A study of sol gel process parameters on CoCuMnOx selective coating characteristics

This research addresses the effect of process parameters on the optical properties of solar selective coating. The proposed study discusses the deposition of a selective coating of CoCuMnOx onto stainless-steel substrate. The coating was prepared via a sol-gel route and dip coating process. Mechanical adhesion between the coating and the substrate was increased through increasing the surface roughness of the substrate. Four parameters were discussed; precursor concentration, withdrawal rate, number of coating layers and the heat treatment temperature. The best achieved absorptivity was 0.906 in the wavelength range of (200–900 nm) and emissivity was 0.116 in the wavelength range of (2.5–25 μm) for a sample with precursor molar ratio divided by 60, 1.5 cm min−1 withdrawal rate, double coating layer and 450 °C heat treatment temperature. Detailed coating characterization was discussed through XRD, EDX and SEM analysis.


Introduction
In the last decade, a lot of scientific research concerned with solar selective coatings have been released due to the awareness of solar energy exploitation as a free and renewable energy resource. Selective coatings are characterized by having high solar absorptivity (α) in the UV range and low thermal emissivity (ε) in the IR range. Solar selectivity is measured by attaining the ratio of absorptivity to emissivity (α/ε). However, a more indicative measurement was proposed by [1] which gives more weight to the absorptivity relative to its importance in the form of α − 0.5 * ε. Selective coating can be used for domestic applications such as space heating and water heating. Moreover, other applications involve water evaporation, water desalination and electricity generation.
Ternary spinel-like oxides such as CuFeMnO 4 , CoCuMnO x and CuCr 2 O 4 are widely used as solar selective coatings [2]. CuFeMnO 4 was used by Shi et al as a photo-thermal coating layer on a cordierite honeycomb ceramic substrate to increase the water evaporation rate for the sake of producing clean water. The prepared sample succeeded to attain 98.2% solar absorptance under the effect of the solar simulator with an intensity of 1000 Wm −2 (AM 1.5) [3]. CoCuMnO x has also shown superior selective characteristics as shown by Vince et al [4]. The coating was deposited over an aluminum substrate and achieved a spectral selectivity α/ε of 0.95/0.04. The effect of the substrate material was investigated by deposition of CoCuMnO x on copper and aluminum substrates [5] and the results indicated that the substrate material affects the optical properties especially the emissivity (α=0.9, ε=0.011 for copper and α=0.9, ε=0.029 for aluminum substrates). Furthermore, Ma et al [6] tested CoCuMnO x to be used as solar absorber in flat plate solar collectors. The coating was applied on stainless steel substrates and achieved α=0.88 and ε=0.12. Kim et al [7] worked on enhancing the efficiency of concentrated solar powers (CSPs) by spray-coating a two-layer tandem structure of (CuFeMnO 4 / CuCr 2 O 4 ) onto Inconel 625 substrates and achieved 90.3% solar to thermal conversion efficiency.
All the previously mentioned efforts in the production of cermet solar selective coatings were prepared through sol-gel approach. This technique converts the solution 'Sol' which usually contains a mixture of metal oxides and solvents into a 'Gel' by the means of poly-condensation of the molecules to form a macromolecular network structure [8]. Different preparation parameters are involved in the nature of the developed jet. For instance, the composition of the starting solution controls the formation of gel shape which can turn into bulk, fiber, or coating film [9]. The type of deposition technique is governed by several constraints like the thickness of the coating layer, the adhesion, the shape and type of substrate, etc. Dip and spray coatings are the commonly used coating methods in sol-gel coating technique, however, spin coating, laminar flow coating and printing are also used [10][11][12][13][14][15]. Controlling sol-gel process parameters has a great role in obtaining enhanced properties for the desired application. For example, Kumaresavanji et al [16] studied precursor molar concentration, viscosity and pH values as effective parameters on structural and magnetic properties of ferromagnetic La2/3Ca1/ 3MnO3 nanotubes. Moreover, Nguyen et al [17] succeeded in applying Anatase-Titania on χ-alumina fibers through sol-gel dip coating. The direct effect of several process parameters such as molar ratio of the precursors in TiO 2 -sols, dip-coating time, drying duration in air, heating processes and number of cyclical repetitions of the process on the structural and morphological characteristics of the coating was found.
In this work, the CoCuMnO x spinel ceramic film was applied onto stainless steel 304 substrate. Stainless steel (SAE 304) is considered the best-known austenitic grade of stainless steel. It has distinctive characteristics represented in forming and welding properties, plus the corrosion/oxidation resistance which makes it an excellent substrate for high temperature solar absorber application. The coating precursor was prepared and deposited through sol-gel dip coating method. The effect of process parameters such as precursor concentration, withdrawal rate of dipping, number of coating layers and heat treatment temperature on optical properties was studied.

