How Are Carbon Nanotubes Made? Methods and Processes

Carbon nanotubes are made by forcing carbon atoms to assemble into hollow cylindrical structures, typically by vaporizing or decomposing a carbon source at high temperatures in the presence of a metal catalyst. The three established methods are arc discharge, laser ablation, and chemical vapor deposition (CVD), with CVD dominating industrial production because it runs at lower temperatures and scales more easily. Each method produces tubes with different characteristics, and the choice of technique, catalyst, and feedstock together determine whether you get single-walled tubes, multi-walled tubes, or a messy mix of both.

Arc Discharge

Arc discharge was the method that put carbon nanotubes on the map in the early 1990s, and it remains the conceptually simplest approach. Two high-purity graphite rods serve as electrodes inside a sealed chamber filled with an inert gas at low pressure. When a current of roughly 50 to 150 amps flows between them at 25 to 40 volts, the gap between the electrodes reaches temperatures high enough to sublimate carbon directly from the anode. A plasma forms in that gap, and carbon atoms reassemble on and around the cathode as nanotubes, soot, and other carbon structures.

The anode can be pure graphite, which tends to produce multi-walled carbon nanotubes, or it can be packed with small amounts of metal catalyst (typically iron, nickel, cobalt, or mixtures of these), which steers the process toward single-walled tubes. The chamber pressure, gas composition, and electrode gap all affect the yield and quality. Operators maintain a constant gap by advancing the anode as it erodes, and a stable plasma can be held for extended reaction times by monitoring the voltage.

Arc-discharge tubes are prized for their high crystallinity because the extreme temperatures anneal out many structural defects. The downside is that the process is hard to scale and produces a lot of byproducts mixed in with the nanotubes, including amorphous carbon and fullerenes, so extensive purification is needed afterward.¹

Laser Ablation

Laser ablation works on a similar principle to arc discharge but uses a focused laser pulse instead of an electric arc to vaporize the carbon target. The original version, developed at Rice University, used a “double-pulse laser oven” setup: a graphite target laced with metal catalyst sits inside a tube furnace heated to around 1,200 °C, and laser pulses blast carbon and catalyst atoms off the target surface. A flow of inert gas, usually helium or argon, sweeps the vaporized material downstream, where it condenses and nucleates into nanotubes on a cooled collector.²

Researchers have experimented with different catalyst metals and concentrations, buffer gas pressures, oven temperatures, and even the porosity of the graphite target to fine-tune the product. Laser ablation tends to produce high-quality single-walled nanotubes with a relatively narrow diameter distribution, which makes them attractive for research applications. But like arc discharge, the batch sizes are small and the equipment is expensive, so laser ablation has mostly stayed in the lab rather than moving to the factory floor.

Chemical Vapor Deposition

Chemical vapor deposition is the workhorse of carbon nanotube manufacturing. Instead of blasting carbon off a solid target, CVD starts with a carbon-containing gas, such as methane, ethylene, acetylene, or carbon monoxide, and decomposes it over catalyst nanoparticles at temperatures typically between 600 and 1,100 °C. The carbon atoms dissolve into or adsorb onto the catalyst particle, and a nanotube nucleates and grows from that particle. The process can run continuously and can be tuned to produce either single-walled or multi-walled tubes depending on the catalyst, temperature, and gas mixture.

Several CVD variants exist. Thermal CVD is the simplest, relying on furnace heat alone to crack the gas. Plasma-enhanced CVD (PECVD) adds an electric field to generate a reactive plasma, which allows the process to run at lower substrate temperatures, a useful feature when growing nanotubes on temperature-sensitive surfaces like glass or flexible electronics.³ Floating-catalyst CVD injects the catalyst precursor (often ferrocene, an iron-containing compound) directly into the gas stream so that catalyst particles form in-flight and nanotubes grow in the gas phase rather than on a substrate. This approach enables continuous production and was instrumental in developing methods to spin nanotube fibers and ribbons straight out of the reactor.⁴

The scalability of CVD is its defining advantage. Fluidized-bed reactors, where catalyst particles are suspended in an upward-flowing gas stream, can produce kilograms or even tonnes of multi-walled nanotubes per run. Most of the carbon nanotubes sold commercially today come from some version of CVD.

