A Spray-Deposited Modified Silica Film on Selective Coatings for Low-Cost Solar Collectors

Abstract

Solar collectors represent an attractive green technology for water heating, where sunlight is efficiently absorbed by a selective coating and the generated heat is transferred to water. In this work, the improvement and scale-up of an electrodeposited black nickel selective coating with a modified silica (MS) film deposited by spray pyrolysis are reported. The MS material was prepared by the sol–gel method using tetraethyl orthosilicate with the addition of n-propyl triethoxysilane to obtain a porous film with an adequate refractive index and enhanced flexibility. The reflectance of electrodeposited selective coatings was characterized with and without the MS film and compared to a commercially available coating of black paint. The MS film increased the solar absorptance from 89% to 93% while maintaining a much lower thermal emittance than the painted coating. The reflectance of the MS film remained unchanged after prolonged thermal treatment at 200 °C (200 h). The fabrication process was scaled up to 193 cm × 12 cm copper fins, which were incorporated in commercial-size flat-plate solar collectors. Three complete collectors of an area of 1.7 m² were fabricated and their performance was evaluated under outdoor conditions. The results show that the electrodeposited selective coating with the MS film outperformed both the commercial black paint system and the system without the modified silica film.

1. Introduction

Solar-to-thermal energy conversion systems have diverse applications and are crucial for the energy transition. They can provide temperatures ranging from 30 °C to 200 °C for domestic water heaters and industrial applications and high temperatures (400–1000 °C) for electricity generation using concentrated solar power [1,2,3]. The efficiency of a thermosolar collector is boosted by the effectiveness of the selective coating used [4,5]. A typical structure of a selective coating includes a glass or metallic substrate coated, if needed, with a metallic layer to prevent heat loss through infrared radiation. This is typically followed by the application of a solar absorber film consisting of metal oxides, cermets, or multilayered materials [6,7]. The absorber must permit the transmission of mid-long infrared (IR) radiation to reduce the thermal emittance [8]. For a selective coating to be industrially viable, it must be low-cost and thermally stable and the fabrication method must be scalable [9]. Electrodeposition and spray pyrolysis represent two attractive techniques because of their low cost, easy scalability, and low energy consumption [10]. In addition, both techniques are sufficiently robust such that small fluctuations in the experimental conditions do not affect the properties of the scaled-up coatings.

Electrodeposition is a method for preparing surface coatings that has made significant advances towards becoming more eco-friendly. A specifically designed black paint is generally used as the solar absorber for copper fin-based solar collectors, but its selectivity is low. An alternative solar absorber is black nickel (BN), deposited via an electroless method, electrodeposition, spray pyrolysis, or chemical conversion [11,12,13,14]. Electrodeposited BN obtained from a chloride-based plating bath presents good optical properties, which are also stable under exposure to moisture [15]. Important features of the BN electrodeposition process are the low cost of precursor salts and the easy maintenance of the bath.

A silica coating (SC) can be applied on top of the selective coating, serving purposes such as increasing solar absorptance and improving chemical and thermal stability [16]. The SC prevents or lowers the rate of oxidation of the metallic particles in the solar absorber in the case of cermet-type materials. BN is a cermet-type material, and a functional SC may increase the solar absorptance and collector performance. Hence, in this work, we integrated silica-based top films in our selective coating [17]. Silica is a well-known antireflective (AR) material with an adequate refractive index; films can be deposited using several methods, such as sputtering, spray pyrolysis, and spin-coating [18,19,20,21,22,23,24,25]. Several studies show that silica AR coatings increase the spectral selectivity of the coatings, while films obtained by sputtering have been reported to increase the solar absorptance specifically [26,27,28,29,30,31,32,33]. There are reports on the effect of a silica film overlayer on the thermal stability of black nickel-based coatings but only a few reports on scale-up techniques to obtain the AR coating. To our knowledge, there are no reports on full-size prototypes detailing the scale-up of electrodeposited BN films covered with SC by spray pyrolysis.

Silica is often obtained using tetraethyl orthosilicate (TEOS) as a sol–gel precursor but, often, the material is not sufficiently flexible [34]. To make silica more flexible, a propyl group can be incorporated into the structure; the resulting material is called modified silica (MS). It has been demonstrated for electrodes in super-capacitors and Li-ion intercalation batteries that the coating flexibility and hydrophobicity can be improved by introducing short-chain organic moieties into the silica sol–gel precursor [35,36]. The propyl group is attractive because of its good thermal stability and hydrophobicity; additionally, it has been reported that it helps the formation of mesopores in the synthesis of titanium silicalite [37]. In the alkoxide precursor TEOS, all four ethoxy groups can hydrolyze (considering complete hydrolysis) and then condense, creating a 3-dimensional oxide network with ethoxy and hydroxyl side groups in the structure if condensation is not complete. When one ethoxy group in the precursor is replaced by an organic group that cannot hydrolyze such as n-propyl, the process of condensation still occurs, but the process and the resulting structures are modified. The stability of MS with ethoxy and n-propyl side groups to UV is high, as reported by J. Feenstra [38], because the organic moieties are not in the backbone of polysiloxane chains. G. Zimmermann [39] demonstrated that UV radiation below 300 nm can rapidly break C-C or C-Si bonds for larger organic groups with more than three carbons, but, for the solar spectrum at sea level, the UV component below 300 nm is almost zero; hence, the organic bond cleavage is not possible. Following this approach, in this work, we used n-propyl triethoxysilane (PTES) as a short-chain organic co-precursor in the TEOS-based sol–gel solution for the preparation of the MS coating. The films were deposited using spray pyrolysis as a cost-effective and scalable method, in line with the requirements for the commercialization of low-temperature, flat-plate solar collectors. Considering that the cover glass used in solar collectors can absorb UV radiation of a wavelength up to 320 nm, thus helping the long-term stability of organic moieties, and that the BN/MS surface needs to be isolated from environmental conditions to extend its durability, we decided to deposit the MS directly on the fins and not on the covering glass plate of the solar collector.

