Transparent Silicone–Epoxy Coatings with Enhanced Icephobic Properties for Photovoltaic Applications

Abstract

Recently, the photovoltaic technology has become very popular as a means to produce renewable energy. One of the problems that are still unsolved in this area of the industry is that photovoltaic panels are subject to a significant loss of efficiency due to the accumulation of dust and dirt. In addition, during the winter season, the accumulation of snow and ice also reduces or stops the energy production. The current methods of dealing with this problem are inefficient and pollute the environment. One way with high potential to prevent the build-up of dirt and ice is to use transparent coatings with self-cleaning and icephobic properties. In this work, the chemical modification of an epoxy–silicone hybrid resin using dually functionalized polysiloxanes was carried out. The icephobic properties (ice adhesion and freezing delay time of water droplets), hydrophobic properties (water contact angle, contact angle hysteresis, and roll-off angle), average surface roughness, and optical properties were characterized. It can be concluded that the performed chemical modification resulted in a significant improvement of the icephobic properties of the investigated coatings: ice adhesion decreased by 69%, and the freezing delay time increased by 17 times compared to those of the unmodified sample. The polysiloxanes also caused a significant reduction in the contact angle hysteresis and roll-off angle. The chemical modifications did not negatively affect the optical properties of the coatings, which is a key requirement for photovoltaic applications.

1. Introduction

Over the past few years, the world of science and industry has focused a great deal of attention on coatings as structural and functional materials. Transparent coatings are of particular interest because of their multitude of applications. These coatings can be used on windows, car windows, aircraft subcomponents, skyscraper roofs, glass facades, and photovoltaic modules [1,2,3].

The photovoltaic technology is one of the most popular and growing means for the production of renewable energy. The technology converts sunlight directly into electricity. Photovoltaics finds a variety of applications such as in solar farms [4], systems to supply power in space [5], remote locations [6], buildings [7], etc. In 2020, there was a record increase of 23% in power generation from solar photovoltaic compared to 2019, translating into a 156 TWh increase in power generation [8,9].

As the market grows, there are new requirements for photovoltaic installations to be increasingly efficient and reliable. However, the performance of photovoltaic panels is directly affected by external atmospheric factors such as dust, pollution, UV radiation, snow, ice, and frost. The accumulation of pollutants on the surface of photovoltaic modules in the form of dust or sediment can reduce their efficiency and power output by up to several percent points. It is estimated that the decrease in output for contaminants with a thickness of 1 µm can be about 10% and for dust with a thickness of 3 µm can be up to 25% [10]. In addition, during winter and in cold regions, snow and ice are other obstacles to maintaining the high efficiency of photovoltaic power generation. The accumulation of snow or frost on installations usually results in a complete halt of power production [11]. The current methods of removing snow from PV installations include manual methods causing energy consumption and the use of polluting de-icing chemicals. However, the chemicals contribute to environmental pollution, and the manual labor includes risk factors of personal injury and the risk of damaging the PV module surfaces with sundry tools [12].

In recent years, a proposed strategy with great potential to deal with surface icing is the use of passive icephobic coatings. In the literature, studies of icephobicity focus on the low ice adhesion strength, the longest possible freezing time for water droplets, and the prevention of ice accumulation. In addition to the aforementioned icephobicity, coatings for photovoltaic panels should also exhibit self-cleaning properties (high value of the water contact angle and low value of the roll-off angle) and durability of their properties over time, while maintaining the required optical properties (transmittance and reflectance) [13]. The reasons limiting the use of durable, transparent coatings with self-cleaning and ice-phobic properties in industrial practice are their high manufacturing cost, difficulties in production scalability, and susceptibility to various types of degradation such as thermal and UV damage [14,15]. None of the solutions currently known from the literature is fully effective in preventing the fouling and icing of photovoltaic panels over the long term.

Epoxy resins and their hybrids are among the most popular resins used. In the available literature, only a few works about transparent coatings based on epoxy resins that exhibit durable self-cleaning and icephobic properties can be found. In [16], an epoxy surface water contact angle of 175° was obtained, but the authors did not perform optical measurements of the investigated coatings. In 2017, Z. Yang et al. [17] presented very promising results on transparent epoxy coatings with superhydrophobic and self-cleaning properties. The work detailed the influence of various factors on the durability of the water contact angle. Unfortunately, the effect of the applied coating modifications on the optical and icephobic properties was not presented. Roppolo et al. [18] also presented transparent epoxy coatings, which can be used as coatings on photovoltaic modules due to the uncomplicated, low-cost, and easily scalable manufacturing method. The modification of these coatings involved the addition of a mica-based mineral filler. In this work, no studies were conducted to evaluate the anti-icing properties.

