This action might not be possible to undo. Are you sure you want to continue?
Analysis of class 1 residual solvents in pharmaceuticals using headspace-programmed temperature vaporization-fast gas chromatography–mass spectrometry
a a Jos´ Luis P´ rez Pav´ n ∗ , Miguel del Nogal S´ nchez, Ma Esther Fern´ ndez Laespada, e e o Carmelo Garc´a Pinto, Bernardo Moreno Cordero ı
Departamento de Qu´mica Anal´tica, Nutrici´ n y Bromatolog´a, Facultad de Ciencias Qu´micas, Universidad de Salamanca, 37008 Salamanca, Spain ı ı o ı ı Received 17 July 2006; received in revised form 28 November 2006; accepted 1 December 2006 Available online 19 December 2006
Abstract A sensitive method is presented for the fast screening and determination of residual class 1 solvents (1,1-dichloroethene, 1,2-dichloroethane, 1,1,1trichloroethane, carbon tetrachloride and benzene) in pharmaceutical products. The applicability of a headspace (HS) autosampler in combination with GC equipped with a programmed temperature vaporizer (PTV) and a MS detector is explored. Different injection techniques were compared. The beneﬁts of using solvent vent injection instead of split or splitless-hot injection for the measurement of volatile compounds are shown: better peak shapes, better signal-to-noise ratios, and hence better detection limits. The proposed method is extremely sensitive. The limits of detection ranged from 4.9 ppt (benzene) to 7.9 ppt (1,2-dichloroethane) and precision (measured as the relative standard deviation) was equal to or lower than 12% in all cases. The method was applied to the determination of residual solvents in nine different pharmaceutical products. The analytical performance of the method shows that it is appropriate for the determination of residual class 1 solvents and has much lower detection limits than the concentration limits proposed by the International Conference on Harmonization (ICH) of Technical Requirements for the Registration of Pharmaceuticals for Human Use. The proposed method achieves a clear improvement in sensitivity with respect to conventional headspace methods due to the use of the PTV. © 2006 Elsevier B.V. All rights reserved.
Keywords: Headspace analysis; Programmed temperature vaporizers; Residual solvents; Contour plots
1. Introduction The volatile organic solvents used in the manufacture of pharmaceutical formulations need to be removed from the ﬁnished product because of their possible health risks and toxicity. The acceptable maximum levels of residual solvents that can be left behind according to worldwide regulatory standards [1–3] were originally derived for patient safety considerations. Guideline Q3C  was adopted by the International Conference on Harmonisation (ICH) of Technical Requirements for Registration of Pharmaceuticals for Human Use in 1997. It classiﬁed residual solvents into three categories according to their toxic potential. Class 1 includes the solvents considered to be the most toxic, such that their use should be avoided
Corresponding author. Tel.: +34 923 29 4483; fax: +34 923 29 4483. E-mail address: email@example.com (J.L. P´ rez Pav´ n). e o
in the production of pharmaceutical products. These chemicals are: benzene, carbon tetrachloride, 1,2-dichloroethane, 1,1-dichloroethene and 1,1,1-trichloroethane (the latter owing to its environmental impact). Class 2 and class 3 residual solvents are considered a lesser risk. The most frequently used technique for the determination of residual solvents in pharmaceutical quality assurance/quality control (QA/QC) is static headspace-gas chromatography [4–11]. Even though the sensitivity provided by headspace methods is in many cases sufﬁcient to achieve the maximum permitted levels, it can be enhanced through the use of dynamic headspace, i.e. purge and trap (P&T) [11,12], although more complex instrumentation is required. Headspace-solid phase microextraction (HS-SPME) [12,13] has also been proposed. Another alternative to explore in order to improve sensitivity, maintaining the simple headspace instrumentation, is the use of a programmed temperature vaporizer inlet (PTV) to inject the samples into the chromatographic column [14–16]. The
0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.12.046
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o
