CoChat

How machine learning can help select capping layers to suppress perovskite degradation

Open access

Published version of a preprint

The DOI matches its Crossref record; 66 of the 66 references it cites were found in a registry or an index.

Abstract

Environmental stability of perovskite solar cells (PSCs) has been improved by trial-and-error exploration of thin low-dimensional (LD) perovskite deposited on top of the perovskite absorber, called the capping layer. In this study, a machine-learning framework is presented to optimize this layer. We featurize 21 organic halide salts, apply them as capping layers onto methylammonium lead iodide (MAPbI 3 ) films, age them under accelerated conditions, and determine features governing stability using supervised machine learning and Shapley values. We find that organic molecules’ low number of hydrogen-bonding donors and small topological polar surface area correlate with increased MAPbI 3 film stability. The top performing organic halide, phenyltriethylammonium iodide (PTEAI), successfully extends the MAPbI 3 stability lifetime by 4 ± 2 times over bare MAPbI 3 and 1.3 ± 0.3 times over state-of-the-art octylammonium bromide (OABr). Through characterization, we find that this capping layer stabilizes the photoactive layer by changing the surface chemistry and suppressing methylammonium loss.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-020-17945-4

Cites

66 of 66 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 The workflow of the study and aging test result over time. a Schematic overview of this study, aimed at developing design rules for capping layer of perovskite solar cells. b the raw image changes for tetrapropylammonium iodide (TPAI)-capped, phenyltriethylammonium iodide (PTEAI)-capped, which have similar molecular weights, and bare MAPbI3 films.
Fig. 1 The workflow of the study and aging test result over time. a Schematic overview of this study, aimed at developing design rules for capping layer of perovskite solar cells. b the raw image changes for tetrapropylammonium iodide (TPAI)-capped, phenyltriethylammonium iodide (PTEAI)-capped, which have similar molecular weights, and bare MAPbI3 films.CC BY · PubMed Central
Fig. 2 Capping-layer precursors and accelerated aging results of best-performing capping-layer material. a A-site cations used in this study and their chemical structure: (1) formamidinium (FA), (2) guanidinium (GA), (3) ethylammonium (EA), (4) dimethylammonium (DMA), (5) iso-propylammonium (iPA), (6) imidazolium (ID), (7) t-butylammonium (tBA), (8) phenylammonium (PhA), (9) benzylammonium (BzA), (10) phenethylammonium (PEA), (11) n-octylammonium (OA), (12) dodecylammonium (DA), (13) tetrapropylammonium (TPA), (14) phenyltriethylammonium (PTEA), and (15) tetrabutylammonium (TBA). b The film structure with 3D MAPbI3 at the bottom and LD capping layer deposited atop. c The time-dependent red and green values of camera images, and d the camera images of pre-degraded and 500 min-degraded films for bare MAPbI3 control material, (11) OABr, used in state-of-the-art high-efficiency devices, and (14) PTEAI, our best-performing material in this study, with scale bar 0.5 cm.
Fig. 2 Capping-layer precursors and accelerated aging results of best-performing capping-layer material. a A-site cations used in this study and their chemical structure: (1) formamidinium (FA), (2) guanidinium (GA), (3) ethylammonium (EA), (4) dimethylammonium (DMA), (5) iso-propylammonium (iPA), (6) imidazolium (ID), (7) t-butylammonium (tBA), (8) phenylammonium (PhA), (9) benzylammonium (BzA), (10) phenethylammonium (PEA), (11) n-octylammonium (OA), (12) dodecylammonium (DA), (13) tetrapropylammonium (TPA), (14) phenyltriethylammonium (PTEA), and (15) tetrabutylammonium (TBA). b The film structure with 3D MAPbI3 at the bottom and LD capping layer deposited atop. c The time-dependent red and green values of camera images, and d the camera images of pre-degraded and 500 min-degraded films for bare MAPbI3 control material, (11) OABr, used in state-of-the-art high-efficiency devices, and (14) PTEAI, our best-performing material in this study, with scale bar 0.5 cm.CC BY · PubMed Central
