Self-Cleaning Solar Mirror Coatings: From the Laboratory Scale to Prototype Field Tests

Featured Application

This work could be applied for producing self-cleaning solar mirrors.

Abstract

In this study, a low-cost, scalable and robust process is proposed as an innovative method for coating solar mirrors with a self-cleaning, transparent in the full solar range and versatile material based on auxetic aluminum nitrides, previously obtained at the laboratory scale. This work presents the scaling-up of the fabrication process from the laboratory to prototypal scale and the preliminary results of outdoor self-cleaning solar mirror field tests in the demonstrative concentrating solar power (CSP) plant ENEASHIP located in Casaccia (Rome) ENEA Research Center. Prototypes with a size of 50 × 40 cm have shown stability in external conditions: no coating degradation occurred during the test campaign. Their washing restores the initial reflectance affected by soiling and the self-cleaning performance allows for the utilization of a reduced quantity of water for cleaning operations with respect to the uncoated glass of back surface mirrors. A similar self-cleaning AlN coating could be utilized on other solar components affected by soiling, such as the glass envelopes in heat-collecting elements, PV panels and other parts where a self-cleaning performance combined with an optical one is required.

1. Introduction

Solar mirrors play a crucial role in concentrating sunlight for various applications, such as solar power generation and industrial processes [1]. However, the efficiency of these mirrors can be significantly compromised by the accumulation of dirt, dust and other contaminants on their surfaces, leading to reduced reflectivity and overall performance [2]. To address this challenge, the development of advanced surface coatings, known as self-cleaning, has gained prominence in the field of solar mirror technology [3]. A lot of materials specifically designed to modulate wettability (for reducing water consumption in washing maintenance operations) and ensure optimal reflectivity and sustained energy capture over time have been proposed on a laboratory scale or on an intermediate scale in the form of small prototypes [4]. None of these coatings have reached the commercial state for different reasons. Passing from a laboratory scale to an industrial one is not a simple task when there are strict optical requisites to be considered; for example, the external positioning of covered mirrors determines the degradation of the exposed layer, and above all, adding a further step to existing production lines of solar mirrors has to be evaluated in economic terms as trade-off between its cost and real benefits in terms of performance [5]. This last aspect of the economic evaluation of water saving in ordinary solar field maintenance operations is difficult because each site has its own peculiarities that make comparison and, in general, absolute evaluation difficult, but which nevertheless allow us to draw an indication of the fact that any water saving, especially in desert areas, is always profitable [6]. In a previous work, the feasibility of changing alumina, that is, the most diffuse last layer in all mirror architectures, with AlN compounds obtained by means of sputtering deposition was investigated [7].

Different auxetic AlN politypoids have been fabricated (from a 52° to 100° water contact angle, WCA, due to the major covalent character of the Al-N bond with respect to the Al-O one) with the purpose of tailoring wettability, preserving optical clarity in the desired solar range and granting durability over time considering the inorganic nature of AlN.

Such promising results have been obtained on a laboratory scale but conceived for a subsequent scaling-up in consideration of the fact that they are produced with an intrinsically scalable technique.

Although research products studied on a laboratory scale exhibit interesting properties, their effective utilization in the real world often requires a lot of applied research, devoted to verifying if the same properties obtained on the small scale can be preserved when dimensional and performance requirements must be satisfied. The first problem to solve is the selection of a fabrication technique that grants the same chemical–physical properties obtained in the lab on the small scale. For example, this occurs in the case of polymeric products, where the coating is obtained at the lab scale by means of spinning, which is clearly a nonscalable technique and, therefore, needs to be changed, e.g., with printing techniques.

This work aims to fabricate inorganic self-cleaning auxetic and transparent aluminum nitride coatings, selected from the family of materials previously developed with sputtering processes on a pre-industrial scale for the purpose of uniformly covering 50 cm × 40 cm prototypes, that can be tested as back surface mirrors, BSMs, in the demonstrative CSP Fresnel plant, named ENEASHIP and located in Casaccia (Rome) ENEA Research Center. In particular, the requirements of such coatings are as follows:

-

High transparence in the solar range, so as not to change the mirror’s specular reflectance required for optimal operation in concentrating solar power plants.

