MICROSCOPY RESEARCH AND TECHNIQUE 27:360-375 (1994)

An Introduction to Biomimetics: A Structural Viewpoint

KEY WORDS

Biological composites, Structural biocomposites, Microarchitecture, Materials design

ABSTRACT

Biomimetics is a newly emerging interdisciplinary field in materials science and engineering and biology in which lessons learned from biology form the basis for novel technological materials. It involves investigation of both structures and physical functions of biological composites of engineering interest with the goal of designing and synthesizing new and improved materials. This paper discusses microarchitectural aspects of some structural biocomposites, presents microstructural criteria for future materials design and processing, and identifies areas of future research. o 1994 WiIey-Liss, Inc. produce soft and hard materials which have properties far beyond those can be achieved in present technological materials. Biological composites are complex in terms of composition and microstructure, but yet highly ordered, containing both inorganic and organic components in an intricate blend (Lowenstam and Weiner, 1989; Simkiss and Wilbur, 1989). Physical properties include optical, magnetic, electronic, and piezoelectric, as well as mechanical. These materials are synthesized at ambient temperatures under atmospheric conditions with the structural build-up completely controlled by the organisms. The unique microstructures and resulting properties in biological composites have been unknown to materials scientists, and may well be inspirational sources in the development of future materials. With this in mind, this research group started investigating structures of biological composites to establish relationships with properties and to obtain processing strategies and microstructural design criteria for the development of new materials with novel properties, a field which is now called biomimetics. (see, for instance, Aksay et al., 1992; Alper et al., 1991; Rieke et al., 1990). Biomimetics may be one of the major ways to produce next generation materials that would meet demands of the technologies of the coming century. We divide biomimetics into two categories (Fig. 1; Sarikaya and Aksay, 1993a). First, by investigating the structures of biomaterials at all possible scales of spatial resolution, the fundamentals of their unique structural designs can be deduced and then mimicked by techniques currently available to materials scientists-an approach we refer to as biomimicking. The second category is mastering the molecular synthesis and processing mechanisms of biomaterials and applying these hitherto unknown methodologies to produce new technological materials superior to those presently

INTRODUCTION Properties of materials are structure sensitive and the basis of materials design and processing, therefore, is to control microstructures so as to achieve desired physical and chemical properties in materials (see, for instance, Baer and Moet, 1991, Fulrath and Pask, 1966, Haasen, 1978). In one of the most widely used structural materials, for example, in low-alloy low-carbon structural steels, a myriad of different microstructures from pearlitic to martensitic are produced through thermal and mechanical treatments that results in countless different types of steel used in many different applications (see, for instance, Zackay and Aaronson, 1962).In a more recent example of high temperature superconductorYBa2Cu307.x, the significance of effects of microstructural features, such as structural characteristics of transformation twins, amounts, types, and distribution of local oxygen ordering, and interface structures and chemistry on the superconducting properties, have been well realized (see, for instance, Jorgensen, et al., 1989). Despite the enormous technological possibilities, however, unlike in steels, some major difficulties are encountered in high temperature superconductors in controlling the structural features during processing. As a consequence, these new materials have been used only in constrained structural forms (like thin films) and, therefore, have so far had limited applications. In these and many other technological materials systems microstructures consist mostly of metastable phases with morphologies that are difficult to control. In cases where nucleation and growth of the secondary phases (such as precipitates) are controlled at the atomic and molecular levels under stringent synthesis conditions, the microstructures are tailored to a large extend to produce certain desired properties. Examples are traditional alloy systems (steels, Al- and Cu-alloys, superalloys, and cemented carbides; Gel1 et al., 1980; Zackay and Aaronson, 1962) and in more recent cases of nanostructral materials (Shinjo and Tanaka, 1987) and in compound semiconductor heterostructures (see, for instance, Shaw et al., 1989). In biological systems, on the other hand, organisms

Received November 15, 1992; accepted in revised form March 15, 1993. Address reprint requests to Dr. Mehmet Sarikaya, Materials Science and Engineering, Roberts Hall, FB-10, University of Washington, Seattle, WA 98195.

0 1994 WILEY-LISS,

INC.

INTRODUCTION TO BIOMIMETICS

361

Biomimetics

neering). Commensurate with the objectives of the symposium and this special issue of the journal, this paper focuses only on the microstructural aspects of biocomposite hard and stiff tissues. The purpose is to bring about various fascinating and unique structural design in biocomposites with the aim of raising interest in microscopists-materials and biological scientists alike-who would be involved in this field in the future.

