A carbon nanotube is a hollow, cylindrical structure made primarily of carbon atoms bonded in a hexagonal network. Its walls have the local arrangement of graphene, an atomically thin sheet of carbon, although nanotubes are normally grown directly rather than manufactured by rolling up sheets. A single-walled carbon nanotube consists of one such cylinder; a multi-walled nanotube contains two or more nested cylinders. Diameters range from less than a nanometre for some single-walled tubes to tens or hundreds of nanometres for larger multi-walled structures, while lengths can extend over many micrometres and, under specialized growth conditions, much farther.
Carbon nanotubes are important because their geometry combines strong carbon–carbon bonding with an unusually large length-to-diameter ratio and a nearly one-dimensional electronic structure. Depending on their atomic arrangement, individual single-walled tubes can behave as metals or semiconductors. They can also be exceptionally stiff, carry large electrical currents, and transport heat efficiently along their axes. These properties have made them subjects of fundamental research and useful ingredients in conductive plastics, battery electrodes, coatings, and experimental electronic devices.
The properties of a nanotube-containing product are not simply those of an ideal, isolated nanotube. Defects, tube length, impurities, aggregation, orientation, and contact resistance often determine practical performance. Consequently, the field encompasses not only atomic structure and quantum physics but also large-scale synthesis, separation, materials processing, measurement, and assessment of environmental and occupational exposure.
Atomic structure and classification
The simplest structural model of a single-walled carbon nanotube begins with graphene’s honeycomb lattice. Each carbon atom bonds to three neighbours through predominantly bonding, leaving electronic states derived from orbitals perpendicular to the local surface. Joining equivalent lattice sites around a circumference defines a cylinder. This construction describes atomic connectivity; it does not imply that a sheet is physically rolled during growth.
The circumference is specified by a chiral vector,
where and are graphene lattice vectors and and are integers. The pair , called the chiral indices, determines the tube’s diameter and the orientation of its hexagons relative to its axis. With graphene’s lattice constant , the ideal diameter is
The corresponding chiral angle is conventionally chosen between and . Tubes with are called zigzag, while those with are armchair. Intermediate arrangements are chiral and can occur in left- and right-handed forms. This chirality concerns the handedness of the atomic lattice, not necessarily a visible helical shape of the tube itself.
Single-walled nanotubes commonly have diameters of roughly 0.7–3 nm, although smaller and larger examples are known. Double-walled nanotubes contain two concentric shells and can combine an accessible outer wall with a relatively protected inner wall. Multi-walled nanotubes contain several to many shells, usually separated by distances close to the approximately 0.34 nm spacing between layers in graphite. Adjacent shells need not have matching rotational symmetry or a shared axial repeat length.
Tube ends may be open or closed. A closed cap generally contains pentagonal carbon rings that provide the curvature needed to close a hexagonal network, connecting nanotube geometry with that of a fullerene. Vacancies, non-hexagonal rings, substitutional atoms, and attached chemical groups create departures from the ideal structure. Five- and seven-membered ring combinations can introduce bends or join regions with different chiral indices.
The distinction between a nanotube and a carbon nanofibre is structural rather than merely dimensional. In many nanofibres, graphitic layers form stacked cones, cups, or platelets instead of continuous cylindrical walls parallel to the fibre axis. Commercial materials may contain mixtures of these morphologies. Identifying the actual wall arrangement is therefore important when comparing products, interpreting measurements, or assessing biological effects.
Discovery and development of the field
Carbon nanotubes emerged from a longer history of research on carbon filaments, soot, and graphitic structures. In 1952, L. V. Radushkevich and V. M. Lukyanovich published electron micrographs of hollow carbon filaments with nanometre-scale diameters. In 1976, Morinobu Endo, A. Oberlin, and T. Koyama described hollow graphitic fibres produced by catalytic decomposition of hydrocarbons. These studies established important structural and growth observations before nanotubes became a distinct international research field.
