Precise control of infrared polarization using crystal vibrations

A natural material has been discovered that exhibits an extreme optical property known as in-plane hyperbolicity. The finding could lead to infrared optical components that are much smaller than those now available.

Hyperbolic materials are highly reflective to light along a certain axis and reflective to light along a perpendicular axis. Typically, one of these axes is in the plane of the material and the other is out of the plane. A material in which both of these axes are in the plane would enable, for example, the manufacture of ultrathin waveplates — optical components that alter the polarization of incident light. Moreover, the reflective behaviour of this material would allow light to be confined and manipulated at extremely small scales (less than one-hundredth the wavelength of the light). In a paper in Nature, Ma et al.1 report the existence of such in-plane hyperbolicity in the natural material molybdenum trioxide.

Many crystals exhibit birefringence, in which their refractive index — a measure of the speed of light in a material — is different along different axes. This property can be used to manipulate the polarization of incident light. The crystal size that is required to achieve sufficient polarization control for practical applications is directly proportional to the wavelength of the incident light and to the strength of the birefringence. Consequently, for light in the mid- to far-infrared regions of the electromagnetic spectrum (with wavelengths of 3–300 micrometres), the crystals typically need to be a few millimetres thick2. To overcome this requirement, a potential solution is to consider materials that exhibit hyperbolicity, which is an extreme form of birefringence.

Hyperbolicity was originally thought to exist only in artificial materials consisting of integrated reflective and transparent domains3. But in 2014, it was observed in the natural material hexagonal boron nitride4,5. The reflective behaviour of both this material and molybdenum trioxide is derived from crystal-lattice vibrations, known as optical phonons, that oscillate in a highly anisotropic (direction-dependent) way. These phonons have relatively long lifetimes (in excess of a picosecond; 1 ps is 10−12 s), which strongly suppresses the absorption of light by the material6. Since the discovery of hyperbolicity in hexagonal boron nitride, a broad array of natural hyperbolic materials has been identified7.

Preliminary investigations of molybdenum trioxide were reported earlier this year8 and showed the existence of hyperbolicity for long-wave infrared light (with wavelengths of 8–14 µm). Ma and colleagues have now demonstrated and characterized in-plane hyperbolicity for the same spectral range. They used this property to confine light to dimensions substantially smaller than its wavelength, through the formation of hybrid light–matter excitations called hyperbolic phonon polaritons. The authors report lifetimes for such polaritons of up to 20 ps, which is about ten times longer than the best values reported for hexagonal boron nitride9.

Because the crystal structure of molybdenum trioxide is highly anisotropic, all three crystal axes, which define the edges of the crystal’s unit cell, have different lengths. Consequently, there is a large difference in the phonon energies associated with these axes and therefore in the corresponding refractive indices — resulting in a birefringence of about 0.31. It should be noted that, earlier this year, a similarly large in-plane birefringence of 0.76 was reported in the natural material barium titanium sulfide for mid-infrared to long-wave infrared light10. However, hyperbolicity was not observed for this material.

The in-plane hyperbolicity of molybdenum trioxide offers opportunities to replace conventional optical components with ones that are much smaller. In particular, using the large in-plane birefringence of this material (or of barium titanium sulfide), infrared waveplates could be constructed from thin slabs that have thicknesses on the order of tens of micrometres (Fig. 1a). Such waveplates could operate in the long-wave infrared, for which commercial waveplates are not widely available and have thicknesses in excess of 1 mm.

Figure 1 | Manipulating infrared polarization. Ma et al.1 show that the material molybdenum trioxide can be used to precisely control the polarization of infrared light. a, Optical components known as waveplates can convert linearly polarized light into circularly polarized light. In the infrared, a waveplate made of a conventional material requires a thickness in excess of 1 millimetre. This material could be replaced with a thin slab of molybdenum trioxide, with a thickness on the order of tens of micrometres. b, Components called polarizers can convert unpolarized light (in which the polarization points in all directions) into linearly polarized light. In the infrared, polarizers made from conventional materials typically need to be thick and use a large grid of metal wires. Such a structure could be replaced with a thin film of molybdenum trioxide that requires essentially no fabrication. c, Nanoscale photonic structures made from conventional materials can emit unpolarized infrared light. But if molybdenum trioxide were used, linearly polarized emission could be achieved.