STRUCTURAL DESIGN IN SOME BIOLOGICAL HARD TISSUES Biological materials of engineering interest are Processing mostly hard and stiff tissues, and small particles (Table 1). In the former case, the structure is composed of an organic matrix (mostly proteins and polysaccharides) with the inorganic material interspersed throughout. New Materials with The formation, morphology, and crystallography of the Tailored Structures inorganic phase are all assumed to be controlled by the organic matrix. In hard tissues such as mollusk shells, the inorganic phase, for example calcite or aragonite Fig. 1. The description of interdisciplinary field of biomimetics. form of CaCO,, is in the form of crystallites of various shapes, size, and morphology that are controlled by the organic matrix which constitutes only less than 5% vol. available-an approach we call bioduplication. The of the composite. In the case of stiff organic tissues, bioduplication approach is much more complex and such as insect cuticles (Bouligand, 19651, all the comwill require a long-term commitment, not only to learn ponents of the composite are organic macromolecules the intricacies of bioprocessing used by organisms but in which matrix is usually composed of proteins and also to develop new strategies to process materials syn- polysacharides. The stiffness comes from fibrillar thetically from the molecular level up with the same polysaccharides, such as collagen and chitin, organized size, shape, complexity, and multifunctionality as the at the nano-, submicro-, micro- or higher scales (Baer et biocomposites. The biomimicking approach, although al., 1992). Finally, some lower organisms, such as bacby no means simple, will require a shorter time com- teria and alga produce fine inorganic particles with unique properties (Frankel and Blakemore, 1991). mitment. These biological materials are listed in Table 1 with One of the most interesting aspects of biological matheir corresponding properties, and will be discussed in terials is that their metastable microstructure is formed from the molecular to macroscales t o attain a some detail in the following examples with emphasis combination of certain physical properties that the or- on the mollusk shells which have relatively simple miganism best functions within its given environment. A crostructures. particular microstructure in an organism is developed Biogenic Small Inorganic Particles through millions of years of evolution. Biocomposites are structured in a hierarchical manner (Baer et al., From the point of obtaining lessons on characteristic 19921, from the molecular to the macro scale, they are materials structures and properties, small inorganic intricately microarchitectured, and they often have particles of biological origin offer analogies with synmultifunctional properties (Currey, 1987; Wainwright thetic nanoparticles and mesoscopic systems. There are et al., 1976). These materials are composites contain- many organisms that produce ultrafine inorganic paring various structural organic macromolecules (pro- ticles in their bodies that perform various functions teins and polysaccharides) and inorganic materials (of- (Frankel and Blakemore, 1991). One most notable exten ceramic crystals or glasses) (Currey, 1987; Frankel ample is iron clusters that form at the center of ferritin and Blakemore, 1991). They can be found from the molecular cages (or vesicles) in organisms (Harrison et most primitive organisms (such as small particles and al., 1989). In some cases, metals clusters are accumuthin films in algae and bacteria) (Frankel and Blake- lated as foreign entities that might otherwise be harmmore, 1991)to invertebrates (shells and cuticles in mol- ful to the host organism, such as CdS in algae (which, lusks [Gregoire, 19721 and insects [Bouligand, 19651, however, have excellent optical properties; Dameron et respectively), and to the highly developed vertebrates al., 1989). Another example is ultrafine magnetic par(such as bones and teeth of mammals) (Glimcher, ticles that are found in bacteria (Blakemore, 1982), as 1981). further discussed below. Biomimetics is a wide area of research and, thereSome species of bacteria that live anaerobically in fore, it requires a close collaboration between biological freshwater and salt swamps move about to seek oxygen (biochemistry, genetics, microbiology, and zoology) and and food by a mechanism that makes use of a string of physical sciences fields (materials sciences, physics, magnetic particles (Fe304 or Fe,S,) as a compass chemistry, chemical, electrical, and mechanical engi- (Frankel and Blakemore, 1991). One of the first puri-

0
Novel Synthesis

362

M.SARIKAYA
TABLE I . A chart showing categorization of various biological composites, their micro-inano-design and physical properties
Materialicomuosite Small particles Ceramiciceramic Ceramicipolymer Polymer/polymer Laminated Fiberimatrix Fiberifiber Liquid crystalline/matrix Examule Bacterial algal Sea-urchin Mollusk Bone dentin Cuticle Tendon Silk Mocus Micro- or nano-level Properties Magnetic Electronic, optical Mechanical (wear resistant) Mechanical (tough, strong), ferroelastic Ferroelastic Ferroelastic Mechanical, ferroelastic, optical Mechanical, optical Mechanical, ferroelastic Mechanical (tensile props.) Rheological

N N
B

N, H N, H N, H N, H N, H N, H N N

N: nano, M: micro, B: both M and N, H: hierarchical

fied strains, in Aquaspirillum magnetotacticum (Blakemore, 1982), for example, each bacterium has about 20-25 particles that are oriented along a string with their magnetization axis along the long axis of the bacterium (Fig. 1).These bacteria have two flagella on each end, and therefore, can move in forward or backward directions, depending on their configuration with respect to the Earths magnetic field. In terms of biomimetic applications, some of the significant materials characteristics of these particles are listed as: (i) they are single crystalline, having no dislocations, twins, or stacking faults; (ii) have a uniform particle size of about 500-600 A, and, thus, are in the single domain region (superparamagnetic); (iii) particle shape is species specific, and can be dodecahedral, cubo-octahedral or hexagonal; (iv) they are aligned in the form of a single string (in some isolated cases, as double strings); (v) they form in a biological sacks, called magnetosomes. In S-rich regions, some species are known to form isomorphic form of magnetite, i.e. Fe,S, (Bazylinski et al., 1991; also see Bazylinski et al., this issue). The formation of magnetic particles within magnetosome membranes is of a great interest in terms of forming small synthetic particles under closely controlled synthesis conditions (Fendler, 1982; also see Yang et al., this issue). Small magnetic particles can be formed synthetically by following several routes, e.g., solution precipitation from precursors, in microemulsions, and using vesicles. In all these cases, however, the particles formed are nonuniform, they are often not fully crystalline, compositionally nonhomogeneous, and, more importantly, in an agglomerated state which imposes problems in processing. Therefore, synthesis of small magnetic, and other inorganic ultrafine, particles via biological routes promises advantages in terms of controlling growth and morphological properties. The most critical issue in the understanding of particle formation in magnetosomes is the mechanism(s) by which organic matrix allows the nucleation and controls the growth of particles. The investigation of structure and composition of the magnetosome membrane and its protein organization, therefore, is essential for understanding transport of ions through the membrane and the early stages of formation or particles. Current knowledge of the membrane is that it is a bi-

layer and is likely to contain proteins that are found in the outer membrane of bacteria (Gorby et al., 1988). Other questions involve details of the early stage of formation of particles (amorphous or crystalline and, possibly, in the hydrated form), selection of their chemistry (Fe30, vs. Fe,S,), the control of their growth, and finally factors that affect particle morphology and size (Mann et al., 1990).

CeramicXeramic Biological Composites In biological composite materials involving inorganic phases, it is usually assumed that organic macromolecules are associated with the composite in a settled way, such as described under the heading Ceramic/ Organic Composites. However, there are also many cases in which mineral appears to be present alone in the structure and, therefore, constitutes the overall sample. For example, in the body and the spine (Fig. 2) of sea-urchin, the mineral which does not appear to contain any organic matrix is a calcitic single crystal (Berman et al., 1988; Veis et al., 1986). The most interesting among the composite structures in sea urchin, however, is its teeth (Brear and Curry, 1976). There are five pieces in the lower center of the body that the organism uses to scrape food from surface of rocks. Cross section at the cutting edge of a tooth exposes a composite of a matrix of amorphous CaC0, with crystalline calcitic CaCO, fibers embedded in it with their long axes perpendicular to the cutting surface as to increase the wear resistance of the tooth (Fig. 3; Sarikaya et al., 1992a). In these biomaterials, location and distribution of an organic matrix, which presumably controls calcite growth, are not clearly known. In regards to biomimetics, therefore, the major questions about sea-urchin skeletal units are the presence, types, and spatial distribution of organic macromolecules. These, constituting less than 1%of the composite, do not appear to form spatially discernible separate phase(s1and which are, then, most likely to be occluded within the inorganic matrix (Berman et al., 1990). If there are proteins, or their fractions, are occluded within the matrix, then these hard tissues may be regarded as molecular composites, i.e., analog of nanocomposites at smaller spatial dimension. Investigation of various skeletal units of sea-urchin in detail by mi-

INTRODUCTION TO BIOMIMETICS

363

Fig. 2. (a,b)TEM images of A. magnetotacticum show string of fine Fe,O, particles (a, arrow).

croscopy and spectroscopy,therefore, is expected to provide better insights into the understanding of these unique structures and new lessons for biomimetics.

