The University of Illinois and Northwestern University have taken artificial
vision to the next level by designing
a fully artificial eyeball, which could one day "plug in" to the
optic nerve for a vision replacement or enhancement.
The project began with Yonggang Huang, Joseph Cummings Professor of Civil and
Environmental Engineering and Mechanical Engineering at Northwestern
University's McCormick School of Engineering and Applied Science, and John
Rogers, the Flory-Founder Chair Professor of Materials Science and Engineering
at the University of Illinois at Urbana-Champaign. They teamed up to create a
naturally curved array of silicon detectors and electronics that can make in
essence a curved camera sensor, which mimics the human eye's design. The
curved surface, like in the eye would act as the focal plane of the camera and
capture an image.
The research was the cover story of the August 7 edition of the journal Nature
and can be read
Optics engineers have long known that the natural solution was the optimal
one. Where cameras have to bounce light through a series of lenses to get
it to form an image on a flat surface, a curved sensor could accept light
passed through a single lens covered aperture akin to the lens and pupil of the
Professor Rogers took over the design of curved surface to print the
electronics on creating a thin elastomeric membrane -- basically rubber -- that
could be stretched flat. After printing electronics on the membrane, it
was unstretched, popping back to a hemispherical configuration.
One critical challenge is that brittle semiconductor materials typically crack
under the stress of curving. To overcome this Professor Rogers and
Professor Huang created an array of electronics so tiny it was unaffected by
the curvature. The array's photodetectors and circuits comprised a 100
micrometer square, comprising a pixel of the device. Similar to buildings
on a curved Earth, the scale of the curvature versus the tiny size of the array
was enough that the silicon went unharmed.
Multiple pixels are connected together via thin metal wires on plastic, which
the pair call "pop-up bridges" as they pop off the rubber surface
when the device snaps back in place. These bridges help to relieve stress
when the substrate returns a spherical shape. They were able to further
relieve stress by sandwiching the silicon devices between two curved layers in
the so-called natural mechanical plane, which minimizes stress.
The method works quite well. When tested after returning to a spherical
shape, 99 percent of the devices still worked. Better yet, the silicon
was only compressed .002 percent -- well below the 1 percent where silicon
devices typically fail and break.
The researchers took initial images from the electronic eye-type camera and
found them to be startlingly clear. When compared to planar camera images
with a similar sensor, the eyeball camera easily triumphed. Professor
Huang stated, "In a conventional, planar camera, parts of the images that
fall at the edges of the fields of view are typically not imaged well using
simple optics. The hemisphere layout of the electronic eye eliminates this
and other limitations, thereby providing improved imaging
To make the final design, the hemispherical membrane plus electronics was
mounted on a hemispherical piece of glass. Then a lens and further
components are added. The end result is a camera roughly the size and
shape of a human eye. The current version is limited to 256 pixels, but
researchers are quickly increasing this number.
This is the first device of this nature that could potentially be used as a
full replacement to the human eyeball. As imaging electronics improve,
the image sensors of the device will only improve, yielding the possibility of
better than human vision resolutions in years to come.
The new device, which has transformed a long-standing science fiction staple
into reality, could eventually see full implant if an interface to the optic
nerve is developed. Before that, it will likely see action in the next
generation of ocular implants. It could also see use in the next
generation of war robots, many of which contain image processing
With either application the revolutionary curved sensor is the key to it all,
as it produces a better image. As Professor
Rogers said, "Optics simulations and imaging studies show that these
systems provide a much broader field of view, improved illumination uniformity
and fewer aberrations than flat cameras with similar imaging lenses.
Hemispherical detector arrays are also much better suited for use as retinal
implants than flat detectors. The ability to wrap high quality silicon
devices onto complex surfaces and biological tissues adds very interesting and
powerful capabilities to electronic and optoelectronic device design, with many
new application possibilities."
The research was funded by the National Science Foundation and the U.S.
Department of Energy.