FEATURE ARTICLE Crystallisation inside fullerene related structures Jeremy Sloan,a,b Jessica Cook,a† Malcolm L. H. Green,*a John L. Hutchisonb and Reshef Tennec aInorganic Chemistry L aboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QR bDepartment ofMaterials, University of Oxford, Parks Road, Oxford, UK OX1 3PH cDepartment ofMaterials and Interfaces, Weizmann Institute, Rehovot 76100, Israel The dierent methods for encapsulating crystalline materials inside fullerene related structures are reviewed.The relationships between the mode of encapsulation and the crystallisation behaviour obtained in each case are described. In particular, the mechanisms of morphological and orientational control of crystallite growth inside carbon nanotubes and the comparative encapsulation behaviours of materials encapsulated by physical and catalytic methods are described and discussed.The encapsulation of defect tungsten oxide structures within inorganic fullerene-like structures are also described. Since the discovery of carbon nanotubes in 1991,1 a number opened carbon nanotubes.2–4 Once encapsulated, the materials can be further modified in situ to give reduced,33 oxidised34 or of researchers have encapsulated both crystalline and noncrystalline materials inside their cavities using either chemical otherwise chemically modified encapsulates.The second strategy involves formation of the encapsulating medium around or physical methods.2–5 Carbon encapsulation has also been achieved in situ by the arc-evaporation of composite carbon the included material.This may be achieved catalytically, either in situ in a carbon arc or by gas-phase deposition onto catalytic electrodes resulting in the formation of sealed carbon coated species contained either within the cores of carbon nanopart- metal particles; or chemically, via formation of the encapsulating material from the outside surface of the encapsulated icles6–12 or as continuous filling, or ‘nanowires’, formed along the internal bore of nanotubes.13–15 Encapsulation has also material.A third strategy can also be identified, although this is a special case restricted to the encapsulation of diamond been achieved via gas-phase deposition of carbon containing species onto catalytic metal particles.16–19 In related develop- only.Banhart and Ajayan36 recently demonstrated that the cores of carbon onions can behave like high pressure cells ments, encapsulates contained within fullerene-like cage structures of the general form MX2 (M=W, Mo; X=S, Se) have in which diamond formation can be induced by irradiation with a high energy (1.25 MeV) electron beam at elevated been synthesised in which the encapsulating material is grown chemically inwards from the surface of finely divided and temperatures.partially reduced material.20–25 Encapsulation research has been directed mainly towards Chemical insertion of materials inside carbon nanotubes the enhancement of the electrical2,6–8,13,15,25,26 and mag- Chemical insertion involves the selective opening of the carbon netic10,27,28 properties of both the encapsulated and encapsulat- nanotubes at their tips, either by refluxing in concentrated ing materials although they are also being considered for nitric acid or, alternatively, by heating in O23 or CO2,25 and applications in fields as diverse as biotechnology,29 and cataly- then precipitating from solution the encapsulated material sis.4,29–32 Benefits arising from this researchhave so far included inside the opened cavities.This can be achieved via a one-step the ability to observe the in situ chemistry of encapsulated procedure4 in which the opening reagent contains a solubilised materials33 and also size limited crystallisation behaviour on metal nitrate that precipitates upon calcination to form an an approaching atomic scale.11,34,35 In this article, it is the encapsulated metal oxide.Alternatively, a two-step pro- latter with which we shall be most concerned. We will attempt cedure37 may be utilised in which the closed nanotubes are to elucidate some of the relationships between modes of first treated with nitric acid and then heated in air to remove encapsulation and the crystallisation behaviour obtained in surface acid groups known to be present on nanotubes opened each case.Some of the more interesting and unusual crystallis- in this way.37 The nanotubes can subsequently be filled by ation behaviour exhibited by encapsulated species will also be stirring with a solution of a metal nitrate or the metal halide, described and discussed. It is hoped that these phenomena