Outdoor performance and durability testing of antireflecting and self-cleaning glass for photovoltaic applications

Available online at www.sciencedirect.com

ScienceDirect Solar Energy 110 (2014) 231–238 www.elsevier.com/locate/solener

Outdoor performance and durability testing of antireflecting and self-cleaning glass for photovoltaic applications Mridul Sakhuja, Jaesung Son, Hyunsoo Yang, Charanjit S. Bhatia, Aaron J. Danner ⇑ Department of Electrical and Computer Engineering, National University of Singapore (NUS), Singapore 117576, Singapore Received 8 April 2014; received in revised form 4 July 2014; accepted 5 July 2014

Abstract Creating an artificial system for imparting self-cleaning and antireflective properties to photovoltaic panels by employing organic and inorganic materials has attracted much attention among researchers recently. This has been done primarily to elevate roughness and modify surface chemistry. However, such materials suffer from durability issues when placed outdoors. Therefore, this paper elaborates upon the self-cleaning and antireflective performance of hydrophilic nanostructured glass substrates. Nanostructured glass substrates were subjected to an outdoor exposure of 12 weeks at different angles of inclination. Their performance was measured in terms of optical transmission, water contact angle, analysis of dust accumulation and application to solar modules as packaging covers. The nanostructured glass samples showed improvement in self-cleaning performance of solar modules, with only an insignificant drop of 0.3% in efficiency relative to a 2% drop in a planar glass solar module over a long term exposure period. Ó 2014 Published by Elsevier Ltd.

Keywords: Nanostructured glass; Antireflection; Self-cleaning; Long term outdoor test

1. Introduction Broadband antireflective (AR) structures that can suppress reflection over a broad spectral range and for oblique angles of incidence are of great research interest (Song et al., 2010; Shimomura et al., 2010; Son et al., 2011; Yang et al., 2011). It is imperative to reduce the reflective losses of light to improve the performance of optical and optoelectronic devices. In this regard, AR structures are important for practical applications such as photovoltaics (Han et al., 2011; Munday and Atwater, 2011), light emitting diodes (Ou et al., 2011) and light sensors (Pergande et al., 2011). The reflection losses at surface interfaces, especially those of solar modules, are detrimental to their energy conversion efficiency. Further, their efficiency is

⇑ Corresponding author. Tel.: +65 65162111.

E-mail address: [email protected] (A.J. Danner). http://dx.doi.org/10.1016/j.solener.2014.07.003 0038-092X/Ó 2014 Published by Elsevier Ltd.

limited by the adhesion of grime (Jiang et al., 2011), as it scatters or absorbs the incoming light. This severely affects their long term performance. It has been seen in literature that the accumulation of dust particles can reduce the net efficiency of solar modules by 32–40% over a first eightmonth period (Nimmo and seid, 1979; Salim et al., 1988) and up to 71.8% over a period of 15 years (Ward and Lof, 1976). Recently, degradation results for cadmium telluride solar modules were reported where outdoor testing was carried out in a desert environment (Qasem et al., 2013). It was observed that the operating efficiency of the modules reduced by 34% over a period of 90 days. Detailed studies regarding the dust accumulation on solar modules have been reported by various researchers. For example, the effect of properties such as dust size and amount of dust, have been well studied (Mani and Pillai, 2010; Mohammad and Fahmy, 1993); Goossens et al. (1993), also studied the effect of wind speed on the accumulation of dust particles over solar panels.

