The remarkable development in photovoltaic (PV) technologies, including materials, cells and modules, over the past five years call for renewed assessments with an eye towards their future progress. We do not restrict such assessments to solar to electrical power conversion efficiencies (PCEs), but also consider many of the factors that affect power output for each cell type. Where appropriate, we note improvements in control over materials and interfaces, and discovery of new properties in materials. The PCE of "champion cells" for all types of PV technology has improved over the past half decade. We analyse and discuss the remarkable progress in cells and modules, based on single crystal -Si, GaInP and InP, and on thin (polycrystalline) films of , esp. CdTe and Cu(In,Ga)Se 2 (=CIGS). In addition, we analyse PV developments of the more recently emerged lead halide perovskites, together with notable improvements in sustainable chalcogenides, organics and quantum dots. By comparing PV cell parameters across technologies, we can appraise how far each technology may progress in the near future, because, even though accurate or revolutionary developments cannot be predicted, often cross-fertilization occurs, making achievements in one cell type an indicator of evolutionary developments in others. This is extremely relevant in the present time, since the common theme of metal halide perovskites has helped to unite previously disparate, technology-focussed strands of PV research. Introduction:Undoubtedly, sunlight is the most abundant, safe and clean energy source for sustainable economic growth. One of the efficient and practical ways to use the sunlight as an energy source is to convert it to electricity using solar cells. An upper limit for light to electrical power conversion efficiency, PCE, by a single junction solar cell (i.e., solar photon energy  electrical energy) is given by the Shockley-Queisser (S-Q) model and formalism 1 . In this formalism there are assumptions, which postulate that all photons with energies above the bandgap create free electrons and holes, with perfectly chargeselective contacts, thus yielding one electron per absorbed photon to the electrical current flow. The S-Q model also stipulates that all electron-hole recombination events, which occur when the solar cell is generating power, are the inverse process to light absorption and therefore radiative -i.e., they result in the re-emission of light. The S-Q limit is based purely on thermodynamic considerations and takes the optical absorption edge (E G ), the solar spectrum and the operating temperature of the solar cell as the only inputs for the PCE calculation. The efficiency of real-world single junction solar cells will always be below the S-Q limit as real material properties come into play, e.g., the absorption edge is not a step function, as assumed by the S-Q model, and real materials have defects, which will lead

The loss from halide-segregation in wide bandgap perovskite solar cells is quantified, revealing that the performance bottleneck currently is, in fact, non-radiative recombination.

Longevity has been a long-standing concern for hybrid perovskite photovoltaics. We demonstrate high-resilience positive-intrinsic-negative perovskite solar cells by incorporating a piperidinium-based ionic compound into the formamidinium-cesium lead-trihalide perovskite absorber. With the bandgap tuned to be well suited for perovskite-on-silicon tandem cells, this piperidinium additive enhances the open-circuit voltage and cell efficiency. This additive also retards compositional segregation into impurity phases and pinhole formation in the perovskite absorber layer during aggressive aging. Under full-spectrum simulated sunlight in ambient atmosphere, our unencapsulated and encapsulated cells retain 80 and 95% of their peak and post-burn-in efficiencies for 1010 and 1200 hours at 60° and 85°C, respectively. Our analysis reveals detailed degradation routes that contribute to the failure of aged cells.

Cs 2 AgBiBr 6 thin film preparation for characterization. The double-perovskite thin films studied in this work were all prepared through sequential vapour deposition. In a vacuumsealed chamber, AgBr (99% Fluka), BiBr 3 (≥98% Sigma Aldrich) and CsBr (99.9% Sigma Aldrich) were placed in separate crucibles and sequentially thermally evaporated onto the substrates. In particular, the standard procedure we optimized evaporated 90 nm of AgBr, 120 nm of BiBr 3 and 150 nm of CsBr to obtain 300 nm of Cs 2 AgBiBr 6 . This basic stack was repeated the necessary number of times to achieve the desired total film thickness. To achieve thicknesses that are not multiples of 300 nm (like the 750 nm reported in the text), we ran the last evaporation cycle depositing half of each precursor thickness, keeping always the same precursors ratio (1:1.3:1.6 AgBr:BiBr 3 :CsBr). After the deposition of the desired thickness, we annealed the samples on a hotplate in air at 250 ºC for 30 minutes. The post-deposition annealing temperature and time were optimised to deliver maximum solar cell performance.Solar cell preparation. FTO or ITO coated glasses were cleaned by sequential sonication in soap, water, acetone and isopropanol. After being dried with a N 2 gun, the substrates were further cleaned by O 2 plasma for 10 minutes. Titanium isopropoxide (140 µl in 1 ml of EtOH) was added to 1 ml of acidic EtOH (10 µl of HCl 2M in 1 ml EtOH), and deposited on the FTO substrates by spincoating at 2000 rpm for 45 sec with 2000 rpm/sec acceleration. Following this, the films were annealed at 150°C for 15 min and 500°C for 30 min. SnO 2 layers were prepared by spincoating at 3000 rpm (200 rpm/sec) for 30 sec of a solution of SnCl 4 ⋅5H 2 O in isopropanol (17.5 mg/ml) on top of the FTO or ITO coated glasses. The so-prepared films were annealed at 100°C for 10 min followed by an annealing at 180°C for 30 min. The SnO 2 and TiO 2 films were placed in the vacuum chamber, and the Cs 2 AgBiBr 6 film was deposited as previously presented. The hole transport material (Spiro-OMeTAD, Lumtec) was dissolved in chlorobenzene (85 mg/ml) and doped with 20 µl of LiTFSI (500 mg/ml in BuOH) and 30 µl of tert-butylpyridine. The solution was then deposited on the active layer by spincoating in air at 2000 rpm (2000 rpm/sec) for 45 sec. The devices were then left overnight in a desiccator in air atmosphere, and then completed by the evaporation of 100 nm silver contacts. All the