Nghiên cứu xây dựng quy trình xác định hàm lượng ammelide (amd) và dicyandiamide (DCD) trong thức ăn chăn nuôi bằng sắc ký lỏng hiệu năng cao đầu dò dad (HPLC-DAD)

Journal of Science Technology and Food 20 (3) (2020) 67-79 67 STUDYING TO BUILD THE DETERMINATION PROCESS OF AMMELIDE (AMD) AND DICYANDIAMIDE (DCD) IN ANIMAL FEED BY HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY (HPLC) WITH A DIODE-ARRAY DETECTOR (DAD) Dang Van Su1*, Phan Thi Thanh Dieu1, Bui Van Tam2 1Ho Chi Minh City University of Food Industry 2National Centre for Veterinary Drugs and Bio-Products Control No.2 *Email: sudv@hufi.edu.vn Received: 12 May 2020; Accepted: 24 July 2020 ABSTRACT The purpose of this study is to built a process for determining the content of fake protein enhancers such as Ammelide (AMD) and Dicyandiamide (DCD) in animal feed in accordance with the sample preparation procedure combined with a solid phase extraction (SPE) purification process and a high-performance liquid chromatography (HPLC) with a diode- array detector (DAD). The content of AMD and DCD in animal feed samples were determined with quantitative limits of 0.1-1.000 ppm, respectively for both substances, meeting AOAC (Association of Official Analytical Chemists) requirements for method validation and requirements for sensitivity, repeatable, linear intervals to be practically applicable. The procedure of determination has been effectively applied at National Centre for Veterinary Drugs and Bio-Products Control No. 2 to control the content of AMD and DCD in the basis of actual animal feed samples. Keywords: Fake protein enhancers, AMD, DCD, animal feed, SPE, HPLC-DAD. 1. INTRODUCTION Researching on determining the content of fake protein enhancers such as Ammelide (AMD) and Dicyandiamide (DCD), has recently become an issue of concern in the food industry. They are characterized by a high content of nitrogen in the molecular formula, so mixed into milk, animal feed to artificially increase the protein content to cope with the product quality controls. Various methods for determining the content of these substances in milk have been developed. Chen et al., proposed the process of determining DCD in milk samples by LC-MS/MS using d-SPE and LLE techniques to clean the sample combined with the internal standard to quantify [1]. MacMahona et al. published the procedure for determining DCD, AMD and some melamine derivatives in infant food samples by LC-MS/MS method with LOQ from 18-162 ppb (depending on substance) [2]. In addition, some authors published the procedures for determining DCD and AMD by conventional methods such as UV [3], GC/MS [4], ion exchange chromatograph [5] or creating complexes and determining by UV-Vis [6]. In general, the above methods required the use of specific chemicals (using internal standards [1], derivatives [6]) and expensive sample preparation techniques and experienced staffs. However, there has not been any announcement of AMD and DCD analysis methods and regulations on their thresholds in animal feed. This is a very complicated matrix because it is a mixture of Dang Van Su, Phan Thi Thanh Dieu, Bui Van Tam 68 many different components such as proteins, fats, antibiotics, minerals and some other components (existing components are available in natural materials). If these components are not removed before quantifying AMD and DCD, they will cause errors in analysis results. The purpose of this study is to develop a process to identify AMD and DCD according to the sample preparation process combined with the cleaning process by solid phase extraction (SPE) and quantification by HPLC-DAD analysis method, which is popularized in many laboratories in Vietnam. 