Organidorganic Biological Composites There are numerous stiff biological tissues, composites of fibrous organic components embedded in a soft organic matrix, that are analogs of fiber- or particlereinforced polymeric composites (Baer, 1986). Tendon, which connects muscle and bone, is a classical example (Baer et al., 1988). It has six discrete levels of structures organized in a hierarchical manner from molecular to centimeter-scale. Silk, found in cocoons of silk moths and webs of spiders, is another structural ma-

terial (Gosline et al., 1986). Its unique structure, consisting of silk fibroin proteins @-pleated sheets) organized in a liquid crystalline fashion in an amorphous protein matrix, is designed to withstand stresses (Gosline et al., 1986) much higher than those encountered by high tensile strength metallic or polymeric fibers (Baer, 1986). One of the major classes of organic biocomposites is formed by cuticles. Arthropods, such as insects, crustaceans, spiders, millipedes, and others, are evolutionarily very successful probably because of their cuticles that cover their body from head to toe (Neville, 1975). The cuticle, therefore, is the skeleton (exoskeleton) of an insect. Its structure resembles that of fiber-rein-

364

M. SARIKAYA

Fig. 4. (a,b) SEM images of an insect cuticle showing helicoidal organization of layers in a ply-wood structure (secondary electron image).

Fig. 3. (a,b) SEM images of a sea-urchin spine (single crystal) showing its intricate structure in a fractured sample. (c) Sea urchin tooth is a composite of crystalline calcite fibers in a n amorphous matrix (etched; secondary electron image).

forced polymer matrix where the fibers are collagen (polysaccharides), and the matrix is mostly proteins (see, for instance, Giraud-Guille, this issue). The composite has a sheet structure in which collagen fibers are arranged in layers. In each layer, the

fibers are oriented parallel to each other. In successive layers, however, there is rotation of parallel fibers only few degrees, i.e., helicoidal. The thickness of each layer may be as small as 5-10 nm and the thickness of the chitin fibrils as thin as 3-5 nm in diameter (Neville, 1975). In Odontotaenius disfunctus, shown in Figure 4 at a higher scale of microstructure, there are two simultaneous clockwise rotations between two sets of sheets (double-helicoid). Therefore, although in each layer the fibers are isotropic, the overall composite structure is anisotropic in the cuticle. This has a significance for the composite material from its mechanics point of view (Gunderson and Schiavone, 1989). The unique microstructure of the insect cuticle, as well as cuticles of other classes, may serve as a lesson for the design of composite materials in which all the components are polymers. Exoskeletons, in addition to serving as light, versatile, and protective armor for insects, also form intricate surface structures which give optical effects (Ghiradella et al., 1972). For example, in butterfly wings, although the origin of color for the most part is pigmentary, some colors such as blues and violets, come from surface structures (Ghiradella et al., 1972; Ghiradella, this issue). The structural details on surfaces can take an intricate and ordered combination of

INTRODUCTION TO BIOMIMETICS

365

layers, scales, and ridges which are arranged to produce optical effects through interference during scattering of light. Coloring through structural effects is an area of study in biomimetics. Furthermore, pigmentation itself is also an area of materials science interest, involving questions such as the origin of pigmentation, the nature of pigments, their size, distribution, and composition within the matrix, and their coloring (possibly, non-linear optical) effects. In cuticles, in summary, there are both biomimicking and bioduplication possibilities for future materials formation. On the biomimicking side, questions involve the types of structural units at the nano- and micro-scales, and their spatial distribution. There are already synthetic polymers that may be used as the matrix (such as diblock copolymers and liquid crystalline polymers; Spontak et al., this issue) as well as fibrous polymers that can be used as the stiff component. On the bioduplication side, some subjects of future investigation are composition and spatial distribution of proteins in the matrix, their possible liquid crystalline order, and their synthesis. The nature of polyscahharides, its molecular structure and composition (including collagen), its structural relationship with the protein matrix, and overall hierarchy of the structure of cuticle are further areas of biomimetic interest.

CeramidOrganic Composites The examples of biological composites containing both ceramic and organic phase(s) in the form of macromolecules, take many different forms (Currey, 1987). They are mainly used as structural materials in skeletons or protective covers for the bodies of organisms (Lowenstam and Weiner, 1989; Simkiss and Wilbur, 1989). These include bones in vertebrates, teeth in fishes and mammals, and shells in mollusks. A representative example of the inner section of many seashells, namely nacre, is discussed below as this hard tissue has been a major area of investigation in the authors laboratory (Sarikaya and Aksay, 1992). Mechanical Properties ofNacre. The nacre structure, mother-of-pearl, is found in many families of mollusks, such as red abalone (Haliotis rufescens; Currey, 1987; Jackson et al., 1988; Sarikaya et al., 1990; Sarikaya and Aksay, 1992; Yasrebi et al., 1990), the gastropod family, cephalopods, such as nautilus (Nautilus pompilius; Gregoire, 1972) and bivalves, such as black-lipped pearl oysters (Pinctada margaritifera; Currey, 1987). A transverse cross-section of the red abalone shell displays two types of microstructures: an outer prismatic layer (calcite) and inner nacreous layer (aragonite). The structure and properties of the nacreous layer are described here as this is the part of the shell that displays an excellent combination of mechanical properties as a result of its highly ordered hierarchical structure (Currey, 1987; Jackson et al., 1988; Sarikaya et al., 1990; Sarikaya and Aksay, 1992; Yasrebi et al., 1990). As shown in Figure 5, nacre is composed of stacked platelets (0.2-0.5 pm thick) that are arranged in brick and mortar microarchitecture with an organic matrix (20-400 nm in thickness) form-

