Formation and Properties of Clay-Polymer Complexes

Dedication

Developments in Clay Science, Vol. 4, Suppl (C), 2012

ISSN: 1572-4352

doi: 10.1016/B978-0-444-53354-8.00018-9

Preface to the First Edition

This book has been in planning since 1968 when I first began to search and abstract the literature on clay–organic systems. The intention then was to produce a critical and comprehensive account of the many and varied aspects of clay–organic chemistry, embracing those reactions in which organic polymers were involved. I still had this idea in mind when the opportunity for writing such a review arose in 1971 with the announcement (by the London publishing house of Adam Hilger Ltd.) of a competition for the Adam Hilger Prize 1972. It would not have been possible, however, for me to enter the competition without the active support and enthusiastic encouragement from the late Dr. M. Fieldes who was then Director of the Soil Bureau. It seems fitting, therefore, that I should dedicate this book to his memory.

On sifting through the published material, it soon became evident that the reactions of clays with organic polymer should not be discussed concurrently with those involving small, well-defined organic species. To do so would not only lead to excessive length but would also result in a certain loss of perspective since, although both systems share a great many features, complex formation between clays and polymers is controlled by variables which are essentially different from those encountered in the adsorption of organic micromolecules.

Once the decision was taken to treat each system separately, it seemed only logical to begin with summarising the extensive data on the behaviour of small organic compounds at clay–mineral surfaces. This work came out in the form of a monograph (Theng, 1974) in which I stated that the clay–polymer interaction would form the subject of a separate book. What I have done here, therefore, is simply to complete the task I set out to do and keep the promise I made to myself. In writing this review, I have been much encouraged by the kind comments which my peers have expressed with regard to my previous effort. Also, the fact that a synthesis of the data on clay–polymer systems has not been previously attempted has provided me with the necessary inspiration at moments when I would have preferred to be in the laboratory.

The systematic study of clay–organic reactions, as a whole, is a young science since the essential crystallinity of the main groups of clay–minerals and the macromolecular concept have not become generally accepted until relatively recently. However, complex formation between clays and the organic constituents of soils has been recognised and described for over a century (Schloesing, 1874). Indeed, the clay–organic interaction has been implicated in the very origin of life on earth; certainly, this process is … as vital to the continuance of life as, and less understood than, photosynthesis … (Jacks, 1973).

Despite its late beginning, much progress had been made over that past two decades in understanding the behaviour of polymeric materials at the solid/solution interface, research being stimulated by the importance and potential of the solid–polymer interaction in agriculture, biology, and industry. It can even be said that theory has outrun practice in the case of uncharged, linear homopolymers. The picture is less clear as regards polyelectrolytes which, of course, are more relevant than their nonionic counterparts to soil and biological systems. For this reason, much of the data on the clay–polymer interaction, in general, are not amenable to a quantitative interpretation. Nevertheless, it seems timely to survey the advances that have so far been made in this subject and to outline the current state of knowledge. In this task, I have been much assisted by the availability of a number of excellent reports on some of the topics under discussion. I have attempted, as much as possible, to present a balanced and coherent account of each topic, pointing out areas of discordance and uncertainty and suggesting ways in which such discrepancies may be resolved. In doing so, I hope to have given something more than just an annotated catalogue of references. How well or otherwise I have succeeded in this direction is for the reader to judge.

The growing literature on clay–polymer systems has made it necessary to be selective as to what aspects and how much, of a given topic, are to be included if the size of the book is to be kept within manageable bounds without sacrificing essential depth and detail. The practical applications of the clay–polymer interaction, for example, are not discussed at great length because much of the information on this topic is published in the form of patents which are difficult to assess and collate. The rapid growth of the subject has also meant that some parts of the book, at least, will become out of date in a very short time. One also had to draw a line as regards the selection of published material which, apart from the few last minute additions, does not go beyond the end of 1976.

For the sake of clarity and convenience, I have divided the contents of the book into three parts. Part I gives an outline of clay–mineral structures, the properties of aqueous clay suspensions, and the behaviour of polymers at clay and mineral surfaces. Part II describes the reactions of clays with synthetic polymers, and Part III deals with complex formation between clays and various classes of naturally occurring polymers. I have tried to write each chapter in Parts II and III in the form of a review paper ending with a rather comprehensive list of references to the original literature. This was done for the benefit of those readers who may wish only to get quickly acquainted with the state of the art in the topic of their own specialty.

Although this book is directed primarily to soil scientists and agronomists, it may also appeal and be of interest to a variety of chemists who, at sometime in their respective professional careers, have to deal with environmental problems in which clays, polymers, and their reaction products feature. Besides serving as a work of reference, the book may be found useful for teaching at the graduate and higher levels of agricultural and soil science courses.

It goes without saying that the reliance of a book of this type on previous publications extends to illustrated and tabulated data. In this connection, I wish to acknowledge the many authors who have kindly permitted me to reproduce figures and tables. Thanks are also due to the draughting staff of the Science Information Division of D.S.I.R. for drawing up the numerous diagrams. I am indebted to the library and typing staff of the Soil Bureau for their unstinting assistance, respectively, in procuring journal articles from other libraries and in converting a long stream of untidily written drafts into neat typescripts. I am very grateful to Dr. R.B. Miller, Director of the Soil Bureau, for allowing me to devote myself entirely to bringing this book to completion over an extended period of time. Last, I would like to thank my family for their forbearance and understanding.

B.K.G. Theng

Lower Hutt

December 1977

References

G.V. Jacks. The biological nature of soil productivity. Soils Fertil. 1973;26:147-150.

Th. Schloesing. Détermination de l’argile dans la terre arable. C. R. Hebdomadaires Séances Acad. Sci. 1874;78:1276-1279.

B.K.G. Theng. The Chemistry of Clay-Organic Reactions. London: Adam Hilger Ltd.; 1974. pp. 343

Developments in Clay Science, Vol. 4, Suppl (C), 2012

ISSN: 1572-4352

doi: 10.1016/B978-0-444-53354-8.00019-0

Preface to the Second Edition

Since the appearance of the first edition of this book in 1979, there have been substantial developments in clay science as a whole, partly owing to the application of modern surface analytical, computational, and spectroscopic techniques (Bergaya et al., 2006). The same can be said about the clay–polymer interaction, especially with regard to the synthesis and characterisation of polymer–clay nanocomposites in conjunction with the actual and potential applications of these ‘hybrid’ materials in industry. As a result, a vast literature on the chemistry of clay–polymer reactions has accumulated over the past three decades or so.

