 |
|
Squeezing light into small spaces a feat of physics
Nothing informs everyday intuition that light has size. It can fill
a whole room and yet appears to slip through any gap. At the scale of
microns (millionths of a meter) or nanometers (billionths), however,
light begins to look large.
Electrical engineers now routinely work at those scales. Such miniaturization
has brought benefits such as powerful microprocessors. But as engineers
have crammed billions of transistors into modern electronics, they’ve
gone below the size, or wavelength, of light. For light to interact with
these devices, and for optical technologies to be similarly miniaturized,
engineers need to squeeze light down to similar scales.
A promising way to “miniaturize” light is to shine it on
specially crafted gratings of “subwavelength” (i.e. smaller
than light’s size) holes in films of opaque material. Important
recent advances in this technology have come from the work of electrical
engineering research associates Peter Catrysse in Associate Professor
Shanhui Fan’s lab and J Provine in Professor Roger Howe’s
lab.
“It is quite unexpected based on traditional diffraction theory
that you could get light through such a small hole,” Catrysse says. “But
in traditional theory you don’t account properly for the properties
of materials. When you do that, you can get light through much smaller
holes than you expected or you can get much more light through a given
sized hole than you expected. That is very exciting for a range of applications.”
Among those applications could be improvements in optical microscopes,
optical data storage (think DVDs beyond Blu-Ray), electronics manufacturing,
solar power cells, and digital imaging. But to make these ideas really
work, engineers have to understand more about how the gratings work and
what materials are best for maximizing the amount of light that can get
through minimal spaces.
Recently—this year in particular—Catrysse, Provine, and
colleagues have made substantial progress on this agenda. Based on their
most recent and promising results, Provine says, the group hopes to make
a sensor capable of monitoring combustion as it occurs in engines. Such
monitoring could lead to gains in vehicle fuel efficiency.
Over the surface and through the holes
Scientists learned about a decade ago that light could shine through
subwavelength holes in a metal film. One mechanism is based on a phenomenon
known as “surface plasmons,” in which light excites waves
of electrons along the front and back surfaces of the film. If the metal
film is not too thick, the waves can couple evanescently through the
holes and excite waves on the opposite surface. The light’s energy
is transmitted once the waves on opposite sides start resonating together.
In 2005 Catrysse, Fan and doctoral student Hocheol Shin showed that
light of particular frequencies can also propagate directly through the
holes. In addition, more gets through that way than does via the surface
plasmon approach. Catrysse says the light now excites “propagating
plasmon” modes inside the hole. The mode that can make it through
depends on the diameter of the hole and the materials used to make the
grating and the holes.
In simulations published in the journal Physical Review B in
2005, the team showed that as much as 80 percent of the light shone on
a metal grating with 50 nanometer radius holes could get through at 442
nanometers (blue/violet) and 520 nanometers (green), wavelengths 9 to
10 times larger than the radius of the holes.
The discovery was exciting, but as the hole radius got smaller, the
range of modes that could get through got smaller, too. Catrysse and
Fan set back to work and in November 2007, they reported in Applied
Physics Letters a way to make gratings much more versatile. They
discovered that in simulations they could widen the range of transmitted
modes (bandwidth) if they varied the materials that they used to fill
the holes. They filled the holes with two different insulators in a concentric
arrangement (like a carrot). The result was high transmission through
50 nanometer radius holes for wavelengths of about 300 nanometers (ultraviolet)
to nearly 700 nanometers (near-infrared). This is almost twice the bandwidth
of holes filled with a single insulator. Moreover, it allows for scaling
of the hole with minimal loss of transmission bandwidth.
Heat ducts
For much of this time, Provine was at UC Berkeley working to make physical
gratings similar to those that Catrysse and Fan were studying theoretically.
In March 2006 Provine moved to Stanford and became the research associate
at the newly established Center for Interfacial Engineering for Microelectromechanical
Systems (CIEMS) led by Howe. Collaborative research into new optical
microelectromechanical systems (MEMS) is a major thrust area at CIEMS.
“When I came here I really had run into a roadblock in terms of
simulating my results to look at the theory behind what I had done,” Provine
says. “I thought right away, well these are the guys I’ve
been reading papers by. I might as well just drop by. Very quickly in
talking with Peter, we found that we had a great deal of overlap.”
Catrysse told Provine that he and Fan were beginning to look into subwavelength
transmission at a scale where plasmon-based transmission doesn’t
work well: the middle of the infrared spectrum. Infrared light, commonly
known as heat, is a longer wavelength than visible light. It is important,
however, in that many molecules of scientific interest radiate in this
spectrum. Developing sensitive molecular sensors therefore requires miniaturizing
mid-infrared light.
Since metals and plasmonics weren’t appropriate, Catrysse had
been simulating gratings made of a different material: silicon carbide
(SiC). When SiC is drilled with a regular array of holes, surface phonon-polaritons
(which are hybrids between vibrations in the crystalline material’s
structure and light) rather than surface plasmons (which are hybrid between
electronic oscillations and light) arise, and allow for some transmission
at mid-infrared wavelengths. Meanwhile, infrared light of specific wavelengths
can also propagate directly through the holes, as it does at the nanoscale
in metals.
Simulations Catrysse and Fan published in Physical Review B in
February 2007 showed that the situation gets complicated, however, at
wavelengths where both surface waves and directly propagating waves are
stimulated. The different transmission mechanisms interfere, leaving
no transmission at all. For example, in a SiC grating with 2.8 micron
holes spaced 10.4 microns apart, more than half the light with wavelengths
between 11.5 and 12.5 microns sailed through, but modes of 11.08 to 11.33
microns were seemingly stopped dead. The reason was the interference.
A particularly important result in the paper, Catrysse says, was parsing
out these ranges of transmission and non-transmission and explaining
why they happen. After all, engineers who want to make real devices for
transmitting mid-infrared light, or perhaps switching transmission on
and off, will need this guidance.
Provine knew he could build an SiC grating through collaborations with
chemical engineering Professor Roya Maboudian and graduate student Christopher
Roper at UC Berkeley (Roper is co-advised by Howe). The team deposited
a polycrystalline SiC film and patterend it into an array of 2.8-micron
radius holes, spaced 10 microns apart, using fabrication tools in the
Berkeley Microlab.
The team exposed the gratings to mid-infrared light and took direct
measurements of the transmission. The curves from the real device weren’t
as ideal as Catrysse’s simulated curves, but they bore a convincing
resemblance. Each, for example, sported two peaks wavelengths of maximum
transmission with almost exactly the same spacing. The group presented
the results at the IEEE/LEOS Conference on Optical MEMS and Nanophotonics,
which was held in Hualein, Taiwan, in August 2007.
SiC is stubbornly durable which gives it good resistance to extreme
heat, Provine says, so an SiC grating could be the core of a sensor that
could be placed right inside the combustion chamber of an engine. Such
a sensor could yield important data in the mid-infrared spectrum, and
contribute to studies of engine efficiency and performance.
It could also be the first in a long line of joint projects on subwavelength
transmission.
“This is a collaboration that has a really extensive future
to it.” Provine says. “This is something we’re all
interested in.”
January 2008
|
|
|
Last Modified: January 8 2008 09:55:39 AM |
|