Chapter 5: Porous Ceramics and Metals by Ice Templating – Ice Templating and Freeze-Drying for Porous Materials and Their Applications

Porous Ceramics and Metals by Ice Templating

5.1 Introduction

Porous ceramics have unique applications as filters for water treatment/hot gases/molten metals, high‐temperature thermal insulation, catalysts or catalyst support, and bioscaffolds for tissue engineering (particularly bone engineering) [1]. These applications require porous ceramics with specific properties including high temperature stability, mechanical strength, interconnected porosity or closed pores, low density, and low thermal conductivity [2]. A variety of fabrication methods have been developed for the preparation of macroporous ceramics because most of these applications would require the presence of macropores (>50 nm by IUPAC definition, and indeed mostly in the micron range) in order to enhance mass transport or to accommodate cells. These methods usually fall under three categories: (i) the replicating approach by impregnation into a preformed structure; (ii) the sacrificial templating method where the templates are removed by washing, pyrolysis or evaporation; and (iii) the direct foaming method [1, 2].

Recently, ice templating and freeze‐drying have been used for fabrication of a wide range of porous structures including polymers, composites, ceramics, and nanostructured materials [35]. Indeed, the most investigated area is probably the preparation of porous ceramics, as evidenced by the large number of publications and reviews [69]. The advantages of ice templating for porous ceramics include: (i) simple processing procedures and hence high versatility; (ii) tuning porosity and pore sizes by the use of various additives and varied freezing conditions; and (iii) different freezing techniques to produce aligned porous structures and monoliths/films. Different terms have been used in literature for this technique, including ice templating, freeze‐drying, freeze casting, cryo‐construction, and cryo‐synthesis. Among them, the terms freeze casting (particularly for ceramics) and ice templating are used most often. The techniques used for the preparation of aligned porous materials may be termed as directional freezing, unidirectional freezing, orientated freezing, directional solidification or ice segregation induced self‐assembly (ISISA). Similarly, the resulting structures may be called aligned porous, parallel microchannels, aligned tubular, aligned channels or honeycomb structures. These structures may differ in their detail but they are usually produced by similar directional freezing techniques.

In this chapter, the focus is on porous ceramics by ice templating, covering formulations, freezing conditions, and freeze casting combined with gelation, which are prepared from ceramic particle slurries with the help of different solvents and additives. This is followed by the preparation of porous silica and metal oxides from the sols, with or without additives. The preparation of porous metals, either by direct freeze casting of metal particles or more often by reduction of porous metal oxides, is then described. The chapter concludes with the applications of porous ceramics fabricated by ice templating.

5.2 Porous Ceramics by Ice Templating

Ceramic particles are formulated as slurries, which are then frozen in controlled ways, depending on the targeted pore structure. The frozen samples are then freeze‐dried (or cryo‐gelled and vacuum dried) to produce green bodies. The green bodies are treated at high temperatures (usually in air) to remove the organic components and to sinter the ceramic particles to generate strong porous ceramics.

The pore structures in ice‐templated ceramics may be classified into homogeneous pores or heterogeneous pores [6]. The homogeneous pores are usually created by slow freezing in a cold chamber or by immersing in a cold liquid (such as liquid nitrogen), with a feature of highly interconnected cellular pores. The heterogeneous pores may be divided into aligned pores or graded pores. Applying a temperature gradient across a liquid sample is critical for the production of aligned pores or graded pores, known as directional freezing. This is often realized by dipping the liquid sample in a suitable container into liquid nitrogen at controlled rates [10, 11] or by the use of cold finger (as shown in Figure 5.1) [12]. For the dipping method, by using the suitable dipping rate, it is possible to maintain a constant temperature gradient at the freezing front and thereby a more uniform aligned structure. For the cold finger approach, the temperature of the cold finger may be controlled precisely by the temperature‐controlling unit. However, because the liquid sample is placed on the temperature‐controlled plate (Figure 5.1), the temperature gradient may change with the progress of the freezing interface, which creates a pore structure of three zones: dense structure, cellular structure, and then transition to lamellar structures [13, 14]. There are other ways of applying directional freezing, some of which are illustrated in Chapter 1. However, the aim is always the same: creating a controlled temperature gradient for the liquid sample.

Figure 5.1 The schematic representation of the experimental set‐up used for directional freezing by cold finger.

Source: Munch et al. 2009 [12]. Reprinted with permission from John Wiley and Sons.

5.2.1 Effect of Formulations Ceramic Particles

A variety of ceramic particles (usually sub‐micron or nanoparticles) have been processed to produce porous ceramics. Among them, alumina particles are the most investigated for the preparation of porous alumina [12, 1420] or as model particles for theoretical freezing studies [2125]. Examples of other particles include silica [10, 2628] or zeolite particles [29], hydroxyapatite [13, 3033], titania [3437], calcium phosphate [38, 39], yttria‐stabilized zirconia (YSZ) and hybrids [4042], perovskite material LSCF [43, 44], Gd‐doped ceria (GDC) [45], SiC [46], silicon nitride [47], perovskite‐type LaCoO3 [48], barium titanate‐hydroxyapatite [49], kaolinite‐silica [50], mullite‐zirconia [51], and alumina‐zirconia [52]. It should be pointed out that this is not an exhaustive list of porous ceramics fabricated by the ice‐templating technique. More examples can be found from the reviews on this topic [69].

Solid content is primarily related to porosity in porous ceramics. As summarized by Deville from the published data, the porosity is inversely proportional to the solid content in the slurries (Figure 5.2) in the concentration range of 5–60 vol% [8]. In general, the higher the ceramic particle concentration, the lower the porosity in the fabricated porous ceramics. The pores become less interconnected as well. If the percentage of the binder and stabilizer are similar, the freeze‐dried green bodies are weaker and are difficult to handle at lower ceramic contents. However, when the ceramic content is too high (>80 wt%), the desired pore structure and the pore interconnectivity may be lost [14]. For slurries with high solid contents (>50 vol%), the rejection of particles can be hindered by the larger number of surrounding particles. This can lead to the transition from cellular pores to dendritic patterns [6]. The solid content can also affect pore shapes when sintering at high temperatures. It has been observed that porous kaolinite–silica composites could generate rounded pores at high temperature of sintering and lead to more closed pores at high solid content. In contrast, for the ceramics prepared from low solid content slurries, the pore shrinkage and thickening of pore walls were heterogeneous, leading to elongated pores [50].

Figure 5.2 The relationship between the porosity in porous ceramics and the concentration in the slurries used to prepare such ceramics by freeze casting.

Source: Deville 2008 [8]. Reprinted with permission from John Wiley and Sons.

Particle size can have significant impacts on pore morphology and porosity and thereby mechanical stability. Deville et al. investigated the influence of alumina particles of different sizes (0.2, 0.4, 1.3 and, 3.4 μm) on nucleation and crystal growth [19]. Large particles provide less nucleation sites, which leads to highly supercooled systems before irreversible nucleation starts. In contrast, small particles have higher surface area and offer more nucleation sites. This results in a less supercooled state. Thus, by comparing the nucleation between large and small particles, a higher temperature gradient is created for large particles. It should be pointed out that this impact of particle size is mainly related to the initial nucleation and the transition to the stable progress of a freezing interface. A longer transition zone has been observed for larger particles whilst the freezing interface velocity is independent of particle sizes [19]. The size effect (1.69 and 3.9 μm) of hydroxyapatite particles on the formed scaffolds was investigated with regard to sintering, shrinkage, porosity, and compression strength [33]. For the cooling rates investigated, it was observed that higher cooling rates led to stronger scaffolds. Larger shrinkage, lower porosity, and higher compressive strength were found during sintering of the scaffolds made of larger particles [33].

The effects of particle sizes can be multiple, depending on particle density, shape, and slurry composition. The ceramic particles should be relatively small so that a stable particle suspension can be formed for the preparation of homogeneous ceramic materials. Owing to their higher density than those of the solvents, large particles may sediment rapidly in slurry. Even if a suspension is macroscopically stable, the movement of large particles when excluded from the freezing front can be limited. As a result, the targeted pore morphology may be lost. This is more significant when preparing ceramic films and membranes. In general, in relation to freezing velocity, it can be inferred that particle size can influence the entrapment of particles in ice as well as the stability of freezing front, which are the main causes for the formation of different ice‐templated morphologies [21]. Solvent

Unsurprisingly, water is the most used solvent when preparing porous ceramics by freeze casting. This is because of its low cost and environment‐friendly nature. However, there are some limitations as well: e.g. poor ability to suspend hydrophobic particles, low vapour pressure of ice, and higher freeze‐drying costs. However, there are also some other solvents that have been frequently used in freeze casting.

tert‐Butyl alcohol (TBA) can freeze below 25 °C and volatilize rapidly above 30 °C. This property is used to prepare porous alumina via freeze‐gel casting. The frozen sample may be thawed, with direct evaporation of TBA [20]. TBA exhibits high melting point (∼25 °C), low surface tension, and high saturation vapour pressure, which ensure fast freeze‐drying at higher drying temperature. TBA also shows low toxicity and has been often used as a co‐solvent with water for pharmaceutics [53]. TBA was used as the solvent to prepare porous alumina with long straight pores. The non‐dendritic morphology was different from the alumina prepared using water and camphene [16]. A sequential slurry casting method was used to prepare three layered alumina‐zirconia composites from the TBA slurries, with the layers being 95/5 alumina/zirconia, 85/15 alumina/zirconia, and 95/5 alumina/zirconia. Aligned microchannels were formed through the layers, with the bottom layer showing the smallest pores at ∼11 μm [54].

Camphene is another commonly used solvent. Camphene is one of three terpenes (the others being menthol and camphor), which can be used as a freezing vehicle for porous ceramics. They are non‐toxic aromatic substances derived from plants [55]. Terpenes are volatile solids that can be readily removed by sublimation at room temperature. Camphene is more widely used than the other two because of its suitable melting point (44–48 °C) and higher vapour pressure in solid state (0.330 kPa at 20 °C) (data from UNEP Publications, YSZ slurry in camphene was formed by ball milling at 60 °C for 20 h and then frozen in a water bath (15 °C). Conveniently, camphene was removed at room temperature to produce the green body. After sintering at 1450 °C for 3 h, YSZ foam with high compressive strength was produced [56]. A similar procedure was applied to fabricate porous ZrB2–SiC [57]. β‐SiAlON particles were suspended in pre‐warmed camphene at 60 °C to fabricate gradient porous ceramic materials [58].

