Solar selective absorber surfaces using spinels

3.2.3.1 CuFeMnOx and CuCoMnOx spinels

During the last decade, spinels deposited on highly reflective metal substrates have attracted considerable interest due to their promising properties as SSAs for solar thermal collectors. The term “spinel” refers to a group of minerals which crystallize in a cubic (isometric) crystal structure. Kaluza et al. [71] have succeeded in synthesising CuFeMnO4

black film spinel SSAs using sol-gel dip-coating and heat-treatment at 500 C. Mn-acetate, Cu- and Fe- chloride precursors were used in a molar ratio of 3:3:1, respectively. To protect the spinel from corrosion, a 3-aminopropyltriethoxy silane (3-APTES) silica precursor was added to the Cu, Mn and Fe sol precursors with molar ratio of (Mn-Cu-Fe):silica = 1: 1. Analytical results showed that the films consisted of two layers: the lower was amorphous SiO2 and the upper was a spinel having the composition of Cu1.4Mn1.6O4. The films exhibited

absorptance values of around  = 0.6 and emittance values of  = 0.29–0.39. The low performance was caused by the difference in the film thickness between the spinel and the silica layer where the absorbing spinel layer film (200 nm) was much thinner than the amorphous SiO2 layer (800 nm). The large thickness of the SiO2 layer increased the thermal

emittance of the composite films due to the strong phonon absorption of the Si-O stretching modes at 1100 cm-1. The absorptance value could reach 0.93 when a base catalyst (NH3)aq

was added to the precursor in the solution preparation, but the thermal emittance value became very high (ε = 0.62) due to the presence of large SiO2 spherical particles (400 – 420

47

Efforts have made to improve the optical performance of the CuFeMnO4 black film

spinels. Kaluza et al. [20] reported that the emittance value could be decreased by substituting silica with zirconium oxide (ZrO2), but the presence of the ZrO2 generated a brown hue color

which caused the absorptance value to drop. This research group subsequently modified the synthesis route using Fe-, Cu-, Mn-acetate precursors. After undertaking thermal hydrolysis steps, they succeeded in making CuFeMn-oxide spinels which did not contain thermally emitting components such as SiO2 or ZrO2, consequently enhancing the spectral selectivity of

the coatings. However, these CuFeMnOx films exhibited a reddish-brown hue, which

originated from the segregated Fe2O3 phase formed during heat treatment at 500 C. As a consequence, the films showed a lower solar absorptance. To address this problem, Fe was substituted with Co and it was expected that even if a segregated Co-oxide phase was formed, the color of the oxides would be black due to the allowed interband transitions of Co-oxide [20].

To prepare CuCoMnOx spinels, Kaluza et al. [20] used an ethanolic sol based on Mn-

acetate and Co- and Cu- chloride precursors. The solution was stirred at 60C until a viscous sol (40 ml) was obtained. A part of the dark greenish viscous sol (6.5 g) was then diluted in a MeOH/H2O mixture (40 g/4.6 g) to obtain optimum viscosity for dip-coating deposition. The

viscosity of the sol solution was further adjusted by the addition of thickening agent hydroxypropylcellulose (HPC), which also contributed to the stability of the solution. The films were dip-coated onto aluminum substrates with a dipping speed of 10 cm/min. To obtain coatings with different spectral selectivities, the film thickness was varied by changing the concentration of the thickening agent and the number of dipping/annealing cycles. Films deposited on aluminum were then annealed at 5000C for either 15 min or 1 h [20]. The best

CuCoMnOx film demonstrated an absorptance of  = 0.9 and an emittance of = 0.05 when 10 wt% of HPC was added to the sol precursor. This result proved that the CuCoMnOx spinel

48

was a promising candidate for a solar absorber coating material. Thermogravimetric analysis showed that the xerogels became crystalline at 3160C while X-ray diffraction analysis

revealed that the coatings and powders consisted of predominantly CuCoMnOx spinels.