The decisive expansion followed Sumio Iijima’s 1991 report of nested graphitic tubes in material produced by an electric arc. High-resolution imaging with an electron microscope revealed concentric cylindrical walls and clarified their atomic-scale organization. The work appeared amid intense interest in fullerenes and encouraged researchers to treat tubular carbon structures as a new class of low-dimensional material. It is therefore widely associated with the beginning of modern nanotube research, although it was not the first observation of hollow nanoscale carbon.
In 1993, single-walled nanotubes were reported independently by Iijima and Toshinari Ichihashi, and by Donald Bethune and colleagues. Their production involved metal-containing arc processes, demonstrating the importance of catalytic metals in controlling carbon assembly. Single-walled tubes were particularly significant because their electronic behaviour could be connected directly to a well-defined atomic lattice rather than to several interacting walls.
Theory advanced alongside synthesis. In 1992, work by groups including Noriaki Hamada, Riichiro Saito, and James Mintmire showed that a tube’s electronic character should depend on its diameter and chiral indices. Measurements reported in 1998 by teams led by Jeroen Wildöer and Teri Wang Odom connected atomic structure with electronic spectra in individual nanotubes, providing experimental support for these predictions. This relationship between geometry and electronic properties became one of the field’s central results.
Production methods then broadened. Laser-vaporization experiments in the 1990s yielded substantial quantities of single-walled material, while catalytic growth from gaseous carbon sources enabled patterned growth and larger-scale production. In 2004, Kenji Hata and colleagues demonstrated water-assisted growth of vertically aligned single-walled nanotube forests. Small amounts of water helped prolong catalyst activity under suitable conditions, allowing tall, dense arrays to form.
Later development increasingly concerned control and integration rather than discovery alone. Researchers developed separation methods for metallic and semiconducting tubes, improved alignment and placement, and fabricated integrated circuits. A carbon-nanotube computer demonstrated in 2013 contained 178 nanotube transistors; a 2019 demonstration used more than 14,000. These milestones showed that device integration was possible, while also exposing the continuing importance of purity, reproducibility, and compatibility with established manufacturing processes.
Electronic and optical behaviour
A nanotube’s electronic properties arise from the combination of graphene’s electronic states and the boundary condition imposed around a cylinder. Only certain electron wavelengths fit around the circumference, dividing the available states into one-dimensional subbands. In a simple model, a tube is metallic when is divisible by three and semiconducting otherwise. This rule describes the principal structure–property relationship, but curvature and other effects introduce important qualifications.
Ideal armchair tubes are metallic in the basic description. Other tubes classified as metallic by the simple rule can develop small gaps because curvature changes orbital overlap. Semiconducting tubes have gaps that scale approximately inversely with diameter; tubes near 1 nm in diameter typically have gaps on the order of several tenths of an electronvolt to about 1 eV. Strain, dielectric surroundings, adsorbed molecules, and interactions between tubes can further alter the measured electronic band structure.
Transport through a short, clean tube can approach ballistic transport, in which electrons traverse the channel with little scattering. For an ideal metallic single-walled tube, the commonly quoted conductance including spin and valley degeneracy is
corresponding to a resistance of approximately . This is a quantum-limited two-terminal value for ideal transmission, not a universal resistance for all nanotubes. Real contacts, disorder, tube length, and the measurement configuration can substantially increase resistance.
Nanotubes also support current densities far above those tolerated by many conventional metal interconnects under comparable nanoscale conditions. Their limits depend on heat dissipation, contact quality, defects, and whether oxygen is present. At high bias, interactions between electrons and lattice vibrations increase scattering, while local heating can cause oxidation or structural failure. A conducting nanotube network introduces another limitation: current must repeatedly cross tube–tube junctions.
Optical absorption and emission reflect the one-dimensional subbands, which produce strong features in the electronic density of states. Coulomb attraction between an excited electron and the hole it leaves behind forms an exciton, so optical transition energies cannot generally be interpreted as independent-electron band gaps. Semiconducting single-walled tubes exhibit characteristic transitions conventionally labelled , , and higher orders. Many emit near-infrared light after optical excitation, with wavelengths dependent on chirality and environment.