conditions are chosen such that the components are retained in the liner by cold trapping, while the solvent is eliminated through the split line. The PTV injector is equipped with an efﬁcient heating and cooling system, in which cooling is carried out by means of air, liquid nitrogen, carbon dioxide, or by electrical systems. By using liners  packed with a selective adsorption material, such as Tenax TA, the range of components that can be trapped in the liner can be signiﬁcantly extended towards the more volatile components. Tenax TA is a porous polymer designed to trap organics without retaining water. In a previous work , we reported the use of headspace generation and fast gas chromatography–mass spectrometry to generate an initial low-resolution chromatogram and contour plots with time and mass/charge ratio axes that would allow the rapid identiﬁcation of residual solvents present in drugs. In the present work, we propose the use of this strategy for the identiﬁcation of toxic ICH class 1 solvents and their determination in pharmaceutical products with a highly sensitive method, using a programmed temperature vaporizer inlet followed by fast capillary gas chromatography coupled to mass spectrometry in the selected ion-monitoring mode acquisition (PTV-fastGC/MS(SIM)). 2. Experimental 2.1. Chemicals The solvents used were purchased from the following sources: methanol was from Merck (Darmstadt, Germany); 1,1-dichloroethene from Supelco (Bellefonte, USA); 1,2dichloroethane and 1,1,1-trichloroethane from Sigma–Aldrich (Sleinheim, Germany); carbon tetrachloride from Panreac (Barcelona, Spain) and benzene from Acros Organics (Geel, Belgium). 2.2. Standard solutions and samples In order to optimize the chromatographic parameters for the identiﬁcation of the solvents, solutions of the ﬁve analytes indicated in Section 2.1 were prepared in ultra-pure water. Likewise, aqueous solutions of these compounds were employed to obtain the calibration curves and detection and quantiﬁcation limits. To perform the measurements, 5 mL of sample was placed in 10 mL vials sealed with silicone septum caps. Each sample was analyzed in triplicate. One liquid (L1) and eight solid (S1–S8) pharmaceutical formulations were analyzed. The liquid sample was prepared by adding 1.5 mL of the formulation plus 3.5 mL of ultra-pure water to the vials. In the case of the solid samples, 5 mL of ultra-pure water was added directly to the samples. Each sample, after homogenization, was analyzed in triplicate. Once the residual solvents had been identiﬁed, they were determined with the standard additions method, using 3.5 mL (liquid samples) or 5 mL (solid samples) of aqueous solutions of the residual solvents corresponding to the different drugs. All samples were prepared in a cold chamber at 5 ◦ C.
2.3. HS-PTV-fast GC–MS measurements 2.3.1. Headspace sampling HS sampling was performed with a model 7694 headspace sampler from Agilent Technologies. This sampler is equipped with a tray for 44 consecutive samples and an oven with positions for 6 sample vials. The oven temperature was kept at 90 ◦ C for 30 min. The sampling system consisted of a stainless steel needle, a 316-SS six-port valve with a 3 mL nickel loop (heated to 100 ◦ C), and two solenoid valves (for pressurization and venting). The headspace sampler was coupled to a PTV injector through a thermostatted transfer line heated to 100 ◦ C. The carrier gas was helium N50 (99.995% pure; Air Liquide). 2.3.2. Programmed-temperature vaporization Different injection techniques were compared (classical splithot injection, classical splitless-hot injection and solvent vent injection). All experiments were carried out in a PTV inlet (CIS4; Gerstel, Baltimore, MD, USA). The CIS-4 inlet was equipped with a 71 mm × 2.0 mm Tenax-TA packed liner. Cooling was carried out by means of CO2 . 184.108.40.206. Split/splitless-hot injection. The injector temperature was kept at 250 ◦ C throughout the analysis time. The split ratio was 1:10. The splitless time was 2.25 min. 220.127.116.11. Solvent vent injection. The injector starting temperature was 5 ◦ C. The vent ﬂow was adjusted to 50.0 mL/min and the vent pressure to 5.00 psi. After 1.70 min, the split valve was closed and the liner was ﬂash-heated at 12 ◦ C/s to 250 ◦ C. The analytes were transferred from the liner to the capillary column (0.60 min). The split valve was then opened and the liner temperature was held at 250 ◦ C for 8.00 min. The experimental conditions are shown schematically in Fig. 1. 2.3.3. Gas chromatography To perform the gas chromatographic measurements, an Agilent 6890 GC device equipped with a DB-VRX capillary column (20 m × 0.18 mm × 1 m) was used. The column oven temperature program involved an initial temperature of 35 ◦ C for 4.50 min; this was increased at a rate of 20 ◦ C/min to 70 ◦ C, then increased at 70 ◦ C/min to 175 ◦ C, and then further increased at 45 ◦ C/min to 240 ◦ C and held for 1.0 min. The latter two temperature ramps are the maximum ones permitted by the instrumental conﬁguration employed. The total chromatographic run time was 10.19 min. The experimental conditions are schematized in Fig. 1. 2.3.4. Mass spectrometry The detector was a quadrupole mass spectrometer (HP 5973 N). It was operated in the electron impact mode using 70 eV ionization voltage. The ion source temperature was 230 ◦ C and the quadrupole was set to 150 ◦ C. The analyses were performed in the scan and SIM modes. 18.104.22.168. Scan detection mode. The m/z range was 25–125 amu, and the abundance threshold value was set to 0. The different
