Present Status and Future Prospects of Perovskite Photovoltaics

comment Present status and future prospects of perovskite photovoltaics Solar cells based on metal halide perovskites continue to approach their theoretical performance limits thanks to worldwide research eforts. Mastering the materials properties and addressing stability may allow this technology to bring profound transformations to the electric power generation industry. Henry J. Snaith

etal halide perovskites have properties that are more challenging to of recombination occurs via non-radiative fascinated the research community control for long-term stability7. However, channels. The external radiative efficiency Mover the past six years, largely if progress in materials and device (ERE) of a PV cell — which is the fraction due to a combination of their high-quality engineering will enable long-term PV of total dark recombination current that optoelectronic properties — unexpected for module deployment, then such improved results in the emission of light — is a good materials processed from solution and at low operational stability, combined with the ease indicator of how close to perfection the cell temperature — and the apparent ease with of formation of high-quality materials and is. Single-digit ERE numbers, such as the which they can be employed as absorber devices, will place perovskites at the ideal ~2% value that can be estimated for the layers in highly efficient photovoltaic (PV) spot to trigger profound transformation in record 26.7% PCE silicon cells9,10, are already devices1. Perovskites typically adopt a 3D the PV market. very good, and the even higher ERE of ~22% crystal structure composed of three primary In this Comment, the present status reached by 28% PCE GaAs cells9 suggests

ions with a stoichiometry of ABX3. For the in research and directions where further that photon recycling (multiple reabsorption all-inorganic compounds, caesium is the research effort is required will be discussed, and reconversion in free charges of photons A-cation, and either lead or tin are the but not comprehensively reviewed, emitted by charge recombination in the B cation, with halides of chlorine, bromine or along with the status and challenges for active layer) starts to contribute to both iodine as the X anion. The ‘hybrid’ perovskite technology transfer and commercialization charge transport and voltage gain (Fig. 1b). compounds, which have shown most of perovskite PVs. In general, high ERE values correspond to

promise so far, have either methylammonium open-circuit voltages Voc approaching the (MA) or formamidinium (FA) as the A Status in research theoretical limits. cation. Some of the most efficient and stable Researchers have demonstrated how to The high luminescence efficiency of compounds even have multiple mixtures of routinely obtain perovskite solar cells with metal halide perovskites was recognized ions at the A and X sites2–5. The rich variety efficiency beyond 20%, through changes early on11. At present, the best perovskite of compositions and the multitude of ways in in materials composition, processing solar cells have an ERE of 1–4%3, and which they can be investigated have fostered conditions and device architectures. photon recycling has been suggested to burgeoning interest in perovskites and Reaching the maximum theoretical occur in these devices12. Another closely related compounds. performance in single- and multi-junction related indicator of loss in a solar cell is the Since the first reports in 2012, metal solar cells and demonstrating stability over ‘operational loss’, which is defined as the halide perovskites were well on their way 25 years require further understanding of difference between the optical band gap to deliver high PV performance1, and now the optoelectronic properties of metal halide of the absorber (in eV) and the maximum single-junction cells are already approaching perovskite devices and of how they can be power point voltage when the cell is power conversion efficiency (PCE) of 23%6. controlled with further chemistry and device operating under sunlight13. In Fig. 1c, the Yet, beyond high efficiency, other critical architecture optimization. operational loss is shown for a number factors are required for a PV ‘concept’ to of different PV technologies, including transition from the laboratory to real-world Solar cells with high external radiative perovskite solar cells; PCE improvements deployment and use, with ambition to efficiency. Miller, Yablonovitch and Kurtz8 have corresponded to marked reduction reach TW-scale production capacity. These popularized the notion that a good solar in loss for this technology. The high ERE include low materials and manufacturing cell must also be a good light-emitting and reduced operational loss indicate that

costs, high stability and independence from diode. When all non-essential charge Voc can reach high values in perovskite rare elements. The major components in recombination paths are eliminated, such devices; however, the PCE is still lower than the perovskite compounds, lead and iodine, as trap-assisted recombination at defect or that of silicon solar cells with comparable are abundant materials, and low cost impurity sites, charges present in the active ERE because other device parameters appears to be preordained. Demonstrating material can only recombine through radiation determining the PV efficiency, such as long-term stability, however, has raised of light. Thus, bright electroluminescence the fill factor and the short-circuit current major concerns5. from a PV cell operating in forward bias density, are affected by resistive losses and In general, there is a sense that the (Fig. 1a), typically measured when the dark optical losses, respectively13. Improving the properties that make perovskites such recombination current equals the short- optical design of the cells is important to remarkable semiconductors, which are yet circuit photocurrent under sunlight, is a move towards and beyond 25% efficiency. to be fully understood, may also be the signature that a smaller and smaller fraction It is also noted that the external radiative