One of the main groups of materials used in the growing fields of icephobicity and hydrophobicity is that of polysiloxanes and their derivatives. These materials have low surface energy, high durability, and low elastic modulus [19]. All these properties provide polysiloxanes with a high potential as anti-icing hydrophobic materials. In many works, coatings exhibiting a low surface energy are being studied for their anti-icing properties [20,21,22,23]. In 2020, Kozera et al. [24] performed a modification of an unsaturated polyester resin with double-organofunctionalized polysiloxanes. This modification resulted in an increase in the water contact angle and a decrease in ice adhesion compared to the unmodified matrix. According to literature reports, the adhesion of ice to surfaces based on polysiloxanes can achieve a value even below 1 kPa [19]. Bessonov et al. [25] also proved that polysiloxanes increase a surface hydrophobicity.

Transparent coatings intended for photovoltaic panel applications should meet several requirements that mainly include icephobic properties, self-cleaning properties, and of course, optical properties. To date, no solution in the literature simultaneously provides all of these properties. The research presented in this work is a first step in the design of coatings for photovoltaic panel applications mainly focusing on obtaining ice-phobic properties. In this work, the modification of a transparent silicone–epoxy resin with original in-lab-synthesized chemical compounds from the group of doubly functionalized polysiloxanes was carried out. The chemical modifiers used in this work contain alkyl and oxirane groups in their structure to ensure compatibility with the selected polymer matrix and icephobic properties. The use of polysiloxanes containing two types of functional groups represents the novelty and uniqueness of this research. This work attempted to test the effect of individual functional groups on the properties of the matrix. The icephobic properties of the produced surfaces were characterized by determining two parameters: ice adhesion and the freezing delay time of water droplets. The dualistic approach in determining the anti-icing properties allowed for a more in-depth analysis of this phenomenon. The hydrophobic properties of the surface were characterized by its water contact angle measured at room temperature and at temperatures below 0 °C, contact angle hysteresis, and roll-off angle. Roughness tests and optical measurements were also included. The relationship between icephobic and hydrophobic properties is presented as well.

2. Materials and Methods

2.1. Coatings

The coating material used in the present study was a silicon–epoxy hybrid resin, SILIKOPON^® ED, from Evonik (Essen, Germany). SILIKOPON^® ED is characterized by good chemical and excellent corrosion resistance. It is suitable for ultra-high solids coatings. Applications include industrial and anti-corrosion coatings, uses in offshore/marine areas, structural steel, rail cars, tank construction, concrete walls and floors, and commercial transport coatings. This resin has a viscosity of 1500 mPas at 25 °C. SILIKOPON^® ED cures at ambient temperature in combination with amino silanes. The hardener used in this paper was Q-sil AMEO (3-aminopropyltriethoxysilane) from Oqema (Mönchengladbach, Germany). The mixing ratio SILIKOPON^® ED/Q-sil AMEO was 4.5:1.

The substances used in the modifier synthesis were silicon compounds (trime-thylsiloxy-terminated polymethylhydrosiloxanes (PMHS)) and olefins (hexene, octane, allyl-glycidyl ether) purchased from Linegal Chemicals (Warsaw, Poland); a solvent (toluene) from Avantor Performance Materials Poland S.A. (Gliwice, Poland); and chloroform-d, toluene-d8, and a Karstedt catalyst from Sigma Aldrich Poland, (Poznan, Poland). In the process, toluene was dried and purified with the MB SPS 800 Solvent Drying System and stored under an argon atmosphere in Rotaflo Schlenk flasks.