Mixing solvents are also used to produce porous ceramics with varied morphologies. This is attributed to the changing composition of the solvents and the different crystal morphologies of each solvent. Water with another polar organic solvent (e.g. TBA and dioxane) is often employed. A TBA‐water system with varying contents of TBA (0, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100 wt%) was used to fabricate porous alumina [59]. With the increasing content of TBA, the prepared alumina showed the change of pore morphologies from lamellar, snowflake, and needle‐like, to hexagonal. Porous alumina with hexagonal pores made from TBA slurry exhibited the best mechanical strength [59]. The TBA‐water system was also employed to produce porous YSZ. In the TBA‐water phase diagram, two eutectic compositions (about 16 and 90 wt% of TBA) were identified. When the slurry was formed in the mix solvent with a TBA content of <16 wt% and frozen under suitable conditions, ice crystals formed first, concentrating the TBA content. As a result, a eutectic structure was formed. This resulted in the generation of large main pores and small pores in the large pore wall. In the other end of the phase diagram, when a slurry with 5 wt% water/95 wt% TBA was processed, the freeze‐dried structure exhibited large TBA‐templated faceted pores and smaller pores templated from the eutectic structure with around 90 wt% water [40]. Dispersant or Stabilizer

Ammonium polymethacrylate (anionic dispersant) is mostly used for aqueous slurries, usually at the concentration of 1 wt% for alumina powder [12, 14, 17, 2123], hydroxyapatite particles [13, 31], β‐tricalcium phosphate particles (β‐TCP, 1–2 μm) [39], YSZ particles [41, 42], and perovskite LSCF particles [43]. Zeta potential and viscosity of aqueous alumina slurries with different dispersants (ammonium polymethacrylate, sodium polymethacrylate, and ammonium polyacrylate) in the concentration range of 0.2–2 wt% were investigated. It was found that the viscosity was irrespective of the dispersant type, with the lowest viscosity at the concentration of ∼0.3 wt%. The concentration range of 0.2–0.4 wt% was found to be optimal for slurries with strong repulsive interactions between the particles. The dispersant concentration could significantly affect the development of a particle depletion zone and formation of ice lenses, which could considerably influence the mechanical strength of the porous ceramics formed [24].

Citric acid was used as a dispersant (1 wt% based on alumina powder) in slurries in TBA [16] or with acrylamide in TBA [20]. Similarly, citric acid was also employed in alumina/zirconia slurries in TBA [54]. Both citric acid (1 wt%) as dispersant and Dynol 604 (0.25 wt%, ethoxylated acetylenic diol) as surfactant were included in the slurries of biphasic calcium phosphate (BCP) particles (sieved under 75 μm) with acrylamide and methylenebisacrylamide in TBA in order to produce porous bioceramics [38]. For the mixing solvent of TBA and water, citric acid was used as the dispersant with polyvinylpyrrolidone (PVP) as binder (>50 wt% TBA). However, for the mixture solvent with <50 wt% TBA, sodium polyacrylate was the dispersant with carboxymethylcellulose (CMC) as binder [59]. In another study where the TBA‐water system was used to prepare porous YSZ, ProxB03 (a polyacrylate ammonium salt) and PEG6M were used as the dispersant and binder for TBA content up to 40 wt%. Only polyvinyl butyral was used as binder without dispersant for TBA content above 70 wt% [40].

Texaphor 963 at a concentration of 5 wt% to the solid powder was added to YSZ–camphene slurry for processing (with no other binder added) [56]. Texaphor 963 was also used for ZrB2-SiC–camphene slurry at the concentration of 3 wt% [57]. The same dispersant was employed at a concentration of 0.6 wt% for TiO2–camphene slurries (solid loading 10–20 wt%). These TiO2 particles were prepared by ethanediamine modification of Degussa P25, which inhibited the anatase‐to‐rutile phase transformation of TiO2 during heat treatment [36]. Binder

Poly(vinyl alcohol) (PVA) acts like a ‘glue’ to bind the ceramic particles together and obtain ice‐templated monolithic materials that can be easily handled [10, 12, 24]. Quite often, PVA also plays the role of additive with the concentrations of 5–10 wt%, leading to the formation of smaller pores in the resulting porous silica or ceramic materials [10, 12]. When PVA is used as a binder, particularly for slurries with high solid contents (e.g. 20–40 vol% solid particles), a low concentration such as 1 wt% may be sufficient for the production of green bodies, as evidenced in the preparation of porous hydroxyapatite [13, 31]. Both the concentrations of PVA as binder and ammonium polyacrylate as dispersant were varied to change the viscosity of the slurries with a subsequent big impact on porosity and compressive strength of the formed ceramic materials [31]. PVA and PVP were both added to silicon slurry (20 vol%) to act as binder and dispersant, respectively [47]. Bentonite clay as inorganic binder and polyethylene glycol (PEG) as sacrificial organic binder were employed in zeolite 13X slurries (particle size 3–5 μm) for the preparation of hierarchically porous silica [29]. PEG (1000 g mol−1) was added to aqueous β‐TCP slurry as organic binder [39]. PEG (1–4 wt%) was used as binder in perovskite LSCF slurries with solid loading of 40–50 wt% [43].

Poly(vinyl butyral) was used as binder in the slurry of alumina–citric acid–TBA [16] and also alumina–zirconia–citric acid–TBA slurries [54]. PVP and CMC were binders for the slurries of alumina–TBA–water with the varying contents of TBA in the system [59]. Additives

Some additives such as sodium chloride, sucrose, trehalose, glycerol, and ethanol are known to depress the solution freezing point (as antifreezers), most likely via the colligative effects [12]. The pore structure of porous alumina changes from lamellar (no additive, trehalose, sucrose) to cellular (gelatin, glycerol), or to a lamellar structure with a bimodal pore width distribution (ethanol). These additives can have varying impacts on interfacial tension and inter‐particle force, leading to varied bridge structures [12]. Glycerol (10 and 20 wt%) was added to 35 vol% alumina slurries, facilitating the formation of more defined porous structures [18]. Ethanol or 1‐propanol was added to aqueous alumina slurries, resulting in increased viscosity and subsequently decreased porosity and increased compressive strength in the sintered ceramics [60]. Glycerol was used as an antifreezer (1 wt% to solid powder) for the preparation of porous silicon [47].

Some polymers were added to the slurries to tune the pore sizes of the freeze‐dried monoliths. However, this does not exclude the polymers from also acting as binders for the monoliths produced [10, 27]. For example, PEG, γ‐cyclodextrin, and dextran were added to 40 wt% aqueous silica colloid suspensions (O‐40). The average channel size of the silica monoliths prepared by directional freezing decreased with the increasing concentration of dextran (MW = 17 500). Remarkably, at the dextran concentration of 10 wt%, the average channel size could be reduced to 180 nm [27]. This is a significant progress because most of the ice‐templated pores are in the micron range. The effect of PEG with different molecular weights (400, 900, 1350, and 3350 g mol−1) at the concentration of 5 wt% (based on water) was investigated on freeze‐cast alumina [61]. The average pore size changed from 10 to 5.5 μm with the increasing molecular weight of PEG. This could be explained by the diffusivity of PEG and constitutional supercooling [61].

Antifreeze proteins (AFPs) have also been used as additives to depress the freezing point of the slurries in order to form homogeneously distributed ice crystals in the whole frozen body [15]. There are two categories of proteins that are reported to depress the freezing temperature: antifreeze glycoproteins (AFGPs) and AFPs (type I, type II, and type III). The mechanism of reducing freezing temperatures from AFPs is non‐colligative because of the presence of thermal hysteresis between freezing temperature and melting temperature. Although the freezing temperature is lowered, the melting temperature has always been very close to 0 °C [62]. Also, the level in depressing the freezing temperature for AFPs at a concentration is considerably higher than those for colligative substances at similar concentrations. It has been postulated that reversible adsorption of AFPs on ice nuclei inhibit ice crystal growth (or allow growth on defined facets), leading to supercooled solutions [62]. Porous alumina was prepared by a gelation freezing method where alumina/gelatin solution was blended at the volume ratio of 10 : 90 with the AFP added into the gelatin solution at weight ratios of 0.25 : 99.75 and 0.50 : 99 : 50 to gelatin. The addition of AFP contributed to the generation of smaller and uniform ice crystals. This was evidenced by the decreasing average pores (from 102.5, 37.8 to 12.8 μm) when the gel was frozen at −20 °C. As shown in Figure 5.3, it is clear that porous alumina prepared with addition of AFP exhibits smaller and uniformly distributed pores, compared to the structure without adding AFP [15]. AFGP was also included in the gelatin system for the preparation of porous mullite‐zirconia insulators [51]. The obtained uniform pore channels led to a mullite insulator with high compressive strength and low thermal conductivity [51].

Figure 5.3 Effect of antifreeze protein (AFP) in the slurry on the pore structure of porous alumina prepared by freezing at −40 °C: (a) the pore structure from the slurry containing no AFP and (b) the pore structure prepared from the slurry containing 0.5 wt% AFP relative to gelatin (99.5 wt%).

Source: Fukushima et al. 2013 [15]. Reprinted with permission from John Wiley and Sons.

However, there are no definite differences between dispersant, binder, and additive for some polymers and surfactants included in the slurries. By definition, a dispersant (or stabilizer) is to help disperse the ceramic particles and form a stable dispersion while the binder is to hold the ceramic particles together and generate a monolithic green body after freeze‐drying. An additive can be anything added into the slurries but usually it is something that can influence pore shape, pore size, and mechanical stability. Some commonly used hydrophilic polymers such as PVA, PVP, CMC, and dextran can play all these roles in aqueous slurries although they are discussed as binders in the earlier sections. The same is true for polymeric surfactants. In general, the amphiphilic property contributes to stabilization whilst molecular weights may have considerable impact as binders and additives. When mixing polymeric and ionic surfactants, it may be difficult to identify their specific roles. For example, uniform silica microspheres were prepared by a modified Stöber method [63]. These silica microspheres were suspended in water at 30 wt% with 5 wt% PVA. In order to investigate the effect of charges on pore morphology, the cationic surfactant cetyltrimethylammonium bromide (CTAB) and anionic surfactant sodium dodecyl sulfate (SDS) were also included in the slurries. It can be observed from Figure 5.4 that a more cellular pore structure is formed with PVA whilst a much improved aligned structure is obtained with the inclusion of SDS and a layer structure with CTAB [26]. Therefore, although they are well‐known ionic surfactants, the role played by SDS or CTAB may be more like additives. These microspheres are mesoporous. The produced monolith could be coated with silica sol and a hierarchically porous silica monolith with improved mechanical stability could be obtained after calcination [26].

Figure 5.4 Effect of charged additives on the pore structure of porous silica prepared by directional freezing: (a) silica colloids with PVA; (b) silica colloids with PVA + anionic surfactant SDS; (c) silica colloids with PVA + cationic surfactant CTAB; and (d) zeta potentials measured in these three aqueous suspensions.

Source: Ahmed et al. 2011 [26]. Reprinted with permission from Royal Society of Chemistry.

5.2.2 Freezing Conditions Modes of Processing and Freezing

  • Directional freezing . This is commonly used for the preparation of aligned porous materials, in contrast to the freezing in a cold chamber or freezer (usually slow) for the production of randomly porous (or homogeneous) materials [6]. The key feature of directional freezing is to apply a temperature gradient across the liquid sample or slurry. The solvent nucleation and the crystal growth start at the cold end and move towards the warm end of the sample, as has been described in other literature [6466]. The directional freezing is usually used for the formation of 3D porous materials [610]. Although it has also been used for the preparation of porous films or membranes [34, 37, 43, 58, 67], some specific techniques have been developed to fabricate ceramic films and membranes.
    Figure 5.5 The schematic diagram for a freeze‐tape‐casting process.

    Source: Sofie 2007 [41]. Reprinted with permission from John Wiley and Sons.

  • Freeze‐tape‐casting. Figure 5.5 shows the schematic representation of a freeze‐tape‐casting process [41]. It is a process to fabricate thin ceramic materials (500–1500 μm in thickness). The process employs a traditional tape‐casting process to cast slurry onto a carrier substrate via a doctor blade apparatus. The slurry on the substrate then moves to the next section on a freezing bed to allow for directional freezing of the slurry. This process has been used to form porous YSZ [41] and GDC on NiO substrate [45] for potential energy applications. It should be noted that thin films can also be prepared by directly applying the slurry onto a substrate, which can be dipped into liquid nitrogen at the controlled rate [34] or placed onto a cold finger for directional freezing [37]. Both cases were demonstrated by the formation of ordered porous TiO2 film but the pore orientation was different; the pores were parallel to the substrate in the dipping approach [34] and perpendicular to the substrate in the cold finger approach [37].
    Figure 5.6 (a) The scheme for surface vacuum freezing. (b) Porous alumina prepared by surface vacuum freezing, h indicating the relative height of the sample.

