Birefringence is the engine behind virtually every liquid crystal optical device. When linearly polarized light passes through a birefringent material, its two orthogonal polarization components travel at different speeds, accumulating a controllable phase difference known as retardation. Voltage-driven liquid crystal retarders exploit this effect to rotate, modulate, and analyze polarization states on demand, and they are ubiquitous in displays, adaptive optics, and imaging systems operating in the visible spectrum. But pushing the same technology into the infrared has proved stubbornly difficult, and the new study offers one of the clearest explanations yet of why—along with a practical way around the problem.

The core issue is one of geometry. Infrared wavelengths are roughly ten times longer than visible wavelengths, and the phase retardation a device delivers is proportional to the product of its birefringence and its thickness. To achieve the same retardation at 3.85 micrometers as at visible wavelengths, a liquid crystal layer would need to be about ten times thicker. That would be an inconvenience on its own, but it becomes a fatal performance problem because the switching speed of a liquid crystal device scales with the square of the layer thickness. A tenfold thicker cell means roughly a hundredfold slower response—far too sluggish for real-time polarization imaging.

A team led by Tomoyuki Sasaki and Hiroshi Ono at Toyama Prefectural University, working with colleagues including Sora Ohara, Kohei Noda, Moritsugu Sakamoto, and Nobuhiro Kawatsuki, reasoned that the better route was to raise the birefringence itself rather than thicken the cell. Their motivation builds on years of work showing that carbon nanomaterials can enhance liquid crystal properties in the visible range. Previous studies had reported that graphene flakes can roughly double the dielectric anisotropy of a nematic liquid crystal at kilohertz frequencies, and that reduced graphene oxide can lift visible-light birefringence by about ten percent at low doping levels. Whether the same trick would work in the infrared, however, was far from obvious.

That uncertainty matters because liquid crystals behave quite differently in the infrared. The molecules absorb strongly at numerous vibrational frequencies across the infrared spectrum, and through the Kramers–Kronig relations those absorption bands reshape the refractive-index dispersion even at wavelengths where absorption itself is negligible. In other words, the optical behavior of a liquid crystal at 3.85 micrometers cannot be safely extrapolated from its behavior in the visible. No one had previously measured the birefringence of graphene-doped liquid crystals anywhere in the infrared, making the Japanese team’s measurements a genuine first.

The experimental recipe combined a commercial nematic mixture, MLC-1902 from Merck, with graphene flakes less than three atomic layers thick and up to a few micrometers across. The flakes were first dispersed in ethanol using a planetary ultrasonic mixer, then blended with the liquid crystal in precisely calculated proportions before the ethanol was evaporated by heating to 120 degrees Celsius. The doped material was loaded into cells made of rubbed polyimide-coated quartz plates, with gaps between roughly 17 and 21 micrometers determined from Fabry–Pérot interference fringes. Birefringence was then measured at two strategically chosen wavelengths—3.85 micrometers in the mid-infrared and 1.55 micrometers in the near-infrared band used by fiber-optic telecommunications—using polarized lasers and a rotating-analyzer technique that recovers the retardation from the polarization azimuth of transmitted light.

The results revealed a striking concentration dependence. At vanishingly small graphene loadings, the birefringence climbed steadily, reaching a maximum at around 4 times 10 to the minus 3 weight percent before declining as more flakes were added. The peak enhancement factors were approximately 1.5 at 3.85 micrometers and 1.4 at 1.55 micrometers, and the team found that birefringence was consistently higher at the longer wavelength across every concentration tested—an indication that the infrared dispersion is governed by nearby molecular vibrational absorption bands rather than a simple Cauchy-like trend. Optical microscope images added a crucial clue: at the lowest concentrations the flakes dispersed cleanly, while at higher loadings they began clustering into aggregates up to tens of micrometers across.

To make sense of this rise-and-fall behavior, the researchers constructed a phenomenological model capturing two competing effects. On one side, π–π stacking interactions between the aromatic rings of the liquid crystal molecules and graphene’s honeycomb lattice enhance the orientational order of the mesophase, and graphene may also modify the molecules’ effective polarizability anisotropy; together these boost birefringence in proportion to an effective order parameter that grows linearly with concentration. On the other side, aggregation at higher loadings carves out disordered regions where the ordering benefit is lost. Modeling the aggregate volume fraction with a nonlinear exponential function and combining the two populations through effective medium theory yielded a simple expression for birefringence versus concentration that fitted both infrared data sets closely, with fitting parameters of similar magnitude at the two wavelengths.

The practical implications are considerable. Because phase retardation depends on birefringence multiplied by thickness, a 50 percent birefringence boost translates directly into cells that are substantially thinner—and therefore dramatically faster—for the same optical function. Faster infrared retarders would in turn benefit infrared polarization imaging, a technique the same group has been developing for applications that include seeing through scattering media and imaging biological tissue, where many materials that scatter visible light strongly become more transparent in the infrared. Notably, the team observed no measurable scattering or absorption losses from the graphene doping at any concentration studied, suggesting the enhancement comes essentially for free.

The authors are appropriately careful about the limits of their findings. The fitted parameters are phenomenological rather than independently measured quantities, the birefringence values were obtained from single samples at each concentration, and the aggregation-based explanation, while consistent with the microscopy, remains a plausible interpretation rather than a quantitatively proven mechanism. The origin of the unusual wavelength dependence, they note, cannot be fully resolved without accompanying infrared absorption spectroscopy, which they flag as a priority for future work. Even so, the central message stands: an optimal, exquisitely small dose of graphene can measurably enhance liquid crystal birefringence in the infrared for the first time, pointing the way toward liquid crystal devices that are fast enough, thin enough, and efficient enough to bring sophisticated polarization control to the infrared frontier.

Subject of Research: Enhancement of infrared birefringence in liquid crystals by graphene flake doping

Article Title: Enhancement of birefringence in liquid crystals with graphene flakes in the infrared region

Article References: Sasaki, T., Ohara, S., Noda, K., Sakamoto, M., Kawatsuki, N., & Ono, H. (2026). Enhancement of birefringence in liquid crystals with graphene flakes in the infrared region. Results in Optics, 25, Article 101135. https://doi.org/10.1016/j.rio.2026.101135

Image Credits: AI Generated

DOI: 10.1016/j.rio.2026.101135

Keywords: liquid crystals, graphene, birefringence, infrared optics, polarization imaging, nematic liquid crystal, phase retardation, nanomaterials, order parameter, π–π stacking, optical retarders, Kramers–Kronig relations