- Email: [email protected]
Determination of trace levels of dimethyl sulfate in the presence of monomethyl sulfate by gas chromatography–mass spectrometry
Journal of Chromatography A, 1289 (2013) 139–144
Contents lists available at SciVerse ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Short communication
Determination of trace levels of dimethyl sulfate in the presence of monomethyl sulfate by gas chromatography–mass spectrometry Claudia Schäfer, Peter Zöllner ∗ Bayer Cropscience AG, Research Technologies, Industriepark Höchst, G836, D-65926 Frankfurt/Main, Germany
a r t i c l e
i n f o
Article history: Received 9 November 2012 Received in revised form 7 March 2013 Accepted 8 March 2013 Available online 13 March 2013 Keywords: Dimethyl sulfate Monomethyl sulfate GC–MS Trace analysis
a b s t r a c t During a routine determination of dimethyl sulfate in technical materials using gas chromatography– mass spectrometry (GC–MS), we found that residual monomethyl sulfate originating from a prior methylation reaction with dimethyl sulfate decomposed in the hot GC injection system to yield dimethyl sulfate and sulfuric acid. This thermal reaction leads to false positive or overestimated residue levels of dimethyl sulfate, accompanied by bad chromatographic peak shapes and poor precision and accuracy values. This short communication describes proper measures to counteract this problem and presents a fast, reliable and validated GC–MS method that is capable of determining dimethyl sulfate residues in the presence of monomethyl sulfate in technical materials using a simple dissolve-and-inject approach. Applying deuterated dimethyl sulfate as internal standard and with a sample weight of 25 mg, a limit of detection of 0.24 mg kg−1 and a limit of quantification of 0.48 mg kg−1 was achieved along with a linear range of 0.48–208.6 mg kg−1 . The method offers a precision of 9.1% and an accuracy of 96.5% at the limit of quantification and a precision of 3.6% and an accuracy of 93.8% at a dimethyl sulfate level of 1 mg kg−1 . © 2013 Elsevier B.V. All rights reserved.
1. Introduction The issue of genotoxic impurities in pharmaceuticals, agrochemicals and intermediates thereof is attracting increasing attention from industry and regulatory agencies. Due to their reactive nature some materials used in technical synthesis, including starting materials, intermediates, reagents and some process related impurities/degradants have been demonstrated as being genotoxic. To ensure that these undesired genotoxic impurities are reduced to an acceptable level [1–3], it is critical to monitor them with high accuracy throughout the whole technical manufacturing process and especially in the final product [4]. Dimethyl sulfate (DMS) is an alkylating reagent commonly used in organic synthesis and in technical manufacturing processes leading to agrochemicals and pharmaceuticals [5]. Due to its potential carcinogenicity [6], DMS levels have to be carefully and accurately monitored. Since human DMS uptake is most likely to occur from contaminated air, a considerable part of the published analytical methods deal with DMS monitoring in air samples [7–13]. Only a very few papers address the problem of DMS determination in more complex matrices, predominantly deriving from technical
∗ Corresponding author. Tel.: +49 69 305 12248; fax: +49 69 305 21802. E-mail address: [email protected] (P. Zöllner). 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.03.007
manufacturing processes or organic synthesis of drug substances and their intermediates [14–20]. Due to the high volatility of DMS, trace quantification methods often rely on gas chromatography in combination with different detection methods including flame photometric detection [13], mass spectrometry [7,8,14–16] and phosphorus–nitrogen selective detection [11]. For liquid chromatography based methods a suitable chemical modification of DMS is always necessary to ensure reliable fluorescence [18] and mass spectrometric detection [19]. Almost all methods include one or more sample preparation steps, either for clean-up and analyte enrichment (e.g. liquid/liquid extraction [15,16]; sampling on solid adsorbents [7–13]) or for final DMS detection (chemical derivatization [18,19]). The development of simple, dissolve-and-inject approaches, though less time-consuming and less error prone, has barely been reported [14]. During DMS methylation of, for example a secondary amine, only one methyl group is transferred from the DMS molecule resulting in the formation of monomethyl sulfate (MMS) (Fig. 1a). MMS is a poor methylating agent and is distinctly less toxic than DMS. It is predominantly removed during product clean-up procedures, however, a minor amount of MMS may still be present in samples. During our investigations, it became evident that any remaining MMS in a sample decomposes in the hot GC injector into DMS and sulfuric acid (Fig. 1b). As a direct consequence of this, DMS levels measured by GC quantification were overestimated in samples also containing MMS. This problem has to the best of our knowledge
140
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
a R1
H N
R2
+ H 3C
CH3
O O CH3 S O O
R1
N
+
R2
O O CH3 S O HO
DMS
MMS
R1 = alkyl or aryl R2 = alkyl or aryl
b
2
O O CH3 S O HO
T > 100 °C
H3C
MMS
O O CH3 S O O DMS
+
O O S OH HO
Fig. 1. (a) Reaction scheme of compound methylation with dimethyl sulfate; (b) reaction scheme of thermal formation of dimethyl sulfate from monomethyl sulfate.
