In an intriguing advance for metasurfaces, researchers at Monash University’s School of Physics and Astronomy in Australia came up with a disordered mosaic design approach—nanostructured materials that perform multiple simultaneous optical functions within a single device.
The idea of a mosaic-style design for metasurfaces came to Haoran Ren, an Australian Research Council Discovery Early Career Research Award Fellow, who leads Monash’s NanoMeta Group, in a completely unexpected way.
“In 2024, while attending the International Conference on Orbital Angular Momentum in South Africa’s Kruger National Park, I noticed an artwork on my hotel room wall,” Ren says. “The image immediately caught my attention with its clusters of colorful dots arranged in what looked like a completely random pattern. Their colors separated them into distinct groups—even though the overall layout was disordered.”
This made him wonder whether it was possible to do something similar for metasurface designs. “What if the metapixels responsible for a particular optical function were placed in a random distribution across the aperture? Multiple such sets could then coexist within the same device, each one selected independently by wavelength and polarization,” Ren says. “If we engineered strong optical selectivity, each function could be isolated without crosstalk. It felt like a powerful way to create a ‘knob’ to access independent photonic functions within a single metasurface.”
He phoned a friend, Changxu Liu, a disordered photonics expert at the University of Exeter, who immediately saw its potential. “Together, we developed a careful plan to test whether randomly arranged metapixels could still perform their intended functions, and to determine the limits of how much disorder the system could tolerate. It led us to a systematic set of simulations and, eventually, a full study,” says Ren.
Metasurfaces with multiple optical functions
Ren’s group at Monash includes Chi Li, a Research Fellow, who specializes in nanophotonics, metasurfaces, and two-dimensional materials, and they translated the visual idea into a physical design framework for metasurfaces—planar arrays of nanoscale structures that control light at subwavelength resolution.
“In a conventional metasurface, the entire aperture is used to implement a single function—such as a lens,” says Li. “We show this isn’t strictly necessary, because a given optical function can be achieved using only a randomly distributed subset of pixels. It allows us to assign different groups of metapixels—each corresponding to a distinct function—and distribute them across the surface in a disordered mosaic pattern. The key requirement is that each group can be selectively addressed, which we achieve by designing the nanostructures to respond to specific wavelengths or polarization states.”
As far as materials, the group uses silicon-based dielectric nanostructures for low optical loss, strong resonant behavior, and compatibility with standard semiconductor fabrication. “In this way, the mosaic concept becomes a practical strategy: Combining disordered spatial sampling with optical selectivity to integrate multiple independent functions within a single metasurface,” says Li.
Proof of concept: An optical metalens
The group’s device is a flat metalens and it focuses by encoding a phase profile using nanoscale structures. “A long-standing challenge within this field is chromatic dispersion—different wavelengths naturally focus at different distances and correcting this typically requires complex dispersion engineering, especially for large-aperture metalenses to be used in practice,” says Li.
Their approach takes a different route. Instead of forcing a single structure to work across all wavelengths, “we assign independent lens functions to different wavelength channels,” Li explains. “Each lens is implemented by a subset of metapixels, and these subsets are randomly distributed across the same aperture. As a result, each wavelength ‘sees’ its own lens, and chromatic aberration is removed without needing any dispersion balancing.”
It significantly reduces design complexity and, importantly, doesn’t compromise lens performance, because each channel operates as an optimized single-wavelength lens.
For polarization and structured light, “we apply the same principle of channel separation,” Li says. “Different metapixel groups selectively respond to different polarization bases and map them into distinct spatial outputs. This enables single-shot measurement of the full polarization state, even when it’s not spatially homogeneous but varies across the beam.”
The benefit is complex light fields—such as vector beams or topological structures—can be analyzed without scanning or multiple optical components to enable more compact, faster, and higher-resolution polarimetric imaging.
Performance gains
One of the most thrilling moment of this study for Ren was “when we realized just how well a disordered metasurface can perform,” he says. “Even when we placed the pixels in a completely random pattern, the function was preserved—not just with 10% of the pixels but even with 1% or less. The key was that this redundancy only works when the pixels are arranged with disorder.”
It was “remarkable because it meant all ‘unused’ pixels within the aperture could be reassigned to other functions, as long as we designed strong optical selectivity so each set could be individually addressed,” Ren says. “In other words: Disorder opened the door to packing multiple independent optical functions into the same metasurfaces without crosstalk.”