The electrical properties of Cu(In,Ga)Se 2 /Mo junctions were characterized with respect of MoSe 2 orientation and Na doping level using an inverse transmission line method, in which the Cu(In,Ga)Se 2 (CIGS)/Mo contact resistance could be measured separately from the CIGS film sheet resistance. The MoSe 2 orientation was controlled by varying the Mo surface density, with the c-axis parallel and normal orientations favored on Mo surfaces of lower and higher density, respectively. The effect of Na doping was compared by using samples with and without a SiO x film on sodalime glass. The conversion of the MoSe 2 orientation from c-axis normal to parallel produced a twofold reduction in CIGS/Mo contact resistance. Measurements of the contact resistances as a function of temperature showed that the difference in CIGS/Mo contact resistance between the samples with different MoSe 2 orientations was due to different barrier heights at the back contact. Comparison between Na-doped and Na-reduced samples revealed that the contact resistance for the Na-reduced system was four times of that of the doped sample, which showed more pronounced Schottky-junction behavior at lower temperature, indicating that Na doping effectively reduced the barrier height at the back contact.

No abstract

Abnormal grain growth (AGG), which occurred during the heat treatment of Pb(Mg 1/3 Nb 2/3 )O 3 -35 mol% PbTiO 3 (PMN-35PT) with excess PbO, was investigated. AGG has been suggested to be the consequence of grain coalescence that results in the formation of ⌺3 coincidence site lattice and low angle grain boundaries. Because of reentrant edges appearing at the ends of these boundaries, the coarsening rate of grains was significantly enhanced and AGG occurred.

Among many artificial neural networks, the research on Spike Neural Network (SNN), which mimics the energy-efficient signal system in the brain, is drawing much attention. Memristor is a promising candidate as a synaptic component for hardware implementation of SNN, but several non-ideal device properties are making it challengeable. In this work, we conducted an SNN simulation by adding a device model with a non-linear weight update to test the impact on SNN performance. We found that SNN has a strong tolerance for the device non-linearity and the network can keep the accuracy high if a device meets one of the two conditions: 1. symmetric LTP and LTD curves and 2. positive non-linearity factors for both LTP and LTD. The reason was analyzed in terms of the balance between network parameters as well as the variability of weight. The results are considered to be a piece of useful prior information for the future implementation of emerging device-based neuromorphic hardware.