- Email: [email protected]
Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies
Available online at www.sciencedirect.com
CERAMICS INTERNATIONAL
Ceramics International ] (]]]]) ]]]–]]] www.elsevier.com/locate/ceramint
Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies Xiang Penga,b, Shunzo Shimaia,c, Yi Suna,b, Guohong Zhoua, Shiwei Wanga,n a
The State Key Laboratory on High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China b University of Chinese Academy of Sciences, Beijing 100049, China c Tokyo University of Agriculture and Technology, Tokyo 183-8538, Japan Received 23 March 2015; received in revised form 11 May 2015; accepted 1 June 2015
Abstract Wet alumina green bodies with a dimension of 400 mm 50 mm 10 mm, respectively gelcast by PIBM (a copolymer of isobutylene and maleic anhydride) and EA (epoxy-amine) gel system, were dried at controlled temperature and humidity. Microstructure evolution and drying behavior of the green bodies with different solids loading and organic network were investigated. Pores among alumina particles became smaller and the constant rate period (CRP) became shorter with the increased solids loading or organic network. Further, the shrinkage of the body using PIBM ceased earlier and was smaller than that of the body using EA gel system. The typical microstructure of the body using PIBM gel system was thin organic networks on the particles and gradually a cocooned structure evolved during drying. While, the body using EA gel system had dense organic networks which evolved into a dense layer and strand-like structure around the particles. Such microstructures played different roles in water transportation and stress relaxation. As a result, the PIBM body was successfully dried without malformation but the EA body was bowing. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: A. Drying; Gelcasting; Microstructure evolution; Alumina
1. Introduction Gelcasting [1,2] is the most promising way to manufacture large and/or complex ceramic components by means of in situ coagulation to create a homogeneous green body using simple equipment and processing techniques [3,4]. However, drying of a wet gelcast green body is particularly troublesome because water covers about half the initial volume of the green body, and it tends to warp and crack caused by the non-uniform capillary pressure in the small pores of the gel network [5]. Drying is such a field that has been frequently re-studied in the past several decades. Scherer et al. presented a classical review on the drying theory of pure gels (e.g. silica gel containing -Si–O–Si- 3D network and water) by a sol–gel n
Corresponding author. Tel.: þ86 21 52414320; fax: þ 86 21 52415263. E-mail address: [email protected] (S. Wang).
processing [5]. For a gelcast ceramic green body, which is composed of ceramic particles and pure gel (organic network and water), the drying process was more complicated. On the one hand, the gelcast ceramic parts with a lower solids loading exhibited a higher drying rate using liquid desiccant drying method because solids loading significantly affected the drying potential [6,7]. On the other hand, Wang et al. gelcast alumina part by a modified free radical polymerization gel system, and demonstrated the graft chains upon heating could quickly release water through the skin layer formed during drying [8]. Moreover, Lewis et al. found that the gelcast alumina layer exhibited a complicated stress evolution because the PVA organic network was not beneficial for stress relaxation [9]. Furthermore, Ma et al. found that the organic network can be adjusted by a proper amount of hydroxyethyl acrylate so that the stresses were reduced in the ceramic green body [10]. For the first time, Ghosal et al. proposed a physical model which
http://dx.doi.org/10.1016/j.ceramint.2015.06.001 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
2
related particles with pure gels together, and assumed that the gel-matrix rupture happened due to the shrinkage of organic network around the particles. The rupture led to the change of channel for water transportation, and thus the drying process turned from constant rate period (CRP) to falling rate period (FRP) [11]. However, such a microstructure evolution has not been evidenced. It is inevitable that the microstructure, composed of solids loading and organic network, evolves during drying process for a gelcast green body. However, microstructure evolution and its influence on drying behavior have not been clarified so far. In this study, we designed wet alumina green bodies with a dimension of 400 mm 50 mm 10 mm for one-dimensional drying and noticeable shrinkage detection during drying. They were gelcast using a copolymer of isobutylene and maleic anhydride (PIBM, hereinafter) [12,13]. For comparison, alumina green body was also cast by a epoxy-amine gel system [14]. The microstructure evolution of the wet green bodies during drying process was demonstrated, and its mechanism was proposed. The relationship between microstructure evolution and drying behaviors was discussed. 2. Experimental procedure 2.1. Materials and samples preparation A commercial alumina powder (AES-11, Sumitomo, Osaka, Japan) with an average particle size of 0.45 μm was used as the raw material, which is shown in Fig. 1. PIBM with an average molecular weight of 55000–65000 (Isobam 104, Kuraray, Osaka, Japan), was used as both dispersant and gelling agent. The process for preparation of the alumina green body using PIBM was similar to a previous report [12]. Isobam 600AF from the same company with a molecular weight of 5500– 6500 was added to increase solids loading [15]. Deionized water, alumina powder and 0.3 wt% PIBM (0.2 wt% Isobam 600AFþ 0.1 wt% Isobam 104, relative to the weight of alumina powder) were mixed by ball-milling to make slurries containing 50–58 vol% solids. Slurries were degassed and cast into a mold with a typical size of 400 mm 50 mm 10 mm, and sealed by a plastic film to prevent water evaporation. The resultant alumina green body was noted as B-PIBM. The
Fig. 1. Scanning Electron Microscope (SEM) image of the raw alumina powder.
