Time-of-Flight MS of Lemon and Lime Oil

Aug 10, 2001 - The high-speed GC separation and MS characterization of lime oil and lemon oil samples using programmable column selectivity and time-o...
39 downloads 6 Views 100KB Size
Download PDF
Anal. Chem. 2001, 73, 4395-4402

Articles

High-Speed GC and GC/Time-of-Flight MS of Lemon and Lime Oil Samples Tincuta Veriotti and Richard Sacks*

Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109

The high-speed GC separation and MS characterization of lime oil and lemon oil samples using programmable column selectivity and time-of-flight mass spectrometry is described. The volatile essential oils are separated on a series-coupled (tandem) column ensemble consisting of a polar trifluoropropylmethyl polysiloxane column and a nonpolar 5% phenyl dimethyl polysiloxane column. Both columns are 7 m long. A 50 °C/min linear temperature ramp from 50 to 200 °C is used, giving an analysis time of ∼2.5 min. A time-of-flight MS with time array detection and automated peak finding and characterization software was used to identify 50 components in lime oil samples and 25 components in lemon oil samples. Despite numerous cases of extensive peak overlap, spectral deconvolution software was very successful in the characterization of most overlapping peaks. For cases where a more complete chromatographic separation is desirable, the tandem column ensemble is operated in the first-column stop-flow mode to enhance the separation of selected overlapping clusters of peaks. A valve between the junction point of the tandem column ensemble and a source of carrier gas at the GC inlet pressure is opened for 2-5-s intervals to stop the flow of carrier gas in the first column. This is used to increase the separation of target component groups that overlap in the ensemble chromatogram without first-column stop-flow operation. This procedure is used to isolate the peak for limonene, the largest peak in the analytical-ion chromatogram of both the lime and lemon oil samples. Essential oils are volatile natural products of botanical origin. The term “essential oil” derives from ancient beliefs that these oily substances contained the essence of the plant. Essential oils are valued for their flavor and fragrance qualities and, in some cases, medicinal properties. Oils from hundreds of plant species are available commercially.1 10.1021/ac010239d CCC: $20.00 Published on Web 08/10/2001

© 2001 American Chemical Society

Essential oils are prepared from plant materials by a number of procedures including steam distillation, vacuum distillation, solvent extraction, cold pressing, and hot pressing. The composition of these oils can vary significantly with place of origin, harvest season, and weather. Oils may be blended, cut with other materials, or otherwise adulterated. Gas chromatography (GC) and GC with mass spectrometry detection (GC/MS) often are used to ascertain quality and purity.2,3 Numerous procedures have been described. Usually sample preparation is minimal, but separation and characterization are relatively slow with analysis times typically from ∼30 min to well over 1 h. In addition, separation conditions may vary greatly for different methods. Lime oil and lemon oil have been extensively studied due to their commercial value. The quality of lemon oil often is ascertained by the total carbonyl content with the cis and trans isomers of citral (geranial and neral, respectively) being the most important flavor-active constituents. Numerous procedures have been described for their determination.4 While most capillary GC and GC/ MS studies of lime and lemon oils involved analysis times of 2030 min, MacLeod used a combination of flow and temperature programming to identify 26 components in lime oil in less than 12 min5 and 20 components in lemon oil is less than 10 min.6 Cartoni et al. used a series-coupled combination of a polar and nonpolar column with a window-diagram optimization procedure to obtain enhanced resolution of several component pairs in a lemon oil sample.7 (1) Formaceck, V.; Kubeczka, K. H. Essential Oils by Capillary Gas Chromatography and Carbon-13 NMR Spectroscopy; John Wiley & Sons: Chichester, U.K., 1982. (2) Masada, Y. Analysis of Essential Oils by Gas Chromatography and Mass Spectrometry; John Wiley & Sons: Chichester, 1976. (3) Lodge, N.; Paterson, V. J.; Young, H. J. Sci. Food Agric. 1984, 35, 447. (4) Gramshaw, J. W.; Sharpe, K. J. Sci. Food Agric. 1980, 31, 93. (5) MacLeod, W. D. J. Agr. Food Chem. 1968, 16, 436. (6) MacLeod, W. D.; McFadden, W. H.; Bugues, N. M. J. Food Sci. 1966, 31, 591. (7) Cartoni, P. G.; Goretti, G.; Monticelli, B.; Russo, M. V. J. Chromatogr. 1986, 370, 93.

