A research team at Aalto University discovered a range of potential problems in stability tests of dye-sensitized and perovskite solar cells, such as those pictured here. [Image: Valeria Azovskaya, Materials Platform, Aalto University]
In the drive to bring low-cost perovskite and dye-sensitized solar cells to market, research has focused on gains in conversion efficiency. Reported efficiencies of perovskite photovoltaics (PVs), for example, have recently exceeded 20 percent. But another key factor in commercialization will be the cells’ stability—their ability to continue to efficiently churn out power over a long useful lifetime. And the “aging tests” that measure such stability are far less common in the literature than studies focused on efficiency gains.
Now, researchers in Finland have found “serious shortcomings” in a large number of reported aging tests for perovskite and dye-sensitized PVs (Energy Environ. Sci., doi: 10.1039/c7ee02670f). Some 50 percent of the aging tests that the team studied, for example, rested on results from only a single sample, making evaluations of statistical significance impossible. And nearly half of the aging tests that the research looked at were conducted entirely in the dark—a far cry from real-world conditions.
Seeking market viability
Much of the excitement surrounding perovskite and dye-sensitized PVs relates to their potentially low production costs. Perovskite PVs, for example, can be produced using low-temperature, solution-processed methods, and dye-sensitized cells also can be manufactured using common materials and relatively low temperatures. And reported stabilities of these cost-effective materials—while not yet at the years of useful life required for market viability—have lately made impressive strides, with some studies reporting stable lifetimes of a thousand hours for perovskite PVs under specific conditions.
To assess the value of these aging tests, the research team at Finland’s Aalto University examined 157 articles published in Web of Science in 2015 and 2016 that included a total of 261 individual aging-test results. (The researchers focused their search on papers that had “stability-related terms” in their titles.) The team then dug into the details of how the tests were actually conducted.
Substantial variations
What the Aalto researchers found was surprising. For starters, according to the team, the sizes of the test cell groups were “alarmingly small” in many of the stability studies, with half the tests presenting aging data for only one cell in a group. The research harvest also revealed substantial variations in the conditions under which the aging tests were done. Only 15 of the 261 tests the Aalto team investigated, for example, were conducted outdoors, where solar cells commonly hang out. More than 60 percent were performed at open-circuit voltage—that is, conditions corresponding to solar-cell storage rather than operation—with only an eighth of the tests done under an operating load more relevant to real-world durability.
Environmental conditions for the tests also ranged widely. Nearly half of the tests took place entirely in the dark—and, while the Aalto researchers acknowledge that dark tests are important, they argue that “the number of dark tests in comparison to illuminated tests is currently out of proportion.” Only around a third of the tests reported parameters such as the intensities of visible and UV light, humidity and temperature. A full 52 percent omitted mention of the intensity of UV light in the test—a significant stressor likely to shorten observed lifetimes of solar cells in actual use.
Seeking community action
To be sure, not all stability tests share these shortcomings. A recent study from the U.S. National Renewable Energy Laboratory (NREL), for example, reported that perovskite cells could retain 94 percent of their efficiency across a thousand hours of continuous operation under the combined stresses of light (including UV), oxygen and relative humidity of 10 to 20 percent (Nat. Energy, doi: 10.1038/s41560-017-0067-y). “During testing, we intentionally stress cells somewhat harder than real-world applications,” noted Joseph Luther, a co-P.I. on the NREL work, “in an effort to speed up the aging.”
Nonetheless, given their findings across a range of other stability studies, the Aalto scientists believe that the stability-testing community at large has some housecleaning to do. The group calls for stability tests to follow a checklist of items ensuring adequate statistical analysis, clear reporting of test parameters and a better effort to implement aging tests in real-world conditions. The authors even suggest “a series of international summits” to help settle on common standards for the tests. “High-quality, well-reported and standardized tests,” concludes team leader Kati Miettunen, “would reinforce the confidence of industry and investors in these technologies.”