not been previously addressed, presumably due to the fact that the presence of MMS residues in previous DMS investigations was not a critical issue (i.e. MMS was not present in air samples [7,13] or was removed prior to GC–MS analysis [15,16]). This short communication describes a fast, reliable and validated GC–MS method that is capable of determining DMS in technical materials, without the need to remove MMS prior to final GC analysis (dissolve-and-inject approach). 2. Materials and methods 2.1. Materials Dichloromethane (99.6%, w/w), dimethyl sulfate (DMS, 99.9%, w/w) and dimethyl sulfate-d6 (DMS-d6, deuterated internal standard, 97.6%, w/w) were purchased from Sigma–Aldrich, Deisenhofen, Germany. The potassium salt of monomethyl sulfate was isolated from a DMS methylation reaction mixture (Bayer Cropscience AG, Frankfurt/Main, Germany). Technically manufactured batches of an N-methylated compound were used as a sample matrix (Bayer Cropscience AG, Frankfurt/Main, Germany). 2.2. Preparation of standard solutions Stock solutions of DMS and DMS-d6 (internal standard) were prepared in dichloromethane, each with a concentration of 1 mg ml−1 . Both stock solutions were diluted 1:100 with dichloromethane (dilution no. 1; 0.01 mg ml−1 ). For spiking experiments dilution no. 1 of DMS was further diluted 1:25 in dichloromethane (dilution no. 2; 0.0004 mg ml−1 ). For the determination of the response factor, a reference mixture of DMS and DMS-d6 was prepared by mixing equal volumes of their dilution no. 1 followed by a 1:25 dilution in dichloromethane (0.0004 mg ml−1 each). All solutions were stable below a temperature of 10 ◦ C for at least one month. 2.3. Instrumentation GC–MS analyses were performed on an Agilent 5973 MSD single quadrupole system equipped with an Agilent 6890 gas chromatograph (Agilent Technologies, Palo Alto, USA) and a Gerstel cold injection system CIS4 (Gerstel, Mühlheim an der Ruhr, Germany). A fused silica DB-XLB capillary column, with a dimenof 30 m × 0.25 mm sion i.d. × 0.25 m film thickness (Agilent Technologies, Palo Alto, USA) was used at a constant helium flow rate of 1.3 ml min−1 . Oven: 40 ◦ C/2 min ⇒ 5 ◦ C min−1 ⇒ 80 ◦ C/0 min ⇒ 100 ◦ C min−1 ⇒ 320 ◦ C/5 min. Injector: 100 ◦ C/6 min ⇒ 200 ◦ C min−1 ⇒ 300 ◦ C/10 min ⇒ 100 ◦ C. Injection volume: 1 l. Split flow: 50 ml min−1 after 1 min operating in the split-less mode. Ion source: 230 ◦ C. Electron impact ionization:
70 eV. The ion source was switched on 4 min after sample injection. Analyte quantification was performed with selected ion monitoring (DMS: m/z 95 and 96; DMS-d6: m/z 98 and 100). Every ion was recorded with a dwell time of 50 ms at low resolution. 2.4. Sample preparation To 25 mg of N-methylated compound was added 50 l of the internal standard solution no. 1 The resulting mixture was dissolved in 1 ml of dichloromethane to give an internal standard concentration of 20 mg kg−1 . For recovery and method sensitivity experiments, appropriate volumes of DMS dilution no. 2 were added. In the course of some of the preliminary experiments the sample solutions were additionally extracted with 2 ml of water in order to remove MMS from the samples. 2.5. Calculation of DMS content The DMS content and recovery were calculated for all analyte/internal standard ion pairs by applying the following equation to calculate the DMS concentration [mg kg−1 ]: ConcDMS [mg kg−1 ] =
ConcDMS-d6 × RFDMS × AreaDMS × 106 ConcSample × AreaDMS-d6
ConcDMS-d6 : spiked amount of internal standard into the sample solution [mg] RFDMS : response factor Area DMS : GC/MS extracted ion chromatogram peak area of DMS ConcSample : weighed sample of technical test item [mg] AreaDMS-d6 : GC/MS extracted ion chromatogram peak area of internal standard To determine the response factor (RF) of DMS a reference mixture containing known amounts of the internal standard DMS-d6 and the analyte DMS in dichloromethane was analysed twice prior to and after each series of samples by GC–MS. The following equation was used to calculate the RF of DMS: RFDMS =
ConcDMS × AreaDMS-d6 ConcDMS-d6 × AreaDMS
ConcDMS-d6 : concentration of internal standard DMS-d6 [mg mL−1 ] RFDMS : response factor AreaDMS : peak area of GC/MS extracted ion chromatogram of DMS ConcDMS : concentration of DMS [mg mL−1 ] AreaDMS-d6 : peak area of GC/MS extracted ion chromatogram of internal standard DMS-d6
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
141
2.6. Method validation
3. Results and discussion
Method sensitivity: the limit of quantification (LOQ) and the limit of detection (LOD) were graphically determined on the basis of signal-to-noise ratios in the presence of the technical sample matrix (S/N = 10:1 and S/N = 3:1, respectively). Linearity: in order to establish the linear working range, eight DMS standard solutions within a concentration range of 0.012 and 5.3 g ml−1 were analysed. They correspond to DMS concentration levels in technical materials of between 0.48 and 208.6 mg kg−1 (based on a sample weight of 25 mg). Prior to injection all calibration solutions were spiked with the internal standard. Accuracy and precision were determined as follows: five samples of the technical test item that did not contain any detectable amounts of DMS were spiked with a DMS concentration of 0.5 mg kg−1 . A sixth sample was left unspiked (only internal standard was added). This set up was then repeated for a DMS concentration level of 1 mg kg−1 . To calculate the recovery values, the content of DMS was determined in the spiked test item. The obtained values were compared with the expected values according to the following equation:
The typical methylation reaction of a secondary amine group with DMS, as carried out in the course of these investigations, is depicted in Fig. 1a. The methylated reaction product (tertiary amine) which precipitated upon cooling the reaction mixture, was filtered off, washed and analysed by GC–MS. Due to the toxicological properties of DMS, 1 ppm was set as the maximum acceptable level in the methylated technical product [2,3,10,11,13]. A simple dissolve-and-inject GC–MS method was set up with an injector temperature of 250 ◦ C enabling the detection of DMS well below this concentration limit. Commercially available deuterated DMS (DMS-d6) was included as internal standard in order to compensate for matrix effects and for variations of detector sensitivity. This analytical approach provided good results even at low DMS levels (Fig. 2a, DMS level below LOD). However, when the solvent from which the methylated reaction product was precipitated was changed from aqueous/polar conditions (approx. 30% water in acetontitrile) to organic/non-polar conditions (<6% water in butanol and acetonitrile) then the detected DMS levels dramatically increased from sub ppm levels up to a 20 to several hundred ppm range. This was accompanied by a bad shape of the DMS peak (as in Fig. 2b) and by poor precision and recovery values. In distinct contrast to this, the good peak shape and peak width of the
Recovery [%] =
(ConcSpiked − ConcUnspiked ) × 100% Concexpected
Concspiked : concentration found in the spiked sample [mg kg−1 ] Concunspiked : concentration found in the unspiked sample [mg kg−1 ] Concexpected : concentration spiked into the sample [mg kg−1 ] The mean recovery of these five recovery experiments gives the accuracy of the method, while the precision is expressed as the mean relative standard deviation of the five recovery experiments from this accuracy value. Method specificity: a minimum of two ions, each for DMS and the internal standard DMS-d6, were monitored in unspiked technical material. No significant interferences in any of the investigated ion traces of the analyte or the internal standard were visible in the GC/MS chromatograms proving that the selectivity of the method was sufficient. The analyte identity was confirmed by the identical chromatographic retention times of DMS and DMS-d6 in the standard solution and in spiked technical material, along with highly selective MS detection of at least two analyte-specific ions and their relative intensity ratio (±15%). The chemical stability of DMS and the internal standard DMSd6 in dichloromethane standard solutions were determined as follows: a fresh reference solution of the internal standard was prepared in dichloromethane and kept below a temperature of 10 ◦ C. On the first day and after 7, 14, 21 and 31 days, a known amount of DMS was added to an aliquot of this solution and the response factor determined by GC–MS. The response factor remained constant (RSD 2.1%) over this time range proving that the internal standard DMS-d6 and, consequently, also DMS are chemically stable over a period of one month. The isotopic stability (H/D exchange) of the labelled internal standard was checked as follows: a solution of DMS-d6 was prepared and kept below a temperature of 10 ◦ C. This solution was injected six times establishing the ion traces of m/z 132, 131, 98 and 97 in the selected ion monitoring mode. This procedure was performed on the first day and after 7, 14, 21 and 31 days. The results were evaluated for any deviations of ion ratios of m/z 132/131 and of m/z 98/97 which would indicate H/D exchange. No H/D exchange could be observed over a period of one month (RSD 1.4%).
