FT-NIR spectroscopy to monitor rennet coagulation in milk with different fat levels

In the last years, the dairy industry has been under high pressure mainly due to the large number of consumers’ requests for healthier products, such as reduced-fat products. Therefore, an improvement of production processes is required in order to maintain high qualitative standards despite the changes in nutritional properties. This work intended to monitor the rennet coagulation process using a FT-NIR system to assess the influence of different factors on the coagulation attitude of milk. An optical fiber was used to continuously monitor 17 coagulation trials set up according to a Box Behnken experimental design with 3 factors (temperature, milk fat content and pH) and 3 levels. From each trial, a dataset (40x2203) was obtained and analyzed, after SNV pretreatment, using interval-PCA in three spectral ranges. The same approach was performed also on a global dataset, containing all the trials (680x2203). From the results, it can be inferred that PC1 in the 12500–9200 cm-1 range shows the liquid-like behavior of milk before coagulation, while in the 6463–5824 cm-1 range it describes the curd formation. Besides, PC2 in 9199–7180 cm-1 range allows detecting and calculating the transition-phase time. Through the Response Surface Methodology it was possible to find out that all the three considered factors had a significant influence on the transition time. In particular, higher coagulation temperature (p<0.01) and milk fat content (p<0.05) decreased transition time, which was instead increased by the increasing of pH (p<0.01). This work demonstrated that interval-PCA models are capable of monitoring in-line the rennet coagulation and estimating the influence of temperature, milk fat content and pH on the process.