Analysis of Semivolatile Pharmaceuticals and Pollutants in Organic

Apr 21, 2009 - Hua Chen, Zhining Xia, Stig Pedersen-Bjergaard, Bo Svensmark and Frants R. Lauritsen*. Department of Chemistry, University of Copenhage...
0 downloads 0 Views 317KB Size
Anal. Chem. 2009, 81, 4010–4014

Analysis of Semivolatile Pharmaceuticals and Pollutants in Organic Micro Extracts Using Hot Cell Membrane Inlet Mass Spectrometry Hua Chen,† Zhining Xia,† Stig Pedersen-Bjergaard,‡,§ Bo Svensmark, and Frants R. Lauritsen§,* Department of Chemistry, University of Copenhagen, Universitetsparken 5, 2100 Copenhagen, Denmark This paper presents the first membrane inlet method that can be used together with field portable mass spectrometers for the analysis of semivolatile pharmaceuticals (pethidine, benzophenone, and cocaine) and environmental pollutants (terbutryne and butylated hydroxyl toluene (BHT)) dissolved in organic micro extracts. A microliter of the organic extract is simply injected into a closed hot cell membrane inlet (hc-MIMS), and an electron ionization mass spectrum of the vaporized semivolatile sample molecules can be recorded shortly thereafter. Detection limits at low picomole quantities or low/sub ng/µL concentrations in the extract are demonstrated for solutes in methanol, ethanol, acetone, and toluene. A linear correlation between analyte concentration and signal was found in the range of 1-100 ng/µL, and the relative standard deviation (RSD) was approximately 10%. As a practical example we demonstrate the detection of cocaine in extracts from dried coca leaves. The analysis of organic micro extracts using hc-MIMS represents a considerable extension of the type and complexity of analytes that can be measured using a field portable MIMS system, since it does not require special and field tedious modifications to the standard MIMS system. In recent years a number of miniaturized field portable mass spectrometers including commercial instruments have been developed. Some of these instruments have total weights at or below 10 kg including batteries.1-3 With field portable instruments it has become possible to perform on-site, real-time analysis of gaseous and liquid samples for their content of gases and volatile organic compounds, with the mapping of gas emissions at volcanoes4 and mapping of chemicals in water using under water * To whom correspondence should be addressed. E-mail: [email protected]. † Current address: College of Chemistry and Chemical Engineering, Chongqing University, Chongqing 400044, China. ‡ Current address: School of Pharmacy, University of Oslo, P.O. Box 1068 Blindern, 0316 Oslo, Norway. § Current address: Faculty of Pharmaceutical Sciences, University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark. (1) Gao, L.; Sugiarty, A.; Harper, J. D.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2008, 80, 7198–7205. (2) Yang, M.; Kim, T. Y.; Hwang, H. C.; Yi, S. K.; Kim, D. H. J. Am. Soc. Mass Spectrom. 2008, 19, 1442–1448. (3) Janfelt, C.; Frandsen, H.; Lauritsen, F. R. Rapid Commun. Mass Spectrom. 2006, 20, 1441–1446. (4) Griffin, T. P.; Diaz, J. A.; Arkin, C. R.; Soto, C.; Curley, C. H.; Gomez, O. J. Am. Soc. Mass Spectrom. 2008, 19, 1411–1418.