232

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

Some coatings and structures based on inorganic (Nakata et al., 2011; Shin et al., 2011) and organic (Sainiemi et al., 2011; Wang et al., 2011) materials have been developed to provide appropriate roughness and surface chemistries for antireflective and self-cleaning effects. In addition, commercial self-cleaning coatings based on titanium dioxide (TiO2) have also been developed commercially (Bioclean, Saint Gobain and Pilkington ActivTM, Pilkington) for glass roofs, building facades and conservatories (Pilkington Glass, 2012; Li et al., 2009). However, the refractive index of TiO2 (n > 2), being greater than that of glass (n = 1.5), typically causes a reduction in transmission. This makes optimization difficult in solar applications where the efficiency is dependent on incident light intensity. Because of disadvantages in optical transmission, these companies have mainly focused on products for the building facßade industry. Nonetheless, there have been some self-cleaning coatings for solar applications reported (Giolando, 2013; Li et al., 2013) which have shown convincing results from laboratory experiments but lacked sufficient information about outdoor performance. Efficient structural morphology on the surface of packaging glass might be an alternative solution which not only provides an antireflection effect but a self-cleaning effect as well, but without introducing any foreign coating. Although outdoor weathering data of different kinds of glass substrates was first reported almost 30 years ago (Shelby et al., 1980), insufficient information is available on the performance of nanostructured glass substrates which can show antireflection effects. In this paper, a practical study on the outdoor performance and durability of planar and nanostructured glass samples is presented. Unlike other methods involving chemical coatings to provide self-cleaning effects, the non-coated hydrophilic glass studied here will be useful for practical outdoor conditions providing antireflection, self-cleaning and consequently, enhanced efficiency of solar modules. 2. Experimental details A non-lithographic technique was used to create nanostructures on the surface of borosilicate planar glass (Borofloat 33, Schott Glass). A schematic diagram of the nanostructuring process is shown in Fig. 1. The nanostructures were fabricated on planar glass by inductively coupled plasma-reactive ion etching (ICP-RIE) using nickel (Ni) nanoparticles as an etching template (Verma et al., 2011; Sakhuja et al., 2012). A 10-nm Ni film deposited on a planar glass sample using an electron beam evaporator was transformed into Ni nanoparticles using a rapid thermal annealing process. The annealing was carried out at a temperature of 600 °C for 5 min to generate randomly distributed Ni nanoparticles on planar glass substrates. The planar glass samples were dry-etched in an SF6 ambience using ICPRIE for different durations, creating nanostructures of different heights varying from 100 nm to 800 nm. The remaining Ni nanoparticles after the etching process were removed

by soaking the etched glass samples in nitric acid. The height of the nanostructures was uniform over any particular etched glass sample. The surface morphology characterization of the nanostructured glass was carried out at an oblique angle of 30° with a field emission scanning electron microscope (SEM, FEI NOVA NanoSEM). A plan view of the Ni nanoparticles on glass substrates after annealing and an oblique view of the nanostructured sample are shown in the SEM images in Fig. 2. Note that the nanostructured glass sample is covered with a 10-nm thin layer of gold to reduce the surface charging effect. The diameter of the Ni nanoparticles was calculated using MATLAB by an image analysis algorithm. The average diameter of the Ni nanoparticles was determined to be 100 nm with a coverage area of 70% of the sample surface. In order to test the practical performance and durability of the planar and nanostructured glass samples, outdoor exposure tests were carried out for a testing period of 3 months (12 weeks) on the roof of a 35-m tall building in Singapore. The planar and nanostructured glass samples were mounted flat (0°) and also inclined at 10° and 20°. The durability of the mounted samples was assessed after the testing period by carrying out the morphological characterization of the nanostructured glass samples with an SEM. Further, the performance analysis was carried out by analyzing the dust accumulation on the samples using an SEM and optical microscope. Optical transmission measurements (over a spectral range of 400–1000 nm) on the outdoor exposed samples were also carried out after every two weeks at normal incidence in a UV–visible spectrophotometer (UV-3700, Shimadzu Corporation). Water contact angle (WCA) measurements were carried out using a contact angle measurement setup (VCA optima contact angle equipment from AST Products) at room temperature. The values to be reported are the average of five measurements made on different areas of every sample. Mini solar modules of area 39.75 cm2 were subsequently prepared using multicrystalline commercial silicon solar cells with both planar glass substrates and nanostructured glass substrates of different heights of nanostructures as their packaging covers. Based on the results obtained from the outdoor exposed glass substrates, the performance of the solar modules was tested for 5 weeks. The inclination angle of the solar modules was chosen based on the best performers from the earlier outdoor exposure experiments on planar and nanostructured glass samples. I–V testing of the solar modules was carried out after every 5 days with an Oriel Xenon arc lamp solar simulator (Newport) with Air Mass (AM) 1.5 filter. The filter is used to reduce spectral mismatch between lamp and solar light of AM 1.5. 3. Results and discussion Fig. 3(a) presents the optical transmission spectrum at normal incidence for planar and nanostructured glass samples. It was seen that nanostructuring of planar glass

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

233

Deposition of thin nickel film using Electron Beam Evaporator

Nickel film (10 nm) Planar Glass

Planar Glass

Annealing of nickel film to obtain nickel nanoparticles Etching of glass substrate and removal of nickel nanoparticles

Nanostructured Glass

Planar Glass

Fig. 1. Schematic of the process to fabricate nanostructured glass samples.