Experimental work
The experimental work is held through four steps which are Base metal pre-processing, Sol-gel preparation, dip coating and samples post-processing. These steps are discussed in the next subsections in details.

Base metal pre-processing
Prior to dip-coating, substrate material has to be specially prepared in order to maintain acceptable adhesion of the coating to the base metal. At first, stainless steel samples of 1.5 mm thickness were cut into strips of 3×10 cm. Then, surface roughness of the samples was increased to 1.53 μm by sandblasting. This ensures enhanced mechanical adhesion between the coating and the substrate. Surface roughness was measured by Time TR110 surface roughness tester. Last step involved ultrasonic cleaning of the samples by immersing them into acetone and sonicating them in ISOLAB ultrasonic bath for 10 min to ensure the removal of any grease or sand residue. Samples were then dried by hot air and kept in a vacuum container ready for the coating process.

Solution preparation
The procedure of preparation of sol-gel coating of CoCuMnO x was previously discussed and shown in details in several previous research [4,5,18]. The main precursor of the solution is based on three main metal oxides; Co II acetate, Cu II nitrate and Mn II acetate. Solvents used were ethylene glycol and absolute ethanol. Four different coatings with different concentrations were prepared by dividing the molar ratios of metal oxides by 50, 60, 70 and 80 respectively.

Dip-coating process
Prior to dip-coating, the gel needed to be re-agitated at proper speed at about 50°C, to maintain its dense viscous form before dipping. As the coating thickness can result in different optical properties, this thickness was varied by changing the withdrawal rates or increasing the number of coating layers. Accordingly, three different withdrawal rates were used in this experiment; 0.5, 1 and 1.5 cm min −1 . In addition, single-layered or doublelayered coatings were also investigated. Intermediate heat treatment was applied on the first layer before the deposition of a new one. This was emphasized in [4] where it was stated that applying the subsequent coatings while the coating was still wet prevents the next coating layer from equally spreading over the whole substrate surface.

Heat treatment
After withdrawing the samples from the solution, they were held in a vertical position then inserted into a muffle furnace for heat treatment at 450°C for 30 min. Figure 1 shows the samples just after the dip coating process and ready for the heat treatment step, and a final sample after heat treatment. Other samples were heated at 700°C for 30 min. It is worthy to mention that the coating after heat treatment showed a homogeneous coating layer and a sufficient adhesion.

Characterization and testing
The optical properties including absorptivity and emissivity were measured at three different spots of the coating area for each sample. Then the average value was obtained for two samples. Absorptivity (α) was measured using T90+UV/VIS spectrometer PG instruments limited (scan range: 0.2-0.9 μm), while Emissivity (ε) was measured by Thermo Finnigan FT-IR spectrometer Nicolet 380 (scan range: 2.5-25 μm).
Absorptivity was calculated through the following equation: [19,20]  where: R is the reflectance obtained experimentally from FTIR in the entire scan wavelength range. P(λ) is the spectral radiance of a black body at a temperature 100°C coherent with the medium temperature applications. λ 1 , λ 2 are the limits of the wavelength range Selectivity factor was obtained using the formula (α −0.5 * ε) [ Coating thickness was determined by taking a section of the coated samples and measuring it through optical microscopic images obtained by OPTIKA B-500-MET optical microscope. A number of points was taken along the section and the average value was determined. Existing phases in the coating were determined by XRD using Bruker d8 Advance. SEM-EDX investigation for some samples of interest was also made using FEI-INSPECT S50 device.   Figure 2(a) shows the effect of changing precursor molar ratio on the absorptivity for different numbers of coating layers and different withdrawal rates. It was observed that the decrease in the precursor concentration increased the absorptivity in most cases for a single coating layer. The increase range was between 1%-2%, depending on the withdrawal rate. As for the double coating layer, increasing the precursor concentration increased the absorptivity. This matches with the emphasis of Zheng et al [22] where they stated that the increase in solar absorption is related to the increase in precursor concentration for the enhanced thickness cases.
As for the emissivity, figure 2(b) shows the effect of changing the precursor molar ratio on the absorptivity for different numbers of coating layers and withdrawal rates. It was observed that there was no significant difference in the emissivity values in most cases. Most of emissivity values were between 0.12 and 0.15.
Due to the slight differences in emissivity values that were found in most cases, the selectivity was significantly affected by the change in the absorptivity. Therefore, the calculated selectivity shown in figure 2(c) had the same trend of the absorptivity that was discussed earlier. Best results for absorptivity, emissivity and selectivity were attained by the precursor concentration of MR/60.

