Japanese Researchers Precisely Align Carbon Nanotubes With Ultrafast Lasers

At the scale of a human hair, alignment means rows, fibers and visible grain. At the scale of a carbon nanotube—a cylinder only a few nanometers wide—it determines whether electrons and light move efficiently, which polarization is absorbed and whether millions of tiny conductors act as a device or as a tangled mat.

A Japanese collaboration has now shown that a finished, initially disordered film can be given different local directions with polarized femtosecond laser pulses. The team brought together Tokyo University of Science, the National Institute of Advanced Industrial Science and Technology (AIST), Osaka University’s SANKEN institute and RIKEN, with support from the Japan Science and Technology Agency and other funders.

The work was led by corresponding authors Daichi Suzuki of AIST and Taishi Nishihara of Tokyo University of Science. Their 13-author paper describes “optical post-processing”: the nanotube film is made first, then a focused beam writes orientation into selected places rather than requiring the entire sheet to be aligned during deposition.

In the best structures, alignment was patterned below one micrometer, extended through films more than 200 nanometers thick and achieved a two-dimensional nematic order parameter of 0.93, where zero is random and one is ideal unidirectional order. The material also acted as a terahertz polarizer with a measured polarization ratio of 24.7.

The Trick: Align by Selective Removal

The phrase “laser alignment” can suggest that light grabs each tube and swivels it into position. That is not what happens. The process is subtractive. It begins with the strong optical anisotropy of a one-dimensional conductor: a nanotube absorbs light differently depending on the relationship between its long axis and the electric field of the polarized beam.

Tubes sufficiently aligned with the laser polarization absorb strongly. The absorbed energy drives rapid ablation. Tubes oriented perpendicular to that polarization absorb far less and remain. What looks afterward like an ordered population is the selected survivor set.

Rotate the polarization and the surviving direction rotates with it. Focus the beam on a small region and the selection happens only there. Direction becomes a writable variable, much as exposure defines a feature in lithography, except that the encoded property is orientation rather than simply presence or absence.

This inversion is the elegance of the method. Instead of forcing billions of nanoscale cylinders to turn, the process asks each tube a binary optical question: does your axis couple to this field strongly enough to be removed?

Why the Pulse Must Be So Short

A femtosecond is one quadrillionth of a second. In that interval light travels roughly 300 nanometers—comparable to the scale of a virus. Such pulses can deposit energy faster than heat spreads appreciably through the surrounding film.

That timing separates selective ablation from ordinary burning. With continuous-wave or nanosecond illumination, heat diffuses outward and the intended directional selectivity is lost as neighboring nanotubes are damaged or removed. With an ultrashort pulse, the most strongly absorbing tubes cross the removal threshold before the local neighborhood equilibrates.

Short duration also brings high peak power without requiring the same total thermal load as a long exposure. This is one reason ultrafast lasers became tools for precision machining, eye surgery and materials processing. The 2018 Nobel Prize in Physics recognized chirped-pulse amplification, the 1985 breakthrough that made intense ultrashort pulses practical without destroying the amplifier.

The study’s physics is nevertheless more specific than “faster is better.” Fluence must sit in a useful window: high enough to remove the selected orientation, low enough to preserve the orthogonal tubes and substrate. Beam focus, overlap, film thickness, nanotube composition and residue all affect that window.

What 0.93 Really Means

The nematic order parameter compresses a distribution of angles into one number. A random in-plane mat averages toward zero. A perfectly parallel population approaches one. The reported 0.93 therefore describes very strong directional order, comparable to leading large-area alignment methods but achieved locally after the film was formed.

It does not mean 93 percent of every tube is geometrically perfect, nor does it measure electronic purity. Carbon nanotube performance also depends on diameter, defects, length, bundling and chirality—the way the graphene lattice wraps around the cylinder, which can make a single-walled tube metallic or semiconducting.

The polarization ratio of 24.7 is a functional test. Terahertz radiation with one polarization was absorbed much more strongly than the orthogonal polarization. Electrical conduction was directional as well, evidence that microscopic orientation translated into device-relevant anisotropy.

The researchers also found a second regime below the ablation threshold. Very weak laser exposure could selectively remove unwanted dispersant and raise the Raman G/D ratio, an indicator of cleaner, lower-defect graphitic material, while retaining a room-temperature process.

From Iijima’s Microscope to a Writable Film

The story returns to Japan in 1991. At NEC’s laboratories, Sumio Iijima used high-resolution electron microscopy to describe needle-like, concentric graphitic tubes in a landmark Nature paper titled “Helical microtubules of graphitic carbon.” The images helped establish carbon nanotubes as a central object of modern nanoscience.

A nanotube can be pictured as a honeycomb sheet of carbon rolled into a cylinder. Multiwalled tubes nest cylinders; a single-walled tube is one cylindrical graphene lattice. The intuitive picture is simple, but the quantum consequences are not. Diameter and wrapping angle determine electronic and optical states.

Researchers rapidly learned that the same geometry giving nanotubes exceptional axial transport also made disorder costly. A solitary tube can conduct brilliantly; a random film forces current across tube-to-tube junctions and directions. In optics, a tangle averages away the polarization response that an ordered array makes useful.

For three decades, the field’s challenge shifted from discovering remarkable properties to controlling populations: grow tubes reproducibly, separate metallic from semiconducting species, choose diameter and chirality, remove contamination, place them where circuits need them, connect them with low-resistance contacts—and align them.