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001 4395

Figure 1. Analytical-ion chromatogram for the high-speed GC/TOFMS of lime oil. Peak numbers correspond to component numbers in Table 1. The insets show expanded analytical-ion chromatogram (A) and extracted-ion chromatograms (B) of some congested regions.

Long analysis time is often the result of the need to achieve a complete separation with relatively broad peaks for characterization with quadrupole MS. Time-of-flight (TOF) MS can be used to obtain reliable extracted-ion chromatograms with high spectral acquisition rates.8 This allows for the characterization of the much narrower GC peaks obtained with high-speed GC methods. Further, the lack of spectral skewing with TOFMS allows for the spectral deconvoution and characterization of overlapping clusters of peaks. This reduces the chromatographic resolution requirements and further decreases analysis time.9 For cases where more complete chromatographic separation is desirable, programmable column selectivity can be used to obtain enhanced separation of some target component pairs with only small increases in analysis time.10-12 Programmable column selectivity is obtained by programming the carrier gas pressure at the junction point of a series-coupled (tandem) ensemble of a polar and a nonpolar column. A useful implementation of programmable column selectivity is obtained with a low-dead-volume valve connecting the column junction point to a source of carrier gas at the GC inlet pressure.13 When the valve is open, the carrier gas flow stops in the first column and speeds up in the second column. A pair of mixture components that are separated by the first column in the ensemble but coelute from the second column can be separated by the ensemble if the valve is opened briefly when the band for one of the components has crossed the junction but the band for the other component is still in the first column. Typically, the valve is opened for a few seconds for each targeted component pair, and this adds a comparable time increment to the analysis time. (8) Leonard, C.; Sacks, R. Anal. Chem. 1999, 71, 5177. (9) Veriotti, T.; Sacks, R. Anal. Chem. 2000, 72, 3063. (10) Smith, H.; Sacks, R. Anal. Chem. 1998, 70, 4960. (11) Grall, A.; Zellers, E. T.; Sacks, R. Environ. Sci. Technol. 2001, 65, 163. (12) Leonard, C.; Sacks, R. Anal. Chem. 1999, 71, 5501. (13) Veriotti, T.; Sacks, R. Anal. Chem. 2001, 73, 3045.

4396

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

For this study, procedures were developed for the high-speed separation and characterization of 50 components in two different lime oil samples and 25 components in three different lemon oil samples. Using series-coupled capillary columns with pressure switching techniques and fast oven temperature programming, analysis times of less than 2.5 min are obtained. EXPERIMENTAL SECTION An HP6890 GC (Hewlett-Packard, Atlanta, GA) equipped with an HP7683 autoinjector and a split/splitless inlet with electronic inlet pressure control was used as an experimental platform. The pressure-programmable column ensemble consists of two 7.0-mlong by 0.18-mm-i.d. columns. The first (polar) column uses a 0.18µm-thick bonded trifluoroprolylmethyl polysiloxane stationary phase (DB-200, J&W Scientific, Inc., Folsom, CA). The second (nonpolar) column uses a 0.18-µm-thick bonded 5% phenyl dimethyl polysiloxane stationary phase (DB-5, J&W Scientific). Pressure programming is provided by a pneumatically operated, low-dead-volume valve (model MOPV-1/50, SGE, Austin, TX) connected between the junction point of the column ensemble and a ballast chamber containing carrier gas at the GC inlet pressure. When the valve is open, the carrier gas pressure at the column junction point is the same as the GC inlet pressure, and the carrier gas flow stops in the first column (stop-flow operation). The pneumatic valve is operated by a 50-55 psig compressed air source connected through an electronically actuated solenoid valve (model GH3412, Precision Dynamics, Phoenix, AZ). The plumbing and the ballast chamber are described in detail in ref 13. The ballast chamber pressure is controlled at the GC inlet pressure by an electronic pressure controller (MKS model 640A, MKS Instruments, Andover, MA). An HP flame ionization detector FID is also connected to the column junction point with an all stainless steel splitter (MT1C56, Valco Instruments, Houston, TX) and a 0.5-m-long, 0.05-mm-i.d. segment of deactivated fused-silica column. The FID is used to