What Happens at the Catalyst Particle

Regardless of whether you use arc discharge, laser ablation, or CVD, the catalyst nanoparticle is where the action happens. The dominant explanation for how nanotubes grow involves carbon atoms dissolving into a molten or near-molten metal particle, reaching a point where the particle is saturated, and then precipitating out as a graphitic cylinder. This is sometimes called the vapor-liquid-solid (VLS) model. Because the catalyst particle is so tiny, both the temperature and the carbon concentration across it stay essentially uniform, and the degree of carbon saturation controls when a nanotube cap nucleates and how fast the tube grows afterward.⁵

Not all catalysts melt during growth, though. Recent work using high-melting-point metals like rhenium has shown that nanotubes can also grow from a solid catalyst particle through a vapor-solid-solid (VSS) mechanism. In that case, carbon atoms travel across the particle’s surface rather than diffusing through a liquid interior. Surface diffusion on a solid particle can be up to roughly 50 times slower than bulk diffusion through liquid iron, which creates a narrow window of conditions where growth succeeds: too little carbon transport and the tube stalls, too much heat and carbon simply wraps around the particle as a graphitic shell instead of extending as a tube.⁶

Whether a nanotube grows from the top of a catalyst particle (tip growth) or from its base (base growth) depends partly on how strongly the particle adheres to the substrate. A weakly anchored particle lifts off and rides at the growing tip; a strongly anchored one stays put while the tube extends upward. The growth mode affects how easy it is to produce aligned nanotube arrays and how the tube eventually terminates.

Why the Catalyst Size Matters

The diameter of the catalyst nanoparticle is the single biggest lever for controlling the diameter of the nanotube that grows from it, but the relationship is not a simple one-to-one match. In-situ observations during CVD growth have confirmed that below a critical particle size of about 2 nanometers, the resulting nanotube is actually wider than the particle. Above that threshold, the tube is narrower. This crossover means that shrinking the catalyst does not always shrink the tube in a proportional way, which complicates efforts to produce batches of nanotubes with a single, precise diameter.⁷

Catalyst composition matters as much as size. Iron, cobalt, and nickel are the traditional choices because they dissolve carbon readily at moderate temperatures and form carbide phases that facilitate nucleation. Alloy catalysts, bimetallic combinations like iron-molybdenum or cobalt-molybdenum, are widely used to narrow the diameter distribution or shift the product toward single-walled tubes. The support material underneath the catalyst (often alumina or silica) also plays a role by stabilizing small particles against sintering, which is the tendency of tiny metal particles to merge into larger, less useful ones at high temperatures.

Controlling Chirality

A carbon nanotube’s properties are not just about its diameter. The way the graphene sheet rolls up, its chirality, determines whether the tube behaves as a metal or a semiconductor. Two tubes with the same diameter but different chiralities can have wildly different electrical properties, which is a serious headache for applications like transistors or sensors that need one type or the other.

Chirality is largely set at the moment of nucleation, so controlling it means controlling what happens at the catalyst surface in the first fraction of a second of growth. Catalyst composition turns out to be a powerful tool here. At 600 °C on a silica support, cobalt catalysts predominantly produce a specific single-walled nanotube type known as (6,5), which has a relatively large chiral angle. Ruthenium catalysts under the same conditions yield a much more mixed population of chiralities. The difference traces back to how much carbon each metal dissolves: cobalt and ruthenium have distinct carbon solubilities, which change the nucleation dynamics and push the growth toward different structural outcomes.⁸

Another strategy is evolutionary selection, where the growth conditions are set up so that tubes of the desired chirality grow fastest and survive longest. By moving the reaction zone along the growing tubes or by reversing the reaction to dissolve the most reactive types, researchers can, in principle, winnow a mixed population down to a single chirality during synthesis rather than after it. The approach draws a loose analogy to natural selection: only the tubes best suited to the conditions persist. Despite its elegance, obtaining a single chirality at will remains a challenge, and progress has been somewhat sporadic.⁹

Purification and Sorting After Synthesis

No synthesis method produces perfectly pure nanotubes. The raw product, sometimes called “as-produced soot,” contains amorphous carbon, graphitic particles, residual metal catalyst, and often a mixture of nanotube types. Cleaning this up is a multi-step process. A common first stage is oxidation in air at a few hundred degrees Celsius, which burns off amorphous carbon faster than it damages the nanotubes. Acid washing with hydrochloric or nitric acid dissolves exposed metal particles.

Stubborn impurities, particularly metal nanoparticles coated in a protective graphitic shell, require more aggressive treatment. One recently demonstrated technique uses a combination of carbon dioxide and Freon gas at around 800 °C. The CO₂ oxidizes the graphitic shell to expose the iron particles underneath, and chlorine from the decomposed Freon reacts with the iron to form volatile metal chlorides that evaporate away in the gas stream. A final mild acid wash removes any remaining traces of metal, yielding highly pure single-walled nanotubes.¹⁰

If the application demands nanotubes sorted by electronic type, a second layer of processing follows purification. Techniques include density-gradient ultracentrifugation, chromatography, and aqueous two-phase separation, all of which exploit subtle differences in how metallic and semiconducting nanotubes interact with surfactants, polymers, or solvents. Single-chirality separation can be achieved through refined versions of these methods, though throughput remains limited.¹¹ For large-diameter single-walled tubes, hydrophobic-interaction chromatography with careful pH adjustment can separate metallic and semiconductor fractions with good yield and purity.¹²

