FusNet: unlocking molecular fusion properties through machine learning

Abstract

Accurate prediction of molecular fusion properties is critical for energy-efficient material design and sustainable process optimization, yet remains challenging due to data scarcity and complex thermodynamic interdependencies. This work introduces machine learning tools to address these gaps by combining expert-curated molecular descriptors with deep learning. By systematically evaluating statistical machine learning algorithms and attention-based architectures, optimized models are identified: a SMILES-augmented Transformer-Convolutional Neural Network for fusion temperature and a graph attention network for fusion enthalpy. Prediction power is further validated experimentally on four structure diverse compounds (γ-butyrolactone, methyl octanoate, N-phenylbenzenesulfonamide, and triethylene glycol dimethyl ether). Interpretability analyses reveal that these models prioritize key structures in molecules: attention in text-based models focuses on key atoms while that in graph models focuses on key chemical bonds, aligning with empirical thermodynamic evidences. By providing rapid, interpretable fusion property predictions, this framework can support the development of low-energy phase-change materials and sustainable solvent systems, advancing data-driven green chemistry.

Graphical abstract

Keywords

fusion temperature / fusion enthalpy / interpretable machine learning / graph attention networks / transformer

References

[1] Biography V , de With G . Melting is well-known, but is it also well-understood. Chemical Reviews, 2023, 123(23): 13713–13795

[2] Du Y , Calzavarini E , Sun C . The physics of freezing and melting in the presence of flows. Nature Reviews: Physics, 2024, 6(11): 676–690

[3] Mathers A , Hassouna F , Klajmon M , Fulem M . Comparative study of DSC-based protocols for API-polymer solubility determination. Molecular Pharmaceutics, 2021, 18(4): 1742–1757

[4] Yani Y , Kanaujia P , Chow P S , Tan R B H . Effect of API-polymer miscibility and interaction on the stabilization of amorphous solid dispersion: a molecular simulation study. Industrial & Engineering Chemistry Research, 2017, 56(44): 12698–12707

[5] Shamshina J L , Rogers R D . Ionic liquids: new forms of active pharmaceutical ingredients with unique, tunable properties. Chemical Reviews, 2023, 123(20): 11894–11953

[6] Matuszek K , Kar M , Pringle J M , MacFarlane D R . Phase change materials for renewable energy storage at intermediate temperatures. Chemical Reviews, 2023, 123(1): 491–514

[7] Saher S , Johnston S , Esther-Kelvin R , Pringle J M , MacFarlane D R , Matuszek K . Trimodal thermal energy storage material for renewable energy applications. Nature, 2024, 636(8043): 622–626

[8] Wu G , Hu C , Cui J , Chen S C , Wang Y Z . Concurrent superhydrophobicity and thermal energy storage of microcapsule with superior thermal stability and durability. ACS Sustainable Chemistry & Engineering, 2017, 5(9): 7759–7767

[9] Florindo C , Romero L , Rintoul I , Branco L C , Marrucho I M . From phase change materials to green solvents: hydrophobic low viscous fatty acid-based deep eutectic solvents. ACS Sustainable Chemistry & Engineering, 2018, 6(3): 3888–3895

[10] Wang N , Tu R , Ma X , Xie Q , Jiang X . Melting behavior of typical thermoplastic materials—an experimental and chemical kinetics study. Journal of Hazardous Materials, 2013, 262: 9–15

[11] Pruksawan S , Chong Y T , Zen W , Loh T J E , Wong J H , Zhu Q , Jin P , Wang F . Sustainable conversion of 3D-printed thermosets into thermoplastics using methacrylic anhydride as a hydrolytic cleavable cross-linker. ACS Applied Polymer Materials, 2024, 6(5): 2777–2788

[12] Trotta J T , Watts A , Wong A R , LaPointe A M , Hillmyer M A , Fors B P . Renewable thermosets and thermoplastics from itaconic acid. ACS Sustainable Chemistry & Engineering, 2019, 7(2): 2691–2701

[13] Niu W , Li Y , Li Q , Wang J , Wang G , Zuo H , She X , Xue Q . Physical and chemical properties of metallurgical coke and its evolution in the blast furnace ironmaking process. Fuel, 2024, 366: 131277

