Kohjiro Hara*,
Sachiko Jonai and
Atsushi Masuda
Research Center for Photovoltaic Technologies, National Institute of Advanced Industrial Science and Technology (AIST), 807-1 Shuku-machi, Tosu, Saga 841-0052, Japan. E-mail: k-hara@aist.go.jp; Tel: +81 942 81 3675
First published on 14th January 2015
Potential-induced degradation (PID) in p-type-based multicrystalline Si photovoltaic (PV) modules was experimentally generated applying −1000 V from an Al plate, which is attached on the front cover glass of the module, to the Si cell at 85 °C for 2 h. The solar energy-to-electricity conversion efficiency (η) of the standard Si PV module significantly decreased after the PID test. In contrast, no degradation was observed in the modules, including a thin polyethylene (PE) film (30 μm thickness) with the copolymer of ethylene and vinyl acetate (EVA) as the encapsulant. It was suggested that the PE film whose volume resistivity is higher than that of EVA prevented the diffusion of Na+ from the front cover glass toward the Si cell, resulting in a suppression of PID because different degradation processes during PID were observed in the EL images for the two modules, including a half PE film. In addition, the Si PV module, including a PE film, demonstrated stable performance after a damp-heat test (85 °C/85% relative humidity) for 4000 h, although the η of the standard module significantly decreased from 16.0% to 7.6% after the test. Our results indicate an attractive and promising low-cost technique for improving the long-term stability of crystalline Si PV modules against potential and damp-heat-induced degradation.
Generally, a random copolymer of ethylene and vinyl acetate (EVA) has been employed as the encapsulant in crystalline Si PV modules because of its low cost.1,2 In long-term environmentally exposed Si PV modules, however, it is known that acetic acid is formed from the EVA encapsulant by a hydrolysis reaction.1–3 Consequently, the corrosion of the bus bar and Ag finger (grid) electrodes occurs, which results in increasing the series resistance of the module. This is supposed to be one of predominant degradation processes of crystalline Si PV modules in addition to the delamination of the EVA encapsulant.3,4 Formation of acetic acid was also observed in the modules after long-term damp-heat (DH) tests under high temperature with high humidity.5,6 Acetic acid evolving from EVA causes the formation of lead acetate (Pb(OOCCH3)2).6 The degradation after the DH test is considered to originate from the increased contact resistance between the Si cells and Ag finger electrodes.7 Therefore, the suppression of the corrosion of the finger electrodes by acetic acid would further improve the long-term stability of the crystalline Si PV modules.
Additionally, another degradation phenomenon, potentially induced degradation (PID), in crystalline Si PV modules has been recently reported in large PV systems in which large numbers of PV modules are serially interconnected. High voltage stress towards partial PV modules seems to cause PID, resulting in significant power loss in the systems.8–14 Metal ions, such as Na+, which are involved in the soda lime front cover glass and diffuse towards the Si cell by high-voltage stress are considered to cause PID.15–20 Several module components, such as the front cover substrate and encapsulant, remarkably influence PID.12,21–23 For example, PID can be basically avoided using Na-free front cover substrates24,25 or encapsulant, which has a high volume resistivity, such as ionomers19 and polyolefin elastomers,22,23 to diminish the migration of Na+ in the modules. Moreover, control over the composition of silicon nitride (SiN) film as the anti-reflecting (AR) coating on the surface of crystalline Si PV cell is also effective in decreasing PID.26,27 In addition to these techniques, several new PID-resistant methods with low-cost would be desirable to prevent PID in outdoor modules.
In order to prevent PID, and consequently further improve the long-term stability of PV modules, we are currently studying the mechanisms of PID and producing low-cost PID-resistant modules. For example, we have reported that TiO2-thin film coated on the inner side of the front cover glass significantly suppressed PID.28 It was suggested that the TiO2 film prevented the diffusion of Na+ from the cover glass towards the Si cell. This result indicates one possibility for low-cost PID-resistant techniques because TiO2 is one of the low-cost metal oxide materials, only a small amount of TiO2 is required and the coating process is simple (such as sol–gel coating).
In this paper, we have focused on introducing a thin polyethylene (PE) film into the standard Si PV module that uses a conventional EVA encapsulant instead of coating the front cover glass. The motivation of this study is that a thin PE film would be effective to suppress PID because it has a higher volume resistivity than that of EVA.29 In addition, the utilization of PE thin film would be able to avoid the increasing production cost because PE is one of the conventional low-cost polymer materials available and a standard lamination process can be used. Furthermore, we demonstrate that the PE film is also effective in improving the long-term stability of the module using the DH test. Our results show an attractive and promising technique to produce low-cost PID and DH-resistant crystalline Si PV modules.
