Chapter 6: Strong and Tough Ceramic Composites via Ice Templating – Ice Templating and Freeze-Drying for Porous Materials and Their Applications

Strong and Tough Ceramic Composites via Ice Templating

6.1 Introduction

6.1.1 Enhanced Applications by Ceramic Composites

The variety of ceramics available offers excellent properties such as mechanical strength, hardness, high temperature stability, optical and electric properties, thermal conductivity, insulation, and biocompatibility [1]. Ceramic composites can be formed between ceramics and metal ceramics and polymer, and between ceramics themselves. These composites provide enhancement in performance by combining the desirable properties of each component and more importantly the possible synergy effect between the components. Ceramics and their composites may be utilized as engineering blocks, in coatings or as porous materials depending on their targeted applications. For example, porous ceramics & composites are used for thermal insulation, filtration, and bioscaffolding [2] while non‐porous composites are used in the manufacture of load‐bearing objects and abrasion‐resistant coatings.

Depending on the applications, different types of ceramic materials may be selected. For medical applications including hip replacement, dental implants and restorations, and bone fillers, alumina was initially used and then gradually replaced by zirconia because of its high strength and high fracture toughness. Yttria‐stabilized zirconia (YSZ) is the standard basic material. However, the presence of water and oxygen can lead to aging problem by penetrating into the gaps due to the trivalent nature of yttria [3]. The development of alumina–zirconia composites and ceria‐ and magnesia‐doped zirconia has been improvised to address this aging issue [3]. In the case of scaffolds for bone engineering and biomedical applications, bioglass and calcium phosphate (CaP)‐based ceramics are mostly used because bioactive hydroxyapatite (HA) can be formed when in contact with the biological fluid [4]. Good mechanical properties are essential for bone tissue engineering. The scaffolds prepared by various approaches are therefore evaluated for their mechanical behaviour in terms of porosity, composition, microstructure, flaws, and reliability [5]. Ultrahigh molecular weight polyethylene (UHMWPE) has been applied as a bearing surface in human joint replacement and artificial bones due to its inert and low‐friction surface and good levels of mechanical properties. Inclusion of HA particles in the UHMWPE matrix leads to improved mechanical & tribological properties and bonding formation between the scaffold and the living bones [6].

Ceramics with high dielectric permittivity (εr) are promising candidates for microwave applications. However, the lack of conformity (or bendable property) in ceramic materials hinders their application for small platforms and system‐on‐packages. Thus, polymer–ceramic composites may be prepared to provide flexible substrates with good permittivity (εr = 20) and low loss, as demonstrated by the composites of barium titanate, Mg‐Ca‐Ti, and Bi‐Ba‐Nd‐Titanate with polydimethylsiloxane (PDMS) [7].

For composite materials, the properties may be simply the addition or combination of the properties from each of the components, or more desirably the significantly improved properties made possible by the synergy effect between the components. It is also possible to generate new properties, which the individual components may not exhibit. For example, multiferroic magnetoelectric materials, showing simultaneous ferroelectricity and ferromagnetism, have attracted scientific interest. However, useful natural materials with multiferroic magnetoelectric properties are rare, either due to weak response or because they occur at low temperatures. The composites formed by combining piezoelectric and magnetic compounds can produce a large magnetoelectric response at room temperature. This is the result of the coupling interaction between piezoelectric and magnetic compounds because neither the piezoelectric phase nor the magnetic phase has the magnetoelectric effect [8]. When a magnetic field is applied to this composite, the magnetic phase changes its shape magnetostrictively. The strain generated leads to an electric polarization in the piezoelectric phase [8].

The high temperature stability and corrosion resistance are probably the main reasons for using ceramics in thermal insulation [2], as thermal exchangers [9] and in thermal barrier coatings (TBCs) [10]. For thermal exchangers, silicon carbide (SiC) is probably the most used ceramic due to its high decomposition temperature (∼2500 °C), good thermal shock resistance, good flexural strength at high temperature, and high thermal conductivity (about four times higher than that of steel) [9]. Silicon nitride (Si3N4, excellent strength and creep resistance but vulnerable to oxidation at high temperatures), alumina (highly resistant under oxidizing and reducing conditions, inexpensive, but comparably low thermal shock resistance), aluminium titanate (low thermal conductivity), and aluminium nitride (good oxidation resistance and thermal stability) have also been used. However, the intrinsic brittleness and lack of reliability limit the applications of these ceramics. Ceramic composites are thus fabricated to address these shortcomings [9]. Low‐thermal conductive ceramic coatings are used in the highly demanding high‐temperature environment of aircraft and industrial gas‐turbine engineering. The TBCs provide considerable temperature reduction from the surface temperature to around 100–300 °C. This allows the operation of gas‐turbine engines at gas temperature well above the melting point of the superalloy (∼1300 °C) [10]. The TBCs are formed from four layers: the substrate, the bond‐coat, the thermally grown oxide, and the ceramic top‐coat. The thermally grown oxide layer is α‐Al2O3 with low oxygen ionic diffusivity, limiting the bond‐coat oxidation. The top‐coat ceramic layer is typically made of YSZ by air‐plasma‐spray deposition and electron‐beam physical‐vapour deposition. YSZ has one of the lowest thermal conductivities at high temperatures (∼2.3 W m−1 K−1 at 1000 °C), a high thermal‐expansion coefficient (∼11 × 10−6 °C−1) to alleviate the stress between the layers, relatively low density (6.4 Mg m−3), a hardness of ∼14 GPa, and a high melting point (∼2700 °C) [10].

For the moveable mechanical parts in satellites and space‐borne systems, the reliability of these parts in terms of lubricant degradation and excessive wear is essential because these systems would remain in space for 10–30 years, exposed to harsh environment and temperature cycling from cryogenic to 400 °C [11]. The tough composite coatings for space tribology may be formed from oxides and dichalcogenides (e.g. PbO/MoS2, ZnO/MoS2, ZnO/WS2), the combination of carbides (TiC, WC)/oxides/dichalcogenides, and also the inclusion of diamond‐like carbon particles in the matrix [11]. In responding to the wide range of temperature variation, the coatings can be improved with the encapsulation of YSZ, nanosized MoS2, and nanodiamonds in a gold matrix [11].

6.1.2 Processing and Mechanical Behaviour of Ceramic Composites

A composite usually consists of a matrix phase and a second/dispersed phase. The second phase can be more than one type of materials. Improving the mechanical properties is essential for nearly all the composites and is paramount for some applications. The matrix can be either polymer or ceramics and the same can be said for the second phase materials as well. The second phase can be a discontinuous phase (e.g. particles) or a continuous phase (e.g. fibres, layered composites). The shape and size of the particles, e.g. spherical particles, irregularly shaped particles, rods, fibres, whiskers, and plates, can have a profound impact on the mechanical behaviour of the resulting composites.

For polymer–ceramics composites where polymer is the matrix, the addition of ceramic particles is often to improve the strength and wear property of the materials [5, 6]. It has been shown that particle loading has a significant impact on Young's modulus while there is little effect from particle/matrix interfacial adhesion [12]. However, for the effect of particle size, there seems to be a critical size (around 30 nm). When the particles are smaller than the critical size, smaller particles in the composites can improve Young's modules more significantly while the effect is not clearly observed when the particles are greater than 30 nm [12].

In this chapter, the focus is on the ceramic composites where the ceramics are the matrix. The second phase can be polymer, metal, or other types of ceramics. Because ceramics usually have high strength but are brittle and the strength or the targeted applications are always compromised by the low toughness, the purpose (or at least one of the main purposes) of making ceramic composites is to improve the fracture toughness. In simple terms, the toughening mechanisms are due to crack‐bridging, crack deflection, pull‐out of the dispersed phase, microcracking, and phase transformation toughening [13]. The interface area of the dispersed phase can exert the impact‐based pull‐out friction, debonding, and crack deflection. Therefore, compared to microparticles, the incorporation of nanoparticles in a composite may contribute more significantly to the enhanced mechanical behaviour [14].

The fracture of brittle materials such as ceramics involves little stress dissipation by processes other than crack extension, leading to failure of the materials. In ceramic glasses and single crystalline ceramics, the fracture toughness can be very low (0.5–2 MPa m½) when compared with metals (15–150 MPa m½). Despite some improvement, the fracture toughness of polycrystalline ceramics is still very low at ≤5 MPa m½. In contrast, high flexure strengths can be achieved for most of the ceramics [13]. A substantial fraction of the strength may be lost for low‐toughness ceramics due to the damage introduced during service. It is therefore highly important to develop toughened ceramics, which is often achieved by fabrication of ceramic composites. Besides the effect of the second phase particles, the shape and size of the matrix ceramic particles can also impact the fracture toughness as it has been shown for elongated alumina particles with SiC whiskers as reinforcement phase [13]. The toughness increases with increasing particle size up to a critical size. In a composite containing tetragonal zirconia, the improved toughness can be attributed to the enhanced matrix toughness and the transformation‐induced toughness [13].

Because elastic bridging by ligaments and pull‐out contribute significantly to the fracture toughness, the dispersed particles with high aspect ratio are usually better for the composite toughness. These include grains, rods, whiskers, platelets, or continuous fibres [1316]. Particularly for the fibres, a big part of the toughness is attributed to the pull‐out of the fibres from the matrix, which depends on frictional sliding resistance, the fibre radius, the fibre volume fraction, and the fibre strength statistics [15]. The statistic fibre strength can be taken as the cumulative number of defects, which fails at stress σ and fibre length L by the Weibull expression [15]:


where σ0 is the stress required to cause one failure, on average, in a fibre of length L0 and m is the variability of measured strengths about the average values.

However, fabrication of long fibre‐reinforced ceramics is expensive and can be usually realized with limited thickness. Also, the most often used carbon and silicon carbide fibres show limited thermal stability in oxidizing environments. This problem may be addressed by the fabrication of laminar ceramic composites, where alternative layers with thicknesses varying from microns to millimetres, maybe also nanometres although it is highly challenging [1618]. Various methods are used conventionally for the fabrication of laminar composites. These include: (i) tape casting: sequentially depositing a suspension on a support either by spreading under a blade (doctor blade process) or by coating followed by solvent evaporation. The layer thickness is in the range of 50–100 µm [16, 17]; (ii) centrifugal casting: achieved by sequential centrifugal casting of slurries with layer thickness down to ∼10 µm [17]; (iii) slip casting: where a suspension with low viscosity is poured into a porous mould. The extraction of the solvent into the pores via wetting and capillary force leads to the build‐up of the solid particle layer. The layer sequence can be achieved by alternating the composition of the suspensions [17]; (iv) electrophoretic deposition: this method is driven by the movement of charged ceramic particles under a direct current (DC) electric field [17]; (v) vapour deposition: essential for the fabrication of nanoscale layers, by magnetron sputtering [18] or plasma spray [10].

Although great efforts have been made to produce strong materials, strength alone is not enough for many applications. Other mechanical properties such as fracture toughness, creeping behaviour, and fatigue tolerance, are also essential. In addition to intrinsic toughness, fracture toughness may be highly important. This is because the ceramic materials cannot be prepared perfectly and some flaws or cracks exist during the fabrication process, which can cause catastrophic failure even under relatively low stress. For bone tissue engineering, depending on different parts of the bone system, the scaffolds with different mechanical properties including strength (failure under a single overload), fracture toughness (failure from a modest load with pre‐existing crack or flaw), and fatigue strength (failure by high number of cycles of small load), should be selected according to the bone types (Figure 6.1) [19].

Figure 6.1 Multiple pathways to mechanical failure of different types of bones. (a) Bone fracture may occur by a single large load, a single modest load with pre‐existing damage in the bone, or accumulated damages by many cycles of loading. (b) Representation showing the likely mode of mechanical failure from different types of bones.

Source: Hernandez and Meulen 2017 [19]. Reprinted with permission from John Wiley and Sons.

However, the strength and toughness in a material are mutually exclusive [20]. When the toughness is increased by fabricating composites, the strength is often decreased. This is because toughness is achieved by adding grains, platelets, or fibres but the resulting interfacial flaws or debonding can reduce the strength [13]. While strength is a stress to resist non‐recoverable deformation under load, toughening is enhanced by crack deflection/bridging from elastic filament and deformation‐induced energy dissipation [20]. It seems there is a trade‐off between strength and toughness for ceramic composites depending on targeted applications. Some strategies have been developed to make strong and tough composites particularly when learning from natural materials [20].

Nature has produced some extraordinary materials that exhibit exceptionally high strength, toughness, and low density. Some of the well‐known examples include bone, bamboo, teeth, and nacre [21, 22]. The common features of these materials are their hierarchical structures from atomic, nanoscale, to macroscale with relevant interfaces. Some of the materials (e.g. nacre, teeth) consist of a very high inorganic solid content and a low content of organic component (proteins or polymers), with the two phases interfacing at the nanoscale. In an effort to make tough composites, strong research activities have been observed by mimicking the structure of nacre [23]. The fabrication methods include the conventional tape casting, slip casting, layer‐by‐layer (LBL) deposition, electrophoretic deposition, mechanical assembly and freeze casting [23].

In this chapter, the emphasis is on the use of the freeze‐casting (ice‐templating) method to fabricate ceramic composites, particularly biomimetic tough composites [24, 25]. Since mechanical properties are usually the mainstay of fabricating ceramics and ceramic composites, we start with the introduction to the commonly used mechanical parameters when discussing ceramic composites. This is followed by porous ceramic composites, different types of tough composites, and finally the summary.

6.2 Mechanical Characterizations of Ceramic Composites

6.2.1 Strength

Strength is the resistance of the material to the onset of permanent deformation [21]. This can be conveniently measured by uniaxial tension (or compression), the shear force, and the torsion force applied. For a block of material, the tensile stress (σ) is defined as the force (F) divided by the cross‐sectional area (A):


Under the tensile stress, the block is deformed or elongated. The deformation can be defined by the strain (ε):


where L is the original length of the block and δL is the extension of the block under the tensile stress [26]. Figure 6.2 shows a representative stress–strain curve.

Figure 6.2 The diagram showing a typical tensile stress–strain profile with some important stress parameters indicated. σE, the elastic limit stress; σY, the tensile yield stress; σB, the ultimate stress; and σF, the fracture stress. The solid straight line indicates the linear part of the stress–strain curve that can be used to calculate the Young's modulus (E). The dotted line indicates the deformation at 0.2%.

The definitions mentioned are for engineering stress or engineering strain, where the original area and length are conveniently used. However, under a constant tensile stress, with the elongation of the block, the cross‐sectional area decreases while the length increases continuously. Therefore, in contrast to engineering strain/engineering stress, true stress and true strain are defined and used where the instantaneous cross‐sectional area and length are used for the calculation. As shown in Figure 6.3, the true stress–strain curve is always higher than the nominal stress–strain curve.

Figure 6.3 The diagram shows the difference between the true stress–strain curve and the nominal stress–strain curves by compression or tension.

