Solar Energy 157 (2017) 559–565

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Experimental study of self-cleaning property of titanium dioxide and T

nanospray coatings in solar applications
⁎
Aslan Gholamia, Ali Akbar Alemrajabib, Ahmad Saboonchib,
a
Isfahan University of Technology, Islamic Republic of Iran
b
Department of Mechanical Engineering, IUT, Isfahan 84156-83111, Islamic Republic of Iran

A R T I C L E I N F O A B S T R A C T

Keywords: Nowadays, the urgent need to replace fossil fuels with renewable energy is no secret. Therefore, researching in
Solar energy this area and developing the infrastructures necessary for the optimized utilization of these types of energy, are
Dust accumulation rising sharply. Among different kinds of renewable energy, due to easy access, solar energy has received special
Transmission coefficient attention and extensive studies have been done, in order to increase the efficiency of the solar collectors and PV
Self-cleaning property
panels. One of the prime problems of solar systems in arid regions is the accumulation of dust on transparent
Titanium dioxide
surfaces, followed by an intensive drop in the transmission coefficient. In the present study, in a 70-day test
Nanocoating
period, up to 22% transmission coefficient loss, caused by the accumulation of dust on the surface, was observed.
Thus, in order to solve this problem, Nanocoatings were deposited on glass samples to create self-cleaning
property. A nanospray was used to create a hydrophobic film on the surface and TiO2 was coated in three
different thickness to create a hydrophilic film on sample surfaces. Test results showed that these surface
modifications have a significant impact in the reduction of dust accumulation and the loss of the transmission
coefficient. Although there was no rain during the test, for samples with surface modification, the transmission
coefficient drops were noticeably reduced. After rain simulation, the results indicate that transmission loss re-
duced from 22% to 0.5% and dust settlement problem was almost completely solved.

1. Introduction
In recent years, global increase of energy prices in one hand and the
decrease of solar system costs as well as attention to environmental
aspects, has made solar energy an attractive solution to overcome the
world energy demands, and in addition to creating solar power plants
and solar farms, numerous projects and researchers in this field have
been done (Arsalis and Alexandrou, 2015; Karaki et al., 2016; Dabaieh
and Elbably, 2015; Gholami et al., 2015). Furthermore, supporting laws
for developing renewable energy, which were passed in many coun-
tries, following by investment tendency, led to many techno-econom-
ical studies and various schemes to optimize photovoltaic systems
(Fabrizio et al., 2010; Petersen and Svendsen, 2012; Ramadhan and
Naseeb, 2011; Lacchini and Dos Santos, 2013; Nottrott et al., 2013;
Shrimali and Jenner, 2013). Locally, a recent study evaluated and
compared the economic aspect of increasing distributed generation
capacity in Iran (Zandi et al., 2017).
Whether using solar collector or photovoltaic panel, the sun radia-
tion should pass through a cover glass before converting to heat or
electricity. So the transmission coefficient1 of the cover glass is as
important for the system efficiency as the other factors such as panel
pitch (β) and orientation (γ), cell temperature, maximum power point
tracking and energy conversion efficiency. Settlement of dust on a glass
cover causes gradual reduction of transmission coefficient (Hegazy,
2001; Gholami et al., 2017), which then results in the overall output
energy reduction.
Due to favorable radiation conditions, including high radiation
power and long daylight, arid regions are more suitable for building
large-scale solar installations. One of the barriers in using solar systems,
either photovoltaic panels or flat collectors, in such areas, is the accu-
mulation of dust on surfaces, which follows by reduction of transmis-
sion coefficient and total efficiency of the solar system as well (He et al.,
2011; Saidan et al., 2016; Appels et al., 2013). Several factors, such as
installation tilt, azimuth angle, dominant wind direction and intensity,
rain and the time in which panels and collectors are exposed to outdoor
conditions without cleaning, affect dust accumulation density on sur-
face and transmittance reduction (Elminir et al., 2006; Sarver et al.,
2013). In reading the literature, it can be understood that dust accu-
mulation strongly depends on the weather conditions (Tanesab et al.,
2015; Paudyal and Shakya, 2016; Ferrada et al., 2015; Abderrezek and

⁎
Corresponding author.
E-mail addresses: aslan.gholami@gmail.com (A. Gholami), rajabi@cc.iut.ac.ir (A.A. Alemrajabi), ahmadsab@cc.iut.ac.ir (A. Saboonchi).
1
Transmission Coefficient or Transmittance is the ratio of the total radiant or luminous flux transmitted by a transparent object to the incident flux.

http://dx.doi.org/10.1016/j.solener.2017.08.075
Received 4 March 2017; Received in revised form 27 June 2017; Accepted 28 August 2017
Available online 14 September 2017
0038-092X/ © 2017 Elsevier Ltd. All rights reserved.
A. Gholami et al. Solar Energy 157 (2017) 559–565

Table 1
Deposition conditions during the coating, using radio frequency sputtering method.

