Título:	Polynomial model with asymmetric curvature induced by excess for mixture design: optimization of green anthocyanin extraction from Cenchrus purpureus using a standardless chromatographic response.
Autoria:	ADRIANO , L. H. C. GERALDO, P. A. NOGUEIRA, C. P. MACHADO, J. C. OLIVEIRA, M. A. L. de
Afiliação:	LUIZ HENRIQUE CANTARINO ADRIANO, UNIVERSIDADE FEDERAL DE JUIZ DE FORA; PATRÍCIA ABRANCHES GERALDO, UNIVERSIDADE FEDERAL DE JUIZ DE FORA; CECÍLIA PINTO NOGUEIRA, UNIVERSIDADE ESTADUAL DE CAMPINAS; JUAREZ CAMPOLINA MACHADO, CNPGL; MARCONE AUGUSTO LEAL DE OLIVEIRA, UNIVERSIDADE FEDERAL DE JUIZ DE FORA.
Ano de publicação:	2026
Referência:	Chemometrics and Intelligent Laboratory Systems, v. 274, 105731, 2026.
Conteúdo:	Canonical Scheff´e mixture models are widely used for binary systems, but their inherent assumption of symmetric component behavior restricts their ability to describe processes in which mixture components fulfill distinct physicochemical roles. This limitation is especially relevant for solvent-based extraction systems involving water and ethanol. In this study, the extraction of anthocyanins from leaves of Cenchrus purpureus. using water:ethanol mixtures were investigated, and an alternative modeling framework, named Asymmetric Curvature Induced by Excess (ACIE), was proposed to overcome this conceptual constraint. Experimental and simulated approach were evaluated in this work. Extractions were carried out by static maceration at room temperature using only water and ethanol, in accordance with Green Chemistry principles. The extracts were analyzed by HPLC-DAD, and a semi-quantitative approach was adopted in which the response variable was defined as the sum of the chromatographic peak areas of the two most reproducible anthocyanin signals. Canonical Scheff´e linear and quadratic models were unable to adequately represent the system, exhibiting lack of fit and limited predictive capability. In contrast, ACIE, by incorporating asymmetric curvatures for modeling the excess of each component, achieved superior statistical performance and a more realistic representation of the response surface. The model predicted an optimal extraction region between 82 and 92% ethanol, consistent with the distinct roles of ethanol in promoting cell wall disruption and water in anthocyanin solubilization. These results show that ACIE effectively addresses a fundamental limitation of Scheff´e-based models relative to asymmetries in binary mixtures and provides a flexible, chemically interpretable framework that may be applicable to other systems characterized by asymmetric component functionality.
Thesagro:	Capim Elefante Antocianina
Palavras-chave:	Mistura de Scheffé Modelagem
Digital Object Identifier:	https://doi.org/10.1016/j.chemolab.2026.105731
Tipo do material:	Artigo de periódico
Acesso:	openAccess
Aparece nas coleções:	Artigo em periódico indexado (CNPGL)