(a) 100 Transmission (%)

(a)

98 96 94 92 90 88 86 84 82 80 400

Planar Glass 100 nm 200 nm 400 nm 800 nm

500

600

700

800

900 1000

Wavelength (nm)

(b)

(b) Water contact angle (degrees)

1 µm

70 60 50 40 30 20 10 0

0

200

400

600

800

Height of nanostructures (nm) Fig. 3. (a) Optical transmission spectra for glass samples of different nanostructure heights (averages of 5 measurements per sample) and (b) variation of water contact angle with the height of nanostructures on glass (averages of 5 measurements per sample).

500 nm Fig. 2. (a) SEM image of Ni nanoparticles on glass after annealing and (b) oblique cross-section view of glass nanostructures after etching and Ni removal.

improves the optical transmission with maximum improvement shown by 200-nm high nanostructures over the whole 400–1000 nm spectral range (Sakhuja et al., 2012). The increase in optical transmission can be attributed to the fact that texturing creates a gradual refractive index change at the boundary of air and glass which suppresses reflection (Muskens et al., 2008). A decrease in the optical transmission was noticed at lower wavelength values. This is due to wavelength dependent scattering losses (Xiu et al., 2008).

The antireflective properties of the nanostructured surfaces have encouraged convergence between self-cleaning structures and antireflective structures (Manca et al., 2009; Lee et al., 2006). WCAs of the planar and nanostructured glasses were measured as shown in Fig. 3(b). It was seen that planar glass exhibited a WCA  60° whereas the WCA reduced upon nanostructuring. The decrease in WCA became profound with an increase in the height of the nanostructures. It has been shown in the literature that the WCA should increase with an increase in the roughness of the substrate according to the Wenzel and Cassie–Baxter wetting models (Gao et al., 2007; Ganjoo et al., 2009), but the fundamental reason of observing an increase in hydrophilicity (or decrease in WCA) here is the capillary effect

234

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

(Zhang et al., 2005; Zorba et al., 2010). When the water droplet comes into contact with the glass nanostructures, water penetrates into the gaps of two adjacent pillars and a thin solid/liquid composite film is formed which makes the water droplet spread easily. On the other hand, the surface wettability is also governed by the surface free energy (Strobel and Lyons, 2012). Therefore, when the solid surface free energy value is equal to the liquid free energy value, it renders the solid to be hydrophilic which is likely the reason for the hydrophilicity shown by the nanostructured glass samples. In order to test the practical long term performance of the planar and nanostructured glass substrates, they were exposed outdoors for a 3-month period. Fig. 4(a)–(c) present optical transmission with respect to the exposure time for planar and nanostructured glass samples for flat

(b) 94

0° mount

Transmission (%)

Transmission (%)

(a) 94 92 90 Planar Glass 100 nm 200 nm 400 nm 800 nm

88 86 84 0

2

4

mounted and oblique mounted samples at a 600-nm wavelength. For flat-mounted samples, it was observed that the optical transmission of planar glass deteriorated from 92.5% to 87%, thus accounting for a 5% decrease over 3 months. In comparison, the nanostructured glass samples showed better performance with the best transmission of 94.5% shown by nanostructured glass with 200-nm high nanostructures. The optical transmission of flat-mounted nanostructured glass samples also decreased after outdoor exposure with the least deterioration of 2.5% shown by the 200-nm high nanostructured sample. Thus, the nanostructured glass samples showed a self-cleaning effect as compared to the planar glass sample. An abrupt decrease in the transmission was observed for the flat-mounted samples between weeks 4 and 6 due to less rainfall (National Environment Agency, Singapore Government) during this