-

Self-cleaning performance.

-

Outdoor durability, mainly understood as persistent adhesion to the substrate even in the presence of adverse meteorological events and chemical–physical instability. A trade-off between self-cleaning performance and optical UV-VIS-NIR transmittance can be reached when the difference in reflectance between the solar mirror as it is and the coated mirror does not exceed 10% in the visible region, and for wavelengths between 850 nm and 2500 nm, a coating transmittance of at least 80% is necessary. When a plant component, like a solar mirror, is modified by means of adding another layer (e.g., a self-cleaning layer), the testing of its performance on-field is crucial. Not even a coating material that is stable indoors or during a climatic chamber testing campaign preserves its performance when positioned outdoors and exposed to adverse and, in any case, variable weather conditions for a long time. For this reason, on-field tests of the solar mirror coating prototypes’ stability and performance are part of this work.

2. Materials and Methods

In this study, AlN layers were deposited by the sputtering technique using a proprietary multicathode sputtering apparatus, named ENEA2 (Figure 1), equipped with a process chamber and load lock chamber. ENEA2 is the apparatus on which they were developed, patented and determined as international industry key technologies in the field of selective thin-film solar coatings for receiver tubes [8].

In fact, it can process tubular substrates with a maximum length of 600 mm, external diameter between 7 and 10 cm (typically 7 cm) and wall thickness between 2 and 5 mm. In the prevailing tubular configuration, the tube–substrate is placed in a horizontal position on a system with a “trolley + pallet” that transports it from the loading chamber to the process chamber.

The tube is made to translate and rotate repeatedly in front of magnetron cathodes (sputtering sources) arranged vertically on both sides of the apparatus. In the process chamber, six cathodes are mounted and arranged in pairs on two opposite sides, specifically, three standard magnetron and three dual magnetron cathodes. The standard magnetron cathodes can be operated in DC and DC pulsed sputtering mode, whereas the dual magnetron cathodes can be operated in bipolar DC pulsed and MF sputtering mode. In the load lock chamber, plasma etching and heating pre-treatments can be performed. The machine is equipped, in the process chamber, with a high-vacuum pumping system. Load lock pumping is carried out by means of a cryo-pump. The general tubular layout of the ENEA2 sputtering plant was changed in this work to a proprietary planar pallet for making deposits on flat panels (e.g., metallized glasses) with a maximum length of 50 cm and a height of up to 40 cm.

The mirror substrate was positioned in front of the target in a planar configuration and the fabrication process was a reactive sputtering with an Ar + N₂ pressure of 3 Pa and a power of 1800 W supplied to the Al cathode.

Silicon substrates for homogeneity tests were purchased from Merk °KGaA, Darmstadt, (Germany) SpA.

Solar mirrors with dimensions of 50 × 40 × 1 cm were purchased from Società Vetraria Biancaldese SpA., Roncade (PV), Italy.

UV-VIS-NIR analysis was performed using a double-beam Perkin-Elmer (Waltham, MA, USA), Lambda 900 instrument, equipped with a 15 cm integrating sphere to measure global spectral reflectance and transmittance.

The thickness of sample layers was measured by means of a surface profiler, TENCOR P-10.

The static water contact angle (WCA) was measured with the direct optical method of drop-shape analysis, by means of the contact angle meter, KRÜSS, Hamburg (Germany) DSA-100.

Tests on-field were performed on prototypes obtained by gluing coated mirror panels on BSM commercial constitutive mirrors (as better explained in the following paragraphs) outdoors in the demonstrative Fresnel plant ENEASHIP, located in Casaccia (Rome) ENEA Research Center. Reflectance measurements on-site were performed by means of a portable D&S R15-USB, averaging 3 values for each 50 × 40 panel and 9 values for the entire solar mirror.

Measurements were carried out on different days at the same points to correctly assess the change in reflectance due to soiling.

