IMEC Mechanically Stacks GaAs/Ge Multijunction Solar Cell
Working toward triple-junction solar cells with efficiencies >40%, IMEC has successfully mechanically stacked a GaAs cell on top of a germanium cell. Unlike monolithic multijunction cells, mechanically stacked cells require no current or lattice matching.
Aaron Hand, Executive Editor -- PV Society, 9/21/2009
In a promising demonstration of its goal to produce mechanically stacked, high-efficiency multijunction solar cells with efficiencies >40%, IMEC is presenting a mechanically stacked GaAs/Ge multijunction solar cell at this week's European Photovoltaic Solar Energy Conference (PVSEC) in Hamburg, Germany.
 |
|
IMEC's mechanically stacked GaAs/Ge cell. |
At the top of the stack is a one-side contacted 4 µm thick GaAs top cell with an efficiency of 23.4%. IMEC has successfully stacked the GaAs cell mechanically onto a germanium bottom cell, which is separately contacted. It has a potential efficiency of 3-3.5%, which is higher than germanium bottom cells in state-of-the-art monolithically stacked InGaP/(In)GaAs/Ge cells.
The mechanically stacked multijunction solar cell is the first demonstration in IMEC's goal to produce mechanically stacked, high-efficiency InGaP/GaAs/Ge triple-junction solar cells. Giovanni Flamand, team manager at IMEC, expects to show a first working triple-junction cell by the beginning of next year.
Multijunction solar cells are traditionally monolithically stacked - a task that is easier, but not without its drawbacks. A typical triple-junction cell begins with a germanium substrate, on top of which is grown a lattice-matched InGaAs layer. On top of that is grown a lattice-matched InGaP layer. Each of the three layers has a sub-cell, and any given current must pass through all three sub-cells. Current matching is used to avoid internal losses in the solar cell.
Stacking the layers mechanically instead removes the need for current matching as well as lattice matching. The germanium cell is made first with separate contacts. IMEC's latest demonstration integrated a GaAs sub-cell on top of that, again with its own contacts so that the current can be extracted separately from each sub-cell, alleviating the need for current matching. Because the materials are not grown on top of one another, there is also no longer a need to match the lattice constant.
This makes the cell very versatile in terms of the materials that can be used, according to Jef Poortmans, photovoltaics program director at IMEC. "Someone pops up and says, ‘Okay, I can make a nice InGaN cell with a bandgap of 2.2 eV.' It would be very difficult to integrate that on top of an existing monolithic multijunction cell because how would you achieve the lattice matching? But in this case, it's possible to do that because it's just a separate layer, which you built on top of the stack, and you don't need to current match anymore, and you don't need to lattice match anymore."
Of course, mechanically stacked cells come with complexities of their own, including knowing how to make the mechanical stack and extracting current from the separate contacts, Poortmans noted. The increased complexity also leads to increased costs. "But that additional cost can be offset by the fact that you have, in principle, slightly higher efficiency, but especially a higher energy yield over the day," he said, explaining that monolithically stacked cells require current matching, which can only be satisfied at one spectral condition. "That spectral condition changes over the day, so that whereas a classic monolithic multijunction cell is limited by this current matching issue, in this stack you don't feel it, or you feel it much less. And the result of that energy yield — not only power efficiency, but energy yield — could be higher, which could offset the potential increases of the processing cost due to the complexity of the process."
Ultimately, IMEC's goal is to produce mechanically stacked triple-junction cells, which are expected to have conversion efficiencies that are 1-2% higher than today's monolithic triple-junction solar cells (>40% with concentrated illumination).
"Now we have integrated the GaAs on top of germanium. But the real breakthrough is when we on top of that can also integrate a thin-film InGaP cell," Poortmans said. The researchers have already made thin-film InGaP cells using the same process as they did for the GaAs cell, he noted. The next step will be to integrate them on top of each other.
IMEC is also looking to use the same principles to stack an InGaP cell on top of a silicon cell, which would cost appreciably less than the germanium and GaAs sub-cells, Poortmans said. Although efficiency would be lower than the triple-junction cell, the lower cost should compensate for that.
R&D Pushes c-Si Solar Cells Through Hurdles
08/31/2009Si and Ge Solar Cell Efficiency Records Set
10/11/2007
