Direct Coupling of Ionic Liquid Based Single-Drop Microextraction and

Dec 22, 2007 - Direct Coupling of Ionic Liquid Based Single-Drop Microextraction and GC/MS ... microextraction (SDME) and its direct coupling to gas c...
0 downloads 0 Views 259KB Size
Download PDF
Anal. Chem. 2008, 80, 793-800

Direct Coupling of Ionic Liquid Based Single-Drop Microextraction and GC/MS Eva Aguilera-Herrador, Rafael Lucena, Soledad Ca´rdenas, and Miguel Valca´rcel*

Department of Analytical Chemistry, Marie Curie Building (Annex), Campus de Rabanales, University of Cordoba, E-14071 Cordoba, Spain

The use of ionic liquids as extracting media in single-drop liquid-phase microextraction (SDME) and its direct coupling to gas chromatography/mass spectrometry (GC/MS) is presented. For this purpose, a new removable interface that enables the introduction of the extracted analytes into the GC system, while preventing the ionic liquid from entering the column, has been developed. The determination of three representative pollutants in water samples has been used as a model analytical problem in order to demonstrate the feasibility of the proposed interface. The analytes (dichloromethane, p-xylene, and n-undecane) were coextracted from the aqueous sample in a 2-µL drop of 1-butyl-3-methylimidazolium hexaflourophosphate. Then, the syringe used to perform the SDME was directly introduced into the interface, which was held at 140 °C in order to achieve a complete volatilization of the target compounds. After the injection, the ionic liquid was retained in the interface, while a carrier gas transferred the volatilized analytes into the GC inlet. The optimization of the operational variables affecting the new interface (temperature, carrier flow rate, sample volume and injection technique) was accomplished. The analytes could be determined with detection limits in the low-nanogram per milliliter concentration range, and the relative standard deviations were between 3.3 and 4.4%. Direct analysis of samples, regardless of their origin, is desirable, although for the majority of cases unfeasible on account of the complexity of the sample matrix, inadequate concentration of the target analytes, or even incompatibility with the detector. Sample treatment techniques are considered a crucial step as they are commonly employed in any analytical methodology for interference removal, analyte separation/preconcentration, or fractionating them into groups.1 Liquid-liquid extraction can be considered a classical technique for sample treatment that has been widely used in different application fields. Its main drawback is the requirement of large amounts of high-purity solvents that are expensive and toxic and result in the production of hazardous waste, the disposal of which is problematic.2 Moreover, the technique is time-consuming, * To whom correspondence should be addressed. Tel/Fax: +34-957-218-616. E-mail: [email protected]. (1) Lucena, R.; Ca´rdenas, S.; Gallego, M.; Valca´rcel, M. Anal. Chim. Acta 2004, 509, 47-54. (2) Psillakis, E.; Kalogerakis, N. Trends Anal. Chem. 2002, 21, 53-63. 10.1021/ac071555g CCC: $40.75 Published on Web 12/22/2007