    Source: Reprinted with permission from Ref. [69].

    Source: Reprinted with permission from Ref. [68].

  • Surface vacuum freezing. In this approach, the freezing process is initiated by introducing a vacuum to the liquid sample, instead of contacting a cold surface below the freezing point. Figure 5.6a shows schematically how the surface freezing by vacuum can be induced [68]. Basically, a drop of liquid on a substrate or a small volume of liquid sample in a vial is placed on a temperature‐controlled shelf in a chamber. The evaporation of water adsorbs heat, which leads to the surface temperature going down below 0 °C, initiating the freezing process at the surface. The surface freeze prevents further water evaporation but the ice sublimation continuously removes heat from the sample, drives down the temperature and completes the freezing (and freeze‐drying) process. This was demonstrated with a drop of aqueous silica colloidal suspension (45 nm, solid content 0.01–1 vol%) on a polytetrafluoroethylene (PTFE) plate placed on a shelf (the temperature is set slightly higher than 0 °C). The appearance of the drop did not change after applying the vacuum for 909 s. It was observed that the fibre morphology was formed at a higher rate of pressure decrease whilst a lower rate of pressure decrease led to the formation of porous sponges. It was recorded that the temperature could go down close to −20 °C [68]. Porous alumina can be similarly prepared by this technique. One such example is shown in Figure 5.6b. Three milliliter alumina slurry in a vial was placed on a shelf at 10 °C, with solid contents of 3.2, 6.4, and 12.8 vol%. The vacuum of 150 mTorr (0.2 mbar) was achieved and maintained during the freeze‐drying process. The alumina monoliths with aligned porous structures could be formed for all these slurries [69]. Similar to the usual directional freezing where the freezing usually starts from the bottom and a dense structure is observed at the bottom, the top surface (where the freezing starts) of the material prepared by surface vacuum freezing exhibits a dense structure followed by well‐defined lamellar structure (Figure 5.6b). However, there is also a dense‐structure layer at the bottom (Figure 5.6b). This may be attributed to the pushing‐back of freezing fronts due to high shelf temperature (different from the case discussed in Ref. [68]) and accumulation of ceramic particles [69]. The evaporation of water could reduce the surface temperature to below −10 °C and further sublimation could result in a fast declining temperature to below −35 °C. This was the cause for the formation of aligned porous ceramics [70].
    Figure 5.7 The pore structure of porous hydroxyapatite prepared by a directional freezing process using the cold finger set‐up described in Figure 5.1.

    Source: Deville et al. 2006 [13]. Reprinted with permission from Elsevier.

  • Graded porous ceramics. This type of ceramic material exhibits a gradient change of pore size and/or porosity across the samples. When aligned porous ceramics are prepared by the cold finger approach or the same temperature gradient cannot be maintained, a dense structure is first formed at the cold end (due to the highest temperature gradient and the entrapment of ceramic particles) followed by a transition zone (cellular structure) and subsequently by stable aligned/lamellar structures. One example is shown in Figure 5.7 for porous hydroxyapatite [13]. Similarly, gradient porous β‐SiAlON could be formed by freeze casting the camphene‐based slurry [58]. To better control or improve the graded porous structure, a two‐sided freezing approach has been employed. The two sides (the top and bottom of the liquid sample) are both cold so that the freezing starts from both surfaces and moves towards the centre. Alternatively, one side can be set as the cold surface while the other side has a higher temperature (maybe well above the freezing point) [20]. This way, the temperature gradient can be controlled higher or lower, which has a significant impact on the pore structure. In a more controlled way, graded porous ceramics can be fabricated by a serial freeze‐casting approach. Slurries of different loading or even of different types of solid particles can be placed in order and then frozen [54]. Alternatively, one may add the second layer after the first layer is frozen. Freeze casting may be also combined with screen‐printing to produce suitable materials for gas permeation [43, 44]. The freeze‐cast approach produces macroporous ceramics while the dense structure results from screen‐printing. The macropores facilitate mass transport but the small pores provide selection/impacts on gas molecules.

    Figure 5.8 Gradient porous PVA–silica composite prepared by freezing the centrifuged emulsion. (a) The evolution of the pore structure in the material, replicated from the centrifuge‐induced distribution of oil droplets in the emulsion. (b–d) The expanded views of the pore structures at different parts of the materials. This pore structure remains after calcination to remove PVA.

    Source: Ahmed et al. 2011 [71]. Reprinted with permission from Royal Society of Chemistry.

  • Centrifugation freezing. For a heterogeneous system (e.g. an emulsion, a slurry), the centrifugation force can be explored for the two phases with different densities in order to fabricate gradient porous materials. An oil‐in‐water emulsion, where cyclohexane was the internal oil phase and PVA/SDS/silica colloids were present in the continuous aqueous phase, was centrifuged to create a gradient distribution of oil droplets. This emulsion was then frozen and freeze‐dried to produce a gradient porous material [71]. As shown in Figure 5.8, from left to right of the image, the number of emulsion‐templated pores increases and the material becomes more porous even with a decreasing degree of ice‐templated pores. Such composite materials can be easily calcined to produce porous silica with increased porosity (due to the presence of large pores) in one direction and increased surface area (due to the dense silica with larger number of mesopores) in the other direction [71]. Although there seems to be no reports on the use of such technique for slurries (to explore the density difference between ceramic particles and the medium), this method may be readily applied to slurries to fabricate gradient porous ceramics.
  • Freeze castingwith magnetic field. When magnetic particles are included in the slurry, the magnetic field can exert an impact on the movement of these particles and thus on the final pore structure. High magnetic fields are used for ceramic processing but it requires special equipment and therefore it is inconvenient to be combined with freeze casting. Fe3O4 nanoparticles are known to respond to weak magnetic field (<0.18 T). Aqueous slurries of hydroxyapatite particles, ZrO2, Al2O3, and TiO2 (10 vol%) were mixed with different concentrations (0–8 vol%) of Fe3O4 nanoparticles (sizes <50 nm) [72]. Fe3O4 nanoparticles moved in the direction of the magnetic field. Particularly, when a rotating magnetic field (0.12 T) was applied around the freezing vial, a spiral dark stripe could be seen for all these freeze‐dried and sintered ceramics except for TiO2 slurry (Figure 5.9a). The composition of Fe in the Fe3O4‐rich region was 25–80 times higher than those in the Fe3O4‐poor region. For TiO2 ceramics, Fe3O4 were distributed quite homogeneously, with the Fe composition of 1.64 wt% in Fe3O4‐rich region versus the Fe composition of 1.61 wt% in Fe3O4‐poor region. This was attributed to the similar density of Fe3O4 and TiO2 particles. However, the alignment of the pore structure could still be considerably affected by the magnetic field due to the loss of pore alignment in the longitudinal direction [72]. By applying a magnetic field transverse to the freezing direction, movement of Fe3O4 nanoparticles could form bridges between the lamellar structures and enhance mechanical stability. This was demonstrated by directional freezing of the slurries containing ZrO2 (10 vol%) and Fe3O4 (0–9 wt%) nanoparticles with a transverse magnetic field (0–0.09 T) [73]. Figure 5.9b shows the bridges formed between the lamellars, with the Fe composition in the bridge about five times higher than that in the lamellars (Figure 5.9c). The compressive strength of the bridged ceramics prepared under the magnetic field of 0.09 T is about five times higher than the ceramics prepared with no magnetic field [73].
    Figure 5.9 Porous ceramics prepared by freeze casting with magnetic fields. (a) The images of different ceramics containing 3 wt% Fe3O4 nanoparticles with a rotating magnetic field of 0.12 T at 0.05 rpm. For each ceramic, the left image is after freeze‐drying and the right image after sintering. (b) The scanning electron microscopy (SEM) image shows a magnetically aligned porous ZrO2 with bridges after sintering formed from 10 vol% ZrO2 + 3 wt% Fe3O4 nanoparticles with a transverse magnetic field of 0.09 T (see the direction of the orange arrow); and (c) element map of this structure.

    Source: Reprinted with permission from Ref. [73].

    Source: Porter et al. 2016 [72]. Reprinted with permission from John Wiley and Sons;

  • Combining freezing with other templates. Other templates may be employed to generate additional pores in an ice‐templated material. Instead of mixing ceramic particles in slurries to make porous hybrid ceramics, sacrificial polymer colloids can be mixed with the ceramic particles. Both types of particles are rejected from the freezing front to form aligned porous materials. Removal of polymer particles by calcination can generate highly interconnected porous ceramic walls. This was demonstrated by a study conducted with silica colloids and poly(styrene‐co‐methacrylate) particles by directional freezing [74]. Pre‐formed porous structures can be also used as sacrificial templates to make porous ceramics [75, 76]. For example, alumina slurry was impregnated into a porous polyurethane sponge and a directional freezing process was then applied. After the removal of polyurethane by sintering, the resulting ceramics exhibited both ice‐templated aligned pores and polymer‐templated pores [77]. Emulsion templating and ice templating are combined to produce porous ceramics that contain both emulsion‐templated and ice‐templated pores [78]. Alternatively, the ice‐templated porous polymers can be used as sacrificial templates to fabricate porous silica/metal oxides [78], hollow TiO2 microtubes and Fe2O3 nanofibres [79], and silica spheres on nanofibres [80]. Freezing Velocities

It is not surprising that freezing conditions are critical for the preparation of porous ceramics by freeze casting. Freezing set‐ups and freezing temperatures translate to freezing direction and freeze velocity that determine whether porous ceramics can be formed and what the pore morphologies are. Freezing velocity, in combination with particle size and solid loading in slurry, is the most important factor in a freeze‐casting process. When the freezing velocity is very low, a planar freezing front pushes all the particles along: i.e. particles are concentrated but no ice‐templated porous ceramics may be formed. When a critical velocity is reached, the freezing front becomes unstable, leading to cellular, dendrite or lamellar crystal growth. At this stage, the particles are excluded from the freezing front and concentrated between the ice crystals [81]. Further increase of freezing velocity can change the ice crystal patterns and the thickness of ice crystals, which transforms to pore size and pore shape in the freeze‐dried materials [8]. However, very fast freezing leads to immediate entrapment of particles in the freezing front and no ice‐templated pores in the ceramics can be obtained after freeze‐drying. Theoretical studies on predicting and measuring critical velocity and wavelength of unstable freezing front (i.e. lamellar thickness) have been carried out, mostly using alumina, silica or clay particles as model particles in dilute slurries. This part has been extensively described in Chapter 2.

In a freeze‐casting process, due to the supercooling requirement for ice nucleation, the initial freezing rate is usually very high and then changes to a steady freezing velocity. This has been well investigated using the cold finger approach by X‐ray radiography and tomography [22, 23]. Such freezing behaviour is reflected in the structure of freeze‐dried ceramics. As demonstrated by porous hydroxyapatite (Figure 5.7) [13] and alumina [14], a dense structure is formed at the start of the freezing process (the bottom of the sample that contacts the cold finger plate) due to the entrapment of particles in the freezing front, followed by a transition zone and then a stable lamellar structure. It is believed that the discussion of the effects of freezing rates on pore structure/pore size is usually based on the steady freezing stage. In general, the faster the freezing rate, the larger the ice crystals and hence the larger pores. There are theoretical equations developed to predict the pore sizes (see Chapter 2). Empirically, the equation can be given as [8]:

where λ is the wavelength of the unstable freezing front (or the lamellae thickness), A is the coefficient related to the slurry system and v is the freezing velocity. For a given system, the data can be fitted into this equation and the values for A and n can be obtained.

