Crystalline Si photovoltaic modules based on TiO₂-coated cover glass against potential-induced degradation

Abstract

Potential-induced degradation (PID) in multicrystalline Si photovoltaic (PV) modules was generated by applying −1000 V from an Al plate attached on the cover glass of the module to the Si cell at 85 °C. The solar energy-to-electricity conversion efficiency of the standard Si PV module remarkably decreased from 15.9% to 0.6% after 2 h of the PID test. Increased concentration of Na species on the surface of the Si cell after the PID test was observed by secondary ion mass spectrometry (SIMS) measurement. Our results indicate that high minus voltage stress toward the Si cell causes the diffusion of metal cations, such as Na⁺, from the front cover glass toward the Si cell, resulting in remarkable decrease in PV performance. PID was significantly prevented by a coating of TiO₂-thin film on the cover glass that suppressed the diffusion of Na⁺, demonstrating an attractive and promising technique for producing low-cost PID-resistant PV modules.

---

Introduction

Over the last decade, photovoltaic (PV) systems, which directly convert solar energy to electricity, have attracted considerable attention as one of the promising clean and renewable energy resources. Larger PV systems consisting of an increasing number of PV modules, the so-called “mega-solar system”, have been constructed to produce greater electric power. The durability of the PV modules and total systems is significantly important for the total power output of the system in addition to the solar energy-to-electricity conversion efficiency (η) and the production cost of the PV modules and systems.

Recently, potential-induced degradation (PID) in Si-based PV modules has been observed and reported especially in large PV systems, where huge numbers of PV modules are serially interconnected. High voltage stress toward the PV modules appears to cause PID, resulting in significant power losses in the systems.^1–6 It has been reported that environmental conditions, such as high temperature and high humidity (or water on the module), are important factors leading to PID.^1,2,7 In addition, several module components, such as the front cover substrate and encapsulant, remarkably influence PID.^1,2,5 The mechanism of degradation of crystalline Si PV modules by PID has been investigated and reported.^8–12 Metal ions, such as Na⁺, which diffuse from the soda lime front cover glass toward the Si cell by high-voltage stress, are considered to cause PID.^8–12

Taking into consideration the mechanism of PID, several PID-resistant techniques have been challenged and reported. For example, PID can be avoided by using Na-free front cover substrates^13,14 or encapsulant whose volume resistivity is high^1,2,5 to diminish migration of Na⁺ in modules. In addition, the control of the composition of silicon nitride (SiN) film as an anti-reflecting (AR) coating on the surface of crystalline Si PV cells is also effective in decreasing PID.^5,15 It has been considered that the increasing conductivity of AR coating releases positive charges, such as Na⁺, resulting in the suppression of PID.¹⁵ In addition to these techniques, new low-cost PID-resistant methods would be important and desirable for preventing PID in outdoor modules.

To prevent PID and consequently improve the long-term stabilities of PV modules, we are currently studying PID in PV modules especially in terms of understanding the PID phenomena and new techniques for the suppression of PID to produce low-cost PID-resistant modules. We have focused on using thin film metal oxide coating, such as TiO₂, on the front cover glass. The motivation of this study is that TiO₂ is one of the low-cost oxide materials, which can be easily coated as thin film on substrates by simple processes such as sol–gel coating and sintering.^16–21 Thus, the utilization of TiO₂ thin film would be a low-cost technique using small amounts of low-cost material and a simple coating process. In this paper, we report the degradation of crystalline Si PV modules by PID, which was easily generated in our laboratory, and the PID-resistant property of TiO₂ thin film coating on the cover glass. Our results demonstrate an attractive and promising technique for producing low-cost PID-resistant PV modules.

Experimental section

Preparation of TiO₂-thin film on glass

The starting material for TiO₂-thin film was an aqueous solution of peroxotitanium complex whose preparation procedure was reported in other papers.^16,17 The solution was coated onto the cover glass (side for Si cell) by the spread or spray technique, and then heat-treated at 200 °C for 1 h after drying: the sintering temperature was changed from 100 °C to 400 °C for the characterization of the TiO₂-thin film by X-ray diffraction (XRD) analysis (X'Pert PRO, PANalytical, Netherlands). The thickness of the TiO₂ film was approximately 50, 100, and 200 nm, which was estimated from the amounts of titanium in the starting material, 12, 23, and 46 μg cm⁻², respectively, measured by fluorescent X-ray analysis. We could not mechanically measure the precise thickness of the TiO₂ film because the surface of the cover glass substrate for the module has embossed coating.