© 2008 American Chemical Society

tedious, and laborious (affecting directly the precision level). Some efforts have been made in order to overcome these problems, liquid-phase microextraction (LPME) being one of the most attractive alternatives.3 Single-drop microextraction (SDME) is an approach evolved from LPME in which the extraction phase is a drop of solvent, usually suspended in the needle of a syringe, direct immersed in a stirred aqueous solution (DI-SDME) or in close contact with its headspace (HS-SDME).4 The coupling between SDME and gas chromatography (GC) has been proved to be a useful technique for the determination of different potential contaminants in samples of environmental concern. In this way, it was successfully applied for the determination of methyl tert-butyl ether,5 chlorobenzenes,6 phenols,7 organophosphorus pesticides,8 amines,9 and halocarbons.10 The main shortcoming of the SDME process is the instability of the drop when an organic solvent is used as extractant. This fact limits the usable volume of the extracting medium, affecting directly the precision and also the sensitivity of the determinations. This limitation is more marked when high-temperature HS-SDME is performed due to the evaporation of the organic solvent during the extraction. Room-temperature ionic liquids (RTIL), which are ionic media resulting from the combination of organic cations and various anions,11 have been proposed as an alternative to these organic solvents due to their low vapor pressure and their high viscosity, which allows the use of larger and more reproducible extracting volumes.12-16 These solvents boast other unique properties, (3) Psillakis, E.; Kalogerakis, N. Trends Anal. Chem. 2003, 22, 565-574. (4) Xu, L.; Basheer, C.; Lee, H. K. J. Chromatogr., A 2007, 1152, 184-192. (5) Yazdi, A. S.; Assadi, H. Chromatographia 2004, 60, 699-702. (6) Vidal, L.; Canals, A.; Kalogerakis, N.; Psillakis, E. J. Chromatogr., A 2005, 1089, 25-30. (7) Saraji, M.; Bakhshi, M. J. Chromatogr., A 2005, 1098, 30-36. (8) Ahmadi, F.; Assadi, Y.; Milani Hosseini, S. M. R.; Rezaee, M. J. Chromatogr., A 2006, 1101, 307-312. (9) Deng, C.; Li, N.; Wang, L.; Zhang, X. J. Chromatogr., A 2006, 1131, 4550. (10) Zhang, T.; Chen, M.; Li, Y.; Liang, P. Chromatographia 2006, 63, 633637. (11) Liu, J. F.; Jonsson, J. A.; Jiang, G. B. Trends Anal. Chem. 2005, 24, 20-27. (12) Liu, J.; Jiang, G. B.; Chi, Y. G.; Cai, Y. Q.; Zhou, Q. X.; Hu, J. T. Anal. Chem. 2003, 75, 5870-5876. (13) Peng, J. F.; Liu, J. F.; Jiang, G. B.; Tai, C.; Huang, M. J. J. Chromatogr., A 2005, 1072, 3-6. (14) Ye, C. L.; Zhou, Q. X.; Wang, X. M. Anal. Chim. Acta 2006, 572, 165-171. (15) Vidal, L.; Psillakis, E.; Domini, C. E.; Grane, N.; Marken, F.; Canals, A. Anal. Chim. Acta 2007, 584, 189-195. (16) Ye, C.; Zhou, O.; Wang, X.; Xiao, J. J. Sep. Sci. 2007, 30, 42-47.