The primary objective of preparing a new edition, therefore, was to update the information and material contained in the first edition. Besides revising the text, I have added many new illustrations and tabulated data. The general arrangement of the contents, however, has essentially been retained in that this volume, like its predecessor, is divided into three parts each of which contains a number of chapters. No changes have been made to the structure of Part I, but an entirely new chapter dealing with polymer–clay nanocomposites has been added to Part II. At the same time, the short chapter on the formation and properties of clay–lignosulphonate complexes has been removed from Part III because very little new material on this topic has been published in the intervening years. As with the first edition, I have attempted to write each chapter in the form of a self-contained, critical review article with its own complement of references to the original literature.

I am indebted to a number of anonymous colleagues for supporting my proposal to Elsevier that a second edition of the book is warranted. I am grateful to Ms. N. Faville for scanning the numerous figures, the reception staff (Palmerston North) for assistance with formatting chapter drafts, figures, and tables, and the library staff (Lincoln) for obtaining a regular stream of journal articles on interlibrary loan. I also wish to acknowledge Drs. M. Calabi-Floody, C.W. Ross, and G. Yuan for providing some diagrams and micrographs as well as authors, scientific societies, and publishers for permission to reproduce figures and tables. Finally, I would like to thank my wife, Judy Theng, for her encouragement, patience, and support.

B.K.G. Theng

Palmerston North

October 2011

Reference

F. Bergaya, B.K.G. Theng, G. Lagaly, editors. Handbook of Clay Science. Amsterdam: Elsevier, 2006. pp. 1224

Developments in Clay Science, Vol. 4, Suppl (C), 2012

ISSN: 1572-4352

doi: 10.1016/B978-0-444-53354-8.00001-3

The Clay Minerals

B.K.G. Theng

Abstract

Here we summarize the basic structures, compositions and surface chemical properties of clay minerals. In terms of the clay–polymer interaction, smectite (e.g., montmorillonite) is the most important mineral species because of its propensity for adsorbing and intercalating organic polymers. Kaolinite, halloysite, illite, palygorskite, sepiolite, allophane and imogolite are less frequently used than smectite. The same comment applies to vermiculite, mica and chlorite. Since the clay–polymer interaction commonly takes place in aqueous media, this chapter also gives a brief account of clay colloid chemistry. Again, montmorillonite holds the limelight because of its interlayer swelling capacity and dispersibility in water.

Keywords

Clay minerals Structures Clay colloid chemistry Montmorillonite

1.1 Structural Aspects

Clay minerals are phyllosilicates (layer silicates) that dominantly make up the fine-grained fraction of soils and sediments. It is now generally accepted that the majority of clay minerals are crystalline. This so-called clay-mineral concept, however, only dates back to the late 1930s by which time the crystallinity and structures of the main groups of phyllosilicates related to clay minerals have been established (Grim, 1968).

In the soil science literature, the term ‘clay’ or ‘clay fraction’ denotes a class of minerals, the particles of which are smaller than 2 μm in equivalent spherical diameter (e.s.d.). Although this value is clearly arbitrary, the ‘non-clay’ constituents, notably feldspar and quartz, are practically concentrated in the >  2 μm e.s.d. fraction (Grim, 1968). Not all of the clay-size minerals may be phyllosilicates, or even crystalline. Indeed, the non-crystalline or short-range order nanoscale particles (<  100 nm) in soil are the most important fraction in terms of their reactivity towards plant nutrients, water, organic compounds and environmental contaminants (Theng and Yuan, 2008). In geology, sedimentology and geo-engineering, the particle size limit is commonly set at <  4 μm, while in colloid science the value of <  1 μm is generally accepted. Thus, ‘clay’ signifies a naturally occurring material of fine-grained texture that becomes plastic when wet and hardens when dried or fired, while the term ‘clay mineral’ denotes a class of hydrated phyllosilicates and minerals that ‘…impart plasticity to clay and which harden upon drying or firing’ (Guggenheim and Martin, 1995). Unlike clay, clay mineral may be synthetic. Nevertheless, these two terms are often used interchangeably. Thus, the ‘clay’ component in such terms as ‘clay–polymer complexes’ and ‘clay–polymer nanocomposites’ is almost invariably identifiable with a clay mineral.

Because of the fine-grained nature of clay minerals, their characterization and identification call for a combination of various techniques. X-ray diffraction, chemical and thermal analyses together with electron microscopy and organic intercalation are routinely used for this purpose. These methods may be supplemented by a range of spectroscopic techniques, including Fourier-transform infrared (FTIR), Mössbauer, nuclear magnetic resonance (NMR), X-ray absorption and X-ray photoelectron spectroscopies (Bergaya et al., 2006; Wilson, 1994). Even then, the occurrence of structural defects, layer stacking disorders and interstratified structures makes for practical difficulties. Perhaps a more intractable analytical problem than ‘microcrystallinity’ is the presence of ‘associated minerals’ such as carbonates, quartz and iron/aluminium hydr(oxides) (Dixon and Schulze, 2002; Guggenheim and Martin, 1995). These mineral ‘impurities’, in the form of either discrete entities or surface-adsorbed species, are difficult to detect by X-ray diffractometry (XRD). This problem arises because the concentrations involved are small and some associated minerals are of short-range order; that is, their structural regularity does not extend beyond ∼  1 nm. As such, the occurrence of associated minerals (including allophane) can often be indicated only indirectly by means of selective dissolution coupled with differential infrared spectroscopy and/or XRD (Campbell and Schwertmann, 1985; Parfitt, 1990; Wada and Greenland, 1970).