ing a gluebetween the platelets (Jackson et al., 1988; Sarikaya et al., 1990; Yasrebi et al., 1990). Fracture toughness (KIc) and fracture strength (uF) of nacreftested in the transverse direction)(Sarikaya et al., 1990; Yasrebi et al., 1990) and those of some of the well-known ceramics and ceramic-based composites (cermets) are plotted in Figure 6. The average K,, and uF values of nacre are some 20-30 times that of synthetically produced monolithic CaCO, (Jackson et al., 1988; Sarikaya et al., 1990; Yasrebi et al., 1990). This result is a major driving force in producing ceramicbased composites (cermets and cerpolys) with better mechanical properties than existing composites through nanoscale lamination based on lessons from biology (Jackson et al., 1988; Sarikaya et al., 1990; Sarikaya and Aksay, 1992; Yasrebi et al., 1990). The investigation on crack propagation behavior in nacre (Sarikaya et al., 1990; Yasrebi et al., 1990) reveals that there is a high degree of tortuosity not seen in the more traditional brittle ceramics, such as Al,03, or in high toughness ceramics, such as ZrO, (see, for example, Dawridge, 1979 and Evans, 1988). The surfaces of fractured samples indicate that a major crack has meandered around the CaCO, layers exposing them through the organic surroundings, resulting in a highly rough fractured surface (Fig. 7). This is similar to that seen in fiber-reinforced ceramic composites where a pull-out mechanism operates (see, for example, Evans and Marshall, 1989, and Tressler and Bradt, 1983). In micrographs recorded a t higher magnifications, either a sliding of the CaCO, layers (Fig. 8a) or organic ligaments between the layers (Fig. 8b) are seen. The latter case, stretching, indicates that the interface between the organic and the inorganic phases is strong and that the organic phase acts as a strong binder (Sarikaya et al., 1990; Sarikaya and Aksay, 1992; Yasrebi et al., 1990). Several toughening mechanisms, therefore, may be proposed (Sarikaya et al., 1990): (i) crack blunting/branching, (ii) microcrack formation, (iii) plate pull-out, (iv) crack bridging (ligament formation), and (v) sliding of CaC03 layers. The high degree of tortuosity seen in crack propagation (Jackson et al., 1988) may be due mainly to crack blunting and branching (Sarikaya et al., 1990; Yasrebi et al., 1990). However, tortuosity (about 40-50%) alone is not a major toughening mechanism in these composites, because it cannot account for the many orders of magnitude increase in toughness. The major toughening mechanisms, therefore, are sliding and ligament formation (Sarikaya et al., 1990; Sarikaya and Aksay, 1992; Yasrebi et al., 1990). Similarities exist between the deformation of a nacreous portion of a sea shell and a metal in the sliding mechanism, and ligament formation in nacre is similar t o bridging by metal in ceramic-metal composites during crack propagation. The results strongly suggest that these complex deformation modes may be the main mechanisms of energy absorption during the propagation of a stable crack. The strength of nacre, on the other hand, may be related to several factors, including the size and structure of the aragonite platelets and the interfaces between the inorganic and the organic components. From the limited thickness of the largest flaw (Griffith,

366

M. SARIKAYA

Fig. 5. TEM image of nacre in an edge-on configuration displays aragonite platelets separated by a thin film of organic matrix. The inset shows the brick and mortar microarchitecture.

INTRODUCTION TO BIOMIMETICS

367

100

200

300

400

Specific Flexural Strength [MPa/(g/cm3>]

19911) based on biomimetic architecture, these have not been as extraordinary as when nacre is compared to monolithic CaCO,. This may be due to limited laminate thicknesses in synthetic composites; thicknesses below 1 pm in the inorganic layers and below 100 nm in the layers of the soft phase, are needed. Secondly, both the inorganic and organic layers have complex structures, in terms of their crystallography, substructures, and morphology. In particular, the organic layer has a complex nanolaminated structure within itself (Watabe, 1965; Weiner and Traub, 1984; Weiner, et al., 1983). Neither the composition of these layers nor their structure have yet been clearly identified. Furthermore, identification of the structural relationship between the organic and the inorganic layers is far from complete and constitutes one the most outstanding problems in biomineralization. Detailed Microstructure of Nacre. It has been impossible to study both the organic and inorganic crystals simultaneously, and the structural relationships between the components of the nacre exist only as Fig. 7. Highly tortuous fracture surface of nacre exposing aragonitic platelets. It is found that the tortuosity is not the main tough- a conjecture (Addadi and Weiner, 1990; Mann, 1988). Bulk studies performed by X-ray diffraction on the naening mechanism. cre revealed that the aragonite platelets are organized with their [OOl] axis perpendicular to the layers (Add1925), i.e., the thickness of the platelet, 0.5 pm, the adi and Weiner, 1990; Mann, 1988). It has been postuincrease in the fracture strength of aragonite would be lated that a and b axes within the layer plane in each about 200 MPa, comparable t o the measured value of platelet are oriented randomly. Furthermore, it was 185-220 MPa in our studies (Sarikaya et al., 1990; assumed from this scheme that each aragonite platelet Sarikaya and Aksay, 1992; Yasrebi et al., 1990).There- grew on the crystallographically related organic temfore, the rule of mixtures (Hull, 1981) may account for plate, which itself had a local random orientation. the value of the measured strength (Sarikaya and Ak- From the composition of the insoluble fraction of the organic matrix, i.e., the inner crystalline sublayers say, 1992). Although improvements were achieved in the me- which contain a high fraction of aspartic and glutamic chanical properties of synthetic laminated composites acid, it might be possible to deduce a self-assembled (cermets [Yasrebi et al., 19901 and cerpolys [Khanuja, structure that is related epitaxially to the aragonite

368

M. SARIKAYA

Fig. 8. SEM images revealing sliding of aragonite platelets (a) and ligament formation of organic phase (b),major toughening mechanisms. Arrows in (a) show direction of sliding, and in (b) they show the organic ligaments formed due to normal stresses.