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008 793

including dual natural polarity, good thermal stability,17 or miscibility with water and organic solvents.18 Additionally, they are regarded as environmentally friendly solvents and are easily synthesized or commercially available. These characteristics have led to an extensive range of applications and investigations in analytical chemistry as recently reviewed.11,19 Some studies about the extractability properties of RTIL have been also reported in the literature,20-22 which supports their consideration as very potential extractants for LPME. However, when RTIL are employed as extractants in SDME, liquid chromatography12-16,23 is preferred to GC as separation technique since this ionic liquid based single-drop microextraction (IL-SDME) is incompatible with GC due to the nonvolatility of the RTIL. Consequently, the utilizations of RTIL in GC are limited to their use as stationary phases,24-26 as coaters of the surface of fibers in solid-phase microextraction,27,28 or as solvents in headspace chromatography.29 Thus, to the best of our knowledge, the direct combination of IL-SDME and GC, which could be a useful tool for different analytical purposes, has not been described before. This paper is focused on the use of ionic liquids as extracting media for the coupling between SDME and GC/MS. For this purpose, a new and removable interface is presented permitting the direct injection of the extraction medium in the chromatograph. The determination of some representative pollutants in water samples has been used as model analytical problem in order to evaluate the potential of the present proposal. EXPERIMENTAL SECTION Reagents. All reagents were of analytical grade or better. Dichloromethane, p-xylene, and n-undecane were purchased from Sigma-Aldrich (Madrid, Spain); trichloromethane, carbon tetrachloride, and methanol (HPLC gradient-grade) were obtained from Panreac (Barcelona, Spain); 1-butyl-3-methylimidazolium hexaflourophosphate ([C4MIM][PF6]) was supplied by Solvent Innovation (Cologne, Germany). Potassium chloride was purchased from Panreac. Stock standard solutions of the analytes (dichloromethane, p-xylene, n-undecane, trichloromethane, carbon tetrachloride) were prepared in methanol at a concentration of 1 g/L and stored in glass-stoppered bottles in the dark at 4 °C. Working solutions were prepared by rigorous dilution of the stocks in the appropriate solvent. Solutions of the analytes were also prepared in [C4MIM](17) Kosmulski, M.; Gustafsson, J.; Rosenholm, J. B. Thermochim. Acta 2004, 412, 47-53. (18) Welton, T. Chem. Rev. 1999, 99, 2071-2083. (19) Anderson, J. L.; Armstrong, D. W.; Wei, G. T. Anal. Chem. 2006, 78, 28922902. (20) Huddleston, J. G.; Willauer, H. D.; Swatloski, R. P.; Visser, A. E.; Rogers, R. D. Chem. Commun. 1998, 16, 1765-1766. (21) Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J. Am. Chem. Soc. 2002, 124, 14247-14254. (22) Liu, J. F.; Chi, Y. G.; Jiang, G. B. J. Sep. Sci. 2005, 28, 87-91. (23) Liu, J. F.; Chi, Y. G; Jiang, G. B.; Tai, C.; Peng, J. F.; Hu, J.-T. J. Chromatogr., A 2004, 1026, 143-147. (24) Armstrong, D. W.; He, L. F.; Liu, Y. S. Anal. Chem. 1999, 71, 3873-3876. (25) Anderson, J. L.; Armstrong, D. W. Anal. Chem. 2003, 75, 4851-4858. (26) Anderson, J. L.; Armstrong, D. W. Anal. Chem. 2005, 77, 6453-6462. (27) Hsieh, Y. N.; Huang, P. C.; Sun, I. W.; Whang, T. J.; Hsu, C. Y.; Huang, H. H.; Kuei, C. H. Anal. Chim. Acta 2006, 557, 321-328. (28) Liu, J. F.; Li, N.; Jiang, G. B.; Liu, J. M.; Jonsson, J. A.; Wen, M. J. J. Chromatogr., A 2005, 1066, 27-32. (29) Andre, M.; Loidl, J.; Laus, G.; Schottenberger, H.; Bentivoglio, G.; Wurst, K.; Ongania, K. H. Anal. Chem. 2005, 77, 702-705.

