A Fast Method for Determination of Melamine in Liquid Milk and Milk Powder by HPLC With UV Detection

Melamine discovered in baby formula and products containing milk has led to one of the largest worldwide food recalls in history. Some manufacturers in China illegally added melamine, an ingredient used in plastics manufacturing, to increase the apparent protein content of low-quality milk. This chemical, shown in Figure 1, can form an insoluble compound in the body, causing kidney stones that can be deadly in infants.

Figure 1 - Structure of melamine.

Factories that produce milk products or use milk products imported from China will need to begin testing for melamine. Melamine has been quantified using enzyme immunoassay (EIA), GC-MS, LC-MS, and HPLC with UV detection.1-4 The Chinese government has approved four methods for melamine testing in milk. Initially three methods were released, including GC-MS, LC-MS, and HPLC-UV methods.5 An HPLC method using a strong cation-exchange (SCX) analytical column and UV detection for liquid milk only was released more recently.6 Derivatization of milk for GC analysis is labor intensive, and the cost of operation of MS detection can be high. The HPLC-UV method is laborious, requiring solidphase extraction (SPE) followed by solvent evaporation. The ion-pairing compound required for this reversed-phase (RP) separation can increase noise and damage columns, and is incompatible with MS detection. The SCX separation has difficulties with interferences, especially when analyzing powdered milk or milk with additives. A simpler, more robust method for testing a wider variety of milk products is necessary.

This article compares results of experiments using the current C18 method and a new, rapid HPLC method performed using an Acclaim® Mixed-Mode WCX-1 column (Dionex Corp., Sunnyvale, CA).7 The column features a mixed-mode silica-based packing material that incorporates both hydrophobic and weak cation-exchange properties, and demonstrates high potential for separating samples that contain a mixture of ionic and neutral compounds, without requiring ion-pairing compounds. Melamine analysis of either liquid or powdered milk is completed within 10 min, about half the time required for the C18 separation. The background noise is significantly reduced using the WCX-1 column, and the mobile phase is also compatible with MS detection. A new sample preparation method that eliminates the SPE and evaporation steps is also demonstrated.

Experimental

Instrumentation

An UltiMate® 3000 HPLC system (Dionex)was used for all separations, and consisted of an HPG-3400A pump, WPS-3000TSL autosampler, TCC-3000 thermostated column compartment, and VWD-3400RS UV-VIS detector (all from Dionex). Chromeleon® 6.80 SP5 Chromatography Management Software (Dionex) was used for instrument control and data collection and processing. Samples were prepared using an IKA® MS1 Minishaker (IKA Works, Guangzhou, China), Kudos® SK3200LH ultrasonic generator (Kudos Ultrasonic Instrumental Co., Shanghai, China), and Anke® TGL-16B centrifuge (Anting Scientific Instrumental Factory, Shanghai, China). Solid-phase extraction was performed using a strata™ -x-c SCX SPE column (Phenomenex, Torrance, CA), and sample drying was performed using an SE-506 nitrogen purge instrument (Shine Tech., Beijing, China).

Chromatographic conditions

Table 1 - Chromatographic conditions

Chromatographic conditions are given in Table 1.

Reagents and standards

Deionized water was produced using a Milli-Q® Gradient A 10 (Millipore, Bedford, MA). HPLC-grade methanol (CH3OH) and acetonitrile (CH3CN) were obtained from Fisher Scientific (Pittsburgh, PA). Analytical-grade ammonium acetate (NH4Ac), acetic acid (HAc), citric acid (C6H8O7 · H2O), and ammonia solution (25%-28%) were obtained from Shanghai Chemical Reagent Co. (Shanghai, China). Sodium 1-octane sulfate (98%) was a Baker Analyzed HPLC Reagent obtained from J.T. Baker (Phillipsburg, NJ). Nitrogen (N2, 99.999%) was from Lumin Gas Works (Shanghai, China). HPLC-grade melamine (99.0%), used as a standard, was acquired from Fluka (Milwaukee, WI).

Samples

Eight powdered milk samples and two liquid milk samples suspected of containing melamine were obtained from manufacturers.