[14] Reddy C D , Zhang Z Q , Msolli S , Guo J , Sridhar N . Impact induced metallurgical and mechanical interlocking in metals. Computational Materials Science, 2021, 192: 110363

[15] Zhou T , Shi H , Sundmacher K . Rational design of ionic liquid phase-change material for efficient thermal energy storage. In: Computer Aided Chemical Engineering 31th eds. European Symposium on Computer Aided Process Engineering. Amsterdam: Elsevier, 2021, 50: 191–196

[16] Wang X , Li W , Luo Z , Wang K , Shah S P . A critical review on phase change materials (PCM) for sustainable and energy efficient building: design, characteristic, performance, and application. Energy and Building, 2022, 260: 111923

[17] Kumar A , Nanda A . In-silico methods of cocrystal screening: a review on tools for rational design of pharmaceutical cocrystals. Journal of Drug Delivery Science and Technology, 2021, 63: 102527

[18] Guo J , Wang A , Liao R , Yan J , Li X , Kang Y , Duan Z , Wang C , Šimůnek J , Yang D . An optimization generator of synthetic DNA fragments for the rational design of environmental tracers. ACS Nano, 2025, 19(9): 9412–9421

[19] Muhieddine M H , Viswanath S K , Armstrong A , Galindo A , Adjiman C S . Model-based solvent selection for the synthesis and crystallisation of pharmaceutical compounds. Chemical Engineering Science, 2022, 264: 118125

[20] Chen J , Zhu F , Qin H , Song Z , Qi Z , Sundmacher K . Rational eutectic solvent design by linking regular solution theory with QSAR modelling. Chemical Engineering Science, 2022, 262: 118042

[21] Hou X J , Yu L Y , He C H , Wu K J . Group and group-interaction contribution method for estimating the melting temperatures of deep eutectic solvents. AIChE Journal, 2022, 68(2): e17408

[22] Wang J , Zhu R , Liu Y , Zhang L . Understanding melt pool characteristics in laser powder bed fusion: an overview of single- and multi-track melt pools for process optimization. Advanced Powder Materials, 2023, 2(4): 100137

[23] Christensen M , Yunker L P E , Adedeji F , Häse F , Roch L M , Gensch T , dos Passos Gomes G , Zepel T , Sigman M S , Aspuru-Guzik A . . Data-science driven autonomous process optimization. Communications Chemistry, 2021, 4(1): 112

[24] Hart G L , Mueller T , Toher C , Curtarolo S . Machine learning for alloys. Nature Reviews: Materials, 2021, 6(8): 730–755

[25] Koutsoukos S , Philippi F , Malaret F , Welton T . A review on machine learning algorithms for the ionic liquid chemical space. Chemical Science, 2021, 12(20): 6820–6843

[26] Zhu X , Polyakov V R , Bajjuri K , Hu H , Maderna A , Tovee C A , Ward S C . Building machine learning small molecule melting points and solubility models using CCDC melting points dataset. Journal of Chemical Information and Modeling, 2023, 63(10): 2948–2959

[27] Liu T , Dong Z , Zhu W , Chen Y , Zhou M , Cui P , Wang Y , Zhu Z . Prediction of the solubility of acid gas hydrogen sulfide in green solvent ionic liquids via quantitative structure-property relationship models based on the molecular structure. ACS Sustainable Chemistry & Engineering, 2023, 11(9): 3917–3931

[28] Risal S , Zhu W , Guillen P , Sun L . Improving phase prediction accuracy for high entropy alloys with machine learning. Computational Materials Science, 2021, 192: 110389

[29] Dijkstra M , Luijten E . From predictive modelling to machine learning and reverse engineering of colloidal self-assembly. Nature Materials, 2021, 20(6): 762–773

[30] Bhandari U , Rafi M R , Zhang C , Yang S . Yield strength prediction of high-entropy alloys using machine learning. Materials Today Communications, 2021, 26: 101871

[31] Mi W , Chen H , Zhu D , Zhang T , Qian F . Melting point prediction of organic molecules by deciphering the chemical structure into a natural language. Chemical Communications, 2021, 57(21): 2633–2636

[32] Hong Q J . Melting temperature prediction via first principles and deep learning. Computational Materials Science, 2022, 214: 111684

