Forum Article pubs.acs.org/IC
Decomposition and Cell Failure Mechanisms in Lead Halide Perovskite Solar Cells Jinli Yang and Timothy L. Kelly* Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, Saskatchewan S7N 5C9, Canada ABSTRACT: Perovskite solar cells have experienced a remarkably rapid rise in power conversion efficiencies, with state-of-the-art devices now competing with multicrystalline silicon and thin-film cadmium telluride in terms of efficiency. Unfortunately, the lead halide perovskite absorbers suffer from a lack of chemical stability and decompose in response to a variety of environmental stimuli. In this Forum Article, we provide a brief overview of the decomposition mechanisms in lead halide perovskite thin films, as well as the processes contributing to cell failure in finished devices. We finish by briefly surveying recent efforts to extend the device lifetime. Ultimately, if perovskite solar cells can be made stable, they will be an exciting, highly complementary addition to existing photovoltaic technologies.
■
INTRODUCTION Ever since the commercialization of photovoltaic devices, relatively few technologies have been able to compete with cells based on crystalline silicon.1 State-of-the-art multicrystalline silicon modules are efficient, cost-effective, and available with 25 year warranties. However, despite these advantages, the energy payback time (the time required for a photovoltaic device to produce the electrical energy originally consumed in its manufacture) for silicon-based solar cells is estimated to be in excess of 2 years.2 As a result, substantial research efforts have been directed toward the development of alternative thinfilm technologies. In 2009, Kojima et al.3 reported the first example of solar cells based on lead halide perovskites [APbX3, where A = CH3NH3+, CH(NH2)2+, or Cs+ and X = I−, Br−, or Cl−]. Since then, perovskite solar cells have jumped to the vanguard of emerging photovoltaic technologies.4−9 A global effort to optimize deposition routes, introduce new cell components, and diversify device configurations10−16 has led to power conversion efficiencies (PCEs) that have increased from less than 4% in 20093 to over 22% in 2016.1 Combined with projected energy payback times that are on the order of months, not years,17 perovskite solar cells are one of the most promising new alternatives to silicon-based devices. In order to realize the full potential of this new photovoltaic technology, perovskite solar cells must compete with silicon not just in terms of efficiency but also in terms of the device lifetime. The standard is high: warranties on the order of tens of years will be required for perovskite modules that will be subjected to extremes of humidity, temperature, and insolation. Unfortunately, it is well-known that perovskite solar cells suffer from inherently poor stability.18−21 Relatively weak Pb−I bonds, combined with high ionic mobility,22,23 make these hybrid organic−inorganic perovskites susceptible to decomposition under a wide variety of environmental stimuli, © XXXX American Chemical Society
including heat, humidity, and light soaking. The problem is exacerbated when the perovskite is combined with other cell components in a working device; contact with components such as the hole-transport material, electron-transport layer, and metal electrode can result in chemical incompatibilities and premature cell failure. Therefore, if perovskite solar cells are to have a commercial future, these degradation and cell failure mechanisms must be systematically identified and eliminated. With this goal in mind, numerous degradation studies have now been carried out on perovskite solar cells. However, analysis of these data has been greatly complicated by a number of factors. Many of these studies have been done at “room temperature” and under “ambient” conditions; because temperature and relative humidity (RH) display dramatic geographical and seasonal variations, the results of these studies are often conflicting or uncertain. Furthermore, the susceptibility of perovskite solar cells to subtle variations in processing conditions has led to tremendous batch-to-batch variability in the cell performance, often even within a single laboratory. As a result, any study examining the effects of environmental conditions or processing methods on the device lifetime is inherently highly correlative; researchers rely on measurements that have been averaged over many devices in order to determine whether such variables are correlated with either improved lifetimes or premature cell failure. Unfortunately, such correlational studies are more familiar to researchers in the life sciences, who are often better equipped to handle questions of significance with analytical rigor.24 Finally, the wide variety of reported device architectures (e.g., mesoporous vs planar, Special Issue: Halide Perovskites Received: May 30, 2016
A
DOI: 10.1021/acs.inorgchem.6b01307 Inorg. Chem. XXXX, XXX, XXX−XXX
Forum Article
Inorganic Chemistry
devices showed a significant decline in performance after even short storage times in ambient air. In order to better elucidate these decomposition mechanisms, more rigorous environmental controls were required. Kamat and co-workers30 used a series of glycerol/water mixtures to effectively fix the RH inside a sample chamber; in this experiment, the equilibrium partial pressure of water is controlled by the mole fraction of water in the glycerol solution. Samples were analyzed by ex situ absorbance spectroscopy and powder X-ray diffraction (pXRD), revealing the presence of an intermediate phase in the decomposition pathway. Concurrently, our research group used a flow-through method to control the RH inside custom-built sample holders for both in situ absorption spectroscopy and in situ grazingincidence wide-angle X-ray scattering (GIWAXS).31 The in situ absorption measurements allowed the rate of perovskite decomposition to be measured as a function of the humidity, and the in situ GIWAXS data revealed the formation of the same intermediate phase as that observed by Christians et al. (Figure 2). This phase was tentatively identified as a dihydrate,
inverted vs conventional, inorganic vs organic interfacial layers, etc.) further complicates the picture,4−9 with many different device components being implicated in cell failure mechanisms. In the elucidation of perovskite decomposition and cell failure mechanisms, some of the most compelling reports have been those making use of in situ techniques. Rather than focusing on a device post-mortemanalyzing a dead device to try to determine which components failedin situ methodologies allow real-time monitoring of how thin films or devices respond to environmental factors such as heat, humidity, and light. Because these studies focus only on a single device, there is no question of batch-to-batch variability; because they provide continuous, real-time data, they can shed light on the early or transient stages of a degradation mechanism. In this Forum Article, we provide a brief overview of the decomposition mechanisms in lead halide perovskites, with particular emphasis on those studies making use of tight environmental controls and in situ methodologies. The origins of cell failure in perovskite solar cells are discussed and broken down into failure caused by structural changes or decomposition of the perovskite active layer, degradation of the interfacial layers, and electrode corrosion. We conclude by summarizing the recent state-of-the-art features in device architectures offering improved perovskite stability and increased device lifetimes.