2.2. Synthesis of the Chemical Modifiers

Mixtures of allyl glycidyl ether (AGE) (0.168 mol) and hexene (HEX) (0.336 mol) in a molar ratio of 1:2 and of allyl glycidyl ether (0.168 mol) and octene (OCT) (0.336 mol) in a molar ratio of 1:2 were added to a solution of trimethylsiloxy-terminated polymethylhydrosiloxanes 992 (30 g, 0.504 mol) in toluene and allyl glycidyl ether (0.100 mol) and hexene (0.403 mol) in a molar ratio of 1:4, and of allyl glycidyl ether (0.100 mol) and octene (0.403 mol) in a molar ratio of 1:4. The mixture was constantly stirred and heated to 70 °C. Then, the Karstedt’s catalyst solution (10^−a eq Pt/mol SiH) was added. The reaction mixture was heated in reflux and stirred until the full conversion of Si–H (controlled by FT-IR). After confirming the complete conversion of the mixture, post-reaction evaporation was performed on a slow-speed vacuum evaporator.

2.3. Analysis of the Chemical Modifiers

Using Bruker Ascend 400 and Ultra Shield 300 spectrometers (both from Bruker Polska, Warszawa, Poland), the nuclear magnetic resonance (NMR) ¹H, ¹³C, and ²⁹Si spectra were recorded (at 25 °C using CDCl₃ as a solvent). Chemical shifts were reported in pm for 1H and 13C concerning the residual solvent (CHCl₃) peaks.

Fourier transform-infrared (FT-IR) spectra were obtained using a Nicolet iS 50 Fourier transform spectrophotometer (Thermo Fisher Scientific, Walsham, MA, USA) equipped with a diamond ATR unit with a resolution of 0.09 cm⁻¹.

2.4. Preparation of the Samples

Four types of in-house synthesized modifiers (MOD) were used. The chemical modifiers were added in the amounts of 2 wt.%. The composition of the prepared samples and the number of modifiers used are shown in

Table 1. Compositions of the prepared epoxy–silicone samples and their modifiers.

Sample No.	MOD Type	PHS	Olefin 1	Olefin 2	Molar Ratio
1 (REF)	-	-	-	-	-
2	MOD 1/2 wt.%	PHS992	AGE	OCT	1:2
3	MOD 2/2 wt.%	PHS992	AGE	OCT	1:4
4	MOD 3/2 wt.%	PHS992	AGE	HEX	1:2
5	MOD 4/2 wt.%	PHS992	AGE	HEX	1:4

. The procedure for preparing the mixtures was as follows. The chemical modifiers were added to the silicone–epoxy resin in appropriate amounts and mixed using a magnetic stirrer. Furthermore, the hardener Q-sil was added in a ratio of 1:4.5 to the resin.

The coatings were manufactured using the spin coating method. The spin coater SPIN150i from POLOS was used. The transparent coatings were applied on glass substrates with the size of 30 mm × 30 mm. The glasses were rinsed with acetone before applying the resin. The spin coating process was divided into four stages: dispensing, spreading, edge bead removal (EBR), and drying. The parameters of each stage are shown in

Table 2. The parameters of the spin coating steps.

Stage	Spin Speed [rpm]	Spin Accel. [rpm/s]	Spin Time [s]
Dispense	100	1000	10
Spread	2000	1000	20
EBR	500	1000	10
Dry	4000	1000	40

. The applied polymer coatings were cured at room temperature for 24 h.

2.5. Determination of the Optical Properties

The optical properties of samples were characterized using a Cary 7000 UV-Vis-NIR Universal Measurement Spectrophotometer (Agilent Inc., Santa Clara, CA, USA). The samples were measured at room temperature in the 350–2500 nm spectral range with an increment of 4 nm. Transmittance and reflectance were measured using an integrating sphere module. The absorptance was calculated assuming the sum of transmittance, reflectance, and absorptance was 100%.

2.6. Determination of Roughness

The roughness tests were performed using a non-contact 3D surface profiler Slynx Sensofar from Sensofar Metrology (Barcelona, Spain). The Ra parameter, i.e., the average roughness of the surface profile was determined. The final values are the average of five different measuring points on the surfaces.

2.7. Determination of Hydrophobicity

The wettability of the surfaces was determined by measuring the water contact angles (WCA), the water contact angles at reduced temperatures, the contact angles hysteresis (CAH), and the roll-off angles (RoA). The measurements were performed using an OCA15 goniometer from DataPhysics Instruments (Filderstadt, Germany) with software SCA20, equipped with a chamber designed for testing at reduced temperatures. The volume of the water droplets for the WCA and CAH measurements was 5 μL, and that for the RoA measurements was 10 μL. The contact angle hysteresis was determined using the needle-in method. The WCA was measured at room temperature and at temperatures of 0 °C, −5 °C, −10 °C, and −15 °C. The final WCA and RoA values are the average of five different measuring points on the surfaces.

