This title appears in the Scientific Report :
2023
Please use the identifier:
http://dx.doi.org/10.2139/ssrn.4309089 in citations.
A Fully Robotic Platform for Optimizing the High-Dimensional Processing Parameter Space of Perovskite Thin-Films
A Fully Robotic Platform for Optimizing the High-Dimensional Processing Parameter Space of Perovskite Thin-Films
We report the design and utilization of a fully automated platform called SPINBOT for the engineering of solution-processed functional thin films. The SPINBOT is capable of performing experiments with high sampling variability through the unsupervised processing of hundreds of substrates with except...
Saved in:
Personal Name(s): | Zhang, Jiyun (Corresponding author) |
---|---|
Liu, Bowen / Liu, Ziyi / Wu, Jianchang / Arnold, Simon / Shi, Hongyang / Osterrieder, Tobias / Hauch, Jens / Wu, Zhenni / Luo, Junsheng / Wagner, Jerrit / Berger, Christian / Stubhan, Tobias / Schmitt, Frederik / Zhang, Kaicheng / Sytnyk, Mykhailo / Heumueller, Thomas / Sutter-Fella, Carolin M. / Peters, Ian Marius / Zhao, Yicheng / Brabec, Christoph | |
Contributing Institute: |
Helmholtz-Institut Erlangen-Nürnberg Erneuerbare Energien; IEK-11 |
Published in: | SSRN eLibrary (2023) |
Imprint: |
[S.l.]
Social Science Electronic Publ.
2023
|
DOI: |
10.2139/ssrn.4309089 |
Document Type: |
Preprint |
Research Program: |
Materials and Interfaces |
Publikationsportal JuSER |
We report the design and utilization of a fully automated platform called SPINBOT for the engineering of solution-processed functional thin films. The SPINBOT is capable of performing experiments with high sampling variability through the unsupervised processing of hundreds of substrates with exceptional experimental control. We demonstrate the unique capabilities of this platform by optimizing metal-halide perovskite films in a complex combinatorial parameter space. After just 5 steps (61 parameter sets), we arrived at a film that, when processed into a solar cell in ambient atmosphere, immediately yields a champion power conversion efficiency (PCE) of 21% with satisfactory performance reproducibility and photo-thermal stability. Exceptional process control and reproducibility was demonstrated by repeating film processing under optimum conditions and achieving a narrow coefficient variance of 0.74% in photoluminescence intensity of perovskite films. Coupling the platform to an Artificial Intelligence Optimizer is the logical next step to create an autonomously operating platform with the capability for inverse design. The combination of fast but “physically rich” proxy measurements like spectral and imaging photoluminescence and very large data sets is key to run experiments at a rate compatible with the rapidly evolving needs of today´s materials science research community. |