As observed from Figure 6.2, the initial part of the stress–strain curve is linear, i.e. the strain is proportional to the stress applied. This normally falls into the elastic deformation region. The elastic deformation means the instantaneous full recovery of the strain when the stress is removed. It is possible that the stress–strain curve may not be linear [26]. The Young's modulus E (elastic modulus) is defined as the slope of the initial linear stress–strain curve. The Young's modulus is often used to describe the material's stiffness, meaning the force required to generate elastic deformation [21, 27]. Another parameter under tensile stress is the Poisson's ratio ν, which is defined as the ratio of the strain parallel to the tensile axis to the strain normal to the tensile axis [27].

The compression stress can be applied like the reverse of the tensile stress. But it is more difficult to obtain reliable stress–strain curves. Under compression, the top and bottom of the specimen should be always parallel and bulging should be avoided. Similar to that for tensile stress, there are also definitions of true compression stress and true strain. The nominal compression can be defined as the compression force divided by the area. Since the area is always increased under compression, the true stress should be lower than the nominal compression stress, as illustrated in Figure 6.3.

When a traction force (T) is applied to the top and bottom surface of a rectangular body specimen, the shear stress τ can be defined as the force (T) divided by the surface area while the strain γ is the ratio of the deformation in the force direction divided by the distance between the top and bottom surface. The shear modulus μ can be defined as:


Along the stress–strain curve in Figure 6.2, σE represents the stress at the elastic limit. Beyond σE, the curve moves into a permanent deformation (plasticity) region where the strain is not recovered when the stress is removed. The tensile yield stress σY is defined as the stress that generates a permanent strain of 0.2% after the stress is removed. The strain increases continuously with the increasing tensile stress until the maximum stress σB (known as ultimate stress). Up to σB, the strain or the elongation under tensile stress is uniform. Beyond σB, the large local elongation occurs, leading to the decrease of stress. This results in eventually the fracture of the specimen, with the relevant fracture stress σF (Figure 6.2). The maximum strain before the fracture, the percentage of elongation or relative change in the cross‐sectional area, is often described as the ductility of the material [21].

Another type of permanent deformation is creep. Figure 6.4 shows the three‐stage creep profile. The strain (permanent deformation) increases rather rapidly under the applied stress at the start (transient creep or primary creep), then slows down in rate and moves into a steady creep region for a long time. The third stage (also known as tertiary creep) is the accelerated rate of deformation before final failure. Usually the two‐stage creep profile occurs in polycrystalline ceramics at low stresses and high temperatures while the tertiary stage can be observed for metals [28].

Figure 6.4 The diagram illustrates the creep behaviour under constant load and temperature.

The mechanical tests are usually carried out following the suggested procedures or standards as proposed by the American Society for Testing and Materials (ASTM) or by other internationally tested methods. The recommendations of ASTM E8‐01 are usually used for tension testing of metallic materials. However, due to the nature of brittle fracture, flexure (bending) tests are usually used to evaluate ceramics and ceramic composites. Both three‐point bending and four‐point bending tests are frequently employed (Figure 6.5). In the three‐point bending, the bending stress is highest at the middle and reduces to zero at the support point. Only a very small part of the specimen is exposed to the high stress. For the four‐point bending test, the central part of the specimen is subjected to uniform bending stress, which is more likely to reflect the real situation. Four‐point flexure testing is more widely used for strength measurement of ceramics when creep is not significant. Typically, quarter‐point (D = L/4) or third‐point (D = L/3) loading is used. Recommended three‐point and four‐point flexure tests can be found in ASTM C 1161‐02C [29].

Figure 6.5 The schematic representation of flexural tests. (a) Three‐point bending test. (b) Four‐point bending test.

6.2.2 Hardness

Hardness is a measure of the material's resistance to permanent deformation when a load is applied on the surface. It can be measured by the extent of penetration by indentation test or indirect methods [29]. The commonly used indentation methods include the Vickers test (the indenter is a square‐based pyramid usually made of diamond), the Krooper test (the indenter with unequal diagonals), and the Brinell test (the spherical indenter) [27, 29].

For the most widely used Vickers test, the hardness (H) can be calculated using the contact area over the four faces of the indenter:

where P is the applied load and 2a is the diagonal length of the indent [29].

In the ceramics industry, the projected area of the indent is often used to calculate the hardness. This gives a small but a significant difference of 7.9%, by comparing the values obtained from Equations (6.5) and (6.6) [29].

Empirically, the Vickers hardness (H) can be linked with the tensile yield stress (σY) in the following equation:


where the coefficient c is in the range of 2.9–3.0 for metals [27, 30]. The value of the coefficient c varies for different types of materials. Thus, one may use the hardness data to estimate the strength of the materials, or vice versa.

Under mild wear conditions, the hardness is directly related to wear. The relationship can be described as:


where wr is the wear rate in volume removed from the surface per unit sliding distance, k is the wear coefficient, P is the force applied against the opposing surface, and H is the hardness. The wear resistance is defined as 1/wr [30].

6.2.3 Fracture Mechanics and Fracture Toughness

In materials, particularly when fabricating engineering materials, the presence of flaws or cracks in these materials is almost unavoidable. The mechanical data obtained from measuring a smooth or ‘perfect’ specimen cannot be rationally applied for these engineering materials. There is also a huge reliability issue among different batches of the materials. In brief, fracture mechanics investigates how the stress and strain develop around a pre‐existing crack in a material and how to improve the fracture resistance. It can be classified into two general categories: linear elastic fracture mechanics (LEFM) and elastic–plastic fracture mechanics (EPFM). LEFM is applied to small cracks in the elastic materials with a small plastic zone while the non‐linear EPFM is usually used for large‐scale conditions within the plastic zone [21, 27].

Fracture mechanics can be characterized by a stress intensity factor K, which is a function of stress, crack size, and boundary conditions. K can be specified based on the modes of crack deformation: KI for mode I deformation (crack opening), KII for mode II deformation (in‐place shear of the crack), and KIII for mode III deformation (out‐of‐plane shear). Because mode I crack deformation is often seen in practical applications, K usually means KI if not otherwise stated. K is generally expressed as:


where F is a dimensionless parameter related to geometry, σ is the applied stress, and a is the crack length. This equation is valid when for LEFM [27].

There are two parameters used for the EPFM materials: (i) J integral, a path‐independent contour integral for a cracked body, equivalent to the energy‐release rate; (ii) crack‐tip opening displacement (CTOD), the plastic deformation after blunting an initial sharp crack.

When the intensity factor K reaches a critical value (Kc), the crack becomes unstable and brittle fracture occurs. Kc is regarded as the fracture toughness for mode I deformation under small‐scale yielding and the plane‐stress state. Similarly, for the EPFM materials, JIC or Jc is the toughness for onset of slow crack growth from an initial crack [27].

Kc is usually measured using a single‐edge notched (SEM) specimen by flexural tests or a compact test. There are several standards to be followed but ASTM E399‐90 is widely used. The specimen may contain other types of notch, e.g. a chevron notch.

The procedures and guidelines for the determination of JIC are available in ASTM E1820. The J–R (J integral – crack Resistance) curve is a plot of J integral values versus the corresponding crack extension values, demonstrating the material's resistance to crack extension.

6.2.4 Toughening Mechanism in Ceramics and Ceramic Composites

In ceramics with high toughness, there is usually more than one type of toughening mechanisms. Conventionally, the mechanisms include [31]: (i) crack deflection that takes place into the local area (quite often the grain boundaries) with lower resistance to crack propagation. The toughening increases with increasing volume fraction of deflecting particles (until a certain volume fraction, ∼20%). Crack deflection may cause partial bridging when a grain/whisker is in the deflection path; (ii) Crack bowing that is the development of a non‐linear crack front for a planar crack when a second‐phase material is introduced into a brittle material. This is different from crack deflection which occurs as out‐of‐plane development of the crack; (iii) Crack tip shielding that can be divided into two basic categories: process zone effects and bridging zone effects. The process zone effects occur around the crack tip. The local compressive or elastic force resulting from the development of crack tip acts to close the crack. The crack bridging is the result of bridging zone effects, occurring at a distance from the crack tip. This effect is usually attributed to the debonding or ligament of the second‐phase in the ceramics; (iv) Contact shielding processes that includes the resistance to pull‐out of fibres or grains, sliding interface friction, etc.; (v) Stress‐induced zone‐shielding processes that is found in tetragonal zirconia transformation, microcracking, compressive stress in the outer layer, and residual stress.

As schematically shown in Figure 6.6 [20], toughening mechanisms may be divided into extrinsic toughening and intrinsic toughening. Intrinsic toughening is inherently related to the plasticity zone of the materials ahead of the crack tip and hinders the initiation and propagation of the cracks. The extrinsic toughening is associated with crack tip shielding via crack bridging and wedging behind or at the crack tip. This type of mechanism works via limiting crack growth and usually shows a rising JR curve [20, 21].

Figure 6.6 The schematic diagram shows the various toughening mechanisms in terms of intrinsic (plasticity) toughening versus extrinsic (shielding) toughening.

Source: Ritchie 2011 [20]. Reprinted with permission from Nature Publishing Group.

Fabrication of ceramic composites by introducing the second phase aims to improve the toughness, mainly via the extrinsic toughening mechanism. As mentioned earlier, discontinuous reinforcing phases including grains, whiskers, and platelets are used to produce toughened ceramics and some equations have been formulated to predict/describe the toughening behaviour [13]. Continuous fibres are found to be more effective in enhancing the toughness of the composites [15, 25]. This has been attributed to crack deflection at the fibre/matrix (preferably weak) interface, better performance in crack bridging by the fibres, and the higher work required for fibre pull‐out after debonding [32]. Laminar composites have been the focus of recent studies in order to obtain strong and tough composites, stimulated by mimicking the exceptional natural materials such as nacre and teeth [17, 2023].

6.3 Porous Ceramic/Polymer Composites

The use of freeze casting for the preparation of porous ceramics including the introduction to different freezing techniques and freeze‐casting parameters is described in Chapter 5. Based on that, it is straightforward to fabricate porous ceramic/polymer composites. This chapter focuses on the composites where the ceramics are the matrix. The composites with dispersed ceramic particles in a polymer matrix are not covered here.

To make porous ceramic/polymer composites, the ceramic particles can be suspended in a polymer solution (the concentration is usually higher than that of polymer only used as binder or stabilizer) and the resulting suspension is then subjected to a freezing process. The frozen samples are freeze‐dried to form porous ceramic composites. For example, aqueous silica colloidal suspension (Ludox HS‐30) is mixed with 5 wt% aqueous poly(vinyl alcohol) (PVA). The resulting suspension is unidirectionally frozen and freeze‐dried, producing an aligned porous PVA/silica composite (Figure 6.7a). PVA helps stabilize the silica colloids in the suspension and then acts like a glue to stick the silica colloids together (Figure 6.7b) [33]. When a surfactant such as sodium dodecyl sulfate (SDS) is added into the PVA/silica suspension, an oil‐in‐water emulsion can be formed by emulsifying an organic solvent such as cyclohexane into the aqueous phase. Freezing the emulsion and then freeze‐drying generates an emulsion‐templated and ice‐templated porous composite material. By varying the volume percentages of droplet phase in the emulsion, it is possible to produce porous structures with the systematically tuned pore morphology and porosity [34]. Figure 6.7c shows the porous structure of PVA/silica prepared from the emulsion with a 75 v/v% internal phase. The pore size distribution measured by Hg intrusion porosimetry exhibits the macropores templated from ice and emulsion droplets and mesopores resulted from the assembly of silica colloids (Figure 6.7d). Instead of directional freezing, the emulsion can be sprayed into liquid nitrogen and then freeze‐dried. This produces highly porous silica/polymer microspheres [35] (Figure 6.7e). Other types of ceramic particles can be readily used as well. For example, ceria nanoparticles in PVA suspensions can be processed to generate composite micron‐sized fibres [33]. Aligned porous columns in capillaries can be prepared by directional freezing of silica colloids (TM‐50) with PVA or poly(diallyldimethylammonium chloride) (PDDA) and show fast separation with relatively low back pressure by high performance liquid chromatography (HPLC) [37]. Mesoporous silica microspheres (∼1 µm in diameter) prepared by a modified Stöber method [38] are employed with PVA to form aligned porous composites [39]. The addition of an anionic surfactant SDS leads to a well‐aligned pore structure while the presence of cationic surfactant cetyltrimethylammonium bromide (CTAB) results in a layered porous composite structure [39].

Figure 6.7 The structures of porous silica–polymer composites. (a) Aligned porous silica (HS‐30 colloids)‐PVA and (b) the magnified area shows the packed silica colloids. (c) The emulsion‐templated and ice‐templated porous silica (HS‐30)‐PVA and (d) the mesopore size distribution measured by N2 sorption. (e) Porous silica‐PVA microspheres prepared emulsion‐spray‐freezing. (f) Porous chitosan‐silica.

Source: Ahmed et al. 2012 [36]. Reprinted with permission from Royal Society of Chemistry.

Source: Reprinted with permission from Ref. [35].

Source: Reprinted with permission from Ref. [34].

Source: Reprinted with permission from Ref. [33].

Freeze‐dried porous polymers can be used as scaffolds for drug delivery. However, the highly interconnected macroporous structures by ice templating always contribute to a significant initial burst release [40]. This issue can be addressed by the use of polymer/mesoporous silica composites as the delivery vehicle. For this purpose, a model drug curcumin is uploaded into mesoporous silica spheres that are then suspended in aqueous chitosan solution also containing curcumin. Freeze‐drying of this suspension generates a porous composite material (Figure 6.7f). A dual‐controlled release profile is achieved with this scaffold – an initial burst release from chitosan (so that a therapeutic drug concentration can be achieved quickly) followed by a steady release from the uploaded mesoporous silica microspheres [36].

Following the same principles as described above, it is possible to produce porous ceramic/polymer composites with different types of ceramic particles. Two type of porous ceramic/polymer composites are described below in detail.

6.3.1 Hydroxyapatite (HA)‐based Composites

These composites are prepared mainly for bone tissue engineering. The biocompatibility and mechanical strength are the most important parameters for porous HA‐polymer composites when used as scaffolds. The mechanical properties and the approximate porosity of the human bones are given in Table 6.1 [5, 41]. In natural bone, ∼70% of the weight is the mineral content of CaP, in the form of carbonated apatite, with about 22% being organic matrix (90% of which is type I collagen) [42]. There are a range of CaP phases. Among them, hydroxyapatite (HA, Ca5(PO4)3OH, Ca:P = 1.67 in mass) and β‐tricalcium phosphate (β‐TCP, Ca3(PO4)2, Ca:P = 1.5 in mass) are mostly used. This is attributed to their osteogenic property and the ability to form bonds with living bones [5, 42]. In addition, bioglasses are also widely used for bone tissue engineering. However, for both CaP and bioglasses, their mechanical strength and particularly low fracture toughness are serious drawbacks [19]. Fabrication of composite scaffolds provides an effective route to improving the mechanical strength, as illustrated in Figure 6.8, based on elastic modulus and compressive strength [41]. By tuning the preparation methods and controlling the compositions of the composite scaffolds, good control of drug delivery from the scaffolds (e.g. proteins, growth factors, antibiotic drugs, etc.) for enhanced bone tissue engineering may be also achieved [41, 42].

