- Email: [email protected]
Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route
G Model
ARTICLE IN PRESS
JECS-10303; No. of Pages 7
Journal of the European Ceramic Society xxx (2015) xxx–xxx
Contents lists available at www.sciencedirect.com
Journal of the European Ceramic Society journal homepage: www.elsevier.com/locate/jeurceramsoc
Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route Manabu Fukushima ∗ , Yu-ichi Yoshizawa National Institute of Advanced Industrial Science and Technology (AIST), 2266-98 Shimo-Shidami, Moriyama-ku, Nagoya 463-8560, Japan
a r t i c l e
i n f o
Article history: Received 1 August 2015 Received in revised form 22 September 2015 Accepted 30 September 2015 Available online xxx Keywords: Insulators Mullite Freeze casting Thermal conductivity Compressive strength
a b s t r a c t Mullite thermal insulators with high porosities of up to 91 vol% were fabricated using a gelation freezing method, resulting in a honeycomb-like morphology with micrometer-sized cells. Mullite particles dispersed gelatin based gels with and without an ice binding protein additives were frozen, dried under vacuum and sintered at 1500 ◦ C. Ice binding proteins could inhibit the formation of large ice crystals and reduce cell size in insulators obtained. The thermal conductivity of the obtained thermal insulators ranged from 0.23 to 0.38 W/mK at room temperature. The compressive strength was measured to be 1.4–21.7 MPa. These properties could be varied by adjusting the process parameters of the gelation freezing. The method proposed here is a promising method of preparing ceramic insulators with very high porosity and improved strength. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Pores are generally treated as fracture origins for structural applications of brittle ceramic materials. However, there have been many industrial applications, in which porosity can be positively taken into account, due to wide range of applications such as refractories, filtrations, biomaterials, catalyst supports, thermal insulators and lightweight structural components. Among these applications, ceramic thermal insulators have recently gained in the importance. By employing thermal insulators with very low conductivity, heat leakage can be decreased in high temperature furnaces and chemical plants, leading to improved energy efficiency and energy savings. Traditional insulating firebricks and ceramic fibrous insulators have been most frequently utilized in various industrial applications mentioned above. The thermal conductivity of the firebrick is relatively high, due to its low porosity, but the applications are mainly structural components such as the lining of high temperature furnace [1,2]. This firebrick is generally fabricated by partial
∗ Corresponding author at: 2266-98 Anagahora, Shimo-Shidami moriyama-ku, Nagoya, Aichi, 463-8560, Japan. Fax: +81 52 736 7405. E-mail addresses: [email protected] (M. Fukushima), [email protected] (Y.-i. Yoshizawa).
sintering, together mixing with sacrificial organic additives to be burned out. Whereas the thermal conductivities of the refractory fibrous insulators are much lower than those of the firebricks, less than or around approximately 0.1 W/mK, they provide insufficient mechanical reliability due to their very high porosity. Furthermore, ceramic fibers are categorized as Group 2B (a possible human carcinogen) as stipulated by the WHO (World Health Organization). Thus, the fibrous insulators have to be carefully reviewed, but the substitutes must be considered. Recent demands on ceramic insulators present two challenges: one is an improved mechanical strength comparable to those of the firebricks, and the other one is lowered thermal conductivity, similar to those of the fibrous insulators. Although a reduction in porosity can result in increased strength, it can also increase the thermal conductivity. Thus, there is a trade-off between mechanical strength and thermal conductivity. It may be possible to overcome the above problems by using a freeze casting method to fabricate thermal insulators, because the porous ceramics prepared by this route have exhibited improved mechanical strength and structural rigidity, despite their very high porosity [3–5]. Typical approaches involve the freeze-drying of aqueous slurry, in which porosity can be created by the formation and sublimation of ice crystals grown in the slurry, followed by sintering. Fukasawa and Ohji reported pioneering works in this field [6–8]. Deville et al. [9–11] studied the anisotropic interface kinetics of the solid/liquid, along with the morphological features of the
http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041 0955-2219/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7 2
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
Fig. 1. Digital micrographs of the polished surfaces of (a–f) sample A, B, C, G, I and L, and SEM micrographs of (g–h) K and G. The images (a–g) and (h) were collected in perpendicular and parallel section to the freezing, respectively.
macro-cellular structure. Combining freeze-drying and gel casting technique has been the other frequently employed approach to fabricate cellular materials, which include the use of solutions and gels of polysaccharides, proteins, gelatin and collagen [12–20]. We have focused on gelation freezing to create unique honeycomb-like microstructures, unlike the dendritic, ellipsoidal, and lamellar microstructures obtained via conventional freeze casting, with having highly interconnected porosities from 79 to 98%
[4,5,21–23]. Due to this honeycomb structure, the mechanical strength of gelation freezing derived porous ceramics can be substantially improved, even though they have higher porosities [22]. The purpose of this study is to fabricate thermal insulators with very high porosity using a gelation freezing method, producing insulators with both high strength and low thermal conductivity. In this article, we report on the fabrication of mullite insulators, and discuss the thermal and mechanical properties of products obtained.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model
ARTICLE IN PRESS
JECS-10303; No. of Pages 7
M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
3
Table 1 Sample descriptions and green porosities as a function of processing parameters. Sample description
Solid loading (vol%)
Water content (vol%)
A B C D E F G H I J K L
7 7 6 6 6 6 6 5 5 5 5 5
90
Freezing temperature (◦ C) −40 −40 −40 −40 −60 −70 −80 −40 −40 −60 −70 −80
91
92
2. Experimental 2.1. Mullite powder synthesis Kaolinite (Eckalite-1, Imerys Japan Co. Ltd., Tokyo, Japan) with an average particle size of 0.4 m and aluminum hydroxide (Higilite 43 M, Showa Denko Co. Ltd., Tokyo, Japan) with an average particle size of 0.7 m were used as raw materials. Mixtures with a weight ratio (Kaolin/aluminum hydroxide) of 45/55 to obtain the stoichiometric ratio for mullite (3Al2 O3 ·2SiO2 ) were blended for 24 h using a ball milling with Al2 O3 balls. After mixing and drying, the powder was calcined at 1400 ◦ C. X-ray diffraction patterns of the calcined powder indicated a well-crystallized mullite phase with traces of ␣-alumina. This powder was used to prepare mullite insulators using a gelation freezing method.
Additive
Green porosity (vol%)
No AFP No AFP AFP AFP AFP No AFP AFP AFP AFP
92.6 92.6 93.2 93.2 93.6 93.7 93.3 95.2 95.1 95.0 95.1 95.8
Table 2 Properties, thermal conductivities and compressive strengths of the obtained thermal insulators. Sample
Porosity (vol%)
A B C D E F G H I J K L
85.2 85.2 89.6 89.4 88.9 89.2 89.2 90.0 91.5 91.1 91.0 90.0
Thermal conductivity (W/mK)
Compressive strength (MPa) 14.3 21.7 2.5 8.8 12.7 11.3 13.1 1.4 8.3 7.0 8.3 8.1
± ± ± ± ± ± ± ± ± ± ± ±
1.6 0.6 0.4 2.4 3.1 2.8 0.4 0.3 1.2 2.0 0.6 0.7
0.34 0.38 0.26 0.29 0.30 0.28 0.24 0.24 0.25 0.23 0.23 0.25
± ± ± ± ± ± ± ± ± ± ± ±
0.01 0.03 0.02 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01
2.2. Fabrication of insulators The obtained mullite powder was pulverized for 3 h using a planetary mill with Al2 O3 pot and Al2 O3 balls to obtain fine mullite particles with an average particle diameter of 0.9 m. The obtained slurry was poured into a warm gelatin solution with the concentration of 3 vol% (Wako pure chemical industries Ltd., Tokyo, Japan) setting at 50 ◦ C. The mixture ratios of mullite powder/gelatin solution were 5/95, 6/94, and 7/93 by volume. As ice binding additives to control the morphology of ice, antifreeze glycoprotein (AFP, Nichirei foods Inc., Chiba, Japan) was used without further purification, which consists of a repeating tripeptide unit of an alanine–alanine–threonine (Ala–Ala–Thr) tripeptide unit whose threonyl OH-group is modified with a disaccharide D-galactosyl-(1,3)-␣-N-acetyl-d-galactosamine with a molecular weight distribution ranging from 2600 to 33,000 [24,25]. The formation of either a polyproline type-II helix or a flexible random coil has been assumed for this protein in the literature [25–27]. AFP was mixed with a gelatin solution in a ratio of 0/100 or 0.25/99.75 by weight, respectively. The mixture was further stirred under vacuum using a planetary homogenizer (ARV-310, Thinky Co. Ltd., Tokyo, Japan). The slurry was placed in a centrifuge container and revolved (800 rpm) while rotating (2000 rpm) under vacuum to defoam the slurry. The prepared slurry was then poured into a plastic mold and kept at 7 ◦ C to obtain mullite powder dispersed gels. The plastic molds containing the gels with various compositions were individually immersed in an ethanol bath setting at −40, −60, −70 and −80 ◦ C, respectively. The contact (immersed) depth with the cooled ethanol was 2 mm from the bottom of the mold. After demolding, the sublimation of ice crystals grown in the frozen gels was carried out in a vacuum freeze drier (Model FDU-2100, Tokyo Rikakikai Co., Ltd., Tokyo, Japan) at −12 to 30 ◦ C and under less than 5 Pa. The resulting disk shaped green bodies were approximately 40 mm in diameter and 18 mm in height. For every composition, at least two specimens were prepared. After the
green bodies dried were sintered at 1500 ◦ C for 2 h, highly porous mullite thermal insulators were obtained. Table 1 summarizes the sample designations, green porosities measured and their processing parameters: the solid loadings and water contents in the initial slurry, freezing temperatures and, AFP additions.
2.3. Sample characterizations The porosity, microstructure, thermal conductivity and compressive strength of the obtained insulators were characterized. Green porosities were measured by the dimension and mixture rule. Open porosities of the insulators obtained were calculated using Archimedes principle with water immersion. The microstructures of polished surfaces were observed by digital microscopy (VHX5000, Keyence Co., Ltd., Osaka, Japan) and scanning electron microscopy (SEM; JEOL, JSM-5600, Tokyo, Japan). The thermal conductivities of the insulators were measured by a hot disk method (TPS1500, Kyoto Electronics Manufacturing Co. Ltd., Kyoto, Japan) with a transient plane source (TPS). A polyimide sensor with a radius of 3.189 mm was used. The output power and measurement time were varied in the ranges of 0.02–0.08 W and 20–40 s, respectively, depending on the thermal characteristics of the samples. All conductivities were measured by using the isotropic modes at room temperature The compressive strength of samples with a diameter of 4.8 mm and a height of 12 mm was measured using the universal testing machine (MTS Systems Corporation, Sintech 10/GL, Minnesota, USA) with a crosshead speed of 0.5 mm/min. Cylindrical specimens were loaded on the top and bottom surfaces. The load was parallel to the freezing direction. Five samples were tested to obtain the average strength.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
4
Fig. 2. Cell sizes and the aspect ratios of cylindrical fine pores (except defects) in the obtained insulators.
3. Results and discussion Table 2 shows the porosities, thermal conductivities and compressive strengths of the obtained thermal insulators. The porosities exhibited around 85, 89 and 91% for samples A–B, C–G and H–L, corresponding to those relative densities of 15, 11, and 9%, respectively. The freezing temperature did not have large effect on the porosity, which was clearly depending on the solid loading. The effect of solid loadings in the initial slurry on porosity suggests full conversion of the water-ice crystals into pores after the sublimation of the ice crystals. The relative density of each insulator was approximately twice the solid loading (7, 6, and 5 vol%) in the initial slurry. Linear shrinkages for samples A–B, C–G and H–L were measured to be around 21, 23 and 27%, respectively. Fig. 1 shows typical digital micrographs of the polished surfaces of (a–f) the samples A, B, C, G, I and L, and SEM micrographs of (g–h) K and G, respectively. Specimens were firstly ground and polished by removing 1.5 mm of the upper end, and the microstructures of the polished surface were observed. The orientations of (a–g) and (h) were perpendicular and parallel to the freezing direction. Honeycomb like structures were observed in all samples, and were comprised of micrometer-sized cylindrical cells. In addition, elongated pore morphologies are visible in Fig. 1(h), in which the cells were portioned by thin cell walls (struts), and the channels appeared a straight, having one-dimensional ordered morphology along the freezing direction. The morphology observed is typical one created by the gelation freezing derived ceramics. The detail pore formation has been reported elsewhere [4,5,21–23,28]. Meanwhile, as shown in Fig. 1(c), large defects were observed in sample C, which was prepared without AFP. Those were found in the sample H as well. These defects are sometimes called as ice lenses due to their shape, which resembles a convex lens, but still including together with micrometer sized cylindrical cells. Cell sizes (but except for the large defects) could be slightly reduced by AFP addition and a decreased freezing temperature, in the following order of sample order A > B, C > G, and I > L. Seen from Fig. 1(g), it was found that there were some porosities remained in the struts, because sintering temperatures to densify mullite ranged from 1700 to 1750 ◦ C [29,30]. The cell sizes and aspect ratios of the cylindrical fine pores (except defects) in the perpendicular plane to the freezing are shown in Fig. 2. Seen from the samples D–G with 6% loading or I–L with 5% loading, the cell size gradually decreased with decreasing freezing temperature, and increased with decreasing solid loading. The diameters of the cells ranged from 24 to 136 m. In the samples A, C and H, which were prepared without AFP, the cell sizes were observed to be much larger. The aspect ratios of the fine pores in all samples were almost constant, regardless of the solid loading in
the initial slurry, the freezing temperature, and whether AFP was added. Coarse defects seen in Fig. 1 represent serious problems, and their inhibition is a key issue to be solved in improving the mechanical reliability of insulators. It is noted that these defects were only observed in the samples C and H (solid loadings less than 6% and without AFP addition). On the other hand, when raw materials with a higher thermal conductivity such as SiC and nickel have been used, such defects have never been formed [22,31]. Therefore, the observed defects are associated with (1) the conductivity of the raw material, (2) the solid loading in the initial slurry, and (3) AFP addition. The thermal conductivity of mullite is relatively closer to that of ice (2.2 W/mK) than that of SiC or nickel. During freezing, there is a smaller temperature difference between the mullite particles and the ice, due to their similar conductivities. Furthermore, a latent heat during freezing is continuously generated by the physical change from water to ice, which has been previously confirmed as temperature increase or the plateau during freezing [22]. The release of latent heat from the ice and heat flow toward the cooling medium through the raw particles and direct conduction by the ice itself, i.e., heat exchange, could locally become insufficient in the present case. Thus, heat accumulation on ice presumably takes place, leading to the formation of coarse ice lenses during freezing. However, in order to prepare insulators, the use of a low conductive material is essential. Therefore, other processing factors are necessary in order to create a defect-free morphology. Lower solid loading can cause the formation of deficient part of particles around the moving ice front during freezing, because of the lower viscosity due to less frequent particle–particle and waterparticle interactions [4], eventually leading to the growth of large ice. When the cell wall is composed of fully packed particles, there should be a threshold value in the initial solid content. If the threshold is below that required to make the packed wall, porosities in the cell wall of green bodies can be formed, in which that can be supported by the results of shrinkage. However, a lower solid loading is very important processing factor to enhance the porosity of the resulting insulators. Thus, the AFP addition is critically required to avoid the formation of defects in the present gelation freezing technique. It is reasonable to consider that AFP addition prevented the formation of the coarse defects (large ice crystals), because of its pronounced effect on ice nucleation and grain growth. The detailed mechanism has been reported by many researchers [24,32–35]. Adsorption inhibition mechanism has been well accepted to retard the grain growth and recrystallization of ice crystals, in which sterically bulky AFP binds to the surface prism planes of ice crystals and further adsorption of water molecules on the surface are strongly inhibited. In fact, the cell size in insulators prepared with AFP addition was observed to be substantially smaller in a diameter, and without any large defects. We can state therefore that AFP needs to provide structural stability for thermal insulators. Fig. 3 shows the thermal conductivity of the obtained insulators, and their effective conductivity calculated by the Series and Parallel models (shown as solid lines) and the Maxwell–Eucken models (shown as dotted lines), as a function of porosity. The thermal conductivities of the prepared insulators were around 0.23–0.38 W/mK, and apparently affected by the porosity. Freezing temperature never affected the conductivities (see Table 2). Insulators can be generally treated as composite materials consisting of two different phases: a solid matrix and air in pores. The thermal conductivity of the air in the pores (0.025 W/mK) is substantially lower than that of the mullite (5 W/mK) [36–38]. Therefore, at higher porosities, the conductivity decreases. Series and Parallel models have been proposed to estimate the effective conductivity of cellular materials, particularly the lower and upper bounds (called as the Wiener bounds). The Series and
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model
ARTICLE IN PRESS
JECS-10303; No. of Pages 7
M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
Fig. 3. Thermal conductivities of the insulators at room temperature, and effective conductivities (room temperature) as a function of porosity, calculated using the Series and Parallel models shown as solid lines and the Maxwell–Eucken models as dotted lines.