preparation of alumina green body using epoxy-amine gel system (noted as B-EA) was similar to that described by Mao et al. [14]. After gelling, the wet green bodies were dried in a commercial dryer (HWS-150, Sumsung, Shanghai, China) with a controlled temperature and relative humidity. 2.2. Measurements In situ drying loss (water removal) and shrinkage was measured by the modified commercial dryer mentioned above. An electronic balance was located on the top of the dryer to avoid the heat and humidity in the chamber (Fig. 2). A ruler was fixed to the sample supporter to measure the length of the sample. The supporter was tilted at 9 1in order to decrease the influence of the friction between the sample and the mold. The weight and length of the sample were recorded at 1 h intervals for the first day and at longer intervals on the following days. To measure the microstructure evolution during air drying, green bodies after different drying durations were firstly quenched in liquid nitrogen and then freeze dried. Microstructure of the green bodies on the fracture surface was evaluated using scanning electron microscopy (S-4800, Hitachi, Tokyo, Japan). Pore size distribution was measured using mercury porosimetry with a poremaster (PoreMaster-33, Quantachrome Corporation, Boynton Beach, FL). 3. Results and discussion 3.1. Influence of solids loading The influence of solids loading on water removal of the BPIBM was evaluated at a temperature of 40 1C and relative humidity (RH) of 60% (Fig. 3a). It's known that the rate of water vaporization (water loss) is the same in CRP stage because it only depends on the surface area and drying conditions. When solids loading increased from 50 to 58 vol %, the CRP duration of drying decreased from 20 to 10 h, and the volume fraction of residual water increased from 8% to 15% at the end of CRP. With further drying, the green body
Fig. 2. Modified dryer for measuring drying loss and shrinkage.
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
3
Fig. 3. (a) Volume fraction of residual water in wet alumina green bodies using PIBM from 50, 54, and 58 vol% solids loading during drying (temperature, 40 1C; RH, 60%); (b) Incremental intrusion versus pore size distribution of the freeze-dried bodies using PIBM from 50, 54, and 58 vol% solids loading when gelled for 2 h.
with a higher solids loading had a lower drying rate and contained more water till 36 h. The shorter CRP stage of the body with a higher solids loading is attributed to several factors. First, there is less water in the green body with a higher solids loading. Second, Fig. 3b shows that the pore (interstice) size among particles decreases from 105 nm to 70 nm when solids loading increases from 50 to 58 vol% before air drying. According to Scherer et al. [5], the small pores are a resistance to water transportation. Transportation through the small pores cannot continually supply enough water to the surface with the drying going on. As a result, water begins to recede from the body surface and then CRP stops. Therefore, the body with a higher solids loading has a shorter CRP. 3.2. Influence of organic network 3.2.1. Water transportation Organic network is one of the key components of the microstructure in a wet green body, which would significantly affect the drying behavior. Fig. 4a illustrats the drying curves of wet alumina green bodies prepared by PIBM and EA gel system, respectively (solid loading, 50 vol%; drying temperature, 40 1C; RH, 60%). The B-PIBM had a CRP of 20 h and 8 vol% water remained at the end of CRP. There was 2 vol% residual water till 36 h. In contrast, the B-EA had a CRP of 10 h and 25 vol% water remained at the end of CRP. There was 9 vol% residual water till 36 h. The difference in CRP of drying (20 and 10 h for the wet bodies using PIBM and EA gel system, respectively) is related to water transportation, which is dependent on their microstructures. Fig. 4b demonstrates that the pore size of the B-EA is 65 nm, smaller than that (105 nm) of the B-PIBM before air drying. As mentioned above, the small pores are a resistance to water transportation. The large pore is helpful for water transportation. As a result, water transportation in the
B-PIBM could compensate for the surface evaporation for a longer time. Therefore, the CRP of the B-PIBM is longer (20 h) and 42 vol% water is removed at this stage (Fig. 4a). However, water transportation in the B-EA is impeded and cannot compensate for the surface evaporation for a long time. As a result, the CRP of the B-EA is shorter (10 h) and only 25 vol% water is removed. 3.2.2. Shrinkage behavior Fig. 5 illustrats that the shrinkage of B-PIBM ceased earlier than B-EA, and the total shrinkage of B-PIBM was 3.4%, smaller than that of the B-EA (4.3%). Compared Fig. 5 with Fig. 4a, the shrinkages of both green bodies finished within their CRP stages although their gel systems were different, which confirms Ghosal's assumption [11]. The capillary forces result in the movement of particles during drying. When the particles move closer to each other, the organic network between them would impede their further movement (resulting in shrinkage decrease). Meanwhile, the organic network shrinks and deforms with the development of stresses. In the present case, the organic network in B-EA is denser than that in B-PIBM because of more organic addition in B-EA (4.5 wt%) than that in B-PIBM (0.3 wt%). That is, the denser organic network in B-EA would have a greater resistance to particle movement but finally have a larger shrinkage than that in B-PIBM. Therefore, the shrinkage process of B-EA lasts longer and its total shrinkage is larger than that of the B-PIBM. Microstructure evolution of the green bodies using different gel systems during drying will be further studied in the next section. 3.2.3. Microstructure evolution Microstructures of the green bodies in different drying durations (corresponding to different shrinkage stages shown in Fig. 5) were exhibited in Figs. 6 and 7. For B-PIBM at the beginning of drying, SEM observation revealed that the
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
4
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
Fig. 4. (a) Volume fraction of residual water in alumina green bodies using PIBM and EA (solids loading, 50 vol%; temperature, 40 1C; RH, 60%); (b) Incremental intrusion versus pore size distribution of the freeze-dried bodies using PIBM and EA gel system from 50 vol% solids loading when gelled for 2 h.