Sample number Coating thickness (nm) Power (W) Reflected power (W) Chamber pressure (μbar) Sputtering rate (A0/s) Substrate rotation (rpm)

1 30 89 2 3.6 2–3.4 10
2 30 91 1 5 1.5–3.9 10
3 30 90 0 4.8 2–3.4 10
4 50 94 1 4.8 2–3.9 10
5 50 92 0 4.8 2–3.4 10
6 50 91 1 4.8 2–3.4 10
7 70 92 0 4.8 2–2.4 10
8 70 92 0 5 2–3.9 10
9 70 93 0 5 2–3.4 10

Fathi, 2017), for example, a single dust storm can reduce the power
output of PV modules by as much as 20% (Adinoyi and Said, 2013).
Recently, Jiang et al. (2016), developed a model to estimate the
cleaning frequency for dirty solar photovoltaic (PV) modules in the
desert environment, based on PV module power performance, dust
density and deposition velocity. They also suggested cleaning of PV
modules every 20 days in desert regions.
In order to receive the optimum sunlight during a year, installation
tilt is usually set the same as the location latitude. Furthermore, in the
northern hemisphere, solar collectors and PV panels are typically faced
toward south (Duffie and Beckman, 1991). Wind and rain conditions
are also depending upon the climate and cannot be controlled. So it
seems that the only way to prevent the transmission coefficient drop is
to remove the dust from the surface.
Transparent surfaces can be cleaned manually, automatically and
naturally (Sayyah et al., 2014). Wind power, gravitational force, and
rain are included as the natural ways which could take settled dust
away from the surfaces. Gaier et al. (1990), Gaier and Davis (1992),
studied the validity of this method on Mars which has huge dust storms
each year. They reported that if panels could be oriented in such a way
that the angle of attack and wind speed are proper, wind will clear off
the dust from the surface of a photovoltaic array. Since the installation
tilt angle of panels on earth is usually set by the latitude of the site, the
proper angle of attack may not be obtained. Furthermore, it is difficult
to rotate large arrays of solar panels, and this may lead to energy
consumption and reduction of total efficiency of the system. In another
work, William et al. Williams et al. (2007), studied the vibration
characterization of self-cleaning solar panels with piezoceramic actua-
tion. Usually, due to high-power consumption, difficulties and expenses
relating to the machinery maintenance, using mechanical methods,
such as brushing, blowing, ultrasonic excitement and vibration, to re-
move dust from the surface of the solar panels, are not efficient.
Electrostatic dust removal method was studied by Clark et al.
(2007), in order to reduce the dust level on the surface, in the lunar
application. The most famous electro dust removal method, based on
the concept of electric curtains, was developed by F.B. Tatom and
collaborators at NASA in 1967. In this method which later was devel-
oped by Masuda at Tokyo University in the 1970s, electrostatic and
dielectrophoretic forces are used to shift and move dust particles. In
recent years, this technology has been used for space applications in the
moon and Mars (Calle et al., 2008, 2009; Atten et al., 2009; Sharma
et al., 2009; Bock et al., 2008; Liu and Marshall, 2010).
Perhaps the most optimal method to remove dirt from surfaces is to
modify the surface to obtain self-cleaning properties. Self-cleaning na-
nofilms which could be coated on the surface for this purpose should be