2. METERIALS AND METHODS 2.1. Materials 2.1.1. Chemicals Ammelide standard (99%, Dr. Ehrenstofer GmbH), cyanoguanidine (98%, Sigma); ammonia solution: 25%, Merck; acetonitrile (ACN): 99.9%, Fisher; methanol: 99.9%, Fisher; formic acid: 98-100%, Merck; ammonium acetate: 98%, Merck; super clean water: 18 MΩ-cm; trifloroacetic acid (TFA): 99%, acros organic. 2.1.2. Standard solutions, samples and mobile phases DCD 1,000 ppm stock standard solution: Accurately weighed about 10 ± 0,1 mg of DCD standard in a 10 mL volumetric flask, add 8 mL of H2O, ultrasonic in 30 minutes, make up to the mark with H2O. This solution is then stored in a refrigerator at 2-8 °C in a light-free condition. AMD 1,000 ppm stock standard solution: Accurately weighed about 10 ± 0,1 mg of AMD standard in a 10 mL volumetric flask, add 50 µL of 25% ammonia solution (because ammelide is sparingly soluble in water and easily soluble in mild alkaline solutions, add 8 mL of H2O, ultrasound in 5 minutes, then make up to the mark with H2O. This solution is then stored in a refrigerator at 2-8 oC in a light-free condition. The analysis standard solution is diluted from stock standard solution with a mixture of solvent ACN - ammonium acetate 10 mM (50:50, v:v). Ammonium acetate solution 10 mM: Accurately weighed about 393.3 mg of ammonium acetate (99%, Merck) into Becher 500 mL, dissolved with 500 mL H2O. Sample extract solution ACN:H2O (50:50, v:v): Dissolved 50 mL of ACN solution into 50 mL H2O, mixed thoroughly with the vortex. Ammonium acetate soluble solution 10 mM - ACN (50:50, v:v): Dissolved 50 mL of 10 mM ammonium acetate into 50 mL of ACN, mixed thoroughly with the vortex. TFA solution 0.1%: Dissolved 100 µL of 99% TFA solution in 100 mL H2O, mixed thoroughly with the vortex. Ammonia solution 5% in MeOH: Dissolved 20 mL of 25% ammonia solution into 80 mL MeOH, mixed thoroughly with the vortex. Mobile phase: Mobile phase A: ammonium acetate 10 mM, adjusted to pH 6.5 with 0.1% formic acid solution; Mobile phase B: ACN. 2.1.3. Blank sample and standard spiked sample Blank sample: The composition of blank sample was similar to the real sample but does not contain analyte or analyte less than the quantitative level of the method being applied. The used blank sample was the “Asian piglet concentrate-feed form” from the Asian company. Studying to build the determination process of ammelide (AMD) and dicyandiamide (DCD) 69 Standard spiked sample: Blank sample was added to a quantity of standard AMD, DCD solution with known concentrations, then mixed well and dried at a temperature of about 60 oC in 8 hours. 2.2. Methods 2.2.1. Analytical process The analytical procedure was based on analytical procedures (FDA LIB 4422, CLG Me.01,..) [7, 8] and studies by Shen et al. [9], Fu & Schreiber [10], Turowski Maciej [11], Krunve et al. [12, 13]: Static phase with Hilic separation column: Inertsil, 5 µm (pore size 100 Å), 4.6 × 250 mm; Flow rate: 0.3-0.5 mL/minute; pH 6-7; Mobile phase solvent (ACN): 50-70; Device used to survey results: HPLC-DAD: wavelength 210 nm; Mobile phase A: amonium acetate 10 mM; Mobile phase B: ACN. 2.2.2. Sample processing procedure Samples of animal feed were ground and homogenized by IKA homogenizer. Weighed 1-2 ± 0.5 g of homogenized sample into a 50 mL centrifuge tube, added exactly 25 mL of ACN:H2O extraction solution (50:50, v:v). Shaked well with Vortex (2500 rpm) in 30 minutes, then centrifuged the entire extract solution (6000 rpm, 4 °C) in 10 minutes. The resulting extract is filtered through a membrane filter (0.45 µm - 25 mm) [7, 8]. Took exactly 20 mL of solution after filtration to clean with solid phase extraction (SPE) with an extraction solvent of ACN:H2O (50:50, v:v), SPE SCX (cation extraction) extraction column 500 mg/3 mL [7, 14, 15]. The solution obtained after cleaning by SPE will be concentrated with nitrogen gas and redissolved with 2 mL of dissolved solution. Injected then into the HPLC system. 