ab c 1.2 SQ — limit loss

1.0 a-Si 9.7% 0.8 14.1% 10.9% CIGS CdTe 17.9% 0.6 20.1% – e 22.1% GaInP 0.4 c-Si GaAs

h+ Operational loss (eV) 0.2

1.01.2 1.4 1.6 1.82.0 Absorption edge (eV)

Fig. 1 | Light emission and voltage losses in perovskite solar cells. a, Photograph of a perovskite PV device emitting light while under forward electrical bias. b, Illustration of photon recycling in a solar absorber layer. External incident light (yellow wavy arrows) generates charge carriers (e− and h+), which can recombine (black dots) to produce new photons (red wavy arrows) that can either be reabsorbed to generate new charge carriers, or are emitted outside the active layer (light-blue). c, Operational loss versus absorption edge of the solar absorber layer for a number of different PV technologies (red squares) and perovskite solar cells (green squares, with the corresponding PCE of the devices). Also shown is the minimum operational loss feasible when a solar cell reaches the Shockley–Queisser (SQ) limit. Reproduced from ref. 3, AAAS (a). Adapted from ref. 8, IEEE (b); and ref. 13, Wiley (c).

efficiency of isolated perovskite films, here intermediate crystalline precursor phase20,21; However, this also results in low measured as the fraction of emitted to however, there is no consensus yet as to what decomposition temperatures, which make absorbed light, is as high as 70%11, thus there are the needed ‘building blocks’ to obtain an these materials intrinsically less thermally is a possibility of achieving much higher ideal crystalline perovskite film. stable than silicon. The use of organic EREs in complete solar cells. Importantly, A-site cations, which can withstand much this indicates that when a perovskite Beyond single-junction solar cells. A more lower temperatures than the inorganic absorber layer is sandwiched between substantial efficiency boost will be obtained counterparts, is also an intrinsic limit; yet, charge-selective contacts in a PV device, by moving to advanced concepts beyond such cations, being larger than any inorganic further recombination losses are introduced. single-junction PV cells. These may include ion in the periodic table, are necessary to Although various methodologies — such multi-exciton generation, singlet fission, form a stable octahedral framework in the as molecular passivation of the perovskite hot-carrier collection and even intermediate 3D perovskite structure28. A few potential surface14 or reducing the interfacial contact band gap cells.22 Results suggesting that routes have emerged to stabilize these between the charge extraction layers some of these processes may occur in frameworks with inorganic ions29, but more and perovskite absorber15 — have shown perovskite films have been reported, work is required to confirm if these, or other progress in reducing such losses, our although these have not yet been translated approaches, can lead to long-term stable present understanding of recombination at into devices with higher PCE values. inorganic perovskites. these heterointerfaces is limited, and there A near-term means to move to higher Aging causes electronic degradation in remains more than 100 mV to be gained in efficiencies is to employ multi-junction silicon arising from the ‘un-passivation’ of 22,23 Voc, by minimizing these losses. concepts , in which perovskite films are dangling bonds on the surface or at grain As discussed above, several combined with silicon or other materials boundaries. For perovskites, the defect optoelectronic parameters indicate that the to expand the absorption spectral range chemistry and physics is still not well average quality of metal halide perovskites (and thus the amount of solar radiation understood, but the defects are thought to films is already high, which is surprising converted into electricity), and convert the have energies close to, or inside, the energy considering that these are usually deposited solar photons into electrical potential energy bands16,27. Therefore, defects emerging through solution processes and have high at a higher voltage24,25. The success of this during aging should be less problematic, densities of crystalline defects — this is approach hinges upon the possibility to tune as suggested by the high efficiency of the reason why these materials have been the optical bandgap of the perovskite layers, perovskite solar cells in the presence 16,17 labelled as defect tolerant . Yet, it is ideally from 2 eV to below 1.2 eV, depending of a relatively large fraction of PbI2 (a expected that minimization of electronic on the absorption properties of the materials degradation product) within the film16. defects through materials improvements they are combined with26. Although bandgap Through substitution of ions and will be important to progress towards tuning has been demonstrated by changing improved contact materials, the present the thermodynamic efficiency limits. An the chemical composition or leveraging generation of perovskite solar cells are important aspect relates to the chemical quantum confinement, at present only cells vastly more stable, and with appropriate processes that occur when depositing with a gap in the 1.5–1.7 eV range deliver encapsulation they may already be close to perovskite films. Although the solutions appropriately high efficiency and stability. requirements for real world deployment2,4, used obviously contain A, B and X which are specified in the International precursors, these ions are incorporated into Stability-related issues. These perovskite Electrotechnical Commission (IEC) 61215 many forms such as colloids, dissolved ions compounds are largely held together standard. However, in order to make the and complexes18; and the solutions also through ionic bonding, and this is likely technology profitable, it is important to contain additives, whose role in the chemical to be one of the factors central to both the have a good level of certainty that the reactions and quality of the synthesized high tolerance to crystalline defects, and product will last 25 years. This requires perovskites needs further understanding19. the ease at which highly crystalline films going significantly beyond the IEC61215 Often crystallization proceeds through an can be fabricated at low temperature27. standard, acquiring more detailed

