Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

CFD simulation and validation of self-cleaning on solar panel surfaces with superhydrophilic coating

CFD simulation and validation of self-cleaning on solar panel surfaces with superhydrophilic coating Solar panel conversion efficiency, typically in the twenty percent range, is reduced by dust, grime, pollen, and other particulates that accumulate on the solar panel. Cleaning dirty panels to maintain peak efficiency, which is especially hard to do for large solar-panel arrays. To develop a transparent, anti-soiling Nano-TiO coating to minimize the need for occasional cleaning is the purpose of this study. In our study, a 2D rainwater runoff model along tilted solar panel surface based on the Nusselt solution was established to have better understanding and predicting the behavior of runoff rain water, especially in contact with solar-panel surfaces with Nano-TiO coating. Our simulation results demonstrate that solar-panel surfaces with Nano-TiO coating create a superhydrophilic surface which cannot hold water, showing features of more pronounced in increasing runoff water film velocity comparing to the uncoated surfaces during raining event resulting in better effect of self-cleaning. Validation of our model was performed on titled solar panels for real time outdoor exposure testing in Switzerland. It is found that the dust particles are not easy to adhere to the coated surfaces of the slides comparing with uncoated surfaces, showing great potential for its use in harsh environmental conditions. This study suggests that superhydrophilic self-cleaning solar panel coating maximize energy collection and increases the solar panel’s energy efficiency. Keywords: Nano-TiO coating; Self-cleaning; 2D runoff model; Solar-panel surfaces Introduction dust accumulation on the tilted glass plates revealed a Solar PV technology is well-proven for producing electri- reduction in plate-transmittance ranging from 64 to city, where the global production has been increasing 370 17 %, for tilt angles ranging from 0° to 60° respectively times than that in 1992 (Kazmerski 2011). The output of a after 38 days of exposure. A reduction of 30 % in useful PV module is usually rated by manufacturers under energy gain was observed by the horizontal collector Standard Test Conditions (STC), where each module is after three days of dust accumulation. Salim et al. tested under a temperature of 25 °C; solar radiation of (1988) indicated that a 32 % reduction in performance 1000 W/m , air mass of 1.5 spectra and wind speed of after eight months occurred under desert conditions in 2 m/s. However, these conditions are different from the KSA. Goossens and Kerschaever (1999) showed that conditions in the practical fields. With the increasing use the deposition of fine dust particles on the cover of PV of PV systems, it is vital to study meteorological parame- modules significantly affects the performance of these ters that affect the performance of these systems such as modules. Katz (2008) reported that the dirt on PV humidity, dust, temperature and wind speed. modules caused a 2 % of power reduction as compared The effect of dust on PV modules performance has to clean PV modules. However, Sayigh (2009) reported been investigated in different ways as can be found in a power decrease of about 11.5 % in a PV module ex- the literature. Wakim (1981) claimed that 17 % of PV posed for only 72 h in Riyadh, Saudi Arabia. power is lost due to dust deposition on PV modules in Kazem et al. (2013) recently conducted experiments Kuwait city. Sayigh et al. (1985) reported the effect of concerning the effects of air pollutants including red soil, ash, sand, calcium carbonate, and silica on the solar * Correspondence: jin.hu@heig-vd.ch power generated. Their results show that the reduction Institute of Thermal Engineering (IGT), University of Applied Sciences of in PV voltage and power strongly depends on pollutant Western Switzerland, Av. des Sports 20, CH-1400 Yverdon-les-Bains, Switzerland type and deposition level. The highest reduction in PV Full list of author information is available at the end of the article © 2015 Hu et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Hu et al. Future Cities and Environment (2015) 1:8 Page 2 of 15 voltage (25 %) is recorded when the ash pollutant is conducted on PV panels under natural environmental used. conditions in Switzerland. A study conducted by Harvard University students showed that energy losses of solar panels due to soiling (of Superhydrophilic and hydrophilic surfaces the surface) vary between 9 and 20 % of the possible Rhykerd et al. (1991) measured ellipsometrically the thick- energy absorption (Wack 1980). This is a significant de- ness of the adsorbed water film on a fused silica surface crease of energy and brings the subject of solar panel and found it ranging from 2.4 to 9.0 nm, depending on the cleaning to attention. However, most solar panels are water vapor pressure. Staszczuk (1985) used gas chroma- placed in regions with difficult accessibility such as a roof. tography to determine the water adsorption isotherm on This combined with high voltage proximity makes clean- quartz at 20 °C and found that about sixteen statistical ing solar panels expensive and difficult. water layers adsorbed from a gas phase saturated with There are numerous ways to clean a surface: from abra- water vapor. Also, similar experiments Janczuk et al. (1983) sive techniques like sandblasting to laser cleaning and using the chromatographic technique showed that about water spraying. Some of the factors of conventional fifteen statistical water layers may adsorb onto a marble methods shall be mentioned: Possible damage caused by surface. Water films with thicknesses from 1.0 to 8.0 nm aggressive cleaning methods that may roughen the surface were also reported for muscovite mica (Perevertaev et al. and the surface is more susceptible to new smudging in 1979). the future; as cleaning is an intensive work, therefore costs Anna Lee et al. (2012) studied the condensation be- a lot for labor work; massive use of chemicals in cleaning haviors of the surfaces with different wettability and will cause environmental problems. roughness. They concluded that the hydrophilic sur- Blossey (2003) claimed two routes to self-cleaning are faces are superior in condensation rate from its early emerging, which work by the removal of dirt by either film stage when the dry surfaces directly face humid air. In or droplet flow. In other words, water film flows, either on their experiments, where the film thickness δ on hydro- hydrophilic surface or water drop flow on hydrophobic philic surface is estimated by the Nusselt theory that surface, are methods to achieve self-cleaning. balances the viscous shear force and the gravitational Nano-TiO transparent coating can make a substrate force (Incropera and DeWitt 2002) surface to be photocatalytic and hydrophilic. UV-radiation "#1 from daylight reacts with dirt and organic deposits, oxi- 4k μðÞ T −T L l s w p dizes them and breaks their adherence to the surface. δ≈ ð1Þ gρρ−ρ h fg There is sufficient evidence to support the removal of organic contaminants (Ohtsu et al. 2009) and bacteria (Dunnill et al. 2009) adsorbed on a TiO surface by the They obtained δ is ~85 μm. photooxidization process. Due to TiO films that exhibit It is commonly accepted that in the vicinity of hydro- hydrophilicity (Ohtsu et al. 2009), surface raindrops philic interfaces water organizes into ice-like water, which spread as a film on the surface, ensuring the loosened dirt project from the surface by a few nanometers (Noguchi particles are carried away from the surface during rainy et al. 2008; Tian et al. 2008; Smith et al. 2004). Yoo et al. weather. (2011) used NMR spectroscopy to study impact of hydro- The phenomenon of soiled and stained facades is serious philic surfaces on interfacial water dynamics. Their find- in the cities. Industrial pollution has increased these prob- ings confirm the existence of highly restricted water layers lems. Especially in densely populated urban areas, building adsorbed onto hydrophilic surfaces and dynamically stable facades are covered by inorganic pollutants like nitrogen water beyond the first hydration layers. Thus, aqueous oxides and organic pollutants like benzene. All these regions on the order of micrometers are dynamically pollutants are destructive for buildings and unhealthy for different from bulk water. people. Nano-TiO coating can be easily applied on fa- cades to reduce the concentration of nitric oxides and Precursor film other toxic substances like benzene which can provoke When a perfectly wetting liquid spreads relatively slowly on respiratory problems and increase smog formation. a clean smooth substrate, a very thin precursor film may Hereby, Nano-TiO coating is very good candidate to propagate ahead of the apparent or macroscopic wetting reduce air pollution in cities. line with the average thickness of the film determined by The purpose of this study is to establish a 2D model of material property of liquid phase and solid wall (Heslot water slide on inclined self-cleaning surface (nano-TiO et al. 1989 and 1990). The first reported observation of an coated superhydrophilic solid surface) for PVT/solar ‘invisible’ film spreading ahead of the edge of a macro- panels, specifically targeting at solar panel surfaces. The scopic drop stems from the pioneering work by (Hardy self-cleaning effect of the super-hydrophilic coating is 1919 and 1936), Hardy was unable to detect its presence Hu et al. Future Cities and Environment (2015) 1:8 Page 3 of 15 directly at that time. Numerous studies have subsequently to completely spread a macroscopic drop (Voinov 1976; confirmed the existence of precursor films using ellipso- Tanner 1979). The precursor which advances at a seem- metry Beaglehole (1989); Bascom et al. (1964), interference ingly faster rate than the normal contact line is indeed microscopy patterns Bascom et al. (1964), and polarized well documented. It has been found that the velocity at reflection microscopy (Ausserre et al.1986).Simultaneous which the precursor film advances depends on the mate- observations of the moving droplet and of the fringe pat- rials of the gas/liquid/solid system and can vary in wide tern by the laser ellipsometry (Ueno and Watanabe 2005) range. Bascom et al. 1964 reported the speed of propaga- reveal the existence of the precursor film, which is ahead tion of the precursor film of squalane on stainless steel is −4 the moving contact line and traveled with varying its about 10 cm/s while indirect measurement for water on profile. glass Marmur & Lelah (1980) suggest that there the pre- If the fluid is perfectly wetting, we have seen that van cursor film velocity is of the order of 10 cm/s. As a result, der Waals forces lead to the formation of a precursor film the existence of precursor film makes the ACL moving on in front of the contact line (de Gennes 1985). Wang the film rather than on a real solid surface. In other words, (2003) in his thesis claimed the surface feature, particu- the ACL moves over the “wet” solid surface instead of the larly micro-structure or roughness, promotes the forma- “dry” solid surface. tion and development of thin precursor film, they claimed that the precursor film spreads much faster than the Precursor film on dry surface movement of the apparent contact line (ACL) on a rough As stated by de Gennes (1985) and Joanny and de Gennes solid surface. Amit Sah (2014) studied precursor film with (1984), long-range forces, for example, van der Waals glass capillaries and cover slips. His experimental results forces, should be taken into account in the spreading imply that the precursor film moving ahead of the contact phenomenon of precursor film. It is found that, at suffi- line controls the wetting behavior. Yuan and Zhao (2010) ciently short time, when the macroscopic droplet still acts using molecular dynamics (MD) simulations to explore as a reservoir, the behavior of the precursor film is diffu- the atomic details and the transport properties of the pre- sive and the radial extension length (l) on a macroscopic cursor film in dynamic wetting. Their results showed that scale follows a universal time dependence (t) of the form: the molecules which finally formed the precursor film came almost from the surface in the initial state of the l ¼ Dt ð2Þ droplet. The fast propagation of the precursor film is owing to continuous and very fast diffusing of the surface where D is the diffusion coefficient of the liquid within water molecules that have the highest self-diffusion coeffi- the precursor film. It is different from the conventional cient to the front of the precursor film allows for fast diffusion coefficient describing the random motion of propagation of the precursor film. Precursor film propa- particles in the bulk liquid phase or on solid substrates. gates fast with low energy dissipation. As a matter of fact, D also depends on the driving forces The typical thickness of the precursor film moving in which cause the film spreading. front of macroscopic body of fluid is in the range of As shown in Fig. 1, the macroscopic wedge of liquid is 500–3000 Å (Bascom et al. 1964; Beaglehole 1989). For a advancing at a constant velocity U on a solid dry surface fixed volume of water droplet, according to Tanner’s laws into an external air. The dynamic contact angle is θ (Tanner 1979), the spreading of the macroscopic part of and the wedge is preceded by a precursor film of length the droplet is rather slow; it would take a very long time l. The dynamic contact angle θ , which depends on the Fig. 1 Precursor film on dry surface Hu et al. Future Cities and Environment (2015) 1:8 Page 4 of 15 advancing velocity U, the liquid/vapor interfacial tension and TK on oxidized silicon wafers. Their analysis revealed γ and the liquid viscosity η through Tanner’s law (Tanner the dependence of D on relative humidity (RH). Overall, 1979). within the range of 20–90 % RH, D varies by more than two orders of magnitude, attaining the values observed for Precursor film on superhydrophilic/hydrophilic surface mesoscopic precursor films. This agrees well with the Hydrophilic surfaces adsorb water from the environment observations made by (Valignat et al. 1998 and Vou’e and the amount of water depositing on the hydrophilic et al. 1999). Such a remarkable enhancement of D