will followed by calcination.This technique is useful for encapsulat- contribute to an understanding of how materials formation ing materials that can interact unfavourably with surface acid can be manipulated at the most intimatescale and, additionally, groups, such as the metal halides. how new types of materials with hybrid physical properties Oxides of the metals Ni,4 U,4 Co,4 Fe,4 Nd,38 Sm,38 Eu,38 can be created.La,38 Ce,38 Y38 and Cd38 have all been encapsulated via the one-step procedure. The mixed-metal oxide FeBiO3 has been Methods of encapsulating materials inside similarly encapsulated38 by calcining a mixture containing an fullerene related materials equimolar solution of iron and bismuth nitrates, concentrated HNO3 and closed nanotubes.Pd,37 Ag,39 Au39 and AuCl39 Two main strategies can be identified for encapsulating mate- have all been encapsulated using the two-step procedure using, rials inside fullerene related structures. The first involves in the case of the latter three, their respective metal halides. In inserting the materials, either physically or chemically, into a modification to the two-step procedure,34 polycrystalline SnO has been encapsulated by mixing opened nanotubes with SnCl2 in acidic solution to which was added a weak base, † Present address: M.I.T., 77 Massachusetts Ave., Room 6-329, Cambridge, MA 02139, USA.resulting in precipitation at a pH of 10.2. J. Mater. Chem., 1997, 7(7), 1089–1095 1089Encapsulated metal oxides can be reduced to their respective Dai et al.61 have deliberately prepared encapsulated nanowires containing copper and germanium by pyrolysis of polycyclic metals by reduction with hydrogen gas at elevated temperatures. Encapsulated Ni metal has been prepared in this aromatic hydrocarbons over finely divided copper and germanium, respectively.fashion.27 Similarly, hydrogen reduction of tubes filled with precipitated KReO4 gives encapsulated crystallites of rhenium metal.38 Encapsulated metal halides can be further modified in Chemical encapsulation situ by treatment with H2S gas at elevated temperatures to MX2–fullerene-like materials, with M=W or Mo and X=S give their respective sulfides.Encapsulated CdS38 and Au2S338 or Se, consist of a network of 2H-MX2 prisms that, in have both been prepared in this way.projection, resemble the graphitic network common to carbon fullerene related structures such as nanotubes (2H refers to the Physical insertion of materials into nanotubes hexagonal symmetry, which repeats every two layers). The relationship between the two types of structures is shown Opened nanotubes can be filled via capillary action using schematically in Fig. 1. Like nanotubes and other fullerene either a low melting, low surface tension metal salt30 or, derived structures, the interlinked 2H-MX2 hexagonal net- alternatively, a eutectic or low melting mixture34 of two works are capable of incorporating dierent alternative components with a resulting surface tension lower than the polyhedra which, depending on the type, impart either positive threshold value of 100–200 mN m-1.30 When a single molten or negative curvature into the structures resulting ultimately component is used, continuous and preferentially orientated in the formation of nanotube and nanoparticle-like structures.filling of the nanotube cavities is invariably obtained, as has Based on these relationships, such materials are often referred been observed for the oxide phases of lead,2,3 vanadium30 and to as the inorganic fullerenes.molybdenum.40 In the case of the latter, subsequent treatment There are now a number of specific and general methods by of the nanotubes filled by MoO3 with hydrogen at 450–500 °C which inorganic fullerenes can be prepared, including: spon- causes reduction to pure and continuously orientated MoO2 taneous room temperature growth from reduced WS3 soot;22,23 filling.40 When a eutectic or low melting mixture of two gas-phase growth from ion beam sputtered W or Mo films;62 components is employed, as in the case of UCl4/KCl,34 continu- pulsed laser evaporation from non-fullerenic MoS2 films;63,64 ous polycrystalline filling is obtained.Encapsulated UCl4 STM induced growth from amorphous finely divided MoS3 formed in this way hydrolyses slowly in air to give continuous particles;65 and gas-phase growth from partially reduced Mo filling with an oxidised product, U(Cl,O)x .34 or W oxides.20–24 Only in the last two cases are filled or partially filled encapsulates obtained whereas in the other Arc encapsulation instances, hollow inorganic fullerene