Table 6.1 Mechanical properties and porosity of human bone.

Source: Data adapted from Refs [5, 41].

Compressive strength (MPa)Tensile strength (MPa)Elastic modulus (GPa)Flexural strength (MPa)Fracture toughness (MPa m½)Porosity (%)
Cortical bone130–18050–15112–18135–1936–85–13
Cancellous bone4–121–50.1–0.5N/AN/A30–90

Figure 6.8 Graph of elastic modulus‐composite strength for porous composites in the context of porous polymers, dense polymers, dense ceramics, and bones.

Source: Rezwan et al. 2006 [41]. Reprinted with permission from Elsevier.

When preparing porous HA‐based composites, biodegradable aliphatic polyesters are often used. These include poly(lactic acid) (PLA), poly(glycolic acid) (PGA), poly(ε‐caprolactone) (PCL), and copolymer poly(lactic‐co‐glycolic acid) (PLGA). PLA is available in three forms: L‐PLA (PLLA), D‐PLA (PDLA, and racemic mixture of D,L‐PLA (PDLLA). The degradation occurs via hydrolysis of ester bond, generally following the order of:


The factors that affect the degradation rates include molecular weight, chemical composition, polydispersity, porosity, crystallinity, and device configuration [41].

For example, nanocrystalline carbonated apatites were prepared by a deposition method. The PLGA–apatite composites were fabricated by freeze casting using dimethyl carbonate as solvent [43]. The weight percentage of apatites in the PLGA/apatite composites varied from 0% to 60%. With the increasing percentage of PLGA, the porosity decreased from 72.1% to 63.9% whilst the Young's modulus increased from 0.16 to 0.34 MPa and then slightly decreased. This was attributed to the aggregation of apatite particles at higher concentration in the original slurries. A higher percentage of apatite in the composites was also found to reduce the in vitro degradation of PLGA [43].

In another study, aqueous CaP slurries with sodium carboxymethyl cellulose (SCMC) as stabilizer were processed by directional freezing and freeze‐drying to produce porous CaP scaffolds. The PLGA–CaP scaffolds were obtained by soaking the CaP scaffolds in PLGA‐CH2Cl2 solution followed by vacuum drying [44]. In order to improve cell attachment and proliferation, the PLGA–CaP scaffolds were treated by NH3 plasma and then exposed to collagen solution and subsequent crosslinking with glutaraldehyde. The collagen‐PLGA–CaP scaffolds showed improved water uptake, porosity, and markedly enhanced cell seeding and growth, while the compressive strength remained unchanged [44].

Other biocompatible natural polymers such as chitosan can also be used to produce porous scaffolds by ice templating [40]. As compared with polyesters such as PLGA, aqueous slurries can be processed via the freeze‐casting approach, avoiding the use of an organic solvent. For example, aligned porous chitosan/HA scaffolds with high porosity (∼85%), large pore sizes (200–500 µm), and interconnected porosity were prepared [45]. MG63 osteoblast‐like cells were found to spread and cluster along the walls and grow well into the tubular pores [45].

Most of the ceramic particles used are either irregularly shaped or spherical. Nacre has high strength and toughness and consists of mainly aragonite platelets and a small fraction of polymer (proteins) [2123]. It is thus impossible to produce strong porous composites with ceramic platelets. For example (although not HA platelets), alumina platelets (diameter 5–10 µm and thickness 300–500 nm) and chitosan and gelatin were used to produce porous scaffolds with nacre‐like walls [46]. This has been compared with large and small spheres with diameters similar to the platelet dimensions, as shown in Figure 6.9. The alumina platelets were assembled into nacre‐like structures by the freezing process, with the polymers holding the platelets together. This nacre‐like porous scaffold showed higher stiffness, strength and toughness by a factor of 1.5–4 than the scaffolds with the same porosity but no nacre‐like structure [46].

Figure 6.9 Porous alumina–polymer composites prepared by freeze casting of alumina platelets and alumina particles. (a) Diagram showing the platelets, the size of alumina particles, and the bimodal particles. (b) Diagram showing how the particles and platelets assemble during the freezing process. (c–f) Show the structures of the composites formed, comparable to the diagram in (b).

Source: Hunger et al. 2013 [46]. Reprinted with permission from Elsevier.

6.3.2 Clay‐based Composites

Clays are usually referred to as natural aluminosilicate colloids, which may also include other minerals and metal oxides [47]. There are two terms usually used: ‘clay’ and ‘clay mineral’. Clays are regarded as natural, with grain sizes smaller than 2 or 4 µm, containing phyllosilicates as their main constituents. Clay minerals can be natural or synthetic, not size limiting and can include both phyllosilicates and non‐phyllosilicates [48]. However, clay minerals are often referred to as clays in literature [48]. Crystalline clays consist of layers of silica tetrahedrons and alumina octahedrons. Accordingly, they can be classified into 1 : 1 layer (one silica layer and one alumina layer), 2 : 1 layer (one alumina layer between 2 silica layers or trioctahedrons), regular mixed layers, and chain‐structure types [47]. There are many types of clay particles. The commonly used clays are given in Table 6.2. The surface chemistry (e.g. charges, polarity, edge surface, and planar surface) of the clay particles when suspended in water or incorporated into a matrix can have significant impact on their applications [49]. Clays have been used for adsorption of dyes [50], heavy metal ions [47], proteins and nucleic acids [51], and drug release [52]. Their potential toxicological impacts have been evaluated [53].

Table 6.2 Properties of commonly used clay particles.

Source: Data adapted from Refs [47, 49].

Crystalline typeThickness of layers unitElectronic chargeaComments
Montmorillonite (Smectite)2 : 1 layer1–2 nm −0.522 to −0.741 (O with no cation)
−0.800 t0 −0.867 (O with K+)
Different cation counterions, weak force between two silica layers permitting water and exchange ions (Na+, )
Kaolinite1 : 1 layer0.72 nm 0.152 (inner H)
0.219 (surface H)
Strong bonding, no layer swelling
BentoniteConsisting of montmorillonite and other crystalline structures
Illite2 : 1 type1.0 nmBonded by K+ ions
Vermiculites2 : 1 typeBonded by Mg2+
Chlorites2 : 1 layer, trioctahedrons1.4 nmMg2+, Al3+, Fe3+, Fe2+ on octahedron sites

aCharges on surface atoms are obtained by the semi‐empirical electronegativity equalisation method (EEM).

Polymer–clay composites are extensively investigated with a view to enhance or facilitate the above‐mentioned applications [50]. In addition to offering new functionalities, the main outcomes of fabricating polymer–clay composites are to enhance their mechanical strength and viscoelastic property and adjust porosity and density [54, 55]. There are various methods that can be used to prepare polymer–clay composites but the ice‐templating method offers a simple and versatile route, particularly with the advantage of producing anisotropic porous structures [24, 55, 56].

Various polymers and clays have been used to fabricate porous composites via the ice‐templating approach. Pectins are complex carbohydrates that can be obtained from plants and fruits. They have been used as coating or packaging for food and pharmaceuticals. Pectins can easily form gels and can be crosslinked by divalent metal ions. To prepare porous pectin–clay composites, pectin and sodium montmorillonite (Na‐MMT) were dissolved/suspended in water and the resulting suspension was then frozen and freeze‐dried. Alternatively, CaCl2 solution could be added to form a pectin/clay gel before freezing [57]. The solid content of Na‐MMT could be adjusted to produce the composites with the moduli in the range of 0.04–114 MPa and the densities of 0.03–0.19 g cm−3 [57]. Casein is a naturally available protein, constituting ∼80% of cow's milk protein. A similar procedure was used to prepare casein‐Na‐MMT composites. The casein‐MMT suspension could be crosslinked using glyceraldehyde before the freeze‐casting procedure. The crosslinking procedure improved the aerogel structure remarkably but with minimal impact on the density [58]. Natural rubber (NR)–MMT composites were prepared by freezing and freeze‐drying of aqueous NR/Na‐MMT suspensions. To prepare the crosslinked aerogel, the freeze‐dried samples were immersed in S2Cl2‐benzene solution and crosslinking by S2Cl2 at −18 or 18 °C for 24 h. The crosslinking procedure resulted in a 26‐fold increase of the compressive modulus to 1.8 MPa [59].

In addition to enhanced mechanical strength, the polymer–clay composites have also exhibited lower thermal conductivity and improved performance in flame retardation. Poly(vinyl alcohol) (PVA, Mn = 31–50 K) composites were prepared with different types of inorganic fillers including silica (12 nm), Na‐MMT, laponite, and halloysite with aqueous suspensions containing 5 wt% PVA and 5 wt% inorganic nanofiller [60]. The suspensions in polystyrene vials were frozen in a solid CO2/ethanol bath (−80 °C) and freeze‐dried with an initial shelf temperature of 25 °C and a condenser temperature of −80 °C. Compared to the PVA aerogel, the addition of laponite resulted in ∼20% decrease in density and an increase of 10% in modulus but this trend was not obvious for other fillers particularly halloysite and silica. The addition of inorganic fillers did not increase the limiting oxygen index values (∼24–25) significantly but resulted in significant decrease in heat release (measured by cone calorimetry), smoke release, and CO production. Figure 6.10 illustrates the heat release rates (HRRs) of the composites with time. It can be seen that the heat release occurs in a very short period for the expanded polystyrene (EPS) packing foam and also significant heat release from PVA only. All four composites exhibited considerably reduced heat release, with the PVA/laponite performing the best [60]. In another effort, the PVA/Na‐MMT suspensions were crosslinked by irradiation using a 60Co source with different absorbed doses before freezing in liquid nitrogen [61]. Higher dosage of irradiation led to higher crosslinking and higher modulus and then the impact became slightly negative. The content of MMT in the composites affected the modulus and density but these were also affected by the absorbed irradiation dosage. However, the higher percentage of MMT in the composites (e.g. 1% PVA–9% MMT) certainly contributed significantly to reduced heat releasing rates [61].

Figure 6.10 The effects of PVOH (polyvinyl alcohol)/clay and PVOH/SiO2 on heat release rates, compared to PVOH only and the commercial expanded polystyrene (EPS).

Source: Chen et al. 2014 [60]. Reprinted with permission from American Chemical Society.

Furfuryl alcohol (FA) may be produced or converted from biomass. It is soluble in water and compatible with the clays with hydroxylated surfaces such as MMT. FA can be readily polymerized in water or in pure monomer form in the manner of head to head or head to tail, catalysed by acid. Homogeneous aqueous suspension of FA and Na‐MMT were polymerized by adding a small amount of sulfuric acid while stirring and at a temperature of 100 °C. The contents were poured into polystyrene vials for freezing in dry ice‐ethanol bath and then freeze‐dried [62]. The crosslinked composite aerogels showed low flammability, withstanding a gas flame for over 20 s without noticeable combustions [62].

Nanofibrillated cellulose (NFC)–MMT composites were prepared by both directional freezing (in PTFE mould with copper bottom, on copper cold finger, precooled to 4 °C, a freeze rate of 10 °C min−1 then applied until reaching −150 °C) and non‐directional freezing (in aluminium cups, pre‐cooled at 4 °C overnight and then freezed in liquid nitrogen) [63]. The freeze‐dried samples were tested by compression. The anisotropic samples were compressed both parallel and perpendicular to the freezing direction. Young's modulus was observed to increase from isotropic composites, to anisotropic testing by perpendicular direction (∼2.5–6 times), then to anisotropic testing by parallel direction (∼10–85 times), depending on the content of NFC. For the yield strength, the order of anisotropic (parallel) > isotropic > anisotropic (perpendicular) was observed for all the samples. For the work to failure, the values for isotropic and anisotropic perpendicular tests were close while it was much higher for anisotropic parallel testing. For each type of material, the increasing content of NFC led to increase of Young's modulus, compressive strength, and toughness. Although NFC is flammable, the addition of MMT improved the heat endurance and shape retention up to 800 °C while the mechanical properties were maintained up to 300 °C [63].

6.4 Porous Ceramic–Ceramic Composites

The objectives of preparing ceramic–ceramic composites usually include obtaining enhanced mechanical properties, combining the useful properties of the ceramic components, and offering a synergy effect between components or new properties [110]. When preparing porous ceramic–ceramic composites by the freeze‐casting method, the ceramic particles can be simply blended and suspended in a suitable solvent. The slurries are then frozen and freeze‐dried. Alternatively, the freeze‐casting method can be used to fabricate porous ceramics. The second ceramic component is then incorporated into the first ice‐templated ceramics by methods such as impregnation or chemical vapour deposition (CVD).

Most of the porous ceramic composites are prepared directly by freeze‐casting of the blended ceramic slurries, e.g. lanthanum strontium manganite (LSM)–YSZ [64, 65], kaolinite–silica [66], Al2O3–ZrO2 [6769], mullite–alumina [70], HfB2–MoSi2 [71], and HA‐based composites [7274]. For the two‐stage method, porous Si3N4 was first formed by freeze casting. A silica sol was impregnated into the pores and the subsequent gelation and solvent replacement and evaporation led to the formation of silica aerogel within the porous Si3N4 [75]. In another example, carbon nanotubes (CNTs) aerogel was first prepared by freeze casting and then exposed to methyltrichloromethane in a CVD furnace. The vapour could readily infiltrate into the porous CNT. After the CVD process, porous CNT/SiC composites were generated [76].

For the Al2O3–ZrO2 composites, when the solid content was increased from 40% to 70%, an increase of compressive strength from 15 to 81 MPa was observed, while the porosity decreased from 74% to 35% [67]. The size of the ceramic particles (associated with their densities) could affect the strength of the ceramics and their relative concentration in the ceramic composites [68]. For the three‐layer Al2O3–ZrO2 composites, a graded porous structure was observed, with the compressive strength in the range of 63–376 MPa for the sintered ceramics [69]. The mullite–alumina with a grade‐layered pore structure also exhibited improved compressive strength [70]. For the kaolinite–silica composites, the improved strength was related to the solid loading and the packing and interfacial contact of silica and kaolinite particles [66]. The aligned porous LSM–YSZ composites combined electronic conduction from LSM and ionic conduction from YSZ while the aligned porous structure could provide a direct percolation pathway for ionic and electronic species [64, 65]. This can be highly beneficial as electrode materials for solid oxide fuel cells. The ultrahigh‐temperature ceramics were demonstrated by the preparation of HfB2–MoSi2 composites via a gel‐casting route. The solid loading could be used to tune the mechanical properties including flexural strength and fracture toughness. The values of the fracture toughness for the sintered ceramics were in the range of 2.18–4.24 MPa [71].

For HA‐based ceramic composites, HA provides the required biocompatibility for bone tissue engineering while the addition of other ceramics can enhance the mechanical stability of the porous scaffolds. In the HA–SiO2 composites, the addition of silica nanoparticles introduced the partial phase transformation of HA to β‐TCP, which improved the thermal stability with less shrinkage after sintering and the attachment and proliferation of human osteoblast‐like cells [72]. With the addition of alumina nanoparticles, both the pore size and compressive strength were increased. This could be attributed to the formation of calcium aluminate phases, increased wall density and thickness, and reduced porosity [73]. With the inclusion of barium titanate in the HA ceramic composites, the piezoelectric effect was introduced to the scaffold, which played an important physiological role in bone growth and fracture healing. For the composites containing 70% and 90% barium titanate, the piezoelectric coefficient d33 was found to be 1.2 and 2.8 pC/N, respectively [74].