Only the salient points concerning the structures and surface chemical properties of representative clay minerals will be mentioned in this chapter. For a detailed discussion of the origin, structures, properties and practical applications of clays and clay minerals together with their characterization and identification by various analytical and instrumental techniques, the reader should consult the series of monographs written or edited by Grim (1968), Greenland and Hayes (1978), Brindley and Brown (1980), Dixon and Weed (1989), Wilson (1994), Velde (1995), Moore and Reynolds (1997) and Bergaya et al. (2006). A number of reviews and books dealing with some particular aspect of clay science are also available, notably that by Weaver and Pollard (1973) and Newman (1987) on the chemistry of clay minerals, by Farmer (1974) on infrared spectroscopy, by Theng (1974) and Yariv and Cross (2002) on the clay–organic interaction, by van Olphen (1977) on colloid chemistry, by Sudo et al. (1981) on electron microscopy and by Sposito (1984) on surface chemistry.

The basic building block in the phyllosilicate structure is an alumino-silicate layer comprising a silica tetrahedral sheet and an alumina octahedral sheet, joined together in certain proportions. A tetrahedron consists of a central Si⁴  + cation coordinated to four oxygens, while in an octahedron the central Al³  + or Mg²  + cation is coordinated to six hydroxyls. A tetrahedral sheet is formed when individual tetrahedra are linked by sharing three corners. As a result, the basal oxygens become approximately coplanar, forming an open hexagonal—in reality ditrigonal—network, while the apical oxygens point in the same direction. On the other hand, individual octahedra in an octahedral sheet are linked by sharing edges (Figure 1.1). When Al³  + is the cation, only two of every three octahedral sites are occupied in order to maintain electrical neutrality. The resultant structure (as found in the mineral gibbsite) is referred to as dioctahedral. On the other hand, when Mg²  + is the central cation, all three octahedral positions are filled. The (brucite) sheet structure that arises is accordingly termed ‘trioctahedral’.

Figure 1.1 Bottom: (A) a silica tetrahedron in which the central silicon ion is coordinated to four oxygens; (B) a tetrahedral sheet formed by linking silica tetrahedra through corner-sharing. Top: (A) an alumina octahedron in which the central aluminium ion is coordinated to six hydroxyls; (B) an alumina octahedral sheet formed by linking octahedra through edge-sharing.

From Grim (1968).

Condensation of one tetrahedral (T) sheet with one octahedral (O) sheet gives rise to a 1:1 (T–O) layer structure (Figure 1.2). Here the tips of the tetrahedral sheet project into a hydroxyl plane of the octahedral sheet, replacing two-thirds of the hydroxyl ions as in kaolinite (Figure 1.3). As a result, one basal plane (surface) of the layer consists of oxygens (from the tetrahedral sheet), while the other is made up of hydroxyls (from the octahedral sheet). Condensation of two tetrahedral sheets with one octahedral sheet through which the octahedral sheet becomes sandwiched between two opposing tetrahedral sheets gives rise to a 2:1 (T–O–T)-type layer structure as exemplified by pyrophyllite (Figure 1.4). Here both basal planes of individual layers are composed of oxygens (belonging to the tetrahedral sheets).

Figure 1.2 Schematic representation of a silica tetrahedral sheet condensing with an alumina octahedral sheet to form a 1:1-type layer structure. The mode of condensation between the two sheets is indicated by the solid rectangles.

Figure 1.3 The structure of kaolinite, viewed along the a-axis, showing the superposition of adjacent 1:1-type layers within a particle. The outer hydroxyls are denoted by A and B, the inner-surface hydroxyls by C and the inner hydroxyl by D. The basal or d(001) spacing of ∼  7   Å is equal to the layer thickness. The unit cell dimension is indicated by the broken rectangle. Note: 1   Å  =  0.1   nm.

Modified after Dixon (1989).

Figure 1.4 The 2:1-type layer structure of pyrophyllite, viewed along to a-axis. The basal or d(001) spacing and unit layer composition are also indicated.

There is considerable scope in these structures for isomorphous substitution, that is, the partial replacement of Si⁴  + (in tetrahedral positions) and/or Al³  + (in octahedral sites), by cations of similar size and coordination but (usually) of lower valency. When this happens, the layer structure acquires a (permanent) negative charge. The net layer charge, x, per formula unit (=  half unit cell), ranging from ∼  0 to ∼  2 electron charges, is therefore an important parameter in the classification of phyllosilicates. In the absence or near absence of isomorphous substitution (x  ∼  0), the interlayer space is essentially empty. For the majority of 2:1-type phyllosilicates, however, x  >  0, and the resultant negative layer charge is balanced by hydrated or non-hydrated cations, occupying interlayer positions. Since the charge-balancing ions or ‘counterions’ are generally exchangeable, the majority of layer silicates have a characteristic cation exchange capacity. The interlayer space may also be filled with a hydroxide sheet to give the chlorite layer type, previously denoted as a 2:1:1-type layer. These points are summarized in Table 1.1.

Table 1.1 Classification of Planar Hydrous Phyllosilicates.

The particles of planar phyllosilicates are made up of a finite number of either 1:1- or 2:1-type layers that are continuous in the a- and b-direction and stacked in a more or less parallel manner in the c-direction perpendicular to the layer planes. Such an arrangement offers scope for layer displacement by rotation and/or translation. Many of the groups listed in Table 1.1 have a number of polytypes or structural varieties with different layer stacking sequences. As might be expected, a clay mineral particle may also consist of different types of layers that are either randomly or regularly interstratified. Many pedogenic clay materials have this mixed-layer structural arrangement.

Guggenheim et al. (2006) listed illite as a dioctahedral species in the interlayer-deficient mica group. Following the proposal by Rieder et al. (1998), however, Guggenheim et al. (2007) subsequently considered illite to be a series name, and hence omitted it from Table 1.1. Similarly, Meunier and Velde (2004) suggested that illite be regarded as a separate group of clay-size, dioctahedral mica-type minerals. Palygorskite and sepiolite are not included in Table 1.1 because these 2:1-type phyllosilicates are non-planar (fibrous). Likewise, allophane is not included in Table 1.1 because this nanosize hydrated alumino-silicate is non-planar (spheroidal) and lacks long-range order. Nevertheless, allophane is commonly regarded as a clay mineral because of past usage. The same argument applies to imogolite, a nanosize tubular alumino-silicate with long-range order along the tubule length. In terms of the clay–polymer interaction, illite, palygorskite, sepiolite, allophane and imogolite are much less frequently used than the smectite group of phyllosilicates. The same comment applies to vermiculite, mica and chlorite. Only a brief description of these minerals will, therefore, be given in the following sections.

