pith. sign in

arxiv: 2606.13060 · v1 · pith:KYNLAJ43new · submitted 2026-06-11 · 💻 cs.LG

A green solvent screening tool for emerging materials via uncertainty aware, transformer enhanced transfer learning

Ioannis Kouroudis , Simon Ternes , Zhaosu Gu , Gohar Ali Siddiqui , Marina Ustinova , Angelo Lembo , Alessio Gagliardi , Aldo Di Carlo This is my paper

Pith reviewed 2026-06-27 07:34 UTC · model grok-4.3

classification 💻 cs.LG
keywords green solventssolubility predictiontransfer learningtransformer modeluncertainty quantificationsolvent screeningmachine learningmaterials chemistry
0
0 comments X

The pith

Transfer learning from QM9 data predicts green solvent properties accurately even when experimental measurements are extremely scarce.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that a pre-trained transformer model on QM9 quantum chemistry targets can be transferred to solubility-related properties using minimal additional data. This yields high-accuracy predictions for Hansen solubility parameters and dielectric constants as baselines, and extends successfully to Gutmann donor and acceptor numbers where data is extremely limited. A sympathetic reader cares because solvent substitution is critical for scaling emerging technologies such as organic photovoltaics and batteries, where greener alternatives must be identified rapidly. The pipeline adds uncertainty quantification so users can judge prediction reliability, and deploys an easy-to-integrate screening tool that ranks candidate solvents. Overall the work augments available solubility descriptor data by orders of magnitude with usable quality.

Core claim

By transferring a pre-trained foundational model on QM9 targets to solubility-related properties with minimal additional data, while integrating uncertainty quantification, the approach achieves high performance on targets such as Gutmann donor and acceptor numbers where experimental data is extremely limited, augments solubility descriptor data by orders of magnitude with high-quality predictions, and supplies a customizable screening tool that rediscovers known green solvents and proposes new candidates.

What carries the argument

Uncertainty-aware transformer-enhanced transfer learning pipeline that adapts QM9 pre-training to experimental solubility targets.

If this is right

  • Baseline predictions reach high accuracy on Hansen solubility parameters and dielectric constant using established databases.
  • High model performance extends to Gutmann donor and acceptor numbers despite extremely limited training data.
  • Solubility descriptor data is augmented by orders of magnitude through high-quality predictions.
  • An easy-to-use, customizable tool ranks and screens solvent substitutes and integrates with high-throughput laboratory workflows.
  • The tool rediscovers known green solvent alternatives and proposes new candidates.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same transfer pipeline could be applied to other sparse experimental datasets in materials chemistry to generate virtual libraries for screening.
  • Uncertainty estimates could be used to prioritize which solvent predictions to validate first in the lab, creating an active-learning loop.
  • Integration of the screening tool into automated synthesis platforms could enable closed-loop discovery of greener process solvents.
  • If the transfer works across additional property classes, it reduces reliance on large experimental datasets for many sustainable-chemistry applications.

Load-bearing premise

A model pre-trained on QM9 quantum chemistry targets transfers to experimental solubility properties using only minimal new data while preserving high accuracy on targets that have very few measurements.

What would settle it

New laboratory measurements of Gutmann acceptor numbers for solvents where the model assigns high confidence yet the measured values differ substantially from the predictions.

Figures

Figures reproduced from arXiv: 2606.13060 by Aldo Di Carlo, Alessio Gagliardi, Angelo Lembo, Gohar Ali Siddiqui, Ioannis Kouroudis, Marina Ustinova, Simon Ternes, Zhaosu Gu.

Figure 1
Figure 1. Figure 1: Machine Learning pipeline used in this work. A SMILES-based filter firstly identifies if the chemical [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Regression curves resulting from the pipeline shown in Fig. 1, for all the relevant parameters for [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: RRMSE for every predicted solubility parameter omitting different components of the pipeline [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison reference solvent molecules with molecules situated at closest Euclidean distance [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
read the original abstract

Accurate prediction of solubility remains a central challenge across materials science and sustainable chemistry. In particular due to emerging technologies like organic and hybrid photovoltaics, batteries, and catalysis, solvent usage is expected to increase significantly within the coming years. Therefore, substituting solvents with greener alternatives is vital. This is where machine learning can have substantial impact. However, the limited data on critical parameters of solubility significantly constraints machine learning efficacy. In this work, we transfer a pre-trained foundational model on QM9 targets to our application with minimal data requirements. Additionally, the pipeline integrates uncertainty quantification, allowing the user to gauge the confidence of the predictions. As baseline, we succeed in predicting the Hansen solubility parameters and Dielectric Constant for which extensive databases exist. Importantly, we achieve high model performance on additional targets, such as Gutmann Donor and Acceptor numbers, where the available data is extremely limited. Overall, we augment data on solubility descriptors by orders of magnitude with high quality predictions. For effective dissemination, we deploy easy-to-use, easily integrateable with high throughput labs, customizable tool for ranking and screening possible solvent substitutes. Finally, we rediscovered known green solvent alternatives and proposed new candidates proving its relevance for finding eco-friendly solvents.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript presents a transfer-learning pipeline that fine-tunes a transformer pre-trained on QM9 quantum-chemical targets to predict solubility descriptors (Hansen parameters, dielectric constant, Gutmann donor/acceptor numbers) with uncertainty quantification; it claims high accuracy on the low-data Gutmann targets, orders-of-magnitude data augmentation, and deploys a screening tool that recovers known green solvents and proposes new candidates.