has surface depends on the relative humidity. It is generally been explained by the fact that at this value of RH the accepted that under ordinary atmospheric conditions, patches of water on the substrate start to overlap and form hydrophilic surfaces adsorb at least a monolayer of water. a tortuous connected structure on which the spreading of For example, a clean glass surface is covered with a molecules encounters a very low friction. This implies on monolayer of adsorbed water at relative humidity of superhydrophilic/hydrophilic surface, the length of precur- around 30–50 % at 20 °C (Razouk and Salem 1948). sor film is much longer than that of the ordinary surfaces. Formation of a water film composed of as many as twenty When viewed at the atomic scale, the precursor film plays molecular layers, or more, may occur at the clean surface an important role in the dynamic wetting process on of high-energy solids, especially at high relative humidities hydrophilic substrate. >90–95 % (Zisman 1965). As showninFig.2,beforethe spreading of liquid on Water appears to have unusual lubricating properties superhydrophilic/hydrophilic surface, the solid surface is and usually gives wearless friction with no stick slip (Raviv already covered with a thin film of water with thickness h et al. 2004). It is also interesting that a 0.25 nm thick due to absorption/condensation on hydrophilic surface. water film between two mica surfaces is sufficient to bring The advancing velocity of the macroscopic wedge of liquid the coefficient of friction down to 0.01−0.02, a value that is not the average velocity u,but rather u′ (Joanny and de corresponds to the unusually low friction of ice. The ef- Gennes 1984). Obviously, if h ≠ 0, u′ > u. fectiveness of water film only 0.25 nm thick to lower the friction force by more than an order of magnitude is at- u ¼ u ð3Þ tributed to the “hydrophilicity” of the mica surface and to ξ−h the existence of a strongly repulsive short-range hydration force between such surfaces in aqueous solutions, which We conclude when the superhydrophilic/hydrophilic effectively removes the adhesion-controlled contribution was previously wetted by a very thin layer of the inner to the friction force (Berman et al. 1998). Clearly, a single water, the advancing velocity of the macroscopic wedge monolayer of water can be a very good lubricant - much of liquid is faster than that of dry solid surface. better than most other monomolecular liquid films (Ruths & Israelachvili 2011). It also stresses the inertial nature of Water spreading dynamics the spreading: a viscocapillary motion would have been When a liquid drop contacts a wettable surface, the liquid much quicker on a pre-wetted substrate, because of the spreads over the solid to minimize the total surface energy. lubricating effect of this water layer (Biance et al. 2004). For low viscous liquids (as water), a power law of the drop Villette et al. (1996) focused specifically on the role of spreading can be observed during almost all the evolution water on the spreading of molecular films of non-volatile varying the relative different contributions (inertia, gravity, liquids PDMS, PDMS with hydroxyl ends (PDMS-OH) viscosity, density, volume, surface tension…etc.). Fig. 2 Precursor film on superhydrophilic/hydrophilic surface Hu et al. Future Cities and Environment (2015) 1:8 Page 5 of 15 Dynamic wetting may proceed with three stages. to n =1/7 (Kavehpour et al. 2002; Oron et al. 1997; Ehrhard About 1 to 100 μs after the drop contacts the surface, 1993). inertia of the moving drop resists the capillary force that This extremely slow dynamics emerges from a balance drives a high speed spreading (in the order of 1 m/s). between surface tension and viscous forces close to the The spreading dynamics of low-viscosity drops (e.g. contact line (Bonn et al. 2009). 0.5 water) follow a power law r = Kt , which is independent of the liquid viscosity and surface wettability (Biance Runoff equations for thin film flow et al. 2004; Bird et al. 2008; Carlson et al. 2011; Chen The model of gravity-driven flow was based on the as- et al. 2011; Winkels et al. 2012; Stapelbroek et al. 2014). sumption that the inclined solid surface is initially covered The quantity r is the spreading radius, t is the spreading with a liquid film. We focus on the gravity-driven film time and K is a coefficient. flow. The solid surface is often initially dry and is gradually With time between ~0.1 to 10 ms, the drop spreading covered by the liquid film as the leading edge of the film speed is still high (~0.1 m/s), wetting is still dominated moves down under the action of gravity. by inertia. However, the surface wettability starts to in- Consider a film of viscous liquid of density ρ and viscosity fluence spreading and the spreading radius grows with μ, flowing down a plane inclined at an angle α, as illustrated time according to another power law r = K′ t (Bird in Fig. 3, and the flow is assumed to be two-dimensional, et al. 2008). K′ is another coefficient and the exponent α with no variations in the direction normal to the plane of is only dependent on the equilibrium contact angle θ . the sketch. Far away from the leading edge the film is ap- eq Experiments of Stapelbroek et al. (2014) reveal a deviation proximately flat and, therefore, the flow is described by the from a pure power-law, the cross-over from the 1/2 power constant-thickness solution, with the average flow velocity law to the final equilibrium radius displays a universal dy- ρgd sin(α/(3 μ)). It is convenient to define the characteristic namics. This cross-over is governed only by the final con- flow velocity by: tact angle, regardless of the details of the substrate. ρgd Bird et al. (2008) considered the energy dissipation U ¼ sinα ð6Þ during spreading and derived a power law which associ- ates α with θ . eq Here, the surface tension σ is assumed constant. At qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi the leading edge of the film, the liquid–gas interface is α ¼ C F ϑ þ cosθ ð4Þ eq eq assumed to be meeting the solid surface at a contact angle θ, as shown in Fig. 3 and the capillary number Experimental study showed that α increases from based on this velocity as: ~0.25 for θ ≈ 120° to ~0.5 for θ ≈ 0° (Bird et al. 2008). eq eq μU After the two inertial stages aforementioned, on Ca ¼ ð7Þ strongly hydrophilic (θ < < ~57.3° or 1 radian) or com- eq pletely wetting surfaces (θ ≈ 0°), a third wetting stage eq was found. In this stage, the wetting speed is much Lubrication approximation of thin liquid film slower than in the previous two stages, it takes a much We model the spreading of a thin liquid film on an inclined longer time for viscous liquids to completely spread and substrate. We use the lubrication approximation of the R < 1. In this long time limit stage, the viscous friction Navier–Stokes equations, and we provide appropriate initial inside the drop is the main source opposing capillarity. and boundary conditions. Within the framework of the The wetting dynamics follows 1/10 power law for small lubrication approximation, the velocity of the fluid is drops of negligible gravity effects (Tanner 1979; Cazabat depth-averaged over the thickness of the film (Greenspan and Cohen-Stuart 1986; Levinson et al. 1988; Chen and 1978). Following this approach, one obtains the average Wada 1989; Rafai et al. 2002): fluid velocity, 0:1 0:1 v ¼ðÞ u; v ð8Þ r ¼ R t ð5Þ μR v ¼ − ½ ∇p−ρgsinα i ð9Þ For largedrops dominating bygravityeffects, theradius 3μ of spreading pattern is to follow Lopez’slaw of a1/8 power law in the gravitational regime for large droplets (Smith; where ∇ =(∂x, ∂y), h is the fluid thickness, p is the pressure, Lopez et al. 1976; Yeo 2008). However, as the radius grows μ is the viscosity, ρ is the density, g is gravity, and α is the beyond the capillary length, the drop changes toward a inclination angle of the plane of the substrate. The coordin- “pancake” shape of constant thickness, curved only at the ate frame is chosen so that i points down the incline, and j rim, and the main driving force is now gravity, leading is the transverse direction in the plane. We note that Eq. 10 Hu et al. Future Cities and Environment (2015) 1:8 Page 6 of 15 Fig. 3 Thin film flows down a plane inclined at an angle α assumes no-slip boundary condition at the fluid/solid inter- the solid surface is pre-wetted with the wetting layer face. The pressure includes the hydrostatic component, and thickness h' (scaled by h) as a regularizing method in this the contribution following from the Laplace-Young bound- work. ary condition at the fluid-air interface. Thus, the lubrication approximation reduces the Navier– Stokes equations to this nonlinear fourth order partial p ¼ −γ∇ h þ ρghcosα ð10Þ differential equations that govern the time evolution of the film thickness h(x, y, t). As we focus on long-scale evolution of liquid films, we Nondimensionalization of thin film equations scale h by the height h' of the precursor film and we We consider a layer of liquid on a plane substrate inclined define the scaled in-plane coordinates and time by to the horizontal at an angle α. The fluid is Newtonian and x t incompressible of density ρ,viscosity μ and surface tension ðÞ x; y; t¼ ; ; where x y t c c γ. We start with the following 2-D lubrication equation (Diez and Kondic 2002): 2 ′ 3 a h ∂h x ¼ ð12Þ ¼ −∇⋅ðÞ hv sinα ∂t 3 2 3 ¼ − ∇ γh ∇∇ h−ρgh ∇hcosα þ ρghsinα i 3μ 3μ a x t ¼ ð13Þ ð11Þ 2 h sinα Here h is the height of the fluid given as a function of qffiffiffiffi x and y. The (x,y) plane is parallel to the substrate and And a ¼ is the capillary length, the velocity scale is ρg the x direction is the direction of the slope. Then, g is chosen naturally as U ¼ ; and the capillary number is de- the acceleration of gravity and i is the unit vector in the μU x direction There are several possibilities to obtain non- fined by ¼ : Using this dimensionless form, Eq. (12) for dimensional variables. h ¼ ; the lubrication equation given above becomes: All the theoretical and computational methods of the spreading drop problem require some regularizing mech- anism - either assumption of a precursor film in front of ∂h ∂ ∂ h ∂h þ h −DðÞ α þ 1 ¼ 0 ð14Þ the apparent contact line (Troian et al. 1989; Bertozzi and   3 ∂t ∂x ∂x ∂x Brenner 1997), or relaxing the no-slip boundary condition at fluid–solid interface (Greenspan 1978; Hocking and Rivers 1982). However, the computational performance of where the single dimensionless parameter DðÞ α ¼ðÞ 3Ca the precursor film model is shown to be much better than cotðÞ α measures the influence of gravity. We note that the that of various slip models. For this reason, in this work, lubrication approximation requires the slope of the free we also use a precursor film model, which assumes that surface to be small. Hu et al. Future Cities and Environment (2015) 1:8 Page 7 of 15 Simulation results and discussion We chose θ =1° as a superhydrophilic surface (e.g. exist- Simulation of water droplet spreading on non-porous flat ing thin water layer, completely wetting); θ =21° as hydro- solid surface philic surface (e.g. n-Octyltriethoxysilane surface, partial 4 3 We consider a spherical droplet of volume V ¼ πr wetting); θ = 117° as hydrophobic surface (e.g. triethoxysi- lane surface). Based on experimental data of Bird et al. deposited on a horizontal wall that enters in contact (2008), for 0.1 ≤ t ≤8 ms, we derived the power-law expo- with a horizontal wettable surface at time t = 0. The ini- nent α is about 0.52 and the coefficient C is 1.46 for super- tial water droplet radius chosen for this simulation is R −1 hydrophilic surface; for hydrophilic surface, α is about 0.45, = 0.82 mm. Water surface tension is σ = 0.0727 Nm ; the coefficient C is equal to 1.21; for hydrophobic surface, α the density of water is equal to ρ =1000 kg/m .To is about 0.3, the coefficient C is equal to 0.69. By applying identify the effects of surface wettability, surfaces with Eqs. 15, 16 and 17, we obtained the simulation results as equilibrium contact angles from ~0° to ~117° were showninFig.4aand b studied. Fig. 4 a Log-log plot of spreading radius r as a function of time t of water drops on three wettable surfaces. Drop radius was R ≈ 0.82 mm. b Water droplet spreads on different non-porous solid surfaces Hu et al. Future Cities and Environment (2015) 1:8 Page 8 of 15 that spreading was dominated by viscous dissipation r t ¼ C ð15Þ (Tanner 1979; Cazabat and Cohen-Stuart 1986). R τ sffiffiffiffiffiffiffiffi Simulation of water droplet sliding on non-porous in- ρR clined flat solid surface τ ¼ ð16Þ Hydrophilic surfaces are superior in condensation rate from its early stage whereas dry surfaces directly face γt 3α humid air (Anna Lee et al. 2012). In this study, we use 1− r ¼ C R ð17Þ ρ pre-wetted superhydrophilic surfaces caused by conden- sation as initial boundary conditions. Figure 4a shows the log-log plot of time t vs spreading A single water droplet sliding on non-porous inclined radius r of water drops spreading on three surfaces with and smooth solid surface (e.g. glass surface) was simulated different wettability. When time is t < 0.1 ms, the wetting under three different surface conditions represented by dif- follows a power law ferent levels of condensation. They are dry glass surface 0.5 r=K t . For 0.1 ≤ t ≤8 ms, the wetting followed a (thin water film thickness = 0 μm) and two pre-wetted ' α power law r = K t . The slope, i.e. α is dependent on the nano-TiO superhydrophilic surfaces with condensed 50 surface wettability and increases from ~0.3 to ~0.5 while and 80 μm thickness thin water film, respectively. To sim- θ decreases from ~117° to ~0°, which was consistent plify the simulation, only the top part of the glass is chosen. eq with previous studies (Bird et al. 2008). Fig. 4a clearly The length of glass is chosen to be 5 cm. The origin of the reveals that the wetting condition has now significant X-axis (x = 0) is set at the top of the glass. The initial height effect on the spreading rate. As the equilibrium contact of the water droplet H is 1.5 mm as shown in Fig. 5a. Sim- angle increases, demonstrates a monotonic decrease in ulations of three cases start with the same initial water the power-law exponent. Drops spread faster on rela- droplet profile. Simulation results are shown in Fig. 5b. tively hydrophilic surfaces than on relatively hydropho- Figure 5b clearly shows that the velocity of the water drop- bic