Water retention was estimated by means of a gravimetric procedure using deionized water with a resistant of 10 MΩ and a technical-scale Giorgio Bormac S.r.l., Carpi (Mo) Italy, XS BALANCE BL 2002 suitable for weighing glass sheets with an area of 10 cm² operating in the temperature range between 10° and 40 °C. Each measurement was performed at a defined temperature and humidity before and 10 s after a washing step, performed by positioning samples on a holder inclined at 70° with respect to the laboratory plane and spraying a defined amount of deionized water on them.

3. Results

To obtain self-cleaning solar mirrors with big dimensions that can be positioned in a real CSP plant, two important aspects have been studied and described in the following paragraphs: the scaling-up of the auxetic AlN magnetron sputtering fabrication process from the laboratory to pre-industrial scale (Section 3.1) and the outdoor testing of fabricated prototypes (Section 3.2).

3.1. Scaling-Up

The scaling-up from the laboratory scale to pre-industrial scale requires the optimization of magnetron sputtering processes to obtain large-scale auxetic AlN coatings with the desired performance in terms of optical and wettability properties by means of robust and repeatable processes.

In a previous work, we described a family of AlN politypoids fabricated by means of magnetron sputtering on 3 × 7 cm mirror substrates, using the tubular geometry of ENEA2 proprietary apparatus, and demonstrated that the auxetic phases simulated by Kilic and colleagues [9] can be experimentally obtained using both a saturation and a transition regimen with the aluminum target, which includes doping the hexagonal AlN lattice with metals (Al or Ag) or oxygen.

Among such an interesting family of materials, the procedure for fabricating a prototype was selecting the material that was the most transparent in UV-VIS-NIR with a higher WCA (°) and then scaling-up the fabrication process to obtain the maximum dimensions possible with proprietary apparatus in an applied research laboratory. In particular, the reactive sputtering of aluminum with nitrogen in the transition was preferred to the saturation regime because the deposition rate is higher and occurs at a lower energy and with lower target consumption while, at the same time, preserving the stability and repeatability required for the production process. So, the starting experimental conditions to scale up were selected as a power supply to the aluminum target of 1500 W, p = 0.1 Pa and a gas flow rate of 200 sccm Ar + 25 sccm N₂; last but not least, the process at the laboratory scale was performed with a rotating holder.

Here, the main goal of the work was to reproduce, on a planar arrangement, the same material properties. In general, the sputtering deposition process involves many parameters, which introduce different complexities. The pre-industrial ENEA 2 plant allows the processing of a substrate with a maximum size of 50 × 40 cm in planar configuration, but to do this, the sample housing has to be replaced, removing the tubular sample holder (designed to fabricate the receiver tubes of a solar CSP plant [10] and utilized for the previous experiments on auxetics) and redefining process conditions such that the deposition on a flat plate reproduces the optical properties of the samples obtained on the small rotating flat substrates typically housed in tubular sample holders. Since Fresnel plants have mirrors with a slightly concave geometry and are equipped with a secondary concentrator, it is still possible to widen the range of the angle of acceptance of the reflected radiation, which translates into less stringent requirements on coating roughness, thickness uniformity and the specularity component.

Figure 2 shows a photo of the ENEA2 planar sample holder with the mirror mounted. This plant modification carried out on a support (the carrier) that must interact with the plasma is not trivial at all. The modification, if not carefully studied, can destabilize the plasma that is generated during the process, triggering discharges that affect uniformity and, in some cases, cause it to shut down. To avoid arcs by keeping the power density and process pressure constant, it was necessary to modulate the process gas flows through a throttle valve. Moreover, the cathode voltage was controlled to ensure process stability instead of the nitrogen content, as is the case in plasma emission monitoring.

So, after investigating different processes for reproducing the desired material, the experimental conditions utilized in such a planar configuration were decided as a power supply to the aluminum target of 1800 W (I = 4.40 A, V = 410 V), a process pressure = 3.0 Pa and gas flows of 200 sccm Ar + 90 sccm N₂.

As the sputtering cathode target length is only 20 cm, it is conceivable that the coverture of the entire panel will not be uniform.

