Friday morning saw an Area 3 session on “Strategies for III-V Cost Reductions,” an important discussion considering several talks throughout PVSC49 on the need for high efficiency sustainable systems. In this session, Jacob Boyer from NREL described state of the art growth of AlInP as passivation layers for HVPE 1J III-V and 2J GaInP/GaAs solar cells with PCE’s as high as 28%. John Mangum, also from NREL, described recent work on high efficiency solar cells grown on spalled germanium without polishing, while Norman Jost from the Insituto de Energia Solar in Madrid highlighted the challenges and progress in the development of micro-CPV modules using conductive ink interconnections presenting > 90% yield on a 132-cell module. Phillip Jahelka from CalTech presented 23.5% GaAs solar cells based upon a low-cost non-vacuum processing technique using zinc diffusion for forming p-type layers with low resistivity and precious metal free n-type contacts.
In Friday morning’s Area 4 session on “Device Physics, Modeling, New/Enhanced Characterization Techniques” Professor Gergely Zimanyi and his collaborators from UC Davis and Arizona State University used molecular dynamics and machine learning to investigate why the open-circuit voltage of a-Si:H/c-Si heterojunction cells degrade over time. This work provided insights into why and how the defect density of these a-Si:H/c-Si interfaces increase over time, including calculations of the energetic barriers associated with the defect generation processes. Xin Chen from Lawrence Berkely National Laboratory presented on the use of supervised machine learning to automate the analysis of electroluminescence images of PV modules, including the detection of cracks and identification of inactive areas originating from the cracks. In this work, Xin and collaborators reported very high precision recall and F1 scores for their segmentation process, showing the promise of the technique. Erin Tonita from SUNLAB at the University of Ottawa presented a more accurate method to determine the spectral albedo for better determination of the energy yield of bifacial PV systems. It was suggested that this method outperforms the current IEC standard and is of critical importance given the bifacial PV currently being installed. The final presentation in this session from Yibo Zhang from the University of Toronto described the materials structure and electrical properties of oxide/c-Si(n) interfaces, including the identification of a hole inversion layer at the n-type c-Si surface.
The Area 6 session on Friday morning discussed “Alternative Halide Perovskite Materials & Solar Cells” including highlights from Tobias Abzieher from NREL who described the development of continuous sublimation of inorganic halide perovskites using vapor processing in three steps: mechano-chemical synthesis, flash sublimation using vapor processing and annealing affording uniform and high-quality films. Niara Wright from Duke University presented work on Ruddlesden-Popper films deposited by a novel pulsed laser evaporation technique, which produced randomly oriented gains which influenced the in-and out of plane conductivity. David Benedikt Ritzer from the Institute of Microstructure Technology in Germany described the optimization of transparency and color using different laser processing conditions and shapes for BIPV applications, and Ajay Singh from the University of Luxembourg presented recent work on lead-free tin-based perovskites using a physical vapor deposition process that produces highly stable Sn4+ free perovskite films.
Friday morning’s Area 7 session on “Modeling and Predicting Energy Yield” saw highlights from Timofey Golubev from ThermoAnalytics Inc, describing the use of artificial neural networks to replace physics-based simulations that enables rapid predictions of energy yield area geographic locations and seasonal variations, and a presentation from Mandy Lewis from the University of Ottawa discussing how shading is localized along racking elements on the outside of rear side bifocal modules and increases non-uniformity, while reflective racking reduces and counteracts shading and non-uniformity. Annie Russell from SUNLAB also described deterministic and independent ways to model mismatch in rear irradiance in bifacial modules that combined with pyranometer data reduces the mean absolute error.
Area 7’s mid-morning session on “Module Materials, Design, and Manufacturing” included presentations from Taizo Masuda from Toyota Motor Corporation describing recent work on a demonstration Toyota Prius equipped with 860 W solar panels, which showed 40 % of the up to 6000 km trip assessed could be self-sufficient and provide a 62 % reduction in CO2, with a driving range strongly influenced by colling and heating. Farhan Rahman from Sandia presented thermal and solid-mechanical modelling of lamination process induced stress in glass-glass and glassbacksheet PV modules that indicated glass-glass configurations have the highest stress of the two configurations, which is formed and/or induced predominantly during lamination processes. Min Ju Yun from the Korea Electrotechnology Research Institute presented work on flexible, tessellated PV modules that are made by combining encapsulated ribbons, giving a stable, reliable module suitable for arbitrarily shaped surfaces.
Friday’s mid-morning session in Area 9 entitled: “Power Electronics and Grid Integration” included highlights from the University of Campinas describing Volt-Var distribution systems with the incorporation of dynamic programming techniques, and work from Sijo Augustine from New Mexico State University discussing a PV storage module integration with a novel control strategy that uses an ZnMnO2 battery technology, which provides benefits over other technologies when considering extreme solar applications. Also discussed in this session was work from the Paderborn University presented by Marius Claus Möller, which tests different high power PV modules with microinverters to compare their performance with a curve-fitting approach.
The final session in Area 10 on Friday morning on “Solar Resource and PV Forecasting” described how solar energy generation is impacted by atmospheric factors such as spectral variations or clouds. The session saw presentations on latest development in methods to include such atmospheric effects in PV performance analysis. Highlights included a presentation from Sophie Pelland from CanmentEnergy who described a study on the comparison of different methods to account for spectral effects on various solar cell technologies. Dr. Pelland suggested that overall spectra derived from different sources agrees, though some sources such as the NSRDB, show biases. Mark Mikofski from DNV presented on the effects of sampling frequency on power predictions. A time-averaged hourly input results in overestimations due to clipping, while Joseph Ranalli from Pen State Hazleton discussed the analysis of cloud advection and how lagging cross correlation improves the ability to discriminate between cloud advection and chaotic movement. Daniel Riley from the University of Alaska Fairbanks presented recent work on albedo measurements and modelling of PV systems in Alaska. A result of this study is an improvement of albedo values given for snowy grounds.
The final oral session of the PVSC 49 in Area 11 on “Large-Scale Deployment” included highlights presented by Arnulf Jager Waldau from the European Commission, Joint Research Centre discussing a delay in passing the European Green Deal, which calls for accelerated greenhouse gas reductions by 2030, and from Billy Stanbery from HelioSource Tech. who discussed how next generation and future disruptive PV technologies provided a 1 to 2 trillion USD dollar market opportunity. Moreover, that manufacturing cost savings of such disruptive technologies could amount to hundreds of $ compared to single junction modules.