cage structures are pro- The in situ arc encapsulation technique consists of packing a duced.STM induced MoS2 growth produces encapsulates hollow graphite anode with an element to be encapsulated and containing amorphous MoS3 filling, while gas-phase growth then proceeding with the conventional Kra�tschmer–Human from the reduced Mo and W oxides produces encapsulates arc deposition experiment. During the arcing process, the tip with partially reduced oxide filling. of the anode and its contents are rendered into the vapour phase and the carbon shell then grows catalytically from the Relationships between mode of formation and condensing species.This technique has now been applied to crystallisation behaviour inside fullerene related nearly half of the elements in the Periodic Table.All of the lanthanides, except Pm, Sm and Eu;7,8,10–12,41–45 all the first- structures row transition metals;11,44–49 the platinum group metals Ru,19 Control over crystallite morphology and orientation in carbon Rh,19 Pd,19,50 Os,19 Ir19 and Pt;19 selected other transition nanotubes metals, such as Y,6,12,45 Au,8 Ta,11,26 Mo,11,45,48 W,11,45,51 Nb,45 Zr45,49 and Re;51 the main-group elements Ge,14 Sn,14 Pb,14 The behaviour of crystalline materials formed by precipitation Sb,14 Bi,14 S,145 Se,14 Te,14,45 B,14,45 Si14,45 and Al;45 and the inside carbon nanotubes allows for the study of their morpho- actinides Th52 and U52 have all been encapsulated by the arc logical control over crystallite formation.Initially, crystallis- method, either as their respective carbides or in elemental ation will proceed according to conventional nucleation and form.With the exception of Cr, Dy, Ni and Gd, which are growth mechanisms but once the size of the encapsulated obtained as continuous filling inside nanotubes,11 all of the crystallite approaches that of the internal diameter of the encapsulates are obtained as single crystals encased inside carbon nanoparticles.However, a mixture of the latter type of filling with continuous filling is also observed in the case of the encapsulated metals or carbides of Gd,10 Y12 and Mn.48 Encapsulation via catalytic growth from solid particles Arc encapsulation, as described above, is a very specific mode of catalytic growth that occurs during co-condensation after both the encapsulated and encapsulating materials have been rendered into the vapour phase.The type of catalytic growth described here pertains to gas-phase53–56 carbon deposition onto catalytic particles. The carbon carrier can vary from hydrocarbon gases, such as methane55 and acetylene,57 to aromatics, including benzene,58 to polymers, such as polyethylene, 59 and even more complex organic species.60,61 Only certain metals eciently promote carbon growth from the gas phase and these include Co,16,18,53 Fe,53 Ni,53–55,57–60 Pd,54 Pt,54 Ti,57 Fig. 1 Structural relationship between graphite and 2H-MX2 struc- W,57 Cu61 or Ge.61 In nearly all cases, the goal has been to tures. In both cases, the fullerene derived structures form networks of produce either nanotubes or modified nanotubes, including interlinked hexagonal units which can form positive or negative single walled nanotubes (SWTs), rather than encapsulated curvature based on the number and type of incorporated polyhedra other than hexagons.species which are essentially a by-product. Recently, however, 1090 J. Mater. Chem., 1997, 7(7), 1089–1095nanotube cavity, then the walls will start to exert control over directions must come to an end once it meets the carbon walls.By contrast, crystallite I may continue growing along its any future growth. The precise nature of the control at this juncture raises some interesting questions. How the capillary current axis until either it meets an obstruction or until the crystallisation process is terminated.In Fig. 2(c) a micrograph walls influence the orientational behaviour of growing crystallites and, secondly, how they interact at the atomic level with of a nanotube with a more regular packing of SnO crystallites can be seen.35 In this case the smaller crystallites observe encapsulated crystalline materialsare problems that are worthy of investigation. apparently random orientations while the larger crystallites (denoted IV), all have their d101 lattice fringes orientated at Polycrystalline SnO forms spherical or ellipsoidal encapsulates with diameters in the range 20–60 A° 35 inside carbon 90° with respect to the nanotube wall.The nanotubescapillaries shown in Fig. 2(b) and (c) clearly exert influence over both the nanotubes with internal