Dynamics of photogenerated nonequilibrium electronic states in ${\mathrm{Ar}}^{+}$-ion-irradiated ${\mathrm{SrTiO}}_3$

A metallic surface is realized on stoichiometric and insulating (100) ${\mathrm{SrTiO}}_3$ by ${\mathrm{Ar}}^{+}$-ion irradiation. The sheet carrier density and Hall mobility of the layer are $\sim 4.0\times {10}^{14}{\text{cm}}^{-2}$ and $\sim 2\times {10}^3{\text{cm}}^2/\text{Vs}$, respectively, at 15 K for the irradiation dose of $\sim 4.2\times {10}^{18}\text{ions}/{\text{cm}}^2$. These samples display ultraviolet light sensitive photoconductivity (PC) which is enhanced abruptly below the temperature ($\approx 100$ K) where ${\mathrm{SrTiO}}_3$ crystal undergoes an antiferrodistortive cubic-to-tetragonal $\left(O_h^1\to D_{4h}^{18}\right)$ structural phase transition. This behavior of PC maps well with the temperature dependence of dielectric function and electric field induced conductivity. The longevity of the PC state also shows a distinct change below $\approx 100$ K. At $T>100$ K its decay is thermally activated with an energy barrier of $\approx 36$ meV, whereas at $T<100$ K it becomes independent of temperature. We have examined the effect of electrostatic gating on the lifetime of the PC state. One nontrivial result is the ambient temperature quenching of the photoconducting state by the negative gate field. This observation opens avenues for designing a solid state photoelectric switch. The origin and lifetime of the PC state are understood in the light of field effect induced band bending, defect dynamics, and thermal relaxation processes.

Figure 1

Temperature dependence of the sheet resistance $R_{\square }$ of three STO (100) slabs irradiated at different dosage of ${\mathrm{Ar}}^{+}$ ions. The solid red lines are best fits to the expression $R_{\square }=R_0+R_1{T}^2+R_2{T}^5$, within the temperature range of 10 K to 180 K. The inset shows mobility $\left(\mu \right)$ and sheet carrier density $\left(n_{\square }\right)$ of the sample irradiated at $\sim 4.2\times {10}^{18}\text{ions}/{\text{cm}}^2$ as a function of temperature in the range 15 K to 80 K.

Figure 2

Source-drain voltage $\left(V_{\text{SD}}\right)$ as a function of current $\left(I_{\text{SD}}\right)$ at different gate fields varying from $-3 \text{kV}/\text{cm}$ to +$3 \text{kV}/\text{cm}$. The data shown in the figure are taken at 20 K for the sample of ion dose $\sim 4.2\times {10}^{18}\text{ions}/{\text{cm}}^2$. Sketch of a typical FET structure is shown in the lower inset. The electrodes are made of Ag/Cr bilayers deposited through a shadow mask. The upper inset shows a zero bias conductance at 20 and 300 K calculated from the slope of the $I_{\text{SD}}\text{−}{V}_{\text{SD}}$ curves at the origin.

Figure 3

Temperature dependence of the percentage change in the resistance of a sample with an ion dose of $\sim 4.2\times {10}^{18}\text{ions}/{\text{cm}}^2$; $\mathrm{\Delta }{R}_{E_g}(\%)=|R(E_g=0)-R(E_g=\pm 3)|/R(E_g=0)\times 100$, due to gate field is shown by black spheres with error bars in the main panel of the figure. The red open circles show Ang et al. (Ref. [37]) data for the temperature dependence of dielectric constant $\left(\epsilon \right)$ of undoped ${\mathrm{SrTiO}}_3$. The $\epsilon$ and $\mathrm{\Delta }{R}_{E_g}$ grow in parallel upon lowering the temperature. Although the positive and negative gate fields change $R_{\square }$ in the opposite manner, the magnitude of $\mathrm{\Delta }{R}_{E_g}$ in two cases differs within 10%. Inset (a) shows the normalized resistance as a function of gate field measured at several temperatures between 300 and 20 K. An illustration of band bending at the interface due to gate field is sketched in the inset (b).

Figure 4

The photoinduced percentage change in resistance as a function of temperature; $\mathrm{\Delta }{R}_P(\%)=[R_0-R(t=t_1)]/R_0\times 100$. Here $t_1$ is the time when UV light is turned off and $R_0$ is the resistance in dark (at $t=0$). The photoinduced conductivity shows an abrupt enhancement below $\sim 100$ K. Inset shows the temporal dependence of the normalized relative change in resistance under and after photoillumination, $\mathrm{\Delta }{R}_N\left(t\right)=[R_0-R\left(t\right)]/R_0$, where $R_0$ is the resistance at $t=0$ and $R\left(t\right)$ is resistance at time $t$, realized at various temperatures.

Figure 5

Time evolution of $\mathrm{\Delta }{R}_N\left(t\right)=[R_0-R\left(t\right)]/R_0$, after switching off photoexposure. The recovery curves show large persistent photoconductivity. The solid black lines are the best fits to the stretched exponential $\mathrm{\Delta }{R}_N\left(t\right)\sim exp[-{\left(\frac{t}{\tau }\right)}^{\beta }]$. Inset shows the $\mathrm{\Delta }{R}_N\left(t\right)$ profiles with fittings below $t=500$ sec on a log-log scale.

Figure 6

ln $\left(\tau \right)$ as a function of 1/T with error bars is plotted in the main graph whereas the temperature dependence of the decay exponent $\beta$ and the error bars are shown in the inset. The $T$ dependence of the decay time constant $\tau$ is fitted to the Arrhenius equation with two different activation energies well separated in temperature at $\sim 100$ K. The fits are shown as solid blue lines.

Figure 7

The recovery processes under negative gate field at 100 K. Black solid line is the photoresponse without any gate field and then the gate field of $-3$ kV/cm was applied during recovery at time $t=0 \left(t_0\right),100 \left(t_1\right),500 \left(t_2\right)$, and $1000 \left(t_3\right)$ seconds in four separate measurements, subsequently after turning off the light at $t_0$. The gate field applied at $t=t_0,t_1,t_2$, and $t_3$ was switched off at $t=t_0^{\prime },t_1^{\prime },t_2^{\prime }$, and $t_3^{\prime }$, respectively. Switching on the gate field (at $t=t_n$, where $n=0$ to 3) leads to an abrupt increase in resistance, while there is a sudden drop in it when the gate field is turned off (at $t=t_n^{\prime }$). The time scale of abrupt change in resistance is of the order of milliseconds.

Figure 8

The acceleration in recovery processes by negative gate field at 300 K. (a) The gate field of $-3$ kV/cm applied during recovery at time $t=0 \left(t_0\right)$ and left on until $t_0^{\prime }$. The $R_{\square }$ marginally exceeds its value in dark during $(t_0^{\prime }-t_0)$ seconds. The sample is allowed to recover its $R_{\square }$ in dark after turning off the gate field and then reexposed under the same photon flux as earlier. This process is repeated for several values of $t_n$. (b) The dependence of the recovery process on the strength of the negative gate field. The fields were applied at the time of switching off the light as marked by the dotted line. The dashed lines indicate the point of turning off the field.

Figure 9

Defect configuration-coordinate diagram indicating the photoexcitation of an electron by energy $E_O$ from the defect into the conduction band, which is then captured in a thermal barrier of energy $E_T$. In order to relax in its ground state, the electron must have enough energy to surmount the barrier. The $E_B$ denotes the net binding energy gained in the process of large lattice relaxation.