In this work, we describe the deposition of the selective coating, consisting of a metallic nickel (Ni) underlayer and BN absorber layer, both with and without the MS overlayer, on 16 cm² substrates. We provide a detailed characterization of the optical properties and thermal stability. In addition, we scaled up the system and here report the fabrication of three complete solar collector prototypes with a total area of 1.72 m² for three different systems: (i) electrodeposited nickel/black nickel selective coating; (ii) electrodeposited nickel/black nickel selective coating with MS overlayer; and (iii) commercial, black paint coating (CBP). The collectors were evaluated under outdoor conditions and the performance of the three systems was compared, highlighting the effect of the MS film.

2. Materials and Methods

Material Fabrication: A potentiostat–galvanostat Gamry Reference 3000 coupled to a Reference 30k booster was employed for galvanostatic electrodeposition. The Cu sheets and fins were washed with detergent and sanded. After the sanding process, the samples were rinsed with deionized water and acetone in an ultrasonic bath. The bright Ni film was electrodeposited from a bath composed of (0.45 M) NiSO₄·6H₂O, (0.32 M) KCl, and (0.32 M) H₃BO₃ at pH 4.5; the current density was −5 mA cm⁻² for 585 s. The average film thickness was 1.0 ± 0.2 mm (see Supporting Information, Figure S1, for thickness measurement details). Later, the samples were rinsed and dried and the BN layer was electrodeposited. The BN layer was electrodeposited from a plating bath with (0.31 M) NiCl₂ and (0.40 M) KCl at pH 6.5, using a two-pulse method of 60 s at −2.6 mA cm⁻², followed by 90 s at −1.4 mA cm⁻²; the average film thickness was 280 ± 50 nm (see Supporting Information, Figure S2). A three-electrode configuration was employed, and rectangular Cu sheet sections of 4 × 6 cm² with a 1 mm thickness were used as substrates; the electrodeposition area was 16 cm². For Cu sheets, the experimental setup can be seen in Figure S3a). Note that both deposition steps corresponded to reduction processes; hence, by convention, a negative current was applied to the Cu substrate. For fins, a two-electrode configuration was employed. The working electrode (cathode) consisted of fins with an active area of 2316 cm² on the front face. For the counter electrode, 11 titanium baskets filled with high-purity nickel plating rounds (99.95%) were used; the baskets were interconnected in series using a copper tube. The plating cell was a 2 m × 36 cm × 30 cm metallic container covered with epoxy paint. The distance between the fin and the baskets was 2 cm and the system was not agitated. A schematic representation of the experimental setup for Cu fins can be seen in the Supporting Information, Figure S3b). To compare the coatings, we prepared samples with commercial black paint that was spray-deposited directly onto the Cu substrate at 10 cm and 3.51 kg cm⁻² to obtain a thickness of about 40.0 ± 3 µm (see Supporting Information, Figure S4). The silica and MS films were deposited on top of the selective coating from a solution prepared by adding 10 mL tetraethyl orthosilicate (TEOS) to 100 mL ethanol at room temperature with agitation at 300 rpm. Then, 2.7 mL of an aqueous HCl solution at pH 2.0 was added slowly to this precursor solution. The molar ratio [H₂O]/[TEOS] was 3.4 to avoid complete TEOS hydrolysis. For preparing the MS, 1 mL n-propyl triethoxysilane (PTES) was added to 10 mL TEOS before adding ethanol. The silica and MS coatings were obtained using spray pyrolysis after 2 h of the solution preparation. The solution was sprayed onto the selective coating or glass slide substrates from a 20 cm distance at a pressure of 3.51 kg cm⁻². The coated substrate was then heated to 200 °C on a hot plate. Two layers were applied to obtain the desired thickness, with a waiting time of 5 min between applications.

The thermal stability of the samples was evaluated using thermal treatments in a furnace (Thermo Scientific, Waltham, MA, USA, model FB1315M) in air, which consisted of one heating cycle for Cu samples at 200 °C for 200 h; the applied heating ramp was 10 °C min⁻¹. The samples are represented by the following notations: Substrate/Metallic interlayer/Absorber layer/SC or MS coating; the thermal treatment is indicated by adding (TT).

Characterization: The electrodeposited films were studied by field-emission scanning electron microscopy (FE-SEM; JEOL JSM-7600F, Tokyo, Japan). X-ray diffraction on the selective coatings was performed using a Siemens D-5000, while a Bruker D8-Advance (Billerica, MA, USA) with Bragg–Brentano geometry was employed to characterize SC and MS powders; both systems used monochromatic Cu-Kα radiation (λ = 1.5418 Å). The film thickness was measured by profilometry using a KLA-Tencor D-120, Chandler, AZ, USA. Attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR) was performed using a Thermo Scientific Nicolet iS5 FTIR spectrometer. Contact angle measurements were performed using 10 μL deionized water droplets, and the angle was calculated from the photograph using Image J software. Atomic force microscopy (AFM) images were obtained using WITec Alpha 300 equipment (Oxford Instruments, Abingdon, UK). X-ray photoelectron spectroscopy (XPS) was performed using K-Alpha Surface Analysis, Thermo Scientific, equipped with an Al Kα X-ray source. Depth profile XPS analysis was performed by eroding the samples with Ar in 20 s steps before each XPS measurement.