Fig. 1 The schematic structures of Si PV modules: (1) the standard module and (2)–(5) PE-introduced Si PV modules. |
Fig. 2 The I–V curves for Si PV modules before and after the PID test by applying −1000 V at 85 °C for 2 h: (a) the standard module 1 and (b) a module 2 with a PE film. |
The volume resistivity of the polymers and the value of leakage current for each module during the PID test are listed in Table 1. The volume resistivity of polymer estimated by applying −1000 V at 25 °C was 1.5 × 1014 Ω cm for EVA and 1.8 × 1017 Ω cm for PE (Table 1). These values are very similar to those reported in other papers.22,28 Leakage current in modules seems to be one of the important factors relating to the degradation of the module by PID.13,19 The value of the leakage current in the module with a PE film (below 0.2 μA) is considerably smaller than that in the standard module with only EVA (ca. 6 μA), as shown in Table 1. Based on these values, we can conclude that a thin PE film whose volume resistivity is higher than that of EVA prevents the migration of ions, such as Na+, from the front cover glass towards the Si cell through the EVA, and consequently leads to a small leakage current and suppression of PID, although the detailed mechanism should be investigated. It was considered that increasing the thickness of the PE film enhances the suppression of PID due to the increased volume resistivity of the PE film.
As shown above, it was reported that the degradation of Si PV modules by PID remarkably depends on the type of encapsulant materials used in the modules. If an alternative encapsulant is employed instead of conventional EVA, however, increasing cost and technical problems in the lamination process in terms of changing conditions and cracking of the Si cells due to the hardness of the new encapsulant might be observed. It should be noted that our module structure, which uses a thin polymer film with a conventional EVA encapsulant, would be able to avoid the problems of remarkably increasing cost of the film and changing the lamination processes. Therefore, our results indicate an attractive and promising low-cost anti-PID technique using a polymer film, although further investigation and optimization are necessary. Kapur et al. employed a thin layer of ionomer (a copolymer of ethylene and methacrylic acid with a thickness of 50–150 μm) in conjunction with the conventional EVA encapsulant to minimize both the cost and module material changes.21 They reported that an ionomer film with EVA also suppressed PID. Their results also indicate that a combination of conventional EVA and a thin polymer film as the encapsulant is useful for decreasing PID with small changes to the module structure and production process.
We have reported that coating the inner side of the front cover glass by a TiO2-thin film significantly suppressed PID.28 We consider that coating the glass with a metal oxide film, as well as introducing a PE film are attractive and promising PID-resistant techniques with low-cost. We cannot simply compare both techniques because each process (i.e., coating and utilization of film) and thickness (ca. 100 nm for TiO2 and ca. 30 μm for PE film, respectively) are quite different.
The PE film might be more effective than the TiO2 film in terms of the suppression of Na+ migration because the PE film is considerably thicker than that of the TiO2 film. It must be important that we can choose several PID-resistant methods (or their combination) with low-cost to prevent PID in outdoor modules.
In order to investigate the detailed mechanism of PID suppression in the module involving a PE film, the PID tests using a half Al plate and half PE-introduced modules were conducted. The half PE film was introduced in two positions of the standard modules, as shown in Fig. 1: the right side between the glass and EVA (module 4), the left side between the glass and EVA (module 5), respectively. The Al plate was attached on the half area of the module (right side) as the electrode for the PID test. We expected that different progress in degradation by PID would be observed by the suppression effect due to the half PE film located in the different positions of the module. The EL images before and after the PID test using a half Al plate for modules 4 and 5, including a half PE film, are shown in Fig. 3a and b. The red square indicates the position of the half Al plate electrode in the modules. A darkened area, in which degradation by PID occurred, was slightly observed in the EL image on the left side of module 4 after 5 h of PID test (Fig. 3a). In the case of module 5, the EL inactive area was first observed only on the right side where the half Al plate was attached, and then gradually expanded towards the left side of the module (Fig. 3b). The entire cell was darkened after 5 h, which was considerably faster than found with module 4.
Fig. 3 The EL images for the Si PV modules including a half PE film before and after the PID test using a half Al plate: (a) module 4 and (b) module 5. |
Fig. 4 shows the schematic diagrams presenting the diffusion of Na+ from the front cover glass in modules 4 and 5 with a half thin PE film during PID. The PE film in the module 4, which is presented at the same position of the Al electrode, would directly suppress diffusion of Na+ from the glass. Therefore, degradation was considerably less than that found in module 5, although Na+ migrated towards the left side of the module slightly during the PID test. In contrast, degradation first occurred in the right side of module 5, and then the degradation gradually expanded towards the left side due to the diffusion of Na+ (Fig. 4b). Therefore, we could succeed in the visualization of different progress of PID in terms of Na+ migration from the front cover glass towards the Si cell using the EL images for two modules involving a half PE film. According to these results, we can conclude that a thin PE film in the module suppresses diffusion of Na+ towards the Si cell, resulting in the prevention of PID.