For studying the properties of thin films at different heights on the holder during the selected sputtering process (labeled AlN_18), five silicon test substrates were positioned at distances of 6 cm from each other inside the frame of the mirror position, as shown in Figure 3.

The thicknesses and water contact angles of such test samples were measured to study the homogeneity of materials obtained on the large area and reported in

Table 1. Thickness and water contact angle, WCA (°), of test samples.

Sample	Thickness (nm)	WCA (°)
S1	780 ± 90	93 ± 3
S2	1586 ± 40	84.2 ± 1.2
S3	1628 ± 5	89 ± 4
S4	1640 ± 7	85 ± 2
S5	1648 ± 8	86 ± 3

.

The transmittance of the test samples deposited on glass substrates is reported in Figure 4.

3.2. Field Positioning and Testing

The ENEASHIP demonstrative CSP plant (see Figure 5) consists of a 36 m long linear solar collector, which, in turn, consists of 425 mirrors (62.5 × 125 × 4 cm), resulting in a total of about 330 m² of reflective surface. For testing self-cleaning mirror prototypes developed in this work and labeled as S_AlN (south AlN-coated mirror) and N_AlN (north AlN-coated mirror), some of the original mirrors were removed and then covered with the samples intended to be studied, indicated in the scheme with the light blue color to distinguish them from their substrate references, labeled as SB (south reference mirror) and SA (north reference mirror), colored in light green.

Bearing in mind that the pre-industrial ENEA 2 sputtering equipment in the applied research laboratory allows for the processing of a 50 × 40 cm substrate, to obtain a larger mirror, it is necessary to manufacture several identical sheets (using the robust and repeatable process described in the previous paragraph) and properly join them. An integration scheme useful for field tests was proposed, consisting of gluing the mirror panels with a size of 50 × 40 cm obtained by sputtering onto the previously removed BSM solar mirror substrates, according to the diagram shown in Figure 6, where the entire surface of the BSM can be covered with three panels with a size of 50 × 40 cm.

The two identical mirror prototypes, S_AlN and N_AlN, fabricated as described, were placed in two different solar Fresnel field positions, as indicated in Figure 5. By means of a portable reflectometer, their reflectance was measured at discrete and fixed points indicated in Figure 6d and compared with the value of adjacent uncovered mirror substrates SB and NB. The averaged initial reflectance values are reported in

Table 2. Prototypes’ averaged reflectance values.

Sample	S_AlN	SB	N_AlN	NB
R%	92.99 ± 0.12	95.57 ± 0.07	93.38 ± 0.19	95.24 ± 0.08

.

Prototype S_AlN was exposed to the external conditions of the plant site, and the results for the first three months of the external measurement campaign are reported in Figure 7. The measurement campaign will continue in the following months.

In Figure 8, the reflectance, R%, of two identical prototypes positioned to the south and north of the plant were compared with the uncovered substrate, taking into account the potentially different exposure.

Figure 9 shows a flow chart of the procedure for estimating the relationship between WCA and water consumption in the cleaning operation of the AlN auxetic coating deposited on glass compared to its substrate and Teflon samples.

A retention of 42% of used water on the washed surface when WCA passes from 46.8° to 89.6° and of 58% when WCA passes from 46.8° to 126.1° can be observed.

4. Discussion

Fabricating new materials at the laboratory scale requires the synthesis and characterization of various samples to obtain the desired physical–chemical properties. As already explained in a previous work, in the case of auxetic AlN, this has led to a careful definition of different experimental conditions. To move toward large-area fabrication of the material, it is necessary to select among these conditions the most profitably scalable. The scaling-up of the sputtering process of auxetic AlN deposition from a laboratory scale to a pre-industrial scale with the proprietary ENEA2 multicathode sputtering apparatus was made possible by changing the tubular sample holder (utilized for obtaining materials at the lab scale) to a planar holder, as shown in Figure 2. With this arrangement, rotation is not possible, so the goal was instead to change the process parameters to obtain the desired film property in a frontal geometry (in terms of the trade-off between optical clarity and WCA). This was made possible, after a long experimental work, by changing the power supplied to the target (from 1500 W to 1800 W) and the total pressure of the process (from 0.1 to 3 Pa). Clearly, on substrates of dimensions larger than the sputtering target, it is important to verify the homogeneity of fabricated coating properties. For this reason, a series of testing substrates were positioned at different heights in the frame of the 50 × 40 cm holder, that is, the maximum area obtainable with such apparatus, as shown in Figure 3. It was observed that only the sample 1 has a slightly little thickness for borders effects influence, while the other samples are comparable in terms of thicknesses and optical transmittance (reported in Figure 4). For the general purpose of adding one layer to existing metallized glasses of BSM solar reflectors, the observed inhomogeneity (<10 nm) can be accepted, especially considering that, when positioned on-field, such mirrors can have a tolerance of several hundreds of microns.