Fig. 5. Shrinkage behavior of alumina green bodies using PIBM, and EA gel system (solids loading, 50 vol%; temperature, 40 1C; and RH, 60%).
particles were homogeneously distributed (Fig. 6a). Compared with the raw powder (Fig. 1), the particle surface in B-PIBM was different though the trace of organic network was not directly observed by the present magnification of SEM. It might result from a small addition of PIBM (0.3 wt% compared to the powder). With further drying and shrinking, particles moved closer to each other and the stretched chains gradually bundled and covered on the surface of particles. Therefore, a thin layer of organic formed and cocooned the alumina particles (Fig. 6b–d). However, the particles in B-EA were surrounded by organic network at the beginning of drying (Fig. 7a). As the drying continued, the interparticle spaces appeared because of the organic network shrinkage [11], as well as the movement of near particles moved closer and formed “particle clusters” (Fig. 7b). With further drying, “organic strands” appeared between the “particle clusters” because of the shrinkage of organic network (arrows marked in Fig. 7c and d).
Based on the above results, microstructure evolution mechanism of the wet green bodies prepared by different gel systems is proposed. The organic network of PIBM forms gradually by twinning and/or bundling PIBM chains through hydrogen bonding and ionic interaction [12]. While, three dimensional networks forms quickly in the B-EA because nucleophilic addition reaction occurs rapidly between epoxy and amine [14]. At the beginning of drying, capillary forces lead to the movement of ceramic particles. When the particles get closer, however, the organic network between particles would impede the particle movement until the stress develops strong enough to deform or collapse the organic network. For B-PIBM with 0.3 wt% organic addition, the organic network is thin and has little resistance to shrinkage, and it easily shrinks to a layer coating and cocoons the alumina particles. However, for B-EA with 4.5 wt% organic addition, the network in B-EA is difficult to collapse and has a great resistance to shrinkage. As a result, the shrinkage process of B-EA lasts longer than that of the BPIBM (Fig. 5). Moreover, the dense network in B-EA severely shrinks which pulls alumina particles together and forms “particle clusters”, resulting in a larger shrinkage than that in the B-PIBM. According to Ghosal et al. [11], ruptures happen due to the shrinkage of organic network coated on the particles, and result in the interparticle spaces (voids). In the present case, the voids in the B-EA appear among “particle clusters” which were pulled by “organic strands”. 3.3. Appearance after drying Fig. 8 showed the dried bodies using (a) PIBM and (b) EA gel system respectively. Compared with B-PIBM without any malformation, the dried B-EA showed a severe warpage. As mentioned above, the pores with a large size in B-PIBM would generate low stress during drying. What is more, the thin organic network in B-PIBM generates little shrinkage and easily collapses. This would be helpful for stress relaxation. Consequently, malformation could not occur in the B-PIBM during
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
5
Fig. 6. Fractured surfaces of the green bodies using PIBM gel system from 50 vol% solids loading after drying for (a) 0 h; (b) 1 h; (c) 3 h; (d) 5 h. (temperature, 40 1C; and RH, 60%).