transparent and they are usually either super-hydrophobic or super-
hydrophilic. Most researchers about super-hydrophobic and hydro-
philic surfaces, are mainly focused on either reducing or increasing the
wetting properties, and still, the question remains that whether it is
possible to use these self-cleaning nanofilms for solar applications or
not. In the current study, to overcome dust accumulation problem, TiO2
nano-coatings and nanospray were used on some of the glass samples
and in a 70-day period, the dust surface density and the transmission
coefficient loss for the coated and uncoated glass samples, were mea-
sured and compared. In the following parts, first, selected coatings will
be introduced briefly. Then, after reporting the climate conditions of
test region, basic specifications of the experiment and the results will be
explained.
2. The hydrophobic and hydrophilic coatings
Super-hydrophobic surfaces such as lotus leaves, show extremely
low wetting property and greatly high hydrophobicity. Nanostructures
of these surfaces could enhance contact angle over 150°. So the rain-
drops, immediately after reaching the surface, would roll on it and wash
away dust particles (Park et al., 2011).
In the current study, the hydrophobic property was added to the
surface of seven glass samples, using a nanospray prepared by
Nanospadan company. To do so, first the glass surfaces were washed by
water and alcohol solution, then the hydrophobic solution was sprayed
on the glass surface. The time required for the stability was 24 h.
The most famous super-hydrophilic film is titanium dioxide which,
in addition to the hydrophilic property, also has a photocatalytic
characteristic. This self-cleaning method has two steps. The first phase
is a photocatalytic process, in which titanium dioxide under the effect
of ultraviolet radiation breaks down organic dust. Then in the next step,
due to the super-hydrophilic property, raindrops, rather than gathering
in one place, spread over the entire surface and wash away the dust
particle (Park et al., 2011).
To coat TiO2 on the surfaces, the equipment at the Laser and Plasma
Research Institute of Shahid Beheshti University were used. TiO2 was
coated on each sample, using radio frequency sputtering method, which
is considered as a physical vapor deposition under vacuum method.
Deposition conditions in the laboratory are given Table 1. It should be
noted that the nanospray is a less permanent surface modification than
titanium dioxide and need to be recoated every two years.
3. Climatology of the test region
Test apparatuses were located on the roof of mechanical en-
gineering department in Isfahan university of technology at 32°43′00″
north latitude and 51°31′45″ East longitude. Therefore required climate
and weather information was extracted from the closest weather station
(32°47′00″ north latitude and 51°67′00″ East longitude).
Due to the rainfall decline in recent years, 84 percent of Isfahan
province is experiencing drought, which 15.8 percent is severe drought.
In addition, by comparing the amount of dust in recent years and the
long-term annual average, a significant increase in the region and the
country becomes apparent (The National Weather Service, 2014-15),
which emphasizes more on the importance of this study.
4. Experimental basic specifications
Given that in the northern hemisphere in order to receive the most

