1362 Markus Himmelsbach1 Thuy Diep Thanh Vo2 1 Institute
of Analytical Chemistry, Johannes Kepler University Linz, Linz, Austria 2 Borealis Polyolefine GmbH, Linz, Austria
Received October 9, 2013 Revised December 3, 2013 Accepted December 4, 2013
Electrophoresis 2014, 35, 1362–1367
Research Article
Determination of melamine impurities by capillary zone electrophoresis with UVand quadrupole time-of-flight mass spectrometric detection A method based on CZE coupled to quadrupole TOF MS (Q-TOF MS), which is suitable to analyze seven common melamine impurities is presented. MS compatible formic acid based BGEs with various concentrations of organic modifiers were investigated using UV detection to determine the best separation system regarding resolution and speed. The optimized BGE consisted of 350 mM formic acid with 10% ACN and migration times were ⬍6 min. The detection limits were considerably improved when ESI MS was applied instead of UV detection. In addition the instrument allows quantification with good linearity and repeatability in the relevant concentration range. The analytical characteristics of the method were evaluated for melamine, guanylurea, diaminotriazine, ureidomelamine, guanlymelamine, melam, methylmelamine, and dimethylmelamine both with UV and MS detection. Several different melamine samples were analyzed und the determined impurity levels were in the range from 0.02 to 2.3%. Keywords: CZE / Melamine / Melamine impurities / quadrupole TOF MS DOI 10.1002/elps.201300493
1 Introduction Melamine (1,3,5-triazine-2,4,6-triamine) is a widely used chemical commonly employed as a monomer in the production of plastic resins, in the manufacturing of plywood, laminates, as composites for molding materials and for casting in electric installation and the fabrication of camping tableware [1]. In recent years this molecule has gained notoriety as it was used for adulteration of milk powder [2] and pet food [3]. For this reason the majority of papers describing the detection and quantitation of melamine and related sub-
Correspondence: Dr. Markus Himmelsbach, Institute of Analytical Chemistry, Johannes Kepler University Linz, Altenberger Straße 69, A-4040 Linz, Austria E-mail: [email protected] Fax: +43-732-2468-8679
Abbreviations: Diaminotriazine, 1,3,5-triazine-2,4-diamine; Dimethylmelamine, N2 ,N2 -dimethyl-1,3,5-triazine-2,4,6-triamine; Guanylmelamine, N-(4-amino-3,6-dihydro-6-imino-1, 3,5-triazin-2-yl)-guanidine; Guanylurea, N-(aminoiminomethyl)-urea; Melam, N2 -(4,6-diamino-1,3,5-triazin-2-yl)-1,3,5triazine-2,4,6-triamine; Melamine, 1,3,5-triazine-2,4,6-triamine; Methylmelamine, N2 -methyl-1,3,5-triazine-2,4, 6-triamine; Ureidomelamine, N-(4,6-diamino-1,3,5-triazin-2yl)-urea
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
stances come from the fields of food and feed analysis [4–7]. Nevertheless, it should be taken into account that melamine still is an important industrial product; so analytical methods for the characterization of this molecule as well as its impurities are needed. Despite this fact only a few reports dealing with melamine analysis in an industrial context can be found in the literature [8–14]. Among them there are also papers describing the use of CZE particularly in combination with mass spectrometric (MS) detection [8,10,14]. These describe investigations on the analysis of methylated melamines in reaction mixtures [10], and melamine formaldehyde resins [8,14]. Thereby CZE succeeded in establishing itself as a viable supplement or even replacement to HPLC methods particularly due to its unique separation mechanism providing substantially different retention/migration patterns [8, 13]. Modern melamine production processes provide high yields of melamine [15] and there are patented industrial purification methods that produce melamine with a purity of at least 99% [16]. Therefore a method suitable for quality control must be able to accurately determine very small amounts of impurities in presence of a huge melamine excess. CZE was investigated regarding its ability to fulfill all the desired requirements. Besides the two most important impurities ureidomelamine (N-(4,6-diamino-1,3,5-triazin-2-yl)-urea) and melam (N2 -(4,6-diamino-1,3,5-triazin-2-yl)-1,3,5-triazine2,4,6-triamine) [16] five other possible impurities were included in this study.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1362–1367
2 Materials and methods 2.1 Samples and reagents Analysis grade 2-propanol and HPLC gradient grade ACN were obtained from BDH Prolabo (Leuven, Belgium). HPLC gradient grade methanol was purchased from JT Baker (Deventer, The Netherlands) and eluent additive for LC-MS grade formic acid was purchased from Sigma-Aldrich (Steinheim, Germany). Analytical standards for quantification of melamine and its impurities melam, N-(4-amino-3,6-dihydro6-imino-1,3,5-triazin-2-yl)-guanidine (guanylmelamine), ureidomelamine, N2 -methyl-1,3,5-triazine-2,4,6-triamine (methylmelamine), N2 ,N2 -dimethyl-1,3,5-triazine-2,4,6triamine (dimethylmelamine), 1,3,5-triazine-2,4-diamine (diaminotriazine) and N-(aminoiminomethyl)-urea (guanylurea) were provided by Borealis Polyolefine (Linz, Austria). 18 M⍀ purified water was obtained from a Milli-Q Reference A+ purification system from Merck Millipore (Vienna, Austria). The structures of the analytes are shown in Fig. 1. Melamine, guanylmelamine, guanylurea, diaminotriazine, methylmelamine, and dimethylmelamine were dissolved in ACN/water/formic acid (50:50:1 v/v/v) giving 1000 mg/L stock solutions. Melam and ureidomelamine were dissolved in ACN/water/formic acid (45:45:10 v/v/v) giving 200 mg/L stock solutions. Melamine samples for quantification were dissolved in ACN/water/formic acid (50:50:1 v/v/v) giving 1000 mg/L solutions. Working standards and injection solutions of the samples were prepared daily prior to use by dilution with ACN/water (1:1 v/v).
2.2 CZE instrumentation and procedures CZE separations were carried out on an Agilent 3D CE system (Agilent, Waldbronn, Germany) with a DAD. Fused silica capillaries (50 m id × 375 m od) were obtained from Polymicro Technologies (Phoenix, AZ, USA). New capillaries were cut to a total length of 75 cm (UV detection) or 60 cm (MS detection), respectively, and conditioned by flush-
1363
ing with 1 M NaOH for 15 minutes, water for 5 min and BGE for 15 min. Before each run the capillary was flushed with BGE for 5 min. The separation voltage was +30 kV and the capillary temperature was 25°C. Sample injection was carried out by hydrodynamic injection (50 mbar for 10 s) and UV detection was performed at 195 nm. Formic acid at various concentrations in water/ACN and water/methanol mixtures was investigated as BGE and the optimized BGE consisted of 350 mM formic acid in water/ACN (90:10 v/v). 2.3 MS detection MS detection was performed on an Agilent 6520 Q-TOF mass spectrometer with an ESI source and an Agilent G1607A coaxial sprayer. All analyses were made in the positive ionization mode. Nitrogen was used as drying gas at a temperature of 250°C and a flow-rate of 4 L/min. Nitrogen nebulizing gas for electrospray was supplied at 3 psi. The sheath liquid consisting of water/iso-propanol (20:80, v/v) with 0.1% formic acid was delivered via an HPLC pump (Agilent) and a 1:100 splitter at a rate of 3 L/min. The voltage set for the MS capillary was 4 kV and the fragmentor was set to 200 V. Scanning mass range was from m/z 50–500 with an acquisition rate of 2 spectra s−1 in the MS mode.