1.1.1 Planar 1:1 Phyllosilicates

1.1.1.1 Kaolinite

Among the many mineral species within the 1:1-type planar phyllosilicate class (Table 1.1), only kaolinite and halloysite have featured in the clay–polymer interaction. A diagram of the layer structure of kaolinite is shown in Figure 1.3. The layer thickness of ∼  0.7 nm also represents the basal or d(001) spacing. When viewed under the electron microscope kaolinite particles appear as pseudo-hexagonal platelets, ranging in thickness from 0.05 to 2 μm, forming ‘book-like’ aggregates (Dixon, 1989) as illustrated in Figure 1.5. With a layer charge (x) close to zero, kaolinite has essentially no interlayer material. Nor does it show interlayer expansion (swelling) in water because the contiguous layers within a particle are strongly held together by Al–OH…O–Si– hydrogen bonding, supplemented by dipole–dipole and van der Waals interactions. Nevertheless, kaolinite can be induced to swell under special conditions, for example, by contacting the mineral with compounds that can form hydrogen bonds with the interlayer surface (e.g. amides, hydrazine, urea) or have high dipole moments, such as dimethylsulphoxide (Lagaly et al., 2006; Theng, 1974).

Figure 1.5 Transmission electron micrograph of kaolinite from Sasso (Italy), showing platy particles with a hexagonal outline forming book-like aggregates.

From Bergaya and Lagaly (2006).

Summation of the anionic and cationic charges in the kaolinite structure, having the unit composition of Al2Si2O5(OH)4, shows that the layer is electrically neutral (Figure 1.3). In reality, however, most kaolinites carry a small amount of negative charge, the origin of which is still a matter of some controversy. This charge may arise from the presence of surface-adsorbed impurities and/or smectite layers (Brady et al., 1996; Jepson, 1984; Ma and Eggleton, 1999). The weight of evidence, however, points to a small amount of isomorphous substitution within the structure (Angel and Hall, 1973; Bolland et al., 1976; Newman et al., 1994; Rengasamy et al., 1975), giving rise to a cation exchange capacity (at pH 7) of 1.4–3.6 cmol/kg. The surface negative charge density of 0.11–0.15 C/m² (Bolland et al., 1980) is comparable to, if not greater than, that of montmorillonite (Lagaly, 2006; Table 1.4). Further, the cation exchange sites in kaolinite are apparently confined to the tetrahedral basal surface rather than being uniformly distributed throughout the whole particle (McBride, 1976; Tombácz and Szekeres, 2006; Weiss and Russow, 1963).

Of greater importance than the probable presence of negatively charged sites at the (siloxane) basal surface, is the variable charge that develops at the edge surface of kaolinite particles as the pH of the ambient solution is varied (Swartzen-Allen and Matijević, 1974; van Olphen, 1977). This pH-dependent charge largely arises from the protonation (under acid conditions) and deprotonation (under alkaline conditions) of ‘under-coordinated’ aluminol groups exposed at particle edges (Schoonheydt and Johnston, 2006). The pH at which the net edge charge is zero is referred to as the point of zero net charge (PZNC). The median PZNC for kaolinite is close to pH 7 (Blockhaus et al., 1997; Herrington et al., 1992; Rand and Melton, 1975). Thus, the edge surface is positively charged when the solution pH is more acidic than the PZNC and negatively charged when the solution pH exceeds the PZNC. In highly alkaline solutions (pH  >  9), the edge silanol groups can also lose their protons and contribute to the overall negative surface charge (Brady et al., 1996; Ma and Eggleton, 1999; Schoonheydt and Johnston, 2006). These points are illustrated in Figure 1.6, based on the early work by Schofield and Samson (1953).

Figure 1.6 Diagram showing the development of positive and negative charges at the edge surface of kaolinite particles in aqueous suspension in response to changes in medium pH. PZC, point of zero charge. Vertical broken lines indicate the extent of the coordination layer or ‘inner Helmholtz plane’ (IHP).

The amphoteric nature of the edge surface of kaolinite and phyllosilicates, in general, has an important bearing on the colloidal properties of clay mineral suspensions (Lagaly, 2006; Tombácz and Szekeres, 2006; van Olphen, 1977). The edge surface in kaolinite makes up 12–36%, but can constitute as much as 50%, of the total particle area of some 11–50 m²/g (Brady et al., 1996; Lee et al., 1991; Lietard et al., 1980), whereas the corresponding value for montmorillonite seldom exceeds 5%. The influence of the edge surface on various chemical processes, such as ligand exchange with inorganic and organic anions as well as colour and polymerization reactions, is, therefore, more pronounced for kaolinite than for montmorillonite (Blockhaus et al., 1997; Solomon, 1968; Theng, 1971; Theng and Walker, 1970; Tombácz et al., 1998). We should also point out that the basal siloxane surface of kaolinite is essentially hydrophobic, whereas the aluminol surface is hydrophilic (Schoonheydt and Johnston, 2006).

1.1.1.2 Halloysite

For the sake of convenience, the structural properties of halloysite are described at this juncture although the particles of halloysite are for the most part non-planar (Joussein et al., 2005). Like kaolinite, halloysite is a dioctahedral 1:1 phyllosilicate (Table 1.1), but the individual layers in (both planar and non-planar) halloysite are, or have been, separated by a monolayer of water molecules (Figure 1.7). From a survey of the literature, Churchman and Carr (1975) have concluded that the presence of interlayer water in its natural state, or during its past history, is the single, most characteristic feature that distinguishes halloysite from kaolinite. Halloysites can occur with varying degrees of hydration, ranging from the fully hydrated form with a unit formula of Al2Si2O5(OH)4·2H2O and a basal spacing of 1.01 nm, referred to as ‘halloysite-(10   Å)’, to the fully dehydrated form with a unit formula of Al2Si2O5(OH)4 and a spacing of about 0.72 nm, termed ‘halloysite-(7   Å)’. The difference of 0.29 nm (in basal spacing) between the two end members corresponds to the thickness of the interlayer water sheet and equals the van der Waals diameter of a water molecule. The interlayer water content of halloysite-(10   Å) is 12.3 wt%, corresponding to two water molecules per formula unit. Halloysite-(10   Å) can irreversibly lose its interlayer water by exposure to dry air, evacuation at room temperature or mild heating in air. This observation and the apparently continuous manner in which halloysite-(10   Å) dehydrates to halloysite-(7   Å) would indicate that the water molecules are weakly held.

