Researchers from EPFL’s Photovoltaics Laboratory and the CSEM PV-center have developed an economically competitive solution to silicon solar cells. They have integrated a perovskite cell directly on top of a standard silicon-based cell, obtaining a record efficiency of 25.2 percent.
Perovskite complements silicon. Perovskite converts blue and green light more efficiently, while silicon is better at converting red and infra-red light. Calculations and experiments show that a 30 percent efficiency should soon be possible.
Tandem devices combining perovskite and silicon solar cells are promising candidates to achieve power conversion efficiencies above 30% at reasonable costs. State-of-the-art monolithic two-terminal perovskite/silicon tandem devices have so far featured silicon bottom cells that are polished on their front side to be compatible with the perovskite fabrication process. This concession leads to higher potential production costs, higher reflection losses and non-ideal light trapping. To tackle this issue, we developed a top cell deposition process that achieves the conformal growth of multiple compounds with controlled optoelectronic properties directly on the micrometer-sized pyramids of textured monocrystalline silicon. Tandem devices featuring a silicon heterojunction cell and a nanocrystalline silicon recombination junction demonstrate a certified steady-state efficiency of 25.2%. Our optical design yields a current density of 19.5 mA cm−2 thanks to the silicon pyramidal texture and suggests a path for the realization of 30% monolithic perovskite/silicon tandem devices.
The idea of putting perovskite — encapsulated against its nominal environmental degradation sensitivity — is a powerful concept for raising PV cell efficiency, perhaps inexpensively. Will it augur in a new age of PV rollout opportunities from the added differential?
Not likely … because the ratio of added efficiency over added manufacturing cost probably won’t beat out the other alternative… “more area, cheaply”. Because the increase in nominal efficiency increase is perhaps +7% or so (absolute) or 7% ÷ 17% —> +40% relative efficiency increase has yet to prove out as actually not costing entirely 40% more for the dual process cells.
Of course in applications where mass minimization and total area minimization are critical — like space communications or instrumentation satellites, or the Space Station — in these cases, hand selecting the very highest areal efficiency raw photovoltaic cells, before assembly into the large panels that are launched and deployed up there is remarkably effective at consolidating a launch-vehicle to the maximal performance achievable within the mass-budget of each space shot.
But that’s not a revolution, is it?
A revolution would be a change from today’s nominal 17% monocrystalline silicon passivated commodity photovoltaic cells to triple that. 50%. That’d be a revolution. Well, tempered against the added cost of the bespoke process needed to accomplish it on a large commercial scale. If the PV cells ’only cost 3× as much, the advantage for delivering 3× the power on 1× the roof area may still be critically important for many organizations. Especially if most installations are not creating dedicated stand-alone PV arrays apart from available building roof opportunities.
For available roof opportunities, many-but-not-most businesses would be quite powerfully motivated to ‘cover the whole roof’ with PV at 3× the efficiency over covering the same roofs at 1× the efficiency. This is because the number of connectors, automatic leveling converters, wiring cables, roof-to-frame environmental seals, and downwind, DC-to-AC converters remains relatively the same. Oh, scaled of course, but still … 3× the power at perhaps only 1.7× the non-PV costs? That’s a win.
But this announcement neither is as revolutionary as 3× the power output efficiency per unit area, nor is it promising the potential to achieve such, not just in the near future, but even much further out. The fundamental physics limits how effectively coupled a photovoltaic (or stack) of ‘’PN junctions’’ can be to the spread-out spectrum of natural light impinging upon them. In other words, if as an example a ‘’red optimized’’ PN junction PV device is illuminated with red light at exactly the right wavelength, such a cell can — with an amount of technological difficulty not to be dismissed — perhaps achieve conversion in excess of 70% efficiency. Really good, in other words! But illuminate that same cell with green light, and its theoretical capture efficiency drops to less than 40%. And 30% for blue. And less for UV. And rapidly-dropping-to-zero for all light that is longer in wavelength than ‘red’.
So the top estimates of multilayer PV cell performance top out near 50% for 4 layer designs, with each successive layer absorbing ‘’just what it best needs’’ and allows the remainder to pass thru without appreciable absorption. This is something to hope for! Thing is… perovskite-over-silicon doesn’t offer the very first high-hope criterion: thin layers of known photoactive perovskites do not pass thru ‘unneeded’ wavelengths efficiently. Big problem: if our enhancement layer(s) absorb a significant amount of light which otherwise would be convertible to electricity, and turn it to heat, well … it isn’t hard to see that that kills the revolutionary possibility which heterogenous multilayer PV cells might ultimately offer.
This is darn interesting, for sure. If it can be implemented cheaply, and environmentally rock-stable and large-scale, well … then there will be a rapid uptake of the process ‘’by the industry’’, and in turn it’ll represent a marketing coup for solar opportunity sales-and-installation firms. Something new to talk about.
However, if it remains a bespoke, relatively expensive, modestly stable, weakly enhancing technology… then it’ll be sequestered to specialty and niche uses.