Optical properties: The total reflectance spectra of the selective coatings were obtained using the spectrophotometer UV–Vis Avantes (Avaspec 2048) and the infrared spectrometer Avantes (Avaspec-Nir 256-2.5), both equipped with the integrating sphere Avantes 50-Ls-HAL. For the UV–Vis–NIR measurements, the Avantes WS-2 white reflective tile was used as the reference. The reflectance spectra obtained in the range of 2.5 to 15 μm were measured using the FTIR spectrometer Perkin Elmer, Waltham, MA, USA (frontier NIR/MIR) equipped with an integrating sphere (PICO, Integrat IR model). A gold standard was used as the reference. All reflectance spectra were obtained at room temperature.

The solar absorptance (α) was defined as the fraction of the solar power absorbed by the selective coating and was calculated by weighting the measured wavelength-dependent reflectance [R(λ)] spectrum of the coatings against the solar radiation spectrum [ISUN(λ)] (ASTM G173-0329), according to Equation (1) [40]. The thermal emittance (ε) was defined by Equation (2) [41] as the fraction of emitted power, as given by the black body radiation spectrum, and was also calculated using the reflectance spectra:

$$\alpha ={\displaystyle \frac{{\int }_{0.3}^{2.5}{I}_{SUN}\left(\lambda \right)\times \left[1-R\left(\lambda \right)\right]d\lambda }{{\int }_{0.3}^{2.5}{I}_{SUN}\left(\lambda \right)d\lambda }}$$

(1)

$$\epsilon ={\displaystyle \frac{{\int }_{2.5}^{15}{I}_{BB}\left(\lambda \right)\times \left[1-R\left(\lambda \right)\right]d\lambda }{{\int }_{2.5}^{15}{I}_{BB}\left(\lambda \right)d\lambda }}$$

(2)

where the black body radiation intensity as a function of wavelength and temperature is given by $I_{BB}\left(\lambda ,T\right)=c_1/\left\{{\lambda }^5\left[e^{\left({\displaystyle \frac{c_2}{\lambda T}}\right)}-1\right]\right\}$ with c₁ = 3.743 × 10⁻¹⁶ W m² and c₂ = 1.4387 × 10⁻² m K. In this work, the reflectance spectra were measured at room temperature, but the thermal emittance was calculated using T = 100 °C in Equation (2) in order to estimate the performance of the coating at a typical, maximum operating temperature of a solar collector for domestic water heating. The spectral selectivity (h) could be used as an indicator of the performance of the coatings and was calculated using h = α − 0.5ε. The optical transmittance of SC and MS films deposited on soda lime glass substrates was examined using an Agilent 8453 spectrophotometer from 190 to 1100 nm. The transmittance of the soda lime glass substrate was used as the baseline.

Collector Performance Characterization: The collectors were experimentally evaluated according to the instantaneous thermal efficiency test described in the protocol ISO 9806:2013 [42]. The collectors connected in series had an aperture area of approximately 1.72 m² and were equipped with two 150 L hot water tanks.

The procedure for evaluating the instantaneous thermal efficiency of solar collectors was carried out by isolating the exposed pipes of the solar collector and placing temperature and flow sensors at its inlet and outlet; this was to measure the temperature differences in the system as well as the mass flow of the fluid at that instant. The flat plate solar collector was mounted on a support structure and the collector angle was set according to the day of the year of each test to obtain solar irradiance perpendicular to the collector plane. Mass flow rate and inlet temperature were kept constant with a recirculation pump and electrical heaters installed in the storage tank for cold water. The outlet water coming from the flat-plate solar collector was stored in the hot water storage tank. The water in the hot water storage tank was returned with the recirculation pump to the cold water storage tank to fix a new temperature and mass flow rate to repeat the process and save water consumption.

According to the protocol to obtain the thermal efficiency of a flat-plate solar collector, the tests were performed at solar noon, using water as the working fluid. During the test, measurements were taken of the inlet (T_i) and outlet (T₀) temperatures of the collector, ambient temperature (T_a), solar irradiance at the collector plane, and mass flow every 10 s. Four instantaneous thermal efficiency points were obtained in four inlet temperature ranges (approximately 28 °C, 38 °C, 55 °C, and 65 °C). With the information obtained from the instantaneous thermal efficiency as a function of temperature difference and solar irradiance, the heat loss transfer coefficients and the collector heat removal factor were obtained. An experimental setup scheme for the collector’s characterization is provided in the Supporting Information, Figure S5, with details on the interconnection of the different elements of the characterization system.