Equation (5.1) only describes the simplified effect of freezing velocity on pore sizes. The true effects are complicated and should be considered in the context of particle size, type, and formulation compositions. Under the conditions of differing ionic strength, particle size, concentration, charges, and solvents, ice lenses may be formed transverse to the lamellar structure. After freeze‐drying and sintering, this defect in structure can considerably weaken the mechanical strength of the porous ceramics [24]. By systematically investigating (and observing) the patterns of ice crystals with those of the ceramic particles under different freezing conditions, diagrams of freezing velocity versus particle size can be constructed based on the stability of the ice freezing front (e.g. particle entrapment, metastable, unstable, stable (particle rejection from planar freezing front) [21]. These diagrams may serve as guides for the fabrication of porous ceramics.

5.3 Porous Ceramics by Gelation‐Freeze‐Casting

The gelation‐freeze‐casting method allows the gelation of ceramic slurry via thermal gelation or polymerization, followed by controlled freezing of the formed gels. This method has two distinct characteristics: (i) the effect on pore structure, which is different from that of polymer additives and (ii) the crosslinking of the system that holds the ceramic particles together and maintains the ice‐templated structure so that the solvent may be removed simply by vacuum or evaporation at room temperature. This means the energy‐consuming freeze‐drying process may not be necessary.

5.3.1 Gelation with Gelatin

This is probably the most used gelation‐freeze‐casting system for the preparation of porous ceramics. Porous ceramics prepared by this method exhibit nearly honeycomb‐like pore channels, unlike the ellipsoidal, dendritic or lamellar structures formed by conventional freeze casting [82]. Typically, the dry ceramic powder is suspended in warm gelatin solution at around 50 °C. The gelation is initiated when the temperature drops down to room temperature (15–20 °C), with the mixing ratio of powder to gelatin in the range of 1/99–10/90. The slurry gel is then frozen under suitable freezing conditions. In addition to the usual freeze‐drying, the frozen gel sample may be dried under vacuum at temperatures of 10–35 °C [82]. Unlike aqueous slurries, the movement of water and ceramic particles is limited in the gel state during the freezing process. The confined movement of water molecules in the gel may lead to the segregated ice nucleation and preferential growth of ice crystals as well as inhibit side branch growth from the main ice crystals, helping to eliminate defect formation in the freezing samples [82].

Porous silicon carbide was prepared via gelation with gelatin. The slurries in the moulds were allowed to gel under atmospheric condition and then frozen in cold ethanol bath at various temperatures (−10 to −70 °C). The frozen gels were then dried under vacuum at 10–35 °C for 24 h [83]. Porous alumina was prepared in a similar way but with AFP as additives. The AFP was added to the gelatin solution at the ratio of 0.25 : 99.75 and 0.50 : 99.50 and mixed with the alumina powders. The addition of the AFP led to homogeneously porous alumina with smaller pore sizes [15]. The AFP and gelation were also used to fabricate highly porous mullite‐zirconia which gave a compressive strength of 11.3 MPa, a porosity of 89.1 vol%, and a thermal conductivity of 0.28 W (m·K)−1 [51].

5.3.2 Photopolymerization (or Photocuring) of Frozen Slurry

Frozen polymerization has been used to prepare porous polymers, which removes the need of freeze‐drying and results in materials with stronger mechanical stability [8486]. This approach has also been used to fabricate porous ceramics. The UV‐curable monomers 4‐hydroxybutyl acrylate and PEG 200 diacrylate and the photoinitiator (a liquid blend containing 2,4,6‐trimethylbenzoylphenyl‐phosphineoxide with a broad absorption peak around 380 nm) were employed to fabricate porous silica and alumina [86]. An emulsion system was also used to form porous alumina microspheres. When AgNO3 was included in the system to prepare TiO2/polyacrylate composites, the in situ reduction of Ag+ ions on the TiO2 surface generated Ag nanoparticles, which triggered the photopolymerization without adding a photoinitiator [86]. In another study, PEG (400) diacrylate was dissolved in the terpenes (menthol, camphor, and camphene) above their melting points. Either silica or alumina powders were mixed into the solution with Irgacure 184 added as the photoinitiator. The suspensions in 1 cm‐high dishes were frozen at room temperature and photopolymerized subsequently at room temperature. The terpene solvents could be removed by sublimation or melting away. The usual calcination and sintering procedure was applied to produce porous ceramics [55].

5.3.3 Polymerization (or Gelation) of Slurry

There are other approaches to polymerizing ceramic slurries. One of the frequently used polymerization systems is with acrylamide. The initial polymerization temperature and rate can be fine‐tuned by combining acrylamide (or N‐isopropylacrylamide) and methylene bisacrylamide (the crosslinker) with the initiator ammonium persulfate (APS) and the catalyst tetramethylethylenediamine (TMEDA) with controlled concentrations [8789]. This polymerization has been used for ceramic slurries. For example, alumina powder was suspended in TBA containing acrylamide to which the initiator APS solution was added just before being frozen. The frozen solid was then heated in an ambient environment at 85 °C. This allowed the melting of TBA, polymerization of acrylamide, and evaporation of TBA in the same process. The polymerization of acrylamide contributed significantly to the strength of the green bodies [20].

In another study, collagen‐apatite gel was formed by incubating the solution in a sealed vial at 25 °C for 1 h and then increased to 40 °C at 0.5 °C min−1 and held at 40 °C for 22.5 h. The formed gel was then left to undergo self‐compression at room temperature for different times before being frozen in a pre‐cooled chamber at −25 °C and then freeze‐dried. The increased self‐compression time resulted in higher density of collagen fibres, reduced lamellar spacing, and increased wall thickness in the freeze‐dried scaffold [30]. Chitosan gels were formed by crosslinking with γ‐glycidoxypropyltrimethoxysilane and then mixed with aqueous suspension of TiO2 particles or TiO2 sols [90]. The subsequent directional freezing process and lyophilization produced the hybrid TiO2‐chitosan scaffolds that were used to degrade methylene blue and orange II dyes [90].

5.4 Porous Ceramics via Cryo‐Sol‐Gelation

Porous ceramics, sometimes cited as inorganic gels, can be obtained from a sol–gel process. The process usually starts with the hydrolysis of an inorganic or metal‐alkoxyl precursor to form a sol. The drying of the gel obtained from the sol is crucial to the porosity and density of the resulting material. The ambient drying process produces a xerogel, which usually exhibits shrinkage, pore structure collapse, and low surface area. To overcome these problems, a supercritical drying process can be applied to produce aerogels or a freeze‐drying process for cryogels [91].

Porous ceramics (or inorganic cryogels) with high surface area and controlled macropore morphologies can be prepared from the inorganic sols. The investigations were initiated with the directional freezing of silicic acid. After thawing, parallel silica fibres with polygonal cross‐sections (about 50 μm in diameter and 15 cm long) were formed. The aging time and concentrations were important factors for the formation of silica fibres [92]. This method has been extended to produce porous silica monoliths with aligned pores or microhoneycomb structures by Mukai et al. [11]. Sodium silicate solutions were adjusted using an ion‐exchange resin (Amberlite IR120B HAG) to give a pH of 3. The pH could also be adjusted by adding a few drops of aqueous ammonium solution while stirring. The obtained clear sol was aged at 303 K and then unidirectionally frozen by dipping a polypropylene tube containing the sol into liquid nitrogen. The frozen sample was then kept in a cold bath at 243 K for 2 h to strengthen the ice‐templated structure and then thawed at 323 K. The thawed sample was soaked in TBA for over 1 day to exchange the water in the gel and then freeze‐dried at 263 K to produce dry porous silica monolith with aligned channels of about 11 μm in width and surface areas in the region of 690–780 m2 g−1 [11]. The hydrosols (after adjusting the pH using the ion‐exchange resin) and the hydrogels (formed after aging the hydrosols with pH < 6) from sodium silicate solutions were further investigated for the preparation of silica gels with a variety of morphologies [93]. The silica gels with aligned channels could be further treated (heated at 800 °C for 2 h, impregnating the structure‐directing agents by soaking and drying at 50 °C, and then crystallized in an autoclave at temperatures of 100–130 °C) to transform into zeolite monoliths [94]. The crystalline zeolite particles (average diameter of 2.5 μm) were clearly seen on the pore surface and on the walls (Figure 5.10). The resulting zeolite monoliths showed type I isotherms of N2 sorption, which was indicative of the presence of micropores, while the ice‐templated macropores were retained [94].

Figure 5.10 Hierarchically porous zeolite monoliths by crystallization treatment of ice‐templated silica gels. (a–c) The ice‐templated silica gel at different magnifications; and (d–f) the macropore structure and the zeolite crystalline particles for the zeolite monoliths.

Source: Mori et al. 2011 [94]. Reprinted with permission from Royal Society of Chemistry.

For the silica sol produced at pH 3, aluminium nitrate was added to achieve the Si/Al ratios of 1.9–9.5. Subsequently, SiO2–Al2O3 monoliths were formed with aligned macropores (10–20 μm and wall thickness 200–500 nm) for the Si/Al ratios of 1.9–3.8 [93]. Al atoms were homogeneously distributed in the monoliths by forming an Al-O-Si polymeric network. Acid sites were confirmed by NH3‐TPD (temperature‐programmed desorption) measurement [95].

Tetraethyl orthosilicate (TEOS) is a common precursor for the preparation of silica via hydrolysis (usually catalysed by acid or base) and sol–gel processes. This can be used to prepare mesoporous silica via surfactant templating [63] or combined with polymer templates to make macroporous silica [75, 96]. The silica sol obtained by hydrolysis of TEOS is not often directly frozen to produce ice‐templated silica. This might be due to the production of ethanol during hydrolysis. Further, the ethanol solutions or gels are not suitable for freeze‐drying because of the very low melting point. In a study to fabricate bioglass from the TEOS silica sol, TEOS was hydrolysed with triethyl phosphate (TEP) in the presence of HNO3 (0.06 M) at room temperature with stirring for 1 day. The resulting sol was mixed with an equal volume of water. Ethanol generated during hydrolysis was removed under reduced pressure at 40 °C 30 mbar. The silica‐TEP sol was mixed with aqueous Ca(NO3)2 solution to achieve the desirable concentration and element molar ratio. The obtained sol was aged for 24 h at room temperature and then subjected to a directional freezing process. The green bodies obtained by freeze‐drying were annealed at 873 K for 5 h [97]. The macroporous bioglass prepared showed good capability for in vitro biomineralization [97].

