Strong Affinity between In and Al Impurities Doped in ZnO

Abstract

In recent years, zinc oxide (ZnO) doped with group 13 elements (Al, Ga, In) as impurity donors is strongly expected for application to functional devices as n-type II–VI compound semiconductors. It is known that the electric conductivity of the semiconductor drastically changes depending on the type, concentration, and introduction methods of impurities. For the precise control of conductivity, therefore, it is essential to investigate the physical and chemical states of the impurities. With respect to the study of the electromagnetic field in the vicinity of dilute impurities, nuclear techniques with radioactive probes are very suited because of their high sensitivity. Taking this advantage, we have applied the timedifferential perturbed angular correlation (TDPAC) method to the investigation of local fields at donor sites in ZnO, employing Cd disintegrated from In, symbolized as 111Cdð 111InÞ hereafter, as the nuclear probe expecting the behavior of In as a donor impurity. In our previous TDPAC studies, we found that local fields at the probe nucleus in ZnO doped with dilute In, Ga and Al ions are distinctly different from that for undoped ZnO. Especially for the Al-doped ZnO, in addition, the doping effect on the 111Cdð 111InÞ probe is prominent even at as low Al concentration as 0.05 at.% (500 ppm). This observation implies a high probability for the local association of the 111Cdð 111InÞ probe and Al ion(s) as a result of their thermal diffusion in ZnO. Directing our interest to the probability of their association, we investigated the dilute limit of Al concentration at which their association presumed above can be observed. The present paper proves the presence of a strong attractive force between In and Al ion(s) in ZnO matrix by showing an experimental evidence that the 111Cdð 111InÞ probe can be locally associated with extremely dilute Al ion(s). Zinc oxide samples doped with various Al concentrations were separately synthesized by a solid-state reaction following our previous work. For the synthesized samples, we found from powder X-ray diffraction patterns that the lattice constants were unchanged from pure ZnO even for the 5000 ppm at.% Al-doped ZnO and new phases were not detected. TDPAC measurements were performed for the 111Cdð 111InÞ probe doped in each sample at the concentration of 100 ppt on the cascade rays with the intermediate state of I 1⁄4 5=2 having a half-life of 85.0 ns. In the present work, observation of TDPAC was made for the directional anisotropy, A22G22ðtÞ, as a function of the time interval between the cascade -ray emissions, t, during which the probe is perturbed by the outer surrounding field. Here, A22 denotes the angular correlation coefficient depending only on the nuclear properties and G22ðtÞ is the time-differential perturbation factor. For easy understanding of the effect of the Al doping on the present perturbation pattern, the TDPAC spectra of 111Cdð 111InÞ in undoped and 500 ppm Al-doped ZnO are firstly cited in Fig. 1 from our previous studies. The spectrum for the undoped ZnO in Fig. 1(a) exhibits an oscillatory structure typical of the electric quadrupole interaction between the probe nucleus and the extranuclear field (hereafter denominated Component A). Although the oscillating pattern reproducible with a single component is not clearly discernible due to the damped structure reflecting a wide distribution of the field at the probe, the static perturbation pattern for the Al-doped ZnO in Fig. 1(b) (denominated Component B) is also ascribable to the electric quadrupole interaction because the sample consists of no magnetic materials; accordingly, the distinct perturbation should be attributed to the change in the charge distribution around the probe by the Al doping. We thus performed least squares fits to the spectra in Fig. 1 with G22ðtÞ expressed as G22ðtÞ 1⁄4 G 22ðtÞ 2;0 þ X3