Fig. 3 Image data parameter extraction, regression results, and Shapley-based importance rank. a Extracted red onset points (in minutes) and their distributions from time-dependent camera data of capping-layer groups and bare MAPbI3 controls, synthesized using 12 different processing conditions. The red line shows the median, and the box shows the upper and lower quartiles. b The cross-validated root mean square error (RMSE) of various machine-learning models, including linear regression, K-nearest neighbors (KNN), random forest regression, gradient boosting regression with decision trees, neural network (multilayer perceptron), and support vector regression (SVR), with non-normalized input. c The comparison of red onset prediction and observation output between linear regression and random forest regression, with the green circle indicating PTEAI-capped film. d The feature importance ranking obtained from the random forest regression algorithm and SHAP library, showing the chemical properties and processing conditions in descending order of importance (rank). The yellow and purple color indicates high and low values of a given feature, respectively.
Fig. 3 Image data parameter extraction, regression results, and Shapley-based importance rank. a Extracted red onset points (in minutes) and their distributions from time-dependent camera data of capping-layer groups and bare MAPbI3 controls, synthesized using 12 different processing conditions. The red line shows the median, and the box shows the upper and lower quartiles. b The cross-validated root mean square error (RMSE) of various machine-learning models, including linear regression, K-nearest neighbors (KNN), random forest regression, gradient boosting regression with decision trees, neural network (multilayer perceptron), and support vector regression (SVR), with non-normalized input. c The comparison of red onset prediction and observation output between linear regression and random forest regression, with the green circle indicating PTEAI-capped film. d The feature importance ranking obtained from the random forest regression algorithm and SHAP library, showing the chemical properties and processing conditions in descending order of importance (rank). The yellow and purple color indicates high and low values of a given feature, respectively.CC BY · PubMed Central
Fig. 4 Crystal structure and morphology characterization for optimum capping layer (PTEAI). a Phenyltriethylammonium (PTEA) has 0 hydrogen-bond donor due to its quaternary amine. Powder X-ray diffraction (XRD) spectra of bare (b) and PTEAI-capped MAPbI3 (c), when they were degraded for 460 and 640 min The purple shaded area indicates the (110) peak of MAPbI3, and the orange shaded area indicates the (001) PbI2 peak. Pre-degraded scanning electron microscopy (SEM) image of bare (d) and PTEAI-capped MAPbI3 (e), scale bar: 200 nm. f Grazing-incidence wide-angle x-ray scattering (GIWAXS) images of PTEAI-capped MAPbI3 presenting the structure on the surface and in the bulk films, with the incident angles of 0.12° and 0.2°, respectively. The green arrow points to LD capping peak, and the blue arrow points to the MAPbI3 peak.
Fig. 4 Crystal structure and morphology characterization for optimum capping layer (PTEAI). a Phenyltriethylammonium (PTEA) has 0 hydrogen-bond donor due to its quaternary amine. Powder X-ray diffraction (XRD) spectra of bare (b) and PTEAI-capped MAPbI3 (c), when they were degraded for 460 and 640 min The purple shaded area indicates the (110) peak of MAPbI3, and the orange shaded area indicates the (001) PbI2 peak. Pre-degraded scanning electron microscopy (SEM) image of bare (d) and PTEAI-capped MAPbI3 (e), scale bar: 200 nm. f Grazing-incidence wide-angle x-ray scattering (GIWAXS) images of PTEAI-capped MAPbI3 presenting the structure on the surface and in the bulk films, with the incident angles of 0.12° and 0.2°, respectively. The green arrow points to LD capping peak, and the blue arrow points to the MAPbI3 peak.CC BY · PubMed Central
Fig. 5 Capping-layer surface chemistry characterization of optimum capping layer (PTEAI). Attenuated total reflection Fourier-transform infrared spectroscopy (ATR-FTIR) with ZnSe crystal of bare (a, c) and PTEAI-capped MAPbI3 (b, d). e The O 1s spectra obtained from x-ray photoelectron spectroscopy (XPS) of bare and PTEAI-capped MAPbI3, and f their corresponding atomic percentages. The purple circle and the green diamond represent the atomic percentages of O 1s for bare and PTEAI-capped MAPbI3 film, respectively.
Fig. 5 Capping-layer surface chemistry characterization of optimum capping layer (PTEAI). Attenuated total reflection Fourier-transform infrared spectroscopy (ATR-FTIR) with ZnSe crystal of bare (a, c) and PTEAI-capped MAPbI3 (b, d). e The O 1s spectra obtained from x-ray photoelectron spectroscopy (XPS) of bare and PTEAI-capped MAPbI3, and f their corresponding atomic percentages. The purple circle and the green diamond represent the atomic percentages of O 1s for bare and PTEAI-capped MAPbI3 film, respectively.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 66 references