2.2.3. Optimization of the HPLC conditions 2.2.3.1. Selection of static phase Based on previous studies on AMD and DCD content determination methods [16-19], the water-based interaction chromatography technique (Hilic) was selected. 2.2.3.2. Investigation of the flow rate, composition of mobile phase and pH Standard solutions with concentrations of 50 ppb (AMD) and 50 ppb (DCD) were used to investigate the effects of flow rate, mobile phase composition and pH. Investigation of the optimal condition of the mobile phase component was conducted on HPLC-DAD: 210 nm wavelength with the following parameters: Flow rate: 0.3-0.5 (low - high); Isocratic running mode; Mobile phase A (ammonium acetate 10 mM): 20-60 (low - high); Mobile phase B (ACN): 80-40 (low - high) and pH: 6-7 (low - high); Based on the research of Srinubabu et al. [20], the experimental model of 2k was selected and arranged according to Table 1. Evaluating the influence of factors based on the result of comparing the ability of separation, retention time, peak area and analyte stability from the obtained results. 2.2.4. Investigation of SPE extraction 2.2.4.1. Selection of extraction solvent The extraction solvent, ACN:H2O (50:50, v:v), was selected according to the references TCVN 9048-2012 [14], FDA LIB 4422 [15], CLG - Melamine 1.0 [7]. Dang Van Su, Phan Thi Thanh Dieu, Bui Van Tam 70 Table 1. Experimental arrangement for investigating the optimal condition of mobile phase composition for HPLC No. % ACN Flow rate pH of mobile phase 1 50 0.3 6 2 70 0.3 6 3 50 0.5 6 4 70 0.5 6 5 50 0.3 7 6 70 0.3 7 7 50 0.5 7 8 70 0.5 7 9 60 0.4 6.5 10 60 0.4 6.5 11 60 0.4 6.5 2.2.4.2. Investigation of SPE extraction procedure The SPE extraction procedure was proposed according to the Phenomenex instructions, including the following steps: Activated the column: added 5 mL MeOH and 5 mL H2O, respectively; Added sample: took exactly 20 mL of sample into the column so that the flow rate of the sample through the column is 2-4 drops/10s; Dried the column: used a vacuum pump to dry the solution contained in the column; Washed impurities: added 5 mL H2O and 5 mL 0,1% TFA, respectively; Dried the column: used a vacuum pump to dry the solution contained in the column; Recovery of analyte: Added 8-10 mL of 5%/MeOH ammonium acetate eluent to the column, dripping speed of 2-4 drops/10 seconds. Because the amount of elution solvent will determine the recovery of elution solvent volumes at 3 levels of 6 mL, 8 mL and 10 mL, with the analyte (AMD and DCD) content at two levels of 5 ppb and 120 ppb will be surveyed; Dried the column: used a vacuum pump to dry the solution contained in the column; All eluents were evaporated in a boiling pot (45-55 °C) combined with blowing nitrogen, then redissolved with 1 mL 10 mM ammonium acetate - ACN (50-50) and then, injected into the HPLC system. The results of the survey were evaluated based on the comparison of the results of the analyte content obtained and the theoretical concentration. 2.2.5. Appraisal method 2.2.5.1. Specificity Following the guidelines of the European Analytical Society, the HPLC-DAD is acceptable to confirm a positive sample [21]. 