US team shrinks photonics with novel layered approach
New research by scientists from Harvard SEAS (School of Engineering and Applied Sciences), University of Texas Austin and the University of California could help make smaller and more efficient chips for telecomms, quantum communication, and other photonic technologies.
The work, published in Nature Nanotechnology, combines layered compound semiconductor materials with nanostructured metasurfaces, creating devices that can mix and transform different colours of light, including at near-infrared wavelengths used in fibre optic networks.
Today’s optical devices rely on nonlinear frequency conversion i.e. generating new frequencies when high-intensity laser beams are passed through a special material. Typically, these materials are crystals such as lithium niobate or GaAs.
The new work is claimed to closed the gap between material design and device design. Instead of just accepting the nonlinear properties of a conventional crystal, the research team designed a layered semiconductor and a metasurface together so that every part of the device, including electronic states, optical fields, and geometry, worked in tandem toward efficient frequency conversion at a given wavelength.
“This work combines, in a creative way, quantum engineering of an underlying material and enhancing its nonlinearity, together with optimised metasurface design,” said Harvard SEAS' Federico Capasso, who runs the lab that led the research. “The overall nonlinear response is greatly enhanced and made usable for free-space optics.”
The team started with crystalline materials of GaAs and AlGaAs, grown as a stack of multi-quantum wells. By carefully choosing the thickness and arrangement of these layers, they engineered electronic energy levels that produce strong interaction between light and matter. Importantly, they did this using a different type of electronic transition than what has traditionally been used, in order to achieve strong nonlinear effects in quantum wells, and enabling the effects to operate at much shorter wavelengths.
To fully unlock the material’s potential, the researchers then patterned its surface with an array of tiny light-shaping nanopillars (metasurfaces), which trap and shape light at subwavelength scales. The Capasso group has strong expertise in metasurface technology.
The nanopillars on the metasurface organise the electromagnetic field in the correct directions within the layered semiconductor while concentrating the field inside the material, increasing the intensity of light within the structure. The researchers also controlled the symmetry of the fields, allowing interactions that would otherwise cancel out.
Together, these effects boost the effective nonlinear conversion of the light to three orders of magnitude higher than what would be observed from the unpatterned wafer, and higher than previously reported values for comparable devices at near‑infrared wavelengths — demonstrating the device’s potential utility in fiber-optic networks.
Collaborators led by Seth Bank at the University of Texas at Austin designed the multi-quantum well material, while the Capasso group designed how the electromagnetic field behaves within that material, “so that it can make the most of the material property,” said Pernille Undrum Fathi, a PhD student in Capasso’s lab and first author of the paper.
The new platform is based on standard compound semiconductor materials and planar nanofabrication, making it compatible with existing semiconductor processes. Because the nonlinear response is so strong, similar devices could be made extremely compact while still converting light efficiently, according to the team.
Potential uses include chip‑scale frequency converters to generate colours of light that are difficult to produce directly with lasers, and sources of entangled photon pairs for photonic quantum communication and computation.
The research received US federal support from an Air Force Office of Scientific Research Multidisciplinary University Research Initiative grant. The work was performed in part at the Harvard University Center for Nanoscale Systems (CNS); a member of the National Nanotechnology Coordinated Infrastructure Network (NNCI), supported by the US National Science Foundation.