Effect of withdrawal rate on the optical properties of CoCuMnO x coated stainless steel
The withdrawal rate is one of the parameters that determines the thickness of the coating and accordingly, the optical properties. Figure 2(a) shows the effect of changing withdrawal rates on the absorptivity of different coatings. In case of single coating layer, it was observed that there was no significant change in the absorptivity with the increase of withdrawal rate in all concentrations proposed. As for the double coating layer, the absorptivity values tended to increase with increasing the withdrawal rate from 0.5 cm min −1 to 1.5 cm min −1 . This increase was about 1% for MR/50, 3% for MR/60, 2% for MR/70 and 1% for MR/80.
The relation between emissivity and withdrawal rate is shown in figure 2(b). The emissivity in the case of single coating layer tended to increase from 0.24 to 0.31 for the precursor concentration (MR/50) as the    figure 2(c). The best result of absorptivity, emissivity and selectivity were reported for the withdrawal rate of 1.5 cm min −1 . Sánchez et al [23] discussed the effect of withdrawal rate on optical properties and found a strong relation between the withdrawal rate and the coating thickness which subsequently affected the optical properties. Similar investigations were also reported in [18].

Effect of number of applied coating layers on the optical properties of coated stainless steel
The absorptivity values increased when a second layer of coating was added in the concentrations of MR/50 and MR/60. However, the concentration of MR/70 and MR/80 behaved in an opposite way as shown in figure 2(a). The biggest change in absorptivity was in the case of precursor concentration (MR/60) and 1.5 cm min −1 withdrawal rate. The absorptivity in this case increased by 4% when a second coating layer was added.
As for the emissivity, shown in figure 2(b), it was observed that the emissivity values tended to decrease slightly in most of the cases when a second coating layer was added.
Combining absorptivity and emissivity values together, an increase in selectivity values in (MR/50) and (MR/60) concentrations was found. On the contrary, decreasing selectivity was observed in (MR/70) and (MR/ 80) as shown in figure 2(c). The best absorptivity, emissivity and selectivity results were 0.906, 0.116 and 0.85 respectively and they were attained by the double coating layer.

Effect of heat treatment on the optical properties of coated stainless steel
Raising the heat treatment temperatures can result in enhancing optical properties. Therefore, the present samples were heated at 450°C and at 700°C. It was found that the absorptivity increased by 5% for (MR/60) when the temperature increased from 450°C to 700°C. This was previously found in [22] where it was stated that absorptivity increased with the temperature increase. However, the molar concentration has to be taken into consideration as the effect of raising the heat treatment temperature was not efficient and declined in case of (MR/70) concentration and even a reduction in absorptivity at (MR/80) was found when heated at 700°C where the absorptivity decreased by 2%. This emphasizes the important role of precursor concentration on the optical properties. Detailed results are illustrated in figure 3(a).
As for the emissivity, figure 3(b), the values dropped by 3% in case of (MR/60) when the heat treatment temperature increased from 450°C to 700°C. In case of (MR/70), a slight decrease in emissivity by 0.3% was found while at (MR/80) an increase in emissivity by 2% was obtained. As a result, selectivity values were affected in return.
The overall selectivity was increased by 6% for (MR/60). This meant that optical properties were enhanced in case of heat treatment as previously mentioned in previous research. However, for (MR/70) the selectivity increased by only 1%, while it decreased by 3% for (MR/80). Illustration for these results is shown in figure 3(c).