560
A. Gholami et al.
Fig. 1. One of the five devices used to hold glass samples in different tilts.
solar radiation possible, fixed collectors and PV panels are generally
installed facing south with a tilt same as the location latitude (Duffie
and Beckman, 1991), in the current study glass samples were installed
at different tilt angles while faced south. The experiment was done
outdoors and for a better simulation, samples were exposed to en-
vironmental conditions. During the tests, great care was taken to avoid
any dust removal except by natural causes, such as wind and rain.
4.1. Sample preparation
For testing, a total of 24 pieces of glass, with dimensions of 100
(mm) × 100 (mm) and thickness of 3 (mm), was required. Of these, 7
pieces of uncoated glass and 7 pieces of samples with hydrophobic
nanospray coating were placed in seven different angles 0°, 15°, 30°,
45°, 60°, 75° and 90°. 9 samples, on which TiO2 was deposited ac-
cording to Table 1 conditions, were also placed in 3 tilt angles of 0°, 45°
and 90°, and a piece, which was cleaned before each measurement, was
used as criteria. To put the pieces of glass in different angles, five de-
vices were designed as shown in Fig. 1. As it can be seen in the Figure,
only one side of the glass samples (the coated side), were exposed to
dust settlement. For more accuracy, a level and a compass were in-
stalled on each device.
4.2. Test details
Before the experiment began, mass of all the samples were carefully
measured and recorded. Experiment was started on May 23rd, 2016 and
over a period of 70 days up until August 1st, 2016, every week dust
level on the surface as well as the transmission coefficient of each
sample was measured. To measure the quantity of dust accumulated on
the glass surfaces, the glass samples were weighed under controlled
conditions using a RADWAG AS 220/C/2 balance with an accuracy of
0.0001 g. The difference between initial mass (clean) and the mass of a
sample at each stage (with settled dust), represented the amount of dust
settlings on the glass. Then the surface density of the dust accumulated
on glass obtained by dividing the dust weight by the surface area of the
glass.
In addition to measuring mass of settled dust on the glass and cal-
culating the dust surface density every week, the amount of radiation
passing through each sample was measured using a LUTRON SPM-
1116SD solar power meter. Then the transmission coefficient for each
sample was calculated as the ratio of the total radiant flux transmitted
by the sample to the incident flux. The transmission coefficient loss (or
the reduction of the transmission coefficient), due to dust settlement,
was determined by comparing the transmission coefficient for the
561
Solar Energy 157 (2017) 559–565
samples and the clean criteria sample.
Dust might not be accumulated uniformly on the sample surfaces.
Non-uniformity deposition of dust usually happens due to the gravity or
wind blow. This may impact on the reported transmission coefficient. In
order to solve this problem for each measurement, transmission coef-
ficient was measured through 9 points of each sample, and the average
was reported. In the following part tests’ results are represented and
analyzed.
Since during the examination, there was no rainfall locality, the
impact of rain on surfaces cleaning and coatings functioning was in-
vestigated, by rain simulation. A rain machine was used for this simu-
lation and with the help of a simple rain gauge, the amount of rainfall
during the rain simulation was measured. In the following part tests’
results are represented and analyzed.
For each number reported in the diagrams and tables, for either
mass or transmission coefficient loss, at least 10 measurements were
taken and the average was reported. The following correlation could be
applied for all the measurements data:
⎛1− max|x i−x avg | ⎞ < 0.99
⎜ ⎟
⎝ x avg ⎠ (1)
So, all the results and reported data, measured during the test, have
at least 99% repeatability.
5. Results and discussion
In the current study, the effect of surface modifications by TiO2
coating and nanospray is investigated. To do so, coated and uncoated
samples exposed to the outdoor conditions for 70 days. The final sur-
face densities of accumulated dust versus tilt angles are illustrated in
Fig. 2.
According to figure, it seems that as the installation tilt (β) in-
creases, generally the accumulation of dust on the surface, due to
gravity, is reduced. The surface modifications, due to the hydrophilic
and hydrophobic properties, show their best performance in the pre-
sence of rain. Despite the fact that through the experiment there was no
rain, still, there was a significant difference in the amount of dust for
coated and uncoated surfaces. This reduction in dust level, might be due
to the decrease in surface adhesion or the surface roughness. It should
be noted that the effects of roughness depends on its scale. If the scale of
the roughness features are small compared to the dust particles, the
adhesion of the dust will go down. But if the roughness features are
similar in size to the particles they can get stuck on them.
The weekly average of accumulated dust surface density could be
calculated by dividing the final surface density by the number of the
weeks (10). This numbers show how much extra dust was accumulated
weekly. As can be obtained from Fig. 2, for samples tilted 45°, the
maximum dust surface deposition was 0.7264 (g/m2) per week for the
uncoated sample, and the minimum dust surface deposition was 0.3889
(g/m2) per week for the sample with 50 nm TiO2. This indicates that the
coating could reduce the amount of dust on the surface up to 46%.
As the dust accumulated on the surface, the transmission coefficient
reduced. The reduction of transmission for each sample after 70 days
versus tilt angle is illustrated in Fig. 3. As mentioned earlier since there
was no rain during the measurement, it can be seen that for the hor-
izontal plates, the coatings had less effect. According to Fig. 3, the best
performance of the surface modifications in no rain conditions was
appeared in samples installed at approximately 30–60° tilt angles. For
example, the transmission coefficient loss caused by dust accumulation,
reduced from about 24% for the uncoated glass to 14% for the sample
with 50 nm dioxide titanium coating. This means that even without
rain, the nano-coating decreased the effect of dust settlement on the
surface about 41%.
In order to accurately study the process of dust settlement on sur-
faces over time and eliminate other factors such as installation tilt and
A. Gholami et al. Solar Energy 157 (2017) 559–565