3 Results and discussion 3.1 Selection of separation conditions In order to separate the melamine related compounds under investigation acidic buffers may be used to protonate the amino groups and facilitate their separation as cations. Organic solvents in the BGE are needed to ensure a sufficient solubility of the investigated analytes in the electrolyte. Investigations regarding separation conditions were done with UV detection, but MS is also considered as detector, so the choice of BGE is limited to volatile electrolytes in order to obtain a good detector response in CE-MS. In previous studies BGEs using ACN as organic modifier and containing
Figure 1. Chemical structures of melamine and melamine impurities.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
1364
M. Himmelsbach and T. D. T. Vo
Electrophoresis 2014, 35, 1362–1367
formic acid were found to be well suited for the separation of melamine-formaldehyde resins [8, 14] and methylmelamines [10]. Based on these results BGEs with formic acid concentrations from 100 to 1000 mM were investigated with a given solvent composition of water/ACN (50:50 v/v) in order to achieve separation of melamine and its impurities, but none of the systems was able to yield a satisfying separation. Even at the highest formic acid concentration of 1000 mM three compounds were comigrating. Varying the amount of acid proved to be important to separate structurally closely related melamine-derivatives such as methylmelamines [10], but did not give the selectivity needed in the case of melamine impurities. To enhance separation the amount of organic modifier was varied in the range from 10 to 50% ACN. Changing organic solvent concentrations of the BGE system affected selectivity clearly more than the formic acid concentration and resolution of peaks improved at lower ACN levels. The most critical peak pairs guanylurea/guanylmelamine and melam/methylmelamine were efficiently separated using 10% ACN. Guanylmelamine experienced the most significant mobility change compared to the other compounds when the organic modifier content was varied. The peak order changed as its migration time decreased from 7.8 min
(50% ACN) to 6.4 min (10% ACN). The influence of ACN percentage is shown in Fig. 2. Due to the low solubility, especially of melam und ureidomelamine in water, no BGEs with less than 10% ACN were investigated. With the given BGE composition of water/ACN (90:10 v/v) the formic acid concentration was reevaluated to determine the lowest acid concentration needed for baseline separation. Increasing acid concentration in the range of 100 to 500 mM formic acid improved resolution and at 350 mM the most critical peak pairs guanylurea/guanylmelamine and melam/methylmelamine were efficiently separated. A further increase did not yield better results as can be seen in Fig. 3. Since the organic modifier plays a key role methanol was investigated in addition to ACN. Selectivity was comparable to ACN but the runs were slower and melam/methylmelamine not that well separated as can be seen from Fig. 3D. Therefore, ACN was used as organic modifier in all following experiments and the final optimized BGE consisted of 350 mM formic acid in water/ACN (90:10 v/v).
Figure 2. Electropherograms obtained for a 10 mg/L standard with UV-detection at 195 nm. (A) BGE: 350 mM formic acid, 50% ACN; (B) BGE: 350 mM formic acid, 30% ACN; (C) BGE: 350 mM formic acid, 20% ACN; (D) BGE: 350 mM formic acid, 10% ACN. Peak numbering: (1) guanylmelamine; (2) guanylurea; (3) diaminotriazine; (4) melamine; (5) melam; (6) methylmelamine; (7) dimethylmelamine; (8) ureidomelamine
Figure 3. Electropherograms obtained for a 10 mg/L standard with UV-detection at 195 nm. (A) BGE: 200 mM formic acid, 10% ACN; (B) BGE: 350 mM formic acid, 10% ACN; (C) BGE: 500 mM formic acid, 10% ACN; D) BGE: 350 mM formic acid, 10% methanol. Peak numbering: (1) guanylmelamine; (2) guanylurea; (3) diaminotriazine; (4) melamine; (5) melam; (6) methylmelamine; (7) dimethylmelamine; (8) ureidomelamine
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.2 CZE using MS detection One major disadvantage of the above-described CZE-UV approach is its poor LOQ, so hyphenation with MS was
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1362–1367
1365
Table 1. Method characteristics calculated for UV and MS detection
Compound
LOQa) UV (w%)
RSDb) UV (%)
LOQa) MS (w%)
RSDb) MS (%)
R2c)
Guanylurea Guanylmelamine Diaminotriazine Melamine Methylmelamine Melam Dimethylmelamine Ureidomelamine
0.40 0.68 0.44 0.32 0.39 0.42 0.44 0.39
2.9 1.3 0.9 3.8 3.4 1.2 2.9 3.5
0.05 0.02 0.02 0.01 0.02 0.01 0.01 0.03
9.0 8.5 9.9 4.5 9.1 8.7 6.4 7.4
0.9986 0.9973 0.9960 0.9961 0.9978 0.9977 0.9987 0.9982
a) LOQ calculated as ten times baseline noise given as weight percent assuming that 100 mg of melamine sample is dissolved in 100 mL solvent. b) relative standard deviations of peak areas for five injections of a 10 mg/L standard (UV) and a 1 mg/L standard (MS). c) Correlation coefficient of the calibration curve in the range from the LOQ to 10 mg/L with MS detection.