The WCA reported in Table 1 differs from the medium value of 87° by a small amount due to thickness. So, the selected process conditions can be considered useful to produce prototypes for the desired application.

Different samples with big dimensions were fabricated by means of the same process, glued onto commercial BSMs of the demonstrative plant substrates, according to the scheme in Figure 6, and positioned in the solar field of ENEASHIP Fresnel plant in two homologous rows to the north and south, as indicated in Figure 5, next to reference mirrors obtained by gluing the same number of uncoated panels to their surfaces. Their reflectance was measured by means of a portable reflectometer following the measurement protocol described in session 2 for a preliminary three-month period. The initial values are reported in Table 2, and it can be noticed that there is only a little difference (R% − R_sub% < 5) with respect to the uncovered solar mirror substrate, thanks to the research work on maximizing the optical transmittance of coatings. The comparison between the north- and south-positioned prototypes’ reflectance in the first 50 days shows a small difference in their soiling, suggesting that each mirror in a field can be subject to this discrete phenomenon to a different extent.

At the same time, the external positioning does not deteriorate the coatings (Figure 7 shows reflectance vs. days for the entire exposure time), even in the presence of Scirocco rain events. Idem washing procedures consisting of spraying distilled water do not degrade self-cleaning layers, and they are able to restore the initial reflectance.

These simple observations are not at all obvious, as many self-cleaning coatings, when positioned outdoors, delaminate from the substrate, especially if they are nano-micro-structured, and suffer erosive damage, which removes the self-cleaning performance.

The relationship between the WCA of the coatings and water consumption for the 10 × 10 cm samples, described in the flow chart in Figure 9, confirms the self-cleaning effect and, in other terms, the possibility of water savings during cleaning procedures. Even the result of the correlation between the static contact angle and the volume of water necessary for washing samples, although carried out on a relatively small sample, represents a unique example in the panorama of sector research.

In general, in the various CSP plants spread throughout the world and operating for the production of electricity, cleaning operations vary. Clearly, the overall washing policy depends on the decision of plant makers. A detailed estimation of how much water can be saved when all mirrors are self-cleaning can be of great interest for plant makers. It is obviously impossible at this stage to have the exact value (because there are not yet commercial self-cleaning solar mirrors nor plants with mirrors that are entirely coated), but in our opinion, a useful approximation can be taken from this experimental work with our estimation that around 50% less water is used when a commercial glass BSM is substituted with an auxetic AlN one.

5. Conclusions

This work proposes a cheap and robust process for depositing, by means of magnetron sputtering, an auxetic transparent aluminum nitride-based self-cleaning coating onto solar mirrors with dimensions of 50 × 40 cm, facing the problem of scaling from lab to pre-industrial scale encountered in previous studies on the promising self-cleaning performance of auxetic materials. Prototypes of the produced self-cleaning solar mirrors were positioned in a demonstrative Fresnel plant located in ENEA, Casaccia (Rome) Research Center, and tested on-field to study their performance. They have shown stability in external conditions: no degradation occurred during the test campaign. Moreover, it is possible to wash them, restoring the initial reflectance affected by soiling. The test campaign will be extended for a longer period.

By addressing the issue of washing mirrors with a smaller amount of water with respect to glass utilized for back surface mirrors, these coatings contribute to the overall efficiency and viability of solar mirror applications in the pursuit of sustainable and clean energy solutions.