9.5 Fig. 2. The final surface density of dust versus tilt angles for
coated and uncoated samples.
South Without
oĂƟng 8.5

Nano-spray
7.5
ρ (gr/m 2)

TiO2-30nm 6.5
d

5.5
TiO2-50nm

4.5

TiO2-70nm
3.5

2.5
0 15 30 45 60 75 90 105
Angle (β °)

orientation, in Figs. 4–6, the dust surface density, measured in each
week, is plotted respectively for installation tilt of 0°, 45° and 90°. It was
found through the course of the experiment that as the time passed,
dust level on the surface increased. It can be seen from the diagrams
4–6, that the rate, at which dust settled on the surface, was faster at the
beginning of the experiment and as time passed it slowed down.
Furthermore, a jump in the dust surface density appeared in the
ninth week which is due to the weather conditions, and the dust storm
occurred in that week. As mentioned earlier and can be seen in figures,
dust accumulation diagrams versus time are strongly depended on the
weather conditions and phenomena during the measurement period, so
the results may not be applicable even for the same location in other
weather conditions, but they give us a vision about the weather con-
ditions through the examination. Again it could be understood from the
above figure that in the absence of rain, coatings showed a more self-
cleaning effect in samples with installation tilt angle of 30–60°.
Since the primary characteristics of the coatings are hydrophilicity
and hydrophobicity, they may have their best performance in the pre-
sence of rain. So, as there was no rainfall in the test region during the
experiment, rain was simulated.
A rain machine was used for rain simulation. Since there might be
physical or chemical reactions between dust particles and the glass, rain
was simulated in two stages. In the first stage, it was raining on the
samples for one hour, which was equal to 0.5 mm rainfall. Then after a
few hours, when all the samples dried under the sunlight, the amount of
remaining dust on the surface and transmission coefficient, were mea-
sured. After that in the second stage, again rain was simulated the same
as first one. Dust surface density and the transmission loss due to dust
deposition measured before and after the rain simulation. The test re-
sults for uncoated and coated samples at different tilt angles are pre-
sented in Table 2.
For uncoated samples, it was observed that after the rain simulation,
in addition to the fact that dust was not completely removed from the
surface, raindrops and dust left some stains on the surface. By com-
paring the amount of transmission loss proportional to dust surface
density before and after the rain simulation, it can be understood that
these stains caused an additional transmission drop. Moreover, using
the test results in Table 2 and comparing the amount of dust washed
away in each rain stage, less dust removal and surface cleaning in the
second rainfall can be seen. This may be caused by the adhesion created
between dust particles and glass surface after the first rainfall. For these
two reasons (creation of stains and increase of stickiness), despite the
two stages of rain, still up to 10% transmission coefficient loss due to
dust settlement can be seen.
In the contrary, for the coated samples, there were no stains left
after the rain simulation, and dust was completely removed from the

26 Fig. 3. The average transmission coefficient loss versus tilt

angles for coated and uncoated samples.
South Without
oaƟng 24
Nano-spray
22

TiO2-30nm 20
Δ τ (%)

18
TiO2-50nm

16
TiO2-70nm
14

12

10
0 15 30 45 60 75 90 105
Angle ( β °)

562
A. Gholami et al. Solar Energy 157 (2017) 559–565

9 Fig. 4. Dust surface density versus time for horizontal

samples.

South Without
7
oaƟng

Nano-spray 6
ρd (gr/m2)

5
TiO2-30nm
4
TiO2-50nm
3
TiO2-70nm
2

0
0 1 2 3 4 5 6 7 8 9 10
Week

8 Fig. 5. Dust surface density versus time for samples with

45° installation tilt.

7
South Without
oaƟng 6
Nano-spray
5
TiO2-30nm
ρd (gr/m2)

4
TiO2-50nm
3
TiO2-70nm
2

0
0 1 2 3 4 5 6 7 8 9 10
Week

6 Fig. 6. Dust surface density versus time for samples with

90° installation tilt.

South Without 5
oaƟng

Nano-spray
4
ρ (gr/m 2)