4010

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

mass spectrometry5 as recent examples. Field applications are normally carried out by the help of either a gas leak or a membrane inlet,6 inlets that are very simple in construction and typically makes direct sample analysis possible. However, these inlets limit the applicability to the analysis of volatile organic compounds (bp. less than 250 °C). To overcome this problem the first examples of electro spray ionization of both extreme hydrophilic and non-volatile organic compounds using small field portable instruments have been demonstrated.1,7 Currently, the mass spectrometric performance with respect to mass resolution and sensitivity for these instruments is not sufficiently good, but this is likely to be changed in a number of years. Hitherto the analysis of semivolatile organic compounds (sVOCs with bp. above 250 °C) and compounds dissolved in organic solutions using MIMS has required special modifications to the membrane inlet and/or special membranes. In most cases these modifications make the use of the technique in the field difficult. For example, to analyze sVOCs it has been necessary to use stimulated desorption of analytes from the membrane surface. Most of these methods are based upon the trap-and-release technique,8,9 where analytes are trapped in a cold (room temperature) membrane over some time (10-20 min) before they are released into the mass spectrometer using various forms of stimulated desorption.10-20 With these methods fairly complex (5) Kibelka, G. P. G.; Short, R. T.; Toler, S. K.; Edkins, J. E.; Byrne, R. H. Talanta 2004, 64, 961–969. (6) Kotiaho, T. J. Mass Spectrom. 1996, 31, 1–15. (7) Keil, A.; Talaty, N.; Janfelt, C.; Noll, R. J.; Gao, L.; Ouyang, Z.; Cooks, R. G. Anal. Chem. 2007, 79, 7734–7739. (8) Leth, M.; Lauritsen, F. R. Rapid Commun. Mass Spectrom. 1995, 9, 591– 596. (9) Matz, G.; Kesners, P. Analysis 1995, 23, M12-M16. (10) Ketola, R. A.; Lauritsen, F. R. Anal. Chem. 1997, 69, 4917–4922. (11) Mendes, M. A.; Eberlin, M. N. Analyst 2000, 125, 21–24. (12) Vallasco, A. P.; Haddad, R.; Eberlin, M. N.; Hoehr, N. F. Analyst 2002, 127, 1050–1053. (13) Lauritsen, F. R.; Rose, J. Analyst 2000, 125, 1577–1581. (14) Aggerholm, T.; Lauritsen, F. R. Rapid Commun. Mass Spectrom. 2001, 15, 1826–1831. (15) Soni, M. H.; Callahan, J. H.; McElvany, S. W. Anal. Chem. 1998, 70, 3103– 3113. (16) Soni, M. H.; Baronavski, A. P.; McElvany, S. W. Rapid Commun. Mass Spectrom. 1998, 12, 1635–1638. (17) Munchmeyer, W.; Walte, A.; Matz, G. Polycyclic Aromat. Hydrocarbons 1996, 9, 299–306. (18) Creaser, C. S.; Weston, D. J.; Smith, B. Anal. Chem. 2000, 72, 2730–2736. (19) Riter, L. S.; Taka´ts, Z.; Cooks, R. G. Analyst 2001, 126, 1980–1984. (20) Creaser, C. S.; Lamarca, D. G.; Freitas dos Santos, L. M.; New, A. P.; James, P. A. Analyst 2003, 128, 1150–1156. 10.1021/ac900437t CCC: $40.75  2009 American Chemical Society Published on Web 04/21/2009

molecules such as R-tocopherol (vitamin E), atrazine, and polyaromatic hydrocarbons with up to five rings have been analyzed. For the analysis of organic solutions alternative membranes, than the normal polydimethylsiloxane (PDMS) membranes, are often used. Two possibilities exist, pervaporation membranes 3,21-25 and microporous membranes.3,26,27 Pervaporation membranes have the advantage that they can be used with almost any standard MIMS system, and they are therefore fully compatible with field portable instruments. However, since they discriminate against the organic solvent, they also reduce the flow of organic sample molecules into the mass spectrometer, and achievable detection limits are generally not very good. The microporous membrane does not show any significant discriminatory effect but lets more or less all molecules enter the mass spectrometer to the same relative extent. The result is a system with a fairly high sensitivity for all organic compounds. Unfortunately, the advantages are gained as a result of a 1000 folds higher flux of molecules into the mass spectrometer, and two stage pumping of the mass spectrometer is generally required. Recently, we introduced the hot cell MIMS (hc-MIMS) idea,28,29 for direct analysis of chemicals liberated from solid materials such as soil, plant materials, plastics, and so on. In this technique the solid material is simply transferred without any pretreatment to a small heated (≈200 °C) measuring cell equipped with a flat sheet membrane interfaced directly to the ion source of an electron ionization mass spectrometer. sVOCs liberated from the solid material penetrates the membrane within 1-2 min and enter the mass spectrometer for analysis. In this fashion practically any solid sample could be screened for its potential liberation of sVOCs with a turnover of 10-15 samples per hour. A major advantage of the hc-MIMS technique is its compatibility with field portable mass spectrometers.28 As described the hc-MIMS system is a very fast screening method, but once an interesting sample has been detected, the question of quantification arises. Standards are not so easily prepared, since the observed signal not only depends upon concentration. It also depends upon morphology and amount of the sample, and whether the chemical is deposited upon the sample surface or embedded in the sample. How do we prepare a representative piece of soil, plastic, or rotten wood with a welldefined content of the chemical in question? To overcome this problem, we have explored the possibility of using the hc-MIMS setup to quantify sVOCs dissolved in organic solutions. The idea presented here is simply to inject a microliter of an organic solution with the chemical of interest into the hot cell in a fashion similar to the injection of samples into a gas chromatograph. The (21) Degn, H.; Bohatka, S.; Lloyd, D. Biotechn. Techn. 1992, 6, 161–164. (22) Favretto, D.; Traldi, P.; Benassi, C. A.; Bettero, A. Biol. Mass Spectrom. 1991, 20, 669–676. (23) Creaser, C. S.; Lamarca, D. G.; Brum, J.; Werner, C.; New, A. P.; Freitas dos Santos, L. M. Anal. Chem. 2002, 74, 300–304. (24) Maden, A. J.; Hayward, M. J. Anal. Chem. 1996, 68, 1805–1811. (25) Kasthurikrishnan, N.; Cooks, R. G.; Bauer, S. Rapid Commun. Mass Spectrom. 1996, 10, 751–756. (26) Lauritsen, F. R.; Kotiaho, T.; Choudhury, T. K.; Cooks, R. G. Anal. Chem. 1991, 64, 1205–1211. (27) Clinton, R.; Creaser, C. S.; Bryant, D. Anal. Chim. Acta 2005, 539, 133– 140. (28) Frandsen, H.; Janfelt, C.; Lauritsen, F. R. Rapid Commun. Mass Spectrom. 2007, 21, 1574–1578. (29) Lauritsen, F. R.; Jensen, A.; Nielsen, C. H. Rapid Commun. Mass Spectrom. 2008, 22, 2334–2340.