Fig. 2. Sample with no detectable DMS after methylation with DMS and precipitation in aqueous solution. Injector temperature: 250 ◦ C, SIM traces DMS: m/z 95, 96; DMS-d6: m/z 98, 100; (a) direct injection of sample solution; (b) same sample fortified with 1000 mg kg−1 MMS without sample clean up and direct injection (“DMS” ≡ DMS produced from MMS in the GC injection system).
142
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
internal standard DMS-d6 remained unaffected (Fig. 2b, note: due to an isotope effect DMS-d6 elutes 0.1 min earlier than its nonlabelled analogue DMS). The addition of nucleophilic compounds to destroy remaining DMS in the reaction mixture prior to analysis did not reduce DMS levels or improve the quality of the GC–MS analysis. During the methylation reaction, the DMS is converted to monomethyl sulfate (MMS) (Fig. 1a). It was postulated that upon changing from aqueous/polar to organic/non-polar solvent conditions a minor part of this MMS remained in the methylated product, as it is more polar than DMS and less soluble in non-polar organic solvents and was, therefore, partly deposited in the precipitating methylated product. In contrast to DMS, MMS has almost no methylation ability and is distinctly less toxic. Its presence is, therefore, of no toxicological relevance, however, it is reported that it decomposes into DMS and sulfuric acid under elevated temperatures above 140 ◦ C (Fig. 1b) [21]. In an attempt to explain the bad peak shape of DMS (but not of DMS-d6), it was proposed that DMS was being formed from MMS in the hot split/splitless injector of the gas chromatograph, thereby raising the detected DMS levels. In order to test this proposal, a sample of the methylated product that contained only DMS below the LOD (Fig. 2a) (methylation was carried out in an aqueous environment) was fortified with 1000 mg kg−1 of the potassium salt of MMS and reinvestigated by GC–MS. As depicted in Fig. 2b a significant, but badly shaped “DMS” peak appeared in the GC–MS chromatogram of the spiked sample proving that the source of these false DMS levels was due to MMS decomposition in the hot GC injection system. As a consequence of these results, a liquid/liquid extraction step with water was implemented in order to remove MMS from the methylated product. As shown in Fig. 3a and b, this procedure removed almost completely MMS from the sample and avoided the formation of DMS from MMS in the injection system. Peak shape and peak width were strikingly improved and only the low DMS level (below 1 ppm) actually present in the sample was detected. Recovery experiments with DMS spiked samples demonstrated that this effect was not caused by decomposition/extraction of DMS during liquid/liquid extraction (recovery: 98%; precision: 4%). However, a major disadvantage of this approach was that the additional sample preparation step was time consuming and might be error prone. Especially in view of the pre-requisites for a fast and simple quality control method this was a critical drawback. Consequently, we attempted to avoid the necessity for removing MMS, by reducing the temperature of the GC injection system to below 140 ◦ C (the temperature at which DMS is reported to be formed from MMS). For this purpose a programmed temperature vaporizer (PTV) injection system was needed, as the injection temperature needs to be raised several minutes after injection in order to remove the methylated product from the injection system. An injector temperature of 100 ◦ C was found to be optimal for the sample injection process and until DMS eluted from the column, since thermal formation of DMS from MMS could only be observed above this injector temperature. Afterwards, the injector temperature was increased to 300 ◦ C for the rest of the run time. Fig. 4 compares two GC–MS chromatograms of a DMS spiked sample analysed either using a hot injector port held at 250 ◦ C (Fig. 4a) or according to the newly developed procedure using an injector temperature of 100 ◦ C (Fig. 4b). In contrast to the results shown in Fig. 4a (overestimated DMS concentration, bad peak shape), the peak shape and peak width in Fig. 4b are strikingly improved and are comparable to those samples which were extracted with water (Fig. 3b). Most importantly, only DMS contamination that was actually present in the sample was detected.
Fig. 3. Sample with 0.9 mg kg−1 DMS after methylation with DMS and precipitation in a non-polar organic solvent; injector temperature: 250 ◦ C, SIM traces DMS: m/z 95, 96; DMS-d6: m/z 98, 100; (a) without sample clean up and direct injection (“DMS” ≡ DMS produced from MMS in the GC injection system) (b) with sample clean up by liquid/liquid extraction with water.
A method validation has been performed for the 100 ◦ C injector temperature modification described above. The results are summarized in Table 1 and prove that the method is able to accurately and precisely quantify DMS down to the sub ppm level. With regard to sensitivity, selectivity and accuracy the method offers comparable or superior validation characteristics to other DMS methods, however, with significantly less effort for sample preparation (Table 1). As potential interfering compound peaks cannot be predicted for other sample matrices, the application of this method to other (technical or synthetic) samples would require a new check of method selectivity. In addition, since this method is a dissolve-and-inject approach, the LOD and LOQ will always depend upon the initial sample weight and this may vary from matrix to matrix, e.g. due to different solubility of samples in dichloromethane or other suitable solvents. The developed method has been routinely applied in numerous technical batches to determine traces of DMS within our laboratories at Bayer. Due to the injection of high concentrations of technical compound, careful cleaning of the injection system is needed when DMS peak shape deteriorates or ghost peaks related to the methylated technical compound show up in the GC–MS chromatograms of other samples (app. every 250 samples). Moreover, no negative impact on other GC–MS analyses, resulting from possible accumulation of sample matrix or sulfuric acid (degradation product of MMS) in the GC–MS System could be observed.