Module fabrication

The standard Si PV module consisted of a front cover glass (Asahi Glass Co., Ltd., soda lime glass, 3.2 mm thickness, 180 mm × 180 mm), two films of commercial EVA (fast-cure type) as the encapsulant, a p-type-based multicrystalline Si cell (Q. Cells Co., 156 mm × 156 mm, 180 μm thickness), and a back sheet whose structure is polyvinyl fluoride (PVF)/polyethylene terephthalate (PET)/PVF. The PV module components (glass/EVA/c-Si cell/EVA/back sheet) were laminated by using a laminator (LM-50x50, NPC Inc.) under vacuum at 135 °C for 15 min. The schematic structures of the standard Si PV module and a module based on the TiO₂-coated cover glass are shown in Fig. 1.

Fig. 1 Schematic structures of Si PV modules: (a) standard module and (b) a module based on the TiO₂-coated cover glass.

PID test and characterization

The PID test was conducted by applying −1000 V from an Al plate (thickness: 0.3–0.5 mm), which was attached on the entire front cover glass of the module as the electrode, to the Si cell using a power supply (Kikusui Electronics Corp., TOS7200) at 85 °C for 2 h. The humidity in the chamber during the PID test was not controlled (ca. 2% at 85 °C).

The η of the Si PV modules before and after the PID test was measured by using an I–V curve measurement system and an AM 1.5G solar simulator (Yamashita Denso Corp., YSS-150A with a 1000-W Xe lamp and an AM filter) as the light source. The electroluminescence (EL) images of the PV modules were measured by an EL measurement system (ITES Co., Ltd.) equipped with a digital camera and a DC power supply (Kikusui, PWR1600M). The secondary ion mass spectrometry (SIMS) analysis of the Si surface before and after the PID test was conducted by Mitsubishi Chemical Group Science and Technology Research Center, Inc. FT-IR absorption spectra of EVA was measured by a Shimadzu FTIR spectrometer (IR Prestige-21) with an ATR system equipped with an ZnSe prism.

Results and discussion

Degradation of a standard Si module by PID test

The I–V curves for a standard Si PV module before and after the PID test (−1000 V at 85 °C for 2 h) are shown in Fig. 2, and the I–V parameters are listed in Table 1. The η of the module remarkably decreased from 15.9% to 0.6% after 2 h of the PID test. The change in the I–V curve suggests that the shunt resistance of the module was decreased, as has been reported in another paper.¹ On the other hand, no change in the PV performance of the module was observed by applying +1000 V (i.e., reverse voltage bias) from the Al plate toward the Si cell at 85 °C for 2 h. These results clearly indicate that PID could be generated in the standard Si module in this experiment with significant decrease in the PV performance, and applying high negative voltage toward the Si cell is one of the essential factors causing PID.

Fig. 2 I–V curves for a standard multicrystalline Si PV module (a) before and (b) after the PID test (−1000 V at 85 °C for 2 h).

Table 1 I–V parameters for the standard Si PV module without the TiO₂ film before and after the PID test by applying −1000 V at 85 °C for 2 h^a

PID test	Isc/A	Voc/V	FF	Pmax/W	η (%)
a Isc: short-circuit current, Voc: open-circuit voltage, FF: fill factor, and Pmax: maximum power output.
Before	8.42	0.61	0.75	3.86	15.9
After	6.42	0.10	0.25	0.15	0.6

---

Hoffmann and Koehl have investigated the effects of PID test conditions on the degradation of PV modules.⁷ They reported that an Al foil, attached on the front cover glass, significantly accelerated PID. The Al foil can directly produce high voltage stress toward the entire front cover glass, accelerating the degradation of the Si cell. Considering this, we expect that our PID test condition using an Al plate attached on the front cover glass at 85 °C is very effective, causing PID in a short time.