Rutherford back scattering (RBS) and transmission electron microscopic (TEM) studies, combined with energy dispersive X-ray spectroscopy (EDXS) measurements, confirmed that Cu, Mn and Co were present in the films in stoichiometric ratios close to that in the initial sols. In addition, CuCoMnOx spinels exhibited relatively weak phonon absorptions at 600

cm-1, i.e. below the peak of black-body thermal radiation [20].

To enhance the absorptance value of CuCoMnOx spinels, Vince et al. [29] attempted

to modify the spinel. They made two different types of films, namely, Ti-doped (up to 30%) CuCoMnOx and undoped CuCoMnOx films. The precursor ratio of Co:Cu:Mn was 1:3:3. The

films were subsequently annealed at 450oC for 15 and 30 min in air. To improve the stability

(weather and abrasion resistance) of the films, two kinds of protective over-coatings were tested: one over-coating was based on polysiloxane resin and the other based on the high- density of silica. Results indicated that undoped CuCoMnOx films with SiOx protective over-

coatings exhibited absorptance values of  = 0.85-0.91 and an emittance value of  < 0.036 after just a single dipping/annealing cycle. All investigated films exhibited poor stability during a boiling-water test (>2 hours) before protective over-coatings were applied. When an over-coating based on high-density silica or polysiloxane resin was applied to either doped or undoped CuCoMnOx films, both of them remained unaffected by the test.

Another shorter and easier method which was used to synthesis CoCuMn-spinel solar selective absorbers was reported by Bayon et al. [130]. Copper, cobalt and manganese nitrates were dissolved in absolute ethanol at various molar ratios. A complexing agent and a wetting additive were also added to stabilize the solution and improve the film adherence. Depositions were performed using dip-coating at different withdrawal rates on aluminium

49

foil, borosilicate glass and stainless steel substrates. The resulting layers were sintered in an oven at 500ºC. A silica AR layer was also deposited by a sol-gel method on top of the spinel absorber. The highest solar absorptance of the resulting film (α = 0.863) was reached with

only one layer of absorber material when the spinel was deposited at 15 cm/min and the molar ratio in solution was 1Cu:0.5Co:1Mn. The solar absorptance was improved to 0.908 when a SiO2 antireflective layer was deposited onto the spinel. Long term stability studies

showed that the CuCoMn-spinel was a very stable material. This study showed that the metallic ratios in the film were very close to the precursor ratios in the dipping solution. XPS measurements have shown that different oxidation states can be found for the metals present in the spinel: Cu+, Cu2+, Co2+, Co3+, Mn2+ and Mn4+ [130]. Although the CoCuMnOx

synthesised via this method is often contaminated by some metal oxides, chlorides, and

oxychlorides, it is better than the co-precipitation method. This is because in the co- precipitation method, it is difficult to control all of the metal cations that precipitate from the solution and which, at the same time, result in composition segregation and low yield [131].

3.2.3.2 CuMnOx spinels/CuMn oxide

A simpler CuMnOx spinel which contains less than three metal components and is

derived from the CuCoMn-spinel also shows the characteristics of a SSA. Bayon et al. [24] reported that CuMn-spinel thin films on aluminium foil synthesised by a sol-gel-like dip- coating method and followed by air-sintering at 5000C could be used as a low temperature

application of SSA. Copper and manganese nitrate precursors were dissolved into the absolute ethanol solutions with the addition of a complexing agent and a wetting additive to stabilize the solution and improve the film adherence. Analysis of the composition showed that the metallic ratio in the film was very close to the ratio in the dip solution and indicated the formation of a spinel-like material with Cu1.5Mn1.5O4 stoichiometry. The annealing time

50

and temperature also influenced the film composition and optical properties. A solid-state redox reaction occurred when temperatures higher than 450oC were applied [78]. The highest

solar absorptance of  = 0.87 is reached by using a one layer film deposited from solutions containing a molar ratio Cu/Mn = 1 and prepared with a withdrawal rate of 20 cm/min. The optical property of the film was dramatically improved by subsequently depositing a SiO2

anti-reflective layer using a sol-gel technique onto the spinel. By optimizing the film thickness of both CuMn-spinel and SiO2 layers, the best absorptance and emittance (at