Figure 1. Chemical structure of the five test compounds.

advantages of using microliter quantities are several: The amount of organic solvent required is limited, preconcentration of the sample is possible by evaporation of the solvent or by the use of microextraction techniques, 30,31 and the small quantity of organic solvent injected into the hot cell makes it possible to use the traditional PDMS membrane. The main advantage of the hc-MIMS system is its compatibility with field portable mass spectrometers, since it only requires a small electrically heated standard membrane inlet. EXPERIMENTAL SECTION Chemicals. To characterize the hot cell MIMS technique for analysis of sVOCs of different origin and use, we have chosen four representative chemicals (Figure 1). Pethidine (CAS R.N. 57-42-1, C15H21NO2, MW: 247) is a strong opioid used against severe pain, benzophenone (CAS R.N. 119-61-9, C13H10O, MW 182) is a UV-blocker used, for example, in cosmetics, terbutryne (CAS R.N. 886-50-0, C10H19N5S, MW 241) is a herbicide of the triazine type, and butylated hydroxy toluene (BHT) (CAS R.N. 128-37-0, C15H24O, MW 220) is a common antioxidant in plastics. A practical demonstration of the technique was performed using dried coca leaves (Mate de Coca, Coca tea, ENACO S.A., Peru) and standards containing cocaine (CAS R.N. 50-36-2, C17H21NO4, MW 303) dissolved in methanol. The chemicals were obtained from Sigma/Aldrich except for the cocaine standard, which was kindly supplied by the Department of Forensic Medicine, Copenhagen University. Design of the Hot Cell/Injection Chamber. The hot cell membrane inlet/ion source system used in this work is very similar in design to the sample cell system we introduced in 1998 for analysis of flavor compounds directly from microbial media.32 A schematic diagram of the inlet is shown in the original paper, and a photo of a hot cell is shown in the paper introducing hcMIMS for direct analysis of soil samples.28 The major difference between the sample cell used here for micro extracts and previous sample cells is the small volume of the hot cell/injection chamber, which is only approximately 250 µL. It has a lid with two 3 cm long 1/16” tubes (1 mm ID). One tube in the lid is meant for insertion of the needle from a micro syringe, and the other tube, (30) Pedersen-Bjergaard, S.; Rasmussen, K. E. J. Chromatogr. A 2008, 1184, 132–142. (31) Pawliszyn, J.; Pedersen-Bjergaard, S. J. Chromatogr. Sci. 2006, 44, 291– 307. (32) Lauritsen, F. R.; Lunding, A. Enzyme Microb. Technol. 1998, 22, 459–465.