Significance. If the transfer-learning benefit and uncertainty calibration are rigorously demonstrated on held-out experimental data, the work would supply a practical, deployable tool for solvent substitution in sustainable materials chemistry.

major comments (3)
  1. [Abstract] Abstract: the central claim that the QM9-pretrained model achieves 'high model performance' on Gutmann donor/acceptor numbers 'where the available data is extremely limited' is unsupported by any numerical metrics, validation protocol, or error bars; without these the transfer-learning assertion cannot be evaluated.
  2. [Abstract] Abstract: the statement that the pipeline 'augment[s] data on solubility descriptors by orders of magnitude with high quality predictions' is load-bearing yet provides neither the baseline database sizes, the augmentation factor, nor an independent quality metric (e.g., external test-set MAE or comparison to a from-scratch model).
  3. [Abstract] Abstract / Methods (inferred): the domain shift from gas-phase QM9 electronic-structure targets to solution-phase experimental Gutmann numbers is not addressed; no ablation, representation-alignment analysis, or comparison against training solely on the target data is mentioned, leaving open whether pre-training confers benefit beyond the limited experimental labels themselves.
minor comments (2)
  1. [Abstract] Abstract: 'constraints machine learning efficacy' should read 'constrains'; 'integrateable' should read 'integrable'.
  2. [Abstract] Abstract: the claim that the tool 'rediscovered known green solvent alternatives' is stated without citing the specific solvents, the literature sources, or quantitative agreement metrics.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments on the abstract and related claims. We address each point below and have revised the manuscript to incorporate quantitative metrics, dataset details, and an ablation study as suggested.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the QM9-pretrained model achieves 'high model performance' on Gutmann donor/acceptor numbers 'where the available data is extremely limited' is unsupported by any numerical metrics, validation protocol, or error bars; without these the transfer-learning assertion cannot be evaluated.

    Authors: We agree that the abstract should include supporting numerical evidence. In the revised version, we have added the test-set MAE, R² values, and uncertainty estimates for the Gutmann donor/acceptor predictions, along with a brief description of the 5-fold cross-validation protocol used on the experimental data. revision: yes

  2. Referee: [Abstract] Abstract: the statement that the pipeline 'augment[s] data on solubility descriptors by orders of magnitude with high quality predictions' is load-bearing yet provides neither the baseline database sizes, the augmentation factor, nor an independent quality metric (e.g., external test-set MAE or comparison to a from-scratch model).

    Authors: We concur that these details are necessary. The revised abstract now specifies the original database sizes (approximately 1000 for Hansen parameters, 500 for dielectric constants, and under 100 for Gutmann numbers), the augmentation to over 10,000 predictions, and references to external test-set MAE along with performance gains versus from-scratch baselines. revision: yes

  3. Referee: [Abstract] Abstract / Methods (inferred): the domain shift from gas-phase QM9 electronic-structure targets to solution-phase experimental Gutmann numbers is not addressed; no ablation, representation-alignment analysis, or comparison against training solely on the target data is mentioned, leaving open whether pre-training confers benefit beyond the limited experimental labels themselves.

    Authors: This is a fair observation on the domain shift. We have added an ablation study to the methods and results sections comparing the transfer-learned model against one trained from scratch on the Gutmann data, showing improved performance from pre-training. A short discussion of the domain shift and the role of the shared transformer architecture has also been included. revision: yes

Circularity Check

0 steps flagged

No circularity: transfer learning pipeline is empirically evaluated on held-out data

full rationale

The paper presents a standard transfer-learning workflow: pretrain on QM9, fine-tune on solubility targets (Hansen, dielectric, Gutmann), and report performance metrics. No equations or claims reduce a prediction to its own training inputs by construction. No self-citation is invoked as a uniqueness theorem or load-bearing justification. The low-data regime is addressed via uncertainty quantification and empirical results rather than definitional equivalence. This is the normal, non-circular case for an ML methods paper.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claims rest on the effectiveness of transfer from QM9 quantum properties to experimental solubility descriptors and on the reliability of uncertainty estimates from the fine-tuned transformer; no new physical entities are postulated.

free parameters (2)
  • fine-tuned transformer weights
    Model parameters are adjusted on the target solubility datasets; the number and values are not stated in the abstract.
  • uncertainty estimation hyperparameters
    Parameters controlling how uncertainty is quantified are chosen during training but not enumerated.
axioms (2)
  • domain assumption Quantum-mechanical properties encoded in QM9 transfer usefully to macroscopic solubility behavior
    Invoked when the pre-trained model is applied to Hansen, dielectric, and Gutmann targets.
  • domain assumption Uncertainty estimates from the transformer reliably indicate prediction confidence on out-of-distribution solvents
    Required for the tool to be usable in screening.

pith-pipeline@v0.9.1-grok · 5777 in / 1449 out tokens · 22052 ms · 2026-06-27T07:34:04.026394+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

90 extracted references · 48 canonical work pages

  1. [1]

    Green and Bio-Based Solvents

    Francisco G. Calvo-Flores et al. “Green and Bio-Based Solvents”. In:Topics in Current Chemistry 376.3 (Apr. 24, 2018), p. 18.issn: 2364-8961.doi:10 . 1007 / s41061 - 018 - 0191 - 6.url:https : //doi.org/10.1007/s41061-018-0191-6(visited on 12/05/2025)

    work page doi:10.1007/s41061-018-0191-6(visited 2018

  2. [2]

    The green solvent: a critical perspective

    Neil Winterton. “The green solvent: a critical perspective”. In:Clean Technologies and Environmental Policy23.9 (Nov. 1, 2021), pp. 2499–2522.issn: 1618-9558.doi:10.1007/s10098-021-02188-8.url: https://doi.org/10.1007/s10098-021-02188-8(visited on 12/05/2025). 13

    work page doi:10.1007/s10098-021-02188-8.url: 2021

  3. [3]

    Recent Advances in Printed Thin-Film Batteries

    Benoit Clement et al. “Recent Advances in Printed Thin-Film Batteries”. In:Engineering13 (June 1, 2022), pp. 238–261.issn: 2095-8099.doi:10 . 1016 / j . eng . 2022 . 04 . 002.url:https : / / www . sciencedirect.com/science/article/pii/S2095809922002053(visited on 10/30/2024)

    2022

  4. [4]

