Pulsed and Steady-State Radiation Effects on Single Junction Si and Multiple Junction GaAs Photocells

Si single junction photocells manufactured by Sandia National Laboratories and commercially available multiple junction GaAs photocells were tested in a pulsed high-dose mixed gamma-neutron environment. The Si photocells were also tested in steady state gamma environment at two different dose rates.


A. Single Junction Si Photocell
The device of interest is a series connected photovoltaic array. When photocells are placed in series the illumination profile becomes a critical parameter [5]. This is due to current limiting caused by the photocell with the weakest illumination. In many applications the photocell is illuminated with a multimode fiber. The illumination profile of the multimode fiber is non-uniform and changes spatially as a function of time. Single junction devices of the same construction as those in the series array are tested in order to prevent radiation induced degradation from being convoluted by time-dependent spatial changes in the illumination profile.
Shown in Fig. 1 is a schematic and SEM cross section of a series connected photocell. These devices were manufactured on SOI wafers. The top-Si is 5 µm thick and the thickness of the oxide is 3 µm. Trench isolation is formed by etching through the entire thickness of the top-Si using a dry reactive ion etching technique. The trench is then filled with SiON and the overburden is removed by chemical-mechanical polishing (CMP). A SiO 2 dielectric is used to passivate the top Si layer as well as provide isolation from the metal lines. SiO 2 is also used to passivate the metal lines. Deposition of the SiO 2 is accomplished by standard plasma deposition. The P/N junction is formed by standard lithographically defined ion implantation which is followed by thermal activation. Dopants used in the n and n+ regions are P and As, respectively. The p+ region dopant is B.

B. Multiple Junction GaAs Photocell
Multiple junction GaAs photocells tested were purchased from a commercial vendor. Dimensions of the device measured 2 mm x 2 mm with a circular active area consisting of six wedge shaped series connected photocells. The die is provided by the manufacturer pre-mounted in a receptacle that is compatible with ST type fiber optic cable connectors [6]. Any further information could not be obtained from the vendor due to issues of propriety. Fig. 2 shows the ST receptacle that the photocell is packaged in and the photocell mounted on a TO header.

III. RADIATION TESTING
Single junction Si photocells were tested in different radiation environments. One environment was a pulsed event comprising of both gamma and neutron radiation. The other environment was steady state gamma from a Co-60 array.
Two radiation test facilities at Sandia National Laboratories were utilized as part of this experiment. The annular core research reactor (ACRR) generated the pulsed gamma-neutron environment and the gamma irradiation facility (GIF) was used to generate the steady-state gamma environment.

A. ACRR Passive Tests
An initial passive test was performed at ACRR to determine if Si and GaAs photocells would survive in a pulsed high neutron fluence environment. ACRR is a research reactor that produces pulsed environment that consists of gamma and neutron radiation. This is the only test in which the GaAs photocells were included.

1) Experiment
The experimental setup in this case involved precharacterizing the photocells under illumination before exposure, wrapping the devices in aluminum foil and exposing them to the ACRR radiation environment, and postcharacterization.
Pre-characterization of the Si photocells consisted of illuminating the devices at 154 mW using a fiber coupled 808 nm laser. The fiber core is 100 µm, has a 0.22 NA, and is terminated with an ST connector. An XYZ translation stage was used to align the photocell to the fiber optic output. Alignment of the GaAs photocells was not required since the devices are prepackaged in an ST receptacle.
A total of 12 Si and 2 GaAs photocells were tested. The Si devices tested have junction areas ranging from 120 µm x 550 µm to 20 µm x 20µm. Two of each junction area was tested. An I-V curve for each photocell was acquired. The value of the short circuit current (I SC ) and the open circuit voltage (V OC ) is of particular interest and will be used as the metrics by which the photocell degradation is measured. After irradiation, the post-characterization data is taken using the same method.
The radiation environment was measured using eight sulfur pellets and eight CaF2 thermo-luminescent dosimeters (TLD). Based on the activation of the sulfur pellets the photocells experienced a 1 MeV (Si) neutron fluence of 4.1x10 14 neutrons/cm 2 . The CaF2 TLDs were saturated indicating a gamma dose in excess of 100s of krad (Si).

2) Results
The single junction Si experienced an average decrease in I SC of 61.7 % and a decrease in V OC in the range of 21.5 % to 12.6 %. Degradation in I SC is independent of the junction area while V OC is not. Larger junction areas result in a larger decrease in V OC . The multiple junction GaAs I SC degraded ~50 % and the V OC decreased by ~17%. The complete results are shown in Table I.