Other types of gels are also freeze‐dried with the main goal being the formation of dry gels with high surface area and high thermal stability. For instance, Co3O4 cryogels with interconnected macroporosity and mesoporosity were prepared and tested as an electrode material by a three‐electrode system [98]. Co(NO3)2 was added to aqueous solution containing citric acid and P123 and the pH was adjusted to 3.33 with ammonia solution. The resulting solution was kept at 70 °C until a dark red gel was formed. The gels were frozen in liquid nitrogen and freeze‐dried, which was further sintered at 300 °C. The ice‐templated channels in the dry Co3O4 gels facilitated electrolyte diffusion, achieving a specific capacitance of 742.3 F g−1 at a potential window of 0.5 V [98]. Manganese oxide cryogels were also synthesized by freeze‐drying the relevant hydrogels [99]. Two routes were used to prepare the hydrogels. The first was by mixing the solutions of disodium fumarate and sodium permanganate. The second route was by mixing solid fumaric acid with aqueous sodium permanganate solution. The gelation occurred from a few minutes to 1 day depending on the route and the compositions in the solutions [99]. In another study, vanadyl triisopropoxide was mixed with the mixing solvent of water/acetone (each of these pre‐cooled in an ice bath for 15 min) and stirred vigorously for 30 s. The gel was formed in about 1 min and aged in a sealed tube for 1–8 days. The aged gel was washed with water, acetone, and solvent‐exchanged with cyclohexane, followed by freezing in liquid nitrogen and freeze‐dried [100]. The freeze‐dried gels were highly porous with specific surface areas up to 250 m2 g−1 [100]. Instead of porous monoliths, porous TiO2 fibres were obtained when the hydrogels were unidirectionally frozen and freeze‐dried [101]. The gels were prepared by hydrolysis of titanium tetraisopropoxide via slow addition of HCl at the ice‐cooled temperature and then the dialysis of the obtained solution. The gels were frozen in liquid nitrogen and aged at 243 K for 24 h. TiO2 fibres were produced after freeze‐drying and calcination at different temperatures [101].

Alumina cryogels produced via a sol–gel process from aluminium sec‐butoxide (ASB) and subsequent freeze‐drying showed higher thermal stability particularly at temperatures above 1000 °C [102]. ASB was added to hot water at 86 °C for hydrolysis and refluxed with vigorous stirring. The subsequent addition of nitric acid produced a clear sol. The sol gelation was initiated by adding urea and then allowed to stand for 24 h at 86 °C. The gel was then frozen in liquid nitrogen and freeze‐dried. The higher stability of this alumina cryogel was attributed to the uniform fibres formed in the gel [102]. Aqueous Pd(NO3)2 solution with ethylenediamine could be added to the sol before adding urea to initiate the gelation. Following the same process, alumina cryogel with uniformly distributed Pd could be produced, which showed high thermal stability and CO oxidation activity [103]. Similarly, cerium nitrate could be added to the sol to prepare CeO2-Al2O3 cryogel [104]. Zirconium dinitrate oxide was dissolved in water before adding ASB and the same procedure was applied to produce tertiary CeO2-ZrO2-Al2O3 cryogel. The effects caused on the thermal stability, pore structure, and catalytic CO and CH4 oxidation activities by adding ethylene glycol to the sol were investigated [104].

5.5 Porous Metals via Ice templating

Porous metals can be obtained directly from metal particles via freeze casting. For example, Ti foams were prepared by directional freezing of aqueous Ti slurries, followed by freeze‐drying and sintering. The sintering was performed under high vacuum at 1150 °C for 8 or 24 h. Longer sintering time was found to reduce the porosity but increase compressive stiffness, strength, and energy absorption [105]. Chemical stability of metal particles at room temperature and moderately elevated temperatures in atmosphere is required for this route. However, the sintering of metal particles at high temperatures is always carried out under inert atmosphere.

More often, porous metals are produced by reducing porous metal oxides that may be fabricated first by the ice‐templating approach. For instance, nickel foams were produced after reduction of freeze‐cast NiO scaffolds. NiO powder (<20 μm) slurry with PVA as binder was directionally frozen using a cold finger. The green body obtained by freeze‐drying was first heated to 300 °C for 2 h to remove PVA and then sintered at 1000 °C under H2 atmosphere containing 5% Ar. The presence of dispersant Darvan 811 could increase the mean pore size (from 10.9 to 14.2 μm) and total porosity (from 51.2% to 62.1%) [106]. Porous Cu with aligned channels (∼100 μm) was fabricated from freeze‐cast camphene‐based CuO. The CuO scaffold was reduced at 300 °C for 30 min and then sintered at 700 °C for 1 h under H2 [107]. CuO powder (∼200 mesh) was suspended in water in the presence of PVA as binder. The freeze‐dried sample was heated to 900 °C under vacuum (<0.04 Pa) to decompose CuO to Cu as well as to sinter the Cu particles into porous Cu [108]. In another study, the porous structure prepared by freeze‐drying of aqueous slurries of CuSO4 with carbon nanotube (CNT) and graphene oxide (GO) were first annealed at 700 °C under N2 to decompose CuSO4 and reduce GO. Further induction of forming gas (5.5 mol% of H2 in N2) could reduce Cu oxides to produce Cu/C composite. The ice‐templated macropores were retained [109]. Porous MoO3 was prepared by the freeze‐casting approach with camphor‐naphthalene as solvent [110]. After freezing in a cold bath at −25 °C, the frozen solvent was removed by sublimation at room temperature with an airflow rate of 0.05 m s−1 for 48 h. The reduction to porous Mo was carried out at 750 °C in H2 atmosphere for 1 h [110].

5.6 Applications of Ice‐templated Ceramics

The ice‐templating method can be used to fabricate highly interconnected porous ceramics with minimal shrinkage. Furthermore, via the change of slurry formulations and the control of the freezing stage, porous ceramics with desired pore morphology and mechanical stability can be produced. These can be highly advantageous for a number of applications.

5.6.1 Filtration/Gas Permeation

Owing to their chemical stability and mechanical stability, porous ceramics are widely used for filtrations. This is achieved by a size‐exclusion effect, i.e. larger particles are excluded from the ceramic membrane while the solvent or smaller particles can move across the membrane [1]. The control of pore size in ice‐templated porous ceramics can be conveniently achieved by changing the solid content in the slurries and the freezing rate. The aligned microchannels created by directional freezing can reduce the operational pressure due to less hindrance to the fluid. More importantly, because of their chemical stability at high temperatures, porous ceramics are widely used for hot gas filtration and molten metal filtrations [1]. The ice‐templated macropores in ceramics are usually not suitable for gas separation. However, this feature can be advantageous in enhancing mass transport when porous ceramics are used for other applications. For example, for oxyfuel technology, the porous ceramic membranes fabricated by freeze‐casting enhanced oxygen permeation [43, 44].

5.6.2 Thermal Insulator

Porous ceramics are extensively used as thermal insulators. The porosity, interconnectivity, and microstructure of the materials can significantly affect thermal conductivity [111]. Closed porosity can prevent convective heat transfer and is usually a good feature for thermal insulation. In general, higher porosity benefits thermal insulation but this should be always considered in conjunction with mechanical stability. This is because higher porosity is usually accompanied by low mechanical strength. For the ice‐templated ceramics, the aligned pores or channels may significantly enhance the efficiency of thermal insulation, as demonstrated in Figure 5.11 [1]. This is attributed to the prevention of cross‐flow convection current due to the aligned pores orthogonal to the heating surface and hence the reduced heat transfer. For example, porous mullite‐zirconia with aligned pores was prepared by gelation freezing. The materials exhibited low thermal conductivities of 0.23–0.33 W (m·K)−1, which were consistent with the values predicted by the Maxwell–Eucken 1 model [51]. The mullite foams were also fabricated using kaolin and Al2O3 slurries, using TBA or water as the dispersing medium [112]. Aligned pores were formed in the mullite and their sizes and porosity could be readily tuned by solid content. At the solid loading of 10 vol%, a porosity of 80% and a compressive strength of 5.6 MPa with TBA as solvent (similarly, 83% porosity and 3.8 MPa strength for the slurry with water) were achieved with a very low thermal conductivity of 0.18 W (m·K)−1 (0.17 W (m·K)−1 for water‐based slurry). For the TBA‐based slurries, when the solid content was increased to 20 vol%, a compressive strength of 49.4 MPa, a porosity of 60.2%, and a thermal conductivity of 0.34 W (m·K)−1 were obtained in the resulting mullite [112].

Figure 5.11 Schematic representation of heat transfer in the randomly porous material (a) and the aligned porous material (b). The blue colour in (b) indicates a better thermal insulating efficiency.

Source: Hammel et al. 2014 [1]. Reprinted with permission from Elsevier.

5.6.3 Bioceramics

Ice‐templating bioceramics have found wide use as scaffolds for tissue engineering. This is due to the highly interconnected macroporosity and good mechanical stability [7]. In order to be used for bone tissue engineering, the materials should support the growth of bone tissues as well as bond to the living bones. These bioceramics usually contain hydroxyapatite or β‐TCP [13, 30, 38, 39, 49]. The bone‐bonding ability can be effectively predicted by mineralization in a simulated body fluid (SBF). Over the years, the recipe of the SBF has evolved to mimic the ionic concentrations in human blood plasma, which has provided reliable results on bone bioactivity [113]. It has been established that a material able to form apatite on its surface can bond to living bone via the apatite layer, as long as the material does not contain the groups or components that induce toxic or antibody reactions. The short period required for apatite formation usually indicates the short period required to achieve bonding to living bones [113].

Porous hydroxyapatite scaffolds produced by freeze casting showed very high compressive strength, up to 145 MPa for 47% porosity or 65 MPa for 56% porosity [13]. Aligned porous barium titanate/hydroxyapatite composites exhibited high piezoelectric coefficients (d33 = 3.1 pC N−1). The 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl tetrazolium bromide (MTT) assay with L929 cells confirmed no cytotoxic effects [49]. The collagen–apatite scaffold was fabricated by a gelation‐freeze‐drying approach. This aligned porous scaffold exhibited a 12‐fold increase in Young's modulus and 2‐fold increase in the compression modules along the aligned direction, compared to the isotropic porous scaffold [30]. This scaffold's biocompatibility towards bone tissue engineering was demonstrated by the attachment and spreading of MC3T3‐E1 osteoblasts [30]. BCP scaffolds consisting of hydroxyapatite and β‐TCP were fabricated by freeze casting using TBA as solvent [38]. The in vitro degradation tests were carried out in Hank's balanced salt solution (HBSS), an extracellular solution with an ionic composition similar to human blood plasma. At the start of immersion in HBSS, the precipitation started to form on the scaffold surface. With the increase of soaking time, β‐TCP was slowly released for the scaffold sintered at 1200 °C whilst α‐TCP was quickly released for the scaffold sintered at 1300 °C. Further, with the soaking time, the particles gradually grew together and formed a dense layer on the BCP scaffold surface [38]. The ice‐templated β‐TCP scaffolds were subjected to cellular colonization evaluation with MG63 human osteoblastic cells [39]. The tests were carried out with the scaffolds fabricated using the poly(methylmethacrylate) (PMMA) beads (200–800 μm) as templates. It was found that the ice‐templated pores >100 μm allowed the cells to grow into the pores and all of the PMMA‐templated pores were colonized by the cells. However, different cell growth behaviours were observed. In the aligned ice‐templated tubular structure, the cells penetrated individually into the scaffold. But for the PMMA‐templated porous structure, the cells were grouped and very few isolated cells were observed. Further, the cell penetration depth (250 μm) in the ice‐templated scaffold was higher than that in the PMMA‐templated scaffold (150 μm) [39].

Macroporous bioglass scaffolds can also be prepared from the silica precursor with the incorporation of biomineral ions. For example, calcium nitrate was introduced into a TEOS sol. Subsequently, the macroporous bioglass was fabricated by a combination of sol–gel and freeze‐drying process [97]. The in vitro biocompatibility test with the SBF demonstrated the formation of a well‐distributed hydroxyapatite nanoparticulate layer [97]. In another interesting approach, silica colloids (HS‐40) were combined with cow's milk (to provide mineral ions and collagen as additional sacrificial templates) to generate hierarchically porous ice‐templated scaffolds [28]. The in vitro biomineralization resulted in the formation of a hydroxyapatite nanolayer on the scaffold surface within 24 h exposure time to the SBF [28].