2.8. Determination of Freezing Delay Time (FDT)

The freezing delay time (FDT) of water droplets, which is one of the parameters describing the icephobicity of a surface, was determined using an OCA15 goniometer (DataPhysics Instruments, Germany) with software SCA20, equipped with a chamber designed for testing at reduced temperatures. The test was conducted at −15 °C and lasted a maximum of 4 h. The volume of the water droplets was 5 μL.

2.9. Determination of Ice Adhesion Strength (IA)

The ice adhesion (IA) was determined using the universal testing machine Zwick/Roel Z050 (from Zwick Roell, Ulm, Germany). During the measurement, the shear strength between the ice and the surface of the sample was determined. The tests of ice adhesion were carried out at room temperature, at a relative humidity of 50%. A detailed description of this method was presented in our earlier work [24].

3. Results

3.1. Characterization of the Polysiloxanes Derivatives

Functionalized polysiloxanes were obtained by hydrosilylation with hexene, octene, and allyl glycidyl ether. The reactions were carried out in the presence of a platinum Karstedt’s catalyst. To confirm the full conversion of the substrates, NMR and FT-IR analyses were performed, and the disappearance of the characteristic signals at 2141 and 889 cm⁻¹ was observed. This disappearance is caused by the stretching and bending of the Si-H group, respectively. Based on the ¹H NMR analysis, the conversion for all compounds was >99%. The structure and purity of the modifiers were also confirmed by ¹H NMR, ¹³C NMR, and ²⁹Si NMR analyses.

Figure 1, Figure 2, Figure 3 and Figure 4 show the schemes of the product structures confirmed by NMR. The following signals were assigned:

Figure 1. The scheme of the structure of the MOD1 modifier (MOD1—poly((methyloctylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

Figure 1. The scheme of the structure of the MOD1 modifier (MOD1—poly((methyloctylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

¹H NMR (400 MHz, CDCl₃): δ (ppm) = 3.70–3.67 (m, position 3), 3.47–3.41 (m, positions 3 and 4), 3.14 (m, position 2), 2.79–2.78 (m, position 1), 2.61–2.60 (m, position 1), 1.66–1.60 (m, position 5), 1.34–1.27 (m, octyl -CH₂-), 0.90–0.88 (m, octyl -CH₃), 0.53–0.50 (m, SiCH₂-), 0.15–0.05 (SiMe, SiMe₃)

¹³C NMR (101 MHz, CDCl₃): δ (ppm) = 74.35, 72.16, 71.53, 69.71, 50.96, 44.43 (AGE), 33.65, 33.58, 32.13, 29.58, 29.51 (OCT), 23.38, 23.23 23.11, (AGE), 22.85, 17.83, 17.55, 14.24 (OCT), 13.66 (AGE), 2.00, 1.61, −0.18, −0.29, −0.54 (SiMe, SiMe₃);

²⁹Si NMR (79.5 MHz, CDCl₃): δ (ppm) = −21.24–(−22.95) (SiMe, SiMe₃).

Figure 2. The scheme of the structure of the MOD2 modifier (MOD2—poly((methyloctylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

Figure 2. The scheme of the structure of the MOD2 modifier (MOD2—poly((methyloctylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

¹H NMR (400 MHz, CDCl₃): δ (ppm) = 3.68–3.65 (m, position 3), 3.48–3.38 (m, positions 3 and 4), 3.13 (m, position 2), 2.78–2.76 (m, position 1), 2.60–2.58 (m, position 1), 1.67–1.59 (m, position 5), 1.34–1.27 (m, octyl -CH₂-), 0.90–0.86 (m, octyl -CH₃), 0.51–0.48 (m, SiCH₂-), 0.13–0.03 (SiMe, SiMe₃)

¹³C NMR (101 MHz, CDCl₃): δ (ppm) = 74.42, 72.17, 71.53, 50.96, 44.45 (AGE), 33.69, 33.52, 32.16, 29.56, 29.47 (OCT), 23.40, 23.26, 23.12, (AGE), 22.87, 17.86, 17.58, 14.25 (OCT), 13.68 (AGE), 2.01, 1.62, −0.16, −0.19, −0.27, −0.53 (SiMe, SiMe₃);

²⁹Si NMR (79.5 MHz, CDCl₃): δ (ppm) = −21.30–(−22.96) (SiMe, SiMe₃).