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o
Table 1 Maximum permitted concentrations, boiling points, retention times and m/z ratios selected for the 5 solvents studied Compound 1,1-Dichloroethene 1,2-Dichloroethane 1,1,1-Trichloroethane Carbon tetrachloride Benzene ICH Limit (ppm) 8 5 1500 4 2 Boiling point (◦ C) 32 83 72 77 80 tR (min) 3.69 5.69 5.73 6.00 6.06 m/z 61, 96 27, 62 97, 99 117, 121 51, 78
Fig. 1. Sequence of events for solvent vent injection.
compounds were identiﬁed by comparison of the experimental spectra with those of the NIST 98 database (NIST/EPA/NIH Mass Spectral Library, version 1.6). 22.214.171.124. SIM detection mode. Three SIM groups were used. The ﬁrst (3.25–5.40 min) contained the two most abundant ions of 1,1-dichloroethene (61, 96). The second (5.40–5.84 min) was formed by two characteristic ions of 1,2-dichloroethane (27, 62) and two of 1,1,1-trichloroethane (97, 99). The last group (5.84–10.19 min) contained four m/z variables, two characteristic ions of carbon tetrachloride (117, 121) and two of benzene (51, 78). The ions were acquired with a dwell time of 100 ms for group 1 and 50 ms for groups 2 and 3, respectively. 2.4. Data analysis Data collection was performed with Enhanced ChemStation, G1701CA Ver. C 00.00 software  from Agilent Technologies. This software was used to calculate S/N ratios. The plotting of the information contained in the chromatograms with contour plots was accomplished using Matlab v 6.5 . 3. Results and discussion 3.1. Preliminary study of HS-PTV-fast GC–MS data 3.1.1. Comparison of the different injection modes In the present work, we studied all the class 1 solvents included in the ICH . The boiling points and maximum permitted limits  for each are shown in Table 1. Initially, we carried out a comparative study of the signal obtained when using different injection techniques: split-hot injection, splitless-hot injection and solvent vent injection. To
do so, we employed a solution of the ﬁve compounds studied in ultra-pure water at concentrations of 24.2, 12.0, 12.0, 12.0 and 3.12 ppb for 1,1-dichloroethene, 1,2-dichloroethane, 1,1,1trichloroethane, carbon tetrachloride and benzene, respectively. The sequence of steps involved when solvent vent injection was used is shown in Fig. 1. In the ﬁrst step, the sample coming from the headspace reaches 100 ◦ C through the transfer line and is injected into the PTV injector, which is at 5 ◦ C, such that the analytes are retained in the liner. The split valve is opened, allowing solvent elimination. The boiling point of the analytes should be higher than that of the solvent. However, even when dealing with compounds with low boiling points, as is the case of the analytes studied here, they can be trapped using liners packed with adequate adsorbents, such as Tenax TA. In the second step, involving transfer of the sample to the column (PTV transfer time), the split valve is closed and the PTV is rapidly heated (12 ◦ C/s) until it reaches 250 ◦ C, the same temperature used in the split and splitless injections. Thus, the analytes are desorbed and transferred onto the column. Finally, the split valve is opened again, and chromatographic separation is accomplished, with the temperature program also shown in Fig. 1. Fig. 2 shows the window of the chromatograms (scan mode) of the most abundant extracted ion for three of the residual solvents with the three injection modes described above: the one with the shortest retention time (1,1-dichloroethene, Fig. 2a), an intermediate one (1,1,1-trichloroethane, Fig. 2b) and the one with the longest retention time (benzene, Fig. 2c). In all three cases, the area under the curve was greater when solvent vent injection was used. The analytical signal increased 6.1, 6.3 and 7.5 times, respectively, with respect to that obtained with splithot injection. On comparing solvent vent injection with splitless injection, the most important observation was an improvement in peak shapes, which were much narrower owing to the fact that the splitless injection of a large volume of headspace is too slow to give sharp peaks (splitless time: 2.25 min). However, in solvent vent injection, there is previously a preconcentration in the liner at low temperature and transfer to the column takes place in a much shorter time (0.6 min). The S/N ratios in the case of the ﬁrst compound (Fig. 2a) were 267, 98 and 61 when solvent vent injection, splitless-hot injection and split-hot injection was used, respectively. In the case of 1,1,1-trichloroethane (Fig. 2b), these ratios were 514, 259 and 114. The last compound, (Fig. 2c) had S/N ratios of 504, 124 and 47 when the above injection modes were employed.
