Coherent emission of light by thermal sources

Thermal radiation is the textbook example of incoherent light. A laser is the textbook example of coherence. That contrast feels so fundamental that most physicists treat it as settled, one of those distinctions so clean it barely needs defending. So when Jean-Jacques Greffet described a thermal source that behaves like a coherent antenna, it forced a genuine reckoning. Either thermal emission can be coherent, or our basic framework for thinking about light and heat needs revision. To understand the puzzle, you need a quick handle on what coherence actually means. Light with a narrow frequency spectrum is temporally coherent. Light with a narrow spatial frequency spectrum — meaning its wavefronts are organized across space, like the beam from a laser — is spatially coherent. A blackbody, by these definitions, is neither. It emits across a broad spectrum in all directions. A laser is both. The gap between them seems absolute. Now consider this: Greffet and colleagues showed that a silicon carbide wafer, heated and patterned with a simple diffraction grating, produces narrow, directional emission lobes in the mid-infrared. Antenna-like lobes. From a hot piece of material. That result doesn't fit the textbook story, which means something in the story was incomplete. The answer lives at the surface. Polar materials like silicon carbide — materials where the lattice carries an electric dipole moment — support a particular kind of surface wave called a surface phonon polariton. These are hybrid excitations, part lattice vibration and part electromagnetic field, bound to the interface and propagating along it. They exist in the mid-infrared, around ten to twelve micrometers, and they carry something unexpected: coherence. Here's why. Inside a hot body, thermal fluctuations drive tiny random currents throughout the material. Each small volume element acts like a random dipole. On its own, such a dipole is uncorrelated with its neighbors — pure disorder. But when that dipole sits near an interface that supports a surface wave, it can excite that surface wave. And the surface wave propagates. Greffet reports that the surface phonon polariton decay length on silicon carbide is on the order of one hundred micrometers. That single number is the key to everything. A surface excitation that lives for one hundred micrometers correlates the field over all the points it passes through. The thermal noise at one spot on the surface becomes linked to the thermal noise one hundred micrometers away because the same surface wave connects them. This was not theoretical speculation. Near-field experiments — scanning an infrared microscope tip across a heated silicon carbide sample held at one hundred seventy degrees Celsius, using a bandpass filter centered at ten point nine micrometers — revealed measurable fringe structures in the field above the surface. The fringes are the signature of spatial coherence: the field is organized. And this coherence exists entirely in the near field, within about ten to one hundred nanometers of the surface, before any grating has been introduced. The grating's role, it turns out, is not to create coherence. It is to extract coherence that was already there. That distinction is the conceptual pivot of Greffet's work, and it's worth sitting with for a moment. The surface of a hot polar material is already threaded with organized, coherent surface waves. They're invisible in the far field because they can't radiate freely, as their wavevectors are too large. The grating solves that problem by supplying a discrete kick of momentum. The grating equation, spoken aloud, says: the angular frequency divided by the speed of light, multiplied by the sine of the emission angle, equals the surface-wave propagation constant minus two pi divided by the grating period. In other words, the grating subtracts just enough momentum from the surface wave to let it escape into free space at a specific angle for a specific frequency. Because different grating lines are illuminated by the same coherent surface wave, they act together — like elements of a phased antenna array — and the diffracted beam is sharp and directional. The long propagation length of the surface wave, those one hundred micrometers, means many grating lines participate coherently. That is exactly why the emission lobes are narrow. The experiment that demonstrated this was a collaboration across several French institutions. Greffet's team at Ecole Centrale designed the structure. Yong Chen's group at L2M and CNRS fabricated it using optical lithography and reactive ion etching. S. Mainguy at CESTA and CEA performed the measurements. They started with a silicon carbide wafer and ruled a periodic grating into its surface, choosing the period to phase match the surface phonon polariton to a target emission angle. A first design based on a perturbative model failed; it didn't produce the expected features. Accurate numerical simulation was required to get the design right. Once they had a working sample, they measured reflectivity first. A deep absorption peak appeared at eleven point thirty-six micrometers for forty-five degree incidence. That total absorption is the signature of the surface resonance. Then, following Kirchhoff's law, which says that emissivity and absorptivity must be equal, they heated the sample and measured its angle-resolved emission spectra. Emissivity peaked at the resonant wavelengths of eleven point zero four, eleven point eighty-six, and eleven point thirty-six micrometers, with the angular position of each peak depending on the detection wavelength. Kirchhoff's law held. But the physics producing those peaks was richer than Kirchhoff ever envisioned. The angle-dependent emission spectrum has a name: the Wolf effect. When a source is spatially coherent, its observed spectrum changes with the direction of observation. A blackbody doesn't do that; it looks the same spectrally in every direction. But the silicon carbide grating did. The peak wavelength shifted as the detection angle changed, because each angle corresponded to a different frequency being diffracted out by the grating equation. You could turn the detector and watch the color change. That is the Wolf effect, and seeing it from a thermal source was the clearest possible demonstration that something coherent was going on. The theoretical framework behind all of this is called fluctuational electrodynamics — a formalism that treats thermal currents as random sources and computes the resulting electromagnetic fields using the full wave equation. Greffet shows that within this framework, the spatial coherence of the near field follows naturally from the dispersion relation of the surface wave. The coherence length of the field is set by the propagation length of the surface phonon polariton, not by some artificial structuring. The grating is a coupler, not a coherence generator. This reframes the entire story of thermal emission. Coherence isn't something you can only get from a laser. It's latent in the thermally excited surface of any polar material that supports surface waves. The disorder of thermal fluctuations is real, but the surface imposes structure on that disorder through the organized propagation of surface phonon polaritons. The practical consequences follow directly. Greffet argues that any incoherent emitter can be made directional by coupling it to a guided mode — a surface wave, a waveguide mode, any mode that propagates and carries phase information over long distances. The grating then converts that guided mode into a beam. This is a genuine generalization of Kirchhoff's law, extending its validity beyond the geometrical optics regime into the wave optics domain where nanostructures live. That opens a path toward ultrathin, nanophotonic thermal emitters — metasurfaces a few hundred nanometers thick that produce directional, polarized, spectrally tailored infrared emission. Sources that would normally require bulky lenses and polarizers could in principle be replaced by a structured surface. The deeper lesson is about where order hides. Thermal sources look disordered from a distance, broad spectrum, all directions, no obvious structure. But at the surface of the right material, coherent surface waves are running continuously, organizing the near field, correlating points that are one hundred micrometers apart. All of that order was there the whole time. Greffet's grating simply gave it somewhere to go. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.