surfaces. At the same time, wetted area on relatively lets on superhydrophilic surface is faster than on uncoated hydrophilic surfaces is larger than that on relatively glass surface at the same inclined angle. The velocity of hydrophobic surfaces. water droplet is faster on a superhydrophilic surface with a On partially wettable surfaces (here θ =21°) and hydro- higher condensation rate. phobic surface (θ =117°), water drops reached equilib- rium after the inertial wetting stage, as shown in Fig. 4a Simulation of water runoff on non-porous inclined flat and b. In contrast, on completely wetting surfaces, a solid surface slower wetting process was observed for t ≥ 8 s. The power A water runoff on non-porous inclined smooth glass law fitting of the data gave a slope of ~0.1, which indicates was simulated under three different surface conditions Fig. 5 a Initial boundary condition of water droplet on vertical glass surface. b Simulation results of water droplet sliding on an inclined glass surface Hu et al. Future Cities and Environment (2015) 1:8 Page 9 of 15 represented by different levels of condensation aforemen- Particles begin to move on the adhered surface when tioned. The origin of the x-axis (x = 0) is set on the top of the combined lift and drag forces produced by the fluid the glass, with x = 5 mm and x = 6 mm. The initial water flow field applied to particles become large enough to 0 1 runoff thickness H is equal to 0.25 mm. Before x ,the counteract the attractive forces (e.g. the gravity, Van der 0 0 water front of the runoff is assumed to have the same Waals force) between the particle and the surface that thickness as H , which can be expressed as a Neumann hold the particle in place. In the study of particles, in lin- boundary condition. The length of the glass is assumed to ear shear flow, Zoeteweij et al. (2009) claimed that the be 25 cm long. Water runoff from the top of the glass is particles are in the boundary layer of the fluid flow; these as shown in Fig. 6a. Figure 6b shows the velocity of the particles will experience a lift force and a drag force water film on superhydrophilic surface was faster than exerted on the particle’s body. Both forces are a function that of an ordinary surface at the same inclined angle. The of fluid velocity at the position of the particle body. At a velocity of the water film is faster on a superhydrophilic certain flow velocity, the particles start to detach from a surface with a higher condensation rate. plate. Among the different possible particle motions (lift, sliding and rotation), particle rotation turns out to be the Effect of fluid flow velocity on the detachment of the responsible mechanism of particle removal. Rotation is adherent particles related to the moment of surface stresses which is also a Aerosol particles cover a size range from 1 nm to function of fluid velocity. Dagaonkar (2012) drew a quite 100 μm in diameter. Particles in the range 0.01 to 10 μm similar conclusion, namely that the lift force is strongly a are most stable in suspension. Particles smaller than function of the flow velocity of the fluid. Specifically, at 0.01 μm in diameter are not stable in the atmosphere; high fluid flow velocities, the contribution of the lift force they will either react with oxygen or tend to coagulate is expected to be higher resulting in the detachment of the into larger units, while those larger than 10 μm readily larger sized particles. settle out in air. Most particles with dimensions greater Grease, dirt particles and other staining materials are than 10 μm require strong air currents to keep them aloft easily attached to a surface of a solar panel. In the ‘photo- (Sharma 1994). Smaller particles < 1 μm float in the air for catalytic’ process of TiO coating, the coating reacts with days or weeks, during this time, they can be transported ultraviolet light to break down the organic dirt on the glass over 1000 km before they deposit. While larger particles and to reduce the adherence of inorganic dirt. In an earlier sediment quite soon and easily onto environmental sur- study, performed by Chabas et al. (2007), on the behaviors faces which can become contaminated because of their of self-cleaning glass in an urban atmosphere, self-cleaning weight (Dagsson-Waldhauserova et al. 2014). glass is found to have an evident self-cleaning effect, even Fig. 6 a Initial boundary condition of water runoff on vertical glass surface. b Simulation results of water runoff sliding on an inclined glass surface Hu et al. Future Cities and Environment (2015) 1:8 Page 10 of 15 when it is not subjected to water. The field study shows panels. Afterwards, they were thoroughly cleaned again that particulate organic matter (POM) was destroyed by a to remove all dust/contaminates on the surfaces. Two percentage of 44–48 % on the self-cleaning surface. solar panels on the right-hand side of Fig. 7 were coated Therefore, we deduce higher water film spreading vel- with nano-TiO coating and other two panels on the ocity on superhydrophilic surface ought to have larger left-hand side of Fig. 7 were uncoated. Then they were detachment force on particles and a better effect on exposed in air for one day, to perform a post-coating test to washing away loose particles. detect if the nano-TiO coating affects solar cell perform- ance. Photocatalysis is more efficient, because TiO parti- Experimental setup cles are finely divided and highly dispersed (as shown in The flow of the liquid film is essential for self-cleaning. Fig. 8a) in order to give the highest surface contact with the To clean a surface, liquid has to transport along the im- surrounding environment. Due to strong wind on the roof purities and finally run off the surface. It is the flowing and relatively good air quality in the test region, hardly any liquid film that carries away the impurities. The purpose contamination was detected on some other test solar of this experiment is to validate our model and to test panels, which are undergoing out-door exposure for one the effect of self-cleaning coating. year. To simulate the effects of heavily polluted rainwater, Four identical solar monocrystalline panels as shown applied to solar cells in a real situation, a “muddy water” in Fig. 7 (Type STP005B-12/DEA) are used for applica- mixture consistingof1.5 gwoodash,representinganash tion of self-cleaning testing. They are all placed vertically pollutant, due to the ash pollutant caused the highest on the test rig and are exposed to sunshine on the roof; reduction in PV voltage (25 %). This was recorded by all of them are facing south as shown in Fig. 7. To make Kazem et al. (2013). In this work, 500 ml water was sprayed comparisons, two solar panels on the right-hand side are onto the surfaces of these four panels numerous times until coated with Nano-TiO (provided by ZIXILAI Co.LtD) they dried (this is shown in Fig. 8b and c). Afterwards, their coating, marked with number 1 and 2 respectively; while performance was measured in the following days as ex- on the left-hand side, two panels are uncoated, marked tremity tests. with number 3 and 4 . Every 20 min, a connected data acquisition system was Experimental results obtaining the following data series for 30 s: date and Reflection and light transmittance of Nano-TiO coating time; global irradiation intensity (W/m ); diffuse irradi- We recently conducted the spectrometric analysis of ation intensity (W/m ); voltage (V); current (A); surface Nano-TiO coating on PV panels. As we may notice in temperature (°C). After temperature compensation, volt- Fig. 9, in the wave length between 400 to 1200 nm, the age U and current I will be used for analyzing output reflection on coated surface seems to be 2 ~ 3 % less than power. that of uncoated surface. Reflections on solar panels need At first four identical solar panels were cleaned thor- to be avoided. Less reflective surface can be important in oughly by water and then dried in air. Furthermore, we increasing a module’s output power. In contrast to light exposed them to air for four days. Their performance reflection, in the same range of wave length, coated PVT was measured to detect differences in power generation panels shows 2 ~ 3 % more light transmittance than that as a pre-coating test for a comparison of these four of uncoated PVT panels. High light transmittance means more photons are absorbed by the solar cells, and more power is generated. We conclude that Nano-TiO2 coating itself has no side effect on the solar cell performance whatsoever. This was very significant since the solar cells could then be coated to improve its other properties with- out damaging the initial raw performance. Applying this Nano-TiO2 coating on the surface of a clean PV module will potentially increase the efficiency comparing with other coatings which most likely reduce the transmissivity of sunlight. The output energy of solar panels The output of a solar panel is usually stated in Watt, and the Watt (the amount of electric power) is deter- mined by multiplying the rated voltage V by the rated amperage I. Both voltage and current were adjusted with Fig. 7 Solar monocrystalline panels test rig for self-cleaning tests the temperature compensation. As our data acquisition Hu et al. Future Cities and Environment (2015) 1:8 Page 11 of 15 Fig. 8 a Nano-sized particles of TiO (photo provided by ZIXILAI Co.Ltd.). b Wood ashes. c Muddy water system obtains data continuously for thirty seconds every In an extremity test, solid particles from evaporated twenty minutes, we use the mean value of I and V within muddy water were located on the panel surface. They these 30s to obtain the power. By integrating the output degraded PV performance very much. However, coated power over the testing period, we obtain the curve of “In- panels generate ~8 % more energy than uncoated panels. tegrated energy output of testing solar panels via time”.As This is shown in Fig. 11c (net increase ~5.4 %). it can be seen from Fig. 10, a high peak on the figure Thanks to two major merits of TiO coating, when a means a higher energy gain (the weather was sunny, warm compound (either an organic soil or a pollutant) is present and dry), while a low peak is related to a low energy gain on the surface of nano-TiO coating, it can be degraded by (the weather was rainy/cloudy and with a relatively high redox reactions involving highly reactive transient species humidity). in thepresenceofUVlight.Then, thedegradation products are either stored in the coating or washed off the surface by rain water. In addition, Nano-TiO coating is also super- Out-door self-cleaning performance comparison hydrophillic and has high water spreading speed; the super Figure 11a shows in a pre-coating test, that panels to be hydrophilicity prevents the formation of water droplets. coated have a slightly better performance (~2.6 % more Water molecules will spread flat immediately to form a thin energy) than uncoated panels. In a post-coating test, as and uniform water film. This means once there is rain, the shown in Fig. 11b, coated panels generated ~6.2 % more rain water forms a uniform film on the coated solar panel energy than uncoated panels (net increase ~3.6 %). This is surface, which accelerates runoff water spreading and flow, due to coated surfaces that show less light reflection and that removes soiling and deters the formation of drying smaller transmittance than uncoated surfaces resulting in marks as well, and finally results in a better energy more generated power. generation. Fig. 9 The spectrometric analysis of Nano-TiO coating on PV panels 2 Hu et al. Future Cities and Environment (2015) 1:8 Page 12 of 15 Fig. 10 Integrated energy output of testing solar panels via time Conclusions on coated surface seems to be 2 ~ 3 % less than that of an The flow of a liquid film is essential for self-cleaning. To uncoated surface; a coated surface shows 2 ~ 3 % more clean a surface, liquid has to transport the impurities light transmittance than an uncoated surface. This implies and finally run off the surfaces. It is the flowing liquid Nano-TiO coating itself doesn’t show any side effect on film that carries away the impurities. the solar cell performance whatsoever. Applying this In this study, a 2D dynamic model water film sliding Nano-TiO coating on a surface of a clean PV/PVT mod- on the inclined non-porous solid surface was established ule will potentially increase the efficiency comparing with that is based on lubrication theory in association with a other coatings which most likely reduce the transmissivity precursor film of a wetting liquid on glass. of sunlight. It is well known that nano-TiO coated glass surfaces In Switzerland, four identical solar monocrystalline display superhydrophilic properties. At the macroscale, by panels are used for the application of self-cleaning test- implementing this property into our simulation model, our ing under natural weather conditions. Pre-coating, post- simulation demonstrates that on superhydrophilic surface, coating and extremity tests were conducted, respectively. in a very short time, water completely wets the substrate Our experimental results reveal that the coated panel and spreads out into a thin film; on inclined glass surfaces generated ~3.6 % more energy than that with uncoated (up to 90°), water droplets/film slide faster on the more panel in a post-coating test. In an extremity test, the hydrophilic surfaces resulting in larger detachment force on coated panel generated ~5.4 % more energy than the particles and better effect on washing away loose particles. one with an uncoated panel in short term test thanks to The spectrometric analysis of Nano-TiO coating on several major merits of the Nano-TiO coating. 