Although its name conjurs connections with the dinosaur realm, fallen
Roman (or Romulan) Empire heroes, or fabled half-human/half-horse
creatures,
Synopsys' Sentaurus
has more in common with virtual reality than paleontology, ancient
history, or Greek mythology. The system's "simulation environment,"
what the company's Ric Borges calls "the cockpit," funnels the
company's TCAD tools (leveraged from years of experience and Giga-reams
of data and code developed for the semiconductor industry) into one
platform to model silicon and nonsilicon-based wafer and thin-film
solar-cell process and device technologies.
"People have used TCAD (technology computer-aided design) for
solar-cell development for awhile," Borges (who's no relation to the
[in]famous labyrinthine Argentine writer, the late
Jorge Luis Borges)
told me during Intersolar North America/Semicon West week, "but now
it's become much more active." And not just for modeling the active
areas of the cells.
The system's ability to accurately model and learn more about the
physics of cell operations, improve process recipes and optimize
conversion efficiencies, and explore new cell designs appeals to a
growing number of photovoltaic researchers and device designers.
Users of the software include teams at the
US National Renewable Energy Laboratory, Germany's
Fraunhofer Institute for Solar Energy Systems ISE
in Freiburg, and other renowned institutions, according to the Synopsys
senior product marketing manager. Dozens of papers have been published
from the various research collaborations using the TCAD tool. Sentaurus
also roams the virtual halls of at least three c-Si solar-cell
companies and three compound semi (III-V and thin film) manufacturers,
he said, with many more interested in the "virtual solar-cell" package.
For c-Si applications, Sentaurus can perform a variety of numerical
simulations on the process itself, Borges explained. By inputting data
and tweaking the parameters from process recipes (including specifics
from diffusion, oxidation, implantation, deposition, and etch steps),
geometries, and optical and electrical information, the tool can create
accurate process, optical, and device models that point to ways to
extend cell lifetimes and enhance cell efficiencies.
Borges described
a study
where gettering was simulated, and a link was established between
optimization of that process (in this case, the dependency of the
gettering time on external quantum efficiency) and improved cell
performance. The impact of surface texturing on increased photon
collection has been modeled using 2- and 3-D geometric "meshing
engines." "Optical generation models" have been run across the spectrum
to evaluate interference effects of different antireflective coatings.
In the area of device simulation, he explained how Sentarus can run
a 2-D cell optimization. Using this approach, one selects the
parameters to be investigated (front and back contact characteristics,
substrate thickness, doping concentrations, recombination velocities,
etc.), "parameterizes" (which I pointed out to Ric is not really a
word...yet) the TCAD model, runs a host of simulations, and then
"visualizes" (which is a word) the influence or impact of each of those
process-variable parameters, displaying a variety of coordinate and
scatter plots, IV and efficiency curves, and tabular data.
In an example that Borges showed, the range of screened parameter
settings included a mean substrate thickness of 250 microns with a
minimum/maximum variation of +/-50 microns, a mean back-contact width
of 2.5 microns with a min/max of +/- 0.5 microns, a mean front-surface
recombination velocity of 1800 cm/second with a minimum variation of
900 cm/sec and a maximum of 3600 cm/sec, and about several other
parameters.
Another simulation cited how bulk doping, lifetime, and
front-surface recombination all had a significant influence on cell
response, while yet another "visualization" showed how the use of a
process window with a "response surface model" revealed that efficiency
and open circuit voltages were a function of bulk doping and substrate
thickness.
"You can vary the parameters, see how to optimize them, as well as
[determining] the areas that don't need to be optimized, then
[establish] the proper process window and optimize the design," Borges
said.
Sentaurus can also be used to optimize existing designs or validate
novel ones for amorphous- and micromorph-silicon, III-V, CIGS, and CdTe
PV applications. Borges admitted that because the modeling of these
non-c-Si cells is tougher, with more material defects, dislocations,
and the like to deal with, the focus has been more on the physics than
the process, such as simulating--and tweaking--the epitaxial structures
of high-efficiency multijunction devices. (Note that the company has a
Webcast on thin-film solar-cell simulation coming up Aug. 20-21.)
Synopsys may be extending its reach deeper into the photovoltaic
manufacturing ecosystem, according to Borges. The company's Sabre
"mechatronic" modeling tools for automotive and other industrial
applications "could be used at the module level with something brought
up and leveraged from solar-cell TCAD." The list of optimization
possibilities includes improvements to overall solar power system
efficiencies (including inverters) as well as reducing string-cell
"weakness" and minimizing the "shade effect" of panel arrays.
Despite the challenges of educating the industry and the research
community to the value of TCAD, training people in its use, and
building the support infrastructure, "so far, the people who have
adopted it, really like it," said Borges. The company has focused teams
working on solar-cell TCAD and its applications in each of the main
regions, including several PhD-level specialists.
With the push to properly implement best practices and further
innovate and significantly reduce costs, not just in the device
architecture and fabrication area but along the entire PV value
chain--from raw material feedstock to grid integration--comprehensive
simulation tools like Sentaurus can play an increasingly vital role in
the solar optimization surge.