Research works printed but not linked: 1

  1. Kingma, D. P. & Ba, J. L. In 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings (International Conference on Learning Representations, ICLR, 2015).

Cited by

See all 189

Metadata

Authors
Hartono, N. T. P., Thapa, J., Tiihonen, A., Oviedo, F., Batali, C., Yoo, J. J., Liu, Z., Li, R., Marrón, D. F., Bawendi, M. G., Buonassisi, T., & Sun, S.
Affiliations
  • Massachusetts Institute of Technology (US)
  • Brookhaven National Laboratory (US)
Corresponding author
  • Buonassisi, Tonio
  • Sun, Shijing
Venue
Nature Communications
Published
2020-08-20
Volume
11
Issue
1
Pages
4172-4172
ISSN
2041-1723 (ISSN-L)
OpenAlex
W3064646198
Cited by
189
References
66
Citations by year
2026: 27 · 2025: 38 · 2024: 35 · 2023: 26 · 2022: 39 · 2021: 21 · 2020: 3
FWCI
7.63
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • ChemRxiv (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-nc)
  • Royal College of Surgeons in Ireland (submittedVersion, cc-by-nc-nd)
Funders
  • National Science Foundation (1605547, W911NF-13-D-0001, 1541959, 1419807, DE-SC0012704, W911NF-13)
  • U.S. Department of Energy (DE‐SC0012704, DEEE0007535, Contract No: DESC0012704, No. DE-SC0012704, SC0012704.)
  • National Aeronautics and Space Administration (DE-SC0012704, NNX16AM70H)
  • Georgia Institute of Technology
  • Harvard University (1541959)
  • Alfred Kordelinin Säätiö
  • Skolkovo Institute of Science and Technology (1913/R)
  • Office of Science (DE-SC-0012704, No. DE-SC0012704, Contract No. DE-SC0012704, SC0012704)
  • Materials Research Science and Engineering Center, Harvard University (DMR1419807)
  • Division of Materials Research (DMR-14-19807, 14-19807)
  • Division of Chemical, Bioengineering, Environmental, and Transport Systems (CBET-1605547)
  • Basic Energy Sciences (DE- SC0012704, SC0012704)
  • Brookhaven National Laboratory (SC0012704, No. DE-SC0012704, DE- SC0012704, Contract No. DE-SC0012704)
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Perovskite Materials and Applications
  • Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
  • Physical Sciences › Materials Science › Materials Chemistry › Ferroelectric and Piezoelectric Materials
Keywords
perovskite solar cells, capping layer, perovskite degradation, environmental stability, methylammonium lead iodide, machine learning, Shapley value, hydrogen bond donors, polar surface area, accelerated aging, surface chemistry

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: How machine learning can help select capping layers to suppress perovskite degradation. https://app.cochat.ai/papers/10.1038/s41467-020-17945-4
The paper
Hartono, N. T. P., Thapa, J., Tiihonen, A., Oviedo, F., Batali, C., Yoo, J. J., Liu, Z., Li, R., Marrón, D. F., Bawendi, M. G., Buonassisi, T., & Sun, S. (2020). How machine learning can help select capping layers to suppress perovskite degradation. Nature Communications. https://doi.org/10.1038/s41467-020-17945-4

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.