In the composites of Si3N4–silica aerogel, Si3N4 is a wave‐transparent material while the mesoporous silica aerogel provides very good thermal insulation. The ice‐templated pores are too large compared to the mean free path of air molecules (∼69 nm) and thus cannot provide efficient thermal insulation. Indeed, the prepared composite with silica aerogel showed low thermal conductivity (0.043 W (m K)−1), low dielectric constant (∼1.6) and loss tangent (∼0.0018) [75].

6.5 Nacre‐like Layered Ceramic–Polymer Composites

6.5.1 Nacre and Nacre‐mimic Composites

It has been long recognized that it is a huge challenge to produce materials with high strength and high toughness because these two properties are usually mutually exclusive [20]. Learning from Nature where extraordinarily strong and tough materials exist [21, 22, 77], it is possible to fabricate such materials by combining minerals/ceramic particles and organic layers such as polymer and proteins. The nanoscale and weak interfaces between minerals and organic layers seem to be the main cause for the toughening effects. The toughening mechanisms include the alleviation of locally high stresses and different extrinsic mechanisms such as crack deflection, bridging, and friction [20, 77].

Among various tough composite materials, nacre and nacre‐mimic layered materials are widely investigated [2123, 78]. Nacre adopts a unique brick‐and‐mortar structure. Nacre consists of 95% volume fraction of inorganic component (aragonite calcium carbonate platelets) and about 5% of organic polymer (chitin and protein) [22, 78]. The inorganic platelets are 8–10 µm wide and 0.4–0.5 µm thick. They are largely stacked/assembled into inorganic layers, with a distance of 20 and 50 nm between them, which is filled with the organic component. There are nanoasperities of 10–30 nm in width (with a distance of 100–200 nm between them) on the surface of the microplatelets. The microplatelets are also connected by mineral bridges [77, 78]. Nacre exhibits exceptional mechanical properties. Its tensile strength and Young's modulus are in the range of 80–135 MPa and 60–70 GPa [23, 78]. A typical tensile stress–strain curve is shown in Figure 6.11 [79]. The hydrated nacre can give a much higher tensile toughness compared to that of the dry nacre. The elongation of the fibril organic polymer/protein is about 100 nm before stiffening. This corresponds well with the inelastic deformation of nacre (∼1%) [23]. The fracture toughness can be as high as 1.5 kJ m−2 [78]. It is obvious that the inclusion of a small percentage of organic layer in nacre improves the strength and toughness of 20–30 times compared to the pure monolithic aragonite [22]. This is attributed to multiple energy dissipation mechanisms, as proposed by Mayer [80] and is summarized below:

  • Crack deflection, delamination, microcracking due to weak interfaces
  • Anchoring, crosslink breaking, polymer chain deformation and unfolding, ligament formation of the organic phase, and crack bridging by organic ligaments
  • Pull‐out of the ceramic phase, friction resistance due to surface toughness/mineral bridges of the platelets, and hole formation at the ends of displaced ceramic phase
  • Plasticizing effects on the polymer layer by moisture
  • Residual stresses to energy absorption.

Figure 6.11 A representative stress–strain curve for nacre.

Source: Song et al. 2008 [79]. Reprinted with permission from American Physical Society.

Different methods have been utilized to prepare layered materials. The traditional methods such as tape casting and slip casting may be difficult to produce the layer thickness into the small micron or nanoscale region [17]. The electrophoretic deposition is slow and is difficult for the preparation of large‐sized materials [23]. The LBL deposition is a versatile method to prepare layered materials with controllable thickness. The method mainly depends on the electrostatic interaction between the two components. For example, the polymer with positive charges, chitosan, and negatively charged clay particles (e.g. MMT) and alumina particles have been used to prepare composite materials [23]. However, the LBL approach is time‐consuming and it is very difficult to generate layered composites with high percentage of inorganic components. Recently, the freeze‐casting approach has been intensively used to fabricate tough nacre–mimic composite materials [2124, 56]. By controlling the slurry compositions and freezing parameters, ceramic particles can be excluded from the freezing front, forming dense ceramic layers between ice crystals [81]. The layer thickness and the spaces between layers can be adjusted by controlling the freezing conditions (and other post‐treatments for the dry porous scaffold, see following discussion). This freezing step is also highly effective in assembling ceramic platelets [46, 82]. The layered porous ceramics can be produced after freeze‐drying and sintering. Subsequently, organic polymers may be incorporated into the pores, resulting in layered composites.

6.5.2 Layered Polymer–Ceramic Composites by Infiltrating Ice‐templated Porous Ceramics

Porous alumina with layered structures can be fabricated via directional freezing. The thickness of the alumina walls and lamellar spacing can be effectively tuned by changing the velocity of the freezing front [81]. This approach is highly versatile. For example, a layered porous HA scaffold was fabricated by freeze‐casting and then filled with an organic phase epoxy resin to create a nacre‐like composite [81]. The HA–epoxy composite has a much higher load‐displacement profile. Indeed, the HA‐based composite exhibited exceptional stiffness (10 GPa), strength (150 MPa), and work of fracture (220 J m−2) [81]. Porous alumina with lamella length up to 70 mm was fabricated from aqueous slurries with solid alumina (particle size D50 = 390 nm) contents of 25–45 vol%. The freeze‐dried green body was sintered at 1550 °C. An epoxy resin was infiltrated into the porous alumina under vacuum and then left for 24 h for the resin to harden [83]. The composite exhibited fracture toughness of 2.92 MPa m½, and Weibull strength of up to 117 MPa [83].

Nacre exhibits a very high inorganic component (∼95%) and a mortar‐and‐brick structure. It is obvious that simply infiltrating a polymer phase into a layered porous ceramic does not meet this requirement. The issues are as follows: (i) the ceramic layer does not allow sufficient infiltration into itself (i.e. a layered composite but not a mortar‐and‐brick structure); (ii) the content of the inorganic component in the composite is too low; (iii) there appears to be only a limited interaction or no interaction at all between the polymer phase and the ceramic phase. To address these issues, Ritchie and co‐authors developed a method to fabricate highly toughened alumina–polymer composites with high yield strength (∼210 MPa) and fracture toughness (∼30 MPa m½) [84, 85]. A procedure was developed to hot press the porous alumina scaffold and condense and break the layered structure. Sucrose was used as an additive in the slurry during the freezing process, which helped to produce a layered porous structure with rough spiked surface. In order to reduce the porosity (and thereby increase the content of the inorganic component in the composite) and create a brick‐and‐mortar structure, the scaffold was firstly infiltrated with paraffin wax and then uniaxially pressed at 80 °C (just melting paraffin wax) at ∼50–100 MPa. The wax phase facilitated the densification of the porous structure at 80 °C and then held the bricks (broken lamellar structure) together. A further thermal treatment at 400 °C (2 h) and 1500 °C (2 h) in air removed the organic phase and promoted further densification. In the final step, the scaffold was pressed isostatically at 1.4 GPa followed by another step of heating treatment (1500 °C, 2 h) to facilitate the formation of inorganic bridges between the bricks [84, 85]. The interaction between the polymer phase and alumina phase was addressed by grafting the polymer to the ceramic bricks. This was achieved by first grafting 3‐(trimethoxysilyl)propyl methacrylate (γ‐MPS) to the alumina washed by a Piranha solution, via the reaction between -OH and Si‐OMe (methoxysilyl group). A partially polymerized methyl methacrylate (MMA) with high content of initiator was then impregnated into the γ‐MPS‐grafted alumina. The impregnated scaffold was annealed at 150 °C under nitrogen for 2 h to complete the polymerization and the grafting. Using this method, PMMA–grafted alumina composite with a brick‐and‐mortar structure and an alumina content of ∼80% has been produced [84, 85].

For comparison, the brick‐and‐mortar composites with non‐grafted PMMA and the lamellar composites with both types of PMMA (grafted and non‐grafted) were also prepared [84, 85]. High yield strengths (120 and 210 MPa) were obtained for non‐grafted and grafted brick‐and‐mortar composites, respectively. Remarkably, these composites showed >1% inelastic deformation before failure. JR curves were used to characterize the toughness behaviour of non‐LEFM observed in these composites. A rising R‐curve demonstrated the extensive extrinsic toughening. The lamellar and brick‐and‐mortar composites could reach fracture toughness of 15 MPa m½ (Jc ∼ 5000 J m−2) and 30 MPa m½ (Jc ∼ 8000 J m−2), respectively, much higher than that by nacre and the homogeneously compressed alumina particles). The effects of strong interface (PMMA–grafted alumina) and weak interface (non‐grafted PMMA–alumina) on fracture toughness (KJ) were investigated. In general, the strong interface in the composites enhanced the fracture toughness (Figure 6.12). The fracture toughness only slightly increased (e.g. 12–16 MPa m½, Table 6.3) for lamellar structures while the values were doubled for the brick‐and‐mortar composites [85]. Multiple toughening mechanisms were observed, including ligament bridging, crack deflection/delamination, void growth, microcracking, and inelastic deformation. Particularly, the cracking and the extent of damage were not localized. In the lamellar composites, the relatively thick PMMA layer still played a structural and load‐bearing role. However, for the brick‐and‐mortar composites, the pull‐out toughening of the bricks was significant, quite like most of the engineering structural ceramics. The unique toughening mechanism was from the sliding between the alumina blocks facilitated by the sub‐micro PMMA film. This also explained the stronger effect of grafting polymer phase on the fracture toughness of the brick‐and‐mortar composites. The grafting could provide a more effective viscoelastic phase that permits but limits the extensive interface sliding [85].

Figure 6.12 The plots of the stress intensity, KJ, as a function of crack extension, show the crack‐resistance behaviour of (a) lamellar alumina–PMMA composites and (b) brick‐and‐mortar alumina–PMMA composites.

Source: Launey et al. 2009 [85]. Reprinted with permission from Elsevier.

Table 6.3 Mechanical properties of layered ceramic composite materials.

SampleFlexural strengthHardnessFracture toughnessaFabrication methodReferences
HA/epoxy150 MPa10 GPa220 J m−2Freeze casting[81]
Alumina/epoxy117140 GPa2.8 MPa m½Freeze casting[83]
Alumina–PMMA (non‐grafted)90 MPa12 MPa m1/2 (KJC)Freeze casting[85]
Alumina–PMMA (grafted)112 MPa12 MPa m1/2 (KJC)Freeze casting[85]
Alumina–PMMA (hot pressed, non‐grafted) 115 MPa15 MPa m1/2 (KJC)Freeze casting, mortar and brick[84, 85]
Alumina–PMMA (hot‐pressed, grafted)210 MPa32 MPa m1/2 (KJC)Freeze casting, mortar and brick[84, 85]
Alumina–cyanate ester300 MPa11 GPaFreeze casting, interlocked structure[86]
HA/PMMA100 MPa20 GPa2075 J m−2Bidirectional freezing, grafting[87]
Aragonite/silk fibroin3 MPa m1/2 (KJC)Mineralization in ice‐templated chitin[88]
Zirconia/epoxy160 MPa4.0 GPaFreeze casting with n‐butanol clathrate hydrate effect[89]
Graphene/Si‐O‐C63 MPa3–3.5 MPa m1/2Graphene network by freeze casting[90]
CNTs/epoxy150 MPa24 GPa4.26 KJ m−2CVD for CNT forest[91]
Alumina/PMMA220 MPa8.5 MPa m½Magnetically assisted slip casting[92]
Zirconia/epoxyShear strength 51 MPaShear modulus 5.5 GPaHelix‐reinforced, magnetic freeze casting[93]
Alumina/Al‐Si300 MPa40 MPa m½Freeze casting[94]
SiC/2024Al931.3 MPa18 MPa m½Freeze casting[95]
SiC/Al‐Si‐Mg722 MPa140 GPaFreeze casting[96]
Alumina/silica‐calcia470 MPa290 GPa22 MPa m½Freezing and hot pressing[97]
Alumina/silica‐alumina650 MPa190 GPa14 MPa m½MASC[92]
Alumina/titania370 MPa150 GPa6.5 MPa m½ (KIC)MASC[98]
Alumina/graphene523 MPa17.66 GPa4.49 MPa m½Solvent evaporation and hot pressing[99]
Si3N4/graphene6.6 MPa m½Solvent evaporation and hot pressing[100]
ZrB2‐SiC/graphene522 MPa12.5 GPa9.45 MPa m½Self‐assembled and hot pressing[101]

aThese values are quoted from relevant references. The data may not clearly indicate whether they are fracture toughness at initial cracking or the maximum fracture toughness. Readers are suggested to read the original references carefully.

Zhao et al. added SCMC (4–9%) to alumina slurries, which facilitated the formation of highly ordered ceramic bridges between alumina layers (9 wt% SCMC in the slurries) (Figure 6.13) [86]. The freeze‐dried scaffolds were sintered at 1600 °C for 4 h before cyanate ester was infiltrated under vacuum and cured at 220 °C in an air‐circulated oven. This method did not employ the hot pressing procedure but produced a nacre–mimic interlocking composite structure. The composite exhibited a flexural strength of up to 300 MPa, a specific strength of 162 MPa (g cm−3)−1 (i.e. the strength divided by the density), and a failure strain of 5% [86]. Since it has been very difficult to produce ceramics with large domains of lamellar structure (≲mm) by the conventional freeze‐casting approach, Bai et al. employed a bidirectional freezing method to fabricate nacre‐like composites as large as 4 × 8 x 25 mm [87]. A PDMS wedge was introduced between the cold finger and the slurry, promoting the formation of long‐range lamellar structure. This was demonstrated by the fabrication of HA/PMMA composites (grafted mortar‐and‐brick structure). The long‐range lamellar structure was pressed uniaxially to give densified scaffolds with porosity of 15–25%. Similarly to that reported previously [84, 85], monomer MMA and initiator 2,2‐azobisisobutyronitrile (AIBN) were infiltrated into the γ‐MPS‐grafted scaffold. The in situ polymerization at 40 °C and aging at 90 °C for another 2 h produced the HA–PMMA composites with high ceramic content (75–85%) and excellent mechanical properties (flexural strength 100 MPa, Young's modulus 20 GPa, and work of fracture up to 2075 J m−2) [87]. In addition to in situ polymerization of the monomers, PMMA may be incorporated into the ceramic scaffolds by emulsion polymerization [102].

Figure 6.13 Addition of SCMC promotes the formation of bridges between alumina layers as shown by SEM images (a–d) with the SCMC concentrations of 4, 5, 7, 9 wt%, respectively, and the schematic description (e).

Source: Zhao et al. 2016 [86]. Reprinted with permission from John Wiley and Sons.