Table 1. Compounds and Their Retention Time for Lime Oil peak no.

compound

tR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

pentane 1,4-cyclohexadiene, 1-methylnonane ocimene (1R)-R-pinene R-pinene 2,3-dimethyl-cyclohexa-1,3-diene 4(10)-thujene camphene 1,4-cyclohexadiene, 1-methyl-4-(1-methylethyl)β-pinene R-myrcene c-terpinene terpinolene isomer limonene terpinene 4-acetate cymene citronellal γ-terpinene limonene oxide, cis 2-methyl-1-nonene-3-yne terpinolene isomer n-nonanal cinnamaldehyde benzene, 1-methyl-2-(1-methylethyl)linalool myrtenol 4-carene, (1S,3R,6R)-(-)limonene oxide, trans 4-terpineol terpinyl acetate nerol R-terpineol 2,3-epoxycarane, (E)1,3-cyclohexadiene, 1-methyl-4-(1-methylethyl)citronellyl acetate trans-farnesene R-bergamotene β-caryophyllene cis-farnesene -selinene β-elemene santalene isoledene valencene isoeugenol germacrene D cadinene 10,12-octadecadiynoic acid spathulenol

13.73 22.05 31.47 33.29 35.70 36.69 37.96 38.89 39.25 40.72 42.97 44.25 46.29 48.57 49.73 50.34 51.22 52.81 53.68 55.73 56.47 57.90 58.53 63.01 64.19 64.95 68.81 72.69 74.73 76.29 78.73 79.61 82.41 89.77 99.13 103.09 110.61 114.21 115.05 120.69 122.41 123.05 124.65 124.65 124.65 125.53 126.73 129.09 135.64 144.81

monitor the effluent from the first column, and ∼8% of the effluent from the first column is split to this detector. A TOFMS (Pegasus II, Leco Corp., St Joseph, MI) was used as the primary detector for the column ensemble. The TOFMS was operated with a spectral acquisition rate of 25 spectra/s with mass range from 35 to 350 m/e. Materials and Procedures. Hydrogen carrier gas is purified by filters for water vapor, oxygen, and hydrocarbons. The GC inlet pressure was 35.0 psig. The inlet temperature was 290 °C. Five different citrus oil samples were investigated. Lemon and lime oils of unknown origin were supplied by Leco Corp. Mexican lime oil, yellow lemon oil, and green lemon oil were supplied by Gritman Essential Oil (Friendswood TX). Temperature programming from 50 to 200 °C at 50 °C/min beginning at the time of injection was used for all samples. Injection size was 0.1 µL with

a split ratio of 150:1. The samples were injected neat for most experiments. For some experiments, the oils were diluted 1:2 with n-pentane. The identity of critical components including neral, geranial, linalool, β-pinene, γ-terpinene, neryl acetate, R-terpineol, limonene, and R-pinene was confirmed by spiking with the pure components. These materials were obtained from Leco Corp. and Fisher Scientific (Pittsburgh, PA). Data processing for the TOFMS was accomplished with software provided by the manufacturer. Processing included automated peak finding, spectral deconvolution of overlapping chromatographic peaks, and component characterization by comparison with the NIST mass spectral database and reverse searches using component lists provided by Leco Corp. A spectral acquisition rate of 25 spectra/s was used for all studies. Signalto-noise threshold for automated peak finding was set at 100:1. Two-point smoothing was used for data processing. Data acquisition for the FID, as well as control of the ballast chamber pressure and operation of the pneumatic valve, was obtained with a 330MHz PC (Dell, OptiPlex GX1) and a 12-bit A/D board (DT-2801, Data Translation, Inc., Marlboro, MA). The interface board was controlled with Labtech Notebook software (Laboratory Technologies, Inc., Wilmington, VA). Grams/32 software (Galactic Industries, Salem, NH) was used for processing FID chromatograms. RESULTS AND DISCUSSION Lime and lemon oils are used for their flavor and fragrance qualities. They are extensively used in soft drinks and confections. Lemon oil is used in a number of household soaps and detergents. Lime oil is obtained by distillation of the juice or whole crushed fruit of Citrus aurantifolia. Lemon oil is obtained by distillation or cold pressing the fresh peel of Citrus limonum. The flavor industry usually uses concentrated, terpeneless oils. The pure essential oils contain ∼90% limonene. Percent levels of citral, R-terpineol, citronellal, linalyl acetate, and linalool as well as other coumarins and terpene hydrocarbons often are present. The pure oils are often cut with lower-cost materials including synthetic limonene or citral. High-Speed Characterization With TOFMS. Figure 1 shows the analytical-ion chromatogram (AIC) for lime oil. The ticalionAIC is found as the sum of the extracted-ion chromatograms for the ions in the found peaks. The AIC provides information similar to the total ion chromatogram but with significantly higher signal-to-noise ratio. The chromatogram is very complex and congested. The analysis is complete in ∼145 s, and the software for automated peak finding found 48 peaks with signal-to-noise ratios greater than 100. The results are shown in Table 1. Component numbers in the table correspond to peak numbers in the AIC. The portion of the AIC in the broken-line box extending from 46 to 60 s is shown on an expanded time scale in inset A. Limonene (peak 15) is the largest peak in the chromatogram. The limonene peak has a significant interference from terpinene 4-acetate (peak 16). Despite extensive overlap, spectral deconvolution by the software was successful. Numerous other peak overlaps are successfully deconvoluted, with the exceptions of component pairs 43/44 and 46/47. Inset B shows extracted-ion chromatograms for these components on an expanded time scale. Chromatograms are shown for m/e of 69, 121, 123, 137, and 161. For successful automated peak finding for Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