Fig. 7. Fractured surfaces of the green bodies using EA gel system from 50 vol% solids loading after drying for (a) 0 h; (b) 1 h; (c) 3 h; (d) 5 h. (temperature, 40 1C; RH, 60%).
drying (Fig. 8a). In fact, the B-PIBM could be dried efficiently without any crack or warpage within the flexible drying conditions (temperature, 30–50 1C; relative humidity, 40–80%). However, the pores with a small size in B-EA would develop large stress. Moreover, the dense organic network generates large shrinkage and does not collapse easily, which is detrimental to the stress relaxation. Therefore, warpage of the B-EA happens after
drying (Fig. 8b). This warpage could not recover because of the strong binding of the dense organic layer and organic strands. 4. Conclusions Wet alumina green bodies (400 mm 50 mm 10 mm) were gelcast with different microstructures (different solids
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
6
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
Acknowledgments This research work is funded by Science and Technology Commission of Shanghai Municipality (14XD1421200).
References
Fig. 8. Dried alumina green bodies (400 mm 50 mm 5 mm) using (a) PIBM and (b) EA gel system (solids loading, 50 vol%; temperature, 40 1C; RH, 60%).
loading and organic networks). The CRP of the bodies using PIBM gel system became shorter with the increase of solids loading from 50 to 58 vol%, which is relevant to the decreasing pore size. Compared with the body with the same solids loading using EA gel system, the body using PIBM had a longer CRP and less residual water after CRP; and finished shrinking in a shorter time and total shrinkage was smaller; and had a thinner organic network which was easier to collapse and helpful for stress relaxation. As a result, it could be dried easily and efficiently without crack or warpage. SEM observation revealed that organic networks by different gel systems led to the different microstructure evolutions, i.e. a thin layer and cocooned structure gradually formed on the particles in the body using PIBM, but a dense layer and strand-like structure formed on the particles in the body using EA gel system. They are responsible for the different drying behaviors mentioned above.
[1] O.O. Omatete, M.A. Janney, R.A. Strehlow, Gelcasting: a new ceramic forming process, Am. Ceram. Soc. Bull. 70 (10) (1991) 1641. [2] A.C. Young, O.O. Omatete, M.A. Janney, P.A. Menchhofer, Gelcasting of alumina, J. Am. Ceram. Soc. 74 (3) (1991) 612–618. [3] F.F. Lange, Powder processing science and technology for increased reliability, J. Am. Ceram. Soc. 72 (1) (1989) 3–15. [4] C. Chuanuwatanakul, C. Tallon, D.E. Dunstan, G.V. Franks, Producing large complex-shaped ceramic particle stabilized foams, J. Am. Ceram. Soc. 96 (5) (2013) 1407–1413. [5] G.W. Scherer, Theory of drying, J. Am. Ceram. Soc. 73 (11) (1990) 3–14. [6] Z.P. Zheng, D.X. Zhou, S.P. Gong, Studies of drying and sintering characteristics of gelcast BaTiO3-based ceramic parts, Ceram. Int. 34 (2008) 551–555. [7] A. Barati, M. Kokabi, M.H.N. Famili, Drying of gelcast ceramic parts via the liquid desiccant method, J. Eur. Ceram. Soc. 23 (2003) 2265–2272. [8] X.F. Wang, R.C. Wang, C.Q. Peng, H.P. Li, B. Liu, Z.Y. Wang, Thermoresponsive gelcasting: improved drying of gelcast bodies, J. Am. Ceram. Soc. 94 (6) (2011) 1679–1682. [9] M.A. Huha, J.A. Lewis, Polymer effects on the chemorheological and drying behavior of alumina-poly (vinyl alcohol) gelcasting suspensions, J. Am. Ceram. Soc. 83 (8) (2000) 1957–1963. [10] L.G. Ma, Y. Huang, J.L. Yang, H.R. Le, Y. Sun, Control of the inner stresses in ceramic green bodies formed by gelcasting, Ceram. Int. 32 (2006) 93–98. [11] S. Ghosal, A. Emami-Naeini, B.S. Ywh-PyngHarn, J.P. Draskovich, Pollinger, a physical model for the drying of gelcast ceramics, J. Am. Ceram. Soc. 82 (3) (1999) 513–520. [12] Y. Yang, S. Shimai, S.W. Wang, Room-temperature gelcasting of alumina with a water-soluble copolymer, J. Mater. Res. 28 (11) (2013) 1512–1516. [13] S. Shimai, Y. Yang, H. Kamiya, S.W. Wang, Spontaneous gelcasting of translucent alumina ceramics, Opt. Mater. Express 3 (8) (2013) 1000–1006. [14] X.J. Mao, S. Shimai, M.J. Dong, S.W. Wang, Gelcasting of alumina using epoxy resin as a gelling agent, J. Am. Ceram. Soc. 90 (3) (2007) 986–988. [15] Y. Sun, S. Shimai, X. Peng, M.J. Dong, H. Kamiya, S.W. Wang, A method for gelcasting high strength alumina ceramics with low shrinkage, J. Mater. Res. 29 (2) (2014) 247–251.