The incorporation of a second organic phase into layered porous ceramics is also effective in the fabrication of composite scaffolds with stronger mechanical properties. For example, PCL‐acetone (1.5 and 5.0 wt%) solution was impregnated into an ice‐templated porous β‐TCP. Because the amount of PCL was small, PCL was only infiltrated into the small pores within the lamellar layers and coated the layer surface as thin film while the macroporosity was retained [103]. Compared to the just freeze‐dried β‐TCP porous scaffold, the porous composites exhibited considerably improved mechanical properties: e.g. flexural strength increased from 3.4 to 5.3 MPa (4.2 vol% PCL) and 5.6 MPa (5.6 vol% PCL) and compressive strength increased from 4.4 to 6.0 MPa (0.7 vol% PCL), 9.5 MPa ( 4.4 vol% PCL) and 10.3 MPa (7.6 vol% PCL) for the samples prepared from the slurries with a solid content of 20 vol% and onset freezing velocity 20 µm s−1 [103].

Furthermore, the concept of harnessing hydrophobic hydration and clathrate hydrate has been explored for the fabrication of stronger composites [89]. This approach is based on the addition of alcohols into aqueous slurries. In dilute aqueous solution, a hydrophobic molecule (such as alcohol with C-H chain) can bond to water molecules and pattern them around itself, a phenomenon called ‘hydrophobic hydration’, as illustrated in Figure 6.14. This phenomenon occurs in the liquid state at or around room temperature with an important parameter being the number of molecules being patterned (nh). The freezing process can transform such structure into a clathrate hydrate (Figure 6.14) [89]. Alcohols, amines, amides, and some polymers such as PVA are potential additives to form clathrates. For example, ethanol, n‐propanol, and n‐butanol were added to aqueous ZrO2 slurries for the freeze‐casting process. The green bodies were sintered at 1350 °C for 3 h in air. The important findings were that the pore area reached a maximum point at a suitable concentration. But the concentrations for maximum pore areas were different for different alcohols, 10 vol% for ethanol, 5 vol% for n‐propanol, 3 vol% for n‐butanol [89]. This suggested that the concentration decreased with the increasing chain length of alcohol molecules. Interestingly, after impregnating the porous ceramics with epoxy to produce the composite materials, higher ultimate compression strength (UCS) and Young's modulus were always observed for ceramics made at such additive concentrations, usually higher than >50%. This is different from the eutectic crystallisation formed from the additive alcohol and water under suitable conditions [104]. While studying the eutectic effect, it was observed that, in addition to the usual ice‐templated macropores, mesopores were generated in the freeze‐dried material and the porosity was characterized by gas sorption study [104]. However, in the case of clathrate hydrate effects, the pore size and area were calculated from SEM images using ImageJ software, assuming ellipse pores with major axis a and minor axis b. The enlarged hydrate structures resulted in enlarged ice crystals and hence larger pores, as evidenced by ImageJ calculation. This also led to larger pore area, contributed positively to the strength of the composites [89].

Figure 6.14 The diagram illustrates hydrophobic hydration and transformation into clathrate hydrates via a freezing path, based on the example of n‐propanol molecule.

Source: Naleway et al. 2016 [89]. Reprinted with permission from Elsevier.

Instead of incorporating an organic phase into a ceramic scaffold to fabricate nacre‐like composites, mineralization in preformed matrix has been reported recently to generate synthetic nacre [88]. In this method, porous chitosan was first prepared by the ice‐templating approach and then transformed into water‐insoluble chitin by acetylation. The fine dispersion of Ca(HCO3)2 (in the presence of polyacrylic acid and Mg2+) was pumped into porous chitin. The chitin layers were assimilated with the gradual mineralization of the matrix. The aragonite phase was obtained by decomposition of Ca(HCO3)2. The nacre‐like composite was formed by silk fibroin infiltration and hot pressing at 80 °C. This synthetic nacre exhibited a structure similar to that of natural nacre, with 91 wt% of aragonite phase, 2–4 µm thick aragonite layers, and 100–150 nm silk fibroin layers [88]. Like natural nacre, a rising crack‐extension profile was observed. This synthetic nacre showed a fracture toughness of ∼3 MPa m½, a specific strength of ∼30 MPa (g cm−3)−1, and a specific toughness of ∼1 MPa m½ (g cm−3) (i.e. toughness divided by the density) [88].

Porous graphene networks were formed by freeze casting aqueous grapheme oxide suspension and subsequently reduced to graphene at 900 °C. A 20% volume shrinkage was observed after this treatment. A pre‐ceramic polymer, polymethyl siloxane, was then infiltrated into the graphene network. The graphene/Si‐O‐C composites were produced by pyrolysis under N2 and heat treatment at 1000 °C and densified using spark plasma sintering (SPS) [90]. The composite contained a network of very thin (20–30 nm) carbon phase and thus was electrically conductive. This could be used to sense the damaging effect in the composite because a rise in voltage could result from microcracking before fracture and the voltage could be recovered if the network connectivity was recovered. The composite showed an initiation toughness of 1.7 MPa m½ and a maximum toughness of 3–3.5 MPa m½ when the rising R‐curve reached the steady state, a calculated work of fracture around 46 J m−2 [90]. Carbon networks could also be formed by other methods and utilized to form strong composites. For example, the CVD technique was used to grow CNT forest perpendicular to woven SiC fabrics [91]. The nanotube‐grown fabrics were infiltrated with epoxy and stacked together to produce layered composites. These composites gave rise to excellent mechanical properties such as flexural strength 150 MPa, flexural modulus 24.3 GPa, and work of fracture 4.26 KJ m−2 for GIC and 140 J m−2 for GIIC [91].

6.5.3 Magnetic Field‐assisted Freeze‐casting for Strong Composites

The application of magnetic fields is convenient to align particles and control their distribution in composites. The only problem is that most ceramic particles are either paramagnetic or only exhibit very weak magnetic properties. Very strong magnetic fields (>1 or even >10 T) are required to control the distribution of feeble magnetic ceramic particles [105, 106]. This problem can be addressed by coating the ceramic particles with superparamagnetic nanoparticles such as iron oxide. Particularly, by using the correct geometry of coated ceramic particles, an ultrahigh magnetic response can be achieved and hence only a weak magnetic field is required to align the particles. For example, alumina platelets (7.5 µm long and 200 nm thick) and calcium sulphate hemihydrate rods (10 µm long, 1 µm thick) were coated with 12 nm iron oxide nanoparticles. A weak magnetic field of 0.8 mT was required to tune the 3D orientation and distribution in a suspension containing polymers or monomers. The composites could be subsequently produced by evaporating the solvent or polymerizing the monomers [105]. The alumina platelets coated with iron oxide nanoparticles were then used in a magnetically assisted slip casting (MASC) process [92]. The rotating magnetic field could be programmed and applied to control the alignment/angle of the assembled platelets. After heating at 500 °C for 3 h to remove the organic binder, the MASC ceramic was uniaxially hot pressed (up to 100 MPa), which increased the volume fractions of the platelets from 35 to 60 vol%. At this stage, the monomer MMA could be infiltrated and the subsequent polymerization led to the production of alumina–PMMA composite. This composite showed a fracture toughness of 8.5 MPa m½, fracture strength of 220 MPa, and specific strength of 76 MPa (g cm−3)−1 [92].

The magnetic field can be combined with a freezing process [106]. When the orientation of the magnetic field is perpendicular to the freezing direction, the magnetic force could facilitate the formation of bridges between the ceramic layers in slurries containing ceramics such as alumina and a small percentage of iron oxide nanoparticles (e.g. 3 or 9 wt%). When a rotating field is applied, ceramic scaffolds with a spiral pattern of iron oxide can be produced [106]. This method was applied to fabricate helix‐reinforced composites with enhanced torsional properties [92]. Aqueous ZrO2 (diameter 0.2–0.5 µm) slurries with solid content of 10 or 20 vol% and 3 wt% (of the solid content) 50 nm Fe3O4 nanoparticles were directionally frozen with a rotating magnetic field of 0.12 T. The green bodies were sintered in air at 1300 °C for 3 h, followed by the infiltration with epoxy [93]. Figure 6.15 shows the ZrO2–epoxy composite containing 40% or 60% zirconia. The spiral patterns can be clearly seen and are affected by the rotating speed of the magnetic field 0.12 T. The dark stripes are concentrated Fe3O4 particles and the distance between the stripes becomes smaller with the increasing rotating speed. The helix angle changes with the variation of rotating speed as well. Inorganic bridges can be clearly seen between the ceramic layers (Figure 6.18d and e). Shear strength up to 54 MPa and shear modulus up to 5.5 GPa were recorded for these helix‐reinforced ZrO2–epoxy composites [93].

Figure 6.15 Helix‐reinforced zirconia–epoxy composites via a magnetically assisted freeze‐casting process. (a–c) Show the images of 40 : 60 composites under a rotating magnetic field of 0.12 T at the speed of (a) 0.05 rpm, (b) 0.20 rpm, and (c) 0.40 rpm. The helix angles are also illustrated in the images. (d–f) Show the structure by SEM imaging of (d) 40 : 60 composite, (e) 60 : 40 composite, and (f) the magnified structure of the 60 : 40 composites.

Source: Porter et al. 2015 [93]. Reprinted with permission from Elsevier.

It is clear that the polymer phase plays a significant role in enhancing fracture toughness in nacre‐like composites. While the organic phase in natural nacre consists of chitin and protein, the synthetic nacres investigated so far mostly contain PMMA and epoxy resin as the organic phase (Table 6.3). The use of different polymers and the role of polymer phase in the energy‐dissipating toughening mechanisms are essential to fabricate toughened composites. Different polymers were thus infiltrated into the porous alumina produced by the MASC approach (silica nanoparticles included to introduce mineral bridges and nanoasperities) [107]. These polymers were carefully selected with different properties: (i) poly(lauryl methacrylate) (PLMA), a soft and weak elastomer; (ii) PMMA, a strong, stiff and brittle thermoplastic; (iii) polyether urethane diacrylate‐co‐poly(2‐hydroxyethyl methacrylate) (PUA‐PHEMA), a polymer of immediate strength and stiffness. The monomers were infiltrated into the hot‐pressed ceramics and polymerized in situ. The flexural strength and fracture behaviour were assessed by three‐point bending. As shown in Table 6.4, compared to the scaffold without polymer, the effects of the polymer phase on the composites are clearly different, depending on the types of the polymers. With the soft polymer PLMA, the mechanical properties are very similar to the scaffold and the fracture toughness even decreases. For the stiff and brittle PMMA, the fracture strength is the highest while fracture toughness is lower than the composite containing PUA‐PHEMA. The tougher polymers also contribute to increased strains at failure while flexural modulus does not change much over the four materials. Based on the KJC‐crack extension graph, the rising JR curves were observed to the crack extension of 0.8 mm and were still not steady for both PMMA‐ and PUA‐PHEMA composites [107].

Table 6.4 Mechanical data of nacre‐like composites of alumina with different polymers.

Source: Niebel et al. 2016 [107]. Adapted with permission from Elsevier.

MaterialsFlexural modulus (GPa)Fracture strength (MPa)Strain at failure (%)Fracture toughnessa (MPa m½)Work of fracture (J m−2)
Alumina scaffold35.0 ± 8.199.7 ± 9.90.27 ± 0.061.88 ± 0.2142.6 ± 8.7
Alumina–PLMA35.3 ± 5.499.9 ± 10.30.25 ± 0.021.43 ± 0.3042.4 ± 7.8
Alumina–PMMA36.6 ± 6.7182.3 ± 13.80.43 ± 0.062.40 ± 0.35110.0 ± 4.2
Alumina–PUA‐PHEMA36.8 ± 3.8168.1 ± 17.50.44 ± 0.033.39 ± 0.28156.2 ± 1.7

aThe values are for fracture toughness at crack initiation.

6.6 Nacre‐like Ceramic/Metal Composites

In order to enhance fracture toughness, composites are usually fabricated from a brittle material scaffold and a second ductile‐phase material. Lamellar composites have shown to be mostly effective in increasing toughness, compared to particles and fibres, for the same volume fraction of ductile reinforce phase [94]. Metals have been widely used as ductile phase to prepare ceramic–metal composites (CMCs). It is thus quite straightforward to fabricate nacre‐like CMCs by replacing polymer with metal. Unlike the use of polymer solution or liquid phase monomers, metal melt will have to be infiltrated into a porous ice‐templated ceramic scaffold.

Layered porous alumina was fabricated by freeze casting and then sintered at 1550 °C for 1 h. A eutectic aluminium–silicon alloy (Al–12Si) was infiltrated into the porous alumina [108]. The elastic constants of the composites were obtained by a wave velocity approach. The composite fabricated from the scaffold with the finer structure exhibited a density of 3.18 ± 0.04 Mg m−3 and an elastic constant (C11) of 240 ± 10 GPa [108]. In another study, the alumina scaffold was heated to 900 °C under a 10−4 Pa vacuum. An argon gas pressure of 70 kPa was introduced to facilitate the infiltration of the molten Al–Si eutectic alloy [94]. The composites contained ∼40% ceramic phase with Al–Si layer thickness of 10 µm and showed an excellent tensile strength of ∼300 MPa and a fracture toughness of 40 MPa m½ [94]. Because metals are usually good thermal conductors, anisotropic thermal conductivity can be achieved for the ceramic–metal composites. In this regard, aligned porous alumina and zirconia scaffolds were fabricated by directional freezing [109]. An aluminium alloy (AlSi10Mg) was infiltrated into the scaffolds at 740 °C using a vacuum pressure metal casting device. Thermal conductivities of 80 W m−1 K−1 (parallel to the freezing direction) and 13 W m−1 K−1 (perpendicular to the freezing direction) were achieved for the zirconia/metal composites [109].

Other CMCs are also prepared by this approach. Lamellar SiC/2024Al composites were fabricated by infiltrating the molten Al alloy into the SiC scaffold at 800 °C. A ram was activated by a 40 ton hydraulic press to force the melt into the scaffold. The highest strength and toughness achieved for the composites were 931.3 MPa and 18.8 MPa m½ (measured by three‐point bending test), respectively [95]. For another SiC composite, after sintering the lamellar porous SiC scaffold at 1200 °C for 2 h in air, the Al–Si–Mg alloy melt was infiltrated into the scaffold at 950 °C under the flowing N2 atmosphere [96]. A new phase of MgAl2O4 was formed to strengthen the interfacial bonding between the alloy phase and the SiC particles. An ultrasonic pulse echo technique was used to measure the elastic strength and modulus. The elastic modulus was found to increase with the volume percentage of SiC (up to 40 vol%), to a value of 163 GPa. The values were nearly the same between the longitudinal direction and the transverse direction. For the composite with 30 vol% SiC, the compressive strength was the highest, 722 ± 35 MPa for longitudinal direction and 555 ± 27 MPa for transverse direction [96].

