Premium
Multivariate Calibration of Semi‐Synthetic Data Sets: Gun Powder Analysis
Author(s) -
Ouellet Nathalie,
Lussier LouisSimon,
Beaupré France
Publication year - 2003
Publication title -
propellants, explosives, pyrotechnics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.56
H-Index - 65
eISSN - 1521-4087
pISSN - 0721-3115
DOI - 10.1002/prep.200390020
Subject(s) - propellant , calibration , multivariate statistics , partial least squares regression , materials science , thermoplastic elastomer , elastomer , analytical chemistry (journal) , chemistry , chromatography , composite material , statistics , polymer , mathematics , organic chemistry , copolymer
Abstract Quantitative analysis of multi‐component mixtures such as propellant powders is not trivial since it usually requires separation of the mixture constituents. Multivariate calibration combined to the use of semi‐synthetic data sets can eliminate the need for standard solutions preparation, and therefore allow the rapid determination of mixtures provided no intermolecular interactions occur in the systems. Multivariate compositional analyses of FTIR spectra of low‐vulnerability (LOVA), high‐energy low‐vulnerability (HELOVA) and energetic thermoplastic elastomer (ETPE) propellant powder systems were performed using the partial least‐squares (PLS) regression algorithm. All constituents except ethyl centralite (EC) were quantified. Concentrations were predicted within 1% error for the major component (1,3,5‐trinitro‐1,3,5‐triazacyclohexane or RDX), and within 5% error for the minor components (between 12 and 2% nominally by weight). LOVA, HELOVA, and ETPE gun powder samples concentrations were estimated and compared to expected compositions.