Yasuhiro Ohbuchi

Evaluation of Interface States in ZnO Varistors by Spectral Analysis of Deep Level Transient Spectroscopy

Interface states in ZnO varistors were studied using spectral analysis of deep level transient spectroscopy (SADLTS) to obtain the emission rate spectrum. We found one interface state in ZnO with the activation energy and the capture cross section distributed around their central values, E0=0.98 eV and σ0=1.8×10-16 cm2, over widths ΔE=0.19 eV and Δσ=2.9×10-17 cm2, respectively.

Photoacoustic Spectra of ZnO-CoO Alloy Semiconductors

Photoacoustic spectra (PA) of Zn1-xCoxO alloy semiconductors sintered as powdered samples were measured to estimate the optical absorption. Photoacoustic spectroscopy (PAS) is insensitive to the light scattering effects on the samples. CoO molecular percentages in the alloy were changed from 0 to 20 mol%. The samples were sintered at 600 or 950°C for 24 h in a quartz glass tube. From the PA spectra in the short wavelength region, we plotted (ρh ν)2 vs h ν and estimated the band-gap energy (Eg) of the alloy semiconductors, where ρ is the PA signal intensity and h ν is the excitation light photon energy. Eg of ZnO-CoO alloy decreases as the CoO content increase and reaches to 2.19 eV in the 20 mol% CoO sample. Although the band-gap variation in the lower-CoO-content samples (CoO content below 10 mol%) is large, it becomes small in the high CoO concentration samples. This result concurred with the result of dissolving CoO in ZnO as obtained by X-ray diffraction measurement.

Characterization of Interface States in Degraded ZnO Varistors.

The degradation mechanism of ZnO varistors was investigated by using spectral analysis of deep-level transient spectroscopy (SADLTS), and using the current–voltage (I–V) characteristics and the capacitance–voltage (C–V) characteristics before and after AC biasing stress. The interface states in the Pr-type ZnO varistors consist of two adjacent levels: T0 and T1 levels. The activation energy ET and the capture cross section σ of the interface states decrease with increasing degradation induced by AC bias stress. The decrease in ET and σ is related to a decrease in the barrier height caused by the degradation. In order to characterize the change in the interface states with degradation, distribution parameters (ΔET, Δσ) are introduced, where the parameters for each level increase after degradation. The degradation is closely related to the distribution of the emission rate of the interface states. The degradation phenomenon is caused by migrated ions or by adsorbed ions such as oxygen. The increase in the distribution parameters upon degradation suggests activation of the oxygen ions at the grain boundaries and nonuniform migration.

Photoacoustic spectra of Sb-doped ZnSe

Photoacoustic (PA) spectra were made on Sb-doped ZnSe samples grown by metalorganic vapor phase epitaxy (MOVPE). The samples deposited at a low temperature under irradiation show the PA spectrum with a sharp edge near the band gap and with three distinct peaks. From the Sb flow rate dependence of PA peaks, two of them seem to be related to Sb impurities. Non-doped sample shows only one peak, which is tentatively ascribed to the deep level associated with Se defects.

The Study of Interface States in ZnO Varistors by Injection Pulse Width Dependence of Transient Response

The injection pulse width dependence of transient response was studied to investigate the interface states in the Bi- and Pr-type ZnO varistors in more detail using isothermal capacitance transient spectroscopy (ICTS). Although the interface states have been considered to be distributed monoenergetically or discretely, two or three different interface states of the emission process were confirmed by varying the injection pulse width at each measurement temperature. For both types of ZnO varistors, the ICTS spectrum overlapping of the interface states consists of two transient responses: the previously reported trap (Trap 1) and Trap 0 which has a faster emission process than Trap 1. For only Bi-type ZnO, Trap 2 which has a slower emission process, was successfully detected at the longer region of the injection pulse widths tW from 1 s to 100 s. This result suggests that the formation of Trap 2 can be attributed to the existence of Bi2O3. In the case of the application of ICTS for the interface states in ZnO varistors, it is necessary to select the optimal injection pulse width which takes into account the emission processes at each measurement temperature.

