Thermal annealing of GaAs concentrator solar cells

The thermal annealing of GaAs concentrator cells for space power after electron irradiation is reported. Results are given for cells annealed at 150 degrees C, 200 degrees C, and 250 degrees C. Isochronal annealing was done for 20 min intervals up to 350 degrees C. For cells irradiated with electrons of energies between 0.7 and 2.3 MeV, the recovery decreases with increasing electron energy. Isothermal and isochronal annealing produce the same recovery. Cells irradiated to 3*10/sup 15/ or 1*10/sup 16/ e/cm/sup 2/ recover to similar unannealed fractions. Significant annealing is seen starting at 150 degrees C, although very long times are required.<<ETX>>

During annealing the bare cells were in a quartz tube in a furnace with dry nitrogen flowing through the tube to prevent any oxidation of the cells. The temperature in the tube was monitored with a thermocoupleand power to the furnace wasadjusted for constant temperature. Performancemeasurementsconsistedof the following: 1. I-V data at 25C and one sun AM0 using an X-25 xenon solar simulator and an appropriate standard cell.
9. I-V data at 25C and 100x AM0 using a pulsed xenon lamp solar simulator and the linear assumption between irradiance and short-circuit current.

Results and Discussion
Isochronal Annealing 2) the same three figures in each column are 1) I/Io, \/'Vo, or data after isochronal annealing;and 3) the un-annealedfraction). The most notable feature of the data is the trend for less annealing as the electron energy increases. Most of the damagecausedby the 0.7 MeV electronscan be annealedcompared to about half for the 2.3 MeV electrons.There appearsto be no major differencein the results of isochronal annealingbetweenthe n/p cellsand the p/n cells. Even though the current degradationis muchlarger than the voltage degradation, the un-annealed fractions after isochronal annealingare in most casesabout the same. Future work is planned using DLTS to investigate the trap levelsand help explain thesetrends.

Isothermal Annealing
Based on the results of the isochronal annealing, we decided to do isothermal annealing at three temperatures, 150C,200C,and 250C. At eachtemperature, a set of six cellswas annealed. Each set consistedof three Varian cells,either p/n or n/p, irradiated to 3 x 1015e/cm2 with 0.7, 1. The results of the isothermal annealing are summarized in tables III and IV. a deeper degradation than the isochronal annealedcells (3 x 10is e/cm2), the unannealed fractions, both for Isc and Pmax,are very similar. For example the l._c un-annealedfractions for the isochronal annealed 1 MeV (3 x 1015e/cm2) cells are approximately 27% and 28_., while for the isothermal annealed1 MeV cells (1 x 101°e /era2) they are 17%,2.5%,25%, 26%, 27%and 30%. This implies that any limit to post-irradiation annealingrecoverymay be independentof fluence level.
Isothermal annealingwasdone at three temperatures,250C,200C, and 150C.It is reasonableto assumethat annealingat higher temperatureswill bring on recovery quicker. This is indeedthe case. Figure 4 showsPmaxrecoveryfor three cellsannealed at 250C. Sincesomeof tile cells in the 250C portion of the experiment had shunting problems, the data in figure4 is for cellsirradiated to different fluences.The lower two curvesare for cells irradiated with 1 MeV electronsto 1 x 101°e/cm" while the upper curve is for a cell irradiated with 0.7 MeV electronsto 3 x 10is e/cm". Note that the annealingis essentiallycompleteafter the 27 hour point, and further annealingto 81 hours has little effect.
For the 200C annealing,more time is requiredto obtain recovery. Figure .5shows normalized Pmaxasa function of annealingtime at 200Cfor the Varian cellsirradiated with three different electron energies.Most of the recovery, occurs betweenabout 10 and 200hours. Howeverthere is moreannealingevenout to 2174hours, which is the last data point.
At 150C,it appearsthat quite a bit of time is requiredfor annealingto take place. Figures 6 and 7 both show annealing results at 150C. Figure 6 shows normalized Pmaxvs. annealingtime for the three cells irradiated to 1 x 10l°e/crn2 with 1 MeV electrons. At the last data point, 5533 hours, there is somesignificant annealing. The cells have recoveredabout one third of their degraded power and the trend appears to indicate much more recovery. Figure 7 shows the normalized I_cratio for three Varian cells irradiated to 3 x 1015e/cm2 with different energy electrons. Again, we are starting to seesomeannealing, especially in the cell irradiated with 0.7 MeV electrons. This cell has recoveredover half its degradedcurrent, while the The time required to reachthe degradation levels involved in this work in spaceis severalyears, dependingon the orbit. This slow degradation rate of cells in spaceis at the heart of the argument for continuousannealing. Even though the annealingis slow, it happensas fast as the induceddegradation, hencecompleteor nearcomplete annealing occurs. Operating temperatures of near 150C can readily be achievedin concentrator arrays, and continuousannealingmay be possible.
For continuous annealingto be proven successful, severalquestions must be answered.Among them are: 1) The annealingcharacteristicsof proton induced damage.
2) The annealingeffectsof irradiating cells in the lab at the annealing temperature.
3) The annealingeffectswhen the cells are irradiated at temperature with a flux similar to those encounteredin space(typically severalorders of magnitude slower than lab experiments).
We intend t.o look at the first two items in future work. Due to the long term nature of the third item, there are no practical experiments which can be performed using particle accelerators. A final answer to the feasibility of continuous annealing may require a flight test, probably in the radiation belts.

Summary
We have performed isochronal and isothermal annealing tests on GaAs concentrator cells which had been irradiated with electrons of various energiesto fluences up to 1 x 1016e/cm2. The results include: l) For cellsirradiated with electronsfrom 0.7 to 2.3 MeV, recoverydecreases with increasing electron energy.
3) \Ve are starting to see some significant annealing a.t 150C although very long times are required. 'The same data after annealing 3) The un-annealed Fraction