4397

Figure 2. Analytical-ion chromatogram for the high-speed GC/TOFMS of lemon oil. Peak numbers correspond to component numbers in Table 2. The insets show expanded analytical-ion chromatogram (A) and extracted-ion chromatograms (B) of some congested regions.

Table 2. Compounds and their Retention Time for Lemon Oil peak no.

compound

tR

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

pentane 2,6-dimethyl-1,3,5,7-octatetraene, E R-pinene camphene β-pinene citronellal c¸ -terpinen limonene m-cymene γ-terpinene terpinolene perilla alcohol linalyl acetate santolina triene R-terpineol 2,3-epoxycarane, (E)neral geranial neryl acetate R-bergamotene geranyl acetate β-caryophyllene santalene limonene dioxide bicyclo[3.1.1]heptane,6,6-dimethyl-2-methylene

13.62 35.66 36.78 39.30 43.26 44.38 46.54 49.74 51.54 54.06 57.98 77.30 78.06 78.98 82.10 83.34 104.04 108.98 111.82 114.34 114.90 115.18 124.82 65.92 95.20

unknown components using a spectral acquisition rate of 25 spectra/s, a minimum peak apex separation of ∼120 ms is required. This is not achieved for these component pairs. Figure 2 shows the AIC for lemon oil. The component list is shown in Table 2. Note that only 25 components were found with a signal-to-noise threshold of 100:1. Component 8 is limonene, which again is the largest peak in the chromatogram. Peak 9 (cymene) overlaps significantly with limonene, but spectral deconvolution was successful, and both peaks were correctly identified. Note that the two citral isomers (peaks 17 and 18) are well separated by the column ensemble in less than 110 s. Inset 4398

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

A shows the AIC for the region extending from 75 to 85 s on an expanded scale. Inset B shows extracted-ion chromatograms for the region containing peaks 18-22. Chromatograms are shown for m/e of 69, 136, 137, and 161. The vertical lines correspond to the peak locations determined by the automated peak finding software. The peaks for all 25 components in Table 2 were found and identified. Figure 3 shows selected portions of the analytical-ion chromatograms for lime oil (Leco) (a), Mexican lime oil (Gritman) (b), lemon oil (Leco) (c), yellow lemon oil (Gritman) (d), and green lemon oil (Gritman) (e). For the lime oil samples, the AIC for the two samples are very similar. The peaks in the region between 100 and 150 s show the greatest distinction for the two samples. Component 50 (spathulenol) is nearly absent in the Mexican lime oil. Several other differences in the ion abundance ratios for these samples also are observed. The AIC for the three lemon oil samples also are very similar except for the low-amplitude peaks in the region from 60 to 100 s. Component 25 is detected only in the yellow lemon oil sample. In addition, the amplitude ratios for peaks 12-16 are very different for the three lemon oils and could be used as a fingerprint for the these oils. These results show that GC/TOFMS with time array detection can rapidly and completely characterize lemon oil samples and differentiate oils from different sources. Enhanced Chromatographic Separation. Two peak pairs in the lime oil sample are not adequately resolved by the column ensemble for successful automated peak finding. More complete chromatographic separation is required. Pressure programming the tunable column ensemble is used to enhance chromatographic separation. Pressure pulses applied to the junction point of the column ensemble have been shown to be useful for increasing the separation by the ensemble for targeted groups of components without adversely affecting the separation of other components in a mixture.13,14 If pressure pulses are obtained by connecting

