Mortality of III-V Multi-Junction Solar Cells of CPV Systems in the Real Field

26th European Photovoltaic Solar Energy Conference and Exhibition; 159-162
Failure of III-V multi-junction solar cells has been acknowledged by experienced CPV system engineers. The frequency and significance in the real field, however, were not reported. For this purpose, 30 kW system consist of 5,250 III-V solar cells and 42 strings was intensively monitored for about three years. Current sensors were applied in all DC strings and observed the change of contribution of each string to detect when failure occurs and how significant to the loss of system output power. The modules that were found containing many dead cells were removed and intensively studied. The damaged cells are found concentrating in a single production lot of receivers and also in a single sorting level, namely, highest Isc bin. This fact suggests there are certainly “weak cells” and “robust cells”. The monitoring to 42 strings reveals that it looks the cells are degraded or dead simultaneously. Another distinct fact is the dead cells are found electrically closed position, rather than physically closed position. These facts imply that some external electrical stress is the direct reason although the robustness of the cell varies. The fact that the internal resistance of CPV array circuit is low means that the Q-value of resonance circuit consist of capacitance from junctions and inductance from interconnects is high, which induces high reverse voltage to the junction due to high frequency resonance by electrical cirtuits. In order to assess the robustness to the electrical stress, a new test is developed applying high voltage spark between the ground and one of the interconnects. Although the test does not apply forward bias current to the cell junctions, the spark kills one or several cells along the interconnects as is seen in the field. The mortality was significantly improved by protecting cells against surge voltage. The cause of field failure of III-V concentrator cells was understood. The infant mortality and field failure of III-V concentrator cells was expected significantly improved by proper protection of the cells. It may be concluded it is important to protect cells from outside electrical stress, rather than expecting III-V cells to be robust and defect free.