Russian Chemical Reviews 69 (1) 35 ± 52 (2000) Methods for preparation of carbon nanotubes E G Rakov Contents I. Introduction II. Electric-arc synthesis III. Laser-assisted synthesis IV. Other methods of graphite vaporisation V. Pyrolysis of hydrocarbons and decomposition of CO VI. Growth of nanotubes by decomposition of metal carbides VII. Other methods VIII. Purification and opening of nanotubes IX. Conclusion Abstract. The most important methods of synthesis and purifica- tion of carbon nanotubes, a new form of material, are described. The prospects for increasing the scale of preparation processes and for more extensive application of nanotubes are evaluated. The bibliography includes 282 references. I. Introduction Immediately after they had been discovered in 1991,1 carbon nanotubes (nanotubulenes, hereinafter, NT) attracted so much attention of scientists of various professions that the discoverer, Iijima, soon became and still remains one of the most cited researchers in the field of nano-sized materials.The reason for this interest is not the unusual structure of these compounds, as in the case of fullerenes (although this also plays a certain role) but rather the prospects that are opened by the properties of NT for materials science. The journal Fullerene Science and Technology, having the word `technology' in its title, was founded in 1992; during the decade which has passed since the discovery of fullerenes (by 1995), 150 patents for their use were granted in the USA;2 reviews dealing with commercialisation of the manufacture and use of fullerenes appeared;3 however, there are still no data on the large- scale implementation of any patented method or, even more so, on the appearance of a new line of research in the materials science.The situation with NT is absolutely different: as early as 1992 ± 1993, the main fields of their potential application took shape and design of the first prototypes of future devices started. Some applications of NT were inherited from fullerenes (electro- des for chemical cells, safe sources of hydrogen in transport devices, optical filters), other applications coincide with those of carbon fibres (high-strength composites), and some fields of application are due to the unique properties of NT themselves (semiconductor devices, field emitters, probes of tunnelling micro- scopes, `quantum wires').It is clear that this is by no means the full E G Rakov D I Mendeleev University of Chemical Technology of Russia, Miusskaya pl. 9, 125047 Moscow, Russian Federation. Fax (7-095) 490 75 23. Tel. (7-095) 948 54 67 Received 18 May 1999 Uspekhi Khimii 69 (1) 41 ± 60 (2000); translated by Z P Bobkova #2000 Russian Academy of Sciences and Turpion Ltd DOI 10.1070/RC2000v069n01ABEH000531 35 36 40 41 42 45 45 46 47 list of currently known applications of NT. It is not without reason that a recent review 4 is entitled `Nanotubes: a revolution in materials science and electronics'. As Buchachenko has noted vividly, carbon nanotubes made it possible `to pass from fine words to impressive deeds'.5 Certainly, fullerenes, too, are finding and will be finding ap- plication (the rapidly progressing chemistry of fullerenes is now at an early stage of development); however, NT as materials have already clearly separated from their three-dimensional relatives. Defect-free carbon NT are formed as a result of rolling of planar atomic network graphite sheets (graphenes) into seamless cylinders with diameters of *1 to 120 ± 150 nm and with lengths of up to hundreds micrometers. There exist three forms of NT: achiral NT of the `armchair' type (two sides of each hexagon are oriented perpendicular to the NT axis), achiral `zigzag' NT (two sides of each hexagon are parallel to the NT axis) and chiral or helical NT (each pair of the hexagon sides is arranged at an angle other than 0 or 90 8 relative to the NT axis).The structures of NT are usually described by means of two indices, n and m, which are related unambiguously to the NT diameter (d ) and the chiral angle (y, characterising the deviation from the `zigzag' configuration and ranging from 0 to 30 8) (Fig. 1). d à a 3Ön2 á m2 á mnÜ , pÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅÅ p where a are the interatomic distances in the planar network, (5,0) (0,0) Zigzag y a1 a2 (5,3) (0,5) (5,5) Armchair Figure 1. Scheme illustrating the structure of idealised NT.36 y a arctan ¢§2n a m 3AAAAAA m p A . Thus achiral NT of the `armchair' type are characterised by indices (n, n), those of the `zigzag' type have indices (n, 0) and chiral ones are described by (n, m).Nanotubes can be single- or multi-walled; the number of walls is theoretically unlimited but normally it does not exceed a few dozen. The distances between the neighbouring shells are close to the interlayer spacing in graphite (0.34 nm); thus, the smallest diameter of carbon NT is *0.7 nm. The diameter of the second and subsequent coaxial atomic layers is stipulated by the diameter of the innermost layer. In this connection, the structure of NT resembles that of onion fullerenes; if the inner shell is C60 , the second shell is C240 , the third one is C540, etc. The spatial accordance of the structures of NT layers, i.e.retention of interatomic distances close to 0.34 nm is possible only provided that the chiral angle changes on passing from layer to layer. The specific structural features of NT can hamper their use as materials �¢ synthesis gives NT with different structures. Multi- walled NT are much more diversified; therefore, more uniform single-walled NT, which, in addition, usually contain fewer defects, are preferred for the design of functional materials. In most cases, the tips of NT are covered by hemispherical or conical caps, which contain not only hexagons but also pentagons, in which the configuration of the carbon atoms is less stable. These caps are somewhat more chemically reactive than the lateral surfaces. The hemispherical caps resemble `halves' of fullerene molecules.Thus the NT (10,10), isolated rather frequently, are capped by halves of spherical C240 fullerene. Synthesis of NT can give a tough precipitate consisting of sintered NT, a somewhat softer precipitate of densely stacked NT, a rubbery material consisting of entangled NT (`paper,' `mats'), a three-dimensional network of long fibres (`lace collar', `web'), or a textured material consisting of parallel or nearly parallel (but arranged at a distance from one another) NT on a substrate (`forest'). Structures looking like a `sea-urchin' and NT as helices (see below) have also been described. Nanotubes tend to form relatively stable aggregates (they are referred to as bundles or `ropes'), in which the axes of individual NTare parallel to one another, the shortest distance between them being *0.32 nm (Fig.2). These aggregates arise due to van der Waals forces. Figure 2. Schematic structure of bundles of single-walled NT. In the majority of cases, the synthesis of carbon NT is accompanied by the formation of other carbon modifications �¢ fullerenes, nanoparticles and amorphous carbon. The yields of NT and side products are determined by the conditions of synthesis; the removal of side products is an important stage in the synthesis of NT. Good prospects for the design of new functional and struc- tural materials are opened by modification of carbon NT, which can be accomplished by the following methods: (1) filling of the inner cavities of NT with substances which alter their electronic, magnetic or mechanical properties; (2) `grafting' of functional groups to the tips of NT; (3) replacement of some of the carbon atoms in NT by atoms of other elements; (4) partial or complete cleavage of double bonds on the lateral surfaces of NT by adding one or another reagent; E G Rakov (5) intercalationsertion) of `guest' atoms or molecules into the intertubular space of NT bundles.In addition, NT can be used as templates in the template syntheses of nanotubular or nano-rod materials for various purposes. The scope of this review is restricted to the methods of synthesis of NT because these methods are important for the development of studies and projects related to NT.They should lay the grounds for the future commercial manufacture. Attention is focussed on recent studies. The publication in 1999 of a Russian review devoted to the structure, electronic properties and morphology of NT6 facili- tated the performance of this task. Earlier Russian reviews dealing with the mechanisms of the formation of fullerenes, nanoparticles and NT7, 8 were also useful; however, they are somewhat less significant because lots of newer studies have appeared. Apart from carbon NT, those containing B and N atoms are also briefly considered in this review. These NT can be synthesised either from carbon NT or simultaneously with them. II. Electric-arc synthesis It is known that an electric arc develops a temperature of up to 4000 8C, its `burning' being accompanied by transfer of a substance between the electrodes.In 1990, electric arc-synthesis with graphite electrodes under an atmosphere of inert gases was first used to prepare fullerenes in relatively large amounts.9 It is among the products of arc synthesis that Iijima discovered carbon NT in 1991.1 Subsequently, this procedure has been modified and used in many laboratories of different countries. 1. Principles of the method The first synthesis of NT in relatively large (gram) amounts was also performed in Japan.10 The electric-arc synthesis was carried out under a helium atmosphere using graphite electrodes �¢ an anode 8 mm in diameter and a cathode 12 mm in diameter �¢ located at a distance of less than 1 mm from each other.The current in the arc reached 100 A (current density *150 A cm72) and the voltage was 10 to 35 V. The rate of the growth of the cathode deposit was *1 mm min71. Some of the graphite evaporated from the anode turned into carbon black and soot, which were deposited on the walls of the reaction chamber, and some was deposited at the cathode. The outer solid layer of the cathode deposit contained sintered NT and nanoparticles, which could not be separated. The purity and yield of NT depended appreciably on the helium pressure�¢the formation ofNTstarted at *13 kPa, and at 66 ¡¾ 332 kPa, the cathode deposit contained only NT and nanoparticles. The optimum pressure was 67 kPa; in this case, *75% of the electrode material consumed was depos- ited at the cathode, the yield of NT, which were accumulated in the inner, black, relatively soft part of the deposit being*25%. The yield of NT attained in an argon flow was much lower than that in a helium flow.Analysis of the results of the first studies on the structure and synthesis of NT10 ¡¾ 14 allowed Ebbesen 15, 16 to note some charac- teristic features. The material formed has a hierarchical structure, in which tens or hundreds of individual multi-walled NT with diameters of 2 ¡¾ 20 nm and nearly equal lengths (microns or tens of microns) are joined into regularly organised bundles resembling ropes. These ropes are combined into fibres with a diameter of*50 mm, while the fibres are joined into threads having even greater diameters (about a millimeter) and visible with the naked eye.The larger the bundle, the more disordered can it be. The formation of columns *50 mm in diameter consisting of NT and arranged parallel to the electrode axis has been reported. 17 This study is notable due to the setup used; it had movable vertical electrodes with a diameter of 19 mm, which (as well as the arc area) could be cooled during the process. (In the conventional arc syntheses, the cathode and anode diameters are different, the anode diameter being smaller than the cathodeMethods for preparation of carbon nanotubes diameter.) Using this setup controlled by a computer, the researchers were able to deposit *90% of the substance at the cathode and to obtain thus high-quality multi-walled NT.The current in the arc of this setup reached 250 ± 300 A, although the current density at the beginning of the process was even lower than the standard value and amounted to 70 ± 80 A cm72. Upon annealing in air at 650 8C, the columns were etched being replaced by channels surrounded by entangled NT. An increase in the electrode diameter normally causes sinter- ing and cracking of NT.Adense greyish product is accumulated in the inner part of the cathode deposit, instead of the easily extractable black material. This phenomenon was initially explained by insufficiently high temperature of the synthesis;18 however, the assumption that the temperature was, conversely, too high seems more reasonable.17 The decrease in the anode diameter from 12.7 to 8.0 mm with the cathode diameter (25.4 mm) and current density (140 A cm72) remaining the same increased the yield of NT.An increase in the pressure of helium (which is believed to quench NT) from 6.6 to 101 kPa also resulted in an almost proportional increase in the rate of growth of the deposit.18 However, accord- ing to Ando and Iijima,19 variation of the yield ofNTas a function of gas pressure passes through a maximum, the position of which corresponds to 7 kPa in the case of Ar orCH4 and to 3 kPa for He. In the opinion of Wang et al.,18 the discharge used in the synthesis of NT is quasi-continuous and has a characteristic interruption frequency of*8 Hz. The discharge is initiated between the closest sections of the electrodes and, after some amount of graphite has been evaporated from the anode (which extends the discharge), it `jumps' to the neighbouring section and thus becomes shorter.Travelling of the arc over the electrode surface changes presum- ably the sites of localisation of NT. Lozovik et al.7 believe that an arc discharge between graphite electrodes has two modes of operation, a noisy and a quiet one, transition from one mode to the other being induced by changes in the current density. A high current density and a low inert gas pressure (noisy mode) result in the predominant formation of fullerenes, while nanoparticles and NT are formed predominantly at a relatively low current density and a high pressure (quiet mode).A thorough investigation of the cross-sections of bundles obtained by the arc method showed that NT do not have, as a rule, a cylindrical shape.20 The cross-sections of NT look like polyhedra or ellipses with numerous defects including edge dislocations. The structures of NT are intermediate between the `Russian doll' model (coaxial seamless cylinders) and the `folded carpet' (`scroll') model. In earlier studies, either model has been preferred, although the methods used in the investigations did not allow one to reveal the differences between them.21 Moreover, it was found that the `scroll' model is too regular to describe many NT and, in some cases, the `papier-mache' model (separate fragments applied onto one another) should be accepted (Fig.3).22 Some disordering of the structure is due to the fact thatNTare formed in the arc under non-equilibrium conditions and to the stress arising upon joining of individual NT into bundles. The presence of a large number of defects in multi-walled NT obtained in an electric arc is confirmed by experiments on their intercala- tion.23, 24 It has been suggested 25 that bundles are formed under the influence of an electric field, their length, diameter and the manner of laying being dependent on the electric field strength. However, in reality, bundles can be formed even without electric fields. Nevertheless, it is beyond doubt that the field has an influence on the morphology of the products of arc synthesis.The yield and shape of NT formed in an arc discharge plasma depend not only on the main arc discharge variables (the voltage between the electrodes, the current strength and density, temper- ature of the plasma) and parameters related to the rge 37 Figure 3. Structure models of multi-walled NT. (a) `russian doll'; (b) `scroll'; (c) `papier-mache'. variables (the pressure and composition of the inert or reacting gas) but also on the gas flow rate, on the dimensions of the reaction chamber, and on the duration and the scale of the process. The configuration of cooling devices if they are present (determines the size and shape of temperature fields) and the power of cooling devices (determines the heat removal), the nature and purity of the