Stimulated-Dielectric-Relaxation Currents in Thin Film Al-CeF3-Al Samples

A new technique, based on stimulated-dielectric-relaxation currents (SDRC), has been successfully used to analyze the defect properties of the Al-Ce${\mathrm{F}}_3$-Al capacitor system. The SDRC technique consists of cooling the system, with a voltage bias $V_d$ applied to the sample electrodes. The voltage bias is then changed to, say, $V_i$, while the sample is at the low temperature, and the current-temperature ($I-T$) characteristics are measured while the sample is being heated at a uniform rate. When $V_d\ne V_i$, the system is in the non-steady-state, and pronounced structure is observed in the $I-T$ characteristics. Depending on the relative magnitudes of $V_d$ and $V_i$, the $I-T$ characteristic exhibits either a single peak or two peaks, and the current can be either positive or negative, or negative at low temperatures and then positive at higher temperatures. However, when $V_d=V_i$, the system is in the steady state at all times, and no such structure in $I-T$ curves is observed. All the measurements are explained on the basis of the SDRC theory presented in the previous papers; in fact, all the salient features of the SDRC theory have been observed. The SDRC characteristics are related to the properties of the system, such as, the trap depth, trapping density, and barrier heights, all of which have been determined. The interfacial barrier heights at the two interfaces are determined to be about 1 eV at the lower electrode and 0.95 eV at the counter electrode. The trap depth and trap density are found to be about 0.68 eV and 3×${10}^{20}$ ${\mathrm{cm}}^{-3\mathrm{}}$, respectively. The energy spread of the occupied trap level is about 1.9×${10}^{-2\mathrm{}}$ eV. It is also found that an electrode-limited-to-bulk-limited transition in the conduction process occurs in these samples at about 3 V. Previous measurements on thin-film insulators are discussed in light of the present observations and other (unpublished) data obtained by the authors. It is concluded that the analyses of previous data should be viewed with caution.