Figure 3. The scheme of the structure of the MOD3 modifier (MOD3—poly((hexylmethylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

Figure 3. The scheme of the structure of the MOD3 modifier (MOD3—poly((hexylmethylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

¹H NMR (400 MHz, CDCl₃): δ (ppm) = 3.68–3.65 (m, position 3), 3.47–3.34 (m, positions 3 and 4), 3.13–3.11 (m, position 2), 2.78–2.76 (m, position 1), 2.59–2.58 (m, position 1), 1.64–1.60 (m, position 5), 1.33–1.26 (m, hexyl -CH₂-), 0.88–0.86 (m, hexyl -CH₃), 0.51–0.48 (m, SiCH₂-), 0.13–0.03 (s, SiMe, SiMe₃)

¹³C NMR (101 MHz, CDCl₃): δ (ppm) = 74.31, 71.53, 50.97, 44.44 (AGE), 33.65, 33.57, 32.12, 29.58, 29.51, (HEX) 23.37, 23.23, 23.18, 23.10 (AGE), 22.85, 17.82, 17.66, 17.55, 14.25 (HEX) 13.66, 13.44 (AGE), 2.01, 1.61, −0.19, −0.29, −0.46 (SiMe, SiMe₃)

²⁹Si NMR (79.5 MHz, CDCl₃): δ (ppm) = −20.99–(−21.26) (SiMe, SiMe₃).

Figure 4. The scheme of the structure of the MOD4 modifier (MOD4—poly((hexylmethylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

Figure 4. The scheme of the structure of the MOD4 modifier (MOD4—poly((hexylmethylsiloxane)-co-(3-glycidoxypropyl)(methyl)siloxane), trimethoxysilyl-terminated).

¹H NMR (400 MHz, CDCl₃): δ (ppm) = 3.70–3.67 (m, position 3), 3.51–3.40 (m, positions 3 and 4), 3.15–3.13 (m, position 2), 2.80–2.78 (m, position 1), 2.62–2.60 (m, position 1), 1.68–1.62 (m, position 5), 1.34–1.30 (m, hexyl -CH₂-), 0.92–0.89 (m, hexyl -CH₃), 0.54–0.51 (m, SiCH₂-), 0.11–0.07 (s, SiMe, SiMe₃);

¹³C NMR (101 MHz, CDCl₃): δ (ppm) = 74.35 (position 4), 71.52 (position 3), 50.95 (position 2), 44.41 (position 1), 33.29, 33.22, 31.81, 23.18, 23.06, 22.79, 17.84, 17.77, 17.57, 14.25 (-CH₂-), 1.97, 1.61, −0.20, −0.31, −0.47 (SiMe, SiMe₃).

²⁹Si NMR (79.5 MHz, CDCl₃): δ (ppm) = −21.28–(−21.36) (SiMe, SiMe₃).

3.2. Optical Properties

The measured optical properties of the fabricated coatings and neat glass are presented in Figure 5. The transmittance of the MOD1, MOD2, MOD3, and MOD4 samples was similar to the transmittance of glass in the tested range (Figure 5a). The differences in values were about 2%. The reference sample had a lower transmittance than the other samples. The reflectance of the measured samples was similar and ranged from 6% to 9% over the tested range (Figure 5b). The reference sample had a broad absorptance band of about 1000 nm. The other measured samples had an absorptance not greater than 4% within the tested range (Figure 5c).

3.3. Roughness of the Surface

Table 3. The hydrophobicity and roughness parameters of the silicone–epoxy coatings.