a b with this growing market. In this respect, 108 lead-based perovskites sit in a remarkably 1.5 ) 7 Cu good position. Figure 2a shows the current –1 10 Si y Pb c-Si annual global production capacity of 1.2 106 elements used in existing PV technologies GaAs 36 1 hour and perovskites . In order to reach 12.5 TW 5 10 35 years (2050) 0.9 PV electricity generation (the dots on the As CIGS Cd 4 1 day graph), GaAs PV cells would require 500 10 1,000 years 0.6 years of today’s gallium production capacity, Se CdTe 3 10 In thin film CdTe PV modules would require Te Ga Ga Perovskite 0.3 1,000 years of today’s tellurium, and thin Current annual production (t 102 film copper indium gallium selenide (CIGS) 1 year Normalized environmental impact s current production would require 400 years of today’s indium. 101 0 101 102 103 104 105 106 107 108 The only commercial PV technology that Si/PK PK/PK can scale to this level is silicon, which for Material required (t) CIGS/PK CZTS/PK 12.5 TW PV electricity generation capacity would require around three years of our Fig. 2 | Material production capacity and environmental impact. a, Annual material production in 2015 current silicon production. In stark contrast, versus material required to deliver 5% (left end of the bar), 50% (circle) and 100% (right end of the bar) perovskites would only take a few days of of a potential 25 TW PV installation for perovskite, silicon, GaAs, CIGS (copper indium gallium selenide) current lead production capacity to scale to 36 and CdTe. Adapted from ref. , Royal Society of Chemistry. b, Normalized environmental impact for the TW level. the manufacturing of different tandem solar cells, where Si/PK = perovskite-on-silicon tandem, Another critical implication of this CIGS/PK = perovskite-on-CIGS tandem, CZTS/PK = perovskite-on-copper zinc tin sulfide tandem, PK/PK = analysis is that we would not have to perovskite-on-perovskite tandem, CZTS = copper zinc tin sulfide, and 1 is the environmental impact for significantly increase our global lead 36 the production of single junction Si solar cells. Adapted from ref. , Royal Society of Chemistry (a); and production, and hence consumption, 37 ref. , Royal Society of Chemistry (b). beyond todays levels, if we moved to 100% lead-based perovskite PV power generation, thus containing the environmental impact understanding of potential failure modes may also enable both the study of new of lead-based perovskite PV modules. that could exist. physics and a better control of perovskite Yet such impact requires more complete surfaces and electronic interfaces4,35. investigation to ensure the sustainability