Mechanism of PID in c-Si PV modules

We analyzed the Si surface of the PID-tested module by SIMS measurement. The SIMS data for Na species in a non-degraded module and a PID-degraded module are shown in Fig. 3. The intensities of Na species, such as Na⁺, Na₂OH⁺, and Na₃CO₃⁺, on the Si surface of the PID-degraded module were considerably higher than those for the non-degraded module. This result strongly suggests that Na⁺ diffused from the front cover glass after applying large minus voltage stress resulted in PID. When a chemically strengthened glass whose Na⁺ in the surface of the glass is replaced by K⁺ is used as the cover front glass instead of a conventional soda lime glass, no degradation of crystalline Si PV module by PID was observed by a PID test carried out by applying −1000 V at 85 °C for 2 h.¹³ This also indicates that Na⁺ is an essential factor resulting in PID, although we cannot deny the influence of other cations included in the cover glass and module components at present.

Fig. 3 SIMS data for Na species on the Si surface: (a) non-degraded module and (b) PID-degraded module.

It has been reported that the diffusion of Na⁺ from the soda lime cover glass toward the EVA and Si cell occurs by applying minus large voltage, and consequently the cations influence the Si cell, resulting in PID.^2,8–11 Hacke et al. reported that a Na-rich precipitate was deposited on the surface of the Si cell (i.e., SiN layer as the AR layer), measured by Auger electron spectroscopy after the PID test, by applying minus high voltage.² In addition, an increase in the Na concentration on the surface of the Si cell in the PID-degraded module was observed by SIMS measurements.^8–11 Based on the increasing Na concentration on the Si surface, Naumann et al. concluded that Na⁺ reaches the AR layer and/or the AR/Si cell interface and interacts with the minus charge in the n-layer of the Si cell, decreasing the band-bending and consequently significantly decreasing the PV performance.^9–11

PID-resistant module based on TiO₂-coated glass

To suppress the migration of Na⁺ from the cover glass, TiO₂-thin film was used to coat the cover glass substrate (only the side for cell). Fig. 4 shows the XRD patterns for the TiO₂ film with changing sintering temperature. The TiO₂ films sintered below 200 °C were assigned as amorphous phase (the peaks near 10° are attributed to the peroxotitanium complex). The formation of crystalline anatase phase, represented by the peaks at 25°, 48°, and 55° was observed at sintering temperatures greater than 300 °C (Fig. 4).

Fig. 4 XRD patterns for the TiO₂ films coated on glass with changing sintering temperature: (a) 100 °C, (b) 200 °C, (c) 300 °C, and (d) 400 °C.

The structure of the Si PV module based on the TiO₂-coated cover glass for PID-resistant module is shown in Fig. 1b. The I–V curves for the modules with TiO₂-coated cover glasses (thickness is 50 nm, 100 nm, and 200 nm, and the sintering temperature is 200 °C, respectively) before and after the PID test by applying −1000 V at 85 °C for 2 h are shown in Fig. 5a–c. The I–V parameters are listed in Table 2; the number of module sample for each TiO₂ thickness was two. Degradation by PID of the module was remarkably suppressed by using TiO₂-coated cover glass compared to the standard module (Fig. 2). Suppression effect improved with increasing thickness of TiO₂ film on the glass, clearly indicating that the suppression of PID is due to the TiO₂ film coated on the glass.

Fig. 5 I–V curves for modules with the TiO₂-coated cover glass before and after the PID test applying −1000 V at 85 °C for 2 h: (a) TiO₂ 50 nm, (b) TiO₂ 100 nm, and (c) TiO₂ 200 nm.

Table 2 The I–V parameters for the modules with a TiO₂-coated glass before and after the PID test by applying −1000 V at 85 °C for 2 h with changing the thickness of TiO₂ film on the cover glass^a

TiO2/nm	PID test	Isc/A	Voc/V	FF	Pmax/W	η (%)
a Isc: short-circuit current, Voc: open-circuit voltage, FF: fill factor, and Pmax: maximum power output.
50	Before	8.34	0.62	0.75	3.88	16.0
	After	8.22	0.60	0.54	2.66	10.9
50	Before	8.37	0.62	0.76	3.91	16.1
	After	8.16	0.59	0.41	2.00	8.2
100	Before	8.23	0.62	0.76	3.85	15.8
	After	8.23	0.62	0.72	3.67	15.1
100	Before	8.16	0.62	0.76	3.81	15.7
	After	8.14	0.62	0.74	3.72	15.3
200	Before	8.20	0.62	0.76	3.83	15.7
	After	8.22	0.62	0.73	3.71	15.2
200	Before	8.19	0.62	0.76	3.83	15.8
	After	8.16	0.62	0.74	3.72	15.3