lattice along the [OOl] projection (Sarikaya and Aksay, 1992; Sarikaya et al., 1992b). In this investigation, each aragonite crystallite was analyzed separately and its crystallographic orientation relationship with respect to its neighbors, both on the same layer and across the thickness of the nacre, was established. Electron microdiffraction with an electron probe size as small as 50 nm diameter in a TEM allowed isolated diffraction patterns from individual aragonite crystal units enabling us to establish the overall crystallography of the inorganic component of the composite in three directions. It was found that adjacent platelets belong to the same [OOlI zone axis, but there is a slight rotation among the platelets about this axis with respect to each other. The question remains whether there is any crystallographic relationship between the a and b axes in platelets on the same layer. The crystallographic relationship among the adjacent platelets in the face-on configuration analyzed by electron diffraction revealed that each platelet is actually twin-related to the one next to it with a twin plane of (110) type of the orthorhombic unit cell. In this scheme, the arrangement of the platelets indicates that all the platelets on the same layer are twin-related whether they share a boundary or not, constituting first generation twins since this twinning takes place a t the largest spatial scale (Fig. 9 a,b). Further analysis indicates that each platelet consists of several domains which are crystallographically coupled (Fig. 9 c,d). Hence, again, diffrac-

tion patterns reveal two superimposed patterns that can be correlated with a twin relationship with (110) twin plane parallel t_o [OOl] direction of the unit cell, i.e., either (110) or (110)variants (Figure 9 e,f). In fact, the patterns recorded from all the domain boundaries show the same twin reflections, indicating that each domain is related to the one next to it by a (110) twin relation. Domains, therefore, constitute the second generation twins. A platelet may have either four, go"-, or six, 60"-, domains. In an ideal hexagonal shaped platelet with six twin-related domains, the angle between each pair of domains would be 60". This is not possible, however, in aragonite lattice, since the outer edges of the platelets are parallel to (110) planes and the angle between each pair of planes-(110) and (110), for example-is 63.5") leaving unaccounted for a 3.5".This discrepancy induces strain during growth into the aragonite matrix and must be accommodated by some structural deformation, such as, in this case, nanometer-scale twins that form on (110) planes. These ultrafine twins are shown in Figure 10(a) formed on (110) planes a t an angle of about 63.5"and are similar to the growth twins in geological minerals. These twins constitute the third generation twins since they occur at the smallest scale. It was found that a portion of the lattice stress created by the 3.5"-strain can also be accommodated by the misalignment of adjacent domains, as observed. However, this misalignment cannot account for all the strain accommodation, as the interfaces between the

Fig. 9. TEM images of aragonite platelets viewed in face-on orientation before (a)and after (b) slight tilting display twin-contrast. Letters A,B,C,D,E, and F indicate aragonite platelets. Similarly, twinning also occurs between domains in a platelet (c-d). Electron diffraction patterns ( e )from the interior of a domain, and (0 from the domain boundary. Arrows in (c) indicate location from which SAD patterns (el and (f) were obtained.

370

M. SARIKAYA

Fig. 10. (a) TEM-BF image of the nanoscale twins within an aragonite domain (Sarikaya and Aksay, 1992). (b)Atomic resolution image of a low angle boundary within a domain, showing that stresses are also accommodated by dislocations (arrows).

domains show a high degree of coherency. Local deformation within domains, such as low angle boundaries, as shown in Figure 10(b), may account for the accommodation of some of the internal stresses. Consequently, these three twin structures cover six orders of magnitude size scale from nanometer to submillimeter, and reveal for the first time a hierarchical structure in a biological hard tissue. The geometrical and crystallographic model of aragonite platelets (Sarikaya and Aksay, 1992; Sarikaya et al., 199213) discussed above is referred to as multiple tiling in mathematics (Grunbaum and Shephard, 1987). It appears that nature utilized this mathematical technique in nacre to form a highly ordered structure that is compatible with both the soft tissue and the crystalline structural constraints of the hard component. Furthermore, recent studies indicate that tiling may also play an important role in providing the overall shape of the nacre and its properties (Sarikaya and Aksay, 1992; Sarikaya et al., 1992b). The constrains that are developed as a result of the commensurate interface between the crystalline conformation of the

organic matrix proteins and crystalline ordering of the ions in the inorganic phase give rise to the hierarchical twin structure in the nacreous section of the mollusk shells. That is, the coupling between organic and inorganic crystalline structures results in a certain morphology of the inorganic phase, retaining the hierarchy of the defect structure in the CaCO, phase. Based on this crystallographic relationship, furthermore, even the growth pattern and shape of the overall shells may be described. There are significant implications of this result in biomimetic design of future materials. If the inorganic crystal units grow under the close scrutiny of the organic matrix, then in producing inorganic materials (thin films, small particles, bulk or laminated composites) all the important structural features of materials, such as shape, size, crystallography, and morphology, may be predicted based on the structural coupling between the organic template used and the knowledge of crystal structure of the desired inorganic phase. Future studies will explore these possibilities both studying the biological composites, such as nacre, and other sections of mollusks and echinoderms, and by