10° mount

92 90 Planar Glass 100 nm 200 nm 400 nm 800 nm

88 86 84

6

8

10

12

0

Exposure time (week)

92 90 Planar glass 100 nm 200 nm 400 nm 800 nm

88 86 84 0

2

4

6

8

10

60 55 50 45 40 35 30 25 20

12

6

8

10

12

Planar glass - 0° inclined Planar glass - 20° inclined 200 nm - 0° inclined 200 nm - 20° inclined

0

Exposure time (week)

2

4

6

8

10

12

Exposure time (week)

(e) 60 Rainfall per week (mm)

4

(d) 65

20° mount

Water contact angle (degrees)

Transmission (%)

(c) 94

2

Exposure time (week)

(f)

50

200 nm

40

Before

30

400 nm

20

200 nm

10 0 0

2

4

6

8

10

12

After

500 nm

Week Fig. 4. (a)–(c) Variation of the optical transmission (at 600-nm wavelength) of flat mounted and inclined planar and nanostructured glass samples with the outdoor exposure time in weeks (averages of 5 measurements per sample), (d) variation of water contact angle for flat mounted and inclined planar and nanostructured glass samples with 200-nm high nanostructures (averages of 5 measurements per sample), (e) rainfall per week (National Environment Agency, Singapore Government) over the testing period and (f) SEM image of nanostructured glass sample with 200-nm high nanostructures before and after the outdoor exposure, respectively.

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

testing period (Fig. 4(e)). The decrease in optical transmission is due to the sticking of dust particles after water droplets evaporate from the surface over time. The planar glass sample mounted at an inclination of 10° also showed similar deterioration in optical transmission as the flat-mounted planar glass sample. The nanostructured glass samples inclined at 10° showed no improvement in the optical transmission and their behavior was similar to the flat-mounted control samples. However, interesting results were obtained with the samples mounted at 20°. The planar glass sample showed similar deterioration in optical transmission as its flat mounted and 10° inclined control samples. However, the nanostructured glass samples showed better performance as compared to their control samples. The trend of decrease in the optical transmission was linear as compared to the flat-mounted and 10° inclined samples with the least deterioration shown by the 200-nm nanostructured sample of 1.5%. The effect of accumulation of dust particles on glass samples can be seen from the variation in the error bar (showing standard deviation) for various data points in the transmission data for a particular sample. It can be seen in Fig. 4(a)– (c) that the variation in optical transmission increased as the exposure time progressed to 12 weeks. The percentage error ((DA/A)  100%, where the data point is represented as A ± DA) in optical transmission for the flat mounted samples (0° mount) ranged from 0.09% to 1% for the first 4 weeks where the samples were being cleaned by rain. However, after the 5th week, the self-cleaning performance of these flat mounted glass samples degraded with percentage errors as high as 1.38% (400 nm at week 12) – 1.5% (planar glass at week 10). This degradation in the optical transmission performance of glass samples was due to sticking of dust particles. On the other hand, a lower degradation was noticed for glass samples mounted at an inclination of 10° and 20°. The percentage errors for the nanostructured glass samples were as low as 0.08% over a period of 12 weeks due to their self-cleaning capability when inclined. However, the percentage error in optical transmission for planar glass was higher than that of nanostructured glass samples but improved when mounted at an inclination, which can be observed in Fig. 4(b) and (c). It is hypothesized that when dust particles come into contact with the nanostructured glass substrates, they are not able to penetrate the nanometer sized pitch between two adjacent nanostructures but sit on the glass nanostructures. Therefore, the dust particles are forced to flow along with the sheeting water which allows the nanostructured glass substrates to exhibit a self-cleaning effect as compared to the planar glass samples where dust particles mix with water and stick to the glass surface. This improvement in the self-cleaning effect can be due to the conditions favoured by the inclination angle. Inclination has a profound effect on the self-cleaning behavior which allows the flowing water to provide a sheeting effect, thus carrying the dust particles away more easily during rains. However, solar modules always have a mismatch