The aim for the nacre–mimetic composites with polymer phase, has always been to increase the content of inorganic (ceramic) component and achieve nanoscale thickness of the polymer layers. However, for the nacre‐like ceramic–metal composites, the content of metal phase is always very high and often higher than the ceramic phase. The mechanical properties are more like a combination of two components and a balance (or compromise) between toughness (elasticity) and strength. In the Al/TiC composite with 15 vol% TiC, the bending strength was 355 MPa, the elastic modulus was 100 GPa, and the toughness was high at KJC 81.0 MPa m½ (measured by three‐point bending test without pre‐fabricated notch). However, for the composite containing 35 vol% TiC, the bending strength increased to 500 MPa while the toughness decreased to 32.9 MPa m½ [110]. This could be explained by non‐localized multiple‐crack propagation in the 15 vol% TiC composites (Figure 6.16a and b). The thick metal layer could effectively dissipate the stress at the crack tip. Further, due to the loose ceramic layers, the metal phase was able to infiltrate between the ceramic particles. This could promote the occurrence of non‐localized cracks. In the case of Al/25 vol% TiC and Al/35 vol% TiC, the thin metal layers were unable to disperse the stress at the crack tip, leading to propagation of cracks and failure (Figure 6.16c and d) [110].

Figure 6.16 The SEM images show the initial cracks of the composite (a and b) Al/15 vol% TiC, (c) Al/25 vol% TiC, and (d) Al/35 vol% TiC after the three‐point bending test.

Source: Wang et al. 2017 [110]. Reprinted with permission from Elsevier.

The mechanical properties of ceramic–metal composites may be enhanced via interfacial reactions [96]. For example, commercial alloy AZ91 (Mg‐9Al‐1Zn) was infiltrated into porous SiC at 650 °C [111]. The SiC scaffold was pre‐oxidized at 1100 °C for 2 h, forming a SiO2 layer. Mg vapour was generated during infiltration and reacted with the oxidized surface, generating a Mg2Si phase. This could enhance the mechanical property of the composite. The flexural and compressive strengths of the composite with 30 vol% ceramics were measured to be 599 ± 44 and 743 ± 20 MPa, respectively [111]. ZL107 alloy (Al‐7Si‐5Cu) was infiltrated into the freeze‐cast lamellar Al2O3–ZrO2 scaffold at 850 °C under a 2 MPa Ar gas. The composite was further heated at 850 °C for a period of 0–120 min. The heating facilitated the reaction between Al and ZrO2 [112]. After heating for about 60 min, the primary products of the reaction, (Al1−m, Sim)3Zr and ZiS2, could develop into bridges between the ceramic layers, thus enhancing the strength and toughness of the composites [112].

6.7 Enamel‐mimic Ceramic–Polymer Composites

The mammalian tooth can be divided into four parts [113]: enamel, dentin, pulp, and cementum. Enamel is the hardest mineralized tissue in the body and consists of 95–97 wt% carbonated HA and <1 wt% of organic materials. Enamel is the uppermost 1–2 mm part of the tooth crown [114]. It is highly strong but may be brittle. The dentin layer has similar composition and mechanical properties as bones. The dentin layer can dissipate the stress from the enamel and prevent it from cracking. The geological or chemical HA has a formula of Ca5(PO4)3(OH) [42]. However, it is carbonated HA, Ca10(PO4, CO3)6(OH)2, that is present in bone and enamel. Carbonated HAs can be classified as type A (carbonate replacing hydroxide) and type B (carbonate replacing phosphate). They can be distinguished by infrared spectroscopy [114]. The compositions of carbonated HA in bones vary because bones serve as a reservoir to maintain homeostasis balance with ions such as Ca2+, Mg2+, and . It is much less the case for enamel because enamel does not resorb and remodel once matured [114].

Enamel is formed in two stages: secretion and maturation. While ameloblast functions in the secretion stage, amelogenin is the dominant organic component in the mature enamel [115]. Amelogenin is hydrophobic and contains C terminus, and can self‐assemble into nanospheres (and then maybe chains and ribbons), which promote the orientated growth of HA crystallites along the c‐axis [114, 115]. Therefore, in the enamel, bundles of HA rods are sandwiched (and maybe with an interweaving structure depending on tooth type [116]) by very small amounts of organic components. The structures of some teeth, with aligned and/or bundled columnar features, are illustrated in Figure 6.17 [117]. In addition to the main component amelogenin, the other organic components in enamel are enamelin (a glycoprotein, strongly adsorbed to enamel crystals) and ameloblastin (also known as amelin or sheathlin) [114]. The strength (elastic modulus) of enamels has been measured by various methods and researchers [115]. The obtained values vary. In general, the elastic modulus is in the region of 30 GPa for a cross‐sectional surface and 10 GPa for an occlusal surface by macroscale tests. For the tests by nanoindentation, the values are in the region of 80–100 GPa [115].

Figure 6.17 The columnar motifs in the teeth of (a) human tooth enamel, (b) beak of Octopus vulgaris; (c–e) tooth enamel from the Odobenidae family (c), from Tyrannosaurus rex (d), and from Albertosaurus spp (e). (f) The structure of the composites fabricated by layer‐by‐layer (LBL) coating of vertical ZnO nanowires.

Source: Yeom et al. 2017 [117]. Reprinted with permission from Nature Publishing Group.

Because enamel cannot resorb and regenerate, tooth implant and tooth restoration are clearly required. It is no surprise that dental materials must have high strength and, ideally, high fracture toughness. Ceramic–ceramic and ceramic–metal composites are conventionally used for dental applications [118]. Alumina and silica‐based ceramics have been traditionally used. In recent years, zirconia has appeared to be an alternative ceramic. Zirconia exists in three crystalline forms: monoclinic (low temperature), tetragonal (>1170 °C), and cubic (>2370 °C). Zirconia exhibits superior mechanical properties than alumina, i.e. excellent flexural strength, fracture toughness, and corrosion resistance, particularly for partially stabilized zirconia [119]. Zirconia is chemically inert. It is very difficult to modify a zirconia surface, which hinders the preparation of strong zirconia composites via polymer grafting. This, however, may be addressed by coating zirconia with silica. A variety of silanol chemistry can then be employed to achieve a better adhesive bonding [119]. Zirconia implants have been examined systemically from clinical data. The short‐term cumulative survival rates and marginal bone loss are promising while more data are required for long‐term predictability [120].

Traditional dental materials are usually composites with ceramic particles randomly distributed in a polymer matrix. To mimic the structure of enamel or dentin, aligned porous ceramics fabricated by directional freezing can be used to produce such composites. Petrini et al. prepared aligned porous alumina with graded porosity which was then infiltrated with epoxy resin [121]. The composite structure was more like that of dentin (composed of 70 wt% HA, 20 wt% of organic materials, and 10 wt% water, with packed tubule structure). Owing to the graded porosity of the alumina scaffold (dense structure at the bottom and well‐aligned porous structure in the upper part), different mechanical properties were obtained for the top and bottom of the composite. For the top part of the composite, the elastic modulus was 9.1 ± 0.3 GPa, flexural strength 213.1 ± 59.9 MPa, and compressive strength 222.4 ± 44.9 MPa. For the bottom part, the elastic modulus was 6.5 ± 2.3 GPa, flexural strength 183.1 ± 49.4 MPa, and compressive strength 140.6 ± 66.6 MPa. Based on these data, it can be inferred that the flexural strength is similar to that of dentin. The mechanical property of the top part of the composite is better than that of the bottom part, despite the high porosity and low ceramic content of the top part. This was attributed to the presence of closed pores in the bottom ceramic [121].

As has been mentioned in Section 6.5, the strategy adopted in the freeze‐casting approach in the preparation of nacre–mimic composites has been to increase the content of the inorganic component and reduce the thickness of the organic layer. Because of the very high percentage of the inorganic component in enamel (up to ∼97 wt%), it is highly challenging to produce such composites via a freeze‐casting approach. In addition to combining hot pressing with freeze casting, other approaches may be employed to fabricate enamel‐like composites. Very recently, inspired by the columnar structure of various types of teeth, vertically arranged ZnO nanowires were fabricated via a hydrothermal approach, followed by LBL deposition of polyelectrolytes between the ZnO nanowires (Figure 6.17f) [117]. The polyelectrolytes used were polyallylamine and polyacrylic acid. This produced a composite with a 67 vol% inorganic component. The procedure was repeated and the composites could be stacked to produce strong composites with the Young's modulus at 39.8 ± 0.9 GPa (by the Berkovich probe) or 1.65 ± 0.06 GPa (by the conical probe). These data are comparable with those of tooth enamel with 85 vol% inorganic component (modulus of 62–108 GPa and hardness of 1.1–4.9 GPa) [117].

6.8 Ceramic–Ceramic Composites with the Second Nanoscale Ceramic Phase

Tough ceramic composites have been fabricated via the introduction of a ductile phase such as polymer and metal into a brittle porous ceramic scaffold. This may reduce the strength of the resulting materials (compared to the ceramics) although there are examples of both strength and toughness being improved for the composites. There is, however, another problem with the use of polymer and metal in the composites. The thermal stability of these composites is limited by the presence of polymer or metal. This could be a serious drawback for high‐temperature applications. Therefore, there is a need to fabricate ceramic–ceramic composites that also demonstrate high strength and high toughness.

Learning from nacre and nacre‐like materials, it has been found that increasing the local density at the interfaces can lead to enhanced toughening, primarily via crack deflection. The increase of local density can be achieved by the use of small nanoscale constituents at the interface. This concept has been demonstrated by Deville and co‐authors [97]. The freezing procedure was used to assemble the anisotropic platelets between the ice crystals. The initial slurry contained alumina platelets (7 µm in diameter and 500 nm in thickness), alumina nanoparticles (100 nm, serving as inorganic bridges and nanoasperities at the platelet surface), and smaller nanoparticles of silica–calcia phase (20 nm, filling the remaining gaps during the sintering stage). The fabrication procedure is illustrated in Figure 6.18. After the pressing step, the material was sintered by field‐assisted sintering at 1500 °C [97]. The resulting composite exhibited the combination of high strength (470 MPa), high toughness (22 MPa m½), and high stiffness (290 GPa). Because the composite only consisted of ceramics, the mechanical properties were retained at a relatively high temperature of 600 °C. The rising KJC‐crack extension was still observed, with the initial toughness (KIC) of 4.7 MPa m½ and maximum toughness (KJC) of 21 MPa m½ [97].

Figure 6.18 Fabrication of all ceramic composites via a freezing path. (a) Self‐organization of particles of different shapes and sizes during the freezing procedure. (b) Schematic representation of the densification of the composites by sintering with pressing. (c) Linear shrinkage of the composites with different components during the densification process.

Source: Bouville et al. 2014 [97]. Reprinted with permission from Nature Publishing Group.

In the MASC process for the preparation of composites, inorganic silica and alumina nanoparticles could be added into the initial slurry to fabricate all ceramic composites. The ceramic–ceramic composite exhibited excellent properties, i.e. flexural modulus of ∼190 GPa, flexural strength of 650 MPa, initial toughness (KIC) of ∼5 MPa m½ and maximum toughness (KJC) of 14 MPa m½ [92]. In a further study, the alumina platelets were coated with titania and then rendered magnetically responsive by adsorption of iron oxide nanoparticles [98]. Titania could be sintered at lower temperatures (e.g. a temperature higher than 1100 °C is necessary to sinter alumina), which could significantly minimize the processing time and material heterogeneity. During the sintering with pressing, titania particles could form in situ mineral bridges between alumina platelets in the temperature window of 900–1100 °C to form strong composites. Indeed, the composite fabricated around 900 °C exhibited the highest toughness (KIC) of ∼6.5 MPa m½ and flexural strength of ∼370 MPa (flexural modulus ∼150 GPa) while the flexural modulus was highest at ∼190 GPa for the composite processed at 1100 °C [98].

Similar methodologies (although the freezing procedure not used) are employed to prepare other types of ceramic composites, mainly graphene–ceramic composites [99101]. Graphene nanoplatelets were suspended with alumina in dimethyl formamide (DMF) by sonication and ball milling. The solvent DMF was removed in an oven at 90 °C for 3 days. The composite was then sintered with a uniaxial pressure of 50 MPa at 1500–1550 °C by the SPS. Graphene platelets overlapping and uniform distribution in alumina matrix were observed. The incorporation of graphene platelets at 0.38 vol% achieved the best improvement in flexural strength (from 400 to 523 MPa) and fracture toughness (3.53 to 4.49 MPa m½) but the hardness was slightly decreased from 18.06 to 17.66 GPa, compared to pure alumina [99]. For the graphene–Si3N4 composite, graphene and Si3N4 were separately suspended in water, stabilized by CTAB and then combined together and sonicated. The water was removed at ∼100 °C and the dry composite was heated in argon at ∼500 °C for 1 h to remove CTAB before being densified by the SPS process at 1650 °C with an applied load of 35 MPa. The presence of the graphene platelets was confirmed by Raman spectroscopy in the composite after the harsh treatment. When graphene platelets were included at 1.5 vol%, the fracture toughness was increased more than twofold (from 2.8 to 6.6 MPa m½) while the hardness was slightly decreased [100]. In another study, graphene oxide and ZrB2 (2 µm) were self‐assembled in water containing PVA to form layered structures. SiC was also added in order to enhance the mechanical property. The layered composite was collected by filtering the solution and then treated by the SPS at 1950 °C for 15 min with a uniaxial load of 30 MPa under high vacuum. The bioinspired graphene/ZrB2–SiC composite exhibited a flexural strength of 522 MPa, fracture toughness of 9.45 MPa m½, tensile strength of 72.1 MPa, and modulus of 12.5 GPa [101].

6.9 Tough and Functional Composites

Functional ceramics are required to fabricate materials for targeted applications. The incorporation of polymer into the ceramic scaffolds usually aims to improve pliability and mechanical properties. The composites may be readily fabricated by mixing ceramic particles and polymer in suspensions or in melts and then processed by solvent evaporation or hot molding techniques [122, 123]. However, most of the functional properties rely on the good contact between the ceramic particles. In a ceramic–polymer composite where the particles are randomly distributed in the polymer matrix, the desired property may be weakened as a result. The aligned porous or layered porous ceramic structures fabricated by ice templating can be highly effective to address this issue while improving the mechanical stability at the same time.

Aligned porous lead‐zirconate‐titanate (PZT) scaffolds were prepared by the freeze‐casting process. Epoxy was then filled into the aligned pores of the PZA scaffold and allowed to cure to generate PZT–epoxy composites [124]. The volume fraction of PZT in the composites varied from 36% to 69%. As shown in Figure 6.19, the longitudinal piezoelectric strain coefficient (d33) is about three times higher than the transverse piezoelectric strain coefficient (−d31) for different PZT fractions. This suggests good contact between PZT particles along the longitudinal (freezing) direction and hence a higher d33. Figure 6.19 also shows that d33 is the highest (∼450 pC/N) at a PZT fraction of 57%. This can be attributed to the dense and well‐defined aligned structure at this composition. Accordingly, the hydrostatic strain coefficient (dh = d33 + 2d31) can be calculated and shows the highest value of 184 pC/N at the PZT fraction of 57%. The mechanical property was evaluated by compression test. The PZT/epoxy composite exhibited an increase of twofold in compressive strength compared to the porous PZT scaffold [124].

Figure 6.19 PZT/epoxy composites show anisotropic piezoelectric strain coefficient along the longitudinal (freezing) direction (d33) and the transverse direction (d31).

