Anal. Chem. 2000, 72, 2774-2779
In-Tube Solid-Phase Microextraction Coupled to Capillary LC for Carbamate Analysis in Water Samples Yanni Gou and Janusz Pawliszyn*
The GuelphsWaterloo Centre for Graduate Work in Chemistry, Department of Chemistry, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada
Recently, the on-line sample preparation technique, intube solid-phase microextraction (SPME), was successfully implemented with a Hewlett-Packard 1100 HPLC system for analysis of carbamates in water samples. This paper describes the coupling of in-tube SPME to capillary LC and explores its utility as a sample preparation method in that format, relative to conventional LC. The HewlettPackard HPLC system was upgraded to a capillary LC system using commercially available accessories from LC Packings. The combination of in-tube SPME with a capillary LC system was expected to build on the merits of both in-tube SPME and the capillary LC to generate a sensitive method with an easy, effective, and efficient sample preparation. Due to the relatively large effective injection volume of the in-tube SPME technique (30-45 µL), on-column focusing was employed in order to achieve good chromatographic efficiency. Excellent sensitivity was achieved with very good method precision. For all carbamates studied, the RSD of retention time was between 0.5 and 0.8% under 4 µL/min microgradient conditions. The RSD of peak area counts was between 1.5 and 4.6%. The detection limits for all carbamates studied were less than 0.3 µg/L and, for carbaryl, just 0.02 µg/L (20 ppt). Compared with the conventional in-tube SPME/LC method, the LODs were lowered for carbaryl, propham, methiocarb, promecarb, chlorpropham, and barban, by factors of 24, 45, 42, 81, 62, and 56, respectively. The optimized method was successfully applied to the analysis of carbamates in surface water samples. This paper demonstrates the coupling of in-tube solid-phase microextraction (SPME) with capillary liquid chromatography (LC) and is intended to describe the advantages and limitations of the method. It builds on previous work describing the implementation of the method for sample preparation coupled to conventional LC. Automation and miniaturization have been two important trends in the development of liquid chromatography over the course of the last two decades. The continuing interest in microcolumn LC reflects the trend in miniaturization that started in the late 1970s.1 Microcolumn LC has been established as a complementary technique to conventional LC, one of the most (1) Krejco´, M. Trace Analysis with Microcolumn Liquid Chromatography; Marcel Dkker, Inc.: New York, 1992; Chapter 2.
2774 Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
powerful separation and analysis tools available. The common cited advantages of the microcolumn LC are as follows: (1) the possibility of reaching very high column efficiency (105-107 theoretical plates) and subsequent resolution of complex mixtures; (2) the ability to work with small volumetric flow rates, thus drastically reducing the consumption of expensive and also environmentally hazardous mobile phases; (3) the improvement in detection performance because of the easy coupling with masssensitive detection devices, such as mass spectrometry, as a result of the strongly reduced chromatographic dilution; and (4) the ability to work with minute sample volumes. Currently, microcolumn LC is almost exclusively performed with slurry-packed columns of various dimensions as a research tool and in routine analysis.2 Microcolumn LC is used frequently as the overall term to describe small-inner diameter packed LC columns. The trend has been to describe 0.50-1.0-mm-i.d. columns as micro-LC, 100500-µm-i.d. columns as capillary LC, and 10-100-µm-i.d. columns as nanoscale LC.3 Because of the growing interest in microcolumn LC, more dedicated systems and system components have recently become commercially available. Many researchers however prefer to upgrade a conventional LC system to a microscale or capillary LC system using appropriate accessories.4 The upgraded LC systems show many advantages over the dedicated microcolumn LC systems, such as high reproducibility, high separation, and detection sensitivity, over a large working range: conventional scale, microscale, capillary scale, and nanoscale. The on-line coupling of in-tube SPME to an upgraded capillary LC system is the goal of this project; however, it presents a