Parallel models are calculated according to the following Eqs. (1) and (2) [39], respectively: ke =
1 (1 − vg )/ks + vg /kg
ke = (1 − vg ) × ks + vg × kg
(1) (2)
where ks , kg , and vg are the thermal conductivities of the solid phase (skeleton), and of air, and the volume fraction of air in the insulator (porosity), respectively. Hashin and Shtrikman derived further narrower bounds of the effective conductivity of a material with two phases, and these bounds always lie between the above Series and Parallel bounds [40]. Proposed Maxwell–Eucken 1 and 2 were computed according to the following Eqs. (3) and (4), respectively: ke = ks
2ks + kg − 2(ks − kg )vg 2ks + kg + (ks − kg )vg
(3)
ke = kg
2kg + ks − 2(kg − ks )(1 − vg ) 2kg + ks + (kg − ks )(1 − vg )
(4)
where ks , kg , and vg represent those described above in the same manner. The difference between the Eqs. (3) and (4) is whether the heat conduction pathway avoids pores or includes an outside solid phase. These models are defined as internal porosity (pores inside materials) or external porosity (pores outside materials), respectively. When pores are dispersed within a continuous condensed solid medium (internal porosity, such as a foam or honeycomb), an optimal heat transfer pathway should avoid the pores. In the literature, heat flux that could avoid pores was depicted as vectors for dispersed spherical pores within a continuous solid (a honeycomb-like structure comprised of spherical pores) on the basis of simulated finite element analysis of simple spherical dispersion model [39]. Furthermore, it is based on an assumption of the model that the pores inside materials do not make any contacts with each other. These restraint assumptions seem to fit with the honeycomb-like microstructures of the present insulators. Seen from the results, measured conductivities were closed to Maxwell–Eucken 1, and located on the side of Parallel models. In the Parallel model, heat flow can be transferred ideally along a direction parallel to materials, so the effective conductivity is larger than that predicted by Maxwell–Eucken 1. The experimental results observed on the side of the Parallel model indicates preferential heat transfer along the struts parallel to cell orientation. Fig. 4 shows the compressive strengths of the insulators obtained and predicted strength values of honeycomb out-of-plane model, in which a load was applied parallel to the cell orientation (the freezing direction). The compressive strength of the insulators ranged from 1.4 to 21.7 MPa, in which those of the insulators C and
5
Fig. 4. Compressive strengths of the obtained insulators and predicted strengths of honeycomb out-of-plane model proposed by Gibson and Ashby.
H were clearly lower than those of the others with similar porosities, again due to the large defects. It was found that the strength of the insulators was affected by porosities (solid loading in the slurry) and defects (AFP addition). Freezing temperature did not have large effect on the strength of the insulators. A honeycomb structure is relatively strong and stiff in the direction parallel to the cell orientation, compared with the in-plane values. This is advantageous for the present honeycomb-like microstructure. The out-of-plane model to predict the compressive strength of a honeycomb with ideal hexagonal cells has been proposed, which can be computed using the Eq. (5) developed by Gibson and Ashby [41]: c =C s
s
(5)
where c , s , C, and s indicate the compressive strength of a cellular solid, strength of a fully dense material, an experimental constant (dotted line computed by the slope of results), density of cellular structure and the fully density, respectively. We used 1310 MPa and 3.16 g/cc as reference values for the compressive strength and density of the fully dense mullite [42,43]. The experimental values were lower than that of predicted line. Compressive strengths of ceramic foams or cellular ceramics have been most frequently expressed as the 1.5 power of relative density, which has been widely utilized for the estimations [42,44–46], while that of a honeycomb is provided as a function of relative density [41]. When a net sectional stress in surface plane perpendicular to cell orientation exceeds fracture strength of a cell wall material ( s ), a honeycomb will be fractured. Thus, this means an upper limiting bound for the strength, as shown by the solid line of Fig. 4, i.e., completely defect-free honeycomb. If there are any defects contained in a honeycomb, it should be fractured at lower stress. As mentioned previously, cell walls observed were porous because of lower solid loading, including some contacting points among cells, and definitely different from ideal dense “beam” used as the assumption, certainly leading to reduced strength of insulators. Thus, it is reasonable to consider that the compressive strength of the insulators obtained can be lower than that of predicted values. We compared the thermal conductivities and compressive strength obtained in our study with those of references, which mostly included silica as additives to improve mechanical strength and reduce thermal conductivity, due to strong neck formed by liquid phase and its lower conductivity, but alumina included on some reports [47–49]. In their cases, the specimens sintered at 1500 ◦ C showed 11 MPa and 0.32 W/mK (porosity 78%), 3 MPa and 0.2W/mK (porosity 74%), and 5 MPa and 0.1 W/mK (porosity 81–84%) for the compressive strength and thermal conductivity, respectively, while the present insulators G exhibited 13.1 MPa and 0.24 W/mK (porosity 89.2%), for the strength and conductivity, respectively. Thus, the present insulators with the honeycomb-like structure showed higher strength than those of literatures, regardless of higher
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
6
porosity. The conductivity was also higher than those of above second and third ones, because they have used loose packed grains to create external porosity or mullite fiber as raw materials to provide narrow heat transfer path [47,50]. However, the comparison with the past studies suggests that carefully tailored external porosity combined with gelation freezing method have a great possibility to give rise to substantially reduced thermal conductivity, and together with improved strength, as future prospective. Work is still ongoing to develop thermal insulators with lowered thermal conductivity and improved compressive strength. 4. Conclusion Mullite thermal insulators were fabricated by a gelation freezing route, and the effects of gelation freezing processing factors on the thermal conductivity and compressive strength of the resulting insulators were investigated. The insulators have been comprised of micrometer sized honeycomb like morphology. The addition of AFP led to a defect-free microstructure and improved strength. The obtained insulators exhibited very low thermal conductivities of 0.23–0.38 W/mK that well agreed with Maxwell–Eucken 1 model. Compressive strength of the insulators ranged from 1.4 to 21.7 MPa, which was lower than that of theoretical prediction proposed by Gibson and Ashby. The proposed process is a simple, effective and ecofriendly method of fabricating highly porous ceramic insulators for use in a wide range of industrial applications. Acknowledgements This research work was performed as a part of Future Pioneering Projects/Research and Development of Thermal Management Materials and Technology by NEDO. The authors gratefully acknowledge their support. References [1] H.W. Russell, Principles of heat flow in porous insulators, J. Am. Ceram. Soc. 18 (1935) 1–5. [2] A.L. Loeb, Thermal conductivity.8. A theory of thermal conductivity of porous materials, J. Am. Ceram. Soc. 2 (1954) 96–99. [3] L. Hu, C.-A. Wang, Y. Huang, C. Sun, S. Lu, Z. Hu, Control of pore channel size during freeze casting of porous YSZ ceramics with unidirectionally aligned channels using different freezing temperatures, J. Eur. Ceram. Soc. 30 (16) (2010) 3389–3396. [4] M. Fukushima, Y.-i. Yoshizawa, Fabrication of highly porous silica thermal insulators prepared by gelation-freezing route, J. Am. Ceram. Soc. 97 (3) (2014) 713–717. [5] M. Fukushima, Y.-i. Yoshizawa, T. Ohji, Macroporous ceramics by gelation-freezing route using gelatin, Adv. Eng. Mater. 16 (6) (2014) 607–620. [6] T. Fukasawa, Z.Y. Deng, M. Ando, T. Ohji, Y. Goto, Pore structure of porous ceramics synthesized from water-based slurry by freeze–dry process, J. Mater. Sci. 36 (10) (2001) 2523–2527. [7] T. Fukasawa, M. Ando, T. Ohji, S. Kanzaki, Synthesis of porous ceramics with complex pore structure by freeze–dry processing, J. Am. Ceram. Soc. 84 (1) (2001) 230–232. [8] T. Fukasawa, Z.Y. Deng, M. Ando, T. Ohji, S. Kanzaki, Synthesis of porous silicon nitride with unidirectionally aligned channels using freeze–drying process, J. Am. Ceram. Soc. 85 (9) (2002) 2151–2155. [9] S. Deville, E. Saiz, R.K. Nalla, A.P. Tomsia, Freezing as a path to build complex composites, Science 311 (5760) (2006) 515–518. [10] S. Deville, E. Saiz, A.P. Tomsia, Ice-templated porous alumina structures, Acta Mater. 55 (6) (2007) 1965–1974. [11] S. Deville, Freeze-casting of porous ceramics: a review of current achievements and issues, Adv. Eng. Mater. 10 (3) (2008) 155–169. [12] R. Chen, C.-A. Wang, Y. Huang, L. Ma, W. Lin, Ceramics with special porous structures fabricated by freeze-gelcasting: using tert-butyl alcohol as a template, J. Am. Ceram. Soc. 90 (11) (2007) 3478–3484. [13] R. Chen, Y. Huang, C.-A. Wang, J. Qi, Ceramics with ultra-low density fabricated by gelcasting: an unconventional view, J. Am. Ceram. Soc. 90 (11) (2007) 3424–3429. [14] D. Koch, L. Andresen, T. Schmedders, G. Grathwohl, Evolution of porosity by freeze casting and sintering of sol–gel derived ceramics, J. Sol–Gel Sci. Technol. 26 (1–3) (2003) 149–152.
