Predicción de solubilidad del ácido poliláctico en solventes convencionales utilizando COSMO-RS.
Loading...
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
La solubilidad es una medida ampliamente utilizada en diversos procesos químicos e industriales. Actualmente, para identificar solutos y solventes compatibles se emplean parámetros empíricos, los cuales solo entregan información cualitativa a temperatura ambiente. Debido a esta limitación, surgió el uso de COSMO-RS para predecir solubilidades de forma cuantitativa y a distintas temperaturas. El objetivo de este estudio es establecer un protocolo de simulación que entregue predicciones precisas a bajo costo computacional.
Se llevaron a cabo pruebas experimentales de solubilidad utilizando ácido poliláctico (PLA) con diferentes solventes. En cloroformo, se obtuvo una solubilidad de 0,161 (𝑚_𝑠𝑜𝑙𝑢𝑡𝑜/𝑚_𝑠𝑜𝑙𝑣𝑒𝑛𝑡𝑒) a 30 °C, 0,260 a 40 °C y 0,301 a 50 °C. Por otro lado, a 40 °C, la solubilidad fue de 0,0497 en 1,4-dioxano y 0,130 en tetrahidrofurano. Las simulaciones computacionales se llevaron a cabo modelando oligómeros de PLA en TmoleX. Se obtuvieron confórmeros representativos con COSMOconf y, finalmente, se realizaron simulaciones de solubilidad en COSMOtherm, utilizando como referencia la solubilidad del PLA en cloroformo a 40 °C para calcular la energía libre de fusión del polímero.
En COSMO-RS se evaluó la precisión de las predicciones de solubilidad respecto a los resultados experimentales, analizando los efectos de diversos parámetros relacionados con la representación geométrica del polímero y las referencias experimentales empleadas, además de explorar la posibilidad de extrapolar las predicciones a otros sistemas.
Finalmente, los parámetros óptimos determinados fueron: el uso de la plantilla BP-TZVP-COSMO para la selección de confórmeros, un oligómero de dos monómeros, la desactivación de grupos terminales y la utilización de la geometría del PLLA como modelo inicial, por representar de forma más precisa el polímero empleado. La aplicación de estos parámetros reportó una mejora, reduciendo el error medio absoluto final a un 13,9 %. Asimismo, se definió el uso de referencias de solubilidad a altas temperaturas para lograr una mayor precisión en las predicciones.
Además, para analizar un solo solvente y su comportamiento frente a la temperatura, es suficiente utilizar el oligómero inicial. En cambio, para determinar solubilidades en varios solventes, es necesario emplear el conjunto completo de confórmeros.
Solubility is a measure widely used in various chemical and industrial processes. Currently, empirical parameters are used to find compatible solutes and solvents, which provide only qualitative information at room temperature. From this limitation arose the use of COSMO-RS, to be able to predict solubilities quantitatively and at different temperatures. The purpose of this study is to establish a simulation protocol that delivers accurate predictions at a low computational cost. Solubility tests were carried out using polylactic acid (PLA) as raw material with different solvents. In chloroform a solubility of 0,161 (𝑚_𝑠𝑜𝑙𝑢𝑡𝑒𝑚_𝑠𝑜𝑙𝑣𝑒𝑛𝑡⁄) at 30 °C, 0,260 at 40 °C and 0,301 a 50 °C. On the other hand, at 40 °C a solubility of 0,0497 and 0,130 were obtained for 1,4-dioxane and tetrahydrofuran respectively. Computational simulations were performed by modeling PLA oligomers in TmoleX, searching for representative conformers in COSMOconf, to finally perform solubility simulations in COSMOtherm. Using as a reference the solubility of PLA in chloroform at 40 °C to obtain the polymer’s free energy of fusion. COSMO-RS sought to evaluate the accuracy of solubility predictions with respect to experimental results, examining the effects of various parameters related to the geometric representation of the polymer and the experimental references used, in addition to evaluating the possibility of extrapolating predictions to other systems. Finally, the parameters evaluated and established were the use of the BP-TZVP-COSMO template as a conformers selection method, a 2-monomer oligomer, deactivation of terminal groups and the use of PLLA geometry as the initial model, being a more accurate representation of the polymer used. Applying these parameters results in an improvement in predictions, reducing the final mean absolute error to 13,9%. Finally, the use of high temperature solubility references was defined to achieve greater prediction accuracy. Furthermore, to analyze only one solvent and study its temperature behavior, it is sufficient to use the initial oligomer. To find solubilities with various solvents, it is necessary to use the entire set of conformers.
Solubility is a measure widely used in various chemical and industrial processes. Currently, empirical parameters are used to find compatible solutes and solvents, which provide only qualitative information at room temperature. From this limitation arose the use of COSMO-RS, to be able to predict solubilities quantitatively and at different temperatures. The purpose of this study is to establish a simulation protocol that delivers accurate predictions at a low computational cost. Solubility tests were carried out using polylactic acid (PLA) as raw material with different solvents. In chloroform a solubility of 0,161 (𝑚_𝑠𝑜𝑙𝑢𝑡𝑒𝑚_𝑠𝑜𝑙𝑣𝑒𝑛𝑡⁄) at 30 °C, 0,260 at 40 °C and 0,301 a 50 °C. On the other hand, at 40 °C a solubility of 0,0497 and 0,130 were obtained for 1,4-dioxane and tetrahydrofuran respectively. Computational simulations were performed by modeling PLA oligomers in TmoleX, searching for representative conformers in COSMOconf, to finally perform solubility simulations in COSMOtherm. Using as a reference the solubility of PLA in chloroform at 40 °C to obtain the polymer’s free energy of fusion. COSMO-RS sought to evaluate the accuracy of solubility predictions with respect to experimental results, examining the effects of various parameters related to the geometric representation of the polymer and the experimental references used, in addition to evaluating the possibility of extrapolating predictions to other systems. Finally, the parameters evaluated and established were the use of the BP-TZVP-COSMO template as a conformers selection method, a 2-monomer oligomer, deactivation of terminal groups and the use of PLLA geometry as the initial model, being a more accurate representation of the polymer used. Applying these parameters results in an improvement in predictions, reducing the final mean absolute error to 13,9%. Finally, the use of high temperature solubility references was defined to achieve greater prediction accuracy. Furthermore, to analyze only one solvent and study its temperature behavior, it is sufficient to use the initial oligomer. To find solubilities with various solvents, it is necessary to use the entire set of conformers.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Químico/a.
Keywords
Solubilidad, Termodinámica, Química cuántica