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
CERAMICS INTERNATIONAL
Ceramics International ] (]]]]) ]]]–]]] www.elsevier.com/locate/ceramint
Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies Xiang Penga,b, Shunzo Shimaia,c, Yi Suna,b, Guohong Zhoua, Shiwei Wanga,n a
The State Key Laboratory on High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China b University of Chinese Academy of Sciences, Beijing 100049, China c Tokyo University of Agriculture and Technology, Tokyo 183-8538, Japan Received 23 March 2015; received in revised form 11 May 2015; accepted 1 June 2015
Abstract Wet alumina green bodies with a dimension of 400 mm 50 mm 10 mm, respectively gelcast by PIBM (a copolymer of isobutylene and maleic anhydride) and EA (epoxy-amine) gel system, were dried at controlled temperature and humidity. Microstructure evolution and drying behavior of the green bodies with different solids loading and organic network were investigated. Pores among alumina particles became smaller and the constant rate period (CRP) became shorter with the increased solids loading or organic network. Further, the shrinkage of the body using PIBM ceased earlier and was smaller than that of the body using EA gel system. The typical microstructure of the body using PIBM gel system was thin organic networks on the particles and gradually a cocooned structure evolved during drying. While, the body using EA gel system had dense organic networks which evolved into a dense layer and strand-like structure around the particles. Such microstructures played different roles in water transportation and stress relaxation. As a result, the PIBM body was successfully dried without malformation but the EA body was bowing. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: A. Drying; Gelcasting; Microstructure evolution; Alumina
1. Introduction Gelcasting [1,2] is the most promising way to manufacture large and/or complex ceramic components by means of in situ coagulation to create a homogeneous green body using simple equipment and processing techniques [3,4]. However, drying of a wet gelcast green body is particularly troublesome because water covers about half the initial volume of the green body, and it tends to warp and crack caused by the non-uniform capillary pressure in the small pores of the gel network [5]. Drying is such a field that has been frequently re-studied in the past several decades. Scherer et al. presented a classical review on the drying theory of pure gels (e.g. silica gel containing -Si–O–Si- 3D network and water) by a sol–gel n
Corresponding author. Tel.: þ86 21 52414320; fax: þ 86 21 52415263. E-mail address: [email protected] (S. Wang).
processing [5]. For a gelcast ceramic green body, which is composed of ceramic particles and pure gel (organic network and water), the drying process was more complicated. On the one hand, the gelcast ceramic parts with a lower solids loading exhibited a higher drying rate using liquid desiccant drying method because solids loading significantly affected the drying potential [6,7]. On the other hand, Wang et al. gelcast alumina part by a modified free radical polymerization gel system, and demonstrated the graft chains upon heating could quickly release water through the skin layer formed during drying [8]. Moreover, Lewis et al. found that the gelcast alumina layer exhibited a complicated stress evolution because the PVA organic network was not beneficial for stress relaxation [9]. Furthermore, Ma et al. found that the organic network can be adjusted by a proper amount of hydroxyethyl acrylate so that the stresses were reduced in the ceramic green body [10]. For the first time, Ghosal et al. proposed a physical model which
http://dx.doi.org/10.1016/j.ceramint.2015.06.001 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
2
related particles with pure gels together, and assumed that the gel-matrix rupture happened due to the shrinkage of organic network around the particles. The rupture led to the change of channel for water transportation, and thus the drying process turned from constant rate period (CRP) to falling rate period (FRP) [11]. However, such a microstructure evolution has not been evidenced. It is inevitable that the microstructure, composed of solids loading and organic network, evolves during drying process for a gelcast green body. However, microstructure evolution and its influence on drying behavior have not been clarified so far. In this study, we designed wet alumina green bodies with a dimension of 400 mm 50 mm 10 mm for one-dimensional drying and noticeable shrinkage detection during drying. They were gelcast using a copolymer of isobutylene and maleic anhydride (PIBM, hereinafter) [12,13]. For comparison, alumina green body was also cast by a epoxy-amine gel system [14]. The microstructure evolution of the wet green bodies during drying process was demonstrated, and its mechanism was proposed. The relationship between microstructure evolution and drying behaviors was discussed. 2. Experimental procedure 2.1. Materials and samples preparation A commercial alumina powder (AES-11, Sumitomo, Osaka, Japan) with an average particle size of 0.45 μm was used as the raw material, which is shown in Fig. 1. PIBM with an average molecular weight of 55000–65000 (Isobam 104, Kuraray, Osaka, Japan), was used as both dispersant and gelling agent. The process for preparation of the alumina green body using PIBM was similar to a previous report [12]. Isobam 600AF from the same company with a molecular weight of 5500– 6500 was added to increase solids loading [15]. Deionized water, alumina powder and 0.3 wt% PIBM (0.2 wt% Isobam 600AFþ 0.1 wt% Isobam 104, relative to the weight of alumina powder) were mixed by ball-milling to make slurries containing 50–58 vol% solids. Slurries were degassed and cast into a mold with a typical size of 400 mm 50 mm 10 mm, and sealed by a plastic film to prevent water evaporation. The resultant alumina green body was noted as B-PIBM. The
Fig. 1. Scanning Electron Microscope (SEM) image of the raw alumina powder.