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4011

which is open to air, prevents an excessive buildup of pressure inside the hot cell caused by rapid vaporization of the solvent during sample injection. An excessive buildup of pressure in the hot cell could lead to membrane rupture and vacuum failure, since we were using an unsupported membrane. The hot cell is mounted on the atmospheric side of a vacuum flange, whereas a closed electron ionization source is built into the vacuum (MS) side of the flange. The hot cell and the back side of the ion source is connected via a 3 mm hole that is covered by a flat and unsupported 125 µm thick silicone membrane of medical grade (Sil-Tec Sheeting, Technical Products Inc., Decatur, GA). The integrated design of the hot cell and ion source makes it possible for sVOCs to pass directly via the membrane from the hot cell and into the ionizing region of the ion source and thereby eliminating condensation effects33,34 caused by interactions of the sVOCs with surfaces in vacuum. Mass Spectrometer. All experiments were conducted using a Balzers (Liechtenstein) QMG 422 single quadrupole mass spectrometer with a 3-lens ion optical transfer system for collection and focusing of ions generated by sources separated from the mass analyzer itself. The mass spectrometer was mounted in a two stage differentially pumped vacuum system, where the ion optical transfer system is isolated from the analyzer part using a large focusing plate with a 3 mm hole. The two sections of the vacuum system were each pumped by a TSU 070 Turbomolecular pump (Pfeiffer Vacuum, Asslar, Germany). The important thing of the setup is that the 3-lens ion optical transfer system terminates directly at the vacuum flange with the fully integrated hot cell membrane inlet/ ion source system. Experimental Procedure. The analysis of micro extracts using hc-MIMS is simply carried out by injecting 0.1-1.0 µL of the extract directly into the small hot cell via one of the tubes in the lid using a micro syringe. Selected ion monitoring of characteristic electron ionization peaks were used to follow the instrumental response to the injection. Typically the sample was left inside the hot cell for 1-2 min before it was sucked out of the cell using a standard diaphragm gas suction pump (1 L/min). In most cases the signal returned to baseline within 2 min, and injections of successive samples could be performed with approximately 5 min interval. In most of the experiments the sample was injected 5-6 times, and the average peak height was used to characterize the various operational parameters. RESULTS AND DISCUSSION Characterization of the hc-MIMS Technique for Analysis of Organic Micro Extracts. Figure 2 shows the result of 10 successive injections of 1 µL of BHT in methanol into the hot cell at 200 °C. Shortly after the injection the signal rose rapidly and reached a maximum after approximately 1 min, where after it slowly decreased as a result of BHT disappearing out of the cell via the membrane and the open tubes. Approximately 1.5 min after the injection the open capillary was connected to the air pump, and as a result the signal disappeared rapidly reaching baseline within 1.5 min. The complete analytical cycle was completed within 3 min. If the peak height is used as a quantitative measure, the relative standard deviation (RSD) of the injections shown in Figure (33) Lauritsen, F. R. Int. J. Mass Spectrom. Ion Process. 1990, 95, 259–268. (34) Hansen, K. F.; Gylling, S.; Lauritsen, F. R. Int. J. Mass Spectrom. Ion Process. 1996, 152, 143–155.

4012

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

Figure 2. Single ion monitoring (m/z 205) of 10 successive injections of 1 µL samples of BHT in methanol (50 ng/µL). Table 1. Relative Peak Heights (Average of 6 Injections) from the Four Test Compounds Following Injection of 1 µL of Samples Containing 50 ng/µL in Water, Methanol, Ethanol, Acetone, and Toluene

BHT Benzophenone Terbutryne Pethidine

water

methanola

ethanol

acetone

toluene

0.26 0.50 0.58 0.43

1.00 1.00 1.00 1.00

0.83 0.27 1.04 1.08

0.45 0.20 1.01 0.96

0.12 0.07 0.70 0.99

a The signals from the individual test compounds are normalized to 1.0 for methanol.

2 was 6%, although in other experiments it was up to 10%. The scattering is most likely caused by problems with reproduction of the manual injection conditions. However, for a field portable analytical method a 10% RSD is generally an acceptable number. Extractions are generally performed into many different solvents, and to study any effect of the solvent upon the hc-MIMS signal we prepared test solutions (50 ng/µL) of the four test compounds in water, methanol, ethanol, acetone, and toluene. These solvents cover the range from extreme hydrophilic (water) to extreme hydrophobic (toluene). Table 1 shows the peak height from the four analytes dissolved in the five solvents. To ease comparison across solvents, the signals from the four different compounds were normalized against the peak height obtained with the analyte dissolved in methanol. Some general observations can be extracted from Table 1. The signal obtained with the analytes dissolved in water was lower than that obtained with the analytes dissolved in methanol in all cases. Apart from benzophenone, that gave relatively low signals in ethanol, acetone, and toluene, the signals obtained from the other three compounds in ethanol and acetone were similar or slightly lower than that obtained when they were dissolved in methanol. Both BHT and benzophenone gave low signals when dissolved in toluene, whereas pethidine and terbutryne gave signals in toluene that were similar to the signals from methanol. The generally high signals obtained when the analytes were dissolved in an organic solvent as compared to water is probably caused by a swelling of the hydrophobic silicone membrane, when it is exposed to concentrated vapors of organic solvent,8 whereas it might contract when exposed to water. Swelling and contraction of the membrane is expected to be particular important in the hot cell MIMS setup used here because the membrane is used without any support. It is therefore expected to bulge into the vacuum, and the extent of this bulging depends upon the softness of the membrane. The organic solvents therefore both soften the polymer structure and increase the active membrane area. In

