Damp-Heat Induced Performance Degradation for InGaP / GaAs / Ge Triple-Junction Solar Cell

We performed accelerated tests on sealed and nonsealed InGaP/InGaAs/Ge triple-junction (TJ) solar cells in a complex high temperature and high humidity environment and investigated the electrical properties over time. The degradation of energy conversion efficiency in nonsealed cells was found to bemore serious than that in sealed cells.The short-circuit current (ISC), opencircuit voltage (VOC), and fill factor (FF) of sealed cells changed very slightly, though the conversion efficiency decreased 3.6% over 500 h of exposure. This decrease of conversion efficiency was suggested to be due to the deterioration of silicone encapsulant. The ISC,VOC, and FF of nonsealed cells decreased with increasing exposure time. By EL and SEM analysis, the root causes of degradation can be attributed to the damage and cracks near the edge of cells induced by the moisture ingress. It resulted in shunt paths that lead to a deterioration of the conversion efficiency of solar cell by increasing the leakage current, as well as decreasing open-circuit voltage and fill factor of nonsealed solar cells.


Introduction
TJ solar cells are fabricated using organometallic vapor phase epitaxy (OMVPE) to deposit subcells of GaInP and GaInAs on Ge substrate [1].A TJ solar cell consists of three individual subcells that are stacked on top of each other to form a series connection.The individual subcells are connected to each other through very thin tunnel junction film of several nanometers.These tunnel junctions are typically very thin to allow the tunneling mechanism to dominate the majority carrier transport [2].Among solar technologies, TJ solar cells have attracted increasing attention owing to their very high conversion efficiencies [3,4] as well as long term stability [5][6][7].TJ solar cells based on III-V semiconductors are being evaluated globally in concentrator photovoltaic (CPV) systems designed to supplement electricity generation for utility companies.III-V TJ solar cells, with demonstrated efficiency over 40% since 2007 [8], strongly reduce the cost of CPV systems and make III-V multijunction concentrator cells the technology of choice for most concentrator systems today.The world's highest energy conversion efficiency solar cells with 44.7% have been demonstrated on June 2013 [9].In fact, the III-V TJ solar cells have nanoscale thin films which can take good advantage of solar light from 300 nm to 1800 nm.
In recent years, there have been many research groups that reported the improvements of III-V triple-junction concentration solar cells using different approaches [10][11][12][13].In addition to pursuing high efficiency for III-V concentration solar cells, one should also understand the degradation mechanism of these devices to maintain their reliability.The reliability of III-V single-junction, dual-junction, and multijunction concentration solar cells has been analyzed by some research groups using different methods and strategies.However, so far, the studies about the reliability on GaInP/GaInAs/Ge concentration solar cells were rather rare and almost focused on clarifying the degradation induced by thermal stress [14][15][16][17].It is believed that the understanding of mechanism of degradation caused by moisture will favor the upgrading qualities of III-V triple-junction concentration solar cells.Once the reliability of III-V triple-junction concentration solar receivers is realized, the prospect of the HCPV systems will be more clearly demonstrated.
The aim of this study is to examine the degradation mechanism of GaInP/GaInAs/Ge triple-junction solar receivers by accelerated damp-heat environment test.These results are considered to be helpful for researchers to understand the causes of degradation and improve the efficiency and stability of TJ solar cells.

Samples Preparation and Test Description
IEC62108 standard is the most commonly used test method for concentrator solar receivers.In this study, III-V InGaP/InGaAs/Ge concentrator solar cells were exposed to 85 ∘ C and 85% relative humidity without electrical bias according to IEC 62108 standard.A photograph of typical concentrator solar receiver used in this study is shown in Figure 1, including concentrator solar cell, protection silicone layer, bypass diode, ceramic substrate, and heat sink.The cells with active area of 0.3025 (0.55 × 0.55) cm 2 were glued to aluminum plates using an adhesive.The schematic structure of the GaInP/GaInAs/Ge triple-junction concentrator solar cell is shown in Figure 2, and the schematic graph of crosssection of silicone sealed cell is shown in Figure 3.Those concentration solar cells are divided into two groups.One is denoted as test  with 20 pieces of cells protected with optical silicone layer.The other is denoted as test  with 20 cells without any protection layer.Test  is designed to provide a contrast to test , which is favorable for analyzing the predominant factor in degradation of solar cells.The experimental time sequence of accelerated degradation test for tests  and  is the same.
Finally, we have performed both dark current versus voltage (DIV) and light current versus voltage (LIV) tests for all solar cells to analyze the variation of their electrical characteristics.We have also visually inspected samples before humidity exposure and after 100, 200, 300, 400, and 500 h of exposure.According to IEC62108 standard, samples should be tested over 2000 h; however, we have observed apparent degradation in nonsealed cells after 500 h of dampheat exposure.We therefore report these results in advance.A solar simulator (WACOM WXS-155S-L2) was used to measure the LIV.In the meanwhile, scanning electron micrographic (SEM) and electroluminescent (EL) examinations were applied to understand the root cause of degradation in this investigation.