113.8/101.0 (2.5/7.5 mg kg−1 ) 1.3/0.8 (5.0/50.0 mg kg−1 )
c
Not available 0.5 mg kg−1
≥0.9999
≥0.9928
0.01–10 mg L−1
0.5–100 mg kg−1
b
13.0/2.0 (0.01/10 mg L−1 )
b
0.002 mg L−1 b
0.006 mg L−1
≥0.9996 1.0–60 mg kg−1
b
0.3 mg kg−1 1.0 mg kg−1
≥0.9998 3.0–45 mg kg−1
b
b
1.0 mg kg−1 3.0 mg kg−1
b
0.02 mg kg−1 b
0.2 mg kg−1 b
≥0.9950
≥0.9970 0.25–50 mg kg
0.2–20 mg kg−1
0.04 mg kg 0.09 mg kg
c −1 c −1
≥0.9999
0.5 (1.0 mg kg−1 ) b 1.34 (3.0 mg kg−1 ) Not available
96.5/93.8 (0.5/1.0 mg kg−1 ) >90 (exact value not available) 102.0 (1.0 mg kg−1 ) 99.1 (3.0 mg kg−1 ) 105.3 (8.0 mg kg−1 ) 98.9 (1.0 mg L−1 ) 9.1/3.6 (0.5/1.0 mg kg−1 ) c 6.5 (1.0 mg kg−1 ) −1
c
0.24 mg kg−1 c
0.48 mg kg−1
Mean recovery [%] RSD [%]
0.48–208.6 mg kg−1
c
Accuracya , c
Correlation coefficient r Linearity rangea , b
143
Fig. 4. Sample with no detectable DMS after methylation with DMS and precipitation in non-polar organic solvent fortified with 0.9 mg kg−1 DMS; SIM traces DMS: m/z 95, 96; DMS-d6: m/z 98, 100; (a) injector temperature: 250 ◦ C (“DMS” ≡ DMS produced from MMS in the GC injection system) (b) injector temperature: 100 ◦ C.
c
100 LC–MS/(MS)
Values are based on the respective sample loadings. Determined in standard solutions. Determined in matrix matched samples.
150 LC–FL
a
25 GC–MS
b
1000 GC–MS
Raman et al. [14] Zheng et al. [15] Hoogerheide and Scott [18] Grinberg et al. [19]
Liquid/liquid extraction Derivatization
Derivatization 5 LC–MS An et al. [17]
None
Derivatization 50 GC–MS Alzaga et al. [16]
None 25 GC–MS This method
Sample weight [mg]
Sample preparation Analytical technique
Derivatization
4. Conclusions
Reference
Table 1 DMS determination in complex matrices (drugs, agrochemicals, technical intermediates, reaction solutions).
LOQa
LODa
Precisiona
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
A reliable and validated GC–MS method for the determination of traces of dimethyl sulfate in DMS methylated technical or synthetic materials is presented. It is based upon a simple dissolve-andinject approach and enables the quantification of DMS even in the presence of the MMS which is generated during the methylation reaction with DMS. Thermal formation of DMS from MMS in the hot GC injector is avoided by initial reduction of the injector temperature to 100 ◦ C enabling detection of actual DMS levels down to a sub ppm level. Apart from the MMS issue, this method is also comparable or superior in terms of sensitivity, accuracy and precision to previously published methods [14–19] (Table 1) and avoids any time consuming sample preparation steps. The use of the stable isotope labelled internal standard (DMS-d6) is regarded as an important measure in order to ensure maximum reliability and robustness for this quantitative GC–MS method. As DMS-d6 is commercially available at low costs, the new method can be easily implemented. Finally, the method can be conveniently adapted to other sample matrices, although method selectivity, LOD and LOQ would need to be verified beforehand. References [1] D.A. Pierson, B.A. Olsen, D.K. Robbins, K.M.D. DeVries, I. Varie, Org. Process Res. Dev. 13 (2009) 285.
144
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
[2] Committee for Medicinal Products for Human Use (CHMP), Europen Medicines Agenca (EMEA), Guidline on the Limits of Genotoxic Impurities, London, UK (EMEA/CHMP/QWP/251344/2006), June 2006. [3] U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Guidance for Industry – Genotoxin and Carcinogenic Impurities in Drug Substances and Products: Recommended Approaches, Silver Spring, MD, USA, 2008, December. [4] D.Q. Liu, M. Sun, A.S. Kord, J. Pharm, Biomed. Anal. 51 (2010) 999. [5] Manfred Müller, Ute Hübsch: dimethyl ether, in: Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2006. [6] Dimethyl sulfate, IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, World Health Organization, International Agency for Research on Cancer, 71, 1999, 575. [7] E. Scobbie, J.A. Groves, Ann. Occup. Hyg. 42 (1998) 201. [8] T. Dublin, H.J. Thöne, J. Chromatogr. 456 (1986) 233. [9] K.S. Sidhu, J. Chromatogr. 206 (1981) 381. [10] L.D. Hansen, V.F. White, D.J. Eatough, Environ. Sci. Technol. 20 (1986) 572. [11] H. Frind, K. Trageser, Z. Fresenius, Anal. Chem. 326 (1987) 517.
[12] S. Fukui, M.M. Morishima, S. Ogawa, Y. Hanazaki, J. Chromatogr. 541 (1991) 459. [13] J.V. Das, K.N. Ramachandran, V.K. Gupta, Micorchem. J. 50 (1994) 51. [14] N.V.V.S.S. Raman, K.R. Reddy, A.V.S.S. Prasad, K. Ramakrishna, Chromatographia 68 (2008) 857. [15] J. Zheng, W.A. Pritts, S. Zhang, S. Wittenberger, J. Pharm. Biomed. Anal. 50 (2009) 1054. [16] R. Alzaga, R.W. Ryan, K. Taylor-Worth, A.M. Lipczynski, R. Szucs, P. Sandra, J. Pharm. Biomed. Anal. 45 (2007) 472. [17] J. An, M. Sun, L. Bai, T. Chen, D.Q. Liu, A. Kord, J. Pharm. Biomed. Anal. 48 (2008) 1006. [18] J.G. Hoogerheide, R.A. Scott, Talanta 65 (2005) 453. [19] N. Grinberg, F. Albu, K. Fandrick, E. Iorgulescu, A. Medvedovici, J. Pharm. Biomed. Anal. 75 (2013) 1. [20] D.P. Elder, A. Teasdale, A.M. Lipczynski, J. Pharm. Biomed. Anal. 46 (2008) 1. [21] Ch. Simmond, Alcohol, Its Production, Properties, Chemistry, and Industrial Applications, Cambridge Scholar Publishing, Cambrigde, 2009.