Figure 4. Extracted ion electropherograms of a 1 mg/L standard solution using optimized ESI-MS conditions.
investigated. All ESI MS-related parameters were carefully optimized to establish a setup that allows the highly sensitive detection of the analytes and at the same time provides stable ESI conditions to ensure reliable results when quantification is needed. A sheath liquid consisting of water/ 2-propanol (20:80 v/v) with 0.1% formic acid was used, which had turned out to be an optimum composition for melamineformaldehyde resins and methylated melamines in previous work [10, 14] and only the most critical parameters like acid concentration and flow rate were investigated. The flow rate investigated in a range from 1 to 6 L/min. At a flow rate of 3 L/min good signal intensities were obtained together with a stable spray and therefore best repeatability. When flow rates higher than 3 L/min were applied signal intensities started to decrease. Furthermore, sheath liquids containing 1% formic acid, 5 mM ammonium formate, and combinations of formic acid and ammonium formate were tested but signal intensities were always lower than with 0.1% formic acid. Nebulizer pressure and drying gas flow rate were varied from 1 to 7 psi and 2 to 10 L/min respectively, whereby the higher nebulizer pressures yielded the highest signals but a negative effect on the separation was observed since the nebulizer created a suction which induced a laminar flow. At a nebulizer pressure of 3 psi the signal intensities were acceptable and separation was maintained. The variation of drying gas flow rates showed only minor changes in signal intensities and a value of 4 L/min was selected. The drying gas temperature was varied from 175 to 350°C also indicating C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
that this parameter does not affect signal intensities significantly; so 250°C was selected. When the MS capillary voltage was raised over a range from 3000 to 4000 V, the signal intensities increased notably. Often a less stable signal is observed at the highest adjustment but in case of melamine impurities this was not the case, therefore the capillary voltage was set to 4000 V. The fragmentor voltage was optimized in a range from 150 to 250 V. Signal intensities remained almost constant from 170 to 225 V followed by a slight decrease when the voltage was increased further; hence 200 V was chosen as the final setting. Figure 4 shows extracted ion electropherograms of a 1 mg/L standard solution using optimized ESI-MS conditions. Employing the optimized ESI MS parameters, calibration was performed with standard mixtures and very good linearity was observed from the LOQ up to 10 mg/L with correlation coefficients in the range from 0.9960 to 0.9987. Repeatability was determined by injecting six standard mixtures in the concentration range from 0.4 mg/L to 10 mg/L five times each. RSDs of the peak areas, calculated from all five injections of the 1 mg/L standard solution, were ⬍10% which is satisfactory for a CZE-MS method without correction by internal standard. Quantification limits were calculated as 10 times the baseline noise giving values between 0.1 mg/L (dimethylmelamine) and 0.5 mg/L (guanylurea). When impurities in melamine samples are analyzed the LOQ should be expressed as weight percent. Assuming that 100 mg of sample is dissolved in 100 mL solvent the resulting LOQs are between 0.01% w/w for dimethylmelamine and 0.05% w/w for guanylurea. The most important method characteristics for MS and UV detection are summarized in Table 1. 3.3 CZE-MS of melamine samples Several melamine samples were analyzed by CZE-ESI-MS and the summarized results are shown in Table 2. Melam www.electrophoresis-journal.com
1366
M. Himmelsbach and T. D. T. Vo
Electrophoresis 2014, 35, 1362–1367
