15C-28 |
Partial least squares method for analysis of milkfat by ATR technique using mid infrared spectroscopy |
A. TAY, S. S. Krishnan, R. K. Singh, and J. P. Gore. Mid Infrared Sensing Diagnostics and Control Consortium, Purdue University, Food Science, 1160 Food Science Bldg., West Lafayette, IN 47907 Mid infrared spectroscopy has been an attractive method due to high sensitivity compared to near infrared (NIR) spectroscopy. In this study a partial least squares (PLS) model was developed to analyze the fat content in milk. PLS is a sophisticated chemometric method that becomes central focus for analysis purposes. Milkfat analyzes conventionally is done by Babcock and Gerber methods. Mid infrared spectroscopy can be used to analyze the fat content of milk and cream. Attenuated total reflectance (ATR) sampling technique is easy to use and fast compared to transmittance sampling technique. The objective of this study is to develop a reliable, accurate and robust chemometric model for analysis of milkfat in milk. Milk samples with known fat concentrations were placed on a multi bounce ZnSe ATR crystal. 64 scans were collected at 4 cm-1 resolution using a Nicolet Nexus 670 infrared spectrometer equipped with DTGS detector. The background was collected before every sample was measured. The chemometric analysis was done by using the TQ analyst 6.0 software program. The PLS model based on the C=O stretching region (1730-1760cm-1) was developed. The correlation coefficient of 0.99 was obtained for fat quantification. The results from this study indicate that PLS model can be used to quantify milkfat in the mid infrared region. The ATR technique is fast, easy to use and can be implemented for in line monitoring of fat composition.
Session 15C, Dairy Foods
|