    Recent Advances and Challenges in Thin-Film Fabrication Techniques for Low-Temperature Solid Oxide Fuel Cells

    Mohammadmehdi Choolaei, Mohsen Fallah Vostakola, and Bahman Amini Horri. “Recent Advances and Challenges in Thin-Film Fabrication Techniques for Low-Temperature Solid Oxide Fuel Cells”. In:Crystals13.7 (July 2023). Publisher: Multidisciplinary Digital Publishing Institute, p. 1008.issn: 2073-4352.doi:10.3390/cryst13071008.url:https://www.mdpi.com/2073- 43...

    work page doi:10.3390/cryst13071008.url:https://www.mdpi.com/2073- 2023

  5. [5]

    Coated and Printed Perovskites for Photovoltaic Applications

    Ian A. Howard et al. “Coated and Printed Perovskites for Photovoltaic Applications”. In:Advanced Ma- terials31.26 (2019). eprint: https://advanced.onlinelibrary.wiley.com/doi/pdf/10.1002/adma.201806702, p. 1806702.issn: 1521-4095.doi:10.1002/adma.201806702.url:https://onlinelibrary.wiley. com/doi/abs/10.1002/adma.201806702(visited on 09/26/2025). [6]Thin ...

    work page doi:10.1002/adma.201806702 2019

  6. [6]

    Volatile Organic Compounds (VOCs) as Environmental Pollutants: Occurrence and Mitigation Using Nanomaterials

    Elena David and Violeta-Carolina Niculescu. “Volatile Organic Compounds (VOCs) as Environmental Pollutants: Occurrence and Mitigation Using Nanomaterials”. In:International Journal of Environ- mental Research and Public Health18.24 (Jan. 2021). Publisher: Multidisciplinary Digital Publishing Institute, p. 13147.issn: 1660-4601.doi:10.3390/ijerph182413147....

    work page doi:10.3390/ijerph182413147.url:https://www.mdpi.com/ 2021

  7. [7]

    A Review on Volatile Organic Compounds (VOCs) as Environmental Pollutants: Fate and Distribution

    Puneeta Pandey and Radheshyam Yadav. “A Review on Volatile Organic Compounds (VOCs) as Environmental Pollutants: Fate and Distribution”. In:INTERNATIONAL JOURNAL OF PLANT AND ENVIRONMENT4.2 (July 31, 2018), pp. 14–26.issn: 2455-202X.doi:10 . 18811 / ijpen . v4i02 . 2.url:https : / / ijplantenviro . com / index . php / IJPE / article / view / 1047(visited ...

    2018

  8. [8]

    Health risk analysis from volatile organic compounds and fine particulate matter in the printing industry

    N. Pongboonkhumlarp and W. Jinsart. “Health risk analysis from volatile organic compounds and fine particulate matter in the printing industry”. In:International Journal of Environmental Science and Technology19.9 (Sept. 1, 2022), pp. 8633–8644.issn: 1735-2630.doi:10.1007/s13762-021-03733-0. url:https://doi.org/10.1007/s13762-021-03733-0(visited on 12/05/2025)

    work page doi:10.1007/s13762-021-03733-0 2022

  9. [9]

    Green solvents, materials, and lead-free semiconductors for sustainable fabrication of perovskite solar cells

    Suresh K. Podapangi et al. “Green solvents, materials, and lead-free semiconductors for sustainable fabrication of perovskite solar cells”. In:RSC Advances13.27 (June 15, 2023). Publisher: The Royal Society of Chemistry, pp. 18165–18206.issn: 2046-2069.doi:10 . 1039 / D3RA01692G.url:https : //pubs.rsc.org/en/content/articlelanding/2023/ra/d3ra01692g(visit...

    2023

  10. [10]

    Green solvent strategies for the sustainable development of perovskite solar cells

    David Sunghwan Lee et al. “Green solvent strategies for the sustainable development of perovskite solar cells”. In:Chemical Communications61.10 (Jan. 28, 2025). Publisher: The Royal Society of Chemistry, pp. 2011–2025.issn: 1364-548X.doi:10.1039/D4CC05454G.url:https://pubs.rsc. org/en/content/articlelanding/2025/cc/d4cc05454g(visited on 12/05/2025)

    work page doi:10.1039/d4cc05454g.url:https://pubs.rsc 2025

  11. [11]

    Organic photovoltaics: Potential fate and effects in the environ- ment

    Yannick-Serge Zimmermann et al. “Organic photovoltaics: Potential fate and effects in the environ- ment”. In:Environment International49 (Nov. 15, 2012), pp. 128–140.issn: 0160-4120.doi:10 . 1016/j.envint.2012.08.015.url:https://www.sciencedirect.com/science/article/pii/ S0160412012001894(visited on 12/05/2025)

    2012

  12. [12]

    Assessing health and environmental impacts of solvents for producing perovskite solar cells

    Rosario Vidal et al. “Assessing health and environmental impacts of solvents for producing perovskite solar cells”. In:Nature Sustainability4.3 (Mar. 2021). Publisher: Nature Publishing Group, pp. 277– 285.issn: 2398-9629.doi:10 . 1038 / s41893 - 020 - 00645 - 8.url:https : / / www . nature . com / articles/s41893-020-00645-8(visited on 12/05/2025)

    2021

  13. [13]

    Toward a More Holistic Framework for Solvent Selection

    Louis J. Diorazio, David R. J. Hose, and Neil K. Adlington. “Toward a More Holistic Framework for Solvent Selection”. In:Organic Process Research & Development20.4 (Apr. 15, 2016). Publisher: American Chemical Society, pp. 760–773.issn: 1083-6160.doi:10.1021/acs.oprd.6b00015.url: https://doi.org/10.1021/acs.oprd.6b00015(visited on 12/06/2025). 14

    work page doi:10.1021/acs.oprd.6b00015.url: 2016

  14. [14]

    What is a green solvent? A comprehen- sive framework for the environmental assessment of solvents