Source: Xu and Wang 2014 [124]. Reprinted with permission from John Wiley and Sons.

The same idea can be used to generate excellent ion conductivity and thermal conductivity in the mechanically enhanced composites. Aqueous Li1+xAlxTi2−x(PO4)3 (LATP) nanoparticle (200–500 nm) slurry was frozen unidirectionally and freeze‐dried to form vertically aligned porous scaffolds [125]. After densification by sintering, solid electrolyte PEO(polyethylene oxide)/PEG(polyethylene glycol) was filled into the porous scaffold. An ion conductivity of 0.52 × 10−4 S cm−1 was achieved, 3.6 times that of the composite with randomly distributed LATP nanoparticles. The tension and compression moduli for the composites were 6.6 and 3.6 MPa, respectively, whilst they were 1.9 and 1.4 MPa for the PEO/PEG sample [125].

Aligned porous BN was prepared via directional freezing of aqueous BN (10 µm microplatelets) slurry with SCMC as stabilizer. In comparison, randomly porous BN was also fabricated by freezing slowly in a freezer. Epoxy/3D‐BN composites were subsequently produced by infiltrating epoxy resin with curing agent and catalyst into the porous scaffolds [126]. Epoxy/BN composites were also prepared by simply mixing BN with epoxy resin and curing. From Figure 6.20a, it can be seen that the thermal conductivity increases with the increase of BN loading in the composites. The epoxy/oriented 3D‐BN composite shows much higher thermal conductivity than the other two composites. The enhancement in thermal conductivity (measured along the freezing direction) compared to epoxy resin is given in Figure 6.20b. At a BN loading of 35 vol%, an enhancement of 2226 for the epoxy/oriented 3D‐BN conductivity (conductivity 4.42 W m−1 K−1) is obtained, nearly three times higher than that of the epoxy/random‐3D‐BN composite and four times higher than that of the epoxy/BN composite made by simple mixing [126]. This can be attributed to the alignment and self‐assembly of BN platelets during directional freezing, similar to that previously reported [97].

Figure 6.20 (a) The plot shows the relationship of thermal conductivity of the epoxy/BN composites with different BN loadings. (b) The thermal conductivity enhancement of different types of epoxy/BN composites based on the thermal conductivity of pure epoxy matrix.

Source: Hu et al. 2017 [126]. Reprinted with permission from American Chemical Society.

6.10 Summary

Ceramics usually have high strength, high thermal stability, and corrosion‐resistant property, and different functional properties depending on the type of ceramic. However, ceramics are usually brittle and this can be catastrophic for some applications and thus must be addressed. One of the main goals in the fabrication of ceramics composites is to improve fracture toughness without negatively impacting their strength significantly. There are other advantages in composites such as combined useful properties, newly generated properties, elasticity, etc.

Ceramic composites can be formed by combining ceramics with ductile phases such as polymer or metal and other ceramic particles (parameters including size, shape and aspect ratio). While the toughness can be enhanced by fabricating such composites, the strength is usually reduced. This is because strength and toughness are mutually exclusive and there is usually a trade‐off between strength and toughness when fabricating ceramic composites. Significant progress has been made recently in preparing strong and tough composites by learning from the wonderful natural materials, such as nacre, bone, and enamel where the hierarchical structure, assembly of inorganic platelets or columnar structure, and most importantly the interfacing between the inorganic phase and the organic phase are the main contributing factors. Different toughening mechanisms have been identified and used in designing and fabricating composite materials.

Nacre‐like composites, with high percentage of assembled inorganic platelets (∼95 vol%) and the nanoscale layer of organic component, have been mostly prepared and investigated. These usually involve the use of alumina platelets to form layered porous scaffolds by directional freezing (freeze casting). The second phase such as polymer or metal is infiltrated into the porous scaffolds. In order to increase the content of solid inorganic phase in the composites, the freeze‐dried scaffolds are often sintered and pressed uniaxially with loads of 30–50 MPa or even higher. When preparing ceramic–ceramic composites (so that they can be used for high‐temperature applications), smaller ceramic nanoparticles (such as silica, titania) can be added into the initial slurries. After a freeze‐casting process or other processes to assemble the platelets and remove the solvent, the resulting composites are usually treated by SPS at temperatures from 900 up to 1950 °C (depending on the type of ceramic) with applied load on the samples.

By learning from Nature, it has been possible to fabricate tough and strong composites. But there is still a long way to go because our fabrication methods employ very harsh conditions and the composites are still inferior if we use the same components from nature. It is thus still very important to understand the detailed structural arrangement and the toughening mechanism in natural materials. From the engineering and manufacturing point of view, one way to go forward is to increase further the percentage of inorganic components in the composites and decrease the thickness of organic component. For example, as the hardest part in the body, enamel contains 97 wt% of inorganic component. It is hugely challenging to fabricate composites with such a high inorganic content while maintaining the hierarchical and delicate organic/inorganic structure.

So far, the commonly used ceramics, such as alumina, HA, SiC, and BN, have been used to fabricate tough composites. The ductile phases have been polymer (PMMA, epoxy) and metal alloys. It is highly desirable to select suitable ceramic particles and polymers to produce functional and tough ceramic composites. Furthermore, the composites may be subjected to etching/micromolding/3D laser lithography to create structural composites with high strength, high toughness, but low density [127].