diameters in the range 20–90 A° .The packing of such crystallites in nanotubes provides some direct morphology and orientation of their encapsulated SnO crystallites. The type of morphological control exhibited by SnO can insight into the mechanisms of control over crystallisation. Fig. 2(a) shows a high-resolution transmission electron micro- sometimes lead to some interesting behaviour inside the capillaries of carbon nanotubes.Fig. 3(a) and (b) show a micrograph graph (HRTEM) of an agglomeration of randomly orientated SnO crystallites inside a nanotube with a large internal diam- and schematic representation of a crystal spiral formed from a chain of single SnO crystallites of approximately equal size eter (ca. 90 A° ).In this instance, the nanotube exerts morphological control over the agglomeration but not over the individual and observed to form inside the capillary of a carbon nanotube. 35 The mode of formation of such a spiral can be explained crystallites. In Fig. 2(b), another micrograph showing a nanotube with a smaller diameter (ca. 35 A° ) can be seen in which completely in terms of the mechanisms of morphological control described above. two well resolved and slightly elongated crystallites I and II reside in the central cavity.A third, less well resolved crystallite Other examples of polycrystalline filling show similar types of morphological and orientational control to SnO. Fig. 4(a) (III) is also visible. Whereas crystallite I has its growth axis (arrowed) aligned parallel to the nanotube axis, crystallite III and (b) show examples of filling with polycrystalline SnO266 and ZrO2,66 respectively. Whereas polycrystalline SnO forms has eectively two growth axes (arrowed) that are both at an angle to the tube axis.Regardless of where the nucleation of small crystallites with diameters approximately equal to the internal diameter and which apparently randomly orientate crystallite II started, the crystal growth in either of the two along the nanotube capillary, polycrystalline ZrO2 crystallites Fig. 3 (a) Spiralling crystal growth, induced by morphological control, observed inside a carbon nanotube capillary. (b) Schematic representation of spiralling crystal growth. Fig. 2 (a) Micrograph illustrating nanotube capillary morphological control over agglomerated SnO crystallites.The crystallites clump together and eectively form a blockage inside the capillary. (b) Micrograph showing oentational control over individual crystallites. I is free to grow along its crystallite axis while the growth of II Fig. 4 (a) Randomly orientated SnO2 crystallites observed inside a is constrained by the nanotube capillary. (c) Random and preferentially orientated SnO crystallites (IV) inside a densely packed carbon nanotube capillary.(b) Preferentially orientated ZrO2 crystallites (arrowed) observed inside a nanotube capillary. nanotube. J. Mater. Chem., 1997, 7(7), 1089–1095 1091are elongated and several are aligned with their lattice fringes et al.30 have noted for V2O5 and Chen et al.40 have similarly noted for MoO3, continuously orientated behaviour is almost parallel to the nanotube axis.In order to attempt to answer the question of how crystalline invariably obtained. The one exception to this is when the filling is achieved using a low melting mixture. In the case of materials interact with carbon nanotube walls, it is instructive to look at well resolved HRTEM images of continuous crystal- UCl4/KCl, continuous polycrystalline filling is obtained.Fig. 6(a) shows an example of continuously orientated MoO3 line filling. Fig. 5(a) shows an elongated Sm2O3 crystallite, that completely fills the internal volume of a nanotube for a distance filling. The spacing of the observed lattice fringes is 3.7 A° , which corresponds to the c lattice repeat of MoO3.40 This of ca. 600 A° .Beneath this image are two enlargements [Fig. 5(b) and (c)] of regions at intervals along the capillary distance is also incommensurate with the atomic periodicity of the carbon nanotube wall along the tube axis (2.3 A° ). Hence where the image contrast reveals the same precise arrangement of the Sm3+ cations (which image much more strongly than crystal growth along the nanotube in this instance is again a function of orientational and morphological control.Fig. 6(b) O2- and resolve as dots) at the interface of the Sm2O3 crystallite with the nanotube wall. The cations closest to the shows an example of polycrystalline filling obtained inside nanotubes filled using a low melting mixture.34 The filling in nanotube wall can be seen