794

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

[PF6] at a concentration of 20 µg/mL and were conveniently stored. Trichloromethane, dichloromethane, and carbon tetrachloride were used for the optimization of the injection process, while dichloromethane, p-xylene, and n-undecane were selected as the target analytes to be extracted in the SDME process. Apparatus. Experiments were carried out by using a Fisons gas chromatograph 8000 interfaced to a Fisons MD800 mass spectrometer and controlled by a computer running MASSLAB software (Thermo-Quest, Madrid, Spain). Gas chromatographic separations were achieved on a fused-silica capillary column (15 m × 0.25 mm i.d.) coated with 5% phenylmethylpolysiloxane (film thickness 0.25 µm) (Supelco, Madrid, Spain). The injection port, transfer line, and detector temperatures were maintained at 300, 250, and 200 °C, respectively, throughout the experiments. The temperature program of the chromatographic oven began at 38 °C, ramped to 90 °C at 15 °C/min, then increased up to 200 °C at 25 °C/min, and was held for 3 min. The splitless column mode and the septum purge split mode (1:22 ratio) was selected for injection. Helium (6.0 grade, Air Liquide, Seville, Spain), at a flow rate of 1 mL/min, regulated by a digital pressure and flow controller, was used as a carrier gas. Electron impact ionization was used with an ionization energy of 70 eV. The measurements were carried out in the full scan mode, monitoring the ions within 45 and 110 m/z. To perform the IL-SDME, a 10-µL microsyringe (Agilent, Palo Alto, CA) was employed in order to introduce 2 µL of the ionic liquid (acceptor phase) in the extraction vial and to form the microdrop. Then, the same syringe was used to inject the extractant into the interface for analysis. A Velp Cientı´fica stirrer (Milan, Italy) was used for agitation of the samples and was located in an oven to reach the required temperature for the extractions. Description of the Proposed Interface. The new interface was built in order to prevent the ionic liquid from entering the capillary column while effectively transferring the extracted analytes into the gas chromatograph. No modification of the GC system was necessary to afford this coupling. The interface developed, depicted in Figure 1, consists of three main integrated components: an injection zone, a removable unit, and a transfer line. The injection zone was a 1/8 -in. Swagelok nut fitted with a poly(dimethylsiloxane) septum and connected to a stainless steel (SS) union tee through which a carrier gas could enter the interface. The SS union tee was connected downstream to a removable unit that consisted of ∼3 cm of 3-mm-i.d. perfluoroalkoxi (PFA) tubing packed with clean cotton. This tube was provided in its extremes with 1/8 -in. Swagelok connectors so that it could be easily removed for cleanup purposes. A transfer line was used to connect the removable unit to the GC inlet. It was made of a custom SS 1/81/ -in. Swagelok reducing union, connected to a 1/ -in. nut 16 16 equipped with a SS needle of 5 cm in length. The needle was passed through the GC septum and remained there during the experiments. The transfer line was designed and constructed so that it rests at the top of the GC inlet covering the entire surface, which facilitated the diffusion of the heat in the GC injector toward the interface. The interface was connected to the carrier gas line using a SS 1/ -in. connector. The flow could be controlled by means of a 8

Figure 1. Schematic diagram of the interface developed for the direct introduction of ionic liquid containing the extracted analytes. SDME, single drop microextraction; PFA, perfluoroalkoxi; SS, stainless steel.

millimeter valve, whereas a two-way valve was used to stop and open the flow as required. The total flow through the capillary column was the sum of the carrier gas from the interface plus the helium supply line of the GC, which allowed the use of the electronic control pressure system of the GC in order to have constant flow rates through the column. The interface was thermally insulated as depicted in Figure 2 in order to hold the temperature required to carry out the desorption of the analytes. The temperature in the interface could be established using the temperature controller of the GC injector. The effective value of this parameter in the interface, which was proportional to the temperature in the GC injector, could be measured and controlled using a thermometer. It was probed that the insulation of the interface was effective and the temperature did not vary once the equilibration time had been reached. The value of this parameter in the interface was always kept below 200 °C to avoid the decomposition of the ionic liquid. For this purpose, the interface was housed in two 10-cm-i.d. cylindrical porcelain units provided with two metal pieces that helped the transference of the heat to the constructed interface. IL-SDME Procedure. The general IL-SDME procedure was performed as follows: 6 mL of an aqueous solution, spiked at a known concentration with the target analytes (dichloromethane, p-xylene, and n-undecane), were placed in a 10-mL glass vial containing a homemade glass coated stirring bar. Potassium chloride was added at a final concentration of 5 g/L as saltingout additive. The glass vial was tightly capped with a silicone septum, and the GC syringe was filled with 2 µL of the ionic liquid acceptor phase. Then, the needle was inserted in the vial through the septum until its tip was located ∼1 cm above the surface of the stirred solution. The plunger was depressed, and a microdrop of the acceptor phase was exposed on the headspace above the aqueous solution at 70 °C for 40 min.

Figure 2. Scheme of the household unit of the interface for the direct introduction of analytes extracted in a drop of ionic liquid into a gas chromatograph.