  1. 1 Rosso, M. (2006). Ceramic and metal matrix composites: routes and properties. J. Mater. Proc. Technol. 175: 364–375.
  2. 2 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.
  3. 3 Chevalier, J. and Gremillard, L. (2009). Ceramics for medical applications: a picture for the next 20 years. J. Eur. Ceram. Soc. 29: 1245–1255.
  4. 4 Boccardi, E., Ciraldo, F.E., and Boccassini, A.R. (2017). Bioactive glass‐ceramic scaffolds: processing and properties. MRS Bull. 42: 226–232.
  5. 5 Johnson, A.J.W. and Herschler, B.A. (2011). A review of the mechanical behaviour of CaP and CaP/polymer composites for applications in bone replacement and repair. Acta Biomater. 7: 16–30.
  6. 6 Macuvele, D.L.P., Nones, J., Matsinhe, J.V. et al. (2017). Advances in ultra high molecular weight polyethylene/hydroxyapatite composites for biomedical applications: a brief review. Mater. Sci. Eng. C 76: 1248–1262.
  7. 7 Koulouridis, S., Kiziltas, G., Zhou, Y. et al. (2006). Polymer–ceramic composites for microwave applications: fabrication and performance assessment. IEEE Trans. Microw. Theory Tech. 54: 4202–4208.
  8. 8 Nan, C., Bichurin, M.I., Dong, S. et al. (2008). Multiferroic magnetoelectric composites: historical perspective, status, and future directions. J. Appl. Phys. 103: 031101.
  9. 9 Sommers, A., Wang, Q., Han, X. et al. (2010). Ceramics and ceramic matrix composites for heat exchangers in advanced thermal systems – a review. Appl. Therm. Eng. 30: 1277–1291.
  10. 10 Padture, N.P., Gell, M., and Jordan, E.H. (2002). Thermal barrier coatings for gas‐turbine engine applications. Science 296: 280–284.
  11. 11 Voevodin, A.A. and Zabinski, J.S. (2005). Nanocomposites and nanostructured tribological materials for space applications. Compos. Sci. Technol. 65: 741–748.
  12. 12 Fu, S., Feng, X., Lauke, B., and Mai, Y. (2008). Effects of particle size, particle/matrix interface adhesion and particle loading on mechanical properties of particulate–polymer composites. Composites B 39: 933–961.
  13. 13 Becher, P.F. (1991). Microstructural design of toughened ceramics. J. Am. Ceram. Soc. 74: 255–269.
  14. 14 Palmero, P. (2015). Structural ceramic nanocomposites: a review of properties and powders' synthesis methods. Nanomaterials 5: 656–696.
  15. 15 Curtin, W.A. (1991). Theory of mechanical properties of ceramic–matrix composites. J. Am. Ceram. Soc. 74: 2837–2845.
  16. 16 Chartier, T., Merle, D., and Besson, J.L. (1995). Laminar ceramic composites. J. Eur. Ceram. Soc. 15: 101–107.
  17. 17 Chan, H.M. (1997). Layered ceramics: processing and mechanical behaviour. Annu. Rev. Mater. Sci. 27: 249–282.
  18. 18 Deng, X., Chawla, N., Chawla, K.K. et al. (2005). Mechanical behaviour of multi‐layered nanoscale metal–ceramic composites. Adv. Eng. Mater. 7: 1099–11108.
  19. 19 Hernandez, C.J. and van der Meulen, M.C.H. (2017). Understanding bone strength is not enough. J. Bone Miner. Res. 32: 1157–1162.
  20. 20 Ritchie, R.O. (2011). The conflicts between strength and toughness. Nat. Mater. 10: 817–822.
  21. 21 Wegst, U.G.K., Bai, H., Saiz, E. et al. (2015). Bioinspired structural materials. Nat. Mater. 14: 23–36.
  22. 22 Libonati, F. and Buehler, M.J. (2017). Advanced structural materials by bioinspiration. Adv. Eng. Mater. 19: 1600787.
  23. 23 Wang, J., Cheng, Q., and Tang, Z. (2012). Layered nanocomposites inspired by the structure and mechanical properties of nacre. Chem. Soc. Rev. 41: 1111–1129.
  24. 24 Porter, M.M., Mckittrick, J., and Meyers, M.A. (2013). Biomimetic materials by freeze casting. JOM 65: 720–727.
  25. 25 Zok, F.W. and Levi, C.G. (2001). Mechanical properties of porous matrix ceramic composites. Adv. Eng. Mater. 3: 15–23.
  26. 26 Wachtman, J.B., Cannon, W.R., and Matthewson, M.J. (2009). Stress and strain (Chapter 1). In: Mechanical Properties of Ceramics, 2nd ed, 1–26. John Wiley & Sons, Inc.
  27. 27 Czichos, H., Saito, T., and Smith, L. (2006). Mechanical properties (Chapter 7). In: Springer Handbook of Materials Measurement Methods, 283–298. Springer Science + Business Media, Inc.
  28. 28 Wachtman, J.B., Cannon, W.R., and Matthewson, M.J. (2009). Types of mechanical behaviour (Chapter 2). In: Mechanical Properties of Ceramics, 2nd ed, 27–34. John Wiley & Sons, Inc.
  29. 29 Wachtman, J.B., Cannon, W.R., and Matthewson, M.J. (2009). Measurements of elasticity, strength, and fracture toughness (Chapter 6). In: Mechanical Properties of Ceramics, 2nd ed, 89–118. John Wiley & Sons, Inc.
  30. 30 Wachtman, J.B., Cannon, W.R., and Matthewson, M.J. (2009). Hardness and wear (Chapter 22). In: Mechanical Properties of Ceramics, 2nd ed, 405–422. John Wiley & Sons, Inc.
  31. 31 Wachtman, J.B., Cannon, W.R., and Matthewson, M.J. (2009). Overview of toughening mechanisms in ceramics (Chapter 10). In: Mechanical Properties of Ceramics, 2nd ed, 189–198. John Wiley & Sons, Inc.
  32. 32 Xia, Z. and Li, L. (2014). Understanding interfaces and mechanical properties of ceramic matrix composites (Chapter 12). In: Advances in Ceramic Matrix Composites (ed. I.M. Low), 267–285. Woodhead Publishing Limited.
  33. 33 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.
  34. 34 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.
  35. 35 Zhang, H., Wang, D., Butler, R. et al. (2008). Formation and enhanced biocidal activity of water‐dispersable organic nanoparticles. Nat. Nanotechnol. 3: 506–511.
  36. 36 Ahmed, A., Hearn, J., Abdelmagid, W., and Zhang, H. (2012). Dual‐tuned drug release by nanofibrous scaffolds of chitosan and mesoporous silica microspheres. J. Mater. Chem. 22: 25027–25035.
  37. 37 Ahmed, A., Myers, P., and Zhang, H. (2012). Preparation of aligned porous silica monolithic capillary columns and their evaluation for HPLC. Anal. Methods 4: 3942–3947.
  38. 38 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.
  39. 39 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.
  40. 40 Qian, L. and Zhang, H. (2010). Green synthesis of chitosan‐based nanofibers and their applications. Green Chem. 12: 1207–1214.
  41. 41 Rezwan, K., Chen, Q.Z., Blaker, J.J., and Boccassini, A.R. (2006). Biodegradable and bioactive porous polymer/inorganic composite scaffolds for bone tissue engineering. Biomaterials 27: 3413–3431.
  42. 42 Bose, S. and Tarafder, S. (2012). Calcium phosphate ceramic systems in growth factor and drug delivery for bone tissue engineering: a review. Acta Biomater. 8: 1401–1421.
  43. 43 Schardosim, M., Soulié, J., Poquillon, D. et al. (2017). Freeze‐casting for PLGA/carbonated apatite composite scaffolds: structure and properties. Mater. Sci. Eng. C 77: 731–738.
  44. 44 He, F., Li, J., and Ye, J. (2013). Improvement of cell response of the poly(lactic‐co‐glycolic acid)/calcium phosphate cement composite scaffold with unidirectional pore structure by the surface immobilization of collagen via plasma treatment. Colloids Surf. B 103: 209–216.
  45. 45 Cai, X., Chen, L., Jiang, T. et al. (2011). Facile synthesis of anisotropic porous chitosan/hydroxyapatite scaffolds for bone tissue engineering. J. Mater. Chem. 21: 12015–12025.
  46. 46 Hunger, P.M., Donius, A.E., and Wegst, U.G.K. (2013). Platelets self‐assembly into porous nacre during freeze casting. J. Mech. Behav. Biomed. Mater. 19: 87–93.
  47. 47 Uddin, M.K. (2017). A review on the adsorption of heavy metals by clay minerals, with special focus on the past decade. Chem. Eng. J. 308: 438–462.
  48. 48 Bergaya, F. and Lagaly, G. (2006). General introductions: clays, clay minerals, and clay science (Chapter 1). In: Handbook of Clay Science, Development in Clay Science, vol. 1 (ed. F. Bergaya, B.K.G. Theng and G. Lagaly), 1–18. Elsevier Ltd.
  49. 49 Schoonheydt, R.A. and Johnston, C.T. (2006). Surface and interface chemistry of clay minerals (Chapter 3). In: Handbook of Clay Science, Development in Clay Science (ed. F. Bergaya, B.K.G. Theng and G. Lagaly), 87–113. Elsevier Ltd.
  50. 50 Ngah, W.S.W., Teong, L.C., and Hanafiah, M.A.K.M. (2011). Adsorption of dyes and heavy metal ions by chitosan composites: a review. Carbohydr. Polym. 83: 1446–1456.
  51. 51 Yu, W.H., Li, N., Tong, D.S. et al. (2013). Adsorption of proteins and nucleic acids on clay minerals and their interactions: a review. Appl. Clay Sci. 80‐81: 443–452.
  52. 52 de Sousa Rodrigues, L.A., Figueiras, A., Veiga, F. et al. (2013). The systems containing clays and clay minerals from modified drug release: a review. Colloids Surf. B 103: 642–651.
  53. 53 Maisanaba, S., Pichardo, S., Puerto, M. et al. (2015). Toxicological evaluation of clay minerals and derived nanocomposites: a review. Environ. Res. 138: 233–254.
  54. 54 Jordan, J., Jacob, K.I., Tannenbaum, R. et al. (2005). Experimental trends in polymer nanocomposites: a review. Mater. Sci. Eng. A 393: 1–11.
  55. 55 Madyan, O.A., Fan, M., Feo, L., and Hui, D. (2016). Enhancing mechanical properties of clay aerogel composites: an overview. Composites B 98: 314–329.
  56. 56 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.
  57. 57 Chen, H., Chiou, B., Wang, Y., and Schiraldi, D.A. (2013). Biodegradable pectin/clay aerogels. ACS Appl. Mater. Interfaces 5: 1715–1721.
  58. 58 Pojanavaraphan, T., Magaraphan, R., Chiou, B., and Schiraldi, D.A. (2010). Development of biodegradable foamlike materials based on casein and sodium montmorillonite clay. Biomacromolecules 11: 2640–2646.
  59. 59 Pojanavaraphan, T., Liu, L., Ceylan, D. et al. (2011). Solution cross‐linked natural rubber (NR)/clay aerogel composites. Macromolecules 44: 923–931.
  60. 60 Chen, H., Wang, Y., and Shiraldi, D.A. (2014). Preparation and flammability of poly(vinyl alcohol) composite aerogels. ACS Appl. Mater. Interfaces 6: 6790–6796.
  61. 61 Chen, H., Liu, B., Huang, W. et al. (2014). Fabrication and properties of irradiation‐cross‐linked poly(vinyl alcohol)/clay aerogel composites. ACS Appl. Mater. Interfaces 6: 16227–16236.
  62. 62 Wang, T., Sun, H., Long, J. et al. (2016). Biobased poly(furfuryl alcohol)/clay aerogel composite prepared by a freeze‐drying process. ACS Sustainable Chem. Eng. 4: 2601–2605.
  63. 63 Donius, A.E., Liu, A., Berglund, L.A., and Wegst, U.G.K. (2014). Superior mechanical performance of highly porous, anisotropic nanocellulose–montmorillonite aerogels prepared by freeze casting. J. Mech. Behav. Biomed. Mater. 37: 88–99.
  64. 64 Lichtner, A.Z., Jauffrès, 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.
  65. 65 Lichtner, A.Z., Jauffrès, D., Roussel, D. et al. (2015). Dispersion, connectivity and tortuosity of hierarchical porosity composite SOFC cathodes prepared by freeze‐casting. J. Eur. Ceram. Soc. 35: 585–595.
  66. 66 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.
  67. 67 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.
  68. 68 Liu, G. and Button, T.W. (2013). The effect of particle size in freeze casting of porous alumina–zirconia composite. Ceram. Int. 39: 8507–8512.
  69. 69 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.
  70. 70 Kim, K.H., Kim, D.H., Ryu, S.C. et al. (2016). Porous mullite/alumina‐layered composites with a graded porosity fabricated by camphene‐based freeze casting. J. Compos. Mater. . doi: 10.1177/0021998316636460.
  71. 71 He, R., Zhang, X., Han, W. et al. (2013). Effects of solids loading on microstructure and mechanical properties of HfB2–20 vol.% MoSi2 ultra high temperature ceramic composites through aqueous gelcasting route. Mater. Des. 47: 35–40.
  72. 72 Blindow, S., Pulkin, M., Koch, D. et al. (2009). Hydroxyapatite/SiO2 composites via freeze casting for bone tissue engineering. Adv. Eng. Mater. 11: 875–884.
  73. 73 Ghazanfari, S.M.H. and Zamanian, A. (2013). Phase transformation, microstructural and mechanical properties of hydroxyapatite/alumina nanocomposite scaffolds produced by freeze casting. Ceram. Int. 39: 9835–9844.
  74. 74 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.
  75. 75 Yang, H., Ye, F., Liu, Q. et al. (2015). A novel silica aerogel/porous Si3N4 composite prepared by freeze casting and sol–gel impregnation with high‐performance thermal insulation and wave‐transparent. Mater. Lett. 138: 135–138.
  76. 76 Han, D., Mei, H., Farhan, S. et al. (2017). Anisotropic compressive properties of porous CNT/SiC composites produced by direct matrix infiltration of CNT aerogel. J. Am. Ceram. Soc. 100: 2243–2252.
  77. 77 Meyers, M.A., McKittrick, J., and Chen, P. (2013). Structural biological materials: critical mechanics‐materials connections. Science 339: 773–779.
  78. 78 Cheng, Q., Jiang, L., and Tang, Z. (2014). Bioinspired layered materials with superior mechanical performance. Acc. Chem. Res. 47: 1256–1266.
  79. 79 Song, F., Zhou, J., Xu, X. et al. (2008). Effect of a negative Poisson ratio in the tension of ceramics. Phys. Rev. Lett. 100: 245502.
  80. 80 Mayer, G. (2005). Rigid biological systems as models for synthetic composites. Science 310: 1144–1147.
  81. 81 Deville, S., Saiz, E., Nalla, R.K., and Tomsia, A.P. (2006). Freezing as a path to build complex composites. Science 311: 515–518.
  82. 82 Bouville, F., Maire, E., and Deville, S. (2014). Self‐assembly of faceted particles triggered by a moving ice front. Langmuir 30: 8656–8663.
  83. 83 Roleček, J., Salamon, D., and Chlup, Z. (2017). Mechanical properties of hybrid composites prepared by ice‐templating of alumina. J. Eur. Ceram. Soc. 37: 4279–4286.
  84. 84 Munch, E., Launey, M.E., Alsem, D.H. et al. (2008). Tough, bio‐inspired hybrid materials. Science 322: 1516–1520.
  85. 85 Launey, M.E., Munch, E., Alsem, D.H. et al. (2009). Designing highly toughened hybrid composites through nature‐inspired hierarchical complexity. Acta Mater. 57: 2929–2932.
  86. 86 Zhao, H., Yue, Y., Guo, L. et al. (2016). Cloning Nacre's 3D interlocking skeleton in engineering composites to achieve exceptional mechanical properties. Adv. Mater. 28: 5099–5105.
  87. 87 Bai, H., Walsh, F., Gludovatz, B. et al. (2016). Bioinspired hydroxyapatite/poly(methyl methacrylate) composite with a nacre‐mimetic architecture by a bidirectional freezing method. Adv. Mater. 28: 50–56.
  88. 88 Mao, L., Gao, H., Yao, H. et al. (2016). Synthetic nacre by predesigned matrix‐directed mineralization. Science 354: 107–110.
  89. 89 Naleway, S.E., Yu, C.F., Hsiong, R.L. et al. (2016). Bioinspired intrinsic control of freeze cast composites: harnessing hydrophobic hydration and clathrate hydrates. Acta Mater. 114: 67–79.
  90. 90 Picot, O.T., Rocha, V.G., Ferraro, C. et al. (2016). Using grapheme networks to build bioinspired self‐monitoring ceramics. Nat. Commun. 8: 14425.
  91. 91 Veedu, V.O., Cao, A., Li, X. et al. (2006). Multifunctional composites using reinforced laminae with carbon‐nanotube forests. Nat. Mater. 5: 457–462.
  92. 92 Ferrand, H.L., Bouville, F., Niebel, T.P., and Studart, A.P. (2015). Magnetically assisted slip casting of bioinspired heterogeneous composites. Nat. Mater. 14: 1172–1181.
  93. 93 Porter, M.M., Meraz, L., Calderon, A. et al. (2015). Torsional properties of helix‐reinforced composites fabricated by magnetic freeze casting. Compos. Struct. 119: 174–184.
  94. 94 Launey, M.E., Munch, E., Alsem, D.H. et al. (2010). A novel biomimetic approach to the design of high‐performance ceramic‐metal composites. J. R. Soc. Interface 7: 741–753.
  95. 95 Liu, Q., Ye, F., Gao, Y. et al. (2014). Fabrication of a new SiC/2024Al co‐continuous composite with lamellar microstructure and high mechanical properties. J. Alloys Compd. 585: 146–153.
  96. 96 Shaga, A., Shen, P., Sun, C., and Jiang, Q. (2015). Lamellar‐interpenetrated Al‐Si‐Ag/SiC composites fabricated by freeze casting and pressureless infiltration. Mater. Sci. Eng. A 630: 78–84.
  97. 97 Bouville, F., Maire, E., Meille, S. et al. (2014). Strong, tough and stiff bioinspired ceramics from brittle constitutes. Nat. Mater. 13: 508–514.
  98. 98 Grossman, M., Bouville, F., Erni, F. et al. (2017). Mineral nano‐interconnectivity stiffens and toughens nacre‐like composite materials. Adv. Mater. 29: 1605039.
  99. 99 Liu, J., Yan, H., and Jiang, K. (2013). Mechanical properties of graphene platelet‐reinforced alumina ceramic composites. Ceram. Int. 39: 6215–6221.
  100. 100 Walker, L.S., Marotto, V.R., Rafiee, M.A. et al. (2011). Toughening in graphene ceramic composites. ACS Nano 5: 3182–3190.
  101. 101 An, Y., Han, J., Zhang, X. et al. (2016). Bioinspired high toughness graphene/ZrB2 hybrid composites with hierarchical architectures spanning several length scales. Carbon 107: 209–216.
  102. 102 Chen, R., Johnson, M.B., Plucknett, K.P., and White, M.A. (2012). Thermal conductivity of tunable lamellar aluminum oxide/polymethyl methacrylate hybrid composites. J. Mater. Res. 27: 1869–1876.
  103. 103 Flauder, S., Sajzew, R., and Müller, F.A. (2015). Mechanical properties of porous β‐tricalcium phosphate composites prepared by ice‐templating and poly(ε‐caprolactone) impregnation. ACS Appl. Mater. Interfaces 7: 845–851.
  104. 104 Borisova, A., De Bruyn, M., Budarin, V.L. et al. (2015). A sustainable freeze‐drying route to porous polysaccharides with tailored hierarchical meso‐ and macroporosity. Macromol. Rapid Commun. 36: 774–779.
  105. 105 Erb, R.M., Libanori, R., Rothfuchs, N., and Studart, A.R. (2012). Composites reinforced in three dimensions by using low magnetic fields. Science 335: 199–204.
  106. 106 Porter, M.M., Yeh, M., Strawson, J. et al. (2012). Magnetic freeze casting inspired by nature. Mater. Sci. Eng. A 556: 741–750.
  107. 107 Niebel, T.P., Bouville, F., Kokkinis, D., and Studart, A.R. (2016). Role of the polymer phase in the mechanics of nacre‐like composites. J. Mech. Phys. Solids 96: 133–146.
  108. 108 Roy, S. and Wanner, A. (2008). Metal/ceramic composites from freeze‐casting ceramic performs: domain structure and elastic properties. Compos. Sci. Technol. 68: 1136–1143.
  109. 109 Hautcoeur, D., Lorgouilloux, Y., Leriche, A. et al. (2016). Thermal conductivity of ceramic/metal composites from preforms produced by freeze casting. Ceram. Int. 42: 14077–14085.
  110. 110 Wang, Y., Shen, P., Guo, R. et al. (2017). Developing high toughness and strength Al/TiC composite using ice‐templating and pressure infiltration. Ceram. Int. 43: 3831–3838.
  111. 111 Zhang, H., Shen, P., Shaga, A. et al. (2016). Preparation of nacre‐like composites by reactive infiltration of a magnesium alloy into porous silicon carbide derived from ice template. Mater. Lett. 183: 299–302.
  112. 112 Guo, R., Shen, P., Lin, S. et al. (2017). High compressive strength in nacre‐inspired Al−7Si−5Cu/Al2O3–ZrO2 composites at room and elevated temperatures by regulating interfacial reaction. Ceram. Int. 43: 7369–7373.
  113. 113 Tamerler, C. and Sarikaya, M. (2008). Molecular biomimetics: genetic synthesis, assembly, and formation of materials using peptides. MRS Bull. 33: 505–512.
  114. 114 Palmer, L.C., Newcomb, C.J., Kaltz, S.R. et al. (2008). Biomimetic systems for hydroxyapatite mineralization inspired by bone and enamel. Chem. Rev. 108: 4754–4783.
  115. 115 He, L.H. and Swain, M.V. (2008). Understanding the mechanical behavior of human enamel from its structural and compositional characteristics. J. Mech. Behav. Biomed. Mater. 1: 18–29.
  116. 116 Fincham, A.G., Moradian‐Oldak, J., and Simmer, J.P. (1999). The structural biology of the developing dental enamel matrix. J. Struct. Biol. 126: 270–299.
  117. 117 Yeom, B., Sain, T., Lacevic, N. et al. (2017). Abiotic tooth enamel. Nature 543: 95–99.
  118. 118 Denry, I. and Holloways, J.A. (2010). Ceramics for dental applications: a review. Materials 3: 351–368.
  119. 119 Thompson, J.Y., Stoner, B.R., Piascik, J.R., and Smith, R. (2011). Adhesion/cementation to zirconia and other non‐silicate ceramics: where are we now? Dent. Mater. 27: 71–82.
  120. 120 Pieralli, S., Kohal, R.J., Jung, R.E. et al. (2017). Clinical outcomes of zirconia dental implants: a systematic review. J. Dent. Res. 96: 38–46.
  121. 121 Petrini, M., Ferrante, M., and Su, B. (2013). Fabrication and characterization of biomimetic ceramic/polymer composite materials for dental restoration. Dent. Mater. 29: 375–381.
  122. 122 Dietze, M., Krause, J., Solterbeck, C.‐H., and Es‐Souni, M. (2007). Thick film polymer–ceramic composites for pyroelectric applications. J. Appl. Phys. 101: 054113.
  123. 123 George, S., Anjana, P.S., Sebastian, M.T. et al. (2010). Dielectric, mechanical, and thermal properties of low‐permittivity polymer–ceramic composites for microelectronic applications. Int. J. Appl. Ceram. Technol. 7: 461–474.
  124. 124 Xu, T. and Wang, C.‐A. (2014). Piezoelectric properties of a pioneering 3‐1 type PZT/epoxy composites based on freeze‐casting processing. J. Am. Ceram. Soc. 97: 1511–1516.
  125. 125 Zhai, H., Xu, P., Ning, M. et al. (2017). A flexible solid composite electrolyte with vertically aligned and connected ion‐conductivity nanoparticles for lithium batteries. Nano Lett. 17: 3182–3187.
  126. 126 Hu, J., Huang, Y., Yao, Y. et al. (2017). Polymer composite with improved thermal conductivity by constructing a hierarchically ordered three‐dimensional interconnected network of BN. ACS Appl. Mater. Interfaces 9: 13544–13553.
  127. 127 Bauer, J., Hengsbach, S., Tesari, I. et al. (2014). High‐strength cellular ceramic composites with 3D microarchitecture. Proc. Natl. Acad. Sci. 111: 2453–2458.