preparation of alumina green body using epoxy-amine gel system (noted as B-EA) was similar to that described by Mao et al. [14]. After gelling, the wet green bodies were dried in a commercial dryer (HWS-150, Sumsung, Shanghai, China) with a controlled temperature and relative humidity. 2.2. Measurements In situ drying loss (water removal) and shrinkage was measured by the modified commercial dryer mentioned above. An electronic balance was located on the top of the dryer to avoid the heat and humidity in the chamber (Fig. 2). A ruler was fixed to the sample supporter to measure the length of the sample. The supporter was tilted at 9 1in order to decrease the influence of the friction between the sample and the mold. The weight and length of the sample were recorded at 1 h intervals for the first day and at longer intervals on the following days. To measure the microstructure evolution during air drying, green bodies after different drying durations were firstly quenched in liquid nitrogen and then freeze dried. Microstructure of the green bodies on the fracture surface was evaluated using scanning electron microscopy (S-4800, Hitachi, Tokyo, Japan). Pore size distribution was measured using mercury porosimetry with a poremaster (PoreMaster-33, Quantachrome Corporation, Boynton Beach, FL). 3. Results and discussion 3.1. Influence of solids loading The influence of solids loading on water removal of the BPIBM was evaluated at a temperature of 40 1C and relative humidity (RH) of 60% (Fig. 3a). It's known that the rate of water vaporization (water loss) is the same in CRP stage because it only depends on the surface area and drying conditions. When solids loading increased from 50 to 58 vol %, the CRP duration of drying decreased from 20 to 10 h, and the volume fraction of residual water increased from 8% to 15% at the end of CRP. With further drying, the green body
Fig. 2. Modified dryer for measuring drying loss and shrinkage.
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
3
Fig. 3. (a) Volume fraction of residual water in wet alumina green bodies using PIBM from 50, 54, and 58 vol% solids loading during drying (temperature, 40 1C; RH, 60%); (b) Incremental intrusion versus pore size distribution of the freeze-dried bodies using PIBM from 50, 54, and 58 vol% solids loading when gelled for 2 h.
with a higher solids loading had a lower drying rate and contained more water till 36 h. The shorter CRP stage of the body with a higher solids loading is attributed to several factors. First, there is less water in the green body with a higher solids loading. Second, Fig. 3b shows that the pore (interstice) size among particles decreases from 105 nm to 70 nm when solids loading increases from 50 to 58 vol% before air drying. According to Scherer et al. [5], the small pores are a resistance to water transportation. Transportation through the small pores cannot continually supply enough water to the surface with the drying going on. As a result, water begins to recede from the body surface and then CRP stops. Therefore, the body with a higher solids loading has a shorter CRP. 3.2. Influence of organic network 3.2.1. Water transportation Organic network is one of the key components of the microstructure in a wet green body, which would significantly affect the drying behavior. Fig. 4a illustrats the drying curves of wet alumina green bodies prepared by PIBM and EA gel system, respectively (solid loading, 50 vol%; drying temperature, 40 1C; RH, 60%). The B-PIBM had a CRP of 20 h and 8 vol% water remained at the end of CRP. There was 2 vol% residual water till 36 h. In contrast, the B-EA had a CRP of 10 h and 25 vol% water remained at the end of CRP. There was 9 vol% residual water till 36 h. The difference in CRP of drying (20 and 10 h for the wet bodies using PIBM and EA gel system, respectively) is related to water transportation, which is dependent on their microstructures. Fig. 4b demonstrates that the pore size of the B-EA is 65 nm, smaller than that (105 nm) of the B-PIBM before air drying. As mentioned above, the small pores are a resistance to water transportation. The large pore is helpful for water transportation. As a result, water transportation in the
B-PIBM could compensate for the surface evaporation for a longer time. Therefore, the CRP of the B-PIBM is longer (20 h) and 42 vol% water is removed at this stage (Fig. 4a). However, water transportation in the B-EA is impeded and cannot compensate for the surface evaporation for a long time. As a result, the CRP of the B-EA is shorter (10 h) and only 25 vol% water is removed. 3.2.2. Shrinkage behavior Fig. 5 illustrats that the shrinkage of B-PIBM ceased earlier than B-EA, and the total shrinkage of B-PIBM was 3.4%, smaller than that of the B-EA (4.3%). Compared Fig. 5 with Fig. 4a, the shrinkages of both green bodies finished within their CRP stages although their gel systems were different, which confirms Ghosal's assumption [11]. The capillary forces result in the movement of particles during drying. When the particles move closer to each other, the organic network between them would impede their further movement (resulting in shrinkage decrease). Meanwhile, the organic network shrinks and deforms with the development of stresses. In the present case, the organic network in B-EA is denser than that in B-PIBM because of more organic addition in B-EA (4.5 wt%) than that in B-PIBM (0.3 wt%). That is, the denser organic network in B-EA would have a greater resistance to particle movement but finally have a larger shrinkage than that in B-PIBM. Therefore, the shrinkage process of B-EA lasts longer and its total shrinkage is larger than that of the B-PIBM. Microstructure evolution of the green bodies using different gel systems during drying will be further studied in the next section. 3.2.3. Microstructure evolution Microstructures of the green bodies in different drying durations (corresponding to different shrinkage stages shown in Fig. 5) were exhibited in Figs. 6 and 7. For B-PIBM at the beginning of drying, SEM observation revealed that the
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
4
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
Fig. 4. (a) Volume fraction of residual water in alumina green bodies using PIBM and EA (solids loading, 50 vol%; temperature, 40 1C; RH, 60%); (b) Incremental intrusion versus pore size distribution of the freeze-dried bodies using PIBM and EA gel system from 50 vol% solids loading when gelled for 2 h.