After the extraction, the drop was retracted and the syringe used to perform the SDME, filled with the ionic liquid containing the extracted analytes, was directly introduced into the interface through the injection zone. The injection point within the syringe needle was far enough downstream of the carrier gas inlet so that the volatilized analytes could not move upstream, preventing dilution of the analytes and carryover from subsequent injections. The temperature in the interface was kept at 140 °C in order to achieve complete volatilization of the target analytes. After the injection, the ionic liquid was retained in the cotton while the carrier gas, at a flow rate of 25 mL/min, transferred the volatilized analytes into the GC inlet. The removable unit was substituted by a clean one every five injections (i.e., 10 µL of ionic liquid had been injected), requiring 1 h of equilibration time in order to allow the new unit to reach the optimum temperature. Meanwhile, the replaced unit can be easily cleaned by washing the tube with a mixture of methanol and water and using a new cotton piece. It should be mentioned here that this equilibration time does not increase the overall analysis time significantly since this temperature is almost reached while the single-drop extractions, which require 40 min, are carried out. Thus, this removable characteristic of the propose interface allows one to make a considerable number of measurements within a working day. RESULTS AND DISCUSSION The use of ionic liquids as solvents for gas chromatography is an interesting alternative for the determination of analytes that can be extracted by these environmentally friendly solvents. Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

795

Organic solvents, which are toxic pollutants, can be avoided by using ionic liquids since the latter have been previously proved to be efficient extractant media.13,15 As a first approach, the potential of the direct injection of the ionic liquid into a gas chromatograph was investigated. For this purpose, the liner of the chromatograph was filled with glass wool and the direct injection of microvolumes of the ionic liquid containing different halocarbons was accomplished. A chromatogram of the separation of the analytes is shown in Figure 3a. Nevertheless, after three injections, the ionic liquid reaches the capillary column, dirtying the system, and even blocking the column. Panels b and c in Figure 3 show the evolution of the chromatograms obtained as the column is getting dirty. When this occurs, the vacuum of the GC/MS must be vented and the apparatus must be stopped to clean the liner or even change the column. Moreover, the system requires nearly 12 h to reach equilibrium again. On a subsequent attempt, the possibility of the recuperation of the single drop was evaluated. A drop of ionic liquid was formed in the headspace at the injector of the gas chromatograph and was held there for 1 min. After the desorption of the target analytes, the drop was intended to be retracted. However, the complete recuperation of the ionic liquid drop was not possible due to the existence of a stream of helium at high flow rates at the top of the GC injector, which destabilized the drop or even removed it from the syringe. Consequently, the direct injection of ionic liquids in the gas chromatograph was proved to be inadequate. With the aim of preventing the ionic liquid from passing into the capillary column an interface, in which the ionic liquid is retained, should be developed. The chromatogram obtained for the target analytes using the previously described interface as a first approach is depicted in Figure 4, where it can be seen that the analytes are properly volatilized and efficiently separated in the capillary column. What is more, the fact that the ionic liquid does not show any peak in the chromatogram allows one to avoid the need of solvent delay required when organic solvents such as methanol are used to perform the extraction and are injected in the GC. This fact allows the determination of analytes within a wide range of polarity/volatility. Therefore, using one chromatographic column, the more polar compounds, which do not have a strong interaction with the stationary phase and reach the detector at a very small retention time, can be recorded and quantitatively determined. This would be impossible in case of using organic solvents that reach the detector at small retention times during which no signal can be recorded. Optimization of the Injection Process. For optimization purposes, a standard of the halocarbons (trichloromethane, dichloromethane, carbon tetrachloride) was prepared in ionic liquid at a concentration of 20 µg/mL. These compounds were selected for the optimization step in order to provide unfavorable conditions for the chromatographic separation since this three substances have a similar nature and a similar behavior in the chromatographic system. First, the way of injection using the interface was studied. In a first step, the recuperation of the drop in the interface was intended. For this purpose, a stopped-flow technique is necessary in order to prevent the carrier gas from displacing the ionic liquid drop. Thus, the syringe, filled with the ionic liquid was introduced into the interface with the helium flow 796