to be arranged in a triangular motif that extends along the wall of the carbon nanotube. Due to a this case is polycrystalline with several crystallites (denoted V in the micrograph) apparently observing the same preferred slight tilt in the crystal, the lattice image in Fig. 5(b) images more strongly as lines which actually represent the (400) planes orientation, analogously to the case of SnO and ZrO2, described above.of Sm2O3, although the motif is still just visible. This motif actually represents the point at which parallel Sm2O3 (222) lattice planes terminate along the nanotube wall and this Formation of products inside catalytically formed carbon cages occurs at precise intervals of 5.46 A° (equivalent to d200 of There is still much uncertainty concerning the mechanism of Sm2O3).The precise arrangement is depicted schematically in formation of materials encapsulated by the arc method. Fig. 5(d). If we now look at a two-dimensional projection of Giuerret-Pie�cort et al.11 claim that the propensity for the a carbon nanotube wall [Fig. 5(e)], we see that the unit cell of formation of long metallic ‘nanowires’ rather than encapsulates hexagonal graphite repeats every 2.13 A° along the axis of the inside nanocapsules is correlated with the existence of an carbon nanotube.This is incommensurate with the periodicity incomplete shell in the most stable ionic state of the element. of the (222) lattice plane terminations [shown schematically Saito et al.67 have indicated that there is an additional corre- in Fig. 5(d)]. Thus, in the case of this Sm2O3 crystallite, the lation for lanthanoids between their volatility and their ability orientation and periodical arrangement must be a function of to form encapsulates.Recently however, Seraphin et al.45 have the gross morphological influence of the nanotube capillary indicated that neither of these models are wholly without during crystallisation rather than any influence due to the exceptions and have advanced their own model, defined in atomic arrangement of the nanotube wall.terms of the interfacial compatibility of the carbide with the When materials are inserted into nanotubes by capillary encapsulating graphitic network. The phenomenon of arc action, the orientational behaviour of the encapsulated crystal- encapsulation is further complicated by the fact that, in some line material is much the same as described above.As Ajayan cases, mixed products are often obtained. In the case of encapsulated manganese, for example, Liu and Cowley47 have observed no fewer than four dierent encapsulated carbides, Mn3C, Mn5C2, Mn7C3 and Mn23C6 , some of which are incommensurate and presumably metastable. Similarly Sloan et al.51 have observed the formation of rhenium metal, hexagonal ReC and an unknown metastable RexCy product inside carbon nanocapsules.A micrograph obtained from the latter Fig. 5 (a) 600 A° long Sm2O3 crystallites observed inside the bore of a Fig. 6 (a) Micrograph showing continuous MoO3 filling formed by carbon nanotube. (b), (c) Enlargements of Sm2O3/carbon interface showing periodic stacking of Sm3+ cations. (d) Schematic represen- the capillary method.The lattice fringes repeat at regular intervals of 3.7 A° which is incompatible with the repeat of the graphitic network tation of termination of Sm2O3 (222) lattice planes on carbon wall. (e) Schematic projection of graphitic nanotube wall showing periodicity (2.3 A° ; see Fig. 5). (b) Polycrystalline UCl4 arranged along the bore of a carbon nanotube obtained via eutectic filling.Some crystals show of graphitic network (repeats every 2.3 A° ). This is incommensurate with the repeat of (222) lattice plane terminations. clear preferred orientations (V). 1092 J. Mater. Chem., 1997, 7(7), 1089–1095product is reproduced here [Fig. 7(a)]. Giuerret-Pie�cort et al.11 proceed via a solution–precipitation mechanism,56,68 which involves absorption of carbon onto the surface of the catalytic have observed the formation of both microcrystalline Yb and ‘spiral’ Dy products in their nanowires, both of which are also particle resulting in the formation of a small amount of interfacial carbide, thus leaving the remainder of the encapsu- indicative of a complex formation process.In view of the complexity of the products obtained in these and other cases late in its native elemental state. The carbon then grows progressively from the element/carbide interface. An example cited in this article, it seems unlikely that one model or explanation alone will suce to account for all of the encapsul- of an encapsulated Ni