2.2.5.2. Investigation of LOD, LOQ LOD detection limit (qualitative limit) is determined according to the method evaluation guidelines of the National Institute for Food Control [22], LOQ was determined by the following formula: LOQ = 3 - 10 × CLOD (1) Studying to build the determination process of ammelide (AMD) and dicyandiamide (DCD) 71 LOD and LOQ of the device were determined as follows: Standard solution with concentration of about 100 ppb or less, injected this solution into DAD; Diluted the concentration of the above solution until a signal of the peak that met the signal / noise requirements (S/N) ≥ 3 - 10 (for substances classified as toxin) and S/N ≥ 3 (for substances not classified as toxic), according to SANCO/825/00 rev.8.1 16/11/2010 [23]; Calculated to determine the LOD, LOQ of the device. The LOD and LOQ of the method were as follows: From the LOD of the device, the amount of standard solution added to the blank sample was calculated so that 1 g of the standard spiked sample contained was equal to the LOD of the device; Homogenized the standard spiked sample according to ISO Guide 35:2017 [24] and EC 657/2002 [25]; Processed sample and injected into the chromatographic system to determine S/N; Increased or decreased the amount of standard solution added to the blank sample according to the results of S/N until it complied with the requirements for the determination of LOD and LOQ according to EC 657/2002 [25]. 2.2.5.3. Investigating linear intervals Standard solution with concentration ranged from 1 - 1,000 ppb, then injected into HPLC system with injection procedure from low to high concentration solution. The standard solution was treated the same as the sample solution. The calibration curve was investigated on the DAD at 210 nm. 2.2.5.4. Investigating repeatability and recovery Based on the guidance of ISO Guide 35:2017 [24] and EC 657/2002 [25], standard spiked sample with known concentration of AMD and DCD standard solution created to investigate the repeatability and recovery. The concentration of standard solution added to animal feed samples is shown in Table 2 and Table 3. Table 2. Concentration of standard solution added to sample to investigate repeatability and recovery of HPLC - DAD Substances Concentration (ppm) Sample 1 Sample 2 Sample 3 AMD 10 50 100 DCD 10 500 100 Table 3. Concentration added to sample for SPE extraction volume survey Substances Concentration (ppb) Sample 4 (SPE) Sample 5 (SPE) AMD 5 120 DCD 5 120 The standard spiked sample was calibrated to determine repeatability at a concentration of 150 ppb of AMD and DCD (3 times). The recovery, the repeatability and the accuracy of the method were evaluated based on a comparison of the obtained results and the theoretical concentration of the analyte. The evaluation was based on the guidance of AOAC Appendix F [26, 27]. Dang Van Su, Phan Thi Thanh Dieu, Bui Van Tam 72 3. RESULTS AND DISCUSSION 3.1. Optimized the HPLC mobile phase conditions Based on previous studies [12, 13], three factors influenced the analysis process were mobile phase solvent (ACN) (50-70), flow rate (TDD) (0.3-0.5 mL/minute) and pH mobile phase (pH 6-7) selected. Table 4. Results of optimization of HPLC mobile phase conditions No. % ACN Flow rate (TDD) pH mobile phase Retention time Peak area AMD DCD AMD DCD 1 50 0.3 6 8.90 10.3 7,902.09 16,867.43 2 70 0.3 6 9.34 10.4 8,644.51 18,938.03 3 50 0.5 6 5.29 6.21 4,770.924 10,227.53 4 70 0.5 6 5.10 6.18 5,133.33 10,323.73 5 50 0.3 7 8.84 10.4 7,968.61 11,273.2 6 70 0.3 7 9.37 10.4 8,654.31 17,328.3 7 50 0.5 7 5.285 6.20 4,810.89 10,246.23 8 70 0.5 7 5.558 6.20 5,081.43 10,638.37 9 60 0.4 6.5 6.63 7.71 6,837.4 14,879.7 10 60 0.4 6.5 6.631 7.70 5,902.57 13,886.85 11 60 0.4 6.5 6.631 7.70 6,905.8 14,922.05 The results in Table 4 showed that the pH mobile phase varied from 6-7, the peak areas of AMD and DCD were not significantly changed. Increasing the amount of ACN from 50% to 70%, the substances were in the column longer. At the high flow rate, the substances output faster, the peak parameters were better than at the low flow rate. According to the survey, the results were stable at a flow rate of 0.4 mL/minute and 60% ACN. The data in Figure 1 and Table 5 showed that the flow rate of the mobile phase significantly affected to the AMD analysis by HPLC. Similarly, results of