Optical behaviour also depends strongly on sample preparation. Bundling commonly quenches semiconducting-tube luminescence by enabling energy transfer to metallic tubes or non-radiative sites. Certain deliberately introduced covalent defects can instead create localized emissive states at longer wavelengths. These effects support research on optical probes and light-emitting devices, while making spectroscopy a sensitive indicator of dispersion, chemical modification, and the local molecular environment.
Mechanical and thermal properties
The axial stiffness of a nanotube derives chiefly from its strong carbon–carbon bonds. Measurements and calculations commonly place the axial Young’s modulus of high-quality nanotubes near 1 TPa, comparable in scale to the in-plane stiffness associated with graphite. Reported tensile strengths for individual high-quality tubes reach tens of gigapascals, with some measurements exceeding 100 GPa. These figures describe particular specimens and test conditions, not guaranteed properties of commercial powders or fibres.
Assigning stress to a single-walled tube requires a convention for its effective cross-sectional area. A wall thickness near graphite’s interlayer spacing is often used, but alternative definitions produce different numerical moduli and strengths from the same measured force. For this reason, axial stiffness expressed as force per strain, together with dimensions and the assumed wall thickness, can be more informative than a modulus quoted alone.
Nanotubes respond very differently to axial tension, compression, bending, and radial loading. A tube that strongly resists stretching may flatten under transverse pressure or buckle under compression. Small-radius tubes can bend elastically through large angles, while larger tubes may develop local kinks or collapse into flattened shapes. Defects often determine the point at which reversible deformation gives way to bond breaking.
Multi-walled structures introduce intershell mechanics. Forces between neighbouring shells are much weaker than the covalent bonds within a shell, and relative sliding can occur. If only an outer wall is effectively gripped during a tensile test, inner walls may contribute less than their total area suggests. Similar limitations affect bundles and yarns: individual tubes can slip past one another unless friction, entanglement, chemical links, or surrounding material transfers load effectively.
Heat transport is dominated largely by lattice vibrations, or phonons. Along a clean individual nanotube, room-temperature thermal conductivity can reach several thousand watts per metre-kelvin in some measurements, although values depend on length, defects, isotope composition, and contact corrections. Short specimens can show partly ballistic heat transport, making an apparent conductivity dependent on specimen length rather than solely on an intrinsic bulk property.
These exceptional axial properties do not automatically produce exceptional bulk materials. A randomly oriented mat conducts heat through many resistive junctions; a polymer composite must transfer stress across a nanotube–polymer interface. Alignment, densification, and interface design are therefore central to engineering performance. The same weak intertube interactions that complicate load transfer can also allow bending, damping, and energy absorption, giving nanotube assemblies behaviour distinct from that of their individual constituents.
Synthesis and growth mechanisms
The principal production routes are arc discharge, laser ablation, and chemical vapour deposition (CVD). They differ in carbon source, temperature, catalyst arrangement, throughput, and product morphology. None automatically yields a single, fully specified nanotube population: as-grown material generally has distributions of diameter, length, wall number, and defect content, together with some unwanted carbon or catalyst residue.
Arc discharge uses a high-current electrical discharge between carbon electrodes, typically in an inert atmosphere. Carbon evaporates from the hot electrode region and condenses into various structures. Suitable conditions produce multi-walled nanotubes, while incorporating catalytic metals can favour single-walled tubes. The method can yield highly crystalline material but also produces particles and other carbon forms that require separation.
Laser ablation vaporizes a carbon target, commonly containing catalytic metals, in a heated environment. The carbon-containing plume cools and forms nanotubes and other products. This method played an important role in producing high-quality single-walled samples for early research. Its equipment and energy requirements differ from those of continuous industrial processes, and controlling the collected product remains important.