Contents lists available at SciVerse ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Short communication
Determination of trace levels of dimethyl sulfate in the presence of monomethyl sulfate by gas chromatography–mass spectrometry Claudia Schäfer, Peter Zöllner ∗ Bayer Cropscience AG, Research Technologies, Industriepark Höchst, G836, D-65926 Frankfurt/Main, Germany
a r t i c l e
i n f o
Article history: Received 9 November 2012 Received in revised form 7 March 2013 Accepted 8 March 2013 Available online 13 March 2013 Keywords: Dimethyl sulfate Monomethyl sulfate GC–MS Trace analysis
a b s t r a c t During a routine determination of dimethyl sulfate in technical materials using gas chromatography– mass spectrometry (GC–MS), we found that residual monomethyl sulfate originating from a prior methylation reaction with dimethyl sulfate decomposed in the hot GC injection system to yield dimethyl sulfate and sulfuric acid. This thermal reaction leads to false positive or overestimated residue levels of dimethyl sulfate, accompanied by bad chromatographic peak shapes and poor precision and accuracy values. This short communication describes proper measures to counteract this problem and presents a fast, reliable and validated GC–MS method that is capable of determining dimethyl sulfate residues in the presence of monomethyl sulfate in technical materials using a simple dissolve-and-inject approach. Applying deuterated dimethyl sulfate as internal standard and with a sample weight of 25 mg, a limit of detection of 0.24 mg kg−1 and a limit of quantification of 0.48 mg kg−1 was achieved along with a linear range of 0.48–208.6 mg kg−1 . The method offers a precision of 9.1% and an accuracy of 96.5% at the limit of quantification and a precision of 3.6% and an accuracy of 93.8% at a dimethyl sulfate level of 1 mg kg−1 . © 2013 Elsevier B.V. All rights reserved.
1. Introduction The issue of genotoxic impurities in pharmaceuticals, agrochemicals and intermediates thereof is attracting increasing attention from industry and regulatory agencies. Due to their reactive nature some materials used in technical synthesis, including starting materials, intermediates, reagents and some process related impurities/degradants have been demonstrated as being genotoxic. To ensure that these undesired genotoxic impurities are reduced to an acceptable level [1–3], it is critical to monitor them with high accuracy throughout the whole technical manufacturing process and especially in the final product [4]. Dimethyl sulfate (DMS) is an alkylating reagent commonly used in organic synthesis and in technical manufacturing processes leading to agrochemicals and pharmaceuticals [5]. Due to its potential carcinogenicity [6], DMS levels have to be carefully and accurately monitored. Since human DMS uptake is most likely to occur from contaminated air, a considerable part of the published analytical methods deal with DMS monitoring in air samples [7–13]. Only a very few papers address the problem of DMS determination in more complex matrices, predominantly deriving from technical
∗ Corresponding author. Tel.: +49 69 305 12248; fax: +49 69 305 21802. E-mail address: [email protected] (P. Zöllner). 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.03.007
manufacturing processes or organic synthesis of drug substances and their intermediates [14–20]. Due to the high volatility of DMS, trace quantification methods often rely on gas chromatography in combination with different detection methods including flame photometric detection [13], mass spectrometry [7,8,14–16] and phosphorus–nitrogen selective detection [11]. For liquid chromatography based methods a suitable chemical modification of DMS is always necessary to ensure reliable fluorescence [18] and mass spectrometric detection [19]. Almost all methods include one or more sample preparation steps, either for clean-up and analyte enrichment (e.g. liquid/liquid extraction [15,16]; sampling on solid adsorbents [7–13]) or for final DMS detection (chemical derivatization [18,19]). The development of simple, dissolve-and-inject approaches, though less time-consuming and less error prone, has barely been reported [14]. During DMS methylation of, for example a secondary amine, only one methyl group is transferred from the DMS molecule resulting in the formation of monomethyl sulfate (MMS) (Fig. 1a). MMS is a poor methylating agent and is distinctly less toxic than DMS. It is predominantly removed during product clean-up procedures, however, a minor amount of MMS may still be present in samples. During our investigations, it became evident that any remaining MMS in a sample decomposes in the hot GC injector into DMS and sulfuric acid (Fig. 1b). As a direct consequence of this, DMS levels measured by GC quantification were overestimated in samples also containing MMS. This problem has to the best of our knowledge
140
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
a R1
H N
R2
+ H 3C
CH3
O O CH3 S O O
R1
N
+
R2
O O CH3 S O HO
DMS
MMS
R1 = alkyl or aryl R2 = alkyl or aryl
b
2
O O CH3 S O HO
T > 100 °C
H3C
MMS
O O CH3 S O O DMS
+
O O S OH HO
Fig. 1. (a) Reaction scheme of compound methylation with dimethyl sulfate; (b) reaction scheme of thermal formation of dimethyl sulfate from monomethyl sulfate.