3. Results

The crystal structure, chemical composition, and morphology of the Ni/BN selective coatings were characterized with and without MS coating. The stagnation temperature in a flat-plate collector can reach temperatures up to about 200 °C; therefore, thermal treatments were chosen to be in air at 200 °C for 200 h, exceeding realistic operating conditions. Figure 1a,b show SEM images of the BN film surface, illustrating the nanostructured and porous morphology of a BN film. Nanoflakes of 10 ± 2 nm thickness and 100 ± 5 nm long were vertically oriented with respect to the substrate, and, besides the nanoflakes, 10 ± 2 nm spherical nanoparticles were observed; the morphology did not change after the thermal treatment. This morphology aids light trapping and is in concordance with previous reports, and the existence of different nickel compounds was expected [43,44]. Figure 1c shows the surface of a BN with the MS coating after the 200 °C thermal treatment, showing a porous material, illustrating that the modified silica film completely covered the black nickel nanoflakes. Interestingly, related to the incorporated functional groups, the MS films exhibited increased hydrophobic character for the thermally treated films. Figure 1d shows that the contact angle was θ = 73° for the sample with PTES and θ = 28° for the sample without PTES. Figure 1e shows the surface AFM images of the SC and MS films treated at 200 °C, illustrating that the MS film presented ordered and periodic pores, like the morphology observed with SEM. The formation of the pores in the MS decreased the film refractive index, approaching the reported ideal value and resulting in the desired anti-reflective effect [45]. The AFM images show that the coatings were composed of silica structures with a sharpened texture, resulting in a low refractive index that defined the antireflective and hydrophobic properties of the coatings related to the propyl moiety in the MS film [46]. The surface average roughness Ra was calculated for the two films from the AFM images; the MS film roughness was 1.1 ± 0.1 nm while 1.4 ± 0.1 nm was obtained for SC. The smaller Ra value for MS indicates that the film had less light scattering, thus contributing to the higher transmittance than that of SC. It was also necessary to consider the thickness of both films, which needed to be adequate to generate the destructive interference effect that contributed to increasing the transmittance; the film thickness for both films can be seen in the Supporting Information, Figure S8. Based on the transmittance results and profilometry, we infer that the thickness of the MS was better, even though both films were deposited under the same conditions.

Figure 2a shows the XRD pattern of the Cu/Ni/BN selective coating before and after thermal treatment at 200 °C for 200 h. Related to the nickel and black nickel films, both Cu from the substrate (ICDD PDF 004-0836) and Ni (ICDD PDF 004-0850) were detected. The electrodeposited films showed a face-centered cubic (fcc) structure for the metallic Ni, and the significant peak intensity corresponding to the (111) plane revealed the preferential growth of the film along this direction. It has been reported that chloride selectively suppresses the growth of certain planes, thus allowing preferential growth [47,48]. No peaks corresponding to nickel oxide were observed, indicating that the NiO_x phases were very thin films and amorphous. The X-ray diffraction pattern for the thermally treated sample was identical to the as-deposited sample, illustrating that the oxides did not crystallize after the thermal treatment, which agrees with the report by Lizama-Tzec [43] related to the presence of amorphous NiO_x on the black nickel surface and metallic nickel in the bottom of the film. In the Supporting Information, Figure S6, the XPS depth profile analysis is shown for samples of BN films without TT, showing the presence of NiO_x. As shown in Figure S6a for the 0 s analysis, a metallic Ni shoulder was detected at 852.8 eV, and the shoulders at 854.2 eV and 856 eV were assigned to Ni²⁺ and Ni³⁺ oxidation states, respectively [49]. After the increasing Ar etching time (in 20 s steps), the analysis showed an intensity increase of the metallic nickel peak, while the Ni²⁺ and Ni³⁺ intensities decreased; this was expected for a cermet material with graded metallic composition [50].

Figure S6b shows the XPS spectra for the oxygen window for the BN sample. The analysis at 0 s shows a wide and pronounced peak centered at 531.7 eV, associated with the oxygen from Ni(OH)₂. The peak covered the energies for the reported species at 532.1 eV (NiOOH) and 532.9 eV (H₂O) [49]. In addition, a less pronounced shoulder centered at 529.8 eV related to NiO was observed; as the successive Ar etching/XPS measurement steps were applied, it was found that the peak for Ni(OH)₂ decreased and the peak assigned to NiO increased. For the XPS analysis, after the etching step at 511 s, the spectra did not present peaks associated with Ni(OH)₂, H₂O, and NiOOH but only showed a weak peak associated with NiO. The previous results indicated that NiO_x coexisted with other chemical species at the BN surface while, at the base of the film, metallic nickel dominated.

Figure 2b presents the X-Ray patterns for silica and modified silica powders. For the two powders, a broadened peak in the 2θ range of 15–30° was observed, corresponding to the amorphous silica matrix. The analysis matches the International Centre for Diffraction Data (ICDD) Powder Diffraction File (PDF) card numbers 050-0057 and 045-0112 assigned to Si₆₄O₁₂₈ and SiO₂, respectively; the patterns are inserted as visual references for the expected peaks. This pattern did not change after exposure to the deposition temperature (200 °C) used in spray pyrolysis, illustrating that the MS remained amorphous. A broadened shoulder over the 2θ range of 3° to 8° was observed for both powders, indicating ordered mesopore networks in the materials; this peak was more intense and sharper for the powder obtained with TEOS/PTES [51]. The ordered porous film could be attributed to the incorporation of a propyl functional group, which replaced some rigid Si-O-Si bonds, resulting in a more ordered material after the drying process. The larger intensity of the peak at low angles could be attributed to a larger number of ordered mesopore networks in the sample with PTES; these results are consistent with reports in the literature [52].