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

stopped (two-way valve closed). The ionic liquid was hold in the needle, and after 1 min of desorption during which the trap is heated but remains closed, the gas was allowed to carry the volatized analytes into the column (two-way valve opened). This initial attempt revealed that the signals obtained were not appropriate due to the dispersion of the analytes in the interface and the introduction of significant variability in the analyte signals. Direct injection of the ionic liquid drop in the cotton of the interface employing stopped-flow analysis revealed the same problems, proving the inadequacy of the stopped-flow technique for our aim. Therefore, the direct injection of the ionic liquid in the interface without stopping the flow was selected as the best injection mode. The possibility of working in the column splitless mode or septum purge splitless mode was tested. The best chromatographic separation of the target analytes was achieved using the column splitless mode while the septum purge was at 22 mL/ min. The effect of the flow rate of the gas stream carrying the analytes through the interface was also investigated. It should be noted that, for an interface like the used in this work, the injection flow rate is particularly important to ensure proper separation of the analytes. The maximum flow allowed is limited by the possibility of displacing the ionic liquid trapped in the cotton of the removable unit of the interface. On the other hand, this flow can be reduced provided that it is enough to carry the volatilized compounds with minimal dispersion in the interface. Thus, the flow rate was evaluated between 10 and 30 mL/min. The smallest flow rates led to the decrease of the signal whereas the efficiency of the transference of the analytes into the capillary column increased when the flow through the interface was increased up to 25 mL/min. The reason for this effect, which is especially acute when the peak height is considered as analytical signal, might be the loss of analytes via the split vent or the septum purge vent valves in the injector of the chromatograph, also increasing the variability in the determination. Consequently, a flow rate of 25 mL/min, which allowed a quantitative and reproducible signal of the target analytes, was selected as optimum. Optimization of the Working Conditions. The principal variables affecting the analysis are the temperature in the interface and the volume of the ionic liquid injected. The temperature of the interface is an essential parameter to be controlled in order to achieve a proper volatilization of the analytes provided that the boiling point of the ionic liquid is not reached. Its value was studied within 80 and 220 °C. An increase of the temperature is not easy to achieve due to the thermal inertia of the tubing in the interface and the limited transfer heat to the desorption zone. Low temperatures were not enough to volatilize the analytes whereas values over 200 °C led to the volatilization of part of the ionic liquid and the appearance of its peak in the chromatogram. Figure 5a shows that, working in the column splitless mode, an unacceptable dispersion of the peaks and a loss of signals occur when low temperatures are applied in comparison with the quantitative volatilization of the target analytes at 140 °C in Figure 5b. When working with the analytes under study, a temperature of 140 °C was found to be sufficient for their volatilization. Analytes of higher boiling points could be determined when an ionic liquid with higher temperature stability is used.

Figure 3. Chromatograms of the separation of halocarbons when the ionic liquid is directly injected into a gas chromatograph. (a) Separation achieved when the ionic liquid is efficiently retained in glass wool placed in the liner; (b) chromatograph showing the ionic liquid reaching the capillary column and the detector; (c) ionic liquid contaminating the system. (1) Carbon tetrachloride; (2) dichloromethane; (3) trichloromethane; IL, ionic liquid.