235

between their lab tested efficiency and practical efficiency when they are mounted outdoors at the equator. Therefore, the obtained results of optical transmission in Fig. 4(a)–(c) were further analyzed with the results in our previous reported work (Sakhuja et al., 2012) where the efficiency of solar modules was tested at different inclination angles. The inclination of solar modules at 10° and 20° decreases their efficiency by 0.5% and 1% respectively due to less illumination as an effect of different effective area of irradiation. Taking into consideration the offset created by optical transmission loss due to dust accumulation over a long term exposure period, the overall efficiency loss would be 4%, 3% and 2.5% for flat mounted, 10° and 20° inclined solar modules. Thus, the self-cleaning effect provided by the 20° inclined nanostructured glass samples increases the overall efficiency of solar modules. WCA measurements were also carried out for the planar and nanostructured glass samples every week after beginning outdoor exposure. Fig. 4(d) presents the variation of WCA versus the exposure time for the flat-mounted and 20° inclined planar glass and 200-nm high nanostructured glass samples. It was seen that the flat-mounted and 20° inclined planar glass samples and 20° inclined nanostructured sample presented a significant variation in the WCA over the testing period. However, the nanostructured sample inclined at 20° presented less fluctuation. The reason behind such fluctuations can be attributed to the capillary effect. The accumulation of dust particles creates a certain roughness on the surface of the sample which leads to a variation in the surface morphology, thus inducing fluctuations in the WCA. Fig. 4(f) presents SEM images of 20° inclined nanostructured sample with 200-nm high nanostructures before and after outdoor exposure. It was observed that there was no change in the height of nanostructures which shows that the glass nanostructures are robust and appropriate for outdoor use. The decrease in the optical transmission and changes in WCA were solely due to the accumulation of dust particles on the surface of the samples. To further understand the reason behind the reduction in the optical transmission for the planar and nanostructured samples, dust analysis was carried out for flatmounted and 20° inclined planar glass and 200-nm high nanostructured glass samples. SEM and optical microscope images were obtained for them and then the images were analyzed using the free software Image J. Fig. 5(a) and (b) present the optical microscope images for 20° inclined planar and 200-nm height nanostructured glass samples, respectively. Fig. 5(d) and (e) present SEM images for 20° inclined planar and 200-nm height nanostructured glass samples, respectively. It can be clearly seen from Fig. 5(a) and (d) that the planar glass samples are dirty as compared to the 200-nm height nanostructured glass inclined at the same angle. Fig. 5(c) and (f) present the dust analysis for flat-mounted (SEM and optical microscope images not shown here) and 20° inclined planar glass and 200-nm height nanostructured glass, respectively. It can be clearly

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

(a)

Planar glass - 20

(b)

200 nm - 20

200 µm

(c)

200 µm

Number of particles

236

1800 1600 1400 1200 1000 800 600 400 200 0

Planar glass - 0°inclined Planar glass - 20° inclined 200 nm - 0° inclined 200 nm - 20° inclined

0

20

40

60

80 100 120

(d)

Planar glass - 20

(e)

200 nm - 20

30 µm

30 µm

(f)

Number of particles

Dust size (µm2) 3000 2500 2000 1500 1000 500

Planar glass - 0° inclined Planar glass - 20° inclined 200 nm - 0° inclined 200 nm - 20° inclined

100 80 60 40 20 0 0.00 0.07 0.15 0.22 0.30

0 0.00 0.05 0.10 0.15 0.20 0.25 0.30

Dust size (µm2) Fig. 5. (a) and (b) Optical microscope images for 20° inclined planar glass sample and nanostructured glass sample with 200-nm high nanostructures after the long term outdoor exposure, (c) number of particles on the surface of planar glass sample and nanostructured glass sample with 200-nm high nanostructures versus the particle/dust size in an area of 0.64 mm2 after the long term outdoor exposure, (d) and (e) SEM images for 20° inclined planar glass sample and nanostructured glass sample with 200-nm high nanostructures captured after the long term outdoor exposure and (f) number of particles on the surface of planar glass sample and nanostructured glass sample with 200-nm high nanostructures versus the particle/dust size in an area of 6400 lm2 after the long term outdoor exposure.