Materials beyond the ABX3 Pb-halide of this technology. As part of the activity perovskites. Since the advent of efficient Technology transfer in perovskite PV within the European project CHEOPS, metal halide PV cells, the prospects of Although addressing the open research SmartGreenScans (https://go.nature. discovering new related compounds has questions described above will help com/2ul5Ggs) have undertaken a redirected new materials discovery30,31. clarifying the technological potential of rigorous life-cycle assessment (LCA) of This is also motivated by the desire to perovskite solar cells, there are challenges perovskite-on-silicon tandem cell and replace lead, due to toxicology concerns. remaining towards market deployment. module production using the pilot facility Partial replacement of lead with tin is Considerations about the availability of in Germany of Oxford Photovoltaics Ltd currently the only route to reduce the materials and their environmental and (https://www.oxfordpv.com/) as a baseline band gap to significantly below 1.5eV, toxicity risks confirm the role of perovskite proxy for the manufacturing process. which is essential for multi-junction cells. PV as strong contenders for large-scale The first results show that emitted lead Although efficient tin and lead–tin PV cells electricity generation. Start-up companies contributes only 0.27% or less to the total have been realized24, the stability of the are in the process of delivering the first freshwater ecotoxicity and human toxicity neat tin-based cells is still extremely low. commercial products, yet more financial (https://go.nature.com/2G5Hp3E). The Other lead-free compounds have also been support to research and stronger interest majority of the negative environmental tested; however, thus far they have not yet from investors is required to overcome or health impacts are related to the delivered PV cells with efficiencies beyond barriers to commercialization. monocrystalline silicon production, with 3%31–33. This suggests that there is something indium also significantly contributing. extraordinarily unique about the group Materials availability and toxicity. Over Similar LCA studies compared the total IV metal halide perovskites. However, it is the past few years, the cost of mainstream environmental impact for the manufacturing still likely that new perovskite families do PV power generation has dropped below of different types of perovskite-based exist, but discovering them is much more the cost of producing electricity from tandem solar cells37. Consistently, as challenging than first anticipated. coal and gas in the sunniest locations in shown in Fig. 2b, silicon dominates the A separate, successful case is the the world. As efficiency increases further total environmental impact, which also extension of perovskites to layered the cost of PV electricity will continue encourages the development of perovskite- structures using larger organic cations. This to fall. It is therefore feasible that PV will on-perovskite tandem cells. has resulted in successful realization of eventually become our primary source of Ruddlesden–Popper phase perovskites, and global power generation, which means Patent landscape and commercial activity. also films with mixed compositions of 2D that the volume and value of today’s PV As perovskite PV research advances, and 3D phases. These approaches can result market is only a fraction of the size it will growing commercial activity is emerging. in both improved long-term stability and become over the next few decades. This This is illustrated in a recent review of the efficiency4,34. The prospect of creating stable implies that raw resources must not limit patent literature. There are over 2,000 filed heterojunctions between 2D and 3D phases the ambition of any PV technology to scale and 300 granted patents in perovskite PVs38.