Alternative and Emerging Feedstocks

Traditional CVD uses fossil-derived gases as the carbon source, but researchers have been exploring greener alternatives. One approach converts biomass into nanotubes by first pyrolyzing plant material to generate a hydrocarbon-rich gas, then feeding that gas into a CVD reactor. A key challenge is that biomass pyrolysis also releases carbon dioxide, which interferes with nanotube growth. Capturing the CO₂ with a calcium-based sorbent before the gas enters the reactor solves this problem and simultaneously produces hydrogen-rich gas as a co-product, making the process potentially carbon-negative depending on the biomass source.¹³

Template-directed synthesis represents a completely different philosophy. Instead of growing nanotubes freely from a catalyst particle, you deposit carbon inside the channels of a porous template, such as anodic aluminum oxide, which has uniform nanoscale pores. Carbon coats the walls of each pore, forming nanotubes whose diameter is dictated by the template rather than by a catalyst. This can produce arrays of uniform, open-ended tubes connected by a shared carbon sheet, a structure that is difficult to achieve any other way.¹⁴ The template must be dissolved away after synthesis, usually with a strong base or acid, which adds cost and waste but gives exceptional structural control.

At the far end of the spectrum sits bottom-up organic chemistry, where researchers try to build nanotube segments molecule by molecule using reactions borrowed from synthetic organic chemistry. The idea is to construct a short “seed” cylinder with a precisely defined chirality and then extend it. Progress has been made in synthesizing the seed structures, but extending them into full-length nanotubes with high efficiency is still an open challenge.

Health and Safety in Manufacturing

Carbon nanotubes are not just another industrial dust. Their shape, long, thin, and biopersistent, draws comparisons to asbestos fibers. Animal inhalation studies have shown that certain types of carbon nanotubes and carbon nanofibers can trigger lung inflammation and tissue remodeling. Human studies of workers who handle these materials have found some mild respiratory and inflammatory effects, though the cross-sectional design of those studies has so far prevented researchers from firmly proving a causal link between exposure and health outcomes.¹⁵

Airborne exposure levels in manufacturing facilities depend heavily on whether the site is a primary producer (making nanotubes from scratch) or a secondary manufacturer (incorporating nanotubes into products like composites or coatings). Surveys have found that primary manufacturing sites generally keep airborne concentrations below proposed recommended exposure limits, while secondary manufacturers, where nanotubes are handled, weighed, mixed, or machined, are more likely to exceed those limits. Visual evidence of nanotube structures in air samples was found at all surveyed sites, with the highest counts at secondary facilities.¹⁶ Engineering controls like enclosed handling systems, local exhaust ventilation, and wet processing are considered the first line of defense, with personal protective equipment as a backup.

From Lab Curiosity to Continuous Fiber

One of the more dramatic demonstrations of how far nanotube manufacturing has come is the direct spinning of nanotube fibers from a CVD reactor. Rather than collecting nanotubes as a powder and trying to assemble them later, this technique draws a continuous fiber or ribbon directly from the hot zone of the furnace while tubes are still forming. A liquid carbon source (often ethanol or a hydrocarbon dissolved in a carrier) is injected alongside an iron catalyst precursor, and a rotating spindle at the reactor exit catches the nanotube aerogel as it forms, pulling it into a continuous thread.⁴

These fibers combine the mechanical strength and electrical conductivity of individual nanotubes, though not yet at the theoretical limits, with the practical form factor of a textile fiber. They are already being explored for lightweight electrical wiring, structural composites, and biomedical electrodes. The key variables are the same ones that govern all CVD processes: temperature, gas flow rate, catalyst loading, and carbon supply, but tuned for a product that must hold together mechanically as it is pulled from the reactor rather than simply collected as loose material.

Why No Single Method Wins

The persistence of multiple synthesis methods, decades after nanotubes were first identified, reflects a real tradeoff rather than a failure to optimize. Arc discharge and laser ablation produce the most structurally perfect tubes, with few defects in the carbon lattice, because their extreme temperatures allow atoms to find their lowest-energy arrangements. CVD produces tubes with more defects but at vastly greater scale and lower cost, and it offers the most flexibility in where and how the tubes grow, whether on a flat wafer for electronics, inside a porous template, or in a free-floating aerogel for fiber spinning.

For many commercial applications, like reinforcing polymers or adding conductivity to battery electrodes, moderate crystallinity is perfectly adequate and CVD’s cost advantage is decisive. For applications where a single chirality or the highest possible conductivity matters, like quantum computing components or high-frequency transistors, the defect-free tubes from arc discharge or carefully controlled CVD remain the preference, often followed by the expensive post-synthesis sorting described earlier. The search for a single method that delivers chirality-pure, defect-free nanotubes at industrial scale and low cost continues, and it is arguably the central unsolved problem in nanotube manufacturing.