significant challenge inherent to the limited injection volume requirement of the capillary LC. In the chromatographic method, the sample volume and the way in which this volume is transferred to the column heavily affect band dispersion and resolution. To maintain chromatographic efficiency on packed capillary and microcolumns, a small-volume injection is necessary, and the theoretical treatment of maximum injection volume is discussed in recent literature.5 From this reference, the maximum injection volume for a 15-cm-length capillary column with 300-µm i.d. and (2) Vissers, J. P. C.; Claessens, H. A.; Cramers, C. A. J. Chromatogr., A 1997, 779, 1507. (3) Chervet, J. P.; Mrsem, M.; Salzmann, J. P. Anal. Chem. 1996, 68, 1552. (4) Grimm, R.; Serwe, M.; Chervet, J. P. LC-GC 1997, 15 (10), 960. (5) Vissers, J. P. C.; de Ru, A. H.; Urem, M.; Chervet, J. P. J. Chromatogr., A 1996, 746, 1. 10.1021/ac990726h CCC: $19.00
© 2000 American Chemical Society Published on Web 05/04/2000
5-µm particle diameter was calculated to be 46 nL for nonretained compounds. In general, a 60-nL injection volume is frequently used for a 300-µm-i.d. capillary column. In the in-tube SPME technique implemented on the HP autosampler, the injection volume is typically over 20 µL, which is significantly larger than the optimal injection volume required by capillary LC. Small injection volumes typically lead to a decrease in sensitivity for chromatographic analyses. This is a common problem for microcolumn LC. The reduced sample capacity leads to limited detection, especially in the case of UV or fluorescence detection.6 Research has been conducted in order to introduce larger sample volumes onto a microcolumn or capillary column. In one approach, solutes are dissolved in a solvent with an elution strength weaker than that of the mobile phase. This technique is known as oncolumn focusing or peak compression sampling. This method allows the sensitivity of the analysis to be increased in proportion to the increase of the volume of sample injected.7 An alternative approach to injecting large sample volumes into the column is to use microprecolumns combined with column-switching techniques. These methods require an additional pump and an injection valve for loading the sample and flushing the precolumn.8 The on-column focusing technique was applied for our research on the on-line coupling in-tube SPME with capillary LC. EXPERIMENTAL SECTION Chemicals and Water Samples. Six carbamates (barban, carbaryl, chlorpropham, methiocarb, promecarb, propham) were purchased from Chem Service (West Chester, PA). All were a purity of g98% and were used as received. Acetonitrile and methanol (HPLC grade quality) were from EM Science (Gibbstown, NJ). Nanopure water was obtained from a Barnstead/ Thermodyne NANOpure ultrapure water system (Dubuque, IA). Surface water was obtained from Red Mill, PA. Single standards for each compound with concentrations of 2 mg/mL were prepared using methanol (HPLC grade) as a solvent. A methanolic standard stock mixture with a concentration of 50 µg/mL for each compound was prepared. Aqueous samples of each compound at concentrations of 500 µg/L were prepared fresh daily by spiking the methanolic stock solution into Nanopure water or surface water. For the limit of detection (LOD) experiments, aqueous samples with concentrations of 1 µg/L for each compound were prepared by spiking with 50 µg/L methanolic stock solutions. Capillary LC system. A HP 1100 HPLC system (HewlettPackard Co., Wilmington, DE) consisting of a binary pump, an on-line vacuum degasser, an autosampler, a thermostated column compartment, and a variable-wavelength UV detector was used. The mobile phases were filtered through a Nylaflo membrane disk filter (0.45 µm, 47 mm) (Pall Gelman Science) and degassed by sonication before usage. A mobile phase of acetonitrile/water (10/ 90, v/v) was used during extraction. After injection, a gradient elution was performed as follows: starting conditions, acetonitrile/ (6) Ling, B. L.; Baeyens, W. J. Microcolumn Sep. 1992, 4, 17. (7) Haddad, P. R. In Selective Sample Handling and Detection in High-performance Liquid Chromatography; Frei, R. W., Zech, K., Eds.; Elsevier: Amsterdam, 1989; Part B, Chapter 2. (8) Nielen, M. W. F.; Frei, R. W.; Brinkman, U. A. Th. In Selective Sample Handling and Detection in High-performance Liquid Chromatography; Frei, R. W., Brinkman, U. A. Th., Eds.; Elsevier: Amsterdam, 1988; Part A, Chapter 1.
