Unusual low-frequency noise in irradiated ${\text{SrTiO}}_3$

We report the transport and noise properties as function of temperature and electric field of ${\text{Ar}}^{+}$ irradiated ${\text{SrTiO}}_3$ exhibiting at its surface a high-mobility electron gas. This system exhibits metallic properties with low-charge densities allowing mobility measurements. The low-frequency $1/f$ noise presents a linear dependence with the mobility. This nonintuitive variation requires a rewriting of the commonly used Hooge model.

Figure 1

(Color online) Sheet resistance at room temperature as a function of the irradiation time. The improvement of the $R_{\text{s}}$ with the irradiation time might be related to an increment of the number of oxygen vacancies created. Inset: sketch showing the position of the contacts in the conducting STO line. Current is applied between contacts 1–2.

Figure 2

Dependence of the mobility as a function of the temperature in a $50\text{−}\mu \text{m}$-large sample irradiated during 10 min. The mobility increase monotonically up to about 10 K and then saturates. In the tested samples there is not a clear dependence on the Ar etching time and the low-temperature mobility. In the right axis is shown the dependence of the number of carriers per surface unit on the temperature for the same sample.

Figure 3

(Color online) (a) The ratio between the Hooge parameter and the surface density of carriers is plotted as function of the sample temperature. Two features could be remarked: first the peak located at $T\approx 28\text{K}$ and increasing of the plotted parameter when $T$ is diminished. The inset shows a zoom to the peak observed at about 30 K in a $50\text{−}\mu \text{m}$-width sample irradiated 15 min. (b) $\alpha$ parameter versus the mobility of a single sample changing the temperature. The dashed line is a linear fit in linear scale, we use instead a logarithmic scale in order to make the plot more visible. The inset shows the same graph in linear scale.

Figure 4

(a) Dependence of $\alpha$ parameter on the mobility when a backgate voltage from $-250$ up to 250 V is applied at 5 K. The linear dependence obtained rule further support our hypothesis that there is an intrinsic relationship between the $1/f$ noise and the mobility in the presented samples. (b) We plot $\gamma$ [Eq. (5)] at different frequencies as function of the temperature. In all the range of temperatures $\gamma$ depends slightly on frequency. The increase in $\gamma$ for $T<50\text{K}$ rules out the DDH model (see text).