The little perovskite solar cell was sweating it out in the “torture chamber.” Well, it wasn’t actually sweating, since a solar cell doesn’t have sweat glands, but if it had, at 185 degrees Fahrenheit and 85% humidity, it’d be sweating.
The 6-inch-by-6-inch cell would be in the chamber – a Xenon Exposure System locker – for 1,000 hours.
“We are trying to find the key stresses,” said dungeon master Michael Owen-Bellini, whose more official title is head of the materials and module group at the National Renewable Energy Laboratory in Golden.
“My job is to break things,” Owen-Bellini explained.
Perovskite solar panels could be the next big thing, if they can hold up. Easier and cheaper to make than the market-dominant photovoltaic silicon cells, just a thin film spread on glass or plastic, they are about 50% more efficient in turning sunlight into electricity.
“The problem is they degrade fairly rapidly, which is different from silicon,” Owen-Bellini said. Still,185 degrees? “It is pretty hot, but not outside the realm of what solar panels might face in the desert.”
For more than seven years NREL researchers have been chasing the perovskite grail and the cells have cleared many hurdles. Their efficiency has gone to 29% from 3%, (silicon cells have an average efficiency of 20%) and the durability of the cells has gone from seconds to months.
“Efficiency, performance, durability, these are the dots we are trying to connect,” said Joe Berry, head of NREL’s perovskite research.
And it isn’t just theoretical. In 2020, NREL created the U.S. Manufacturing of Advanced Perovskites consortium to bring the technology to market, working with three universities and six thin-film solar companies, mostly start-ups.
Companies in Europe and China say they are already on the way to marketing perovskite products. Oxford PV, the United Kingdom company that set a cell efficiency record at 29.5%, said it aims to start production of a tandem perovskite-silicon cell in 2022.
“We are right at the cusp of commercialization,” said Laura Schelhas, the consortium’s executive director and an NREL group research manager.
But to paraphrase Huckleberry Finn, “We’ve been there before.”
Some 15 years ago there was a wave of venture capital investment in hot, new thin-film solar companies, only to see the bottom drop out of the market as the price of silicon cells plummeted.
The General Electric Co. scrapped plans for a $300 million thin-film factory in Aurora and Abound Solar, which used technology developed by Colorado State University at its Loveland factory, declared bankruptcy in 2012.
The most notorious thin-film casualty was Fremont, California-based Solyndra, which filed for bankruptcy in 2011 wiping out $1 billion in private investment and leaving taxpayers on the hook for $535 million in federal loan guarantees.
Walk, then run to simple manufacturing
There are lessons to be learned from the Abounds and Solyndras, researchers and federal energy officials say, and perovskites also do not have one key weakness of the earlier thin film technologies.
“We need to make sure we are not trying to run and walk at the same time,” said Garrett Nilsen, acting director of the Solar Energy Technology Office in the U.S. Department of Energy. “Early on they tried to invest in large-scale thin film technologies that still had R&D issues.”
DOE has provided $60 million in grants and financial backing for the development of perovskites.
So even as the consortium is promoting commercial development, Schelhas said NREL is continuing to do the “science behind the down-the-road application, trying to understand the fundamental chemistry and physics.”
Perovskites start with the advantage of being easy to manufacture. The solar cells can be made using existing inkjet printing or roll-to-roll technologies. One company working with NREL in the consortium, Energy Materials Corp. is trying to retool an old plant in Rochester, N.Y., to manufacture perovskite solar films where Kodak photo film once was made.
By comparison silicon cell manufacturing requires refining under high heat, infusing it with other materials, industrial milling machines to precisely slice it into wafers, and a clean room to assemble the cells into a solar panel.
Perovskites’ greatest edge, however, is efficiency. The earlier thin films – such as cadmium telluride or copper indium gallium selenide – had efficiency levels of 7% to 18%.
The idea was that even though they delivered less electricity, the thin films would be far cheaper than silicon panels. But when the price of the more efficient silicon cells dropped to 20 cents a watt in 2020 from $3.40 a watt in 2008, those thin films lost any advantage they might have had.
Perovskites have reached 29% efficiency in lab tests with a theoretical ceiling of 31%. A perovskite-silicon tandem might be able to get to 45%, more than double today’s solar panels.
Perovskite’s power is in the nature of its structure, for perovskites aren’t a thing but rather a shape, a geometry.
The first perovskite – calcium titanium oxide – was discovered in 1839 by German scientist Gustav Rose with additional work being done on the crystal structure by its namesake, Russian mineralogist Lev Perovski.
Any combination of elements – represented by the chemical formula ABX₃ – that create a structure the same as calcium titanium oxide is a perovskite.
It is the same as having a triangle of cheese, a triangle of wood and a triangle of pizza (crust trimmed). They are all triangles.
The particular perovskites of interest are the so-called metal halide compounds, usually with lead or tin and iodine or bromine. What researchers have found is that these perovskite compositions can generate electrons when they are hit by the sun.
To demonstrate the ease, Schelhas cooked up a methyl ammonium lead-iodide perovskite in an NREL lab and gave a small vial of it and an artist’s paint brush to a visitor and had him apply the solution to a square of glass sitting under a bright light and an attached to a voltage meter.
As the clear solution was brushed onto the glass it dried into a black coating and the voltage meter started to jump. “It’s like magic,” Schelhas said.
And since the solution can be fiddled with, perovskites can be “tuned” to capture different wavelengths of light.
