Camera Lens Works Like An EyeTracy Staedter chats with John Rogers, professor of materials science and engineering at the University of Illinois at Urbana-Champaign
John Rogers Rogers and his team are making rigid semiconductor chip materials do the impossible: bend.
An Eye for Design Rogers's electronic eye camera has an array of digital sensors that curve just like the natural light sensors curve in the back of the human eye.
Curved Image Unlike a flat lens, a curved lens captures an image in a way that maintains focus at the edges. It can also capture a wider field of view than a flat lens.
4:02 PM imtracynotstacy: hello
jarogers11: hi. things seem to be working
imtracynotstacy: this will be fun. a lot of researchers that agree to do the IM interview say it's their first time!
jarogers11: yes, my first. i'm quite primitive when it comes to things like instant messaging, and so on; so far, so good... imtracynotstacy: okay, well let me start out by asking you where you are.....what city?
jarogers11: i'm here in champaign, illinois, the location of the university of illinois.
imtracynotstacy: is that corn country? jarogers11: yes, corn everywhere...makes it easy for the students to stay focused on research. a big plus, in my view.
imtracynotstacy: ha ha
so what is your team's research?
4:05 PM jarogers11:
we're a materials science group that is interested in figuring out how
to make electronic devices do things that are impossible when
integrated in the standard fashion on semiconductor wafers.
imtracynotstacy: and I understand you just had a paper published in the journal Nature recently. What was that about? jarogers11:
yes, in this most recent paper, we demonstrated the first working
camera that is based on a key design principle of the human eye -- a
very simple imaging optic combined with a detector array on a
hemispherical substrate (i.e. an artificial retina) imtracynotstacy: an imaging optic....okay....like a lens? And then a detector array? What is that?
4:09 PM jarogers11:
yes, just a very simple lens, like the one in the human eye. the
detector array serves as the equivalent of the CCD chip in a
conventional digital camera. the key difference, which is important
when the optics of the system are considered, is that our detector
array is integrated with a curved, hemispherical substrate. conventional
devices are built on a flat surface. imtracynotstacy: flat is not like the eye, obviously. okay so that opens up 2 big questions
why make it curved?
And how did you do it?
4:13 PM jarogers11:
yes, good. for the first, it turns out that a simple lens forms an
image that is not flat. in other words, when projected onto a flat
detector chip, the image loses sharpness as one moves out toward the
periphery of the field of view.
by making the detector curved, we can capture the image in a way that maintains focus in these peripheral regions. one can capture a wider angle field of view with a curved detector than with a planar one. imtracynotstacy: ok so HOW do you make curved jarogers11:
yes, traditional cameras avoid this field curvature problem by adding
complexity to the lenses. we show that one can accomplish the same
outcome with simple lenses by making the detector curved. as
for how we make the detector curved, the process involves two key
ideas. first, we make the detector array thin and tiled, such that each
individual detector pixel is made of silicon, but nearest neighbor
pixels are connected via extremely tiny 'ribbon' cables -- thin metal
wires with thin layers of plastic on top and bottom.
in this layout, the detector array takes the form of an interconnected mesh that can be stretched/compressed. This mechanics allows the initially planar system to be conformally wrapped onto the surface of a hemisphere. imtracynotstacy: are the ribbon cables in between really stretching? like elastic? 4:21 PM jarogers11:
well, not exactly. they have the overall effect of making the detector
array stretchable, but the cables themselves are not stretching.
instead, their end-to-end lengths simply change due to a
buckling process.
the point is that this type of mechanical response accommodates all of the strains necessary. the silicon, which is brittle and fragile, does not experience any significant level of strain during the deformation. 4:24 PM imtracynotstacy:
I see. I'm sort of visualizing those "sheets" of small ceramic tiles
that you can buy at Home Depot for redoing your floor or back splash. The
tiles are rigid but because they are evenly spaced on some kind of mesh
backing, you could theoretically tile a cuved surface.
jarogers11: yes, pretty good analogy! imtracynotstacy:
okay, so each of these little rigid semiconductor "tiles" is detecting
light and producing a pixel's worth of image data, right?