INTRODUCTION TO BIOMIMETICS

37 1

investigating biomineralization in model synthetic systems using engineered macromolecular assemblies. Interface Between Organic-Inorganic Phases in Nacre. As described by others (Nakahara et al., 19821, the organic matrix is thought to have a sandwich form, with chitin in the middle (as the framework macromolecules), acidic macromolecules surrounding it (as nucleator macromolecules; Sarikaya and Aksay, 1993a,b), and soluble macromolecules forming the outer layers adjacent to the platelets. These models are based mostly on indirect investigation of the characteristics of the biological macromolecules and extracted from the shell after decalcification. Macromolecules are divided into two groups: as those soluble in weak acids, and those that are not. In the former case, the proteins extracted are found to contain aspardic and glutamic acids, known to be major components of structural proteins that form sheet-like structures, such as in P-pleated sheets, that are found in many organisms as a structural protein. A survey of literature on the composition of the proteins, in fact, reveals that on the average, the composition of the organic matrix of the shell is about 25% aspardic acid, 6% glutamic acid, 8% serine, 25% glycine, 10% alamine, and the rest constituting various amino acids (see references in Sarikaya and Aksay, 1993a,b). There is also speculation of the composition of the insoluble portion of the organic matrix, but is probably much less accurate. It is possible, based on many investigations, that this portion of the matrix may constitute polysacharides in addition to proteins. The major flaw in all these studies have been that the macromolecules extracted from the shell have not been specific to a certain section in the shell through the thickness. In fact, the shell is often pulverized, with both the calcitic (prismatic) and aragonitic (nacre) sections intermixed before decalcification. The organic macromolecules from each of these sections are expected to have different stoichiometric compositions and, hence, different conformations with respect to each other, this will, in turn, greatly affect the structure, crystallography, and the geometry of the inorganic crystals forming. The analysis which gives a mixture of all the biological macromolecules in a shell, therefore, has not so far been adequate in describing the true composition and the structure of the macromolecules that make up the nacre structure. Nevertheless, based on the findings so far, and limited direct analysis, the organic matrix may actually have a sandwich structure containing three organic sublayers, each with its own unique composition and molecular conformation (Addadi and Weiner, 1990; Nakahara et al., 1982; Watabe, 1965; Weiner and Traub, 1984; Weiner et al., 1983). According to this scheme, mentioned above, the central portion of the organic matrix is composed of chitin (polysacharides), which is surrounded on both sides by P-pleated proteins. The layer next to the inorganic is then composed of acidic proteins. In this scheme, while the chitin provides the structural (mechanical) stability t o the composite forming a back-bone (framework macromolecules), the P-pleated sheets act as the substrate and provides the biomineralization sites (nucleator macro-

molecules). The acidic proteins are there to fill in the gaps between the organic matrix and the aragonite crystals. Neither the composition of these layers nor their structure has yet been clearly investigated (a preliminary result obtained using an atomic force microscope (AFM)is presented in Figure 11which shows the details of the organic layer). Therefore, the establishment of the structural (stereochemical) relationship between the organic and the inorganic layers, a critical issue in biomineralization, have to wait until full investigation of the organic matrix. It may be possible to directly study the organic macromolecules in the TEM if proper molecular markers are developed. For this, isolation of each of the major macromolecules from the nacre organic matrix is necessary. Once this is done, development of antibodies and proper stains will follow. These stains, then, can be used in ultramicrotomed thin sections to identify locations and concentrations of each of the macromolecules, a procedure requiring a long term investigation. A preliminary work is being carried out in this group to reveal structures of the macromolecules in samples that are either ultramicrotomed or low temperature ion-beam milled (Gunnison et al., 1992). Two such micrographs are displayed in Figure 12. Figure 12a reveals layered structure of the organic matrix in an ultramicrotomed section. The contrast difference is due to differential staining of the organic sub-layers by uranyl acetate due to their differences in composition and structure. A detailed microstructure of the interface between organic layer and aragonite is shown in Figure 12b. The image was taken from a sample that was low-temperature ion-milled and then slightly etched with gold citrate (Liu et al., 1992). The preferential etching of the aragonite platelets gives them a saw-tooth appearance. The light acid used that causes this feature also dissolves away the acid soluble portion of the organic sublayers. As a result, central sub-layers of the organic matrix are exposed. This is the part we think is made up of polysaccharides since they are known to be resistant to such treatment. Further studies are underway to correlate these characteristic features related to the structures of aragonite and the organic matrix. It is clear from the point of biomimetics, i.e., utilizing lessons from biology in developing technological materials, that it is essential to understand what the function of the organic matrix in nacre is in controlling the nucleation of the inorganic crystals, their shape formation and their growth. In future crystal engineering, i.e., in making nanostructures and laminated composites based on biological hierarchical composites, by using either synthetic or biological macromolecules, the first requirement is to understand the mechanism of inorganic-organic interactions in biocomposites. The living systems, closely controlling and using these interactions, can produce extraordinary and highly ordered structures that have unprecedented physical properties. Detailed further investigations of interfaces to establish these organictinorganic interactions, especially in a hard tissue like nacre having a relatively simple structure, is well warranted for potential biomimetic applications.

372

M.SARIKAYA

Fig. 11. AFM image of cleaved surface of the nacre section of a red abalone shell: The low magnification image in (a)shows the general view of the aragonite platelets with the platelet boundaries indicated by arrows. The image in (b) shows the detail of an interface between aragonite platelets (arrows). Corrugations in (b) are possibly due to

plastic deformation of the organic matrix during cleavage. Local ordering of corrugations at smaller length scale is discernible in the high magnification images ( c ) and (d).The origin of these corrugations, that have a separation of about 5 nm, is not yet known. Arrows in ( c ) and (d) show periodic organization of macromolecules.

structure of the organic matrix and its structural relaCONCLUDING REMARKS AND tionship with the inorganic phase are still limited. FurFUTURE STUDIES Biological composites were discussed in the context ther investigation, therefore, is necessary in nacre, and of their unique micro- and nano-architecture which re- other similar biomaterials reviewed here before their sult in excellent properties. Despite considerable effort possible biomimetic applications. Based on the results in the field, our understanding of the mechanisms that from nacre, one of the widely investigated hard tissue, operate in nacre making it a tough and strong compos- some of the major issues for future biomimetic research ite, and current knowledge of the composition and may be stated as follows: 1. On mechanical properties,

INTRODUCTION TO BIOMIMETICS

373

Fig. 12. Structure of the organic matrix in nacre of abalone. The sample in (a) was ultramicrotomed,fixed, and stained. At least five sub-layers are exposed in the organic layers, possibly due to their differential staining (arrows).(b) The image was taken from a sample

that was ion-beammilled and slightly acid etched; it reveals insoluble (central) portion of the organic layer db (dark bands) in the organic matrix and the saw-tooth (st) appearance of the etched aragonite (C)/ matrix (M) interface.