Fig. 4 The schematic structures of the Si PV modules including a half PE film showing the diffusion of Na+ from the front cover glass during the PID test: (a) module 4 and (b) module 5. |
Fig. 6a and b show the EL images of the standard module 1 and the PE-introduced module 3, respectively, before and after the DH test for 4000 h. The degradation-darkened part was observed in the outer cell area of the standard module 1 after 3500 h and the EL inactive area expanded towards the center of the cell after 4000 h (Fig. 6a). Peike et al. have reported a similar degradation change in the EL images by DH test.6 In contrast, no remarkable change was observed in module 3, including a PE film, while the EL inactive area was slightly observed near the bus bar electrodes after 4000 h.
Fig. 6 The EL images for the Si PV modules before and after the DH test: (a) the standard module 1 and (b) module 3 with a PE film. |
The ATR-FT-IR absorption spectra of EVA in the standard module 1 and module 3 having a PE film after the DH test (4000 h) are shown in Fig. 7a and b, respectively. The EVA samples were removed from a corner of the modules and both the surfaces of the glass side and back sheet side were measured. The absorption spectrum of neat EVA is also shown in Fig. 7a and b as reference. The strong peaks, which are assigned as carbonyl CO (1737 cm−1) and C–O (1237 cm−1),3 decreased after the DH test for 4000 h for the EVA in the standard module 1, as shown in Fig. 7a. Instead, a peak at 1550 cm−1, which was attributed to the carboxylate salt COO−,3 was observed only in the EVA of the glass side. These changes in the absorption spectra indicate that acetic acid was formed from EVA during or after the DH test, as previously reported in other papers.1–3,5,6 We consider the reason why the peak of carboxylate was not observed in the EVA of the bask sheet side is that acetic acid formed on the back sheet side would egress from the module through the TPT back sheet.
The degradation mechanism of the module in the DH test is discussed. Schematic diagrams are shown in Fig. 8a and b. First, water vapor would predominantly ingress into the module through the TPT back sheet, which has been reported by Kempe,30 whereas the water ingression rate depends on the back sheet material. No degradation was observed in a double-glass module, which has a glass at the rear side instead of a TPT back sheet, after 4000 h of DH test (data not shown). This result obviously indicates that the ingression of water vapor from the edge side of the module is not predominant until 4000 h, although it probably influences degradation after the 4000 h of DH test. After the ingression of water, acetic acid is formed gradually by the hydrolysis reaction of EVA at 85 °C. After the concentration of acetic acid between the glass and Si cell was increased and reached a particular content, corrosion of the finger electrode of the front side would begin, which increases the series resistance of the Si cell,6,7 and consequently decreases solar cell performance, as shown in Fig. 5a. This mechanism was supported by the result that the EL inactive area was first observed on the outer Si cell in the EL image for the standard module 1 after the DH test of 3500 h (Fig. 6a), which has been reported in another paper.6
Fig. 8 The degradation mechanism of the module under the DH conditions: (a) the standard module 1 and (b) module 3 with a PE film. |
Interestingly, the carboxylate peak at 1550 cm−1 was also observed in the EVA of the glass side for the PE-introduced module 3, as shown in Fig. 7b. This suggests that acetic acid was also formed from EVA in module 3, including a thin PE film, whereas no significant degradation occurred after the DH test for 4000 h (Fig. 5b and 6b). At present, we consider that the permeability of acetic acid through the PE film would be quite low, resulting in the prevention of corrosion of the finger electrodes by acetic acid (i.e., the PE film protects the electrodes from acetic acid), whereas some content of water might be able to penetrate into the PE film, as shown in Fig. 8b.
Therefore, we can conclude that the introduction of a thin PE film into the standard Si PV module can improve the long-term stability of the module against not only PID, but also DH. If optimization of the PE film, such as thickness, additives and reforming of the surface conditions is conducted, the long-term stability of the modules would be further improved. When a thin PE film is introduced into the Si PV module, the adhesive strength between the PE film and cover glass or Si cell must be investigated for long-term stability of the modules. No delamination of the encapsulant involving a thin PE film was observed after the DH test for 4000 h. This indicates that the adhesive strength between the PE film and Si cell is sufficient.
This journal is © The Royal Society of Chemistry 2015 |