2 2 glass surface was conducted by our partner in Germany; Our entire experiment resulted in a most important in the wave length between 400 to 1200 nm, the reflection conclusion: the coated PV panels distinctively displayed Fig. 11 a Pre-coating test result. b Post-coating test result. c Extremity test result Hu et al. Future Cities and Environment (2015) 1:8 Page 13 of 15 − 1 better “self-cleaning” properties when muddy water was U, Adimensionalion factor for the velocity (m. s ) − 1 applied. In our cases of extremity test, our PV panel im- Characteristic flow velocity (m. s ) proved performance on an average of over 5 % compared V, Voltage (V) to the non-coated ones. As nano-superhydrophilic coat- V, Volume (m ) ings are easily applied on glass substrates by common v = (u,v), Velocity vector − 1 coating techniques, like spraying, dipping, flooding, spin- v, Velocity component in the x direction (m. s ) ning etc., − considering that large solar power plants cost x, x*, X coordinate (m) and dimensionless x coordinate billions of dollars – this performance boost would save tangentially to the surface support millions of dollars. Many surfaces would be easier to clean x , Adimensionalion factor for x-axis (m) and maintain if performed hydrophilic. α, Angle of the inclined plane(rad) This research experiment proved that superhydrophilic α, Exponent of the power-law − 1 surfaces have useful self-cleaning mechanisms, particularly γ, Air-water surface tension (N. m ) − 1 − 1 on the glass surfaces of the solar panels. With the weath- μ, μ , Dynamic viscosity of liquid (kg. m .s ) − 3 ering that many solar panels undergo, this self-cleaning ρ, Density of the water (kg. m ) − 3 process may greatly improve the performance and main- ρ , Density of air (kg. m ) tenance of solar cells and be a key towards Green Energy θ , Dynamic contact angle (°) in solar cell development. Our research in self-cleaning θ , θ , Equilibrium contact angle (°) e eq surfaces can be applied to many other fields as well. τ, Inertial time scale (s) Competing interests The authors declare that they have no significant competing financial, Nomenclature professional or personal interests that might have influenced the a, Capillary length (m) performance or presentation of the work described in this manuscript. μU Ca ¼ , Capillary number of the runoff equation Authors’ contributions C, Prefactor for the power-law 2 − 1 JH and NB carried out the simulation; JH designed and conducted the D, Diffusion coefficient (m .s ) experiments. OS and SR supervised the research work. All authors participated in D(α), Dimensionless parameter for the runoff equation the sequence alignment and drafted the manuscript of paper “CFD simulation and validation of self-cleaning on solar panel surfaces with superhydrophilic δ, film thickness on (m) − 2 coating” for Future Cities and Environment Journal, doi:10.1186/s40984-015-0006-7. g, Gravitational acceleration (m. s ) All authors read and approved the final manuscript. h’, Adimensionalion factor for the film thickness (m) H , Initial droplet or runoff film thickness (m) Acknowledgements The study is supported by EU HERB project. Special thanks to Mr. Morey h, h*, Film thickness (m) and dimensionless film Philippe for outdoor experimental setup and data acquisition. The thickness spectrometric analysis of Nano-TiO coating by Mr. Dietrich Schneider at h , Latent heat of condensation at the bulk air Stuttgart University of Applied Sciences is also acknowledged. fg − 1 temperature T∞. (J. kg ) Author details I, Current (A) 1 Institute of Thermal Engineering (IGT), University of Applied Sciences of i, Unit vector of the x-axis Western Switzerland, Av. des Sports 20, CH-1400 Yverdon-les-Bains, Switzerland. Department of Architecture and Built Environment, University j, Unit vector of the y-axis of Nottingham, Nottingham, UK. K, K’, Coefficients of the power law K ¼ , Kapitza number Received: 14 July 2015 Accepted: 14 July 2015 μU − 1 − 1 k , Thermal conductivity of condensate (W. m .K ) l, radial extension length (m) References L, Adimensionalion factor for the surface support (m) Ausserre D, Picard AM, Leger L (1986) Existence and role of the precursor film in the spreading of polymer liquids. Phys Rev Lett 57:2671–2674 L , Length of plate (m) Bascom WD, Cottington RL, Singleterry CR (1964) In Contact Angle, Wettability, ∇ =(∂x, ∂y), Nabla operator and Adhesion. Fowkes FM (ed) ACS, Washington, DC, p 355–379. R , Drop radius (m) Beaglehole D (1989) Profiles of the precursor of spreading drops of siloxane oil on glass, fused silica, and mica. J Phys Chem 93(2):893–899 R , the Reynolds number Berman A, Drummond C, Israelachvili J (1998) Amontons’ law at the molecular r, Spreading radius (m) level. Tribol Lett 4:95–101 t, t*, Time (s) and dimensionless time Bertozzi AL, Brenner MP (1997) Linear stability and transient growth in driven contact lines. Phys Fluids 9:530 Biance A-L, Clanet C, Quere D (2004) First steps in the spreading of a liquid T , Vapor saturation temperature (297 K) droplet. Phys Rev E 69:016301 T , Wall temperature (K) Bird JC, Mandre S, Stone HA (2008) Short-time dynamic of partial wetting. Phys − 1 Rev Lett 100:234501 u, Average velocity (m. s ) Blossey R (2003) “Self-cleaning surfaces — virtual realities”. Nat Mater 306 − 1 Velocity component in the x direction (m. s ) Bonn D, Indekeu J, Meunier J, Rolley E (2009) Reviews of Modern Physics − 1 u', Advancing velocity (m. s ) 81(05):739–805 Hu et al. Future Cities and Environment (2015) 1:8 Page 14 of 15 Carlson A, Do-Quang M, Amberg G (2011) Dissipation in rapid dynamic wetting, frequency generation spectroscopy. Phys Chem Chem Phys 10:4987–4993, J Fluid Mech 682:213–240 PubMed:18688544 Cazabat A-M, Cohen-Stuart MA (1986) Dynamics of wetting: effects of surface Ohtsu N, Masahashi N, Mizukoshi Y, Wagatsuma K (2009) Hydrocarbon decomposition roughness. J Phys Chem 90:5845 on a hydrophilic TiO2 surface by UV irradiation: spectral and quantitative analysis Chabas A, Lombardo T, Cachier H, Pertuisot MH, Oikonomou K, Falcone R, Verità, using in situ XPS technique. Langmuir 25(19):11586–11591 M. Geotti-Bianchini, F. (2008) Behaviour of self-cleaning glass in urban Oron A, Davis SH, Bankoff SG (1997) Long-scale evolution of thin liquid films. Rev atmosphere. Build Environ 43(12):2124–2131 Mod Phys 69(3):931 Chen J-D, Wada N (1989) Wetting dynamics near the edge of a spreading drop. Perevertaev VD, Metsik MS, Golub LM (1979) Dependence of the heat of Phys Rev Lett 62:3050 adsorption of water-molecules on the activity of the surface of muscovite Chen L, Auernhammer GK, Bonaccurso E (2011) Short time wetting dynamics on crystals. Colloid J USSR 41(1):124–126, Phys., 12:1473 – 1484, 1979 soft surfaces. Soft Matter 7:9084–9089 Rafai S, Sarker D, Bergeron V, Meunier J, Bonn D (2002) Superspreading: aqueous surfactant drops spreading on hydrophobic surfaces. Langmuir Dagaonkar M (2012) Effect of fluid flow, solution chemistry and surface morphology 18(26):10486–10488 of fibrous material on colloid filtration. J Eng Fibers Fabrics 7(3):62 Dagsson-Waldhauserova P, Arnalds O, Olafsson H, Skrabalova L, Sigurdardottir Raviv U, Perkin S, Laurat P, Klein J (2004) Fluidity of water confined down to GM, Branis M, Hladil J, Skala R, Navratil T, Chadimova L, von Lowis of Menar subnanometer films. Langmuir 20:5322–5332 S, Thorsteinsson T, Carlsen HK, Jonsdottir I (2014) Physical properties of Razouk RI, Salem AS (1948) The adsorption of water vapor on glass surfaces. suspended dust during moist and low wind conditions in Iceland. Icel Agric J Phys Chem 52(7):1208–1227 Sci 27(25):25–39 Rhykerd CL, Cushman JH, Low PF (1991) Application of multiple-angle-of- de Gennes P-G (1985) Wetting: statics and dynamics. Rev Mod Phys 57:827 incidence ellipsometry to the study of thin fluid films adsorbed on Diez JA, Kondic L (2002) Computing threedimensional thin film flows including surfaces. Langmuir 7(10):2219–2229 contact lines. J Comput Phys 183:274–306 Ruths M, Israelachvili JN (2011) Chapter 13, surface forces and nanorheology of Dunnill CWH, Aiken ZA, Pratten J, Wilson M, Morgan DJ, Parkin IP (2009) Enhanced molecularly thin films. In: Bhushan B (ed) Nanotribology and photocatalytic activity under visible light in N-doped TiO2 thin films produced nanomechanics. Springer-Verlag, Berlin Heidelberg by APCVD preparations using antibacterial films. J Photochem Photobiol Sah A (2014) Precursor film: a key driver to determine wetting behavior in the A207(2–3):244–253 vicinity of surface heterogeneity. Soft Matter 10:3890–3896 Ehrhard P (1993) Experiments on isothermal and nonisothermal spreading. Salim A, Huraib F, Eugenio N (1988) PV power-study of system options and J Fluid Mech 257:463 optimization. In: Proceedings of the 8th European PV Solar Energy Goossens D, Kerschaever EV (1999) Aeolian dust deposition on photovoltaic solar Conference. Florence, Italy cells: the effects of wind velocity and airborne dust concentration on cell Sayigh A (2009) Worldwide progress in renewable energy. Renewable Energy. performance. Solar Energy 66:277–289 Sciverse Science Direct. 20 November 2009:1-5. Greenspan HP (1978) On the motion of a small viscous droplet that wets a Sayigh A, Al-Jandal S, Ahmed H (1985) Dust effect on solar flat surfaces devices surface. J Fluid Mech 84:125 in Kuwait. In: Proceedings of the workshop on the physics of Hardy WB (1919) The spreading of fluids on glass. Phil Mag 38:49–55 Non-conventional energy sources and materials science for energy. Triest, Hardy WB (1936) Collected works. University Press, Cambridge Italy, pp 353–367 Sharma BK (1994) Air pollution, ISBN: 81-8283-014-1. Heslot F, Cazabat AM, Levinson P (1989) Dynamics of wetting of tiny drops: ellipsometric study of the late stages of spreading. Phys Rev Lett Smith SH (1969) On initial value problems for the flow in a thin sheet of viscous 62:1286–1289 liquid. J Appl Math Phys (ZAMP), 20, 556–560 Heslot F, Cazabat AM, Levinson P (1990) Experiment on wetting on the Smith JD, Cappa CD, Wilson KR, Messer BM, Cohen RC, Saykally RJ (2004) scale of nanometers: influence of the surface energy. Phys Rev Lett Energetics of hydrogen bond network rearrangements in liquid water. 65:599–601 Science 306:851–853 [PubMed: 15514152] Hocking LM, Rivers AD (1982) The spreading of a drop by capillary action. J Fluid Stapelbroek BBJ, Jansen HP, Kooij ES, Snoeijer JH, Eddi A (2014) Universal Mech 121:425 spreading of water drops on complex surfaces. Soft Matter 10:2641 Incropera FP, DeWitt DP (2002) Fundamentals of heat and mass transfer, 5th edn. Staszczuk P (1985) Application of the chromatographic step profile method for Wiley, Hoboken, NJ determination of water film pressure and surface freeenergy of quartz. Janczuk B, Chibowski E, Staszczuk P (1983) Determination of surface free-energy Chromatographia 20(12):724–728 components of marble. J Colloid Interface Sci 96(1):1–6 Tanner LH (1979) The spreading of silicone oil drops on horizontal surfaces. J Phys D Appl Phys 12:1473–1484 Joanny JF, de Gennes PJ (1984) Static structure of wetting films and contact lines. C R Acad Sci II 299:279–283 Tian C, Ji N, Waychunas GA, Shen YR (2008) Interfacial structures of acidic and Katz GB (2008) Effect of Dust on Solar Panels. Available www.gregorybkatz.com/ basic aqueous solutions. J Am Chem Soc 130:13033–13039, PubMed: Home/effect-of-dust-on-solar-panels. 18774819 Kavehpour HP, Ovryn B, McKinley GH (2002) Evaporatively-driven marangoni Troian SM, Herbolzheimer E, Safran SA, Joanny JF (1989) Fingering instabilities of instabilities of volatile liquid films spreading on thermally conductive driven spreading films". Europhys Lett 10:25 substrates. Coll Surf A 206:409–423 Ueno I, Watanabe T (2005) Dynamic behavior of precursor liquid film in Kazem HA, Khatib T, Sopian K, Buttinger F, Elmenreich W, Albusaidi AS (2013) front of moving droplet on solid, 16th International Symposium on Effect of dust deposition on the performance of multi-crystalline Transport Phenomena, ISTP-16, Prague. photovoltaic modules based on experimental measurements. Int J Renew Valignat MP, Oshanin G, Villette S, Cazabat A-M, Moreau M (1998) Molecular Energy Res 3:4 weight dependence of spreading rates of ultrathin polymeric films. Phys Rev Lett 80:5377–5380 Kazmerski L (2011) Photovoltaic: history, technology, markets, manufacturing, Villette S, Valignat MP, Cazabat AM, Jullien L, Tiberg F (1996) Wetting on the applications, and outlook, 84th International Seminar in Brighton, Renewable Energy Policy, Security, Electricity, Sustainable Transport, Water molecular scale and the role of water. A case study of wetting of hydrophylic silica surfaces. Langmuir 12:825–830 Resources/Management and Environment, Brighton, UK, 3–9. Lee A, Moon M-W, Lim H, Kim W-D, Kim H-Y. (2012) Water harvest via dewing. Voinov OV (1976) Hydrodynamics of wetting. Fluid Dynamics 11:714–721 Langmuir 28(27):10183–10191. Vou’e M, Valignat MP, Oshanin G, Cazabat A-M (1999) Dissipation processes at Levinson P, Cazabat A-M, Cohen-Stuart MA, Heslot F, Nicolet S (1988) The the mesoscopic and molecular scale. The case of polymer films. Langmuir spreading of macroscopic droplets. Rev Phys Appl 23:1009 15:1522–1527 Lopez J, Miller CA, Ruckenstein E (1976) Spreading kinetics of liquid drops on Wack KO (1980) “Cleaning system for solar energy collectors”, Physics Abstract solids. J Colloid Interface Sci 53:460–461 Science Marmur A, Lelah MD (1980) The dependence of drop spreading on the size of Wakim F (1981) Introduction of PV power generation to Kuwait. Kuwait Institute the solid surface. J Colloid Interf Sci 78(262):265 for Scientific Researchers, Kuwait City Noguchi H, Hiroshi M, Tominaga T, Gong JP, Osada Y, Uosaki K (2008) Interfacial Wang XD (2003) Contact angle hysteresis and dynamic wetting behavior on water structure at polymer gel/quartz interfaces investigated by sum rough solid surfaces, Ph.D. Thesis. Tsinghua University, China Hu et al. Future Cities and Environment (2015) 1:8 Page 15 of 15 Winkels KG, Weijs JH, Eddi A, Snoeijer JH (2012) Initial spreading of low-viscosity drops on partially wetting surfaces. Phys Rev E 85(5):055301 Yeo L (2008) Wetting and Spreading, Encyclopedia of Microfluidics and Nanofluidics, pp 2186–2196 Yoo H, Paranji R, Pollack GH (2011) Impact of hydrophilic surfaces on interfacial water dynamics probed with NMR spectroscopy. J Phys Chem Lett 2(6):532–536 Yuan Q, Zhao Y-P (2010) Precursor film in DynamicWetting, electrowetting, and electro-elasto-capillarity. Phys Rev Lett, Prl 104:246101 Zisman WA (1965) Improving the performance of reinforced plastics. Ind Eng Chem 57:26–34 Zoeteweij ML, van der Donck JCJ, Versluis R (2009) Particle removal in linear shear flow: model prediction and experimental validation. J Adh Sci Technol 23(6):899–911 Submit your manuscript to a journal and benefi t from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the fi eld 7 Retaining the copyright to your article Submit your next manuscript at 7 springeropen.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Future Cities and Environment Springer Journals

CFD simulation and validation of self-cleaning on solar panel surfaces with superhydrophilic coating

Download PDF
15 pages

Loading next page...