particle formed in the presence of a Ni- Harshaw catalyst at 780 °C, according to conditions specified ation behaviour observed in arc deposited products.A particular practical diculty is the fact that there is at present no by Tsang et al.,55 is shown in Fig. 7(b). eective way of observing in situ the encapsulation process. Perhaps a better approach is to consider each case individually, Novel layered and defect structures observed inside inorganic taking into consideration the complex kinetic and thermo- fullerenes dynamic mechanisms that can be obtained within a particular system.The encapsulation mechanism of sulfide and oxide particles The natures of the encapsulated products obtained in the encapsulated inside inorganic fullerenes is relatively easy to case of gas-phase deposition onto catalytic particles are much interpret for the simple reason that the encapsulating material simpler to interpret than those of species formed by arc co- is formed by consumption of the outside of the encapsulated deposition.In general, encapsulation has been proposed to material. Thus, the growth mechanisms can be interpreted in terms of progressive growth from the exterior.The three dierent mechanisms for STM induced MoS2 growth from MoS3, gas-phase growth of MoS2 from condensing MoO3-d vapour and gas-phase reaction with solid WO3-d are depicted schematically in Fig. 8((b) and (c), respectively. In the case of MoS2 induced growth in the STM, encapsulation results in the formation of amorphous MoS3 encapsulates only. In this instance, the morphology of the resulting encapsulates will be influenced by the morphology of the amorphous precursors.In the case of the formation of Mo and W oxide encapsulates, the first occurs wholly in the gas phase whereas the second occurs as a gas–solid reaction with the oxide remaining in the solid phase. The encapsulation mechanism of MoO3-d will presumably be similar to that exhibited by arc deposited encapsulates (see above), with the whole process occurring during condensation from the vapour phase.The morphology of the encapsulated products will therefore be Fig. 7 (a) Microstructure of the encapsulated metastable RexCy prod- determined by the extent of reaction and the conditions of uct formed via arc co-deposition. (b) Micrograph showing Ni encapsulated by catalytic formation of carbon from the Ni-Harshaw catalyst.condensation. Feldman and coworkers24,69 have observed a Fig. 8 Schemes showing: (a) mechanism of formation of MoS2 induced by STM from MoS3, resulting in the formation of amorphous MoS3 encapsulates; (b) mechanism of gas-phase reaction of H2S with MoO3-d to form encapsulated oxide and empty MoS2 inorganic fullerenes; (c) mechanism of gas–solid reaction of H2S with solid WO3-d to form encapsulated oxide and empty WS2 inorganic fullerenes.All mechanisms eventually produce empty nested inorganic fullerenes (adapted from Feldman et al.69). J. Mater. Chem., 1997, 7(7), 1089–1095 1093variety of products, from MoS2 fullerenes, to MoS2 nanotubes, to MoS2 encapsulated MoO3-d, obtained precisely by varying these conditions. In the case of the formation of WS2 encapsulated WO3-d, however, the situation is made more interesting by the fact that encapsulation occurs via reaction of H2S gas with the reduced WO3 solid in a situation analogous to the gas-phase carbon deposition onto solid catalytic particles, described above.Under these conditions, the overall morphology of the encapsulated species will be determined by the morphology of the reduced WO3-d precursor particles.An example of this phenomenon is shown in the micrograph in Fig. 9(a). The encapsulate in this case is a ‘bent’ crystallite of WO3-d encapsulated by a ‘skin’ of WS2 fullerenic material. Closer inspection of the crystallite reveals that the bend in the crystallite is due to several grain boundaries (arrowed) that occur as defects within the crystal.A further interesting feature of these encapsulates is that, as partially reduced WO3, there Fig. 10 Layered and lamellar reduced tungsten oxide structure is always the potential that other types of defects and structural observed inside WS2 nested fullerene features may be observed inside encapsulates. This is in fact now the case.70 In Fig. 9(b), an example of an encapsulated structure of the Wadsley defect type is shown. The shear planes sharing WO3 octahedra network partially collapsing to form observed in this encapsulate arise as a result of the corner- shear planes comprised of edge-sharing WO3 octahedra. A schematic representation of the encapsulated