Table 2. Results for the analysis of five industrial melamine samples
Content (%) Compound
Sample 1 Sample 2 Sample 3 Sample 4 Sample 5
Guanylurea Guanylmelamine Diaminotriazine Methylmelamine Melam Dimethylmelamine Ureidomelamine
of Analytical Chemistry, Johannes Kepler University Linz, Linz, Austria 2 Borealis Polyolefine GmbH, Linz, Austria
Received October 9, 2013 Revised December 3, 2013 Accepted December 4, 2013
Electrophoresis 2014, 35, 1362–1367
Research Article
Determination of melamine impurities by capillary zone electrophoresis with UVand quadrupole time-of-flight mass spectrometric detection A method based on CZE coupled to quadrupole TOF MS (Q-TOF MS), which is suitable to analyze seven common melamine impurities is presented. MS compatible formic acid based BGEs with various concentrations of organic modifiers were investigated using UV detection to determine the best separation system regarding resolution and speed. The optimized BGE consisted of 350 mM formic acid with 10% ACN and migration times were ⬍6 min. The detection limits were considerably improved when ESI MS was applied instead of UV detection. In addition the instrument allows quantification with good linearity and repeatability in the relevant concentration range. The analytical characteristics of the method were evaluated for melamine, guanylurea, diaminotriazine, ureidomelamine, guanlymelamine, melam, methylmelamine, and dimethylmelamine both with UV and MS detection. Several different melamine samples were analyzed und the determined impurity levels were in the range from 0.02 to 2.3%. Keywords: CZE / Melamine / Melamine impurities / quadrupole TOF MS DOI 10.1002/elps.201300493
1 Introduction Melamine (1,3,5-triazine-2,4,6-triamine) is a widely used chemical commonly employed as a monomer in the production of plastic resins, in the manufacturing of plywood, laminates, as composites for molding materials and for casting in electric installation and the fabrication of camping tableware [1]. In recent years this molecule has gained notoriety as it was used for adulteration of milk powder [2] and pet food [3]. For this reason the majority of papers describing the detection and quantitation of melamine and related sub-
Correspondence: Dr. Markus Himmelsbach, Institute of Analytical Chemistry, Johannes Kepler University Linz, Altenberger Straße 69, A-4040 Linz, Austria E-mail: [email protected] Fax: +43-732-2468-8679
Abbreviations: Diaminotriazine, 1,3,5-triazine-2,4-diamine; Dimethylmelamine, N2 ,N2 -dimethyl-1,3,5-triazine-2,4,6-triamine; Guanylmelamine, N-(4-amino-3,6-dihydro-6-imino-1, 3,5-triazin-2-yl)-guanidine; Guanylurea, N-(aminoiminomethyl)-urea; Melam, N2 -(4,6-diamino-1,3,5-triazin-2-yl)-1,3,5triazine-2,4,6-triamine; Melamine, 1,3,5-triazine-2,4,6-triamine; Methylmelamine, N2 -methyl-1,3,5-triazine-2,4, 6-triamine; Ureidomelamine, N-(4,6-diamino-1,3,5-triazin-2yl)-urea
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
stances come from the fields of food and feed analysis [4–7]. Nevertheless, it should be taken into account that melamine still is an important industrial product; so analytical methods for the characterization of this molecule as well as its impurities are needed. Despite this fact only a few reports dealing with melamine analysis in an industrial context can be found in the literature [8–14]. Among them there are also papers describing the use of CZE particularly in combination with mass spectrometric (MS) detection [8,10,14]. These describe investigations on the analysis of methylated melamines in reaction mixtures [10], and melamine formaldehyde resins [8,14]. Thereby CZE succeeded in establishing itself as a viable supplement or even replacement to HPLC methods particularly due to its unique separation mechanism providing substantially different retention/migration patterns [8, 13]. Modern melamine production processes provide high yields of melamine [15] and there are patented industrial purification methods that produce melamine with a purity of at least 99% [16]. Therefore a method suitable for quality control must be able to accurately determine very small amounts of impurities in presence of a huge melamine excess. CZE was investigated regarding its ability to fulfill all the desired requirements. Besides the two most important impurities ureidomelamine (N-(4,6-diamino-1,3,5-triazin-2-yl)-urea) and melam (N2 -(4,6-diamino-1,3,5-triazin-2-yl)-1,3,5-triazine2,4,6-triamine) [16] five other possible impurities were included in this study.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1362–1367
2 Materials and methods 2.1 Samples and reagents Analysis grade 2-propanol and HPLC gradient grade ACN were obtained from BDH Prolabo (Leuven, Belgium). HPLC gradient grade methanol was purchased from JT Baker (Deventer, The Netherlands) and eluent additive for LC-MS grade formic acid was purchased from Sigma-Aldrich (Steinheim, Germany). Analytical standards for quantification of melamine and its impurities melam, N-(4-amino-3,6-dihydro6-imino-1,3,5-triazin-2-yl)-guanidine (guanylmelamine), ureidomelamine, N2 -methyl-1,3,5-triazine-2,4,6-triamine (methylmelamine), N2 ,N2 -dimethyl-1,3,5-triazine-2,4,6triamine (dimethylmelamine), 1,3,5-triazine-2,4-diamine (diaminotriazine) and N-(aminoiminomethyl)-urea (guanylurea) were provided by Borealis Polyolefine (Linz, Austria). 18 M⍀ purified water was obtained from a Milli-Q Reference A+ purification system from Merck Millipore (Vienna, Austria). The structures of the analytes are shown in Fig. 1. Melamine, guanylmelamine, guanylurea, diaminotriazine, methylmelamine, and dimethylmelamine were dissolved in ACN/water/formic acid (50:50:1 v/v/v) giving 1000 mg/L stock solutions. Melam and ureidomelamine were dissolved in ACN/water/formic acid (45:45:10 v/v/v) giving 200 mg/L stock solutions. Melamine samples for quantification were dissolved in ACN/water/formic acid (50:50:1 v/v/v) giving 1000 mg/L solutions. Working standards and injection solutions of the samples were prepared daily prior to use by dilution with ACN/water (1:1 v/v).
2.2 CZE instrumentation and procedures CZE separations were carried out on an Agilent 3D CE system (Agilent, Waldbronn, Germany) with a DAD. Fused silica capillaries (50 m id × 375 m od) were obtained from Polymicro Technologies (Phoenix, AZ, USA). New capillaries were cut to a total length of 75 cm (UV detection) or 60 cm (MS detection), respectively, and conditioned by flush-
1363
ing with 1 M NaOH for 15 minutes, water for 5 min and BGE for 15 min. Before each run the capillary was flushed with BGE for 5 min. The separation voltage was +30 kV and the capillary temperature was 25°C. Sample injection was carried out by hydrodynamic injection (50 mbar for 10 s) and UV detection was performed at 195 nm. Formic acid at various concentrations in water/ACN and water/methanol mixtures was investigated as BGE and the optimized BGE consisted of 350 mM formic acid in water/ACN (90:10 v/v). 2.3 MS detection MS detection was performed on an Agilent 6520 Q-TOF mass spectrometer with an ESI source and an Agilent G1607A coaxial sprayer. All analyses were made in the positive ionization mode. Nitrogen was used as drying gas at a temperature of 250°C and a flow-rate of 4 L/min. Nitrogen nebulizing gas for electrospray was supplied at 3 psi. The sheath liquid consisting of water/iso-propanol (20:80, v/v) with 0.1% formic acid was delivered via an HPLC pump (Agilent) and a 1:100 splitter at a rate of 3 L/min. The voltage set for the MS capillary was 4 kV and the fragmentor was set to 200 V. Scanning mass range was from m/z 50–500 with an acquisition rate of 2 spectra s−1 in the MS mode.