The optimum refractive index for the antireflection coating of the glass could be calculated using Equation (3):

n_f = (n₀ n_s)^1/2 ≈ 1.22

(3)

where n_s ≈ 1.5 is the refractive index of the glass or other substrate material and n₀ is the refractive index of air. The lowest refractive indices known for dielectric materials are near 1.35; therefore, the value of n_f is unreachable for conventional compact single AR coatings [53]. However, using mesoporous or microporous silica films for this objective with pore sizes smaller than the visible wavelengths, the refractive index of the medium is dependent on the film porosity (or film density). High porosity must be introduced in films to approach the refractive index of ~1.22 using silica with a refractive index of ~1.5. The expected effect of AR coating on the selective coating was to decrease the reflection in the solar range and increase the solar absorptance. Note that Equation (3) is only valid for non-absorbing film substrates, which was not the situation for the BN coating; however, the estimated refractive index of the BN absorbing film was about 1.7 [54].

The MS coatings were evaluated using transmittance spectroscopy. Figure 3 shows that the transmittance for an MS film in the 320–1100 nm range is about 101.2% on average, an increase of about 1.2% with respect to the glass slide without the AR coating. The transmittance exceeded 100%, demonstrating that the refractive index of the MS films was smaller than that of the glass substrate. As a result, the thin MS films presented an excellent antireflective effect, and films with similar properties were deposited on the selective coatings. The insert shows an image of a glass slide with the right half treated with the AR coating and the left half untreated. The difference in reflection is noticeable, in agreement with the lower transmittance of the glass (which was used as the 100% reference point). The results of transmittance spectroscopy and the image of the glass substrate exemplify the expected anti-reflective effect of the modified silica on selective coatings. Based on G. Helsch’s [55] reported values, it was estimated that the porosity of the MS-coated films should be around 30% to exhibit an anti-reflective effect over a wide wavelength range, as observed for the TEOS/PTES film [56]. It is worth highlighting that in the range of 320–400 nm, the transmittance of the MS coating was around 2% above the reference value, showing that the film did not have a long-term problem due to cleavage of the Si-C or C-C bonds associated with solar spectrum UV radiation. On the other hand, MS showed a higher intensity for the X-ray shoulder at 6° and was less rough than SC; therefore, the refractive index of this film was closer to the ideal theoretical value of 1.22, resulting in a better anti-reflective effect.

Figure 4 shows the infrared absorbance spectra for films obtained with TEOS and TEOS/PTES precursor solutions with and without thermal treatment at 200 °C for 200 h. For graphic clarity, in Figure 4a the peaks are labeled and in Figure 4b they are only indicated with arrows. All peaks in Figure 4a,b were found at the same wavenumber values, except those indicated below. In both figures, a close-up is inserted in the range of 1200 cm⁻¹ to 3800 cm⁻¹. In Figure 4b, an additional close-up is inserted in the range of 2820 cm⁻¹ to 3050 cm⁻¹. The difference found between the spectra of films obtained with TEOS and TEOS/PTES, both as-deposited, was an increase in the relative intensity of the peaks that were assigned to Si-ethoxy bonds and siloxane groups. An increase in intensity for the peaks at 2873 cm⁻¹ and 2963 cm⁻¹ assigned to the vibrational modes of Si-CH₂-CH₂-CH₃ was found for the film obtained with PTES (see Table 1).

In the comparison of the spectra between the TEOS and TEOS/PTES films, both thermally treated, a decrease in the intensity of the peaks associated with Si-ethoxy, ethyl, and siloxane groups was found. At 1720 cm⁻¹, C=O of acetaldehyde was observed, formed by the loss of the hydrogen of the carbon of the group -OH of ethanol; this band was not affected by the TT. For the film with PTES, the presence of the peak assigned to C-Si at 775 cm⁻¹ was detected overlapping with the peak at 785 cm⁻¹.

In the 2852–2975 cm⁻¹ range, peaks assigned to the symmetric and antisymmetric vibrations of the -CH₂ and -CH₃ groups were detected. After thermal treatment, the peak at 2873 cm⁻¹ did not disappear for samples prepared with PTES; this peak could be assigned to the -CH₃ group of n-propyl. The peak at 2963 cm⁻¹ was present in all four samples. The peak at 2975 cm⁻¹ was assigned to -CH₃ of ethanol and disappeared after the thermal treatment. In the 3000–3500 cm⁻¹ range, a wide band assigned to the -OH group was observed.

On the other hand, a slight decrease in the relative intensity of the 1023 cm⁻¹ and 947 cm⁻¹ bands was observed for the thermally treated samples, indicating a small degree of densification. The temperature of treatment at 200 °C was low, and the existence of remanent bands related to Si-ethoxy groups indicated that the TEOS and PTES did not hydrolyze completely. This was because to obtain dense, purely inorganic silica, it is necessary to increase the heat treatment temperature to about 300 °C, as was demonstrated by San Vicente [57].

As reported by Que [58], the presence of ethoxy groups after TT gives greater flexibility to the silica film, which is useful in the application of solar collectors due to the constant heating and cooling cycles. It has been reported that the presence of alkyl groups helps to modulate the hydrophobicity of the silica film due to the non-polarity of the alkyl groups [59].

For the samples, the infrared spectrum showed a shoulder at 1620 cm⁻¹ due to the formation of Si-OH groups at the removal of ethoxy groups during the condensation reaction. The addition of PTES increased the hydrophobicity of the silica film surface due to the presence of the n-propyl group (Si-CH₂-CH₂-CH₃). Figure 1d shows an example of the differences between the contact angle for a sample treated with PTES TT and a silica TT sample [60,61,62,63,64].

Table 1. Assignment of the FTIR-ATR peaks observed for the silica and MS films.