including: microstructure-property correlations, including the coupling between the properties of the individual organic and inorganic components and their size and configuration (nanocomposite, size, or mesoscopic effects); micromechanisms of toughening and strengthening. 2. On biomimicking: new microstructural design strategies in future synthetic materials that requires investigation of: nucleation, interface structures and properties, shape formation and its relation with the macromolecular conformation, biomineralization (controlled by the organic substrate), new materials fabrication techniques to mimick microarchitecture of biological composites. 3. On bioduplication: novel synthesis and processing strategies in manufacturing new materials, that require investigation of: macromolecular studies: localization, identification, purification, cloning, and expression of proteins, selfassembly of proteins, biomineralization (controlled by the genetically designed and self-assembled macromol-

ecules), growth control from the molecular to macroscale, evolution of microarchitecture, hierarchy, and shape. It is clear that a thorough understanding of the structural organization and relationships between the units that make up a biological composite is the first step in biomimetics. The structural organization, which is often hierarchical, range from the molecular scale (10-l' m) to macro scale (lo-' m). The analysis of biological structures, therefore, require an arsenal of techniques that provides structural information at each level of dimensional hierarchy. This hierarchy ranges from the types, compositions, and conformation of the organic macromolecules and nano- and microstructures of biogenic inorganic phases, including aspects of inorganiclorganic interfacial structures, to overall macro-architecture of a biological tissue, and its functionalities. The investigation of structures at each dimensional scale may require one or more tech-

374

M. SARIKAYA Glimcher, J. (1981) On the form and function of bone: From molecules to organs. In: The Chemistry and Biology of Mineralized Biological Tissues: Wolffs Law Revisited. A. Veis, ed. Elsevier, New York, pp. 617-673. Gorby, Y.A., Beveridge, T.J., and Blakemore, R.P. (1988) Characterization of the bacterial magnetosome membrane. J. Bacteriol., 170: 834-841. Gosline, M., DuMont, M.E., and Denny, M.W. (1986) Structure and properties of spiders' silk. Endeavour, 10:37-43. GrBgoire, C. (1972) Structure of the molluscan shell. In: Chemical Zoology, M. Florkin and M. Scheer, eds. Academic Press, New York, pp. 45-102. Griffith, A. (1925) The phenomena of rupture and flow in solids. Phil. Trans. CCXXI, A:163-198. Griinbaum, B., and Shephard, G.C. (1987)Tilings and Patterns. W.H. Freeman, New York. Gunderson, S.L., and Schiavone. R.C. (1989) The insect exoskeleton: A natural structural material. JOM, 41:80-82. Gunnison, K., Sarikaya, M., Liu, J., and Aksay, LA. (1992) Structuremechanical property relationships in a biological ceramic-polymer composite: Nacre. In: Hierarchically Structured Materials, MRS Symp. Proc., Vol. 255. LA. Aksay, E. Baer, M. Sarikaya, and D.A. Tirrell, eds. Materials Research Society, Pittsburgh, PA, pp. 171183. Haasen, P. (1978) Physical Metallurgy. Cambridge University Press, London. Harrison, P.M., Artymiuk, P.J., Ford, G.C., Lawson, D.M., Smith, J.M.A., Treffry, A., and White, J.L. (1989) Ferritin: Function and structural design of a n iron-storage protein. In: Biomineralization: Chemical and Biochemical Perspectives. S.Mann, J . Webb, and R.J.P. Williams eds. VCH, Weinheim, Germany, pp. 257-294. Hull, D. (1981) An Introduction to Composite Materials. Cambridge University Press, London. Jackson, A.P., Vincent, J.F.V., and Turner, R.M. (1988) The mechanical design of nacre. Proc. R. SOC. London (Biol), B234:415-440. Jorgensen, J.D., Kitazawa, K., Tarascon, J.M., Thomson, M.S., and Torrance, J.B., eds. (1989) High Temperature Superconductors:Relationships Between Properties, Structure, and Solid-state Chemistry, Vol. 156. Materials Research Society, Pittsburgh, PA. Khanuja, S. (1991) Processing of Laminated B,C-Polymer Laminated Composites. MS Thesis, University of Washington. Liu, J., Sarikaya, M., and Aksay, LA. (1992) A hierarchically structured model composite: A TEM study of the hard tissue of red abalone. In: Hierarchically Structured Materials, MRS Symp. Proc., Vol. 255. I.A. Aksay, E. Baer, M. Sarikaya, and D.A. Tirrell, eds. Materials Research Society, Pittsburgh, PA, pp. 9-17. Lowenstam, H.A., and Weiner, S. (1989) On Biomineralization. Oxford University Press, New York. Mann, S. (1988) Molecular recognition in biomineralization. Nature, 33:119-123. Mann, S., Sparks, H.C., and Wade, W.J. (1990) Crystallo-chemical control of iron oxide biomineralization. In: Iron Biominerals. R.B. Frankel and R.P. Blakemore, eds. Plenum, New York, pp. 21-49. Nakahara, H., Bevelander, G., and Kakei, M. (1982) Electron microscopic and amino acid studies on the outer and inner shell layers of Haliotis rufescens. VENUS Jpn. J. Malac., 41:33-46. Neville, A. (1975) The Biology of Arthropod Cuticle. Springer and Verlag, Berlin. Rieke, P.C., Calvert, P.D., and Alper, M., eds. (1990) Materials Synthesis Using Biological Processes. Vol. 174. Materials Research Society, Pittsburgh, PA. Sarikaya, M., and Aksay, I.A. (1992) Nacre of abalone shell: A natural multifunctional nanolaminated ceramic-polymer composite material. In: Results and Problems in Cell Differentiation in BiopolyCase ed. Springer and Verlag, Amsterdam, pp. 1-25. mers, S. Sarikaya, M., and Aksay, I.A. (1993a) Nacre: Properties, crystallography, morphology, and formation. In: Design and Processing of Materials by Biomimetics. M. Sarikaya and I.A. Aksay eds. American Institute of Physics, Washington, D.C., pp. 35-86. Sarikaya, M., and Aksay, I.A. eds. (1993b) Design and Processing of Materials by Biomimetics. American Institute of Physics, Washington, D.C. Sarikaya, M., Gunnison, K.E., Yasrebi, M., and Aksay, LA. (1990) Mechanical Property-Microstructural Relationships in Abalone Shell. Vol. 174. Materials Research Society, Pittsburgh, PA, pp. 109-116. Sarikaya, M., Liu, J., and Aksay, LA. (1992a) Structure of the seaurchin teeth. unpublished research.

niques which might be specialized or involve all modes of microscopical analysis including imaging, diffraction, and spectroscopy, using light, X-rays, neutrons, or electrons as primary radiation. It will be possible only then to correlate structures at different length scales including their functions, and achieve design rules for synthesis of future biomimetic materials with tailored structures and properties.

