Theoretical Study of Absolute Entropy, Entropy of Formation, and Gibbs Energy of Formation of Two Novel Macromolecules Obtained by the Solid State

TL;DR

This paper introduces a new method for synthesizing diimidetricarboxylic acids using solid-state reactions and evaluates their thermochemical properties using machine learning.

Contribution

The study demonstrates that machine learning can predict thermochemical properties as accurately as traditional DFT methods.

Findings

A solid-state synthesis route for diimidetricarboxylic acids was developed with good yields and easy purification.

Thermochemical properties like enthalpy, entropy, and Gibbs energy of formation were calculated using machine learning.

Machine learning predictions matched the accuracy of DFT methods for these properties.

Abstract

In this study, we present a solid-state reaction that
enables the
novel synthesis of diimidetricarboxylic acids (DITAs) in good yields
without using harmful solvents. The synthesis route offers advantages,
including satisfactory efficiency and easy product purification by
sublimation. The reaction utilizes trimellitic anhydride (TMA) with
3,4- and 3,5-diaminobenzoic acid to obtain the desired DITAs. Products
were characterized using differential scanning calorimetry (DSC),
Fourier transform infrared spectroscopy (FTIR), proton nuclear magnetic
resonance (1H NMR), and mass spectrometry (MS). The enthalpies
of formation in the gas phase of DITAs (Δ
f

H°) were calculated, occupying absolute
entropies (S°), entropies of formation (Δ
f

S°), and Gibbs energies
of formation (Δ
f

G°). These properties were determined using the hybrid method
B3LYP functional and the computational technique machine learning.
Our theoretical studies revealed that it is possible to predict thermochemical
properties with machine learning as accurately as with density functional
theory (DFT).

Figures