DCD were shown in Figure 2 and Table 5. Figure 1. Pareto frequency chart of AMD effects Studying to build the determination process of ammelide (AMD) and dicyandiamide (DCD) 73 Figure 2. Pareto frequency chart of DCD effects Table 5. Variant values of AMD and DCD Substances DCD AMD Source F-Ratio P-Value F-Ratio P-Value A: % ACN 27.00 0.0351 1.69 0.3232 B: flow rate 191.99 0.0052 71.15 0.0138 C: pH 17.18 0.0536 0.00 0.9713 AB 21.22 0.0440 0.25 0.6657 AC 6.67 0.1229 0.01 0.9339 BC 20.67 0.0451 0.00 0.9607 ABC 0.00 0.9843 Lack-of-fit 8.11 0.1098 0.04 0.8667 R-squared (%) 93.9871 97.2903 R-squared (adjusted for Degree of freedom) (%) 84.9678 90.9678 Standard Error of Est, 586.137 560.514 Mean absolute error 599.492 127.002 Factors with a P-Value < 0.05 significantly affected the analysis results (peak area). Accordingly, the variant values in Table 5 showed: For AMD: flow rate of the mobile phase (P-Value = 0.0138) significantly effected on the analysis results; For DCD: % ACN (P-Value = 0,0351) and the mobile phase flow rate (P-Value = 0.052) have a significant influence on the analysis results. Lack-of-fit parameters with P-Value = 0.1098 (> 0.05) and R-squared > 90% at α = 95% showed that 2k model was suitable for experimental design. For AMD, the prediction equation: Area = 8,600.55 + 66,8136*%ACN – 12,746.3*TDD + 116.262*pH – 42.2666*%ACN*TDD – 0.198763*%ACN*pH + 306.751*TDD*pH - 8.78914*%ACN*TDD*pH (2) For DCD, the prediction equation: Area = 97,598.0 – 206.027*%ACN – 93,914.8*TDD – 15,675.6*pH - 954.671*%ACN*TDD + 107.011*%ACN*pH + 18,843.3*TDD*pH (3) Where: TDD: Flow rate of mobile phase (mL/minute) pH: pH of mobile phase Dang Van Su, Phan Thi Thanh Dieu, Bui Van Tam 74 The results in Table 6 showed the optimization of parameters of HPLC technique for AMD and DCD analysis. Table 6. Results of optimization of HPLC specifications for AMD and DCD Substances DCD AMD Factors Low High Optimal values Optimal values ACN 50.0 70.0 60.7355 61.3798 Flow rate 0.3 0.5 0.387795 0.378455 pH 6.0 7.0 6.51702 6.51686 To simply the installation of device, the following parameters were proposed: %ACN: 60%; Flow rate: 0.4 mL/minute: pH: 6.5. Ammonium acetate 10 mM - ACN (50:50, v:v) was chosen as the solvent, pH 6.5. AMD standard solutions (50 ppm) and DCD (50 ppm) were used to verify optimum results of HPLC. The verification results were shown in Table 7. Table 7. Compared results after optimization Values Peak area of DCD Peak area of AMD The predicted optimal value 5,905.82 14,922.05 The actual optimal value 5,878.58 5,778.81 5,818.74 14,915.4 13,254.6 14,275.8 Tstat 2.775 Tcritical (α = 0.05) 4.302 The statistical results showed that there was almost no difference (tcritical > tstat; α = 0.05) between the actual value and the predicted optimal value. Therefore, the selected parameters can be applied in real sample. 3.2. Investigation of sample cleaning procedure by SPE Table 8 and Table 9 showed the survey results of AMD and DCD contents at different elution solvent volumes. Table 8. Survey results of AMD content at different elution solvent volumes AMD (5 ppb) AMD (120 ppb) The volume of eluting solvent 6 mL 8 mL 10 mL 6 mL 8 mL 10 mL AMD content detected (ppb) 2.0559 4.1396 3.8067 101.0632 113.3745 121.5610 1.6520 4.4079 4.3899 103.5953 112.0980 116.3981 2.0453 4.3627 4.3631 102.5272 118.2784 121.3872 Studying to build the determination process of ammelide (AMD) and dicyandiamide (DCD) 75 Table 9. Survey results of DCD content at different elution solvent volumes DCD 5 ppb DCD 120 ppb The volume of eluting solvent 6 mL 8 mL 10 mL 6 mL 8 mL 10 mL DCD content detected (ppb) 2.3425 4.5084 5.5057 82.2926 119.6097 129.3401 2.2589 4.7848 4.0333 107.1688 102.9985 107.8945 3.0764 4.8819 5.5259 71.6622 103.1706 108.7572 Table 10. Results of the recovery