Figure 3. Single ion monitoring (m/z 228) of 1 µL samples of terbutryne in methanol (50 ng/µL ) recorded at different temperatures. The injections started at 40 s, and the hot cell was flushed with laboratory air at the arrows.

addition to the signal enhancing modifications to the membrane, when organic solvents are used, an opposite effect might arise as the result of excessive pressure in the closed ion source. We noticed that the total pressure in the analyzer part of the instrument increased with more than a factor of 10 to 2 × 10-6 mbar with the organic solvents. With the differential pumping used in our instrument the pressure inside the closed ion source is estimated to be in the 10-3 mbar range. At this high pressure an attenuation of the electron ionization peaks is to be expected. When the hot cell MIMS technique was used for direct analysis of solid materials we observed that the signal-to-noise ratio increased and the rise and fall times decreased with temperature.29 We expected to see the same behavior here with injection of micro extracts into the hot cell. To test this, standard solutions of the four test compounds in methanol were all injected into the hot cell at temperatures of 175, 200, 225, and 240 °C. Figure 3 shows the result for terbutryne. At 175 and 200 °C the peak rose too slowly for a maximum to be reached before the hot cell was cleaned by flushing air through it. At 225 and 240 °C the peak rose quickly to a maximum and then slowly decreased again as the sample molecules disappeared from the hot cell either via the membrane into the mass spectrometer or out via the two 3 cm long tubes in the lid. The relative increase in signal intensity for benzophenone, BHT, pethidine, and terbutryne was found to be a factor of 1.6, 1.6, 2.3, and 8.6, respectively, when the temperature rose from 175 to 240 °C. We consider the hc-MIMS technique as a head space analytical technique, even though there are two tubes in the lid that makes it possible for the solvent and analyte molecules to escape from the cell. Apart from a fast blow of solvent out of the cell in connection with the injection, the long tubes ensure a slow diffusion of sample molecules out of the cell during analysis. We therefore expected that the sensitivity should increase linearly with the injection volume. To test this, different volumes of a standard solution of BHT dissolved in water, methanol, and toluene were injected into the hot cell. The result is shown in Figure 4. In the case of toluene we got the expected linear correlation between injection volume and sensitivity up to the maximum tested volume of 1 µL. To avoid any risks of membrane rupture as a result of membrane swelling, when exposed to concentrated solvent vapors, we decided not to operate with volumes higher than 1 µL. With this limitation we never experienced any membrane rupture even though we have injected more

Figure 4. Recorded peak height from BHT as a function of sample volume. ( BHT dissolved in toluene, 9 BHT dissolved in methanol, and 2 BHT dissolved in water. For easy comparison between the solvents each solvent series was normalized to 1 for the maximum recorded peak height. Table 2. Calibration Data and Detection Limits for the Four Test Compounds in Methanol detection squared calibration limit ion monitored correlation range (S/N ) 3), compound (m/z) coefficient, R2 (ng/µL) (ng/µL) BHT Benzo phenone Terbutryne Pethidine