observed that the planar glass samples are affected more by the accumulation of dust as compared to nanostructured glass sample which accounts for the deterioration of their optical transmission behavior over long term outdoor exposure. The phenomenon of deterioration of optical transmission can be explained on the basis of Mie scattering and Rayleigh scattering (Taylor, 2011; Saleh and Teich, 2007; Hecht, 2002). The pattern of scattering depends on the wavelength of light and the particle size. It can be seen from Fig. 5(f) that the size of dust particles on the flatmounted and 20° inclined planar glass samples is mainly distributed around 0.05 lm2 (250 nm in diameter). Rayleigh scattering is applicable to small nanometer-sized particles, scattering light both in the forward and backward directions, whereas Mie scattering has no size effect and converges to the limit of geometric optics for large particles, thus scattering light mostly in the forward direction; forward scattering increasing with larger particles. Since the dominant particle size on the planar glass samples is in the nanometer range, light is scattered according to the Rayleigh criterion, thus reducing the optical transmission for planar glass samples. On the other hand, large sized particles are predominant on the nanostructured glass samples which scatter light mostly in the forward direction according to Mie scattering. In this manner, the optical transmission of nanostructured glass samples is less affected compared to planar glass samples. We now discuss a subsequent experiment where nanostructured glasses were then tested as packaging covers for the solar cells. It has been observed that inclination of glass samples is required for optimum performance.

For this purpose, mini solar modules of area 39.75 cm2 were fabricated with planar glass and nanostructured glasses with different heights of nanostructures as their packaging covers. The solar modules were tested outdoors at an inclination of 20° (best performance from optical transmission results). Fig. 6(a) and (b) represent the variation of the short circuit current density and efficiency of the solar modules with different packaging covers respectively as a function of the exposure time. Each data point represents a single module. It can be seen that the short circuit current density of the planar glass solar module reduces prominently by 5 mA/cm2 after an outdoor exposure of 5 weeks. A significant drop of 2% was also observed in the efficiency. This reduction in the solar module parameters is obviously due to the accumulation of dust particles on the planar packaging glass. In comparison, the nanostructured glass solar modules showed better performance. Among the nanostructured samples, the best performance was observed for the nanostructured glass solar module with 200-nm high nanostructures on the packaging glass cover with a reduction in short circuit current density and efficiency by 1 mA/cm2 and 0.3% respectively; thus providing self-cleaning and antireflective effects with increased power conversion efficiency for the underlying solar cell. 4. Conclusion The self-cleaning behavior, outdoor performance and durability of planar and nanostructured glass samples have been systematically investigated over an outdoor exposure period of 12 weeks. It was observed that inclination was

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

Short circuit current density (mA/cm2)

(a)

39 38 37 36 Planar glass 100 nm 200 nm 400 nm 800 nm

35 34 33 32 0

5

10

15

20

25

30

35

Exposure time (days)

Efficiency (%)

(b) 16.5 16.0 15.5 15.0

Planar glass 100 nm 200 nm 400 nm 800 nm

14.5 14.0 13.5 0

5

10

15

20

25

30

35

Exposure time (days) Fig. 6. (a) Variation of short circuit current density with exposure time for planar and nanostructured glass solar modules (1 measurement per sample) and (b) variation of efficiency with exposure time for planar and nanostructured glass solar modules (1 measurement per sample).

required for the hydrophilic nanostructured glass samples for lower dust accumulation and self-cleaning effect. The nanostructured glass sample with 200-nm high nanostructures on its surface provided superior antireflective and self-cleaning effects compared to a planar glass sample over the testing period. This particular nanostructured glass sample also provided the best performance when tested as the packaging cover of a solar module, with reduction in efficiency by only 0.3% over a testing period of 5 weeks. Thus, this practical demonstration of the self-cleaning and antireflective performance of nanostructured glass samples can find potential application in outdoor optoelectronic devices such as solar modules without any application of chemical treatment or coatings to the system. Acknowledgement This work is supported by Singapore National Research Foundation (NRF) grant number NRF2008EWTCERP02-032. References Ganjoo, S., Azimirad, R., Akhavan, O., Moshfegh, A.Z., 2009. Persistent superhydrophilicity of sol-gel derived nanoporous silica thin films. J. Phys. D Appl. Phys. 42, 025302–025308. Gao, X., Yao, X., Jiang, L., 2007. Effects of rugged nanoprotusions on the surface hydrophobicity and water adhesion of anisotropic micropatterns. Langmuir 23, 4886–4891.