Wavenumber (cm−1)	Origin	Remark	Reference
775	Si-C	from Si-CH2-CH2-CH3	[65]
785	Si-O	overlapped, -Si-ethoxy	[61]
810	Si-O	siloxane groups: Si-O-Si, overlapped	[61,62]
880	C-O	from ethanol	[60]
947	Si-O	remnant shoulder Si-ethoxy, non-hydrolyzed	[60]
1023	Si-O	Si-ethoxy, non-hydrolyzed	[60]
~1060	O-Si-O	shoulder overlapped, stretching vibrations	[61]
1170	Si-O	remnant shoulder Si-ethoxy, non-hydrolyzed	[61]
1298	-CH	ethoxy, low intensity in TT samples	[63]
1372–1496	-CH	ethyl, low intensity in the TT samples	[63]
1620	Si-OH	condensation reaction of Si-OH	[64,66]
1720	C=O	all samples	[67]
2852	-CH2	symmetric stretching vibration	[61]
2873	-CH3	symmetric stretching vibration related to Si-CH2-CH2-CH3	[61,63]
2923	-CH2	asymmetric stretching vibration	[61]
2963	-CH3	asymmetric stretching vibration related to Si-CH2-CH2-CH3	[61,63]
2975	-CH3	from ethanol	[61,63]
~3000–3500 (3316)		hydroxyl group stretching; adsorbed water	[64]

Table 1. Assignment of the FTIR-ATR peaks observed for the silica and MS films.

Wavenumber (cm−1)	Origin	Remark	Reference
775	Si-C	from Si-CH2-CH2-CH3	[65]
785	Si-O	overlapped, -Si-ethoxy	[61]
810	Si-O	siloxane groups: Si-O-Si, overlapped	[61,62]
880	C-O	from ethanol	[60]
947	Si-O	remnant shoulder Si-ethoxy, non-hydrolyzed	[60]
1023	Si-O	Si-ethoxy, non-hydrolyzed	[60]
~1060	O-Si-O	shoulder overlapped, stretching vibrations	[61]
1170	Si-O	remnant shoulder Si-ethoxy, non-hydrolyzed	[61]
1298	-CH	ethoxy, low intensity in TT samples	[63]
1372–1496	-CH	ethyl, low intensity in the TT samples	[63]
1620	Si-OH	condensation reaction of Si-OH	[64,66]
1720	C=O	all samples	[67]
2852	-CH2	symmetric stretching vibration	[61]
2873	-CH3	symmetric stretching vibration related to Si-CH2-CH2-CH3	[61,63]
2923	-CH2	asymmetric stretching vibration	[61]
2963	-CH3	asymmetric stretching vibration related to Si-CH2-CH2-CH3	[61,63]
2975	-CH3	from ethanol	[61,63]
~3000–3500 (3316)		hydroxyl group stretching; adsorbed water	[64]

3.1. Optical Properties of the Selective Coatings

The optical properties of the selective coatings of Ni/BN on the Cu substrates, with and without the MS coating and before and after thermal treatment, were characterized using reflectance spectroscopy. Figure 5a displays the total reflectance spectra of Cu/Ni/BN and Cu/Ni/BN/MS coatings after thermal treatments at 200 °C.

Table 2. Solar absorptance, thermal emittance, and spectral selectivity values were calculated for the selective coatings.

	Solar Absorptance α (%)	Thermal Emittance ε100 °C (%)	Spectral Selectivity h = α − 0.5 ε (%) [68]
Cu/Ni/BN	89.0 ± 0.2	12.0 ± 0.4	83.0 ± 0.4
Cu/Ni/BN TT	90.0 ± 0.1	11.0 ± 0.2	85.0 ± 0.2
Cu/Ni/BN/MS	93.0 ± 0.1	13.0 ± 0.3	86.5 ± 0.3
Cu/Ni/BN/MS TT	93.0 ± 0.1	13.0 ± 0.2	86.5 ± 0.2
Cu/CBP	92.0 ± 0.1	88.0 ± 0.2	48.0 ± 0.2
Cu/CBP TT	91.0 ± 0.3	90.0 ± 0.1	46.0 ± 0.4

shows the calculated values for solar absorptance and thermal emittance for these films. The as-deposited samples show the spectrum of a selective coating, with low reflectance in the visible and near-infrared and a large reflectance in the mid-long infrared. After thermal treatment at 200 °C, the reflectance spectrum was essentially the same, with increased solar absorptance. The thermally treated coating exhibited solar absorptance of 90.0 ± 0.1% and low thermal emittance of 11.0 ± 0.2%. The thermal treatment therefore improved the selectivity of the coating from 83.0 ± 0.4% to 85.0 ± 0.2%. The changes observed in the 300–2000 nm wavelength range were due to an increase in the BN film roughness from 27 ± 2 nm to 31 ± 2 nm, resulting in a slight increase in absorptance. Roughness profiles related to the BN films with and without TT are provided in Figure S7.

The reflectance spectra of the Cu/Ni/BN/MS sample after thermal treatment at 200 °C for 200 h compared to the spectrum for the Cu/Ni/BN TT coating show that there was an increase of 3% in solar absorptance in the spectral range of 300–2500 nm, while the increase in thermal emittance was slight at 2%. The spectra for the as-deposited MS samples and the thermally treated samples were identical, indicating no subsequent changes in the coating due to the prolonged thermal treatment time related to the decomposition of remaining functional groups or the densification of silica particles. In the samples with the MS film, the presence of the spray-deposited silica was signaled in the spectrum by the well-defined signal at 9430 (1060 cm⁻¹) nm. In fact, the thermal emittance was somewhat larger, mainly because of the absorption peak at 9430 nm. The absorption bands at about 9430 nm were assigned to the stretching vibration of Si-O-Si groups from the modified silica film, in accordance with the FTIR results.