[33] Ayres L B , Bandara M , McMillen C D , Pennington W T , Garcia C D . eutXG: a machine-learning model to understand and predict the melting point of novel X-bonded deep eutectic solvents. ACS Sustainable Chemistry & Engineering, 2024, 12(30): 11260–11273

[34] Klimanova O , Miryashkin T , Shapeev A . Accurate melting point prediction through autonomous physics-informed learning. Physical Review B, 2023, 108(18): 184103

[35] Bergström C A S , Norinder U , Luthman K , Artursson P . Molecular descriptors influencing melting point and their role in classification of solid drugs. Journal of Chemical Information and Modeling, 2003, 43(4): 1177–1185

[36] Nigsch F , Bender A , Van Buuren B , Tissen J , Nigsch E , Mitchell J B O . Melting point prediction employing k-nearest neighbor algorithms and genetic parameter optimization. Journal of Chemical Information and Modeling, 2006, 46(6): 2412–2422

[37] Tetko I V , Sushko Y , Novotarskyi S , Patiny L , Kondratov I , Petrenko A E , Charochkina L , Asiri A M . How accurately can we predict the melting points of drug-like compounds. Journal of Chemical Information and Modeling, 2014, 54(12): 3320–3329

[38] Galeazzo T , Shiraiwa M . Predicting glass transition temperature and melting point of organic compounds via machine learning and molecular embeddings. Environmental Science: Atmospheres, 2022, 2(3): 362–374

[39] Dai Z , Wang L , Lu X , Ji X . Melting points of ionic liquids: review and evaluation. Green Energy & Environment, 2024, 9(12): 1802–1811

[40] Benjamin B V L , Luciano A A . Reviewing challenges of predicting protein melting temperature change upon mutation through the full analysis of a highly detailed dataset with high-resolution structures. Molecular Biotechnology, 2021, 63(10): 863–884

[41] Jiang X B , Xiao B B , Lan R , Gu X Y , Sheng H C , Zhang X H . Modeling the melting temperature, melting entropy, and melting enthalpy of freestanding metallic nanoparticles. Materials Chemistry and Physics, 2020, 241: 122280

[42] Semnani A , Keshavarz M H . Using molecular structure for reliable predicting enthalpy of melting of nitroaromatic energetic compounds. Journal of Hazardous Materials, 2010, 178(1-3): 264–272

[43] Acree W Jr , Chickos J S . Phase transition enthalpy measurements of organic compounds: an update of sublimation, vaporization, and fusion enthalpies from 2016 to 2021. Journal of Physical and Chemical Reference Data, 2022, 51(4): 043101

[44] Chickos J S , Gavezzotti A . Sublimation enthalpies of organic compounds: a very large database with a match to crystal structure determinations and a comparison with lattice energies. Crystal Growth & Design, 2019, 19(11): 6566–6576

[45] Fajar A T N , Hanada T , Hartono A D , Goto M . Estimating the phase diagrams of deep eutectic solvents within an extensive chemical space. Communications Chemistry, 2024, 7(1): 27

[46] Breiman L . Random forest. Machine Learning, 2001, 45(1): 5–32

[47] Karpov P , Godin G , Tetko I V . Transformer-CNN: Swiss knife for QSAR modeling and interpretation. Journal of Cheminformatics, 2020, 12(1): 17

[48] Jian Y , Wang Y , Barati Farimani A . Predicting CO₂ absorption in ionic liquids with molecular descriptors and explainable graph neural networks. ACS Sustainable Chemistry & Engineering, 2022, 10(50): 16681–16691

[49] Witanowski M , Biedrzycka Z , Sicinska W , Grabowski Z . A study of solvent polarity and hydrogen bonding effects on the nitrogen NMR shieldings of N-nitramines and Ab initio calculations of the nitrogen shieldings of C-nitro, N-nitro and O-nitro systems. Journal of Molecular Structure, 2002, 602–603: 199–207

[50] Ma̧kosza M , Wojciechowski K . Nucleophilic substitution of hydrogen in heterocyclic chemistry. Chemical Reviews, 2004, 104(5): 2631–2666

RIGHTS & PERMISSIONS

Higher Education Press