3.2.2.1. Metal particles embedded in dielectric matrix
Many researchers in the field of SSA synthesis have investigated cermet selective absorbers using various synthesising methods. This is because the cermet structure is unique and is also one of the highest performance selective surfaces [75]. However, to the best of our knowledge, the synthesis of this type of absorber using sol-gel methods is relatively scarce. Eisenhammer et al. [122] patented the idea of metal/conductive particles in alumina, with either Al65Cu20Ru15 in alumina or TiN in alumina, as a SSA. Each composite was obtained by
mixing the conductive particles with an alumina matrix sol precursor. The alumina sol precursor was prepared by dissolving niobium chloride (NbCl5) in butanol and mixing with
sodium butoxide (Na(OBu)n) under reflux conditions. This produced Nb(OBun)5 which was
subsequently mixed with glacial acetic acid to form the alumina sol precursor. Eisenhammer et al. [122] also investigated another route to prepare the alumina sol precursor by mixing boehmite with HNO3 at 550C. For the synthesis of quasicrystal Al65Cu20Ru15 conductive
38
particles in alumina film, the particles were mixed with the alumina sol precursor solution which was then sprayed onto a copper substrate and heat-treated at 6000C. For synthesis of
TiN conductive particles in alumina film, the particles were dispersed into the alumina sol precursor solution, then coated on the copper substrate by centrifugation (spin) and finally heat-treated at 6000C. The Al65Cu20Ru15 alumina layer had a thickness of 110 nm and a
volume fraction of 30%, whereas the TiN-alumina layer had a thickness of 130 nm and a volume fraction of 20% [122]. However, they did not show any absorptance and emittance values, but from the curves created in their patent, these two SSAs can be categorized as having comparable selectivity values.
Bostrom and co-researchers [22, 37, 39, 51, 64, 92, 123, 124] have synthesised nickel nanoparticles embedded in an alumina ceramic matrix (Ni-Al2O3) thin film on a smooth and
highly specular aluminium substrate using a sol-gel-like method. They reported that although the sol-gel methods have been known to fabricate a wide variety of materials for many decades, it was only in the last few years that the solution-chemistry or sol-gel science was found to be a suitable method to produce nanoparticle composites appropriate for thermal solar absorber applications [20, 22]. Precursor solutions of nickel and pure amorphous Al2O3
in different proportions were mixed to control the nickel to alumina ratio in the final absorbing films [22]. Film deposition was conducted via spin-coating at 3700 rev/min for 20 seconds before the film was heat-treated to temperatures of 550-580oC in an oxygen-free
glass tube. During the heat treatment, solvents were evaporated and the only substances left in the final film coating were alumina and metallic nickel [22, 51]. The thin films produced were smooth and homogeneous with nickel content of up to 80% of the volume fraction. The optimal single layer of coating had a nickel content of 65 volume %, a thickness of 100 nm and particle size between 5-10 nm. This absorbing layer showed a solar absorptance of α =
39
layer on top of the absorbing layer enhanced the performance of the absorber. The optimum anti-reflection-coated sample reached a solar absorptance of = 0.93 and a thermal emittance of = 0.04. These results showed that the Ni-Al2O3 cermet film had excellent spectrally selective optical properties. They suggested that for constructing a Ni-Al2O3 absorber layer
more efficiently, the bottom part of the layer should have high nickel content while at the top it should have minimum nickel content [22]. The use of a rough aluminium surface as a substrate was also implemented in this research, but the results were less satisfactory than the smooth substrate.