Effect of coating layer thickness on optical properties
The term 'thickness sensitive spectrally selective' (TSSS) in paint coatings emphasizes the relation between coating thickness and optical properties. This relation was investigated by Ma et al [24] at different annealing temperatures and the results revealed an increase in coating thickness which resulted in lower selectivity. At 500°C, for example, the thickness range was between 2.41 and 5.16 μm, and the selectivity decreased by almost 50%. Figure 4 shows the effect of average thickness on optical properties. As it was mentioned in earlier research, dip coating in sol-gel technique did not result in a precise thickness as it was controlled by a large number of parameters. Therefore, several sample coating thicknesses were measured and the average values were illustrated.    thickness. It can be concluded that the thinner the coating layer, the better the optical properties obtained. Best selectivity value obtained was 0.85 for a thickness of 4.96 μm.

Coating characterization [XRD analysis and SEM investigation]
In an attempt to investigate the resulting compounds in each coating, XRD analysis was applied on five samples with different conditions. Table 1 shows the different conditions in each case that will be discussed through this section. Figures 5-7 show the XRD analysis of the five cases. The analysis showed that in all these cases, the following compounds are formed: cobalt copper, copper di-manganese oxide, copper oxide and manganese oxide.
Case 1. Shown in table 1 and figure 5, represents the optimum sample that attained the best optical properties. the average crystallite size in this case was 38 nm.
By raising the heat treatment temperature from 450°C to 700°C (Case 2 and Case 3 -MR/60), the average crystallite size of the coating compounds apparently increased from 41.07 nm to 77.53 nm with 88.8% relative increase as indicated in figure 6 and table 1. This was accompanied by an increase in absorptivity, decrease in emissivity and significant increase in selectivity from 0.78 to 0.84. The noticeable increase in the average crystallite size can be attributed to the agglomerations of fine particles which in turn led to an increase in the quantity of the fine pores that boosted light entrapment and consequently led to absorptivity enhancement as it was emphasized in [6].
As for raising the heat treatment temperature from 450°C to 700°C (Case 4 and Case 5 -MR/80), the average crystallite size increased from 33.4 nm to 54.23 nm with 62.4% relative increase as shown in figure 3(c). This was accompanied by a decrease in absorptivity, increase in emissivity and increase in selectivity from 0.82 to 0.79. This indicated that two opposite factors that can affect the optical properties were present: the precursor concentration and heat treatment temperature. The good crystallinity at a higher precursor concentration was previously introduced in [24,25]. Furthermore, the effect of increasing precursor concentration in enhancing the crystallinity was discussed in [26]. Accordingly, this validates the presented work. As tabulated in table 1, the crystallite size of case 2 which had precursor concentration of (MR/60) was larger than case 4 which had precursor concentration of (MR/80). Likewise, the crystallite size of case 3 was larger than case 5. This meant that decreasing the precursor concentration caused a direct decrease in the crystallite size of the coating compounds. Figure 8(a) shows SEM image of the sample with the best results of conditions MR/60 precursor concentration, 1.5 cm min −1 withdrawal rate and double coating layer. This case achieved 0.906 absorptivity, 0.116 emissivity and 0.85 selectivity. The shown image illustrates grooves and pores that play an important role in light trapping as discussed earlier. In addition, figure 8(b) illustrates the EDX analysis which shows the elements of the coating (Co, Cu, Mn and O) and the base metal elements (Fe, Ni and Cr).

Conclusions
Through this study, CoCuMnO x spinel ceramic film was applied onto stainless steel 304 through sol-gel dip coating method and proved to be successful as an effective solar selective coating. Several parameters were discussed that included precursor concentration, withdrawal rate, number of coating layers and heat treatment temperature. Best absorptivity, emissivity and selectivity measurements were 0.906, 0.116 and 0.85, respectively and was considered as the optimum case. The process parameters that resulted in this optimum case were MR/ 60 precursor concentration, 1.5 cm min −1 withdrawal rate, double coating layer and 450°C heat treatment temperature. The increase of coating thickness resulted in deterioration of selectivity. The optimum sample with the best optical properties had the smallest coating thickness which was 4.96 μm. Coating characterization proved that the improvement in optical properties was accompanied with an increase in the average crystallite size of the deposited compounds. This work has shown the significance of the tested parameters on the optical properties of solar selective coatings. It also emphasizes the role of the synergy of the different parameters in order to achieve best results.