    Christian Capello, Ulrich Fischer, and Konrad Hungerb¨ uhler. “What is a green solvent? A comprehen- sive framework for the environmental assessment of solvents”. In:Green Chemistry9.9 (Aug. 28, 2007). Publisher: The Royal Society of Chemistry, pp. 927–934.issn: 1463-9270.doi:10.1039/B617536H.url: https://pubs.rsc.org/en/content/articlelanding/2007/gc/b61...

    work page doi:10.1039/b617536h.url: 2007

  15. [15]

    Green and Sustainable Solvents in Chemical Processes

    Coby J. Clarke et al. “Green and Sustainable Solvents in Chemical Processes”. In:Chemical Reviews 118.2 (Jan. 24, 2018). Publisher: American Chemical Society, pp. 747–800.issn: 0009-2665.doi:10. 1021/acs.chemrev.7b00571.url:https://doi.org/10.1021/acs.chemrev.7b00571(visited on 12/05/2025)

    work page doi:10.1021/acs.chemrev.7b00571(visited 2018

  16. [16]

    CHEM21 selection guide of classical- and less classical-solvents

    Denis Prat et al. “CHEM21 selection guide of classical- and less classical-solvents”. In:Green Chem- istry18.1 (2016). Publisher: Royal Society of Chemistry, pp. 288–296.doi:10 . 1039 / C5GC01008J. url:https : / / pubs . rsc . org / en / content / articlelanding / 2016 / gc / c5gc01008j(visited on 12/05/2025)

    2016

  17. [17]

    Tools and techniques for solvent selection: green solvent selection guides

    Fergal P. Byrne et al. “Tools and techniques for solvent selection: green solvent selection guides”. In: Sustainable Chemical Processes4.1 (May 23, 2016), p. 7.issn: 2043-7129.doi:10.1186/s40508-016- 0051-z.url:https://doi.org/10.1186/s40508-016-0051-z(visited on 12/05/2025)

    work page doi:10.1186/s40508-016- 2016

  18. [18]

    Alternative Solvents: Shades of Green

    James H. Clark and Stewart J. Tavener. “Alternative Solvents: Shades of Green”. In:Organic Process Research & Development11.1 (Jan. 1, 2007). Publisher: American Chemical Society, pp. 149–155. issn: 1083-6160.doi:10.1021/op060160g.url:https://doi.org/10.1021/op060160g(visited on 12/06/2025)

    work page doi:10.1021/op060160g.url:https://doi.org/10.1021/op060160g(visited 2007

  19. [19]

    LoDoPaB-CT, a benchmark dataset for low-dose computed tomography reconstruction,

    P. Llompart et al. “Will we ever be able to accurately predict solubility?” In:Scientific Data11.1 (Mar. 18, 2024). Publisher: Nature Publishing Group, p. 303.issn: 2052-4463.doi:10.1038/s41597- 024 - 03105 - 6.url:https : / / www . nature . com / articles / s41597 - 024 - 03105 - 6(visited on 12/23/2025)

    work page doi:10.1038/s41597- 2024

  20. [20]

    Findings of the Challenge To Predict Aqueous Solubility

    Anton J. Hopfinger et al. “Findings of the Challenge To Predict Aqueous Solubility”. In:Journal of Chemical Information and Modeling49.1 (Jan. 26, 2009). Publisher: American Chemical Society, pp. 1–5.issn: 1549-9596.doi:10.1021/ci800436c.url:https://doi.org/10.1021/ci800436c (visited on 12/23/2025)

    work page doi:10.1021/ci800436c.url:https://doi.org/10.1021/ci800436c 2009

  21. [21]

    Solubility parameters

    Allan F. M. Barton. “Solubility parameters”. In:Chemical Reviews75.6 (Dec. 1, 1975). Publisher: American Chemical Society, pp. 731–753.issn: 0009-2665.doi:10.1021/cr60298a003.url:https: //doi.org/10.1021/cr60298a003(visited on 12/05/2025)

    work page doi:10.1021/cr60298a003.url:https: 1975

  22. [22]

    Solubility, solvency, and solubility parameters

    John W. Van Dyk, Harry L. Frisch, and Dao T. Wu. “Solubility, solvency, and solubility parameters”. In:Industrial & Engineering Chemistry Product Research and Development24.3 (Sept. 1, 1985). Pub- lisher: American Chemical Society, pp. 473–478.issn: 0196-4321.doi:10.1021/i300019a028.url: https://doi.org/10.1021/i300019a028(visited on 12/05/2025)

    work page doi:10.1021/i300019a028.url: 1985

  23. [23]

    Application of Hansen solubility parameters in the eutectic mixtures: difference between empirical and semi-empirical models

    Cl´ audio C. Fernandes et al. “Application of Hansen solubility parameters in the eutectic mixtures: difference between empirical and semi-empirical models”. In:Scientific Reports15.1 (Jan. 31, 2025). Publisher: Nature Publishing Group, p. 3862.issn: 2045-2322.doi:10.1038/s41598-025-87050-3. url:https://www.nature.com/articles/s41598-025-87050-3(visited o...

    work page doi:10.1038/s41598-025-87050-3 2025

  24. [24]

    Hansen.Hansen Solubility Parameters: A User’s Handbook, Second Edition

    Charles M. Hansen.Hansen Solubility Parameters: A User’s Handbook, Second Edition. 2nd ed. Boca Raton: CRC Press, June 15, 2007. 544 pp.isbn: 978-0-429-12752-6.doi:10.1201/9781420006834

    work page doi:10.1201/9781420006834 2007

  25. [25]

    Revisiting Hansen Solubility Parameters by Including Thermodynamics

    Manuel J. Louwerse et al. “Revisiting Hansen Solubility Parameters by Including Thermodynamics”. In:ChemPhysChem18.21 (2017). eprint: https://chemistry-europe.onlinelibrary.wiley.com/doi/pdf/10.1002/cphc.201700408, pp. 2999–3006.issn: 1439-7641.doi:10 . 1002 / cphc . 201700408.url:https : / / onlinelibrary . wiley.com/doi/abs/10.1002/cphc.201700408(visite...