 
/lp/springer-journals/cfd-simulation-and-validation-of-self-cleaning-on-solar-panel-surfaces-B3cvOk4rhI

References (79)

Publisher
Springer Journals
Copyright
Copyright © 2015 by The Author(s)
Subject
Energy; Energy Efficiency (incl. Buildings); Renewable and Green Energy; Energy Technology; Landscape/Regional and Urban Planning
eISSN
2363-9075
DOI
10.1186/s40984-015-0006-7
Publisher site
See Article on Publisher Site

Abstract

Solar panel conversion efficiency, typically in the twenty percent range, is reduced by dust, grime, pollen, and other particulates that accumulate on the solar panel. Cleaning dirty panels to maintain peak efficiency, which is especially hard to do for large solar-panel arrays. To develop a transparent, anti-soiling Nano-TiO coating to minimize the need for occasional cleaning is the purpose of this study. In our study, a 2D rainwater runoff model along tilted solar panel surface based on the Nusselt solution was established to have better understanding and predicting the behavior of runoff rain water, especially in contact with solar-panel surfaces with Nano-TiO coating. Our simulation results demonstrate that solar-panel surfaces with Nano-TiO coating create a superhydrophilic surface which cannot hold water, showing features of more pronounced in increasing runoff water film velocity comparing to the uncoated surfaces during raining event resulting in better effect of self-cleaning. Validation of our model was performed on titled solar panels for real time outdoor exposure testing in Switzerland. It is found that the dust particles are not easy to adhere to the coated surfaces of the slides comparing with uncoated surfaces, showing great potential for its use in harsh environmental conditions. This study suggests that superhydrophilic self-cleaning solar panel coating maximize energy collection and increases the solar panel’s energy efficiency. Keywords: Nano-TiO coating; Self-cleaning; 2D runoff model; Solar-panel surfaces Introduction dust accumulation on the tilted glass plates revealed a Solar PV technology is well-proven for producing electri- reduction in plate-transmittance ranging from 64 to city, where the global production has been increasing 370 17 %, for tilt angles ranging from 0° to 60° respectively times than that in 1992 (Kazmerski 2011). The output of a after 38 days of exposure. A reduction of 30 % in useful PV module is usually rated by manufacturers under energy gain was observed by the horizontal collector Standard Test Conditions (STC), where each module is after three days of dust accumulation. Salim et al. tested under a temperature of 25 °C; solar radiation of (1988) indicated that a 32 % reduction in performance 1000 W/m , air mass of 1.5 spectra and wind speed of after eight months occurred under desert conditions in 2 m/s. However, these conditions are different from the KSA. Goossens and Kerschaever (1999) showed that conditions in the practical fields. With the increasing use the deposition of fine dust particles on the cover of PV of PV systems, it is vital to study meteorological parame- modules significantly affects the performance of these ters that affect the performance of these systems such as modules. Katz (2008) reported that the dirt on PV humidity, dust, temperature and wind speed. modules caused a 2 % of power reduction as compared The effect of dust on PV modules performance has to clean PV modules. However, Sayigh (2009) reported been investigated in different ways as can be found in a power decrease of about 11.5 % in a PV module ex- the literature. Wakim (1981) claimed that 17 % of PV posed for only 72 h in Riyadh, Saudi Arabia. power is lost due to dust deposition on PV modules in Kazem et al. (2013) recently conducted experiments Kuwait city. Sayigh et al. (1985) reported the effect of concerning the effects of air pollutants including red soil, ash, sand, calcium carbonate, and silica on the solar * Correspondence: jin.hu@heig-vd.ch power generated. Their results show that the reduction Institute of Thermal Engineering (IGT), University of Applied Sciences of in PV voltage and power strongly depends on pollutant Western Switzerland, Av. des Sports 20, CH-1400 Yverdon-les-Bains, Switzerland type and deposition level. The highest reduction in PV Full list of author information is available at the end of the article © 2015 Hu et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Hu et al. Future Cities and Environment (2015) 1:8 Page 2 of 15 voltage (25 %) is recorded when the ash pollutant is conducted on PV panels under natural environmental used. conditions in Switzerland. A study conducted by Harvard University students showed that energy losses of solar panels due to soiling (of Superhydrophilic and hydrophilic surfaces the surface) vary between 9 and 20 % of the possible Rhykerd et al. (1991) measured ellipsometrically the thick- energy absorption (Wack 1980). This is a significant de- ness of the adsorbed water film on a fused silica surface crease of energy and brings the subject of solar panel and found it ranging from 2.4 to 9.0 nm, depending on the cleaning to attention. However, most solar panels are water vapor pressure. Staszczuk (1985) used gas chroma- placed in regions with difficult accessibility such as a roof. tography to determine the water adsorption isotherm on This combined with high voltage proximity makes clean- quartz at 20 °C and found that about sixteen statistical ing solar panels expensive and difficult. water layers adsorbed from a gas phase saturated with There are numerous ways to clean a surface: from abra- water vapor. Also, similar experiments Janczuk et al. (1983) sive techniques like sandblasting to laser cleaning and using the chromatographic technique showed that about water spraying. Some of the factors of conventional fifteen statistical water layers may adsorb onto a marble methods shall be mentioned: Possible damage caused by surface. Water films with thicknesses from 1.0 to 8.0 nm aggressive cleaning methods that may roughen the surface were also reported for muscovite mica (Perevertaev et al. and the surface is more susceptible to new smudging in 1979). the future; as cleaning is an intensive work, therefore costs Anna Lee et al. (2012) studied the condensation be- a lot for labor work; massive use of chemicals in cleaning haviors of the surfaces with different wettability and will cause environmental problems. roughness. They concluded that the hydrophilic sur- Blossey (2003) claimed two routes to self-cleaning are faces are superior in condensation rate from its early emerging, which work by the removal of dirt by either film stage when the dry surfaces directly face humid air. In or droplet flow. In other words, water film flows, either on their experiments, where the film thickness δ on hydro- hydrophilic surface or water drop flow on hydrophobic philic surface is estimated by the Nusselt theory that surface, are methods to achieve self-cleaning. balances the viscous shear force and the gravitational Nano-TiO transparent coating can make a substrate force (Incropera and DeWitt 2002) surface to be photocatalytic and hydrophilic. UV-radiation "#1 from daylight reacts with dirt and organic deposits, oxi- 4k μðÞ T −T L l s w p dizes them and breaks their adherence to the surface. δ≈ ð1Þ gρρ−ρ h fg There is sufficient evidence to support the removal of organic contaminants (Ohtsu et al. 2009) and bacteria (Dunnill et al. 2009) adsorbed on a TiO surface by the They obtained δ is ~85 μm. photooxidization process. Due to TiO films that exhibit It is commonly accepted that in the vicinity of hydro- hydrophilicity (Ohtsu et al. 2009), surface raindrops philic interfaces water organizes into ice-like water, which spread as a film on the surface, ensuring the loosened dirt project from the surface by a few nanometers (Noguchi particles are carried away from the surface during rainy et al. 2008; Tian et al. 2008; Smith et al. 2004). Yoo et al. weather. (2011) used NMR spectroscopy to study impact of hydro- The phenomenon of soiled and stained facades is serious philic surfaces on interfacial water dynamics. Their find- in the cities. Industrial pollution has increased these prob- ings confirm the existence of highly restricted water layers lems. Especially in densely populated urban areas, building adsorbed onto hydrophilic surfaces and dynamically stable facades are covered by inorganic pollutants like nitrogen water beyond the first hydration layers. Thus, aqueous oxides and organic pollutants like benzene. All these regions on the order of micrometers are dynamically pollutants are destructive for buildings and unhealthy for different from bulk water. people. Nano-TiO coating can be easily applied on fa- cades to reduce the concentration of nitric oxides and Precursor film other toxic substances like benzene which can provoke When a perfectly wetting liquid spreads relatively slowly on respiratory problems and increase smog formation. a clean smooth substrate, a very thin precursor film may Hereby, Nano-TiO coating is very good candidate to propagate ahead of the apparent or macroscopic wetting reduce air pollution in cities. line with the average thickness of the film determined by The purpose of this study is to establish a 2D model of material property of liquid phase and solid wall (Heslot water slide on inclined self-cleaning surface (nano-TiO et al. 1989 and 1990). The first reported observation of an coated superhydrophilic solid surface) for PVT/solar ‘invisible’ film spreading ahead of the edge of a macro- panels, specifically targeting at solar panel surfaces. The scopic drop stems from the pioneering work by (Hardy self-cleaning effect of the super-hydrophilic coating is 1919 and 1936), Hardy was unable to detect its presence Hu et al. Future Cities and Environment (2015) 1:8 Page 3 of 15 directly at that time. Numerous studies have subsequently to completely spread a macroscopic drop (Voinov 1976; confirmed the existence of precursor films using ellipso- Tanner 1979). The precursor which advances at a seem- metry Beaglehole (1989); Bascom et al. (1964), interference ingly faster rate than the normal contact line is indeed microscopy patterns Bascom et al. (1964), and polarized well documented. It has been found that the velocity at reflection microscopy (Ausserre et al.1986).Simultaneous which the precursor film advances depends on the mate- observations of the moving droplet and of the fringe pat- rials of the gas/liquid/solid system and can vary in wide tern by the laser ellipsometry (Ueno and Watanabe 2005) range. Bascom et al. 1964 reported the speed of propaga- reveal the existence of the precursor film, which is ahead tion of the precursor film of squalane on stainless steel is −4 the moving contact line and traveled with varying its about 10 cm/s while indirect measurement for water on profile. glass Marmur & Lelah (1980) suggest that there the pre- If the fluid is perfectly wetting, we have seen that van cursor film velocity is of the order of 10 cm/s. As a result, der Waals forces lead to the formation of a precursor film the existence of precursor film makes the ACL moving on in front of the contact line (de Gennes 1985). Wang the film rather than on a real solid surface. In other words, (2003) in his thesis claimed the surface feature, particu- the ACL moves over the “wet” solid surface instead of the larly micro-structure or roughness, promotes the forma- “dry” solid surface. tion and development of thin precursor film, they claimed that the precursor film spreads much faster than the Precursor film on dry surface movement of the apparent contact line (ACL) on a rough As stated by de Gennes (1985) and Joanny and de Gennes solid surface. Amit Sah (2014) studied precursor film with (1984), long-range forces, for example, van der Waals glass capillaries and cover slips. His experimental results forces, should be taken into account in the spreading imply that the precursor film moving ahead of the contact phenomenon of precursor film. It is found that, at suffi- line controls the wetting behavior. Yuan and Zhao (2010) ciently short time, when the macroscopic droplet still acts using molecular dynamics (MD) simulations to explore as a reservoir, the behavior of the precursor film is diffu- the atomic details and the transport properties of the pre- sive and the radial extension length (l) on a macroscopic cursor film in dynamic wetting. Their results showed that scale follows a universal time dependence (t) of the form: the molecules which finally formed the precursor film came almost from the surface in the initial state of the l ¼ Dt ð2Þ droplet. The fast propagation of the precursor film is owing to continuous and very fast diffusing of the surface where D is the diffusion coefficient of the liquid within water molecules that have the highest self-diffusion coeffi- the precursor film. It is different from the conventional cient to the front of the precursor film allows for fast diffusion coefficient describing the random motion of propagation of the precursor film. Precursor film propa- particles in the bulk liquid phase or on solid substrates. gates fast with low energy dissipation. As a matter of fact, D also depends on the driving forces The typical thickness of the precursor film moving in which cause the film spreading. front of macroscopic body of fluid is in the range of As shown in Fig. 1, the macroscopic wedge of liquid is 500–3000 Å (Bascom et al. 1964; Beaglehole 1989). For a advancing at a constant velocity U on a solid dry surface fixed volume of water droplet, according to Tanner’s laws into an external air. The dynamic contact angle is θ (Tanner 1979), the spreading of the macroscopic part of and the wedge is preceded by a precursor film of length the droplet is rather slow; it would take a very long time l. The dynamic contact angle θ , which depends on the Fig. 1 Precursor film on dry surface Hu et al. Future Cities and Environment (2015) 1:8 Page 4 of 15 advancing velocity U, the liquid/vapor interfacial tension and TK on oxidized silicon wafers. Their analysis revealed γ and the liquid viscosity η through Tanner’s law (Tanner the dependence of D on relative humidity (RH). Overall, 1979). within the range of 20–90 % RH, D varies by more than two orders of magnitude, attaining the values observed for Precursor film on superhydrophilic/hydrophilic surface mesoscopic precursor films. This agrees well with the Hydrophilic surfaces adsorb water from the