Super-resolution microscopy and empirically validated autocorrelation image analysis discriminates microstructures of dairy derived gels

The food industry must capitalise on advancing technologies in order to optimise the potential from emerging ingredient technologies. These can aid in product optimisation and provide quantitative empirical data to which there is a fundamental physical understanding. Super-resolution microscopy provides a tool to characterise the microstructure of complex colloidal materials under near native conditions. Coherent Anti-Stokes Raman Scattering (CARS) microscopy was used to show the presence of fluorescent dye required for imaging does not affect gel microstructure and super-resolution Stimulated Emission Depletion (STED) microscopy is used to image four dairy derived gels. Image analysis has been developed based on 2D spatial autocorrelation, and a model that extracts parameters corresponding to a typical length of the protein domains and the inter pore distance. The model has been empirically validated through the use of generated images to show the fitting parameters relate to precise physical features. The fractal dimension is extracted from Fourier space analysis. The combination of STED microscopy and image analysis is sensitive enough to significantly differentiate samples based on whether gels were made from fresh or reconstituted milk, and whether gelation was induced through acidification or rennet addition. Rheometry shows that the samples exhibit different macroscopic behaviours, and these differences become increasingly significant with time. Samples can be differentiated earlier in the gelation process with imaging as compared to rheometry. This highlights the potential of STED imaging and image analysis to characterise the size of protein domains, pore spacing and the fractal dimensions of microstructures to aid product optimisation.