defect structure is shown in Fig. 9(c). Previously, this type of defect structure has only been observed in reduced single crystals of reduced WO3.71 Another new type of encapsulated structure is shown in Fig. 10. Inside a bilayer of WS2 material, a novel layered tungsten oxide can clearly be seen. This is the first example of a complex layered structure formed inside a fullerene related structure. A more detailed analysis of the structure of this encapsulate will appear elsewhere.70 Concluding remarks The types of encapsulation behaviour discussed in this article raise the prospect of manipulating materials formation at the most intimate scale.The mechanisms of morphological control of nanotube capillaries over crystallite formation contribute to an understanding of how encapsulates such as nanowires can form inside nanotubes.The main benefit of catalytic encapsulation of materials, either in situ or by gas-phase deposition onto catalytic particles, is in the coating of reactive or airsensitive species. In the case of in situ formed materials, an additional benefit lies in the fact that many of the encapsulated products are metastable and are often not readily obtainable by other methods.The observation of defect and layered materials inside inorganic fullerenes presents a wholly new perspective in encapsulation technology. These materials represent, for the first time, the first realistic possibility for incorporating complex layered or defect structures, with one or more useful properties, inside a wholly dierent type of structure, with completely dierent properties, thus generating a new class of hybrid materials. The authors are indebted to Dr.Edman S. C. Tsang, of Reading University, for supplying the encapsulated MoO3 specimen, to Dr. Jens Hammer and Rufus Heesom, of Oxford University, for preparing the encapsulated rhenium carbide specimen, to Dr. Andrew P. E. York, of the Inorganic Chemistry Laboratory, for the provision of the Ni-Harshaw specimen and also to Moshe Homyonfer and Yishai Feldman, of the Weizmann Institute, who prepared the WO3-d encapsulated WS2 specimen.Fig. 9 (a) External morphology of encapsulate induced by defect structure of reduced oxide encapsulate. WS2 skin follows the bend in the WO3-d crystal induced by grain boundaries (arrowed). (b)Wadsley References defect structure incorporated inside WS2 nested fullerene.(c) Schematic representation of (b) showing how the partial collapse of WO3 network 1 S. Iijima, Nature (L ondon), 1991, 354, 56. 2 P.M. Ajayan and S. Iijima, Nature (L ondon), 1993, 361, 333. leads to the formation of shear planes. The total number of shear planes depicted is 4, but there are no less than 18 inside the imaged 3 P.M. Ajayan, T. W. Ebbesen, T. Ichihashi, S. Iijima, K. Tanigaki and H. Hiura, Nature (L ondon), 1993, 362, 522. structure. 1094 J. Mater. Chem., 1997, 7(7), 1089–10954 S. C. Tsang, Y. K. Chen, P. J. F. Harris and M. L. H. Green, Nature 39 A. Chu, J. Cook, R. Heesom, J. L. Hutchison, M. L. H. Green and (L ondon), 1994, 372, 159. J. Sloan, Chem.Mater., 1996, in press. 5 S.C. Tsang, J. J. Davis, M. L. H. Green, H. A. O. Hill, Y. C. Leung 40 Y. K. Chen, M. L. H. Green and S. C. Tsang, Chem. Commun., and P. J. Sadler, J. Chem. Soc., Chem. Commun., 1995, 2579. 1996, 2489. 6 S. Seraphin, D. Zhou, J. Jiao, J. C. Withers and R. Loufty, Nature 41 Y. Yosida, S. Shida, T. Oshuna and N. Shiraga, Abstr. FallMeeting (L ondon), 1993, 362, 503. Phys. Soc. Jpn., 1993, 2, 414. 7 Y. Saito, T. Yoshikawa, M. Okuda, N. Fujimoto, K. Sumiyama, 42 M. Tomita, Y. Saito and T. Hayashi, Jpn. J. Appl. Phys., 1993, K. Suzuki, A. Kasuya and Y. Nishina, J. Phys. Chem. Solids, 1993, 32, L280. 54, 1849. 43 M. Ata, Y. Kijima, A. J. Hudson, H. Imoto, N. Matsuzawa and 8 D. Ugarte, Chem. Phys. L ett., 1993, 209, 99. N. Takahashi, Adv.Mater., 1994, 6, 590. 9 P.M. Ajayan, C.Colliex, J. M. Lambert, P. Bernier, L. Barbedette, 44 T. Hayashi, S. Hirono, M. Tomita and S. Umemura, Nature M. Tence� and O. Stephan, Phys. Rev. L ett., 1994, 72, 1722. (L ondon), 1996, 372, 772. 10 S. Subramoney, R. S. Ruo, D. C. Lorents, B. Chan, R. Malhorta, 45 S. Seraphin, D. Zhou and J. Jiao, J. Appl. Phys., 1996, 381, 772. M. J. Dyer and K. Parvin, Carbon, 1994, 32, 507. 46 N. S. Kopelev, V. Chechersky, A. Nath, Z. L. Wang, E. Kuzmann, 11 C. Guerret-Pie�court, Y. Le Bouar, A. Loiseau and H. Pascard, B. S. Zhang and G. H. Via, Chem.Mater., 1995, 7, 1419. Nature (L ondon), 1994, 372, 761. 47 M. G. Liu and J. M. Cowley, Carbon, 1995, 33, 749. 12 J. W. Cowley and M. Liu,Micron, 1994, 25, 53. 