ΔvapHm-RDKit: Standard Vaporization Enthalpy Database

<p>We present a database to Predict Standard Vaporization Enthalpies through Intelligible Features.</p><p>This database presents a variety of molecules of diverse chemical families with known experimental&nbsp;Standard Vaporization Enthalpy&nbsp;(Δvap<i>H</i>m°) values. Entries were collected from the <a href="https://webbook.nist.gov/chemistry/">NIST WebBook Library</a>&nbsp;through a web scraping routine, considering only&nbsp;Δvap<i>H</i>m° experimental values, ranging from 15 kJ mol-1&nbsp;to 400 kJ mol-1. Experimental value scrapping was automatically made, without duplicates, by selecting the first listed value (either a from single or average of values as determined by NIST). The database includes each value's original literature, along with experimental uncertainties (when available).&nbsp;</p><p>Each compound is represented by its molecular family, SMILES string and respective InChlKey &amp;&nbsp;InChl identifiers. The curated database consists of 1781 molecules over 15 different chemical families. The chemical descriptors used to feed our model were generated using <a href="http://www.rdkit.org/">RDKit</a>&nbsp;version 2022.09.4, running on top of Python 3.9. Descriptors were calculated&nbsp;from the "MolFromSmiles" function in "RDKIT.Chem" as descriptors with non-numerical values were removed. The descriptors encode significant chemical information and are used to present physicochemical characteristics of compounds, building a&nbsp;relationship between structure and&nbsp;Δvap<i>H</i>m° determinations.</p><p>Through Machine Learning regression algorithms, we have developed&nbsp;models that were able to make accurate&nbsp;Δvap<i>H</i>m°&nbsp;predictions, based on the information encoded in each chemical feature. Our models can be accessed through <a href="https://github.com/jfcaetano/EnthVapML">GitHub</a>.</p>