Fig. 5. Shrinkage behavior of alumina green bodies using PIBM, and EA gel system (solids loading, 50 vol%; temperature, 40 1C; and RH, 60%).
particles were homogeneously distributed (Fig. 6a). Compared with the raw powder (Fig. 1), the particle surface in B-PIBM was different though the trace of organic network was not directly observed by the present magnification of SEM. It might result from a small addition of PIBM (0.3 wt% compared to the powder). With further drying and shrinking, particles moved closer to each other and the stretched chains gradually bundled and covered on the surface of particles. Therefore, a thin layer of organic formed and cocooned the alumina particles (Fig. 6b–d). However, the particles in B-EA were surrounded by organic network at the beginning of drying (Fig. 7a). As the drying continued, the interparticle spaces appeared because of the organic network shrinkage [11], as well as the movement of near particles moved closer and formed “particle clusters” (Fig. 7b). With further drying, “organic strands” appeared between the “particle clusters” because of the shrinkage of organic network (arrows marked in Fig. 7c and d).
Based on the above results, microstructure evolution mechanism of the wet green bodies prepared by different gel systems is proposed. The organic network of PIBM forms gradually by twinning and/or bundling PIBM chains through hydrogen bonding and ionic interaction [12]. While, three dimensional networks forms quickly in the B-EA because nucleophilic addition reaction occurs rapidly between epoxy and amine [14]. At the beginning of drying, capillary forces lead to the movement of ceramic particles. When the particles get closer, however, the organic network between particles would impede the particle movement until the stress develops strong enough to deform or collapse the organic network. For B-PIBM with 0.3 wt% organic addition, the organic network is thin and has little resistance to shrinkage, and it easily shrinks to a layer coating and cocoons the alumina particles. However, for B-EA with 4.5 wt% organic addition, the network in B-EA is difficult to collapse and has a great resistance to shrinkage. As a result, the shrinkage process of B-EA lasts longer than that of the BPIBM (Fig. 5). Moreover, the dense network in B-EA severely shrinks which pulls alumina particles together and forms “particle clusters”, resulting in a larger shrinkage than that in the B-PIBM. According to Ghosal et al. [11], ruptures happen due to the shrinkage of organic network coated on the particles, and result in the interparticle spaces (voids). In the present case, the voids in the B-EA appear among “particle clusters” which were pulled by “organic strands”. 3.3. Appearance after drying Fig. 8 showed the dried bodies using (a) PIBM and (b) EA gel system respectively. Compared with B-PIBM without any malformation, the dried B-EA showed a severe warpage. As mentioned above, the pores with a large size in B-PIBM would generate low stress during drying. What is more, the thin organic network in B-PIBM generates little shrinkage and easily collapses. This would be helpful for stress relaxation. Consequently, malformation could not occur in the B-PIBM during
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
5
Fig. 6. Fractured surfaces of the green bodies using PIBM gel system from 50 vol% solids loading after drying for (a) 0 h; (b) 1 h; (c) 3 h; (d) 5 h. (temperature, 40 1C; and RH, 60%).
Fig. 7. Fractured surfaces of the green bodies using EA gel system from 50 vol% solids loading after drying for (a) 0 h; (b) 1 h; (c) 3 h; (d) 5 h. (temperature, 40 1C; RH, 60%).
drying (Fig. 8a). In fact, the B-PIBM could be dried efficiently without any crack or warpage within the flexible drying conditions (temperature, 30–50 1C; relative humidity, 40–80%). However, the pores with a small size in B-EA would develop large stress. Moreover, the dense organic network generates large shrinkage and does not collapse easily, which is detrimental to the stress relaxation. Therefore, warpage of the B-EA happens after
drying (Fig. 8b). This warpage could not recover because of the strong binding of the dense organic layer and organic strands. 4. Conclusions Wet alumina green bodies (400 mm 50 mm 10 mm) were gelcast with different microstructures (different solids
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001
6
X. Peng et al. / Ceramics International ] (]]]]) ]]]–]]]
Acknowledgments This research work is funded by Science and Technology Commission of Shanghai Municipality (14XD1421200).