electrode materials and several other parameters that can hardly be quantitatively evaluated are also signifi- cant.10, 15 ± 19, 26 The main parameters governing the yield of NT � the rates of NT growth and quenching � depend on numerous variables which are far from being fully taken into account in experimental studies.This accounts for some contra- dictions, hampers comparison of the results and often makes them poorly reproducible. The absence of theoretical models (related to the setup geometry), which makes scaling impossible, might be due to the same reasons. The mechanism of the formation ofNTin an arc discharge has not yet been unambiguously elucidated, although it is discussed in many studies. There exist two main opposing models.According to one model, growth of NT advances due to the addition of carbon atoms or fragments from the vapour phase to the dangling bonds at the tips of open NT, whereas according to the other model, they add to the topological defects in the caps of closed NT. In recent years, preference is given to the former model, because the occurrence of `edge-to-edge' (`lip ± lip') interactions has been proved. They prevent the formation of caps in multi- walled NTdue to the formation of `fluctuating' (closure ± rupture) C7C bonds at the tips of two or three neighbouring coaxial NT.27, 28 The role of the electric field in the mechanism of the formation of NT is obviously overstated in some publications. It has been shown 16, 29 that this role is far from being paramount, the NT growth involving simultaneously charged (C+ etc.) and neutral (C2 etc.) carbon species.The difference between the main carbon sources might account for the formation of two types of con- densed products, NT and polyhedral particles. It is also unlikely that the tips of NT growing at the cathode in an arc discharge are opened due to strong electric fields.30 The experimental studies considered above describe the preparation of multi-walled NT. The formation of single-walled NT was first observed by Ajayan and Iijima,31 but this study has not been adequately appreciated, later studies being regarded as the pioneering ones.32, 33 Both groups of researchers cited 32, 33 performed the synthesis in the presence of catalytic additives.This opened up a new chapter in the history of NT. 2. The influence of catalytic additives The pioneering studies 32, 33 somewhat contradict each other in details but coincide in the main point stating that the introduction of small amounts (1% ± 2%) of transition metals such as Fe, Co, Ni or their mixtures into the graphite anode influences appreci- ably the shape and the yield of NT and, in some cases, also localisation of the NT-containing product in the reaction chamber. Subsequently, admixtures of Li, Cu, Ag, Zn, Cd, B, Al, In, Y, La, lanthanides, Si, Ge, Sn, Ti, Hf, Pb, Sb, Bi, S, Se, Cr, W, Mn, Ru, Pd, Pt, mixtures of two metals or of a metal with a nonmetal and several carbides and oxides were tested.28, 34 ± 37 It was found that vaporisation of an anode containing Co, Co7Ni, Co7Y, Co7Fe, Ni, Ni7Y, Ni7Lu and Ni7Fe gives deposits looking38 like a lace collar or a soft belt being formed around the cathode deposit.They contain single-walled NT mixed with amorphous carbon and metal particles, the NT having diameters of 1.2 ± 1.4 nm and being joined into more or less ordered bundles. The tips of single-walled NT are closed and contain no metal particles. Upon evaporation of graphite anodes with Ni7Y and Co7Y admixtures, the content of single-walled NT in the deposits reaches 70%± 90%. 37 The yield of NT also markedly increases when a Co ± Pt mixture is used. 38 Some catalysts (Cu; Cu with Ni, Pt, Yor Fe; Ni; Ni with Y, Lu or Fe) caused the formation of a `web', which hung between the cathode and the walls of the chamber. This web, as well as the deposit on the chamber walls, contained fullerenes, amorphous carbon, flat graphite particles and small amounts of single-walled NT.In some cases, a rubber-like deposit was formed on the walls; it could be detached as pieces or ribbons (Fig. 4).39 3 1 2 5 4 7 6 8 9 10 13 11 12 14 16 15 Figure 4. Setup for electric arc synthesis and localisation of various products; (1) to vacuum pump; (2) filter; (3) viewing window; (4) cathode; (5) nozzle for the removal of the cooling water; (6) graphite electrodes; (7) web-like precipitate; (8) cathode deposit; (9) rubber-like deposit around the cathode; (10) filling material (a mixture of powdered graphite with a metal); (11) fullerene- and NT-containing soot on the walls; (12) anode; (13) nozzle for the supply of cooling water; (14) voltmeter; (15) helium inlet; (16) power supply.The addition of sulfur to aCo catalyst also resulted in a greater amount of the web-like product and higher NT yields and in a quite noticeable scatter in the NT diameters (from 1 to 6 nm).40 Evaporation of Co and S simultaneously with graphite changed the shapes of the resulting products and led to the evolution of multi-walled NT in the middle of the cathode deposit, multi- and single-walledNTtogether with many other particles in the rubber- like deposit around the cathode, and single-walled NT in the web- like material.39 Kiang et al.40 believe that the multi- and single-walled NT are formed via different mechanisms and that Co and S induce the formation of single-walled NT directly in the gas phase and prevent the NT tips from closing.Since sulfur itself does not catalyse the formation of NT, its role seems to reduce to the stabilisation of dangling bonds. The addition of Bi and Pb also causes an increase in the NT diameter. The carbon ring structures formed in the gas phase are believed to play an important role in the mechanism of formation of single-walled NT. The most stable ring structures are those containing 10 to 40 carbon atoms, which form ComCn clusters together with Co atoms. It is these clusters that function as catalysts, while S, Bi and Pb stabilise the ring structures.41 The results of studies on the catalysed arc synthesis of single- walled NT are summarised in Table 1.42 It can be seen that the diameter of NT obtained in this way varies from 0.6 to 6 nm.E G Rakov Table 1. Characteristics of NT prepared by arc discharge synthesis in the presence of catalysts.42 Composition of crystallites a Diameter of NT /nm Catalyst average limits >0.6 >0.6 0.7 ± 1.6 0.80, 1.05 b 0.6 ± 1.3 0.7 ± 0.8 0.9 ± 2.4 1.3, 1.5 b 0.6 ± 1.8 7 1.2 ± 1.5 7 0.6 ± 1.3 0.7 ± 0.8 0.9 ± 3.1 1.7 1.3 ± 1.8 1.2 ± 1.3 1.0 ± 6.0 1.3, 1.5 b Fe Fe Co Co Ni Ni Fe+Ni Fe+Ni Co+Ni Co+S 0.8 ± 5.0 1.2, 1.5 b 7 Fe3C 7Co Co in a graphite shell Ni in polyhedral particles 7777Co and Co in polyhedral particles 7 0.7 ± 4.07 7 *2 1.1 ± 1.7 7 Co+Bi Co+Pb Co+Pt Y CoPt YC2 in polyhedral particles Cu in polyhedral particles 1 ± 4 >2 Cu None 77 7 a The crystallites are located inside NT or polyhedral particles (see Section III.3).b The curve for the distribution of NT over diameters exhibits two maxima. The mechanism of the catalytic action of metals in the formation of single-walled NT implies adsorption of carbon atoms on the surface of metal particles and their free migration over the surface to the base of a growing NT.43 It was shown experimentally that the catalytic reaction can follow either of two routes, depending on the size of the metal catalyst particles.If the particle size (average diameter) is several tens of nanometers, which is much greater than the NT diameter (*1 nm)ny closed NT sprout from one particle. If the particle size is not larger than the NT diameter, the particle moves together with the growing NT tip. The results of molecular dynamics modelling 43 confirmed the probability of the former pathway, involving migration of adsorbed carbon atoms to the NT base and the `root' growth of NT. It should be borne in mind, however, that this model is of limited utility. It does not take into account the fact that NT formed in arc synthesis contain lots of defects and that a stream of carbon atoms directed towards growing NT has the greatest intensity near the NT tips.In addition, the conditions of vapour condensation in different sections of a setup, as well as in different setups are nonequivalent, which results in the formation of particles with dissimilar morphologies. Nevertheless, the modelling made it possible to explain why multi-walled NT are not formed in the presence of catalytic additives (this would require cooperative processes, associated with consistent interaction of numerous particles), why single- walled NT always have relatively small diameters (the growth of NT is initiated on the surface bulges of metal particles and the diameter of these bulges is small with respect to the height), and why mixed catalysts are often more efficient than one-component ones (this is due to the enhancement of carbon adsorption, to the change in the activation energy of the NT growth, and to the formation of a surface with a large number of bulges).Apparently, single-walled NT are even more prone to form bundles, which are combined in `ropes' with a two-dimensional triangular crystal lattice.37, 44, 45 Journier and Bernier,26 who surveyed studies on catalytic arc synthesis, have rightly stated that the yield and the shape of NT can differ substantially even when the same metal is used, depending on the metal concentration, pressure of the inert gas, current strength (or density), and the geometry of the arc setup.Methods for preparation of carbon nanotubes Aspecial role in the arc synthesis ofNTis played by boron, the introduction of which as B, B2O3 orBNinduces the formation of a large number of well graphitised long (up to 20 mm) NT with boron-containing caps.46 In this case, B4C crystals, large full- erenes and BC3 nanotubes are produced in addition to carbon NT.The presence of boron in the gas phase decreases appreciably the concentration of C2 fragments in the plasma zone and the contents of fullerenes in the wall deposit.47 Since the C2 clusters are the main source of condensed structures, the introduction of boron can change the mechanism of the formation of NT. Note that the arc setup used in the study cited 47 differed only slightly in geometry and electric parameters from the conventional setups used for the synthesis of carbon NT [anode and cathode diameters 6 mm (a cathode with a diameter of 200 mm was also tested), distance between the electrodes 1 mm, current 80 or 150 A, voltage 25 or 40 V, respectively, gas pressure 27 ± 67 kPa].In some studies the formation of carbon NT, separate areas of which were made of BN has been noted. Under certain conditions, for example, upon evaporation of BN or mixtures of B or BN with graphite, boron carbides, nitride, and carbonitrides can form `self-contained' NT.48 ± 51 The use of HfB2 (see Refs 52, 53) or ZrB2 (see 54, 55) electrodes for the electric arc synthesis in a stream of N2 results in the formation of BN nanotubes. If BC4N serves as the electrode material,56, 57 NT containing B, C and N (or BN) are produced; when the electrodes are made of graphite with the addition of BN,48 the process yields NT containing BC2N and BC3 .However, in the latter case, the composition of NT is nonuniform along the length of the tubes. Unique multi-walled NT have been prepared by Suenaga et al.,58 who also used HfB2 and graphite as the electrode materials; several inner shells consisted of carbon, several central shells (embracing the carbon shells) were made of BN, and some outer shells were again carbon. The mechanism of formation of these NT (they evolved together with polyhedral particles, which also contained layers with different compositions) is totally unclear. The inner layers might grow upon diffusion from an NT tip; however, the probability of this process is low. Since the electronic properties of C and BN layers are different, this result brought about the idea of designing electronic devices with radial hetero- junctions.Studies on the catalytic synthesis have much in common with the studies on the synthesis of NT the inner cavity of which is filled with one or another substance. 3. Preparation of filled nanotubes Filling of the inner cavity of NT (encapsulation) is a route to the development of a great number of novel nanomaterials of various classes and functions (materials with specific electronic, magnetic, optical, or mechanical properties, catalysts, sorbents). Some of them could be used as a type of `NT ± filler' nanocomposite, while other require that the carbon shell be removed (burned out). Particular attention is devoted to the preparation of `quantum wires', i.e., conducting materials with a diameter of several nanometers, whose conductivity is close to truly one-dimensional conductivity.The possibility of using NT and nano-rods in nanotechnological devices is being discussed. The first study on filling NT was performed in 1993.59 Mixtures of NT with Pb were annealed in air at *400 8C. In the presence of O2, the molten metal opened the tips of NT, removed the hemispherical caps and was sucked into the inner cavities of the NT by capillary action. However, it has been reported 60 that lead is able to fill NT only at pressures of 1000 to 10 000 atm; hence, under the conditions described above, NT were filled by lead compounds rather than by lead metal.Indeed, it has been shown that heating of a mixture of Pb3O4 with NT for 9 h at 700 8C results in opening and filling of the NT.61 Encapsulation of molten V2O5 , PbO, Bi2O3 (see Refs 62, 63), MoO3 (see Ref. 64) and AgNO3 (see Refs 65, 66) into open NT has also been reported. Even before the discovery of NT, the arc method has been used to prepare endohedral fullerenes 67 and nanoparticles encap- 39 sulated into a multilayered graphite shells.68, 69 Therefore, it comes as no surprise that filling of NT can occur directly during the arc synthesis. Ordinary and filled NT are obtained under different conditions, in particular, filled NT are formed at lower temperatures (1000 to 2000 8C). Yttrium carbide was the first substance detected inside NT prepared in an arc.70 ± 72 According to an estimate,35 yttrium is encapsulated into NT more easily than many other metals, although this feature can hardly be related to the enthalpy of formation of YC2 .Later, carbides of other lanthanides and Mn have also been encapsulated into NT.73 ± 75 Carbides have been formed when metal-doped graphite anodes have been used because carbides are the species most stable in the presence of carbon. However, in some cases, elementary substances have also been detected inside NT, for example, Mn,76 Cu, Ge 77 ± 79 and even small amounts of Y.72 These elements either do not form stable carbides or are liberated in an arc discharge under an H2 atmosphere. Thus copper forms no stable carbides and does not catalyse the formation of NT or fullerenes; therefore, the mechanism of its encapsulation differs from the mechanism of insertion of other metals.78 Evaporation of a graphite anode in a hydrogen plasma in the presence of copper gives rise to polycyclic aromatic hydro- carbons, which form graphite shells around Cu particles.78, 79 The formation of filled NT upon joint evaporation of pyrene and Cu or Ge (this was done using a Cu anode with a cavity filled with pyrene and aWcathode) 80 was considered 78, 79 as direct evidence supporting the above statement.Metals with different volatilities show different behaviour during encapsulation.81 Some substances fill long NT giving continuous crystals with lengths of up to 1 mm (Cr, Ni, Sm, Gd, Dy, Yb, S, Se, Ge, Sb); other substances fill only short NT (Al, Bi, Te), and some other elements form inclusions separated from one another inside NT (Co, Fe, Pd).36, 82 ± 85 In a study of encapsula- tion of high-melting transition metal carbides (NbC, TaC, MoC), an exceptionally stable face-centred form of the carbide MoC was discovered. 85 Nanotubes containing B4C in the inner cavity have also been prepared. 