Paper published on: Solar Energy Materials & Solar Cells, volume 144 (2016), pages 608-615 DOI: 10.1016/j.solmat.2015.09.068 http://www.sciencedirect.com/science/article/pii/S0927024815005103
Optical properties of ZrB2 porous architectures Elisa Sani1*, Elena Landi2, Diletta Sciti2, Valentina Medri2 1
2
CNR-INO National Institute of Optics, Largo E. Fermi, 6, I-50125 Firenze, Italy CNR-ISTEC,Institute of Science and Technology for Ceramics, Via Granarolo 64, I-48018 Faenza (Italy) * corresponding author:
[email protected]
Abstract Porous ceramic materials are currently used as volumetric sunlight absorbers in concentrating solar power systems. As the efficiency of thermodynamic cycles rapidly increases with the operating temperature, the favorable characteristics of so-called ultra-high temperature ceramics (UHTCs) can be successfully exploited in novel solar absorbers. The present work reports, for the first time to the best of our knowledge, on optical properties and microstructural analysis of novel ice-templating porous ZrB2 UHTCs, to evaluate their potential as volumetric solar receivers. We demonstrate that the different complex structures that can be obtained with the freeze casting technique show promising optical properties. The idea of conjugating an highly tailorable morphology, useful for optimizing gas fluxes and heat exchanges between absorber and gas, to the spectral selectivity which is a characteristics of ZrB2 can be a promising route for increasing the efficiency of thermal solar systems. Keywords: borides; freeze casting; optical properties; solar absorbers; concentrating solar power.
1
Paper published on: Solar Energy Materials & Solar Cells, vol. 144 (2016), 608-615 http://dx.doi.org/10.1016/j.solmat.2015.09.068 1. Introduction A general rule for concentrating solar power systems is that the efficiency rapidly increases with increasing working temperature. Thus, for solar thermal exploitation, a large effort is put for raising the system operating temperatures by developing novel solutions [1-4]. For the plant approach with central solar tower, a critical parameter for temperature increasing is the choice of the receiver material, which is devoted to absorb the sunlight collected by the mirrored surface and to efficiently transfer the energy to the thermal exchange medium, keeping at minimum radiative losses due to thermal emission. Porous ceramics are presently used as volumetric absorbers of solar radiation for application in concentrating solar power systems. Air at relatively low temperature is drawn through the porous absorber toward the back, gradually being heated by convection from the solid absorber. The gas temperature at the front face of the absorber should then be relatively low, while the temperature at the back can be much higher. To date, existing volumetric absorbers produce high temperature at the front face, but they do not yet provide the level of performance needed for efficient collection and conversion of the solar radiation. The needed superior performance of high-temperature volumetric absorbers can be achieved by optimization of the morphology and optical properties of the porous absorber. This includes the creation of gradients of absorber structure and composition to allow fine control of optical absorption rate and convective heat transfer. Besides metals and metallic alloys [5], few ceramic materials have been studied up to now for this application, basically belonging either to the family of silicon carbide (SiC, a grey semiconductor with good sunlight absorption and high oxidation resistance) and SiC-based materials [3, 6] or being white ceramics like alumina [4] or cordierite [6], which are characterized by very high thermal stability, oxidation resistance and high refractoriness, but with non-optimal sunlight absorption properties due to their white color. Zirconium, hafnium and tantalum boride, nitride, and carbide-based materials and composites are called ultra-high temperature ceramics, UHTCs, because they combine outstanding robustness and refractoriness (melting temperatures above 3000°C) [7] and are widely recognized as the unique materials for harsh service environments, especially for aerospace, rocket propulsion and energy applications. Very recently, different UHTCs have been proposed for solar thermal absorbers [8-12] thanks to their spectral selectivity properties and low thermal emittance. ZrB2 has already proved to be a suitable material for this application from several points of view [13-15]. The ultra-high melting point of ZrB2, together with the unique combination of good thermal conductivity and chemical stability appear intriguing for employing it in high temperature novel solar furnaces, where the front surface of the absorber can even reach temperatures as high as 1700K [16]. Moreover, both room-temperature reflectance spectra and high-temperature emittance favourably compare ZrB2 against other ceramics used as volumetric absorbers such as SiC. In this work we characterize, for the first time to the best of our knowledge, ZrB2 porous architectures obtained by freeze-casting of aqueous suspensions [17, 18]. With this technique, porous structure with main unidirectionally oriented pores can be produced [19] . Changing the slurry characteristics (for example powder composition, solid load, type and amount of dispersant) or the temperature gradients during freezing, the macro-structures can be tailored. The ice network results in macro-pores mainly aligned in the same direction or arranged into randomly distributed dendritic forms. Several samples are investigated including both monophasic ZrB2 and composites containing SiC particles, and their structural and optical properties are assessed and compared.