The Alignment Bottleneck

Conventional alignment is usually decided before or during film formation. Tubes can be guided by flow, shear, stretching, electric or magnetic fields, crystalline surfaces, controlled growth or liquid-crystal-like concentration. Each route can make impressive large-area order.

But a sheet aligned in one direction is not the same as a chip containing many directions. Existing processes often operate on millimeter or larger areas and bind orientation to deposition. Making a neighboring region point 45 degrees elsewhere may require another transfer, mask, growth step or mechanical manipulation.

That matters because modern devices are heterogeneous. A polarization camera needs neighboring pixels with different analyzing axes. On-chip optical routing may need local directional elements. Logic layouts need channels that turn and connect rather than one global grain.

Post-processing changes the order of operations. Manufacture a uniform film, then encode local direction where needed. It is analogous to the historical power of lithography: a material becomes more useful when patterning is separated from bulk synthesis.

Four Directions Inside One Pixel

The team demonstrated closely packed regions oriented at 0, 45, 90 and 135 degrees within a micrometer-scale area—the four analyzer directions commonly used to reconstruct linear polarization. They also wrote letters spelling “AIST,” striped photonic structures and a circularly varying radial polarizer.

These are more than decorative demonstrations. A division-of-focal-plane camera places different polarization analyzers beside one another, allowing a single exposure to estimate how the incoming field is oriented. That avoids rotating a filter and losing temporal information.

At terahertz frequencies, polarization can reveal structure that ordinary visible imaging misses. Potential research targets include anisotropic materials, stress, fiber orientation, coatings and concealed features. CNT films are attractive because their one-dimensional electronic response naturally couples strongly to polarization.

The radial pattern shows a second strength: orientation need not be limited to four straight axes. If beam position and polarization can be coordinated reliably, local vector patterns become design elements for polarizers, routing and field shaping.

Why Terahertz Needs Better Pixels

Terahertz radiation occupies the band between microwaves and infrared. It is non-ionizing and sensitive to molecular motion and free carriers, yet sources, detectors and compact optical components remain less mature than their visible-light counterparts.

Traditional polarization analysis can require multiple exposures or mechanically changing an element. A static mosaic of locally oriented CNT pixels could encode several polarization channels at once. RIKEN’s terahertz expertise in the collaboration connects the material process to that system-level opportunity.

A high extinction or polarization ratio is necessary but not sufficient. A practical sensor also needs sensitivity, low noise, calibrated pixel matching, fast readout, environmental stability and integration with electronics. Fabrication variability across a large array could matter more than the best single patterned region.

The paper therefore supplies a material building block, not a finished imager. Its significance is that local orientation and functional terahertz anisotropy were demonstrated in the same platform.

Could It Help Replace Silicon?

Semiconducting single-walled nanotubes are attractive transistor channels because carriers can move efficiently through an extremely thin one-dimensional body. In principle, they offer electrostatic control and low-voltage operation at dimensions where conventional silicon scaling becomes increasingly difficult.

Alignment improves the current path and allows dense parallel channels. Local post-processing could help define directional networks after coating, potentially reducing the number of separately aligned transfers needed for a complex layout.

But alignment is only one gate in a long manufacturing corridor. Metallic tubes can short a logic device; tube-to-tube junctions and contacts add resistance; placement density must be controlled; processing must fit semiconductor contamination rules; and every step must work across wafers with high yield.

The laser method is also subtractive, so it trades material for order. Researchers will need to quantify how much network connectivity is lost, how quickly large areas can be written, whether multiplexed beams or scanning can scale, and how repeated processing affects the substrate and contacts.

A Cleaner Form of Precision

There is an appealing economy in making the material perform its own selection. The beam’s polarization is both instruction and test; the tube’s anisotropic absorption determines its fate. No nanoscale tweezer must locate and rotate each cylinder.

The collaboration also illustrates how a mature research ecosystem operates. AIST contributed sensing and nanocarbon expertise, Osaka University materials implementation, RIKEN terahertz photonics, and Tokyo University of Science physical analysis, with public and foundation funding sustaining the work before a product exists.

Nature Communications notes that the currently available article is an unedited early version that will undergo further production editing. The paper is open access, with source data and supplementary materials, allowing other laboratories to probe the processing window and reproduce the claims.

The most compelling achievement is not that carbon nanotubes have acquired another extraordinary property. Their directionality was already known. It is that the researchers turned an old limitation—the tubes absorb light according to their axis—into a local manufacturing rule.

Thirty-Five Years After Discovery

Carbon nanotubes have often seemed like a future material waiting for manufacturing to catch up. Their strength, transport and optical response are impressive in a single object; technology requires those objects to become a controlled population.

The 2026 work narrows that gap by making orientation programmable after a film exists. It does so at a scale compatible with microdevices and preserves strong directional electrical and terahertz behavior.

There is still a distance from “AIST” written in nanotubes to a commercial polarization camera, and a greater distance to replacing silicon logic. Yet progress in materials engineering often arrives this way: not through one final invention, but by removing a bottleneck that prevented many designs from being attempted.

In 1991, Iijima’s microscope made the tube visible. Thirty-five years later, polarized flashes are being used to decide which tubes remain—and what direction an entire device might take.