Making news headlines this week the 2008 baby milk scandal similarly themes as the 2007 pet food recalls and part of a larger topic of poor Chinese business practises. In this scandal, milk has been found to be deliberately spiked with the compound melamine as a result of which many infants have become ill and some even tragically died. Chinese farmers dilute milk they sell to processing plants for an additional profit and as milk is primarily tested for nitrogen content, the farmers make up for the loss of nitrogen by adding a very nitrogen-rich compound to the milk which is melamine.

The journalists are very brief on the chemical details but trying to get some chemical background information is not that easy and certainly the scientific literature is not very helpfull (my library does not exactly subscribe to such journals as journal of food protection or journal of diary science). So with the combined power of search engines, wikipedia and some educated guesses this is a milk / melamine update.

The nitrogen content of milk is made up of casein protein and whey protein both of which give the nutritional value. Non-protein nitrogen sources are ammonia, urea, creatine, creatinine, uric acid, orotic acid, peptides, hippuric acid and several amino acids. A typical milk sample will contain 87% water, true protein (TP) 3% , crude protein (CP) 3.10% and casein 2.6% link. These figures are surprisingly stable regardless of cow brand or cow nationality.

The melamine contamination scandal concerns both regular milk and milk powder. One liquid milk sample was found to contain 8.4 milligrams of melamine per kg Link. With a crude protein content of 3.1% per kilogram milk, it is theoretically possible to disguise up to 40% dilution with an added 3 gram of melamine which is also the maximum solubility of melamine in water.

The Kjeldahl method is very indiscriminate and simply tests for all nitrogen present. Crude protein (CP) is the Kjeldahl N multiplied by 6.38 and includes protein and non-protein sources. Apparently the Kjeldal method is is the only method used in Chinese quality labs and that's where all the trouble started. At least two methods exist that specifically target TP and not used by the Chinese: Udy dye colorimetry and infrared reflectance.

The Udy dye (Udy is the name of the company selling the method) (Link) is the
azo dye Acid Orange 12 and a close cousin of methyl orange which specifically binds to protein causing it to precipitate. The remaining azo dye is then determined by colorimetry. The website states the method will discriminate against any non-protein source but it would be interesting to examine if it interacts with melamine.

Another option is examination of melamine content directly, preferably by one cheap and simple method that can also be used those people affected the worst. A very old method published in 1954 is simple paper chromatography (J.E. Milks DOI (eluent BuOH/EtOH/water 4:1:1 staining with dimethylaminobenzaldehyde) but that method cannot be very sensitive. The US FDA has a protocol ased on GC-MS for the quantification of melamine, ammeline, ammelide and cyanuric Acid link. Other methods based exist based on LC-MS. On the other side of the technology spectrum (and lot easier to use than any MS technique), at least two companies supply tests based on enzyme-linked immunosorbent assay ELISA: Romer labs link and Abraxis link. All these methods have been developed in response of the pet food scandal but not necessarily optimized for milk. There is much ado about molecular sensors in the scientific literature but surprisingly melamine is not a target.

Even if Chinese milk quality control will eventually catch up with melamine the nagging question remains why the Chinese milk plants apparently did not simple test for water content of milk supplies coming into the plant. Measuring water content is easy and although drying out seems straightforward (as anyone who left pan of milk on the stove for too long a can attest to)
freezing point depression appears to be the preferred method Link . The addition of 10% water to milk will universally reduce the freezing point of the mixture from -0.510°C to 0.459°C. And even if the plant did not check water content in the laboratory beforehand the manager would be surprised to find much less milk powder production than he had bargained for.

Update (25 September 2008). Reuters report that poor quality cow feed (causing low nitrogen content) is at the heart of the problem (Link). Also cited local costs for testing for melamine in milk: 145 US dollars.
Correction: Changed calorimetry to colorimetry (many thanks to vigilant reader).
Correction: taking 3.1% crude protein and 3 g/L solubility it is possible to dilute up to 40% (again thanks to reader comment)