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o
roethane (Fig. 3a) and 44-fold higher for carbon tetrachloride (Fig. 3b) when the SIM mode was used in data acquisition as compared with the scan mode. Similar results were obtained for the other analytes. Accordingly, it was decided to perform the identiﬁcation and determination of the ﬁve residual solvents in SIM mode. 3.1.3. Generation of contour plots for rapid identiﬁcation of the analytes Fig. 4 shows the chromatogram obtained on analyzing the solution containing all ﬁve solvents. The retention times are shown in Table 1. 1,1-Dichloroethene appears completely separated from the others and the other four are seen in two partially overlapping groups (b and c). The partial overlapping of some of the compounds seen in Fig. 4a does not prevent their individual chromatographic determination. The determination of all the solvents is possible using the chromatograms extracted from speciﬁc ions, as may be seen for the pairs of compounds 1,2-dichloroethane/1,1,1trichloroethane (Fig. 4b) or carbon tetrachloride/benzene (Fig. 4c). The information contained in these chromatograms can be visualized by using contour plots with time and mass/charge ratio axes. Fig. 5a corresponds to a plot of the information contained in the chromatogram shown in Fig. 4a. Thus, it is possible to visualize all the information in a single ﬁgure: signal intensity, retention time, and the mass spectrum. Fig. 5a shows the plot corresponding to a data matrix where each row represents the analysis time and each column contains the intensities for the m/z ratios.
Fig. 2. Extracted ion chromatograms for m/z 61 (a), 97 (b) and 78 (c) when the three different injection modes studied were used: classical split-hot injection, classical splitless-hot injection and solvent vent injection. A solution of the ﬁve compounds studied in ultra-pure water at concentrations of 24.2, 12.0, 12.0, 12.0 and 3.1 ppb for 1,1-dichoroethene, 1,2-dichloroethane, 1,1,1-trichloroethane, carbon tetrachloride and benzene, respectively was used.
The improved peak shape and higher peak response of measurements with the solvent vent injection mode resulted in higher signal-to-noise ratios and hence lower detection limits. The shift in the retention times for each compound when all three injection modes were used is mainly due to the short delay in liner heating (12 ◦ C/s) on using the solvent vent-based system. In the other two cases, the temperature of the injection port was maintained constant at 250 ◦ C. In light of the results obtained, we decided to use the solvent vent injection technique. 3.1.2. Comparison of the scan and SIM data acquisition modes Fig. 3 shows the window of the chromatograms obtained for the previous sample when the m/z 62 (Fig. 3a) and m/z 117 (Fig. 3b) ratios characteristic of 1,2-dichloroethane and carbon tetrachloride, respectively, were used. In both cases, the analytical signal obtained in SIM mode and in scan mode was compared. The S/N ratio was 35-fold higher in the case 1,2-dichlo-
Fig. 3. Extracted ion chromatograms for m/z 62 (a) and 117 (b) when SIM and scan mode were used. A solution of the ﬁve compounds studied in ultra-pure water at concentrations of 24.2, 12.0, 12.0, 12.0 and 3.1 ppb for 1,1dichoroethene, 1,2-dichloroethane, 1,1,1-trichloroethane, carbon tetrachloride and benzene, respectively was used.