Evaluation of ZnO Varistors by Photoacoustic Spectroscopy

ZnO varistors degraded under various conditions were evaluated by photoacoustic spectroscopy (PAS). The degradation where the grain boundary is damaged by DC bias stress is more than that by AC bias stress. PAS, however, reveals that the interior of the grains of the sample degraded by AC bias stress is much more damaged than that by the DC bias stress. The PA signal intensity at a wavelength of more than 500 nm increases and the dispersion of the spectrum decreases throughout the wide wavelength range considered with the degradation time. In particular, the decrease of the spectral dispersion below 500 nm is caused by the change of the electronic states at the interface, that is, the increase of the recombination center of the space charge. The annealing effect on the degradation of ZnO varistors was also studied. The PA spectrum of the sample annealed in N2 gas corresponds to those of the sample degraded by DC and AC bias stress for a long time. This suggests that the degradation of ZnO varistors is closely related to the release of oxygen from both ZnO grain interior and grain boundaries.

Deep levels in Sb-doped ZnSe fabricated by metalorganic vapor-phase epitaxy

Abstract Sb-doped ZnSe samples were deposited on the (001)GaAs substrate by metalorganic vapor-phase epitaxy (MOVPE). Isothermal capacitance transient spectroscopy (ICTS) and spectral analysis of deep-level transient spectroscopy (SADLTS) were used to characterize deep levels of Sb-doped ZnSe. The p-type sample grown by MOVPE at 490°C in the darkness shows three ICTS peaks. Three deep levels were observed in the N-doped ZnSe deposited by MOVPE. Using the SADLTS, we can estimate the activation energy and the capture-cross section distributions of that hole traps. We also examined samples that were photoassist-deposited at lower temperature. The non-doped ZnSe thin films were also measured to check the effects of light irradiation during the deposition. We could get only n-type samples and the light irradiation generates the new level of the electron traps. Sb doping generates other new levels. The levels that correspond to trap E1 in the light-irradiated Sb-doped samples are constructed from two adjacent levels in SADLTS, and one new level near trap E1 can be observed in SADLTS.

Evaluation of Sb2O3-doped ZnO varistors by photoacoustic spectroscopy

Photoacoustic (PA) spectra for the Sb2O3-doped ZnO varistors were measured to study the Sb2O3 doping effects on their nonradiative properties and deep levels, where the Sb2O3-doped ZnO varistors had the highly nonohmic current-voltage feature. For the samples without the Sb2O3 doping, the PA spectra are almost flat between 750 nm and 950 nm. When the ZnO varistors were doped with Sb2O3, a decrease in the intensities of the PA signals above 800 nm with increasing wavelength is observed. The Sb2O3 doping generates the spinel phase, which suppresses the concentration of Co dissolution into the ZnO grains; the PA peaks of Co in ZnO observed between 800 nm and 1000 nm appear to be low. The small peak at approximately 950 nm is observed for the Sb2O3-doped samples. This peak corresponded to the deep level caused by the Sb2O3 doping. The analysis of this level helps in the estimation of the properties of ZnO varistors.

Pinning Effect by Interface States in Pr-type ZnO Varistors

The interface states in Pr-type ZnO varistors, which consist of two adjacent levels, T1 and T0, were studied using spectral analysis of deep-level transient spectroscopy (SADLTS) to characterize the pinning effect of the Fermi level by two interface states and the bias dependence on the emission process of the interface states. The measurements were carried out by changing the applied steady-bias voltage for the injection pulse. When the steady-bias voltage is below 60% of the breakdown voltage, the obtained values of the activation energy ET and the capture cross section σ of the T1 level are almost constant. The Fermi level is perfectly pinned by the T1 level. In contrast, the Fermi level is never pinned by the T0 level since ET and σ of the T0 level vary with increasing steady-bias voltage and the T0 level has a faster emission process.