ACKNOWLEDGMENTS This work is supported by the Air Force Office of Scientific Research, Grant AFOSR-91-0281 and by a University Research Initiative through Army Research Office, Grant DAAL03-92-G-0241.

REFERENCES
Addadi, L., and Weiner, S.(1990) Interaction between acidic macromolecules and structured crystal surfaces, stereochemistry and biomineralization. Mol. Cryst. Liq. Cryst., 13:305-322. Aksay, LA., Baer, E., Sarikaya, M., and Tirrell, D.A., eds. (1992) Hierarchically Structured Materials. Vol. 255. Materials Research Society, Pittsburgh, PA. Alper, M., Calvert, P.D., Frankel, R., Rieke, P., and Tirrell, D.A., eds. (1991) Vol. 218. Materials Research Society, Pittsburgh, PA. Baer, E. (1986) Advanced polymers. Sci. Am., 255:178-190. Baer, E., and Moet, A., eds. (1991) High Performance Polymers. Hanser, Munich. Baer, E., Cassidy, J.J., and Hiltner, A. (1988) Hierarchical structure of collagen and its relationship to the physical properties of tendon. In: Collagen: Biochemistry and Biomechanics. M.E. Nimi, ed. CRC Press, New York, pp. 177-199. Baer, E., Hiltner, A., and Morgan, R.J. (1992) Biological and synthetic hierarchical composites. Physics Today, October: pp. 60-67. Bazylinski, D.A., Frankel, R.B., Garrat-Reed, A., and Mann, S. (1991) Biomineralization of iron sulfides in magnetotactic bacteria from sulfidic environments. In: Iron Biominerals. R.B. Frankel and R.P. Blakemore, eds. Plenum Press, New York, pp. 239-256. Berman, A. et al. (1988) Interaction of sea-urchin skeleton macromolecules with growing calcite crystals-A study of intercrystalline proteins. Nature, 3 3 1 5 4 6 4 4 8 , Berman, A,, Addadi, L., Leiserowitz, L., Weiner, S., Nelson, M., and Kvik, A. (1990) A synchrotron X-ray study of a unique proteincalcite composite material. Science, 250:664-667. Blakemore, R.P. (1982) Magnetotactic bacteria. Ann. Rev. of Microbiol., 36:217-238. Bouligand, Y. (1965) Sur une architecture torsadee repandue dans de nombreuses cuticules d'arthropodes. C R Hebd. Sceances Acad. Sci., 26:3665-3668. Brear, K., and Curry, J.D. (1976) Structure of the sea-urchin tooth. J . Mat. Sci., 11:197?-1978. Currey, J.D. (1987) Biological composites. J . Mat. Edu., 9:118-296. Dameron, C.T., Reese, R.N., Mehra, R.K., Kortan, A.R., Carrol, P.J., Steigerwald, L.E., Brus, M.L., and Winge, D.R. (1989) Biosynthesis of cadmium sulfide quantum semiconductor crystallites. Nature, 338596-597. Dawridge, W. (1979) Mechanical Behavior of Ceramics. Cambridge Univ&ity Press, London. Evans, A.G. (1988) High toughness ceramics. J . Mat. Sci. Eng., A1051 106165-75. Evans, A.G., and Marshall, D.A. (1989) The mechanical behavior of ceramic matrix composites. Acta. Met., 37:2567-2583. Fendler, J.H. (1982) Membranemimetic Chemistry. Wiley-Interscience, New York. Frankel, R., and Blakemore, R.P. eds. (1991) Iron Biominerals. Plenum, New York. Fulrath, R.M., and Pask, J.A., eds. (1966) Ceramic Microstructures. University of California Press, Berkeley. Gell, M., Duhl, D.H., and Giamei, A.F. (1980) High Temperature Superconductors: Relationships Between Properties, Structure, and Solid-state Chemistry. ASM-International, Metals Park, OH. Ghiradella, H., Aneshansley, D., Eisner, T., Silberglied, R.E., and Hinton, H.E. (1972) Ultraviolet reflection of male butterfly: Interference color caused by thin-layer elaboration of wing scales. Science, 178:1214-1217.

INTRODUCTION TO BIOMIMETICS
Sarikaya, M., Liu, J., and Aksay, LA. (1992b) unpublished results. Shaw, D.W., Bean, J.C., Keramidas, V.G., and Peercy, P.S., eds. (1989) Epitaxial Heterostructures. Vol. 198. Materials Research Society, Pittsburgh, PA. Shinjo, T., and Takada, T., eds. (1987) Metallic Superlattices, Elsevier, Amsterdam. Simkiss, K., and Wilbur, K.M. (1989) Biomineralization, Academic Press, New York. Tressler, R.E., and Bradt, R.C., eds. (1983) Deformation of Ceramic Materials. Plenum, New York. Veis, D.J., Albinger, T.M., Clohisy, J., Rahima, M., Sabsay, B., and Veis, A. (1986) Matrix proteins of the teeth of the sea urchin Lytechirzus uuriegutus. J.-Exp. Zool., 240:35-46. Wainwright, S., Briggs, W.D., Currey, J.D., and Gosline, J.M. (1976) Mechanical Design in Organisms. John Wiley and Sons, New York. Watabe, N. (1965) Studies in Shell Formation: Crystal-Matrix Rela-

375

tionships in the Inner layers of Molluscan Shells. J . Ultrastr. Res. 12:351-370. Weiner, S., and Traub, W. (1984) Macromolecules in mollusc shells and their functions in biomineralization. Philos. Trans. R. SOC. Lond. Biol. B304:425-434. Weiner, S., Talmon, Y., and Traub, W. (1983) Electron diffraction studies of molluscan shell organic matrices and their relationship to the mineral phase. Int. J . Biol. Macromol., 5325-328. Yasrebi, M., Kim, G.H., Milius, D.L., Sarikaya, M., and Aksay, LA. (1990) Biomimetic processing of ceramics and ceramic-based composites. In: Better Ceramics Through Chemistry-IV. Vol. 180. C.J. Brinker, D.R. Ulrich, and B.J. Zelinski, eds. Materials Research Society, Pittsburgh, PA, pp. 625-635. Zackay, V.F., and Aaronson, H.I., eds. (1962) Decomposition of Austenite by Diffusional Processes. Interscience, New York.