Zero-Bias Tunneling Anomalies—Temperature, Voltage, and Magnetic Field Dependence

The change of the "zero-bias tunneling anomaly" in $\mathrm{T}\mathrm{a}-I-\mathrm{A}\mathrm{l}$, $\mathrm{S}\mathrm{n}-I-\mathrm{S}\mathrm{n}$, and silicon $p-n$ junctions in magnetic fields of 0-40 kG has been measured from 4.2 to 1°K ($I$=insulator). The strongest magnetic field dependence was observed in $\mathrm{S}\mathrm{n}-I-\mathrm{S}\mathrm{n}$ junctions, where the conductance peak is depressed and splits into two peaks when the field increases from 0 to 40 kG at 1.5°K. The results have been compared with the theoretical model proposed by Appelbaum and Anderson, who calculated the interaction between the tunneling electrons and magnetic impurities at and near the boundary of the insulating layer. The data indicate that the second-order term in the theory is sufficient to fit the magnetic field dependence, and the third-order term fits the conductance peak at zero bias. Measurements in a magnetic field thus determine the $g$ and $S$ values of the magnetic impurities. We have found that the number of impurities trapped inside $\mathrm{S}\mathrm{n}-I-\mathrm{S}\mathrm{n}$ junctions can vary with the method of preparation. The magnitude of the zero-bias anomaly, the superconducting $I-V$ characteristics for $V<\mathrm{}2\mathrm{}{\Delta }_0$, and the background conductance at very high bias are all affected by the number of impurities present. Giant anomalies in $\mathrm{C}\mathrm{r}-I-\mathrm{A}\mathrm{g}$ and $\mathrm{C}\mathrm{r}-I-\mathrm{P}\mathrm{b}$ junctions illustrate a different effect in which a very strong field-independent conductance dip is observed at zero bias. We conclude that small conductance peaks are explained by the Appelbaum-Anderson model, but the giant resistance anomalies are unexplained at present.