not been previously addressed, presumably due to the fact that the presence of MMS residues in previous DMS investigations was not a critical issue (i.e. MMS was not present in air samples [7,13] or was removed prior to GC–MS analysis [15,16]). This short communication describes a fast, reliable and validated GC–MS method that is capable of determining DMS in technical materials, without the need to remove MMS prior to final GC analysis (dissolve-and-inject approach). 2. Materials and methods 2.1. Materials Dichloromethane (99.6%, w/w), dimethyl sulfate (DMS, 99.9%, w/w) and dimethyl sulfate-d6 (DMS-d6, deuterated internal standard, 97.6%, w/w) were purchased from Sigma–Aldrich, Deisenhofen, Germany. The potassium salt of monomethyl sulfate was isolated from a DMS methylation reaction mixture (Bayer Cropscience AG, Frankfurt/Main, Germany). Technically manufactured batches of an N-methylated compound were used as a sample matrix (Bayer Cropscience AG, Frankfurt/Main, Germany). 2.2. Preparation of standard solutions Stock solutions of DMS and DMS-d6 (internal standard) were prepared in dichloromethane, each with a concentration of 1 mg ml−1 . Both stock solutions were diluted 1:100 with dichloromethane (dilution no. 1; 0.01 mg ml−1 ). For spiking experiments dilution no. 1 of DMS was further diluted 1:25 in dichloromethane (dilution no. 2; 0.0004 mg ml−1 ). For the determination of the response factor, a reference mixture of DMS and DMS-d6 was prepared by mixing equal volumes of their dilution no. 1 followed by a 1:25 dilution in dichloromethane (0.0004 mg ml−1 each). All solutions were stable below a temperature of 10 ◦ C for at least one month. 2.3. Instrumentation GC–MS analyses were performed on an Agilent 5973 MSD single quadrupole system equipped with an Agilent 6890 gas chromatograph (Agilent Technologies, Palo Alto, USA) and a Gerstel cold injection system CIS4 (Gerstel, Mühlheim an der Ruhr, Germany). A fused silica DB-XLB capillary column, with a dimenof 30 m × 0.25 mm sion i.d. × 0.25 m film thickness (Agilent Technologies, Palo Alto, USA) was used at a constant helium flow rate of 1.3 ml min−1 . Oven: 40 ◦ C/2 min ⇒ 5 ◦ C min−1 ⇒ 80 ◦ C/0 min ⇒ 100 ◦ C min−1 ⇒ 320 ◦ C/5 min. Injector: 100 ◦ C/6 min ⇒ 200 ◦ C min−1 ⇒ 300 ◦ C/10 min ⇒ 100 ◦ C. Injection volume: 1 l. Split flow: 50 ml min−1 after 1 min operating in the split-less mode. Ion source: 230 ◦ C. Electron impact ionization:
70 eV. The ion source was switched on 4 min after sample injection. Analyte quantification was performed with selected ion monitoring (DMS: m/z 95 and 96; DMS-d6: m/z 98 and 100). Every ion was recorded with a dwell time of 50 ms at low resolution. 2.4. Sample preparation To 25 mg of N-methylated compound was added 50 l of the internal standard solution no. 1 The resulting mixture was dissolved in 1 ml of dichloromethane to give an internal standard concentration of 20 mg kg−1 . For recovery and method sensitivity experiments, appropriate volumes of DMS dilution no. 2 were added. In the course of some of the preliminary experiments the sample solutions were additionally extracted with 2 ml of water in order to remove MMS from the samples. 2.5. Calculation of DMS content The DMS content and recovery were calculated for all analyte/internal standard ion pairs by applying the following equation to calculate the DMS concentration [mg kg−1 ]: ConcDMS [mg kg−1 ] =
ConcDMS-d6 × RFDMS × AreaDMS × 106 ConcSample × AreaDMS-d6
ConcDMS-d6 : spiked amount of internal standard into the sample solution [mg] RFDMS : response factor Area DMS : GC/MS extracted ion chromatogram peak area of DMS ConcSample : weighed sample of technical test item [mg] AreaDMS-d6 : GC/MS extracted ion chromatogram peak area of internal standard To determine the response factor (RF) of DMS a reference mixture containing known amounts of the internal standard DMS-d6 and the analyte DMS in dichloromethane was analysed twice prior to and after each series of samples by GC–MS. The following equation was used to calculate the RF of DMS: RFDMS =
ConcDMS × AreaDMS-d6 ConcDMS-d6 × AreaDMS
ConcDMS-d6 : concentration of internal standard DMS-d6 [mg mL−1 ] RFDMS : response factor AreaDMS : peak area of GC/MS extracted ion chromatogram of DMS ConcDMS : concentration of DMS [mg mL−1 ] AreaDMS-d6 : peak area of GC/MS extracted ion chromatogram of internal standard DMS-d6
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
141
2.6. Method validation
3. Results and discussion
Method sensitivity: the limit of quantification (LOQ) and the limit of detection (LOD) were graphically determined on the basis of signal-to-noise ratios in the presence of the technical sample matrix (S/N = 10:1 and S/N = 3:1, respectively). Linearity: in order to establish the linear working range, eight DMS standard solutions within a concentration range of 0.012 and 5.3 g ml−1 were analysed. They correspond to DMS concentration levels in technical materials of between 0.48 and 208.6 mg kg−1 (based on a sample weight of 25 mg). Prior to injection all calibration solutions were spiked with the internal standard. Accuracy and precision were determined as follows: five samples of the technical test item that did not contain any detectable amounts of DMS were spiked with a DMS concentration of 0.5 mg kg−1 . A sixth sample was left unspiked (only internal standard was added). This set up was then repeated for a DMS concentration level of 1 mg kg−1 . To calculate the recovery values, the content of DMS was determined in the spiked test item. The obtained values were compared with the expected values according to the following equation:
The typical methylation reaction of a secondary amine group with DMS, as carried out in the course of these investigations, is depicted in Fig. 1a. The methylated reaction product (tertiary amine) which precipitated upon cooling the reaction mixture, was filtered off, washed and analysed by GC–MS. Due to the toxicological properties of DMS, 1 ppm was set as the maximum acceptable level in the methylated technical product [2,3,10,11,13]. A simple dissolve-and-inject GC–MS method was set up with an injector temperature of 250 ◦ C enabling the detection of DMS well below this concentration limit. Commercially available deuterated DMS (DMS-d6) was included as internal standard in order to compensate for matrix effects and for variations of detector sensitivity. This analytical approach provided good results even at low DMS levels (Fig. 2a, DMS level below LOD). However, when the solvent from which the methylated reaction product was precipitated was changed from aqueous/polar conditions (approx. 30% water in acetontitrile) to organic/non-polar conditions (<6% water in butanol and acetonitrile) then the detected DMS levels dramatically increased from sub ppm levels up to a 20 to several hundred ppm range. This was accompanied by a bad shape of the DMS peak (as in Fig. 2b) and by poor precision and recovery values. In distinct contrast to this, the good peak shape and peak width of the
Recovery [%] =
(ConcSpiked − ConcUnspiked ) × 100% Concexpected
Concspiked : concentration found in the spiked sample [mg kg−1 ] Concunspiked : concentration found in the unspiked sample [mg kg−1 ] Concexpected : concentration spiked into the sample [mg kg−1 ] The mean recovery of these five recovery experiments gives the accuracy of the method, while the precision is expressed as the mean relative standard deviation of the five recovery experiments from this accuracy value. Method specificity: a minimum of two ions, each for DMS and the internal standard DMS-d6, were monitored in unspiked technical material. No significant interferences in any of the investigated ion traces of the analyte or the internal standard were visible in the GC/MS chromatograms proving that the selectivity of the method was sufficient. The analyte identity was confirmed by the identical chromatographic retention times of DMS and DMS-d6 in the standard solution and in spiked technical material, along with highly selective MS detection of at least two analyte-specific ions and their relative intensity ratio (±15%). The chemical stability of DMS and the internal standard DMSd6 in dichloromethane standard solutions were determined as follows: a fresh reference solution of the internal standard was prepared in dichloromethane and kept below a temperature of 10 ◦ C. On the first day and after 7, 14, 21 and 31 days, a known amount of DMS was added to an aliquot of this solution and the response factor determined by GC–MS. The response factor remained constant (RSD 2.1%) over this time range proving that the internal standard DMS-d6 and, consequently, also DMS are chemically stable over a period of one month. The isotopic stability (H/D exchange) of the labelled internal standard was checked as follows: a solution of DMS-d6 was prepared and kept below a temperature of 10 ◦ C. This solution was injected six times establishing the ion traces of m/z 132, 131, 98 and 97 in the selected ion monitoring mode. This procedure was performed on the first day and after 7, 14, 21 and 31 days. The results were evaluated for any deviations of ion ratios of m/z 132/131 and of m/z 98/97 which would indicate H/D exchange. No H/D exchange could be observed over a period of one month (RSD 1.4%).