Paper published on: Solar Energy Materials & Solar Cells, vol. 144 (2016), 608-615 http://dx.doi.org/10.1016/j.solmat.2015.09.068 2. Materials and Methods 2.1 Materials The starting powders were commercial products: ZrB2 powder (H.C. Starck grade B, Karlsrhue, Germany), β-SiC powders (H.C. Starck BF12, Karlsrhue, Germany), α-Si3N4 powder (Baysinid, Bayer, Leverkusen, Germany). Different aqueous suspensions were prepared using pure ZrB2 (labelled as M=“monophasic”) or the composite (vol%): 76 ZrB2–20 SiC– 4 Si3N4, (labeled as C=”composite”), where SiC is added as a reinforcing phase whilst Si3N4 is just a sintering additive to enhance densification [20]. Two commercial ammonium poly-acrylates, namely Duramax D3005 and Dolapix PC33, with different molecular weights and pH range of activity were used as dispersants [21]. Duramax D3005 is already known to promote a better stabilization of the ZrB2-SiC slurry compared to Dolapix PC33 [22]. In order to modify the porous structure (lamellar or dendritic) according to the effect of the slurry formulation or the cold transmission [17], the powder suspensions were produced by varying the dispersants type and amount and/or the solid loading (35-48 vol%), while the slurries were cast into plastic or metallic cylindrical molds with different diameter/height ratios (1 - 4), as reported in Table 1. Further details concerning the influence of the processing parameters on the developed micro-macrostructures were previously reported [17, 18]. Sample label
Dispersant
Solid loading (vol%)
mold
Type
(wt%)
Type Diameter/height
M1 M2 M3
Duramax Dolapix Dolapix
2.9 4 4
35 45 45
metal metal plastic
4 1 1
C1 C2 C3 C4 C5
Duramax Dolapix Dolapix Dolapix Dolapix
5.8 6 6 4 6
38 48 48 48 48
plastic plastic plastic plastic plastic
4 1 4 1 2
Table 1: Powder suspensions and molds details. The slurries were freeze cast and dried (Edwards Mod.MFD01, Crawley, UK) using a cooling temperature of -40°C and chamber vacuum value of 10 Pa. Samples were pressureless sintered at 2100° C for 1 h in flowing argon. The morphological and microstructural features of the as-produced materials were observed by SEM (E-SEM FEI Quanta 200, FEI Company). Pore size distribution and open porosity in the range 0.0058 - 100 μm were analysed by mercury intrusion porosimetry, MIP (Thermo Finnigan Pascal 140 and Thermo Finnigan Pascal 240). The bulk density of the samples was determined by weight-to-volume ratio and the per cent values of the total porosity was estimated as [1- (bulk density/theoretical density)] x 100. The theoretical density values used for M and C samples were 6.1 and 5.4 g/cm3, respectively. For comparison, a pure dense ZrB2 sample was also included in the investigation. 2.2 Optical characterization Optical reflectance spectra at room temperature in the 0.25-2.5 µm wavelength region were
Paper published on: Solar Energy Materials & Solar Cells, vol. 144 (2016), 608-615 http://dx.doi.org/10.1016/j.solmat.2015.09.068 acquired using a double-beam spectrophotometer (Lambda900 by Perkin Elmer) equipped with a 150-mm diameter integration sphere for the measurement of the hemispherical reflectance. The spectra in the wavelength region 2.5-15.5 μm have been acquired using a Fourier Transform spectrophotometer (FT-IR "Excalibur" by Bio-Rad) equipped with a gold-coated integrating sphere and a liquid nitrogen-cooled detector. In all cases the reflectance spectra are acquired for quasi-normal incidence angle. 3. Results 3.1 Structural characterization The structural and textural characteristics of monophasic and composite samples under investigation are listed in Table 2. Typical macro and microstructural features are shown in Figs. 1-3, where images of the top surface at different magnification are reported for each sample listed in Table 2. Sample
Top Surface details
Sample Sample §Open §Pore size density Total porosity distribution Pores’ Main Lamella 3 porosity vol% Most frequent shape and pore width thickness g/cm (mm) grouping vol% pore sizes (μm) (μm) (μm) and relative Solid wideness contributes (%)
Sample Size label Φ, h
M1
30, 7
Lamellar domains
5-20
5-7
2.62
57
-
M2
15, 15
Lamellar domains
20-30
15-30
3.10
49.2
39.1
M3
15, 15
Lamellar domains
30-60
30-100
3.11
48.9
36.9
C1
30, 7
Tree-like
4000-5000
-
48.2
39.1
C2
15, 15
Thin Lamellar random
20-40
100-800
3.31
38.7
37.8
C3
30, 7
Small tree-like
1000-2000 (tree) 20-50 (single pore)
-
3.33
38.3
37.8
C4
15, 15
Lamellar domains
20-60
100-400
3.32
38.5
32.1
C5
30, 15
Lamellar domains
20-60
50-300
3.20
40.7
37.1
2.80
Bimodal 1-2µm, 50% 20-30 µm; 50 % Bimodal 1-2 µm, 65 % (100 µm), 35% Bimodal 0.8 µm, 17% 30 µm, 83 % Bimodal 0.9 µm, 33% 10 µm, 67% Bimodal 0.9 µm, 45% 10 µm, 55% Bimodal 0.8 µm, 25% 20-50 µm, 75% Bimodal 0.9 µm, 33% 10 µm, 67%
Table 2: List of investigated materials (M= pure ZrB2; C= ZrB2-SiC composite) and structural characteristics. Top surface details were assessed by SEM investigations. Sample density was measured as weight-to-volume ratio (bulk density) and total porosity calculated as [1- (bulk density/theoretical density)] x 100. §Data obtained by MIP (detection range 0.0058 -100 μm).