water 10/90 ramp to 60/10 in 1 min, then hold. An Acurate microflow processor (LC Packings [USA] Inc., San Francisco, CA) was placed between the pump and the injection valve. The pump flow rate was set at 400 µL/min, and the microflow after the split was 4 µL/min. A 15 cm × 300 µm capillary column (LC Packings) packed with Hypersil C18, 3-µm dp material, was connected to a 1 cm × 300 µm guard column (LC Packings) packed with Hypersil C18, 5-µm dp material and to the injection valve by means of suitable fittings. The UV detector was equipped with a 35-nL, 8-mm path length, U-shaped UZ -View capillary flow cell (LC Packings). The wavelength for detection was 220 nm. The HP ChemStation software (Hewlett-Packard) was used for system control, data acquisition, and evaluation. The capillary LC instrument setup is shown in Figure 1. System Setup for In-Tube SPME/Capillary LC. To install in-tube solid-phase microextraction on the system, a 30-cm section of GC capillary, (Omegawax 250, i.d. 0.25 mm, film thickness 0.25 µm. Supelco, Bellefonte, PA), was connected to the injection needle assembly of a HP 1100 HPLC system by means of a 2-cm sleeve of 1/16 in. o.d. × 0.02 in. i.d. PEEK tubing and standard 1/ -in. HPLC fittings. This was used as the in-tube SPME capillary 16 to extract analytes from aqueous samples. The other end of the capillary was connected to a 60-cm section of polar deactivated fused-silica tubing (i.d. 0.25 mm. Supelco) using similar fittings and a zero dead volume union. The silica tubing was subsequently connected to the metering device. The 30-cm-long capillary had a total internal volume of 15 µL, and the volume of the stationary phase was 59 nL. The schematic of the in-tube SPME setup for the capillary LC system is also shown in Figure 1. Automated In-Tube SPME Technique. The Omegawax 250 GC capillary was previously demonstrated as optimal for the extraction of six carbamates from aqueous solutions.9 In the first step of the in-tube SPME technique, extraction is performed by a dynamic absorption/desorption process where sample is aspirated from and dispensed back to a sample, through the Omegawax extraction capillary. This is performed until an equilibrium level of extraction has been achieved between analyte concentrations in the sample and the extraction phase inside the capillary. The whole extraction process was automatically controlled by the autosampler software. In the case of the HP 1100 system, it was the injector program of the HP ChemStation software that controlled this step. In the second step, analytes are desorbed from the extraction capillary and transferred to the separation column by mobile-phase flow. During the extraction process, the six-port injection valve is in the bypass (load) position, and so the extraction capillary is isolated from the mobile-phase flow and is under normal pressure. Initially, 30 µL of sample is drawn into the capillary. Next, 25 µL is ejected back into the sample. This is followed by 10 more draw/ eject cycles of 25 µL each, before 30 µL of sample is finally ejected back to the sample. In the second step (desorption), the six-port valve is switched to the main pass (inject) position, and the mobile phase passes through the capillary at high pressure, from the Accurate microflow processor, at 4 µL/min. Analytes desorb from the extraction capillary into the mobile phase and are transferred to the µ-Guard column and a capillary column for separation. (9) Gou, Y.; Tragas, C.; Lord, H.; Pawliszyn, J. J. Microcolumn Sep. 2000, 12 (3), 125.
Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
2775
Figure 1. Schematic of in-tube SPME/capillary LC as implemented on the Hewlett-Packard 1100 LC system
RESULTS AND DISCUSSION On-Column Focusing. The in-tube SPME technique is a dynamic absorption/desorption process which is achieved by moving a sample in to and out of a capillary. The mobile phase that follows the sample inside the capillary can desorb the analytes from the coating during each eject step and concentrate them in the capillary. To apply the on-column focusing technique, the analytes must be dissolved in a more weakly eluting mobile phase. To optimize focusing, the system was conditioned with various weakly eluting mobile phases for 30 min, after which time extraction and injection were performed, the mobile-phase composition was ramped to final conditions (acetonitrile (ACN)/H2O 50/50) over 1 min, and the resulting chromatograms were compared for peak shape and resolution. Figure 2 illustrates the chromatograms of the analytes when dissolved in six different initial mobile-phase compositions. Serious peak distortion and band broadening occur when the sample is dissolved in the eluting mobile phase, ACN/water (50/50, v/v). The chromatographic efficiency increases gradually as the dissolving mobile phase gets weaker. An excellent peak shape was achieved, for both late- and early-eluting peaks, when ACN/water (10/90, v/v) and ACN/water (0/100, v/v) were used. Therefore, ACN/water (10/90, v/v) was selected as the optimal extraction condition for our further studies, as it is not desirable to use only water as the mobile phase for a C18 column. These experiments proved that on-column focusing results in efficient chromatographic performance when the in-tube SPME is combined with the capillary LC system. Investigation of Separation Conditions. For samples dissolved in a weak eluting solvent, gradient elution allows the injection of a large sample volume without any adverse effect on the band broadening. Under these conditions, the sample undergoes on-column focusing at the column inlet and the injection of 2776
Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
Figure 2. Effect of on-column focusing for a large-volume injection of ∼30 µL on a 300-µm-i.d. column for the chromatographic separation of six carbamates. Initial mobile-phase composition is specified for each chromatogram. Gradient elution to a final composition of ACN/H2O 50/50 was performed in each case.
