significant fraction of visible light goes to waste at solar farms, as current
generation cells generally can only take advantage of a very specific band of
the spectrum. Researchers have spent the last decade cooking up cells
that could capture the entire spectrum and operate at higher efficiency, therefore lowering
costs. But those cells were hard to manufacture.
Now a research collaboration between researchers in the Solar Energy
Materials Research Group in the Materials
Sciences Division (MSD) at the U.S. Department of Energy’s Lawrence
Berkeley National Laboratory (Berkeley Lab), and a corporate
partner -- Sumika
Electronics Materials, Inc. in Phoenix, Arizona -- may have
discovered [press release] the key to mass producing these
more efficient cells.
The key was in cooking up the right semiconductor.
I. Finding the Right Material
Ideally a semiconductor would absorb the entire visible and non-visible
spectrum of light, including infrared, the visible colors (white light), and
ultraviolet light. Unfortunately no single semiconductor can do that.
Wladek Walukiewicz [profile],
head of Berkley Lab's MSD team explains, "Since no one material is
sensitive to all wavelengths, the underlying principle of a successful
full-spectrum solar cell is to combine different semiconductors with different
Many semiconductors used in the industry are actually blends of two or more
elements -- one or more of which are semiconducting elements like silicon or
gallium. The blend composition (how much of each element) determines what
wavelength of light is absorbed. What Dr. Walukiewicz's team discovered
in 2002 was that wiring layers of semiconductor could create a full spectrum
cell with different blends (and thus different absorption wavelengths) in
series. That study used blends of indium and gallium in indium gallium
Then in 2004, his team discovered a way to make a single alloy have three bands
(a valence band, an "intermediate" band, and a conduction band), thus
absorbing the full spectrum in a single layer of alloy. The alloy used
was zinc (plus manganese) and tellurium, doped with oxygen. While
the results were good, the resulting material was too expensive and difficult
Now the team has cooked a semiconductor blend that they feel is ready for prime
time. Like the zinc/manganese/tellurium/oxygen blend, the new mix is a
tri-band device. Taking gallium arsenide and replacing some of the
arsenic atoms with nitrogen, to form the third band, form the semiconductor
blend. This replacement can be done using metalorganic chemical
vapor deposition (MOCVD), one of the most common methods of fabricating
The findings are a beautiful example of how research that is far from
production can iteratively evolve and mature into something commercially
viable. It is also a testament to theory pointing the way for
The team used a special model developed in lab, called the band
anticrossing model. That model states, among other things, that the
element replaced in the alloy must be from the same group in Mendeleev’s
original periodic table. It also gives clues as to what percent to use.
The original zinc/manganese/tellurium/oxygen blend was a so-called
"II-VI" semiconductor alloy, while the new blend
(gallium/arsenic/nitrogen) is a "III-V" alloy.
II. Preparing the Blend for Production
Though the researchers had a good idea what material and what percent of it
they wanted to use for this study, the key to building a working device was in
doping the semiconductor so as to isolate the intermediate band.
Describes Dr. Walukiewicz, "The intermediate band must absorb light,
but it acts only as a stepping stone and must not be allowed to conduct charge,
or else it basically shorts out the device."
By adding additional doped layers the researchers were able to essentially
isolate the central layer, and assured it was passing on its energy and not
Part of the verification that the device worked was taken from when current was
applied to it. When a solar material is subject to current, it reverses
from absorbing photons, to emitting them (goes from a photodiode to a
light-emitting diode). The resulting LED admitted emitted three peaks in
the visible range and a peak in the infrared range. This means that not
only does the device promise high efficiency absorbance of sunlight, but it may
also be promising as a light-emitting diode for applications such as lighting and lasers.
It was a long path to the full-spectrum cell, but the results in
"Engineering the Electronic Band Structure for Multiband Solar Cells"
certainly offer a convincing argument that a commercially viable solution has
finally been reached. The paper was written by team members Nair Lopez,
Lothar Reichertz [profile], Kin Man
Yu [profile], and Wladyslaw
Walukiewicz, plus Sumika Electronic's Ken Campman, and published [abstract] in the January 10, 2011 edition of the
peer-reviewed journal Physical Review Letters.
quote: My neighbors are even more extreme they have a 2kw windmill and only 800 watts of PV, and have only had to turn on a back up generator a few times in 8 years.
quote: Maybe instead of attacking solar all the time we should attack energy sucking lifestyles that make it non-viable.
quote: Didn't you just make a point that solar was viable just to turn it around and make a point that it isn't?
quote: Those of us in northern parts of the US receive only 1/4 the total solar energy in the winter as we do in the summer.
quote: I see to many armchair-pundits on this site discrediting solar every chance they get.
quote: New tech is great, but solar already made sense for people a long time ago and only continues to do so
quote: At some point, perhaps we will face reality and build more nuclear reactors.
quote: The resulting LED admitted emitted three peaks in the visible range and a peak in the infrared range.
quote: A significant fraction of visible light goes to waste at solar farms, as current generation cells generally can only take advantage of a very specific band of the spectrum.