Figure 3. Analytical-ion chromatograms showing details of lime oil and lemon oil samples: (a) lime oil; (b) Mexican lime oil; (c) lemon oil; (d) yellow lemon oil; (e) green lemon oil. Peak numbers correspond to component numbers in Tables 1 and 2.

Figure 4. Column 1 stop flow for enhanced separation of limonene (15) and terpinene 4-acetate (16) in lime oil: (a) FID chromatogram from column 1 with arrow showing time of stop-flow initiation; (b) analytical-ion chromatogram with no stop flow; (c) analytical-ion chromatogram with 5.0-s stop flow.

the column junction point through a low-dead-volume valve to a source of carrier gas at the GC inlet pressure, carrier gas flow in the first column stops when the valve is open. Typically, the valve is opened for an interval from 1 to 5 s for each targeted component pair. The advantages and implementation of first-column stop-flow operation have been described.13 To achieve increased ensemble separation of a targeted component pair, it is necessary that the component bands be (14) Veriotti, T.; Sacks, D. Anal. Chem. 2001, 73, 813.

completely separated by the first column in the ensemble. Figure 4a shows the FID signal for the detector monitoring the effluent from the first column for a lime oil sample diluted 1:2 with n-pentane, and Figure 4b shows the AIC from the column ensemble. This portion of the chromatograms contains the limonene peak (peak 15). Note that peak 15 is well separated from peak 16 (terpinene 4-acetate) on the first column, but the peak pattern changes from the column ensemble with the result that these peaks show considerable overlap in the AIC. In the AIC, Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

4399

Figure 5. Column 1 stop flow for enhanced separation of R-bergamotene (38) and β-caryophyllene (39) and for santalene (43) and isoledene (44) in lime oil: (a) FID chromatogram from column 1 with arrows showing times for stop-flow initiation; (b) analytical-ion chromatogram with no stop-flow; (c) analytical-ion chromatogram with two 5-s stop-flow intervals.

peak 16 appears as a shoulder on the falling edge of peak 15. Dilution of the sample with n-pentane was necessary in order to obtain a complete separation of components 15 and 16 in the first column. The neat injection used for Figure 1 resulted in a very broad limonene peak due to column overload, which was not adequately separated from component 16 in the first column. In Figure 4c, the AIC is shown for the case where the flow in the first column is stopped for 5.0 s beginning 30.5 s after injection. This is indicated by the vertical arrow in the FID chromatogram of Figure 4a. Note that the band for component 15 has completely crossed the column junction and is in the second column when the valve is opened. Without the stop-flow pressure pulse, the band for component 16 reaches the junction ∼32 s after injection, and thus, it is still in the first column when the valve is opened. The result of the pressure pulse is an increase in the ensemble retention time approximately equal to the pulse duration. With the stop-flow pulse the separation is complete, and baseline resolution is obtained for all target components. The ensemble peak separations for components 16-19 and 22 have changed only slightly from the pulse, but the separation of 15 and 16 has increased by an amount nearly equal to the pulse width. Parts a and b of Figure 5 show the FID chromatogram from the first column and the AIC from the column ensemble, respectively, for the lime oil sample in the region containing peaks 38-50. Several overlapping pairs or groups of peaks are observed in the AIC. These include 38/39, 41/42, and 43/44/45/46/47. The peaks for component pairs 43/44 and 46/47 are most significant since their peak apex separations are inadequate for automated peak finding and spectral deconvolution in an unknown mixture using a spectral acquisition rate of 25 spectra/s. Peak 38 (R-bergamotene) and peak 39 (β-caryophyllene) are well separated by the first column, and a 5.0-s-long stop-flow pulse initiated ∼84 s after injection results in complete separation by 4400