    work page doi:10.1002/cphc.201700408 2017

  26. [26]

    Comparing and Correlating Solubility Parameters Governing the Self-Assembly of Molecular Gels Using 1,3:2,4-Dibenzylidene Sorbitol as the Gelator

    Yaqi Lan et al. “Comparing and Correlating Solubility Parameters Governing the Self-Assembly of Molecular Gels Using 1,3:2,4-Dibenzylidene Sorbitol as the Gelator”. In:Langmuir30.47 (Dec. 2, 2014). Publisher: American Chemical Society, pp. 14128–14142.issn: 0743-7463.doi:10.1021/la5008389. url:https://doi.org/10.1021/la5008389(visited on 12/05/2025). 15

    work page doi:10.1021/la5008389 2014

  27. [27]

    Kamlet et al

    Mortimer J. Kamlet et al. “Linear solvation energy relationships. 23. A comprehensive collection of the solvatochromic parameters, .pi.*, .alpha., and .beta., and some methods for simplifying the generalized solvatochromic equation”. In:The Journal of Organic Chemistry48.17 (Aug. 1, 1983). Publisher: American Chemical Society, pp. 2877–2887.issn: 0022-326...

    work page doi:10.1021/jo00165a018.url:https: 1983

  28. [28]

    Empirical parameters for donor and acceptor properties of solvents

    V. Gutmann. “Empirical parameters for donor and acceptor properties of solvents”. In:Electrochimica Acta21.9 (Sept. 1, 1976), pp. 661–670.issn: 0013-4686.doi:10 . 1016 / 0013 - 4686(76 ) 85034 - 7. url:https : / / www . sciencedirect . com / science / article / pii / 0013468676850347(visited on 12/05/2025). [30]Stenutz Chem Tables.url:https://www.stenutz....

    1976

  29. [29]

    Utilizing data-driven optimization to automate the parametrization of kinetic Monte Carlo models

    Ioannis Kouroudis, Manuel Gosswein, and Alessio Gagliardi. “Utilizing data-driven optimization to automate the parametrization of kinetic Monte Carlo models”. In:The Journal of Physical Chemistry A127.28 (2023), pp. 5967–5978

    2023

  30. [30]

    Machine learning and optoelectronic materials discovery: a growing synergy

    Felix Mayr et al. “Machine learning and optoelectronic materials discovery: a growing synergy”. In: The Journal of Physical Chemistry Letters13.8 (2022), pp. 1940–1951

    2022

  31. [31]

    Synthesizer: Chemistry-Aware Machine Learning for Precision Control of Nanocrys- tal Growth

    Nina A Henke et al. “Synthesizer: Chemistry-Aware Machine Learning for Precision Control of Nanocrys- tal Growth”. In:Advanced Materials(2025), e09472

    2025

  32. [32]

    Rapid data-efficient optimization of perovskite nanocrystal syntheses through machine learning algorithm fusion

    Carola Lampe et al. “Rapid data-efficient optimization of perovskite nanocrystal syntheses through machine learning algorithm fusion”. In:Advanced Materials35.16 (2023), p. 2208772

    2023

  33. [33]

    Finding new perovskite halides via machine learning

    Ghanshyam Pilania et al. “Finding new perovskite halides via machine learning”. In:Frontiers in Materials3 (2016), p. 19

    2016

  34. [34]

    Machine-learning-accelerated perovskite crystallization

    Jeffrey Kirman et al. “Machine-learning-accelerated perovskite crystallization”. In:Matter2.4 (2020), pp. 938–947

    2020

  35. [35]

    Optoelectronic perovskite film characterization via machine vision

    M Harth et al. “Optoelectronic perovskite film characterization via machine vision”. In:Solar Energy 262 (2023), p. 111840

    2023

  36. [36]

    Data-driven frameworks to robustly predict solubility parameter of diverse polymers

    Raouf Hassan and Mohammad Reza Kazemi. “Data-driven frameworks to robustly predict solubility parameter of diverse polymers”. In:Scientific Reports15.1 (2025), p. 31157

    2025

  37. [37]

    The prediction of donor number and acceptor number of electrolyte solvent molecules based on machine learning

    Huaping Hu et al. “The prediction of donor number and acceptor number of electrolyte solvent molecules based on machine learning”. In:Journal of Energy Chemistry98 (Nov. 1, 2024), pp. 374–382. issn: 2095-4956.doi:10.1016/j.jechem.2024.06.050.url:https://www.sciencedirect.com/ science/article/pii/S2095495624004625(visited on 12/06/2025)

    work page doi:10.1016/j.jechem.2024.06.050.url:https://www.sciencedirect.com/ 2024

  38. [38]

    Easy and fast prediction of green solvents for small molecule donor-based organic solar cells through machine learning

    Asif Mahmood, Yahya Sandali, and Jin-Liang Wang. “Easy and fast prediction of green solvents for small molecule donor-based organic solar cells through machine learning”. In:Physical Chemistry Chemical Physics25.15 (2023), pp. 10417–10426

    2023

  39. [39]

    A Bayesian approach to predict solubility parameters

    Benjamin Sanchez-Lengeling et al. “A Bayesian approach to predict solubility parameters”. In:Ad- vanced Theory and Simulations2.1 (2019), p. 1800069

    2019

  40. [40]

    A tool for identifying green solvents for printed electronics

    Christian Larsen et al. “A tool for identifying green solvents for printed electronics”. In:Nature Com- munications12.1 (2021), p. 4510

    2021

  41. [41]

    MACE: Higher order equivariant message passing neural networks for fast and accurate force fields

    Ilyes Batatia et al. “MACE: Higher order equivariant message passing neural networks for fast and accurate force fields”. In:Advances in neural information processing systems35 (2022), pp. 11423– 11436

    2022

  42. [42]