relatively large sample volumes becomes possible. The 10% (% ACN) initial mobile-phase strength was restricted by the extraction
conditions in the on-column focusing studies. The 50% (% ACN) final strength was initially set since it was the optimal isocratic condition used for the in-tube SPME/LC methods described previously.9,10 The value of the retention factor k for each band in isocratic elution is important for controling the HPLC separation. The average value of k in gradient elution, defined as k*, determines the sample resolution and bandwidth. Values of k* can be estimated using eq 2,11 where tG is gradient time (min), F
k* ) 87tGF/Vm(∆% B)S
Table 1. Estimated Injection Volume (µL) with Different Lengths of Omegawax 250, i.d. 0.25 mm, Film 0.25 µm length (cm)
void vol (µL)
coating vol (nL)
draw/eject vol (µL)
est injection vol (µL)
15 30 60
7.38 14.8 29.5
30 59 118
18 25 40
23-38 30-45 45-60
(2)
is the flow rate (mL/min), Vm is the column dead volume (mL), ∆% B is the difference between the initial and the final % B values, and S is a property of the analyte of interest. For samples with molecular weights of 100-500 (the compounds under study here belong to this category), S ≈ 4, and eq 2 can be approximated as
k* ) 20tGF/Vm(∆% B)
(3)
where k* is proportional to the gradient time tG (min). The longer the gradient time, the larger the k* value. Because of the similarity between the isocratic and the gradient elutions, larger values of k* should lead to the same effects as larger values of k. As k* increases, (1) resolution Rs increases, (2) bands become broader with a corresponding reduction in peak height, and (3) run times become longer. Based on our experimental conditions, the column dead time was found to be ∼5 min, the microflow rate was 4 µL/ min, and the initial and final B values were 10 and 50%, respectively. The gradient time was calculated as less than 2 min in order to have a good compromise in resolution Rs, peak height, and run time (0.5 < k* < 20). Experimentally it was observed that a 1-min gradient time produced the best chromatographic efficiency (data not shown). The gradient range refers to the difference between the initial and final % B. The larger the range, the shorter the run time. Gradient separations of the six carbamates under different final ACN levels, from 50% to 70% were assessed. As the gradient steepness increased from 40 to 60%/min,, k* decreased from 10 (50% ACN) to 6.7 (70% ACN), and the retention time for the lasteluting band changes from 64 to 50 min. The 60% ACN produced a shorter run time and high chromatographic efficiency without sacrificing the resolution between the critical bands. With 70% ACN, even higher peak height (high sensitivity) was achieved within a shorter time. However, the resolution Rs for the last two eluting bands is ∼1.5. For trace analysis, it is desirable that Rs be higher than this. Therefore, a final 60% ACN was chosen as the optimal condition. Effect of Methanol. In the conventional in-tube SPME/HPLC method, it was found that preloading of the extraction capillary with methanol can double the extraction efficiency of carbaryl from an aqueous sample. A volume of 4 µL of methanol was selected to preload into the capillary before extraction based on experimental optimization.9 It was expected that methanol would have the same effect on the extraction of analytes in the capillary LC system. (10) Gou, Y.; Eisert, R.; Pawliszyn, J. Automated in-tube SPME /HPLC for carbamate pesticides analysis, J. Chromatogr., A 2000, 873, 137. (11) Snyder, L. R.; Kirkland, J. J.; Glajch, J. L. Practical HPLC Method Development; John Wiley & Sons: New York, 1997; Chapter 8.