The effect of the volume of the ionic liquid injected was also evaluated. A high injected volume can be beneficial for decreasing

the limits of detection since a bigger drop can be used in the SDME procedure. However, this volume is limited by the length Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

797

Figure 4. Chromatogram of the separation of halocarbons when the proposed interface is used. (1) Carbon tetrachloride; (2) dichloromethane; (3) trichloromethane.

of the removable unit. The system developed in this work, which uses a removable unit with a 3-cm-length tubing, allows one to inject a volume of ionic liquid within 0.5-10 µL. When this volume is exceeded in this system, the ionic liquid is likely to reach the transfer line. However, this volume could be increased by increasing the length of the removable unit. Nevertheless, the extractions were conducted by using 2 µL of the ionic liquid in order to increase the number of possible measurements. The liquid volume injected in the interface was proved to have negligible influence on the desorption efficiency of the analytes. As a matter of fact, there is a linear relationship between the signal obtained for the target analytes and the volume injected as explained in more detail in a following section. Prospective Application. An exemplary application of the use of the propose interface in IL-SDME coupled to GC was performed. Substances of different nature were selected in order to evaluate the applicability of the proposed interface. Thus, three analytes were selected as representative pollutants that could appear in water environmental samples: dichloromethane as a halocarbon, p-xylene as a aromatic hydrocarbon, and n-undecane as an aliphatic hydrocarbon. The extraction conditions were selected as follows: A sample volume of 6 mL in a 10-mL glass vial was fixed as optimum as it provides the best sensitivity. Concerning the temperature and equilibration time, they were studied within the intervals 50-85 °C and 20-60 min. The temperature has a double effect on the extraction process; on the one hand, an increase in the temperature favored the enrichment of the gaseous phase with the analytes and thus their preconcentration on the ionic liquid drop; on the other hand, values over 75 °C produced a negative effect on the sensitivity, which was ascribed to the losses of the analytes from the drop due to thermal desorption. Based on these results, a temperature of 70 °C was selected as compromise. The extraction time was then set at 40 min to ensure the reproducibility between measurements. Finally, the use of a homemade glass-coated stirring bar was mandatory to minimize losses of the analytes by adsorption on the polymeric material during the extraction. Potassium chloride was selected 798

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

as salting-out additive to favor the analytes’ release from the sample to the headspace. A chromatogram obtained under these conditions for the extraction of the target analytes at an intermediate concentration of the calibration interval is depicted in Figure 6. The peak of each analyte is separately represented and has been extracted using three specific ions for each analyte. As for the linearity of the injector, the relation between the volume of ionic liquid injected (0.5-2 µL) having the same concentration of the analytes and the signal obtained was found to be linear up to the 2 µL studied and the correlation coefficient was higher than 0.99 for the three substances. This range of linearity is suitable for the aim of the interface presented in this work. Experiments were performed to assess the efficiency of the desorption of the analytes from the ionic liquid retained in the cotton of the interface, including their transference in the carrier gas toward the GC. A set of extractions of an aqueous standard containing the representative compounds selected (dichloromethane, p-xylene, n-undecane) at a 400 ng/mL concentration level were conducted and analyzed using the injection procedure described above and working under the optimized conditions. Analytical precision of the interface (repeatability) was calculated for seven replicates of standards at the same concentration level (400 ng/mL) and is expressed as relative standard deviation in Table 1. When peak area is used as analytical signal, the value obtained is within 3 and 4%, which it is slightly better than that obtained for the peak height (5-7%). It should be taken into account that this value would include the deviation of the SDME procedure and has been obtained without an internal standard. This precision proves that the combination of IL-SDME with GC/ MS offers the possibility of determining the extracted analytes in a more reproducible way than that achieved when organic solvents are used to form the drop of extraction (according to the bibliography values of relative standard deviation can reach 1213%).6,7 That is because ionic liquids boast a high stability at the temperatures used to perform the SDME whereas irreproducible losses of organic solvent take place under these conditions.

Figure 5. Chromatograms showing the effect of the interface temperature on separation and detection: (a) 80 and (b) 140 °C. Peaks of interest: (1) carbon tetrachloride; (2) dichloromethane; (3) trichloromethane.