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o
Analytical Procedures: Methodology) : DL = 3.3σ S
where σ is the standard deviation of peak response corresponding to an S/N ratio of approximately 3, and S is the slope of the calibration curve. DLs as low as 4.9 ppt for benzene and 5.1 ppt for carbon tetrachloride were obtained. The DLs for the other analytes were similar.
Fig. 4. (a) Total ion current chromatogram for a laboratory-prepared solution containing the ﬁve solvents studied in ultra-pure water. (b) Extracted ion chromatograms corresponding to m/z 62 and 97. (c) Extracted ion chromatograms corresponding to m/z 117 and 78. A solution of the ﬁve compounds studied in ultra-pure water at concentrations of 24.2, 12.0, 12.0, 12.0 and 3.1 ppb for 1,1dichoroethene, 1,2-dichloroethane, 1,1,1-trichloroethane, carbon tetrachloride and benzene, respectively was used.
In this plot, it is also possible to establish speciﬁc zones for each of the solvents studied. In Fig. 5b, the zones used to identify the solvents are marked with black symbols. Finally, it is possible to compile a template that has the columns tR (column 4) and m/z (column 5) from Table 1 as coordinates. Use of this template allows rapid identiﬁcation of the residual solvents present in the samples. Fig. 5c shows the template generated. 3.2. HS-PTV-fast GC–MS method evaluation The linearity of the method was evaluated from triplicate injections of a series of standard solutions prepared for each analyte over the concentration range studied. The variables used in the calibrations were the area under the curve of 1,1-dichloroethene, 1,2-dichloroethane, 1,1,1-trichloroethane, carbon tetrachloride, and benzene in the extracted ion chromatogram for m/z 61, 62, 97, 117 and 78, respectively. All the calibrations showed good linear behavior, with coefﬁcient of determination (R2 ) values above 0.99. The intercept included zero in all cases. For the set of ﬁve solvents, repeatability (n = 10) was evaluated at a level corresponding to an S/N ratio of approximately 3. The repeatability was satisfactory, with an RSD equal to or less than 12%. In many cases, the RSD was lower than 10%. The detection limits (DLs) were estimated using the following equation based on ICH guideline Q2B (Validation of
Fig. 5. (a) Contour plots with time and mass/charge ratio axes for a laboratoryprepared solution containing the ﬁve solvents studied in ultra-pure water. (b) Zones selected for each solvent studied. (c) Template generated for the identiﬁcation of solvents in pharmaceutical products.
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o
The quantitation limits (QLs) were estimated using the following equation based on ICH guideline Q2B (Validation of Analytical Procedures: Methodology) : QL = 10σ S
where σ is the standard deviation of peak response corresponding to an S/N ratio of approximately 3, and S is the slope of the calibration curve. For all the solvents, the QLs are well below the ICH limits. The results for linearity, precision, and detection and quantitation limits are summarized in Table 2. When compared with related methods, the results obtained in this study are highly satisfactory, particularly in terms of simpliﬁcation of the sample preparation step. According to the literature, the use of HS-GC-FID  for determination of benzene provided a detection limit of 100 ppb, with 2.7% of precision (RSD %). HS-GC–MS, without a programmed temperature vaporizer  achieved limits of detection equal to 8 ppb and 42 ppb for benzene and carbon tetrachloride, respectively, with RSD values of 7.2% and 13%. If preconcentration steps are added to the method, better detection limits are accordingly achieved, but they are in the same order of magnitude that the ones achieved in the present work. Thus, solid phase microextraction coupled to GC–MS  provided limits of detection of 10 ppt for 1,2-dichloroethane and benzene, with RSD values of 2% (at a 100 ppb concentration level). 3.3. Analysis of pharmaceutical products The detection of residual solvents by using low-resolution chromatograms and visual inspection of the contour plots was carried out in nine drugs. In six of them (S3–S8), none of the compounds studied was detected. The characteristic zones of these compounds in the template used did not show any signal. In the liquid formulation (L1), from the contour plot of the chromatogram only one residual solvent was detected: benzene. Fig. 6a shows the signal obtained on analyzing this drug, on which the previously generated template had been superimposed. Visualization of the template favors the rapid detection of any of the solvents studied. Only the zone of benzene showed a signal. Even though it is also possible to observe a signal in the zone of 1,2-dichloroethane, this corresponds to only one of its characteristics m/z ratios, such that it cannot be attributed to that analyte.