TiO2-30nm
3
d

TiO2-50nm
2
TiO2-70nm

0
0 1 2 3 4 5 6 7 8 9 10
Week

563
A. Gholami et al. Solar Energy 157 (2017) 559–565

Table 2 Several factors, such as mounting tilt angle, the orientation of panels
Rain simulation test’s results. and the dominate wind direction and other weather conditions at the
site, affect the settlement of dust on the surface. Generally, in the
Sample Before rain tests Rain#1 (1 h & 0.5 mm Rain#2 (2 h & 1 mm
rain) rain) northern hemisphere, the solar panels and collectors are facing towards
the south. To optimize receiving radiation annually, the mounting tilt
ρd (g/m2) Δτ (%) ρd (g/m2) Δτ (%) ρd (g/ Δτ angle is usually selected the same as the latitude of the installation.
m2) (%) Given that the weather conditions at the site, also cannot be controlled,
Uncoated-0° 8.4289 21.64 4.0141 14.23 2.4459 10 the only way to reduce the accumulation of dust, is to modify the glass
Uncoated-15° 7.5832 21.5 3.5178 12.2 2.3324 10 cover on the surface, in order to have self-cleaning property.
Uncoated-30° 7.5734 21.64 3.4217 11.7 2.2741 9.6 For this purpose in the current study, on the surface of several glass
Uncoated-45° 7.2643 21.23 3.2341 10.5 2.1102 9.4 samples, either hydrophobic nanofilms (nanospray) or hydrophilic na-
Uncoated-60° 6.5929 21 3.0714 9.87 1.9402 9
Uncoated-75° 6.5832 20.89 2.8537 9.34 1.9138 8.4
nofilms (titanium dioxide) were coated. The samples were exposed to
Uncoated-90° 5.2181 19 2.209 7.41 1.7408 8.5 the outdoor conditions over a 70-day test course and dust surface
Nanospray-0° 7.5014 21.5 2.385 7.25 1.6341 7 density, as well as transmission coefficients for coated and uncoated
Nanospray-15° 6.4948 21 2.0117 6 0.8742 2.5 samples, were measured. The results confirmed that deposition of the
Nanospray-30° 5.0362 18 1.9782 5.7 0.7813 2
nanofilm coatings on the surface, even without any rainfall, causes a
Nanospray-45° 4.7014 17 1.9054 5.5 0.7576 2
Nanospray-60° 4.3757 15 1.8248 5.5 0.6829 2 significant (up to 41%) decrease in the accumulation of dust on the
Nanospray-75° 4.2959 15 1.7022 5 0.5934 1.8 surface.
Nanospray-90° 3.5991 11.5 1.2345 4 0.5561 1.8 For the uncoated samples, it was observed that after the first rain
TiO2-30 nm-0° 7.0862 21.34 1.566 5 0.3734 1 simulation, accumulated dust particles were not washed away, due to
TiO2-30 nm-45° 4.2498 15 1.2639 4 0.2564 0.8
TiO2-30 nm-90° 4.0002 14 0.7863 3 0.2284 0.5
the adhesion increase between them and the surface. Moreover, the
TiO2-50 nm-0° 7.0625 21.2 1.3741 4.8 0.2436 0.5 stains that remained on the surface after the rain, led to a noticeable
TiO2-50 nm-45° 3.8894 12.84 1.1874 3.7 0.1975 0.3 reduction in the transmission coefficient. On the contrary, for the
TiO2-50 nm-90° 3.6389 11.75 0.6541 2.5 0.1343 0.1 coated samples after the rain simulation, the surfaces were almost
TiO2-70 nm-0° 7.1408 21.34 1.4582 5 0.2884 0.5
completely clean and dust particles did not stick to them and also there
TiO2-70 nm-45° 4.6691 16.68 1.2236 4.2 0.2111 0.4
TiO2-70 nm-90° 3.6419 12 0.7152 2.5 0.1566 0.2 was no stain left. Among the coated samples, titanium dioxide coating
with the thickness of 50 (nm) showed the best performance. Titanium
dioxide nanofilms in addition to better self-cleaning property, are more
modified surfaces. As reported in Table 2, after the rain simulation, permanent while nanospray coatings should be renewed every two
these surfaces almost became as clean as their initial state. The best years.
surface self-cleaning property was observed respectively for samples The current study, like any other study, needs further research in
with 50 (nm), 70 (nm) and 30 (nm) thickness of TiO2 coatings and after some areas to reduce the problems caused by dust accumulation on the
them the nanospray coating. surfaces of the solar equipment. Since the experiment for this article
The difference between the transmission coefficient loss for coatings was done in a 70-day period, a long-term experiment should be done to
with 50 nm and 70 nm TiO2, for 90° cannot be distinguished clearly in see if it could obtain more accurate results. In addition, the current
Fig. 3. However according to Table 2, by comparing the dust surface study did not optimize the thickness of the coatings, and further re-
