An approximate analysis of losses in the diode as it is used in the rectenna element also indicated that the capacitance across the junction should be reduced if the diode were to be operated at higher impedance level. Two major modifications in the diode were therefore indicated if it were to be used at a lower power level - a decrease in the junction capacitance and in the voltage drop across the junction. Another area of interest was the further progressive reduction of the epitaxial layer thickness into the "punch through" region to establish the impact of this upon the efficiency of the diode. A search of other possible junctions reveals that the Schottky barrier voltage for a GaAs-W barrier is about 25% lower than that of the GaAs-Pt barrier. An objective of this contract was to build some diodes with the GaAs-W barrier and to evaluate them. Now tungsten is a desirable material from most viewpoints but it is more difficult to get it to bond to GaAs than it is platinum. Fortunately, the development effort concerned with getting a satisfactory mechanical and a satisfactory electrical bond between GaAs and Tungsten was the subject of an ongoing effort supported by another contract. The diode development work for this contract benefited from that experience. The voltage current characteristic of a GaAs-W diode constructed under this contract is compared with that of a GaAs-Pt diode specified for the RXCV element. ( Figure 4-1. ) Both diodes had approximately the same reverse breakdown voltage and the same junction capacitance so a direct comparison can be made. The GaAs-Pt diode was a diode that had been used as a standard during the construction of the JPL RXCV rectenna. It is noted that the slope of the voltage-current characteristic in the forward direction of the two diodes are essentially the same, indicating that the series resistance of the diodes are very similar. The essential difference in the two diodes is in the Schottky barrier voltage and that this difference as measured by the intercept of the tangents of the voltagecurrent characteristic with the zero-current axis is about 0.2 volts. This is the value that one would expect. The impact that the reduced junction voltage would have upon improvement in efficiency of the rectenna element was examined by putting the diodes in the ground-plane test fixture ( Table 4-1) and then running a comparison test of overall efficiency and diode losses. The results indicate that the reduction in diode losses are just about what would be anticipated from the reduction in the Schottky-barrier voltage. The major importance of the comparison tests between the GaAs-W and the GaAs-Pt diodes is to confirm what was expected - a slight improvement in efficiency at the higher power output levels with a greater improvement as the DC voltage output was reduced. There were no surprises. As we sought to improve the efficiency at the lower microwave input levels, however, some unexpected difficulties were encountered. As the DC voltage on the diode is increased for a given input power by using a higher value of DC load resistance, the loss in the series resistance during the conduction period goes down while the loss in the series resistance during the non-conduction period increases rapidly due to the charging current flowing into the junction capacitance.
RkJQdWJsaXNoZXIy MTU5NjU0Mg==