Table 2 Figures of merit for the HS-PTV-fast GC–MS method Solvent Linearity Concentrations (ppb) 1,1-Dichloroethene 1,2-Dichloroethane 1,1,1-Trichloroethane Carbon tetrachloride Benzene 0–24.2–48.4–72.6–96.8–121 0–12.0–24.0–36.0–48.0–60.0 0–12.0–24.0–36.0–48.0–60.0 0–12.0–24.0–36.0–48.0–60.0 0–3.12–6.24–9.36–12.5–15.6 R2 0.9914 0.9913 0.9926 0.9909 0.9908 12 10 7.7 4.5 5.7 6.1 7.9 6.3 5.1 4.9 19 24 19 15 15 Precision RSD (%) Detection limit (ppt) Quantitation limit (ppt)
Fig. 6. Superimposition of the template on the contour plots for the L1 (a) liquid formulation and S1 (b) and S2 (c) solid pharmaceuticals.
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o Table 3 Analytes found in the pharmaceutical products containing residual solvents Pharmaceuticals Residual solvents (ppb) 1,1-Dichloroethene L1 S1 S2 ND: not detected. ND ND ND 1,2-Dichloroethane ND 1.1 ± 0.1 1.0 ± 0.3 1,1,1-Trichloroethane ND ND ND Carbon tetrachloride ND ND ND
Benzene 4.0 ± 0.3 ND ND
The benzene content was determined with HS-PTV-fast GC–MS. To overcome matrix effects, standard additions quantitation was used. The benzene contents and the conﬁdence interval for 95% probability was 4.0 ± 0.3 ppb (Table 3). This concentration is well below the maximum permitted for benzene (2 ppm). In S1 and S2, only 1,2-dichloroethane was found. The corresponding contour plots are shown in Fig. 6b and c, respectively. The 1,2-dichloroethane contents and the conﬁdence interval for 95% probability were 1.1 ± 0.1 ppb and 1.0 ± 0.3 ppb for S1 and S2, respectively (Table 3). Both pharmaceutical products had 1,2-dichloroethane concentrations much lower than the maximum permitted (5 ppm). To check the possibilities of the methodology for those compounds that were not detected, two of the formulations studied (one liquid and the other solid) were spiked with them. In L1, 1,1dichloroethene, 1,2-dichoroethane, 1,1,1-trichloroethane and carbon tetrachloride were added. S2 was spiked with 1,1dichloroethene, 1,1,1-trichloroethane carbon tetrachloride and benzene. Again, standard additions quantitation was used. The predicted versus added concentrations are shown in Table 4. These results conﬁrm the applicability of the proposed methodology for the quantiﬁcation of all ﬁve class 1 residual solvents. 3.4. Time of analysis The volatile generation time used in the present work was 30 min. However, since the instrumental conﬁguration employed had an oven with six positions for heating samples simultaneously, this headspace generation could be overlapped and
the injection interval between samples could be considerably reduced (5 min). The methodology required about 10 min for the temperature program to be completed and to ensure complete elution of the compounds present in the sample injected. Additionally, about 10 min were necessary to measure the next sample since the column had to be cooled down from the ﬁnal temperature attained (240 ◦ C) to initial conditions of 35 ◦ C. Considering the time invested in re-establishing the initial conditions, the analysis time per sample (after the ﬁrst 30 min) was therefore in the region of 20 min. After the ﬁrst 30 min, it is possible to analyse three samples per hour. 4. Conclusions The proposed methodology has been successfully applied in different types of pharmaceutical products. The advantage of this approach over others described previously  is that in the present case, a PTV inlet was used. Important beneﬁts of using solvent vent injection instead of classical-hot injection, such as better peak shapes and better signal-to-noise ratios, were found. Using solvent vent injection, it was possible to carry out screening and quantitative analysis for residual class 1 solvents at the low ppt level. It should be emphasized that the method showed good precision and accuracy. The use of headspace generation for introducing the sample and standard additions procedure as a quantiﬁcation technique provided satisfactory results and minimized the matrix effect. An important advantage of the methodology used here is that no prior treatment of the sample is required, thus minimizing the creation of analytical artifacts and the errors associated with this step of the analytical process. The importance of using automatic conﬁgurations for cryogenic focusing coupled to static headspace sampling has been recently pointed out by Kolb and Ettre in their monography on headspace-gas chromatography . Only a few reports on this coupling have been published, mainly as application notes, by instrumentation companies . Acknowledgments We acknowledge the ﬁnancial support of the DGI (Project CTQ2004-01379/BQU) and the Consejer´a de Educaci´ n y Culı o tura of the Junta de Castilla y Le´ n (Project SA057A05) for this o research.