3 Results and discussion 3.1 Selection of separation conditions In order to separate the melamine related compounds under investigation acidic buffers may be used to protonate the amino groups and facilitate their separation as cations. Organic solvents in the BGE are needed to ensure a sufficient solubility of the investigated analytes in the electrolyte. Investigations regarding separation conditions were done with UV detection, but MS is also considered as detector, so the choice of BGE is limited to volatile electrolytes in order to obtain a good detector response in CE-MS. In previous studies BGEs using ACN as organic modifier and containing
Figure 1. Chemical structures of melamine and melamine impurities.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
1364
M. Himmelsbach and T. D. T. Vo
Electrophoresis 2014, 35, 1362–1367
formic acid were found to be well suited for the separation of melamine-formaldehyde resins [8, 14] and methylmelamines [10]. Based on these results BGEs with formic acid concentrations from 100 to 1000 mM were investigated with a given solvent composition of water/ACN (50:50 v/v) in order to achieve separation of melamine and its impurities, but none of the systems was able to yield a satisfying separation. Even at the highest formic acid concentration of 1000 mM three compounds were comigrating. Varying the amount of acid proved to be important to separate structurally closely related melamine-derivatives such as methylmelamines [10], but did not give the selectivity needed in the case of melamine impurities. To enhance separation the amount of organic modifier was varied in the range from 10 to 50% ACN. Changing organic solvent concentrations of the BGE system affected selectivity clearly more than the formic acid concentration and resolution of peaks improved at lower ACN levels. The most critical peak pairs guanylurea/guanylmelamine and melam/methylmelamine were efficiently separated using 10% ACN. Guanylmelamine experienced the most significant mobility change compared to the other compounds when the organic modifier content was varied. The peak order changed as its migration time decreased from 7.8 min
(50% ACN) to 6.4 min (10% ACN). The influence of ACN percentage is shown in Fig. 2. Due to the low solubility, especially of melam und ureidomelamine in water, no BGEs with less than 10% ACN were investigated. With the given BGE composition of water/ACN (90:10 v/v) the formic acid concentration was reevaluated to determine the lowest acid concentration needed for baseline separation. Increasing acid concentration in the range of 100 to 500 mM formic acid improved resolution and at 350 mM the most critical peak pairs guanylurea/guanylmelamine and melam/methylmelamine were efficiently separated. A further increase did not yield better results as can be seen in Fig. 3. Since the organic modifier plays a key role methanol was investigated in addition to ACN. Selectivity was comparable to ACN but the runs were slower and melam/methylmelamine not that well separated as can be seen from Fig. 3D. Therefore, ACN was used as organic modifier in all following experiments and the final optimized BGE consisted of 350 mM formic acid in water/ACN (90:10 v/v).
Figure 2. Electropherograms obtained for a 10 mg/L standard with UV-detection at 195 nm. (A) BGE: 350 mM formic acid, 50% ACN; (B) BGE: 350 mM formic acid, 30% ACN; (C) BGE: 350 mM formic acid, 20% ACN; (D) BGE: 350 mM formic acid, 10% ACN. Peak numbering: (1) guanylmelamine; (2) guanylurea; (3) diaminotriazine; (4) melamine; (5) melam; (6) methylmelamine; (7) dimethylmelamine; (8) ureidomelamine
Figure 3. Electropherograms obtained for a 10 mg/L standard with UV-detection at 195 nm. (A) BGE: 200 mM formic acid, 10% ACN; (B) BGE: 350 mM formic acid, 10% ACN; (C) BGE: 500 mM formic acid, 10% ACN; D) BGE: 350 mM formic acid, 10% methanol. Peak numbering: (1) guanylmelamine; (2) guanylurea; (3) diaminotriazine; (4) melamine; (5) melam; (6) methylmelamine; (7) dimethylmelamine; (8) ureidomelamine
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.2 CZE using MS detection One major disadvantage of the above-described CZE-UV approach is its poor LOQ, so hyphenation with MS was
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 1362–1367
1365
Table 1. Method characteristics calculated for UV and MS detection
Compound
LOQa) UV (w%)
RSDb) UV (%)
LOQa) MS (w%)
RSDb) MS (%)
R2c)
Guanylurea Guanylmelamine Diaminotriazine Melamine Methylmelamine Melam Dimethylmelamine Ureidomelamine
0.40 0.68 0.44 0.32 0.39 0.42 0.44 0.39
2.9 1.3 0.9 3.8 3.4 1.2 2.9 3.5
0.05 0.02 0.02 0.01 0.02 0.01 0.01 0.03
9.0 8.5 9.9 4.5 9.1 8.7 6.4 7.4
0.9986 0.9973 0.9960 0.9961 0.9978 0.9977 0.9987 0.9982
a) LOQ calculated as ten times baseline noise given as weight percent assuming that 100 mg of melamine sample is dissolved in 100 mL solvent. b) relative standard deviations of peak areas for five injections of a 10 mg/L standard (UV) and a 1 mg/L standard (MS). c) Correlation coefficient of the calibration curve in the range from the LOQ to 10 mg/L with MS detection.