survey The volume of eluting solvent AMD (5 ppb) AMD (120 ppb) 6 mL 8 mL 10 mL 6 mL 8 mL 10 mL Recovery 38.35 % 87.31 % 78.66 % 85.3 % 95.5 % 78.7 % DCD (5 ppb) DCD (120 ppb) Recovery 50.95 % 95.2 % 92.97 % 72.53% 90.5 % 96.11 % The analytical results in Table 10 showed that the amount of elution solvent was 6 mL, the recovery efficiency varied between 38% and 85%. At the volumes of solvent respectively 8 mL and 10 mL, the results showed that there was not significant difference. Therefore, the recommended elution volume was 8 mL. 3.3. Validation of analytical methods 3.3.1. Specificity / selection The results of determination of AMD and DCD content of animal feed samples "Concentrated feed for pigs from training - finishing" from Asian company were determined that AMD and DCD content were negative. 3.3.2. The limit of detection (LOD) and the limit of quantitation (LOQ) The results of the limit of detection (LOD) and the limit of quantitation (LOQ) for AMD and DCD were presented in Table 11. Table 11. LOD, LOQ survey results for LC-DAD Substances Concentration (ppm) Ratio S/N Number of injections AMD LOD 0.05 5.01 7 LOQ 0.1 14.18 DCD LOD 0.05 3.25 LOQ 0.1 10.16 LOD and LOQ of AMD and DCD of LC-DAD method were 0.05 and 0.1 ppm, respectively. These were similar to these of the previous published with another methods [2, 3]. Dang Van Su, Phan Thi Thanh Dieu, Bui Van Tam 76 3.3.3. Results of investigating linear intervals Results of investigating linear intervals for HPLC-DAD in Table 12 showed that AMD and DCD have linear range of 0.1-50 ppm. Table 12. Investigating linear intervals AMD and DCD Substances Concentration of standard solution (ppm) Correlation coefficients DCD 0.1 0.9998 0.5 1 5.0 10 50 AMD 0.1 0.9999 0.5 1 5.0 10 50 3.3.4. Accuracy, repeatability and recovery Table 13 and Table 14 showed the results of the investigation of repeatability and recovery of standard spiked animal feed samples. According to the AOAC (app-f) documentation of the validity of the method, the results were on completely responsive. Recoveries ranged from 92.4% to 98.2% that were relatively higher than those in previous studies (84.6%-96.8% - DCD analysis with HPLC-UV) [3], (61.4%-117.2% - AMD analysis with GC-MS/MS) [4]. Table 13. Concentration added to the sample to investigate recovery, repeatability Substances Concentration (ppm) Sample 1 Sample 2 Sample 3 AMD 1 5 10 DCD 1 5 10 Studying to build the determination process of ammelide (AMD) and dicyandiamide (DCD) 77 Table 14. Accuracy, repeatability and recovery Substances Level spiked (ppm) Assay concentration (ppm) Results AOAC requirements Sample 1 Sample 2 Sample 3 Average Recovery % RSD (%) Recovery % RSD (%) DCD 1 0.92 0.96 0.95 0.94 94.33 5.43 60 - 115 < 21 5 4.87 4.93 4.94 4.91 98.2 0.77 10 9.21 9.29 9.23 9.24 92.4 0.45 80 – 110 < 15 AMD 1 0.93 0.89 0.97 0.93 93.0 4.30 60 - 115 < 21 5 4.83 4.81 4.84 4.83 96.5 0.23 10 9.11 9.35 9.26 92.4 92.4 1.31 80 – 110 < 15 4. CONCLUSION The study has established a process to identify AMD and DCD according to the sample preparation process combined with SPE technique and quantification by HPLC-DAD. This proposed analytical procedure included sample pretreatment using SPE SCX (cation extraction) extraction column 500 mg/3 mL with an extraction solvent of ACN:H2O (50:50, v:v) at elution volume of 8 mL, followed by pre-concentration using nitrogen gas and analysis with HPLC-DAD. Static phase with Hilic separation column: Inertsil, 5 µm (pore size 100 Å), 4.6 × 250 mm; Flow rate: 0.4 mL/minute; pH 6.5; Mobile phase solvent (ACN): 60%; Wavelength 210 nm. The method is capable of quantifying AMD and DCD content in real animal feed samples with quantitative limits of 0.1-1.000 ppm, respectively, for both substances, meeting the requirements of AOAC for method