquid (NFL) behavior is found in the electrical resistivity. The high sample quality allows access to the intrinsic electronic properties via THz spectroscopy. For frequencies below 0.6 THz, the conductivity is Drude-like and can be modeled by FL concepts; for higher frequencies, non-Drude behavior is found, which is inconsistent with FL predictions. This establishes ${\mathrm{CaRuO}}_3$ as a prime example of optical NFL behavior in the THz range.

Figure 1

(a) Temperature dependence of the electrical resistivity of a 77-nm thin film of ${\mathrm{CaRuO}}_3$. (b) Topography of a $(1\times 1)\text{−}\mu {\mathrm{m}}^2$ square as determined by room-temperature STM. (c) Corresponding height profile along the direction indicated by the blue line from right to left in (b).

Figure 2

(a) SdH oscillations as obtained by subtraction of a second-order polynominal from the isothermal magnetoresistance of ${\mathrm{CaRuO}}_3$ at 50 mK for two different thin films at two different field ranges for $B\parallel (110)$. (b) Angular dependence of the oscillation frequencies (closed squares indicate additional frequency in tilted field) as determined between 25 and 33 T compared to the local-density approximation (LDA) prediction (lines). (c) Temperature dependence of the SdH amplitude for sample D12. Solid line indicates fit to the Lifshitz-Kosevich formula yielding an effective mass of $4.4m_e$.

Figure 3

(a) Electrical resistivity of ${\mathrm{CaRuO}}_3$ vs $T^{3/2}$ at zero field and a field of 9 T applied transverse to the film plane. Dotted straight line indicates $T^{3/2}$ dependence. (b) Low-temperature (between 50 mK and 4 K) electrical resistivity as $\rho -{\rho }_0(B)$ vs $T^2$ for various fields. Dotted black lines indicate Fermi liquid behavior. (c) Coefficient $A$ from $\rho -{\rho }_0=A{T}^2$ behavior vs magnetic field. (d) Phase diagram indicating the existence range of Fermi liquid $T^2$ behavior in the electrical resistivity, given by the maximal temperature below which the data and straight lines in Fig. 3 overlap within noise level. (e) Temperature dependence of the electrical resistivity exponent $d\ln (\rho -{\rho }_0)/d\ln T$ on a logarithmic temperature scale.

Figure 4

THz optical response of a 40-nm-thick ${\mathrm{CaRuO}}_3$ film grown on (110) ${\mathrm{NdGaO}}_3$. (a) Real and imaginary parts of optical conductivity at 15 K (symbols). Dashed lines are a combined Drude fit to ${\sigma }_1$ and ${\sigma }_2$. For comparison, data by Kamal et al. are plotted as full lines [7]. (b) Frequency and temperature dependence of ${\sigma }_1/{\sigma }_{\text{dc}}$, compared to FL model [37]. (c) Frequency and (d) temperature dependence of real part ${\rho }_1$ of resistivity [proportional to optical scattering rate $\mathrm{\Gamma }(\omega )$]. Dashed line in (c) is a FL prediction for 5 K based on dc data below $T_{\mathrm{FL}}$.