Modelling the process of milk-whey syneresis during fresh cheese production by Near InfraRed Spectroscopy

The aim of this work was to test the suitability of Near InfraRed Spectroscopy for modelling the process of milk-whey syneresis during the cheese-making of an Italian fresh cheese, named Crescenza cheese. The applicability of NIRS for this purpose was tested at the lab scale and it was verified at the pilot scale level. Crescenza cheese is produced only from pasteurised whole cow's milk; thermophylic lactic acid bacteria were used as starters. It was chosen as model of fresh cheese because it covers, at this moment, more than 40% of Italian fresh cheese market. Furthermore, whey is a reliable source for a number of high quality and biologically-active proteins, carbohydrates and minerals. Nine lab tests, using 3L of milk for each assay, were carried out simulating the cheese-making process. Six whey sampling points were individuated as critical point of this process, after clotting and curd cutting. NIR spectra were collected, at each sampling point, by an InfraAlyser 500 spectrometer (Bran+Luebbe) in trasflectance mode in the range 1100-2500 nm at room temperature. FT-NIR spectra were also collected by the NIR-Lab 200 spectrometer (Buchi) in trasflectance mode in the range 10000-4000 cm-1 at room temperature. NIR data were processed by applying Principal Component Analysis (PCA) to NIR and FT-NIR data one by one in order to verify their distribution on a plane defined by two PCs. Factorial Discriminant Analysis (FDA) was also applied to classify the whey samplings according the evolution of syneresis process. The loadings plot allowed the identification of absorption bands mainly responsible for the process trend; they were associated to water-water, water-protein and water-fat interactions. PLS1 regression, applied to scores distribution obtained by statistical analyses made before (PCA, FDA), allowed the identification of the principal critical points during syneresis process. The transposition of this model to pilot scale (5 cheese-making processes;150L each), confirmed the suitability of NIRS suggesting the potential use of this technique as on-line routine control. The monitoring of whey syneresis process can indirectly give interesting information about the related cheese-making process, occurring in a more accurate control of cheese production in real-time. The same approach can be extended to monitor the whey syneresis of other dairy production. The collected information could also be used to define the milk whey quality and composition.