environment observations made by (Valignat et al. 1998 and Vou’e and the amount of water depositing on the hydrophilic et al. 1999). Such a remarkable enhancement of D has surface depends on the relative humidity. It is generally been explained by the fact that at this value of RH the accepted that under ordinary atmospheric conditions, patches of water on the substrate start to overlap and form hydrophilic surfaces adsorb at least a monolayer of water. a tortuous connected structure on which the spreading of For example, a clean glass surface is covered with a molecules encounters a very low friction. This implies on monolayer of adsorbed water at relative humidity of superhydrophilic/hydrophilic surface, the length of precur- around 30–50 % at 20 °C (Razouk and Salem 1948). sor film is much longer than that of the ordinary surfaces. Formation of a water film composed of as many as twenty When viewed at the atomic scale, the precursor film plays molecular layers, or more, may occur at the clean surface an important role in the dynamic wetting process on of high-energy solids, especially at high relative humidities hydrophilic substrate. >90–95 % (Zisman 1965). As showninFig.2,beforethe spreading of liquid on Water appears to have unusual lubricating properties superhydrophilic/hydrophilic surface, the solid surface is and usually gives wearless friction with no stick slip (Raviv already covered with a thin film of water with thickness h et al. 2004). It is also interesting that a 0.25 nm thick due to absorption/condensation on hydrophilic surface. water film between two mica surfaces is sufficient to bring The advancing velocity of the macroscopic wedge of liquid the coefficient of friction down to 0.01−0.02, a value that is not the average velocity u,but rather u′ (Joanny and de corresponds to the unusually low friction of ice. The ef- Gennes 1984). Obviously, if h ≠ 0, u′ > u. fectiveness of water film only 0.25 nm thick to lower the friction force by more than an order of magnitude is at- u ¼ u ð3Þ tributed to the “hydrophilicity” of the mica surface and to ξ−h the existence of a strongly repulsive short-range hydration force between such surfaces in aqueous solutions, which We conclude when the superhydrophilic/hydrophilic effectively removes the adhesion-controlled contribution was previously wetted by a very thin layer of the inner to the friction force (Berman et al. 1998). Clearly, a single water, the advancing velocity of the macroscopic wedge monolayer of water can be a very good lubricant - much of liquid is faster than that of dry solid surface. better than most other monomolecular liquid films (Ruths & Israelachvili 2011). It also stresses the inertial nature of Water spreading dynamics the spreading: a viscocapillary motion would have been When a liquid drop contacts a wettable surface, the liquid much quicker on a pre-wetted substrate, because of the spreads over the solid to minimize the total surface energy. lubricating effect of this water layer (Biance et al. 2004). For low viscous liquids (as water), a power law of the drop Villette et al. (1996) focused specifically on the role of spreading can be observed during almost all the evolution water on the spreading of molecular films of non-volatile varying the relative different contributions (inertia, gravity, liquids PDMS, PDMS with hydroxyl ends (PDMS-OH) viscosity, density, volume, surface tension…etc.). Fig. 2 Precursor film on superhydrophilic/hydrophilic surface Hu et al. Future Cities and Environment (2015) 1:8 Page 5 of 15 Dynamic wetting may proceed with three stages. to n =1/7 (Kavehpour et al. 2002; Oron et al. 1997; Ehrhard About 1 to 100 μs after the drop contacts the surface, 1993). inertia of the moving drop resists the capillary force that This extremely slow dynamics emerges from a balance drives a high speed spreading (in the order of 1 m/s). between surface tension and viscous forces close to the The spreading dynamics of low-viscosity drops (e.g. contact line (Bonn et al. 2009). 0.5 water) follow a power law r = Kt , which is independent of the liquid viscosity and surface wettability (Biance Runoff equations for thin film flow et al. 2004; Bird et al. 2008; Carlson et al. 2011; Chen The model of gravity-driven flow was based on the as- et al. 2011; Winkels et al. 2012; Stapelbroek et al. 2014). sumption that the inclined solid surface is initially covered The quantity r is the spreading radius, t is the spreading with a liquid film. We focus on the gravity-driven film time and K is a coefficient. flow. The solid surface is often initially dry and is gradually With time between ~0.1 to 10 ms, the drop spreading covered by the liquid film as the leading edge of the film speed is still high (~0.1 m/s), wetting is still dominated moves down under the action of gravity. by inertia. However, the surface wettability starts to in- Consider a film of viscous liquid of density ρ and viscosity fluence spreading and the spreading radius grows with μ, flowing down a plane inclined at an angle α, as illustrated time according to another power law r = K′ t (Bird in Fig. 3, and the flow is assumed to be two-dimensional, et al. 2008). K′ is another coefficient and the exponent α with no variations in the direction normal to the plane of is only dependent on the equilibrium contact angle θ . the sketch. Far away from the leading edge the film is ap- eq Experiments of Stapelbroek et al. (2014) reveal a deviation proximately flat and, therefore, the flow is described by the from a pure power-law, the cross-over from the 1/2 power constant-thickness solution, with the average flow velocity law to the final equilibrium radius displays a universal dy- ρgd sin(α/(3 μ)). It is convenient to define the characteristic namics. This cross-over is governed only by the final con- flow velocity by: tact angle, regardless of the details of the substrate. ρgd Bird et al. (2008) considered the energy dissipation U ¼ sinα ð6Þ during spreading and derived a power law which associ- ates α with θ . eq Here, the surface tension σ is assumed constant. At qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi the leading edge of the film, the liquid–gas interface is α ¼ C F ϑ þ cosθ ð4Þ eq eq assumed to be meeting the solid surface at a contact angle θ, as shown in Fig. 3 and the capillary number Experimental study showed that α increases from based on this velocity as: ~0.25 for θ ≈ 120° to ~0.5 for θ ≈ 0° (Bird et al. 2008). eq eq μU After the two inertial stages aforementioned, on Ca ¼ ð7Þ strongly hydrophilic (θ < < ~57.3° or 1 radian) or com- eq pletely wetting surfaces (θ ≈ 0°), a third wetting stage eq was found. In this stage, the wetting speed is much Lubrication approximation of thin liquid film slower than in the previous two stages, it takes a much We model the spreading of a thin liquid film on an inclined longer time for viscous liquids to completely spread and substrate. We use the lubrication approximation of the R < 1. In this long time limit stage, the viscous friction Navier–Stokes equations, and we provide appropriate initial inside the drop is the main source opposing capillarity. and boundary conditions. Within the framework of the The wetting dynamics follows 1/10 power law for small lubrication approximation, the velocity of the fluid is drops of negligible gravity effects (Tanner 1979; Cazabat depth-averaged over the thickness of the film (Greenspan and Cohen-Stuart 1986; Levinson et al. 1988; Chen and 1978). Following this approach, one obtains the average Wada 1989; Rafai et al. 2002): fluid velocity, 0:1 0:1 v ¼ðÞ u; v ð8Þ r ¼ R t ð5Þ μR v ¼ − ½ ∇p−ρgsinα i ð9Þ For largedrops dominating bygravityeffects, theradius 3μ of spreading pattern is to follow Lopez’slaw of a1/8 power law in the gravitational regime for large droplets (Smith; where ∇ =(∂x, ∂y), h is the fluid thickness, p is the pressure, Lopez et al. 1976; Yeo 2008). However, as the radius grows μ is the viscosity, ρ is the density, g is gravity, and α is the beyond the capillary length, the drop changes toward a inclination angle of the plane of the substrate. The coordin- “pancake” shape of constant thickness, curved only at the ate frame is chosen so that i points down the incline, and j rim, and the main driving force is now gravity, leading is the transverse direction in the plane. We note that Eq. 10 Hu et al. Future Cities and Environment (2015) 1:8 Page 6 of 15 Fig. 3 Thin film flows down a plane inclined at an angle α assumes no-slip boundary condition at the fluid/solid inter- the solid surface is pre-wetted with the wetting layer face. The pressure includes the hydrostatic component, and thickness h' (scaled by h) as a regularizing method in this the contribution following from the Laplace-Young bound- work. ary condition at the fluid-air interface. Thus, the lubrication approximation reduces the Navier– Stokes equations to this nonlinear fourth order partial p ¼ −γ∇ h þ ρghcosα ð10Þ differential equations that govern the time evolution of the film thickness h(x, y, t). As we focus on long-scale evolution of liquid films, we Nondimensionalization of thin film equations scale h by the height h' of the precursor film and we We consider a layer of liquid on a plane substrate inclined define the scaled in-plane coordinates and time by to the horizontal at an angle α. The fluid is Newtonian and x t incompressible of density ρ,viscosity μ and surface tension ðÞ x; y; t¼ ; ; where x y t c c γ. We start with the following 2-D lubrication equation (Diez and Kondic 2002): 2 ′ 3 a h ∂h x ¼ ð12Þ ¼ −∇⋅ðÞ hv sinα ∂t 3 2 3 ¼ − ∇ γh ∇∇ h−ρgh ∇hcosα þ ρghsinα i 3μ 3μ a x t ¼ ð13Þ ð11Þ 2 h sinα Here h is the height of the fluid given as a function of qffiffiffiffi x and y. The (x,y) plane is parallel to the substrate and And a ¼ is the capillary length, the velocity scale is ρg the x direction is the direction of the slope. Then, g is chosen naturally as U ¼ ; and the capillary number is de- the acceleration of gravity and i is the unit vector in the μU x direction There are several possibilities to obtain non- fined by ¼ : Using this dimensionless form, Eq. (12) for dimensional variables. h ¼ ; the lubrication equation given above becomes: All the theoretical and computational methods of the spreading drop problem require some regularizing mech- anism - either assumption of a precursor film in front of ∂h ∂ ∂ h ∂h þ h −DðÞ α þ 1 ¼ 0 ð14Þ the apparent contact line (Troian et al. 1989; Bertozzi and   3 ∂t ∂x ∂x ∂x Brenner 1997), or relaxing the no-slip boundary condition at fluid–solid interface (Greenspan 1978; Hocking and Rivers 1982). However, the computational performance of where the single dimensionless parameter DðÞ α ¼ðÞ 3Ca the precursor film model is shown to be much better than cotðÞ α measures the influence of gravity. We note that the that of various slip models. For this reason, in this work, lubrication approximation requires the slope of the free we also use a precursor film model, which assumes that surface to be small. Hu et al. Future Cities and Environment (2015) 1:8 Page 7 of 15 Simulation results and discussion We chose θ =1° as a superhydrophilic surface (e.g. exist- Simulation of water droplet spreading on non-porous flat ing thin water layer, completely wetting); θ =21° as hydro- solid surface philic surface (e.g. n-Octyltriethoxysilane surface, partial 4 3 We consider a spherical droplet of volume V ¼ πr wetting); θ = 117° as hydrophobic surface (e.g. triethoxysi- lane surface). Based on experimental data of Bird et al. deposited on a horizontal wall that enters in contact (2008), for 0.1 ≤ t ≤8 ms, we derived the power-law expo- with a horizontal wettable surface at time t = 0. The ini- nent α is about 0.52 and the coefficient C is 1.46 for super- tial water droplet radius chosen for this simulation is R −1 hydrophilic surface; for hydrophilic surface, α is about 0.45, = 0.82 mm. Water surface tension is σ = 0.0727 Nm ; the coefficient C is equal to 1.21; for hydrophobic surface, α the density of water is equal to ρ =1000 kg/m .To is about 0.3, the coefficient C is equal to 0.69. By applying identify the effects of surface wettability, surfaces with Eqs. 15, 16 and 17, we obtained the simulation results as equilibrium contact angles from ~0° to ~117° were showninFig.4aand b studied. Fig. 4 a Log-log plot of spreading radius r as a function of time t of water drops on three wettable surfaces. Drop radius was R ≈ 0.82 mm. b Water droplet spreads on different non-porous solid surfaces Hu et al. Future Cities and Environment (2015) 1:8 Page 8 of 15 that spreading was dominated by viscous dissipation r t ¼ C ð15Þ (Tanner 1979; Cazabat and Cohen-Stuart 1986). R τ sffiffiffiffiffiffiffiffi Simulation of water droplet sliding on non-porous in- ρR clined flat solid surface τ ¼ ð16Þ Hydrophilic surfaces are superior in condensation rate from its early stage whereas dry surfaces directly face γt 3α humid air (Anna Lee et al. 2012). In this study, we use 1− r ¼ C R ð17Þ ρ pre-wetted superhydrophilic surfaces caused by conden- sation as initial boundary conditions. Figure 4a shows the log-log plot of time t vs spreading A single water droplet sliding on non-porous inclined radius r of water drops spreading on three surfaces with and smooth solid surface (e.g. glass surface) was simulated different wettability. When time is t < 0.1 ms, the wetting under three different surface conditions represented by dif- follows a power law ferent levels of condensation. They are dry glass surface 0.5 r=K t . For 0.1 ≤ t ≤8 ms, the wetting followed a (thin water film thickness = 0 μm) and two pre-wetted ' α power law r = K t . The slope, i.e. α is dependent on the nano-TiO superhydrophilic surfaces with condensed 50 surface wettability and increases from ~0.3 to ~0.5 while and 80 μm thickness thin water film, respectively. To sim- θ decreases from ~117° to ~0°, which was consistent plify the simulation, only the top part of the glass is chosen. eq with previous studies (Bird et al. 2008). Fig. 4a clearly The length of glass is chosen to be 5 cm. The origin of the reveals that the wetting condition has now significant X-axis (x = 0) is set at the top of the glass. The initial height effect on the spreading rate. As the equilibrium contact of the water droplet H is 1.5 mm as shown in Fig. 5a. Sim- angle increases, demonstrates a monotonic decrease in ulations of three cases start with the same initial water the power-law exponent. Drops spread faster on rela- droplet profile. Simulation results