237

Giolando, D.M., 2013. Nano-crystals of titanium dioxide in aluminium oxide: A transparent self-cleaning coating applicable to solar energy. Sol. Energy 97, 195–199. Goossens, D., Offer, Z., Zangvil, A., 1993. Wind tunnel experiments and field investigations of eolian dust deposition on photovoltaic solar collectors. Sol. Energy 50, 75–84. Han, K.S., Shin, J.H., Yoon, W.Y., Lee, H., 2011. Enhanced performance of solar cells with antireflection layer fabricated by nano-imprint lithography. Sol. Energy Mater. Sol. C. 95, 288–291. Hecht, E., 2002. Optics, fourth ed. Addison-Wesley, London. Jiang, H., Lu, L., Sun, K., 2011. Experimental investigation of the impact of airborne dust deposition on the performance of solar photovoltaic (PV) modules. Atmos. Environ. 45, 4299–4304. Lee, D., Rubner, M.F., Cohen, R.E., 2006. All-nanoparticle thin-film coatings. Nano Lett. 6, 2305–2312. Li, J., Lu, Y., Lan, P., Zhang, X., Xu, W., Tan, R., Song, W., Choy, K.L., 2013. Design, preparation, and durability of TiO2/SiO2 and ZrO2/SiO2 double-layer antireflective coatings in crystalline silicon solar modules. Sol. Energy 2013, 134–142. Li, Y., Zhang, J., Zhu, S., Dong, H., Jia, F., Wang, Z., Sun, Z., Zhang, L., Li, Y., Li, H., Xu, W., Yang, B., 2009. Biomimetic surfaces for highperformance optics. Adv. Mater. 21, 4731–4734. Manca, M., Cannavale, A., Marco, L.D., Arico, A.S., Cingolani, R., Gigli, G., 2009. Durable superhydrophobic and antireflective surfaces by trimethylsilanized silica nanoparticles-based sol-gel processing. Langmuir 25, 6357–6362. Mani, M., Pillai, R., 2010. Impact of dust on solar photovoltaic (PV) performance: research status, challenges and recommendations. Renew. Sustain. Energy Rev. 14, 3124–3131. Mohammad, S.E.-S., Fahmy, M.H., 1993. Effect of dust with different physical properties on the performance of photovoltaic cells. Sol. Energy 51, 505–511. Munday, J.N., Atwater, H.A., 2011. Large integrated absorption enhancement in plasmonic solar cells by combining metallic gratings and antireflection coatings. Nano Lett. 11, 2195–2201. Muskens, O.L., Diedenhofen, S.L., Weert, M.H.M., Borgstrom, M.T., Bakkers, E.P.A.M., Rivas, J.G., 2008. Epitaxial growth of aligned semiconductor nanowire metamaterials for photonic applications. Adv. Func. Mater. 18, 1039–1046. Nakata, K., Sakai, M., Ochiai, T., Murakami, T., Takagi, K., Fujishima, A., 2011. Antireflection and self-cleaning properties of a moth-eye-like surface coated with TiO2 particles. Langmuir 27, 3275–3278. Document on “Daily values of Meteorological elements in 2012”. National Environment Agency, Singapore Government. . Nimmo B., Seid A.M.S., 1979. Effect of dust on the performance of thermal and photovoltaic flat plate collectors in Saudi Arabia. Preliminary results. In: Proceedings of the Second Miami International Conference on Alternative Energy Sources. Ou, Y., Corell, D.D., Dam, H.C., Petersen, P.M., Ou, H., 2011. Antireflective sub-wavelength structures for improvement of the extraction efficiency and color rendering index of monolithic white light-emitting diode. Opt. Express 19, A166–A172. Pergande, D., Geppert, T.M., Rhein, A., Schweizer, S.L., Wehrspohn, R.B., Moretton, S., Lambrecht, A., 2011. Miniature infrared gas sensors using photonic crystals. J. Appl. Phys. 109, 083117–083123. Information on Pilkington Self-Cleaning Glass, 2012, . Qasem, H., Betts, T.R., Gottschalg, R., 2013. Spatially-resolved modelling of dust effect on cadmium telluride photovoltaic modules. Sol. Energy 90, 154–163. Sainiemi, L., Jokinen, V., Shah, A., Shpak, M., Aura, S., Suvanto, P., Franssila, S., 2011. Non-reflecting silicon and polymer surfaces by plasma etching and replication. Adv. Mater. 23, 122–126. Sakhuja, M., Son, J., Verma, L.K., Yang, H., Bhatia, C.S., Danner, A.J., 2012. Omnidirectional study of nanostructured glass packaging for

