PT-symmetric thermal emitters offer a way to control the spectrum of heat-generated light by engineering how resonators exchange energy and lose it to their surroundings. In a demonstrated design, a silicon resonator array is coupled to a tungsten substrate through a magnesium fluoride (MgF2) spacer. Changing the spacer thickness adjusts the coupling, shifting the system between distinct optical phases and changing its thermal-emission spectrum.
The approach is relevant to engineers exploring infrared sources for sensing and energy conversion. Its central lesson is that emission depends not only on the materials and resonant frequencies, but also on loss, coupling and the relative phase of the interacting modes.
Why thermal emitters are non-Hermitian systems
A thermal emitter must exchange energy with its environment to radiate. That exchange involves optical loss, making the emitter an open system. In non-Hermitian physics, this openness is represented in the system’s eigenvalues: unlike those of an ideal closed system, they can have imaginary components associated with damping.
This framework provides design options beyond simply selecting a material with a desired resonance. In particular, a system combining a nearly lossless resonator with a strongly lossy one can exhibit passive parity-time (PT) symmetry. The two resonators behave like coupled oscillators, and their optical response depends on both the coupling strength and the damping of the lossy resonator.
As coupling changes, the system can pass through an exceptional point, where its eigenmodes meet before evolving into a different regime. The transition affects the mode structure and thermal-emission spectrum. In the reported system, emission is strong in both the PT-symmetric and broken-PT-symmetry phases, but reduced near the exceptional point.
PT-symmetric thermal emitters built from silicon and tungsten
The device pairs cylindrical silicon disks with a tungsten ground plane, separated by MgF2. Silicon acts as the comparatively low-loss, high-quality-factor resonator, while surface plasmon polaritons on tungsten provide the lossy resonance. The spacer thickness controls the interaction between these resonators.
The disks are arranged in a periodic array. This metasurface design suppresses radiative damping of an individual silicon resonator and supports a Friedrich–Wintgen quasi-bound state in the continuum (quasi-BIC). That mode helps localize the resonance, allowing the silicon elements to be treated as individual resonators coupled to the plasmonic substrate rather than as a radiatively dominated array.
For a thermal-emitter design, this distinction matters: the resonator array must retain the local mode behavior needed for the coupled-resonator system to display passive PT symmetry. Isolated silicon resonators did not show the same behavior in the reported implementation.
Spacer thickness tunes the optical phase
Increasing the MgF2 thickness weakens the coupling between the silicon and tungsten resonators. The measured spectra show how this changes the modes. With no spacer, two prominent emission peaks appear. As the spacer becomes thicker, the peaks move together and merge near the exceptional point. With further increases in thickness, the system enters the broken-PT-symmetry phase, where a single prominent mode remains in the measured spectral range.
At a measured operating temperature of 700 °C, the broken-phase emission peak reached a reported quality factor of 37 while maintaining nearly constant peak emissivity. This illustrates how coupling can be used to modify spectral selectivity without changing the basic material pair.
Simulations also distinguish the two regimes through energy storage over time. In the PT-symmetric phase, strong coupling produces energy exchange between the silicon and tungsten resonators, described as Rabi oscillations. In the broken phase, that exchange is absent and the energy behavior is instead monotonic. Field analysis shows that the silicon resonator can store much of the electromagnetic energy in the broken phase, while more than 98% of thermal emission originates from the tungsten substrate in both phases.
Internal phase shapes far-field emission
The relative phase of coupled modes adds another design variable. The device supports horizontal and vertical resonant modes, whose interactions with their image modes differ. The horizontal mode has an out-of-phase image, while the vertical mode has an in-phase image. This phase difference affects damping and the way the modes evolve as coupling changes.
When measurements are taken at an angle that allows horizontal and vertical modes to interact, the system exhibits additional coupling and a more complex spectrum. The results show that internal phase can influence the frequency distribution of far-field thermal radiation. More broadly, phase relationships in coupled resonators can be considered when designing for effects such as superradiance or subradiance, although these possibilities require further investigation in such systems.
Design implications for infrared sources
For engineers and system designers, the work illustrates several practical considerations:
- Treat loss as a design parameter. The contrast between a low-loss semiconductor resonator and a lossy plasmonic mode is essential to the passive PT-symmetric behavior.
- Use spacing to control coupling. The dielectric spacer provides a direct way to adjust resonator interaction and tune the spectral response.
- Preserve localized resonances. The quasi-BIC metasurface design helps limit radiative damping and retain the behavior of the individual resonators.
- Account for collection geometry and mode phase. Measurement angle can reveal coupling between horizontal and vertical modes, while their phase relationships affect damping and spectral features.
The demonstrated structure was fabricated on tungsten with deposited Al2O3, MgF2 and silicon layers, then patterned into silicon disks. Thermal emission was characterized at high temperature using infrared spectroscopy. These details underline that translating the concept into a device involves both optical design and fabrication choices that preserve the intended resonator interactions.
Conclusion
PT-symmetric thermal emitters show how non-Hermitian physics can be applied to control heat-generated radiation. In the silicon–MgF2–tungsten design, spacer thickness tunes coupling and drives changes in the emission spectrum, while resonator phase and quasi-BIC mode localization further shape the response. For infrared-source development, the practical implication is clear: loss, coupling, phase and geometry can be designed together as tools for tailoring selective thermal emission.