[15] S.R. Mukai, H. Nishihara, H. Tamon, Formation of monolithic silica gel microhoneycombs (SMHs) using pseudosteady state growth of microstructural ice crystals, Chem. Commun. 7 (2004) 874–875. [16] H. Nishihara, S.R. Mukai, D. Yamashita, H. Tamon, Ordered macroporous silica by ice templating, Chem. Mater. 17 (3) (2005) 683–689. [17] C.M. Pekor, P. Kisa, I. Nettleship, Effect of polyethylene glycol on the microstructure of freeze-cast alumina, J. Am. Ceram. Soc. 91 (10) (2008) 3185–3190. [18] C. Pekor, B. Groth, I. Nettleship, The effect of polyvinyl alcohol on the microstructure and permeability of freeze-cast alumina, J. Am. Ceram. Soc. 93 (1) (2010) 115–120. [19] E. Landi, F. Valentini, A. Tampieri, Porous hydroxyapatite/gelatine scaffolds with ice-designed channel-like porosity for biomedical applications, Acta Biomater. 4 (6) (2008) 1620–1626. [20] H.F. Zhang, I. Hussain, M. Brust, M.F. Butler, S.P. Rannard, A.I. Cooper, Aligned two- and three-dimensional structures by directional freezing of polymers and nanoparticles, Nat. Mater. 4 (10) (2005) 787–793. [21] M. Fukushima, M. Nakata, Y. Yoshizawa, Fabrication and properties of ultra highly porous cordierite with oriented micrometer-sized cylindrical pores by gelation and freezing method, J. Ceram. Soc. Jpn. 116 (1360) (2008) 1322–1325. [22] M. Fukushima, M. Nakata, Y. Zhou, T. Ohji, Y.I. Yoshizawa, Fabrication and properties of ultra highly porous silicon carbide by the gelation-freezing method, J. Eur. Ceram. Soc. 30 (14) (2010) 2889–2896. [23] M. Fukushima, S. Tsuda, Y.-i. Yoshizawa, Fabrication of highly porous alumina prepared by gelation freezing route with antifreeze protein, J. Am. Ceram. Soc. 96 (4) (2013) 1029–1031. [24] Y. Yeh, R.E. Feeney, Antifreeze proteins: structures and mechanisms of function, Chem. Rev. 96 (2) (1996) 601–617. [25] M.M. Harding, P.I. Anderberg, A.D.J. Haymet, Antifreeze’ glycoproteins from polar fish, Eur. J. Biochem. 270 (7) (2003) 1381–1392. [26] D. Slaughter, G.L. Fletcher, V.S. Ananthanarayanan, C.L. Hew, Antifreeze proteins from the sea raven, Hemitripterus-americanuS—further evidence for diversity among fish polypeptide antifreezes, J. Biol. Chem. 256 (4) (1981) 2022–2026. [27] C.H.C. Cheng, L.B. Chen, Evolution of an antifreeze glycoprotein, Nature 401 (6752) (1999) 443–444. [28] K. Hamamoto, M. Fukushima, M. Mamiya, Y. Yoshizawa, J. Akimoto, T. Suzuki, Y. Fujishiro, Morphology control and electrochemical properties of LiFePO4/C composite cathode for lithium ion batteries, Solid State Ionics 225 (2012) 560–563. [29] G.S. Perry, Microwave dielectric properties of mullite, Trans. J. Br. Ceram. Soc. 72 (6) (1973) 279–283. [30] B.L. Metcalfe, J.H. Sant, Synthesis, microstructure and physical properties of high-purity mullite, Trans. J. Br. Ceram. Soc. 74 (6) (1975) 193–201. [31] M. Fukushima, Y.-i. Yoshizawa, Fabrication of highly porous nickel with oriented micrometer-sized cylindrical pores by gelation freezing method, Mater. Lett. 153 (2015) 99–101. [32] J.A. Raymond, A.L. Devries, Adsorption inhibition as a mechanism of freezing resistance in polar fishes, Proc. Natl. Acad. Sci. U. S. A. 74 (6) (1977) 2589–2593. [33] C.A. Knight, A.L. Devries, L.D. Oolman, Fish antifreeze protein and the freezing and recrystallization of ice, Nature 308 (5956) (1984) 295–296. [34] G.L. Fletcher, M.H. Kao, R.M. Fourney, Antifreeze peptides confer freezing resistance to fish, Can. J. Zool.-Rev. Can. De Zool. 64 (9) (1986) 1897–1901. [35] R.L. McKown, G.J. Warren, Enhanced survival of yeast expressing an antifreeze gene analogue after freezing, Cryobiology 28 (5) (1991) 474–482. [36] L.M. Russell, L.F. Johnson, D.P.H. Hasselman, R. Ruh, Thermal conductivity/diffusivity of silicon carbide whisker reinforced mullite, J. Am. Ceram. Soc. 70 (10) (1987) C226–C229. [37] T.M. Kyaw, Y. Okamoto, K. Hayashi, Thermal conductivity of mullite-zirconia composites, J. Ceram. Soc. Jpn. 103 (12) (1995) 1289–1292. [38] T.W. Clyne, I.O. Golosnoy, J.C. Tan, A.E. Markaki, Porous materials for thermal management under extreme conditions, Philos. Trans. R. Soc.-Math. Phys. Eng. Sci. 364 (1838) (2006) 125–146. [39] J.K. Carson, S.J. Lovatt, D.J. Tanner, A.C. Cleland, Thermal conductivity bounds for isotropic, porous materials, Int. J. Heat Mass Transfer 48 (11) (2005) 2150–2158. [40] Z. Hashin, S. Shtrikman, A Variational approach to the theory of the effective magnetic permeability of multiphase materials, J. Appl. Phys. 33 (10) (1962) 3125. [41] Lorna J. Gibson, M. F. Ashby, Cellular Solids: Structure and Properties 2nd Edition, Cambridge University Press, (1999). [42] Y.W. Kim, H.D. Kim, C.B. Park, Processing of microcellular mullite, J. Am. Ceram. Soc. 88 (12) (2005) 3311–3315. [43] H. Ledbetter, S. Kim, D. Balzar, S. Crudele, W. Kriven, Elastic properties of mullite, J. Am. Ceram. Soc. 81 (4) (1998) 1025–1028. [44] P. Colombo, E. Bernardo, Macro- and micro-cellular porous ceramics from preceramic polymers, Compos. Sci. Technol. 63 (16) (2003) 2353–2359. [45] F.M. Varela-Feria, J. Martinez-Fernandez, A.R. de Arellano-Lopez, M. Singh, Low density biomorphic silicon carbide: microstructure and mechanical properties, J. Eur. Ceram. Soc. 22 (14–15) (2002) 2719–2725. [46] J.M. Tulliani, L. Montanaro, T.J. Bell, M.V. Swain, Semiclosed-cell mullite foams: preparation and macro- and micromechanical characterization, J. Am. Ceram. Soc. 82 (4) (1999) 961–968.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
[47] Z. Hou, H. Du, J. Liu, R. Hao, X. Dong, M. Liu, Fabrication and properties of mullite fiber matrix porous ceramics by a TBA-based gel-casting process, J. Eur. Ceram. Soc. 33 (4) (2013) 717–725. [48] H. Qian, X. Cheng, H. Zhang, R. Zhang, Y. Wang, Preparation of porous mullite ceramics using fly ash cenosphere as a pore-forming agent by gelcasting process, Int. J. Appl. Ceram. Technol. 11 (5) (2014) 858–863.
[49] R.F. Zhang, J.J. Feng, X.D. Cheng, L.L. Gong, Y. Li, H.P. Zhang, Porous thermal insulation materials derived from fly ash using a foaming and slip casting method, Energy Build. 81 (2014) 262–267. [50] L. Gong, Y. Wang, X. Cheng, R. Zhang, H. Zhang, Porous mullite ceramics with low thermal conductivity prepared by foaming and starch consolidation, J. Porous Mater. 21 (1) (2014) 15–21.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
7
ARTICLE IN PRESS
JECS-10303; No. of Pages 7
Journal of the European Ceramic Society xxx (2015) xxx–xxx
Contents lists available at www.sciencedirect.com
Journal of the European Ceramic Society journal homepage: www.elsevier.com/locate/jeurceramsoc
Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route Manabu Fukushima ∗ , Yu-ichi Yoshizawa National Institute of Advanced Industrial Science and Technology (AIST), 2266-98 Shimo-Shidami, Moriyama-ku, Nagoya 463-8560, Japan
a r t i c l e
i n f o
Article history: Received 1 August 2015 Received in revised form 22 September 2015 Accepted 30 September 2015 Available online xxx Keywords: Insulators Mullite Freeze casting Thermal conductivity Compressive strength
a b s t r a c t Mullite thermal insulators with high porosities of up to 91 vol% were fabricated using a gelation freezing method, resulting in a honeycomb-like morphology with micrometer-sized cells. Mullite particles dispersed gelatin based gels with and without an ice binding protein additives were frozen, dried under vacuum and sintered at 1500 ◦ C. Ice binding proteins could inhibit the formation of large ice crystals and reduce cell size in insulators obtained. The thermal conductivity of the obtained thermal insulators ranged from 0.23 to 0.38 W/mK at room temperature. The compressive strength was measured to be 1.4–21.7 MPa. These properties could be varied by adjusting the process parameters of the gelation freezing. The method proposed here is a promising method of preparing ceramic insulators with very high porosity and improved strength. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Pores are generally treated as fracture origins for structural applications of brittle ceramic materials. However, there have been many industrial applications, in which porosity can be positively taken into account, due to wide range of applications such as refractories, filtrations, biomaterials, catalyst supports, thermal insulators and lightweight structural components. Among these applications, ceramic thermal insulators have recently gained in the importance. By employing thermal insulators with very low conductivity, heat leakage can be decreased in high temperature furnaces and chemical plants, leading to improved energy efficiency and energy savings. Traditional insulating firebricks and ceramic fibrous insulators have been most frequently utilized in various industrial applications mentioned above. The thermal conductivity of the firebrick is relatively high, due to its low porosity, but the applications are mainly structural components such as the lining of high temperature furnace [1,2]. This firebrick is generally fabricated by partial
∗ Corresponding author at: 2266-98 Anagahora, Shimo-Shidami moriyama-ku, Nagoya, Aichi, 463-8560, Japan. Fax: +81 52 736 7405. E-mail addresses: [email protected] (M. Fukushima), [email protected] (Y.-i. Yoshizawa).
sintering, together mixing with sacrificial organic additives to be burned out. Whereas the thermal conductivities of the refractory fibrous insulators are much lower than those of the firebricks, less than or around approximately 0.1 W/mK, they provide insufficient mechanical reliability due to their very high porosity. Furthermore, ceramic fibers are categorized as Group 2B (a possible human carcinogen) as stipulated by the WHO (World Health Organization). Thus, the fibrous insulators have to be carefully reviewed, but the substitutes must be considered. Recent demands on ceramic insulators present two challenges: one is an improved mechanical strength comparable to those of the firebricks, and the other one is lowered thermal conductivity, similar to those of the fibrous insulators. Although a reduction in porosity can result in increased strength, it can also increase the thermal conductivity. Thus, there is a trade-off between mechanical strength and thermal conductivity. It may be possible to overcome the above problems by using a freeze casting method to fabricate thermal insulators, because the porous ceramics prepared by this route have exhibited improved mechanical strength and structural rigidity, despite their very high porosity [3–5]. Typical approaches involve the freeze-drying of aqueous slurry, in which porosity can be created by the formation and sublimation of ice crystals grown in the slurry, followed by sintering. Fukasawa and Ohji reported pioneering works in this field [6–8]. Deville et al. [9–11] studied the anisotropic interface kinetics of the solid/liquid, along with the morphological features of the
http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041 0955-2219/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7 2
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
Fig. 1. Digital micrographs of the polished surfaces of (a–f) sample A, B, C, G, I and L, and SEM micrographs of (g–h) K and G. The images (a–g) and (h) were collected in perpendicular and parallel section to the freezing, respectively.
macro-cellular structure. Combining freeze-drying and gel casting technique has been the other frequently employed approach to fabricate cellular materials, which include the use of solutions and gels of polysaccharides, proteins, gelatin and collagen [12–20]. We have focused on gelation freezing to create unique honeycomb-like microstructures, unlike the dendritic, ellipsoidal, and lamellar microstructures obtained via conventional freeze casting, with having highly interconnected porosities from 79 to 98%
[4,5,21–23]. Due to this honeycomb structure, the mechanical strength of gelation freezing derived porous ceramics can be substantially improved, even though they have higher porosities [22]. The purpose of this study is to fabricate thermal insulators with very high porosity using a gelation freezing method, producing insulators with both high strength and low thermal conductivity. In this article, we report on the fabrication of mullite insulators, and discuss the thermal and mechanical properties of products obtained.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model
ARTICLE IN PRESS
JECS-10303; No. of Pages 7
M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
3
Table 1 Sample descriptions and green porosities as a function of processing parameters. Sample description
Solid loading (vol%)
Water content (vol%)
A B C D E F G H I J K L
7 7 6 6 6 6 6 5 5 5 5 5
90
Freezing temperature (◦ C) −40 −40 −40 −40 −60 −70 −80 −40 −40 −60 −70 −80
91
92
2. Experimental 2.1. Mullite powder synthesis Kaolinite (Eckalite-1, Imerys Japan Co. Ltd., Tokyo, Japan) with an average particle size of 0.4 m and aluminum hydroxide (Higilite 43 M, Showa Denko Co. Ltd., Tokyo, Japan) with an average particle size of 0.7 m were used as raw materials. Mixtures with a weight ratio (Kaolin/aluminum hydroxide) of 45/55 to obtain the stoichiometric ratio for mullite (3Al2 O3 ·2SiO2 ) were blended for 24 h using a ball milling with Al2 O3 balls. After mixing and drying, the powder was calcined at 1400 ◦ C. X-ray diffraction patterns of the calcined powder indicated a well-crystallized mullite phase with traces of ␣-alumina. This powder was used to prepare mullite insulators using a gelation freezing method.
Additive
Green porosity (vol%)
No AFP No AFP AFP AFP AFP No AFP AFP AFP AFP
92.6 92.6 93.2 93.2 93.6 93.7 93.3 95.2 95.1 95.0 95.1 95.8
Table 2 Properties, thermal conductivities and compressive strengths of the obtained thermal insulators. Sample
Porosity (vol%)
A B C D E F G H I J K L
85.2 85.2 89.6 89.4 88.9 89.2 89.2 90.0 91.5 91.1 91.0 90.0
Thermal conductivity (W/mK)
Compressive strength (MPa) 14.3 21.7 2.5 8.8 12.7 11.3 13.1 1.4 8.3 7.0 8.3 8.1
± ± ± ± ± ± ± ± ± ± ± ±
1.6 0.6 0.4 2.4 3.1 2.8 0.4 0.3 1.2 2.0 0.6 0.7
0.34 0.38 0.26 0.29 0.30 0.28 0.24 0.24 0.25 0.23 0.23 0.25
± ± ± ± ± ± ± ± ± ± ± ±
0.01 0.03 0.02 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.01 0.01
2.2. Fabrication of insulators The obtained mullite powder was pulverized for 3 h using a planetary mill with Al2 O3 pot and Al2 O3 balls to obtain fine mullite particles with an average particle diameter of 0.9 m. The obtained slurry was poured into a warm gelatin solution with the concentration of 3 vol% (Wako pure chemical industries Ltd., Tokyo, Japan) setting at 50 ◦ C. The mixture ratios of mullite powder/gelatin solution were 5/95, 6/94, and 7/93 by volume. As ice binding additives to control the morphology of ice, antifreeze glycoprotein (AFP, Nichirei foods Inc., Chiba, Japan) was used without further purification, which consists of a repeating tripeptide unit of an alanine–alanine–threonine (Ala–Ala–Thr) tripeptide unit whose threonyl OH-group is modified with a disaccharide D-galactosyl-(1,3)-␣-N-acetyl-d-galactosamine with a molecular weight distribution ranging from 2600 to 33,000 [24,25]. The formation of either a polyproline type-II helix or a flexible random coil has been assumed for this protein in the literature [25–27]. AFP was mixed with a gelatin solution in a ratio of 0/100 or 0.25/99.75 by weight, respectively. The mixture was further stirred under vacuum using a planetary homogenizer (ARV-310, Thinky Co. Ltd., Tokyo, Japan). The slurry was placed in a centrifuge container and revolved (800 rpm) while rotating (2000 rpm) under vacuum to defoam the slurry. The prepared slurry was then poured into a plastic mold and kept at 7 ◦ C to obtain mullite powder dispersed gels. The plastic molds containing the gels with various compositions were individually immersed in an ethanol bath setting at −40, −60, −70 and −80 ◦ C, respectively. The contact (immersed) depth with the cooled ethanol was 2 mm from the bottom of the mold. After demolding, the sublimation of ice crystals grown in the frozen gels was carried out in a vacuum freeze drier (Model FDU-2100, Tokyo Rikakikai Co., Ltd., Tokyo, Japan) at −12 to 30 ◦ C and under less than 5 Pa. The resulting disk shaped green bodies were approximately 40 mm in diameter and 18 mm in height. For every composition, at least two specimens were prepared. After the
green bodies dried were sintered at 1500 ◦ C for 2 h, highly porous mullite thermal insulators were obtained. Table 1 summarizes the sample designations, green porosities measured and their processing parameters: the solid loadings and water contents in the initial slurry, freezing temperatures and, AFP additions.
