Turning Loss Into a Steering Wheel for Light in Photonic Crystals

For more than a century, engineers and physicists have treated loss in optical and photonic systems as an enemy to be conquered. Every attenuation of a signal, every photon absorbed before it reaches its destination, has traditionally been counted as a failure of design. A new experimental study from a team of Chinese researchers now argues that this long-standing intuition misses something profound. In open photonic systems, where energy is free to leak into the surrounding environment, loss is not merely a nuisance appended to an otherwise ideal wave. It is a fundamental property of the wave itself, encoded in the imaginary part of a mode’s complex eigenfrequency, and it can be shaped, structured and exploited as deliberately as frequency or momentum.

The research, led by Yihao Yang and Hongsheng Chen at Zhejiang University together with collaborators from the National University of Defense Technology and Beijing Institute of Technology, was published in Science Bulletin. The work demonstrates a phenomenon the authors call nonreciprocal dynamic degeneracy splitting, a mouthful of terminology that describes a strikingly simple idea. Two waves traveling in opposite directions through the same material, at exactly the same real frequency, can possess dramatically different lifetimes. One survives long enough to carry energy across the sample; the other dies almost immediately. The consequence is that the material itself becomes directional, transmitting light one way while suppressing it the other, without any change in the real-frequency band structure that conventional wave physics would use to describe it.

To understand why this matters, it helps to revisit how physicists normally describe wave propagation in periodic media such as photonic crystals. The standard tool is the band structure, a map of how frequency depends on momentum. Equifrequency contours drawn on this map tell designers which propagation channels exist at a given operating frequency, and much of modern photonics consists of engineering those contours to achieve effects like negative refraction, self-collimation or lensing. What this framework leaves out, however, is time. A real frequency tells you where a mode sits on the spectrum, but says nothing about how long that mode persists once excited. That information lives in the imaginary part of the eigenfrequency, the decay rate, and it is precisely this quantity that the new study learns to control.

The experimental platform was a gyromagnetic photonic crystal operating in the microwave regime, a choice that gives researchers the fine control needed to sculpt both dispersion and dissipation. The team achieved their asymmetric lifetime landscape through what they describe as sequential symmetry breaking, a layered strategy in which each step contributes a distinct ingredient. First, a static magnetic field was applied to break time-reversal symmetry, the fundamental prerequisite for any nonreciprocal response in which forward and backward directions behave differently. Second, a ceramic pillar was displaced from its symmetric position, reshaping the coupling between guided modes and rotating p-like modes and thereby producing an anisotropic, hyperbolic-like dispersion. Third, and most unconventionally, microwave absorbers were deliberately introduced to add loss to the system.

That last step is where the counterintuitive magic happens. Naively, adding absorbers everywhere should simply damp all modes equally, killing the signal without creating any directionality. The key insight is that different Bloch modes overlap differently with the lossy regions of the structure. A mode whose field profile concentrates where the absorber sits will decay quickly; a mode whose field avoids those regions will survive. Because the displaced pillar has already made the mode structure anisotropic, this overlap becomes momentum-dependent, converting uniform local material loss into an asymmetric lifetime landscape across momentum space. Loss, in other words, stops being a scalar nuisance and becomes a structured, direction-sensitive filter.

The researchers then demonstrated how this lifetime asymmetry translates into directional transport. Using a phased array of dipole sources, they excited specific Bloch momenta within the crystal. The source array plays the role of choosing which momentum channel is launched, while the nonreciprocal dynamic degeneracy splitting determines whether that channel can actually propagate. In both simulations and microwave near-field measurements, long-lived bulk modes sailed across the sample, while their counter-propagating partners, sitting at the same real frequency but blessed with much shorter lifetimes, decayed rapidly into the absorbers. Fourier-space analysis of the measured fields confirmed that the directional response tracked the momentum-dependent lifetime distribution rather than arising from boundary states, the mechanism behind many previously demonstrated nonreciprocal transport schemes.

Having established directional transmission, the team extended the same principle to directional imaging, with results that are among the most striking in the study. Hyperbolic-like bulk dispersion provides the negative refraction needed to bring diverging rays back together and form a focus, a property long exploited in superlensing proposals. But negative refraction alone typically suffers from reflected and counter-propagating components that blur and weaken the image. Lifetime asymmetry solves this by dynamically suppressing those unwanted components: they exist on paper, but they decay before they can contribute. At an operating frequency of 7.8 gigahertz, the forward configuration produced a clear focal spot with an experimentally measured full width at half maximum of approximately 0.81 wavelengths, a resolution approaching the diffraction limit of the setup.

The contrast between the two propagation directions was unambiguous. When the array was reversed, no comparable focus formed at all, because the bulk channels responsible for focusing decayed far too strongly in that direction. The researchers reported forward and backward imaging efficiencies of 0.495 and 0.026 respectively, with an isolation ratio of up to 4.8 decibels. It is worth emphasizing what makes this different from earlier approaches to reflection-suppressed negative refraction, such as those based on three-dimensional Weyl surface states. The filtering mechanism demonstrated here operates entirely through the bulk states of a two-dimensional photonic crystal, requiring no engineered surfaces and no three-dimensional architecture, which could make it considerably simpler to implement in practical devices.

The conceptual payoff of the work extends beyond the specific demonstration. The authors frame their advance as a move from band engineering toward what they call complex-band dynamical engineering, a framework in which the real and imaginary parts of a photonic band structure are designed together as a unified whole. Conventional dispersion engineering asks one question: which propagation routes are allowed by the band structure? Nonreciprocal dynamic degeneracy splitting adds a second, equally important question: which of those allowed routes survive long enough to actually carry energy? By answering both questions simultaneously, designers gain a channel-selection rule written directly into dissipation, something previously unavailable in Hermitian photonics where lifetimes are effectively infinite for all modes.

The researchers suggest that the concept is not tied to their particular platform. It could be extended to other lattice geometries, to multiband systems where several families of modes coexist, and to synthetic dimensions where internal degrees of freedom play the role of momentum. Potential applications include mode-selective transport, in which only chosen channel families pass through a device; nonreciprocal frequency conversion; and frequency-domain routing for information processing. Mingyu Tong and Yuze Hu are co-first authors of the study, with Hongsheng Chen and Yihao Yang serving as corresponding authors. If the broader vision holds, the dusty old engineering habit of fighting loss at every turn may give way to a far more elegant philosophy: stop asking how to eliminate dissipation, and start asking where, in the vast landscape of momentum and frequency, you would like it to stand guard.

Subject of Research: Nonreciprocal dynamic degeneracy splitting in gyromagnetic photonic crystals for directional photon transport and imaging

Article Title: Loss-enabled steering of photon propagation

Article References: Loss-enabled steering of photon propagation. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: photonics, photonic crystals, non-Hermitian physics, loss engineering, nonreciprocity, dynamic degeneracy splitting, negative refraction, directional imaging, gyromagnetic media, microwave near-field measurement, complex band structure, Zhejiang University