100oC) values achieved were 94% and 6%, respectively. These results showed that it was

possible to obtain a very good selective absorber with only two layers (absorber layer and anti-reflective coating) from cheap materials and by using a simple dip coating deposition method [24]. Although the optical performance of this spinel oxide solar absorber was quite promising, it was still not high enough to be competitive in the market. The absorptance of this absorber surface could be improved to 0.95 by introducing an additional CuMn-oxide absorber layer (a total of 3 layers) [38]. Thermal stability and humidity tests were conducted based on the method developed by the International Energy Agency (IEA) within the Solar Heating and Cooling (SHC) Program Task X for low-temperature SSAs [88, 91]. The results of a preliminary up-scaling study revealed that it was possible to deposit CuMn-oxide absorbers on large-area substrates and that they could be a good alternative to the materials present today in the market, not only in terms of optical properties but also in terms of long term durability [38].

Overall, the general strategy to implement sol-gel methods for the synthesis of absorber-reflector tandem structures (non organic binder) suitable for SSA materials is shown in Fig. 3.1. The absorptance, emittance and selectivity of various SSAs produced by sol-gel methods to date are summarized in Table 3.1.

51

Figure 3.1. General strategy for synthesising metal oxide/spinels (route A) and metal/carbon particles embedded in non-organic matrix/binder (route B) solar selective absorbers.

Table 3.1. Summary of absorptance (α) and emittance (ɛ) of various SSA materials produced by sol-gel methods

Sol-gels SSA Materials and Substrates α ɛ Reference Metal Oxide Based Absorber

Bare CuO on aluminium 0.93 0.11 (80oC) [95] CuO-SiO2 on stainless steel 0.92 0.2 [93] Black cobalt on galvanized iron 0.91 0.12 (100oC) [104]

Cobalt oxide on stainless steel 0.93 0.14 (100oC) [106] CoFeO on stainless steel 0.94 0.2 (100oC) [108] Cobalt oxide on stainless steel 0.77 0.2 [107] Cobalt oxide-nickel oxide on mild steel 0.9 0.1 (80oC) [101] Black cobalt on stainless steel 0.88 0.12 [102] Black cobalt-tin oxide on nickeled stainless steel 0.72 0.037 (100oC) [40]

Cobalt oxide-copper oxide on stainless steel 0.84 0.28 [119] Ruthenium oxide on the ASTM grade 2 titanium 0.74 0.12 [120] Nickel oxide - alumina on aluminium 0.92 0.03 [8]

Cermet based absorber

Nickel-alumina cermet on aluminium 0.97 0.05 [39]

Carbon-silica on glass 0.94 0.15 [127]

Carbon-NiO on aluminium 0.84 0.04 [6]

Carbon-ZnO on aluminium 0.71 0.06 [6]

Spinels based absorber

CuCoMnOx on aluminium 0.9 0.05 [20]

CuCoMnOx-SiOx on aluminium 0.91 0.036 [29]

52 3.3. Effect of Silica Thickness

Various SSAs, whether synthesised by sol-gel or other methods, often involve the incorporation of silica to improve their selectivity or durability. The deposition of a silica layer, especially silica as an AR layer, usually necessitates a sol-gel technique even though the absorber film was deposited by other methods than sol-gel. In this review, the use of silica (SiO2) either as anti-reflection (AR) layer, matrix or underlayer has been mentioned. However, the use of silica as a protecting agent (matrix or underlayer) of the absorber film has had an unfavourable influence on the optical performance. High emittance values are the consequence of the incorporation of silica as a matrix and/or an underlayer because the silica absorbs too much solar radiation in IR range [20, 32, 71, 93], while silica as an AR layer has a more positive effect because it can improve absorptance with a non-significant influence on the increase of the surface emittance value [24, 29, 92]. Silica as an AR layer is frequently synthesised thinner than the silica as a matrix or underlayer, so, in the construction of a SSA protective layer (matrix or underlayer) involving silica, the protective layer thickness should be an important factor to be optimized. The AR layer or other protective upper coatings should normally be within 50-70 nm or in the scale of tens of nanometers [132].

53