References
Fig. 8. Dried alumina green bodies (400 mm 50 mm 5 mm) using (a) PIBM and (b) EA gel system (solids loading, 50 vol%; temperature, 40 1C; RH, 60%).
loading and organic networks). The CRP of the bodies using PIBM gel system became shorter with the increase of solids loading from 50 to 58 vol%, which is relevant to the decreasing pore size. Compared with the body with the same solids loading using EA gel system, the body using PIBM had a longer CRP and less residual water after CRP; and finished shrinking in a shorter time and total shrinkage was smaller; and had a thinner organic network which was easier to collapse and helpful for stress relaxation. As a result, it could be dried easily and efficiently without crack or warpage. SEM observation revealed that organic networks by different gel systems led to the different microstructure evolutions, i.e. a thin layer and cocooned structure gradually formed on the particles in the body using PIBM, but a dense layer and strand-like structure formed on the particles in the body using EA gel system. They are responsible for the different drying behaviors mentioned above.
[1] O.O. Omatete, M.A. Janney, R.A. Strehlow, Gelcasting: a new ceramic forming process, Am. Ceram. Soc. Bull. 70 (10) (1991) 1641. [2] A.C. Young, O.O. Omatete, M.A. Janney, P.A. Menchhofer, Gelcasting of alumina, J. Am. Ceram. Soc. 74 (3) (1991) 612–618. [3] F.F. Lange, Powder processing science and technology for increased reliability, J. Am. Ceram. Soc. 72 (1) (1989) 3–15. [4] C. Chuanuwatanakul, C. Tallon, D.E. Dunstan, G.V. Franks, Producing large complex-shaped ceramic particle stabilized foams, J. Am. Ceram. Soc. 96 (5) (2013) 1407–1413. [5] G.W. Scherer, Theory of drying, J. Am. Ceram. Soc. 73 (11) (1990) 3–14. [6] Z.P. Zheng, D.X. Zhou, S.P. Gong, Studies of drying and sintering characteristics of gelcast BaTiO3-based ceramic parts, Ceram. Int. 34 (2008) 551–555. [7] A. Barati, M. Kokabi, M.H.N. Famili, Drying of gelcast ceramic parts via the liquid desiccant method, J. Eur. Ceram. Soc. 23 (2003) 2265–2272. [8] X.F. Wang, R.C. Wang, C.Q. Peng, H.P. Li, B. Liu, Z.Y. Wang, Thermoresponsive gelcasting: improved drying of gelcast bodies, J. Am. Ceram. Soc. 94 (6) (2011) 1679–1682. [9] M.A. Huha, J.A. Lewis, Polymer effects on the chemorheological and drying behavior of alumina-poly (vinyl alcohol) gelcasting suspensions, J. Am. Ceram. Soc. 83 (8) (2000) 1957–1963. [10] L.G. Ma, Y. Huang, J.L. Yang, H.R. Le, Y. Sun, Control of the inner stresses in ceramic green bodies formed by gelcasting, Ceram. Int. 32 (2006) 93–98. [11] S. Ghosal, A. Emami-Naeini, B.S. Ywh-PyngHarn, J.P. Draskovich, Pollinger, a physical model for the drying of gelcast ceramics, J. Am. Ceram. Soc. 82 (3) (1999) 513–520. [12] Y. Yang, S. Shimai, S.W. Wang, Room-temperature gelcasting of alumina with a water-soluble copolymer, J. Mater. Res. 28 (11) (2013) 1512–1516. [13] S. Shimai, Y. Yang, H. Kamiya, S.W. Wang, Spontaneous gelcasting of translucent alumina ceramics, Opt. Mater. Express 3 (8) (2013) 1000–1006. [14] X.J. Mao, S. Shimai, M.J. Dong, S.W. Wang, Gelcasting of alumina using epoxy resin as a gelling agent, J. Am. Ceram. Soc. 90 (3) (2007) 986–988. [15] Y. Sun, S. Shimai, X. Peng, M.J. Dong, H. Kamiya, S.W. Wang, A method for gelcasting high strength alumina ceramics with low shrinkage, J. Mater. Res. 29 (2) (2014) 247–251.
Please cite this article as: X. Peng, et al., Correlation between microstructure evolution and drying behavior of gelcast alumina green bodies, Ceramics International (2015), http://dx.doi.org/10.1016/j.ceramint.2015.06.001