4:32 PM jarogers11:
yes, that's right. each pixel also has some basic electronics built
that facilitate automated image readout to a computer. the other part
of the process, after forming the stretchable/compressible detector
array, involves performing the geometry transformation in a
reproducible, high yield manner. for this part of the process, we use a
thin membrane of a soft elastomer, shaped into the form of the
hemisphere that matches the size of the final, electronic eye camera.
mounting such an elastomer membrane onto a stage that pulls it taught
in the radial direction stretches it out from its initially
hemispherical shape to a flat drumhead. transferring the detector mesh,
which is fabricated in a planar layout on a silicon wafer, to this
drumhead, followed by release accomplishes the planar-hemispherical
transformation. the final step involves transferring the hemispherical
detector array to a rigid hemispherical substrate. imtracynotstacy: so what would (or could) you ultimately do with this curved camera?
4:38 PM jarogers11:
the cameras that we can build now are quite good for an academic type
of effort, but many more pixels would be required to make them
commercially competitive. theoretical analysis and experimental
feasibility tests indicate the ability to use these same kinds of ideas
on more sophisticated detector arrays. we're currently working in that
direction. a successful outcome would be a class of camera that
achieves good performance with very low cost, lightweight, simple
imaging optics. such systems could be valuable in applications where
weight and size and cost are very important. the military has
long-standing interests in these types of systems and, in fact, the us
department of defense has funded various research efforts toward this
goal, off and on, for the last twenty years. consumer possibilities
also exist, but likely the more advanced, cutting edge applications
would be the most natural fit, at least in the early stages.
4:40 PM imtracynotstacy: digital cameras today are already pretty lightweight and low cost. How might what you're working on compare? jarogers11:
yes, i agree that the bar is pretty high. for wide field of view and
other characteristics, however, simple lenses and hemispherical
detectors offer a good combination. more generally, the ability to
relax the planar design constraint on the detector could open up new
ways to balance optics and lenses with detector geometry.
4:43 PM imtracynotstacy: okay, so how did you get into this line of research?
4:51 PM jarogers11:
we've been interested for >10 years in ways to achieve electronics
with mechanical properties and geometrical layouts that are impossible
with conventional rigid, planar and brittle semiconductor wafers. about
three years ago, we published a paper demonstrating the ability to make
silicon stretchable by structuring it into ultrathin, wavy geometries
--
i.e. in this configuration, the silicon acts mechanically like accordion bellows to enable an effective type of stretchability. this kind of idea directly led us to the concepts for achieving hemispherical cameras. more generally, we think that these types of systems enable integration of electronics with the human body in ways that are inconceivable with wafer-based systems. the hemispherical detector arrays might, for example, be more naturally integrated with the human eye as retinal implants, compared to the planar chips that presently represent the main focus of work in this area. we are currently exploring several different biomedical devices that similarly exploit stretchablility electronics. 4:52 PM imtracynotstacy: I can see how this device could potentially work well in a retinal implant, since it is based on the eye already.
4:56 PM jarogers11:
yes, it's a natural type of integration. the other devices that we are
working on now involve electronics that conform to the brain, for the
purpose of monitoring spatial/temporal patterns of electrical activity,
and to the heart to provide not only a similar type of monitoring
capability but also a means to pace the contractions in very
sophisticated ways. in these and other areas, we work closely with
doctors and researchers at medical schools -- we build the stretchable
electronics to their specs; they do the evaluation. imtracynotstacy: sounds great. I usually ask researchers a couple of questions
that don't have anything to do with their current
work
for example.....what is the most unusual thing in your office or lab?
4:59 PM jarogers11:
the most unusual thing is easy -- i have one of the world's only
electronic eye cameras sitting here on the shelf above my desk. other
than that, i have a very unique flute-like instrument that one of my
students brought me from istanbul. and no, i don't know how to play
it...
imtracynotstacy: what's an electronic eye camera? jarogers11: very funny...
imtracynotstacy: :) And what does one do in the middle of Illinois for fun?
jarogers11: do research, with some of the best students and senior colleagues that you'll find anywhere! what could be better than that?imtracynotstacy: And what did you eat for lunch?
5:03 PM jarogers11: some 'iffy' chinese food from a place nearby my office. things were/are busy today.
imtracynotstacy: why are they are so busy? jarogers11: just the usual. it's tough keeping up with the students, and besides, classes start next week.
imtracynotstacy: well, in that case, I'll let you get back to your work. Thank you so much for your time.
jarogers11: thanks!
5:05 PM imtracynotstacy: have a good day
jarogers11: bye! |
advertisement
Download the Tech WidgetWhat's On Now
|
-
our sites
-
video
mobile
-
shop
-
stay connected
-
corporate
