Solar cell developers look beyond silicon
Peterborough, New Hampshire, United States February 20, 2006
By Chappell Brown, EE Times
A new slant on compound semiconductors has produced a high-efficiency
photovoltaic process that might beat existing technologies in cost
of production as well.
HelioVolt Corp. (Austin, Texas) has developed a process based on
rapid thermal annealing and anodic bonding that allows high-performance
copper-indium-gallium-selinide (CIGS) films to be deposited on just
about any substrate. Founder and photovoltaic pioneer Billy Stanbery
claims the process can dramatically shorten manufacturing time and
reduce the thermal budget by a factor of 10 to 100. The process
could allow a new class of materials for building integrated photovoltaics
that serve, for example, as a robust coating on external building
materials or on interior furnishings like curtains, to turn buildings
into self-powered photovoltaic plants.
The breakthrough came out of X-ray diffraction studies of CIGS films
under different annealing schedules. Stanbery, who began his photovoltaic
research at Boeing in 1978, was trying to understand why those films
had such a high efficiency in generating electrons from incident
light. What he found - dense, nanostructured domains that act as
p-n junctions - is now known as the Stanbery model and has shifted
the photovoltaic community away from silicon.
Shell and Honda, both of which have substantial photovoltaic development
groups, recently announced that they were focusing their efforts
exclusively on copper-indium-selinide (CIS) systems, the parent
system for CIGS. Shell currently has an 80 megawatt system based
on silicon solar cells, but announced last week that it will be
devoting all of its billion-dollar R&D budget to CIS-based thin-film
panels.
As a compound semiconductor, these systems have a direct bandgap,
unlike silicon, and are therefore inherently more efficient at converting
photons to electric current. "In 1983 I began work on CIS thin films,"
said Stanbery. "At that time it was our group at Boeing and another
group at ARCO who were developing the technology."Shell inherited
its current photovoltaic effort through a few corporate transfers,
and due to the long development period by both the Boeing and ARCO
groups, Stanbery is very familiar with Shell's approach. The exact
nature of Honda's photovoltaic efforts is less well-known, he said.
In 2001, Stanbery founded HelioVolt to develop his own solar-cell
manufacturing process based on CIGS. Last year he received $8 million
in venture funding from New Enterprise Associates (Menlo Park, Calif.).
"CIS is the most absorbing semiconductor known to man and it allows
us to get good results with very thin films," he said. "Thin-film
photovoltaics have their roots in IC technology and really resemble
large-area display technology."
While CIGS thin films hold the record for efficiency, it is not
that advantage alone that is generating the excitement about using
them as basic photovoltaic material.
"As I studied CIGS systems, I began to realize that it had big advantages
for manufacturing," Stanbery said. "Silicon solar cells are essentially
printed-circuit boards. You build the solar cells on silicon wafers
and then solder and wire them together in panels. As a thin-film
technology, CIGS can use the type of lithography and deposition
processes used in integrated circuits."
Stanbery's key discovery was that rapidly annealed CIGS films were
actually a mixture of two phases. Past analyses had started with
slowly annealed films and the X-ray techniques did not pick up on
the second phase, which is not as prominent with slow annealing
schedules. Stanbery found that with more careful observations and
rapidly annealed films, two phases, an alpha and a beta, were tightly
intermixed at the nanoscale. The second phase was not showing up
because it did not have long-range order that would be revealed
easily with X-ray diffraction. "The structure is actually a lot
like the high-critical-temperature superconductors and represents
a new class of materials with complex intermixed phases," he said.
The reason a tightly intertwined structure consisting of two phases
increases the films' conversion efficiency has to do with the role
of domain walls. When photons strike a semiconductor, they generate
an electron and hole, which often simply recombine to either generate
a quantum of lattice vibration, called a phonon, or another photon.
That means the electron is not available to contribute to a current.
In the CIGS system, the holes and electrons become separated by
the domain walls between the alpha and beta phases and do not recombine
easily, making more electrons available.
The rapid annealing of very thin films is the ideal way to create
this intermixed phase material, which makes them ideal for high-throughput
manufacturing. "There are several levels of advantages that you
get out of this," Stanbery said. "On the first level, the films
do not use very much material. That is significant because over
half the cost of silicon solar cells is in the silicon itself. At
the next level, you have a shorter value chain. The silicon cells
have to be wired together and many additional packaging costs occur.
With CIGS films, we eliminate most of that."
Stable, efficient
Moreover, Stanbery maintains that only CIS-based systems address
the two principal disadvantages of all thin-film techniques: long-term
stability and efficiency. "The perception among buyers is that thin-film
systems do not last long and have low efficiency, and frankly there
is a lot of justification for that," he said. CIS, he went on, "is
the most efficient of any photovoltaic thin-film technology and
it is beginning to overlap with silicon in that area." Moreover,
"CIS is the only thin-film technology that has been demonstrated
to be intrinsically stable," he said, though "they may corrode due
to external factors if not properly packaged.
"All the other thin-film approaches have inherent instabilities,
but CIS, like silicon, has no inherent degradation mechanism that
has been identified," Stanbery said.
Copyright
© 2006 CMP Media LLC.
|