86 The operation conditions used in these studies 70 ± 83 did not differ much from those used for the synthesis of NT.Thus Setlur et al.78 used electrodes with a diameter of 10 mm located at a distance of 0.25 to 2.0 mm; the current was 100 A, the voltage was 20 V and the gas pressure ranged from 1.3 to 6.7 kPa.In another series of studies,82 ± 84 the corresponding parameters were 9 mm, 1 mm, 100 ± 110 A, 20 ± 30 V and 60 kPa. Sulfur plays a special role in the formation of filled NT.84 Sulfur can be contained in a graphite anode as a minor impurity; it promotes filling of NT with other elements. In some cases, sulfur enters NT as a sulfide but more frequently it is not present in the filling material. This effect of sulfur has been explained 84 by the fact that sulfur forms clusters with carbon in the gas phase; this ensures transportation of sulfur to the growing NT, where it promotes graphitisation of NT and reconstruction of a metal surface, resulting in the exposure of catalytically more active crystallographic planes. At the final stages of the process, sulfur is removed from the graphite shell of the NT. These results are similar to the data reported in earlier publications.39, 40 In some cases, sulfur suppresses the formation of single-walled NT.The technique of filling NT under the conditions of arc synthesis suffers from a severe drawback �this process is hardly controllable. The yield of filled NT and the composition, structure and morphology of encapsulated substances cannot be governed in the majority of cases. 4. Modifications of arc synthesis The use of an alternating-current arc and electrodes with identical diameters changes substantially the pattern of the process, namely, the products are formed on the chamber walls rather than at the electrodes, new forms of NT being produced together with the known ones.87, 88 The yields and forms of some products40 also depend on whether the electrodes are mounted in the vertical or horizontal position.A series of publications 19, 89 ± 91 deal with the formation of multi-walled NT under a methane atmosphere. This process differs from the conventional one (performed under an inert gas) in the fact that, under some conditions, the formation of NT is not accompanied by evolution of fullerenes or nanoparticles. At high partial pressures of methane, relatively thick multi-walled NT are formed and at low pressures (1 ± 3 kPa), thin and long multi- walled NT are produced. The optimal conditions for the forma- tion of thin NT for the vertical mounting of electrodes with a diameter of 6 mm are the following: methane pressure 2.7 kPa, current 30 A.Later, the authors of the above series of studies arrived at the conclusion that, since CH4 decomposes in an arc to give C2H2 and H2 , the synthesis of NT should better be performed in a stream of H2 (see Refs 92 ± 95).{ Some advantages of using H2 have been mentioned above in the consideration of copper encapsulation. A certain role in the formation of NT can also be attributed to the high heat capacity of H2 , owing to which it can efficiently quench NT.96 When the process is performed at `plasma' temperatures, the fact that the energy of ionisation (the first ionisation potential) of H2 is half that of He also becomes significant. Rather long NT are formed even at an H2 pressure of 7 kPa; their yield at a pressure of 13 kPa is comparable with the yield attained in the atmosphere of He at 66 kPa.Many NT prepared in a flow of H2 have a narrow inner channel; they are open and usually high-quality NT (having few nanoparticles on the outer surface). The thinnest and longest multi-walled NT were prepared at H2 pressures of 6.7 ± 11 kPa. The arc synthesis in an atmosphere of H2 differs appreciably from the synthesis in inert or hydrocarbon media because the temperature developed in the atmosphere of H2 is higher. This provided grounds for assuming that C+ ions participate in the formation of NT in anH2 medium.92 It is also assumed that either H2 molecules or H atoms add to the dangling bonds of growing NT and thus prevent closure of these bonds and, in addition, eliminate the possibility of the formation of amorphous carbon.92 The addition of H2 is evidenced by the fact that an increase in the partial pressure of H2 results in higher yields of open NT.The mechanism involved may be similar to the mechanism of the formation of diamond from the gas phase. Particles of various natures are produced; however, under a hydrogen atmosphere, the most reactive ones are completely or partially gasified (etched). Yokomichi et al.97 have attempted to prepare NT in a CF4 medium. It has been assumed that the F atoms formed upon decomposition of CF4 would attach to the dangling bonds and prevent them from closing, thus influencing the process of NT formation. This assumption largely proved to be correct; in any case, no fullerenes were formed together with NT.The synthesis was carried out using a setup with electrodes 5 mmin diameter at a current of 20 ± 60 A and a CF4 pressure ranging from 2.7 to 53 kPa. Under the optimal conditions (20 ± 40 A, 6.7 ± 13.3 kPa), multi-walled NT were synthesised with an outer and inner diameters of 20 and 5 nm, respectively; they did not differ much from the NT obtained in a flow of He. The researchers cited 97 studied the influence of the main parameters of the process (medium, gas pressure and current) on the yield of NT and polyhedral particles and on the length of NT and compared their data with those reported in other publica- tions; the results were summarised in a table (Table 2).In particular, it follows from this table that the processes occurring in an atmosphere of CF4 and CH4 are fairly similar. However, the data of Table 2 should be treated with caution because this way of presenting the information does not show, for example, that in { Subsequently, the same conclusions have been drawn by other inves- tigators.96 E G Rakov Table 2. Influence of the increase in the gas pressure and in the current on the main characteristics of electric-arc synthesis of NT under atmospheres of various gases.97 Length of NT Medium Yield of NT Yield of nano- particles decreases does not change does not change does not change increases does not change increases does not change decreases increases decreases decreases CF4 He CH4 H2 some cases, the plots for the variation of the yield vs current pass through a maximum.Koprinarov et al.98 have attempted to devise a continuous electric arc process and to make up for the consumption of graphite due to evaporation by supplying CH4 . For this pur- pose, they used the `reverse method', i.e. the area of the anode in the setup used was an order of magnitude greater than the area of the cathode. The products of electric-arc synthesis were found to contain fullerenes, polyhedral particles and single-walled and multi-walled NT. The nanotubes had different lengths and structures. Some of them were partly single-walled and partly multi-walled structures; this indicates that various sections of NT grew at different rates.The `reverse method' offers no advantages over the conventional method. Carbonised coal can be used as the electrode material instead of graphite;99 however, the presence of non-volatile impurities causes contamination of NT; hence, in the majority of studies, high-purity graphite is the material of choice. The main drawback of the arc method is low output. This is demonstrated by a publication 37 in which synthesis in aHe flow in the presence of Ni and Y introduced in a graphite anode gave only 2 g of NT over a period of 2 min. After that, the setup had to be disassembled to replace electrodes and take the produc out. III. Laser-assisted synthesis Fullerenes were first obtained by the laser method in 1985; however, only 10 years later, this method was employed to prepare NT.100, 101 The first setup was a 50-cm quartz tube with a diameter of 2.5 cm containing a graphite rod 1.25 cm in diameter arranged along the tube axis.The tube was evacuated and simultaneously heated to 1200 8C. After that, Ar was supplied (pressure 66.5 kPa, linear gas velocity 0.2 ± 2.0 cm s71). The target was exposed to a laser beam with a wavelength of 532 nm (Nd laser), a pulse frequency of 10 Hz, a pulse power of 250 mJ, and a pulse duration of 10 ns. The surface of the target was scanned by a laser spot with a diameter of 3 or 6 mm. The products of evaporation (multi-walled NT and nanoparticles) were collected on a cooled copper finger, on the walls of the tube, and on the reverse side of the graphite target (Fig.5). 1 4 3 2 Figure 5. Scheme of the laser setup; (1) furnace heated to 1200 8C; (2) neodymium laser; (3) graphite target; (4) water-cooled collector.Methods for preparation of carbon nanotubes The daily output of the first setup was up to 80 mg of a nanomaterial containing NT. The process had to be terminated because the tube was clogged near the target with a web-like material which also contained NT. By using a tube with a diameter of 3.8 cm, all other conditions remaining almost the same, another group of researchers attained a higher yield of a rubbery material containing single-walled NT. In this case, the output was 200 mg per experiment (3 ± 5 h).102 It was noted that distribution of the NT over diameters depends on the radiation wavelength (either 532 or 1064 nm).102 An increase in the diameter of the reaction tube to 5 cm was even more effective.This setup can afford *1 g of the material (containing 60%± 90% of NT) per 24 h.103 Subsequently, one laser was replaced by two lasers (wave- lengths 532 and 1064 nm), which emitted alternately, the interval between the pulses being 42 ns, and the power of a pulse was increased to 490 and 550 mJ, respectively.103 This made it possible to `knock down' the plums formed. Subsequently the tube diameter was doubled once again; in addition, alternation of the side of the target exposed to radiation and beam scanning were used.As a consequence, the yield of products containing 40%± 50% of NT reached 20 g during 48 h of continuous operation.103 Yudasaka et al.104 used a modified setup with two targets, one made of a graphite powder and the other made of a metal or an alloy. Study of the influence of the pressure of Ar and the furnace temperature on the yield and properties of the laser ablation product has shown that at pressures below 13 kPa only amor- phous carbon is formed, while NT appear in the products, together with amorphous carbon, above 26 kPa.105 An increase in the furnace temperature from 780 to 1050 8C resulted, in the presence of catalysts, in an increase in the average diameter of single-walled NT from 0.8 to 1.51 nm, although the yield of NT at low temperatures was relatively low.106 The yield and the shape of NT prepared by the laser method are determined by fewer parameters than those of the NT obtained by the electric-arc synthesis.Perhaps, this is why markedly higher yields of NT have been attained in the laser- assisted synthesis. The temperature of the area from which graphite is evaporated and the temperature gradient in the gas phase near this area can be considered to be the crucial factors. Unfortunately, in the experiments described above these param- eters have been determined insufficiently definitely if at all (the measurements were hampered by the beam scanning over the target surface). Instead, properties influencing the temperature of evaporation were measured (the furnace temperature, laser radiation power, diameter of the light spot on the target, the velocity of the spot migration, the gas pressure, and the gas flow rate).The introduction into graphite of small amounts of catalysts led to the formation of single-walled NT, which, unlike those obtained in arc, were only slightly covered by amorphous carbon particles.101, 107 Mixtures of Co with Ni (0.6 at.% each) and of Co with Pt (0.6 at.% and 0.2 at.%) proved to be the best catalysts. When these mixtures were used, the yield of single-walled NT exceeded 70%, which is tens or hundreds times higher than the yields attained using individual metals as the catalysts. High yields have also been obtained with mixtures of Ni with Pt.100 The use of a mixture of two noble metals, Rh and Pd, mixed with graphite in a laser-assisted synthesis afforded single-walled NTwith diameters of 1.0 ± 1.5 nm.108 Only mixtures of Cu with Ni were found to be substantially less active than copper itself.100 The mechanism of the catalytic formation of single-walled NT upon laser ablation proposed by Thess et al.107 was defined as the `scooter' mechanism.According to this mechanism, individual atoms of Ni, Co or other catalytically active metals are adsorbed on the open tips of the curved Cn (n450) graphene fragments and `scoot' around the tips, thus promoting removal of all carbon structures except for the energetically preferred ones. 41 Clusters comprising 10 to 100 closed-tip nanotubes formed `ropes', 10 to 50 of which composed bundles.A more detailed study of the `ropes' showed that they are polycrystals, the sizes of separate grains ranging from 10 to 100 nm (the typical size is 10 to 20 nm); their cross-sections are oval and the predominant aspect ratio is*3.109 Some of the NT are clustered into 5 ± 15 nm thick `ropes' forming rings with diameters of 300 ± 500 nm.110 The products of laser-assisted synthesis were found to contain an interesting type of NT, so-called `nano-peas', which are chains composed of spherical C60 molecules located inside single-walled NT with a diameter of 1.3 ± 1.4 nm.111 The laser method also makes it possible to prepare NT from BN;112 ± 115 however, the experimental conditions employed in this case differed sharply from those used to prepare carbon NT; specifically, high and superhigh pressures were used.The process induced by laser heating of hexagonal BN in a diamond anvil at a high pressure under an atmosphere of N2 has an interesting feature, namely, cubic BN appears at the bottom of the NT formed.115 The mechanism of formation of NT from BN upon laser evaporation at superhigh pressures might include a stage of surface diffusion of molecules from the bottom of an NT to its tip, which implies restriction of the NT length.116 Laser ablation of a complex target (C, Si, BN) at near- atmospheric pressure accompanied by chemical reactions can be used to synthesise axial NT. This type of NT with diameters of several tens of nanometers and with lengths of up to 50 mm contained a core of b-SiC, an intermediate layer of amorphous SiO and an outer shell of BN and C.117 The study cited aimed at designing coaxial nano-sized electronic devices with heterojunc- tions.In addition to the neodymium laser, a CO2 laser has also been tested for the synthesis of carbon NT.118 IV. Other methods of graphite vaporisation The conclusion that fullerenes and NT are formed in the vapour phase, irrespective of the way used to vaporise graphite or another form of carbon, was drawn fairly long ago and has been confirmed in a number of publications. Thus Ebbesen et al.119 proved, using isotope dilution, that NT (as well as fullerenes) are formed exactly from carbon vapour. Apart from the arc- and laser-induced evaporation of graphite considered above, NT are also prepared using resistivity vaporisation and vaporisation induced by elec- tron or ion beams or by sunlight. 1.Resistivity vaporisation Heating a 0.5-mm graphite foil by Joule heat in vacuo (1076 Pa) and cooling the resulting vapour to730 8C resulted in NT being deposited on the surface of single-crystalline graphite.120 ± 123 The rate of deposition of NT with a diameter of 1 ± 4 nm was 0.5 As71. The NT were capped and combined in arrays. This method permits the preparation of particles having diverse shapes � single-walled and multi-walled NT, NT bundles and nano-cones, the yield of single-walled NT ranging from several percent to 80%± 90%. 