Study On Rapid Detection Of Adulterated Milk By Low-field Nuclear Magnetic Resonance Technology

The quality of dairy products has been a prominent problem with the development of dairy industry. The rapid, objective and effective technology to detect adulterated milk is the urgent needs for a modern dairy industry and the relevant regulatory authorities, and it has also been one of the research hot spots in the dairy industry.The low-field nuclear magnetic resonance(LF-NMR) technology has many advantages, such as rapid detection, friendly to environment and easy to operate. It has a special advantage in researching food nature that is closely related to water and become an important means that research the status and distribution of water in food.In this study, different forms of adulterated milk samples and milk products were chosen as the research objects. Preliminary exploration has been done for the application of LF-NMR technology in rapid testing adulterated milk and three chemometrics methods were used to analyze and process the testing dates, in order to provide some experimental basis for its application in detecting adulterated milk for dairy products. The main research work is as follows:(1) Rapid detection of adulterated milk based on LF-NMRPure milk samples and different forms of adulterated milk samples were analyzed by the LF-NMR machine. Milk samples adulterated with water, salt, soybean milk, urea or melamine, reconstituted milk and UHT sterilized milk samples were tested. The testing datas were analyzed and processed by chemometrics methods. The results showed that pure milk and different kinds of adulterated milk were distinguished in Principle Component Analysis(PCA) scorer plot. There were also effective identifications among pure milk, reconstituted milk and their mixed milk. The application of LF-NMR in adulterated milk with different additive proportions showed the similar results, too. The Soft Independent Modeling of Class Analogy/Statistical Isolinear Multi-Category Analogy (SIMCA) discrimination model of milk was set up. Six kinds of adulterated milk samples were prepared to validate the model, while five replicates of each sample were prepared. The discrimination accuracies of six kinds of adulterated milk by SIMCA model were100%. The prediction models of adulterants in qualitative analysis based on Partial Least Squares (PLS) were established. The discrimination accuracies of milk adulterated with water, soybean milk, urea, melamine, salt, reconstituted milk was respectively100%,100%,100%,100%,56%,77%.(2) Study on the detection limits of LF-NMR on several adulterants in milkDifferent forms of adulterated milk samples were analyzed by the LF-NMR machine. Raw milk adulterated with water, soybean milk, urea or melamine, reconstituted milk and raw milk samples were tested. The testing dates were analyzed and processed by chemometrics methods. After the testing dates were analyzed and processed by PCA, the detection limits of adulterant in milk can be determined according to its distribution on PCA scorer plot. The detection limits are as follows:water3%-5%, soybean milk5%-7%, reconstituted milk15%-20%, urea0.02%-0.04%, melamine0.0062%-0.0124%. The SIMCA model of raw milk was built to determine the detection limits of adulterant in milk. The detection limits are as follows:water3%-5%, soybean milk5%-7%, reconstituted milk15%-20%, urea0.01%-0.02%, melamine0.0062%-0.0124%. The PLS-DA model of adulterated milk was built to determine the detection limits of adulterant in milk. The detection limits are as follows:water1%-3%, soybean milk1%-3%, reconstituted milk15%-20%, urea0.01%-0.02%, melamine0.0031%-0.0062%.(3) The establishment of LF-NMR datebase of raw milk and some common adulterated milkThe database of raw milk and three different adulterated milk(adulterated with water, adulterated with soybean milk, adulterated with reconstituted milk)was established initially. The standard drtermination maps of adulterants were obtained(water,soyben milk, reconstituted milk) in milk.