Seen from above, a sheet of silicon has been textured with an array of tiny inverted-pyramid shapes, so small that they correspond to the wavelengths of light and can efficiently trap light waves. Image: Anastassios Mavrokefalos
purified silicon represents up to 40% of the overall costs of conventional
solar-cell arrays—so researchers have long sought to maximize power output
while minimizing silicon usage. Now, a team at Massachusetts Institute of
Technology (MIT) has found a new approach that could reduce the thickness of
the silicon used by more than 90% while still maintaining high efficiency.
secret lies in a pattern of tiny inverted pyramids etched into the surface of
the silicon. These tiny indentations, each less than a millionth of a meter
across, can trap rays of light as effectively as conventional solid silicon
surfaces that are 30 times thicker.
new findings are being reported in Nano Letters in a paper by MIT
postdoctoral student Anastassios Mavrokefalos, professor Gang Chen, and three
other postdoctoral and graduate students, all of MIT’s Department of Mechanical
see our method as enhancing the performance of thin-film solar cells,”
Mavrokefalos says, but it would actually work for any silicon cells. “It would
enhance the efficiency, no matter what the thickness,” he says.
student Matthew Branham, a coauthor of the paper, says, “If you can
dramatically cut the amount of silicon [in a solar cell] … you can potentially
make a big difference in the cost of production. The problem is, when you make
it very thin, it doesn’t absorb light as well.”
operation of a solar cell occurs in two basic steps: First, an incoming
particle of light, called a photon, enters and is absorbed by the material,
rather than reflecting off its surface or passing right through. Second,
electrons knocked loose from their atoms when that photon is absorbed then need
to make their way to a wire where they can be harnessed to produce an
electrical current, rather than just being trapped by other atoms.
most efforts to increase the ability of thin crystalline silicon to trap
photons—such as by creating a forest of tiny silicon nanowires on the surface—also
greatly increase the material’s surface area, increasing the chance that
electrons will recombine on the surface before they can be harnessed.
new approach avoids that problem. The tiny surface indentations—the team calls
them “inverted nanopyramids”—greatly increase light absorption, but with only a
70% increase in surface area, limiting surface recombination. Using this
method, a sheet of crystalline silicon just 10 um thick can absorb light as
efficiently as a conventional silicon wafer 30 times as thick.
could not only reduce the amount of expensive, highly purified silicon needed
to make the solar cells, Mavrokefalos explains, but also cut the weight of the
cells, which in turn would reduce the material needed for frames and supports.
The potential cost savings are “not only in the cell material, but also in the
installation costs,” he says.
addition, the technique developed by Mavrokefalos and his colleagues uses
equipment and materials that are already standard parts of silicon-chip
processing, so no new manufacturing machinery or chemicals would be required. “It’s very easy to fabricate,” Mavrokefalos says, yet “it attacks big
create the tiny dents, the researchers use two sets of overlapping laser beams
to produce exceptionally tiny holes in a layer of material—called a photoresist—that
is deposited on top of the silicon. This interference lithography technique is
scalable to a large area. Following several intermediate steps, a chemical
called potassium hydroxide is used to dissolve away parts of the surface that
were not covered by the photoresist. The crystal structure of silicon leads
this etching process to produce the desired pyramidal shapes in the surface,
far, the team has only carried out the first step toward making the new type of
solar cells, producing the patterned surface on a silicon wafer and
demonstrating its improvement in trapping light. The next step will be to add
components to produce an actual photovoltaic cell and then show that its efficiency
is comparable to that of conventional solar cells. The expectation is that the
new approach should produce energy-conversion efficiencies of about 20%—compared
to 24% for the best current commercial silicon solar cells—but this remains to
be proved in practice.