a Top assignees Some examples of the larger initiatives in perovskite solar patents Perovskite PV research include the US Office Oxford Photovoltaics Ltd of Naval Research (ONR) Air Force Office Sekisui Chemical Co. Ltd of Scientific Research (AFOSR) programmes Fujifilm Corp. on organic PVs (https://go.nature. Hunt Energy Enterprises LLC Huazhong University of Science and Technology com/2GchMtG), which have refocused Korea Research Institute of Chemical Technology much of their funding towards perovskite- IG Chem Ltd based PVs, and the US SunShot initiative Bohai University (https://go.nature.com/2DVzKyC), which op assignees T Ecoale Polytechnique Fedaeral de Lausanne funds a number of perovskite PV projects. Tianjin Vocational Institute There is also significant activity in Asia, Institute of Physics Chinese Academy of Sciences Okinawa Institute of Science and Technology School such as the support provided by the Japanese Chengdu New Keli Chemical Science Co. Ltd New Energy and Industrial Technologies Commonwealth Scientific and Industrial Research Organisation Development Organisation (NEDO) (https://go.nature.com/2G9n3Xa). In the 0 20 40 60 80 100120 Published patents UK, the total funding of Solar Technologies by the engineering and physical sciences bc research council (EPSRC) amounts to approximately £32 million (which is 0.66% of the EPSRCs portfolio), of which approximately half is related to perovskite solar cell research (https://go.nature. com/2G5LxAy). From the European Union, there is presently one running, and four commencing medium-scale collaborative projects focussed on perovskite PVs, which amounts to around 25 million euros, in addition to around 20 individual grants from the Marie Curie Fellowship, or European Fig. 3 | Patents and PV technologies under development. a, List of top 14 assignees of perovskite solar Research Council, amounting to another cell patents. Figure reproduced from ref. 38. b,c, Photographs of: a 156 × 156 mm2 pseudo-square wafer ten million euros. If we compare this to perovskite-on-silicon tandem cell (b), courtesy of Oxford Photovoltaics Ltd; and a 160 cm2 flexible energy storage, in UK EPSRC provides perovskite module processed via roll-to-roll coating (c), courtesy of Soliance Solar Research. around £110 million in funding, including the recently launched Faraday Institution, which accounts for £55 million. The storage A chart of the number of patents assigned, facility in Germany, which was previously activities are presently mainly focussing on per top assignees is shown in Fig. 3a. Of a CIGS pilot production facility. Other battery storage for the automotive industry, the top ten assignees of perovskite patents, companies, such as Saule Technologies with a £246 million commitment for the six are commercial entities (a mixture of (http://sauletech.com), and the Soliance Faraday Challenge. On nuclear fusion start-ups and large corporations) with the consortium (https://solliance.eu/), which research, the UK has committed £86 million remainder being universities or research includes Greatcell Solar (www.greatcellsolar. for advancing national research (http:// institutes. Figure 3a also illustrates the global com) and Panasonic (www.panasonic. www.ccfe.ac.uk/), and on a European scale nature of perovskite PV commercialization, com), are developing lightweight, flexible the investment is two orders of magnitude with top assignees distributed between and potentially semi-transparent perovskite larger, with the total ITER project (https:// Europe, USA, Japan, China, Korea modules (Fig. 3c), for a broad range of www.iter.org/) estimated to cost in excess of and Australia. applications including building-integrated 14 billion euros. There are diverse technical directions PVs, the automotive industry and low-light In comparison to the scale of the being pursued to commercialize perovskite indoor applications. fusion project, solar PVs, and specifically solar technologies, with some companies perovskites, are receiving less than 1% of the being more transparent than others. Oxford Support for research and development. investment in research and development. Photovoltaics Ltd is targeting perovskite- Major initiatives to support perovskite PV However, there exists a tangible prospect for on-silicon tandem solar cells (Fig. 3b). The research and development are presently perovskite PV to contribute to the radical rationale is that this technology capitalises lacking in ambition and scale. This may transformation of the US$100 billion PV upon the existing well-developed PV value in part be due to the relatively short time industry and the multi-trillion global power chain, and through collaborating with passed since perovskite PV has shown industry in the near to medium term. The existing silicon PV manufacturers, should such promise, and an inherent inertia and opportunity to capture future value from enable the rapid scaling to large volume time lag in strategic investment decisions this industry — by individual countries manufacturing and address the largest in research. Although over 3,000 academic and multinational initiatives investing in markets. However, there is no hiding from journal publications (in 2017 alone) do advancing research in perovskite PV — has the key requirement to deliver 25 years suggest that many researchers are working not yet been realized. Despite the fact that stability, along with very high efficiency on perovskite PVs and the related areas, real first-generation products may enter and high yield in manufacturing. Oxford the funding of this research is through the market over the next few years, there Photovoltaics Ltd are presently scaling up a collection of small grants, rather than remains the opportunity for major further towards pilot-scale manufacturing in a large ambitious multilateral projects. innovation over the next few decades.