Figure 5b shows the reflectance spectra for the Cu/CBP coating, characterized by low reflectance throughout the spectrum. The coating exhibited a large solar absorptance of 92.0 ± 0.1%. This was coupled with a lack of selectivity due to the calculated high thermal emittance of 88.0 ± 0.2%. The spectra changed slightly after the thermal treatment, resulting in lower solar absorptance and higher thermal emittance. The results show that the highest spectral selectivity of h = 86.5 ± 0.3% was obtained for the full coating before the thermal treatment, related to the increase in solar absorptance to α = 93.0 ± 0.1% and combined with a minimal increase in thermal emittance to ε = 13.0 ± 0.3%. In addition, the coatings exhibited excellent thermal stability at 200 °C. No peeling was detected in the MS coating, and the optimized transmittance of the MS coating helped to improve the solar absorptance of the selective coating.

3.2. Scale-Up and Implementation in the Solar Collector

The deposition process of the modified silica material was scaled up and implemented to provide the anti-reflective film on Ni/BN selective coatings electrodeposited onto Cu fins of 193 cm × 12 cm. We assembled solar collectors consisting of seven fins each, corresponding to a total aperture area of about 1.7 m².

Three solar collectors were fabricated to compare the performance of a commercial system with a black paint film (Cu/CBP) to collectors with the Ni/BN selective coating, both with and without the modified silica AR coating. The collectors were evaluated under outdoor conditions in Jiutepec, Morelos, Mexico; Figure 6 shows an image of the three collectors tested.

The collectors were evaluated by measuring the temperature differences in the system for a certain mass flow rate of the fluid, which could be related to the efficiency, η, of the collector (see experimental section for details). The collector efficiency was determined by the radiant power absorbed and the heat losses related to conduction, convection, and thermal emittance. If there were no losses, which corresponded to the situation where the temperature difference was negligibly small, the maximum efficiency, η₀, was given by the solar absorptance, α, corrected for the collector cover transmittance, τ, and multiplied by a heat removal factor, F_r, which described the heat transfer to the collector fluid and which depended on the fluid properties, flow rate, and characteristics of the collector–fluid interface. For sufficiently small temperature differences, the total efficiency decreased linearly with increasing temperature differences according to the following relation [69,70,71]:

$$\eta =F_r\alpha \tau -F_r{U}_L\left({\displaystyle \frac{Ti-T_a}{G}}\right)={\eta }_0-{\eta }_1\left({\displaystyle \frac{T_i-T_a}{G}}\right),$$

(4)

where U_L is the heat loss transfer coefficient, G is the solar irradiance at the collector plane, and (Ti $-$ Ta) is the difference between the temperature of the fluid at the inlet of the collector and the ambient temperature. If the efficiency was graphed versus (Ti $-$ Ta)/G, the intercept on the y-axis corresponded to the maximum collection efficiency, and the slope of the approximately straight line, η₁ = −F_r U_L, represented the total rate of heat loss from the collector to the environment [72,73]. Note that F_r corresponds to the heat transfer to the collector fluid, while U_L provides a quantification of the heat losses associated with conduction, convection, and thermal emittance. Figure 7 shows the instantaneous thermal efficiency curves obtained under outdoor conditions for the Cu/Ni/BN, Cu/Ni/BN/MS, and Cu/CBP collectors.

Figure 7 and

Table 3. Performance characterization results for the evaluated solar collectors.

Collector	α	ε100 °C	τ	η0	η1	Fr	(ατ)	UL
Cu/Ni/BN	0.89	0.12	0.91	0.68	3.9	0.83	0.81	4.7
Cu/Ni/BN/MS	0.93	0.13	0.91	0.70	4.2	0.84	0.83	5.0
Cu/CBP	0.92	0.88	0.91	0.70	6.8	0.84	0.83	8.1

show that the maximum collector efficiency improved upon adding the MS film onto the Cu/Ni/BN selective coating, which was in accordance with the higher solar absorptance measured and was the same as for the commercial Cu/CBP collector. The efficiency decreased approximately linearly with the temperature difference, in agreement with Equation (4). An important observation was that the slope for the Cu/CBP collector was significantly larger than for the Cu/Ni/BN systems. Considering that the collector–fluid heat transfer was the same for all systems measured (i.e., F_r was the same), this indicated a more significant value for U_L describing the heat losses under operating conditions for the Cu/CBP collector. Since all other materials and interfaces were the same, this was most likely due to the large thermal emittance of the CBP film. Note that the slope for the Cu/Ni/BN/MS collector was slightly larger than for the system without the modified silica coating, which also correlated with a slightly larger thermal emittance. However, in the temperature difference range evaluated, the system with the modified silica coating had much better total performance due to the higher solar absorptance. These results also demonstrate that the characterization of the optical properties of the 4 cm × 4 cm samples can provide a good indication of the performance of scaled-up systems in a complete solar collector. On the other hand, based on the literature, it is thought that MS contributes to less heat loss since amorphous materials do not have a structural order, so there is no continuous crystalline network in which the atoms can vibrate, so that phonons propagate. As a result, for the phonon, the mean free paths are restricted to the interatomic spacing and the effective thermal conductivity of MS remains low.