2.3. Sample characterizations The porosity, microstructure, thermal conductivity and compressive strength of the obtained insulators were characterized. Green porosities were measured by the dimension and mixture rule. Open porosities of the insulators obtained were calculated using Archimedes principle with water immersion. The microstructures of polished surfaces were observed by digital microscopy (VHX5000, Keyence Co., Ltd., Osaka, Japan) and scanning electron microscopy (SEM; JEOL, JSM-5600, Tokyo, Japan). The thermal conductivities of the insulators were measured by a hot disk method (TPS1500, Kyoto Electronics Manufacturing Co. Ltd., Kyoto, Japan) with a transient plane source (TPS). A polyimide sensor with a radius of 3.189 mm was used. The output power and measurement time were varied in the ranges of 0.02–0.08 W and 20–40 s, respectively, depending on the thermal characteristics of the samples. All conductivities were measured by using the isotropic modes at room temperature The compressive strength of samples with a diameter of 4.8 mm and a height of 12 mm was measured using the universal testing machine (MTS Systems Corporation, Sintech 10/GL, Minnesota, USA) with a crosshead speed of 0.5 mm/min. Cylindrical specimens were loaded on the top and bottom surfaces. The load was parallel to the freezing direction. Five samples were tested to obtain the average strength.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
4
Fig. 2. Cell sizes and the aspect ratios of cylindrical fine pores (except defects) in the obtained insulators.
3. Results and discussion Table 2 shows the porosities, thermal conductivities and compressive strengths of the obtained thermal insulators. The porosities exhibited around 85, 89 and 91% for samples A–B, C–G and H–L, corresponding to those relative densities of 15, 11, and 9%, respectively. The freezing temperature did not have large effect on the porosity, which was clearly depending on the solid loading. The effect of solid loadings in the initial slurry on porosity suggests full conversion of the water-ice crystals into pores after the sublimation of the ice crystals. The relative density of each insulator was approximately twice the solid loading (7, 6, and 5 vol%) in the initial slurry. Linear shrinkages for samples A–B, C–G and H–L were measured to be around 21, 23 and 27%, respectively. Fig. 1 shows typical digital micrographs of the polished surfaces of (a–f) the samples A, B, C, G, I and L, and SEM micrographs of (g–h) K and G, respectively. Specimens were firstly ground and polished by removing 1.5 mm of the upper end, and the microstructures of the polished surface were observed. The orientations of (a–g) and (h) were perpendicular and parallel to the freezing direction. Honeycomb like structures were observed in all samples, and were comprised of micrometer-sized cylindrical cells. In addition, elongated pore morphologies are visible in Fig. 1(h), in which the cells were portioned by thin cell walls (struts), and the channels appeared a straight, having one-dimensional ordered morphology along the freezing direction. The morphology observed is typical one created by the gelation freezing derived ceramics. The detail pore formation has been reported elsewhere [4,5,21–23,28]. Meanwhile, as shown in Fig. 1(c), large defects were observed in sample C, which was prepared without AFP. Those were found in the sample H as well. These defects are sometimes called as ice lenses due to their shape, which resembles a convex lens, but still including together with micrometer sized cylindrical cells. Cell sizes (but except for the large defects) could be slightly reduced by AFP addition and a decreased freezing temperature, in the following order of sample order A > B, C > G, and I > L. Seen from Fig. 1(g), it was found that there were some porosities remained in the struts, because sintering temperatures to densify mullite ranged from 1700 to 1750 ◦ C [29,30]. The cell sizes and aspect ratios of the cylindrical fine pores (except defects) in the perpendicular plane to the freezing are shown in Fig. 2. Seen from the samples D–G with 6% loading or I–L with 5% loading, the cell size gradually decreased with decreasing freezing temperature, and increased with decreasing solid loading. The diameters of the cells ranged from 24 to 136 m. In the samples A, C and H, which were prepared without AFP, the cell sizes were observed to be much larger. The aspect ratios of the fine pores in all samples were almost constant, regardless of the solid loading in
the initial slurry, the freezing temperature, and whether AFP was added. Coarse defects seen in Fig. 1 represent serious problems, and their inhibition is a key issue to be solved in improving the mechanical reliability of insulators. It is noted that these defects were only observed in the samples C and H (solid loadings less than 6% and without AFP addition). On the other hand, when raw materials with a higher thermal conductivity such as SiC and nickel have been used, such defects have never been formed [22,31]. Therefore, the observed defects are associated with (1) the conductivity of the raw material, (2) the solid loading in the initial slurry, and (3) AFP addition. The thermal conductivity of mullite is relatively closer to that of ice (2.2 W/mK) than that of SiC or nickel. During freezing, there is a smaller temperature difference between the mullite particles and the ice, due to their similar conductivities. Furthermore, a latent heat during freezing is continuously generated by the physical change from water to ice, which has been previously confirmed as temperature increase or the plateau during freezing [22]. The release of latent heat from the ice and heat flow toward the cooling medium through the raw particles and direct conduction by the ice itself, i.e., heat exchange, could locally become insufficient in the present case. Thus, heat accumulation on ice presumably takes place, leading to the formation of coarse ice lenses during freezing. However, in order to prepare insulators, the use of a low conductive material is essential. Therefore, other processing factors are necessary in order to create a defect-free morphology. Lower solid loading can cause the formation of deficient part of particles around the moving ice front during freezing, because of the lower viscosity due to less frequent particle–particle and waterparticle interactions [4], eventually leading to the growth of large ice. When the cell wall is composed of fully packed particles, there should be a threshold value in the initial solid content. If the threshold is below that required to make the packed wall, porosities in the cell wall of green bodies can be formed, in which that can be supported by the results of shrinkage. However, a lower solid loading is very important processing factor to enhance the porosity of the resulting insulators. Thus, the AFP addition is critically required to avoid the formation of defects in the present gelation freezing technique. It is reasonable to consider that AFP addition prevented the formation of the coarse defects (large ice crystals), because of its pronounced effect on ice nucleation and grain growth. The detailed mechanism has been reported by many researchers [24,32–35]. Adsorption inhibition mechanism has been well accepted to retard the grain growth and recrystallization of ice crystals, in which sterically bulky AFP binds to the surface prism planes of ice crystals and further adsorption of water molecules on the surface are strongly inhibited. In fact, the cell size in insulators prepared with AFP addition was observed to be substantially smaller in a diameter, and without any large defects. We can state therefore that AFP needs to provide structural stability for thermal insulators. Fig. 3 shows the thermal conductivity of the obtained insulators, and their effective conductivity calculated by the Series and Parallel models (shown as solid lines) and the Maxwell–Eucken models (shown as dotted lines), as a function of porosity. The thermal conductivities of the prepared insulators were around 0.23–0.38 W/mK, and apparently affected by the porosity. Freezing temperature never affected the conductivities (see Table 2). Insulators can be generally treated as composite materials consisting of two different phases: a solid matrix and air in pores. The thermal conductivity of the air in the pores (0.025 W/mK) is substantially lower than that of the mullite (5 W/mK) [36–38]. Therefore, at higher porosities, the conductivity decreases. Series and Parallel models have been proposed to estimate the effective conductivity of cellular materials, particularly the lower and upper bounds (called as the Wiener bounds). The Series and
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model
ARTICLE IN PRESS
JECS-10303; No. of Pages 7
M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
Fig. 3. Thermal conductivities of the insulators at room temperature, and effective conductivities (room temperature) as a function of porosity, calculated using the Series and Parallel models shown as solid lines and the Maxwell–Eucken models as dotted lines.