It is also important to note that Outlook 17. Yin, W.-J., Shi, T. & Yan, Y. Unusual defect physics in CH3NH3PbI3 perovskite solar cell absorber. Appl. Phys. Lett. 104, government investment into research There has been tremendous research 063903 (2014). should not only be made into pre- activity in metal halide perovskites for 18. Yan, K. et al. Hybrid halide perovskite solar cell precursors: commercial technologies, but also into PV applications over the past six years, colloidal chemistry and coordination engineering behind device processing for high efciency. J. Am. Chem. Soc. 137, technologies that support growing industry. which has resulted in many discoveries 4460–4468 (2015). In Europe, research projects supporting and advances. Although expanding upon 19. Noel, N. K. et al. Unveiling the Infuence of pH on the silicon PVs are unlikely to be funded in the lead halide material set has proven crystallization of hybrid perovskites, delivering low voltage loss the next rounds. However, there remain photovoltaics. Joule 1, 328–343 (2017). challenging, improving the quality and 20. Moore, D. T. et al. Crystallization kinetics of organic–inorganic many prospects to innovate even with stability of the lead halide perovskites trihalide perovskites and the role of the lead anion in crystal this well-established silicon technology; and devices has brought the technology growth. J. Am. Chem. Soc. 137, 2350–2358 (2015). by also investing in schemes that make tangibly close to commercial readiness. 21. Wang, F., Yu, H., Xu, H. & Zhao, N. HPbI3: A New Precursor Compound for Highly Efcient Solution-Processed Perovskite the environment economically favourable There is now growing industrial momentum Solar Cells. Adv. Functional Mater. 25, 1120–1126 (2015). for commercialization, value would be and it is likely that real PV applications 22. Green, M. A. & Bremner, S. P. Energy conversion approaches pulled back to the countries that make the employing metal halide perovskites will be and materials for high-efciency photovoltaics. Nat. Mater. 16, 23 (2016). investment in research and development. demonstrated over the next few years, which 23. Green, M. A. Tird generation photovoltaics: solar cells for 2020 will represent a second phase, or era, for this and beyond. Physica E Low Dimens. Syst. Nanostruct. 14, Barriers to commercialization. There research community. ❐ 65–70 (2002). are hurdles to overcome on the path to 24. Eperon, G. E., Horantner, M. T. & Snaith, H. J. Metal halide perovskite tandem and multiple-junction photovoltaics. Nat. Rev. commercial success. The most significant Henry J. Snaith Chem. 1, 0095 (2017). ones are probably not technical, but relate Department of Physics, University of Oxford, 25. Lal, N. N. et al. Perovskite tandem solar cells. Adv. Energy Mater. to practicalities of business, where a key Oxford, UK. https://doi.org/10.1002/aenm.201602761 (2017). 26. Horantner, M. T. et al. Te potential of multijunction perovskite issue is raising finance. Scaling up from very e-mail: [email protected] solar cells. ACS Energy Lett. 2, 2506–2513 (2017). promising research results to a technology 27. Kim, J., Lee, S.-H., Lee, J. H. & Hong, K.-H. Te role of ready for manufacturing requires an order Published online: 23 April 2018 intrinsic defects in methylammonium lead iodide perovskite. of magnitude increase in investment. https://doi.org/10.1038/s41563-018-0071-z J. Phys. Chem. Lett. 5, 1312–1317 (2014). 28. Filip, M. R., Eperon, G. E., Snaith, H. J. & Giustino, F. Steric Many venture firms, especially from the engineering of metal-halide perovskites with tunable optical band USA, remain cautious of investing in PV References gaps. Nat. Commun. 5, 5757 (2014). technology due to the previous surge in PV 1. Green, M. A., Ho-Baillie, A. & Snaith, H. J. Te emergence of 29. Swarnkar, A. et al. Quantum dot–induced phase stabilization of perovskite solar cells. Nat. Photon. 8, 506 (2014). α-CsPbI3 perovskite for high-efciency photovoltaics. Science 354, investment in the mid 2000s, where the likes 2. Bush, K. A. et al. 23.6%-efcient monolithic perovskite/silicon 92–95 (2016). of MiaSole, Nanosolar and Solyndra either tandem solar cells with improved stability. Nat. Energy 2, 30. Brandt, R. E. et al. Searching for “defect-tolerant” photovoltaic filed for bankruptcy or sold for a fraction 17009 (2017). materials: combined theoretical and experimental screening. 39 3. Saliba, M. et al. Incorporation of rubidium cations into perovskite Chem. Mater. 