205 182

0.999 0.999

0.5-100 0.5-100

0.3 0.3

228 247

0.969 0.983

2.0-200 5.0-200

1.0 2.0

than 2000 samples during this work. In the case of methanol we observed an increase in signal with sample volume that was faster than linear. We expect that this behavior reflects an increased swelling of the membrane as the concentration of methanol inside the hot cell increases. Methanol is a more polar compound than toluene, and it does not swell the membrane nearly as effective as toluene. Water as solvent was surprising, since it showed an increase in sample signal at low injection volumes (below 0.4 µL), where after the signal decreased relatively fast with increasing sample volume. We think that two opposing mechanisms explain this behavior. At low volumes the increase in analyte concentration inside the hot cell with increasing injection volume dominates and the signal rises, whereas at higher sample volumes a contraction of the hydrophobic silicone membrane, when exposed to the polar water molecules reduce diffusion of sample through the membrane. Linearity and detection limits are important characteristics of an analytical technique. To test hc-MIMS for these parameters standard solutions of the four test compounds dissolved in methanol were injected into the hot cell (1 µL samples, 225 °C). In all cases a linear relationship was found over 2-3 orders of magnitude (see Table 2). The detection limits (signal/noise ) 3) were found to be 0.3 ng/µL (≈ 2 picomole) for BHT and benzophenone, 1 ng/µL (≈ 4 picomole) for terbutryne, and 2 ng/ µL (≈ 8 picomole) for pethidine. The four test compounds are similar in the way that they all have large hydrophobic groups and one or two small hydrophilic parts. The difference in detection limits is likely to be the result of differences in size (molecular weight) and vapor pressure. The hc-MIMS technique is an all round technique intended for many applications, such as analysis of pharmaceuticals in extracts from blood and urine at emergency rooms in hospitals, on-site screening of plant materials for their content of potentially active compounds, and detection of contaminants in environmental samples. In some cases our detection limits are more than sufficient as exemplified below by the analysis of cocaine in coca leaves, and in other cases it will not be sufficient, Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4013

thereby revealing the presence of cocaine in the leaves. The spectrum is distorted by a large number of ions at lower masses, which we expect are the results of other compounds in the leaves and pyrolysis of the leaves in the 200 °C hot cell. The spectrum of the extract (Figure 5b) shows a much cleaner spectrum with an almost perfect match to the standard spectrum (Figure 5c), except for extra peaks at m/z 256 and m/z 199 revealing the presence of another unidentified sVOC in the leaves. The direct analysis of solid samples using hc-MIMS is very convenient for fast screening of samples, but it is a qualitative analysis that only gives a very rough estimate of concentrations. Via the use of small extracts as demonstrated here a good quantitative analysis is obtained. Using single ion monitoring of m/z 303 following injections of the extract and comparison of the peak heights with that of standards (640 ng/µL and 320 ng/µL), we estimated the concentration of cocaine in the extract to be 600 ± 80 ng/µL. With 0.1 g of coca leaves extracted into 0.50 mL methanol this correspond to 3.0 ± 0.4 mg cocaine per g coca leaves. This result is in very good agreement with published values for the cocaine content in Mate de Coca tea (approximately 3 mg/ g).35 The uncertainty in the quantitative estimations is higher than normal for laboratory analysis, but for applications in the field they are fully acceptable.

Figure 5. EI-MS spectra recorded from the following: (a) Dried leaves from Mate de Coca tea poured directly into the hot cell without any treatment, (b) extract of leaves from Mate de Coca tea in methanol injected into the hot cell, and (c) standard solution of cocaine in methanol (640 ng/µL) injected into the hot cell.

for example for trace level detection of contaminants in the environment. Analysis of Dried Coca Leaves Using hc-MIMS. As a practical demonstration of the possibilities in hc-MIMS for analysis of micro extracts, we have chosen to analyze dried coca leaves from the famous Mate de Coca tea. To do this, 0.1 g of coca leaves were simply mixed with 0.5 mL of methanol in a 1.5 mL HPLC sample vial and shaken for 5 min by hand. The extract was then analyzed by injecting 1 µL samples of the extract into the hot cell. For comparison Figure 5a-c shows the following: (a) The EIMS spectrum recorded from untreated coca leaves that were simply poured into the hot cell, (b) the EI-MS spectrum recorded from the extract of the coca leaves, and (c) the EI-MS spectrum recorded following injection of 1 µL samples of a standard solution of cocaine in methanol (640 ng/µL). The spectrum of the untreated coca leaves (Figure 5a) clearly shows characteristic ions of cocaine at m/z 303 (molecular ion), m/z 272, and m/z 182

4014

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

CONCLUSION This paper demonstrates that hc-MIMS can be used for quantitative analysis of semivolatile pharmaceuticals and environmental contaminants in organic micro extracts (sub microliter samples). The technique is very suited for field analysis for several reasons: (a) hc-MIMS is compatible with standard MIMS, (b) it can be used with field portable instruments, (c) the analytical procedure is very simple, and (d) only very small amounts of solvents are required. ACKNOWLEDGMENT This work was supported by the Danish Science Research Council and the Danish Carlsberg Foundation. Hua Chen was supported by a travel grant from the China Scholarship Council of Chinese Education Ministry.

Received for review February 27, 2009. Accepted April 6, 2009. AC900437T (35) Strano-Rossi, S.; Colamonici, C.; Botre`, F. Anal. Chim. Acta 2008, 606, 217–222.