B. ACRR Active Tests
Two active tests were performed at ACRR with the Si photocells illuminated at 150 mW. The purpose of the active tests was to measure photocell recovery, if any, immediately after the radiation pulse.
1) Experiment Illuminated and unilluminated photocells of the same junction area were placed side by side. The purpose of the unilluminated photocell was to monitor for radiation generated current within the photocell. Photocell parameters of interest are I SC and V OC . I SC was monitored by measuring the voltage generated across a potentiometer using an oscilloscope. Load resistance was adjusted to ensure that the photocell was biased around 250 mV, well within the linear region of the photocell I-V curve. V OC was determined by recording the I-V curve before and immediately after the radiation pulse. Additional fiber optic cables were placed alongside those used to illuminate the photocells. These fibers are used to account for any reductions in photocell performance due to radiation induced fiber darkening or fluctuations in laser diode output power. High OH content fused silica fibers were used to minimize this effect. Fig. 3 Fig. 3. Schematic of test setup used for active radiation testing of single junction Si photocells.
The first test was performed on photocells with junction areas of 120 µm x 550 µm and 40 µm x 60 µm. Neutron fluence was measured by placing four sulfur pellets and one nickel foil alongside each pair of illuminated and unilluminated photocells. Gamma dose was measured using four TLDs also placed alongside the photocells. The photocells were exposed to a mean 1 MeV (Si) fluence of 2.9x10 13 neutrons/cm 2 and a total gamma dose of 8.8 krad (Si) radiation environment.
Photocells with junction areas of 120 µm x 550 µm and 20 µm x 20 µm were tested next. The 1 MeV (Si) neutron fluence and gamma dose was measured to be 4.2x10 14 neutrons/cm 2 and 170 krad (Si), respectively, using the same method as mentioned above.
2) Results Photocells that were exposed to the first low energy radiation pulse had a reduced I SC of 30 % and 21% for the large and small junction area photocells, respectively. V OC was reduced by 11% and 4% for the large and small junction area cells, respectively. In Fig. 4 a plot of the response from the two illuminated photocells and the output from the active neutron detector are shown.
The second test exposed the photocells to a high energy radiation pulse and reduced the I SC of both devices by 70%. V OC was decreased by 26% and 14%, respectively, in the large and small junction photocells. Fig. 5 shows the response of the photocells to the high energy radiation pulse.  Neither radiation generated current within the photocell nor fiber darkening was observed. Data was recorded for twenty minutes after the radiation pulse to observe any recovery in device performance. None was evident.

C. GIF Tests
Testing of Si single junction photocells was also performed at GIF in order to determine susceptibility to steady-state gamma radiation. The gamma source is a Co-60 array.

1) Experiment
The experimental setup used at GIF was identical to that used at ACRR (Fig. 3) with the exception of the data acquisition system. A multi-channel curve tracer was used to simultaneously record the I-V curves of all of the photocells. Data was recorded over a period of 18 hours. After four hours of exposure the alignment of the test setup inside of the Co-60 array began to change thereby invalidating any data after that point. Change in alignment was due to the degradation of previously unknown plastic components that were present inside of the alignment fixture.
Two sets of tests, using photocells with junction areas of 120 µm x 550 µm and 20 µm x 20 µm, were run simultaneously at different dose rates. Different dose rates were achieved by placing one setup in the center of the Co-60 array and the other setup inside of a box shielded with 1/8" thick lead at the back of the test cell. Dose rate was determined by exposing four TLDs, placed at the same location as that of the photocells, for 5 minutes and dividing the total dose by the total time of exposure. The TLDs measured a dose rate of 800 rad (Si)/s and 5.33 rad (Si)/s in the center of the Co-60 array and inside of the lead shielded box, respectively.
2) Results Fig. 6 shows the normalized photocell I SC as a function of dose for both tests. Device degradation appears to be a function of photocell junction area, dose rate and total dose. The high dose rate environment shows that photocell I SC degrades by 30 % and 60% for the large and small photocells, respectively, at a total dose of 12 Mrad (Si). Large and small junction area photocells exposed at the low dose rate had I SC reductions of 25 % and 65 %, respectively, at a total dose of 350 krad (Si). Even though the total dose, at the lower dose rate, was two orders of magnitude lower the devices suffered comparable degradation. These results suggest that the photocells are susceptible to enhanced low dose rate effects (ELDRS). When the test results are plotted as a function of time in (fig. 7) there appears to be a strong time dependent degradation in I SC . Data was recorded for 30 minutes after the gamma exposure had ended. There was no device recovery.

IV. SUMMARY
Single junction Si and multiple junction GaAs photocells were tested in a pulsed high-dose mixed gamma-neutron environment at ACRR and a steady-state gamma environment at GIF.
Testing in a pulsed gamma-neutron environment has shown that a 1 MeV (Si) neutron fluence of 3.9x10 13 neutrons/cm 2 and a total gamma dose of 170 krad (Si) degrades the I SC in Si photocells by ~30 % independent of the photocell junction area. Reduction in V OC ranges from 10 % to 4 % with larger junction area devices suffering the greater degradation. A 1 MeV (Si) neutron fluence of 4.2x10 14 neutrons/cm 2 and a total gamma dose of 170 krad (Si) degraded the photocell I SC by ~70% independent of junction area and reduced the V OC by 26 % and 14%, respectively, in the large and small junction areas. GaAs photocells had a reduction of ~50 % and ~20 % in I SC and V OC respectively.
Testing at GIF showed that the Si photocell I SC degradation in a steady-state gamma environment is a function of dose rate, junction area, and total dose. I SC degradation of 25 % occurred in a 5.33 rad (Si)/s dose rate with a total dose of 300 krad (Si) for photocells with the larger junction area. Smaller junction area photocells in the same environment had an I SC degradation of 65 %. Photocells exposed to a total gamma dose of 12 Mrad (Si) at a rate of 800 rad (Si)/s saw degradation in I SC of 30 % and 60 %, respectively, for the large and small junction area photocells. These Si photocells exhibited an increased sensitivity to a low dose rate environment. The large area photocells survived better in both the high and low dose rate environments compared to the same photocells with a smaller junction area.