29, 4667–4674 (2017). of the investment that had attracted . solar cells improves photovoltaic performance. Science 354, 31. Hoye, R. L. Z. et al. Perovskite-inspired photovoltaic materials: However, it is the opinion of the author that 206–209 (2016). toward best practices in materials characterization and investors should be more optimistic today, 4. Wang, Z. P. et al. Efcient ambient-air-stable solar cells with 2D– calculations. Chem. Mater. 29, 1964–1988 (2017). as the PV industry is now in a completely 3D heterostructured butylammonium-caesium-formamidinium 32. Giustino, F. & Snaith, H. J. Toward lead-free perovskite solar cells. lead halide perovskites. Nat. Energy 2, 17135 (2017). ACS Energy Lett. 1, 1233–1240 (2016). different position as compared to the 5. Asghar, M. I., Zhang, J., Wang, H. & Lund, P. D. Device stability of 33. Greul, E., Petrus, M., Binek, A., Docampo, P. & Bein, T. Highly period between 2008 and 2012. The cost of perovskite solar cells — a review. Renew. Sustain. Energy Rev. 77, stable, phase pure Cs2AgBiBr6 double perovskite thin flms electricity from PV is now cheaper than any 131–146 (2017). for optoelectronic applications. J. Mater. Chem. A 5, 6. Green, M. A. & Ho-Baillie, A. Perovskite solar cells: the birth of a 19972–19981 (2017). other source in many places in the world, new era in photovoltaics. ACS Energy Lett. 2, 822–830 (2017). 34. Quan, L. N. et al. Ligand-stabilized reduced-dimensionality and the downward trend in cost is set to 7. Manser, J. S., Saidaminov, M. I., Christians, J. A., Bakr, O. M. & perovskites. J. Am. Chem. Soc. 138, 2649–2655 (2016). continue over the next decade. Therefore, Kamat, P. V. Making and breaking of lead halide perovskites. 35. Hu, Y. et al. Hybrid perovskite/perovskite heterojunction solar Acc. Chem. Res. 49, 330–338 (2016). cells. ACS Nano 10, 5999–6007 (2016). PV is set to become a major component 8. Miller, O. D., Yablonovitch, E. & Kurtz, S. R. Strong Internal and 36. Jean, J., Brown, P. R., Jafe, R. L., Buonassisi, T. & Bulovic, of global electricity generation and hence External Luminescence as Solar Cells Approach the Shockley– V. Pathways for solar photovoltaics. Energy Environ. Sci. 8, the market scale is assured. This makes Queisser Limit. IEEE J. Photovoltaics 2, 303–311 (2012). 1200–1219 (2015). strategic investment feasible; however, the 9. Green, M. A. Radiative efciency of state-of-the-art photovoltaic 37. Celik, I. et al. Environmental analysis of perovskites and other cells. Prog. Photovoltaics Res. Appl. 20, 472–476 (2012). relevant solar cell technologies in a tandem confguration. Energy largest investments in PV at present go into 10. Yoshikawa, K. et al. Silicon heterojunction solar cell with Environ. Sci. 10, 1874–1884 (2017). deployment and investment in solar farms, interdigitated back contacts for a photoconversion efciency over 38. Perovskite photovoltaic: A review of the patent landscape. which have predictable rates of return, 26%. Nat. Energy 2, 17032 (2017). Cintelliq https://go.nature.com/2IGsIR9 (2018). 11. Deschler, F. et al. High photoluminescence efciency and optically 39. Wesof, E. Rest in peace: the list of deceased solar companies. rather than investments into development pumped lasing in solution-processed mixed halide perovskite GreenTechMedia (6 April 2013). of new technology where the risk is much semiconductors. J. Phys. Chem. Lett. 5, 1421–1426 (2014). 12. Pazos-Outón, L. M. et al. Photon recycling in lead iodide higher. Another non-technical issue, which perovskite solar cells. Science 351, 1430–1433 (2016). Acknowledgements also affects the ability to raise investment, 13. Nayak, P. K. & Cahen, D. Updated assessment of possibilities and H.S. is funded by the EPSRC, UK and the European is how to develop an appropriate business limits for solar cells. Adv. Mater. 26, 1622–1628 (2014). Union’s Horizon 2020 framework programme for research model, which will create value from the PV 14. Noel, N. K. et al. Enhanced photoluminescence and solar cell and innovation under grant agreement no. 653296 of performance via lewis base passivation of organic inorganic lead the CHEOPS project. H.S. thanks P. Nayak for providing industry, in the event of technical success. halide perovskites. ACS Nano 8, 9815–9821 (2014). adaptations to Fig. 1c. Existing mainstream PV manufacturers do 15. Wolf, C. M. et al. Reduced interface-mediated recombination not make a reliable profit margin; hence, new for high open-circuit voltages in CH3NH3PbI3 solar cells. Adv. Mater. 29, 1700159 (2017). Competing interests approaches to creating value, or a change in 16. Steirer, K. X. et al. Defect tolerance in methylammonium lead H.S. is co-founder and Chief Scientific Officer of Oxford the structure of the industry, have to emerge. triiodide perovskite. ACS Energy Lett. 1, 360–366 (2016). Photovoltaics Ltd.