2. Electron- or ion-beam vaporisation The first experiments on the preparation of NT by electron-beam vaporisation of high-purity (99.99%) reactor grade graphite in vacuo (1073 Pa) and deposition on various substrates (Si, quartz, graphite, ceramics, anodised Al) were performed in Russia back in 1992.124 The resulting condensate was a 0.01 ± 10-mm thick film consisting of NT with a diameter of *1 nm.Separate NT were joined in fibres*5 nmin diameter, while the fibres were clustered in cables with a diameter of 10 ± 30 nm. The tubular texture was retained throughout the whole film thickness; moreover, variation of the angle between the direction of the stream of carbon particles and the substrate surface resulted in the formation of inclined42 textures.Kosakovskaya et al.124 measured a large number of film characteristics and noted anisotropy of film properties. Nanotubes have also been formed on bombardment of high- purity graphite with Ar+ ions with an energy of 60 keV in a high vacuum at a normal angle of incidence.125 Polyhedral nano- particles were isolated together with NT. Exposure of hexagonalBNto an electron beam gave NT;126 as in the study described above,115 in this case, too, the formation of seeds of cubic boron nitride at an initial stage was assumed. The irradiation with electrons with a current density of 10 ± 20 A cm72 at a voltage of 300 kV resulted in annealing of most of defects and in almost complete ordering of theBNwalls of the NT.112 An electron beam with a current density of 150 Acm72 at an accelerating voltage of 400 kV ensured the formation of multi-walled coaxial BN clusters.127 Irradiation of boron-containing carbon with electrons gave carbon NT doped with boron.128 The role of the gas phase in this process and in other processes described above is obscure; therefore, the term `vaporisation' is used only conditionally. 3. Sunlight-induced vaporisation The studies dealing with the use of solar concentrators for the preparation of NT have been briefly surveyed in a publication.26 These studies were carried out only in France.Asetup producing a temperature of about 3000K at the focus was developed. Evaporation of pure graphite gave only a small amount of soot, whereas vaporisation of a graphite powder with catalysts gave rise to NT.The type of NT formed (single-walled, multi-walled or bamboo-like) and the amount of impurities were governed by the catalyst and the gas pressure in the process, i.e., by the same parameters as in the case of electric-arc and laser-assisted syntheses, described above and studied much more extensively. In the presence of Co, single-walled NT with a diameter of 1 ± 2 nm were found in the soot deposit and NT bundles with diameters of more than 20 nm containing no admixture of amorphous carbon were found in the web-like deposit. V. Pyrolysis of hydrocarbons and decomposition of CO 1. Catalytic pyrolysis Catalytic pyrolysis of hydrocarbons had been used to prepare carbon fibres even before the discovery of NT and fullerenes.Nanotubes were first synthesised by this method in 1993.129 The course of the process is influenced by the temperature, the overall pressure, the initial hydrocarbon used and its partial pressure, the nature and characteristics of the catalyst (first of all, the particle size, which could determine the diameter of the NT), and the nature of the catalyst support. The pyrolysis can result in layers of amorphous carbon or graphite formed around catalyst particles, carbon fibres and multi-walled or single-walled NT.26 The length of NT and the degree of their coverage by amorphous carbon particles depend on the duration of the process. a. Pyrolysis of acetylene Acetylene is used most often for pyrolysis. Jose -Yacama n et al.129 carried out acetylene pyrolysis at 500 8C and atmospheric pres- sure over a graphite catalyst containing 2.5% of Fe.The concen- tration of acetylene in the diluent (N2) was 9%. It was noted that graphite particles are formed initially around the Fe particles and this is followed by growth of NT with diameters of 5 ± 20 nm and lengths of up to 50 mm. Nine different catalysts (Fe, Co, Ni and Cu supported on three materials � flake graphite, SiO2 or HY zeolite) have been assayed;130, 131 the effects of the flow rate, the temperature and duration of the pyrolysis on the yield and quality of NT have been studied. An increase in the temperature from 500 to 800 8C influences only slightly the length of NT but increases the yield of amorphous carbon (amorphous carbon constitutes up to 50% of the deposit even under the optimal conditions).Longer E G Rakov pyrolysis times result in higher relative yields and greater lengths and diameters of NT. The replacement of the inert diluting gas by H2 barely influences the growth of NT. In all cases, the structure of NT was defective. A graphite coating was formed around the catalyst particles and multi-walled NT were `extruded' from it. Testing of Fe catalysts prepared by various procedures showed 132 that Fe supported on SiO2 provides the best results. The greatest yield ofNTwas attained when C2H2 was pyrolysed at 700 8C. The outer and inner diameters ofNTwere, on the average, 10 ± 20 and 5 ± 8 nm, respectively. As a further development of the study cited,132 pyrolysis of C2H2 over Fe particles in mesoporous SiO2 has been carried out.133 The catalyst was prepared by hydrolysis of tetraethoxysi- lane in a solution of Fe(NO3)3 followed by reduction with an H27N2 mixture at 550 8C.The diameter of NT (*30 nm) obtained at 700 8C from an acetylene ± nitrogen gas mixture (9% of C2H2) was nearly equal to the diameter of pores in SiO2 . It has been assumed that NT with a smaller diameter can also be obtained in this way. Each NT contained 40 to 50 shells and was located at a distance of *100 nm from the closest neighbours; thus, the deposit looked like a `forest' consisting of nanotube `stems' parallel to one another. The rate of the growth of NT was about 25 mm h71.In the initial stages, NT were free from the amorphous carbon impurity; however, as the duration of the pyrolysis and, correspondingly, the length of NT increased (the NT grew to a length of 50 ± 100 mm over a period of 2 ± 5 h), the `stems' bent and amorphous carbon appeared on their surface. The main obstacle hampering the conducting of larger-scale process is the difficulty of manufacturing substrates with large dimensions because they are subject to shrinkage and cracking. An improved method for the synthesis of a supported Fe catalyst has been proposed.134, 135 The method was based on the deposition of a film of a Fe(NO3)3 -containing gel onto a 30 ± 50 mmquartz plate, the removal of excess water and other solvents and drying at 80 8C.As this was done, the gel cracked into pieces with an area of 5 ± 20 mm2 each. These pieces were calcined in vacuo and then the iron was reduced to give small (5 ± 50 nm) iron particles distributed uniformly over the surface. Pyrolysis of a C2H2±N2 mixture containing 9% of acetylene at 600 8C and a pressure of 24 kPa in the presence of this catalyst made it possible to obtain NT in very high yields. The outer diameter of NT amounted to 30 ± 40 nm and the inner diameter was 10 ± 15 nm. They consisted of 10 ± 30 coaxial layers. Individ- ual NT were located at distances of *100 nm from one another. The rate of the growth was 30 ± 40 mm h71 and the length of NT (attained in 48 h) was *2 mm. The greatest area coated by NT was 15 mm2. No formation of polyhedral particles was detected.The layer of NT was easily separable from the substrate. A mechanism consisting in the NT growth on the side of a free tip, which is covered by a catalyst particle, has been proposed for this process.135 The micrographs of NT obtained by the trans- mission electron microscopy, which are presented in a study cited,135 show an irorticle at the tip of the NT and a top- shaped cavity at theNTbase. These results are at variance with the data of Jose -Yacama n et al.129 It has been considered initially that the reaction mechanism should include the formation and decomposition (upon supersaturation) of `active' metal carbides. This mechanism was consistent with the fact that the growth of NT, which has stopped after cooling of the carbonised catalyst, is resumed after the subsequent heating.The root growth of NT can proceed by a mechanism described previously and assuming the surface diffusion of carbon atoms to the bulges on the catalyst surface;43 in this case, dehydrogenation of C2H2 on the catalyst surface to give H2 rather than mere adsorption of carbon can be the first stage of pyrolysis. A mechanism involving immersion of small Cn clusters into the caps of NT has also been considered. It should be stated that catalytic pyrolysis of the same hydrocarbon (C2H2) present in the same concentration in N2 (9%) at similar temperatures (500 and 600 8C) can follow absolutely differentMethods for preparation of carbon nanotubes mechanisms depending on the pressure in the system, the method used to prepare the catalyst and other factors.Pyrolysis of C2H2 over Fe-containing catalysts supported on various materials including SiO2 has been considered in several other publications.136 ± 139 It was noted that, in addition to the NT described above, helical and even branched NT can be formed. Apart from iron, cobalt also exhibits high activity in the pyrolytic decomposition of C2H2.136 The materials used as supports for this metal include SiO2 , 137, 140, 141 NaY zeolite 137 and Al2O3 .142 The procedure for the preparation of Co catalysts, in particular, the pH of precipitation of the metal salt from a solution, influences appreciably the quality of NT.141 Biro et al.143 have classified NT into two groups: `raft-like arrays' with a diameter of 1 nm and single ones with a diameter of 10 nm.Remarkable NT with a coil structure were also isolated upon pyrolytic decomposition of C2H2 over Co. Nanotubes of this type have been studied in detail.144 It was shown by high-resolution electron microscopy that the coils consist of several straight sections, the junctions of which can contain disordered areas. The convolutions closely adjoin one another, the NT forming them being `flattened', i.e. compressed along the coil axis. A certain catalytic activity in the formation of NT from C2H2 is displayed by Ni andMnoxides, while platinum metals proved to be poor catalysts.136 However, a catalyst promoting the formation of very long NT was prepared from Pt5(CNC8H9)10.Catalyst particles can be introduced into a support by impregnating it with aqueous solutions of salts followed by decomposition of the salts, by chemical vapour deposition using organometallic compounds, or by electrodeposition. The size of metal particles can range from several nanometers to tens of nanometers. Zeolites, especially NaY zeolite, make good supports for the catalysts of acetylene pyrolysis. 137 The synthesis described in Refs 133 ± 135 should actually be categorised as a template synthesis. Membranes prepared by anodic oxidation of Al are also good templates for the catalytic synthesis of NT.139 These membranes contain parallel closely packed hexagonal channels (pores) running throughout the whole bulk.By varying the conditions of anodic treatment, one can control the pore diameter (2 ± 500 nm), the membrane thickness (50 ± 500 mm) and the pore density (107 ±1012 cm72). The Al2O3 membranes are transparent and stable up to a temperature of at least 1000 8C; the chemical properties of their surface can change upon the addition of hydroxy groups. Recently a template synthesis of NT by pyrolysis of C2H2 in the presence of a mixed catalyst has been performed.145 Commer- cial zeolite samples were impregnated with aqueous solutions of Co and V acetates and used in the process immediately after drying. The pyrolysis was carried out at 700 8C. This procedure yielded well graphitised and relatively thin NT (10 ± 12 shells, outer diameter 8 ± 10 nm, inner diameter 2.5 ± 2.8 nm).The role of the vanadium additive is unknown; the mechanism proposed in the study cited 145 and involving the transition of metal com- pounds to the gas phase seems doubtful. b. Pyrolysis of methane, ethylene and propylene It has been noted in a publication 137 devoted to the pyrolysis of acetylene that the use of methane, ethylene or propylene instead of acetylene either does not give NT or gives them in low yields. However, in some studies, conditions have been selected under which pyrolysis of these compounds does yield NT. Thus a deposit consisting of NT to an extent of 90% was obtained from CH4 at 600 8C in the presence of the NixMg17xO catalyst, prepared from Ni and MgO, in which some Ni atoms were located on the catalyst surface as aggregates.146, 147 The synthesis of NT with a length of >20 mm from a CH4±H2 mixture in the presence of Fe particles has been reported.148 An interesting point in this study 148 is that NT were synthesised using a microwave setup (Fig.6), widely employed for the preparation of diamond films. However, the 43 1 2 3 4 6 5 7 89 10 Figure 6. Setup for plasma-activated chemical deposition from the gas phase. The pressure of the mixture of CH4 and H2 at the inlet is 0.13 Pa, the substrate temperature is 370 ± 950 8C; (1) microwave antenna; (2) quartz window; (3) gas inlet; (4) viewing window; (5) plasma discharge; (6) to the optical emission spectrometer; (7) substrate; (8) an energy-sensitive graphite stand; (9) drive to a stepper motor; (10) to a vacuum pump.synthesis of single-walled NT carried out at Stanford University (USA) appears the most impressive.149, 150 In this study, methods of nanotechnology were employed. The catalyst was applied to a silicon plate as micron-sized islets. To this end, square holes were made in a poly(methyl methacrylate) film by electron-beam lithography and then the catalyst precursor was deposited onto the film from a methanol solution containing Fe(NO3)3 , MoO2(acac)2 and Al2O3 particles. Subsequently the solvent and the film were removed, the catalyst was heated in argon, and, finally, pyrolysis of CH4 was carried out. During pyrolysis, very straight NT with diameters of 1 ± 3 nm and lengths of up to 20 mmcontaining no topological defects were formed on the catalyst islets over a period of 10 min.Some NT served as bridges between neighbouring catalyst islets and thus formed electric contacts. An increase in the process duration to several hours resulted in thicker NT. This method is rather straightforward and easily reproducible; it can be used to prepare NT on substrates having large areas. The CoSi2 formed upon the deposition of cobalt onto silicon also catalyses the formation of NT from CH4 .151 The use of ethylene for the template catalytic synthesis of single-walled NT on porous Al2O3 plates has been described.152, 153 Pyrolysis of propylene 132, 137, 139 and pyrolysis of polyethylene over a Ni catalyst have been reported.154, 155 c.Pyrolysis of benzene and other aromatic compounds Benzene can also serve as the initial hydrocarbon for the synthesis of NT; this synthesis can be carried out using the equipment designed for the preparation of carbon fibres fromC6H6 .156 It was noted that the growth of NT and carbon fibres follow different mechanisms; indeed, catalyst particles were detected at the tips of the fibres, whereas the NT usually had conical caps free from inclusions. The preparation of NT from a C6H6±H2 mixture over an Fe catalyst requires as a rule lower partial pressures of C6H6 and lower temperatures than the synthesis of fibres. In some cases, the growth of NT is replaced by the growth of solid fibres. Solid fibres and NT prepared from the same hydrocarbon in single-type equipment differ markedly in mechanical properties � the fibres are easily broken on bending, whereas NT exhibit flexibility and elasticity.Pyrolysis of C6H6 over Co supported on SiO2 affos simultaneously bent and helical NT and fibres,157 the ratio of the yields of NT and fibers being determined by the size of the catalyst particles and the composition of the reaction mixture. A catalyst containing a higher proportion of Co (5 mass %) and consisting of larger particles (11 nm) results in a higher proportion of fibres than a catalyst with a lower metal content (0.5 mass %) and a44 medium particle size (7 nm). The replacement of the gas diluent H2 by N2 increases the yield of NT.The helical NT resulting from the pyrolysis of C6H6 differ from those formed upon the pyrolysis ofC2H2 (see above 144); they are multi-walled NT, the distance between the shells being intermediate between the corresponding distances in ordinary and turbostratic graphites. The shells themselves are largely disordered.157 Among other reasons, the formation of such structures might be due to the addition of hydrogen (the diluent of benzene) to the dangling bonds of NT. However, the catalyst particles exert a greater influence. This hypothesis is consistent with the model proposed by Amelinckx et al.,158 which assumes that the growth ofNTprogresses on the side of a free tip having an encapsulated catalyst particle. Deposits of NT with a definite texture can be obtained from gaseous C6H6.Thus pyrolysis of C6H6 on a profiled substrate with scratched parallel grooves gives rise toNT arranged parallel to the substrate surface and growing from the upper edges to the centres of the grooves.159 It was suggested that irregularities on the upper edges of the grooves exhibit enhanced catalytic activity. A similar expedient has been used in the pyrolysis of 2-amino- 4,6-dichloro-s-triazine.160, 161 A 10- to 100-nm thick Co film was sprayed onto a SiO2 plate. Then 1 to 20 mm wide tracks spaced 100 ± 200 mmapart were etched on the substrate using a laser. The substrate was placed in a furnace the coating downwards and pyrolysis was carried out at 950 8C. This gave rise to NT bundles having a strictly identical length and very close outer diameters (30 ± 50 nm).The NT obtained in this way contained up to 5% of nitrogen. When the plate was placed in the furnace the coating upwards, the situation changed sharply; strongly twisted NT were formed.161 This outcome has not yet been adequately interpreted. The researchers believe that etching of the Co film gives rise to energy-saturated metal or metal oxide clusters, which are depos- ited along the track edges. Hydrogen chloride, arising upon decomposition of triazine, can convert Co into volatile CoCl2 and thus influence the catalytically active particles at the tips of NT. The growth of NT is terminated when the catalyst species have lost activity or have been completely evaporated as the chloride.Pyrolysis of tris(aminotriazine) (melamine) and cyanuric acid over etched Co, Ni, or Fe films occurred in a similar way. A mixture of gaseous phenylacetylene and thiophene with He 162 and a mixture of 2-methyl-1,20-dinaphthyl ketone with He 163 in the presence of a Ni catalyst have also been used for the pyrolytic synthesis of NT. It is noted in both studies that at the first stage of pyrolysis, Ni particles are coated with a graphite layer; this is followed by the nucleation and growth of NT. The temperature of choice is 700 8C. Yudasaka et al.163 proposed an original procedure for the preparation of the catalyst. They sputtered a metal film on a quartz glass substrate and heated it in vacuo. The size of the Ni particles thus formed was determined by the thickness of the deposited film; in the optimal case, it was 20 ± 30 nm.It has been concluded 162 that NT are `extruded' from Ni3C particles; Yudasaka et al.163 noted an important role of the shape of Ni nanoparticles. 