Figure 1.7 Diagrammatic sketch of the halloysite-(10   Å) structure as proposed by Hendricks and Jefferson (1938) showing the presence in the interlayer space of a single sheet of water molecules interacting with each other and with opposing silicate layers through hydrogen bonding.

Modified after Grim (1968).

As compared with kaolinite, the stacking of successive 1:1-type layers in halloysites is disordered in both the a- and b-directions. Further, halloysite particles are more often tubular than platy (Figure 1.8), while spheroidal forms are not uncommon (Dixon, 1989). The iron content is an important factor determining the b-dimension and particle shape of halloysite (Churchman and Theng, 1984; Joussein et al., 2005). Here again, however, there is a continuous gradation in degree of mutual layer displacement, crystallinity and particle morphology. Polymeric nanocomposites with halloysite (cf. Chapter 7) have almost invariably been obtained using the long tubular form. The lengths of halloysite tubes may vary between 0.02 and 30 μm, while their widths range from 0.05 to 0.2 μm (Joussein et al., 2005). Small tubes with a width of <  80 nm contain central cylindrical pores of 5–15 nm in diameter, but such pores are largely absent from large tubes with a width of >  1000 nm (Churchman et al., 1995).

Figure 1.8 Transmission electron micrograph of tubular halloysite from Sasso (Italy). The sample also contains some platy kaolinite particles.

From Bergaya and Lagaly (2006).

The cation exchange capacity of halloysites, ranging from 2 to 60 cmol/kg (Joussein et al., 2005), may arise from a small amount of Al³  + for Si⁴  + substitution in the tetrahedral sheet, and of Fe²  + for Al³  + substitution in octahedral sites. In addition, there is some non-stoichiometric substitution of Fe³  + for Al³  +, giving rise to vacancies in the octahedral sheet (Soma et al., 1992). Solid-state ²⁷Al NMR spectroscopy, however, shows only limited substitution of Al³  + for Si⁴  + in halloysites (Newman et al., 1994). Thus, the high charge (and potassium selectivity) of some halloysites may be due to the presence of 2:1 phyllosilicate contaminants (Takahashi et al., 2001). Rheological measurements by Theng and Wells (1995) indicate that the PZNC of the edge surface of halloysite ranges from 6.0 to 7.1 depending on particle shape.

Bailey (1990) has proposed that halloysite contains some interlayer cations to balance the negative layer charge. By analogy with smectites, these cations would attract their complement of hydration water and hence account for the formation of halloysite-(10   Å). Alternatively, Joussein et al. (2005) have proposed that the ability of halloysite to hold water may be likened to that of kerolite, a hydrated variety of talc (Table 1.1). Like halloysite, kerolite has a disordered layer stacking and a small negative layer charge. Also as in halloysite, most of the water in kerolite is weakly bound to the basal and edge surface of the particles (Miller et al., 1991).

As might be expected, a wide variety of polar, non-ionic organic compounds can intercalate into halloysite-(10   Å) by displacing the interlayer water. In common with kaolinite, halloysite-(7   Å) can also intercalate organic molecules that are highly polar, or capable of forming hydrogen bonds with the interlayer surface. Since the rate of intercalation is generally much faster than that for kaolinite, intercalation methods have been developed for distinguishing halloysite-(7   Å) from kaolinite (Churchman et al., 1984; Theng et al., 1984). Because of the many factors involved, however, no unique test is ideally suited for making this differentiation (Joussein et al., 2005). Halloysite differs from montmorillonite in that it does not intercalate more than a monolayer of water and organic molecules, whereas montmorillonite can form multiple layers (Lagaly et al., 2006; Theng, 1974). Whether halloysite can intercalate organic polymers, however, is an open question. In common with kaolinite (cf. Chapter 7), halloysite would be expected to facilitate the in situ intercalative polymerization of monomers that can displace pre-intercalated (guest) compounds, but hardly, any information is available on this point.

1.1.2 Planar 2:1 Phyllosilicates

1.1.2.1 Smectite

Of the seven groups within the planar 2:1-type phyllosilicate class, only the minerals in the talc–pyrophyllite group have no layer charge (x  ∼  0) (Table 1.1). Thus, summation of the anionic and cationic charges in a unit formula of Al2Si4O10(OH)2 for dioctahedral pyrophyllite shows that its layer structure is electrically neutral (Figure 1.4). Likewise, the layer structure of trioctahedral talc with a unit composition of Mg3Si4O10(OH)2 is uncharged. As a result, both pyrophyllite and talc have no (hydrated) exchangeable cations in the interlayer space, and their surface is essentially hydrophobic. By the same token, pyrophyllite and talc do not show interlayer swelling in water since the contiguous layers in a particle are strongly held together by van der Waals and electrostatic attractive forces (Giese, 1975).

Of the other six groups in the class, the smectites are the most important in terms of the clay–polymer interaction. Within the smectite group, dioctahedral montmorillonite has held the limelight (Table 1.1). This is because montmorillonite particles can expand in water, making its extensive interlayer surface area (700–800 m²/g) accessible to a wide range and variety of guest molecules. Because of its high cation exchange capacity, montmorillonite also serves as nature’s own model for a planar, negatively charged colloid (van Olphen, 1977). The interactions of montmorillonite with small, well-defined organic compounds have been reviewed at regular intervals (Lagaly et al., 2006; Mortland, 1970; Raussell-Colom and Serratosa, 1987; Theng, 1974; Yariv and Cross, 2002), while its interactions with organic polymers are discussed in the following chapters of this book.

