Detection of Milk Adulteration with Volatile Fingerprints

The odors of cow, buffalo, and goat milk differ due to variations in their volatile organic compound (VOCs) profiles. Buffalo and goat milk, which are more expensive than cow milk, can be adulterated by mixing with inexpensive cow's milk. The detection of adulteration in expensive milks with cow's milk poses significant methodological challenges, given the high similarity in composition and physicochemical properties exhibited by these two milks. Furthermore, the scarcity of robust and universally applicable markers leads to prolonged and costly analytical procedures, thereby complicating fast, easy and cost-effective quantification, particularly at low adulteration levels. In this study, an environmentally friendly easy, rapid and cost-effective milk adulteration detection method was proposed based on the VOC fingerprinting. For this purpose, only the retention times (R t ) and area values of all detected VOCs in milk types were compared without applying any qualification or quantification step. Sampling parameters were optimized using Taguchi Experimental Design, and the optimized method was applied to all adulterated and non-adulterated samples. Varying statistical methods were used to identify possible VOC discriminative retention times (dRts) of indicators for the identification of milk type and adulteration rate. The PCA results reveal that the milk types as 85.38% of the total variance. Subsequent one-versus-rest receiver operating characteristic (ROC) analysis validated the discriminative power of the selected volatile markers, achieving classification accuracies of 72.7%, 86.9%, and 89.2% for the authentication of cow, goat, and buffalo milks, respectively. The proposed method appears to demonstrate acceptable analytical sensitivity and could be potentially be used to estimate the adulteration percentage of cow milk in buffalo and goat milks at trace levels, reaching down to 0.2%.