Table 4 Predicted concentrations and conﬁdence intervals (95% probability) for the solvents added in the different pharmaceutical products Pharmaceutical products Compound Added concentration (ppb) 4.1 4.1 4.1 4.1 4.1 2.6 1.0 1.0 Predicted concentration (ppb) 5 5 5 4 4 2.5 1.1 1.1 ± ± ± ± ± ± ± ± 1 1 1 1 1 0.8 0.3 0.3
1,1-Dichloroethene 1,2-Dichloroethane 1,1,1-Trichloroethane Carbon tetrachloride 1,1-Dichloroethene 1,1,1-Trichloroethane Carbon tetrachloride Benzene
J.L. P´ rez Pav´ n et al. / J. Chromatogr. A 1141 (2007) 123–130 e o  C. Witschi, E. Doelker, Eur. J. Pharm. Biopharm. 43 (1997) 215.  C.B. Hymer, Pharm. Res. 20 (2003) 337.  C.C. Camarasu, M. Mezei-Sz¨ ts, G. Bert´ k Varga, J. Pharm. Biomed. 18 u o (1998) 623.  W. Engewald, J. Teske, J. Efer, J. Chromatogr. A 842 (1999) 143.  J. Janssen, Sample Introduction Techniques for Capillary Gas Chromatography, 2001. Gerstel, M¨ lheim/Ruhr. u  M.V. Russo, Chromatographia 39 (1994) 645.  B. Kolahgar, E. Pfannkock, Tecnical Note 36, Gerstel, M¨ lheim/Ruhr. u  J.L. P´ rez Pav´ n, M. del Nogal S´ nchez, C. Garc´a Pinto, M.E. Fern´ ndez e o a ı a Laespada, B. Moreno Cordero, Anal. Chem. 78 (2006) 4901.  Enhanced ChemStation, G1701CA, Version C00.00, Agilent Technologies, 1999.  The Mathworks Inc., MATLAB® , version 6.5, Natick, MA, USA, 2000.  International Conference on Harmonization (ICH) of Technical Requirements for the Registration of Pharmaceuticals for Human Use, Q2B: Validation of Analytical Procedures: Methodology, Step 4, November 1996.  B. Kolb, L.S. Ettre, Static Headspace-Gas Chromatography, Wiley Interscience, Hoboken, New Jersey, 2006.
 International Conference on Harmonization (ICH) of Technical Requirements for the Registration of Pharmaceuticals for Human Use, Q3C: Impurities: Guideline for Residual Solvents, Step 4, July 1997.  European Pharmacopoeia 4th Edition, Identiﬁcation and Control of Residual Solvents, Council of Europe, Strasbourg, France 2002.  United States Pharmacopeia method 467. Organic Volatile Impurities. 2004.  T.K. Chen, J.G. Phillips, W. Durr, J. Chromatogr. A 811 (1998) 145.  P. Sandra, F. David, R. Sz¨ cs, Trends Anal. Chem. 21 (2002) 662. u  K. Urakami, A. Higashi, K. Umemoto, M. Godo, J. Chromatogr. A 1057 (2004) 203.  R. Otero, G. Carrera, J.F. Dulsat, J.L. F´ bregas, J.J. Claramunt, J. Chroa matogr. A 1057 (2004) 193.  K. Fliszar, J.M. Wiggins, C.M. Pignoli, G.P. Martin, Z. Li, J. Chromatogr. A 1027 (2004) 83.  S. Klick, A. Sk¨ ld, J. Pharm. Biomed. Anal. 36 (2004) 401. o  M.J. Rocheleau, M. Titley, J. Bolduc, J. Chromatogr. B 805 (2004) 77.