The layer structure of montmorillonite was deduced by Hofmann et al. (1933) on the basis of its similarity to that of pyrophyllite. Unlike pyrophyllite, however, montmorillonite has a moderate amount of isomorphous substitution in its structure, giving a layer charge (x) of 0.2–0.6 (Table 1.1). Further, the stacking of successive layers in a montmorillonite particle is more or less random (Hendricks, 1942; Maegdefrau and Hofmann, 1937; Marshall, 1935). The Hofmann–Endell–Wilm–Marshall–Maegdefrau–Hendricks structure, or ‘Hofmann structure’ for short, with a structural formula of M+y·nH2O (Al2  −  yMgy)(Si4)O10(OH)2 is generally accepted. In this ideal composition, the negative layer charge is entirely due to isomorphous substitution (of Al³  + for Mg²  +) in the octahedral sheet and balanced by hydrated monovalent cations (M+y·nH2O) in the interlayer space (Figure 1.9). In reality, however, montmorillonite has some isomorphous substitution in the tetrahedral sheet and contains about 3% by weight of structural iron. Indeed, 15–50% of the total layer charge in the ‘Chambers-type’ montmorilonite is tetrahedral in origin (Schultz, 1969).

Figure 1.9 The ‘Hofmann’ layer structure of montmorillonite, showing isomorphous substitution in the octahedral and tetrahedral sheets, and the presence of exchangeable cations and water molecules in the interlayer space. The basal or d(001) spacing varies according to the nature of the interlayer cation and the ambient relative humidity. Under the conditions of complete dehydration and layer collapse, the basal spacing is ∼  0.95   nm.

In this context, we should mention that the magnitude of the negative layer charge in smectites is an average over the whole particle. In other words, the value of x (Table 1.1) commonly varies (between certain limits) from layer to layer. Indeed, only few montmorillonite samples show a homogeneous charge distribution (Lagaly and Weiss, 1976). The layer charge heterogeneity has implications for the synthesis of polymer–montmorillonite nanocomposites (cf. Chapter 7). Further, the (negative) basal charge of smectites is independent of suspension pH since it arises from isomorphous substitution within the structure. With reference to Figure 1.6, the edge surface of smectite particles also carries a small pH-dependent charge, amounting to less than 10% of the total charge or cation exchange capacity (García-García et al., 2009; Ras et al., 2007).

The other important members of the smectite group that feature in the clay–polymer interaction are dioctahedral beidellite and trioctahedral hectorite (Table 1.1). The unit formula (composition) of the ideal beidellite structure may be written as M+x·nH2O (Al2)(Si4  −  xAlx)O10(OH)2, indicating that the negative layer charge arises from partial substitution of Al³  + for Si⁴  + in the tetrahedral sheet. The layer structure of hectorite is based on that of talc, but hectorite is negatively charged due to some substitution of Li+ for Mg²  + in the octahedral sheet. Thus, hectorite with the ideal composition of M+x·nH2O (Mg3  −  xLix)(Si4)O10(OH)2 may be regarded as the trioctahedral analogue of montmorillonite. Some naturally occurring hectorites have fluoride replacing hydroxyl in the structure. Synthetic fluorohectorite has often been used in the preparation of polymeric nanocomposites (cf. Chapter 7). A synthetic hectorite with a unit formula of Na+0.27(Mg2.73Li0.27)(Si4)O10(OH)2 is marketed under the trade name of Laponite.

1.1.2.2 Vermiculite

Vermiculites may occur as both micro- and macro-size particles. Fine-grained vermiculite-like minerals or ‘clay vermiculites’ may be either dioctahedral or trioctahedral and are widely distributed in soils of different climatic regions (Douglas, 1989). On the other hand, macroscopic vermiculites from which most of our information on structural and surface properties has derived are invariably trioctahedral and apparently do not occur in soil (Walker, 1975).

The 2:1 (T–O–T) layer of vermiculite is similar in structure to that of biotite in that Mg²  + is the dominant cation occupying octahedral positions. Some substitution of Fe³  + and/or A1³  + for Mg²  + may occur, giving rise to an excess of positive charge in the octahedral sheet. The extensive replacement of Si⁴  + by Al³  + in the tetrahedral sheet gives rise to a net negative layer charge (x) of 0.6–0.9 per formula unit (Table 1.1) and a cation exchange capacity of 150–250 cmol/kg. In naturally occurring vermiculites, the counterions balancing the negative layer charge are commonly magnesium. These points are illustrated by the composition of two well-known and much investigated vermiculites, namely, Mg²  +0.32·4.32H2O (Mg2.36Fe³  +0.48A10.16)(Si2.72A11.28)O10(OH)2 for the specimen from Kenya and Mg²  +0.47·4.72H2O (Mg2.83Fe³  +0.01A10.15)(Si2.86A11.14)O10(OH)2 representing the specimen from Llano, Texas (Douglas, 1989).

Figure 1.10 shows a structural diagram of Kenya vermiculite, deduced by Mathieson and Walker (1954) from single-crystal X-ray diffraction analysis. The interlayer Mg²  + ions are positioned midway between every two opposing silicate layers and occupy one-third of the possible cation exchange sites. Under ambient conditions of humidity and temperature, the interlayer space also contains two superposed sheets of water molecules. These are arranged in a regular hexagonal network around each Mg²  + ion and interact mutually and with the surface oxygens of opposing silicate layers by means of hydrogen bonding, giving a basal spacing of 1.43–1.44 nm.

Figure 1.10 Diagram showing the structure of a two-sheet hydrate of magnesium–vermiculite according to Mathieson and Walker (1954). Q and R denote silicate half layers.

As in smectite, the basal spacing of vermiculite is dependent on the nature of the inorganic counterion and the relative humidity of the atmosphere (Walker, 1975). For practical purposes, however, vermiculites may occur in one of three stable hydration states: fully dehydrated, partially hydrated and fully hydrated. The respective basal spacings are 0.9–1.0, ∼  1.2, and 1.4–1.5 nm corresponding to the presence in the interlayer space of zero, one sheet, and two sheets of water molecules. Among the inorganic forms of vermiculite, only the lithium-saturated variety shows extensive interlayer swelling in water (Section 1.2). Vermiculites saturated with n-propyl- and n-butyl-ammonium ions behave similarly when the particles are immersed in water or dilute solutions of the corresponding n-alkylammonium chlorides (Garrett and Walker, 1962; Walker, 1960). In single crystals of vermiculite, the individual silicate layers maintain their mutually parallel orientation although separated by several tens of nanometres (Garrett and Walker, 1967). The large interlayer separations, obtained by this means, would allow and facilitate the intercalation of polymers from solution, but surprisingly little information is available on this topic.