    Directional message passing for molecular graphs

    Johannes Gasteiger, Janek Gross, and Stephan Gunnemann. “Directional message passing for molecular graphs”. In:arXiv preprint arXiv:2003.03123(2020)

    work page arXiv 2003

  43. [43]

    E (3)-equivariant graph neural networks for data-efficient and accurate inter- atomic potentials

    Simon Batzner et al. “E (3)-equivariant graph neural networks for data-efficient and accurate inter- atomic potentials”. In:Nature communications13.1 (2022), p. 2453

    2022

  44. [44]

    Transfer learning for materials informatics using crystal graph convolu- tional neural network

    Joohwi Lee and Ryoji Asahi. “Transfer learning for materials informatics using crystal graph convolu- tional neural network”. In:Computational Materials Science190 (2021), p. 110314. 16

    2021

  45. [45]

    Predicting materials properties with little data using shotgun transfer learn- ing

    Hironao Yamada et al. “Predicting materials properties with little data using shotgun transfer learn- ing”. In:ACS central science5.10 (2019), pp. 1717–1730

    2019

  46. [46]

    Transfer learning using attentions across atomic systems with graph neural networks (TAAG)

    Adeesh Kolluru et al. “Transfer learning using attentions across atomic systems with graph neural networks (TAAG)”. In:The Journal of Chemical Physics156.18 (2022)

    2022

  47. [47]

    AUGUR, a flexible and efficient optimization algorithm for identification of optimal adsorption sites

    Ioannis Kouroudis et al. “AUGUR, a flexible and efficient optimization algorithm for identification of optimal adsorption sites”. In:npj Computational Materials11.1 (2025), p. 136. [50]Compsol database (revised, 2024). Sept. 24, 2024.doi:10.60893/figshare.jpr.26871184.v1.url: https://aip.figshare.com/articles/dataset/Compsol_database_revised_2024_/2687118...

    work page doi:10.60893/figshare.jpr.26871184.v1.url: 2025

  48. [48]

    Version 5.0.0

    International Union of Pure and Applied Chemistry (IUPAC).Solubility. Version 5.0.0. 2025.doi: 10 . 1351 / goldbook . S05740.url:https : / / doi . org / 10 . 1351 / goldbook . S05740(visited on 12/23/2025)

    2025

  49. [49]

    Critical Assessment of the Hildebrand and Hansen Solubility Parameters for Polymers

    Shruti Venkatram et al. “Critical Assessment of the Hildebrand and Hansen Solubility Parameters for Polymers”. In:Journal of Chemical Information and Modeling59.10 (Oct. 28, 2019). Publisher: American Chemical Society, pp. 4188–4194.issn: 1549-9596.doi:10.1021/acs.jcim.9b00656.url: https://doi.org/10.1021/acs.jcim.9b00656(visited on 12/05/2025)

    work page doi:10.1021/acs.jcim.9b00656.url: 2019

  50. [50]

    Hansen theory applied to the identification of nonhazardous solvents for hybrid perovskite thin-films processing

    Azin Babaei et al. “Hansen theory applied to the identification of nonhazardous solvents for hybrid perovskite thin-films processing”. In:Polyhedron147 (June 1, 2018), pp. 9–14.issn: 0277-5387.doi: 10.1016/j.poly.2018.03.003.url:https://www.sciencedirect.com/science/article/pii/ S0277538718301232(visited on 12/23/2025)

    work page doi:10.1016/j.poly.2018.03.003.url:https://www.sciencedirect.com/science/article/pii/ 2018

  51. [51]

    Determination of Hansen Solubility Pa- rameters of Ionic Liquids by Using Walden Plots

    Yumi Kobayashi, Sonoko Tokishita, and Hideki Yamamoto. “Determination of Hansen Solubility Pa- rameters of Ionic Liquids by Using Walden Plots”. In:Industrial & Engineering Chemistry Research 59.32 (Aug. 12, 2020). Publisher: American Chemical Society, pp. 14217–14223.issn: 0888-5885.doi: 10 . 1021 / acs . iecr . 0c01947.url:https : / / doi . org / 10 . 1...

    2020

  52. [52]

    Mutual Lewis Acid–Base Interactions of Cations and Anions in Ionic Liq- uids

    Markus Holzweber et al. “Mutual Lewis Acid–Base Interactions of Cations and Anions in Ionic Liq- uids”. In:Chemistry (Weinheim an Der Bergstrasse, Germany)19.1 (Jan. 2, 2013), pp. 288–293.issn: 0947-6539.doi:10 . 1002 / chem . 201201978.url:https : / / pmc . ncbi . nlm . nih . gov / articles / PMC3576480/(visited on 12/06/2025)

    2013

  53. [53]

    Understanding and applying the donor number of electrolytes in lithium metal batteries

    Pan Zhou, Yong Xiang, and Kai Liu. “Understanding and applying the donor number of electrolytes in lithium metal batteries”. In:Energy & Environmental Science17.21 (Oct. 29, 2024). Publisher: The Royal Society of Chemistry, pp. 8057–8077.issn: 1754-5706.doi:10.1039/D4EE02989E.url:https: //pubs.rsc.org/en/content/articlelanding/2024/ee/d4ee02989e(visited o...

    work page doi:10.1039/d4ee02989e.url:https: 2024

  54. [54]

    Gutmann Donor and Acceptor Numbers for Ionic Liquids

    Matthias Schmeisser et al. “Gutmann Donor and Acceptor Numbers for Ionic Liquids”. In:Chemistry – A European Journal18.35 (2012). eprint: https://chemistry-europe.onlinelibrary.wiley.com/doi/pdf/10.1002/chem.201200584, pp. 10969–10982.issn: 1521-3765.doi:10.1002/chem.201200584.url:https://onlinelibrary. wiley.com/doi/abs/10.1002/chem.201200584(visited on ...

    work page doi:10.1002/chem.201200584 2012

  55. [55]

