-
StatusOngoing
-
Status date2011-11-22
In the light of a move to solar cells with several junctions and ever increasing efficiency in the space photovoltaic area, mechanical defects in multijunction cells are a topic of some concern. On the one hand, a cell crack that does propagate during the in orbit life can have a significant impact on the reliability of the array due to the large voltage contribution of each cell. On the other hand, a cell replaced unnecessarily represents a large cost factor. Up to now, however, little work has been performed in this area, mostly for Si solar cells.
The aim of the CIEL (Cell Inspection Criteria based on Electroluminescence) project was therefore to study experimentally on a large sample basis the evolution of cracks in TJ cells under simulated loads as well as their associated end of life electrical impact. These results can then serve as the basis of a set of well founded acceptance criteria for mechanical cell defects begin of life.
The evolution of cell cracks during the in orbit lifetime of an array was simulated experimentally, based on "real world" production defects, which were grouped together in strings of similar type and size.
The number of instances were crack growth occurred (crack growth events, CGE) was identified as the main parameter that had to be analyzed, not the length by which a crack propagates.
The size of the original crack and the number of thermal cycles were identified as the main factors that determine crack growth.
The structure was found to govern the direction of crack growth. For the EADS Astrium open weave faceskin design a growth direction parallel to the cell gridfingers was confirmed.
Based on quantitative laws on crack propagation and on a quantitative understanding of the electrical impact of cell cracks, new crack acceptance criteria were established for TJ cells used in GEO representative environments up to 2000 cycles. They were chosen to have the same acceptable consequence for the end of life performance of a cell than the present criteria. In this way the CIEL results were not used in absolute terms but in a relative fashion to provide a transformation between the different crack criteria. The in orbit experience gained with cells inspected according to the present criteria is utilized in this way, while the basis of the new crack criteria is sound. They are based on real world crack patterns and the knowledge of their end of life behavior.
Replacing the visual inspection and the current crack criteria by an ELM inspection together with the newly established criteria increases the reliability of a solar array considerably, also owing to the three times better detection efficiency of the new method.
The problem of crack growth was treated statistically. The main loads encountered in orbit, which are thermally induced, were simulated by coupon testing. 5 miniature solar arrays, featuring a total of ? 200 cells with representative mechanical defects, were built and subjected up to 30000 simulated eclipse phases. In addition ? 400 cell cracks on actual flight panels were monitored throughout sine vibration and acoustic noise testing to cover the mechanical loads during launch of a solar array.
The exact path of a crack in the TJ cells was visualized by the Electroluminescence Method (ELM) (for further details refer to C. G. Zimmermann, J. Appl. Phys. 100, 023714, 2006). By recording the spatial distribution of the radiation emitted from the individual subcells under forward biased conditions, a high resolution image of the crack pattern is obtained (Fig. 1). In addition, electrically inactive cell areas are revealed by the absence of radiation. To document cracks with similar resolution by standard optical techniques, high magnification optical microscopy would have to be used, an approach unsuitable to larger areas. The standard visual inspection, on the other hand, has been found to suffer from detection efficiencies of only 30%. ELM is thus a crucial feature of this project.
Fig. 1: Typical ELM (middle) cell image with mechanical defect. 
click for larger image
While crack propagation was tracked entirely by ELM, the mechanisms resulting in electrical degradations were analyzed by pulsed solar simulator measurements in combination with the ELM data.
The project test and inspection sequence is summarized in Fig. 2.
Fig. 2: Test and inspection plan of the CIEL project. 
click for larger image
The project has been completed. The impact of a given configuration of mechanical defects on the electrical performance and thus the reliability of a triple junction solar cell throughout launch and 2000 eclipse phases can be predicted quantitatively. The relevant functional dependencies were derived. Apart from shunt resistances, which did develop only in few well defined cases, the electrical impact of cell cracks originates from an interruption of the front side contact network. Continued thermal cycling weakens the gridfinger intersected by the crack line.
The probability of gridfinger interruption has been determined and thus the associated loss in cell area can be calculated based on the original extend of the crack. The original size of the crack also determines the number of crack growth events that are to be expected. Taking into account the preferred vertical growth direction observed on EADS Astrium substrates, an extrapolation to a likely crack pattern after 2000 cycles is possible. Only in case this pattern intersects the current collecting bar of a cell area without an alternative electrical connection, an additional electrical impact results, which is at worst again proportional to the separated cell area.
More details can be found in C. G. Zimmermann, "The impact of mechanical defects on the reliability of solar cells in aerospace applications", IEEE Trans. Device and Materials Reliability, Vol. 6, No 3, 2006 (in press). A set of criteria has been developed that define which defects can be accepted begin of life.