5.6.4 Electric/electrode Materials

Ceramics with photo‐sensitive or electrical properties are used in these types of applications, one example of which is porous titania. Titania films with aligned microgrooves on alumina substrate were fabricated by a directional freezing process. The freeze‐dried films were heat treated at 1000 °C in air for 1.5 h and then at the same temperature for 3 h but under a reducing H2/Ar (5% H2) atmosphere [34]. This process generated oxygen vacancies in the rutile TiO2, resulting in electrically conducting substoichiometric TiO2 phase (known as Magnéli phase). For the TiO2 film with aligned microgrooves, the electrical resistance was found to be dependent on the orientation of the grooves. This electrical property was explored for electrical simulation. When a simulating electronic circuit was connected to the conducting patterned TiO2 film, charge‐balanced biphasic stimulus pulses with tuneable current amplitude and frequency were delivered [34]. In another study, TiO2 film with near vertical microchannels on fluorine‐doped tin oxide (FTO) glass was prepared by freeze casting and sintered in air at 480 °C for 30 min [37]. This film was then treated by soaking in the N719 (a type of Ru‐based dye)‐ethanol solution and sensitized for about 14 h at 60 °C. The sensitized TiO2 film was used as a photoanode to fabricate a dye‐sensitized solar cell (DSSC) device. A 13% higher photocurrent density was observed, when compared with the TiO2 film with randomly porous structure prepared by the conventional doctor blade method. This higher efficiency was attributed to the longer residence time of the light in the pores by reflecting several times from wall to wall and leading to a stronger absorption by the dye molecules [37].

Because of the stability at high temperatures, porous ceramics are used as electrodes for solid oxide fuel cells (SOFCs). Usually, a ceramics paste is applied and sintered to a solid electrolyte via a screening or tape‐casting procedure. The isotropic pore structure generated by this procedure exhibits a highly tortuous pathway and a longer transport time, resulting in low efficiencies [42]. This isotropic pore structure is also not ideal for the anisotropic stresses experienced during the cell's lifetime [114]. Porous ceramics with anisotropic porosity, such as NiO‐YSZ and lanthanum strontium manganite (LSM)‐YSZ composites, can be readily fabricated by a directional freeze‐casting process [42, 114].

YSZ electrolyte has been widely used for SOFC due to its superior chemical stability and oxygen‐ionic conductivity over a range of temperatures and oxygen partial pressures. However, the ion conductivity of YSZ decreases when the operation temperature is decreased. There is always a drive to find better‐performing materials at low or intermediate operation temperatures because this can considerably reduce the operation cost and improve the SOFCs stability and safety. GDC has proven to be a good electrolyte suitable for intermediate temperature (500–600 °C) [45]. In a study of the fabrication of hierarchically aligned macroporous anode‐supported SOFC, the aligned macroporous NiO‐GDC anode was first fabricated by a freeze‐drying tape‐casting process. A drop‐coating process was then applied to deposit GDC electrolyte slurry on the freeze‐dried NiO‐GDC. The resulting material was then sintered at 1400–1500 °C for 12 h. An SOFC cell was fabricated with the NiO‐GDC as the anode, the dense GDC film as the electrolyte and the LSCF‐GDC as the cathode. A high cell powder output of 1021 W cm−2 at 600 °C was obtained, a value almost two times higher than that achieved with similar cell materials but different NiO‐GDC pore structure [45].

5.6.5 Catalysis

Porous ceramics can act as a catalyst or as a support for catalyst. The ice‐templated interconnected macropores and, particularly, the aligned pores produced by directional freezing can facilitate mass transport or diffusion of reagent molecules to and from the active sites. The thermal stability of the ceramics can make them highly useful for high‐temperature catalytic reactions. For example, alumina cryogels with high thermal stability were fabricated from ASB via a sol–gel and subsequent freeze‐drying process [102]. Pd nanoparticles could be introduced into this alumina cryogel when palladium nitrate was included in the sol during the preparation stage. Compared to other types of dry gels (including xerogel), superior thermal stability of Pd nanoparticles was observed for the alumina cryogel, as demonstrated by the Pd dispersion in alumina (measured by CO chemisorption) when heated in air in the temperature range of 500–900 °C. The dispersion percentage was significantly decreased when the heating temperature was higher than 800 °C. The size of Pd nanoparticles was estimated to be about 2.5 nm after heating at 500 °C and about 3.0 nm after heating at 800 °C. These particle sizes were in good agreement with the slightly larger pore diameters. The uniformly distributed finer Pd nanoparticles in alumina cryogel resulted in high performance for CO oxidation, either after heating at 500 °C or 800 °C [103].

When a catalyst‐active component is included, the porous ceramics may be directly used as catalysts. Ceria is known to be an effective component to purify exhaust gases or act as promoter for methane reforming. Thus, binary ceramics CeO2-Al2O3 and ternary ceramics CeO2-ZrO2-Al2O3 were fabricated similarly by the sol–gel and freeze‐drying process. The addition of ethylene glycol could affect the pore structure and thermal stability of the ceramics. Transmission electron microscopy (TEM) imaging demonstrated the distribution of CeO2 nanoparticles throughout the cryogel. The cryogels were heated in O2-N2 and H2-N2 atmospheres separately at 500 °C for 1 h. For the CO oxidation, heating in H2-N2 atmosphere and addition of ethylene glycol during preparing gave lower T50 (the temperature required to achieve 50% of CO conversion). Similar impacts were observed for CH4 oxidation but were not as obvious [104]. This was attributed to the high dispersion of smaller CeO2 nanoparticles in the ceramic matrix under various treatments [104]. Another example was the preparation of LaCoO3 (perovskite‐type material) powder prepared by spray‐freezing/freeze‐drying and its performance for oxidation of methane and CO at low temperatures [48]. The LaCoO3 catalyst showed higher catalytic activity compared to double the amount of Pt/Al2O3 (prepared by a conventional impregnation method) under the same conditions (reaction temperatures of 300–440 °C). For CO oxidation, similar catalytic activity of LaCoO3 was observed, compared to five times the amount of Pt/Al2O3 catalyst [48].

TiO2 is well known as a photocatalyst for various catalytic reactions [115]. Porous TiO2 ceramics were synthesized by the freeze‐casting method using camphene as solvent [36]. TiO2 powders (Degussa P25) were treated with ethanediamine before processing. This procedure inhibited the anatase‐to‐rutile phase transformation during heat treatment (calcined at 800 °C). The slurry with a solid content of 15 wt% was optimal for the preparation of highly efficient and floating catalyst for photodegradation of atrazine and thiobencarb in water. The total organic carbon (TOC) removal efficiency was 95.7% and 96.7% for atrazine and thiobencarb (5 mg ml−1 solution), respectively. The catalyst could be re‐used at least six times without any obvious impact on its catalytic efficiency. Other pollutants were also tested effectively with this TiO2 catalyst [36]. The hybrid TiO2–chitosan scaffolds were fabricated by a gelation‐freeze‐casting method. The photodegradation of methylene blue and Orange II dyes was investigated under irradiation with a UV lamp at λ = 365 nm. Different TiO2 nanoparticles were used in this study and their effects on the efficiency of photodegradation were elaborated.

5.7 Summary

Ice templating or freeze casting is a simple and versatile process to produce a wide range of porous ceramics. It can be applied to a variety of particle slurries (either in water or organic solvent or a mixture of solvent) or inorganic sols. The porosity, pore size, and pore interconnectivity, in relation to mechanical strength, can be readily tuned by varying the slurry compositions and the freezing conditions. The most distinct advantage may be the preparation of aligned porous ceramics by the unique directional freezing process. These anisotropic pore structures may facilitate directional transport, guide cell growth, and provide unique properties that are suitable for thermal insulators. The combination of freeze casting with different gelation processes offers additional control on pore morphology and the mechanical strength of the resulting materials.

Ice‐templated porous ceramics have been employed for many applications, some of which are described in this chapter. The future development of porous ceramics by this technique may be more application‐orientated. Depending on the requirement of the targeted applications, e.g. large pores with interconnectivity, closed pores, unidirectional microchannels of desirable width, mechanical stability, high temperature stability, and light sensitive property, the ice‐templating method may be employed to fabricate the porous ceramics that meet these requirements. Additional templating, e.g. nanocasting, colloidal templating or emulsion templating, may be combined with ice templating to produce hierarchical pore structures that mimic biostructured materials including bones and bamboos [116]. Indeed, the layered porous ceramics made by freeze casting can be further processed with organic polymer infiltration, thus generating tough composites, mimicking the structure of nacre. More discussion on this topic is the focus of Chapter 6.