Results and Discussion
The collected average performance data of sealed and bare solar cells are shown in Table 1.It can be seen that the energy conversion efficiency of sealed cells degraded by 3∼4% after 500 h exposure.However,  oc and FF except  sc changed only very slightly in sealed cells.On the surface of the sealed cells, no significant morphological change was observable.Figure 4 shows the typical dark - curves of solar cells which were measured sequentially after each 100 h time interval.In general, the dark current of III-V solar cells as a function of bias voltage can be formulated by double exponential terms [18] as described in  Here,  diff () denotes the diffusion current in the neutral region and  scr () is related to the generation-recombination current in the depletion region.In general, the dark current is dominated by recombination current at low voltage and by diffusion current at high voltage.Since the decrease of FF was very small, this implied that both shunt and series resistances did not change apparently.From Figure 4, it is noteworthy that almost all sealed cells showed similar dark - characteristics.No obvious difference can be recognized between these - curves implying that both  diff and  scr did not change even after 500 h of exposure.In other words, there were no significant defects generated, which would increase the recombination current [19].We believe that there is no deterioration of electrical characteristics in sealed cells.Hence, the losses of conversion efficiency were suggested to be due to encapsulant material.These results are similar to that reported previously by McIntosh et al. [20].They reported that water moisture can diffuse into silicone materials, then the water molecules will scatter incident light and decrease the transmission coefficient of silicone materials, thus decrease the short-circuit current and conversion efficiency.
As for bare solar cells, the average energy conversion efficiency of 20 cells after 500 h of exposure degraded by 5∼6%, which is higher than the data of sealed cells.It is worth mentioning that there are two cells that are denoted as S1 and S2 and degraded over 10% on energy conversion efficiency.The light - and dark - curves of S1 are shown in Figures 5(a) and 5(b), respectively.The variation of  SC was not apparent over the damp-heat test.The degradation of  OC was also not apparent before 400 h exposure; however,  OC decreased significantly after 500 h of exposure.From the EL images of 400 h and 500 h exposed samples, as shown in Figure 6, some different color zones near the edge of cells can be observed.It reveals that some defects were generated after 500 h of damp-heat test.Evidently, these defects increase  scr (), therefore, result in the elevation of  dark and a significant decrease in  OC .
Figures 7(a) and 7(b) show the light - curves and dark - curves of S2 cell before and after 100, 200, 300, 400, and 500 h of exposure in damp-heat environment, respectively.It exhibits a linear (ohmic) behavior after 500 h of exposure.It is well known that a lot of commercially produced silicon solar cells have a problem caused by local short circuits called shunts [21].These shunts may lead to a deterioration of the conversion efficiency of solar cells by increasing the leakage current and decreasing both the opencircuit voltage and the fill factor of the solar cells.Shunts in solar cells often localized at the cell edges.Some of these edge shunts have a linear - (ohmic) behavior and some of them have a nonlinear - (diode-like).To analyze the cause of the sudden degradation after 500 h of exposure, we performed SEM and EDX examination for S2. Figure 8 shows the EL images of S2 after 400 h and 500 h of exposures.
No obvious damage was found in the EL image of 400 h exposed sample; however bright zones, as marked by dashedrectangle, appeared in the 500 h of exposed sample.We inspected the bright zone with SEM and found that a particle existed in the zone, which was furthermore identified by EDX to be a segregated particle composed mainly of Au.Therefore, the bright zone exhibited low resistance, and leakage current resulted in shunts.The SEM micrograph of Figure 9 also shows a crack at the edge, as marked by a circle.Figure 10 shows an enlarged SEM micrograph of crack region.We furthermore performed EDX analysis and found that region a is the Ge substrate, regions b and c are GeO x , and region d is the epitaxial layer.The crack initiated in the nanoscaled epilayer, possibly induced by in-diffused moisture and part of Ge was oxidized into GeO x .Both the EL and SEM analysis suggested that the shunt and leakage paths occurred in the edge of cells and resulted in the degradation.

Conclusions
We performed accelerated tests on both sealed and bare TJ concentrator solar cells in a damp-heat environment and investigated the degradation of electrical characteristics over time.Our results indicated that silicone encapsulant really protects solar receiver and resists moisture ingress; therefore, the performance of sealed cells is much better than unsealed cells.However, careful selection of silicone material should be adopted to avoid the deterioration of silicone from moisture attack and to preserve the reliability.For bare cells, we found that  sc ,  oc , and FF and conversion efficiency degraded after long time exposure to damp-heat environment.Both EL and SEM analysis indicated that the causes of degradation can be attributed to the damage and cracks in the edge of cells.The damage and cracks resulted in the leakage paths and increased recombination current.In the near future, we will investigate solar cells with different silicone materials and increase the damp-heat test time in order to clarify the degradation mechanism of silicone-sealed solar cells and to understand the lifetime under the influence of damp-heat environment.

Figure 1 :Figure 2 :
Figure 1: Picture of concentrator solar receiver used in this study.

Figure 3 :Figure 4 :
Figure 3: Schematic graph of cross-section of concentrator solar receiver.

Figure 5 :Figure 6 :
Figure 5: (a) Light - and (b) dark - curves of bare cell S1 measured sequentially after each step of damp-heat exposures.