Dataset on the enthalpy of mixing in binary liquids

Two versions of the dataset are available:
- Version 1 includes data on 375 binary liquids and machine learning predictions in 2415 binary systems, as detailed in our 2024 publication [1].
- Version 2 expands on this with data on 433 binary liquids, including 29 data obtained from ab initio molecular dynamics simulations. It also includes the updated predictions in 2415 binary liquids. Further details can be found in our 2026 publication [2].

Both versions are organized as follows:
- The "Dataset" folder contains data on the enthalpy of mixing collected from Calphad assessments in composition domains where the models are supported by experimental measurements, as well as additional data from ab initio molecular dynamics for Version 2.
- The "Predictions" folder contains machine learning predictions given as Redlich-Kister polynomials in 2415 binary liquids generated by 70 elements. These predictions are compared with those from the Miedema model in both tables and figures where data are also plotted when available.

For more information, and to use this dataset, please refer to these publications:
[1] G. Deffrennes, B. Hallstedt, T. Abe, Q. Bizot, E. Fischer, J-M. Joubert, K. Terayama, and R. Tamura, Data-driven study of the enthalpy of mixing in the liquid phase, Calphad 87 (2024) 102745, https://doi.org/10.1016/j.calphad.2024.102745
[2] Q. Bizot, R. Tamura, G. Deffrennes, Active Learning for Predicting the Enthalpy of Mixing in Binary Liquids Based on Ab Initio Molecular Dynamics, Comput. Mater. Sci. 266 (2026), 114568, https://doi.org/10.1016/j.commatsci.2026.114568