Figure 3. Chromatograms of six carbamates with extraction using different lengths of capillary: (a) 30, (b) 15, and (c) 60 cm.
Experimentally it was determined that when 3-10 µL of methanol was preloaded, the first two eluting peaks showed progressively more distortion (tailing). The rest of the peaks remained intact compared to the run without preloaded methanol. The effect on the early-eluting peaks was possibly due to an enhanced “partial elution” at the top of the column before the mobile-phase elution process began. Therefore, it was not beneficial to preload methanol for the capillary LC studies. Capillary Length. Different lengths of capillary thus have different void volumes and volumes of coating (stationary phase). The total draw/eject sample volume is determined by the total volume of the needle assembly and the extraction capillary. As described previously, the sample injection volume is determined by the volume of mobile phase required to completeely desorb the analytes.9 Three different lengths of capillary, (15, 30, and 60 cm) were investigated; the corresponding injection volumes (µL) are listed in Table 1. Figure 3 shows the chromatograms of the six carbamates with the different lengths of capillary. It should be noted that the focusing time for the 15- and 30-cm capillaries was the same but was increased for the 60-cm capillary. There was no band broadening for any of the bands when 60 cm of capillary was used. This indicates that injection of a larger amount of sample (∼60 µL) is possible using the on-column focusing technique. However, the retention time for the firsteluting peak was ∼64 min, which was 26 min longer than when a 30-cm capillary was used. The delay in the elution when the injection volumes were increased is a common phenomenon with both the conventional LC and the micro-LC when the on-column focusing technique is applied.2 As expected, the extraction sensitivity was increased with increasing length of capillary. When a 15-cm capillary was used, extraction sensitivity was lower than when a 30-cm capillary was used. However, the retention time of the first peak was similar to that found with a 30-cm capillary. Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
2777
Table 2. Precision and Limit of Detection (LOD) of the In-Tube SPME/Capillary LC for Six Carbamates LOD (µg/L) precision (% RSD) no.
compound
RT (min)
area counts
1 2 3 4 5 6
carbaryl propham methiocarb promecarb chlorpropham barban
0.5 0.7 0.7 0.8 0.8 0.8
1.7 1.5 1.7 3.5 2.1 4.6
in-tube SPME/ in-tube capillary SPME/ sensitivity LC LC gain 0.02 0.26 0.16 0.19 0.21 0.18
0.53 12 7.0 15 13 10
24 45 42 81 62 56
Therefore, the 30-cm capillary was selected as the optimal capillary length for our further applications. Precision. The reproducibility of the retention time under microgradient elution conditions was a concern during the method development. The reproducibility was determined with six replicate samples (500 ng/mL). Table 2 shows the relative standard deviations (RSD) for both retention time and peak area counts for the six carbamates. For all the carbamates studied, the RSD of retention time was between 0.5 and 0.8% with the 4 µL/min microgradient conditions. The RSD of peak area counts was between 1.5 and 4.6%. Thus, the precision of the in-tube SPME/ capillary LC method can be considered excellent. A more complete discussion of the accuracy and validation of the general method of in-tube SPME for carbamate analysis has been published previously for standard LC. Sensitivity. Column dispersion is directly related to column dimensions. A reduced dispersion results in a gain in the final mass sensitivity. This is the main advantage of micro-LC. The maximum theoretical gain in sensitivity by changing to a smaller column diameter can be calculated by the downscale factor f, which is proportional to the ratio of the square of the two column diameters, according to eq 4,11 where dconv and dmicro are the
Figure 4. Chromatograms of six carbamates extracted and analyzed by (a) in-tube SPME/LC and (b) in-tube SPME/capillary LC. In-tube SPME conditions: extraction capillary, 30-cm Omegawax 250; sample, six carbamates of concentrations of 500 µg/L each spiked into Nanopure water; extraction conditions, draw/eject sample volume 25 µL, draw/eject steps 10, draw/eject speed 250/300 µL/min. Capillary LC: column, 15 × 300 µm, 3-µm C18; gradient, ACN/water from 10/90 to 60/40 in 1 min, then held; 4 µL/min. Conventional LC: column, 10 × 8 mm, 4-µm C18; isocratic ACN/water 50/50, 1 mL/ min.