238

M. Sakhuja et al. / Solar Energy 110 (2014) 231–238

solar modules. Prog. Photovoltaics Res. Appl.. http://dx.doi.org/ 10.1002/pip. 2276. Saleh, B.E.A., Teich, M.C., 2007. Fundamentals of Photonics, second ed. Wiley, Hoboken. Salim A.A., Huraib F.S., Eugenio N.N., 1988. PV power – Study of system options and optimization. In: Proceedings of the 8th European Photovoltaic Solar Energy Conference. Shelby, J.E., Vitko, J., Pantano, C.G., 1980. Weathering of glasses for solar applications. Sol. Energy 3, 97–110. Shimomura, H., Gemici, Z., Cohen, R.E., Rubner, M.F., 2010. Layer-bylayer-assembled high-performance broadband antireflection coatings. ACS Appl. Mater. Interfaces. 2, 813–820. Shin, J.H., Han, K.S., Lee, H., 2011. Anti-reflection and hydrophobic characteristics of M-PDMS based moth-eye nano-patterns on protection glass of photovoltaic systems. Prog. Photovoltaics Res. Appl. 19, 339–344. Son, J., Verma, L.K., Danner, A.J., Bhatia, C.S., Yang, H., 2011. Enhancement of optical transmission with random nanohole structures. Opt. Express 19, A35–A40. Song, Y.M., Jang, S.J., Yu, J.S., Lee, Y.T., 2010. Bioinspired parabola subwavelength structures for improved broadband antireflection. Small 6, 984–987. Strobel, M., Lyons, C.S., 2012. An essay on contact angle measurements. Plasma Processes Polym. 8, 8–13.

Taylor, J.M., 2011. Scattering theory, optical binding phenomena: observations and mechanism. Springer Theses, 11–49. Verma, L.K., Sakhuja, M., Son, J., Danner, A.J., Yang, H., Zeng, H.C., Bhatia, C.S., 2011. Self-cleaning and antireflective packaging glass for solar modules. Renew. Energ. 36, 2489–2493. Wang, H., Xue, Y., Ding, J., Feng, L., Wang, X., Lin, T., 2011. Durable, self-healing superhydrophobic and superoleophobic surfaces from fluorinated-decyl polyhedral oligomeric silsesquioxane and hydrolyzed fluorinated alkyl silane. Angew. Chem. Int. Ed. 50, 11433–11436. Ward, J.C., Lof, G.O.G., 1976. Long term (18 years) performance of a residential solar heating system. Sol. Energy 18, 301–306. Xiu, Y., Zhang, S., Yelundur, V., Rohtagi, A., Hess, D.W., Wong, C.P., 2008. Superhydrophobic and low light reflectivity silicon surfaces fabricated by hierachical etching. Langmuir 24, 10421–10426. Yang, W., Yu, H., Tang, J., Su, Y., Wan, Q., Wang, Y., 2011. Omnidirectional light absorption in thin film silicon solar cell with dual anti-reflection coatings. Sol. Energy 85, 2551–2559. Zhang, X.T., Sato, O., Taguchi, M., Einaga, Y., Murakami, T., Fujishima, A., 2005. Self-cleaning particle coating with antireflection properties. Chem. Mater. 17, 696–700. Zorba, V., Chen, X., Mao, S.S., 2010. Superhydrophilic TiO2 surface without photocatalytic activation. Appl. Phys. Lett. 96, 093702– 093704.