Avaliação do uso de laser-induced breakdown spectroscopy (LIBS) e wavelength dispersive x- ray fluorescence (WDXRF) na análise direta de couros naturais e sintéticos

Abstract

The aim of this doctoral thesis was the development of analytical procedures for the qualitative and quantitative analysis of leather samples using LIBS and WDXRF techniques for direct analysis of solid samples. The first part of this work reports the development of classification models for the prediction of quality parameters of bovine and ovine leather tanned with Cr. In total, 375 leather samples were analyzed by LIBS and three classification models were developed using reference values obtained by physical-mechanical tests (official methods to certify the quality of leathers): K-nearest neighbors (KNN), soft independent modelling of class analogy (SIMCA) and discriminant analysis by least squares method (PLS-DA). The models had satisfactory results with correct prediction percentages ranging from 75 (SIMCA) to 80% (PLS-DA) for the calibration samples and from 72 (SIMCA) to 81% (KNN) for the validation samples. The second part deals with the establishment of a sample preparation procedure for leathers, in which digestion with HNO3 and H2O2 was used. In the experimental procedure 60 samples were evaluated and the concentration range of the elements determined using inductively coupled plasma optical emission spectrometry (ICP OES) were as follows (mg kg-1): Al: 94-1,011; As: <LOD-13; Ba: <LOD-40; Ca: <LOD-3,348; Cr: 10-23,407; Cu: <LOD-15; Fe: 32-831; Mg: <LOD-1,126; Ni: <LOD-5; Pb: <LOD-25; Sb: 31-318; Sr: <LOD-9; Ti: <LOD-6; Zn: <LOD-186 and Cd was at concentration levels below the limit of detection: (<LOD). The third part describes the determination of Al, Ca, Cr, Fe, Ti and Zn using WDXRF. Calibration models were developed with reference concentrations determined by ICP OES obtained in the previous study. In spite of attempts made to develop calibration models for the direct determination of Al, Ca, Cr, Fe, Ti and Zn, it was only possible to obtain satisfactory calibration models for Cr with the following figures of merit: a) calibration curve: [Cr]= (43.9 ± 1.0) x + (19,583 ± 14,600); b) linear correlation coefficient of 0.9901; c) predictive capacity based on the standard validation error (SEV) of 628 mg kg-1; c) LOD of 30 and limit of quantification (LOQ) of 100 mg kg-1; d) type of signal used: height; e) concentration range of Cr determined: SEV at 21,638 mg kg-1. The fourth part proposed the evaluation of synthetic leather and nine samples were analyzed using ICP OES, LIBS and WDXRF combined with principal component analysis (PCA). The concentrations of the elements determined in samples of synthetic leather by ICP OES were as follows (mg kg-1): Ca: 770-112,233; Cr: 13-24,732; Mg: 165-18,834; Ba: <LOD-4,606; Pb: 2.8-2,484; Al: 147-1,607; Fe: 130-1,015; Zn: <LOD-528; Sb: 22-262; Cu: <LOD-163; As: <LOD-121; Sr: <LOD-104; Ti: <LOD-48; Cd: 4.1-27; Mn: <LOD-27, Co: <LOD-3.9 and Ni was at concentration levels <LOD in all samples. The PCA evaluated the samples regarding the chemical composition obtained quantitatively by the concentrations of ICP OES and qualitatively by the LIBS or WDXRF spectra. The proposed methods present high analytical frequency, simplicity and are in accordance with the principles of green chemistry.