validation. The method has been effectively applied at the National Centre for Veterinary Drugs and Bio-Products Control No.2 to control AMD and DCD content in animal feed sample background. The method satisfies the requirements of sensitivity, repeatability, linear intervals to be practical in application. In the future, this method may be proposed to further research and develop into a Vietnamese Standard. However, more research is needed to assess the factors that influence the analysis process from more feed sample bases to expand the applicability of the method to serve the needs of analysis as well as contribute to the analysis, support the quality management of animal feed, contribute to ensuring food safety and hygiene. REFERENCES 1. Chen X.H., Zhou L.X., Zhao Y.G., Pan S.D., Jin M.C. - Application of nanoring amino-functionalized magnetic polymer dispersive micro-solid-phase extraction and ultra fast liquid chromatography–tandem mass spectrometry in dicyandiamide residue analysis of powdered milk, Talanta 119 (2014) 187-192. 2. MacMahon S., Begley T.H., Diachenko G.W., Stromgren S.A. - A liquid chromatography-tandem mass spectrometry method for the detection of economically motivated adulteration in protein-containing foods, Journal of Chromatography A 1220 (2012) 101-107. 3. Qiu H., Sun D., Gunatilake S.R., She J., Mlsna T.E. - Analysis of trace dicyandiamide in stream water using solid phase extraction and liquid chromatography UV spectrometry, Journal of Environmental Sciences 35 (2015) 38-42. Dang Van Su, Phan Thi Thanh Dieu, Bui Van Tam 78 4. Miao H., Fan S., Wu Y.N., Zhang L., Zhou P.P., Li J.G., Chen H.J., Zhao Y.F. - Simultaneous determination of melamine, ammelide, ammeline, and cyanuric acid in milk and milk products by gas chromatography-tandem mass spectrometry, Biomedical and Environmental Sciences 22 (2) (2009) 87-94. 5. Ono S., Funato T., Inoue Y., Munechika T., Yoshimura T., Morita H., Rengakuji S., Shimasaki C. - Determination of melamine derivatives, melame, meleme, ammeline and ammelide by high-performance cation-exchange chromatography, Journal of Chromatography A 815 (2) (1998) 197-204. 6. Yu H., Tang Y., Zhang G., Wang Z., Gao R. - A new chemiluminescence method for determination of dicyandiamide based on the N-bromosuccinimide-merbromin- cetyltrimethylammonium bromide system, Journal of Luminescence 157 (2015) 327-332. 7. United States Department of Agriculture - CLG-MEL1.01: Determination and Confirmation of Melamine by LC/MS/MS, Effective: 12/02/2008. 8. Benvenuti M., O’Connor A. - Melamine, ammeline and cyanuric acid analysis by UPLC/MS/MS and UPLC/PDA, Waters Corporation, USA (2007). 9. Shen Y., Han C., Zhou X., Chen X., Huang F., Zhu Z. - Microwave-assisted extraction and determination of dicyandiamide residue in infant formula samples by liquid chromatography-tandem mass spectrometry, Journal of Dairy Science 96 (11) (2003) 6877-6882. 10. Fu F., Schreiber A. - LC-MS/MS analysis of emerging food contaminants. Quantitation and identification of dicyandiamide in milk and other protein-rich food (2013). 11. Turowski M., Deshmukh B. - Direct chromatographic method for determination of hydrogen cyanamide and dicyandiamide in aqueous solutions, Analytical Letters 37 (9) (2003) 1981-1989. 12. Kruve A., Rebane R., Kipper K., Oldekop M-L., Evard H., Herodes K., Ravio P., Leito I. - Tutorial review on validation of liquid chromatography–mass spectrometry methods: part I, Analytica Chimica Acta 870 (2015) 29-44. 13. Kruve A., Rebane R., Kipper K., Oldekop M-L., Evard H., Herodes K., Ravio P., Leito I. - Tutorial review on validation of liquid chromatography–mass spectrometry methods: part II, Analytica Chimica Acta 870 (2015) 8-28. 