---

Fig. 6a–c show the EL images for the standard module without TiO₂ coating and the modules with TiO₂-coated cover glass (the sintering temperature is 200 °C) before and after the PID test (−1000 V at 85 °C for 2 h). For the standard module, the EL image perfectly disappeared after the PID test. Darkened parts were partially observed in the module with 50 nm of TiO₂ film, indicating that PID partially occurred (Fig. 6b). On the other hand, no change in the EL image was observed for the module with 100 nm of TiO₂ film after the PID test. These results and Fig. 5a suggest that the thinner TiO₂ film (50 nm) is insufficient to completely suppress the diffusion of Na⁺, and more than 100 nm of TiO₂ film is necessary for the suppression of Na⁺ diffusion in this PID test condition.

Fig. 6 EL images for a standard module and the modules with the TiO₂-coated cover glass before and after the PID test applying −1000 V at 85 °C for 2 h: (a) standard without TiO₂ film, (b) TiO₂ 50 nm, and (c) TiO₂ 100 nm.

It has been reported that Na⁺ diffused from the soda lime glass reacts with TiO₂ thin film coating on the glass substrate, producing either sodium titanate or a brookite phase.^18–22 Therefore, our results suggest that the TiO₂ film reacts with Na⁺ diffused from the front cover glass by applying minus high voltage. As a result, the diffusion of Na⁺ from the cover glass to the Si cell was prevented, resulting in the suppression of remarkable degradation by PID. Thin films of silicon-based materials, such as SiO₂ and SiN_x, also demonstrate the blocking property of Na⁺ diffusion from the soda lime glass.^21,22 At present, the difference between SiO₂ and TiO₂ in their ability to suppress the effect of PID is unclear. We observed that a SiO₂/TiO₂ composite film coated on the cover glass also significantly prevented PID, similar to the pure TiO₂ film. Thus, we consider that both SiO₂ and TiO₂ thin films have similar ability to suppress PID; however, more detailed investigation is necessary.

Influence of TiO₂ film on I–V performance and long-term stability of the module

Effects of the TiO₂ film coating on the I–V performance and durability of the module are important. For example, when the TiO₂-coated glass was employed as the cover glass, the short circuit current I_sc decreased up to 1–3%, which depends on the thickness of the TiO₂ film, as shown in Tables 1 and 2 This would be caused by the reflection, scattering, and absorption of the incident light by the TiO₂ film. To maintain the I_sc, the TiO₂ film conditions, such as film thickness and composition, should be optimized further.

When crystalline TiO₂ film was used to coat the glass by sintering at more than 400 °C, similar suppression effect of PID was observed. However, if we use the crystalline anatase TiO₂ film for the coating material, the EVA encapsulant is decomposed by the crystalline TiO₂ film under UV-light irradiation because of its high photocatalytic property assigned by ATR-FT-IR absorption analysis (data is not shown). Therefore, amorphous TiO₂ film should be employed to maintain the durability of the module. We observed that the amorphous TiO₂ film sintered below 200 °C does not have high photocatalytic activity.

In addition, it has been reported that the photocatalytic activity of the Na-contaminated TiO₂ film is lower than that of the TiO₂ film.^21,22 Thus, the reaction between Na⁺ and the TiO₂ film not only suppresses PID, but also decreases the photocatalytic activity of the pure TiO₂ film, which consequently results in maintaining the long-term stability of the module.

Conclusions

PID in multicrystalline Si PV modules could be easily generated by the PID test by applying −1000 V at 85 °C for 2 h using an Al plate. The PV performance of the modules significantly decreased. The SIMS data of PID-damaged modules suggested that Na⁺, which is diffused from the front cover glass (soda lime glass) toward the Si cell by applying high minus voltage, causes PID. Degradation by PID was remarkably prevented by using TiO₂-coated cover glass due to the suppression of Na⁺ diffusion from the glass. Our results demonstrate one of the attractive and promising techniques for producing low-cost PID-resistant Si PV modules because TiO₂ is a low-cost material and the amount of TiO₂ used for the thin film is quite small.

