Determination of moisture content in nylon 6,6 by near‐infrared spectroscopy and chemometrics

The effects of moisture on the morphology and mechanical properties of polyamides have been extensively studied by a number of researchers. However, the assessment of water content in the resins has been carried out by thermal or thermogravimetric methods, which are destructive. In the present work partial least‐squares (PLS) calibration models based on near‐infrared (NIR) spectroscopy were produced in order to predict the moisture content of nylon 6,6. Differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and the loss‐on‐drying (LOD) method were used as reference methods. TGA, LOD, DSC, and NIR analysis were performed in parallel, and the obtained data were used for multivariate calibration purposes. Data pretreatment techniques such as derivation and multiplicative scattering correction (MSC) successfully eliminated the baseline offset present in the raw spectra and compensated for differences in thickness and light scattering of the analyzed samples. Calibration models were validated by full cross validation with the help of a test set. A comparison of the prediction ability of PLS models based on pretreated data was also done. NIR spectroscopy is a rapid and nondestructive method for the determination of moisture in recycled nylon. The moisture content can be predicted with a RMSEP = 0.05 wt %. © 2003 Wiley Periodicals, Inc. J Appl Polym Sci 87: 2165–2170, 2003