Parallel models are calculated according to the following Eqs. (1) and (2) [39], respectively: ke =
1 (1 − vg )/ks + vg /kg
ke = (1 − vg ) × ks + vg × kg
(1) (2)
where ks , kg , and vg are the thermal conductivities of the solid phase (skeleton), and of air, and the volume fraction of air in the insulator (porosity), respectively. Hashin and Shtrikman derived further narrower bounds of the effective conductivity of a material with two phases, and these bounds always lie between the above Series and Parallel bounds [40]. Proposed Maxwell–Eucken 1 and 2 were computed according to the following Eqs. (3) and (4), respectively: ke = ks
2ks + kg − 2(ks − kg )vg 2ks + kg + (ks − kg )vg
(3)
ke = kg
2kg + ks − 2(kg − ks )(1 − vg ) 2kg + ks + (kg − ks )(1 − vg )
(4)
where ks , kg , and vg represent those described above in the same manner. The difference between the Eqs. (3) and (4) is whether the heat conduction pathway avoids pores or includes an outside solid phase. These models are defined as internal porosity (pores inside materials) or external porosity (pores outside materials), respectively. When pores are dispersed within a continuous condensed solid medium (internal porosity, such as a foam or honeycomb), an optimal heat transfer pathway should avoid the pores. In the literature, heat flux that could avoid pores was depicted as vectors for dispersed spherical pores within a continuous solid (a honeycomb-like structure comprised of spherical pores) on the basis of simulated finite element analysis of simple spherical dispersion model [39]. Furthermore, it is based on an assumption of the model that the pores inside materials do not make any contacts with each other. These restraint assumptions seem to fit with the honeycomb-like microstructures of the present insulators. Seen from the results, measured conductivities were closed to Maxwell–Eucken 1, and located on the side of Parallel models. In the Parallel model, heat flow can be transferred ideally along a direction parallel to materials, so the effective conductivity is larger than that predicted by Maxwell–Eucken 1. The experimental results observed on the side of the Parallel model indicates preferential heat transfer along the struts parallel to cell orientation. Fig. 4 shows the compressive strengths of the insulators obtained and predicted strength values of honeycomb out-of-plane model, in which a load was applied parallel to the cell orientation (the freezing direction). The compressive strength of the insulators ranged from 1.4 to 21.7 MPa, in which those of the insulators C and
5
Fig. 4. Compressive strengths of the obtained insulators and predicted strengths of honeycomb out-of-plane model proposed by Gibson and Ashby.
H were clearly lower than those of the others with similar porosities, again due to the large defects. It was found that the strength of the insulators was affected by porosities (solid loading in the slurry) and defects (AFP addition). Freezing temperature did not have large effect on the strength of the insulators. A honeycomb structure is relatively strong and stiff in the direction parallel to the cell orientation, compared with the in-plane values. This is advantageous for the present honeycomb-like microstructure. The out-of-plane model to predict the compressive strength of a honeycomb with ideal hexagonal cells has been proposed, which can be computed using the Eq. (5) developed by Gibson and Ashby [41]: c =C s
s
(5)
where c , s , C, and s indicate the compressive strength of a cellular solid, strength of a fully dense material, an experimental constant (dotted line computed by the slope of results), density of cellular structure and the fully density, respectively. We used 1310 MPa and 3.16 g/cc as reference values for the compressive strength and density of the fully dense mullite [42,43]. The experimental values were lower than that of predicted line. Compressive strengths of ceramic foams or cellular ceramics have been most frequently expressed as the 1.5 power of relative density, which has been widely utilized for the estimations [42,44–46], while that of a honeycomb is provided as a function of relative density [41]. When a net sectional stress in surface plane perpendicular to cell orientation exceeds fracture strength of a cell wall material ( s ), a honeycomb will be fractured. Thus, this means an upper limiting bound for the strength, as shown by the solid line of Fig. 4, i.e., completely defect-free honeycomb. If there are any defects contained in a honeycomb, it should be fractured at lower stress. As mentioned previously, cell walls observed were porous because of lower solid loading, including some contacting points among cells, and definitely different from ideal dense “beam” used as the assumption, certainly leading to reduced strength of insulators. Thus, it is reasonable to consider that the compressive strength of the insulators obtained can be lower than that of predicted values. We compared the thermal conductivities and compressive strength obtained in our study with those of references, which mostly included silica as additives to improve mechanical strength and reduce thermal conductivity, due to strong neck formed by liquid phase and its lower conductivity, but alumina included on some reports [47–49]. In their cases, the specimens sintered at 1500 ◦ C showed 11 MPa and 0.32 W/mK (porosity 78%), 3 MPa and 0.2W/mK (porosity 74%), and 5 MPa and 0.1 W/mK (porosity 81–84%) for the compressive strength and thermal conductivity, respectively, while the present insulators G exhibited 13.1 MPa and 0.24 W/mK (porosity 89.2%), for the strength and conductivity, respectively. Thus, the present insulators with the honeycomb-like structure showed higher strength than those of literatures, regardless of higher
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
6
porosity. The conductivity was also higher than those of above second and third ones, because they have used loose packed grains to create external porosity or mullite fiber as raw materials to provide narrow heat transfer path [47,50]. However, the comparison with the past studies suggests that carefully tailored external porosity combined with gelation freezing method have a great possibility to give rise to substantially reduced thermal conductivity, and together with improved strength, as future prospective. Work is still ongoing to develop thermal insulators with lowered thermal conductivity and improved compressive strength. 4. Conclusion Mullite thermal insulators were fabricated by a gelation freezing route, and the effects of gelation freezing processing factors on the thermal conductivity and compressive strength of the resulting insulators were investigated. The insulators have been comprised of micrometer sized honeycomb like morphology. The addition of AFP led to a defect-free microstructure and improved strength. The obtained insulators exhibited very low thermal conductivities of 0.23–0.38 W/mK that well agreed with Maxwell–Eucken 1 model. Compressive strength of the insulators ranged from 1.4 to 21.7 MPa, which was lower than that of theoretical prediction proposed by Gibson and Ashby. The proposed process is a simple, effective and ecofriendly method of fabricating highly porous ceramic insulators for use in a wide range of industrial applications. Acknowledgements This research work was performed as a part of Future Pioneering Projects/Research and Development of Thermal Management Materials and Technology by NEDO. The authors gratefully acknowledge their support. References [1] H.W. Russell, Principles of heat flow in porous insulators, J. Am. Ceram. Soc. 18 (1935) 1–5. [2] A.L. Loeb, Thermal conductivity.8. A theory of thermal conductivity of porous materials, J. Am. Ceram. Soc. 2 (1954) 96–99. [3] L. Hu, C.-A. Wang, Y. Huang, C. Sun, S. Lu, Z. Hu, Control of pore channel size during freeze casting of porous YSZ ceramics with unidirectionally aligned channels using different freezing temperatures, J. Eur. Ceram. Soc. 30 (16) (2010) 3389–3396. [4] M. Fukushima, Y.-i. Yoshizawa, Fabrication of highly porous silica thermal insulators prepared by gelation-freezing route, J. Am. Ceram. Soc. 97 (3) (2014) 713–717. [5] M. Fukushima, Y.-i. Yoshizawa, T. Ohji, Macroporous ceramics by gelation-freezing route using gelatin, Adv. Eng. Mater. 16 (6) (2014) 607–620. [6] T. Fukasawa, Z.Y. Deng, M. Ando, T. Ohji, Y. Goto, Pore structure of porous ceramics synthesized from water-based slurry by freeze–dry process, J. Mater. Sci. 36 (10) (2001) 2523–2527. [7] T. Fukasawa, M. Ando, T. Ohji, S. Kanzaki, Synthesis of porous ceramics with complex pore structure by freeze–dry processing, J. Am. Ceram. Soc. 84 (1) (2001) 230–232. [8] T. Fukasawa, Z.Y. Deng, M. Ando, T. Ohji, S. Kanzaki, Synthesis of porous silicon nitride with unidirectionally aligned channels using freeze–drying process, J. Am. Ceram. Soc. 85 (9) (2002) 2151–2155. [9] S. Deville, E. Saiz, R.K. Nalla, A.P. Tomsia, Freezing as a path to build complex composites, Science 311 (5760) (2006) 515–518. [10] S. Deville, E. Saiz, A.P. Tomsia, Ice-templated porous alumina structures, Acta Mater. 55 (6) (2007) 1965–1974. [11] S. Deville, Freeze-casting of porous ceramics: a review of current achievements and issues, Adv. Eng. Mater. 10 (3) (2008) 155–169. [12] R. Chen, C.-A. Wang, Y. Huang, L. Ma, W. Lin, Ceramics with special porous structures fabricated by freeze-gelcasting: using tert-butyl alcohol as a template, J. Am. Ceram. Soc. 90 (11) (2007) 3478–3484. [13] R. Chen, Y. Huang, C.-A. Wang, J. Qi, Ceramics with ultra-low density fabricated by gelcasting: an unconventional view, J. Am. Ceram. Soc. 90 (11) (2007) 3424–3429. [14] D. Koch, L. Andresen, T. Schmedders, G. Grathwohl, Evolution of porosity by freeze casting and sintering of sol–gel derived ceramics, J. Sol–Gel Sci. Technol. 26 (1–3) (2003) 149–152.