density, it can be seen that TiO2 coating with 50 nm thickness showed searches are needed in this area.
better surface modification effects. In the samples tilted 90°, in com-
parison to 45°, the difference between the accumulated dust surface References
densities is quite low (of the order of magnitude of 10−1). In calculating
the transmission coefficient, the amount of radiation passing through Abderrezek, M., Fathi, M., 2017. Experimental study of the dust effect on photovoltaic
each sample was of the order of magnitude of 103. This results in low panels’ energy yield. Sol. Energy 142, 308–320.
Adinoyi, Muhammed J., Said, Syed A.M., 2013. Effect of dust accumulation on the power
differences between transmission coefficient losses of the coatings. outputs of solar photovoltaic modules. Renew. Energy 60, 633–636.
Both hydrophobic and hydrophilic films showed a significant re- Appels, R., Lefevre, B., Herteleer, B., Goverde, H., Beerten, A., Paesen, R., De Medts, K.,
duction in transmission coefficient loss. However according to the re- Driesen, J., Poortmans, J., 2013. Effect of soiling on photovoltaic modules. Sol.
Energy 96, 283–291.
sults in Table 2, after the rain simulation reduction in transmission Arsalis, A., Alexandrou, A.N., 2015. Parametric study and cost analysis of a solar-heating-
coefficient loss for samples coated with nanospray and TiO2, was of the and-cooling system for detached single-family households in hot climates. Sol. Energy
order of magnitude of 1 and 10−1, respectively. So the results indicate 117, 59–73.
Atten, P., Pang, Hai Long, Reboud, J.L., 2009. Study of dust removal by standing-wave
that TiO2 coatings were more effective in creating self-cleaning prop- electric curtain for application to solar cells on mars. IEEE Trans. Ind. Appl. 45 (1),
erty. As mentioned previously, the nanospray is a less permanent sur- 75–86.
face modification than TiO2 and need to be recoated every two years. Bock, J.P., Robison, J.R., Sharma, R., et al., 2008. An efficient power management ap-
proach for self-cleaning solar panels with integrated electrodynamic screens. In: Proc.
Despite these disadvantages, since nanospray coatings have a lower
ESA Annual Meeting on Electrostatics; Paper O2.
price and are easier to use, they might be a more interesting candidate Calle, C.I., McFall, J.L., Buhler, C.R., et al., 2008. Dust particle removal by electrostatic
for large-scale PV installation. and dielectrophoretic forces with applications to NASA exploration missions. In: Proc.
The results from these experiments show that using the suggested ESA Annual Meeting on Electrostatics; Paper O1.
Calle, C.I., Buhler, C.R., McFall, J.L., et al., 2009. Particle removal by electrostatic and
surface modifications could solve the problem of dust accumulation on dielectrophoretic forces for dust control during lunar exploration missions. J.
any glass cover, such as the ones used for PV panels. Electrostat. 67, 89–92.
Clark, P.E., Minetto, F.A., Keller, J., et al., 2007. Finding a dust mitigation strategy that
works on the lunar surface, The 38th Lunar and Planetary Science, Texas.
Dabaieh, M., Elbably, A., 2015. Ventilated Trombe wall as a passive solar heating and
6. Conclusion and recommendations cooling retrofitting approach; a low-tech design for off-grid settlements in semi-arid
climates. Sol. Energy 122, 820–833.
Duffie, J.A., Beckman, W.A., 1991. Solar engineering of thermal processes, 2dn ed. John
To full fill energy needs and reduce pollution, greater uses of solar Wiley & Sons, Inc., New York.
systems are inevitable. One of the problems in using solar systems in the Elminir, H.K., Ghitas, A.E., Hamid, R.H., El-Hussainy, F., Beheary, M.M., khaled, M.A.,
arid climate with dust storm, such as Iran, is dust accumulation on the 2006. Effect of dust on the transparent cover of solar collectors. Energy Convers.
Manage. 47, 3192–3203.
surface which causes a noticeable transmission coefficient loss, fol- Fabrizio, E., Corrado, V., Filippi, M., 2010. A model to design and optimize multi-energy
lowed by an obvious reduction in performance of the system.