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

the column ensemble. This is shown in the AIC of Figure 5c. The pulse initiation time is indicated by the vertical arrow at 85 s in the FID chromatogram. Note that the 1-s difference between the pulse initiation time and the location of the arrow in Figure 5a is the result of the 1-s transport time in the capillary tube connecting the column junction point to the FID.13 Component pairs 43/44 and 46/47 also are separated by the first column. A stop-flow pulse initiated when components 43 and 47 have crossed the junction but 44, 45, and 46 are still in the first column will enhance the separation of component pairs 43/44 and 46/47. This is shown in Figure 5c. Note that the actual initiation time for the pulse was 97.1 s after injection. This later time is required since components 39-47 all cross the column junction point later than shown in the FID chromatogram because their migration in the first column stopped for 5 s during the first stop-flow pulse used to enhance the separation of component pair 38/39. The separation quality using the two stop-flow pulses is substantially greater than for the case without the pulses. With the two pulses, all target components with signal-to-noise ratios greater than 100:1 are sufficiently separated that the automated peak finding and spectral deconvolution found and correctly identified all target compounds. The three largest peaks from components 38, 39, and 43 are reasonably free from interferences. Only component pairs 41/42 and 44/45 have chromatographic resolution less than 1.0. The components for each of these pairs have similar retention on both the polar and the nonpolar columns, and thus, no manipulation of the ensemble junction point pressure will be successful at completely separating these components. If it is necessary to achieve a complete separation of these component pairs, a different column ensemble or an ensemble of greater resolving power is required. Figure 6 shows how a sequential pair of stop-flow pulses can be used to obtain complete separation of R-bergamotene (peak 20), geranyl acetate (21), and β-caryophyllene (22) in lemon oil.

Figure 6. Column 1 stop-flow for enhanced separation of R-bergmotene (20), geranyl acetate (21) and β-caryophyllene (22) in lemon oil: (a) FID chromatogram from column 1 with arrows showing times for stop-flow initiation; (b), extracted-ion chromatograms with no stop flow; (c) extracted-ion chromatograms with 3.0-s stop flow at 84.5 s; (d) extracted-ion chromatograms with 5.0-s stop flow at 84.5 s; (e) extracted-ion chromatograms with 2.0-s stop flow at 87.7 s; (f) extracted-ion chromatograms with 5.0-s stop flow at 84.5- and 2.5-s stop flow at 92.7 s.

Note that R-bergamotene corresponds to peak 38 and β-caryophyllene corresponds to peak 39 in lime oil, and thus, these components can easily be separated with a stop-flow pulse beginning ∼84 s after injection as described by Figure 5 for lime oil. However, in lemon oil, geranyl acetate elutes from the column ensemble between the other two components, and the situation is more complex. Figure 6a shows the FID chromatogram from the first column, and Figure 6b shows extracted-ion chromatograms from the column ensemble without stop-flow operation. Extracted-ion chromatograms are shown for m/e of 69, 136, 137, and 161. Since peaks 20, 21, and 22 are all well separated from each other by the first column, enhanced separation quality can be achieved with firstcolumn stop flow. The extracted-ion chromatograms in Figure 6b show that the peaks from these three components overlap in the ensemble chromatogram, forming a more complex structure. The peak for component 17 is not seen in Figure 6b since it elutes before the time window presented in Figure 6. The peak-free regions in the FID chromatogram between peaks 20 and 22 and between 22 and 17 represent appropriate stop-flow initiation times. Figure 6c shows extracted-ion chromatograms for a single 3.0s-long stop-flow pulse initiated 85 s after injection. This time is indicated by a vertical arrow in the FID chromatogram. Peak 20 is well separated from 21 and 22. However, peak 20 coelutes with peak 19 in the ensemble chromatogram. Note that the separation of peaks 21 and 22 has not changed significantly since both bands were in the same column (first column) when the stop-flow pulse was initiated. An increase in the stop-flow pulse width to 5.0 s, as shown in Figure 6d, further retards the migration of components 19, 21, and 22, with the result that peak 20 elutes between peaks 18 and 19. Figure 6e shows extracted-ion chromatograms for a single 2.0s-long stop-flow pulse initiated 88.5 s after injection, as indicated