Fig. 2. Sample with no detectable DMS after methylation with DMS and precipitation in aqueous solution. Injector temperature: 250 ◦ C, SIM traces DMS: m/z 95, 96; DMS-d6: m/z 98, 100; (a) direct injection of sample solution; (b) same sample fortified with 1000 mg kg−1 MMS without sample clean up and direct injection (“DMS” ≡ DMS produced from MMS in the GC injection system).
142
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
internal standard DMS-d6 remained unaffected (Fig. 2b, note: due to an isotope effect DMS-d6 elutes 0.1 min earlier than its nonlabelled analogue DMS). The addition of nucleophilic compounds to destroy remaining DMS in the reaction mixture prior to analysis did not reduce DMS levels or improve the quality of the GC–MS analysis. During the methylation reaction, the DMS is converted to monomethyl sulfate (MMS) (Fig. 1a). It was postulated that upon changing from aqueous/polar to organic/non-polar solvent conditions a minor part of this MMS remained in the methylated product, as it is more polar than DMS and less soluble in non-polar organic solvents and was, therefore, partly deposited in the precipitating methylated product. In contrast to DMS, MMS has almost no methylation ability and is distinctly less toxic. Its presence is, therefore, of no toxicological relevance, however, it is reported that it decomposes into DMS and sulfuric acid under elevated temperatures above 140 ◦ C (Fig. 1b) [21]. In an attempt to explain the bad peak shape of DMS (but not of DMS-d6), it was proposed that DMS was being formed from MMS in the hot split/splitless injector of the gas chromatograph, thereby raising the detected DMS levels. In order to test this proposal, a sample of the methylated product that contained only DMS below the LOD (Fig. 2a) (methylation was carried out in an aqueous environment) was fortified with 1000 mg kg−1 of the potassium salt of MMS and reinvestigated by GC–MS. As depicted in Fig. 2b a significant, but badly shaped “DMS” peak appeared in the GC–MS chromatogram of the spiked sample proving that the source of these false DMS levels was due to MMS decomposition in the hot GC injection system. As a consequence of these results, a liquid/liquid extraction step with water was implemented in order to remove MMS from the methylated product. As shown in Fig. 3a and b, this procedure removed almost completely MMS from the sample and avoided the formation of DMS from MMS in the injection system. Peak shape and peak width were strikingly improved and only the low DMS level (below 1 ppm) actually present in the sample was detected. Recovery experiments with DMS spiked samples demonstrated that this effect was not caused by decomposition/extraction of DMS during liquid/liquid extraction (recovery: 98%; precision: 4%). However, a major disadvantage of this approach was that the additional sample preparation step was time consuming and might be error prone. Especially in view of the pre-requisites for a fast and simple quality control method this was a critical drawback. Consequently, we attempted to avoid the necessity for removing MMS, by reducing the temperature of the GC injection system to below 140 ◦ C (the temperature at which DMS is reported to be formed from MMS). For this purpose a programmed temperature vaporizer (PTV) injection system was needed, as the injection temperature needs to be raised several minutes after injection in order to remove the methylated product from the injection system. An injector temperature of 100 ◦ C was found to be optimal for the sample injection process and until DMS eluted from the column, since thermal formation of DMS from MMS could only be observed above this injector temperature. Afterwards, the injector temperature was increased to 300 ◦ C for the rest of the run time. Fig. 4 compares two GC–MS chromatograms of a DMS spiked sample analysed either using a hot injector port held at 250 ◦ C (Fig. 4a) or according to the newly developed procedure using an injector temperature of 100 ◦ C (Fig. 4b). In contrast to the results shown in Fig. 4a (overestimated DMS concentration, bad peak shape), the peak shape and peak width in Fig. 4b are strikingly improved and are comparable to those samples which were extracted with water (Fig. 3b). Most importantly, only DMS contamination that was actually present in the sample was detected.
Fig. 3. Sample with 0.9 mg kg−1 DMS after methylation with DMS and precipitation in a non-polar organic solvent; injector temperature: 250 ◦ C, SIM traces DMS: m/z 95, 96; DMS-d6: m/z 98, 100; (a) without sample clean up and direct injection (“DMS” ≡ DMS produced from MMS in the GC injection system) (b) with sample clean up by liquid/liquid extraction with water.
A method validation has been performed for the 100 ◦ C injector temperature modification described above. The results are summarized in Table 1 and prove that the method is able to accurately and precisely quantify DMS down to the sub ppm level. With regard to sensitivity, selectivity and accuracy the method offers comparable or superior validation characteristics to other DMS methods, however, with significantly less effort for sample preparation (Table 1). As potential interfering compound peaks cannot be predicted for other sample matrices, the application of this method to other (technical or synthetic) samples would require a new check of method selectivity. In addition, since this method is a dissolve-and-inject approach, the LOD and LOQ will always depend upon the initial sample weight and this may vary from matrix to matrix, e.g. due to different solubility of samples in dichloromethane or other suitable solvents. The developed method has been routinely applied in numerous technical batches to determine traces of DMS within our laboratories at Bayer. Due to the injection of high concentrations of technical compound, careful cleaning of the injection system is needed when DMS peak shape deteriorates or ghost peaks related to the methylated technical compound show up in the GC–MS chromatograms of other samples (app. every 250 samples). Moreover, no negative impact on other GC–MS analyses, resulting from possible accumulation of sample matrix or sulfuric acid (degradation product of MMS) in the GC–MS System could be observed.
113.8/101.0 (2.5/7.5 mg kg−1 ) 1.3/0.8 (5.0/50.0 mg kg−1 )
c
Not available 0.5 mg kg−1
≥0.9999
≥0.9928
0.01–10 mg L−1
0.5–100 mg kg−1
b
13.0/2.0 (0.01/10 mg L−1 )
b
0.002 mg L−1 b
0.006 mg L−1
≥0.9996 1.0–60 mg kg−1
b
0.3 mg kg−1 1.0 mg kg−1
≥0.9998 3.0–45 mg kg−1
b
b
1.0 mg kg−1 3.0 mg kg−1
b
0.02 mg kg−1 b
0.2 mg kg−1 b
≥0.9950
≥0.9970 0.25–50 mg kg
0.2–20 mg kg−1
0.04 mg kg 0.09 mg kg
c −1 c −1
≥0.9999
0.5 (1.0 mg kg−1 ) b 1.34 (3.0 mg kg−1 ) Not available
96.5/93.8 (0.5/1.0 mg kg−1 ) >90 (exact value not available) 102.0 (1.0 mg kg−1 ) 99.1 (3.0 mg kg−1 ) 105.3 (8.0 mg kg−1 ) 98.9 (1.0 mg L−1 ) 9.1/3.6 (0.5/1.0 mg kg−1 ) c 6.5 (1.0 mg kg−1 ) −1
c
0.24 mg kg−1 c
0.48 mg kg−1
Mean recovery [%] RSD [%]
0.48–208.6 mg kg−1
c
Accuracya , c
Correlation coefficient r Linearity rangea , b
143
Fig. 4. Sample with no detectable DMS after methylation with DMS and precipitation in non-polar organic solvent fortified with 0.9 mg kg−1 DMS; SIM traces DMS: m/z 95, 96; DMS-d6: m/z 98, 100; (a) injector temperature: 250 ◦ C (“DMS” ≡ DMS produced from MMS in the GC injection system) (b) injector temperature: 100 ◦ C.
c
100 LC–MS/(MS)
Values are based on the respective sample loadings. Determined in standard solutions. Determined in matrix matched samples.
150 LC–FL
a
25 GC–MS
b
1000 GC–MS
Raman et al. [14] Zheng et al. [15] Hoogerheide and Scott [18] Grinberg et al. [19]
Liquid/liquid extraction Derivatization
Derivatization 5 LC–MS An et al. [17]
None
Derivatization 50 GC–MS Alzaga et al. [16]
None 25 GC–MS This method
Sample weight [mg]
Sample preparation Analytical technique
Derivatization
4. Conclusions
Reference
Table 1 DMS determination in complex matrices (drugs, agrochemicals, technical intermediates, reaction solutions).
LOQa
LODa
Precisiona
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
A reliable and validated GC–MS method for the determination of traces of dimethyl sulfate in DMS methylated technical or synthetic materials is presented. It is based upon a simple dissolve-andinject approach and enables the quantification of DMS even in the presence of the MMS which is generated during the methylation reaction with DMS. Thermal formation of DMS from MMS in the hot GC injector is avoided by initial reduction of the injector temperature to 100 ◦ C enabling detection of actual DMS levels down to a sub ppm level. Apart from the MMS issue, this method is also comparable or superior in terms of sensitivity, accuracy and precision to previously published methods [14–19] (Table 1) and avoids any time consuming sample preparation steps. The use of the stable isotope labelled internal standard (DMS-d6) is regarded as an important measure in order to ensure maximum reliability and robustness for this quantitative GC–MS method. As DMS-d6 is commercially available at low costs, the new method can be easily implemented. Finally, the method can be conveniently adapted to other sample matrices, although method selectivity, LOD and LOQ would need to be verified beforehand. References [1] D.A. Pierson, B.A. Olsen, D.K. Robbins, K.M.D. DeVries, I. Varie, Org. Process Res. Dev. 13 (2009) 285.
144
C. Schäfer, P. Zöllner / J. Chromatogr. A 1289 (2013) 139–144
[2] Committee for Medicinal Products for Human Use (CHMP), Europen Medicines Agenca (EMEA), Guidline on the Limits of Genotoxic Impurities, London, UK (EMEA/CHMP/QWP/251344/2006), June 2006. [3] U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Guidance for Industry – Genotoxin and Carcinogenic Impurities in Drug Substances and Products: Recommended Approaches, Silver Spring, MD, USA, 2008, December. [4] D.Q. Liu, M. Sun, A.S. Kord, J. Pharm, Biomed. Anal. 51 (2010) 999. [5] Manfred Müller, Ute Hübsch: dimethyl ether, in: Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2006. [6] Dimethyl sulfate, IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, World Health Organization, International Agency for Research on Cancer, 71, 1999, 575. [7] E. Scobbie, J.A. Groves, Ann. Occup. Hyg. 42 (1998) 201. [8] T. Dublin, H.J. Thöne, J. Chromatogr. 456 (1986) 233. [9] K.S. Sidhu, J. Chromatogr. 206 (1981) 381. [10] L.D. Hansen, V.F. White, D.J. Eatough, Environ. Sci. Technol. 20 (1986) 572. [11] H. Frind, K. Trageser, Z. Fresenius, Anal. Chem. 326 (1987) 517.
[12] S. Fukui, M.M. Morishima, S. Ogawa, Y. Hanazaki, J. Chromatogr. 541 (1991) 459. [13] J.V. Das, K.N. Ramachandran, V.K. Gupta, Micorchem. J. 50 (1994) 51. [14] N.V.V.S.S. Raman, K.R. Reddy, A.V.S.S. Prasad, K. Ramakrishna, Chromatographia 68 (2008) 857. [15] J. Zheng, W.A. Pritts, S. Zhang, S. Wittenberger, J. Pharm. Biomed. Anal. 50 (2009) 1054. [16] R. Alzaga, R.W. Ryan, K. Taylor-Worth, A.M. Lipczynski, R. Szucs, P. Sandra, J. Pharm. Biomed. Anal. 45 (2007) 472. [17] J. An, M. Sun, L. Bai, T. Chen, D.Q. Liu, A. Kord, J. Pharm. Biomed. Anal. 48 (2008) 1006. [18] J.G. Hoogerheide, R.A. Scott, Talanta 65 (2005) 453. [19] N. Grinberg, F. Albu, K. Fandrick, E. Iorgulescu, A. Medvedovici, J. Pharm. Biomed. Anal. 75 (2013) 1. [20] D.P. Elder, A. Teasdale, A.M. Lipczynski, J. Pharm. Biomed. Anal. 46 (2008) 1. [21] Ch. Simmond, Alcohol, Its Production, Properties, Chemistry, and Industrial Applications, Cambridge Scholar Publishing, Cambrigde, 2009.