564
A. Gholami et al.
systems in buildings at the design concept stage. Renew. Energy 35 (3), 644–655.
Ferrada, P., Araya, F., Marzo, A., Fuentealba, E., 2015. Performance analysis of photo-
voltaic systems of two different technologies in a coastal desert climate zone of Chile.
Sol. Energy 114, 356–363.
Gaier, J., Davis, P., 1992. Effect of particle size of Martian dust on the degradation of
photovoltaic cell performance, NASA TM-105232.
Gaier, J., Davis, P., Marabito, M., 1990. Aeolian removal of dust types from photovoltaic
surfaces on Mars. In: 16th AIAA/NASA/ASTM/IES Space Simulation Conference. NM:
Albuquerque.
Gholami, A., Eslami, Sh., Alam Rajabi, A.A., 2015. Photovoltaic and solar water heater
system design and economic analysis using RETScreen software for solar energy park.
In: Second International Conference and Exhabition on Solar Energy.
Gholami, A., Saboonchi, A., Alemrajabi, A.A., 2017. Experimental study of factors af-
fecting dust accumulation and their effects on the transmission coefficient of glass for
solar applications. Renew. Energy 112, 466–473.
He, Gaofa, Zhou, Chuande, Li, Zelun, 2011. Review of self-cleaning method for solar cell
array. Proc. Eng. 16, 640–645.
Hegazy, A.A., 2001. Effect of dust accumulation on solar transmittance through glass
covers of plate-type collectors. Renew. Energy 22, 525–540.
Jiang, Y., Lu, L., Lu, H., 2016. A novel model to estimate the cleaning frequency for dirty
solar photovoltaic (PV) modules in desert environment. Sol. Energy 140, 236–240.
Karaki, S., Brisbane, T.E., Loef, G.O.G., 2016. Performance of the solar heating and
cooling system for CSU solar house III, summer season 1983 and Winter Season 1983-
84. Guide to the Records of the Solar Energy Applications Laboratory and Solar
Village. Finding aid: < http://lib.colostate.edu/archives/findingaids/university/
ucsv.html > .
Lacchini, C., Dos Santos, J.C.V., 2013. Photovoltaic energy generation in Brazil-Cost
analysis using coal-fired power plants as comparison. Renew. Energy 52, 183–189.
Liu, G.Q., Marshall, J.S., 2010. Effect of particle adhesion and interactions on motion by
traveling waves on an electric curtain. J. Electrostat. 68, 179–189.
Nottrott, A., Kleissl, J., Washom, B., 2013. Energy dispatch schedule optimization and
cost benefit analysis for grid-connected, photovoltaic-battery storage systems. Renew.
Energy 55, 230–240.
Park, Y.B., Im, H., Im, M., et al., 2011. Self-cleaning effect of highly water-repellent
microshell structures for solar cell applications. J. Mater. Chem. 21, 633–636.
565
Solar Energy 157 (2017) 559–565
Paudyal, B.R., Shakya, S.R., 2016. Dust accumulation effects on efficiency of solar PV
modules for off grid purpose: a case study of Kathmandu. Sol. Energy 135, 103–110.
Petersen, S., Svendsen, S., 2012. Method for component-based economical optimisation
for use in design of new low-energy buildings. Renew. Energy 38 (1), 173–180.
Ramadhan, M., Naseeb, A., 2011. The cost benefit analysis of implementing photovoltaic
solar system in the state of Kuwait. Renew. Energy 36 (4), 1272–1276.
Saidan, M., Albaali, A.G., Alasis, E., Kaldellis, J.K., 2016. Experimental study on the effect
of dust deposition on solar photovoltaic panels in desert environment. Renew. Energy
92, 499–505.
Sarver, T., Al-Qaraghuli, A., Kazmerski, L.L., 2013. A comprehensive review of the impact
of dust on the use of solar energy: history, investigations, results, literature, and
mitigation approaches. Renew. Sustain. Energy Rev. 22, 698–733.
Sayyah, A., Horenstein, M.N., Mazumder, M.K., 2014. Energy yield loss caused by dust
deposition on photovoltaic panels. Sol. Energy 107, 576–604.
Sharma, R., Wyatt, C.A., Zhang, Jing, et al., 2009. Experimental evaluation and analysis
of electrodynamic screen as dust mitigation technology for future Mars missions.
IEEE Trans. Ind. Appl. 45 (2), 591–596.
Shrimali, G., Jenner, S., 2013. The impact of state policy on deployment and cost of solar
photovoltaic technology in the US: a sector-specific empirical analysis. Renew.
Energy 60, 679–690.
Tanesab, J., Parlevliet, D., Whale, J., Urmee, T., Pryor, T., 2015. The contribution of dust
to performance degradation of PV modules in a temperate climate zone. Sol. Energy
120, 147–157.
The National Weather Service, 2014–15. The Yearbook of the National Center of Drought
and Crisis Management, The Ministry of Roads and Urban Development, 2014–15 (in
Persian).
Williams, R.B., Tanimoto, R., Simonyan, A., et al., 2007. Vibration characterization of
self-cleaning solar panels with piezoceramic actuation. In: Collection of Technical
Papers - 48th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and
Materials Conference, pp. 512–520.
Zandi, M., Bahrami, M., Eslami, S., Gavagsaz-Ghoachani, R., Payman, A., Phattanasak,
M., ... Pierfederici, S., 2017. Evaluation and comparison of economic policies to in-
crease distributed generation capacity in the Iranian household consumption sector
using photovoltaic systems and RETScreen software. Renewable Energy 107,
215–222.