14. TCVN 9048:2012 (ISO/TS 15495:2010) - Milk, milk products and infant formulae - Guidelines for the quantitative determination of melamine and cyanuric acid by liquid chromatography - tandem mass spectrometry (LC-MS/MS). 15. Smoker M., Krynitsky A.J. - Interim method for determination of melamine and cyanuric acid residues in foods using LC-MS/MS: Version 1.0, Laboratory Information Bulletin LIB No.4422, U.S Food and Drug Administration, 2008. 16. Ibáñez M., Sancho J.V., Hernández F. - Determination of melamine in milk-based products and other food and beverage products by ion-pair liquid chromatography– tandem mass spectrometry, Analytica Chimica Acta 649 (1) (2009) 91-97. 17. Domenico Montesano, Oriella Gennari, Carmen Festa, Franco Zollo, Serenella Seccia, Stefania Albrizio - A simple HPLC-DAD Method for the analysis of melamine in protein supplements: validation using the accuracy profiles, Journal of Chemistry 2013, ID 239342, 7 pages. 18. Hua Wu, Rong An, Yao Xiao, Zhixu Zhang, Ping Li - Quantitative Liquid ChromatographyAnalysis of Melamine in Dairy Products Using Agilent's 1120 Compact LC and 1200 Rapid Resolution LC, Agilent Technologies Co., Ltd., Beijing, China, 2010. Studying to build the determination process of ammelide (AMD) and dicyandiamide (DCD) 79 19. Joachim Weiss - Handbook of ion chromatography, Wiley-VCH Verlag GmbH & Co. KgaA (2016). 20. Srinubabu G., Raju ChA., Sarath N., Kumar P.K., Rao J.V. - Development and validation of a HPLC method for the determination of voriconazole in pharmaceutical formulation using an experimental design, Talanta 71 (3) (2007) 1424-1429. 21. Gowik P. - Criteria and requirements of commission decision 2002/657/EC, Bulletin - International Dairy Federation 383 (2003) 52-56. 22. National Institute for Food Control - Evaluation of methods in chemical analysis and microbiology analysis, Publishing House for Science & Technology, Ha Noi, 2010 (in Vietnamese). 23. European Commission, Directorate General Health and Consumer Protection - Guidance document on pesticide residue analytical methods, SANCO/825/00 rev.8.1 (2010). 24. ISO Guide 35:2017 - Reference materials - General and statistical principles for certification (2017). 25. Gowik P. - Criteria and requirements of commission decision 2002/657/EC, Bulletin- International Dairy Federation 383 (2003) 52-56. 26. AOAC International - Guidelines for standard method performance requirements AOAC Official Methods of Analysis, Appendix F (2016) 1-18. 27. AOAC SMPR 2015.008 - Standard method performance requirements for alkaloids of Mitragyna speciosa, OAC International, 2015. https://members.aoac.org/AOAC_Docs/SPDS/SMPR2015_008.pdf. TÓM TẮT NGHIÊN CỨU XÂY DỰNG QUY TRÌNH XÁC ĐỊNH HÀM LƯỢNG AMMELIDE (AMD) VÀ DICYANDIAMIDE (DCD) TRONG THỨC ĂN CHĂN NUÔI BẰNG SẮC KÝ LỎNG HIỆU NĂNG CAO ĐẦU DÒ DAD (HPLC-DAD) Đặng Văn Sử1*, Phan Thị Thanh Diệu1, Bùi Văn Tâm2 1Trường Đại học Công nghiệp Thực phẩm TP.HCM 2Trung tâm Kiểm nghiệm Thuốc Thú y Trung ương II *Email: dvsu@hufi.edu.vn Mục đích của nghiên cứu này là xây dựng quy trình xác định hàm lượng các các chất tăng đạm giả Ammelide (AMD) và Dicyandiamide (DCD) trong thức ăn chăn nuôi (TACN), theo quy trình chuẩn bị mẫu kết hợp với quá trình làm sạch bằng kỹ thuật chiết pha rắn (SPE) và định lượng bằng phương pháp phân tích sắc ký lỏng hiệu năng cao đầu dò DAD (HPLC-DAD). Hàm lượng AMD và DCD trong mẫu TACN được xác định với giới hạn định lượng LOD và LOQ của AMD, DCD có nồng độ là 0.05 và 0.1 ppm, đáp ứng được yêu cầu của AOAC (Association of Official Analytical Chemists) về thẩm định phương pháp và về độ nhạy, độ lặp lại, khoảng tuyến tính để có thể ứng dụng vào thực tế. Quy trình xác định đã được áp dụng hiệu quả tại Trung tâm Kiểm nghiệm Thuốc Thú y TW II để kiểm soát hàm lượng AMD và DCD trong nền mẫu TACN thực tế. Từ khóa: Chất tăng đạm giả, AMD, DCD, thức ăn chăn nuôi, SPE, HPLC-DAD.