To compare the AR thermal stability results obtained on the 16 cm² samples with the results of the copper fins with the selective coating used in the solar collector, we sectioned one of the copper fins with the Ni/BN/MS coating, which was deposited by an identical process, and characterized the surface with ATR-FTIR. The measurement was performed in three different zones on the fin surface, and spectra were identical for the three zones, illustrating the reproducibility.

Figure 8 shows ATR-FTIR spectra obtained from sections of Ni/BN/MS TT-coated copper fins fabricated under equal conditions to those used for the evaluated collector. The spectra look like those obtained for glass slide substrates, as observed in Figure 4.

In comparison with Figure 4, the positions of some peaks were displaced as detailed below: as shown in Figure 8, for the sample, a band at 775 cm⁻¹ was detected, which was assigned to C-Si and was overlapping with the band at 785 cm⁻¹ assigned to Si-O of Si-ethoxy. The main peak associated with Si-ethoxy was detected at 1020 cm⁻¹. These peaks, compared to those detected in Figure 4 on glass, were close to each other. At 1170 cm⁻¹ [74].

The close-up inserted spectra in the 1200–3800 cm⁻¹ wavenumber range show that the spectra are essentially the same as those in Figure 4a for TEOS/PTES film TT, without significant shifts in the peak positions, except for peaks assigned to H₂O, C=O and the peak assigned to -CH at 1372 cm⁻¹. The peak at 1298 cm⁻¹ related to the -CH vibration of the ethoxy group was detected without changes. The changes in intensity observed were in the range of 2800–3000 cm⁻¹; however, the symmetric and asymmetric stretching vibrations related to the -CH₃ group from Si-CH₂-CH₂-CH₃ were detected. The peak assigned to CH₃ at 2975 cm⁻¹ of ethanol disappeared. In the 3000–3600 cm⁻¹ range, a widened and pronounced band was detected, indicating the existence of additional OH- groups on the BN surface; this was useful because it improved the adherence of MS at the black nickel surface. This result demonstrates the compatibility between MS and BN surfaces.

Based on these observations, we can infer that for the PTES-containing films, even after prolonged thermal treatment, peaks corresponding to the alkyl groups were present, indicating that the MS films were thermally stable. In addition, the surface morphology observed in the AFM image of the MS was regulated by steric effects [75]. This conclusion also correlates with the observation that the films showed increased hydrophobicity after thermal treatment. In addition, the films did not exhibit any cracking or peeling after the prolonged thermal treatment, which was related to the improved flexibility caused by the un-hydrolyzed ethoxy moieties and the intentionally incorporated propyl functional groups.

Finally, for a new system to be economically viable, it is essential to perform a cost–benefit analysis.

Table 4. Performance in terms of the average useful heat generated, Q_U, and cost analysis results for the three solar collectors fabricated and characterized in this work.

Collector	QU (J)	Cost (USD)	Unit Cost (USD/J)
Cu/CBP	840	93.5	0.111
Cu/Ni/BN	894	95.6	0.107
Cu/Ni/BN/MS	990	105	0.106

shows the costs of the different coatings used in the three collectors fabricated in this work in relation to the generated heat. We calculated the average useful heat, Q_U, using the following equation:

Q_U = η_av G A_c

(5)

where η_av corresponds to the average collector efficiency and A_c is the collector aperture area (1.72 m²). Table 4 shows that the solar collectors with the Ni/BN selective coating, both with and without MS film, produced more useful heat than the painted system. The fabrication of these more complex systems was somewhat more expensive, but the unit cost of the heat produced was lowest for the Cu/Ni/BN/MS system, providing an excellent motivation to consider the manufacture of this system. Detailed information about the cost estimate can be consulted in Tables S1 and S2 in the Supporting Information.

Another benefit of the higher efficiency of the Cu/Ni/BN/MS collector was that it required less collector area than a commercial collector. This is an important aspect of sustainable manufacturing and reduces the collector’s environmental impact. The area is an important parameter in solar collectors; just as in solar cells, efficiency is dependent on the device area. The preparation of selective coatings in small areas can give better properties than selective coatings for full-size solar collectors due to several factors implicit in the material preparation methodology and more optimal processes that can be applied in smaller areas. Therefore, being able to synthesize selective coatings in real-size prototypes and deposit a film with anti-reflective properties is a good achievement.

4. Conclusions

In this work, we evaluated the effects of implementing a modified silica (MS) film on top of electrodeposited Ni/BN selective coatings, both at the laboratory scale and in full-sized 1.72 m² solar collectors. We fabricated three collectors based on copper fins: one with commercial black paint (CBP) as an absorber and two with an electrodeposited black nickel selective coating, of which one was covered with an MS film. The MS film was deposited by spray pyrolysis from a [PTES]/[TEOS] molar ratio 0.09 precursor solution, resulting in a conformal mesoporous film that did not show change under thermal treatments up to 200 °C. The fabrication method was shown to be scalable up to collector size. The incorporation of the MS film increased the solar absorptance of the coating and improved the performance of a complete, 1.72 m² solar collector under real outdoor conditions. The Cu/Ni/BN/MS system maintained the advantage of electrodeposited Ni/BN selective coatings in terms of a much lower thermal emittance than coatings based on black paint. The novel system also demonstrated stability under heat treatments, illustrating the technical viability. A cost analysis indicated that the higher efficiency of the new collector resulted in a lower unit cost of heat produced compared to the commercial black paint system. Finally, higher efficiency is an important aspect of sustainable manufacturing as it decreases the amount of material needed for a given amount of heat and thus the carbon footprint.