Paper published on: Solar Energy Materials & Solar Cells, vol. 144 (2016), 608-615 http://dx.doi.org/10.1016/j.solmat.2015.09.068
Figure 1 Top surface images of the investigated monophasic samples (in the insert) and their micro-macrostructures (main images): a) M1, b) M2, c) M3. In d): high magnification SEM micrograph evidencing the microporosity.
Paper published on: Solar Energy Materials & Solar Cells, vol. 144 (2016), 608-615 http://dx.doi.org/10.1016/j.solmat.2015.09.068
Monophasic samples generally had a much higher degree of total porosity (49-57%) compared to composite samples (39-48%) (Table 2). This was mainly due to absence of sintering aid, and/or different water content of the slurry. In addition pore volume, morphology and dimension were heavily affected by the starting composition, casting conditions and densification, as detailed in previous papers [17, 18]. M samples showed main unidirectional pore channels separated by ZrB2 ceramic plates, both oriented in the freezing direction (Fig.1). The wide channel-like pores visible in the SEM images (Fig.1) contribute in lowering the density of the whole M samples and consequently in increasing their total porosity, while they can be out of the detection range of MIP. For this reason, for each sample the open porosity value is generally lower than the corresponding total porosity (Tab. 2), since it does not account for all the macro- and ultramacro-pores larger than 100 µm, which are particularly evident in the SEM images of monophasic samples (Fig.1). Lamellae as well as channel-like pores are grouped and oriented in different directions which reflect the freezing directions. The ceramic walls are constituted by a highly porous microstructure due to lack of the sintering aid in M samples (Fig. 1d). Both the average width of pores and the wall thickness increased in the order M1C4 =C1 could be defined, where C1 looks more like C4, showing a lower content of SiC. 3.2 Room temperature reflectance spectra 3.2.1 Monophasic samples Fig.4 shows the reflectance spectra of pure-ZrB2 samples, comparing freeze-cast pellets to a fully dense isotropic specimen. The spectra differ each other mainly for the absolute value of the reflectance, while retaining the qualitative S-shape which is typical of ZrB2 material and indicates its intrinsically solar-selective characteristics. It is interesting to observe that the reflectance of freeze-cast samples decreases as both the pore width and lamella thickness increase. In fact M1, which has the thinnest channel width (5-20 µm) and thinnest lamellae (5-7 µm) shows the highest reflectance among freeze-cast samples across all the investigated range. On the contrary M3, which has both the thickest lamellae (30-100 µm) and largest pores (30-60 µm) shows the lowest reflectance at all wavelengths. M2 is an intermediate sample, with intermediate pore sizes in the range 20-30 µm and lamella thickness values of 15-30 µm. Its spectrum lies in between those of M1 and M3 for wavelengths shorter than about 5 µm, while it is similar to that of the smoother M1 for longer light wavelengths. Based on the relative contribution of the two pore size modes, calculated from MIP data (Table 2), small pores approximatively amount to 24 and 19.5 vol% respectively in M3 and M2 (corresponding to 65% and 50% of open porosity in Table 2). The higher microporosity combined with the wider solid surface exposed in M3 due to the presence of thicker lamellae can explain the reflectance results at short wavelengths. In fact, due to the presence of wider solid portions in the top surface of M3, the open pore distribution and thus the amount of micrometric pores (C2>C5>C4 =C1) is clearly in agreement with the optical reflectance signal related to SiC presence. Further, C5 and C2 are not so different from C3 concerning the pore size distributions (Table 2), which can explain the overlapping of the reflectance curves at short wavelengths (