Furthermore, using the material’s in-plane hyperbolicity, polarizers — components that extinguish undesired polarizations of incident light — could be made from simple 1-µm-thick films (Fig. 1b). Previously, polarizers needed to be thicker and typically required a large grid of metal wires to be patterned on their surface. The remarkable properties of molybdenum trioxide could therefore greatly reduce both the size and the cost of optical components, offering broad applicability in thin, compact infrared devices.

Beyond conventional optics, the properties of molybdenum trioxide could lead to advances in the realm of nanophotonics, which focuses on confining light to nanoscale dimensions. In the long-wave infrared, where the hyperbolicity of this material is observed, nanoscale light confinement necessarily implies defeating the diffraction limit — the usual restriction that light cannot be squeezed into dimensions much smaller than its wavelength. Molybdenum trioxide can beat this limit and, as a result, presents opportunities for producing improved infrared-emitting devices.

For instance, heating nanoscale photonic structures made from materials that can support polaritons can produce light of one or more specific frequencies — rather than light of a broad range of frequencies that that emitted by, for example, conventional light bulbs. Such structures provide an optical source that is akin to light-emitting diodes, but that can be designed to operate anywhere in the infrared. The emitted light from these photonic structures is usually unpolarized (Fig. 1c). It is only through the use of materials that exhibit in-plane hyperbolicity that light of a single, pure polarization can be generated.

Finally, hyperbolic materials such as molybdenum trioxide could serve as the basis for hyperlenses — lenses that produce magnified images of objects smaller than the wavelength of the imaging light. They could also be used in heterostructures (structures in which layers of different materials are combined) to make nanophotonic components that have controllable properties11,12.

Ma and colleagues have demonstrated that, once again, nature has more in store for us than we thought. The future of nanophotonics was once considered to be in the realization of artificial materials, but this study and others in the past few years have demonstrated that, in many cases, the best approach for finding advanced materials is to look among the vast array of natural materials. The results of these studies offer substantial advances in the fields of infrared optics and nanophotonics that could enable infrared imaging and detection to become as ubiquitous as its visible counterpart — a vision that would enable imaging through smoke for first responders, near-instant medical diagnostics and enhanced chemical spectroscopy.

Nature 562, 499-501 (2018)


  1. 1.

    Ma, W. et al. Nature 562, 557–562 (2018)

  2. 2.

    Suslikov, L. M., Gadmashi, Z. P., Kovach, D. Sh. & Slivka, V. Yu. Opt. Spectrosc. 53, 283–287 (1982).

  3. 3.

    Poddubny, A., Iorsh, I., Belov, P. & Kivshar, Y. Nature Photon. 7, 948–957 (2013).

  4. 4.

    Dai, S. et al. Science 343, 1125–1129 (2014).

  5. 5.

    Caldwell, J. D. et al. Nature Commun. 5, 5221 (2014).

  6. 6.

    Caldwell, J. D. et al. Nanophotonics 4, 44–68 (2015).

  7. 7.

    Korzeb, K., Gajc, M. & Pawlak, D. A. Optics Express 23, 25406–25424 (2015).

  8. 8.

    Zheng, Z. et al. Adv. Mater. 30, 1705318 (2018).

  9. 9.

    Giles, A. J. et al. Nature Mater. 17, 134–139 (2018).

  10. 10.

    Niu, S. et al. Nature Photon. 12, 392–396 (2018).

  11. 11.

    Li, P. et al. Nature Mater. 15, 870–875 (2016).

  12. 12.

    Folland, T. G. et al. Preprint available at (2018).

Download references

Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.


Sign up to Nature Briefing

An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing