The Parr 1341 plain jacket calorimeter was assembled and calibrated with certificated benzoic acid pellets (NIST Standard Material Reference 39j), the energy equivalent, ε(calor) = (9930.1 ± 2.2) J K⁻¹, was obtained from five standardization tests. The static bomb adiabatic calorimeter in the oxygen atmosphere was tested with acetanilide and after was used to calculate the standard molar enthalpies of formation in solid phase at 298.15 K, of 4,4'-methylenedi(phenylene isocyanate) and of three polyamide-imides chemically synthesized.

The standard molar enthalpy of formation for trimellitic acid (TMAc) in the crystalline phase at 298.15 K, Δ_fH_m°(cr), was calculated experimentally from the enthalpy of combustion through combustion calorimetry experiments. Likewise, the standard molar enthalpy of sublimation was determined from the standard molar enthalpy of fusion and from the standard molar enthalpy of vaporization from differential scanning calorimetry and thermogravimetry, respectively. Subsequently, the standard molar enthalpies of formation in the gas-phase at 298.15 K, Δ_fH_m°(g), were calculated. The enthalpies of formation for TMAc, hemimellitic, and trimesic acids were predicted using multiple linear regression (MLR) with a nonreplacement evaluation technique. MLR was applied to the data set that allowed estimating these thermochemical properties with an R² greater than 0.99. This model was used to compare the predicted and experimental results for benzene carboxylic acids.

Experimentally, the standard molar enthalpy of formation in the crystalline phase at 298.15 K, Δ_fH_m°(cr) for 7-methoxy-4-methylcoumarin (7M4MC) was calculated by traditional linear regression, which was obtained by combustion calorimetry. Similarly, the standard molar enthalpy of sublimation was determined through the standard molar enthalpy of fusion and by the standard molar enthalpy of vaporization, from differential scanning calorimetry and thermogravimetry, respectively; lately using these results, the standard molar enthalpy of formation in the gas phase was calculated at 298.15 K, Δ_fH_m°(g). In addition ML was used to predict the standard molar enthalpy of formation in the gas phase for the 7M4MC, constructing an experimental data set containing three kinds of functional groups: esters, coumarins, and aromatic compounds. The procedure was performed by using multiple linear regression algorithms and stochastic gradient descent with a R² of 0.99. The obtained models were used to compare those predicted values versus experimental for coumarins, resulting in an average error rate of 9.0%. Likewise, four homodesmic reactions were proposed and predicted with the multiple linear regression algorithm of ML obtaining good results.

The prediction of thermochemical properties such as the standard enthalpy of combustion is essential for the design and evaluation of energetic materials. In this study, the prediction of this thermochemical property is proposed through a QSPR strategy that combines machine learning and chemical graph theory. The data set consisted of 3477 organic compounds. SMILES codes were used for each molecule to construct their molecular graphs, from which topological indices such as Estrada, Wiener, and Gutman, as well as centrality measures, were calculated. These descriptors served as predictors in supervised learning models, with tree-based ensemble models showing the best performance. The best-performing model, random forest, achieved the following metrics on the test set: R ² = 0.9810, MAE = 287.5988 kJ·mol^–1, MAPE = 0.1048, RMSE = 551.9050 kJ·mol^–1, and RMSLE = 0.1933. Interpretability analysis using SHAP confirmed that the Estrada and Gutman indices were the most influential variables in the predictions. In addition, the same random forest model was trained using 210 molecular descriptors obtained from RDKit, yielding slightly better metrics: R ² = 0.9927, MAE = 142.2272 kJ·mol^–1, MAPE = 0.0484, and RMSE = 342.0464 kJ·mol^–1, and RMSLE = 0.1172. Moreover, specific models were developed for different families of compounds, achieving R ² ≈ 0.99 in all cases. Finally, a clustering analysis using the K-Means algorithm in the space defined by the topological indices enabled the identification of latent molecular patterns, providing a novel framework for organizing and analyzing chemical space. This work demonstrates the potential of combining supervised and unsupervised learning methods with chemical graph theory to enable accurate, robust, and scalable prediction of thermochemical properties such as combustion enthalpy.