What question did this study set out to answer?

To develop a simple QSPR model that predicts lower heating value and molar volume using only the molecular formula.

May 15, 2026Industrial & Engineering Chemistry Research0 citationsOpen Access

Simple QSPR for Calculating Combustion Metrics by Hand

DWDarren L. Williams TGTommy R. Gilpin VWVictoria L. White

Key Points

To develop a simple QSPR model that predicts lower heating value and molar volume using only the molecular formula.
A QSPR model was created using the molecular formula to predict combustion properties.
1174 ΔHlow and 4865 Vm values were used for training via linear regression.
Models were validated against literature values, demonstrating high accuracy (R2 = 0.9979 for ΔHlow and R2 = 0.9943 for Vm).
The QSPR model accurately predicts lower heating value with high correlation (R2 = 0.9979).
The molar volume predictions also show high correlation (R2 = 0.9943).
The model's simplicity allows it to be applied across diverse fields, demonstrating versatility in various thermochemical analyses.

Abstract

Quantitative structure–property relationships (QSPRs) provide a framework for predicting molecular properties from structural features. This work proposes a QSPR model of novel simplicity that can be executed using a standard hand-held calculator. It uses only the molecular formula (CaHbNcOdFeClfBrgIh) to predict lower heating value (ΔHlow(kJ/mol)) and molar volume (Vm(mL/mol)). With these two parameters and the molar mass, other properties are easily computed such as density (ρ(g/mL)), specific heat of combustion (ΔHmass(kJ/g)), and combustion energy density (ΔHvol(kJ/mL)). Linear regression was used to train these QSPR models against 1174 ΔHlow and 4865 Vm values from the CRC Handbook of Chemistry and Physics. The regression produced this model for lower heating value: ΔHlow(kJ/mol) = 431.4(1.3)a + 88.85(79)b + 98.2(5.4)c – 161.2(2.8)d – 193.6(4.4)e – 51.5(7.4)f – 36(16)g – 80(26)h. A second regression produced this model for molar volume: Vm(mL/mol) = 16.03(38) + 7.01(6)a + 4.80(3)b + 3.71(23)c + 5.31(11)d + 9.78(11)e + 17.56(23)f + 19.34(40)g + 27.25(88)h. These models perform well against the literature training set values for ΔHlow (R2 = 0.9979) and Vm (R2 = 0.9943). The molecular formula cannot account for structural isomers, which introduces some expected limitations. To demonstrate versatility, this model is applied to many unrelated fields such as macro-nutrient energy content, combustible sources of energy, fire suppression analysis, and the energy content of volumetric fuel blends. Overall, these results demonstrate that a simple QSPR model using only the molecular formula can serve as a computationally inexpensive tool for industrial chemistry use and related thermochemical analyses.

AIに質問

Bookmark

View Full Paper