Further investigations by Bostrom and co-researchers [92] focused on improving the selectivity and durability of the nickel-alumina cermet and enhancing the performance of the AR coatings. They reported that the performance of the nickel-alumina selective absorber thin film system was improved if a three-layer system was applied. This system was composed of an 80% nickel and 20% alumina film with thickness of 103 nm at the base (first layer), a 40% nickel – 60% alumina film with the thickness of 59 nm in the middle (second layer) and a silica/hybrid-silica film with the thickness of 90 nm at the top (third layer/AR layer). This optimal three-layer system showed a solar absorptance value of 0.97 and a thermal emittance value of 0.05 [37, 39, 92, 124]. These results were comparable to commercial products. These synthesis processes were simple and cost-effective but the nickel-alumina solution was unstable and agglomerated to form precipitates within 24 hours, thus reducing the reproducibility of this system, even though the stability can be enhanced for up to one week in a methanol solution [51]. The calcination step also required strictly oxygen-free conditions, which was troublesome. This absorber has been industrially produced on a pilot scale since 2009 and the company is working on having a full-scale process in the near future.
40
Another effort to improve the nickel-alumina SSA coatings synthesised using a sol- gel-like method was carried out by Nejati [32]. Nejati used nickel nitrate and alumina powder as precursors. Nickel nitrate was first dissolved in distilled water or ethanol, then, while stirring, alumina powder was gradually added. The prepared mixture was then dispersed mechanically using a dissolver and ultrasonication. To avoid agglomeration, the temperature was strictly controlled and different additives such as a wetting agent; a coupling agent and a dispersing agent were added to the suspension before dispersion. Cleaned aluminium substrates were then dip-coated in the suspension at different speeds. The wet films were dried for 30 minutes at 120oC and then quickly annealed for 1 hour at 450oC in a hydrogen
atmosphere. Nejati found that the mechanical properties of a pure Ni-Al2O3 cermet composite
layer and the substrate were poor and the layers were easily removed during the tape test. Nejati did not use only TEOS as a source of silica for the AR layer but also used it to improve the bonding ability between the absorber thin film and the substrate (the silica was also used as an underlayer). Adhesion and scratch resistance of the thin film was improved significantly. The silica network formed after the addition of TEOS also enhanced the solar absorption by lowering the effective refractive index of the film. However, although the addition of the silica AR layer increased the solar absorptance value, it also increased the emittance value slightly. The best result was shown by a sample with absorptance value of = 0.94 and emittance value of = 0.11 [32]. Based on accelerated ageing and humidity studies, Nejati estimated that the nickel-alumina absorber was suited for glazed collector applications such as domestic solar water heaters operating at low temperatures. Due to the promising optical performance and good thermal and humidity stability, the developed absorber film could compete with sputtered absorber films [32].
41
3.2.2.2 Carbon particles in dielectric matrix
3.2.2.2.1. Carbon particles in silica
Katumba et al. [41] outlined the reasons for studying carbon-in-silica tandem selective solar absorbers. Firstly, both carbon and silica are abundant, environmentally-friendly and stable materials. Secondly the sufficiently small size of carbon particles, approximately 10 nm or less, have a high absorption cross-section for UV-VIS radiation [41, 125]. Finally, carbon-silica composites could be synthesised easily via sol-gel techniques.
Mastai et al. [126] introduced a new concept for the design of carbon-silica based SSA materials. This group showed that porous carbon-silica hybrid nanocomposites have SSA characteristics. The synthesis of this composite involves a sol-gel-like method to perform a direct carbonization in the nanoconfinement of porous silica leading to the formation of nano-sized amorphous carbon particles. Materials used included sugar as a precursor of carbon, and cyclodextrins (CD) and polystyrene-polyethylene oxide (SE) as precursors of CD-based silica and SE-based silica, respectively. In such a structure, solar radiation was absorbed and transferred into heat without infrared (IR) re-emission. The carbon nanoparticles contributed to high absorptance and thermal stability, whereas silica contributed a transparent matrix and binder material. Especially in the case of CD-based silica, the overall processes were ideal because of cheap and “green chemistry” conditions. Also, sugar was easily available and non-toxic. This composite was obtained under one-pot synthesis conditions with the elimination of water. No removal or addition of any further chemical was necessary to obtain the non-toxic carbon-containing silica. In addition, leaching of the final material was practically impossible and if that did happen it would only release materials that were already abundant in nature [126].