48 J. P. Hare, W. K. Hsu, H. W. Kroto, A. Lappas, K. Prassides, 13 Y.Murakami, T. Shibata, K. Okuyama, T. Arai, H. Suematsu and M. Terrones and D. R. M. Walton, Chem.Mater., 1996, 8, 6. Y. Yoshida, J. Phys. Chem. Solids, 1993, 54, 1861. 49 S. Bandow and Y. Saito, Jpn. J. Appl. Phys., 1993, 32, L1677. 14 A. Loiseau and H. Pascard, Chem. Phys. L ett., 1996, 256, 246. 50Wang, J. Am. Chem. Soc., 1994, 116, 397. 15 A. Loiseau, Fullerene Sci.T echnol., 1996, 4, 1263. 51 J. Sloan, J. Cook, J. Hammer, R. Heesom and M. L. H. Green, in 16 C. H. Kiang, W. A. Goddard, R. Beyers, J. R. Salem and preparation. D. S. Bethune, J. Phys. Chem., 1994, 98, 6612. 52 H. Funasaka, K. Sugiyama, K. Yamamoto and T. Takahashi, 17 S. Amelinckx, X. B. Zhang, D. Bernaerts, X. F. Zhang, V. Ivanov J. Appl. Phys., 1995, 78, 5320. and J. B.Nagy, Science, 1994, 265, 235. 53 J. R. Rostrup-Nielsen, Catal. Sci. T echnol., 1984, 5, 1. 18 D. Bernaerts, X. B. Zhang, X. F. Zhang, S. Amelinckx, 54 R. Lamber, N. Jaeger and G. Schulz-Eklo, Surf. Sci., 1988, 197, G. Vantendeloo, J. Vanlanduyt, V. Ivanov and J. B. Nagy, Philos. 402. Mag. A, 1995, 71, 605. 55 S. C. Tsang, J. B. Claridge and M. L. H. Green, Catal. T oday, 1995, 19 Y.Saito, K. Nishikubo, K. Kawabata and T. Matsumoto, J. Appl. 23, 3. Phys., 1996, 80, 3062. 56 P. E. Nolan, D. C. Lynch and A. H. Cutler, Carbon, 1996, 34, 817. 20 R. Tenne, L. Margulis, M. Genut and G. Hodes, Nature (L ondon), 57 S. Motojima, S. Asakura and M. Hirata, Mater. Sci. Eng. B, 1995, 1992, 360, 444. 34, L9. 21 L. Margulis, G. Salitra and R. Tenne, Nature (L ondon), 1992, 365, 58 M.Endo, K. Takeuchi, K. Kobori, K. Takahashi, H. W. Kroto and 113. A. Sarkar, Carbon, 1995, 33, 873. 22 M. Hershfinkel, L. A. Gheber, V. Volterra, J. L. Hutchison, 59 N. A. Kiselev, J. Sloan, D. N. Zakharov, E. F. Kukovitskii, L. Margulis and R. Tenne, J. Am. Chem. Soc., 1994, 116, 1914. J. L. Hutchison, J. Hammer and A. S. Kotosonov, in preparation. 23 L. Margulis, R.Tenne and S. Iijima, Microsc. Microanal. 60 M. Yudasaka, R. Kikuchi, T. Matsui, Y. Ohki, S. Yoshimura and Microstruct., 1996, 7, 87. E. Ota, Appl. Phys. L ett., 1995, 67, 2477. 24 Y. Feldman, E. Wasserman, D. J. Srolovitz and R. Tenne, Science, 61 J. Y. Dai, J. M. Lauerhaas, A. A. Setlur and R. P. H. Chang, Chem. 1995, 267, 222. Phys. L ett., 1996, 258, 547. 25 S. C. Tsang, P. J.F. Harris and M. L. H. Green, Nature (L ondon), 62 L. Margulis, J. L. Hutchison and R. Tenne, in New Horizons for 1993, 362, 520. Materials, ed. P. Vincenzini, Techna Sci, Italy, 1995, pp. 301–308. 26 Y. Yosida, Appl. Phys. L ett., 1994, 64, 3048. 63 M. S. Donley, N. T. McDevitt, T. W. Hass, P. T. Murray and 27 Y. Saito and M. Masuda, Jpn. J. Appl. Phys., Part 1, 1994, 34, 5594. J. T. Grant, T hin Solid Films, 1989, 168, 335. 28 J. Jiao, S. Seraphin, X. K. Wang and J. C. Withers, J. Appl. Phys., 64 Q. Zhang, R. B. Huang, Z. Y. Liu and L. S. Zheng, Chem. J. Chin. 1996, 80, 103. Univ., 1995, 16, 1624. 29 M. Freemantle, Chem. Eng. News, 1996, 74, 62. 65 M. Homyonfer, Y. Matsai, M. Hershfinkel, V. Volterra, 30 P. M. Ajayan, O. Stephan, P. Redlich and C. Colliex, Nature J. L. Hutchison and R. Tenne, J. Am. Chem. Soc., 1996, 118, 7804. (L ondon), 1995, 375, 564. 66 R. Heesom, Part II Thesis, Oxford, 1996. 31 S. M. Csicsery, Stud. Surf. Sci. Catal., 1995, 94, 1. 67 Y. Saito, M. Okuda, T. Yoshikawa, A. Kasuya and Y. Nishina, 32 R. Tenne, Adv.Mater., 1995, 7, 965. J. Phys. Chem., 1994, 98, 6696. 33 E. G. Bithell, A. Rawclie, S. C. Tsang, M. J. Goringe and 68 R. Lamber, N. Jaeger and G. Schulz-Eklo, Surf. Sci., 1988, 197, M. L. H. Green, Inst. Phys. Conf. Ser., 1995, 147, 361. 402. 34 J. Cook, J. Sloan, A. Chu, M. Zwiefka-Sibley, M. L. H. Green and 69 Y. Feldman, G. L. Frey, M. Homyonfer, V. Lyakhovitskaya, J. L. Hutchison, J. Solid State Chem., submitted. L. Margulis, H. Cohen, G. Hodes, J. L. Hutchison and R. Tenne, 35 J. Sloan, J. Cook, R. Heesom, M. L. H. Green and J. L. Hutchison, J. Am. Chem. Soc., 1996, 118, 5362. J. Cryst. Growth, 1997, in press. 70 J. Sloan, J. L. Hutchison, R. Tenne, M. Homyonfer, Y. Feldman 36 F. Banhart and P. M. Ajayan,Nature (L ondon), 1996, 382, 433. 37 R. M. Lago, S. C. Tsang, K. L. Lu, Y. K. Chen and M. L. H. Green, and T. Tsirlana, in preparation. J. Chem. Soc., Chem. Commun., 1995, 1355. 71 R. J. D. Tilley, Chem. Scr., 1979, 14, 147. 38 Y. K. Chen, A. Chu, J. Cook, M. L. H. Green, P. J. F. Harris, R. Heesom, M. Humphries, J. Sloan, S. C. Tsang and J. C. F. Turner, J.Mater. Chem., 1997, in press. Paper 7/00035I; Received 2nd January, 1997 J. Mater. Chem., 1997, 7(7), 1089–1095 1095