2. Pyrolysis in the presence of a `floating catalyst' A conventional supported catalyst is coated, sooner or later, by a layer of the products of pyrolysis of hydrocarbons and is thus deactivated. The introduction into the system of catalyst precur- sors as volatile compounds which decompose to give catalytically active species directly in the reaction area could, in principle, permit one to avoid the catalyst deactivation and to bring the process of pyrolysis closer to a continuous process.There are only a few publications in which this approach has been employed. Thus decomposition of Ni phthalocyanine at 700 and 800 8C has resulted in the synthesis of NT containing some nitrogen.164 These NT were multi-walled particles with large (*200 nm), small (10 nm) or variable diameters, their length reaching 6 mm. It has been assumed 164 that the mechanism of E G Rakov the NT growth includes encirclement (encapsulation) of Ni particles with a carbon layer, graphitisation of this layer, the change in the shape of Ni particles, the appearance of seeds and the growth of NT. The Ni particle inside theNTcan either migrate from the substrate to the growing tip or remain in a particular section of the NT, forming a node point. Decomposition of Fe and Co phthalocyanines is also accom- panied by the formation of NT, whereas Cu phthalocyanine gives neither NT nor graphitised particles.Ferrocene has been used as the `floating catalyst' in the pyrolysis of thiophene.165 Unlike the study cited above,164 in this case, single-walled NT were produced. The process was carried out at 1100 ± 1200 8C for 1 to 30 min. This yielded a large amount of long and relatively thick `ropes' and ribbons coloured silvery black. Some of the ribbons were semitransparent and not attached to the surface. The longest `ropes' were 3 ± 4 cm long and had a diameter of 0.1 mm, while the ribbons were up to several millimeters wide. Each `rope' and each ribbon consisted of several thousands of vaguely oriented threads, comprised by bundles of well oriented closely packed single-walled NT.The diameter of the bundles ranged from several to forty nanometers, the average diameter of the NT being 1.7 nm. It was found that the bundles are formed in the reaction area and carried away with the gas stream; outside the area they stick together forming `ropes' or ribbons. Pyrolysis of ferrocene and its mixtures with C2H2 has also been studied.166 3. Decomposition of CO Thermal decomposition of CO (disproportionation to CO2 and carbon) differs sharply from the pyrolysis of hydrocarbons from the thermodynamic viewpoint; at atmospheric pressure and low temperatures (300 ± 750 K), the equilibrium yield of carbon is nearly quantitative, while at higher temperatures and lower pressures, it diminishes.Conversely, the yield of carbon in the pyrolysis of C2H2 and CH4 increases as the temperature rises and the pressure is reduced and approaches a quantitative yield at 1250 ± 1500 K. In this respect, carbon monoxide may seem to be a less convenient starting compound for the synthesis of NT. How- ever, for kinetic reasons, a hydrocarbon is more difficult to heat to a temperature above 800 ± 900 8C (before it is brought in contact with the catalyst) than CO; hence, carbon monoxide offers some advantages. The first study 167 on the catalytic decomposition of CO to give NT was carried out in 1995, i.e. much later than NT were synthesised by other methods. Decomposition of a CO±CO2 gas mixture (20% of CO) was carried out at 500 8Cover the Ni/Al2O3 catalyst.A graphitised shell was formed around the Ni particle; nanotubes protruded out of this shell. The introduction of hydro- gen into the gas mixture changed appreciably the morphology of the deposit. Thermal decomposition of neat CO over a Ni ± Co catalyst supported on Al2O3 was performed at 1200 8C at a pressure somewhat higher than atmospheric. 168 This process gave rise to single-walled NT. The preparation of an active molybdenum catalyst used for the synthesis of NT from CO was an important achievement. The catalyst was prepared by impregnation of alumina with a methanol solution of bis(acetylacetonato)dioxomolybdenum fol- lowed by heating to 200 8C.The diameter of NT formed in the presence of this catalyst was 1 ± 5 nm. Catalyst particles with sizes of several nanometers were detected on the tips of many of the NT. This provided grounds for proposing a mechanism of the growth, which was called the `skull-cap' mechanism. According to this mechanism, a catalyst particle promotes dehydrogenation of the hydrocarbon molecules deposited on the particle from the gas phase. Carbon diffuses to the open tip of the NT (where the catalyst particle is located) over the catalyst surface or through the bulk and is embedded in the NT structure. It is the coatingMethods for preparation of carbon nanotubes consisting of chemisorbed carbon atoms on the surface of the catalyst particle that is referred to as `skull-cap'.It prevents the formation of dangling bonds. Micrographs of single-walled NT with catalyst particles located on the open tip and having a size equal to the NT diameter were reported.168 Testing of the Ni ± MgO, 146, 169 Co ±MgO170 and Ni ± AlPO4 169 systems as catalysts for decomposition of CO gave positive results. Not only multi-walled NT147 but also single-walled NT containing no deposit of amorphous carbon on the external surface have been synthesised from CO. None of the proposed mechanisms 171 is able to explain the influence of the nature of the initial gas on the morphology of NT and other structures. 4. Synthesis of nanotubes containing B ±C± N, C±N and B±N It has been noted above that the catalytic pyrolysis of triazine gives rise to nitrogen-containing NT.160 The pyrolysis of CH3CN.BCl3 at 900 ± 1000 8C over a Co powder afforded fibres andNTof the composition BxCyNz with various morphologies.172 Nanotubes of the composition C38N, the properties of which differed markedly from those of carbon NT, were obtained from pyridine in the presence of a Co catalyst; the nanotubes BC28N were synthesised from the adduct (CH3)3N.BH3. 173 All the nanotubes thus prepared were multi-walled. Since the yield of NT in the pyrolysis of the adduct NH3 .BH3 was very low, it was proposed that coating of metal particles with graphite is a necessary initial stage in the formation of boron- and nitrogen- containing NT. The data available to date are still inadequate for drawing definite conclusions about the reaction mechanisms, although the shape of some NT provided grounds for speaking about the unusual phenomenon of periodic motion of the catalyst particles located inside the NT along the NT axis.172 Previously a similar mechanism had been described for purely carbon NT with a bamboo-like structure.174 VI.Growth of nanotubes by decomposition of metal carbides When performing arc synthesis, researchers have noticed that particles of catalysts (carbides) are first covered by an envelope of several graphitised layers, which serve as the source for the growth of NT.175 The mechanism of formation of such a structure (which has been termed `sea-urchin') includes the formation of super- saturated solutions of carbon in a metal or metal carbide, subsequent segregation of carbon from these solutions and the `root' growth of NT.`Sea-urchins' were formed on particles of YC2 , 71, 175 LaC2 , 176 GdC2 , 177 Ni3C, 178 TiC 179 and other carbides. Heating of silicon carbonitride under a static atmosphere of N2 at 1400 8C or in a flow of N2 at 1850 8C resulted in the synthesis of multi-walled NT with a diameter of 10 ± 25 nm and with a length of up to 1 mm.180 According to a publication which appeared almost simultaneously,181 laser ablation of a-SiC resulted in sublimation of silicon and the formation of NT on its surface. Soon this process was even more simplified; in particular, laser treatment was replaced by resistivity heating of powdered 182 or single-crystalline 183 silicon carbide.Heating of powdered silicon carbide in vacuo at 1600 ± 1700 8C allowed preparation of NT on a relatively large area over a period of 10 ± 15 min. These NT grew at right angles to the external surface of the powder; they were substantially shorter than those obtained by the arc method but longer than those prepared by the laser ablation of SiC. Even more uniform NT (`a forest') were synthesised at a temperature of 1700 8C and a pressure of 1.361072 Pa on a 36560.34 mm plate. They grew to a length of 0.15 mm over a period of 30 min and were distinguished by a fairly regular mutual orientation. Apparently, this procedure is the simplest way of producing cathodes for field electron emitters.45 Small amounts of NT were formed in Fe7Ni and Fe7Ni7Co alloys containing carbon 184, 185 and in complex solid solutions containing Fe3C. 186 During subsequent studies, this procedure could be modified for the preparation of NT- containing composites possessing specific properties. VII. Other methods 1. Synthesis in the flame Combustion of hydrocarbons in oxygen is employed for the synthesis of diamonds (see, for example, Ref. 187) and full- erenes.188 Nanotubes can also be prepared in this way.189 It is of interest that the C:O ratios, which were equal to 1.06, 1.07 and 0.86 ± 1.00 for the combustion of C2H2, C2H4 and C6H6, respec- tively, are close to the optimal values for the synthesis of diamond. However, these processes were carried out under different conditions � in the synthesis of diamond, the inner cone of the flame was directed at a substrate cooled to a particular temper- ature, whereas in the preparation of NT, catalysts were used.The data on the synthesis of NT in flames are too scarce to analyse them, as has been done for the synthesis of diamonds (see, for example, Refs 190, 191). It is possible that NT, like diamonds, can be synthesised using any carbon compound, no matter what other chemical elements (halogens, sulfur, nitrogen, phosphorus, silicon, boron, etc.) are contained in the compound. 2. Electrolysis of molten salts Nanotubes can also be synthesised without participation of a gas phase, namely, in ionic salt melts. The electrolysis of molten LiCl at temperatures above 600 8C in a cell with carbon anode and cathode causes a heavy erosion of the cathode and appearance of sludge in the melt, which can be washed by water and toluene after cooling the melt.192 In addition to spheroidal and polyhedral particles, multi-walled (2 ± 10 walls) NT with a diameter of 2 ± 10 nm were detected in the sludge.The current density has a strong influence on the yield of NT. The yield and the quality of NT also depend substantially on the temperature of the melt and the nature of the salt used (LiCl, NaCl or KCl).193 The electrolysis should not be carried out for a long time because the sludge can short circuit the electrodes. The electrolysis of a LiCl ± SnCl2 melt resulted in the synthesis of NT filled with b-Sn, i.e.a nanowire.85 3. Chlorination of carbides and other methods According to a publication,194 back in the 1960s, multi-walled NT were obtained by chlorination of SiC and TaC at 800 ± 850 8C; the electron micrograph of these NT was published in 1978 in a book.195 Presumably, the coke that had been prepared using a graphite electric heater (with a temperature of at least 2500 8C) and employed for the synthesis of carbides, served as the source of NT.There also exist other methods for the synthesis of NT, for example, pyrolysis of the powdered polymer prepared by poly- esterification of oxalic acid and ethylene glycol,26 and the reaction of metallic Cs with nanoporous carbon, prepared from a mixture of polyfurfuryl alcohol and polethylene glycol, carried out at 50 8C.196 A method for the synthesis of NT similar to that described in Ref.148, has been proposed by KuÈ ttel et al. 197 The researchers prepared NT in a microwave discharge using a setup for the plasma synthesis of diamond coatings and a CH4 (2%) ±H2 (98%) mixture at a pressure of 0.4 kPa; Ni or Fe islets were applied preliminarily on the substrate, the temperature of the substrate being only slightly higher than the optimum temperature for the synthesis of diamonds (900 ± 1000 8C). The deposit had a `spa- ghetti' structure and consisted of NT with a diameter of 20 ± 60 nm and a length of up to 100 mm entangled with onenother and firmly attached to the substrate. The catalyst particles remained near the NT `roots' during the synthesis.46 The same researchers were able to perform the synthesis ofNT using the `classical' setup used for chemical vapour deposition of diamonds, namely a hot-filament reactor.Since setups similar to that used by KuÈ ttel et al. 197 are widely employed in dozens of laboratories in many countries and do not need to be modified for switching to the preparation of NT, further development of these methods for the synthesis of NT may be expected. Recently, syntheses ofNTby heating carbon black with boron at 2200 8C and by detonation of 2,4,6-triazidotriazine have been described.198 An interesting feature of the latter method (explosive synthesis) is that it permits preparation of NT with a record- breaking inner diameter equal to 80 ± 120 nm (although they are formed in a very low yield).VIII. Purification and opening of nanotubes Detailed investigaton into the properties of NT and some of their practical applications require individual and uniform open-tipNT containing no impurities. However, they are usually obtained as bundles consisting of NT with different lengths, capped at one tip and contaminated with impurities. Therefore, techniques for purification and opening of NT are as important as the methods used to synthesise them. Unlike fullerenes, neither individual NT nor, all the more so, bundles are soluble in any solvent, which complicates the problem of their purification. Nevertheless, several techniques which permit more or less successful purification of NT have been proposed.The methods are based on the following characteristic features of NT. The materials that typically contaminate NT (fullerenes, polyhedral graphitised particles, amorphous carbon) are more reactive than NT and some of them (for example, fullerenes) are soluble in organic solvents. Sections ofNTwith a higher density of defects also exhibit higher reactivity than defect-free NT. This refers first of all to the caps at the NT tips, which contain not only six-membered but also five-membered carbon rings. The carbon atoms in these rings are more reactive. The sections on the lateral surfaces of bent NT possess similar properties because kinks would be impossible without insertion of either five-membered (positive curvature) or seven-membered (negative curvature) carbon rings in a net consisting of standard six-membered rings.Finally, enhanced reactivity is also characteristic of the atoms at the edge dislocations of NT (the `scroll' or `papier-mache' structure), in which dangling bonds are concentrated together with defects arising upon replacement of carbon atoms by atoms of other elements. Multi-walled NT, which are normally more defective than single-walled ones, are more reactive, whereas NT annealed at high temperatures (annealing leads to elimination of defects) are less reactive. Nanotubes filled with metals or metal carbides differ in density from empty NT. The methods used to purify nanotubes can be divided into three groups�chemical, physicochemical and mechanical.1. Chemical methods The simplest method for opening NT is selective oxidation of the caps, which can be carried out by gases, melts or aqueous solutions. The gaseous oxidants used for this purpose are O2 (see Ref. 62), air,199 CO2 (see Ref. 200) and oxygen plasma.201 The maximum rates of oxidation with air are attained at 420 8C for C60, at 645 8C for graphite and at 695 8C for NT or nano- particles.199 Oxidation with O2 or with air proceeds most efficiently at 650 ± 750 8C. Even at 550 8C, a gas flow containing 1% of O2 causes disordering of the outer shells of NT and formation of wells with a diameter of 2 ± 10 nm on their surface; as a consequence, thin NT swell up.202 At 750 8C, gasification readily occurs at the sites of cracks, defects or deformations.E G Rakov After opening of NT, the oxidation slows down, whereas oxidation of nanoparticles goes on until they are completely eliminated.203 However, complete elimination of nanoparticles requires that more than 99% of the initial material be oxidised. Meanwhile, even after 95% oxidation, a sample contains only 10%± 20% of the initial NT. This is due to the fact that not only caps but also lateral walls of the NT are oxidised. As this takes place, multi-walled NT become thinner and some of them are completely gasified. To purify multi-walled NT in air, heating by IR radiation can be used.204 In particular, a 0.1-mm thick spongy paste with an area of *10 mm2 consisting of multi-walled NT was obtained in this way from the products of arc synthesis in an H2 flow.In some cases, it is sufficient to heat the material in air at 500 ± 600 8C for 30 min.205 Purification can also be carried out in hydrogen plasma.201 However, in practice, purification of multi-walled NT from amorphous carbon is carried out by treatment in an H2±N2 mixture at 900 8C.137 The method of oxidation in melts 59, 61 has not been further developed. In the case of single-walled NT, oxidation in aqueous solutions is apparently the most important method. The most frequently encountered procedure is refluxing in concentrated (60% ± 70%) HNO3 . 109, 203, 206 ± 208 The products of catalytic synthesis are freed simultaneously from the inevitable metal impurities.After refluxing for 4.5 h, the loss of mass in the material obtained by the arc method was less than 2%; as a result, 80% of NT became open.206 Refluxing of 10 g of the initial material produced by the laser ablation method in 1 litre of a 2 ± 3 M solution of HNO3 for 45 h resulted in a 70% loss of weight.108 Apart from treatment with nitric acid or after this treatment, NT are made to react with mixtures of HNO3 with H2SO4 and of H2SO4 withH2O2 (see Ref. 103). In some cases,H3IO5 and HClO4 (see Ref. 209) or solutions based on concentrated HCl 210 have been used instead of HNO3. Other compounds used as oxidants include H2O2, K2Cr2O7 , KMnO4 , solutions of Ru and Os chlorides in a solution of NaIO3, etc.Some of the oxidants (acidic solutions of H2O2 and K2Cr2O7) proved to be much less selective and efficient than HNO3 , some other (solutions of KMnO4 containing MnO2 or CrO3) exhibited medium activity, and a group of reagents (solutions of Ru and Os chlorides in a solution of NaIO3) exhibited a very high reactivity and ensured opening of up to 80% ±90% of NT at 100 8C.211, 212 Oxidation in solutions can be combined with filling of NT (for this purpose, a soluble metal salt is added to the solution 206, 210) or with chemisorption of metals on the NT surface.213 Chemisorp- tion is due to the fact that upon treatment with an acid, the surface of NT is coated with the acidic groups COOH, which can react with metal ions. The quantity of the Pd2+ ions absorbed was found to be strictly correlated with the concentration of acidic groups on closed or open NT.213 An especially high density of acidic groups is attained by using a mixture ofH2SO4 withHNO3 ; these groups promote deposition of finely dispersed metal clusters on the NT surface (see, for example, Ref.214). The data on wetting, filling and coating of NT have been briefly surveyed in a publication.215 Foreign particles can be removed using organic solvents, for example, toluene, carbon disulfide and other solvents, while metal particles can be removed with the aid of acids. In a pioneering study dealing with purification of NT,216 chemical modification of NT by grafting dichlorocarbene to the graphene wall at a double bond has been proposed.In a later publication,217 preparation of soluble NT has been reported. The researchers attached a long-chain amide in place of the carboxy groups at the tips of cut single-walled NT. This was attained by treatment of NT with thionyl chloride (70 8C, 24 h) and then with octadecylamine (90 ± 100 8C, 96 h). The product was readily soluble in chloroform, dichloromethane, aromatic compoundsMethods for preparation of carbon nanotubes and CS2 . These studies open up new prospects for the develop- ment of more facile methods for the pification, separation, study, description and application of NT.218 2. Physicochemical and mechanical methods In recent years, a number of physicochemical and mechanical techniques for the purification of NT have been proposed.They are briefly described below. A physicochemical method, chroma- tography, has been used to purify both multi-walled 219 and single- walled NT.220 The initial NT are dispersed in aqueous media, the dispersions being stabilised by adding surfactants. After separa- tion of the material into fractions using a column with a porous glass packing (with an average diameter of pores of 300 nm), NT are isolated from the individual fractions by centrifuging. Thus not only impurities are removed but, in addition, NT of different lengths are separated. `Cutting' of NT into 100 ± 300-nm long sections has been reported.221 The foundations of an electrophoretic method for the purifi- cation of NT (as a suspension in isopropyl alcohol) have been outlined in a study.222 Various mechanical procedures for the purification of NT are known.They include sonication, microfiltration and centrifug- ing.103, 199, 206 ± 208, 223 ± 225 Many of these procedures are labour- consuming and include large numbers of stages; in some cases, they are used only in combination with chemical methods. Sonication increases the density of defects, especially, in multi- walled NT. The ability of individual NT to form bundles (`ropes') manifests itself not only during the synthesis but also in the purification � in some cases, the purified material contains bundles with greater diameters than the initial material. Under some conditions of sonication in acids, the bundles of NT assume a ring shape with a diameter of 0.25 ± 0.55 mm.226 IX. Conclusion Particular examples of combined strategies for the purifica- tion of single-walled NT can be found in the literature.103, 208 Methods for the synthesis of NT are fewer in number, judging by reviews, 3, 7 than methods for the synthesis of fullerenes. Never- theless, the main range of these methods has already taken shape and now the question is what of these methods are best suited for the large-scale synthesis of NT. Different methods can only be compared conventionally because in most cases, the necessary characteristics are not fully available or not reported at all. First of all, methods that require vacuum or elevated pressures should be distinguished from the procedures of synthesis at ambient pressure.The former are less productive. The vaporisation of graphite induced by electron or ion beams and the use of diamond anvils also imposes some restrictions on the productivity. It is also expedient to classify as a separate group those methods that cannot be easily made continuous, for example, electrolysis or explosive synthesis. Many investigators agree in the opinion that the most widely used electric-arc and laser-ablation methods are applicable only on a laboratory scale, whereas pyrolysis should become the most important commercial method.131, 156, 165, 227 This method is similar in many respects to the pyrolytic method long employed for the production of carbon fibres both in the presence and in the absence of catalysts.The necessary equipment is relatively simple. It is the pyrolytic method that was used to prepare the longest known thread-like bundles of NT and macroscopic ribbons (`mats') of interwoven NT.165 The advantages of the pyrolytic method are manifested especially clearly in the production of ordered structures by deposition of NT onto a smooth substrate with a supported catalyst or onto a porous template. Neither arc nor laser methods can give such structures. Mechanical expedients have been developed which allow laying of NT of any origin parallel to one another (for example, dispersion in a matrix made of a polymeric 47 resin followed by cutting thin slices of the composite 228 or microfiltration of a suspension in ethanol through specific ceramic filters followed by transfer onto a plastic surface 229).However, they are all much more complicated and much less elegant than catalytic pyrolysis and cannot in full measure compete with it. Ordered structures fromNTprepared by the pyrolytic method are best suited for template synthesis, i.e. application of metals, for example, Ni or Co on their surface,142, 230, 231 for deposition of oxides,63 carbides 232 ± 235 or gallium nitride,236 for the preparation of catalysts, sorbents, membranes for electrochemical cells,152 field emitters and artificial muscles (actuators). Pyrolysis of hydrocarbons in flame 188 and in plasma 196 are also promising because both methods are facile and can be implemented in a continuous mode.Diamond coatings are already prepared on an industrial scale by means of plasma devices; the principal (perhaps, the only) condition for using them for the synthesis of NT is the presence of an appropriate catalyst on the substrate. However, it may happen that NT meant for particular purposes will be prepared by other methods which have not yet been tested, for example, in plasma jet devices. However, the use of solar energy concentrators for the preparation of NT can hardly be expected in Russia, because the intensity of solar radiation in Russia is relatively low and the level of development of this method does not surpass that for other methods of heating.237 Studies of carbon NT have provoked interest in the prepara- tion of nanotubes consisting of other inorganic compounds.Tubular structures are possible for a fairly large number of inorganic compounds of various classes.238 Studies on the syn- thesis of NT from BCx , BN, BCxNy and CNx are only at the beginning; however, in this case, too, pyrolytic methods would apparently prove to be the most convenient. The problem of the synthesis of NT with identical geometries might be solved using the method proposed by Smalley, namely, selection of homogeneous seed NT and their further growth by pyrolysis of hydrocarbons (http://www.dtic.mil/dusdst/agenda/ agenda31999.html). A complicated synthetic task would be routine preparation of NT with reproducible heterojunctions (variation of diameter, chirality or chemical composition); some enthusiasts believe that this could serve as the basis for the nascent nanoelectronics.The design of nanomechanical devices in which NT are expected to serve as important and abundant parts will also require great efforts. * * * Since this manuscript was prepared for the publication, new journal papers and Internet sites devoted to the synthesis of carbon NT have appeared (or become available). Thus the mechanism of formation of carbon NT, nanowires and nanoparticles in an electric arc burning under a hydrogen atmosphere has been described. The participation of polycyclic aromatic hydrocarbons in the synthesis of NT has been con- firmed. A direct correlation between the ability of H2 to remove graphenes from an anode and the yield of NT has been eluci- dated.239 It was found that only three platinum metals, Rh, Pd and Pt, induce the formation of single-walled NT in an electric arc, whereas mixtures of metals (Ru ± Pd, Rh ± Pd, Ru ± Rh, Ru ± Pt and Pd ± Pt) exhibit low or zero activity; only a Rh ± Pt mixture (1 : 1 or 5 : 2) is active in this process.240 At a helium pressure of 78 kPa over a Rh ± Pt catalyst, NT with approximately equal diameters (1.280.07 nm) were formed.A10 ± 12 fold decrease in the helium pressure resulted in the growth of nanotubes of different diameters (0.7 ± 1.3 nm).A novel method for the synthesis of NT in a pulsed arc setup has been developed.241 The highest yield of NT was attained in a48 flow of Ar or Kr at temperatures of >1000 8C and pulse duration of >1 ms.Interesting results concerning growing of NT by the electric- arc method have been obtained in a study by Chang et al.,242 who used an anode with a diameter of 6 mm and a 10-mm thick disk 30 mm in diameter as a cathode. Both electrodes were intensely cooled with water and the electrolysis was carried out in a self- sustained mode (interelectrode distanc3 ± 4 mm, voltage 20 V, current 55 ± 65 A, helium pressure 67 kPa). The cathode deposit was formed on an area with a diameter close to the anode diameter. The morphology of the deposit differed appreciably from the standard morphology. A bed of chaotically entangled slightly bent NT (diameter 10 ± 40 nm, length >30 mm), the axes of which were oriented randomly, was located under a thin tough shell. These NT had no defects, contained no amorphous carbon impurity and could be purified from the nanoparticle impurity by merely keeping in air (850 8C, 1 h).The intense cooling of the electrodes and gases near the cathode as well as the large distance between the electrodes, which influences the mode of the current passage, is assumed to be largely responsible for the sharp change in the morphology and properties of the arc synthesis products. A detailed study of some parameters of the laser-induced thermal synthesis of NT has been reported.243, 244 In particular, it was found that the amount of carbon being vaporised from the target is determined by the intensity of laser radiation, while the course of the chemical reaction yielding NT is governed by the temperature in the furnace.The rate of evaporation of catalysts (Ni and Co) is affected by both parameters. The laser radiation intensity has a slight influence on the diameter of NT, while a decrease in the furnace temperature results in smaller diame- ters.243 The increase in the delay between single laser pulses from 0.1 to 120 s has almost no influence on the structure of single- walled NT but decreases the yield of the web-like product.244 It has been shown theoretically 245 that the use of shorter pulses (picoseconds instead of nanoseconds) and higher frequencies (tens of megahertz) brings the synthesis of NT closer to a quasi- continuous mode.It has been shown 246 that the laser-assisted synthesis of NT in the presence of Ni and Co can also be performed in a flow of N2 . The application of a continuous-wave CO2 laser to the synthesis of NT without additional furnace heating of the target is a method new in principle.247 In this study, a vertical setup was used with a rotating target moving along its axis (a rod 6 mm in diameter) and made of a mixture of graphite with a catalyst. A hot zone with a diameter of*1 cm was formed around the laser spot; this resulted in local heating of the gas flow and prevented the plasma from cooling too fast. The synthesis was carried out under Ar. When the radiation power was 250 Wand the diameter of the laser spot was *1 mm, the vaporisation rate reached 200 mg h71.The yield of single-walled NT in the presence of Ni ±Y or Ni ±Co catalysts was*20% based on the amount of the vaporised material. French investigators have successfully tested a prototype of a 1000-kW solar concentrator for the synthesis of NT (http://www.uiuc.edu/cnrs/Cnrspresse/en352a3.htm). The formation of NT on irradiation of graphite with high- energy Ne+, Kr+ and Xe+ ions has been studied.248 Numerous publications in 1998 ± 1999 were devoted to the pyrolytic synthesis of NT. For instance, the influence of the substrate on the catalytic pyrolysis of C2H2 in the presence of Co or Fe has been considered. 