The interlayer inorganic cations in vermiculite may be exchanged for organic cations, such as monomethylammonium, dimethylammonium, tetramethylammonium, tetramethyl-phosphonium and anilinium (Slade et al. 1987; Vahedi-Faridi and Guggenheim, 1997, 1999a,b). As the organic cations are packed in a certain fashion, the stacking order of silicate layers within a particle (‘crystallite’) tends to increase, giving rise to sharp reflections in the X-ray diffraction pattern. Other organically modified forms of vermiculite have been used in the preparation of polymeric nanocomposites (cf. Chapter 7).

1.1.2.3 Mica, Illite and Chlorite

Like vermiculite, macroscopic forms of (true) mica and chlorite are rarely found in soils, if at all. However, fine-grained micaceous minerals occur in nearly all soils. Indeed, in terms of their abundance and distribution, these minerals are the most important soil constituents (Fanning et al., 1989; von Reichenbach and Rich, 1975). Clay chlorites have also been reported to occur in some soils.

Muscovite and phlogopite are two well-known species within the true mica group. The structure (unit formula) of an ideal muscovite may be written as K+ Al2(Si3Al)O10(OH)2, indicating that one of every four Si⁴  + in the tetrahedral sheet is replaced by A1³  +. The resultant negative layer charge is balanced by interlayer K+ ions, giving a layer thickness or basal spacing of 1 nm (Figure 1.11) that remains unchanged on treatment with ethylene glycol (or glycerol), and heating to 500   °C. The K+ ions are in 12-fold coordination, situated centrally on the lines joining the centres of the ditrigonal network of oxygen ions of opposing silicate layers. Muscovite is dioctahedral, whereas phlogopite (K+·Mg3(Si3A1)O10(OH)2) is trioctahedral. The interlayer cations in mica are generally not exchangeable although they may be replaced by alkylammonium ions under special conditions (Mackintosh et al., 1971). Organically modified mica and fluormica have been used in the synthesis of polymeric nanocomposites (Tamura et al., 2008; Uno et al., 2009; Yokoyama et al., 2009).

Figure 1.11 The layer structure of muscovite (mica) viewed along the a-axis with interlayer potassium ions balancing the negative layer charge, giving a basal or d(001) spacing of 1.0 nm.

The term ‘illite’ denotes a group of clay-size or fine-grained micaceous minerals with a basal spacing of ∼  1 nm that does not increase on treatment with ethylene glycol (Grim et al., 1937). A more precise description, proposed by Środoń and Eberl (1984), is that of ‘a non-expanding, dioctahedral, aluminous, potassium mica-like mineral’ that occurs in the <  4 μm fraction of soils and sediments. An ‘average’ composition for illite, derived from chemical analyses of a large number of samples, may be written as (Ca²  +0.05Na+0.03K+0.63) (Al1.55Fe³  +0.21Fe²  +0.04Mg0.28) (Si3.41Al0.59)O10(OH)2 (Weaver and Pollard, 1973). This shows that much of the isomorphous substitution (of Al³  + for Si⁴  +) occurs in the tetrahedral sheet with some replacement of Al³  + by Mg²  +, Fe²  + and Fe³  + taking place in octahedral positions. Thus, illite is structurally similar to muscovite but has a smaller negative layer charge and less potassium than muscovite.

Besides showing a variety of chemical compositions, many illite samples are essentially mixed-layer structures, with smectite usually making up the other component although the presence of interstratified smectite layers is difficult to detect when their proportion is less than 10%. Since illite contains more water than macroscopic mica, the terms ‘illite’, ‘hydro-mica’ and ‘hydrous mica’ have been used interchangeably in the literature. The nature and origin of the ‘extra’ water in illite are not well understood although part, if not all, of it may be present as hydronium (H3O+) ions, occupying interlayer K+-deficient positions (Loucks, 1991; Nieto et al., 2010).

If we accept the above composition as being representative of illite, the negative layer charge (x) of 0.76 per formula unit falls within the range given by vermiculite (Table 1.1). Unlike vermiculite, however, the interlayer counterions in illite are practically non-exchangeable. Accordingly, the cation exchange capacity of illite is much smaller than its layer charge would suggest. In this respect, and in showing a basal spacing of ∼  1.0 nm that is unaffected by heating, illite is more akin to mica and less like vermiculite. In this context, we should mention that when dry illite is wetted, its layer structure contracts (along the c-axis) by ≤  0.02 nm, probably due to deprotonation of interlayer hydronium ions (Kühnel et al., in preparation).

In terms of layer charge (x-value), the brittle mica may be regarded as representing 2:1-type phyllosilicates at the other end of the scale. Here, there is extensive isomorphous substitution (x  ∼  1.8–2.0) and the negative layer charge is balanced by interlayer non-hydrated divalent cations (Table 1.1). Because of the strong electrostatic attraction between successive layers, these minerals do not swell in water and their interlayer cations are not exchangeable.

The majority of chlorites have a trioctahedral layer structure, the particles being made up of successive 2:1 mica-type layers that are interleaved with a continuous magnesium hydroxide sheet. Since the thickness of the interlayer sheet is ∼  0.4 nm, the basal spacing of chlorite is close to 1.4 nm (Figure 1.12). This value is comparable to the d(001) spacing of a two-layer hydrate of Mg²  +-vermiculite depicted in Figure 1.10. Unlike vermiculite, however, the basal spacing of chlorite remains unchanged on treatment with ethylene glycol or heating up to 500 °C.

Figure 1.12 The structure of chlorite projected on the ac plane showing the presence of a hydroxide sheet between two repeating (successive) 2:1-type layers. Note: 1   Å  =  0.1   nm.