The possibility of appearance of memory effects occurring between analyses was also checked. For this purpose, 2 µL of ionic liquid without analytes was injected in the interface after the analysis of the same volume of the IL with the analytes extracted. The system does not seem to show noticeable memory effects for the compounds under study. Consequently, our results demonstrate that the proposed interface is effective in quantitatively transferring the target analytes into the GC, offering the possibility of combining IL-SDME with GC/MS to perform the analysis in a more reproducible way than that achieved when organic solvents are used.

The calibration graphs for the response of the target analytes (dichloromethane, p-xylene, n-undecane) were constructed by extracting five working standards of the mixture of the substances prepared in distilled water at different concentrations (between 25 and 400 ng/mL). The 2 µL of the ionic liquid used for the extraction was then injected in the interface, and the analytes were transferred to the GC to be determined. In the present method, peak area was used as analytical signal. The measured signals were related to the concentration of the analytes in the samples by a polynomial expression. This curve relation could be a consequence of the small volume of ionic liquid used to perform the SDME. It is likely that the extraction capacity of 2 µL of the Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

799

Figure 6. Chromatogram of the separation of the target analytes at 20 ng/mL extracted in the ionic liquid [C4MIM][PF6] and determined using the proposed interface coupled to a GC/MS. (1) Dichloromethane; (2) p-xylene; (3) n-undecane.

Table 1. Figures of the Merit of the Proposed Direct Coupling for Its Application in the Extraction of Representative Analytes Using 2 µL of [C4MIM][PF6] RSDa (%) analyte

R2

LOD (ng/mL)

area mode

height

dichloromethane p-xylene n-undecane

0.993 0.984 0.998

5.6 15.6 7.9

3.4 4.4 3.3

7.0 5.6 5.9

a

Relative standard deviation.

microdrop of ionic liquid is reaching its maximum at the upper concentrations tested. Higher volumes of the extracting medium could accept higher concentrations of the analytes and the relation would not be a curve. Calculating detection and quantitation limits from a nonlinear signal-concentration response is difficult. In order to simplify the calculations, we fitted a linear relationship to the lower portion of the calibration curve and used the standard deviation of the intercept (Sa). The limits of detection (LOD) were calculated as three times the standard deviation of the intercept divided by the slope of the graph. The obtained values are shown in Table 1, in all cases being in the lower nanograms per milliliter range. These lowest detectable concentrations were similar to those obtained using the criterion adopted by the U.S. Environmental Protection Agency (U.S. EPA), which revolves around a parameter called the method detection limit.30 It should be noted that these detection limits are reached for the analytes under (30) U.S. EPA. Guidance on Evaluation, Resolution and Documentation of Analytical Problems Associated with Compliance Monitoring, Washington, DC, 1993.

800 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

nonoptimized conditions of extraction since the aim of the present work was not to optimize the extraction process but to prove the feasibility of the constructed interface for the determination of analytes at low concentration ranges. CONCLUSIONS A new interface for the direct coupling between IL-SDME and GC/MS has been successfully developed. Using this interface, no ionic liquid can reach the GC system dirtying the chromatographic column and a proper volatilization and subsequent transference of the substances to the GC is achieved. Therefore, the advantages of the use of ionic liquids with SDME, avoiding the irreproducibility associated with the use of organic solvents to form the drop, are added to the analytical possibilities of GC/ MS. Furthermore, using the proposed interface analytes of an extensive range of polarity/volatility can be determined on account of the fact that no solvent delay is necessary in the detection step. Moreover, this fact allows the determination and quantitation of substances with a small retention time, which would not be possible if organic solvents like methanol are used as extractants and then are injected in the GC system. In addition, compounds with a stronger interaction with the column, that is, n-undecane, can also be properly volatilized and separated using the same chromatographic column. ACKNOWLEDGMENT Financial support from the Spanish DGICyT (Grant CTQ200401220) is gratefully acknowledged. Received for review July 24, 2007. Accepted October 30, 2007. AC071555G