Sample No.	MOD Type	Core of MOD	Olefin 1	Olefin 2	Molar Ratio	Ra [nm]	WCA [°]	CAH [°]	RoA [°]
1 (REF)		-	-	-	-	2 ± 0.1	80 ± 1	22 ± 2	75 ± 8
2	MOD1	PWS992	AGE	OCT	01:02	12 ± 1.1	84 ± 1	9 ± 1	50 ± 3
3	MOD2	PWS992	AGE	OCT	01:04	21 + 1.9	89 ± 1	7 ± 2	55 ± 9
4	MOD3	PWS992	AGE	HEX	01:02	5 + 0.5	88 ± 2	9 ± 1	35 ± 6
5	MOD4	PWS992	AGE	HEX	01:04	7 + 0.4	90 ± 1	11 ± 2	32 ± 14

presents the values of the Ra parameter, which indicates the average roughness of the surface profile. The reference sample obtained a value of 2 ± 0.1 nm. Figure 6a shows a visualization of the surface of the reference sample acquired with the profilometer. Analyzing the obtained results, it can be seen that the chemical modification with polysiloxanes increased the surface roughness of the silicone–epoxy samples. The range of the Ra values for the modified samples was from 5 nm to 21 nm. The highest increase in the Ra values was recorded for the MOD2-modified sample (Figure 6b). Furthermore, it can be seen that the polysiloxanes containing the OCT functional group displayed higher Ra parameter values compared to the polysiloxanes containing the HEX functional group. In addition, a higher amount of OCT/HEX functional groups also resulted in higher Ra parameter values compared to a lower amount of these groups.

3.4. Hydrophobic Properties

The surface hydrophobicity of the produced samples was determined by measuring three parameters: water contact angle (WCA), contact angle hysteresis (CAH), and roll-off angle (RoA). The values of these parameters are presented in Table 3.

The reference sample, an unmodified epoxy–silicone resin, obtained a WCA of 80 ± 1° and hence may be considered hydrophilic (WCA < 90°). Analyzing the WCA values of the coatings after the chemical modification, it can be concluded that the double functionalization of polysiloxanes caused an increase in the WCA values. Nevertheless, the increase was not significant. The range of the WCA values for the modified samples was from 84° to 90°. Taking into account the measurement uncertainty, it can be concluded that the samples modified with the MOD2, MOD3, and MOD4 additives presented surfaces at the boundary between hydrophilicity and hydrophobicity, obtaining WCA values close to 90°. The lowest increase in WCA compared to the reference sample was recorded for the sample modified with MOD1. Figure 7 shows a picture of a water droplet applied on the surface of a MOD4-modified sample, which achieved the highest WCA value.

The unmodified sample achieved a CAH value of 22 ± 2°. The chemical modifications caused a significant decrease in this parameter. All modified samples obtained similar results, and the CAH range was from 7° to 11°. The lowest value was recorded for the sample modified with MOD2 polysiloxane, for which CAH decreased by 59% compared to that of the reference coating.

The RoA parameter also significantly decreased after modification with polysiloxanes. The reference sample obtained an RoA of 75 ± 8°, and the modified samples achieved RoA values in the range from 32° to 55°. The lowest RoA values were obtained for MOD3- and MOD4-modified samples, with a reduction of about 55% compared to the RoA of the reference sample.

In summary, double-functionalized polysiloxanes showed increased water contact angle values (despite not crossing the boundary between hydrophobicity and hydrophilicity) and a significant reduction in contact angle hysteresis and roll-off angle values for all types of modifications.

Water Contact Angle at Temperatures Below 0 °C

Figure 8 shows the values of WCA measured at four temperatures below 0 °C to determine how a temperature reduction affected the surface hydrophobicity.

Analyzing the results, it can be seen that the reference sample as well as all modified samples recorded a decrease in the WCA values at reduced temperatures compared to the values determined at room temperature. In addition, for some of the modifications, a relationship can be observed, i.e., the lower the test temperature, the lower the WCA values. The highest reduction in WCA, i.e., a reduction in the hydrophobic properties of the surface, was recorded for the sample modified with MOD3; at the temperatures of −10 °C and −15 °C, the WCA value decreased by 27% compared to the WCA value at room temperature. On the other hand, the highest hydrophobic stability at low temperatures was shown by the sample modified with polysiloxane MOD1, obtaining only a decrease in WCA equal to about 5% compared to the room temperature value. Moreover, for this sample, no further decrease in WCA was observed with the decreasing temperature, the WCA values determined at temperatures from 0 °C to −15 °C being either the same or close to each other. The reference sample also experienced a low decrease in the WCA values at reduced temperatures, as the reduction was equal to 8% compared to the WCA determined at room temperature. Other samples modified with the MOD2 and MOD4 additives showed a similar decrease in WCA of about 10% compared to the values determined at room temperature. Fu et al. [26] also observed that the wettability changes with ambient temperature and that the higher the WCA at ambient temperature, the higher the decrease in the value of this parameter at reduced temperatures. In addition, He et al. [27] found that for a modified superhydrophobic surface at −10 °C, the WCA value would decrease by about 8° compared to the value at room temperature. Moreover, some reports proposed that the surface free energy increases at low temperatures, which is related to the decrease in WCA [28,29].