by a vertical arrow in the FID chromatogram. Since the bands for components 20 and 22 have migrated across the junction prior to the pulse, their separation has not changed significantly from the no-pulse case. However, the band for component 21 is still in the first column at the time of the pulse, and thus, its elution from the column ensemble is delayed by nearly the pulse width. Figure 6f shows extracted-ion chromatograms for a two-pulse sequence including a 5.0-s-wide pulse initiated 84.5 s after injection and 2.5-s-wide pulse initiated 92.7 s after injection. Note that the second pulse is delayed relative to the 87.7-s initiation time used for Figure 6e because the first stop-flow pulse delays the elution of component 22 from the first column. The use of the two-pulse sequence results in a significantly enhanced separation relative to the no-pulse case, and all target components are separated with a resolution greater than 1.0. Figure 7 compares the stop-flow pulse requirements for the complete separation of components 38 and 39 in lime oil and components 20, 21, and 22 in lemon oil. Parts a and b of Figure 7 show the AIC with no stop-flow pulses for lime oil and lemon oil, respectively. Components 38 and 39 in the lime oil correspond to 20 and 22 in the lemon oil. Parts c and d of Figure 7 show the AIC for lime oil and lemon oil, respectively, for a single 5.0-swide stop-flow pulse initiated 85.0 s after injection. Components 38 and 20 elute at the same time for the lime and lemon oils, respectively; the same is true for components 39 and 22. Figure 7e shows the AIC for the lemon oil sample with the two-pulse sequence described for Figure 6f. In this case, all target compounds in the lemon oil are separated. CONCLUSIONS High-speed separation and characterization of lime and lemon oils in a time frame of ∼2.5 min/sample has been demonstrated. With TOFMS using time array detection, peaks for all 25 Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

4401

Figure 7. Analytical-ion chromatograms showing enhanced separation of R-bergmotene (38) and β-caryophyllene (39) in lime oil and R-bergamotene (20), geranly acetate (21), and β-caryophyllene (22) in lemon oil: (a) lime oil with no stop flow; (b) lemon oil with no stop flow; (c) lime oil with 5.0-s stop flow at 84.5 s; (d) lemon oil with 5.0-s stop flow at 84.5 s; (e) lemon oil with 5.0-s stop flow at 84.5- and 2.5-s stop flow at 92.7 s.

components found in lemon oil with signal-to-noise ratios greater than 100:1 are found and successfully identified using automated peak finding software. Lime oil is considerably more complex than lemon oil, and two peak pairs are inadequately separated by the column ensemble for the application of automated peak finding and deconvolution with a spectral acquisition rate of 25 spectra/ s. With two sequential stop-flow pulses, the ensemble separation of these component pairs was enhanced with the result that all peaks were found and complete characterization was achieved. The separation quality for some overlapping peak pairs or peak clusters can be substantially improved with column-one stop-flow operation. Of particular significance is the complete separation of limonene, the major component in both lime and lemon oils, with a single stop-flow pulse. The general criteria for enhanced ensemble separation of a component pair with stop-flow operation are that the component bands must be completely separated by the first column and the stop-flow pulse must be initiated after one of the components has migrated across the column junction but the other component is still in the first column. For multiple stop-flow pulses, the occurrence of the each pulse requires that subsequent pulses be delayed, since all components

4402

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

in the first column during a pulse are delayed in arriving at the column junction point. For isothermal operation of the first-column stop-flow system, the delay for the second pulse is equal to the duration of the first pulse and so on. This is because band migration rates after completion of a stop-flow pulse are the same as they were just prior to the pulse. For temperature-programmed GC, column heating continues during each pulse with the result that component migration rates in the first column are greater after completion of the pulse than just prior to its initiation. This effect increases with increasing pulse duration. ACKNOWLEDGMENT The authors gratefully acknowledge Leco Corp., St. Joseph, MI, for use of the Pegasus II TOFMS and for the lemon and lime oil samples and Gritman Essential Oil, Friendswood TX, for the Mexican lime oil, yellow lemon oil, and green lemon oil samples.

Received for review February 27, 2001. Accepted June 29, 2001. AC010239D