There are seven wavelengths from high-energy violet to low-energy red. Silicon absorbs light from the red end of the spectrum, but perovskites can harvest a broader array including high-energy blue photons. (Those are the photons that give the sky its color.)
It is, however, the very nature of the geometry that creates the resilience challenges. “People have made progress on stability, but there are three major issues,” said Letian Dou, a chemical engineering professor at Purdue University whose research of perovskite stability is funded by DOE grants.
Chemically the compounds can be broken down by heat, moisture and even the air. They may become electronically unstable because the ions in the crystal move and the materials degrade quickly under some particular flows of current.
They are also mechanically vulnerable as repeated heating (think of a hot day) and cooling (think of a cold night) may cause cracking and delamination.
The two major approaches to dealing with the stability issues are improving the packaging and encapsulation of the cell to protect the perovskites, or to develop a more durable compound, which Dou said is not an easy task.
“It is very hard to find new materials,” he said. “Halide perovskites were a lucky discovery.”
From the point of view of an industrial product, however, the wide variety of perovskite compounds in itself poses a problem.
Sure, a triangle of pizza is pizza and a triangle, but it could be topped with pepperoni and cheese, or maybe sausage, but pies have also been garnished with coconut, peanuts, and squid.
“The compositional variability is so wide and the manufacturing process and conditions all influence the reliability for the cells and modules,” said Joshua Stein, director of a new perovskite test center at Sandia National Laboratories in Albuquerque. “It’s like complicated cooking or baking. It’s an art form.”
The way companies are testing the performance and durability of their perovskite cells also varies widely, as there are as of yet no standardized test. “Right now, it’s like the Wild West,” Stein said.
Silicon is silicon and the way to test whether a silicon cell is durable is well established.
But with the wide variety of perovskite compounds being employed and no established testing protocol, it will be hard to evaluate the cells and assure that they can compete with silicon and that in turn will make it hard to finance the technology on an industrial scale.
To solve the problem and answer the questions, DOE turned to NREL and Sandia with a $15 million testing project.
“DOE realized they need the horsepower of a couple of national labs to figure this out. It is a complicated problem,” Stein said.
Starting early next year perovskite cells will be put out in Sandia’s high desert solar testing facility. The 7-acre site is home to about 1,000 different kinds of solar modules covering a host of technologies from a host of manufacturers.
“Perovskites are being tested in laboratories, but there has been very little outdoor testing side by side with other technologies,” Stein said.
In the lab most of the testing has been on square-inch cells. The 6-inch-by-6-inch perovskite Sandia test cells, 360 in all, are being made by the universities of North Carolina, Washington and Toledo, all members of the consortium.
Stein said Sandia is also in talks with eight companies to place their perovskite devices at the test site.
“A key for solar technology is how much electricity it can produce outdoors,” Stein said. But just as important as the cells’ success will be their failure.
“What we don’t know about perovskites are what are the important failure modes that we need to develop tests for,” Stein said.
When the researchers find the perovskite’s pressure points outdoors, they will have a guide to developing accelerated indoor testing to measure a cell’s sturdiness.
Those tests and testing protocols will be developed by Owen-Bellini, Schelhas and the research group at NREL.
“We have to know what the key stressors are; we can’t just test blindly,” Schelhas said. “You can’t test the way you test for silicon.”
For example, some perovskite cells have shown better performance in the morning, with decline in output as the day goes on and there is some evidence that the cells recover in the dark.
If so, one of the silicon durability tests – endless hours of light – is probably not appropriate for perovskites.
The 1,000-hour test a perovskite cell was going through in Owen-Bellini’s lab is a silicon protocol.
“Once we’ve identified the most important stresses, we can speed it up because you can’t wait 20 years to see if the cell is still working,” Schelhas said.
And getting beyond 20 years is vital. “DOE requires technology to last for 30 years or more for cost reasons,” DOE’s Nilsen said. “It needs to be durable for at least 20 years.”
And perhaps in a nod to the checkered past of thin film solar, part of the Sandia testing process is an offer to companies of a performance verification and bankability assessment done by the engineering firm Black & Veatch.
“Successful companies will have a bankability report with which they can go to investors,” Stein said.
Perovskites could help get U.S. into the solar game
A major reason for the collapse of the U.S. thin film solar industry a decade ago was the rise of China as the major producer of inexpensive silicon cells. With $41 billion in government subsidies, China went from making almost none of the world’s solar equipment in 2000 to 60% by 2012.
In 2019, China produced about 80% of the world’s solar panels and in 2018 the administration of former President Donald Trump imposed tariffs on Chinese solar imports, charging dumping and unfair trade practices.
Those tariffs are set to expire in February 2022 and the Biden administration is reviewing a request by domestic solar cell manufacturers to extend them.
“Our hope perovskites provide an avenue, specifically for U.S. manufacturers to get in the PV game,” Schelhas said. “Perovskite would be manufactured in the U.S. and there would be supply chain opportunities. Where do we get the glass? Where do we get the polymers?”
One of the U.S. companies working with NREL in the consortium is Boston-based CubicPV. “Labs like NREL are national treasures,” said Frank van Mierlo, CubicPV’s CEO. “They are absolutely essential to winning the battle.”
CubicPV says it is set to produce a tandem perovskite-silicon cell by 2025. “We believe a tandem cell can deliver 30% efficiency,” Mierlo said. “Eventually all cells will be tandem cells.”
There is still a way to go and problems to solve and cash to raise before then.
“It might become one of the dominant technologies,” DOE’s Nilsen said. “If the stars were to align it would be a very lost cost, domestically manufactured solar project.”