PerspectiveSOLAR CELLS

Perovskite solar cells must come of age

See allHide authors and affiliations

Science  26 Jan 2018:
Vol. 359, Issue 6374, pp. 388-389
DOI: 10.1126/science.aar5684

Perovskite solar cells (PSCs) have reached peak performances rivaling those of established technologies that have been painstakingly optimized for decades (13). Their high power outputs and low production costs have attracted serious industry attention from established companies and have led to the founding of multiple start-up companies (4). However, for commercial products, long-term stability is crucial. Thus, for perovskites to succeed, an informed discussion on a standard for stability data is required.

Characterizing perovskites is notoriously difficult because of the “hysteresis phenomenon” caused by mobile ions (see the figure, top panel) that may lead to the discrepancy between scans of the current density J for forward and reverse sweeps of voltage V (5, 6). As shown in the bottom left figure panel, traditional fast JV scans (>100 mV/s) may overestimate power outputs. Slow scan speeds of ∼10 mV/s are needed, together with maximum power point (MPP) tracking, where the solar cell is operated under full illumination and load-resembling working conditions (not illustrated in the figure).

New aging standards for perovskite solar cells

Mobile ions cause hysteresis effects in measurements of these solar cells that can lead to artifacts in current density–voltage (JV) scans. Proposed stability measurements require full temperature and atmosphere control as well as maximum power point (MPP) tracking under constant illumination.

GRAPHIC: C. BICKEL/SCIENCE

Before the hysteresis phenomenon was fully appreciated, JV curves were frequently taken naïvely employing fast scan rates or preconditioning partially used for other materials (e.g., amorphous silicon), such as light soaking or prebiasing, altering ion migration substantially. From these measurements, “efficiency” values were extracted that were incomparable among research groups because of incomplete reporting of the precise testing conditions. The ambiguity in measuring efficiency, one of the most fundamental parameters, stifled the development of the research field. This is also the reason why certified perovskite efficiencies are still being classified as “not stabilized” (1).

With the JV question being slowly settled, there is an evident parallel with respect to stability measurements of perovskites. For example, using elevated temperature and light cycling for aging are well known for silicon; these tests are frequently conducted under nonoperational conditions (e.g., in the dark), because for silicon solar cells, such measurements have been established as a sufficient proxy for long-term stability (7). However, PSCs do not have established metrics for long-term stability extrapolation because perovskites are unusual semiconductors exhibiting ion migration that can take many hours to reach stable, steady-state conditions. Indeed, accelerated aging protocols for PSCs need to be developed as a function of standard aging parameters. Until then, all aging tests must be recorded and justified meticulously. Previous aging routines (e.g., for organic or silicon photovoltaics) serve better as guidelines than as a mandatory standard.

The suggested aging routine for PSCs in the figure (bottom right panel) acknowledges the nature of perovskite semiconductors. Because of the effects of ion migration, obtaining meaningful long-term stability data for PSCs requires testing under real working conditions. This entails an emphasis on using MPP tracking, where the solar cell is operated under illumination and load (3, 5). Ion migration may cause an initial or “reversible loss” component. Thus, any long-term data require at least 100 hours, ideally even more, to distinguish the effects of short-term ion migration and long-term degradation. In this sense, “hot” or “cold” drawer tests—i.e., dark storage at ambient or elevated temperature, where the solar cell is not generating power—are unsuited to analyzing short-term versus long-term degradation (8).

Importantly, PSCs may recover after an overnight resting period without load and illumination. Therefore, measurements should be taken before aging (open circle in bottom right figure panel) and after aging, including a resting period in the dark (solid circles in the bottom right figure panel) to determine the “reversible loss.” Devices with different starting performances vary largely: A low-efficiency device may be considerably less stable than a high-efficiency counterpart (see red and black lines in the bottom right figure panel). Hence, PSC aging routines need to disclose all data with absolute rather than relative (or “normalized”) values and ideally provide repeated measurements to ensure the validity of the data. Bare perovskite layers are very sensitive to extrinsic degradation, e.g., ambient moisture or oxygen (9). These can be remedied with encapsulation or, alternatively, an inert nitrogen atmosphere resembling “perfect encapsulation” (10, 11), but these procedures and conditions must be stated as well. The outlined aging routine is only a starting point and may well change as the understanding of aging mechanisms in PSCs is improved. However, it is along these parameters that a calibration curve for accelerated aging can be developed to start the process of establishing a standard.

The power outputs from PSCs are sufficiently high to make them serious contenders for sustainable energy sources. In the coming years, PSCs will undergo relentless scrutiny for one of the most important topics in terms of market viability: long-term stability. Unfortunately, characterizing PSCs is challenging because of ion mobility. These issues already cast serious doubt on the reliability of efficiency measurements from the early period of perovskite research. A critical self-reflection on aging standards is needed to establish a basis for norms and common procedures with the chief aim of providing a calibration curve that fully captures the long-term stability of PSCs. Aging routines that account for the hour-long settling time of mobile ions, as outlined here, can then allow for the development of accelerated aging routines to guide industrial development.

References

View Abstract

Stay Connected to Science

Navigate This Article