  1. 1 Hammel, E.C., Ighodaro, O.L.‐R., and Okoli, O.I. (2014). Processing and properties of advanced porous ceramics: an application based review. Ceram. Int. 40: 15351–15370.
  2. 2 Studart, A.R., Gonzenbach, U.T., Tervoort, E., and Gauckler, L.J. (2006). Processing routes to macroporous ceramics: a review. J. Am. Ceram. Soc. 89: 1771–1789.
  3. 3 Qian, L. and Zhang, H. (2011). Controlled freezing and freeze drying: a versatile route for porous and micro−/nano‐structured materials. J. Chem. Technol. Biotechnol. 86: 172–184.
  4. 4 Deville, S. (2013). Ice‐templating, freeze casting: beyond materials processing. J. Mater. Res. 28: 2202–2219.
  5. 5 Gutiérrez, M.C., Ferrer, M.L., and del Monte, F. (2008). Ice‐templated materials: sophisticated structures exhibiting enhanced functionalities obtained after unidirectional freezing and ice‐segregation‐induced self‐assembly. Chem. Mater. 20: 634–648.
  6. 6 Li, W.L., Lu, K., and Walz, J.Y. (2012). Freeze casting of porous materials: review of critical factors in microstructure evolution. Int. Mater. Rev. 57: 37–60.
  7. 7 Wegst, U.G.K., Schecter, M., Donius, A.E., and Hunger, P.M. (2010). Biomaterials by freeze casting. Phil. Trans. R. Soc. A 368: 2099–2121.
  8. 8 Deville, S. (2008). Freeze‐casting of porous ceramics: a review of current achievements and issues. Adv. Eng. Mater. 10: 155–169.
  9. 9 Liu, R., Xu, T., and Wang, C.‐A. (2016). A review of fabrication strategies and applications of porous ceramics prepared by freeze‐casting method. Ceram. Int. 42: 2907–2925.
  10. 10 Zhang, H., Hussain, I., Brust, M. et al. (2005). Aligned two‐ and three‐dimensional structures by directional freezing of polymers and nanoparticles. Nat. Mater. 4: 787–793.
  11. 11 Mukai, S.R., Nishihara, H., and Tamon, H. (2004). Formation of monolithic silica gel microhoneycombs (SMHs) using pseudosteady state growth of microstructural ice crystals. Chem. Commun. 874–875.
  12. 12 Munch, E., Saiz, E., and Tomsia, A.P. (2009). Architectural control of freeze‐cast ceramics through additives and templating. J. Am. Ceram. Soc. 92: 1534–1539.
  13. 13 Deville, S., Saiz, E., and Tomsia, A.P. (2006). Freeze casting of hydroxyapatite scaffolds for bone tissue engineering. Biomaterials 27: 5480–5489.
  14. 14 Deville, S., Saiz, E., and Tomsia, A.P. (2007). Ice‐templated porous alumina structures. Acta Mater. 55: 1965–1974.
  15. 15 Fukushima, M., Tsuda, S., and Yoshizawa, Y. (2013). Fabrication of highly porous alumina by gelation freezing route with antifreeze protein. J. Am. Ceram. Soc. 96: 1029–1031.
  16. 16 Li, D. and Li, M. (2012). Preparation of porous alumina ceramic with ultra‐high porosity and long straight pores by freeze casting. J. Porous. Mater. 19: 345–349.
  17. 17 Zheng, J., Salamon, D., Lefferts, L. et al. (2010). Ceramic microfluidic monoliths by ice templating. Microporous Mesoporous Mater. 134: 216–219.
  18. 18 Tallón, C., Moreno, R., and Nieto, M.I. (2009). Shaping of porous alumina bodies by freeze casting. Adv. Appl. Ceram. 108: 307–313.
  19. 19 Deville, S., Maire, E., Lasalle, A. et al. (2010). Influence of particle size on ice nucleation and growth during the ice‐templating process. J. Am. Ceram. Soc. 93: 2507–2510.
  20. 20 Chen, R., Wang, C.‐A., Huang, Y. et al. (2007). Ceramics with special porous structures fabricated by freeze‐gelcasting: using tert‐butyl alcohol as template. J. Am. Ceram. Soc. 90: 3478–3484.
  21. 21 Deville, S., Maire, E., Bernard‐Granger, G. et al. (2009). Metastable and unstable cellular solidification of colloidal suspensions. Nat. Mater. 8: 966–972.
  22. 22 Deville, S., Maire, E., Lasalle, A. et al. (2009). In situ X‐ray radiography and tomography observations of the solidification of aqueous alumina particle suspensions. Part I: initial instants. J. Am. Ceram. Soc. 92: 2489–2496.
  23. 23 Deville, S., Maire, E., Lasalle, A. et al. (2009). In situ X‐ray radiography and tomography observations of the solidification of aqueous alumina particle suspensions. Part II: steady state. J. Am. Ceram. Soc. 92: 2497–2503.
  24. 24 Lasalle, A., Guizard, C., Maire, E. et al. (2012). Particle redistribution and structural defect development during ice templating. Acta Mater. 60: 4594–4603.
  25. 25 Bareggi, A., Maire, E., Lasalle, A., and Deville, S. (2011). Dynamics of the freezing front during the solidification of a colloidal alumina aqueous suspension: in situ X‐ray radiography, tomography, and modelling. J. Am. Ceram. Soc. 94: 3570–3578.
  26. 26 Ahmed, A., Clowes, R., Myers, P., and Zhang, H. (2011). Hierarchically porous silica monoliths with tuneable morphology, porosity, and mechanical stability. J. Mater. Chem. 21: 5753–5763.
  27. 27 Nishihara, H., Iwamura, S., and Kyotani, T. (2008). Synthesis of silica‐based porous monoliths with straight nanochannels using an ice‐rod nanoarray as a template. J. Mater. Chem. 18: 3662–3670.
  28. 28 Onna, D., Minaberry, Y., and Jobbágy, M. (2015). Hierarchical bioglass scaffolds: introducing the “milky way” for template bioceramics. J. Mater. Chem. B 3: 2971–2977.
  29. 29 Ojuva, A., Akhtar, F., Tomsia, A.P., and Bergström, L. (2013). Laminated adsorbents with very rapid CO2 uptake by freeze‐casting of zeolites. ACS Appl. Mater. Interfaces 5: 2669–2676.
  30. 30 Xia, Z., Villa, M.M., and Wei, M. (2014). A biomimetic collagen–apatite scaffold with a multi‐level lamellar structure for bone tissue engineering. J. Mater. Chem. B 2: 1998–2007.
  31. 31 Zhang, Y., Zhou, K., Bao, Y., and Zhang, D. (2013). Effects of rheological properties on ice‐templated porous hydroxyapatite ceramics. Mater. Sci. Eng. C 33: 340–346.
  32. 32 Zhou, K., Zhang, Y., Zhang, D. et al. (2011). Porous hydroxyapatite ceramics fabricated by an ice‐templating method. Scr. Mater. 64: 426–429.
  33. 33 Zamanian, A., Farhangdoust, S., Yasaei, M. et al. (2014). The effect of particle size on the mechanical and microstructural properties of freeze‐casted macroporous hydroxyapatite scaffolds. Int. J. Appl. Ceram. Technol. 11: 12–21.
  34. 34 Romeo, H.E., Trabadelo, F., Jobbagy, M., and Parra, R. (2014). 2D‐ice template titanium oxide films as advanced conducting platforms for electrical stimulation. J. Mater. Chem. C 2: 2806–2814.
  35. 35 Nezafati, N., Hafezi, M., Zamanian, A., and Naserirad, M. (2015). Effect of adding nano‐titanium dioxide on the microstructure, mechanical properties and in vitro bioactivity of a freeze cast merwinite scaffold. AIChE 31: 550–556.
  36. 36 Xing, Z., Zhou, W., Du, F. et al. (2014). Facile synthesis of hierarchical porous TiO2 ceramics with enhanced photocatalytic performance for micropolluted pesticide degradation. ACS Appl. Mater. Interfaces 6: 16653–16660.
  37. 37 Wang, X., Yang, B., Zhou, K. et al. (2015). Ordered porous TiO2 films obtained by freezing and the application in dye sensitized solar cells. Curr. Appl. Phys. 15: 662–668.
  38. 38 Kim, D.‐H., Kim, K.‐L., Chun, H.‐H. et al. (2014). In vitro biodegradable and mechanical performance of biphasic calcium phosphate porous scaffolds with unidirectional macropore structure. Ceram. Int. 40: 8293–8300.
  39. 39 Meurice, E., Bouchart, F., Hornez, J.C. et al. (2016). Osteoblastic cells colonization inside beta‐TCP macroporous structures obtained by ice‐templating. J. Eur. Ceram. Soc. 36: 2895–2901.
  40. 40 Guizard, C., Leloup, J., and Deville, S. (2014). Crystal templating with mutually miscible solvents: a simple path to hierarchical porosity. J. Am. Ceram. Soc. 97: 2020–2023.
  41. 41 Sofie, S.W. (2007). Fabrication of functionally graded and aligned porosity in thin ceramic substrates with the novel freeze‐tape‐casting process. J. Am. Ceram. Soc. 90: 2024–2031.
  42. 42 Lichtner, A.Z., Jauffres, D., Martin, C.L., and Bordia, R.K. (2013). Processing of hierarchical and anisotropic porosity LSM‐YSZ composites. J. Am. Ceram. Soc. 96: 2745–2753.
  43. 43 Gaudillere, C., Garcia‐Fayos, J., and Serra, J.M. (2014). Enhancing oxygen permeation through hierarchically‐structured perovskite membranes elaborated by freeze‐casting. J. Mater. Chem. A 2: 3828–3833.
  44. 44 Gaudillere, C., Garcia‐Fayos, J., Balaguer, M., and Serra, J.M. (2014). Enhanced oxygen separation through robust freeze‐cast bilayered dual‐phase membranes. ChemSusChem 7: 2554–2561.
  45. 45 Chen, Y., Zhang, Y., Baker, J. et al. (2014). Hierarchically oriented macroporous anode‐supported solid oxide fuel cell with thin ceria electrolyte film. ACS Appl. Mater. Interfaces 6: 5130–5136.
  46. 46 Han, D., Mei, H., Xiao, S. et al. (2017). Porous SiCnw/SiC ceramics with unidirectionally aligned channels produced by freeze‐drying and chemical vapour infiltration. J. Eur. Ceram. Soc. 37: 915–921.
  47. 47 Kim, D.S. and Kim, D.K. (2015). Hierarchical structure of porous silicon nitride ceramics with aligned pore channels prepared by ice‐templating and nitridation of silicon powder. Int. J. Appl. Ceram. Technol. 12: 921–931.
  48. 48 Lee, S.H., Lee, J.Y., Park, Y.M. et al. (2006). Complete oxidation of methane and CO at low temperature over LaCoO3 prepared by spray‐freezing/freeze‐drying method. Catal. Today 117: 376–381.
  49. 49 Zhang, Y., Chen, L., Zeng, J. et al. (2014). Aligned porous barium titanate/hydroxyapatite composites with high piezoelectric coefficients for bone tissue engineering. Mater. Sci. Eng. C 39: 143–149.
  50. 50 Li, W., Lu, K., and Walz, J.Y. (2013). Effects of solids loading on sintering and properties of freeze‐cast kaolinite‐silica porous composites. J. Am. Ceram. Soc. 96: 1763–1771.
  51. 51 Fukushima, M. and Yoshizawa, Y. (2016). Fabrication of highly porous honeycomb‐shaped mullite‐zirconia insulators by gelation freezing. Adv. Powder Technol. 27: 908–913.
  52. 52 Liu, G., Zhang, D., Meggs, C., and Button, T.W. (2010). Porous Al2O3–ZrO2 composites fabricated by an ice template method. Scr. Mater. 62: 466–468.
  53. 53 Vessot, S. and Andrieu, J. (2012). A review on freeze drying of drugs with tert‐butanol (TBA) + water systems: characteristics, advantages, drawbacks. Dry. Technol. 30: 377–385.
  54. 54 Choi, H.J., Yang, T.Y., Yoon, S.Y. et al. (2012). Porous alumina/zirconia layered composites with unidirectional pore channels processed using a tertiary‐butyl alcohol‐based freeze casting. Mater. Chem. Phys. 133: 16–20.
  55. 55 Tomeckova, V. and Halloran, J.W. (2012). Porous ceramics by photopolymerization with terpene‐acrylate vehicles. J. Am. Ceram. Soc. 95: 3763–3768.
  56. 56 Hong, C., Zhang, X., Han, J. et al. (2010). Camphene‐based freeze‐cast ZrO2 foam with high compressive strength. Mater. Chem. Phys. 119: 359–362.