CVD decomposes carbon-containing gases over catalyst particles, commonly containing iron, cobalt, nickel, or combinations of metals. Feedstocks include methane, ethylene, acetylene, and carbon monoxide, with growth temperatures frequently lying between roughly 600 and 1,200 °C depending on the system. Catalyst particles can be supported on a substrate, embedded in a powder, or introduced into a flowing reactor. Floating-catalyst processes can continuously generate nanotube-containing aerogels that are collected as sheets or fibres.
Growth is commonly described through carbon adsorption, decomposition, transport, and incorporation at the catalyst–nanotube interface. Carbon may move across a particle’s surface or through its interior, with the relative importance depending on catalyst composition and reaction conditions. In base growth, the catalyst remains attached to the support; in tip growth, it is lifted as the tube elongates. These categories describe observable geometries rather than a single universally applicable atomic mechanism.
Catalyst size strongly influences tube diameter, but diameter control is not equivalent to chirality control. Catalyst structure, carbon chemical potential, nucleation kinetics, support interactions, and growth termination all affect the result. Excess carbon can encapsulate and deactivate particles; controlled amounts of oxidizing species can sometimes remove disordered carbon and extend catalyst lifetime.
Industrial production must additionally manage heat and mass transfer, catalyst recovery, consistency, and reactor fouling. Yield may be reported per unit catalyst, per reactor volume, or per operating time, and these measures are not interchangeable. High carbon conversion or large powder output does not by itself demonstrate suitability for electronic devices, where even a small unwanted population of metallic tubes or particles can be consequential.
Purification, separation, and assembly
Freshly synthesized nanotubes commonly coexist with amorphous carbon, graphitic particles, catalyst metals, and support material. Purification may combine oxidation, acid treatment, filtration, centrifugation, and thermal processing. Each step changes the material as well as removing impurities. Oxidation can preferentially consume disordered carbon but can also attack nanotube ends and defects; strong chemical treatment may shorten tubes or introduce oxygen-containing groups.
Dispersion presents a separate problem. Nanotubes attract one another through van der Waals forces and readily form bundles or entangled networks. Surfactants, selected solvents, and wrapping polymers can stabilize individual tubes in liquid. Ultrasonication helps separate aggregates, but prolonged or intense treatment can break tubes. The resulting length distribution may be very different from that of the original growth product.
Functionalization can be covalent or non-covalent. Covalent attachment modifies the carbon lattice and can improve compatibility with a surrounding polymer, provide chemical recognition sites, or strengthen interfacial bonding. It may also disrupt electronic conduction and introduce mechanical weak points. Non-covalent adsorption preserves the lattice more directly but can leave insulating residues, alter surface chemistry, or become unstable under changed processing conditions.
For single-walled tubes, separation by electronic type is especially important. Density-gradient ultracentrifugation, gel chromatography, aqueous two-phase extraction, and selective polymer wrapping can enrich metallic, semiconducting, or particular chiral species. Their selectivity arises from differences in how tube structures interact with surfactants, polymers, stationary phases, or liquid environments. High purity can be achieved, but recovery yield, cost, length retention, and removal of processing chemicals remain significant considerations.
Assembly transforms separated or as-grown tubes into usable forms. Filtration can produce thin mats often called buckypaper. Direct spinning from a reactor or drawing from aligned forests can produce fibres and sheets. Solution-based methods include coating, printing, dielectrophoretic placement, and alignment by flow or electric fields. The orientation distribution, density, porosity, and number of junctions determine properties as strongly as the characteristics of individual tubes.
Conductivity in a composite often follows a percolation transition: an interconnected network appears above a critical filler concentration. Long, well-dispersed tubes can establish connectivity at comparatively low loading, but insulating surface layers and poor junctions can raise the required concentration. Mechanical reinforcement follows a different set of constraints, including interface strength and effective tube length. A treatment that improves dispersion may therefore help one property while degrading another, making processing a balance among competing requirements rather than a single purification or mixing step.