■
DECOMPOSITION MECHANISMS IN PEROVSKITE THIN FILMS Humidity. One of the most obvious issues with the methylammonium lead iodide semiconductors commonly used in perovskite solar cells is a lack of stability with respect to water. Any researcher who has left thin films of CH3NH3PbI3 on the bench for more than a day has witnessed this decomposition process first-hand, as the bright-yellow color of PbI2 replaces the deep-brown color of the perovskite (Figure 1). Similarly, liquid water will destroy a perovskite film
Figure 2. (a) UV−vis spectra, acquired at 15 min intervals, of a CH3NH3PbI3 film exposed to flowing N2 gas with RH = 98 ± 2%. (b) Normalized absorbance at 410 nm as a function of time for perovskite films exposed to various relative humidities. Data at 50% and 20% RH were acquired once per 24 h. The temperature was 22.9 ± 0.5 °C for all measurements. Reprinted with permission from ref 31. Copyright 2015 American Chemical Society.
which had been structurally characterized in 1987;32 however, further work by Leguy et al.33 better identified it as a monohydrate phase, which was structurally characterized by Kanatzidis and co-workers.34 These studies reveal why the perovskite is so susceptible to moisture: hydrogen bonding between water molecules and the methylammonium cations leads to transformation of the 3D perovskite structure into 2D sheets separated by water molecules and, finally, into isolated PbI64− octahedra (Figure 3). The first step of this process is
Figure 1. CH3NH3PbI3 thin films on glass: (a) a freshly prepared film; (b) the same film after storage on the laboratory bench for 5 days.
in a matter of seconds, as the methylammonium iodide is rapidly leached out of the film.25,26 However, early efforts to elucidate the underlying mechanism of this humidity-induced decomposition were largely speculative; samples were frequently stored under “ambient” conditions, which vary significantly from lab-to-lab and from day-to-day. Additionally, humidity appeared to have a dual effect on CH3NH3PbI3 films: higher levels of humidity appeared to benefit the initial growth of perovskite thin films15,27,28 yet ultimately led to more rapid CH3NH3PbI3 decomposition.29 As a result, unencapsulated
Figure 3. Tetragonal structure of methylammonium lead iodide and its associated hydrate phases. Structures were visualized using VESTA.37 Decomposition is thought to proceed via hydration to CH3NH3PbI3· H2O, followed by the formation of (CH3NH3)4PbI6·2H2O and PbI2.30,31,33 B
DOI: 10.1021/acs.inorgchem.6b01307 Inorg. Chem. XXXX, XXX, XXX−XXX
Forum Article
Inorganic Chemistry reversible, making the identification of CH3NH3PbI3·H2O by pXRD elusive; either fast measurements (e.g., minimal sample handling, fast scan rates, etc.)30,33 or in situ methods31,35 are required to characterize the monohydrate before it reverts back to the perovskite under ambient conditions. Additional X-ray photoelectron spectroscopy (XPS) studies have suggested that hydration of CH3NH3PbI3 is followed by the loss of CH3NH2 and HI and the eventual formation of various amorphous leadcontaining byproducts (e.g., PbCO3, Pb(OH)2, and PbO).36 These studies also evaluated the role of oxygen in the decomposition of CH3NH3PbI3 thin films.30,31 The observation was that oxygen appeared to play little role in the decomposition process, with the results being similar in both air and nitrogen atmospheres. Other studies have examined the effect of other solvents and reagents, with similar results. Zhao and Zhu found that ammonia vapor would rapidly (but reversibly) decompose CH3NH3PbI3 films to a colorless phase.38 Although no degradation mechanism was proposed, given the number of other solvate phases that are now known,34 it would not be surprising if a similar mechanism were active here. Light. Initial studies suggested that perovskite solar cells could last several weeks without encapsulation if kept at a RH of 200 °C, with decomposition complete by 250 °C.48 Yet, despite this apparent intrinsic stability, Snaith and co-workers have, nonetheless, observed substantial decomposition occurring at temperatures