    Correlation Between Solubility Parameters and Dielectric Constants

    Anthony N. Paruta, Bartley J. Sciarrone, and Nicholas G. Lordi. “Correlation Between Solubility Parameters and Dielectric Constants”. In:Journal of Pharmaceutical Sciences51.7 (July 1, 1962), pp. 704–705.issn: 0022-3549.doi:10.1002/jps.2600510726.url:https://www.sciencedirect. com/science/article/pii/S002235491533608X(visited on 12/05/2025)

    work page doi:10.1002/jps.2600510726.url:https://www.sciencedirect 1962

  56. [56]

    Relationship between Hansen Solubility Parameters and Lewis Acid-Base Parameters of Solvents

    Baoli Shi. “Relationship between Hansen Solubility Parameters and Lewis Acid-Base Parameters of Solvents”. In:Journal of Macromolecular Science, Part B47.1 (Dec. 1, 2007). Publisher: Taylor & Francis eprint: https://doi.org/10.1080/00222340701746192, pp. 174–179.issn: 0022-2348.doi: 10.1080/00222340701746192.url:https://doi.org/10.1080/00222340701746192(v...

    work page doi:10.1080/00222340701746192 2007

  57. [57]

    Earle Waghorne

    W. Earle Waghorne. “Solvent Acidity and Basicity Scales: Analysis of Catalan’s SB and SA Scales and Gutmann’s Acceptor Number and Comparison with Kamlet and Taft’sβandαSolvent Scales, Gutmann’s Donor Number and Abraham’s B and A Solute Scales”. In:Journal of Solution Chemistry 53.5 (May 1, 2024), pp. 747–760.issn: 1572-8927.doi:10.1007/s10953-024-01382-8....

    work page doi:10.1007/s10953-024-01382-8.url:https: 2024

  58. [58]

    Rational design of Lewis base molecules for stable and efficient inverted perovskite solar cells

    Chongwen Li et al. “Rational design of Lewis base molecules for stable and efficient inverted perovskite solar cells”. In:Science379.6633 (Feb. 17, 2023). Publisher: American Association for the Advancement of Science, pp. 690–694.doi:10.1126/science.ade3970.url:https://www.science.org/doi/10. 1126/science.ade3970(visited on 12/06/2025)

    work page doi:10.1126/science.ade3970.url:https://www.science.org/doi/10 2023

  59. [59]

    Lewis Acid–Base Adduct Approach for High Efficiency Perovskite Solar Cells

    Jin-Wook Lee, Hui-Seon Kim, and Nam-Gyu Park. “Lewis Acid–Base Adduct Approach for High Efficiency Perovskite Solar Cells”. In:Accounts of Chemical Research49.2 (Feb. 16, 2016). Publisher: American Chemical Society, pp. 311–319.issn: 0001-4842.doi:10.1021/acs.accounts.5b00440. url:https://doi.org/10.1021/acs.accounts.5b00440(visited on 12/06/2025)

    work page doi:10.1021/acs.accounts.5b00440 2016

  60. [60]

    PubChem.PubChem.url:https://pubchem.ncbi.nlm.nih.gov/(visited on 12/23/2025)

    2025

  61. [61]

    Pencil and Paper Estimation of Hansen Solubility Parameters

    Didier Mathieu. “Pencil and Paper Estimation of Hansen Solubility Parameters”. In:ACS Omega3.12 (Dec. 31, 2018). Publisher: American Chemical Society, pp. 17049–17056.doi:10.1021/acsomega. 8b02601.url:https://doi.org/10.1021/acsomega.8b02601(visited on 12/06/2025)

    work page doi:10.1021/acsomega 2018

  62. [62]

    ISSN: 0887-8978 Publisher: The Book and Paper Group of the American Institute for Conservation

    John Burke.Solubility Parameters: Theory and Application. ISSN: 0887-8978 Publisher: The Book and Paper Group of the American Institute for Conservation. 1984.url:https://cool.culturalheritage. org/coolaic/sg/bpg/annual/v03/bp03-04.html(visited on 12/06/2025)

    1984

  63. [63]

    Quantum chemistry structures and properties of 134 kilo molecules

    Raghunathan Ramakrishnan et al. “Quantum chemistry structures and properties of 134 kilo molecules”. In:Scientific data1.1 (2014), pp. 1–7

    2014

  64. [64]

    Online chemical modeling environment (OCHEM): web platform for data storage, model development and publishing of chemical information

    Iurii Sushko et al. “Online chemical modeling environment (OCHEM): web platform for data storage, model development and publishing of chemical information”. In:Journal of computer-aided molecular design25.6 (2011), pp. 533–554

    2011

  65. [65]

    Gaussian processes for regression

    Christopher Williams and Carl Rasmussen. “Gaussian processes for regression”. In:Advances in neural information processing systems8 (1995)

    1995

  66. [66]

    Christopher K. I. Williams and Carl Edward Rasmussen.Gaussian Processes for Machine Learning. Vol. 2. 3. MIT press Cambridge, MA, 2006

    2006

  67. [67]

    Gpytorch: Blackbox matrix-matrix gaussian process inference with gpu acceler- ation

    Jacob Gardner et al. “Gpytorch: Blackbox matrix-matrix gaussian process inference with gpu acceler- ation”. In:Advances in neural information processing systems31 (2018)

    2018

  68. [68]

    Constant-time predictive distributions for Gaussian processes

    Geoff Pleiss et al. “Constant-time predictive distributions for Gaussian processes”. In:International Conference on Machine Learning. PMLR. 2018, pp. 4114–4123

    2018

  69. [69]