The absorptance and emittance values for the SE-based carbon-silica composite were
42
absorptance value was 0.92 and emittance value was 0.13. All samples showed good stability under the influence of humidity and high temperatures. Based on the nature of the components involved in this composite, it could be assumed that long-term stability of the samples was likely to be high. The degradation of solar thermal absorber coatings which is usually caused by thermal oxidation of metal particles did not happen in this composite [126].
In separate but related research, Katzen et al. [127] created a carbon-silica nanocomposite film selective absorber on a glass substrate. The film was prepared using the sol-gel spin-coating method. The silica sol preparation was followed by CD-based silica- carbon composite preparation. β-methylated cyclodextrin (2 g) was dissolved in 3 g of
aqueous HCl (pH 2) and 4 g tetramethylorthosilane (TMOS) were added while stirring until a homogeneous solution was produced (within a few minutes). The films were deposited by spin-coating at 4000 rpm for 1–2 min. The films were dried at room temperature and annealed in an oven under nitrogen (95%) flow at a temperature of 850 K (increasing at 20 K/min intervals). All films prepared by this method contained approximately 15% carbon. It was found that the best thin film silica-carbon nanocomposite (thickness 1000 nm) showed α
= 0.94 and ε = 0.15. The films showed good stability under the influence of humidity, as they
were held above a water bath at 1000C for 5 hours and in a high temperature environment
(250–300oC) for 48 h [127].
Another sol-gel method used to synthesis carbon-silica thin film composites for solar selective absorbers was reported by Katumba et al. [41]. The processes consisted of using a silica-carbon precursor sol, which was spin-coated onto a metal (specular and rough aluminum and stainless steel) substrate and carbonizing it in an inert atmosphere. Samples were made from silica sols based on acid-catalysis of TEOS and water that were impregnated with sucrose (SUC) as the carbon precursor. Four categories of samples were studied. These were the tetraethyl-orthosilicate only (TEOS-only), methyl trimethoxysilane (MTES), acetic
43
acid anhydride (Ac2O) and soot (SOOT) samples. In this case, MTES and Ac2O functioned as
organic modifiers of inorganic silica.
The spin-coating technique produced films with very flat surfaces and uniform thicknesses in the 1 μm range. The fine structure showed homogeneous mixing of the carbon
and silica in the TEOS-only samples, while the addition of both MTES and Ac2O resulted in
the segregation of silica and carbon at the nano-scale. However, the addition of 20 wt % MTES or 15 wt % Ac2O to the TEOS-only sols helped to reduce the cracks in the TEOS-only
samples. The samples with 20 wt% MTES had a solar absorptance of α = 0.74 and thermal
emittance of ε = 0.30 while the corresponding values for samples with 15 wt % Ac2O had α =
0.81 and ε = 0.44 [41]. The addition of soot did not yield a net advantage.
3.2.2.2.2. Carbon particles in ZnO, NiO and TiO2 matrices
Carbon nanoparticles dispersed in ZnO and NiO dielectric matrices on aluminium substrates, to be used as SSAs, have been prepared by Katumba et al. [6]. The sol-gel-like method used to prepare these samples was closely related to the method of Liu et al. [128]. Appropriate amounts of zinc acetate dihydrate and nickel acetate tetrahydrate were separately dissolved in 50 ml of anhydrous ethanol and stirred by a magnetic stirrer at room temperature. Diethanolamine (DEA) was added as a chelating agent in a way that the molar ratio of each type of acetate to DEA was maintained at 1:1. These solutions formed the ZnO and NiO precursors. Sucrose was dissolved in distilled water in the mass ratio 1:1 prior to mixing with the matrix precursor solutions. This constituted the carbon precursor solution. The oxide and carbon precursor solutions were mixed and stirred again. After a period of stirring, 1 g of polyethylene glycol (PEG) was added to the ZnO and NiO matrix precursor sols. The resultant solution was stirred further until the formation of a sol which was immediately spin-coated onto pre-cleaned aluminium substrates. The PEG was used as a
44
structure-directing template. The spin-coated samples were then calcined in a tube furnace with nitrogen-flow at 550oC for 1 h to carbonize the carbon precursor and also to dry and
solidify the oxide matrix. This method ensured an even distribution of the carbon nanoparticles in the oxide matrices [6]. The absorptance and emittance values achieved were
α = 0.84 and ε = 0.04 for C-NiO and α = 0.71 and ε = 0.06 for C-ZnO. SEM analysis revealed
a smooth surface for both C–ZnO and C–NiO samples, but other C–NiO samples showed dendritic characteristics. The coatings contained amorphous carbon embedded in nanocrystalline ZnO or NiO matrices. Explorations with a Selected Area Electron Diffraction (SAED) instrument showed that a small amount of Ni grains of 30 nm diameter also existed in the NiO matrix. Both C-ZnO and C-NiO also had grain sizes for the carbon clusters in the range 55–62 nm and a crystallite size of 6 nm as indicated by Raman spectroscopy [129]. The accelerated ageing tests in a weather chamber with a high relative humidity environment of 95% and a temperature of 450C for 600 h showed that the C–NiO sample maintained better
performance than the C–ZnO sample [6].