249 The influence of the natures of the catalyst (Fe2O3 , CoO, NiO or their mixtures) and the support (Al2O3 or SiO2 with a specific surface area of 100 and 300 m2 g71, respectively) on the formation of single-walled NT in the pyrolysis of pure CH4 has been studied.250 Relying on the fact that the tips of growing NT contained no catalyst particles, the researchers 250 concluded that the process followed a `root' growth mechanism. The highest yield of individual NT was attained with the Fe2O3/ Al2O3 catalyst. E G Rakov The pyrolysis ofmetallocenes and Fe(CO)5 mixed withC2H2 or C6H6 has been studied.251, 252 It was found that the content of the amorphous carbon impurity can be diminished by introducing H2 into the gas phase. The conduct of the pyrolysis of C2H2 by means of a hot filament with plasma activation permitted the substrate temper- ature to be decreased to 650 8C.253, 254 Yet another novelty was the use of an NH3 admixture, which was found to function as a catalyst.The diameter of the NT produced in this way was dictated by the thickness of the Ni film sprayed preliminarily onto a glass substrate and varied from 20 to 100 nm. The growth rate was several times higher than that attained in earlier studies (see, for example, Ref. 133) and was equal to 120 mm h71. Pyrolysis of naphthalene vapour in the presence of Cr(CO)6 vapour with `anode field activation' was performed for the first time.255 The process at a voltage of 4 ± 6 kV, an anode temper- ature of 1100 ± 1200 8C and a vapour pressure of*8 Pa afforded carbon NT with Cr nano-rods protruding from their inner cavity.These rods had a strictly invariant diameter (*10 nm) along the full length (up to 0.5 mm). The attempts to accomplish a similar process with Mo(CO)6 or W(CO)6 failed. Yet another modification of the pyrolytic synthesis of NT includes two-stage heating in vacuo (first at 350 ± 450 8C and then at 500 ± 800 8C) of tripropylamine introduced in the channels of AlPO4-5 single crystals.256 The NT thus formed had equal diameters and equal lengths; however, they proved to be unstable with respect to HCl. Nanotubes consisting of a new promising material, carbon nitride, have been prepared on a graphite substrate from a C2H2±N2 mixture in a microwave plasma using a bias poten- tial.257 A peculiar method for the synthesis of B,N-containing NT is based on the replacement of C atoms in carbon NT by B and N atoms upon the reaction of NT with B2O3 vapour diluted with N2 .258 The preparation of multi-walled NT filled with molten UCl4 or its mixtures with KCl has been described.259 A method for purification of NT (prepared by the pyrolysis of C2H2 over Co introduced into NaY zeolite) has been described, which involves dissolution of impurities in hydrofluoric acid and subsequent oxidation of NT with a solution of KMnO4 or air.260 Afairly comprehensive study 261 is devoted to the behaviour of NT treated with a mixture of concentrated HNO3 and H2SO4 . The presence of acid functional groups on the surface of these NT makes it possible to prepare viscoelastic gels; on drying, solid materials or films with specific structures and properties can be obtained.Under particular conditions, the dispersions can be converted into a material reminiscent of liquid crystals. Fullerenes have been synthesised by irradiation of C6H6±O2 orC6H6±N2Omixtures with aCO2 laser in the presence of SF6.262 This method might also enable the synthesis of NT. The preparation of Langmuir ± Blodgett films from matrix- diluted single-walled NT and application of the films onto various surfaces using micelle-like aggregates has been reported.263 Soluble carbon NT coated with poly(phenylacetylene) have been synthesised.264 Many of the studies mentioned above were reported at the conference on commercialisation of the achievements in the large- scale production of carbon NT (USA, Washington, April 1999; http://www.knowledgefoundation.com/carbon.html). The pros- pects for the investigation and use of NT-based materials were also discussed at the conference of the Materials Research Society (USA, Boston, December 1998), which has been described in a brief review.197 Studies on the application of NT in electronics were surveyed by Johnson at the International Conference on Solid-State Circuits (USA, San-Francisco, January 1999; http:// www.eet.com/story/OEG19990217S0045). Several new promising fields of application of NT have appeared quite recently. The greatest interest was aroused by the development of a prototype of artificial muscles, actuators, activated by application of a low electric potential, which wasMethods for preparation of carbon nanotubes reported by an international group of scientists 265 ± 268 (see also M Fox Infoseek News May 21, 1999; http://www.mpg.de/news99/ news26_99.html; and http://www.msnpc.com/news/271534.asp).A new route to the design of NT-based parts of electronic devices (amplifiers, switches, logical elements) was discovered.269 The possibility of isotope separation with the aid of NT270 and of usingNTfor nanolithography 271 was demonstrated. An electrode manufactured from a single NT was developed and tested.272, 273 Recent reviews and issues of journals devoted to NT are also worth mentioning.274 ± 282 References 1. S Iijima Nature (London) 354 56 (1991) 2.Carbon 33 1011 (1995) 3. J C Withers, R O Loutfy, T P Lowe Fullerene Sci. Technol. 5 1 (1997) 4. M Terrones, W K Hsu, H W Kroto, D R M Walter Top. Curr. Chem. 199 189 (1999); Chem. Abstr. 130 127 022 (1999) 5. A L Buchachenko Usp. Khim. 68 99 (1999) [Russ. Chem. Rev. 68 85 (1999)] 6. A L Ivanovskii Usp. Khim. 68 119 (1999) [Russ. Chem. Rev. 68 103 (1999)] 7. Yu E Lozovik, A M Popov Usp. Fiz. Nauk 167 751 (1997) a 8. A V Eletskii Usp. Fiz. Nauk 167 945 (1997) a 9. W KraÈ tschmer, L D Lamb, K Fostiropoulos, D R Huffman Nature (London) 347 354 (1990) 10. T W Ebbesen, P M Ajayan Nature (London) 358 220 (1992) 11. N Hamada, S Sawada, A Oshiyama Phys. Rev. Lett. 68 1579 (1992) 12. R Saito,M Fujita, G Dresselhaus, M S Dresselhaus Appl.Phys. Lett. 60 2204 (1992) 13. K Tanaka, K Okahara,M Okada, Y Yamabe Chem. Phys. Lett. 191 469 (1992) 14. T W Ebbesen, H Hiura, Y Ochiai, S Matsui, K Tanigaki Chem. Phys. Lett. 209 83 (1993) 15. T W Ebbesen Ann. Rev. Mater. Sci. 24 235 (1994) 16. T W Ebbesen Phys. Today 49 26 (1996) 17. D T Colbert, J Zhang, S M McClure, P Nikolaev, Z Chen, J H Hafner, D W Owens, G Kotula, C B Carter, J H Weaver, A G Rinzler, R E Smalley Science 266 1218 (1994) 18. X K Wang, X W Lin, V P Dravid, J B Ketterson, R P H Chang Appl. Phys. Lett. 62 1881 (1993) 19. Y Ando, S Iijima Jpn. J. Appl. Phys. 32 L107 (1993) 20. S Q Feng, D P Yu, G Hu, X F Zhang, Z Zhang J. Phys. Chem. Solids 58 1887 (1997) 21. S C Tsang, P de Oliveira, J J Davis,M L H Green, H A O Hill Chem.Phys. Lett. 249 413 (1996) 22. O Zhou, R M Fleming, D W Murphy, C H Chen, R C Haddon, A P Ramirez, S H Glarum Science 263 1744 (1994) 23. V Z Mordkovich, M Baxendale, S Yoshimura, R P H Chang Carbon 34 1301 (1996) 24. V Z Mordkovich, M Baxendale,M Yudasaka, S Yoshimura, J-Y Daic, R P H Chang Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A 310 159 (1998); Chem. Abstr. 129 114 120 (1998) 25. A Srivastava, A K Srivastava, O N Srivastava Appl. Phys. Lett. 72 1685 (1998) 26. C Journet, P Bernier Appl. Phys. A, Mater. Sci. Process. 67 1 (1998) 27. N M Buongiorno, C Brabec, A Maiti, C Roland, J Bernholc Phys. Rev. Lett. 80 313 (1998) 28. J-C Charlier, A De Vita, X Blase, R Car Science 275 646 (1997) 29. E G Gamaly, T W Ebbesen Phys.Rev. B, Condens. Matter 52 2083 (1995) 30. J Bernholc, C Brabec,M Nardelli, N Buongiorno, A Maiti, C Roland, B I Yacobson Appl. Phys. A, Mater. Sci. Process. 67 39 (1998) 31. P M Ajayan, S Iijima Nature (London) 358 23 (1992) 32. S Iijima, T Ichichashi Nature (London) 363 603 (1993) 33. D S Bethune, C H Kiang,M S de Vries, G Gorman, R Savoy, J Vazquez, R Beyers Nature (London) 363 605 (1993) 34. S Seraphin, D Zhou Appl. Phys. Lett. 64 2087 (1994) 35. S Seraphin J. Electrochem. Soc. 142 290 (1995) 36. A Loiseau, H Pascard Chem. Phys. Lett. 256 246 (1996) 49 37. C Journet,W K Maser, P Bernier, A Loiseau, M L de la Chapelle, S Lefrant, P Deniard, R Lee, J E Fischer Nature (London) 388 756 (1997) 38. J M Lambert, P M Ajayan, P Bernier, J M Planeix, V Brotons, B Coq, J Castaing Chem.Phys. Lett. 226 364 (1994) 39. C-H Kiang, M S Dresselhaus, R Beyers, D S Bethune Chem. Phys. Lett. 259 41 (1996) 40. C-H Kiang, W A Goddard III, R Beyers, J R Salem, D S Bethune J. Phys. Chem. 98 6612 (1994) 41. C-H Kiang, W A Goddard III Phys. Rev. Lett. 76 2515 (1996) 42. Chin-Hwa Kiang,W A Goddard III, R Beyers,D S Bethune Carbon 33 903 (1995) 43. A Maiti, C J Brabec, J Bernholc Phys. Rev. B, Condens. Matter 55 R6097 (1997) 44. C Journet,W K Maser, P Bernier, A Loiseau, M L de la Chapelle, P Deniard, S Lefrant, R Lee, J E Fischer, in Molecular Nanostruc- ture. (Proceedings of the International Winterschool on Electronic Properties of Novel Materials) (Singapore: World Scientific, 1998) p.386 45. P Bernier, W Maser, C Journet, A Loiseau, M L de la Chapelle, S Lefrant, R Lee, J E Fischer Carbon 36 675 (1998) 46. M Terrones, W K Hsu, S Ramos, R Castillo, H Terrones Fullerene Sci. Technol. 6 787 (1998) 47. H Lange, A Huczko, P Byszewski, E Mizera, H Shinohara Chem. Phys. Lett. 289 174 (1998) 48. Z Weng-Sieh, K Cherrey, N G Chopra, X Blase, Y Miyamoto, A Rubio,M L Cohen, S G Louie, A Zettl, R Gronsky Phys. Rev. B, 49. O Stephan, P M Ajayan, C Colliex, Ph Redlich, J M Lambert, 50. N G Chopra, R J Luyken, K Cherrey, V H Crespi,M L Cohen, 51. M Terrones,W K Hsu,H Terrones, J P Zhang, S Ramos, J P Hare, Condens. Matter 51 11 292 (1995) P Bernier, P Lefin Science 266 1683 (1994) S G Louie, A Zettl Science 269 966 (1995) R Castillo, K Prassides, A K Cheetham, H W Kroto, 55. Y Saito,M Maida J.Phys. Chem. A 103 1291 (1999) D R M Walton Chem. Phys. Lett. 259 568 (1996) 52. A Loiseau, F Willaime, N Demoncy, G Hug, H Pascard Phys. Rev. Lett. 76 4737 (1996) 53. A Loiseau, F Willaime, N Demoncy, N Schramchenko, G Hug, C Colliex, H Pascard Carbon 36 743 (1998) 54. Y Saito,M Maida, T Matsumoto Jpn. J. Appl. Phys. 38 159 (1999) 56. Ph Redlich, J Loeffler, P M Ajayan, J Bill, F Aldinger,M RuÈ hle Chem. Phys. Lett. 260 465 (1996) 57. Y Shimizu, Y Moriyoshi, S Komatsu, T Ikegami, T Ishigaki, T Sato, Y Bando Thin Solid Films 316 178 (1998) 58. K Suenaga, C Colliex, N Demoncy, A Loiseau, H Pascard, F Willaime Science 278 635 (1997) 59. P M Ajayan, S Iijima Nature (London) 361 333 (1993) 60.E Dujardin, T W Ebbesen, H Hiura, K Tanigaki Science 265 1850 (1994) 61. R Seshadri, A Govindaraj, H Aiyer, R Sen, G N Subbanna, A R Raju, C N R Rao Curr. Sci. 66 839 (1994) 62. P M Ajayan, T W Ebbesen, T Ichihashi, S Iijima, K Tanigaki, H Hiura Nature (London) 362 522 (1993) 63. P M Ajayan, O Stephan, Ph Redlich, C Colliex Nature (London) 375 564 (1995) 64. Y K Chen,M L H Green, S C Tsang Chem. Commun. 2489 (1996) 65. D Ugarte, A Chaà telain, W A de Heer Science 274 1897 (1996) 66. D Ugarte, T Stoeckli, J M Bonard, A ChaItelain,W A de Heer Appl. Phys. A, Mater. Sci. Process. 67 101 (1998) 67. R D Johnson,M S de Vries, J R Salem, D S Bethune, C S Yannoni Nature (London) 355 239 (1992) 68. R S Ruoff, D C Lorents, B Chan, R Malhotra, S Subramoney Sci- ence 259 346 (1993) 69. M Tomita, Y Saito, T Hayashi Jpn.J. Appl. Phys. 32 L280 (1993) 70. S Seraphin, D Zhou, J Jiao, J C Withers, R Loufty Nature (London) 362 503 (1993) 71. S Seraphin, D Zhou, J Jiao, J C Withers, R Loufty Appl. Phys. Lett. 63 2073 (1993) 72. J M Cowley,M Liu Micron 25 53 (1994) 73. S Subramoney, R S Ruoff, D C Lorents, B Chan, R Malhotra, M J Dyer, K Parvin Carbon 32 507 (1994) 74. M Liu, J M Cowley Carbon 33 225 (1995) 75. M Liu, J M Cowley Carbon 33 749 (1995)5076. P M Ajayan, C Colliex, J M Lambert, P Bernier, L Barbedette, M Tence , O Ste phan Phys. Rev. Lett. 72 1722 (1994) 77. X Lin, X K Wang, V P Dravid, R P H Chang, J B Ketterson Appl. Phys. Lett. 64 181 (1994) 78.A A Setlur, J M Lauerhaas, J Y Dai, R P H Chang Appl. Phys. Lett. 69 345 (1996) 79. J Y Dai, J M Lauerhaas, A A Setlur, R P H Chang Chem. Phys. Lett. 258 547 (1996) 80. A A Setlur, J Y Dai, J M Lauerhaas, R P H Chang Carbon 36 721 (1998) 81. Y Saito Carbon 33 979 (1995) 82. C Guerret-Pie court, Y Le Bouar, A Loiseau, H Pascard Nature (London) 372 761 (1994) 83. A Loiseau Fullerene Sci. Techn. 4 1263 (1996) 84. N Demoncy, O Ste phan, N Brun, C Colliex, A Loiseau, H Pascard Eur. Phys. J. B 4 147 (1998) 85. M Terrones, W K Hsu, A Schilder, H Terrones, N Grobert, J P Hare, Y Q Zhu,M Schwoerer, K Prassides, H W Kroto, D R M Walton Appl. Phys. A, Mater. Sci. Process. 66 307 (1998) 86. D Zhou, S Seraphin, J C Withers Chem.Phys. Lett. 234 233 (1995) 87. L Zhu, H Zeng, G-M Hao, R-S Sheng Gaodeng Xuexiao Huaxue Xuebao 19 350 (1998); Chem. Abstr. 128 251 992 (1998) 88. H Zeng, L Zhu, G Hao, R Sheng Carbon 36 259 (1998) 89. Y Ando Jpn. J. Appl. Phys. 32 L1342 (1993) 90. X Zhao,M Wang,M Ohkohchi, Y Ando Jpn. J. Appl. Phys. 35 Pt. 1 4451 (1996) 91. X Zhao,M Wang,M Ohkohchi, Y Ando Meijo Daigaku Sogo Kenkyusho Kiyo 1 7 (1996); Chem. Abstr. 129 155 957 (1998) 92. X Zhao,M Ohkohchi,M Wang, S Iijima, T Ichihashi, Y Ando Carbon 35 775 (1997) 93. X Zhao,M Ohkohchi, Y Ando Meijo Daigaku Sogo Kenkyusho Kiyo 2 11 (1997); Chem. Abstr. 128 289 308 (1998) 94. X Zhao, Y Ando Meijo Daigaku Sogo Kenkyusho Kiyo 3 1 (1998); Chem. Abstr. 129 155 959 (1998) 95. X Zhao, Y Ando Meijo Daigaku Rikogakubu Kenkyu Hokoku 38 151 (1998); Chem.Abstr. 128 272 319 (1998) 96. X K Wang, X W Lin, V P Dravid, J B Ketterson, R P H Chang Appl. Phys. Lett. 66 2430 (1995) 97. H Yokomichi, M Matoba, H Sakima,M Ichiashi, F Sakai Jpn. J. Appl. Phys. 37 Pt. 1 6492 (1998) 98. N S Koprinarov,M V Marinov, G V Pchelarov, M A Konstantinova Chem. Phys. Lett. 285 1 (1998) 99. J S Qin, Y Zhou, L N Wang, S C Tsang Carbon 36 465 (1998) 100. T Guo, P Nikolaev, A G Rinzler, D Tomanek, D T Colbert, R E Smalley J. Phys. Chem. 99 10 694 (1995) 101. T Guo, P Nikolaev, A Thess, D T Colbert, R E Smalley Chem. Phys. Lett. 243 49 (1995) 102. C Bower, S Suzuki,K Tanigaki,O Zhou Appl. Phys. A, Mater. Sci. Process. 67 47 (1998) 103. A G Rinzler, J Liu, H Dai, P Nikolaev, C B Huffman, F J Rodriguez-Macias, P J Boul, A H Lu, D Heymann, D T Colbert, R S Lee, J E Fischer, A M Rao, P C Eklund, R E Smalley Appl.Phys. A, Mater. Sci. Process. 67 29 (1998) 104. M Yudasaka, T Komatsu, T Ichihashi, S Iijima Chem. Phys. Lett. 278 102 (1997) 105. M Yudasaka, T Komatsu, T Ichihashi, Y Achiba, S Iijima J. Phys. Chem. B 102 4892 (1998) 106. S Bandow, S Asaka, Y Saito, A M Rao, L Grigorian, E Richter, P C Eklund Phys. Rev. Lett. 80 3779 (1998) 107. A Thess, R Lee, P Nikolaev, H Dai, P Petit, R Jerome, C Xu, Y H Lee, S G Kim, A G Rinzler, D T Colbert, G E Scruseria, D Toma nek, J E Fischer, R Smalley Science 273 483 (1996) 108. H Kataura, A Kimura, Y Ohtsuka, S Suzuki, Y Maniwa, T Hanyu, Y Achiba Jpn.J. Appl. Phys. 37 Pt. 2 L616 (1998) 109. J Lefebvre, R Antonov, A T Johnson Appl. Phys. A, Mater. Sci. Process. 