are shown in Fig. 5b. tively hydrophilic surfaces than on relatively hydropho- Figure 5b clearly shows that the velocity of the water drop- bic surfaces. At the same time, wetted area on relatively lets on superhydrophilic surface is faster than on uncoated hydrophilic surfaces is larger than that on relatively glass surface at the same inclined angle. The velocity of hydrophobic surfaces. water droplet is faster on a superhydrophilic surface with a On partially wettable surfaces (here θ =21°) and hydro- higher condensation rate. phobic surface (θ =117°), water drops reached equilib- rium after the inertial wetting stage, as shown in Fig. 4a Simulation of water runoff on non-porous inclined flat and b. In contrast, on completely wetting surfaces, a solid surface slower wetting process was observed for t ≥ 8 s. The power A water runoff on non-porous inclined smooth glass law fitting of the data gave a slope of ~0.1, which indicates was simulated under three different surface conditions Fig. 5 a Initial boundary condition of water droplet on vertical glass surface. b Simulation results of water droplet sliding on an inclined glass surface Hu et al. Future Cities and Environment (2015) 1:8 Page 9 of 15 represented by different levels of condensation aforemen- Particles begin to move on the adhered surface when tioned. The origin of the x-axis (x = 0) is set on the top of the combined lift and drag forces produced by the fluid the glass, with x = 5 mm and x = 6 mm. The initial water flow field applied to particles become large enough to 0 1 runoff thickness H is equal to 0.25 mm. Before x ,the counteract the attractive forces (e.g. the gravity, Van der 0 0 water front of the runoff is assumed to have the same Waals force) between the particle and the surface that thickness as H , which can be expressed as a Neumann hold the particle in place. In the study of particles, in lin- boundary condition. The length of the glass is assumed to ear shear flow, Zoeteweij et al. (2009) claimed that the be 25 cm long. Water runoff from the top of the glass is particles are in the boundary layer of the fluid flow; these as shown in Fig. 6a. Figure 6b shows the velocity of the particles will experience a lift force and a drag force water film on superhydrophilic surface was faster than exerted on the particle’s body. Both forces are a function that of an ordinary surface at the same inclined angle. The of fluid velocity at the position of the particle body. At a velocity of the water film is faster on a superhydrophilic certain flow velocity, the particles start to detach from a surface with a higher condensation rate. plate. Among the different possible particle motions (lift, sliding and rotation), particle rotation turns out to be the Effect of fluid flow velocity on the detachment of the responsible mechanism of particle removal. Rotation is adherent particles related to the moment of surface stresses which is also a Aerosol particles cover a size range from 1 nm to function of fluid velocity. Dagaonkar (2012) drew a quite 100 μm in diameter. Particles in the range 0.01 to 10 μm similar conclusion, namely that the lift force is strongly a are most stable in suspension. Particles smaller than function of the flow velocity of the fluid. Specifically, at 0.01 μm in diameter are not stable in the atmosphere; high fluid flow velocities, the contribution of the lift force they will either react with oxygen or tend to coagulate is expected to be higher resulting in the detachment of the into larger units, while those larger than 10 μm readily larger sized particles. settle out in air. Most particles with dimensions greater Grease, dirt particles and other staining materials are than 10 μm require strong air currents to keep them aloft easily attached to a surface of a solar panel. In the ‘photo- (Sharma 1994). Smaller particles < 1 μm float in the air for catalytic’ process of TiO coating, the coating reacts with days or weeks, during this time, they can be transported ultraviolet light to break down the organic dirt on the glass over 1000 km before they deposit. While larger particles and to reduce the adherence of inorganic dirt. In an earlier sediment quite soon and easily onto environmental sur- study, performed by Chabas et al. (2007), on the behaviors faces which can become contaminated because of their of self-cleaning glass in an urban atmosphere, self-cleaning weight (Dagsson-Waldhauserova et al. 2014). glass is found to have an evident self-cleaning effect, even Fig. 6 a Initial boundary condition of water runoff on vertical glass surface. b Simulation results of water runoff sliding on an inclined glass surface Hu et al. Future Cities and Environment (2015) 1:8 Page 10 of 15 when it is not subjected to water. The field study shows panels. Afterwards, they were thoroughly cleaned again that particulate organic matter (POM) was destroyed by a to remove all dust/contaminates on the surfaces. Two percentage of 44–48 % on the self-cleaning surface. solar panels on the right-hand side of Fig. 7 were coated Therefore, we deduce higher water film spreading vel- with nano-TiO coating and other two panels on the ocity on superhydrophilic surface ought to have larger left-hand side of Fig. 7 were uncoated. Then they were detachment force on particles and a better effect on exposed in air for one day, to perform a post-coating test to washing away loose particles. detect if the nano-TiO coating affects solar cell perform- ance. Photocatalysis is more efficient, because TiO parti- Experimental setup cles are finely divided and highly dispersed (as shown in The flow of the liquid film is essential for self-cleaning. Fig. 8a) in order to give the highest surface contact with the To clean a surface, liquid has to transport along the im- surrounding environment. Due to strong wind on the roof purities and finally run off the surface. It is the flowing and relatively good air quality in the test region, hardly any liquid film that carries away the impurities. The purpose contamination was detected on some other test solar of this experiment is to validate our model and to test panels, which are undergoing out-door exposure for one the effect of self-cleaning coating. year. To simulate the effects of heavily polluted rainwater, Four identical solar monocrystalline panels as shown applied to solar cells in a real situation, a “muddy water” in Fig. 7 (Type STP005B-12/DEA) are used for applica- mixture consistingof1.5 gwoodash,representinganash tion of self-cleaning testing. They are all placed vertically pollutant, due to the ash pollutant caused the highest on the test rig and are exposed to sunshine on the roof; reduction in PV voltage (25 %). This was recorded by all of them are facing south as shown in Fig. 7. To make Kazem et al. (2013). In this work, 500 ml water was sprayed comparisons, two solar panels on the right-hand side are onto the surfaces of these four panels numerous times until coated with Nano-TiO (provided by ZIXILAI Co.LtD) they dried (this is shown in Fig. 8b and c). Afterwards, their coating, marked with number 1 and 2 respectively; while performance was measured in the following days as ex- on the left-hand side, two panels are uncoated, marked tremity tests. with number 3 and 4 . Every 20 min, a connected data acquisition system was Experimental results obtaining the following data series for 30 s: date and Reflection and light transmittance of Nano-TiO coating time; global irradiation intensity (W/m ); diffuse irradi- We recently conducted the spectrometric analysis of ation intensity (W/m ); voltage (V); current (A); surface Nano-TiO coating on PV panels. As we may notice in temperature (°C). After temperature compensation, volt- Fig. 9, in the wave length between 400 to 1200 nm, the age U and current I will be used for analyzing output reflection on coated surface seems to be 2 ~ 3 % less than power. that of uncoated surface. Reflections on solar panels need At first four identical solar panels were cleaned thor- to be avoided. Less reflective surface can be important in oughly by water and then dried in air. Furthermore, we increasing a module’s output power. In contrast to light exposed them to air for four days. Their performance reflection, in the same range of wave length, coated PVT was measured to detect differences in power generation panels shows 2 ~ 3 % more light transmittance than that as a pre-coating test for a comparison of these four of uncoated PVT panels. High light transmittance means more photons are absorbed by the solar cells, and more power is generated. We conclude that Nano-TiO2 coating itself has no side effect on the solar cell performance whatsoever. This was very significant since the solar cells could then be coated to improve its other properties with- out damaging the initial raw performance. Applying this Nano-TiO2 coating on the surface of a clean PV module will potentially increase the efficiency comparing with other coatings which most likely reduce the transmissivity of sunlight. The output energy of solar panels The output of a solar panel is usually stated in Watt, and the Watt (the amount of electric power) is deter- mined by multiplying the rated voltage V by the rated amperage I. Both voltage and current were adjusted with Fig. 7 Solar monocrystalline panels test rig for self-cleaning tests the temperature compensation. As our data acquisition Hu et al. Future Cities and Environment (2015) 1:8 Page 11 of 15 Fig. 8 a Nano-sized particles of TiO (photo provided by ZIXILAI Co.Ltd.). b Wood ashes. c Muddy water system obtains data continuously for thirty seconds every In an extremity test, solid particles from evaporated twenty minutes, we use the mean value of I and V within muddy water were located on the panel surface. They these 30s to obtain the power. By integrating the output degraded PV performance very much. However, coated power over the testing period, we obtain the curve of “In- panels generate ~8 % more energy than uncoated panels. tegrated energy output of testing solar panels via time”.As This is shown in Fig. 11c (net increase ~5.4 %). it can be seen from Fig. 10, a high peak on the figure Thanks to two major merits of TiO coating, when a means a higher energy gain (the weather was sunny, warm compound (either an organic soil or a pollutant) is present and dry), while a low peak is related to a low energy gain on the surface of nano-TiO coating, it can be degraded by (the weather was rainy/cloudy and with a relatively high redox reactions involving highly reactive transient species humidity). in thepresenceofUVlight.Then, thedegradation products are either stored in the coating or washed off the surface by rain water. In addition, Nano-TiO coating is also super- Out-door self-cleaning performance comparison hydrophillic and has high water spreading speed; the super Figure 11a shows in a pre-coating test, that panels to be hydrophilicity prevents the formation of water droplets. coated have a slightly better performance (~2.6 % more Water molecules will spread flat immediately to form a thin energy) than uncoated panels. In a post-coating test, as and uniform water film. This means once there is rain, the shown in Fig. 11b, coated panels generated ~6.2 % more rain water forms a uniform film on the coated solar panel energy than uncoated panels (net increase ~3.6 %). This is surface, which accelerates runoff water spreading and flow, due to coated surfaces that show less light reflection and that removes soiling and deters the formation of drying smaller transmittance than uncoated surfaces resulting in marks as well, and finally results in a better energy more generated power. generation. Fig. 9 The spectrometric analysis of Nano-TiO coating on PV panels 2 Hu et al. Future Cities and Environment (2015) 1:8 Page 12 of 15 Fig. 10 Integrated energy output of testing solar panels via time Conclusions on coated surface seems to be 2 ~ 3 % less than that of an The flow of a liquid film is essential for self-cleaning. To uncoated surface; a coated surface shows 2 ~ 3 % more clean a surface, liquid has to transport the impurities light transmittance than an uncoated surface. This implies and finally run off the surfaces. It is the flowing liquid Nano-TiO coating itself doesn’t show any side effect on film that carries away the impurities. the solar cell performance whatsoever. Applying this In this study, a 2D dynamic model water film sliding Nano-TiO coating on a surface of a clean PV/PVT mod- on the inclined non-porous solid surface was established ule will potentially increase the efficiency comparing with that is based on lubrication theory in association with a other coatings which most likely reduce the transmissivity precursor film of a wetting liquid on glass. of sunlight. It is well known that nano-TiO coated glass surfaces In Switzerland, four identical solar monocrystalline display superhydrophilic properties. At the macroscale, by panels are used for the application of self-cleaning test- implementing this property into our simulation model, our ing under natural weather conditions. Pre-coating, post- simulation demonstrates that on superhydrophilic surface, coating and extremity tests were conducted, respectively. in a very short time, water completely wets the substrate Our experimental results reveal that the coated panel and spreads out into a thin film; on inclined glass surfaces generated ~3.6 % more energy than that with uncoated (up to 90°), water droplets/film slide faster on the more panel in a post-coating test. In an extremity test, the hydrophilic surfaces resulting in larger detachment force on coated panel generated ~5.4 % more energy than the particles and better effect on washing away loose particles. one with an uncoated panel in short term test thanks to The spectrometric analysis of Nano-TiO coating on several major merits of the Nano-TiO coating. 