Figure 4. Extracted ion electropherograms of a 1 mg/L standard solution using optimized ESI-MS conditions.
investigated. All ESI MS-related parameters were carefully optimized to establish a setup that allows the highly sensitive detection of the analytes and at the same time provides stable ESI conditions to ensure reliable results when quantification is needed. A sheath liquid consisting of water/ 2-propanol (20:80 v/v) with 0.1% formic acid was used, which had turned out to be an optimum composition for melamineformaldehyde resins and methylated melamines in previous work [10, 14] and only the most critical parameters like acid concentration and flow rate were investigated. The flow rate investigated in a range from 1 to 6 L/min. At a flow rate of 3 L/min good signal intensities were obtained together with a stable spray and therefore best repeatability. When flow rates higher than 3 L/min were applied signal intensities started to decrease. Furthermore, sheath liquids containing 1% formic acid, 5 mM ammonium formate, and combinations of formic acid and ammonium formate were tested but signal intensities were always lower than with 0.1% formic acid. Nebulizer pressure and drying gas flow rate were varied from 1 to 7 psi and 2 to 10 L/min respectively, whereby the higher nebulizer pressures yielded the highest signals but a negative effect on the separation was observed since the nebulizer created a suction which induced a laminar flow. At a nebulizer pressure of 3 psi the signal intensities were acceptable and separation was maintained. The variation of drying gas flow rates showed only minor changes in signal intensities and a value of 4 L/min was selected. The drying gas temperature was varied from 175 to 350°C also indicating C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
that this parameter does not affect signal intensities significantly; so 250°C was selected. When the MS capillary voltage was raised over a range from 3000 to 4000 V, the signal intensities increased notably. Often a less stable signal is observed at the highest adjustment but in case of melamine impurities this was not the case, therefore the capillary voltage was set to 4000 V. The fragmentor voltage was optimized in a range from 150 to 250 V. Signal intensities remained almost constant from 170 to 225 V followed by a slight decrease when the voltage was increased further; hence 200 V was chosen as the final setting. Figure 4 shows extracted ion electropherograms of a 1 mg/L standard solution using optimized ESI-MS conditions. Employing the optimized ESI MS parameters, calibration was performed with standard mixtures and very good linearity was observed from the LOQ up to 10 mg/L with correlation coefficients in the range from 0.9960 to 0.9987. Repeatability was determined by injecting six standard mixtures in the concentration range from 0.4 mg/L to 10 mg/L five times each. RSDs of the peak areas, calculated from all five injections of the 1 mg/L standard solution, were ⬍10% which is satisfactory for a CZE-MS method without correction by internal standard. Quantification limits were calculated as 10 times the baseline noise giving values between 0.1 mg/L (dimethylmelamine) and 0.5 mg/L (guanylurea). When impurities in melamine samples are analyzed the LOQ should be expressed as weight percent. Assuming that 100 mg of sample is dissolved in 100 mL solvent the resulting LOQs are between 0.01% w/w for dimethylmelamine and 0.05% w/w for guanylurea. The most important method characteristics for MS and UV detection are summarized in Table 1. 3.3 CZE-MS of melamine samples Several melamine samples were analyzed by CZE-ESI-MS and the summarized results are shown in Table 2. Melam www.electrophoresis-journal.com
1366
M. Himmelsbach and T. D. T. Vo
Electrophoresis 2014, 35, 1362–1367
Table 2. Results for the analysis of five industrial melamine samples
Content (%) Compound
Sample 1 Sample 2 Sample 3 Sample 4 Sample 5
Guanylurea Guanylmelamine Diaminotriazine Methylmelamine Melam Dimethylmelamine Ureidomelamine