3.5. Icephobic Properties

Figure 9 presents the obtained values of ice adhesion and freezing delay time for the investigated coatings. These two parameters allowed determining the anti-icing behavior of the epoxy–silicone samples.

3.5.1. Ice Adhesion Strength

The ice adhesion determines the force needed for ice to break away from the surface. The lower the ice adhesion value, the more the surface is icephobic. The reference sample without chemical modification achieved the IA value of 178 ± 33 kPa. The samples modified with polysiloxanes obtained IA values in the range from 55 kPa to 118 kPa; so, all chemical modifications caused a decrease in the value of this parameter. The MOD4 sample showed the highest decrease, as the reduction in the IA values was as much as 69% compared to the IA of the reference sample. The MOD3 sample also achieved a very large decrease in IA values similar to the MOD4 sample, as the reduction was equal to 62% compared to the IA of the unmodified coating. The other modified MOD1 and MOD2 samples also recorded positive results, as the IA values for these samples were reduced by 34% and 35%, respectively, compared to the reference value. Another important aspect is that the MOD3 and MOD4 samples recorded IA values below 100 kPa. It is considered that surfaces exhibit low ice adhesion when the detachment force for ice is lower than 100 kPa. From such surfaces, ice is detached under the influence of natural forces such as wind, gravity, or ambient vibration [30]. The effectiveness of improving the anti-icing properties by using the chemical modification of the polymer matrix with double-functionalized polysiloxanes was proven in a work [31], achieving an IA reduction of 51% compared to the IA of the unmodified epoxy resin.

3.5.2. Freezing Delay Time

The FDT recorded for the reference sample was 3 min. The other modified coatings showed values for this parameter from 14 min to 50 min. Thus, it can be concluded that the addition of functionalized polysiloxanes increased the FDT and consequently improved the anti-icing properties of the surface of the epoxy–silicone coatings. The longest FDT was recorded for the sample MOD4, which also obtained the lowest ice adhesion value. The FDT for this sample increased 17 times compared to that of the unmodified sample; the FDT for the MOD 4 sample was 50 min. The other samples also showed a significant increase in the FDT, as the time values for these samples increased 5, 6, and 8 times. A previous work [32] also observed an increase in the FDT after modification with polysiloxanes of a polymer resin.

3.5.3. Summary of the Icephobic Properties

Comparing the results of the ice adhesion and the freezing delay time of water droplets, a direct correlation between these parameters can be seen. The lower the ice adhesion value the sample showed, the higher the value of the freezing delay time. Among the samples tested, the sample MOD4 showed the highest icephobic properties. This sample displayed the lowest ice adhesion and the longest freezing delay time for water droplets.

4. Discussion

4.1. Influence of the Chemical Groups on Wettability and Icephobicity

Modifications with polysiloxanes functionalized with the AGE and HEX functional groups yielded the best icephobic results (the lowest ice adhesion value and the highest freezing delay time of water droplets). The modifications with polysiloxanes functionalized with the AGE and OCT functional groups achieved inferior results compared to those obtained using the AGE and HEX groups, but also achieved a significant improvement in the anti-icing properties compared to the unmodified sample. In addition, the higher content of the HEX functional group relative to the AGE functional group (i.e., 4:1) made it possible to obtain a significantly higher freezing delay time of water droplets and a slightly lower ice adhesion value compared to samples with a lower content of the HEX functional group relative to the AGE group (i.e., 2:1). In the case of modifications with polysiloxanes with the OCT functional group, the above relationship was not observed, and the values of the parameters of the icephobic properties were similar to each other. In summary, the polysiloxanes functionalized with the HEX functional group provided better icephobic properties to the epoxy–silicone surface compared to the polysiloxanes functionalized with the OCT functional group. In addition, the sample with the higher content of HEX functional groups recorded the best icephobic properties.