  57. 57 Du, J., Zhang, X., Hong, C., and Han, W. (2013). Microstructure and mechanical properties of ZrB2–SiC porous ceramic by camphene‐based freeze casting. Ceram. Int. 39: 953–957.
  58. 58 Hou, Z., Ye, F., Liu, L., and Liu, Q. (2012). Fabrication of gradient porous β‐SiAlON ceramics via a camphene‐based freeze casting process. Mater. Sci. Eng. A 558: 742–746.
  59. 59 Tang, Y., Qiu, S., Wu, C. et al. (2016). Freeze cast fabrication of porous ceramics using tert‐butyl alcohol–water crystals as template. J. Eur. Ceram. Soc. 36: 1513–1518.
  60. 60 Zeng, J., Zhang, Y., Zhou, K., and Zhang, D. (2014). Effects of alcohol additives on pore structure and morphology of freeze‐cast ceramics. Trans. Nonferrous Metals Soc. China 24: 718–722.
  61. 61 Pekor, C. and Nettleship, I. (2014). The effect of the molecular weight of polyethylene glycol on the microstructure of freeze‐cast alumina. Ceram. Int. 40: 9171–9177.
  62. 62 Yeh, Y. and Feeney, R.E. (1996). Antifreeze proteins: structures and mechanism of function. Chem. Rev. 96: 601–617.
  63. 63 Ahmed, A., Clowes, R., Willneff, E. et al. (2010). Synthesis of uniform porous silica microspheres with hydrophilic polymer as stabilizing agent. Ind. Eng. Chem. Res. 49: 602–608.
  64. 64 Yoon, H.‐J., Kim, U.‐C., Kim, J.‐H. et al. (2010). Macroporous alumina ceramics with aligned microporous walls by unidirectionally freezing foamed aqueous ceramic suspensions. J. Am. Ceram. Soc. 93: 1580–1582.
  65. 65 Zhang, H. and Cooper, A.I. (2007). Aligned porous structures by directional freezing. Adv. Mater. 19: 1529–1533.
  66. 66 Zhang, H., Long, J., and Cooper, A.I. (2005). Aligned porous materials by directional freezing of solution in liquid CO2. J. Am. Chem. Soc. 127: 13482–13483.
  67. 67 Qian, L., Ahmed, A., Glennon‐Alty, L. et al. (2015). Patterned substrates fabricated by a controlled freezing approach and biocompatibility evaluation by stem cells. Mater. Sci. Eng. C 49: 390–399.
  68. 68 Inasawa, S., Katayama, T., and Yamaguchi, Y. (2016). Surface freezing and surface coverage as key factors for spontaneous formation of colloidal fibers in vacuum drying of colloidal suspensions. Soft Matter 12: 7663–7669.
  69. 69 Groβberger, S., Fey, T., and Lee, G. (2016). Vacuum‐induced surface freezing to produce monoliths of aligned porous alumina. Materials 9: 983.
  70. 70 Groβberger, S., Fey, T., and Lee, G. (2017). Freezing kinetics of vacuum‐induced surface directional freezing in a glass vial. Chem. Eng. Sci. 167: 154–160.
  71. 71 Ahmed, A., Smith, J., and Zhang, H. (2011). Gradient porous materials by emulsion centrifugation. Chem. Commun. 47: 11754–11756.
  72. 72 Porter, M.M., Yeh, M., Strawson, J. et al. (2012). Magnetic freeze casting inspired by nature. Mater. Sci. Eng. A 556: 741–750.
  73. 73 Porter, M.M., Niksiar, P., and McKittrick, J. (2016). Microstructural control of colloidal‐based ceramics by directional solidification under weak magnetic fields. J. Am. Ceram. Soc. 99: 1917–1926.
  74. 74 Kim, J.‐W., Tazumi, K., Okaji, R., and Ohshima, M. (2009). Honeycomb monolith‐structured silica with highly ordered, three‐dimensionally interconnected macroporous walls. Chem. Mater. 21: 3476–3478.
  75. 75 Zhang, H., Hardy, G.C., Khimyak, Y. et al. (2004). Synthesis of hierarchically porous silica and metal oxide beads using emulsion‐templated scaffolds. Chem. Mater. 16: 4245–4256.
  76. 76 Zhang, H. and Cooper, A.I. (2005). Emulsion‐templated hierarchically porous silica beads using silica nanoparticles as building blocks. Ind. Eng. Chem. Res. 44: 8707–8714.
  77. 77 Han, J., Hu, L., Zhang, Y., and Zhou, Y. (2009). Fabrication of ceramics with complex porous structures by the impregnate‐freeze‐casting process. J. Am. Ceram. Soc. 92: 2165–2167.
  78. 78 Qian, L., Ahmed, A., Foster, A. et al. (2009). Systematic tuning of pore morphologies and pore volumes in macroporous materials by freezing. J. Mater. Chem. 19: 5212–5219.
  79. 79 Qian, L., Willneff, E., and Zhang, H. (2009). A novel route to polymeric sub‐micron fibers and their use for inorganic structures. Chem. Commum. 3946–3948.
  80. 80 Ahmed, A., Clowes, R., Willneff, E. et al. (2010). Porous silica spheres in macroporous structures and on nanofibers. Phil. Trans. R. Soc. A 368: 4351–4370.
  81. 81 Waschkies, T., Oberacker, R., and Hoffmann, M.J. (2011). Investigation of structure formation during freeze‐casting from very slow to very fast solidification velocities. Acta Mater. 59: 5135–5145.
  82. 82 Fukushima, M., Yoshizawa, Y., and Ohji, T. (2014). Macroporous ceramics by gelation‐freezing route using gelatin. Adv. Eng. Mater. 16: 607–620.
  83. 83 Fukushima, M., Nakata, M., Zhou, Y. et al. (2010). Fabrication and properties of ultra highly porous silicon carbide by the gelation‐freezing method. J. Eur. Ceram. Soc. 30: 2889–2896.
  84. 84 Barrow, M., Eltmimi, A., Ahmed, A. et al. (2012). Frozen polymerization for aligned porous structures with enhanced mechanical stability, conductivity, and as stationary phase for HPLC. J. Mater. Chem. 22: 11615–11620.
  85. 85 Barrow, M. and Zhang, H. (2013). Aligned porous stimuli‐responsive hydrogels via directional freezing and frozen UV initiated polymerization. Soft Matter 9: 2723–2729.
  86. 86 de Hazan, Y. (2012). Porous ceramics, ceramic/polymer, and metal‐doped ceramic/polymer nanocomposites via freeze casting of photo‐curable colloidal fluids. J. Am. Ceram. Soc. 95: 177–187.
  87. 87 Zhang, H. and Cooper, A.I. (2002). Synthesis of monodisperse emulsion‐templated polymer beads by oil‐in‐water‐in‐oil (O/W/O) sedimentation polymerization. Chem. Mater. 14: 4017–4020.
  88. 88 Zhang, H. and Cooper, A.I. (2007). Thermoresponsive “particle pumps”: activated release of organic nanoparticles from open‐cell macroporous polymers. Adv. Mater. 19: 2439–2444.
  89. 89 Gun'ko, V.M., Savina, I.N., and Mikhalovsky, S.V. (2013). Cryogels: morphological, structure and adsorption characterization. Adv. Colloid Interf. Sci. 187–188: 1–46.
  90. 90 Suwanchawalit, C., Patil, A.J., Kumar, R.K. et al. (2009). Fabrication of ice‐templated macroporous TiO2‐chitosan scaffolds for photocatalytic applications. J. Mater. Chem. 19: 8478–8483.
  91. 91 Eychmüller, A., Ziegler, C., Wolf, A. et al. (2017). Modern inorganic aerogels. Angew. Chem. Int. Ed. 56: 13200–13221.
  92. 92 Mahler, W. and Bechtold, M.F. (1980). Freeze‐formed silica fibers. Nature 285: 27–28.
  93. 93 Mukai, S.R., Nishihara, H., and Tamon, H. (2008). Morphology maps of ice‐templated silica gels derived from silica hydrogels and hydrosols. Microporous Mesoporous Mater. 116: 166–170.
  94. 94 Mori, H., Aotani, K., Sano, N., and Tamon, H. (2011). Synthesis of a hierarchically micro‐macroporous structured zeolite monolith by ice‐templating. J. Mater. Chem. 21: 5677–5681.
  95. 95 Nishihara, H., Mukai, S.R., Fujii, Y. et al. (2006). Preparation of monolithic SiO2–Al2O3 cryogels with inter‐connected macropores through ice templating. J. Mater. Chem. 16: 3231–3236.
  96. 96 Zhang, H., Hardy, G.C., Rosseinsky, M.J., and Cooper, A.I. (2003). Uniform emulsion‐templated silica beads with high pore volume and hierarchical porosity. Adv. Mater. 15: 78–81.
  97. 97 Minaberry, Y. and Jobbágy, M. (2011). Macroporous bioglass scaffolds prepared by coupling sol–gel with freeze drying. Chem. Mater. 23: 2327–2332.
  98. 98 Wang, X., Sumboja, A., Khoo, E. et al. (2012). Cryogel synthesis of hierarchically interconnected macro−/mesoporous Co3O4 with superb electrochemical energy storage. J. Phys. Chem. C 116: 4930–4935.
  99. 99 Yang, J. and Xu, J.J. (2003). Influence of synthesis conditions on the electrochemical properties of nanostructured amorphous manganese oxide cryogels. J. Power Sources 122: 181–187.
  100. 100 Sudant, G., Baudrin, E., Dunn, B., and Tarascon, J.‐M. (2004). Synthesis and electrochemical properties of vanadium oxide aerogels prepared by a freeze‐drying process. J. Electrochem. Soc. 151: A666–A671.
  101. 101 Mukai, S.R., Nishihara, H., Shichi, S., and Tamon, H. (2004). Preparation of porous TiO2 cryogel fibers through unidirectional freezing of hydrogel followed by freeze‐drying. Chem. Mater. 16: 4987–4991.
  102. 102 Osaki, T., Yamada, K., Watari, K. et al. (2012). Alumina cryogels with superior thermal stability for catalyst supports. J. Sol‐Gel Sci. Technol. 61: 268–274.
  103. 103 Osaki, T., Yamada, K., Watari, K. et al. (2012). Palladium‐alumina cryogel with high thermal stability and CO oxidation activity. Catal. Lett. 142: 95–99.
  104. 104 Osaki, T. Effect of ethylene glycol on structure, thermal stability, oxygen storage capacity, and catalytic CO and CH4 oxidation activities of binary CeO2–Al2O3 and ternary CeO2–ZrO2–Al2O3 cryogels. J. Sol‐Gel Sci. Technol. 82: 133–147.
  105. 105 Li, J.C. and Dunand, D.C. (2011). Mechanical properties of directionally freeze‐cast titanium foams. Acta Mater. 59: 146–158.
  106. 106 Jo, H., Kim, M.J., Choi, H. et al. (2016). Morphological study of directionally freeze‐cast nickel foams. Metall. Mater. Trans. E 3: 46–54.
  107. 107 Oh, S., Chang, S., and Suk, M. (2012). Microstructure of porous Cu fabricated by freeze‐drying process of CuO/camphene slurry. Trans. Nonferrous Metals Soc. China 22: s688–s691.
  108. 108 Ran, H., Feng, P., Liu, Z. et al. (2015). Complex‐shaped porous Cu bodies fabricated by freeze‐casting and vacuum sintering. Metals 5: 1821–1828.
  109. 109 Charnvanichborikarn, S., Worsley, M.A., Bagge‐Hansen, M. et al. (2014). Ice templating synthesis of low‐density porous Cu–C nanocomposites. J. Mater. Chem. A 2: 18600–18605.
  110. 110 Oh, S.‐T., Kim, Y.D., and Suk, M.‐J. (2015). Freeze drying for porous Mo with different sublimable vehicle compositions in the camphor‐naphthalene system. Mater. Lett. 139: 268–270.
  111. 111 Li, D. and Li, M. (2012). Porous Y2SiO5 ceramic with low thermal conductivity. J. Mater. Sci. Technol. 28: 799–802.
  112. 112 Wang, Z., Feng, P., Wang, X. et al. (2016). Fabrication and properties of freeze‐cast mullite foams derived from coal‐series kaolin. Ceram. Int. 42: 12414–12421.
  113. 113 Kokubo, T. and Takadama, H. (2006). How useful is SBF in predicting in vivo bone bioactivity? Biomaterials 27: 2907–2915.
  114. 114 Moon, J., Hwang, H., Awano, M., and Maeda, K. (2003). Preparation of NiO–YSZ tubular support with radially aligned pore channels. Mater. Lett. 57: 1428–1434.
  115. 115 Schneider, J., Matsuoka, M., Takeuchi, M. et al. (2014). Understanding TiO2 photocatalysis: mechanisms and materials. Chem. Rev. 114: 9919–9986.
  116. 116 Wegst, U.G.K., Bai, H., Saiz, E. et al. (2015). Bioinspired structural materials. Nat. Mater. 14: 23–36.