Intrinsic degradation factors:Ionic movementOnce environmental (e.g., moisture or UV) andother external factors (e.g., thermally unstablespiro-OMeTAD) have been addressed, intrinsicfactors associated with the perovskite can be moreproperly assessed. Ionic movement in perovskiteshas been shown to lead to fast (seconds to minutes;Fig. 2B) and slow (minutes to hours; Fig.4E) performance degradation, namely hysteresis(37) and reversible losses (78, 79), respectively. Inthe latter, devices suffer from decreased PCEduring aging and recover to the initial value afterdark storage for a few hours. The reversible losseshave been attributed to lattice deformation, andhence to halide migration within the perovskite,when the device warms up as a result of theinfrared component of the sunlight (79). Alternatively,the losses have been assigned to migrationof cations at a slower rate than halides (Fig.4E) (78). A comprehensive study investigating thereversible losses on different device architecturesprepared with different charge-selective contactsis still lacking. From the few studies reported,migration of ions within the perovskite seemsto have a milder impact on the device performancefor planar p-i-n (Fig. 4F) (24, 26, 70). Regardless,it is important to establish how organic and inorganiccontacts affect the accumulation of ionsat interfaces in the long term, because mild reversiblelosses might be an obstacle for long-termstability.Shelf versus maximum powerpoint stabilityLong-term stability can be assessed in variousways, including storing devices on a shelf in thedark (“shelf test,” Fig. 4C), MPPT in a controlledlab environment, and outdoor tests. The stressby intrinsic and extrinsic degradation parameters(illumination, voltage, current, temperature,atmosphere) varies considerably among thesetests. Therefore, depending on how the devicesare measured, different performance metrics canbe extracted. For instance, stability curves (PCEversus time) can seem substantially better whenextracted from periodically collected currentvoltagecurves (because of hysteresis) rather thancontinuous MPPT (Fig. 4G) (11, 22). Establishinga robust protocol to measure operational longtermstability may require the adoption of conventionalMPPT (rather than the shelf test) forperovskite-specific phenomena to simulate conditionscloser to field operation.Opportunities and challengesPSCs have made remarkable advances in just afew years, in part by borrowing expertise fromother more established fields, such as organic ordye-sensitized solar cells, for contact layers andarchitectures. Perovskites keep surprising withnew applications such as x-ray detection, andhave displayed impressive new properties suchas long-lived hot carriers, which promise toallow exceeding the Shockley-Queisser limit inPSCs (80). The remarkably low loss in potentialrecently reported by PSCs (21) shows that thissolution-processed solar cell could soon approachother state-of-the-art technologies, such as Sior GaAs.Now that efficiencies are beyond 20%, theperovskite community must focus on long-termstability. More robust testing procedures areneeded to properly assess PSC stability. MPPTunder light soaking and high temperatures andhumidity is warranted, as is, for instance, outdoorfield testing in order to correlate acceleratedindoor testing to real working conditions. Acceleratedtesting procedures designed in the pastfor Si or organic PV are currently being appliedto PSCs. However, it is likely that some of thesetesting conditions are either too strict or too mildto give a current estimation of how the PCE willdegrade with time, and therefore they need to berevisited by the community. Regardless, longtermstability is still one of the key issues thatimpedes rapid commercialization of PSCs anddraws skepticism from established solar cell technologies.To become a player in the power market,PSCs must be able to last for at least 20 years withminimal degradation; in order to do so, both intrinsicand extrinsic degradation in the perovskitedevice stack must be offset.