[15] S.R. Mukai, H. Nishihara, H. Tamon, Formation of monolithic silica gel microhoneycombs (SMHs) using pseudosteady state growth of microstructural ice crystals, Chem. Commun. 7 (2004) 874–875. [16] H. Nishihara, S.R. Mukai, D. Yamashita, H. Tamon, Ordered macroporous silica by ice templating, Chem. Mater. 17 (3) (2005) 683–689. [17] C.M. Pekor, P. Kisa, I. Nettleship, Effect of polyethylene glycol on the microstructure of freeze-cast alumina, J. Am. Ceram. Soc. 91 (10) (2008) 3185–3190. [18] C. Pekor, B. Groth, I. Nettleship, The effect of polyvinyl alcohol on the microstructure and permeability of freeze-cast alumina, J. Am. Ceram. Soc. 93 (1) (2010) 115–120. [19] E. Landi, F. Valentini, A. Tampieri, Porous hydroxyapatite/gelatine scaffolds with ice-designed channel-like porosity for biomedical applications, Acta Biomater. 4 (6) (2008) 1620–1626. [20] H.F. Zhang, I. Hussain, M. Brust, M.F. Butler, S.P. Rannard, A.I. Cooper, Aligned two- and three-dimensional structures by directional freezing of polymers and nanoparticles, Nat. Mater. 4 (10) (2005) 787–793. [21] M. Fukushima, M. Nakata, Y. Yoshizawa, Fabrication and properties of ultra highly porous cordierite with oriented micrometer-sized cylindrical pores by gelation and freezing method, J. Ceram. Soc. Jpn. 116 (1360) (2008) 1322–1325. [22] M. Fukushima, M. Nakata, Y. Zhou, T. Ohji, Y.I. Yoshizawa, Fabrication and properties of ultra highly porous silicon carbide by the gelation-freezing method, J. Eur. Ceram. Soc. 30 (14) (2010) 2889–2896. [23] M. Fukushima, S. Tsuda, Y.-i. Yoshizawa, Fabrication of highly porous alumina prepared by gelation freezing route with antifreeze protein, J. Am. Ceram. Soc. 96 (4) (2013) 1029–1031. [24] Y. Yeh, R.E. Feeney, Antifreeze proteins: structures and mechanisms of function, Chem. Rev. 96 (2) (1996) 601–617. [25] M.M. Harding, P.I. Anderberg, A.D.J. Haymet, Antifreeze’ glycoproteins from polar fish, Eur. J. Biochem. 270 (7) (2003) 1381–1392. [26] D. Slaughter, G.L. Fletcher, V.S. Ananthanarayanan, C.L. Hew, Antifreeze proteins from the sea raven, Hemitripterus-americanuS—further evidence for diversity among fish polypeptide antifreezes, J. Biol. Chem. 256 (4) (1981) 2022–2026. [27] C.H.C. Cheng, L.B. Chen, Evolution of an antifreeze glycoprotein, Nature 401 (6752) (1999) 443–444. [28] K. Hamamoto, M. Fukushima, M. Mamiya, Y. Yoshizawa, J. Akimoto, T. Suzuki, Y. Fujishiro, Morphology control and electrochemical properties of LiFePO4/C composite cathode for lithium ion batteries, Solid State Ionics 225 (2012) 560–563. [29] G.S. Perry, Microwave dielectric properties of mullite, Trans. J. Br. Ceram. Soc. 72 (6) (1973) 279–283. [30] B.L. Metcalfe, J.H. Sant, Synthesis, microstructure and physical properties of high-purity mullite, Trans. J. Br. Ceram. Soc. 74 (6) (1975) 193–201. [31] M. Fukushima, Y.-i. Yoshizawa, Fabrication of highly porous nickel with oriented micrometer-sized cylindrical pores by gelation freezing method, Mater. Lett. 153 (2015) 99–101. [32] J.A. Raymond, A.L. Devries, Adsorption inhibition as a mechanism of freezing resistance in polar fishes, Proc. Natl. Acad. Sci. U. S. A. 74 (6) (1977) 2589–2593. [33] C.A. Knight, A.L. Devries, L.D. Oolman, Fish antifreeze protein and the freezing and recrystallization of ice, Nature 308 (5956) (1984) 295–296. [34] G.L. Fletcher, M.H. Kao, R.M. Fourney, Antifreeze peptides confer freezing resistance to fish, Can. J. Zool.-Rev. Can. De Zool. 64 (9) (1986) 1897–1901. [35] R.L. McKown, G.J. Warren, Enhanced survival of yeast expressing an antifreeze gene analogue after freezing, Cryobiology 28 (5) (1991) 474–482. [36] L.M. Russell, L.F. Johnson, D.P.H. Hasselman, R. Ruh, Thermal conductivity/diffusivity of silicon carbide whisker reinforced mullite, J. Am. Ceram. Soc. 70 (10) (1987) C226–C229. [37] T.M. Kyaw, Y. Okamoto, K. Hayashi, Thermal conductivity of mullite-zirconia composites, J. Ceram. Soc. Jpn. 103 (12) (1995) 1289–1292. [38] T.W. Clyne, I.O. Golosnoy, J.C. Tan, A.E. Markaki, Porous materials for thermal management under extreme conditions, Philos. Trans. R. Soc.-Math. Phys. Eng. Sci. 364 (1838) (2006) 125–146. [39] J.K. Carson, S.J. Lovatt, D.J. Tanner, A.C. Cleland, Thermal conductivity bounds for isotropic, porous materials, Int. J. Heat Mass Transfer 48 (11) (2005) 2150–2158. [40] Z. Hashin, S. Shtrikman, A Variational approach to the theory of the effective magnetic permeability of multiphase materials, J. Appl. Phys. 33 (10) (1962) 3125. [41] Lorna J. Gibson, M. F. Ashby, Cellular Solids: Structure and Properties 2nd Edition, Cambridge University Press, (1999). [42] Y.W. Kim, H.D. Kim, C.B. Park, Processing of microcellular mullite, J. Am. Ceram. Soc. 88 (12) (2005) 3311–3315. [43] H. Ledbetter, S. Kim, D. Balzar, S. Crudele, W. Kriven, Elastic properties of mullite, J. Am. Ceram. Soc. 81 (4) (1998) 1025–1028. [44] P. Colombo, E. Bernardo, Macro- and micro-cellular porous ceramics from preceramic polymers, Compos. Sci. Technol. 63 (16) (2003) 2353–2359. [45] F.M. Varela-Feria, J. Martinez-Fernandez, A.R. de Arellano-Lopez, M. Singh, Low density biomorphic silicon carbide: microstructure and mechanical properties, J. Eur. Ceram. Soc. 22 (14–15) (2002) 2719–2725. [46] J.M. Tulliani, L. Montanaro, T.J. Bell, M.V. Swain, Semiclosed-cell mullite foams: preparation and macro- and micromechanical characterization, J. Am. Ceram. Soc. 82 (4) (1999) 961–968.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
G Model JECS-10303; No. of Pages 7
ARTICLE IN PRESS M. Fukushima, Y.-i. Yoshizawa / Journal of the European Ceramic Society xxx (2015) xxx–xxx
[47] Z. Hou, H. Du, J. Liu, R. Hao, X. Dong, M. Liu, Fabrication and properties of mullite fiber matrix porous ceramics by a TBA-based gel-casting process, J. Eur. Ceram. Soc. 33 (4) (2013) 717–725. [48] H. Qian, X. Cheng, H. Zhang, R. Zhang, Y. Wang, Preparation of porous mullite ceramics using fly ash cenosphere as a pore-forming agent by gelcasting process, Int. J. Appl. Ceram. Technol. 11 (5) (2014) 858–863.
[49] R.F. Zhang, J.J. Feng, X.D. Cheng, L.L. Gong, Y. Li, H.P. Zhang, Porous thermal insulation materials derived from fly ash using a foaming and slip casting method, Energy Build. 81 (2014) 262–267. [50] L. Gong, Y. Wang, X. Cheng, R. Zhang, H. Zhang, Porous mullite ceramics with low thermal conductivity prepared by foaming and starch consolidation, J. Porous Mater. 21 (1) (2014) 15–21.
Please cite this article in press as: M. Fukushima, Y.-i. Yoshizawa, Fabrication and morphology control of highly porous mullite thermal insulators prepared by gelation freezing route, J Eur Ceram Soc (2015), http://dx.doi.org/10.1016/j.jeurceramsoc.2015.09.041
7