Characterization and measurement
Reliable characterization requires several complementary techniques because no single measurement establishes structure, purity, dimensions, electronic type, and performance simultaneously. A dark powder labelled “carbon nanotubes” may contain very different proportions of nanotubes, residual metal, non-tubular carbon, and surface chemicals. Even two samples with similar average diameters can behave differently because their length distributions, wall counts, or defect populations differ.
Transmission electron microscopy resolves tube walls, cap structures, catalyst particles, and local defects. High-resolution images can distinguish single-, double-, and multi-walled structures, while electron diffraction can provide crystallographic information. Scanning electron microscopy is especially useful for forests, networks, fibres, and larger-scale morphology. Both methods sample limited regions, and specimen preparation or electron-beam exposure can modify delicate structures.
Atomic force microscopy measures deposited tubes and networks through their surface topography. It can estimate individual-tube heights and lengths, although substrate interactions, adsorbates, tip geometry, and tube deformation affect apparent dimensions. Scanning tunnelling microscopy can connect atomic arrangement with local electronic states, but requires suitable conductive configurations and careful interpretation of the electronic contribution to image contrast.
Raman spectroscopy is widely used because nanotubes have distinctive vibrational signatures. The radial breathing mode, in which atoms move approximately inward and outward together, is especially useful for many single-walled tubes. Its frequency varies approximately inversely with diameter:
where the empirical constants depend on environment and calibration. The graphitic G band reflects carbon bond vibrations, while the defect-activated D band responds to disorder, edges, and other symmetry-breaking features. A D-to-G intensity ratio is not a universal defect concentration: excitation wavelength, resonance, tube population, and measurement conditions all matter.
Optical absorption and photoluminescence provide complementary information about single-walled species. Photoluminescence excitation maps can distinguish many semiconducting chiralities, but metallic tubes are not represented in the same way, and bundled or defective semiconducting tubes may emit weakly. Resonance effects likewise mean that a Raman spectrum collected with one laser wavelength does not sample every tube equally.
Thermogravimetric analysis tracks mass changes during heating and can help estimate combustible carbon and inorganic residue. Elemental analysis and plasma-based methods identify residual metals, while surface-sensitive spectroscopy probes chemical groups. Electrical, tensile, and thermal tests then characterize functional behaviour. Meaningful comparisons require reporting preparation history, humidity or atmosphere where relevant, geometry, contact arrangements, and statistical variation. Otherwise, differences attributed to intrinsic nanotube quality may instead arise from sampling or test configuration.
Technological uses and device integration
One of the most established roles of nanotubes is as a conductive additive. In plastics and coatings, a sparse nanotube network can reduce electrical resistance while preserving much of the host material’s processability. Applications include antistatic components, electrically dissipative surfaces, and electromagnetic shielding. Performance depends on network continuity and contact resistance, so a high intrinsic conductivity of individual tubes does not guarantee an effective formulation.
In lithium-ion batteries, nanotubes can connect active particles and help maintain electrical pathways as electrodes expand, contract, or crack. Their high aspect ratio can provide connectivity with relatively small additive quantities. They are generally auxiliary conductive components rather than the principal lithium-storage material in such uses. Research also examines nanotube-supported catalysts and porous electrodes for fuel cells and electrochemical energy storage, where accessible surface, electrolyte wetting, and chemical stability are important.
Structural uses exploit nanotubes as fillers, interfacial modifiers, or assembled fibres. Potential benefits include stiffness, toughness, fatigue resistance, electrical monitoring, and multifunctionality, but results depend strongly on dispersion and load transfer. Very strong individual tubes have not translated automatically into bulk fibres with the same specific strength. Junctions, tube ends, misalignment, and sliding remain central constraints. Nanotube sheets and coatings also serve in lightweight resistive heaters and flexible conductive structures.