Titanium dioxide (TiO2) has also been used as the host for carbon particles for SSA
applications. Rincon et al. [42] synthesised carbon blacks (CB) and carbon nanotubes (CNT) embedded in a TiO2 matrix deposited on polished stainless-steel substrates. The use of CNT
could bring interesting optical properties to the composite because it is highly anisotropic. These researchers used a chemical method based on sol–gel techniques. They first prepared a TiO2 matrix where the large refractive index of TiO2 was decreased by increasing the
porosity of the oxide with the addition and subsequent thermal elimination of polyethylene glycol (PEG). The CB-TiO2 composite films were prepared by dissolving CB in an alcohol
solution prior to the addition of titanium tetraisopropoxide and then adding a minor amount of dispersing agent. Then the solution was stirred for 24 h at room temperature. Films were fabricated on stainless-steel substrates by dip-coating. Different dipping speeds and a number
45
of dipping and drying cycles were combined with a final annealing temperature to control the thickness and microstructure of the deposited film.
The CNT-TiO2 film was prepared using a layer-by-layer strategy whereby the first
layer contained the TiO2-PEG coating and the second layer contained the CNT-PEG film.
The mixture of PEG, CNT and dispersing agent was sonicated for 90 min to obtain an ink, 30 ml of 2-propanol was added and the emulsion was centrifuged for 30 min. The supernatant liquid was discarded and the nanotubes were re-dissolved in water and centrifuged again. In the third centrifugation, the nanotubes could no longer be precipitated and the ink was stable for weeks. This ink was kept under vigorous stirring for 2 days prior to being coated on the TiO2 substrates (first layer). Five to fifteen immersions/drying cycles produced TiO2 film
thicknesses in the range of 150–650 nm after annealing. Once the first layer was obtained, it was rinsed in distilled water and then in a mild solution of NH4OH before being rinsed again
in distilled water. This sequence eliminated the excess PEG and HCl from the sol–gel bath. The TiO2-coated stainless-steel substrates were laid on a plane tilted at 8º to the horizontal
where a fixed amount of the CNT emulsion was cast and let dry for 24 h. Further drying at 1000C for 24 h, followed by annealing in N2 at 4000C for 15 min, eliminated any remnants of
PEG and additive [42]. The research group found that the sol-gel-like method investigated in their work proved successful in producing coatings with good spectrally selective properties, but they did not give any specific absorptance and emittance values. The coating system was thermally stable and free of corrosion problems due to the hydrophobic nature of carbon.
Despite many advantages in the use of carbon absorbers in the matrices described above, the optical performance of these absorber materials has not yet reached a satisfactory level. As such, there are still many opportunities for further research to improve their optical properties and durability before commercialization. In addition to that, the processes involved
46
are relatively long and cumbersome since an inert environment is required in the carbonization process.