Evaluation of microfluidic sample preparation methods for the analysis of milk contaminants

In lab-on-chip for food analysis, dealing with large samples with complex composition (i.e. the matrix) and target analytes potentially in very low concentrations is one of major issues that needs to be tackled. In order to increase the analytical accuracy of a method, it is often required to process the sample before quantification of the chemical specie under examination. In milk case, the presence of fat, proteins and many other components in a complex phase equilibrium poses particular challenges for analysis. Caseins, in particular, are known as blocking agent for surfaces, and therefore may interact with sensor surfaces. Many methods are available for microfluidic processing [1], but they are usually developed for small samples (often blood). Scaling up processing rate to deal with 1-100ml samples in minutes is not practical in most cases. In milk, casein is usually found in micelles the size of about 50-300nm [2], formed by hydrophobic caseins complexed by calcium ions and enclosed by partially hydrophilic k-casein. Since isoelectric point of casein is 4.6, at milk pH there is an intrinsic negative charge, which can be used to separate it from the sample. Other options include size and density separation in microfluidic structure as for instance acoustic separation. In this paper, we present the development of a high-throughput electrophoretic separation system for milk proteins allowing the separation of casein at flow rates in the range of mls/min, by exploiting the intrinsic charge of casein at milk pH (i.e. 6.6). The device is intended as a step for sample preparation to increase the accuracy of detection of contaminants in milk samples. Design, estimated performances and implementation of the device for the experimental evaluation of the approach are reported. We preliminary focused on BSA, lactoglobulins and caseins as model molecules. An electrophoretic mobility of -1.39µm cm V-1 s-1 was measured for casein in conditions as close as possible to real sample (typical ionic strength of milk 0.08M, pH 6.6). The range of flow rate suitable for separation can be estimated by the velocity of electrophoretic drag and flow velocity. Modelling of device provided the evaluation of optimal working conditions and design (potential and flow rate); with a maximum applied potential of 1.4V, the expected fractionation of proteins between the two outlets is reported in Fig. 1, where 0.5 is equivalent to no separation. The device was implemented with platinum electrodes lithographed on silica glass wafers with semi-transparent grid pattern for inspection. In order to increase the throughput, we designed a separation chamber with large cross section (Fig. 2), where electrodes are facing on top and bottom walls of a chamber with width 4mm and channel height 510µm. Electrodes are 28 mm long. Electric field applied drags charged particle toward the upper or lower walls depending on charge. The fluidic chamber was fabricated by 3D printing in PVC and assembled using two shells machined to house the electrode chips (Fig. 3). Experimental demonstration and optimization of the separation procedure is ongoing.