67 71 (1998) 110. J Liu, H Dai, J H Hafner, D T Colbert, R E Smalley, S J Tans, C Dekker Nature (London) 385 780 (1997) 111. B W Smith, M Monthioux, D E Luzzi Nature (London) 396 323 (1998) 112. D Golberg, Y Bando, M Eremets, K Takemura, K Kurashima, T Tamiya, H Ysa Chem. Phys. Lett. 279 191 (1997) E G Rakov 113. D P Yu, X C Sun, C S Lee, I Bello, S T Lee, H D Gu, K M Leung, G W Zhou, Z F Dong, Z Zhang Appl. Phys. Lett. 72 1966 (1998) 114. H Yusa,M I Emerets, D Golberg, K Takemura, K Kurashima, Y Bando, in Proceedings of International Conference�AIRAPT-16 and HPCJ-38�on High Pressure and Technology, Kyoto, 1997 p.1057 115. T Tamiya, Y Bando, D Golberg,M Eremets, K Takemura, H Yusa Inst. Phys. Conf. Ser., 153 589 (1997); Chem. Abstr. 128 175 379 (1998) 116. O A Louchev Appl. Phys. Lett. 71 3522 (1997) 117. Y Zhang, K Suenaga, C Colliex, S Iijima Science 281 973 (1998) 118. E Munos, A M Benito, L C Estepa, J Fernandez, Y Maniette, M T Martinez, G F de la Fuente Carbon 36 525 (1998) 119. T W Ebbesen, J Tabuchi, K Tanigaki Chem. Phys. Lett. 191 336 (1992) 120. M Ge, K Sattler Science 260 515 (1993) 121. M Ge, K Sattler J. Phys. Chem. Solids. 54 1871 (1994) 122. M Ge, K Sattler Appl. Phys. Lett. 65 2284 (1994) 123. K Sattler Carbon 33 915 (1995) 124. Z Ya Kosakovskaya, L A Chernozatonskii, E F Fedorov Pis'ma Zh. Eksp. Teor. Fiz. 56 26 (1992) 125.Z Wang, F Zhu,W Wang,M Ruan Phys. Lett. A 242 261 (1998) 126. C Collazo-Davila, E Bengy, C Leslie, L D Marks Appl. Phys. Lett. 72 314 (1998) 127. F Banhart,M Zwanger, H-J Muhr Chem. Phys. Lett. 231 98 (1994) 128. D Golberg, Y Bando, K Kurashima, T Sasaki Appl. Phys. Lett. 72 2108 (1998) 129. M Jose -Yacama n,M Miki-Yoshida, L Rendo n, J G Santiesteban Appl. Phys. Lett. 62 657 (1993) 130. V Ivanov, J B Nagy, P Lambin, A Lukas, X B Zhang, X F Zhang, D Bernaerts, G Van Tendeloo, S Amelinckx, J Van Landuyt Chem. Phys. Lett. 233 329 (1994) 131. V Ivanov, A Fonseca, J B Nagy, A Lukas, P Lambin, D Bernaerts, X B Zhang Carbon 33 1727 (1995) 132. K Hernadi, A Fonseca, J B Nagy, D Bernaerts, A A Lukas Carbon 34 1249 (1996) 133. W Z Li, S S Xie, L X Qian, B H Chang, B S Zou, W Y Zhou, R A Zhao, G Wang Science 274 1701 (1996) 134.Z W Pan, S S Xie, B H Chang, C Y Wang, L Lu, W Liu, W Y Zhou, W Z Li, L X Qian Nature (London) 394 631 (1998) 135. Z W Pan, S S Xie, B H Chang, L F Sun,W Y Zhou, G Wang Chem. Phys. Lett. 299 97 (1999) 136. T E MuÈ ller, D G Reid, W K Hsu, J P Hare, H W Kroto, D R M Walton Carbon 35 951 (1997) 137. A Fonseca, K Hernadi, P Piedogrosso, J-F Colomer, K Mudhopadhyay, R Doome, S Lazarescu, L P Biro, Ph Lambin, P A Thiry, D Bernaerts, J B Nagy Appl. Phys. A, Mater. Sci. Process. 67 11 (1998) 138. W Li, S Xie, L X Qian, Y Zhang, W Lin, B Chang, W Zhou, R Zhao, C Fu, G Wang Dianzi Xianwei Xuebao 17 243 (1998) 139. G L Hornya k, J J Schneider, N Czap, K M Jones, F S Hasoon, M J Heben in The 4th International Conference on Nanostructure 145.K Mudhopadhyay, A Koshio, N Tanaka, H Shinohara Jpn. J. Materials `NANO'98' (Abstracts of Reports), Stockholm, 1998 p. 84 140. H Gaucher, R Pellenq, S Bonnamy, F Beguin, V Nalimova, Y Grillet, in Molecular Nanostructure (Proceedings of the Interna- tional Winterschool on Electronic Properties of Novel Materials) (Singapore: World Scientific, 1998) p. 395; Chem. Abstr. 129 9098 (1998) 141. K Hernadi, A Fonseca, P Piedigrosso, M Delwaux, J B Nagy, D Bernaerts, J Riga Catal. Lett. 48 229 (1997) 142. J Li,M Moskovits, T L Haslett Chem. Mater. 10 1963 (1998) 143. L P Biro, J B Nagy, Ph Lambin, S Lazarescu, A Fonseca, P A Thiry, A A Lucas, in Molecular Nanostructure (Proceedings of the International Winterschool on Electronic Properties of Novel Materials) (Singapore: World Scientific, 1998) p.419; Chem. Abstr. 129 102 458 (1998) 144. D Bernaerts, X B Zhang, X F Zhang, S Amelinckx, G Van Tendeloo, J Van Landuyt, V Ivanov, J B Nagy Philos. Mag. A 71 605 (1995) Appl. Phys. 37 L1257 (1998) 146. P Chen, H-B Zhang, G-D Lin, Q Hong, K-R Tsai Carbon 35 1495 (1997)Methods for preparation of carbon nanotubes 148. L C Qin, D Zhou, A R Krauss, D M Gruen Appl. Phys. Lett. 72 147. P Chen, H-B Zhang, G-D Lin, K-R Tsai Gaodeng Xuexiao Huaxue Xuebao 19 765 (1998); Chem. Abstr. 129 197 061 (1998) 3437 (1998) 149. J Kong, H T Soh, A M Cassell, C F Quate, H Dai Nature (London) 395 878 (1998) 150.R Baum Chem. Eng. News 76 (45) 11 (1998) 151. J M Mao, L F Sun, L X Qian, Z W Pan, B H Chang,W Y Zhou, G Wang, S S Xie Appl. Phys. Lett. 72 3297 (1998) 152. G Che, B B Lakshmi, C R Martin, E R Fischer, R S Ruoff Chem. Mater. 10 260 (1998) 153. G Che, B B Lakshmi, E R Fisher, C R Martin Nature (London) 393 346 (1998) 154. L A Chernozatonskii, E F Kukovitskii, A L Musatov, A B Ormont, K R Izraeliants, S G L'vov Carbon 36 713 (1998) 155. N A Kiselev, J Sloan, D N Zakharov, E F Kukovitskii, J L Hutchison, J Hammer, A S Kotosonov Carbon 36 1149 (1998) 156. M Endo, K Takeuchi, K Kobori, K Takahashi, H W Kroto, A Sarkar Carbon 33 873 (1995) 157. G Dia z,M Benaissa, J Santiesteban, M Jose'-Yacama'n Fullerene Sci. Technol. 6 853 (1998) 158. S Amelinckx, X B Zhang, D Bernaerts, X F Zhang, V Ivanov, J B Nagy Science 265 635 (1994) 159.T Arie, S Akita, Y Nakayama J. Surf. Anal. 4 277 (1998) 160. M Terrones, N Grobert, J Olivares, J P Zhang, H Terrones, K Kondratos,W K Hsu, J P Hare, P D Townsend, K Prassides, A K Cheetham, H W Kroto, D R M Walton Nature (London) 388 52 (1997) 161. M Terrones, N Grobert, J P Zhang, H Terrones, J Olivares, W K Hsu, J P Hare,A K Cheetham,H W Kroto,D R M Walton Chem. Phys. Lett. 285 299 (1998) 162. M Jose'-Yacama'n, H Terrones, L Rendo'n, J M Domõ nguez Carbon 33 669 (1995) 163. M Yudasaka, R Kikuchi, Y Ohki, S Yoshimura, E Ota Appl. Phys. Lett. 67 2477 (1995) 164. M Yudasaka, R Kikuchi, Y Ohki, S Yoshimura Carbon 35 195 (1997) 165. H M Chen, F Li, X Sun, S D M Brown, M A Pimenta, A Marucci,G Dresselhaus,M S Dresselhaus Chem.Phys. Lett. 289 177. S Subramoney, R S Ruoff, D C Lorents, R Malhotra Nature 602 (1998) 166. C N R Rao, R Sen, B C Satishkuma Govindaraj Chem. Commun. 1525 (1998) 167. P E Nolan, M J Schabel, D C Lynch, A H Cutler Carbon 33 79 (1995) 168. H Dai, A G Rinzler, P Nikolaev, A Thess, D T Colbert, R E Smalley Chem. Phys. Lett. 260 471 (1996) 169. A Govindaraj,R Sen,A K Santra, B V Nagaraju Mater. Res. Bull. 33 663 (1998) 170. A A Khassin, T M Yurieva, V I Zaikovskii, V N Parmon React. Kinet. Catal. Lett. 64 63 (1998) 171. R Saito, G Dresselhaus, M S Dresselhaus Physical Properties of Carbon Nanotubes (London: Imperial College, 1998) 172. M Terrones, A M Benito, C Mantega-Diego,W K Hsu, O I Osman, J P Hare, D G Reid, H Terrones, A K Cheetham, K Prassides, H W Kroto, D R M Walton Chem.Phys. Lett. 257 576 (1996) 173. R Sen, B C Satishkumar, A Govindaraj, K R Harikumar, G Raina, J-P Zhang, A K Cheetham, C N R Rao Chem. Phys. Lett. 287 671 (1998) 174. Y Saito, T Yoshikawa J. Cryst. Growth 134 154 (1993) 175. D Zhou, S Seraphin, S Wang Appl. Phys. Lett. 65 1593 (1994) 176. Y Saito,M Okuda,M Tomita, T Hayashi Chem. Phys. Lett. 236 419 (1995) (London) 366 637 (1993) 178. Y Saito,M Okuda, N Fujimoto, T Yoshikawa,M Tomita, T Hayashi Jpn. J. Appl. Phys., A 33 Pt. 2 L526 (1994) 179. Y Saito, T Matsumoto, K Nishikubo Carbon 35 1757 (1997) 180. Y L Li J. Mater. Res. 12 1678 (1997) 181.M Kusunoki, M Rokkaku, T Suzuki Appl. Phys. Lett. 71 2620 (1997) 182. H Takikawa, R Miyano,M Yatsuki, T Sakakibara Jpn. J. Appl. Phys., A 37 L187 (1998) 51 183. M Kusunoki, J Shibata, M Rokkaku, T Hirayama Jpn. J. Appl. Phys., A 37 L605 (1998) 184. V P Val'chuk, A B Ormont, L A Chernozatonskii Phys. Lett. A 200 171 (1995) 185. L A Chernozatonskii, V P Val'chuk, N A Kiselev, O I Lebedev, A B Ormont, D N Zakharov Carbon 35 749 (1997) 186. Ch Laurent, A Peigney, A Rousset J. Mater. Chem. 8 1263 (1998) 187. E V Grigoryev, V N Savenko, D V Sheglov, A V Matveev, V A Cherepanov, A S Zolkin, V A Kolesov Carbon 36 581 (1998) 188. J B Howard, J T McKinnon, Y Makarovsky, A L Lafleur, M E Johnson Nature (London) 352 139 (1991) 189. J B Howard, K Das Chowdhury, J B Vander Sande Nature (London) 370 603 (1994) 190.P K Bachmann, D Leers, H Lydtin Diam. Relat. Mater. 1 1 (1991) 191. E G Rakov Dokl. Akad. Nauk 349 350 (1996) b 192. W K Hsu, J P Hare, M Terrones, H W Kroto, D R M Walton, P J F Harris Nature (London) 377 687 (1995) 193. G Z Chen, X Fan, A Luget, M S P Shaffer, D J Fray, A H Windle J. Electroanal. Chem. 446 1 (1998) 194. H P Boehm Carbon 35 581 (1997) 195. G R Millward, D A Jefferson, in Chemistry and Physics of Carbon Vol. 14 (Eds P L Walker, P A Thrower) (New York: Marcell Dekker, 1978) 196. MG Stevens, S Subramoney, H C Foley Chem. Phys. Lett. 292 352 (1998) 197. O M KuÈ ttel, O Groening, C Emmenegger, L Schlapbach Appl. Phys. Lett. 73 2113 (1998) 198. R Dagani Chem.Eng. News 77 (2) 31 (1999) 199. L S K Pang, J D Saxby, S P Chatfield J. Phys. Chem. 97 6941 (1993) 200. S C Tsang, P J F Harris, M L H Green Nature (London) 362 520 (1993) 201. K Morishita, Y Takarada, K Uchida,M Yumura Tanso 180 245 (1997); Chem. Abstr. 128 208 004 (1998) 202. K Morishita, T Takarada Carbon 35 977 (1997) 203. T W Ebbesen, P M Ajayan, H Hiura, K Tanigaki Nature (London) 367 519 (1994) 204. Y Ando, X Zhao,M Okohchi Jpn. J. Appl. Phys. 37 L61 (1998) 205. S Bandow, S Asaka, X Zhao, Y Ando Appl. Phys. A, Mater. Sci. Process. 67 23 (1998) 206. S C Tsang, Y K Chen, P J F Harris, M L H Green Nature (London) 372 159 (1994) 207. E Dujardin, T W Ebbesen, A Krishnan, M M J Treacy Adv. Mater. (Weinheim) 10 611 (1998) 208.K B Shelimov, R O Esenaliev, A G Rinzler, C B Huffman, R E Smalley Chem. Phys. Lett. 282 429 (1998) 209. K Morishita, T Takarada Kagaku Kogaku Ronbunshu 24 670 (1998); Chem. Abstr. 129 155 965 (1998) 210. J Sloan, J Hammer,M Zwiefka-Sibley, M L H Green Chem. Commun. 347 (1998) 211. K C Hwang Chem. Commun. 173 (1995) 212. H Hiura, T W Ebbesen,K Tanigaki Adv. Mater. (Weinheim) 7 275 (1995) 213. R M Lago, S C Tsang, K L Lu, Y K Chen, M L H Green Chem. Commun. 1355 (1995) 214. R Yu, L Chen, Q Liu, J Lin, K-L Tan, S C Ng, H S O Chan, G-Q Xu, T S A Hor Chem. Mater. 10 718 (1998) 215. T W Ebbesen J. Phys. Chem. Solids 57 951 (1996) 216. Y Chen, R C Haddon, S Fang, A M Rao, P C Eklund,W H Lee, E C Dickey, E A Grulke, J C Pendergrass, A Chavan, B E Haley, R E Smalley J.Mater. Res. 13 2423 (1998) 217. R C Haddon, P C Eklund Science 282 95 (1998) 218. S S Wong, E Joselevich, A T Wooley, C L Cheung, C M Lieber Nature (London) 394 52 (1998) 219. G S Duesberg,M Burghard, J Muster, G Philipp, S Roth Chem. Commun. 435 (1998) 220. G S Duesberg, J Muster, V Krstic, M Burghard, S Roth Appl. Phys. A, Mater. Sci. Process. 67 117 (1998) 221. J Liu, A G Rinzler, H Dai, J H Hafner, R K Bradley, P J Boul, A Lu, T Iverson, K Shelimov, C B Huffman, F Rodriguez- Macias, Y-S Shon, T R Lee, D T Colbert, R E Smalley Science 280 1253 (1998) 222. K Yamamoto, S Akita, Y Nakayama J. Phys. D, Appl. Phys. 31 (8) L34 (1998)52 223. J-M Bonard, J-P Salvertat, T Stora, F Maier, T Stockli, L Forro, W A de Heer, A Chaà telain, in Molecular Nanostructure. (Proceedings of the International Winterschool on Electronic Properties of Novel Materials) (Singapore: World Scientific, 1998) p.410 224. S Bandow, A M Rao, K A Williams, A Thess, R E Smalley, P C Ecklund J. Phys. Chem. B 101 8839 (1997) 225. J-M Bonard, T Stora, J-P Salvertat, F Maier, T Stockli, C Duschl, L Forro,W A de Heer, A Chaà telain Adv. Mater. (Weinheim) 9 827 (1997) 226. R Martel, H R Shea, P Avouris Nature (London) 398 299 (1999) 227. C Niu, E K Siche, R Hoch, D Moy, H Tennent Appl. Phys. Lett. 70 1480 (1997) 228. P M Ajayan, O Stephan, C Colliex, D Trauth Science 265 1212 (1994) 229. W A de Heer, W S Basca, A Chaà telain, T Gerfin, R Humphrey- Baker, L Forro, D Ugarte Science 268 845 (1995) 230. Q Li, S Fan, C Sun,W Liang Jpn. J. Appl. Phys. 36 L501 (1997) 231. H Han, S Fan, Q Li, C L Zhang Jpn. J. Appl. Phys. 37 L1085 (1998) 232. H Dai, E W Wong, Y Z Lu, S Fan, C M Lieber Nature (London) 375 769 (1995) 233. W Han, S Fan,W Liang, B Gu, D Yu Chem. Phys. Lett. 265 374 (1997) 234. W Han, S Fan, Q Li, B Gu, D Yu Wuji Cailiao Xuebao 12 774 (1997); Chem. Abstr. 129 18 741 (1998) 235. G W Meng, L D Zhang, in The 4th International Conference on Nanostructured Materials `NANO'98' (Abstracts of Reports), Stockholm, 1998 p. 88 236. W Han, S Fan, Q Li, Y Hu Science 277 1287 (1997) 237. D Laplaze, P Bernier, W K Mazer, G Flamant, T Guillard, A Loiseau Carbon 36 685 (1998) 238. E G Rakov Zh. Neorg. Khim. 44 1827 (1999) c 239. A A Setlur, J M Lauerhaas, J Y Dai, R P H Chang Springer Ser. Mater. Sci., 33 43 (1998); Chem. Abstr. 130 132 600 (1999) 240. Y Saito, Y Tani, N Miyagawa, K Mitsushima, A Kasuya, Y Nishina Chem. Phys. Lett. 294 593 (1998) 241. T Sugai, H Omote, S Bandow, N Tanaka, H Shinohara Jpn. J. Appl. Phys. 38 (4B) L477 (1999) 242. B H Chang, S S Xie, W Y Zhou, L X Qian, Z W Pan, J M Mao, W Z Li J. Mater. Sci. Lett. 17 1015 (1998) 243. M Yudasaka, T Ichihashi, S Iijima J. Phys. Chem. B 102 10 201 (1998) 244. M Yudasaka, T Ichihashi, T Komatsu, S Iijima Chem. Phys. Lett. 299 91 (1999) 245. E G Gamaly, A V Rode Proc. SPIE�Int. Soc. Opt. Eng. 3343 Pt. 2 847 (1998); Chem. Abstr. 130 131 348 (1999) 246. Y Zhang, H Gu, S Iijima Appl. Phys. Lett. 73 3827 (1998) 247. W K Maser, E Munoz, A M Benito,M T Martinez, G F de la Fuente, Y Maniette, E Anglaret, J-L Sauvajol Chem. Phys. Lett. 292 587 (1998) 248. L P Biro, G I Mark, J Gyulai, K Havancszak, S Lipp, Ch Lehrer, L Frey, H Ryssel Nucl. Instrum. Methods Phys. Res., Sect. B 147 142 (1999); Chem. Abstr. 130 162 158 (1999) 249. A Siska, K Hernadi, I Kiricsi, I Rojik, J B Nagy AIP Conf. Proc. 442 20 (1998); Chem. Abstr. 130 57 668 (1999) 250. J Kong, A M Kassel, H Dai Chem. Phys. Lett. 292 576 (1998) 251. C N R Rao, A Govindaraj, R Sen, B C Satishkumar Mater. Res. Innovations 2 128 (1998); Chem. Abstr. 130 133 022 (1999) 252. B C Satishkumar, A Govindarai, R Sen, C N R Rao Chem. Phys. Lett. 293 47 (1998) 253. Z P Huang, J W Xu, Z F Ren, J H Wang,M P Siegal, P N Provencio Appl. Phys. Lett. 73 3845 (1998) 254. Z F Ren, Z P Huang, J W Xu, J H Wang, P Bush,M P Siegal, P N Provencio Science 282 1105 (1998) 255. F Okama, T Hayashi, Y Fujimoto J. Appl. Phys. 84 1626 (1998) 256. Z K Tang, H D Sun, J Wang, J Chen, G Li Appl. Phys. Lett. 73 257. S L Sung, S H Tsai, C H Tseng, F K Chiang, X W Liu, H C Shin 258. W Han, Y Bando, K Kurashima, T Sato Appl. Phys. Lett. 73 3085 259. J Sloan, J Cook, A Chu,M Zwiefka-Sibley, M L H Green, 2287 (1998) Appl. Phys. Lett. 74 197 (1999) (1998) J L Hutchison J. Solid-State Chem. 140 83 (1998) E G Rakov 260. J-F Colomer, P Piedigrosso, I Willems, C Journet, P Bernier, G Van Tendeloo, A Fonseca, J B Nagy J. Chem. Soc., Faraday Trans. 98 3753 (1998) 261. M S P Shaffer, X Fan, A H Windle Carbon 36 1603 (1998) 262. R Alexandrescu, X Armand,M Cauchetier, N Herlin, S Petcu, I Voicu Carbon 36 1285 (1998) 263. V Krstic, G S Duesberg, J Muster,M Burghard, S Roth Chem. Mater. 10 2338 (1998) 264. B Z Tang, H Xu Macromolecules 32 2569 (1999) 265. R H Baughman, C Cui, A A Zakhidov, Z Iqbal, J N Barisci, G M Spinks, G G Wallace, A Mazzolini, D Rossi, A G Rinzler, O Jaschinski, S Roth,M Kertesz, in The 13 International Winterschool on Electronic Properties of Novel Materials. Science and Technology of Molecular Nanostructures, Kirchberg, 1999 p. 17 266. R H Baughman, C Cui, A A Zakhidov, Z Iqbal, J N Barisci, G M Spinks, G G Wallace, A Mazzolini, D De Rossi, A G Rinzler, O Jaschinski, S Roth,M Kertesz Science 284 1340 (1999) 267. O Inganas, I Lundstrum Science 284 1281 (1999) 268. M Brennan News Week 77 (21) 6 (1999) 269. A A Farajian, K Esfarjani, Y Kawazoe Phys. Rev. Lett. 82 5084 (1999) 270. Q Wang, S R Chandra,D S Sholl, J K Johnson Phys. Rev. Lett. 82 956 (1999) 271. H Dai, N Franklin, J Han Appl. Phys. Lett. 73 1508 (1998) 272. J K Campbell, L Sun, R M Crooks J. Am. Chem. Soc. 121 3779 (1999) 273. Chem. Eng. News 77 (17) 23 (1999) 274. H Hoffschulz, J W Otto, H Eickenbusch Zukuenftige Technol. 25 (i-vi) 1 (1998); Chem. Abstr. 129 103 306 (1998) 275. C Laurent, E Flahaut,A Peigney, A Rousset New J. Chem. 22 1229 (1998) 276. Ceramics Jpn. 33 (6) (1998) 277. Appl. Phys. A (1 5) 67 (1998) 278. J Hu, T W Odom, C M Lieber Acc. Chem. Res. 32 435 (1999) 279. P M Ajayan Chem. Rev. 1787 (1999) 280. R Nesper, H-J Muhr Chimia 52 571 (1998) 281. S Subramoney Adv. Mater. (Weinheim) 10 1157 (1998) 282. R E Smalley, B I Yakobson Solid State Commun. 107 597 (1998) a�Physics-Uspekhi (Engl. Transl.) b�Dokl. Chem. Technol., Dokl. Chem. (Engl. Transl.) c�Russ. J. Inorg. Chem. (Engl. Tr