Regarding the hydrophobic properties, it can be said that the polysiloxanes functionalized with the AGE and HEX functional groups were characterized by higher WCA values and lower RoA values compared to analogous polysiloxanes functionalized with the AGE and OCT functional groups. It should be added that the WCA values were not diametrically different from each other, as the three modifications yielded very similar WCA values; only the sample modified with polysiloxane functionalized with the AGE and OCT groups in a ratio of 1:2 obtained a lower WCA value of 84°. It should be added that this value is not diametrically different from the other values of the modified samples, which also did not cross the boundary between hydrophobicity and hydrophilicity. As for the RoA values, the differences between the above polysiloxanes were significant. Comparing the WCA and RoA values of samples modified with polysiloxanes functionalized with the same functional groups but in different molar ratios, it can be concluded that they were very close to each other. The CAH values of all modified samples were in a similar range. In summary, the polysiloxanes functionalized with the HEX functional group provided better hydrophobic properties to the epoxy–silicone surface compared to the polysiloxanes functionalized with the OCT functional group.

Among the samples tested, the highest stability of the WCA values at reduced temperature was recorded for the sample modified with polysiloxane functionalized with the AGE and OCT groups in a ratio of 1:2, while the lowest stability was recorded for the sample modified with polysiloxane functionalized with the AGE and HEX groups in a ratio of 1:2. Polysiloxanes with higher ratios of the AGE functional group to the OCT/HEX groups showed very similar relationships. Thus, it can be concluded that polysiloxanes functionalized with the OCT functional group lead to a better stability of the WCA values of silicone–epoxy surfaces compared to polysiloxanes functionalized with the HEX functional group.

4.2. Relationship between Wettability and Icephobicity

In many reports in the literature, hydrophobic and icephobic properties are related [33,34,35,36,37]. Figure 10 shows the relationship between ice adhesion and the hydrophobic parameters water contact angle and roll-off angle. The values of contact angle hysteresis were similar for all chemical modifications; so, they will not be collated in the following discussion.

Analyzing the relationship between IA and WCA, it can be concluded that as the value of WCA increased, the value of IA decreased. This was related to the change in surface free energy. In previous works [26,38], it was proven that this relationship has a strong influence on the icephobic properties. Only the MOD2-modified sample showed a deviation from this relationship, but it also recorded a significant decrease in the IA values compared to the reference sample. This sample obtained a very similar IA value as the MOD1-modified sample, despite obtaining a higher WCA value. This was likely influenced by the surface roughness. The MOD2-modified sample recorded twice the value of the Ra parameter, which may have offset the effect of the increased water contact angle, i.e., reduced surface energy. It was proven in many works that an increased roughness can significantly affect ice adhesion through the formation of a mechanical connection between the emerging ice and the rough substrate (ice anchoring) [37,38,39]. The second relationship between the icephobic and hydrophobic parameters was that as the value of RoA decreased, the value of IA also decreased.

The lowest possible values of IA and RoA and the highest value of WCA are key features that coatings intended for photovoltaic panels, among others, must present. The respective values of these parameters guarantee the surface’s icephobicity. In further studies, the authors of this work will focus on surface texturization to obtain higher values of the WCA and lower values of the RoA parameters to achieve self-cleaning properties for the investigated coatings.

5. Conclusions

• The silicone–epoxy coatings produced in the entire tested spectral range showed similar optical properties (T, R, A) to those of glass. They can be potentially used in photovoltaic panels.
• The chemical modification with polysiloxanes resulted in an increase in surface roughness compared to the unmodified coating.
• The chemical modification with polysiloxanes contributed to an increase in hydrophobicity, with most samples achieving WCA values close to 90°.
• The use of polysiloxanes resulted in a significant reduction in the CAH and RoA values compared to those of the unmodified coating.
• The WCA values at reduced temperatures decreased for all applied modifications compared to the WCA values determined at room temperature. In addition, it was observed that the lower the ambient temperature, the lower the WCA value.
• Double-functionalized polysiloxanes in the epoxy–silicone matrix led to a significant improvement in the anti-icing properties, a reduction in ice adhesion, and an increased freezing delay time of water droplets in comparison to the unmodified silicone–epoxy resin.
• The synergistic effect of the AGE and HEX functional groups at the polysiloxane core yielded better icephobic and hydrophobic results compared to those obtained with the AGE and OCT functional groups when using a silicone–epoxy matrix.
• Two relationships were observed for the modifications used. As the WCA values increased and as the RoA values decreased, the IA values were reduced.