Sample preparation: solid-phase extraction method

Milk powder samples #1-5 (approximately 2 g of each) were accurately weighed into 50-mL centrifuge tubes, and 15 mL of aqueous trichloroacetic acid (1%, v/v) and 5 mL acetonitrile were added to each. After 1 min of vortex shaking, samples were placed in an ultrasonic bath for 30 min, and then shaken again for 10 min. After 10 min of centrifugation (setting = rpm ≥ 10,000), the supernatants were passed through filter paper into 25-mL volumetric flasks. Samples were brought to volume with 1% aqueous trichloroacetic acid.

Prior to use, the SPE column was activated by passing 3 mL CH3OH and 5 mL H2O through in turn. Sample extracts were diluted with deionized (DI) water (5 mL of extract and 5 mL of water), and transferred onto the activated SCX SPE column. The SPE column was washed with 3 mL methanol and 3 mL water, respectively; then samples were eluted with 6 mL of aminated methanol solution (mixture of 5 mL ammonia solution and 95 mL methanol). The collected eluents were dried with N2 at 50 °C. The residues were dissolved in 1 mL of mobile phase and vortex mixed for 1 min. Prior to injection, solutions were filtered through a 0.2-µm Millex®-HV filter (Millipore).

Sample preparation: new method

Milk powder samples #6-8 (approximately 1 g each) were accurately weighed into 15-mL centrifuge tubes, and 10 mL water was added to each. After 1 min of vortex shaking, each was placed in an ultrasonic bath for 30 min. To each was added 1 mL dilute HAc (3%, v/v), and sample solutions were stored at 4 °C for at least 30 min. After 15 min of centrifugation (setting =  rpm ≥ 10,000), supernatants were transferred to 10-mL volumetric flasks and were brought to volume with DI water. Prior to injection, the solutions were filtered through a 0.2-µm Millex-HV filter.

Liquid milk samples #9-10 (approximately 10 mL) were accurately measured into 15-mL centrifuge tubes, diluted with 1 mL dilute HAc (3%, v/v), and stored at 4 °C for at least 30 min. The remainder of the procedure was the same as that for milk powder samples #6-8.

Spiked milk powder and liquid milk sample preparation

A 1000-µg/mL melamine stock standard was prepared by accurately weighing approximately 100 mg of melamine into a 100-mL volumetric flask and bringing to volume with aqueous methanol (50%, v/v). The stock standard was diluted appropriately to prepare working standards with concentrations of 0.05, 0.1, 0.2, 0.5, 1.0, 2.0, 5.0, 20, 25, 40, 50, and 100 µg/mL for calibration.

For milk powder samples #1-5, spiked samples were prepared by adding 20 µL of the stock standard solution of melamine to the 50-mL centrifuge tubes together with the weighed sample, and then following the first procedure above. For milk powder samples #6-8, spiked samples were prepared by adding 40 µL of the stock standard solution of melamine to the 15-mL centrifuge tubes together with the weighed sample, and then following the second procedure above. For liquid milk samples #9-10, spiked samples were prepared by adding 40 µL of the stock standard solution of melamine to the 15-mL centrifuge tubes together with the measured sample, and then following the second procedure above.

Results and discussion

Optimized chromatographic conditions

Melamine is a hydrophilic compound that is poorly retained on a typical RP column (e.g., C18 or C8 column). Most RP methods for melamine use an ion-pairing reagent, such as octane sulfate. With an ion-pairing reagent in the mobile phase, melamine is well retained. However, the ion-pairing reagent may coat the RP stationary phase, changing the retention property of the RP column, which may not be desirable if the column is used for other methods. The RP-PIC (paired ion chromatography) method is also not compatible with MS detection. Therefore, separation of the cationic melamine was attempted on the Acclaim Mixed-Mode WCX-1 column using an ammonium acetate buffer as the eluent to prevent column damage and make the method compatible with MS.

The Acclaim Mixed-Mode WCX-1 column features a mixed-mode silica-based packing material that incorporates both hydrophobic and weak cation-exchange properties. Mobile phase pH affects the charge and hydrophobicity of the stationary phase. At a pH below the pKa of the stationary phase carboxylate group, the hydrophobic interaction is the primary retention mechanism. At a pH above the pKa of the stationary phase carboxylate group, both cation-exchange and hydrophobic interactions contribute to retention, depending on the structures of analytes. Experiments revealed that melamine is poorly retained when the pH is lower than 3.5. At a pH higher than 5.0, melamine becomes deprotonated and is also poorly retained.