    Expert Review for GHS Classification of Chemicals on Health Effects

    Takeshi Morita and Kaoru Morikawa. “Expert Review for GHS Classification of Chemicals on Health Effects”. In:Industrial Health49.5 (2011), pp. 559–565.doi:10.2486/indhealth.MS1267

    work page doi:10.2486/indhealth.ms1267 2011

  70. [70]

    doi:https://doi.org/10.1016/B978-0-12- 164902-9.50034-4

    “Framework for Solvent Recovery, Reuse, and Recycling In Industries”. In:Computer Aided Chemical Engineering. Vol. 47. ISSN: 1570-7946. Elsevier, Jan. 1, 2019, pp. 199–204.doi:10.1016/B978-0-12- 818597-1.50032-1.url:https://www.sciencedirect.com:5037/science/chapter/bookseries/ abs/pii/B9780128185971500321(visited on 12/06/2025)

    work page doi:10.1016/b978-0-12- 2019

  71. [71]

    Water Analysis: Emerging Contaminants and Current Is- sues

    Susan D. Richardson and Thomas A. Ternes. “Water Analysis: Emerging Contaminants and Current Is- sues”. In:Analytical Chemistry83.12 (June 15, 2011). Publisher: American Chemical Society, pp. 4614– 4648.issn: 0003-2700.doi:10.1021/ac200915r.url:https://doi.org/10.1021/ac200915r(visited on 12/05/2025)

    work page doi:10.1021/ac200915r.url:https://doi.org/10.1021/ac200915r(visited 2011

  72. [72]

    Computer-aided design and solvent selection for organic paint and coating formulations

    Markus Enekvist et al. “Computer-aided design and solvent selection for organic paint and coating formulations”. In:Progress in Organic Coatings162 (Jan. 1, 2022), p. 106568.issn: 0300-9440.doi: 10.1016/j.porgcoat.2021.106568.url:https://www.sciencedirect.com/science/article/ pii/S0300944021004392(visited on 12/06/2025)

    work page doi:10.1016/j.porgcoat.2021.106568.url:https://www.sciencedirect.com/science/article/ 2022

  73. [73]

    Development of a solvent sustainability guide for the paints and coatings industry

    Laura Pilon et al. “Development of a solvent sustainability guide for the paints and coatings industry”. In:Green Chemistry26.18 (Sept. 16, 2024). Publisher: The Royal Society of Chemistry, pp. 9697– 9711.issn: 1463-9270.doi:10 . 1039 / D4GC01962H.url:https : / / pubs . rsc . org / en / content / articlelanding/2024/gc/d4gc01962h(visited on 12/06/2025)

    2024

  74. [74]

    Green solvents for sustainable organic synthesis: state of the art

    Roger A. Sheldon. “Green solvents for sustainable organic synthesis: state of the art”. In:Green Chemistry7.5 (May 3, 2005), pp. 267–278.issn: 1463-9270.doi:10.1039/B418069K.url:https: //pubs.rsc.org/en/content/articlelanding/2005/gc/b418069k(visited on 06/01/2026). 18

    work page doi:10.1039/b418069k.url:https: 2005

  75. [75]

    Safer bio-based solvents to replace toluene and tetrahydrofuran for the bio- catalyzed synthesis of polyesters

    Alessandro Pellis et al. “Safer bio-based solvents to replace toluene and tetrahydrofuran for the bio- catalyzed synthesis of polyesters”. In:Green Chemistry21.7 (2019), pp. 1686–1694.doi:10.1039/ C8GC03567A.url:https : / / pubs . rsc . org / en / content / articlelanding / 2019 / gc / c8gc03567a (visited on 06/01/2026)

    2019

  76. [76]

    Improved biodiesel production from waste cooking oil with mixed methanol–ethanol using enhanced eggshell-derived CaO nano-catalyst

    Yeshimebet Simeon Erchamo et al. “Improved biodiesel production from waste cooking oil with mixed methanol–ethanol using enhanced eggshell-derived CaO nano-catalyst”. In:Scientific Reports11.1 (Mar. 23, 2021), p. 6708.issn: 2045-2322.doi:10.1038/s41598-021-86062-z.url:https://www. nature.com/articles/s41598-021-86062-z(visited on 06/01/2026)

    work page doi:10.1038/s41598-021-86062-z.url:https://www 2021

  77. [77]

    The Influence of Solvent Choice on the Extraction of Bioactive Compounds from Asteraceae: A Comparative Review

    Ji-Eun Lee et al. “The Influence of Solvent Choice on the Extraction of Bioactive Compounds from Asteraceae: A Comparative Review”. In:Foods13.19 (Jan. 2024), p. 3151.issn: 2304-8158.doi:10. 3390/foods13193151.url:https://www.mdpi.com/2304-8158/13/19/3151(visited on 06/01/2026)

    2024

  78. [78]

    Beyond a solvent: triple roles of dimethylformamide in organic chemistry

    Majid M. Heravi, Mahdieh Ghavidel, and Leyla Mohammadkhani. “Beyond a solvent: triple roles of dimethylformamide in organic chemistry”. In:RSC Advances8.49 (Aug. 2, 2018), pp. 27832– 27862.issn: 2046-2069.doi:10 . 1039 / C8RA04985H.url:https : / / pubs . rsc . org / en / content / articlelanding/2018/ra/c8ra04985h(visited on 06/01/2026)

    2018

  79. [79]

    Local Structure and Polar Order in Liquid N-Methyl-2-pyrrolidone (NMP)

    Nadir S. Basma et al. “Local Structure and Polar Order in Liquid N-Methyl-2-pyrrolidone (NMP)”. In:The Journal of Physical Chemistry B122.38 (Sept. 27, 2018), pp. 8963–8971.issn: 1520-6106.doi: 10 . 1021 / acs . jpcb . 8b08020.url:https : / / doi . org / 10 . 1021 / acs . jpcb . 8b08020(visited on 06/01/2026)

    2018

  80. [80]

    New solvent seeks to replace NMP

    “New solvent seeks to replace NMP”. In:Chemical & Engineering News().url:https://pubs.acs. org/doi/pdf/10.1021/cen-09722-buscon1(visited on 06/01/2026)

    work page doi:10.1021/cen-09722-buscon1(visited 2026

Showing first 80 references.