Electronics makes particularly demanding use of semiconducting single-walled tubes. In a field-effect transistor, a gate changes the carrier population in a nanotube channel and thereby controls current. Small channel dimensions, strong electrostatic control, and favourable transport properties motivate research on logic circuits and flexible electronics. Integration requires consistent placement, suitable contacts, stable threshold voltages, and extremely effective exclusion or disabling of metallic tubes that could prevent switching off. Demonstrations of circuits establish feasibility but do not remove these manufacturing requirements.
Nanotubes also function as sensing elements. Adsorbed molecules can change carrier concentration, scattering, contact barriers, or the dielectric environment, producing measurable changes in electrical or optical signals. Such sensitivity supports sensor research for gases, chemical species, strain, and biological analytes. It also creates cross-sensitivity: humidity, contaminants, and temperature can resemble the desired signal. Selective coatings, reference devices, and calibration are therefore part of the device architecture, not merely optional additions.
Other applications exploit geometry or surface interactions. Sharp nanotube tips can emit electrons under strong electric fields, supporting development of field-emission sources. Aligned nanotubes and nanotube-containing membranes are investigated for selective transport and water treatment, although membrane defects, support resistance, fouling, and manufacturing cost influence practical performance. Near-infrared fluorescence enables experimental imaging and molecular sensing, while nanotube-based delivery systems remain a research area whose biological behaviour depends on formulation. Across these uses, commercial deployment is strongest where mixed tube populations are acceptable; applications requiring atomically uniform tubes face substantially stricter production and quality-control demands.
Exposure, environmental behaviour, and unresolved limitations
Carbon nanotubes are not a single substance with one exposure profile. Length, diameter, rigidity, wall number, surface chemistry, aggregation, residual metals, and persistence all influence how a material interacts with biological systems. A tightly bound nanotube additive in a solid polymer differs from a freely dispersible powder, and both differ from a suspension prepared for laboratory testing. Exposure assessment must distinguish the manufactured product from particles released during handling, machining, weathering, or disposal.
Inhalation has received particular attention because some nanotube forms are respirable and persistent. Animal studies have shown that certain materials can cause pulmonary inflammation, granulomatous responses, and fibrosis. Long, rigid, durable structures may be difficult for immune cells to engulf completely, a response often described as frustrated phagocytosis. These findings do not establish that every nanotube formulation produces the same effects; they identify material characteristics and exposure conditions important to hazard evaluation.
The International Agency for Research on Cancer classified the specific multi-walled material MWCNT-7 as possibly carcinogenic to humans, Group 2B. Other multi-walled nanotubes and single-walled nanotubes were placed in Group 3, meaning not classifiable as to carcinogenicity to humans on the available evidence, rather than demonstrated harmless. The distinctions reflect the evidence for particular materials and should not be generalized into a uniform classification for all nanotubes.
In 2013, the United States National Institute for Occupational Safety and Health recommended an exposure limit of of respirable elemental carbon as an eight-hour time-weighted average for carbon nanotubes and carbon nanofibres. This was an occupational recommendation intended to reduce risk, not a finding that all exposure below that concentration is without effect. Measurement is complicated by background carbon particles and by the need to identify whether sampled carbon actually consists of nanotubes.
Environmental behaviour also varies with formulation. Nanotubes can aggregate, associate with natural organic matter, bind other chemicals, or accumulate in sediments. Oxidation and biological processes can alter some nanotube materials under particular conditions, but laboratory degradation does not establish rapid disappearance in every environment. Residual catalysts and attached chemicals can contribute to measured effects, making well-characterized controls essential in nanotoxicology and ecological studies.
The technical limitations of the field are closely connected to these characterization issues. Manufacturers and researchers still face trade-offs among output, purity, tube length, structural uniformity, processing intensity, and cost. Life-cycle comparisons must account for reactor energy, feedstocks, purification reagents, product lifetime, and end-of-life treatment rather than infer environmental benefit from low material weight alone. Standardized terminology, reference materials, reproducible measurements, and material-specific exposure data are necessary for distinguishing the behaviour of an ideal nanotube from that of the diverse products sold under the same name.