2 2 glass surface was conducted by our partner in Germany; Our entire experiment resulted in a most important in the wave length between 400 to 1200 nm, the reflection conclusion: the coated PV panels distinctively displayed Fig. 11 a Pre-coating test result. b Post-coating test result. c Extremity test result Hu et al. Future Cities and Environment (2015) 1:8 Page 13 of 15 − 1 better “self-cleaning” properties when muddy water was U, Adimensionalion factor for the velocity (m. s ) − 1 applied. In our cases of extremity test, our PV panel im- Characteristic flow velocity (m. s ) proved performance on an average of over 5 % compared V, Voltage (V) to the non-coated ones. As nano-superhydrophilic coat- V, Volume (m ) ings are easily applied on glass substrates by common v = (u,v), Velocity vector − 1 coating techniques, like spraying, dipping, flooding, spin- v, Velocity component in the x direction (m. s ) ning etc., − considering that large solar power plants cost x, x*, X coordinate (m) and dimensionless x coordinate billions of dollars – this performance boost would save tangentially to the surface support millions of dollars. Many surfaces would be easier to clean x , Adimensionalion factor for x-axis (m) and maintain if performed hydrophilic. α, Angle of the inclined plane(rad) This research experiment proved that superhydrophilic α, Exponent of the power-law − 1 surfaces have useful self-cleaning mechanisms, particularly γ, Air-water surface tension (N. m ) − 1 − 1 on the glass surfaces of the solar panels. With the weath- μ, μ , Dynamic viscosity of liquid (kg. m .s ) − 3 ering that many solar panels undergo, this self-cleaning ρ, Density of the water (kg. m ) − 3 process may greatly improve the performance and main- ρ , Density of air (kg. m ) tenance of solar cells and be a key towards Green Energy θ , Dynamic contact angle (°) in solar cell development. Our research in self-cleaning θ , θ , Equilibrium contact angle (°) e eq surfaces can be applied to many other fields as well. τ, Inertial time scale (s) Competing interests The authors declare that they have no significant competing financial, Nomenclature professional or personal interests that might have influenced the a, Capillary length (m) performance or presentation of the work described in this manuscript. μU Ca ¼ , Capillary number of the runoff equation Authors’ contributions C, Prefactor for the power-law 2 − 1 JH and NB carried out the simulation; JH designed and conducted the D, Diffusion coefficient (m .s ) experiments. OS and SR supervised the research work. All authors participated in D(α), Dimensionless parameter for the runoff equation the sequence alignment and drafted the manuscript of paper “CFD simulation and validation of self-cleaning on solar panel surfaces with superhydrophilic δ, film thickness on (m) − 2 coating” for Future Cities and Environment Journal, doi:10.1186/s40984-015-0006-7. g, Gravitational acceleration (m. s ) All authors read and approved the final manuscript. h’, Adimensionalion factor for the film thickness (m) H , Initial droplet or runoff film thickness (m) Acknowledgements The study is supported by EU HERB project. Special thanks to Mr. Morey h, h*, Film thickness (m) and dimensionless film Philippe for outdoor experimental setup and data acquisition. The thickness spectrometric analysis of Nano-TiO coating by Mr. Dietrich Schneider at h , Latent heat of condensation at the bulk air Stuttgart University of Applied Sciences is also acknowledged. fg − 1 temperature T∞. (J. kg ) Author details I, Current (A) 1 Institute of Thermal Engineering (IGT), University of Applied Sciences of i, Unit vector of the x-axis Western Switzerland, Av. des Sports 20, CH-1400 Yverdon-les-Bains, Switzerland. Department of Architecture and Built Environment, University j, Unit vector of the y-axis of Nottingham, Nottingham, UK. K, K’, Coefficients of the power law K ¼ , Kapitza number Received: 14 July 2015 Accepted: 14 July 2015 μU − 1 − 1 k , Thermal conductivity of condensate (W. m .K ) l, radial extension length (m) References L, Adimensionalion factor for the surface support (m) Ausserre D, Picard AM, Leger L (1986) Existence and role of the precursor film in the spreading of polymer liquids. Phys Rev Lett 57:2671–2674 L , Length of plate (m) Bascom WD, Cottington RL, Singleterry CR (1964) In Contact Angle, Wettability, ∇ =(∂x, ∂y), Nabla operator and Adhesion. Fowkes FM (ed) ACS, Washington, DC, p 355–379. R , Drop radius (m) Beaglehole D (1989) Profiles of the precursor of spreading drops of siloxane oil on glass, fused silica, and mica. J Phys Chem 93(2):893–899 R , the Reynolds number Berman A, Drummond C, Israelachvili J (1998) Amontons’ law at the molecular r, Spreading radius (m) level. Tribol Lett 4:95–101 t, t*, Time (s) and dimensionless time Bertozzi AL, Brenner MP (1997) Linear stability and transient growth in driven contact lines. Phys Fluids 9:530 Biance A-L, Clanet C, Quere D (2004) First steps in the spreading of a liquid T , Vapor saturation temperature (297 K) droplet. Phys Rev E 69:016301 T , Wall temperature (K) Bird JC, Mandre S, Stone HA (2008) Short-time dynamic of partial wetting. Phys − 1 Rev Lett 100:234501 u, Average velocity (m. s ) Blossey R (2003) “Self-cleaning surfaces — virtual realities”. Nat Mater 306 − 1 Velocity component in the x direction (m. s ) Bonn D, Indekeu J, Meunier J, Rolley E (2009) Reviews of Modern Physics − 1 u', Advancing velocity (m. s ) 81(05):739–805 Hu et al. Future Cities and Environment (2015) 1:8 Page 14 of 15 Carlson A, Do-Quang M, Amberg G (2011) Dissipation in rapid dynamic wetting, frequency generation spectroscopy. Phys Chem Chem Phys 10:4987–4993, J Fluid Mech 682:213–240 PubMed:18688544 Cazabat A-M, Cohen-Stuart MA (1986) Dynamics of wetting: effects of surface Ohtsu N, Masahashi N, Mizukoshi Y, Wagatsuma K (2009) Hydrocarbon decomposition roughness. J Phys Chem 90:5845 on a hydrophilic TiO2 surface by UV irradiation: spectral and quantitative analysis Chabas A, Lombardo T, Cachier H, Pertuisot MH, Oikonomou K, Falcone R, Verità, using in situ XPS technique. Langmuir 25(19):11586–11591 M. Geotti-Bianchini, F. (2008) Behaviour of self-cleaning glass in urban Oron A, Davis SH, Bankoff SG (1997) Long-scale evolution of thin liquid films. Rev atmosphere. Build Environ 43(12):2124–2131 Mod Phys 69(3):931 Chen J-D, Wada N (1989) Wetting dynamics near the edge of a spreading drop. Perevertaev VD, Metsik MS, Golub LM (1979) Dependence of the heat of Phys Rev Lett 62:3050 adsorption of water-molecules on the activity of the surface of muscovite Chen L, Auernhammer GK, Bonaccurso E (2011) Short time wetting dynamics on crystals. Colloid J USSR 41(1):124–126, Phys., 12:1473 – 1484, 1979 soft surfaces. Soft Matter 7:9084–9089 Rafai S, Sarker D, Bergeron V, Meunier J, Bonn D (2002) Superspreading: aqueous surfactant drops spreading on hydrophobic surfaces. Langmuir Dagaonkar M (2012) Effect of fluid flow, solution chemistry and surface morphology 18(26):10486–10488 of fibrous material on colloid filtration. J Eng Fibers Fabrics 7(3):62 Dagsson-Waldhauserova P, Arnalds O, Olafsson H, Skrabalova L, Sigurdardottir Raviv U, Perkin S, Laurat P, Klein J (2004) Fluidity of water confined down to GM, Branis M, Hladil J, Skala R, Navratil T, Chadimova L, von Lowis of Menar subnanometer films. Langmuir 20:5322–5332 S, Thorsteinsson T, Carlsen HK, Jonsdottir I (2014) Physical properties of Razouk RI, Salem AS (1948) The adsorption of water vapor on glass surfaces. suspended dust during moist and low wind conditions in Iceland. Icel Agric J Phys Chem 52(7):1208–1227 Sci 27(25):25–39 Rhykerd CL, Cushman JH, Low PF (1991) Application of multiple-angle-of- de Gennes P-G (1985) Wetting: statics and dynamics. Rev Mod Phys 57:827 incidence ellipsometry to the study of thin fluid films adsorbed on Diez JA, Kondic L (2002) Computing threedimensional thin film flows including surfaces. Langmuir 7(10):2219–2229 contact lines. J Comput Phys 183:274–306 Ruths M, Israelachvili JN (2011) Chapter 13, surface forces and nanorheology of Dunnill CWH, Aiken ZA, Pratten J, Wilson M, Morgan DJ, Parkin IP (2009) Enhanced molecularly thin films. In: Bhushan B (ed) Nanotribology and photocatalytic activity under visible light in N-doped TiO2 thin films produced nanomechanics. Springer-Verlag, Berlin Heidelberg by APCVD preparations using antibacterial films. J Photochem Photobiol Sah A (2014) Precursor film: a key driver to determine wetting behavior in the A207(2–3):244–253 vicinity of surface heterogeneity. Soft Matter 10:3890–3896 Ehrhard P (1993) Experiments on isothermal and nonisothermal spreading. Salim A, Huraib F, Eugenio N (1988) PV power-study of system options and J Fluid Mech 257:463 optimization. In: Proceedings of the 8th European PV Solar Energy Goossens D, Kerschaever EV (1999) Aeolian dust deposition on photovoltaic solar Conference. Florence, Italy cells: the effects of wind velocity and airborne dust concentration on cell Sayigh A (2009) Worldwide progress in renewable energy. Renewable Energy. performance. Solar Energy 66:277–289 Sciverse Science Direct. 20 November 2009:1-5. Greenspan HP (1978) On the motion of a small viscous droplet that wets a Sayigh A, Al-Jandal S, Ahmed H (1985) Dust effect on solar flat surfaces devices surface. J Fluid Mech 84:125 in Kuwait. In: Proceedings of the workshop on the physics of Hardy WB (1919) The spreading of fluids on glass. Phil Mag 38:49–55 Non-conventional energy sources and materials science for energy. Triest, Hardy WB (1936) Collected works. University Press, Cambridge Italy, pp 353–367 Sharma BK (1994) Air pollution, ISBN: 81-8283-014-1. Heslot F, Cazabat AM, Levinson P (1989) Dynamics of wetting of tiny drops: ellipsometric study of the late stages of spreading. Phys Rev Lett Smith SH (1969) On initial value problems for the flow in a thin sheet of viscous 62:1286–1289 liquid. J Appl Math Phys (ZAMP), 20, 556–560 Heslot F, Cazabat AM, Levinson P (1990) Experiment on wetting on the Smith JD, Cappa CD, Wilson KR, Messer BM, Cohen RC, Saykally RJ (2004) scale of nanometers: influence of the surface energy. Phys Rev Lett Energetics of hydrogen bond network rearrangements in liquid water. 65:599–601 Science 306:851–853 [PubMed: 15514152] Hocking LM, Rivers AD (1982) The spreading of a drop by capillary action. J Fluid Stapelbroek BBJ, Jansen HP, Kooij ES, Snoeijer JH, Eddi A (2014) Universal Mech 121:425 spreading of water drops on complex surfaces. Soft Matter 10:2641 Incropera FP, DeWitt DP (2002) Fundamentals of heat and mass transfer, 5th edn. Staszczuk P (1985) Application of the chromatographic step profile method for Wiley, Hoboken, NJ determination of water film pressure and surface freeenergy of quartz. Janczuk B, Chibowski E, Staszczuk P (1983) Determination of surface free-energy Chromatographia 20(12):724–728 components of marble. J Colloid Interface Sci 96(1):1–6 Tanner LH (1979) The spreading of silicone oil drops on horizontal surfaces. J Phys D Appl Phys 12:1473–1484 Joanny JF, de Gennes PJ (1984) Static structure of wetting films and contact lines. C R Acad Sci II 299:279–283 Tian C, Ji N, Waychunas GA, Shen YR (2008) Interfacial structures of acidic and Katz GB (2008) Effect of Dust on Solar Panels. Available www.gregorybkatz.com/ basic aqueous solutions. J Am Chem Soc 130:13033–13039, PubMed: Home/effect-of-dust-on-solar-panels. 18774819 Kavehpour HP, Ovryn B, McKinley GH (2002) Evaporatively-driven marangoni Troian SM, Herbolzheimer E, Safran SA, Joanny JF (1989) Fingering instabilities of instabilities of volatile liquid films spreading on thermally conductive driven spreading films". Europhys Lett 10:25 substrates. Coll Surf A 206:409–423 Ueno I, Watanabe T (2005) Dynamic behavior of precursor liquid film in Kazem HA, Khatib T, Sopian K, Buttinger F, Elmenreich W, Albusaidi AS (2013) front of moving droplet on solid, 16th International Symposium on Effect of dust deposition on the performance of multi-crystalline Transport Phenomena, ISTP-16, Prague. photovoltaic modules based on experimental measurements. Int J Renew Valignat MP, Oshanin G, Villette S, Cazabat A-M, Moreau M (1998) Molecular Energy Res 3:4 weight dependence of spreading rates of ultrathin polymeric films. Phys Rev Lett 80:5377–5380 Kazmerski L (2011) Photovoltaic: history, technology, markets, manufacturing, Villette S, Valignat MP, Cazabat AM, Jullien L, Tiberg F (1996) Wetting on the applications, and outlook, 84th International Seminar in Brighton, Renewable Energy Policy, Security, Electricity, Sustainable Transport, Water molecular scale and the role of water. A case study of wetting of hydrophylic silica surfaces. Langmuir 12:825–830 Resources/Management and Environment, Brighton, UK, 3–9. Lee A, Moon M-W, Lim H, Kim W-D, Kim H-Y. (2012) Water harvest via dewing. Voinov OV (1976) Hydrodynamics of wetting. Fluid Dynamics 11:714–721 Langmuir 28(27):10183–10191. Vou’e M, Valignat MP, Oshanin G, Cazabat A-M (1999) Dissipation processes at Levinson P, Cazabat A-M, Cohen-Stuart MA, Heslot F, Nicolet S (1988) The the mesoscopic and molecular scale. The case of polymer films. Langmuir spreading of macroscopic droplets. Rev Phys Appl 23:1009 15:1522–1527 Lopez J, Miller CA, Ruckenstein E (1976) Spreading kinetics of liquid drops on Wack KO (1980) “Cleaning system for solar energy collectors”, Physics Abstract solids. J Colloid Interface Sci 53:460–461 Science Marmur A, Lelah MD (1980) The dependence of drop spreading on the size of Wakim F (1981) Introduction of PV power generation to Kuwait. Kuwait Institute the solid surface. J Colloid Interf Sci 78(262):265 for Scientific Researchers, Kuwait City Noguchi H, Hiroshi M, Tominaga T, Gong JP, Osada Y, Uosaki K (2008) Interfacial Wang XD (2003) Contact angle hysteresis and dynamic wetting behavior on water structure at polymer gel/quartz interfaces investigated by sum rough solid surfaces, Ph.D. Thesis. Tsinghua University, China Hu et al. Future Cities and Environment (2015) 1:8 Page 15 of 15 Winkels KG, Weijs JH, Eddi A, Snoeijer JH (2012) Initial spreading of low-viscosity drops on partially wetting surfaces. Phys Rev E 85(5):055301 Yeo L (2008) Wetting and Spreading, Encyclopedia of Microfluidics and Nanofluidics, pp 2186–2196 Yoo H, Paranji R, Pollack GH (2011) Impact of hydrophilic surfaces on interfacial water dynamics probed with NMR spectroscopy. J Phys Chem Lett 2(6):532–536 Yuan Q, Zhao Y-P (2010) Precursor film in DynamicWetting, electrowetting, and electro-elasto-capillarity. Phys Rev Lett, Prl 104:246101 Zisman WA (1965) Improving the performance of reinforced plastics. Ind Eng Chem 57:26–34 Zoeteweij ML, van der Donck JCJ, Versluis R (2009) Particle removal in linear shear flow: model prediction and experimental validation. J Adh Sci Technol 23(6):899–911 Submit your manuscript to a journal and benefi t from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the fi eld 7 Retaining the copyright to your article Submit your next manuscript at 7 springeropen.com

Journal

Future Cities and Environment – Springer Journals

Published: Aug 22, 2015

There are no references for this article.