Acknowledgements

We acknowledge Mr M. Inoue and Ms. S. Jonai from AIST for their support in the experiments. We also thank Dr H. Miyauchi from Mitsubishi Rayon Co., Ltd. for supporting SIMS measurement.

References

1. S. Pingel, O. Frank, M. Winkler, S. Daryan, T. Geipel, H. Hoehne and J. Berghold, Proceedings of the 35th IEEE Photovoltaic Specialists Conference (IEEE PVSC), Honolulu, 2010, p. 2817.
2. P. Hacke, K. Terwillinger, R. Smith, S. Glick, J. Pankow, M. Kempe, S. Kurtz, I. Bennett and M. Kloos, Proceedings of the 37th IEEE PVSC, Seattle, 2011, p. 814.
3. M. Schütze, M. Junghänel, O. Friedrichs, R. Wichtendahl, M. Scherff, J. Müller and P. Wawer, Proceedings of the 26th European Photovoltaic Solar Energy Conference and Exhibition (EU PVSEC), Hamburg, 2011, p. 3097.
4. H. Nagel, A. Metz and K. Wangemann, Proceedings of the 26th EU PVSEC, Hamburg, 2011, p. 3107.
5. S. Koch, J. Berghold, O. Okoroafor, S. Krauter and P. Grunow, Proceedings of the 27th EU PVSEC, Frankfurt, 2012, p. 1991.
6. H.-C. Liu, C.-T. Huang, W.-K. Lee and M.-H. Lin, Energy Power Eng., 2013, 5, 455.
7. S. Hoffmann and M. Koehl, Prog. Photovoltaics, 2014, 22, 173.
8. J. Bauer, V. Naumann, S. Groβer, C. Hagendorf, M. Schütze and O. Breitenstein, Phys. Status Solidi RRL, 2012, 6, 331.
9. V. Naumann, C. Hagendorf, S. Grosser, M. Werner and J. Bagdahn, Energy Procedia, 2012, 27, 1.
10. V. Naumann, D. Lausch, S. Groβer, M. Werner, S. Swatek, C. Hagendorf and J. Bagdahn, Energy Procedia, 2013, 33, 76.
11. V. Naumann, D. Lausch, A. Hähnel, J. Bauer, O. Breitenstein, A. Graff, M. Werner, S. Swatek, S. Groβer, J. Bagdahn and C. Hagendorf, Sol. Energy Mater. Sol. Cells, 2014, 120, 383.
12. P. Lechner, D. Sanchez, D. Geyer and H.-D. Mohring, Proceedings of the 27th EU PVSEC, Frankfurt, 2012, p. 3152.
13. M. Kambe, K. Hara, K. Mitarai, S. Takeda, M. Fukawa, N. Ishimaru and M. Kondo, Proceedings of the 28th EU PVSEC, Paris, 2013, p. 2861.
14. T. Kajisa, H. Miyauchi, K. Mizuhara, K. Hayashi, T. Tokimitsu, M. Inoue, K. Hara and A. Masuda, Jpn. J. Appl. Phys., 2014, 53, 092302.
15. K. Mishina, A. Ogishi, K. Ueno, T. Doi, K. Hara, N. Ikeno, D. Imai, T. Saruwatari, M. Shinohara, T. Yamazaki, A. Ogura, Y. Ohshita and A. Masuda, Jpn. J. Appl. Phys., 2014, 53, 03CE01.
16. H. Ichinose, M. Terasaki and H. Katsuki, J. Ceram. Soc. Jpn., 1996, 104, 715.
17. H. Ichinose, M. Taira, S. Furuta and H. Katsuki, J. Am. Ceram. Soc., 2003, 86, 1605.
18. H.-J. Nam, T. Amemiya, M. Murabayashi and K. Itoh, J. Phys. Chem. B, 2004, 108, 8254.
19. I. N. Kuznetsova, V. Blaskov, I. Stambolova, L. Znaidi and A. Kanaev, Mater. Lett., 2005, 59, 3820.
20. C. Ohara, T. Hongo, A. Yamazaki and T. Nagoya, Appl. Surf. Sci., 2008, 254, 6619.
21. J. Zita, J. Maixner and J. Krysa, J. Photochem. Photobiol., A, 2010, 216, 194.
22. E. Aubry, J. Lambert, V. Demange and A. Billard, Surf. Coat. Technol., 2012, 206, 4999.

---