From Grim (1968).

As in vermiculite, there is much substitution of A1³  + for Si⁴  + in the tetrahedral sheet of the chlorite structure. Although the resultant negative charge is partly offset by substitution of A1³  + for Mg²  + in octahedral sites, most of this charge is balanced by intercalation of a positively charged brucite-like sheet (Barnhisel and Bertsch, 1989). The composition of this sheet may be written as [(Mg3  −  xAlx) (OH)6]x+, indicating partial substitution of Al³  + for Mg²  +. Strong interaction of the interlayer hydroxide sheet with opposing silicate layers through electrostatic attraction and hydrogen bonding would account for the stability of chlorite to heating and its failure to expand in water and organic liquids. Further, the requirement for (long) hydrogen bonding between the hydroxyl groups of the interlayer sheet and the oxygens of opposing silicate layers restricts the number of possible layer stacking arrangements (Bailey and Brown, 1962).

1.1.3 Non-planar Phyllosilicates

1.1.3.1 Sepiolite and Palygorskite

Sepiolite and palygorskite (‘attapulgite’) are hydrous magnesium silicates with a fibrous particle morphology (Figure 1.13) and a characteristic dehydration pattern. Individual fibres, measuring about 1 μm in length and 0.01 μm in diameter, are made up of a bundle of fibrils. Each fibril consists of 2:1-type layers, arranged in ribbons (also referred to as ‘chains’) along the fibril (a-) axis, with the apical oxygens of the silicon tetrahedra in adjacent ribbons pointing in opposite directions. For this reason, sepiolite and palygorskite have been referred to as ribbon- or chain-layer silicates. Being joined at the corners through oxygen ions, the rectangular ribbons alternate with channels of similar size and shape, running the full length of the fibril. Under ambient conditions of humidity and temperature, these channels contain free (‘zeolitic’) water and exchangeable counterions (Brigatti et al., 2006; Singer, 1989).

Figure 1.13 Transmission electron micrograph of fibrous palygorskite from Southern Georgia.

After Bergaya and Lagaly (2006).

Figure 1.14 shows the sepiolite structure as proposed by Brauner and Preisinger (1956) having an ideal composition of Mg8Si12O30(OH)4(OH2)4·8H2O. The unit cell with a b- and c-dimension of 2.7 and 1.34 nm, respectively, contains four ‘bound’ and eight zeolitic water molecules The former are coordinated to octahedral magnesium ions, exposed at the ribbon edge. In practice, however, not all eight octahedral sites may be occupied. In addition, there is always some substitution of Fe³  + and Fe²  + for Mg²  + in the octahedral sheet, and of A1³  + and Fe³  + for Si⁴  + in tetrahedral positions. The resultant net negative charge is usually balanced by extraneous Ca²  + ions, giving a cation exchange capacity of about 20–45 cmol/kg (Weaver and Pollard, 1973).

Figure 1.14 The structure of sepiolite according to Brauner and Preisinger (1956). The unit cell dimension is indicated by the solid rectangle. At the edge of the mineral fibre, the bound water molecules interact with neighbouring silanol groups (1) or oxygen ions (2), as suggested by Serna et al. (1975).

Figure 1.15 shows the structure of palygorskite as proposed by Bradley (1940). It has the ideal composition of (Mg,Al)5Si8O20(OH)2(OH2)4·4H2O, indicating that the mineral is intermediate between di- and trioctahedral, whereas sepiolite is nearly fully trioctahedral. Recent measurements by Garcia-Romero and Suárez (2010), however, indicate no compositional gap exists between sepiolite and palygorskite. A more important structural difference is that each ribbon in palygorskite consists of five octahedral sites, giving rise to a b-dimension of 1.8 nm as compared with ∼  2.7 nm in sepiolite. Accordingly, the channel dimensions of palygorskite (0.37 nm  ×  0.64 nm) are appreciably smaller than those of sepiolite (0.37 nm  ×  1.06 nm) (Singer, 1989). In addition, palygorskite has relatively less isomorphous substitution within the structure, giving rise to a cation exchange capacity of 5–30 cmol/kg (Weaver and Pollard, 1973).

Figure 1.15 The structure of palygorskite, showing the unit cell dimension (solid rectangle) according to Bradley (1940). Circles with dark outlines represent zeolitic water molecules. Note: 1   Å  =  0.1   nm.

After Grim (1968).

Besides being characterized by the fibrous appearance under the electron microscope, the differential thermal analysis curves of sepiolite and palygorskite are diagnostic for their occurrence in admixtures with planar layer silicates. The zeolitic water in sepiolite can be removed by mild heating (<  200 °C in air). At the completion of this step, the bound water molecules interact with neighbouring hydroxyls or oxygen ions through hydrogen bonding as indicated by the broken lines, marked (1) and (2), in Figure 1.14. The second step corresponds to the loss of half of the bound water, that is, one molecule from each ribbon edge site. When this occurs (∼  300 °C in air or ∼  175 °C under vacuum), the network of hydrogen bonds is disrupted, allowing the ribbons to rotate and the crystal to fold. This process leads to a marked reduction in surface area and a collapse of the channel structure (Dandy and Nadiye-Tabbiruka, 1975). At this stage of dehydration, the original structure can be restored by allowing the crystal to take up water at ambient temperature and humidity. The third dehydration step occurs between 380 and 680 °C in air when the other half of the bound water is lost to yield the ‘anhydride’. This form can only be rehydrated by exposing the crystal to water vapour at elevated temperatures and pressures, and then often incompletely (Hayashi et al., 1969; Nagata et al., 1974; Serna et al., 1975; Van Scoyoc et al., 1979). The fourth step, which occurs between 680 and 900 °C, can be identified with dehydroxylation when the original structure is destroyed and a new crystalline or amorphous phase is formed (Brigatti et al., 2006; Hayashi et al., 1969; Preisinger, 1959).

Palygorskite behaves similarly on heating, but the loss of bound water and hence crystal folding occur at a lower temperature than in sepiolite. Further, step 3 may not always be clearly discernible in the differential thermal and thermogravimetric analysis patterns because the high-temperature dehydration step in palygorskite may overlap with structural