(4)
Figure 5. Chromatograms of the six carbamates in surface water. Peak identifications: (1) carbaryl; (2) propham; (3) methiocarb; (4) promecarb; (5) chlorpropham; (6) barban. Inset is an enlarged chromatogram of the impurities present in the surface water sample. In-tube SPME conditions: 30-cm Omegawax 250 mixture of six carbamates with concentrations of 500 µg/L each, draw/eject sample volume 25 µL, draw/eject steps 10. Capillary LC: column, 15 × 300 µm, 3-µm C18; gradient, ACN/water from 10/90 to 60/40 in 1 min, then hold, 4 µL/min.
diameters of the conventional and microcolumns, respectively. The conventional column used for the in-tube SPME/LC method was a 8 × 100 mm RCM cartridge column. The diameter of the capillary column used for this study was 0.3 mm. Thus, the downscale factor f was just over 700. However, this value is only valid for direct comparison when the column length and detector are the same. Therefore, we cannot expect to see exactly the sensitivity gain in our method. It is useful to compare the in-tube SPME/capillary LC method with the in-tube SPME conventional LC system by comparing the peak area counts for the six carbamates studied under the two methods. For this comparison, a 30-cm Omegawax 250 capillary was used as the extracting capillary and the sample concentration was 500 µg/L for each compound. The chromatograms are shown in Figure 4. The peak areas increased 76-fold for carbaryl, 173fold for propham, 121-fold for methiocarb, 166-fold for promecarb, 122-fold for chlorpropham, and 134-fold for barban. The limits of detection with the in-tube SPME/capillary LC method were determined (S/N ) 3) using 1 µg/L sample
concentrations with two replicates. LODs for the six carbamates studied using in-tube SPME/capillary LC are also shown in Table 2. For most carbamates, the LOD was between 0.16 and 0.26 µg/ L, and for carbaryl in particular, the LOD was ∼0.02 µg/L (20 ppt). As compared to the LODs found using the conventional intube SPME/LC method described previously,9 a significant reduction in LOD was found. Compared with the in-tube SPME/ LC method, the LODs for carbaryl, propham, methiocarb, promecarb, chlorpropham, and barban were lowered by factors of 24, 45, 42, 81, 62, and 56, respectively, using the in-tube SPME/ capillary LC method. Analysis of Carbamates in Surface Water Samples. The in-tube SPME/capillary LC method was applied for analysis of carbamates in surface water samples. The chromatograms are shown in Figure 5. Some unknown compounds were present in the surface water sample as evidenced by the unknown peaks. In general, extraction efficiency ranged from 15 to 34% higher in surface water samples relative to Nanopure water samples, likely due to matrix effects. This highlights the need to ensure
f ) (dconv/dmicro)2
2778 Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
unknowns are contained in a matrix the same as that used for calibrations or that internal standards are used. CONCLUSIONS In-tube solid-phase microextraction was successfully coupled to a converted capillary LC system. On-column focusing was used to concentrate the analytes from the large sample volume prior to separation and analyses. The analytes were desorbed into an initial noneluting mobile phase before a gradient elution was performed. The gradient time and the final composition of the mobile phase for the gradient elution were optimized on the basis of chromatographic principles. With the in-tube SPME/capillary LC method, the limits of detection achieved for all the carbamates studied were less than 0.3 µg/L and for carbaryl it was 0.02 µg/L (20 ppt). Compared to the conventional in-tube SPME/LC method, the sensitivities for carbaryl, propham, methiocarb, promecarb, chlorpropham, and barban were enhanced by factors of 24, 45, 42, 81, 62, and 56, respectively. Longer analysis times were
necessary for this in-tube SPME/capillary LC method due to the requirements of on-column focusing and gradient elution. The running time for each sample was ∼60 min, which was longer than the conventional in-tube SPME/LC method which required less than 20 min. In addition, a reconditioning of the system between runs with the initial mobile phase was also required to achieve reproducible results. This paper reports on a first approach toward coupling the in-tube SPME sample preparation method to a capillary LC system. This preliminary study has shown that the analysis sensitivity can be greatly enhanced, which would present a very promising method for the determination of analytes in trace or microtrace levels.
Received for review July 2, 1999. Accepted March 27, 2000. AC990726H
Analytical Chemistry, Vol. 72, No. 13, July 1, 2000
2779
