Ultrathin Metalenses with Computational Processing Eliminate Chromatic Aberrations

Lenses act as lifesavers for researchers and photographers. As lenses reflect and refract light, they can be used in imaging systems that steer finding through the microscope and conserve history using cameras.

However, glass-based lenses used at present are huge and cannot be miniaturized. Futuristic technologies (for example, tiny microscopes and ultrathin cameras) mandate the use of lenses produced by using an innovative range of materials.

Researchers from the University of Washington (UW) have now successfully integrated two distinctive imaging techniques—a kind of lens developed for nanoscale interaction with lightwaves, along with powerful computational processing—to develop full-color images. The process has been described in a paper published in Science Advances on February 9, 2018.

The ultrathin lens developed by the researchers is part of a category of engineered objects called metasurfaces. Metasurfaces are two-dimensional equivalents of metamaterials, which are fabricated materials with chemical and physical characteristics not usually seen in nature. A metasurface-based lens, or metalens, includes flat microscopically patterned material surfaces developed to interact with lightwaves. Until now, images captured by using metalenses produce images of high clarity, at best, for just small slices of the visual spectrum. However, the metalens developed by UW researchers, when combined with computational filtering, produces full-color images with very few aberrations over the visual spectrum.

Our approach combines the best aspects of metalenses with computational imaging - enabling us, for the first time, to produce full-color images with high efficiency.

Arka Majumdar, Senior Author and UW Assistant Professor of Physics and Electrical Engineering

In contrast to fabricated silicone or glass, metalenses include repeated arrays of nanometer-scale structures, such as fins or columns. If these structures are suitably distributed at these minuscule levels, they can interact with individual lightwaves with higher accuracy, unlike conventional lenses. As metalenses are also very thin and small, they occupy considerably less space when compared to the huge lenses of high-resolution microscopes and cameras. The manufacturing process of metalenses is same as the semiconductor fabrication process adopted for fabricating computer chips.

Metalenses are potentially valuable tools in optical imaging since they can be designed and constructed to perform well for a given wavelength of light. But that has also been their drawback: each type of metalens only works best within a narrow wavelength range.

Shane Colburn, Lead Author and UW Doctoral Student in Electrical Engineering

In experiments yielding images by using metalenses, the ideal wavelength range thus far has been highly narrow, that is, at best with a width of about 60 nm with higher efficiency. However, the width of the visual spectrum is 300 nm.

Prevalent metalenses in general yield high-precision images within their narrow ideal range, for example, an all-red image or an all-green image. In the case of scenes consisting of colors outside of that ideal range, the images seem to be blurred, with low resolution and other flaws called “chromatic aberrations.” While imaging a rose in a blue vase, a red-optimized metalens may image the red petals of the rose with fewer aberrations; however, the blue vase and the green stem will be unresolved blotches, with higher chromatic aberrations.

Majumdar and his colleagues proposed that when a single metalens can form a consistent form of visual aberration in an image over all visible wavelengths, the aberrations for all wavelengths can be resolved later by adopting computational filtering algorithms. In the case of the rose in the blue vase, this kind of metalens will take an image of the blue vase, red rose, and green stem all with chromatic aberrations of similar nature, which can be resolved afterward by means of computational filtering.

The researchers engineered and developed a metalens with the surface covered by minute, nanometer-wide silicon nitride columns. These columns were adequately tiny to diffract light across the whole visual spectrum, including wavelengths of 400-700 nm.

Essentially, the team designed the positioning and size of the silicon nitride columns in the metalens such that it exhibits a “spectrally invariant point spread function.” Typically, this attribute assures that for the complete visual spectrum; the image will include aberrations that can be outlined by the same kind of mathematical formula. As this formula will be the same irrespective of the light’s wavelength, the team can adopt the same computational processing to “correct” the aberrations.

The team then developed a prototype metalens depending on their design and investigated the performance of the metalens when integrated with computational processing. “Structural similarity,” a metric describing the extent to which two images of the same scene share structure, luminosity, and contrast, is a standard measure of image quality. The amount of chromatic aberrations in an image is inversely proportional to the structural similarity it has with the other image.

The UW researchers discovered that on using a traditional metalens, they could accomplish a structural similarity of 74.8% while comparing blue and red images of the same pattern. In contrast, when the new metalens design coupled with computational processing was used, the structural similarity increased to 95.6%. However, the total thickness of their imaging system is only 200 mm, which is nearly 2000 times smaller than prevalent cell phone cameras.

This is a substantial improvement in metalens performance for full-color imaging—particularly for eliminating chromatic aberrations

Alan Zhan, Co-Author and UW Doctoral Student in Physics

Moreover, in contrast to several other metasurface-based imaging systems, the strategy of the UW researchers is not impacted by the light’s polarization state, that is, the orientation of the electric field in the three-dimensional space in which the lightwaves are moving.

The researchers stated that their technique would act as a roadmap toward developing a metalens, and creating additional computational processing steps, with the ability to capture light more efficiently, and also enhance resolution and give a sharper contrast. This may bridge the gap in realizing tiny, futuristic imaging systems.

The UW, an Intel Early Career Faculty Award, and an Amazon Catalyst Award funded the study.